Changes0100644000113000011300000000017107355467174011047 0ustar kripkripRevision history for Perl extension OpenIsis. 0.01 Wed May 23 08:04:14 2001 - original version; created by h2xs 1.19 Demo.jsp0100644000113000011300000001173307557537772011172 0ustar kripkrip<%@page import="org.openisis.*,java.io.IOException,java.util.Date"%> <%! static final String EMPTY = ""; static Db[] dbs = new Db[2]; static String dberr = EMPTY; static { try { String[] args = new String[] { "-v", "error", "-encoding", "Cp850" }; dbs[0] = Db.open( "/var/db/cds/cds", args ); args = new String[] { "-v", "error", "-format", "aligned" }; dbs[1] = Db.open( "/var/db/unesb/unesb", args ); } catch (IOException e) { e.printStackTrace(); dberr = e.toString(); } }; static String NV ( String x ) { return null==x ? EMPTY : x; } static int ROWS = 5; // number of rows to show at once %> <% String search = NV( request.getParameter( "search" ) ); int off = 0; int rows = ROWS; int dbnum = -1; try { off = Integer.parseInt( request.getParameter( "off" ) ); } catch ( Exception e ) {} try { dbnum = Integer.parseInt( request.getParameter( "db" ) ); if ( dbs.length <= dbnum ) dbnum = -1; } catch ( Exception e ) { } if ( 0 > dbnum ) try { dbnum = Integer.parseInt( (String)session.getAttribute( "db" ) ); } catch ( Exception e ) { dbnum = 0; } else session.setAttribute( "db", ""+dbnum ); Db db = dbs[dbnum]; try { rows = Integer.parseInt( request.getParameter( "rows" ) ); } catch ( Exception e ) {} int[] set = (int[])session.getAttribute( "set" ); %> OpenIsis JSP demo <%= dberr %>
narrow previous search rows at once in
Note: $,"",*,+,^,(F),(G),..,$$,/(tag) are supported. Queries are processed straight left to right.

<% if ( EMPTY != search ) { long then = (new Date()).getTime(); int[] nset = db.search( search, Db.QRY_SIMPLE ); long now = (new Date()).getTime(); if ( null == nset ) { %> sorry, no results for <%=search%> <% session.removeAttribute( "set" ); return; } else { %> got <%=nset.length%> rows for <%=search%> in <%=(now - then)/1000.%> seconds.
<% if ( null == set || null == request.getParameter("narrow") ) set = nset; else { // and sets set = Db.AND( set, nset ); %> <%=set.length%> rows remaining after combine with previous search
<% } session.setAttribute( "set", set ); } } // EMPTY != search if ( null == set ) return; if ( 0 < off ) { int boff = off-rows; if ( boff < 0 ) boff = 0; %> <<rows <%=boff+1%> to <%=boff+rows%> <% } int end = off + rows; if ( end >= set.length ) end = set.length; else { %> rows <%=end+1%> to <%=end+rows%>>> <% } %>

rows <%=off+1%> to <%=end%>

<%-- start output --%> <% if ( 0 == dbnum ) { for ( int i=off; i

MFN : <%= r.getRowId() %>
<%= r.v("CONFERENCE : ",null,12,m,0,0,r13,s) %> <%= r.v("TITLE : ",null,24,m,0,0,r13,s) %> <%= r.v("AUTHOR(S) : ",null,14,m,0,0,"; "+r13+" ",s) %> <%= r.v("EDITION : ",null,25,m,0,0,r13,s) %> <%= r.v("IMPRINT : ",null,26,m,0,0,r13,s) %> <%= r.v("COLLATION : ",null,30,m,0,0,r13,s) %> <%= r.v("SERIES : ",null,44,m,0,0,r13,s) %> <%= r.v("NOTES : ",null,50,m,0,0,r13,s) %> <%= r.v("CORP. BODY : ",null,71,m,0,0,r13,s) %> <%= r.v("ADDED CONF : ",null,72,m,0,0,r13,s) %> <%= r.v("OTHER TITLE: ",null,74,m,0,0,r13,s) %> <%= r.v("OTHER LANG : ",null,76,m|'*',0,0,r13,s) %> <%= r.v("KEYWORDS : ",null,69,m,0,0,r13,s) %> <% } } else { %> <% for ( int i=off; i <% continue; } int id = r.getRowId(); int l = r.getLen(); %> <% for ( int j=0; j <% if ( '^' == f.val.charAt(0) ) { Field[] s = f.split(); if ( null == s ) continue; for ( int k=0; k <% } // for subfields } // if subfields } // for fields } // for rows %>
       
skip bad row <%=set[i]%> as <%=i+1%> of <%=set.length%>
row <%=id%> with <%=l%> fields as <%=i+1%> of <%=set.length%>
<%=id%> <%=f.tag%>   <%=f.html()%>
<%=id%> <%=f.tag%> <%=(char)s[k].tag%> <%=s[k].html()%>
<% } %> Header.txt0100644000113000011300000000163107644371740011501 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ LGPL.txt0100644000113000011300000006347607355467175011075 0ustar kripkrip GNU LESSER GENERAL PUBLIC LICENSE Version 2.1, February 1999 Copyright (C) 1991, 1999 Free Software Foundation, Inc. 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed. [This is the first released version of the Lesser GPL. It also counts as the successor of the GNU Library Public License, version 2, hence the version number 2.1.] Preamble The licenses for most software are designed to take away your freedom to share and change it. By contrast, the GNU General Public Licenses are intended to guarantee your freedom to share and change free software--to make sure the software is free for all its users. This license, the Lesser General Public License, applies to some specially designated software packages--typically libraries--of the Free Software Foundation and other authors who decide to use it. You can use it too, but we suggest you first think carefully about whether this license or the ordinary General Public License is the better strategy to use in any particular case, based on the explanations below. When we speak of free software, we are referring to freedom of use, not price. Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for this service if you wish); that you receive source code or can get it if you want it; that you can change the software and use pieces of it in new free programs; and that you are informed that you can do these things. To protect your rights, we need to make restrictions that forbid distributors to deny you these rights or to ask you to surrender these rights. These restrictions translate to certain responsibilities for you if you distribute copies of the library or if you modify it. For example, if you distribute copies of the library, whether gratis or for a fee, you must give the recipients all the rights that we gave you. You must make sure that they, too, receive or can get the source code. If you link other code with the library, you must provide complete object files to the recipients, so that they can relink them with the library after making changes to the library and recompiling it. And you must show them these terms so they know their rights. We protect your rights with a two-step method: (1) we copyright the library, and (2) we offer you this license, which gives you legal permission to copy, distribute and/or modify the library. To protect each distributor, we want to make it very clear that there is no warranty for the free library. Also, if the library is modified by someone else and passed on, the recipients should know that what they have is not the original version, so that the original author's reputation will not be affected by problems that might be introduced by others. Finally, software patents pose a constant threat to the existence of any free program. We wish to make sure that a company cannot effectively restrict the users of a free program by obtaining a restrictive license from a patent holder. Therefore, we insist that any patent license obtained for a version of the library must be consistent with the full freedom of use specified in this license. Most GNU software, including some libraries, is covered by the ordinary GNU General Public License. This license, the GNU Lesser General Public License, applies to certain designated libraries, and is quite different from the ordinary General Public License. We use this license for certain libraries in order to permit linking those libraries into non-free programs. When a program is linked with a library, whether statically or using a shared library, the combination of the two is legally speaking a combined work, a derivative of the original library. The ordinary General Public License therefore permits such linking only if the entire combination fits its criteria of freedom. The Lesser General Public License permits more lax criteria for linking other code with the library. We call this license the "Lesser" General Public License because it does Less to protect the user's freedom than the ordinary General Public License. It also provides other free software developers Less of an advantage over competing non-free programs. These disadvantages are the reason we use the ordinary General Public License for many libraries. However, the Lesser license provides advantages in certain special circumstances. For example, on rare occasions, there may be a special need to encourage the widest possible use of a certain library, so that it becomes a de-facto standard. To achieve this, non-free programs must be allowed to use the library. A more frequent case is that a free library does the same job as widely used non-free libraries. In this case, there is little to gain by limiting the free library to free software only, so we use the Lesser General Public License. In other cases, permission to use a particular library in non-free programs enables a greater number of people to use a large body of free software. For example, permission to use the GNU C Library in non-free programs enables many more people to use the whole GNU operating system, as well as its variant, the GNU/Linux operating system. Although the Lesser General Public License is Less protective of the users' freedom, it does ensure that the user of a program that is linked with the Library has the freedom and the wherewithal to run that program using a modified version of the Library. The precise terms and conditions for copying, distribution and modification follow. Pay close attention to the difference between a "work based on the library" and a "work that uses the library". The former contains code derived from the library, whereas the latter must be combined with the library in order to run. GNU LESSER GENERAL PUBLIC LICENSE TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION 0. This License Agreement applies to any software library or other program which contains a notice placed by the copyright holder or other authorized party saying it may be distributed under the terms of this Lesser General Public License (also called "this License"). Each licensee is addressed as "you". A "library" means a collection of software functions and/or data prepared so as to be conveniently linked with application programs (which use some of those functions and data) to form executables. The "Library", below, refers to any such software library or work which has been distributed under these terms. A "work based on the Library" means either the Library or any derivative work under copyright law: that is to say, a work containing the Library or a portion of it, either verbatim or with modifications and/or translated straightforwardly into another language. (Hereinafter, translation is included without limitation in the term "modification".) "Source code" for a work means the preferred form of the work for making modifications to it. For a library, complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the library. Activities other than copying, distribution and modification are not covered by this License; they are outside its scope. The act of running a program using the Library is not restricted, and output from such a program is covered only if its contents constitute a work based on the Library (independent of the use of the Library in a tool for writing it). Whether that is true depends on what the Library does and what the program that uses the Library does. 1. You may copy and distribute verbatim copies of the Library's complete source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice and disclaimer of warranty; keep intact all the notices that refer to this License and to the absence of any warranty; and distribute a copy of this License along with the Library. You may charge a fee for the physical act of transferring a copy, and you may at your option offer warranty protection in exchange for a fee. 2. You may modify your copy or copies of the Library or any portion of it, thus forming a work based on the Library, and copy and distribute such modifications or work under the terms of Section 1 above, provided that you also meet all of these conditions: a) The modified work must itself be a software library. b) You must cause the files modified to carry prominent notices stating that you changed the files and the date of any change. c) You must cause the whole of the work to be licensed at no charge to all third parties under the terms of this License. d) If a facility in the modified Library refers to a function or a table of data to be supplied by an application program that uses the facility, other than as an argument passed when the facility is invoked, then you must make a good faith effort to ensure that, in the event an application does not supply such function or table, the facility still operates, and performs whatever part of its purpose remains meaningful. (For example, a function in a library to compute square roots has a purpose that is entirely well-defined independent of the application. Therefore, Subsection 2d requires that any application-supplied function or table used by this function must be optional: if the application does not supply it, the square root function must still compute square roots.) These requirements apply to the modified work as a whole. If identifiable sections of that work are not derived from the Library, and can be reasonably considered independent and separate works in themselves, then this License, and its terms, do not apply to those sections when you distribute them as separate works. But when you distribute the same sections as part of a whole which is a work based on the Library, the distribution of the whole must be on the terms of this License, whose permissions for other licensees extend to the entire whole, and thus to each and every part regardless of who wrote it. Thus, it is not the intent of this section to claim rights or contest your rights to work written entirely by you; rather, the intent is to exercise the right to control the distribution of derivative or collective works based on the Library. In addition, mere aggregation of another work not based on the Library with the Library (or with a work based on the Library) on a volume of a storage or distribution medium does not bring the other work under the scope of this License. 3. You may opt to apply the terms of the ordinary GNU General Public License instead of this License to a given copy of the Library. To do this, you must alter all the notices that refer to this License, so that they refer to the ordinary GNU General Public License, version 2, instead of to this License. (If a newer version than version 2 of the ordinary GNU General Public License has appeared, then you can specify that version instead if you wish.) Do not make any other change in these notices. Once this change is made in a given copy, it is irreversible for that copy, so the ordinary GNU General Public License applies to all subsequent copies and derivative works made from that copy. This option is useful when you wish to copy part of the code of the Library into a program that is not a library. 4. You may copy and distribute the Library (or a portion or derivative of it, under Section 2) in object code or executable form under the terms of Sections 1 and 2 above provided that you accompany it with the complete corresponding machine-readable source code, which must be distributed under the terms of Sections 1 and 2 above on a medium customarily used for software interchange. If distribution of object code is made by offering access to copy from a designated place, then offering equivalent access to copy the source code from the same place satisfies the requirement to distribute the source code, even though third parties are not compelled to copy the source along with the object code. 5. A program that contains no derivative of any portion of the Library, but is designed to work with the Library by being compiled or linked with it, is called a "work that uses the Library". Such a work, in isolation, is not a derivative work of the Library, and therefore falls outside the scope of this License. However, linking a "work that uses the Library" with the Library creates an executable that is a derivative of the Library (because it contains portions of the Library), rather than a "work that uses the library". The executable is therefore covered by this License. Section 6 states terms for distribution of such executables. When a "work that uses the Library" uses material from a header file that is part of the Library, the object code for the work may be a derivative work of the Library even though the source code is not. Whether this is true is especially significant if the work can be linked without the Library, or if the work is itself a library. The threshold for this to be true is not precisely defined by law. If such an object file uses only numerical parameters, data structure layouts and accessors, and small macros and small inline functions (ten lines or less in length), then the use of the object file is unrestricted, regardless of whether it is legally a derivative work. (Executables containing this object code plus portions of the Library will still fall under Section 6.) Otherwise, if the work is a derivative of the Library, you may distribute the object code for the work under the terms of Section 6. Any executables containing that work also fall under Section 6, whether or not they are linked directly with the Library itself. 6. As an exception to the Sections above, you may also combine or link a "work that uses the Library" with the Library to produce a work containing portions of the Library, and distribute that work under terms of your choice, provided that the terms permit modification of the work for the customer's own use and reverse engineering for debugging such modifications. You must give prominent notice with each copy of the work that the Library is used in it and that the Library and its use are covered by this License. You must supply a copy of this License. If the work during execution displays copyright notices, you must include the copyright notice for the Library among them, as well as a reference directing the user to the copy of this License. Also, you must do one of these things: a) Accompany the work with the complete corresponding machine-readable source code for the Library including whatever changes were used in the work (which must be distributed under Sections 1 and 2 above); and, if the work is an executable linked with the Library, with the complete machine-readable "work that uses the Library", as object code and/or source code, so that the user can modify the Library and then relink to produce a modified executable containing the modified Library. (It is understood that the user who changes the contents of definitions files in the Library will not necessarily be able to recompile the application to use the modified definitions.) b) Use a suitable shared library mechanism for linking with the Library. A suitable mechanism is one that (1) uses at run time a copy of the library already present on the user's computer system, rather than copying library functions into the executable, and (2) will operate properly with a modified version of the library, if the user installs one, as long as the modified version is interface-compatible with the version that the work was made with. c) Accompany the work with a written offer, valid for at least three years, to give the same user the materials specified in Subsection 6a, above, for a charge no more than the cost of performing this distribution. d) If distribution of the work is made by offering access to copy from a designated place, offer equivalent access to copy the above specified materials from the same place. e) Verify that the user has already received a copy of these materials or that you have already sent this user a copy. For an executable, the required form of the "work that uses the Library" must include any data and utility programs needed for reproducing the executable from it. However, as a special exception, the materials to be distributed need not include anything that is normally distributed (in either source or binary form) with the major components (compiler, kernel, and so on) of the operating system on which the executable runs, unless that component itself accompanies the executable. It may happen that this requirement contradicts the license restrictions of other proprietary libraries that do not normally accompany the operating system. Such a contradiction means you cannot use both them and the Library together in an executable that you distribute. 7. You may place library facilities that are a work based on the Library side-by-side in a single library together with other library facilities not covered by this License, and distribute such a combined library, provided that the separate distribution of the work based on the Library and of the other library facilities is otherwise permitted, and provided that you do these two things: a) Accompany the combined library with a copy of the same work based on the Library, uncombined with any other library facilities. This must be distributed under the terms of the Sections above. b) Give prominent notice with the combined library of the fact that part of it is a work based on the Library, and explaining where to find the accompanying uncombined form of the same work. 8. You may not copy, modify, sublicense, link with, or distribute the Library except as expressly provided under this License. Any attempt otherwise to copy, modify, sublicense, link with, or distribute the Library is void, and will automatically terminate your rights under this License. However, parties who have received copies, or rights, from you under this License will not have their licenses terminated so long as such parties remain in full compliance. 9. You are not required to accept this License, since you have not signed it. However, nothing else grants you permission to modify or distribute the Library or its derivative works. These actions are prohibited by law if you do not accept this License. Therefore, by modifying or distributing the Library (or any work based on the Library), you indicate your acceptance of this License to do so, and all its terms and conditions for copying, distributing or modifying the Library or works based on it. 10. Each time you redistribute the Library (or any work based on the Library), the recipient automatically receives a license from the original licensor to copy, distribute, link with or modify the Library subject to these terms and conditions. You may not impose any further restrictions on the recipients' exercise of the rights granted herein. You are not responsible for enforcing compliance by third parties with this License. 11. If, as a consequence of a court judgment or allegation of patent infringement or for any other reason (not limited to patent issues), conditions are imposed on you (whether by court order, agreement or otherwise) that contradict the conditions of this License, they do not excuse you from the conditions of this License. If you cannot distribute so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not distribute the Library at all. For example, if a patent license would not permit royalty-free redistribution of the Library by all those who receive copies directly or indirectly through you, then the only way you could satisfy both it and this License would be to refrain entirely from distribution of the Library. If any portion of this section is held invalid or unenforceable under any particular circumstance, the balance of the section is intended to apply, and the section as a whole is intended to apply in other circumstances. It is not the purpose of this section to induce you to infringe any patents or other property right claims or to contest validity of any such claims; this section has the sole purpose of protecting the integrity of the free software distribution system which is implemented by public license practices. Many people have made generous contributions to the wide range of software distributed through that system in reliance on consistent application of that system; it is up to the author/donor to decide if he or she is willing to distribute software through any other system and a licensee cannot impose that choice. This section is intended to make thoroughly clear what is believed to be a consequence of the rest of this License. 12. If the distribution and/or use of the Library is restricted in certain countries either by patents or by copyrighted interfaces, the original copyright holder who places the Library under this License may add an explicit geographical distribution limitation excluding those countries, so that distribution is permitted only in or among countries not thus excluded. In such case, this License incorporates the limitation as if written in the body of this License. 13. The Free Software Foundation may publish revised and/or new versions of the Lesser General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns. Each version is given a distinguishing version number. If the Library specifies a version number of this License which applies to it and "any later version", you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Library does not specify a license version number, you may choose any version ever published by the Free Software Foundation. 14. If you wish to incorporate parts of the Library into other free programs whose distribution conditions are incompatible with these, write to the author to ask for permission. For software which is copyrighted by the Free Software Foundation, write to the Free Software Foundation; we sometimes make exceptions for this. Our decision will be guided by the two goals of preserving the free status of all derivatives of our free software and of promoting the sharing and reuse of software generally. NO WARRANTY 15. BECAUSE THE LIBRARY IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY FOR THE LIBRARY, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE LIBRARY "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE LIBRARY IS WITH YOU. SHOULD THE LIBRARY PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION. 16. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR REDISTRIBUTE THE LIBRARY AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE LIBRARY (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE LIBRARY TO OPERATE WITH ANY OTHER SOFTWARE), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. END OF TERMS AND CONDITIONS How to Apply These Terms to Your New Libraries If you develop a new library, and you want it to be of the greatest possible use to the public, we recommend making it free software that everyone can redistribute and change. You can do so by permitting redistribution under these terms (or, alternatively, under the terms of the ordinary General Public License). To apply these terms, attach the following notices to the library. It is safest to attach them to the start of each source file to most effectively convey the exclusion of warranty; and each file should have at least the "copyright" line and a pointer to where the full notice is found. Copyright (C) This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA Also add information on how to contact you by electronic and paper mail. You should also get your employer (if you work as a programmer) or your school, if any, to sign a "copyright disclaimer" for the library, if necessary. Here is a sample; alter the names: Yoyodyne, Inc., hereby disclaims all copyright interest in the library `Frob' (a library for tweaking knobs) written by James Random Hacker. , 1 April 1990 Ty Coon, President of Vice That's all there is to it! MANIFEST0100644000113000011300000000007507355467174010710 0ustar kripkripChanges Makefile.PL MANIFEST OpenIsis.pm OpenIsis.xs test.pl Makefile0100644000113000011300000003753307665623703011224 0ustar kripkrip# #/* # openisis - an open implementation of the CDS/ISIS database # Version 0.8.x (patchlevel see file Version) # Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # see README for more information #EOH */ # # # $Id: Makefile,v 1.85 2003/05/30 09:43:31 kripke Exp $ # the Makefile # TARGETS=$(BINARY) # some require explicit ST or MT build, other targets work with both # O switchable directory for binaries; one of # sto for single threaded unix (used by Perl, PHP, Tcl) # mto for multi threaded unix (used by Java, Tcl) # win for gates of hell (used by general failure) # # sto or win STO=sto # mto or win MTO=mto # any O=sto MTFLAGS=-D_REENTRANT MTLIBS=-lpthread ifneq ($(WIN32),) STO=win MTO=win O=win # don't support MT on windows -- Java or PHP must connect server MTFLAGS= MTLIBS= EXE=.exe endif ifneq ($(MT),) O=mto endif BINARY=$(O)/openisis$(EXE) JTARGETS=openisis.jar $(MTO)/libopenjsis.so # object files in lib LOBJS=lstr.o ldb.o lrec.o lfmt.o luti.o lio.o lqry.o lbt.o lcs.o lses.o lfdt.o lstb.o ldsp.o lcli.o lenc.o # object files in server lib LSVOBJS=$(MTO)/lsv.o # classes in jar ifeq ($(J2EE),) EECLASSES= else # additional utilities requiring at least basic j2ee libs javax.mail, org.xml # you might have to add something like that to your CLASSPATH # /opt/tomcat/common/lib/mail.jar:/opt/tomcat/common/lib/activation.jar:/opt/tomcat/common/lib/xerces.jar EECLASSES=Mail RecToXml XmlToRec XmlTest endif CLASSES=Field Rec Db NativeDb PureDb Test Charset Fdt \ PlainSerializer Utils $(EECLASSES) XSUB=perl/blib/arch/auto/OpenIsis/OpenIsis.so # standard make defines CC=gcc RANLIB=ranlib # you may want to try jikes JC=javac # figure out where to find JNI ... ifeq ($(JNIINC),) ifeq ($(JAVAHOME),) JAVAHOME=/usr/java/jdk1.3 endif JNIINC=-I$(JAVAHOME)/include -I$(JAVAHOME)/include/linux endif # running openisis OPENISIS=$(BINARY) RUNJAVA=LD_LIBRARY_PATH=$(MTO) java -classpath openisis.jar:$(CLASSPATH) ifneq ($(WIN32),) JTARGETS=openisis.jar openjsis.dll # Note that console output redirection doesn't work well # older wine 20010731 often croaks "FIXME:pthread_rwlock_rdlock" # since 200111*, console output didn't work at all # some versions should support wine -console tty OPENISIS=wine -- win/openisis include make.mingw tclsubdir=win endif # DBG= ifeq ($(NDEBUG),) DBG=-g else DBG=-DNDEBUG endif # ARCH=-march=i586 # compile CPPFLAGS= # note that -O is necessary for -Wuninitialized warnings # -fstack-check ??? ifeq (,$(WIN32)) PIC=-fPIC else PIC= endif CFLAGS=$(DBG) -O2 -Wall -W -Wshadow -pedantic -Wno-long-long $(PIC) # ############################# link ################################## # common linker flags LDFLAGS= # link statically from openisis.a # do not use -static, since we don't want to slurp in libc LDSTATIC=$(LDFLAGS) # create shared object LDSHARED=$(LDFLAGS) -shared -Wl,-init,openIsisInit,-fini,openIsisFini # link using shared object - the default LDDYNAM=$(LDFLAGS) SYSLIBS= ifeq (solaris,$(OSTYPE)) # rt for nanosleep SYSLIBS=-lrt -lsocket -lnsl endif LDLIBS=$(SYSLIBS) # DB to use -- define DB for db/DB/DB, or DBDIR for other path ifeq ($(DB),) DB=cds endif ifeq ($(DBDIR),) DBDIR=db/$(DB)/$(DB) endif # single and multi threaded objects STOBJS=$(LOBJS:l%=sto/l%) # mto or win MTOBJS=$(LOBJS:l%=$(MTO)/l%) # ############################# targets ################################## # # standard default target all: $(TARGETS) .PHONY: all clean run demo lstbtest lrectest perl jar java jdoc tcl tk # do not delete intermediates .PRECIOUS: org/openisis/%.java %.class %.h clean: rm -rf sto mto win *.jar *.tmp *.exp org/openisis/*.class core jdoc Makefile.old testout.txt [ ! -f perl/Makefile ] || make -C perl clean make -C tcl clean sto/libopenisis.a: $(STOBJS) $(AR) $(ARFLAGS) $@ $^ $(RANLIB) $@ $(MTO)/libopenisis.a: $(MTOBJS) $(AR) $(ARFLAGS) $@ $^ $(RANLIB) $@ sto/libopenisis.so: $(STOBJS) $(CC) $(LDSHARED) -o $@ $^ $(MTO)/libopenisis.so: $(MTOBJS) $(CC) $(LDSHARED) $(LDLIBS) $(MTLIBS) -o $@ $^ $(MTO)/liboisrv.a: $(LSVOBJS) $(AR) $(ARFLAGS) $@ $^ $(RANLIB) $@ sto/openisis: sto/openisis.o sto/libopenisis.a $(CC) $(DBG) $(LDSTATIC) $(LDLIBS) -o $@ $^ $(MTO)/openisis$(EXE): $(MTO)/openisis.o $(MTO)/libopenisis.a $(CC) $(DBG) $(LDSTATIC) $(LDLIBS) $(MTLIBS) -o $@ $^ shared: $(O)/openisis.o $(O)/libopenisis.so $(CC) $(DBG) $(LDDYNAM) -o $(O)/openisis $^ win32: @ $(MAKE) --no-print-directory WIN32=sorry # ########################## demos and tests ############################### # $(O)/demo: $(O)/demo.o $(O)/libopenisis.a $(CC) $(DBG) -o $@ $^ $(O)/demo -db $(DBDIR) 42 $(O)/lrectest: $(O)/lrectest.o $(O)/libopenisis.a $(CC) $(DBG) -o $@ $^ $(O)/lrectest $(O)/lstbtest: $(O)/lstbtest.o $(O)/libopenisis.a $(CC) $(DBG) -o $@ $^ $(O)/lstbtest $(STO)/oipl: oipl.c $(CC) $(CPPFLAGS) $(CFLAGS) -rdynamic -ldl -o $@ $^ ois: $(MTO)/ois : $(MTO)/ois: $(MTO)/ois.o $(MTO)/libopenisis.a $(MTO)/liboisrv.a $(CC) $(LDSTATIC) $^ $(LDLIBS) $(MTLIBS) -o $@ run: $(BINARY) $(OPENISIS) -v $${V:-warn} -db $(DBDIR) $(FMT) 2>&1 | less dump: $(BINARY) $(OPENISIS) -v $${V:-warn} -db $(DBDIR) $(FMT) recode: $(BINARY) $(OPENISIS) -v $${V:-warn} -db $(DBDIR) $(FMT) | recode 850/..Latin-1 check: $(BINARY) $(OPENISIS) -v $${V:-warn} -db $(DBDIR) -check search: $(BINARY) $(OPENISIS) -v $${V:-warn} -search 'development$$' -db $(DBDIR) $(FMT) 2>&1 | less searchdump: $(BINARY) $(OPENISIS) -v $${V:-warn} -search '$$' -db $(DBDIR) $(FMT) 2>&1 terms: $(BINARY) $(OPENISIS) -v $${V:-warn} -terms '$$' -db $(DBDIR) $(FMT) 2>&1 | less test: $(BINARY) $(OPENISIS) -scan development -fmt prop -db $(DBDIR) >testout.txt $(OPENISIS) -terms '$$' -fmt prop -db $(DBDIR) >>testout.txt $(OPENISIS) -search 'devel$$' -fmt mfnf -db $(DBDIR) >>testout.txt diff testout.txt testres.txt crash: $(BINARY) time $(OPENISIS) -v $${V:-warn} -crash 'Z$$' -db db/unesb/unesb searchtest: $(BINARY) $(OPENISIS) -v $${V:-warn} -fmt mfnf -search 'a$$' -db $(DBDIR) $(FMT) >testout.txt diff testout.txt searchtest.txt pftest: $(BINARY) $(OPENISIS) -v $${V:-warn} -search 'Dacca' -db $(DBDIR) \ -pft "$${PFT:-`cat test.pft`}" shsearch: shared LD_LIBRARY_PATH=$(O) $(OPENISIS) -v $${V:-debug} -search development -db $(DBDIR) $(FMT) 2>&1 | less # take your time :) time: $(BINARY) free time $(OPENISIS) -perf 100000 -db $(DBDIR) >/dev/null free # time results for the unesb perf test differ dramatically # depending on the working of operating system caching # you may achive about 0.40 sec (12.000 recs per sec) # but immediatly after timing the cds, # the same test may take more than 40 sec utime: $(BINARY) free time $(OPENISIS) -perf 5000 -db db/unesb/unesb >/dev/null free time $(OPENISIS) -fmt mfnf -search 'k$$' -db db/unesb/unesb >/dev/null # ############################# perl ################################## # # perl: $(XSUB) : : ++++++++++++++++++++++++++++++++++++++++++++++++++++++ : +++ as root in the perl subdir, run 'make install' +++ : ++++++++++++++++++++++++++++++++++++++++++++++++++++++ ptest: $(XSUB) @$(MAKE) -C perl test perl/Makefile: perl/Makefile.PL cd perl && perl Makefile.PL # perl's Makefile doesn't know all deps $(XSUB): perl/OpenIsis.pm perl/OpenIsis.xs openisis.h $(STO)/libopenisis.a perl/Makefile @-[ ! -f perl/OpenIsis.c ] || rm perl/OpenIsis.c @$(MAKE) -C perl # standalone tree with OpenIsis module perltree: $(XSUB) @-[ -d perl/tmptree/auto/OpenIsis ] || mkdir -p perl/tmptree/auto/OpenIsis cp -p perl/OpenIsis.pm perl/tmptree cp -p $(XSUB) perl/tmptree/auto/OpenIsis - strip perl/tmptree/auto/OpenIsis/OpenIsis.so pdemo: perltree perl/demo.pl # ############################# tcl ################################## # tcl: $(O)/libopenisis.a @$(MAKE) -C tcl bin/isish tk: $(O)/libopenisis.a @$(MAKE) -C tcl bin/wisish tcltk: tcl tk ttest: tcl @$(MAKE) -C tcl test # ############################# java ################################## # java: $(JTARGETS) : jtest: java $(RUNJAVA) \ org.openisis.Test -v $${V:-warn} -db $(DBDIR) -encoding Cp850 \ -row 42 \ -search development \ -terms '' \ -mpu -row 130 \ -mhl -hta -row 130 \ -m1 -row 130 \ >testout.txt diff testout.txt jtest.txt jcrash: java $(RUNJAVA) \ org.openisis.Test -v $${V:-warn} -dbpath db/unesb -db unesb \ -crash Z # print in cds1.pft format jcds1: java $(RUNJAVA) \ org.openisis.Test -v $${V:-warn} -db $(DBDIR) -encoding Cp850 \ -m1 -dump jtime: java $(RUNJAVA) \ org.openisis.Test -search 'a$$' -v $${V:-warn} \ -db db/unesb/unesb >/dev/null jdump: java $(RUNJAVA) \ org.openisis.Test -v 9 -db $(DBDIR) 2>&1 | less jwrite: java $(RUNJAVA) \ org.openisis.Test -v 9 -db $(DBDIR) -write jxml: java ifeq ($(J2EE),) echo please be shure to have org.xml. stuff in your CLASSPATH $(MAKE) J2EE=1 jxml else $(RUNJAVA) \ -Dorg.xml.sax.driver=org.apache.xerces.parsers.SAXParser \ org.openisis.XmlTest $${testxml:-test.xml} endif # -overview overview.html jdoc: - mkdir jdoc javadoc \ -sourcepath . \ -encoding ISO-8859-1 \ -J-Xmx64m \ -d jdoc \ -use \ -version \ -author \ -windowtitle 'org.openisis source code documentation' \ -doctitle 'org.openisis source code documentation' \ -header 'org.openisis source code documentation' \ -footer 'this is public material' \ -bottom 'for your eyes only' \ -stylesheetfile jdocstyle.css \ -charset "iso-8859-1" \ org.openisis charset: java mkdir -p db/charset $(RUNJAVA) \ org.openisis.Charset -utf8seq >db/charset/charset.seq mx seq=db/charset/charset.seq create=db/charset/charset -all now tell=1 mx db/charset/charset fst='1 1 MHL,V1' \ ln1=db/charset/charset.lk1 ln2=db/charset/charset.lk2 +fix -all now tell=1 ifload db/charset/charset db/charset/charset.lk1 db/charset/charset.lk2 \ +fix tell=1 -reset -balan $(RUNJAVA) \ org.openisis.Charset # make additional ee classes j2ee: openisis.jar $(EECLASSES:%=org/openisis/%.class) : # openisis.jar: $(foreach cls,$(CLASSES),org/openisis/$(cls).class) openisis.jar: $(CLASSES:%=org/openisis/%.class) jar cvf $@ org/openisis/*.class %.class: %.java $(JC) -classpath .:$(CLASSPATH) $< %.h: %.class javah -jni -o $@ $(subst /,.,$*) $(MTO)/openjsis.o: openjsis.c @- [ -d $(MTO) ] || mkdir $(MTO) $(CC) -c $(CPPFLAGS) $(CFLAGS) $(JNIINC) -o $@ $< $(MTO)/libopenjsis.so: $(MTO)/openjsis.o $(MTO)/libopenisis.a $(CC) $(LDSHARED) -o $@ $^ %.dll: $(CC) -mdll -o junk.tmp -Wl,--base-file,$*.tmp $($*_deps) dlltool --dllname $@ --base-file $*.tmp \ --output-exp $*.exp --input-def $*.def dlltool --dllname $@ --input-def $*.def \ --output-lib lib$*.a gcc -mdll -mconsole -o $@ $($*_deps) -Wl,$*.exp,--enable-stdcall-fixup # dependences of win32 dlls # note on cross-builds: make sure you have the proper jni headers included - # those supplied for the target platform openjsis_deps=openjsis.o $(LOBJS) openjsis.dll: $(openjsis_deps) # ############################# rules ################################## # # $@ target $^ all prerequisites $? newer $< first $* stem # mto/%.o: %.c @- [ -d mto ] || mkdir mto $(CC) -c $(CPPFLAGS) $(MTFLAGS) $(CFLAGS) -o $@ $< # sto or win $(STO)/%.o: %.c @- [ -d $(STO) ] || mkdir $(STO) $(CC) -c $(CPPFLAGS) $(CFLAGS) -o $@ $< # maintenance targets ... most need a posix shell ci: clean : +++ enter checkin message +++ -read msg && ci -l -m"$$msg" * Version: $(BINARY) echo "OpenIsis `$(OPENISIS) -version`" >Version ./lscvs -lr | sort >>Version tar: Version @-[ -d tar ] || mkdir tar ./lscvs -r | sort | tar czvfT tar/openisis.`$(OPENISIS) -version`.tgz - snap: clean @-[ -d tar ] || mkdir tar tar czvf tar/oi`date +%y%m%d`.tgz . --exclude CVS --exclude tar --exclude WEB-INF headers: for f in *.h *.c */*.h */*.c perl/*.xs org/openisis/*.java; do \ cat Header.txt >$$f.tmp; \ sed -e '1,/^EOH/d' $$f >>$$f.tmp; \ mv -f $$f.tmp $$f; \ done for f in `grep -rl '^#EOH' .`; do \ head -n 1 $$f >$$f.tmp; \ sed -e 's/^/#/' Header.txt >>$$f.tmp; \ sed -e '1,/^#EOH/d' $$f >>$$f.tmp; \ echo $$f; \ diff $$f $$f.tmp; \ mv -f $$f.tmp $$f; \ done # please do not try to make deps w/o sed, gcc and mv ... deps: gcc -MM `./lscvs |grep '\.c$$' |sort` >deps.tmp sed -n -e '1,/^# generated deps/p' Makefile >Makefile.tmp sed -e '/^[^ ]/s/^/mto\//' >Makefile.tmp sed -e '/^[^ ]/s/^/$$(STO)\//' >Makefile.tmp rm deps.tmp mv -f Makefile Makefile.old mv Makefile.tmp Makefile -diff Makefile Makefile.old # generated deps -- do not edit below this line mto/demo.o: demo.c openisis.h mto/ioperf.o: ioperf.c mto/lbt.o: lbt.c lio.h luti.h loi.h openisis.h lll.h lbt.h ldb.h mto/lcli.o: lcli.c openisis.h loi.h lio.h luti.h lll.h lcli.h mto/lcs.o: lcs.c luti.h loi.h openisis.h lio.h lll.h lcs.h mto/ldb.o: ldb.c lstr.h lio.h luti.h loi.h openisis.h lll.h lbt.h ldb.h \ lcs.h lfdt.h mto/ldsp.o: ldsp.c openisis.h loi.h ldb.h luti.h lll.h ldsp.h lfdt.h mto/lenc.o: lenc.c openisis.h ldsp.h mto/lfdt.o: lfdt.c openisis.h loi.h lfdt.h luti.h mto/lfmt.o: lfmt.c ldb.h luti.h loi.h openisis.h lll.h mto/lio.o: lio.c luti.h loi.h openisis.h lses.h lio.h lll.h lstr.h mto/lqry.o: lqry.c ldb.h luti.h loi.h openisis.h lll.h mto/lrec.o: lrec.c ldb.h luti.h loi.h openisis.h lll.h lio.h lcs.h mto/lrectest.o: lrectest.c openisis.h loi.h lfdt.h luti.h mto/lses.o: lses.c lses.h loi.h openisis.h lio.h luti.h lll.h mto/lstb.o: lstb.c openisis.h loi.h lcli.h ldb.h luti.h lll.h ldsp.h \ lfdt.h mto/lstbtest.o: lstbtest.c openisis.h loi.h ldb.h luti.h lll.h mto/lstr.o: lstr.c lstr.h luti.h loi.h openisis.h mto/lsv.o: lsv.c lsv.h loi.h openisis.h lio.h luti.h lll.h lses.h mto/luti.o: luti.c openisis.h luti.h loi.h mto/mlock.o: mlock.c mto/oipl.o: oipl.c mto/ois.o: ois.c lsv.h loi.h openisis.h lio.h luti.h lll.h lses.h mto/openisis.o: openisis.c openisis.h lio.h luti.h loi.h lll.h mto/openjsis.o: openjsis.c openisis.h org/openisis/NativeDb.h mto/tpl.o: tpl.c ldb.h luti.h loi.h openisis.h lll.h $(STO)/demo.o: demo.c openisis.h $(STO)/ioperf.o: ioperf.c $(STO)/lbt.o: lbt.c lio.h luti.h loi.h openisis.h lll.h lbt.h ldb.h $(STO)/lcli.o: lcli.c openisis.h loi.h lio.h luti.h lll.h lcli.h $(STO)/lcs.o: lcs.c luti.h loi.h openisis.h lio.h lll.h lcs.h $(STO)/ldb.o: ldb.c lstr.h lio.h luti.h loi.h openisis.h lll.h lbt.h ldb.h \ lcs.h lfdt.h $(STO)/ldsp.o: ldsp.c openisis.h loi.h ldb.h luti.h lll.h ldsp.h lfdt.h $(STO)/lenc.o: lenc.c openisis.h ldsp.h $(STO)/lfdt.o: lfdt.c openisis.h loi.h lfdt.h luti.h $(STO)/lfmt.o: lfmt.c ldb.h luti.h loi.h openisis.h lll.h $(STO)/lio.o: lio.c luti.h loi.h openisis.h lses.h lio.h lll.h lstr.h $(STO)/lqry.o: lqry.c ldb.h luti.h loi.h openisis.h lll.h $(STO)/lrec.o: lrec.c ldb.h luti.h loi.h openisis.h lll.h lio.h lcs.h $(STO)/lrectest.o: lrectest.c openisis.h loi.h lfdt.h luti.h $(STO)/lses.o: lses.c lses.h loi.h openisis.h lio.h luti.h lll.h $(STO)/lstb.o: lstb.c openisis.h loi.h lcli.h ldb.h luti.h lll.h ldsp.h \ lfdt.h $(STO)/lstbtest.o: lstbtest.c openisis.h loi.h ldb.h luti.h lll.h $(STO)/lstr.o: lstr.c lstr.h luti.h loi.h openisis.h $(STO)/lsv.o: lsv.c lsv.h loi.h openisis.h lio.h luti.h lll.h lses.h $(STO)/luti.o: luti.c openisis.h luti.h loi.h $(STO)/mlock.o: mlock.c $(STO)/oipl.o: oipl.c $(STO)/ois.o: ois.c lsv.h loi.h openisis.h lio.h luti.h lll.h lses.h $(STO)/openisis.o: openisis.c openisis.h lio.h luti.h loi.h lll.h $(STO)/openjsis.o: openjsis.c openisis.h org/openisis/NativeDb.h $(STO)/tpl.o: tpl.c ldb.h luti.h loi.h openisis.h lll.h README0100644000113000011300000000207207644370673010434 0ustar kripkripopenisis - an open implementation of the ISIS database Version 0.8.x (microversion see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA This software is dedicated to the memory of Eckart Dietrich. This software is inspired by (but contains no code of) the iAPI Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. See iAPI.txt for what it contains. Version0100664000113000011300000002333207665654623011133 0ustar kripkripOpenIsis 0.8.9 Changes 1.1.1.1 Sun Sep 30 00:59:40 2001 Demo.jsp 1.12 Tue Oct 29 17:06:02 2002 Header.txt 1.8 Mon Apr 7 21:55:44 2003 LGPL.txt 1.1.1.1 Sun Sep 30 00:59:41 2001 MANIFEST 1.1.1.1 Sun Sep 30 00:59:40 2001 Makefile 1.85 Fri May 30 09:43:31 2003 README 1.1 Mon Apr 7 21:46:35 2003 Version 1.18 Fri May 30 13:15:31 2003 WEB-INF/web.xml 1.1 Mon Feb 4 19:54:31 2002 any2txt 1.4 Tue Apr 8 00:20:52 2003 demo.c 1.10 Tue Apr 8 00:20:52 2003 doc.jsp 1.6 Tue Jan 7 15:25:03 2003 doc/.index 1.12 Thu May 8 14:04:21 2003 doc/AboutUs.txt 1.4 Wed Apr 30 17:50:17 2003 doc/Api.txt 1.8 Fri May 2 14:25:32 2003 doc/Bindings.txt 1.2 Wed Apr 30 17:50:17 2003 doc/Builtin.txt 1.1 Wed Apr 30 17:50:17 2003 doc/CdsIsis.txt 1.1 Wed Apr 30 17:50:17 2003 doc/Client.txt 1.1 Wed Apr 30 17:50:17 2003 doc/Collating.txt 1.2 Fri May 2 15:41:16 2003 doc/Concurrency.txt 1.6 Wed Apr 30 17:50:17 2003 doc/Config.txt 1.3 Wed Apr 30 17:50:17 2003 doc/Contact.txt 1.3 Wed Apr 30 17:50:17 2003 doc/CsTables.txt 1.3 Wed Apr 30 17:50:17 2003 doc/Cuba.txt 1.6 Wed Apr 30 17:50:17 2003 doc/CubaPHP.txt 1.2 Wed Apr 30 17:50:17 2003 doc/DbFiles.txt 1.1 Mon May 26 10:52:00 2003 doc/FactSheet.txt 1.1 Wed Apr 30 17:50:17 2003 doc/GetIt.txt 1.10 Fri May 30 13:11:54 2003 doc/History.txt 1.10 Fri May 30 13:11:36 2003 doc/HowTo.txt 1.1 Wed Apr 30 17:50:17 2003 doc/Inverting.txt 1.3 Thu May 8 13:35:20 2003 doc/IsisIntro.txt 1.1 Wed Apr 30 17:50:17 2003 doc/IsisLinks.txt 1.7 Wed Apr 30 17:50:17 2003 doc/Java.txt 1.1 Wed Apr 30 17:50:17 2003 doc/Meta.txt 1.4 Wed Apr 30 17:50:17 2003 doc/MikiDoc.txt 1.2 Wed Apr 30 17:50:17 2003 doc/News.txt 1.24 Fri May 30 13:14:18 2003 doc/OpenIsis.txt 1.2 Wed Apr 30 17:50:17 2003 doc/OpenSource.txt 1.1 Wed Apr 30 17:50:17 2003 doc/OverView.txt 1.5 Wed Apr 30 17:50:17 2003 doc/PHP.txt 1.1 Wed Apr 30 17:50:17 2003 doc/PatchWork.txt 1.1 Wed May 14 15:20:25 2003 doc/RdbConv.txt 1.4 Wed Apr 30 17:50:17 2003 doc/SeeIt.txt 1.5 Wed Apr 30 17:50:17 2003 doc/Selbstdarstellung.txt 1.2 Wed Apr 30 17:50:17 2003 doc/Serialized.txt 1.7 Fri May 30 07:00:13 2003 doc/Server.txt 1.3 Wed Apr 30 17:50:17 2003 doc/Status.txt 1.5 Wed Apr 30 17:50:17 2003 doc/Struct.txt 1.5 Wed Apr 30 17:50:17 2003 doc/Syspar.txt 1.1 Wed Apr 30 17:50:17 2003 doc/Tcl.txt 1.14 Fri May 2 14:25:35 2003 doc/ToDo.txt 1.6 Wed Apr 30 17:54:04 2003 doc/UeberSicht.txt 1.4 Wed Apr 30 17:50:17 2003 doc/UniStat.txt 1.2 Fri May 2 15:54:45 2003 doc/Unicode.txt 1.2 Thu May 8 14:04:39 2003 doc/Views.txt 1.2 Mon May 5 11:00:44 2003 doc/WritePerf.txt 1.6 Wed Apr 30 17:50:17 2003 doc/XReunion.txt 1.1 Wed Apr 30 17:50:17 2003 doc/btree.txt 1.6 Wed Apr 30 17:50:17 2003 doc/charsets.html 1.4 Mon May 5 10:55:35 2003 doc/encoding.txt 1.2 Wed Apr 30 17:50:17 2003 doc/formatting.txt 1.8 Wed Apr 30 17:50:17 2003 doc/fst.txt 1.2 Wed Apr 30 17:50:17 2003 doc/gizmo.txt 1.2 Wed Apr 30 17:50:17 2003 doc/limits.txt 1.3 Wed Apr 30 17:50:17 2003 doc/longidx.txt 1.2 Wed Apr 30 17:50:17 2003 doc/mx.txt 1.2 Wed Apr 30 17:50:17 2003 doc/sfindex.html 1.1 Wed Apr 30 17:50:17 2003 doc/threading.txt 1.7 Wed Apr 30 17:50:17 2003 doc/trad-fmt.txt 1.2 Wed Apr 30 17:50:17 2003 doc/unirec.txt 1.6 Wed Apr 30 17:50:17 2003 doc/whatabout.txt 1.8 Wed Apr 30 17:50:17 2003 doc/writing.txt 1.3 Wed Apr 30 17:50:17 2003 doc/xmlisis.txt 1.6 Wed Apr 30 17:50:17 2003 foo.oip 1.1 Mon Mar 17 14:08:37 2003 fulltext 1.3 Mon Feb 24 20:32:49 2003 iAPI.txt 1.1.1.1 Sun Sep 30 00:59:40 2001 idm.jsp 1.4 Fri Dec 13 21:30:27 2002 img/Makefile 1.2 Tue Apr 8 00:20:53 2003 img/isi.cursor 1.1 Mon Jan 27 19:05:28 2003 img/isi.pgm 1.1 Mon Jan 27 18:33:46 2003 img/logo.pgm 1.1 Mon Jan 27 19:23:03 2003 index-es.html 1.2 Mon Sep 2 17:42:02 2002 index.html 1.32 Mon Oct 21 10:24:16 2002 ioperf.c 1.3 Tue Apr 8 00:04:49 2003 isocvt 1.1 Mon Feb 4 19:54:31 2002 jdocstyle.css 1.1 Mon Mar 18 21:23:39 2002 jtest.bat 1.1 Wed Aug 28 14:47:13 2002 jtest.txt 1.2 Mon Mar 18 13:56:10 2002 lbt.c 1.35 Wed May 28 15:11:55 2003 lbt.h 1.12 Tue May 27 11:03:30 2003 lcli.c 1.13 Mon May 19 13:56:47 2003 lcli.h 1.8 Tue May 6 21:01:21 2003 lcs.c 1.8 Tue Apr 8 00:20:52 2003 lcs.h 1.3 Fri Apr 11 12:17:43 2003 ldb.c 1.92 Fri May 30 11:28:15 2003 ldb.h 1.19 Tue Apr 8 00:20:52 2003 ldsp.c 1.31 Mon May 19 13:56:48 2003 ldsp.h 1.3 Mon May 19 13:56:48 2003 lenc.c 1.1 Tue Apr 29 11:38:29 2003 lfdt.c 1.23 Mon May 19 13:56:48 2003 lfdt.h 1.10 Thu May 8 13:13:21 2003 lfmt.c 1.14 Tue Apr 8 00:20:52 2003 lio.c 1.42 Fri May 30 11:29:13 2003 lio.h 1.22 Fri May 30 07:00:13 2003 lll.h 1.3 Fri May 30 11:28:52 2003 loi.h 1.30 Tue May 27 10:28:59 2003 lqry.c 1.7 Tue Apr 8 00:20:52 2003 lrec.c 1.53 Fri May 30 07:00:13 2003 lrectest.c 1.12 Fri May 2 14:25:30 2003 lscvs 1.2 Fri Dec 13 15:42:42 2002 lses.c 1.18 Mon May 26 10:56:01 2003 lses.h 1.8 Thu Apr 10 13:43:44 2003 lstb.c 1.40 Mon May 19 13:56:48 2003 lstbtest.c 1.21 Fri May 9 13:43:13 2003 lstr.c 1.13 Tue Apr 8 00:20:53 2003 lstr.h 1.15 Tue Apr 8 00:20:53 2003 lsv.c 1.12 Fri Apr 11 12:21:15 2003 lsv.h 1.10 Tue Apr 8 00:20:53 2003 luti.c 1.47 Mon May 5 10:55:34 2003 luti.h 1.30 Mon Apr 28 18:00:58 2003 make.mingw 1.1 Thu Nov 28 14:18:03 2002 mlock.c 1.4 Tue Apr 8 00:04:50 2003 msgdb/key.mst 1.1 Wed May 14 10:16:06 2003 msgdb/key.oxi 1.2 Mon May 19 13:56:51 2003 msgdb/key.xrf 1.1 Wed May 14 10:16:06 2003 oip.h 1.2 Tue Apr 8 00:20:53 2003 oipc 1.2 Tue Apr 8 00:20:53 2003 oipl.c 1.2 Tue Apr 8 00:20:53 2003 ois.c 1.8 Fri Apr 11 11:09:29 2003 oister 1.5 Tue Apr 8 00:20:53 2003 openisis.c 1.66 Fri May 30 13:09:04 2003 openisis.h 1.89 Fri May 30 13:15:07 2003 openjsis.c 1.17 Tue Apr 8 00:20:53 2003 openjsis.def 1.4 Wed Mar 5 19:59:33 2003 org/openisis/Charset.java 1.3 Tue Apr 8 00:20:53 2003 org/openisis/Db.java 1.6 Tue Apr 8 00:20:53 2003 org/openisis/Fdt.java 1.5 Tue Apr 8 00:20:53 2003 org/openisis/Field.java 1.5 Tue Apr 8 00:20:53 2003 org/openisis/Log.java 1.1 Sat Apr 12 14:48:21 2003 org/openisis/Mail.java 1.4 Tue Apr 8 00:20:53 2003 org/openisis/NativeDb.java 1.5 Tue Apr 8 00:20:53 2003 org/openisis/PlainSerializer.java 1.8 Mon May 5 10:55:35 2003 org/openisis/PureDb.java 1.4 Tue Apr 8 00:20:53 2003 org/openisis/Rec.java 1.4 Tue Apr 8 00:20:53 2003 org/openisis/RecToXml.java 1.2 Tue Apr 8 00:04:50 2003 org/openisis/Test.java 1.4 Tue Apr 8 00:20:53 2003 org/openisis/Utils.java 1.2 Tue Apr 8 00:20:53 2003 org/openisis/XmlTest.java 1.3 Tue Apr 8 00:04:50 2003 org/openisis/XmlToRec.java 1.2 Tue Apr 8 00:04:50 2003 perl/Makefile.PL 1.4 Fri Apr 11 12:21:15 2003 perl/OpenIsis.pm 1.3 Tue Apr 8 00:20:53 2003 perl/OpenIsis.xs 1.5 Tue Apr 8 00:20:53 2003 perl/demo.pl 1.4 Tue Apr 8 00:20:53 2003 perl/encoding.pl 1.1 Wed Feb 26 11:38:18 2003 perl/encodings 1.1 Wed Feb 26 11:38:44 2003 perl/test.pl 1.3 Tue Apr 8 00:17:40 2003 php/Isis.php 1.4 Fri May 23 15:49:56 2003 php/Isis/Db.php 1.2 Fri May 23 17:18:49 2003 php/Isis/File.php 1.3 Mon May 26 10:56:02 2003 php/Isis/Http.php 1.2 Thu May 22 13:48:48 2003 php/Isis/Rec.php 1.8 Fri May 23 17:18:49 2003 php/Isis/Server.php 1.2 Mon May 26 14:55:48 2003 php/echo 1.1 Wed May 21 14:57:01 2003 php/file 1.1 Fri May 23 15:43:01 2003 php/foo 1.8 Fri May 23 15:49:56 2003 php/php.cgi 1.1 Sun May 18 19:58:11 2003 php/query 1.1 Fri May 23 17:17:32 2003 searchtest.txt 1.1 Sat Jan 5 01:27:15 2002 tcl/Makefile 1.20 Fri May 30 09:43:49 2003 tcl/README 1.2 Wed Apr 30 17:13:09 2003 tcl/addmsgkey.tcl 1.2 Fri May 23 15:53:43 2003 tcl/configure 1.5 Sat Apr 12 18:56:47 2003 tcl/demo.cgi 1.1 Tue Apr 29 19:01:14 2003 tcl/detail.tsp 1.10 Thu May 8 12:03:49 2003 tcl/dexml 1.1 Mon Apr 28 14:19:22 2003 tcl/fdt21.tcl 1.3 Fri May 9 13:42:36 2003 tcl/fullinv 1.2 Thu May 8 13:16:14 2003 tcl/isish.c 1.5 Tue May 27 10:30:00 2003 tcl/lib/mlcm.db.tcl 1.3 Mon May 26 10:50:23 2003 tcl/lib/mlcm.gui.tcl 1.4 Mon May 26 10:50:23 2003 tcl/lib/mlcm.lang.tcl 1.3 Mon May 19 13:56:53 2003 tcl/lib/mlcm.menu.tcl 1.5 Mon May 19 13:56:53 2003 tcl/lib/mlcm.meta.tcl 1.15 Mon May 26 10:50:23 2003 tcl/lib/mlcm.prof.tcl 1.6 Fri May 23 15:53:45 2003 tcl/lib/mlcm.util.tcl 1.5 Fri May 23 15:53:45 2003 tcl/list.tsp 1.6 Thu May 8 11:54:06 2003 tcl/lp 1.6 Mon Feb 17 11:28:17 2003 tcl/lp2db 1.2 Mon Feb 17 11:28:17 2003 tcl/lpdb 1.1 Mon Jan 27 20:51:11 2003 tcl/maketcltk 1.14 Mon Apr 28 18:01:05 2003 tcl/md 1.2 Sun Apr 20 14:09:34 2003 tcl/mf 1.1 Wed May 14 15:42:02 2003 tcl/mlcm 1.7 Mon May 26 10:50:23 2003 tcl/oits.c 1.11 Tue May 27 10:30:00 2003 tcl/oitsrc 1.1 Wed Mar 26 13:36:18 2003 tcl/oitstest 1.1 Fri Mar 28 19:07:49 2003 tcl/openIsis.tcl 1.10 Wed May 14 14:49:38 2003 tcl/openisistcl.087.c 1.2 Tue Apr 8 00:20:54 2003 tcl/openisistcl.c 1.66 Fri May 23 15:53:45 2003 tcl/openisistcl.h 1.10 Mon Apr 28 18:01:06 2003 tcl/sample.fsp 1.3 Thu May 8 12:04:53 2003 tcl/test/readall.tcl 1.2 Thu May 8 13:12:49 2003 tcl/test/test.tsp 1.2 Mon Apr 28 18:01:07 2003 tcl/test/test1.087.tcl 1.1 Thu Mar 6 16:01:47 2003 tcl/test/test1.tcl 1.21 Fri May 30 07:00:21 2003 tcl/test/test2.tcl 1.4 Mon Apr 28 18:01:07 2003 tcl/test/test4.087.tk 1.1 Thu Mar 6 16:01:47 2003 tcl/test/test5.087.tcl 1.1 Thu Mar 6 16:01:47 2003 tcl/test/testcli.tcl 1.3 Fri May 23 15:53:46 2003 tcl/test/testfdt.tcl 1.1 Sat May 3 16:40:47 2003 tcl/test/testrec.tcl 1.7 Mon May 12 21:31:08 2003 tcl/test/testsrv.tcl 1.3 Thu May 15 18:59:20 2003 tcl/test/teststb.tcl 1.16 Fri May 9 13:42:38 2003 tcl/tisis 1.12 Thu May 8 12:09:35 2003 tcl/tkhtml/Makefile 1.3 Sat Apr 12 18:23:25 2003 tcl/tspr.c 1.5 Tue May 27 10:30:00 2003 tcl/wisish.c 1.7 Tue May 27 10:30:00 2003 test.pft 1.1 Mon Jun 24 11:22:26 2002 testres.txt 1.4 Fri May 30 07:00:13 2003 tomcat 1.4 Mon Mar 18 16:53:00 2002 tpl.c 1.9 Tue Apr 8 00:20:53 2003 unicode.jsp 1.3 Mon Mar 4 23:14:53 2002 unistat 1.2 Fri May 2 15:54:31 2003 WEB-INF/web.xml0100664000113000011300000000027507427563367012112 0ustar kripkrip OpenIsis any2txt0100755000113000011300000000654107644412744011114 0ustar kripkrip#!/bin/bash #/* # openisis - an open implementation of the CDS/ISIS database # Version 0.8.x (patchlevel see file Version) # Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # see README for more information #EOH */ # # CVSID='$Id: any2txt,v 1.4 2003/04/08 00:20:52 kripke Exp $' # # utility to convert several document types to text # used for mail archiving # similar to anytopnm # # assumptions: # filename is a name of the file in the current directory # we are allowed to use any filename.* # # return code: # the lowest valued bit denotes that there is a plain version in filename.txt # # 0 file was fine text and is unmodified # 1 file was sort of (non-iso,long-lined) text, plain version is filename.txt # 2 file was formatted html text # 3 file was formatted html text, plain version is filename.txt # 4,5 likewise for pdf # 6,7 likewise for rtf # 8,9 likewise for doc (M$-w0rd) # 16,17 file was mail text # 32 gif image # 34 jpg image # 36 bmp image # 38 png image # 40 tiff image # 46 other image # ?? other # 124 unknown # 126 error # verbose=false err () { echo "$*" >&2 exit 126 } dewindoofy () { tr '\200-\237' 'E?,f".+#^%S<Ö?Z??'\'\''""*--"Ts>ö?zY' } echo "any2txt $* ($PWD)" >&2 while [ 1 -lt $# ]; do # shift options case "$1" in -v) verbose=true;; -m) mtype="$2"; shift;; esac shift done # if $verbose; then # : #else # no output expected; so any output is error exec >&2 #fi filename="$1" [ -f "$filename" -a -r "$filename" ] || err "'$filename' not readable" ftype="$(file -b "$filename")" code=124 case "$ftype" in *"mail text"*) sed -e '1,/^$/d' <"$filename" >"$filename".txt code=17 ;; *HTML*) # maybe ASCII or ISO-8859 HTML document text, so check before ASCII code=2 lynx -dump -force_html -nolog "$filename" >"$filename".txt && code=3 ;; *"Non-ISO extended-ASCII"*"long lines") # assume windows stuff dewindoofy <"$filename" | fmt -s >"$filename".txt code=1 ;; *"Non-ISO extended-ASCII"*) # assume windows stuff dewindoofy <"$filename" >"$filename".txt code=1 ;; ASCII*text*"long lines"|ISO-8859*text*"long lines") fmt -s <"$filename" >"$filename".txt code=1 ;; ASCII*text*|ISO-8859*text*) code=0 ;; PDF*) code=4 pdftotext -raw "$filename" - >"$filename".txt && code=5 ;; "Rich Text Format"*) code=6 unrtf --nopict --html "$filename" | lynx -dump -force_html -nolist /proc/self/fd/0 >"$filename".txt && code=7 ;; "Microsoft Office Document"*) code=8 antiword "$filename" >"$filename".txt && code=9 ;; *"GIF image"*) code=32;; *"JPEG image"*) code=34;; *"PC bitmap"*) code=36;; *"PNG image"*) code=38;; *"TIFF image"*) code=40;; *image*) code=46;; esac $verbose && echo "$filename: $ftype: $code" >&2 exit $code demo.c0100644000113000011300000000314207644412744010637 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: demo.c,v 1.10 2003/04/08 00:20:52 kripke Exp $ demo main file */ #include /* NULL */ #include /* printf */ #include /* atol */ #include "openisis.h" int main ( int argc, const char **argv ) { int rowid = atoi(argv[argc-1]); int db, i; OpenIsisRec *rec; if ( 0L == rowid ) { printf( "usage: demo -db \n\tprints row with number \n" ); return 1; } openIsisLog( 0, NULL ); db = openIsisOpen( NULL, argv, argc ); rec = openIsisReadRow( db, rowid ); for ( i=0; ilen; i++ ) printf( "<%d>%s\n", rec->field[i].tag, openIsis2Html(rec->field[i].val, rec->field[i].len), rec->field[i].tag ); return 0; } /* demo main */ doc.jsp0100644000113000011300000001170307606570517011035 0ustar kripkrip<%@ page import="java.io.*, java.util.*, org.openisis.*" session="false" %><%! // need at least Field.html static File _dir ; // have no reliable ServletConfig before init static long timestamp; static String[] files; static synchronized void refresh ( ServletContext a ) throws IOException { if ( null == _dir ) _dir = new File( a.getRealPath("doc") ); files = _dir.list( new FilenameFilter() { public boolean accept( File _dir, String name ) { return name.endsWith( ".txt" ); } } ); Arrays.sort( files ); timestamp = _dir.lastModified(); } // mini wiki-style doc static String mikidoc ( Writer o, BufferedReader in ) throws IOException { String notes = null; String l; boolean inli = false; boolean br = false; // break on any line boolean code = false; // code mode while ( null != (l = in.readLine()) ) { int len = l.length(); boolean isli = false; char c = 0 == len ? 0 : l.charAt(0); int tab = 1 < len && '\t' == l.charAt(1) ? 2 : 1; if ( code && '$' != c ) l = Field.html( l ); else switch( c ) { case 0: l = "
"; break; case '$': // toggle code l = (code = !code) ? "

" : "
"; break; case '*': // head String bg = "#CCCCCC"; String anchor = ""; if ( 1==tab && 2 < len && ' ' != l.charAt(1) ) { /* make anchor */ int i = l.indexOf( '\t' ); if ( 1 < i ) { tab = i+1; anchor = " "; } } l = "" +anchor+l.substring(tab) +""; break; case '-': if ( l.startsWith("---") ) { // rule l = "
"; break; // rule :) } // new list item l = (inli ? "" : "" ); o.write( l ); if ( br && ! l.endsWith( "
" ) ) o.write( "
\n" ); else o.write( '\n' ); inli = isli; } return notes; } static String mikidoc ( Writer o, String file ) throws IOException { file = _dir+File.separator+file; try { return mikidoc( o, new BufferedReader( new InputStreamReader( new FileInputStream( file ) ) ) ); } catch (IOException e) { o.write( "" ); } return null; } %><% if ( null == _dir || timestamp != _dir.lastModified() ) refresh( application ); String right = null; String file = request.getParameter("file")+".txt"; boolean sides = null == request.getParameter("print"); String name = ""; int f = files.length; while ( 0 < f-- && !files[f].equals(file) ) ; file = 0 > f ? "OverView.txt" : files[f]; name = file.substring( 0, file.length()-4 ); %> OpenIsis - <%=name%> <% if ( sides ) { %> <%-- left: the index --%> <%-- middle: the content --%>
<% mikidoc( out, ".index" ); %>
print
plaintext
<% } // sides %>
<%= name %>
<% right = mikidoc( out, file ); %>
<% if ( sides ) { %> <% if ( null != right ) { %> <%-- right: notes --%>
<%= right %>
<% mikidoc( out, right+".txt" ); %>
<% } %>
<% } // sides %> doc/.index0100644000113000011300000000175207656461745011441 0ustar kripkrip! > OverView Isi * organization > UeberSicht der OpenIsis Verein > OpenSource open source ; > InfoSys information systems > AboutUs about us > http://wiki.openisis.org/i/view our Wiki * software > OpenIsis OpenIsis > FactSheet at a glance > Status current status > SeeIt sample databases > GetIt download > http://sourceforge.net/projects/isis/ @ sourceforge > IsisLinks links ... * Isis databases > IsisIntro Isis principles > whatabout Isis compared > unirec the Isis record > Struct embedded structures > xmlisis XML and RDF > RdbConv ISIS and RDBs > Unicode * OpenIsis details > History > HowTo installation > Api the API > Client > Server > Bindings languages > Tcl > Java > Collating > Inverting > Views > btree new B-L-Tree index > Concurrency > threading multi-threading > CsTables Latin1 and Cp850 > MikiDoc doc structure * traditional > Meta FDT/metadata > formatting PFT > fst FST > Syspar doc/AboutUs.txt0100664000113000011300000000231307654006331012430 0ustar kripkripThe board of the OpenIsis society is - Erik Grziwotz (chair) pgp - Gabi Rohmann pgp - Ingo Struck pgp - Thomas Sonnemann OpenIsis.org is sponsored by > http://www.sferix.com/ sferix GmbH and > http://www.allmaxx.de/ allmaxx.de As a student's site, allmaxx supports open software with a focus on education and knowledge management. We closely cooperate with the > http://www.oc4s.org/ open community for science . Currently the site is maintained by > mailto:erik@openisis.org Erik Grziwotz and > mailto:paul@openisis.org Klaus Paul Ripke. The new PHP version was written by > mailto:braulio@openisis.org Braulio José Solano Rojas. Volunteers are very welcome. Openisis sources are available via CVS at > http://sourceforge.net/projects/isis/ SourceForge side by side with Franck Martin's PHP isis project. Thanks, Franck! ... Contact doc/Api.txt0100644000113000011300000001171407654477734011605 0ustar kripkripNote: the follwing list might be somewhat out of date, but probably you get the picture. There is a major overhaul of the API from 0.8.7 to 0.9 underway: The concept of a session is introduced to allow for safe and efficient multithreaded concurrency. Nearly every API call has a Session parameter, and basic system services like memory allocation and stdio are bound to sessions. All but the control calls may be called from any session in any thread (but from no single session in more than one thread). All but the database calls will moreover execute in parallel without any interlocking. The new Api is organised in the following sections: - M et al: memory & utilities - S: session & stream IO - R: record - D: db IO - F: fdt utilities - N: network stubs - C: control the control calls are the only calls w/o a Session parameter, since they MUST NOT be called from any but the default session There is also an > Concurrency overview on concurrency and synchronization in OpenIsis. * naming conventions All calls besides some common utilities start with the letter of their respective section followed by an uppercase letter. Functions are mixed case with initial lower, types are mixed case with initial upper and macros and enum values are all upper, variables are all lower case. For convenience, there are two versions of the "OpenIsis namespace": - in the official version as given by openisis.h, every call starts with openIsis followed by an uppercase letter. Every type starts with OpenIsis, and every macro with OPENISIS_. - inside the "OpenIsis namespace", as given by loi.h, the leading openIsis is cut of to avoid cluttering, while preserving the case rules. As an example, OpenIsisSession and openIsisMAlloc may be used as Session and mAlloc with loi.h. * transition The old (0.8.7) interface is retained via macros refering to the default session. During the transition process, some calls will also get a redesign with respect to other parameters, for example, by replacing an argv/argc list by a Record. It should, however, be reasonably safe to refer to the old interface. The following list is neither complete nor fixed, but rather a roadmap. The mapping of old to new calls is not always one-to-one. * M et al: memory & utilities $ mAlloc mFree mDup toHtml openIsis2Html utf8Chk openIsisValidUTF8 $ * S: session & stream IO $ Session OpenIsisSession Stream OpenIsisStream sMsg openIsisMsg sOpen openIsisSopen sGets openIsisSGets sReadln openIsisSReadln sGetr openIsisReadStream $ * R: record $ Field OpenIsisField Rec OpenIsisRec rSplit openIsisReadField rDup openIsisClone rMsg openIsisPrintf rFmt openIsisFmt rSel (openIsisFmt) $ * D: db IO $ dRead openIsisReadRow dRaw openIsisReadRaw dFmt openIsisRead dWrite openIsisWrite dMaxId openIsisMaxRowid dScan openIsisScan dQuery openIsisQuery dWhere (openIsisQuery) dTerm openIsisTerm dIndex (openIsisTerm) dIndex openIsisIdxLoop $ * C: control $ cInit openIsisInit cOpen cClose cSet openIsisLog cSession openIsisSesGet cDOpen openIsisOpen cDCheck openIsisCheck cDClose openIsisClose cXOpen openIsisIdxOpen cXBadd openIsisIdxAdd $ In the multithreaded server, sessions are actually created by the multiplexer, which need not be in the same thread as the default session. *modules library modules This, too, describes the should-be rather than current state: - uti: Memory and other basic utilities - io: internal operating system access - ses: Sessions and Streams - fdt: fdt utils and global system fdts - fmt: the rSel formatting - rec: other R tools - stb: Stubs - str: internal structure - cs: character sets - bt: index X access - qry: the dWhere query - db: most Db and Control * initialization order: - cOpen: default cs - cOpen: io fd 0-2 (windoze) - cOpen: default session w/o params, pulls memory - cOpen: default session w/ params - cDOpen: cOpen - cDOpen: str - cDOpen: database - cSession: other sessions *ISOC a note on ISO C It became obvious that there is little we can use from the system, partly due to compatibility problems, partly due to the multithreaded environment we need. stdio doesn't work at all and malloc heap management needs to be wrapped. Thus the use of libc (at least outside of lio.c) is limited more or less to string.h (for strlen, memcpy, memcmp). I found that in many places the sole reason to include any system header, namely stddef.h, is NULL and size_t. We require an ISO C (a.k.a. ANSI C a.k.a. STDC) compiler anyway. Studying the standard, I found that the use of NULL is completely obsolete in ISO C, since the relation between "NULL pointers", the 0 literal and conditional expressions is well defined _by the compiler_. What stddef.h does, is only _suggested_, and looking at gcc's stddef.h, it is a rather desperate attempt to arrange for all that should be _despite_ any system environment. The same holds for size_t, ssize_t, off_t, ptrdiff_t, you name it: it's a broken interface to a broken "standard" library, so we don't use it anymore. --- $Id: Api.txt,v 1.8 2003/05/01 10:06:32 mawag Exp $ doc/Bindings.txt0100644000113000011300000000425407654006331012607 0ustar kripkripoverview of features, platforms and operation modes supported by the the various language bindings * dimensions The bindings are categorized according to whether they support - client sessions All versions of OpenIsis will (some day) support client sessions. Within the database server, however, they are not used (in order to keep it simple, safe and responsive). Where client sessions are used, one can always use multiple client sessions to connect to several servers. multiple sessions from one IP to the same server may or may not be allowed by the server. - multiple threads This is useful only where multiple users are served, i.e. in the server and within a Java Servlet Engine. Multithreaded database access also requires the use of server sessions. - asynchronous operation This is useful only for applications build around an event loop, actually the server and the Tcl/Tk GUI. Of course any asynchronous connection can be easily used synchronous. Asynchronous processing is supported only on a communication channel, i.e. when accessing a remote database. - charset support While any version of OpenIsis will handle any ASCII compatible encoding (including UTF-8) transparently, only Tcl and Java have built-in support for converting charsets. * tabular overview Here, "openisis" denotes the command line utility and server is the openisis server process. $ | *system* | *client* | *threads* | *async* | *charsets* | | Perl | yes | no | no | no | | Java | yes | yes | no | yes | | Tcl/Tk | yes | no | yes | yes | | openisis | yes | no | no | planned | | C-API | yes | yes | yes | planned | | server | no | yes | yes | planned | $ * notes - The Tcl/Tk binding (and thus, of course, the C-API) is the only one actively supported for windows. While Perl and Java should work, we won't bother. - Where a Tcl/Tk GUI wishes to operate asynchronously on a local database, it should start a server process. - Neither Perl nor Tcl are meant to be used in a multiuser environment. --- $Id: Bindings.txt,v 1.2 2003/02/14 17:30:32 kripke Exp $ doc/Builtin.txt0100644000113000011300000000201207654006331012446 0ustar kripkripoverview of OpenIsis builtin metadata * system parameters This FDT includes the fields specified for the > Syspar SYSPAR.PAR file, augmented by system related (process wide) parameters for the OpenIsis tools in the 700 range. * database parameters These are the fields specified for the > Syspar dbname.PAR and .EXP files, augmented by per database parameters for OpenIsis in the 800s. This also includes the metadata fields - worksheets - formats - selections - field specified in the > Meta FDT although in OpenIsis, databases and FDTs do not need to have a strict one-to-one relation (but typically have). * the internal request, query and formatting language This contains the whole story of transforming one record into another. All field ids are five digit (greater than 9999), and structure tags are greater than 16383. * the other protocols language This FDT will contain a collection of commonly used headers from SMTP, NNTP, HTTP and similar stuff. --- $Id: Builtin.txt,v 1.1 2003/01/10 15:34:42 kripke Exp $ doc/CdsIsis.txt0100664000113000011300000000277707654006331012425 0ustar kripkripoverview of isis software The mother of all isis software is a DOS version of "MicroISIS" as an integrated system with textual user interface. There is a BSD version of "CDS/ISIS" which also runs under linux up to some 2.2.x kernels (current 2.4 kernels do not support the iBCS module for COFF binaries). Then there are several versions of "WinISIS" (M$-Windows only, but runs under linux/wine). A shared library version "isis.dll" of functions to access an isis db from your code exists, despite it's name, also in a linux version ("isilux"); however, you need some pretty special libs to make it run. A set of command line tools ("cisis") performs tasks like importing ISO2709 bibliographic databases, inverting (index building) etc. The thing next to an isis database server is "wwwisis", which runs as CGI or from the command line and performs most isis tasks (win and lin versions). However, it actually runs per request, not as a server itself, and thus cannot provide concurrency control. This "official" isis software, which is maintained by > http://www.unesco.org/webworld/isis/isis.htm Unesco and/or > http://www.bireme.br/isis/ Bireme , is accompanied by a couple of independent developments, some of which are in the public domain. > http://web.tiscali.it/javaisis/ Javaisis is an AWT-based GUI (3.5 uses SWING) and a corresponding server, which in turn uses wwwisis. > http://www.jezuici.krakow.pl/soft/iapi/ Robert Janusz wrote a C-lib (iAPI) from scratch, which was the starting point for the openisis software. doc/Client.txt0100664000113000011300000000605107654006331012267 0ustar kripkripthe OpenIsis client API * overview The client API is the set of calls needed to contact an OpenIsis > Server server. Although the server may run in process, it is usually connected via the network, so this is called the "network API", indicated by the letter "N". With the exeption of some local control and utility calls, all calls are bound to a session and asynchronous in nature. The API may, however, be set for synchronous operation, meaning that attempts to retrieve results will block until results are available. Asynchronous operation may not be avialable on all connections. * basics From another point of view, there is only ONE call actually talking to the server, which is split in two halves: - nSend sends a request record to a server In synchronous mode, this will also call nRecv, so that the result record is available immediately. - nRecv receives all or part of a response from the server. In asynchronous mode, this may or may not give a response. At any time, there may be only one request active on a session. The session holds record buffers for both the request and response, which will be overwritten by any successive request. Records embedded within the response will be stripped out, so that they are available separately. * the request The request record may contain the following fields: - database This required field contains a database name. - view a named format to use. The view "#" selects rowids. - rowid Selects the record with the given rowid to be added to the result list. - query Selects the records matching query to be added to the result list. - terms selects a record containing terms to be added to the response. May contain subfields f (from inclusive) and t (to exclusive). - meta selects a record containing the FDT (for the given language). - maxid - a skip rowid - output size (number of records or terms) - a record to write The client session prepares the request with a field containing the session id and a request serial number. Since a client session is commonly used against one database only, it will remember the database used on the first call and automatically prepend a database field to any successive call, if necessary. * the response The response record is made up of: - session id and serial - error - number processed items - embedded result records * client processing sequence a request is processed in the following steps: - create a new request record with session id, serial and database - fill in request fields - send the request by nSend. A server connection is opened as needed. In the asynchronous case, the sending socket is registered against some select loop in order to trigger writing, whenever the socket is writable. - receive the response by nRecv. Asynchronous handling analogous to writing. - completing the response in the session's response buffer may trigger some callback like calling some TCL code In synchronous mode, the complete send, receive and callback sequence will be finished on return of the initial send call. --- $Id: Client.txt,v 1.1 2003/01/06 16:05:06 kripke Exp $ doc/Collating.txt0100664000113000011300000001500007654510634012765 0ustar kripkripThis document discusses the issues involved in collating entries for the index (inverted file). Collating starts after preliminary steps of inverting like selecting data from a record (or other sources), including extraction of keywords and the like, which are described in > Inverting . Collating involves a string transformation, so that the resulting strings - are collation keys, i.e. yield the desired ordering on binary comparision (a la strcmp). - are normalized, i.e. some variants of the same contents are mapped to a canonical base The same transformation is applied to both entries before written to the index and keys before searched in the index. A desired property of this transformation (mapping) is reversibility to a legible representation of the original key, so that index contents can be listed. However, an alternative is to use the original record contents for listing, where this can be easily identified (e.g. from a thesaurus). Since collating involves looking up characters, it can also detect word boundaries. The traditional means provided by the ISISAC.TAB and ISISUC.TAB tables do not extend well to multibyte character sets like UNICODE, so we explore alternatives. * basic collation A basic variant using ISO-C/POSIX means (ctype.h, string.h) would be - using toupper for normalization - using isalpha to detect word boundaries - using strxfrm to create collation keys This fallback approach has several disadvantages: - normalization is very limited - it depends on a "locale" which is a global state of a process, so working with databases with different settings is not multithreading safe. - the dependency from the locale is not specified precisely - there is no standard way to customize settings - it is not revertible * full unicode collation Using a library such as Big Blue's > http://oss.software.ibm.com/icu/userguide/Collate_Intro.html ICU , we can use the full unicode collation algorithm ( > http://unicode.org/unicode/reports/tr10/ UCA ), including the customization options described there. - use the given encoding to convert input to unicode - uppercase and canonicalize input, especially decompose composite characters - remove most diacritics (but not the ring on the A when in sweden) - use the alphabetic character property to detect word boundaries - get a collation key This approach is very powerful. It provides prepared data for virtually all scripts, encodings and locales without the need for customization. It provides for three or four levels of sorting and can even handle french accent ordering (where accents later in the otherwise same word have higher preceedence). Sooner or later, OpenIsis will provide a means to link to ICU. However, this still has a couple of disadvantages - it requires a library that alone is several times larger than OpenIsis. The used algorithms are very complex, so it can not be expected to perform all too well on low hardware ressources. - customization for special normalizations is difficult - reverting a compressed collation key might not always be possible, at least it is a pretty non-trivial task (and not supported by ICU) * isis custom collation We devise a relatively simple scheme, that combines the mappings used for normalization, collation key creation and word boundary detection into one step. It does (by now) not provide multi-level ordering, since the discriminators for secondary levels (like case and diacritics) are typically discarded anyway during normalization. Features are - efficiency - easy customization - independence of encoding - mapping of sequences The mapping is specified by an isis record, that could be stored in an separate file or be part of embedded header data. In any way, the mappings are described using the actual characters, in the encoding of the database, rather than code numbers. This way, simple explicit mappings will survive a recoding (ranges, however, may be broken). Fields used in the customization are mapping rules, containing one or more "items" separated by TAB characters. An item is - a single character if it is a single character (sic!) - a literal sequence if the first one is a quote '"' (34) - a range is it consists of three characters, and the second one is a dash '-' (45) - an enumeration else. A dash at second position or quote at first position can be avoided by permutation or by using multiple rules. The fields in the mapping (record) map the second and following items to the first one. Let n the number of characters in the target (the first item), m those of the source, where a sequence is regarded as a single character. - if n >= m, the source characters are mapped to the first m corresponding target characters - if n < m, the first n source characters are mapped to the corresponding target characters, other source characters are mapped to the last character of the target (or a blank, if n=0). Sequences of characters mapping to the last one are squeezed as with the -s option of the "tr" tool, i.e. produce only one character. All but "map" fields describe a final rule, which assigns a character class (letter,word,other,null) and one or a range of collation values to the listed items. Collation values are assigned in the order of rules (fields). - map do NOT assign a collation value, but process the result recursively. - ignore assigns character class null and no collation value. ignored characters are silently discarded and don't break words (like soft hyphen, combining diacritical marks). - letter assigns character class letter, i.e. characters that are part of words. - word assigns character class word, i.e. characters that are themselves words. - other assigns character class other, i.e. characters that are not part of words and are discarded in word-split mode. The builtin default rule maps all characters to a blank, which thus always has the lowest collation value. In any case, the longest matching sequence is used, and map rules should (?) take preceedence over final rules. TODO: detailled preceedence of conflicting rules. The actual collation key is a sequence of collation values, encoded as unsigned characters, if less than 256, or UTF-8 else. (A fixed-number-of-bits encoding could be used alternatively). The reverse mapping is simple: for each collation value, there is a corresponding final rule that produced this value or a range containing this value. The corresponding character (or sequence) in the first item of this rule is used. Simple example for traditional spanish collation: $ letter A-L a-l letter LL ll letter M-Z m-z map AEIOUN ÁÉÍÓÚÑ áéíóúñ $ doc/Concurrency.txt0100644000113000011300000003150607654006331013344 0ustar kripkripSeveral concurrency issues arise whenever a database is accessed simultaniously be multiple processes or threads (lightweight processes sharing all their system ressources like memory and open files, including file positions). * multiprocess (MP) environments MP environments are distinguished according to - whether all processes are readonly or there are one or more writers - whether processes are single-shot (i.e. open, work, exit like CGI scripts, including PHP in CGI mode) or resident (like PHP module living in Apache 1.x/Unix childs). Note that PHP module in a multithreaded server is not multiprocess. Within a readonly environment, there is not much of a problem. Each process may read and cache file contents independent of each other. So the rest of this section discusses read/write access. In the presence of writers, there are some problems: - at least the actual writing accesses must be *strictly* mutually exclusive - it must be ensured that readers do not use old cached data (or at least use it in a well controlled manner) - it must be ensured that changed data is written and read in a consistent way These problems are addressed in reverse order: - the data structures used in OpenIsis are designed so that a consistent way of reading and writing can be defined. For example, the XRF pointer to a new or changed record is written after the record, so readers will not see an invalid pointer. However, this will work only where the operating system guarantees such semantics for file reads/writes. This again may depend on the filesystem and will not hold for most network file systems. - learning which cached blocks are outdated is not possible with reasonable effort. One possible approach is to not cache at all, i.e. resort to the operating system cache, which hopefully is properly synchronized (see above). - A simple and well supported means of mutual exclusion is the use of exclusive file locks. flock-(BSD-)style locks are sufficient; we do not need locking of file regions nor locking over NFS, which is not reliable anyway (and it's not much better with SMB). The most easy and reliable solution is to completely encapsulate any access --from database open to close-- within an exclusive lock. That way there clearly is no inconsistent cache. - for single-shot processes, this solution is reliable and does not incurr too much cost: exit will release any lock and the processes may not benefit from caching anyway. - for resident processes, the need to open and close on any request is more of a disadvantage, and it is a problem how to guarantee that any lock is released after processing. Possible perfomance enhancements that might be implemented one day: - readers could use shared locks. however, this gives a risk of writer starvation. - if all data access methods are carefully checked and a reasonable local file system is used, non-caching readers could get by without locking at all. - another, quite complex approach is to share cache memory between processes, similar to ORACLE's SGA. This would also help in guaranteeing consistent read-write-sequences. To summarize the multiprocess issues: - readonly access is fine - DO NOT TRY TO WRITE ON A NETWORK DRIVE (or at least make sure it is accessed only by one host at a time) - the best solution for multiple processes is to contact a server for writing instead of doing it themselves - for PHP as module in read/write mode, we have to rely on register_shutdown_function to close any db * multithreaded environments OpenIsis is designed to run multithreaded. Multithreading is used only within some sort of server (like database, web or servlet engine) in order to run multiple requests from multiple clients in parallel. MT environments are distinguished according to - whether they support active dispatching of requests to threads - whether they support parallel IO. Besides the basic calls for parallel IO (like pread,pwrite, or ReadFileEx "overlapped" IO in Win speak, which is missing on Win 9x/Me), this also requires condition variables (like pthread_cond_wait/broadcast, which are rather difficult to emulate on Win 9x/Me in the absence of SignalObjectAndWait) and should include memory mapping (like mmap,msync, which is working poorly on Win 9x/Me). All threads of a single process share the same cache, so dirty caches are not an issue here. Synchronization is cheaper and more easy to use. However, this great performance benefit comes at a price: While there are a few utilities without any side effects (i.e. proper FORTRAN functions), not only access to the database and it's cache, but any access to system ressources like files or the memory heap must be carefully checked for possible collisions and, when in doubt, must be synchronized -- even in a readonly environment. * Session synchronization Our strategy is to share as little as possible between threads and to protect all that must be shared (basically the database) by a single lock. The means to give each thread it's own, unshared environment is the SESSION. A session represents a single client accessing the database. (At least this is the idea, but depends on the dispatcher's abilities, see below). The session may hold result sets from previous queries, some authentication info from the client and other temporary data. In a standalone environment like the Tk GUI not connected to a server, there is only one session, the "default session" (session id 0). In a database or web-server, however, there may exist several sessions on behalf of several users at the same time. Requests from each session are serialized by some dispatcher, so that each session is accessed by at most one thread at a time. Consequently, in an environment with one session only, there also is only one thread used to access the database. To summarize, from a session's point of view, the world is single threaded. Each session has a private memory heap and even it's own IO stream buffers stdin, stdout and stderr (as streams 0,1,2) and need not care about how it is connected and to whom. Since the dispatcher guarantees that no session is accessed by more than one thread at a time, dynamic memory, streaming IO and other session ressources can be used without further interlocking. * dispatching requests and locking sessions Due to the dual nature of a session as both representing a user and serving as object of synchronization, dispatching requests has two tasks: - ensuring serialized (single-threaded) access - finding the session bound to a given user While the former is crucial in MT environments, the latter is used only if - the environment identifies a user session in the first place - the session object's ability to keep state (like result sets) is used We distinguish two cases of when and how dispatching is done: - passive/late dispatching: In most environments we have to get the session from within a thread dedicated to that request. The dispatcher is implemented as a call, accessing a session pool protected by some mutex. - active/early dispatching: Within the database server, the proper session can be looked up before a thread is allocated for a request. Here, the dispatcher is an active component, probably running in a thread on it's own (thus not requiring a mutex on the session pool). That way several requests on the same session may be queued (or discarded) without consuming any thread ressources. This should yield better performance under high load and somewhat better protection against denial of service. There are also two different situations with regard to the scope of synchronization: - per request: The session is "locked" (somehow marked as busy) until processing the request has finished. Locking is done by the dispatcher, and unlocking must be performed on exit, e.g. using register_shutdown_function in PHP. For the passive dispatcher, if some user session id is used to locate an existing session and there is already a request executing in this session, the current thread has to wait. - per use: In a high level language, i.e. Java, basic synchronization is achieved by having a Java object representing the session and marking the appropriate methods as synchronized. Note that unless we promise that sessions actually will remember some state, a simple dispatcher may decide to operate on a session pool of size 1 (one), containing only the default session, thus ruling out any parallel operation. * Configuration synchronization Operations that change the overall system state like opening a database are allowed for session 0 only. Consequently, IO (logging) and memory associated with such operations is bound to the default session. Databases may be marked for exclusive use by session 0 for example during a lengthy batch index update or in order to perform structural changes like modifying the FDT. On the other hand, the worker sessions need some confidence that configuration is not going to change while they are in the midth of processing a request. Therefore, any database that is somehow accessed by a session, is marked as used by the session and marked as unused when the session is released. This protects the database from being closed or put in exclusive ("single-user") mode and thus also configuration from being changed. Note that a request for the database need not be the same as the original user request. For a database server, the request for a database operation is all that is known, thus clients issuing several remote requests won't get no guarantee that the DB is unchanged between database accesses (regardless of the environment they are running in). When accessing a local database, the scope of locking depends on the environment as described above. An explicit lock on a local database might be provided for Java (to be unlocked in a finally clause). However, the situation is not as bad as it might look, since there are complex database accesses, bundling several operations into one. A standard example is to perform a query and not only obtain a result set, but also the contents of the first n records, like with a Z39.50 piggybacked "present". For remote databasse access, this is the most efficient operation mode anyway. * Database synchronization All database ressources like master file and index have associated in memory structures like a cache. These structures must not be accessed by more than one thread at once and are therefore protected by a mutex (some "mutual exclusion" object like a critical section). Again, there are two modes to distinguish: - basic mutex The database (actually all databases) are locked when starting an access like reading or writing a record or searching an index, and unlocked when done. Since there is not very much and especially no IO happening outside the database access, it doesn't make much sense to allow parallel access in the first place and we will rather resort to a one-session environment. - parallel IO This is the interesting case to be discussed now Parallel IO aims at using the time one thread has to wait for an IO operation to complete in order to let another thread use the CPU and possibly start additional IOs. Therefore, the mutex is released during IO. In certain situations like thread A wishing to access a cache page being read by another thread B, A has to wait on a condition which will be signaled by B after returning from the IO. The mutex and condition are implemented by an OpenIsisLockFunc, which may map it to a pthread mutex and associated condition variable. This is very similar to the concept of a monitor as implemented by Java's synchronized blocks. The mutual exclusion could be made even more finegrained by using one mutex per database and another one for global structures. With parallel IO, however, the mutex is locked only during CPU use and released during IO, so this, while adding overhead, would hardly increase concurrency on a single CPU system. On a Windoze box capable of basic mutex only, on the other hand, you would probably not access multiple databases anyway. * Summary by environments The following gives an overview of simple approaches to be used in basic implementations: - PHP/Apache1.x/Unix, any CGI: Multiple processes use mutual exclusion based on file locking. Database must be closed after request. Actually, file locking is performed always on database open/close, without asking whether there might be other processes. - PHP/MT/windoze: Uses trivial dispatcher, requests fully synchronized on default session. - PHP/MT/Apache2.0: May use real dispatcher, once the MT-Apache is stable. - Java: May use non-trivial dispatcher, if it provides LockFunc. - OpenIsis server: Uses active dispatcher / > Server multiplexer * Notes on PHP For various PHP run modes, see > http://www.php.net/manual/en/features.persistent-connections.php As of Feb.03, several extensions are > http://www.php.net/manual/en/faq.obtaining.php listed as being NOT thread-safe! --- $Id: Concurrency.txt,v 1.6 2003/02/18 18:10:20 kripke Exp $ doc/Config.txt0100644000113000011300000000723407654006331012260 0ustar kripkripaccessing configuration data * overview Conceptually, all configuration data is bound to a database. Databases are organized in SCHEMAS, with one schema representing local databases and other schemas corresponding to remote server stubs. While many configuration options may be treated likewise for local and remote schemas, some options, such as file paths, are meaningful only within the local schema and should not be distributed by servers. * schema and database access Schemas are accessed by either names or numbers (one byte), the local schema has empty name and number 0. A remote schema's name may be an URL like "host" or "host:port" (possibly one day even with protocol like "z3950://...") or a symbolic name, typically referring to some local configuration source for host and other parameters. Within each schema, databases are accessed by either simple names like "cds" or numbers (two bytes). A main entry with empty name and number 0 holds options which are global to a schema and/or serve as defaults for databases. In addition to the configuration data, each schema entry may contain data related to the actual implementation, e.g. file id's for a local database or sockets in a remote schema's main entry. * configuration parameters A database's configuration can for the most part be considered one record with nested subrecords. For efficient access, however, several logical subrecords may be available as records on their own and/or in a parsed representation. A configuration entry includes - general parameters record These include traditional standard system and database parameters, encoding names, named texts to be interpreted as "printformats" (PFT) and so on. This record's metadata is > Builtin - a parsed version of the FDT - a parsed version of the FST - optional character tables These are corresponding to the ISISAC.TAB, > http://www.cindoc.csic.es/isis/21-5.htm ISISUC.TAB, SRT-files and recoding tables as of > Syspar 106, 107. - optional named views The sort of "printformat" acting as database views. - stopword list (STW) Several of these are used only for indexing and thus are usually not needed in a remote database description. Some environments may use additional data like work sheets (FMT). * implementation notes - mapping of names to numbers is relative to each process. between processes, only names may be exchanged. - most ressources are accessed using defaulting, i.e. if a parameter is not found in the entry of a given database, it is searched for in the schema's main entry. - in the traditional (WinISIS) implementation of local schema configuration data, named PFTs, FMTs and other are looked up as files, which are not actually relative to a database, but to either the schema's standard directory or a tool directory as of dbname.par 10. While the sharing of such ressources in the standard directory corresponds to placing them in a schema's main entry, other --possibly shared-- tool directories will (some day) be implemented by additional database entries, which may be referred to by other databases. * synchronization All configuration ressources can be modified only by the default session. Other sessions can assume them to be immutable for the lifetime of one request. If the default session wants to change a configuration, the database has to be remounted in single user mode, inaccessible to other sessions. This might involve some waiting for other sessions to finish their current request. To support this, each database entry has a mount status and a count of sessions using it (not counting the default session). For details see the paper on > Concurrency . --- $Id: Config.txt,v 1.3 2003/02/13 16:58:45 kripke Exp $ doc/Contact.txt0100664000113000011300000000020607654006331012440 0ustar kripkrip write or phone $ OpenIsis Verein Stralauer Allee 23a 10245 Berlin +49-30-29048706 $ Germany, RFA, Alemania, Duitsland, Tyskland ... doc/CsTables.txt0100644000113000011300000000663407654006331012556 0ustar kripkrip* ASCII and Latin1 (ISO-8859-1) the ASCII table including C0 conrols (first 32 characters) and the Latin1 table created by Charset -tab -enc ISO-8859-1 (i.e. the identity table, looks like Latin1, if this is the encoding used to see this document). $ ... 0 1 2 3 4 5 6 7 8 9 A B C D E F 0 NUL STX SOT ETX EOT ENQ ACK BEL BS HT LF VT FF CR SO SI 16 DLE DC1 DC2 DC3 DC4 NAK SYN ETB CAN EM SUB ESC FS GS RS US ... 0 1 2 3 4 5 6 7 8 9 A B C D E F 32 ! " # $ % & ' ( ) * + , - . / 48 0 1 2 3 4 5 6 7 8 9 : ; < = > ? 64 @ A B C D E F G H I J K L M N O 80 P Q R S T U V W X Y Z [ \ ] ^ _ 96 ` a b c d e f g h i j k l m n o 112 p q r s t u v w x y z { | } ~ DEL 128 (C1 controls ... 140 ... ommited) 160  ¡ ¢ £ ¤ ¥ ¦ § ¨ © ª « ¬ ­ ® ¯ 176 ° ± ² ³ ´ µ ¶ · ¸ ¹ º » ¼ ½ ¾ ¿ 192 À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï 208 Ð Ñ Ò Ó Ô Õ Ö × Ø Ù Ú Û Ü Ý Þ ß 224 à á â ã ä å æ ç è é ê ë ì í î ï 240 ð ñ ò ó ô õ ö ÷ ø ù ú û ü ý þ ÿ $ Note: there is a similar charset, Cp1252 (Windows Latin-1), which differs only in the range 128-159, where several special characters like the EURO sign and typograhic quotes (like `` ´´ ,,) are added. Note: character 173 is the soft-hyphen (shy) used in text-processing to indicate an optional hyphenation. Depending on how you read this text, it might not be displayed. * Cp437 and Cp850 (DOS LatinUS a.k.a. "OEM Charset" and Latin1) representation of Cp850 in Latin1 created by Charset -tab -enc Cp850. The blanks represent line-drawing and other characters which cannot be represented in Latin1 (java actually prints question marks). Position 240 has the shy, the last char (255) is the nonbreaking space (nbsp, Latin1 160). $ ... 0 1 2 3 4 5 6 7 8 9 A B C D E F 128 Ç ü é â ä à å ç ê ë è ï î ì Ä Å 144 É æ Æ ô ö ò û ù ÿ Ö Ü ø £ Ø × 160 á í ó ú ñ Ñ ª º ¿ ® ¬ ½ ¼ ¡ « » 176 Á Â À © ¢ ¥ 192 ã Ã ¤ 208 ð Ð Ê Ë È Í Î Ï ¦ Ì 224 Ó ß Ô Ò õ Õ µ þ Þ Ú Û Ù ý Ý ¯ ´ 240 ­ ± ¾ ¶ § ÷ ¸ ° ¨ · ¹ ³ ²   $ According to the default ISISAC.TAB, ranges 128-154 and 160-165 are alpha. This fits Cp437, which is the same for these ranges, but contains even more linedrawing in the higher positions. Most confusing, it has some greek letters for technical use in row "224", with the beta just at position 225, where it was replaced by the similar looking german sz ligature. Less surprising, the default ISISUC.TAB is also made for Cp437. the unicode mapping of Cp850 as created by Charset -ctab $ 199, 252, 233, 226, 228, 224, 229, 231, 234, 235, 232, 239, 238, 236, 196, 197, 201, 230, 198, 244, 246, 242, 251, 249, 255, 214, 220, 248, 163, 216, 215, 402, 225, 237, 243, 250, 241, 209, 170, 186, 191, 174, 172, 189, 188, 161, 171, 187, 9617,9618,9619,9474,9508, 193, 194, 192, 169,9571,9553,9559,9565, 162, 165,9488, 9492,9524,9516,9500,9472,9532, 227, 195,9562,9556,9577,9574,9568,9552,9580, 164, 240, 208, 202, 203, 200, 305, 205, 206, 207,9496,9484,9608,9604, 166, 204,9600, 211, 223, 212, 210, 245, 213, 181, 254, 222, 218, 219, 217, 253, 221, 175, 180, 173, 177,8215, 190, 182, 167, 247, 184, 176, 168, 183, 185, 179, 178,9632, 160, $ See also a > encoding list of encodings supported by Java and some > /doc/charsets.html notes on the use of charsets with ISIS . Roman Czyborra compiled an > http://czyborra.com/charsets/codepages.html illustrated overview over the most commonly used codepages. --- $Id: CsTables.txt,v 1.3 2002/12/02 18:20:56 kripke Exp $ doc/Cuba.txt0100644000113000011300000005670207654006331011731 0ustar kripkriproadmap for the OpenIsis development Mapa para el desarrollo de OpenIsis * status - read-only engine - very basic versions of query/formatting - multiple language bindings: PHP,Perl,Java - webapps work fine (neither complex queries nor formatting) This is the status since summer 2002, one year after the development started. While multithreading ability would be useful when running inside a Java servlet engine, webforms do not usually create complex queries, do not refer to previous results and the graphical layout is anyway driven by HTML code inside a JSP rather than by formatting language constructs like "FONTS". * Estado actual - motor de sólo-lectura - consultas y formateo aún básicos - múltiples asociaciones (bindings) en los siguientes lenguajes: PHP, Perl, Java - aplicaciones web funcionan bien (sin consultas complejas ni formateo) Este es el estado desde verano 2002, un año despúes de haber comenzado el desarrollo. Mientras que sería útil incluír la habilidad de manejar multihilos dentro de un motor de servlets Java, formularios en web generalmente no requieren la realización de consultas complejas, no necesitan acceso a resultados anteriores y el diseño gráfico ya está incluído en el código HTML dentro de un JSP y no requiere de instrucciones de formateo como "FONTS". * prepared - Tcl/Tk GUI client - Lehmann/Yao index with collation support - highly efficient multithreading server Since midth of 2002, we've been preparing the next step, which shall lead to a (non web based) graphical user interface supporting simultaneous write access. In order to support even very demanding applications, we sacrificed the simplicity and convenience of direct file access in favor of a server based approach. After several tests and studies on how this might be integrated within the various environments the design phase is mostly finished and much of the code has already been written. The index structure is borrowed from Postgres/the GiST. * en preparación - cliente con interfaz gráfica en Tcl/Tk - Indice Lehmann/Yao con alfabetización flexible - servidor multihilos altamente eficiente Desde mediados de 2002 hemos preparado el próximo paso que deberá conducir a una interfaz gráfica (no basada en web) que posibilite acceeso a escritura simultáneo. A fin de dar soporte incluso a aplicaciones con altas exigencias, sacrificamos la simplicidad y conveniencia de acceso directo a archivos en favor de un enfoque basado en un servidor. Después de muchas pruebas y estudios sobre cómo podría ser integrado en los diferentes entornos, prácticamente hemos finalizado con la fase del diseño y ya hay mucho código escrito. La estructura del índice es prestado de Postgres/the GiST. * aims - multi-purpose database and server technology - robustness, performance and scalability - exploring application areas with tools While OpenIsis clearly has it's roots in CDS/ISIS, we believe the principles of IIF(ISO2709)-based databases to be very well suited for a wide range of applications, not only in the bibliographical domain. The more programmers feel ISIS is their database of choice, the more will contribute improvements. Maybe one day even Koha and Gnuteca will switch over from MySQL/Postgres. Therefore we strictly focus on an industrial-strength implementation of the core issues and on making them available to techies in various ways rather than building elaborated applications offering easy end user access. * Objetivos - base de datos y tecnología de servidor multipropósito - robustez, rendimiento y escalabilidad - exploración de diferentes áreas de aplicación con las herramientas Aunque las raíces de OpenIsis provienen de CDS/ISIS, creemos que los principios de bases de datos basados en el formato de intercambio de información (ISO 2709) se adaptan muy bien a una amplia gama de aplicaciones, no sólo en el ámbito bibliográfico. Cuanto más programadores eligan a ISIS como base de datos, mayores serán los avances contribuidos. Quizas incluso Koha y Gnuteca se pasen de MySQL/Postgres a ISIS. Como consecuencia nos hemos concentrado en una poderosa implementación de las funciones básicas y a ofrecerlos a desarrolladores de diferentes formas, en lugar de construir aplicaciones completas para el usuario final. * variants - standalone - client and multithreaded server - no multi-process writing, single-shot writers possible The most simple variant is working for some time now. In a read-only application, several clients may access the database simultaneously. (This *may* even work with *one* writer, but details are system dependent, so don't rely upon this). No two writing processes may have overlapping access to the database, since this would require fine-grained file locking (which has lots of problems) and rule out agressive caching (which has lots of benefits). However, where the complete access to the database, from opening to closing, is strictly serialized by one single lock, multi-process write access is possible (CGI, PHP). The approach most stable yet giving highly parallelized access is a server acting on behalf of a bunch of clients, as used by every RDBMS. * Variantes - estación de trabajo independiente - cliente y servidor multihilos - sin escritura multi-proceso, funciones de escritura monoestables La variante más sencilla ya funciona desde algún tiempo. Es una aplicación de sólo lectura, varios clientes pueden acceder una base de datos en forma simultánea. (Esto incluso *puede* funcionar con *un* cliente que grabe datos sobre ella, sin embargo los datalles dependen del sistema de implementación, por lo que ésto no deberá ser tomado al pie de la letra). * application scenarios - multi-purpose or customized Tk application - PHP or Perl webapp readonly or client - Java webapp standalone, server or client - multi-protocol C server (Z39.50, HTTP, IMAP) The basic GUI version will try to mimik the basic features of WinISIS. The scripting abilities of Tcl/Tk can be used for customizing applications. For experienced users facing complex tasks, this could provide a much more demanding, yet also more flexible alternative to worksheets and printformats. As with today's RDBMS, all language environments can act as client to a database server. Java, due to it's one-multithreaded-process architecture, may also incorporate the multisession server code locally (via JNI). The server proper is not limited to it's native database access protocol, but may support a host of session based TCP protocols, e.g. Z39.50 using the Yaz toolkit. * Escenarios de aplicación - aplicaciones multipropósito o a medida con Tk - aplicaciones web en PHP o Perl de sólo-lectura o clientes - aplicaciones Java independientes, cliente o servidor - servidor multi-protocolo en C (Z39.50, HTTP, IMAP) La versión con interfaz gráfica reimplementará las características básicas de WinISIS. La facilidad de Tcl/Tk para integrar scripts puede aprovecharse para aplicaciones a medida. Para usuarios experimentados que se enfrentan a tareas complejas, ésto es una alternativa más exigente, pero también más flexible que hojas de trabajo y formatos de impresión (PFT). * server architecture - multithreading - multisession - various embedded languages possible - usually not itself client The server opens one session for every client to hold references to previous query results. All requests on behalf of one client are serialized. As a consequence, IO and memory, which are also bound to sessions, can be accessed safely within the server without additional locking. The basic server language, yet another PFT-lookalike, can be augmented by embedded versions of the several bindings similar to the well known PASCAL formatting exits. While the server could also itself access remote databases, this is not recommended since it may hamper availability. * Arquitectura del servidor - multihilos - multisessión - posibilidad de usar varios lenguajes incrustados - generalmente no es un cliente en sí El servidor abre una sessión para cada cliente para almacenar referencias a los resultados obtenidos por consultas anteriores. Todas las solicitudes de un cliente se serializan, como consecuencia el acceso a los archivos y la memoria también se ligan a una sesión y pueden accederse en forma segura dentro del servidor sin bloquéo adicional. El lenguaje del servidor es solamente otro tipo de PFT y puede aumentarse incrustando versiones de las diferentes asociaciones similar a las conocidas salidas de formato de Isispascal. Mientras que el servidor también podría acceder bases de datos remotas, ésto no se recomiendo ya que puede haber inconvenientes con la avabilidad. * client architecture - TCP or local sockets - asynchronous in event based environment (GUI,server) - synchronous else The client connects to the server by means of TCP or the faster Unix sockets. The basic programmatic interface will, after sending a request to the server, wait until the response is available completely. In some environments like the Tk GUI however, it is possible for the client to stay responsive while waiting, display some progress bar and then show results as they arrive. * Arquitectura de clientes - TCP o sockets locales - entorno asincrónico basado en eventos (GUI, servidor) - otros sincrónicos El cliente se contecta al servidor mediante TCP o los sockets de Unix que son más rápidos. La interfaz de programación básica espera -después de haber enviado una solicitud al servidor- hasta que la respuesta esté disponible en forma completa. Sin embargo algunos entornos como la interfaz gráfica de Tk pueden incluír algún tipo de barra de progreso y mostrar los resultados en la medida que ingresan a fin de mantener el cliente en espera. * relation to technologies - use pros, avoid cons - bring ISIS power to other techniques - but stay independent of single concept like XML There is no single perfect programming environment for all kinds of tasks and all sorts of programmers. In order to support the widest possible range of techniques, we must not tie OpenIsis to any particular one. Using C++, for example, causes problems when linked into another environment which fails to properly call constructors and destructors or was compiled with another version of "io.h". While relying on a C++ based XML-parser is perfectly ok for a given application, it would not be such a good idea for a multi-purpose library. Even C-runtime features must be used with caution. * Relación con tecnologías - aprovechar los pros, evitar los contra - llevar el poder de ISIS a otras tecnologías - pero permanecer independiente de conceptos aislados como XML No hay un entorno de programación perfecto para todo tipo de tareas y todos los programadores. Para dar soporte al rango más amplio de tecnologías no debemos atar a OpenIsis a ninguna en particular. Usando C++ por ejemplo, aparecen problemas cuando se enlazan en otro entorno que falla al llamar correctamente a constructures y destructores o cuando ha sido compilado con otra versión de "io.h". Mientas que un parser XML basado en C++ es perfectamente factible en una aplicación dada, no sería tan bueno para una librería multi-propósito. Incluso características del librería de runtime de C (libc) deben emplearse con cautela. * Java features - works well in servlet engines - full-fledged language with well-designed utilities - mature and portable multithreading support - JVM license problems, buggy GUI toolkits The various improvements over the C++/Smalltalk heritage made Java the language of choice for object oriented programming in large-scale software projects. Servlet engines are the most elaborate environment for complex web-publishing tasks. Unfortunately, the GUI toolkits AWT and Swing are poorly designed and implemented. The Sun JVM and JDK are subject to export restrictions, and the free JVMs are not very mature. * Características de Java - funciona bien en motores servlet - lenguaje completo con funciones bien diseñadas - soporte para multihilos maduro y portable - problemas con la licencia, toolkits GUI deficientes (AWT) Las diferentes mejores sobre la herencia de C++/Smalltalk hacen de Java el lenguaje de elección para programación orientada a objetas en proyectos de software de gran escala. Motores servlets ofrecen el entorno más elaborado para tareas complejas en la publicación web. Desafortunadamente los toolkits de interfaces gráficas AWT y Swing están pobremente diseñados e implementadas. La máquina virtual Java y JDK de Sun están sujetas a restricciones en cuanto a su exportación y otras máquinas virtuales Java no están demasiado maduras. * Java application - multithreaded network application with in-process server - synchronous client in pure Java - integration based on JDBC or "enterprise beans" The database frontend for "Institut der Didaktik der Mathematik" is a simple example how to take advantage of HTTP session handling and javax.mail classes to create an order system. Remote database access can be implemented in pure Java (like Oracle's "thin client"), however, asynchronous communication is not supported by java.net.Socket. The wealth of existing Java based solutions might be basis for integration e.g. with SOAP. * Uso de Java - aplicación multihilos en red con servidor de procesamiento interno - clientes sincrónicos en Java puro - integración basada en JDBC o "enterprise beans" El interfaz de base de datos para el "Instituto de didáctica de la matemática" es un ejemplo simple de cómo se puede aprovechar las ventajas de un handling de sesión HTTP y de clases javax.mail para crear un sistema de solicitud bibliográfica. Se puede implementar el acceso a bases de datos remotas en Java puro (como el "cliente liviano" de Oracle), sin embargo java.net.Socket no ofrece soporte para comunicación asincrónica. La abundancia de soluciones existentes basadas en Java pueden ofrecer una base para una mayor integración, por ejemplo con SOAP. * Tcl/Tk features - best portable GUI toolkit - very tight integration with ISIS - easy scripting for customization - not too widespread use As a GUI toolkit, Tk is unparalleled in portability and ease of use. While it can also be used as PerlTk, PythonTk and even JavaTk, Tcl is Tk's native language and, among the mentioned, the most easy to learn. Since Tcl was conceived as Tool Command Language, features of the C library are easily and efficiently made accessible. While Tcl is twelve years old and very mature, it is not that well known and does not offer such an enormous amount of third-party libraries to use. * Características de Tcl/Tk - el mejor juego de herramientas GUI multiplataforma - integración muy estrecha con ISIS - scripting sencillo para requerimientos particulares - uso no demasiado extendido Como toolkit de interfaz gráfica, Tk no tiene igual en cuanto a portabilidad y facilida de uso. Mientras que también puede ser usado como PerlTk, PythonTk e incluso JavaTk, Tcl es el lenguaje nativo de Tk y entre los mencionados, el de más facil aprendizaje. Debido a que Tcl fué concebido como un lenguaje de control de herramientas, se accede con facilidad y eficiencia a las características de la librería en C. Mientras que Tcl ya tiene doce años y es muy maduro, no es tan conocido y no ofrece una cantidad tan enorme de librerías adicionales. * Tcl/Tk application - graphical user interface - asynchronous singlethreaded client - integrated in server Tcl's most important application is to provide a portable Tk-based GUI, which, due to it's nature as a script, can be tailored to the needs of each library or augmented by additional functionality like Z39.50 (indexdata wrote a Tcl/Tk client based on Yaz). This GUI provides synchronous access to local databases and asynchronous remote access, however, runs itself in a single thread. Tcl is also well suited as a safe server extension language, with one interpreter per session. * Uso de Tcl/Tk - interfaz de usuario gráfica - cliente asincrónico de proceso simple (sólo hilo) - integrado con el servidor La aplicación de Tcl más importante es de proveer una interfaz gráfica portable y basada en Tk que, debido a su naturaleza de script, puede ser hecho a medida de cada biblioteca o aumentada con funcionalidades adicionales como Z39.50 (indexdata escribió un cliente Tcl/Tk basado en Yaz). Esta interfaz gráfica provee acceso sincrónico a bases de datos locales y acceso remoto asincrónico, sin embargo corre en sí mismo como un sólo hilo. * Perl features - pathologically eclectic rubbish lister - best for the lazy, impatient and hubristic - unreadable one-way code The experienced programmers preferred quick-and-dirty language. There is rarely a situation in life for which not somebody wrote some Perl Module. Usually, however, you can use it but not read it. * Características de Perl - "pathologically eclectic rubbish lister" - lo mejor para vagos, impacientes y presuntuosos - código unidireccional ilegible El lenguaje rápido-y-sucio preferido por programadores experimentales. Se encuentra raramente una ocasión en la vida donde alguién NO haya escrito un módulo Perl. Sin embargo generalmente lo puede usar, pero no leerlo. Nota1: (Listador de tonterías patológicamente eclécticas). * Perl application - converting - reports - integration - cgi You wouldn't start writing a complex application in Perl, would you? One year later it's nearly impossible to go any further. However, behind and between the scenes Perl is as useful as for small CGI based web frontends. * Uso de Perl - conversión - informes - integración - cgi Ustedes no comenzarían escribir una aplicacion compleja en Perl, ¿no es cierto? Un año más tarde es casi imposible de ir más allá. Sin embargo, atrás y entre escenas Perl es tan práctico para pequeños cgi-frontends para web. * PHP features - easy for beginners - commonly available on servers - with Apache, runs multiprocess * Características de PHP - fácil para principiantes - generalmente disponible en los servidores - multiproceso con Apache * PHP application - lightweight web apps - integration with existing PHP code - for writing, should be used as client * Uso de PHP - Aplicaciones web livianas - integración con código PHP pre-existente - para escritura, debe ser utilizado como cliente * Mail and MIME - mailheaders easily stored as ISIS records - IMAP queries similar to Type-1/ISIS - well established concept for handling multimedia MIME is not exactly a programming language, but an environment worthwhile to support. While media attached to an ISIS DB is locally best stored in separate files (just like RDBMS store their "BLOBS"), it's type and attributes should be described according to the MIME schema. Even where bibliographical rules for such descriptions apply, MIME is best suited to give some client app hints on how to display or further process such data. In the long term one could envision basing the cyrus IMAP server on ISIS to hold the metadata. * Mail y MIME - Encabezamientos mail se almacenan fácilmente como registros ISIS - consultas IMAP tienen similitudes con consultas tipo-1/ISIS - concepto establecido para manejar multimedia MIME no es exactamente un lenguaje de programación, pero un entorno que merece soporte. Mientras que es preferible almacenar archivos multimedia atachados localmente a una base de datos ISIS en forma separada (justo como las bases relacionales almacenan sus "BLOBS"), el tipo y atributos de estos archivos deberían describirse acorde al esquema MIME. Incluso cuando se aplican reglas bibliográficas en estas descripciones, MIME ofrece la mejor forma de informar a las aplicaciones cliente como visualizar o procesar estos datos. A largo plazo se podría enfocar de basar el servidor IMAP cyrus en ISIS para almacenar los metadatos. * XML - structures easily flattened to ISIS records - compact storage, fast access - can search for all attributes During the last years virtually every software "moved" to XML, actually putting fancy new cloths on their same old data. Since bibliographers were among the early supporters of XML's anchestor SGML it is not surprising that these cloths can easily be stripped off in order to store structures whose XML representation is available into an ISIS DB. Each opening tag produces a field with subfields holding it's attributes. Fields representing the childs are followed by a field denoting the closing tag. * XML - estructura que se adapta fácilmente a registros ISIS - almacenamiento compacto, acceso rápido - permite consultas por todos los atributos Durante los ultimos años practicamente todo "se movió" hacia XML colocando bonitas ropas nuevas sobre los mismos datos viejos. Debido a que entre los primeros en impulsar SGML -el ancestro de XML- estuvieron bibliógraficos, no es de sorprender que estas ropas pueden quitarse facilmente para almacenar datos en bases de datos ISIS, mientras que las estructuras XML siguen accesibles. Cada etiqueta de apertura genera un campo con subcampos que almacenan sus atributos. Los nodos hijos se representan en campos a los que le sigue un campo que significa las etiquetas de cierre. * supported platforms - linux fully supported - other unixes incl. MacOS X - windows w/o multithreading As of today, linux/intel is the ideal platform. On other unixes, we may have to deal with big endian processors and support for POSIX threads may vary. Windows 9x/Me lacks sufficient multithreading support. NT4/00/XP has the necessary calls, but difficult to use. * Soporte para plataformas - soporte completo para linux - otros unixes incluyendo MacOS X - windows sin multi-procesamiento Actualmente linux/intel representa la plataforma ideal. Sobre otros unixes tenemos que manejar procesadores big endian y el soporte para subprocesos POSIX puede variar. Windows 9x/Me carece de suficiente soporte para multihilos. NT4/00/XP incluya las llamadas necesarias, pero son dificiles de usar. * proposed file formats - plaintext MST - XRF 4/2, 5/3, 7/5 or as seems fit Besides the legacy MST and XRF, OpenIsis will support additional formats. A plaintext MST is very similar to ISO2709, representing a record as a series of lines. There are neither limits on field length, record length or database size nor any alignment or byte order incompatibilities. In most cases it will even occupy slightly less space. The record length should be stored together with the position in the XRF instead of the MST, so access requires less IO and the number of bytes used for position/length can be allocated as needed. * Formatos de archivo propuestos - archivo maestro plano (MST) - archivo de referencias cruzadas (XRF) 5/3, 7/5 o como parezca ajustarse Mas allá de los formatos heredados MST y XRF, OpenIsis ofrecerá soporte a formatos adicionales. Un MST en un archivo plano es muy similar a ISO 2709, representando un registro como una serie de lineas. Tampoco hay limitacion de longitud de campo, longitud de registro o tamaño de la base de datos, ni alineamiento o incompatibilidades de orden de bytes. En muchos casos, incluso ocuparía un espacio levemente menor. * application projects - common serials database - commercial applications (web shop and content management) - several web based * Proyectos de aplicaciones - base de datos de publicaciones periódicas - aplicaciones comerciales (comercios en-línea y gestión de contenidos) - varios basados sobre aplicaciones web * roadmap 2003 - 02: finish index - winter: standalone GUI (HTML formatting) - spring: server with GUI client - Java and Perl overhaul - reworked formatting - full Type-1 queries * Objetivos para 2003 - febrero: terminar índice - marzo: aplicación GUI para estaciones de trabajo independiente (sobre formato HTML) - primavera: servidor con cliente GUI - revisión de las asociaciones con Java y Perl - retrabajar formateo - soporte completo para consultas tipo 1 --- $Id: Cuba.txt,v 1.6 2003/02/10 17:15:39 kripke Exp $ doc/CubaPHP.txt0100644000113000011300000001212107654006331012264 0ustar kripkripPHP Extension for OpenIsis (a personal experience with the OpenIsis project) Extensión PHP para OpenIsis (una experiencia personal con el proyecto OpenIsis) * history of the PHP extension - Problem: creation of a library module for PostNuke - CGI has a high overhead - PHP easier to maintain than WWWISIS - OpenIsis the better solution: C-lib We needed to develop a library module for PostNuke using ISIS databases. CGI has a high overhead associated with creating the script process and then executing the image of the CGI script (not for high availability). Also PHP is easier to learn than WWWISIS and its formatting language, therefore the code is easier to maintain for new employees (better software engineering). After evaluating several solutions the most attractive was to create a PHP extension capable of using the OpenIsis library. * historia de la extensión PHP - Problema: crear un módulo para bibliotecas en PostNuke - CGI tiene muchos costos - PHP es más fácil de mantener que WWWISIS - OpenIsis la mejor solución: C-lib Necesitábamos desarrollar un módulo para bibliotecas en PostNuke utilizando bases de datos ISIS. CGI tiene un costo muy alto asociado con la creación del proceso y luego la ejecución de su imágen (no es para alta disponibilidad). Además PHP es más fácil de aprender que WWWISIS y su lenguaje de formateo, por lo tanto el código es más fácil de mantener para empleados nuevos (mejor ingeniería de software). Luego de evaluar varias soluciones la más atractiva fue la de crear una extensión de PHP capaz de utilizar la biblioteca de funciones OpenIsis. * achievements - API for reading ISIS databases using PHP - PHP extension for Unix and Win32 - PECL including of the extension After one month of development we had a first functional prototype of the extension with all the basic functions. Since, the development has been improved, has been corrected of bugs, and most important runs on different platforms using PHP. Also there has been an effort to include the extension into PECL. * logros - API para leer bases de datos ISIS utilizando PHP - Extensión PHP para Unix y Win32 - Inclusión en PECL de la extensión Luego de un mes de desarrollo tuvimos un primer prototipo funcional de la extensión con todas las funciones básicas. Desde entonces, el desarrollo ha sido mejorado, corregido de errores, y aún más importante corre en diferentes plataformas utilizando PHP. También ha habido un esfuerzo para incluir la extensión en PECL. * future of the PHP extension - Extension for PHP5 - New OO API for the extension, with procedimental backward compatibility - Add writing of ISIS databases The new version of PHP will feature: New object oriented model, Integration with external object oriented models, such as COM and Java, Exception handling and Stack tracing. Therefore upgrading the extension to this model will be a priority, since PHP will be suitable for larger, high-traffic, mission-critical web applications. Also the writing of ISIS databases will be included. * futuro de la extensión PHP - Extensión para PHP5 - Nueva API OO para la extensión, con compatibilidad procedimental - Agregar la escritura de bases de datos ISIS La nueva versión de PHP tendrá características como: un nuevo modelo orientado a objetos, integración con modelos externos orientados a objetos como COM y Java, manejo de excepciones y rastreo de la pila. Por lo tanto actualizar la extensión a este modelo será una prioridad, dado que PHP será adecuado para aplicaciones web de misión crítica, de gran tamaño y de alto tráfico. Además se incluirá la escritura de bases de datos ISIS. * BIG lesson learned (my own experience) - Open Source development can lead to human development (IMHO) - Open Source can be one of the different tools used for human development. Since Software Development is not an extensive activity but an intensive one, it can be very productive with less resources. There is a very strong cicle between learning and development when we participate in Open Source projects. The feature "Open" can attract people from around the world to our development. If this is true also for large scale projects we must use this approach for our Latinamerican countries (we need human development). * GRAN lección aprendida (mi propia experiencia) - El desarrollo de Código Abierto puede llevar al desarrollo humano - http://www.bsolano.com/images/opensource.gif El Código Abierto puede ser una de las diferentes herramientas utilizadas para el desarrollo humano. Dado el Desarrollo de Software no es una actividad extensiva, sino más bien una actividad intensiva, puede ser muy productiva con menos recursos. Hay un fuerte ciclo entre el aprendizaje y el desarrollo de software cuando se participa en un proyecto de Código Abierto. La característica "Abierto" puede atraer muchas personas de alrededor del mundo hacia nuestros desarrollos. Si esto es cierto también para proyectos a gran escala debemos utilizar este enfoque en nuestros países latinoamericanos (necesitamos desarrollo humano). --- $Id: CubaPHP.txt,v 1.2 2003/02/10 19:51:43 kripke Exp $ doc/DbFiles.txt0100664000113000011300000000213307664370720012365 0ustar kripkriporder of opening database files * checking options - try to read an options file .opt or .OPT if found, merge options with those passed, and use the successful case for extensions - else, if the last path component contains an uppercase ascii and no lowercase ascii, use uppercase extensions - Options may select to open only index or only data. - For both index and data, either traditional or new files are used during normal operation, with new files preferred. However, both sets may be open during copying. - if no read/write mode is explicitly selected, readonly is used if a TXT or MST file exists and is readonly. (By index file permissions in index-only mode). * checking data - if TXT exists, use it, done - if traditional MST/XRF exist, and they are writable or writing was not explicitly selected, use them as is, done - create TXT and copy any available readonly data If for a TXT no valid PTR exists, create one. * checking index - if OXI exits, use it, done - if traditional index files exist, open readonly - else, OR if writing is wanted create OXI for writing, copy old index doc/FactSheet.txt0100664000113000011300000000412607654006331012720 0ustar kripkripthe OpenIsis database at a glance * characteristics - highly flexible data structure: potentially unlimited number of data fields ("columns") in record - highly efficient storage: unused data fields consume no space - natural modelling - ultra fast access: logically related data that would be artificially separated in a relational DB is stored in a single record - highly flexible index structure: index entries associated with a record are under full application control, can even be derived from associated text documents of any format. * standards - Z39.2 (IIF) based data structure record structure corresponds to ANSI Z39.2 (ISO 2709, "Information Interchange Format") - Z39.50 based retrieval the query language resembles Z39.50 "Type-1" queries as used by all major libraries (known from WAIS and most search engines). - all bibliographic data models like MAB and the various MARC standards can be represented in IIF and thus in OpenIsis - XML data is canonically mapped to IIF for highly efficient storage and retrieval * interfaces - variety of programming languages Java, Perl, Tcl, PHP, C-library for all compiler languages - web applications using Java, Perl or PHP as wanted - graphical user interface for Windows and Unix (incl. Mac OS X) based on Tk - usage in-process or as server - Z39.50 protocol server planned using indexdata's Yaz * performance benchmarks on an 800MHz Pentium III with standard IDE disk - 100.000 records/sec read in random order - 50.000 entries/sec index lookup - 15.000 records/sec write access with no index entries - 3.000 records/sec write access with ten index entries per record * limits There is a choice between traditional and 21st century database formats. - size of data field and record traditional: 32KB, 21st: no limit - number of records and database size traditional: 16 million records of 500 MB total size, 21st: no limit. logical database can span several physical databases (virtually no limit). - index size traditional: ~1 TB, 21st: 32 TB per physical index. logical index can span several physical indexes (virtually no limit). doc/GetIt.txt0100644000113000011300000000363307665654232012101 0ustar kripkripAvailable for download: * Sources The > /tar/openisis.0.8.9.tgz Current version 0.8.9 includes tons of new stuff, most of which is working pretty stable. See > History for version notes. Also available are the older versions > /tar/openisis.0.8.7.5.tgz 0.8.7.5 , > /tar/openisis.0.8.7.tgz 0.8.7 (linux only), > /tar/openisis.0.8.6.tgz 0.8.6 stable and > /tar/openisis.0.8.5.2.tgz 0.8.5.2 > /tar/PHP-OpenIsis.1.0.tar.gz PHP-OpenIsis.1.0 has sources for the PHP binding. All are in tar.gz format and include some documentation. See > HowTo installation notes Alternative download from > http://sourceforge.net/project/showfiles.php?group_id=11257 sourceforge There also is a more or less 0.8.6-based version > /tar/VisualC_adapted.zip adapted for M$ VC by > mailto:dubyk@lsl.lviv.ua Serhij Dubyk * Databases > /tar/cds.tgz CDS (~53KB, 160 rows, needed for tests), > /tar/test.tgz test (~40KB, a writable, Latin-1 encoded copy of CDS, used by new Tcl/Tk demos), > /tar/unesb.tgz unesb (about 15 MB, 58.000 rows) * Binaries > http://sourceforge.net/project/showfiles.php?group_id=11257&release_id=136977 PHP-OpenISIS.zip contains windows binaries of Braulio's release 1.0 of the PHP-binding, including openisis.dll 0.8.6 > /tar/builds/win32/openisis-0860-win.zip 0.8.6/win32 (mingw32 cross-build) includes the openisis.dll and Java stuff. should work on all win32 platforms from win95b on. Note: If you want to use java you will need the JDK1.3.x The current windows version is not yet thread-safe, so using it with a Java servlet engine under windows requires some care. > /tar/Perl_Win32_library.zip A perl DLL for Win32 (based on 0.8.6, I think) by > mailto:dubyk@lsl.lviv.ua Serhij Dubyk > /tar/builds/solaris/solaris-20020212.tar.gz 0.8.4/solaris ( build on solaris 5.8 ). Please > mailto:erik@openisis.org mail us if you need a more up-to-date solaris binary. Note: the JDK shipped with solaris 5.7 / 5.8 ( java 1.2 ) should work. doc/History.txt0100664000113000011300000002006207665654210012517 0ustar kripkrip* 0.8.9 May 2003 New > Serialized plaintext database passes all test. It's typically more than 10% smaller than the traditional, has the same format on all platforms and no limits on field or record size. It can be fixed with a text editor; the pointer file is rebuild automatically. By default, it uses 4/3/1 pointers (4GB DB size, 16MB records). See > DbFiles for how and when it's used. Create a text db using $ sto/openisis -db db/cds/cds >db/cds/cds.txt $ New plain PHP package features flexible "v" operator, direct access to the plaintext database (using 4/4/0 pointers) and isis database server access. The OpenIsis Tcl server oits can be used like $ bin/oits -db 1234 test/testsrv.tcl $ a plain server is to follow soon ... * 0.8.8 April 2003 Sort of a preview to stable version 0.9. The C-API seems to be mostly stable now (but don't use the RecStream), and there are not much known bugs (deleting index entries is broken; fixed in current CVS version). Yet, of course, features are going to be added ... The Tcl-API now gives a good idea of what it's going to be, but still will undergo some changes during the following month. After that has settled down, Java, Perl and PHP will receive their upgrades. Although there may be some changes until version 0.9, it's recommended to start playing around with the Tk demo GUI, tisis, and it's formatting TSPs (Tcl Server Pages), that also can be used as CGIs (see demo.cgi). Use > /tar/test.tgz this little test database (a copy of CDS recoded to Latin-1, with OpenIsis index and a procedure to select index entries). Also try dexml, which uses the TclXML package to convert *ANY* valid XML input to ISIS records w/o loosing information! * 0.8.7.5 April 2003 Last version of the 0.8.7.x developer series. It contains all the developments of winter 03. Version 0.8.7.5 is the last to contain all that devel stuff. Version 0.8.8 will have explicit passing of Session removed, i.e. change the 0.8.7.x API, but still be compatible for the 0.8.6 calls. * Perl improvements February 2003 Try 'make pdemo' to see encoding support (Cp850 to Latin1, for other see perl/encoding.pl), a sample standalone perl install tree, support for 'indicators' (initial unnamed subfields) and a non-nonsense way of spitting out XML. * 0.8.7.4 January 2003 Experimental FDT support and preliminary (old-style), but pretty usable Tcl bindung by Matthias included (see tcl/test for how it works, please contact us before using it for serious work). Of course several fixes... * 0.8.7.2/3 December 2002 No longer uses stdio but direct threadsafe buffered io (*snprintf still used). Every session has their own buffered in, out and err streams, which may be accessed w/o interlocking. * 0.8.7.1 November 2002 The IO module is rewritten to use native windows calls for WIN32. This also fixes some problems with read-only databases. Error checking needs to be added; however, under wine, it passes the standard test. adjust make.mingw to fit your MINGW installation, then use $ make WIN32=sigh test $ * October 19th 2002 The OpenIsis society ("Verein") has been founded with 15 members. Chairman is Erik Grziwotz, other board members are Gabi Rohmann, Ingo Struck and Thomas Sonnemann. * 0.8.7 October 2002 Version 0.8.7 supports writing of DOS/WinIsis masterfiles and xrf. This currently works fine for a single process on Linux. See [17]writing.txt for details. Current TODOs: interlock multiprocess writing (PHP, Perl CGI) and fix the windows and solaris versions. Besides writing support, there is a new streaming record reader, which groks various formats like the SYSPAR.PAR, email headers and property files, so you can fill your db from such textfiles. Next step: new indexing engine. * August 2002 No new software yet. Still very busy doing metawork. We are preparing to set up an organisation to support OpenIsis development with much more momentum and a company for professional services like help on large scale ISIS installations. paperwork on [18]the universal ISIS record * July 2002 some paperwork on [19]What is it about ISIS that makes it ISIS? * 0.8.6 June 2002 This version supports basic formatting. While most, especially graphical, features of WinISIS or CISIS formatting are not yet supported (which are typically not used in a web environment anyway), there is support for repeated subfields as declared by MARC for many fields. See [20]formatting notes for details. The perl binding supports formatting (see the test.pl), enhanced versions for all those languages are to follow soon. * PHP June 2002 [21]Braulio José Solano Rojas from Costa Rica created a PHP binding, which can be seen in action at [22]galileo. Available as [23]download or from sourceforge module php-openisis. Also, the [24]Institut Teknologi Bandung of Indonesia switched it's web index to a PHP/OpenIsis based version. * 0.8.5.2 March 2002 Much enhanced PERL binding. See the OpenIsis.pm included in the sources. * 0.8.5.1 March 2002 Java now has support for basic formatting modes like MHL, various HTML-safety modes (like escaping all non-ASCIIs), a [25]Vn-field-selector-style method and several nice utils. Indentation is not properly handled, since there is no easy common solution in HTML. Will build tools for a choice of standard strategies later ... * 0.8.5 March 2002 Finally some implementation of the query language ([26]demo). All the operators are there (including /(tag), but not /(t1,t2...)). Every attempt is made to limit the potential costs of even extremely stupid queries like "$"^"$", so no historical (#n) or intermediate results (for precedence) are stored. Queries are processed strictly left to right on a buffer of 1000 hits. * 0.8.4 January 2002 Nearly complete rewrite of search code with support for NEAR conditions. Fixed alignment problems in IFP, now works with the unesb db (-format aligned) and the cds db as distributed with winisis. The cds db we had with previous versions (from an old CDS distribution) has a mixed format: aligned leaf files, others unaligned. Although support for a mixed format is easily achieved with openisis, it will not be included unless somebody has a need for it (send us mail). The JSP demo now supports [27]searching the unesb db with over 58.000 rows (note the hosting server is a 500MHz Celeron). Searching is limited to 1000 postings, usually resulting in a somewhat lower number of rows (where rows have multiple matching postings). The lowest row number (MFN) that was cut off is recorded, and it is possible (not yet in the JSP) to repeat the search starting from that rowid. * 0.8.3 October 2001 First truly usable release, since we have true search by index now (as prefix or complete word). Search gives a list (array) of sorted MFNs; arithmetic on those lists (and, or, not) is straightforward. The JSP [28]demo shows how a query is refined (narrowed) iteratively by ANDing it with a second query. And thanks to Veronica Lencinas and colleagues, we have a version of this document [29]en espanol. * September 2001 Not a new release yet, but maintenance and testing. New structure logging mechanism. Sources are available via CVS at [30]sourceforge. Windows version openisis.exe running. * 0.8.2 August 2001 openisis now under [31]LGPL, no legacy code. Conversion from file structures governed by abstract dynamic description rather than C-structs, so we can support different file layouts, much larger databases, big endian processors and more. Simple full-scan search available in C-Lib and Java. Given a random read throughput of about 30.000 Records/sec on a lame 300 MHz Notebook this seems to be of practic use. jsp demo under [32]development ;). * 0.8.1 June 2001 Java native interface version available. Java package org.openisis has Db, Rec, Field, Test. NativeDb implemented in libopenjsis.so. Subfield splitting and htmlifying in pure java. * 0.8.0 May 2001 Subfield splitting and htmlifying. Everything also available from perl as an xsub. Record shows up as hash, handy but no repeated fields. Makefile.PL, test.pl etc. * 0.7.9 May 2001 First version: static C-Library libopenisis.a for reading records by rowid(Mfn). Executable "openisis" does test. Logging, argumentparsing, Makefile, demo etc. ... ToDo doc/HowTo.txt0100664000113000011300000000543607654006331012117 0ustar kripkrip Start by downloading the Software. Unpack everything in some arbitrary directory. For the tests, you will also need some isis database, which must be located as files db/cds/cds.*, see downloads. Make sure filenames are lowercase. If you are on Windows, you should either get yourself the cygwin environment with tools like gmake and gcc or volunteer as a porter and start writing the Makefile for your make and compiler. Erik has build a Windows version using mingw and Linux gcc as crosscompiler. If you are on Linux, everything is fine. Ports to MAC OS X and other UNIXes should be no problem. Type "make" and enjoy the compiler messages. (If your make complains, e.g. on BSD, try "gmake"). Type "make demo" and enjoy your first open isis record. (You installed a db/cds/cds.*, didn't you? It has more than 42 rows?) Type "make run" and watch the guts of your db passing by. Type "make test", there should be no difference between the testout.txt and the testres.txt as provided (using the cds database from winisis and the 15 MB 58.000+ row unesb db, see downloads). Type "make time" to measure performance, subsequent tries are usually much faster. My 800 MHz P3 random-reads more than 179.000 records a second, once the files are in the system cache. Typical values: $ time ./openisis -perf 1000000 -db db/cds/cds >/dev/null real 0m5.655s user 0m3.650s sys 0m2.000s time ./openisis -perf 100000 -db db/unesb/unesb >/dev/null real 0m0.991s user 0m0.670s sys 0m0.320s time ./openisis -fmt mfnf -search 'k$' -db db/unesb/unesb >/dev/null 860 rows for k real 0m0.044s user 0m0.040s sys 0m0.000s $ Type "make perl" to build the perl stuff; some perl 5.* must be installed beforehand. Type "make java" or, if you just can't get enough, "make jdump", to see it all happen in your shiny new JDK1.3 Java VM. Some 1.2.* JDKs should do, but tell the Makefile to not look in /usr/java/jdk1.3 by setting JAVAHOME. * installing openisis libopenisis.a can be linked with your code; no installation necessary. You may wan't to install the 'openisis' binary somewhere in your path for the fun of it; go ahead, just copy, no magic registry entries. To install the perl stuff for general availability in your /usr/lib/perl5 or whatever, cd to the perl subdir (after "make perl") and issue "make install" (as root or otherwise legitimated). After that, try "perldoc OpenIsis" and the demo.pl script. Java, like perl, needs to dynamically slurp both some stuff in the own language and a native shared object. The former is openisis.jar, set your CLASSPATH to include it, or specify when invoking java like in the Makefile. The latter is libopenjsis.so on linux (yes, it's jsis). The system dynamic linker must be able to find it; see > /jdoc/org/openisis/NativeDb.html NativeDb.java for details. ... GetIt doc/Inverting.txt0100644000113000011300000001175607656456430013037 0ustar kripkripThis document discusses various issues of indexing (creating entries for the inverted file). * the inverted file The inverted file maps keys to (sorted) sequences of values. Although an application can store arbitrary information in values, in standard use, they denote a hit (where the key was found in a record). A hit is a structure of five numbers, specifying (in that order) a database number, a rowid (mfn), tag, occurence (of this tag in it's record) and position (of this "word" in it's field). The inverted file is implemented as a B-Link-Tree (Lehmann/Yao) with variable length keys of up to 255 bytes. The length of values is fixed for a given index, but can be configured on index creation in the range 8 to 23 bytes (8+n, with 4 bits for n). The 8 byte minimum configuration (corresponding to the traditional IFP) uses - 2 bytes pos - 1 byte occ - 2 bytes tag - 3 bytes mfn A ninth byte is used for the mfn, bytes 10 and 11 for a database number, additional bytes have no standard usage in hits (by now). Occurence and position are typically used for "near" conditions in queries. Where this is not needed, they could also be used, for example, to apply weights to the hits. * creation of index entries Traditional ISIS systems based indexing on - a "print format" to create text from a record - an "indexing technique" to extract entries. The various indexing techniques split the text at newlines, subfields, word boundaries, angle brackets or other delimiters. OpenIsis uses a different approach based on an "index record", which is created either explicitly by application code or using > Views a view. Most split up like extracting keywords from between angle brackets or other delimiters is done during creation of this record, containing, in the simplest case, one field per index entry. The only additional transformation that may be applied is word split to be done during > Collating , because the definition of which characters constitute words is configured together with the collation info. In the index record, fields with a non-negative tag are treated as index entries according to the current mode. All fields with negative tags give processing instructions as follows (denoted with symbolic tag names): - index f[ields] [startocc] an xmode field with value 'f' (or some prefix of 'fields') sets default indexing mode, where each value (of a non-negative field) is used as literal index entry. The occurence is set to startocc or 0, incremented for runs of fields with the same tag, and reset to 0 on tag change. - index w[ords] [startpos] sets word mode, assuming fields to contain pre-split words. Instead of the occ, the pos is set, incremented and reset. - index s[plit] sets word autosplit mode. Index entries are split into words according to collation info; each word increments the pos, each field resets the pos and increments or resets the occ as in fields mode. If the index has no collation info, all characters but the well-known ASCII non-letters are assumed to be word characters. - index a[dd]|d[el] sets add (default) or del mode for index entries. - index m[fn] mfn sets the mfn to use - index n where n is a number, sets max key length to n (default is 30) - hit [+|-][[dbn.]mfn.]tag.[occ].[pos]|tag[.pos] key declares explicit hit data for key. With a leading + or -, explicitly add or del the key regardless of mode. Unspecified fields are determined as usual from the environment. Since search results can be sorted by increasing pos, this field can also be used to specify a weight (starting from 0 for a 100% match). - fst viewspec declares a view line to create fields from the current data record, see > Views The index fields can also be used with a tag and value in their field value, in which case the given mode applies to that tag and value only. (TODO) While view-based indexing is yet to be done (as of May 03), application controlled indexing is working and robust (see the save method in openIsis.tcl and sample.fsp). * special values The entries created according to the index record are then subject to a > Collating collation transformation to create index keys. Finally they are handed to the B-Tree for insertion. If the key is already there, it may have an associated special value, which does not denote a hit, but rather processing instructions. Special values are recognized by a mfn of 0. - a completely blank value of all zeroes marks a stopword. no other values for this key are added to the index. (This is implemented). - a value of mfn 0, first byte of tag 1, marks a ref. The rest of the value and some successive values contain an alternate key, under which the entry is made (and searched). Since this has some difficulties, it will probably take some time until this is implemented... A multilingual (English, French, Italian, Spanish and Portuguese) > http://www.bib.wau.nl/isis/docum.txt stopword file can be loaded like $ tr '[a-z]' '[A-Z]' http://www.niso.org/standards/resources/Z39-2.pdf Z39.2 ) * subfields There is a convention to encode multiple fields in one by separating them with a '^' followed by one character tagging the subfield. So the field value '^afoo^bbar^bbaz' represents a field having one 'a' subfield with value 'foo' and two 'b' subfields 'bar' and 'baz'. An other separator char may be used, e.g. ASCII character 31 ("Unit Separator") is used in the > http://www.loc.gov/marc/specifications/specrecstruc.html MARC standard. * formatting There is a formatting language, with literal text, field and subfield variables, if-else branches (on field existance) and for loops (over field repetitions) (roughly speaking). * indexing An index is build by converting a row into a list of words (optionally applying formats) and stuffing every word, qualified by the position of it's occurence in the row, into a B+-Tree (which is actually spread to six files). Searching for a word or word prefix is possible with or without qualifying the position (field). Since all fields can be combined into one index, it is usually not necessary (but possible) to set up multiple indexes. * queries A query language allows for combination of word lookups using and, or and not(without) operators. This is very similar to the "Type-1" query of > ftp://ftp.loc.gov/pub/z3950/official/part1.txt Z39.50. * usage While isis lacks most features of RDBMS like complex relations between different entities, it's flexibility comes in handy for many catalogues and directories with highly varying records and one single level of substructure, which today are usually modelled in XML documents rather than table rows. In other words, isis is an ideal storage for many XML applications. The flexible indexing mechanism combines the best of full text searching and structured retrieval. doc/IsisLinks.txt0100664000113000011300000000474207654006331012766 0ustar kripkrip> http://sourceforge.net/projects/isis/ openisis and PHP isis at sourceforge * isis core sites: > http://www.unesco.org/webworld/isis/isis.htm Unesco > http://www.unesco.org/webworld/isis/ new Unesco site > http://www.bireme.br/isis/ Bireme * documentation: > http://www.cindoc.csic.es/isis/indice.htm THE BOOK CDS/ISIS reference manual incl. data formats (en espanol), > http://www.vista.gov.vn/VistaEnglish/VistaWeb/learn/Isis/Eng/ WinISIS manual * standards: - ISO2709 "Information Interchange Format", a.k.a. ANSI/NISO > http://www.niso.org/standards/resources/Z39-2.pdf Z39.2 - Machine Readable Catalogues > http://www.loc.gov/marc/specifications/specrecstruc.html (US) MARC 21 , > http://lcweb.loc.gov/marc/ overview , references at the > http://www.oasis-open.org/cover/marc.html Cover Pages - Z39.50 > ftp://ftp.loc.gov/pub/z3950/official/ official spec , overview at > http://www.oclcpica.org/content/45/pdf/z3950_handbook_paper.pdf OCLC|Pica , links at > http://indexdata.dk/z3950/ indexdata , makers of excellent free Z39.50 software. * people and projects: > http://www.jezuici.krakow.pl/soft/iapi/ Robert Janusz' iAPI , Kafkas Caprazli's > http://members.aol.com/cdsisis/ EVERYTHING about CDS/ISIS > http://www.oss4lib.org/ open source software for libraries > http://web.tiscali.it/javaisis/ javaisis > http://isisonline.lib.itb.ac.id/ Institut Teknologi Bandung IsisOnline in Indonesia * user groups: > http://www.bib.wau.nl/isis/ Netherlands / international > http://www.axp.mdx.ac.uk/~alan2/isisplus.htm UK (ISIS PLUS) > http://www.isisnetz.de/ Germany (isisnetz) * non-ISIS If you do not care for ISIS compatibility but need a fully integrated library system, then > http://koha.org/ koha might be for you. Unlike OpenIsis, which provides flexible tools and components to be used in a wide variety of applications, koha is strictly focused on everything a library needs, including circulation and the like. However, it has a plain table structure based on MySQL, which can not support sophisticated queries as ISIS can. There are > http://www.zlb.de/bibliothek/fachbereiche/infdienste/fremdspr/englisch/ lots of mailinglists on library issues in many languages. * yet other ways to find a book Not all books are living in libraries! You may try to > http://bookcrossing.com/ catch one that has been set free. * and don't forget ... ... to visit the > http://moffem.de/ Museum of Faulty Error Messages :) doc/Java.txt0100644000113000011300000000437107654006331011733 0ustar kripkripreview of the OpenIsis Java interface * Java, Server and Sessions Approaching version 0.9 of OpenIsis, giving the concepts needed for a multisession server, we also have to revise the options for the Java binding. When speaking about Java, we mostly have Java Servlet Engines in mind. Therefore we assume that we are running multithreaded, and that we have little control about the selection of threads. A client's request already has a thread dedicated to it, and this thread cannot do anything else before this request is served, i.e. requests can't be somehow queued or postponed. Therefore asynchronous access to a database is mostly pointless. All we can do is to synchronize access by blocking one thread while another accesses the same ressource. To not reduce concurrency to complete mutual exclusion, we clearly need database sessions. Moreover we have to assume that we have multiple users, which might be represented by Http sessions. Where database sessions are configured to keep results sets, they should be bound to Http sessions. Otherwise we might pick an unused database session from a pool. * Java on a local db The situation of a local database accessed from a servlet engine is similar to that of a dedicated database server in that we have multiple sessions and multiple threads. The multiplexer, ensuring that there are no two requests running in parallel on a given session, is implemented in Java by synchronizing on the Java object representing each session. The creation of a session is synchronized on the default session. A servlet engine accessing a local db may not only create cute looking HTML pages (by means of JSPs), but may also provide a general purpose database server, speaking a HTTP wrapped variant of the server protocol. * Java as client of a remote db On the other hand, a database session might be a client session bound to some OpenIsis server. Since the protocol used for communication with a server is pretty simple, and we will use synchronous access anyway, we are planning to implement this in pure Java. This will remove the need for JNI, which has some problems especially where servlet engines deploy multiple class loaders for what they think are "web apps". --- $Id: Java.txt,v 1.1 2003/01/07 19:28:25 kripke Exp $ doc/Meta.txt0100644000113000011300000001101107654006331011725 0ustar kripkripdescription of OpenIsis metadata * the FDT According to the CDS/ISIS manual, the field definition table is displayed like $ 44 Serie 300 X R vz $ and contains for each field tag: - a field description (up to 30 characters) - maximum length (1 to 1650) - field type: X for alphanum, A for strictly alpha (not including space), N for numeric (decimal digits), P for pattern (see below). - repeatability indicator R meaning field is repeatable, N else - format (subfield list or pattern) For a P field, this gives a COBOL-PIC-style pattern consisting of X (alnum), A (alpha), 9 (num) and literal letters, e.g. 99-999/AA to allow for input like 35-674/XE. For other field types, this is a list of legal subfields, e.g. vz to allow for input like foo^vbar^zbaz. Patterns are not supported by winisis as of Version 1.4. The actual file format looks like $ Series vz 44 300 0 1 $ which is 30 characters name, 20 characters subfield/pattern, the field tag, maximum length, a number encoding the type (0 alphanum, 1 alpha, 2 numeric, 3 pattern), and a number 1 for repeatable / 0 for non repeatable fields. Moreover, the FDT defines the available worksheets W (.fmt), printformats F (.pft) and field selections S (.fst). Each such definition is on a line starting with the key letter and a colon, and followed by (blank padded) 6 character fields of file basenames. $ F:THES THES1 THES2 $ * field names It is common practice to use long field descriptions like "Corporate Bodies" or "Govt. Publications No.", which may be nice looking but are not well suited for technical use. Therefore the OpenIsis metadata contains an extra field name in it's metadata record (see below). When sourcing metadata from an FDT file, OpenIsis will derive field names by lowercasing the descriptions and replacing all runs of non-alphanums by a single '_', yielding "govt_publications_no". * ISO A special kind of alphanum, using the ISO2709/Z39.2 subfield delimiter ASCII 31 (CTRL-_) has symbolic name I and code 10. * enum A special kind of numeric, where the possible values are named, has symbolic name E and code 12. * table Table fields use a fixed set of subfields separated by TAB without the subfield letters. It has symbolic name T and code 14. * structures In order to describe > Struct substructures (i.e. embedded records), we introduce a new field type with symbolic name S and code 15. Since fields introducing structures may contain subfields (attributes), the other type properties are interpreted just like for 'X'. * subfield names and descriptions As an extension to be able to give names to subfields, which is especially required when converting from/to SGML-style long attributes, we use a second set of fields by adding 16 to the type. The tag is the tag of the field for which this description is relevant, or 0, if it could apply to several fields. For the subfields, the first letter of the subfield spec is the subfield described. The rest may be a pattern for type 19 (= 16+3). Subfield specs may have length and repeatability. * enum values Since subfields may be anything, but not structures, we can use code 31 for the values attached to an enum with symbolic name V. The tag and (first char of) subfield are the tag (and subfield) for which this description is relevant, or 0, if it could apply to several fields. The numeric value is held in the length. * compatibility When exporting the metadata to FDT format, S fields are written as X, E as N, and F and V entries are omitted. * representation as record The metadata record uses TAB as subfield delimiter. One field of metadata is used to describe each field. It has the following subfields: - i id (tag) - t type (as digit) - n name - s subfields - l length - r repeatability (not present if not repeatable) - d description - c child legal child id (repeatable) There are other tags reserved for (repeatable) fields to hold the names of available worksheets, printformats and field selections. * example A simple description for HTML's td tag could look like $ td abchnrvw 99 4000 15 1 align a 99 1 30 0 center a 99 0 31 0 left a 99 1 31 0 .. bgcolor b 99 1 1 0 .. valign v 99 1 30 0 .. $ Note the list of > Builtin builtin metadata --- $Id: Meta.txt,v 1.4 2003/01/17 21:40:47 kripke Exp $ doc/MikiDoc.txt0100644000113000011300000000171607654006331012371 0ustar kripkripMikidoc is a mini pseudo- > http://c2.com/cgi/wiki WikiWiki -style documentation convention. Unlike real Wikis, it is much simpler and less oriented towards the web and web-editing of content, more towards the content. Mikidoc currently support the following constructs: Empty (really empty) lines give a <br>, lines starting with - $ toggle code mode - * are a head - --- give a <hr> - - start a list topic - TAB continue a list topic - > are links; the link body may follow the ref after a TAB - ... denote a footnote document displayed to the right - ! set break-after-every-line mode (pseudo pre) - ; are ignored Todo: support *bold*, _italic_ and MikiLinks. Write converter to texinfo (left as an exercise to the reader). That's all, folks! > http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/isis/openisis/doc.jsp?rev=HEAD&content-type=text/vnd.viewcvs-markup get the code ... (4 KB :) --- $Id: MikiDoc.txt,v 1.2 2002/12/12 14:44:23 kripke Exp $ doc/News.txt0100644000113000011300000000253007665654452012000 0ustar kripkrip> History more ... * 0.8.9 May 2003 Nearly stable version with > History plaintext masterfile, preliminary server and new PHP support * May 2003 > http://wiki.openisis.org/i/view/Php/PhpDemos fancy new PHP stuff * May 2003 Principles of a > PatchWork patchworked database . Paper on > Collating terms in the Unicode environment, plus interesting > UniStat stats on the Unicode Character Database. Proposals for > Inverting and > Views * 0.8.8 April 2003 Semi-stable version with a GUI, XML-support and > History lots of other new features. * 0.8.7.5 April 2003 Last version of the 0.8.7.x developer series. * April 2003 detailled description of the > Serialized record format * PoC March 2003 nearly all ideas now have passed their > Status#devel Proof-of-Concept * idxall March 2003 now supports writing with application-controlled index update at > WritePerf reasonable performance * Ukrainian translation of the 0.8.6 > /index-uk.html index page by > mailto:dubyk@lsl.lviv.ua Serhij Dubyk , who also did some > GetIt Windows stuff. * Perl improvements Feb 2003 The reworked perl demo provides encoding support and no-nonsense-XML. * new index improvements Feb 2003 The B-L-Tree now supports update/delete at about 30.000 ops/second. * X. Reunion Feb 2003 > XReunion X. regional meeting of ISIS distributors in la Havana. > History more ... doc/OpenIsis.txt0100664000113000011300000000634107654006331012604 0ustar kripkripBy now, OpenIsis is mainly used for web publishing of Isis databases. Web developers use the Java, Perl or PHP binding for this. There also is a command line version which serves as a demonstration and test of the various features and can be used as a utility for data import/export and similar tasks. A graphical, standalone user interface similar to UNESCO's WinIsis will be developed during 2003, after the Tcl/Tk binding is finished. See > Status current status * developing OpenIsis In general, there are no plans to reimplement every piece of code ever written for isis. To be of practical value, OpenIsis has to maintain compatibility in the format of the database files anyway. So, one may use winisis or whatever existing import scripts to create and maintain the database, yet deploy OpenIsis' perl interface to run powerful reports and the Java Native Interface to allow queries from a Servlet based web application. OpenIsis will focus on providing tools rather than applications. For example, there will be no attempt to mirror the full functionality of winisis. To achieve this, OpenIsis provides access from the most important programming languages: Java and PHP for the web, Perl for the scripts and Tcl/Tk for platform independent GUIs (partly DONE). All others can, of course, link the lib. * Roadmap for the development of OpenIsis -- as of September 2002 After more than a year of OpenIsis development, experimenting with various options and environments and some feedback, we learned something about what is important to the community and what is not. * findings and goals - A fairly portable C library that can be plugged into Java, Perl, PHP or Tcl proved to be valuable and estimated by several developers. There was not much demand for a pure Java version. - There is high demand for a GUI version similar to WinISIS (cross-plattform, of course) with a safe multi-user capability. - This in turn requires a simple but fast and robust server. Multithreading can greatly improve the throughput. - Users are used to incompatible file formats, formatting and so on. There is little of a standard to follow. So building a next generation system clearly is more important than sticking to the (which?) legacy too much. - The power buried in the similarity between ISIS records and other structures like, for example, E-Mails is not leveraged. - The existing versions of the formatting language are the result of a process of growth, adding feature after feature. The logic is mixed with the presentation (e.g. HTML literals). * conclusions - The multiformat support will be dropped in favour of an efficient support of the "DOS/WinISIS" (packed little endian) structures. - The database kernel has to support basic thread-safety for use in Java and other server environments. - Tcl will be the scripting language of choice: Tcl/Tk for portable GUIs and a limited set of Tcl commands for server side scripting. * flavours - The core library will not contain any particular external technology, so it is as lightweight as possible and can be linked into Java, Perl, PHP, Tcl and the like. - There are binaries in plain, with Tcl and with Tcl/Tk. Any of these can act standalone, as client, as server, or both. volunteers are welcome ! ... ToDo doc/OpenSource.txt0100664000113000011300000000572307654006331013140 0ustar kripkripSo why are we writing the openisis software? Because Isis is not open source software, it's not even free software, and that leads to a whole bunch of problems. * problems resulting from closed software - Availability (in theory) Versions of the software exist for some operating systems, library versions and languages. For other environments, there is no version of the software, and there is not much one can do about it. - Availability (in practice) You may download most the software, but it's partly protected with passwords, which you have to order at some national distributor. You have to pay some fee and/or declare some good reasons, why you want to use the software. Then you have to wait. In germany, for example, it didn't work at all for some time, until the newly founded > http://www.isisnetz.de/ IsisNetz remedied the situation. - Availability (in legal terms) Some parts of the software are accompanied by different documents stating some license terms, others are not. Terms seem to be pretty different between countries. One can not easily figure out, what exactly might be allowed usage. - Availability (of documentation) Some documentation is available in english, some only in portugese, espanol or italiano. Only a small part is downloadable at all, most is paperware. - Bugfixing There is no way one can fix a bug, and not much one can do about having somebody fix it. - Extending The only way one could write a Binding for perl or Java would be using the isis.dll. There are problems with regard to required additional libraries (especially some C++ stuff), there are no statements about thread safety, unicode compatibility and so on. As a consequence, it's practically impossible to write a state-of-the-art web application based on an isis db. - Improving Many users develop useful ideas for improvements from practice. Their expertise is lost as they are not able to turn it into improved software. - Enabling While open source software enables people all over the world to shape their tools themselves, closed software lets them in dependency. * benefits of open software To address these problems we feel the need for an open source implementation of isis. Of course it would be best to have all of the existing isis code under one or the other form of open license (GPL, LGPL, artistic or similar as appropriate). On the other hand, an independent secondary implementation has advantages in it's own right. It may have a different focus and develop strengths in one aspect while another aproach performs better in other situations. For example, openisis will have some support for multithreading and unicode, which is paid for by a certain overhead. A rewrite by developers with a different background might introduce new ideas which finally, after having had their indepent test bed, help improve the standard. OpenIsis as a software to access isis databases is and will be freely available for everybody with full sourcecode, no fee, no restrictions. doc/OverView.txt0100644000113000011300000000471107654006331012616 0ustar kripkripThe OpenIsis society was founded in October 2002 to further the development and dissemination of open source information systems. We support libraries, archives and other public organizations like social movements to manage and retrieve information in a high quality way according to bibliographical principles. Most important is the > OpenIsis database system which was developed by our members since May 2001. OpenIsis is the open source member of the > IsisIntro CDS/ISIS > CdsIsis software family. * who needs quality information systems? Well ... everybody who has quality information. Traditional examples are libraries, where librarians invest much knowledge and experience to make you find the book you need. While the large libraries have a long track of using and developing bibliographic databases, many organizations like worker's unions, charity organizations and social movements have smaller archives of their publications, which ought to be as well accessible. Even for the purely virtual web-archives, newsgroups and mailing lists, it is increasingly understood, that the power of search engines needs to be augmented by quality metadata to be retrieved reliably. Dublin Core and the Resource Description Framework are examples of such approaches based on bibliographical principles. * who needs open source information systems? Many individuals, organizations and institutions can't afford the commercial bibliographic database systems, where even small site licenses start at about 10.000 US$. Therefore, > http://www.unesco.org/webworld/isis/ UNESCO's CDS/ISIS , which is available for a nominal fee, is in widespread use, where the budgets are low. Our > http://www.gnu.org/philosophy/free-sw.html free software is not only free of charge, but also free of restrictions. It enables the users to build on it, to improve it, to adopt it to their needs rather than being dependent on the original developers. To close the "digital gap", access to *both* the content and the software is needed. * what about the "Information Society"? Even the techies and librarians at rather well-equiped universities in the US or Europe, which are working on new ways to publish and manage free content, like the > http://www.openarchives.org/ Open Archives Initiative or the > http://oc4s.org/ Open Community for Science , do not have access to the code of existing commercial systems. Following the Linux Economy, we may find that open source is the basis for progress ... ... News doc/PHP.txt0100644000113000011300000000603107654006331011474 0ustar kripkripconsiderations on the design of a PHP interface This paper is based on the requirements of > Concurrency * why not to export sessions From the C-API's point of view, the use of sessions is a MUST in a multithreaded environment, which PHP may be. This, however, is the server-side of the PHP engine. The PHP page, as seen by the programmer, is a client to the database: it does NOT handle multiple requests, but one! In other words, the page is already executing within the context of a given request, including a given thread. Since a request is not meant to operate on several sessions, but rather has one session as an attribute of it's context, nothing would be gained by making the session visible ("export") to the PHP programmer by calls like read(ses,...) or ses->read(...). Moreover, the currently most used environments, i.e. MP Apache 1.x/Unix and some Windoze boxes, do not allow/require the use of multiple sessions, anyway. Finally, the only reason to explicitly select a session based on a user's session id, would be to access this session's state as #1-style result sets or authentication information. Both will not be implemented in the near future and both will rarely be needed in web environments. Yet, it might be made available when needed by means of an additional call that "selects" that session as the one to be associated with the request context. * strategy Therefore, I would suggest the following strategy: - In the first step, stick to the default session by using a static session variable like openisis_ses0 and auto-initializing it like the OPENISIS_SES0 macro does. - For the MP version register a function using register_shutdown_function (is there a corresponding C-API to register a C function ?) that closes all open DBs in order to release their flocks. (Berlin will provide such a "close all"). - On Windoze and Apache 2, make sure to actually run single-threaded by using EnterCriticalSection/pthread_mutex_lock on a static CRITICAL_SECTION/PTHREAD_MUTEX_INITIALIZER and leaving/unlocking it in a registered shutdown_function - Some day, in a better age, replace the static session variable with a thread specific one by using pthread_key_create/pthread_setspecific or maybe PHP's TSRM. C.f. man pthread_mutex_lock, > http://msdn.microsoft.com/library/en-us/dllproc/base/critical_section_objects.asp critical Windoze * notes - is there a call in the PHP API that can be relied upon to be the first one used, in order to do initialization? Probably opening a DB? - In the MT version, closing databases is not enforced, thus successive requests will open already opened databases. This is ok with the C-API's DOpen, which will happily return the same handle. - Once additional worker sessions/threads are supported, we'll need additional open/close calls, which do neither actually open or close a database, but access and release (and possibly flush). The MT and MP usage could then be made to look very similarm with details governed by a few config settings. --- $Id: PHP.txt,v 1.1 2003/02/13 20:36:23 kripke Exp $ doc/PatchWork.txt0100644000113000011300000001616607660457271012773 0ustar kripkripprinciples of a patchworked database: how OO-style inheritance works for data * references It is quite common for records in a database to refer in some way to other records. In fact, it seems to be such an interesting property, that the main classification of databases is along the lines of how they handle references: - in a hierarchical DB, every record has 1 or 0 parents, and any number of childs - in a networked DB, every record has a variable number of references (typically fix for a given type of record) - in a relational DB, every record has any number of references, which are determined by queries. References are constructed on the fly; if they are backed by some index and the query optimizer is able to figure that out, this works even for larger DBs. While basically every model can be emulated on top of every other, it is a matter of performance and convenience (i.e. performance counted by hours of programming) how well a given model of data access supports a given pattern of data usage (that's why you should stop thinking that RDBMS solve all problems). The magic working behind references is always the same: either a (more or less) physical key of the referred record or some logical key is stored in the referring record, the latter requiring translation by an index (typically B-Tree). In general, the key might as well be a RDF URI, resolved by a http or other server. The "physical" key is known as "master file number" (MFN) in ISIS databases, as "ROWID" or similar in most RDBMS. Using a ROWID is a gross violation of the principles of relational databases (Codd's axioms), and most RDBMS cannot retain a ROWID reference via export/import (i.e. DBs using such references can neither be reliably exported nor backed up, short to a DB shutdown and file backup, as stated by a major vendor). Hierarchical and networked DBs, on the other hand, have much stronger support for the fast physical key, but typically still allow for logical keys to reach via an index on any record. They just don't sport an elaborate and "standardized" query language like SQL giving convenient access to what does and doesn't work. It's up to the application programmer, rather than the query optimizer designer, to follow the paths of well supported references. ISIS databases are clearly in the latter (older) branch of the family of databases, believing more in the application programmer's knowledge than in artificial intelligence. Besides the conceptually simple and ultra-fast MFN reference, it also has a very flexible indexing mechanism, allowing for references to even be based on fulltext. We want to present a view on data, that has some interesting properties and (while it can be emulated on any DBMS) is particularily well supported on ISIS DBs: * the patch relation We propose a relation between records that transcedes a mere reference (independent of whether the underlying reference is based on a physical or logical key): - there is a patch operator that constructs a new record from referring and referred-to record - there is a diff operator that constructs a new record from referring and referred-to record, in such a way that the patch operator, applied to the latter and the result, will construct the referring record Basically, a very small record can tell a story like - I am related to that record - I am very similar to that record, but - I have these and those fields added/changed/deleted Everybody who is accustomed to tools like RCS or CVS (based on RCS) will get the idea of such a relation: depending on what is most commonly needed, either one or the other version can be stored much more compactly as just the diff to the other one. Or, if both are referred to oftenly, you can store them both at full life size, but put only the differences in the index. If the idea of applying diff and patch to database records seems strange, have a look at > Serialized a plain text representation for ISIS records. This paper does include a means to store patch "scripts" even at database master file level, thereby providing very efficient support for cross-referring patch records. However, the patch relation can as well very efficiently be implemented on top of a standard ISIS database. Some meta data entries (per field in the FDT) might be used to set convenient defaults for interpretation of references and patches: - to which database an entry in field x referres to - index tag and/or prefix to be used on lookup - whether field occurences in the patch record are additive or overriding In a way, every reference can be regarded as defining a patch relation: If you resolve the reference in a table join, you are actually creating a new virtual record, comprised of some fields of the referred and referring records. However, in the much more flexible ISIS data model, the creation of a new record is much more flexibel. * inheritance by patching One way to use patching is to think of a patch relation as being an inheritance relation: not by class, but by object! A record can inherit data fields from one or several other records, which still exists as independent entities. When the inheritance is resolved, some data from the parents is copied to the new virtual record, some is dropped or overwritten. An inheritance relation does not only apply to logically subordinate child records, it can also be used for versioning, with a successor inheriting from it's predecessor, while both are still available as independent entities. * example: the serial killer Management of serials is known as a notoriuosly hard problem. So let's have a look at how we could make a data model using record inheritance. We start out simple, with - one record for the serial, - one for each issue, inheriting from the serial - one for each copy, inheriting from the issue The issue adds a number, date, table of contents and so on to the serial. The copy adds some id to the issue, and maybe a note that it's in a poor state, since somebody poured coffee over it. But then things can get a little bit more complicated: - if the serial changes it's name or other attributes, a successor may inherit from the original one, and following issues inherit from the successor (while earlier issues still belong to the original) - if two serials are joined, a successor may inherit from both - sometimes two otherwise independent serials are printed together, so the actual issues inherit from both - an issue might have a reprint, which inherits from the original issue, changing the date and adding a new preface - if somebody ripped an article out of a given copy, that copy might adjust it's issues table of contents - if copies of several issues are bound together, you actually have only one copy, inheriting from several issues To summarize, a lot of complicated situations can be modelled in a quite natural fashion when thinking in terms of inheritance through patch relations, and these, in turn can be easily and efficiently be implemented based on ISIS record and database structures. If you think of modelling that in a relational database, you will probably find that you end up with > RdbConv mimicking the ISIS record. --- $Id: PatchWork.txt,v 1.1 2003/05/14 15:20:25 kripke Exp $ doc/RdbConv.txt0100644000113000011300000001327507654006331012412 0ustar kripkripRepresenting ISIS data in a relational database If you consider setting up a table with columns "author", "title" and so on, don't do it. The following structure gives you at least some of the benefits of ISIS, because it's shaped more closely to the ISIS data model. * table structure It is not very difficult to think of a table structure for lossless representation of the contents of an ISIS masterfile: $ CREATE TABLE Mst ( mfn number, -- master file number fld number, -- counting fields in this masterfile record tag number, -- the field tag val varchar(2048) -- field's content ) $ Notes: - a row in this table does NOT hold an ISIS record, but one field of an ISIS record - each ISIS record is represented by all table rows with the same mfn - you may omit the column "fld" if you don't care for the ordering - you may choose other datatypes for the field tag, for example number(3) or char(3), depending on the compatibility required (e.g. MARC) - you may choose other dimensioning for the val column. ISIS can hold up to 32KB (signed short) in one field, Oracle has a maximum of 4KB on the varchar type. * accessing the data Retrieving row 123 is easily done with $ SELECT tag, val WHERE mfn = 123 ORDER BY fld $ It is suggested that you fill the val column with native ISIS data as of "MPL" proof mode. Then you have to extract subfields and perform other substitutions for output. This can be achieved for example using the v-method of the > /jdoc/org/openisis/Field.html Field Java class (it's a static method of about 100 lines, which you may grab and port to your system). An alternative would be to pre-split the subfields into a structure like $ CREATE TABLE Mst ( mfn number, fld number, tag number, sub char(1), -- subfield code val varchar(2048) -- subfield's content ) $ A typical keywords field like $ $ should be split into three subfields using any subfield code, say '>'. The advantage of the pre-split approach is that it makes both display and indexing easier. The disadvantage is, that you'll loose some information, albeit not too much. * searching Searching is also relatively easy, as long as you ask simple questions ;). To get a list of the mfns with a field 24 containing Twain, use $ SELECT DISTINCT(mfn) FROM Mst WHERE val = 'Twain' AND tag = 24 $ For this to work efficiently, you have to put an index on val. You may than search for 'Twain' with or without the restriction on a certain tag, just like in ISIS. You can also use right truncation and restrict on a couple of tags: $ SELECT DISTINCT(mfn) FROM Mst WHERE val LIKE 'Twain%' AND tag IN (201,401) $ Right truncation should still use the index on val, thus be reasonable efficient (depending on how smart your RDB is). With the pre-split structure as above, this applies to any subfield. If you want to find Twain anywhere within the field's values, you need $ SELECT DISTINCT(mfn) FROM Mst WHERE val LIKE '%Twain%' $ This will perform a pretty costly full table scan, so you may go fetch yourself a coffee meanwhile. While you may use SQLs OR predicate to combine several terms as alternatives, the ISIS AND does NOT map to SQLs AND, but to an INTERSECTion on the mfn lists. It's actually pretty much the same as ISIS does internally (but more costly). * building an index Still, this is lightyears away from the possibilities of an ISIS db. The index on val does not do word splits, does not use stopwords, is not case insensitive, does not apply character or word conversion and so on. To come anywhere near the features of ISIS, you have to do just what ISIS does: set up a separate "Inverted File". $ CREATE TABLE Index ( val varchar(30), -- index entry mfn number, -- record containing entry tag number, -- tag of field containing entry occ number, -- field is occ'th occurence of field tag in record pos number -- entry is pos'th word in field ) $ To fill this table, you could of course go and reimplement all the features of an ISIS .FST in, say, Oracle's PL/SQL and recreate one record's entries from within an INSERT/UPDATE trigger. Much easier, given you had a true ISIS database in the first place, is to load the .LK1 and .LK2 text files created during full index update. Some ISIS tools also allow you to create these files separately. Searching is performed against the Index table (which of course needs an RDB index at least on val), not the Mst table. The latter is only used to retrieve the full content for a given mfn, or for full table scans. * so why then use ISIS in the first place? There is a long list of reasons, here go a few of them: - You don't want to reimplement the .FST, the full formatting language, all the nifty details of the various indexing techniques and so on. Actually, if you DO want, you're very welcome in the OpenIsis team ;) - Not only the index building, but also data entry is done much more comfortable using ISIS tools. - Any queries but the most simple examples given above will not only be somewhat complex to write in SQL, but will also execute much slower than in a native ISIS implementation (consider publishing with OpenIsis). - Especially the NEAR operators can't easily be expressed in SQL. Although the index structure given above contains the necessary data in occ and pos, advanced SQL features like nested queries are needed in the appropriate SQL query. You need to have an advanced (read: expensive) RDBMS and a deep and thorough understanding of it's query optimizer and hinting system for this to run in any reasonable time. - So why then use a RDBMS in the first place? See also > whatabout a comparision of ISIS and other database systems and > unirec notes on the flexibility of the ISIS data model. --- $Id: RdbConv.txt,v 1.4 2002/11/26 16:57:19 kripke Exp $ doc/SeeIt.txt0100664000113000011300000000341507654006331012063 0ustar kripkrip * demo databases the very simplistic > ../Demo?search=development standard Demo searches the small (160 record) cds database or the larger (~ 58.000 records) unesb. Based on the OpenIsis java binding, the whole interface is done with only 200 lines of code. There is also a > http://galileo.or.cr/php_isis/ PHP version available at galileo foundation of Costa Rica, written by > mailto:braulio@openisis.org Braulio José Solano Rojas . a special sample was set up to search for > ../unicode?search=&prefix=1&action=search unicode characters . You can also list the > ../unicode?search=&prefix=1&action=search&terms=1 index . Note that the ability to display fonts like arabic or georgian depends on your browser. See > /doc/charsets.html notes on character sets. * real world examples > http://databases.unesco.org/xtrans/xtra-form.html UNESCO Index Translationum , the International Bibliography of Translations with more than 1,300,000 references. > http://isisonline.lib.itb.ac.id/ IsisOnline databases by Institut Teknologi Bandung of Indonesia. > http://catalog.dml.or.id/credit.php (Credit) The > http://knjiznice.ffzg.hr/search/Search.cgi WebPAC of the Zagreb University's faculty of philosophy. See also Dobrica Pavlinusic's > http://webpac.sourceforge.net/ paper Library of the > ../idm Institut für Didaktik der Mathematik of the University Bielefeld, Germany. > http://www.bv.umsanet.edu.bo/home.htm Biblioteca Virtual Universitaria de Bolivia ("Este sitio ha sido desarrollado principalmente con OpenIsis"). > http://www.funde.org/cendoc.html Fundación Nacional para el Desarollo , El Salvador (still using very early and slow version, but works). > http://www.nied.edu.na/resource%20centre/openisis_lookup.dwt the National Institute for Educational Development of Namibia. doc/Selbstdarstellung.txt0100664000113000011300000001007707654006331014555 0ustar kripkripDer OpenIsis Förderverein für open source Informationssysteme Die digitale Informationsvermittlung und -archivierung sind im Zeitalter der sogenannten Wissensgesellschaft zur Basis des Zugangs zu den Wissensvorräten der Welt geworden. Der OpenIsis Verein will mit der Entwicklung und Förderung von open source Informationssystemen einen Beitrag zu einem gerechten Zugang zu Wissen und Information leisten, besonders durch Unterstützung von Projekten im öffentlichen, sozialen und kulturellen Bereich. Dazu vermittelt der OpenIsis Verein Kenntnisse über den Einsatz von open source Informationssystemen auf der Basis bekannter Studien wie auch eigener Forschungsarbeiten. Mit Publikationen und der Präsenz auf Kongressen und Tagungen betreibt der OpenIsis Verein intensive Öffentlichkeitsarbeit. Er engagiert sich in weltweiter Kooperation mit öffentlichen und gemeinnützigen Institutionen und Organisationen. In Deutschland arbeitet der OpenIsis Verein bisher mit dem Tübinger open source Projekt "Open Community for Science" (OC4) und Isisnetz e.V., dem Deutschland-Distributor der UNESCO für das Bibliotheksprogramm CDS/ISIS. Der OpenIsis Verein unterstützt beispielsweise BibliothekarInnen in Argentinien, Costa Rica, Indonesien und Kroatien in der Entwicklung von Informationssystemen. Das erste große Projekt des Vereins ist die Entwicklung und der Einsatz des gleichnamigen Bibliotheks- und Dokumentationsprogramms OpenIsis. * Das Programm OpenIsis OpenIsis ist eine freie Software für Bibliotheken und andere Einrichtungen und Organisationen, die Wissensbestände archivieren und verfügbar machen wollen. Es knüpft an das seit über 30 Jahren entwickelte System CDS/ISIS an, das auf internationalen bibliothekarischen Standards basiert, und bis heute von der UNESCO weltweit vertrieben wird. Der Quellcode von CDS/ISIS ist nicht zugänglich, eine offene Weiterentwicklung oder Anpassung daher nicht möglich. Seit April 2001 arbeitet eine Gruppe, gegründet von den Berliner Programmierern Klaus Ripke und Erik Grziwotz, an einer freien Version von CDS/ISIS. Der Quellcode von OpenIsis ist offen. Wie bei Linux kann jede und jeder an dieser Software mitarbeiten; AnwenderInnen überall auf der Welt können das Programm an ihre Bedürfnisse bei der Archivierung und Digitalisierung von Bibliotheksbeständen anpassen. Weiterentwicklungen sind schon in Costa Rica, Indonesien und Kroatien entstanden. * OpenIsis - Digitale Spaltung - Wissensgesellschaft CDS/ISIS wird von UNESCO, GTZ und anderen als wichtiger Schritt in Richtung eines chancengleichen Zugangs zu Informationen und zur Vernetzung von Wissensbeständen gefördert. Mit abnehmenden Budgets der Bibliotheken wächst die Bedeutung der Fernleihe und damit auch des Zugriffs auf andere Kataloge über das Internet oder durch Datenaustausch. Für digitale Katalogisierung in Lateinamerika oder Afrika kommt aber nur ganz oder nahezu kostenfreie Software in Betracht. Außerdem müssen lokale Versionen - nicht nur in den Kolonialsprachen - von den AnwenderInnen selbst erstellt werden können. OpenIsis leistet diese Wissensvernetzung und ist gleichzeitig kostenfrei. Die Software trägt damit dazu bei, die digitale Spaltung zwischen Nord und Süd zu überwinden und steht für eine Demokratisierung des Wissens. * Förderung von open source und freiem Wissen Wir freuen uns über jede weitere Unterstützung, im technischen und vor allem im bibliothekarischen, archivarischen oder dokumentarischen Bereich um unser Know-how zu ergänzen. Dem Verein sind bisher Programmierer, WebdesignerInnen, Sozial- und WirtschaftwissenschaftlerInnen und JournalistInnen beigetreten, die aktiv an der Entwicklung, Verbreitung und Anwendung von OpenIsis arbeiten. Mit Ihrem Beitritt können Sie persönlich einen besonderen Beitrag für die Entwicklung von freiem Wissen leisten. Nähere Information über den Verein und das Programm OpenIsis finden Sie unter http://openisis.de/ Gerne beraten wir im deutschsprachigen Raum Einrichtungen im öffentlichen und gemeinnützigen Bereich in Fragen der Verwaltung, Digitalisierung und Archivierung ihrer Dokumentationen sowie im Wissensmanagement. - Stand: November 2002 doc/Serialized.txt0100644000113000011300000002307007665600575013156 0ustar kripkripISIS records serialized Serialization means to convert the internal representation of an ISIS record to a sequence of bytes ("octets") suitable to be stored in a file or transferred via a network. The serialization format described here is used by OpenIsis for both database master files and network communications. * design goals The serialization format should be - easy to use for programmers and tool writers - efficient in execution time and space used - robust a broken masterfile should be fixed using a text editor - versatile can be used for a variety of applications using a variety of tools - without limits in number and size of records and fields * basic format In general, a record is serialized by - serializing meta information - serializing the fields in order - appending a blank line Fields are serialized as - the field tag printed using ASCII decimal digits (optionally preceeded by a minus sign, if negative tags are allowed) - a (horizontal) TAB character (ASCII value 9, ^I) - the field value - a newline character (ASCII value 10, ^J) Metadata is serialized in the same way, using special tags according to the needs of the environment. Two situations are distinguished: - "soft" metadata, which may and should be accessible as part of the record. This is encoded by convention using negative tags. An example of this is HTTP and other MIME-style communication, where the MIME headers like "User-agent" or "Date" are encoded in such a way, while content data like GET or POST parameters should be mapped to positive IDs. - "hard" metadata, which must not interfere with the record contents in order for the environment to work properly. This is encoded using a single non-digit character instead of tag digits. An example of this is the MFN in a master file and information regarding record deletion or update. The final blankline may be omitted, where only a single record is contained in an otherwise delimited byte sequence. A reader should support a lazy mode, allowing the TAB to be omitted, where unambigous (the field value does not start with a digit or a TAB). Writers, however, are strongly urged to write the TAB. Tags with leading zeros are allowed (typically with %03d) and must not be interpreted as octal using atoi. * newline conventions Two ways are supported to deal with newline characters in field values: - in "text mode", newlines are replaced with vertical tabs (ASCII 11, ^K) on serialization and vice versa on deserialization. - in "binary mode", newlines are replaced as newline-TAB sequences. The advantages of text mode over binary mode are - it is slightly faster than the binary translation - the serialized records do not need more space than the internal representation (whereas the binary serialization might need nearly twice as much in worst case) - it is easily used with line-oriented utilities like grep or sed, since each field is contained within one line The binary mode (which resembles MIME continuation lines) has the advantage of not loosing vertical tab characters that might have been contained in the original field values. It is fully transparent and can be used to store any binary data like images with an average overhead of 0.4% (as compared to +33% needed by BASE64 encoding). The OpenIsis server automatically detects binary mode, if the client uses a continuation line. * masterfile format A basic masterfile consists of a blank-line separated series of records. The first record is the "controlling record", containing descriptive information such as the newline convention, the subfield separator and the character encoding. All of this is optional; the masterfile might just start with a blank line. The MFNs are then assigned implicitly in order, starting from 1. There is no distinction between empty (consecutive blanklines) and deleted records. There is absolutely no redundant information contained. Masterfile compression creates this state (however, it may choose to use special meta lines for long ranges of deleted records, see below -- but those are inefficient in the Xref, anyway). Such a masterfile can be very easily created by any tool (like Perl and such). The Xref file can be easily, fast and reliably recreated. When writing to the database, information is ALWAYS appended to the end. There is NO OVERWRITING of any data, ever, period. That way data CANT BE DESTROYED by any operation (one could advise the operating system to set a mandatory read lock on that), and all changes are easily traced using tail -f. A binary mode masterfile starts with a "broken" continuation line containing a single TAB. * basic mode writing In metalines, all numbers are given in decimal digits and multiple items are separated by TABs. The optional timestamps are an arbitrary prefix of generalized time format YYYYMMDDhhmmssttt... as of 'date +%Y%m%d%H%M' plus milliseconds ttt, optionally followed by up to 14 characters to create a unique request id. Write operations use the following meta line (preceeding record data): - W mfn [oldpos [timestamp]] followed by new record data denotes a write of record mfn. Depending on the needs of the environment, the byte offset of the last version might be added in order to support access to old versions (e.g. for delayed index update). oldpos may be given as position[.length[.fields]] A W with no data following is mostly equivalent to a delete. If an otherwise written mfn is higher than the highest known, the highest known (and thus the implicit counter) is set to this. A lazy reader should not require meta lines to be followed by a blank line, where unambigous. For writers, however, the blank line is strongly recommended. A reasonable size limit on metalines is 127(+newline), since - 22 = 1+1+20 for operator, tab and mfn (64 bits ~ 20 digits) - 43 = 1+20+1+10+1+10 for tab position.length.fields - 32 = 1+17+14 for tab, milliseconds time and id totals to 97, so we have some room left. * advance mode For advanced space efficiency at the cost of read increased access time, the following lines may be used: - D mfn [oldpos [timestamp]] a record entry consisting of only a D line denotes the deletion of record number mfn. Basically equivalent to a write with no data following, but a little bit more explicit. - I mfn [timestamp] (set id) override the implicit MFN counter, e.g. after a series of deleted rows or just to be explicit. Basically equivalent to a write with no old pos, but a little bit more explicit. Recommended when appending records after some deletes. - C mfn oldpos [timestamp] introduces a series of patch commands specifying how record mfn was changed. Software writing the masterfile will typically choose to write full updates and MUST provide a switch to forcedly do so, in order to be compatible with basic mode readers. However, the patch command language is particularily useful with server operations, to avoid the need for read-write sequences with some sort of locking. * the patch language The patch commands are lines starting with special characters like +, -, ~ and so on, followed by (an optional TAB and) field addresses, TAB and field data. The simplest case is the '+' command, meaning that it's data is to be appended to the record. The + and TAB may be omitted (both, in order to not be confused with a continuation line). In other words, the add command may look exactly like an ordinary field line. A series of '=' commands works exactly like the set operation in an OpenIsis Tcl record. Especially, field indexes and subfields are supported. The '-' command resembles the del operation. A detailled description of the "patch language" is to be done. example $ C 1234 = 24 foo = 24 bar 25 baz $ changes record 1234 by setting the first to occurences of field 24 to foo and baz, respectively, deleting any other occurences of field 24, and appending a field 25 with value baz. * the pointer file The pointer file is an array of n-Byte (n >= 6) entries, the ith entry referencing mfn i, similar to the traditional .XRF (crossref). The n=k+l+m bytes specify two or three numbers (in native byte order) of up to 8 bytes each: - the first k bytes (k >= 4) give the position of the record (or it's last update or change entry) - the next l bytes (l >= 2) give the length of the record (excluding the last field's terminating newline and following blank line) - the final m bytes (m >= 0) give the number of fields. If m is 0, or all bits in a field number are set (for large records), the reader has to determine the number fo fields by inspecting the record. The first six bytes of the first entry describe the detailled layout. Four bytes are the "magic number" containing the ASCII characters "ISIX". Two bytes are the number (m*256 + l*16 + k) in native byte order. The minimum case k=4, l=2 imposes the limits of traditional ISIS. Actually the lower limits are not enforced; in a very specialised application one might want to use k/l/m = 3/1/0. Recommended are at least eight bytes as k/l/m = 4/3/1 or 4/4/0 or, with large file support, 5/3/0. However, when any of the limits is reached, or an unsupported combination or byte order is found, the Xref can easily be recreated with greater values. The 12 byte pointer with k/l/m = 6/4/2 will be enough for even gigantic databases of a quarter Petabyte (262.144 Gigabyte). The number of fields is redundant, but as an optimization may make live a little bit easier for a reader. If the pointer structure has m>0 (typically m=1), a value of 0 must be stored if the number of fields exceeds the representable range and a reader should be prepared to figure it out itself in that case. --- $Id: Serialized.txt,v 1.7 2003/05/29 18:06:01 kripke Exp $ doc/Server.txt0100644000113000011300000001352607654006331012322 0ustar kripkripoperating an OpenIsis server with one or more threads * server interface From the OpenIsis API, all db IO calls can be redirected to a server. However, only the basic database access is done by the server, convenience functions like formatting are processed locally. There would be no point in processing stream IO or record utilities remotely. While database control calls are generally NOT accessible via the wire, some of them have counterparts for managing the server session. * the multisession server A multisession server deploys multiple threads to process requests from multiple sessions in parallel. Each session occupies at most one thread at a time (and, of course, vice versa), but threads and sessions are not (necessarily) bound to each other during their lifetime. Rather there is a dispatcher monitoring all session's input streams and allocating threads as needed. However, there is one thread bound to the default session, which is thus not subject to dispatching. * principles of communication The server communicates with it's clients by sending messages over streams. While the server's IO multiplexer engine is designed to allow for a variety of applications, the database server API's messages are OpenIsis records. There are field numbers and subfield codes assigned to each parameter of a request. Parameters which are records themselves are encapsulated transparently, i.e. by giving their length in advance. In basic communication mode (i.e. not using a wrapper protocol like HTTP), these records are then transmitted as lines consisting of a field number printed using ASCII decimal digits, followed by a TAB, followed by the field's contents and a newline (NL). Each NL within the contents is converted to NL-TAB and on the receiving side, each line starting with a TAB is treated as continuation of the previous field. Transmission ends with an empty line. * session streams The streams connected to a session are usually backed by some kind of socket. Since requests carry a session id, sessions may be multiplexed onto one or more sockets, possibly using one connection per request or even UDP datagrams. Each response, however, is sent on the socket where the corresponding request was received on and in the order request went in. The default session is always bound 1:1 to a communication channel, not subject to multiplexing (i.e. session id 0 is implied by and allowed only on this channel). Yet this channel may be a named pipe or localhost or unix socket, which does not have a permanent client connection and may be controlled by the multiplexer. Typically, it's client side is a Tcl belonging to the same process, either a tclsh as controller of a multisession server or a wish as the sole client within a standalone application. On Unix, this server thread is most easily connected to by a pair of pipes (using Tcl_CreateFileHandler). Under Windows, things are much more complicated ... In general, the marshalling to a stream might of course be skipped when using inter thread communication. Either side might just prepare the record to the session's request or response buffer, resp., and then ring some bell like Tcl_ThreadQueueEvent and Tcl_ThreadAlert. * the server's heart A server enters multisession mode by starting up the multiplexer. The multiplexer may run in a separate thread or in the same one as the default session, depending on the means by which the default session is connected to it's controlling input. The multiplexer starts up some number of worker threads, listens on a TCP port and then sits in a loop - selecting the server socket and all already open connections - accepting new connections - writing available responses to writable sockets - reading the first few bytes from readable sockets (from which it not already has a partly request) to check for session ids (SID). If a request does not start with a SID, it belongs to (i.e. optionally creates) the session bound to this socket. - deciding from which (readable) sockets more should be read. Input belonging to sessions already in process is left alone. - reading available input, checking for complete requests - if the multiplexer handles default session's input and there is a request available, it is processed in this thread - queing complete requests for processing - checking the response queue for completed requests, collecting responses and freeing threads. Requests producing large amounts of output (similar to piggybacked Z39.50 present requests) may create partly responses before beeing finished. - assigning from the request queue to available worker threads Note that this multiplexer architecture is in no way special to OpenIsis database access, but suited and intended to drive a session oriented, but otherwise multi-purpose application server. In the context of serving an OpenIsis database, the server is essentially stateless. It may at any time decide to drop inactive sessions, unless it was talked into (configured for) promising to keep result sets. * the single threaded server The single threaded version of the server has only one thread and typically only one session, i.e. the default session. Note that single threaded means there is only one thread accessing the database. There may still be other threads in the process, e.g. to care for the user interface. A single session server is not very useful in itself. However, it allows some user interface to be programmed using the server API regardless of whether a local or remote DB is accessed, and to discouple even local processing from the UI. Another variant of a single threaded server is multisession, i.e. running a session multiplexer, but not dispatching jobs to a worker thread pool. This may be useful as minimal server on platforms with no or broken multithreading support, especially Win 9x/Me (which lacks the NT call SignalObjectAndWait). --- $Id: Server.txt,v 1.3 2003/02/14 17:30:33 kripke Exp $ doc/Status.txt0100644000113000011300000000463107654006331012334 0ustar kripkripstatus of OpenIsis development * stable version Version 0.8.6 marks the first version of OpenIsis to contain basic implementations of all major features required for web based publishing of catalogues: - simple queries as typically composed by web search forms - basic formatting (field access and modes) with detailled rendering left to the web frontend code (JSP, PHP ...) All features present in 0.8.6 are and will be supported with the same syntax. In most situations you should use the most recent version, but not rely on the new features. If in doubt, please contact us. *devel developer version: use with caution! Since summer 2002 we are developing the various 0.8.7.x versions. They include - writing capability - new B-L-Tree index - server side sessions - threadsafe buffered IO - preliminary Tcl binding - XML conversion Most new features where planned carefully and prepared by a lot of paperwork. However, as long as the number is 0.8.7.*, all features introduced since 0.8.6 are considered experimental and might still need to be changed. So before you are going to base a lot of new work on these features, it's recommended that you contact us in order to learn detailled up-to-date status. Update: as of March 03, we have the core of a multi-threaded server engine (request dispatcher, session handling) running, a Tcl/Tk GUI demo using TSPs (Tcl Server Pages) and Java code to convert to/from XML. While basically all is as planned, we may remove the Sessions from the API before blessing it 0.8.8. Update: as of April 03, Version 0.8.7.5 is the last to contain all that devel stuff. Version 0.8.8 will have explicit passing of Session removed, i.e. change the API. * plans for 2003 (... not necessarily in exactly that order) - within short, the new index should support immediate update (DONE) based on basic formatting (TODO) - by end of winter, a local implementation of the server interface should be basis for a first Tk based GUI (Demo DONE) - in spring, a single threaded but multi-session server with GUI client should give proof of all those concepts, marked by version 0.9 (Tests DONE) - during summer, the server should actually run multithreaded and Java, Perl and PHP will get a client binding - once the formatting has gotten a complete overhaul, we may consider OpenIsis "grown-up" and put version 1.0 under the christmas tree ... ToDo --- $Id: Status.txt,v 1.5 2003/04/07 16:15:50 kripke Exp $ doc/Struct.txt0100644000113000011300000000764407654006331012344 0ustar kripkripembedded structures in ISIS records * structures In order to support substructures (i.e. embedded records), we introduce a new field type with symbolic name S and code 9. A structure consists of a continous range of fields within a record, started by some field to introduce the structure. There are basically three ways to denote the boundaries of structures: - embraced where a special field is used to denote the structures end. This resembles SGML-style notations, where each opening tag is matched by a closing tag. This is relatively easy and recommended for every day use. - by length where the number of fields (not childs) belonging to the structure is given in the initial field. This allows for safe embedding regardless of the structure's contents and is thus used in contexts where full generality is needed like when embedding result records within a server's response. - implicit where the number of childs is fixed. An example of this is the parse tree of a query, where the structure "AND" has exactly two childs (which in turn might be structures). This is used mostly for internal structures like parsed queries or formats, which are not meant to be exchanged. Where used, Implicit fields and their number of childs are typically recognized by the field tag range -1..-99 reserved for this purpose. The field introducing a structure might have any subfields just like other fields, similar to the attributes that might be assigned to a tag in SGML applications like HTML. The first (unnamed) subfield ("indicator" in MARC speech) however is reserved: - a plus sign '+' as first character indicates explicity opening a structure - a minus sign '-' as first character indicates an empty structure (containing no childs) - an empty value indicates explicity closing a structure - an initial numeric value (of decimal digits) gives the number of fields to follow. - an initial character @A-Z gives the number of childs to follow (@=0,A=1,B=2...) The opening sign or digit may be followed by auxiliary information. An at sign @ followed by a name (alphanum) indicates the structure to be of the type name, typically described in file name.fdt. * conventions While the intented usage of structures might be specified in more detail in the > Meta Field Definition Table , the schema is meant to be as standalone as possible. The extend of structures by length or braces can be safely determined if you just know that you want the given field to be regarded as structure. For structures of fixed number of childs (meant for internal use), it is necessary to recognize whether a following field is itself a structure. In this context, typically one of two modes is used: * the MIME processing mode for processing list-style content, assumes that negative tags denote structures, while positive contain plain data. * in XML processing mode, everything but the 0 tag (text node) is a structure. * examples A typical HTML table definition starting with $ ... $ will be compacted to, say, $ 100 +^w100%^p0^s0^m0^h0^t0^l0^b0 101 + 102 +^vtop^w160 0 this is the textbody 103 - 0 of the td node 102 101 ... $ For a detailed description of the transformation, see > xmlisis the XML-ISIS doku A six field result record might be embedded within a response like $ 4 6@cds 24 Hydrological achievements and social problems ... $ Assuming we gave tag -20 to "OR" (and 0 to a literal), the query "plant OR water" might be parsed to $ -20 B 0 plant 0 water $ "frog AND (plant OR water)" might look like, if -21 is assigned to "AND" $ -21 B 0 frog -20 B 0 plant 0 water $ For implicit tags, the number of childs is redundant (fixed per tag in a given use) and will typically be omitted. --- $Id: Struct.txt,v 1.5 2003/04/07 13:12:42 kripke Exp $ doc/Syspar.txt0100644000113000011300000001140307654006331012325 0ustar kripkripthe system and database parameters of Winisis (collected from various sources) * syspar.par - 0 redirection path to other syspar file - 1 program path pascal files - 2 menus path the MNlltt menu databases and ISISA/UC.TAB - 3 msg path the MSll message databases - 4 work path for temporary files - 5 db path permanent DB files ANY, CNT, IFP, L01, L02, N01, N02, MST, XRF, FDT, FST, FMT, PFT, STW and SRT; c.f. the dbname.par - 6 default db - 7 language (DOS) - 8 repeatable fields sep string to separate reps on entry; default '%' - 9 startup sequence (DOS) e.g. "scds^M" to select CSD database on startup - 10 editor insert mode (DOS) (bool); default 0 - 11,12 line drawing characters (DOS) - 13 extended memory (DOS); values 0,64k,128k,192k - 14 network mode default 0 (single user); 1 is full network, 2 restricted network - 101 language; default en - 102 menu (tt in db name); default DF (SH has short version) - 103 display search progress (bool); default 0 - 104 restrict dictionary (bool) default 0; display only terms occuring in selected field - 105 auto display results (bool) default 0; immediately open result display after search - 106 name of DOS to Windows ("OemToAnsi") character conversion table file in the menu dir with 256 codes, one per line. use supplied file TXT_TXT.TAB for identitiy conversion. - 107 name of Windows to DOS character conversion table - 108 right to left script (bool) default 0; set 1 for arabic, hebrew ... - 109 field edit font format fontname[,size] like Arial Narrow,24 - 110 formatting defaults a PFT expression (not file name) containing any of fonts,cols,cf,f,fs to preselect fonts and colours - 111 format edit font; like 109 for the format editor - 112 printer setup last used printer name/port; auto saved by Winisis - 113 print index separator; default ", " - 115 small font used by dictionary, search expressions - 116 dialog font for message dialog windows - 117 system font for some other texts - 120 display toolbar (bool); default 1 - 121 display toolbar help (bool); default 1 - 122 last main window position; auto saved by Winisis - 123 show empty fields in worksheet (bool); default 0 - 124 restricted shell mode (bool) and pos mode[,left,top,right,bot]; mode 1 restricts to default db w/o toolbar - 125 restricted databases (bool) default 0; restricts access to databases in fields 1001-1005 - 130 auto IF update (bool); default 0 - 131 word length to paginate non-latin scripts - 132 display pascal error (bool); default 0 - 133 windows "large fonts" (bool); default 0 - 134 data entry field width (pixel) - 135 search results statusbar (0..20); default 0 - 136 window zoom (80..160); default 120 - 137 confirm close (bool); default 1 - 138 allow dictionary dump - 140 display log. del. recs (bool); default 1 - 141 background bitmaps dir - 142 highlight fgcolor[,bgcolor] - 144 force tabular display (bool); default 0; see 980 - 148 plugins dir - 150 date format; YMD or DMY; default is M-D-Y - 152 close print dialog (bool); default 0 - 155 deny repeated subfields (bool); default 0 - 156 validate even unmodified records (bool); default 0 - 160 startup plugin name - 161 exit plugin name - 170 username for GetUsername() format; max 9 chars - 902 hypertext debugger (bool); default 0 - 905 status for new/upd records default 0,0 (???) - 918 export separators replace (ISO) export separators 29 (GS/RT for records) and 30 (RS/FT for fields). Value is two printable chars literally. - 939 logfile name - 980 tabular display on multiple results (bool); default 0 - 981 pft name for search tips - 1001-1005 recently used masterfiles filenames may be followed by *** and a descriptive text The DOS version also allows to assign character sequences to function keys using syntax Kn=seq, where n is between 59 and 118. See section 02-2-15 (PARÁMETROS DE DEFINICIÓN DE TECLAS DE FUNCIÓN) of the manual. * dbname.par This file is searched for where the masterfile is expected, i.e. in the db path, and may specify locations for various database files: - 1 XRF - 2 MST - 3 CNT - 4 N01 - 5 N02 - 6 L01 - 7 L02 - 8 IFP - 9 ANY - 10 others: FDT, FST, FMT, PFT, STW and SRT - 0 network switch like syspar.par field 14 * dbname.exp This one is also searched in the db path. It contains settings to finetune data display. - 500 pft name for right results panel - 502 pft expression for left results panel - 504 descending mfns (bool); default 1 - 506 allow pft selection (bool); default 1 - 550 FST tags prefix (repeatable) each value is "tag prefix", e.g. "24 TI=" to hide prefix from terms list - 560 FST tags descriptions (repeatable) each value is "tag text", e.g. "240 Words from title" for FST tags. Language sensitive runs of description may be introduced by "0 ll", where ll is the language code like EN, FR. --- $Id: Syspar.txt,v 1.1 2003/01/10 14:33:15 kripke Exp $ doc/Tcl.txt0100644000113000011300000002305207654477737011617 0ustar kripkripdesign of the Tcl interface * sessions and records The `objects´ in the OpenIsis Tcl interface are tcl commands using the same generic syntax like "array", "info", "interp" or the Tk widgets: $ name option ?arg arg ...? $ Like slave interpreters created by "interp create" or widgets created by "button", "entry" and the like, the name of the command can be choosen freely. There are two basic `objects´ to consider when using OpenIsis: - a record is a command giving access to it's fields. It provides access to single fields, a selection of fields, formatting or looping over the records contents. - a session creates new records from scratch or by fetching them from a database. A session also implements all record commands, giving access to it's configuration. Moreover a session has two embedded records, the request and the result. Finally, new sessions may be created. There is one builtin command introduced by OpenIsis giving access to the default session: the command "openIsis". If you do not intend to work with multiple sessions, you can switch the command openIsis to a remote session. * synchronous vs. event based processing Sessions within this context are always client sessions, and all but the default session (which is local) are connected to a remote server doing the actual work. Unlike the corresponding sessions on the server(s), which are dispatched to multiple threads in order to process multiple client's requests in parallel, the client sessions do essentially nothing but communicate with the server, which is handled pseudo-asynchronously within one thread by means of fileevents, compare > Client client basics. A session may be configured for event based processing by configuring a command to be executed upon request completion. Client sessions are asynchronous in nature and the server socket can (on Unix) easily be bound to a Tcl filehandler. If no command is specified, each request will block until completed. The default session, on the other hand, runs in the same process. Although it is possible to detach processing to a slave thread, integration is more demanding and system dependent and may not be available until some later version. Nonetheless, you may specify a command which will be executed after the completion. * the record options Following is a short outline of the options of the record command. Wherever a field parameter is used, it may be specified as either a decimal number or, if a dictionary is available for this record, as a field name. Field names may also be preceeded by a dash for an option style look. - del ?field ...? All values for all or the specified fields are removed. - get ?field ...? Returns a list of all or the specified field's values. If the one and only argument is -tags, returns a plain list with alternating tag and value elements. If the one and only argument is -tagnames and the record has a fdt, returns a plain list with alternating tagname and value elements. If the -nodefaults option is not given and one argument is a list of two elements, the first element specifies the field name (or id) and the second gives the default value if the field is not present. - add field value ?field value ...? Adds the specified pairs of fields and values. - set field ?value field value ...? With only a single field specified, returns the first value of the single specified field. If values are given, each field is set to the given value, overriding any previous value in the same place. In detail, for every sequence of the same field in the list, the values are assigned to existing occurences of this field in the record. Remaining occurences in the record are deleted. Remaining occurences in the parameter list are added. - do ?options? varname body loops body over fields of the record, assigning each value in turn to varname. body may use break and continue. options include: -fieldid varname to assign each field id to varname, -fieldname varname to assign each field name to varname, -field field to select only occurences of field, -fields fieldlist to select only occurences of fields in fieldlist - clone ?options? newname ?field value ...? clone this record to newname (which must not be a valid option). If newname is "-", some name starting with "openIsis" is choosen. Following field value pairs are applied as with set. options include: -empty to make an empty clone (sharing the same dictionary) - fmt ?options? format apply ("print") format. - done have the record go away. same effect as with Tcl's rename. - serialize set result to serialized record - deserialize string deserialize record from string - wrap -tag tagnumber | tagname ?-number count? ?-done? embed wrap record given by embed command into this record with length field tag. if count is given, append count as tag value and embed record to this record. if -done is given, have the embed record gone away. - db ?options? if no options are given, create a record command for the metadata of the db this record belongs to. otherwise apply options to a temporary metadata record. - fdt ?options? if no options are given, create a record command for the fdt of this record. otherwise apply options to a temporary fdt record. - rowid return mfn of record, useful only for records embedded in a sessions read or query response - .path ?option arg ...? access embedded record specified by path, where each component is of the form field or field[occ]. Without parameters, return 0 or 1 depending on availability of embedded rec. Example: rec .4[2] set -foo bar * the session options Note that a session supports the record options, applied to it's configuration. - db db ?options? if no options are given, create a record command for the metadata of db. otherwise apply options to a temporary metadata record. - fdt db ?options? if no options are given, create a record command for the fdt of db. otherwise apply options to a temporary fdt record. db names -syspar, -scheme, -fd, -fdt, -dbpar, -request, -response may be used to access built in system fdts. - new -schema name ?options? create a new scheme with configuration specified by options. - new ?-db db? ?name? create new record (for db) db names -syspar, -scheme, -fd, -fdt, -dbpar, -request, -response may be used to create a non-db record with built in system fdt. - req ?-db db? ?-param val? Apply the -param val settings to the request record (as with set), then send it to the server. See the > Client common client description for an overview of parameters. preset db for data record of an insert or update request, if specified. - recv Force wait for an response. Does nothing on a synchronous connection or if the response is already there. - .req ?-db db? ?options? access request record - .res ?option ...? ("response/result") With further parameters, apply them to the result record (error, if not response there). Without parameters, return 0 or 1 depending on availability of response. The following fields can be set in the request record: - type - open, close a local database - mount a remote database - ls get a list all available (opened) databases - maxrow get maxrowid of database - query issue a query to database - read read one record from database - insert append new record to database - update update existing record on database - delete (not implemented yet) - db database name different from sessions default db - rowid mfn in read or update requests - config embedded configuration record when opening a local database - flen, fd fdt length and embedded fd records when opening a local database - rec embedded data record in insert or update requests - idx embedded index record in insert or update requests - flags bitwise-or of 1 (retrieve meta data of database with this request) and 2 (fetch records in addition to row ids in query request) - mode mode in query request - skip skip mfn in query request - size limit result set to size in query request - key key value in query request Dependent on the request type, the response may contain these fields: - db database name(s) - flen, fd database fdt length and embedded fd records - config embedded configuration database record - rowid mfn(s) (read, query, insert requests) - rec embedded data record(s) (read, query requests) - error, error2 error codes - msg error message if error is non-zero - total total size of query result set - size actual size of delivered result set * the session configuration - server Hostname - port - command command to be executed upon completion (async, if possible) - encoding encoding used by the server (default under windows is Cp850) - db (database) Name of default database on the server to be used for calls that don't specify one. - url specify server, port and db in url style - loglevel - logfile * see also > http://mini.net/tcl/969 Arts and Crafts of Tcl-Tk Programming Minimum used Tcl stuff: > http://sourceforge.net/projects/tcl Tcl/Tk 8.3.5 (for use with TclX, else 8.4.2 is ok) > http://www.hwaci.com/sw/tkhtml/ Tkhtml "2.0" The full story includes: > http://sourceforge.net/projects/tclx TclX 8.3.5 > http://sourceforge.net/projects/tix Tix 8.1.4 > http://sourceforge.net/projects/blt BLT 2.4z > http://sourceforge.net/projects/incrtcl [Incr Tcl] 3.2.1, [Incr Widgets] 4.0.1 > http://sourceforge.net/projects/tcllib Tcllib 1.3, BWidgets 1.4.1 > http://sourceforge.net/projects/tclxml tclXml/Dom/Xslt 2.6 Downloads: $ for dir in tcl tclx tix blt incrtcl tcllib tclxml oratcl; do http://belnet.dl.sourceforge.net/sourceforge/$dir done $ http://www.hwaci.com/sw/tkhtml/download.html --- $Id: Tcl.txt,v 1.14 2003/05/02 12:28:04 mawag Exp $ doc/ToDo.txt0100664000113000011300000000060507654006674011727 0ustar kripkripunderway ... - finish Tcl binding - create Tk-based GUI - utilities for > unirec nested records - XML I/O-conversion utilities - implement server version * next steps (not necessarily in that order) - update language bindings Perl, Java ... - rework formatting based on structured representation of format as a record - rework query processing based on structured representation of query doc/UeberSicht.txt0100644000113000011300000000562507654006331013112 0ustar kripkripDer OpenIsis Förderverein für open source Informationssysteme wurde im Oktober 2002 gegründet. Wir unterstützen Bibliotheken, Archive und Bildungseinrichtungen in Deutschland und anderen Ländern beim qualifizierten Informationsmanagement mit Programmierung und der Bereitstellung von Servern. Unser zentrales Projekt ist das Datenbanksystem > OpenIsis , seit Mai 2001 entwickelt von den Vereinsgründern. OpenIsis ist die freie ("open source") Variante im Kreis der > IsisIntro CDS/ISIS > CdsIsis software Familie . * wer benötigt "Informationssysteme"? Jeder, der qualitativ hochwertige Informationen hat! Bibliothekare investieren seit Alexandria ihr Wissen und ihre Erfahrung, damit das gesuchte Buch gefunden wird. Die großen Bibliotheken nutzen und entwickeln bibliographische Datenbanken seit langem; diese Möglichkeiten sind aber vielen kleineren Archiven wie denen von Wohlfahrtsorganisationen, Gewerkschaften oder sozialen Bewegungen nicht zugänglich. Selbst im reinen Online-Bereich (Web-Archive, Newsgroups, Mailing-Listen) weicht angesichts der zunehmenden Masse die Begeisterung darüber, etwas zu finden, dem Verständnis für die Notwendigkeit, das Gesuchte zuverlässig zu finden, auch und gerade wenn es selten referenziert wird. Das "Dublin Core" Metadaten-Schema und das Resource Description Framework (RDF) sind Beispiele neuerer Ansätze auf der Basis bibliographischer Prinzipien. * wer benötigt "open source" Informationssysteme? Die Kosten für brauchbare kommerzielle Systeme im fünf- bis sechsstelligen Euro-Bereich sind nur für wenige Institutionen tragbar. Nicht ohne Grund also ist das > http://www.unesco.org/webworld/isis/ CDS/ISIS System der UNESCO (gegen Schutzgebühr) dort sehr weit verbreitet, wo die Mittel knapp sind. Unsere > http://www.gnu.org/philosophy/free-sw.html freie Software ist nicht nur kostenlos, sondern vor allem frei von Beschränkungen. Die Nutzer können darauf aufbauen, sie verbessern oder an ihre Bedürfnisse anpassen, ohne auf die ursprünglichen Entwickler (uns) angewiesen zu sein. Wir sind überzeugt, dass ein wesentlicher Beitrag zur Minderung des "digital gap" nur geleistet werden kann, wenn nicht nur Zugang zu Wissen durch ein Informationssystem, sondern auch Zugang zu dem System selbst gewährleistet wird. * was hat das mit der "Informationsgesellschaft" zu tun? Auch die Bibliothekare und Techniker an den gut ausgestatteten Universitäten in den USA und Europa finden ja keineswegs die Unterstützung der Industrie, wenn sie an neuen Wegen arbeiten, der alten Idee des > http://www.informatik.uni-leipzig.de/~graebe/projekte/imd-mat.html Gemeinguts Wissen neues Leben einzuhauchen, wie etwa die > http://www.openarchives.org/ Open Archives Initiative oder unsere Freunde von der Tübinger > http://oc4s.org/ Open Community for Science . Ein Pfad zu gesellschaftlicher Entwicklung jenseits von reinen Verwertungsinteressen wird vielleicht von der "Linux Ökonomie" vorgezeichnet ... ... News doc/UniStat.txt0100664000113000011300000003217007654512305012444 0ustar kripkripunicode block statistics of general character categories, decomposition and uppercase mappings based on Blocks.txt and UnicodeData.txt > http://unicode.org/Public/UNIDATA/ * decomposition mappings see > http://unicode.org/Public/UNIDATA/UCD.html#Character_Decomposition_Mappings $ (canonical) 1926 circle 229 compat 673 final 240 font 1038 fraction 16 initial 171 isolated 238 medial 82 narrow 122 noBreak 5 small 26 square 205 sub 24 super 100 vertical 25 wide 104 $ * major category/block table Categories are letter, mark, numeric, punctuation, symbol, separator and other. Additional columns give number of characters which have an uppercase and canonical decomposition mapping, resp. Final columns give begin and end, block length and name. $ Let Mar Num Pun Sym Sep Oth upC deC beg end len block 52 0 10 23 9 1 33 26 0 0000 007F 128 Basic Latin 65 0 6 5 18 1 33 32 53 0080 00FF 128 Latin-1 Supplement 128 0 0 0 0 0 0 63 108 0100 017F 128 Latin Extended-A 183 0 0 0 0 0 25 75 91 0180 024F 208 Latin Extended-B 96 0 0 0 0 0 0 19 0 0250 02AF 96 IPA Extensions 36 0 0 0 44 0 0 0 0 02B0 02FF 80 Spacing Modifier Letters 0 107 0 0 0 0 5 1 4 0300 036F 112 Combining Diacritical Marks 113 0 0 2 5 0 24 56 26 0370 03FF 144 Greek and Coptic 239 6 0 0 1 0 10 119 52 0400 04FF 256 Cyrillic 16 0 0 0 0 0 32 8 0 0500 052F 48 Cyrillic Supplementary 78 0 0 8 0 0 10 38 0 0530 058F 96 Armenian 30 47 0 5 0 0 30 0 0 0590 05FF 112 Hebrew 147 41 20 9 5 0 34 0 8 0600 06FF 256 Arabic 34 28 0 14 0 0 4 0 0 0700 074F 80 Syriac 39 11 0 0 0 0 14 0 0 0780 07BF 64 Thaana 66 26 10 3 0 0 23 0 11 0900 097F 128 Devanagari 52 19 16 0 3 0 38 0 5 0980 09FF 128 Bengali 51 16 10 0 0 0 51 0 6 0A00 0A7F 128 Gurmukhi 52 20 10 0 1 0 45 0 0 0A80 0AFF 128 Gujarati 53 17 10 0 1 0 47 0 5 0B00 0B7F 128 Oriya 35 14 12 0 8 0 59 0 4 0B80 0BFF 128 Tamil 51 19 10 0 0 0 48 0 1 0C00 0C7F 128 Telugu 53 19 10 0 0 0 46 0 5 0C80 0CFF 128 Kannada 52 16 10 0 0 0 50 0 3 0D00 0D7F 128 Malayalam 59 20 0 1 0 0 48 0 4 0D80 0DFF 128 Sinhala 57 16 10 3 1 0 41 0 0 0E00 0E7F 128 Thai 40 15 10 0 0 0 63 0 0 0E80 0EFF 128 Lao 47 74 20 20 32 0 63 0 17 0F00 0FFF 256 Tibetan 47 15 10 6 0 0 82 0 1 1000 109F 160 Myanmar 79 0 0 1 0 0 16 0 0 10A0 10FF 96 Georgian 240 0 0 0 0 0 16 0 0 1100 11FF 256 Hangul Jamo 317 0 20 8 0 0 39 0 0 1200 137F 384 Ethiopic 85 0 0 0 0 0 11 0 0 13A0 13FF 96 Cherokee 628 0 0 2 0 0 10 0 0 1400 167F 640 Unified Canadian Aboriginal Syllabics 26 0 0 2 0 1 3 0 0 1680 169F 32 Ogham 75 0 3 3 0 0 15 0 0 16A0 16FF 96 Runic 17 3 0 0 0 0 12 0 0 1700 171F 32 Tagalog 18 3 0 2 0 0 9 0 0 1720 173F 32 Hanunoo 18 2 0 0 0 0 12 0 0 1740 175F 32 Buhid 16 2 0 0 0 0 14 0 0 1760 177F 32 Tagbanwa 54 31 20 6 1 0 16 0 0 1780 17FF 128 Khmer 129 4 10 11 0 1 21 0 0 1800 18AF 176 Mongolian 29 24 10 2 1 0 14 0 0 1900 194F 80 Limbu 35 0 0 0 0 0 13 0 0 1950 197F 48 Tai Le 0 0 0 0 32 0 0 0 0 19E0 19FF 32 Khmer Symbols 108 0 0 0 0 0 20 0 0 1D00 1D7F 128 Phonetic Extensions 246 0 0 0 0 0 10 121 245 1E00 1EFF 256 Latin Extended Additional 218 0 0 0 15 0 23 97 229 1F00 1FFF 256 Greek Extended 0 0 0 60 2 16 34 0 2 2000 206F 112 General Punctuation 2 0 17 4 6 0 19 0 0 2070 209F 48 Superscripts and Subscripts 0 0 0 0 18 0 30 0 0 20A0 20CF 48 Currency Symbols 0 27 0 0 0 0 21 0 0 20D0 20FF 48 Combining Diacritical Marks for Symbols 43 0 0 0 32 0 5 0 3 2100 214F 80 Letterlike Symbols 0 0 49 0 0 0 15 16 0 2150 218F 64 Number Forms 0 0 0 0 112 0 0 0 6 2190 21FF 112 Arrows 0 0 0 0 256 0 0 0 38 2200 22FF 256 Mathematical Operators 0 0 0 5 204 0 47 0 2 2300 23FF 256 Miscellaneous Technical 0 0 0 0 39 0 25 0 0 2400 243F 64 Control Pictures 0 0 0 0 11 0 21 0 0 2440 245F 32 Optical Character Recognition 0 0 82 0 78 0 0 26 0 2460 24FF 160 Enclosed Alphanumerics 0 0 0 0 128 0 0 0 0 2500 257F 128 Box Drawing 0 0 0 0 32 0 0 0 0 2580 259F 32 Block Elements 0 0 0 0 96 0 0 0 0 25A0 25FF 96 Geometric Shapes 0 0 0 0 145 0 111 0 0 2600 26FF 256 Miscellaneous Symbols 0 0 30 14 130 0 18 0 0 2700 27BF 192 Dingbats 0 0 0 6 22 0 20 0 0 27C0 27EF 48 Miscellaneous Mathematical Symbols-A 0 0 0 0 16 0 0 0 0 27F0 27FF 16 Supplemental Arrows-A 0 0 0 0 256 0 0 0 0 2800 28FF 256 Braille Patterns 0 0 0 0 128 0 0 0 0 2900 297F 128 Supplemental Arrows-B 0 0 0 28 100 0 0 0 0 2980 29FF 128 Miscellaneous Mathematical Symbols-B 0 0 0 0 256 0 0 0 1 2A00 2AFF 256 Supplemental Mathematical Operators 0 0 0 0 14 0 242 0 0 2B00 2BFF 256 Miscellaneous Symbols and Arrows 0 0 0 0 115 0 13 0 0 2E80 2EFF 128 CJK Radicals Supplement 0 0 0 0 214 0 10 0 0 2F00 2FDF 224 Kangxi Radicals 0 0 0 0 12 0 4 0 0 2FF0 2FFF 16 Ideographic Description Characters 9 6 13 27 8 1 0 0 0 3000 303F 64 CJK Symbols and Punctuation 89 2 0 0 2 0 3 0 27 3040 309F 96 Hiragana 94 0 0 2 0 0 0 0 31 30A0 30FF 96 Katakana 40 0 0 0 0 0 8 0 0 3100 312F 48 Bopomofo 94 0 0 0 0 0 2 0 0 3130 318F 96 Hangul Compatibility Jamo 0 0 4 0 12 0 0 0 0 3190 319F 16 Kanbun 24 0 0 0 0 0 8 0 0 31A0 31BF 32 Bopomofo Extended 16 0 0 0 0 0 0 0 0 31F0 31FF 16 Katakana Phonetic Extensions 0 0 50 0 191 0 15 0 0 3200 32FF 256 Enclosed CJK Letters and Months 0 0 0 0 256 0 0 0 0 3300 33FF 256 CJK Compatibility 6582 0 0 0 0 0 10 0 0 3400 4DBF 6592 CJK Unified Ideographs Extension A 0 0 0 0 64 0 0 0 0 4DC0 4DFF 64 Yijing Hexagram Symbols 20902 0 0 0 0 0 90 0 0 4E00 9FFF 20992 CJK Unified Ideographs 1165 0 0 0 0 0 3 0 0 A000 A48F 1168 Yi Syllables 0 0 0 0 55 0 9 0 0 A490 A4CF 64 Yi Radicals 11172 0 0 0 0 0 12 0 0 AC00 D7AF 11184 Hangul Syllables 0 0 0 0 0 0 896 0 0 D800 DB7F 896 High Surrogates 0 0 0 0 0 0 128 0 0 DB80 DBFF 128 High Private Use Surrogates 0 0 0 0 0 0 1024 0 0 DC00 DFFF 1024 Low Surrogates 0 0 0 0 0 0 6400 0 0 E000 F8FF 6400 Private Use Area 361 0 0 0 0 0 151 0 349 F900 FAFF 512 CJK Compatibility Ideographs 56 1 0 0 1 0 22 0 34 FB00 FB4F 80 Alphabetic Presentation Forms 591 0 0 2 2 0 93 0 0 FB50 FDFF 688 Arabic Presentation Forms-A 0 16 0 0 0 0 0 0 0 FE00 FE0F 16 Variation Selectors 0 4 0 0 0 0 12 0 0 FE20 FE2F 16 Combining Half Marks 0 0 0 32 0 0 0 0 0 FE30 FE4F 32 CJK Compatibility Forms 0 0 0 21 5 0 6 0 0 FE50 FE6F 32 Small Form Variants 140 0 0 0 0 0 4 0 0 FE70 FEFF 144 Arabic Presentation Forms-B 162 0 10 30 23 0 15 26 0 FF00 FFEF 240 Halfwidth and Fullwidth Forms 0 0 0 0 2 0 14 0 0 FFF0 FFFF 16 Specials 88 0 0 0 0 0 40 0 0 10000 1007F 128 Linear B Syllabary 123 0 0 0 0 0 5 0 0 10080 100FF 128 Linear B Ideograms 0 0 45 2 10 0 7 0 0 10100 1013F 64 Aegean Numbers 31 0 4 0 0 0 13 0 0 10300 1032F 48 Old Italic 26 0 1 0 0 0 5 0 0 10330 1034F 32 Gothic 30 0 0 1 0 0 1 0 0 10380 1039F 32 Ugaritic 80 0 0 0 0 0 0 40 0 10400 1044F 80 Deseret 48 0 0 0 0 0 0 0 0 10450 1047F 48 Shavian 30 0 10 0 0 0 8 0 0 10480 104AF 48 Osmanya 55 0 0 0 0 0 9 0 0 10800 1083F 64 Cypriot Syllabary 0 0 0 0 246 0 10 0 0 1D000 1D0FF 256 Byzantine Musical Symbols 0 30 0 0 181 0 45 0 13 1D100 1D1FF 256 Musical Symbols 0 0 0 0 87 0 9 0 0 1D300 1D35F 96 Tai Xuan Jing Symbols 932 0 50 0 10 0 32 0 0 1D400 1D7FF 1024 Mathematical Alphanumeric Symbols 42711 0 0 0 0 0 9 0 0 20000 2A6DF 42720 CJK Unified Ideographs Extension B 542 0 0 0 0 0 2 0 542 2F800 2FA1F 544 CJK Compatibility Ideographs Supplement 0 0 0 0 0 0 128 0 0 E0000 E007F 128 Tags 0 240 0 0 0 0 0 0 0 E0100 E01EF 240 Variation Selectors Supplement 0 0 0 0 0 0 65536 0 0 F0000 FFFFF 65536 Supplementary Private Use Area-A 0 0 0 0 0 0 65536 0 0 100000 10FFFF 65536 Supplementary Private Use Area-B 90547 941 612 370 3754 21 142187 763 1926 $ BMP: nonblock 4112 unassigned 2247 * detailled block stats see > http://unicode.org/Public/UNIDATA/UCD.html#General_Category_Values $ Basic Latin b0000 l128 Lu26 Ll26 Nd10 Zs1 Cc33 Pc1 Pd1 Ps3 Pe3 Po15 Sm6 Sc1 Sk2 Latin-1 Supplement b0080 l128 Lu30 Ll35 No6 Zs1 Cc32 Cf1 Pi1 Pf1 Po3 Sm4 Sc4 Sk4 So6 Latin Extended-A b0100 l128 Lu63 Ll65 Latin Extended-B b0180 l208 Cn25 Lu90 Ll84 Lt4 Lo5 IPA Extensions b0250 l96 Ll96 Spacing Modifier Letters b02B0 l80 Lm36 Sk44 Combining Diacritical Marks b0300 l112 Cn5 Mn107 Greek and Coptic b0370 l144 Cn24 Lu52 Ll60 Lm1 Po2 Sm1 Sk4 Cyrillic b0400 l256 Cn10 Lu120 Ll119 Mn4 Me2 So1 Cyrillic Supplementary b0500 l48 Cn32 Lu8 Ll8 Armenian b0530 l96 Cn10 Lu38 Ll39 Lm1 Pd1 Po7 Hebrew b0590 l112 Cn30 Lo30 Mn47 Po5 Arabic b0600 l256 Cn29 Lm3 Lo144 Mn40 Me1 Nd20 Cf5 Po9 So5 Syriac b0700 l80 Cn3 Lo34 Mn28 Cf1 Po14 Thaana b0780 l64 Cn14 Lo39 Mn11 Devanagari b0900 l128 Cn23 Lo66 Mn18 Mc8 Nd10 Po3 Bengali b0980 l128 Cn38 Lo52 Mn9 Mc10 Nd10 No6 Sc2 So1 Gurmukhi b0A00 l128 Cn51 Lo51 Mn12 Mc4 Nd10 Gujarati b0A80 l128 Cn45 Lo52 Mn13 Mc7 Nd10 Sc1 Oriya b0B00 l128 Cn47 Lo53 Mn8 Mc9 Nd10 So1 Tamil b0B80 l128 Cn59 Lo35 Mn3 Mc11 Nd9 No3 Sc1 So7 Telugu b0C00 l128 Cn48 Lo51 Mn12 Mc7 Nd10 Kannada b0C80 l128 Cn46 Lo53 Mn5 Mc14 Nd10 Malayalam b0D00 l128 Cn50 Lo52 Mn4 Mc12 Nd10 Sinhala b0D80 l128 Cn48 Lo59 Mn5 Mc15 Po1 Thai b0E00 l128 Cn41 Lm1 Lo56 Mn16 Nd10 Po3 Sc1 Lao b0E80 l128 Cn63 Lm1 Lo39 Mn15 Nd10 Tibetan b0F00 l256 Cn63 Lo47 Mn71 Mc3 Nd10 No10 Ps2 Pe2 Po16 So32 Myanmar b1000 l160 Cn82 Lo47 Mn10 Mc5 Nd10 Po6 Georgian b10A0 l96 Cn16 Lu38 Lo41 Po1 Hangul Jamo b1100 l256 Cn16 Lo240 Ethiopic b1200 l384 Cn39 Lo317 Nd9 No11 Po8 Cherokee b13A0 l96 Cn11 Lo85 Unified Canadian Aboriginal Syllabics b1400 l640 Cn10 Lo628 Po2 Ogham b1680 l32 Cn3 Lo26 Zs1 Ps1 Pe1 Runic b16A0 l96 Cn15 Lo75 Nl3 Po3 Tagalog b1700 l32 Cn12 Lo17 Mn3 Hanunoo b1720 l32 Cn9 Lo18 Mn3 Po2 Buhid b1740 l32 Cn12 Lo18 Mn2 Tagbanwa b1760 l32 Cn14 Lo16 Mn2 Khmer b1780 l128 Cn14 Lm1 Lo53 Mn20 Mc11 Nd10 No10 Cf2 Po6 Sc1 Mongolian b1800 l176 Cn21 Lm1 Lo128 Mn4 Nd10 Zs1 Pd1 Po10 Limbu b1900 l80 Cn14 Lo29 Mn9 Mc15 Nd10 Po2 So1 Tai Le b1950 l48 Cn13 Lo35 Khmer Symbols b19E0 l32 So32 Phonetic Extensions b1D00 l128 Cn20 Ll54 Lm54 Latin Extended Additional b1E00 l256 Cn10 Lu120 Ll126 Greek Extended b1F00 l256 Cn23 Lu69 Ll122 Lt27 Sk15 General Punctuation b2000 l112 Cn15 Zs14 Zl1 Zp1 Cf19 Pc3 Pd6 Ps3 Pe1 Pi5 Pf3 Po39 Sm2 Superscripts and Subscripts b2070 l48 Cn19 Ll2 No17 Ps2 Pe2 Sm6 Currency Symbols b20A0 l48 Cn30 Sc18 Combining Diacritical Marks for Symbols b20D0 l48 Cn21 Mn20 Me7 Letterlike Symbols b2100 l80 Cn5 Lu27 Ll12 Lo4 Sm6 So26 Number Forms b2150 l64 Cn15 Nl36 No13 Arrows b2190 l112 Sm27 So85 Mathematical Operators b2200 l256 Sm256 Miscellaneous Technical b2300 l256 Cn47 Ps2 Pe2 Po1 Sm32 So172 Control Pictures b2400 l64 Cn25 So39 Optical Character Recognition b2440 l32 Cn21 So11 Enclosed Alphanumerics b2460 l160 No82 So78 Box Drawing b2500 l128 So128 Block Elements b2580 l32 So32 Geometric Shapes b25A0 l96 Sm10 So86 Miscellaneous Symbols b2600 l256 Cn111 Sm1 So144 Dingbats b2700 l192 Cn18 No30 Ps7 Pe7 So130 Miscellaneous Mathematical Symbols-A b27C0 l48 Cn20 Ps3 Pe3 Sm22 Supplemental Arrows-A b27F0 l16 Sm16 Braille Patterns b2800 l256 So256 Supplemental Arrows-B b2900 l128 Sm128 Miscellaneous Mathematical Symbols-B b2980 l128 Ps14 Pe14 Sm100 Supplemental Mathematical Operators b2A00 l256 Sm256 Miscellaneous Symbols and Arrows b2B00 l256 Cn242 So14 CJK Radicals Supplement b2E80 l128 Cn13 So115 Kangxi Radicals b2F00 l224 Cn10 So214 Ideographic Description Characters b2FF0 l16 Cn4 So12 CJK Symbols and Punctuation b3000 l64 Lm7 Lo2 Mn6 Nl13 Zs1 Pd2 Ps10 Pe11 Po4 So8 Hiragana b3040 l96 Cn3 Lm2 Lo87 Mn2 Sk2 Katakana b30A0 l96 Lm3 Lo91 Pc1 Pd1 Bopomofo b3100 l48 Cn8 Lo40 Hangul Compatibility Jamo b3130 l96 Cn2 Lo94 Kanbun b3190 l16 No4 So12 Bopomofo Extended b31A0 l32 Cn8 Lo24 Katakana Phonetic Extensions b31F0 l16 Lo16 Enclosed CJK Letters and Months b3200 l256 Cn15 No50 So191 CJK Compatibility b3300 l256 So256 CJK Unified Ideographs Extension A b3400 l6592 Cn10 Lo6582 Yijing Hexagram Symbols b4DC0 l64 So64 CJK Unified Ideographs b4E00 l20992 Cn90 Lo20902 Yi Syllables bA000 l1168 Cn3 Lo1165 Yi Radicals bA490 l64 Cn9 So55 Hangul Syllables bAC00 l11184 Cn12 Lo11172 High Surrogates bD800 l896 Cs896 High Private Use Surrogates bDB80 l128 Cs128 Low Surrogates bDC00 l1024 Cs1024 Private Use Area bE000 l6400 Co6400 CJK Compatibility Ideographs bF900 l512 Cn151 Lo361 Alphabetic Presentation Forms bFB00 l80 Cn22 Ll12 Lo44 Mn1 Sm1 Arabic Presentation Forms-A bFB50 l688 Cn93 Lo591 Ps1 Pe1 Sc1 So1 Variation Selectors bFE00 l16 Mn16 Combining Half Marks bFE20 l16 Cn12 Mn4 CJK Compatibility Forms bFE30 l32 Pc5 Pd2 Ps9 Pe9 Po7 Small Form Variants bFE50 l32 Cn6 Pd2 Ps3 Pe3 Po13 Sm4 Sc1 Arabic Presentation Forms-B bFE70 l144 Cn3 Lo140 Cf1 Halfwidth and Fullwidth Forms bFF00 l240 Cn15 Lu26 Ll26 Lm3 Lo107 Nd10 Pc2 Pd1 Ps5 Pe5 Po17 Sm11 Sc5 Sk3 So4 Specials bFFF0 l16 Cn11 Cf3 So2 Linear B Syllabary b10000 l128 Cn40 Lo88 Linear B Ideograms b10080 l128 Cn5 Lo123 Aegean Numbers b10100 l64 Cn7 No45 Po2 So10 Old Italic b10300 l48 Cn13 Lo31 No4 Gothic b10330 l32 Cn5 Lo26 Nl1 Ugaritic b10380 l32 Cn1 Lo30 Po1 Deseret b10400 l80 Lu40 Ll40 Shavian b10450 l48 Lo48 Osmanya b10480 l48 Cn8 Lo30 Nd10 Cypriot Syllabary b10800 l64 Cn9 Lo55 Byzantine Musical Symbols b1D000 l256 Cn10 So246 Musical Symbols b1D100 l256 Cn37 Mn22 Mc8 Cf8 So181 Tai Xuan Jing Symbols b1D300 l96 Cn9 So87 Mathematical Alphanumeric Symbols b1D400 l1024 Cn32 Lu443 Ll489 Nd50 Sm10 CJK Unified Ideographs Extension B b20000 l42720 Cn9 Lo42711 CJK Compatibility Ideographs Supplement b2F800 l544 Cn2 Lo542 Tags bE0000 l128 Cn31 Cf97 Variation Selectors Supplement bE0100 l240 Mn240 Supplementary Private Use Area-A bF0000 l65536 Cn2 Co65534 Supplementary Private Use Area-B b100000 l65536 Cn2 Co65534 $ doc/Unicode.txt0100644000113000011300000001371207656461767012463 0ustar kripkrip some notes on the use of charsets with ISIS * what are charsets? Since computers can store nothing but numbers, but we want them to store characters, there has to a table telling which character is stored as which number, or, vice versa, which number is to display and print as which character. such tables are called charsets. Since the smallest unit of number storage is a byte, which can hold 256 different numbers from 0 to 255, many charsets are based on one byte and thus can hold up to 256 characters. such charsets are called one-byte-charsets . For many scripts, like the various versions of latin, greek, cyrillic, hebrew and arabic, 256 characters are more than enough. For others, namely chinese, japanese and korean ( > http://czyborra.com/charsets/cjk.html CJK ) scripts with several thousand characters, it's not enough. The modern > http://czyborra.com/charsets/vietnamese.html vietnamese script is based on latin letters but needs a vast amount of accented letters, so 256 isn't enough. Those scripts don't get by with one byte per character, so they need multi-byte-charsets, where two or more bytes are needed to encode one character. * what is UNICODE > http://czyborra.com/unicode/standard.html UNICODE is a big multi-byte-charset designed to include all > http://czyborra.com/unicode/characters.html characters needed in the world (over 40.000 by now), even for some ancient languages. The problems having several charsets are a) you have to know which charset is used in a given text, b) computer systems need to be aware of all possible charsets and c) it's not possible to have a text or database contain characters which are encoded in different charsets. Having all text in unicode solves those problems. Check out > http://www.unicode.org/iuc/iuc10/x-utf8.html this sample page - with a 21st century browser like Mozilla 5 (Netscape 6) you will see most or all of the letters. * ASCII-compatible charsets and encodings Many charsets use the numbers 0 to 127 in the same way: to represent the basic set of latin characters defined by > http://czyborra.com/charsets/iso646.html ASCII . Whenever there's a byte with a number in that range, this byte has the meaning of the corresponding ASCII-character. For example, the number 43 always is a plus sign +, which is important if a query expression is scanned for such characters. All > http://czyborra.com/charsets/iso8859.html ISO-8859-x charsets are ASCII-compatible. Older > http://czyborra.com/charsets/cyrillic.html Cyrillic charsets are NOT compatible with ASCII. Some of the eastern multi-byte-charsets are, some are not. Some of the multi-byte-charsets have different encodings, that is, there is only one table mapping numbers to letters, but distinct ways to use multiple bytes to express such a number, some of which use the numbers in the ASCII-range only for ASCII characters, others don't. UNICODE has two widely used encodings, > http://czyborra.com/utf/#UTF-8 UTF-8 and UTF-16 (UCS-2). UTF-8 is ASCII-compatible, UTF-16 is not. * so what about ISIS - the ISIS database format itself is capable of storing anything and thus can store text in any charset/encoding. tools like biremes mx may store and retrieve (by MFN) text in nearly any encoding (but depending on how the programming is done, UTF-16 may not work because it may use bytes with value 0). - the ISIS query and formatting language depends on special ASCII-characters having special meaning and therefore will require an ASCII-compatible encoding. All the ISO-8859-x charsets will do as will UTF-8 encoded unicode (although some care must be taken when multiple bytes representing one character are cut off in the midth). At least in theory, mx and wwwisis are able to search for records in any ASCII-compatible encoding including UTF-8 unicode (given carefull web-programming). - winisis doesn't know about the possibility of one character having multiple bytes. It will work with any ASCII-compatible one-byte-charset , as long as it doesn't have to know what it does. That is, if your computer has some preferred charset installed, you will see all characters displayed according to that charset, and a character possibly entered as the german ä could show up as greek delta :). No support for multi-byte-charsets, especially not unicode. - Like any Java software, > http://web.tiscali.it/javaisis/ JavaISIS is - in theory - able to handle unicode characters and even to do the transformation between unicode and most of the other charsets. Some limitations may result from the underlying wwwisis. In practice, version 3.5 claims to give "Multi-language encoding support", but unfortunately it's in beta since March 2001 (sources made available in Feb 2002). - openisis supports any charset and with it's Java-binding, especially unicode and all the conversions. openisis alone can do it on the web, and in combination with JavaISIS (once new sources are available) also with a winisis-like interface. * some other resources on unicode To see all those characters, you need fonts to tell your display or printer how they look like. Here's a > http://www.hclrss.demon.co.uk/unicode/fonts.html very fine page on how to acquire and install those fonts (and some more advice). James Kass has a > http://home.att.net/~jameskass/scriptlinks.htm long list of high quality links related to Unicode. If you for some reason have to waste your time with M$ products, you may want to check out > http://www.microsoft.com/typography/fonts/ this page . Especially there's the one-size(23 MB)-fits-all fat font > http://office.microsoft.com/downloads/2000/aruniupd.aspx Arial Unicode MS (TM, (c), ... expect the worst) containing nearly all unicode glyphs, which is also included with newer Windoze and/or Ophice versions. See > UniStat statistics about how characters are distributed amongth Unicode. For example, the only scripts using uppercase/lowercase are those derived from Greek (i.e. Latin, Cyrillic, Armenian and Georgian). --- $Id: Unicode.txt,v 1.2 2003/05/08 14:04:39 kripke Exp $ doc/Views.txt0100644000113000011300000002230107655442134012146 0ustar kripkripUsing views in OpenIsis. A "view", like a VIEW in SQL, creates new, typically temporary records based on existing ones by means of some transformation like selecting a subset of the available fields (a projection), retagging fields or manipulating field values. As general concept, a view can be implemented using any algorithm in any of the available programming languages to create new records (and need not only refer to record contents, but may also access other ressources like files). In a more narrow sense, however, a view is a special kind of transformation defined by a "view record". The fields of a view record have tags as they should appear in the target, typically some valid tags of the source plus, for example, index control tags, if the view describes indexing. In the following, the term "alphanumeric" denotes any ASCII letter or digit, or any non-ASCII character. "Word character" denotes any alphanumeric, hyphen '-' or underscore '_'. The value can have one of several forms: - if it is empty, the tag is passed to the source record's v command (see below). - if it starts with a %, the rest of the value (w/o the %) is passed to the source record's v command. If the tag is not 0, '=tag;' is prepended. - if the value starts with any word character, it is used literally. - if it starts with a quote, the rest of the value is used literally (w/o the quote). If the value's last character is a quote, it is discarded. - if it starts with an @, the rest of the value names a view to be included - if it starts with an &, the rest of the value is the name of an extension exit to call - if it starts with an {, the rest of the value is a script to be executed in the host language (after stripping an optional } as last character) - any other form (i.e. starting with other ASCII punctuation) is reserved for future use Example: the view $ 24 70 $ is a simple projection selecting fields 24 and 70 from the source. * the v command is described here as an abstract command. It is available in the C-API as well as from the language bindings, possibly with language specific variations. It resembles the core concepts of traditional formatting, including access to and looping over fields and subfields, selecting substrings and attaching optional literals. It is sort of the record's printf. Like printf, and unlike traditional formatting, it neither supports flow control nor screen rendering. It takes a source and target record plus a string specifying a format. Depending on the language environment, the source and/or target may be implicit. If the format starts with '=tag;', where tag is a tag, this gives the tag used in the target and as default. Otherwise, tags from the source are used in the target and default is *. The first (next) character is then checked for an encoding mode, see below. The format is a series of output specifications, consisting of a field tag (word characters, either numerical or by field name), selectors and modifiers. The special tag * selects all fields. Each spec may contain several subspecs, separated by commas, using the same child context (otherwise, specs and subspecs are the same). So the format is spec[;spec...], and a spec is spec[,subspec...]. The general operation of the v command is to loop over the record until the last occurence was seen for all tags. In the nth repetition, for each tag in any spec, the (n+i)th occurence of a field with this tag is used, where i is an offset given by an occurence selector. Determine whether this is the last occurence. For every iteration, a new output field is started, and the format is processed as follows: - loop over the (main) specifications - loop over childs (or use the given field) - loop over subspecs - loop over subfields (or use the whole field) - apply decoding - apply substring - apply encoding - attach literals - append the result to the target record Each spec starts with an optional decoding mode, optionally followed by a tag, optionally followed by a child selector, optionally followed by a subfield selector, optionally followed by string modifiers, optionally intermingled with occurence selectors and literals: - , starts a new subspec - ; starts a new spec with default context reset to the last tag seen - . starts a child selector - ^% start a subfield selector - ([ start an occurence selector - /~"'`|+ start a literal - : starts a substring selector - & calls an extension - { evaluates a script * encoding mode One of the following operators as first character of the format can select an output "encoding": - ? outputs a 1, if the selected entitity exists, 0 else - ! the opposite of ? - & applies HTML encoding - % applies URL encoding The test encodings ?! inhibit normal processing; they immediatly return after checking the first occurence of the the first tag. For example, using a default of all tags (*), the format consisting solely of a '?' checks wether a record is empty. More special characters (but not the '*') may be designated in the future, so a format should always start with a tag (possibly explicit *). * decoding mode An uppercase character before the tag may denote a decoding mode: $ - H heading mode: ^x is replaced as ';' for x=a, ',' for x=b..i, '.' for others angle brackets are removed (>< replaced by '; '), or evaluates to a - D data mode: in addition to heading mode, if there is no explicit literal after this field, append ' ', if it ends in "punctuation", or '. ' else. - X index mode like heading, but evaluates to nothing and to b - M traditional For compatibility, specs reading MHx or MDx (x = L or U) set heading or data mode, resp., as default processing (before substringing). The case directive is ignored. $ * child selector If a tag is immediatly followed by a dot '.' and optional tag, field context is switched, for this spec and following specs separated by ',', to loop over the childs with the given tag. Tag defaults to 0, selecting text nodes in the canonical XML representation. A * selects all childs, a second . recursively selects all childs. * subfield selectors The primary subfield selector is the hat '^', followed by one character. It can produce multiple items, like repetitions of a subfield or keywords. If the selector character is - alphanumeric select the (repetitions of the) subfield tagged with this character. - an opening pairing brace i.e. one of '(','{','[' or the angle bracket '<', words between pairs of this brace are selected (commonly keywords). - a * selects the part up to the first subfield delimiter - a space selects naive words as sequences of alphanum - a ) selects parts between TABs (array mode) - other punctuation like / or | selects parts between pairs of this character The percent sign '%' (think printf) works basically like the hat, but - removes quotes surrounding values - by default treats the TAB as subfield delimiter - if followed by a punctuation character or space, treats this plus surrounding whitespace as delimiter, not separating within quotes. - if followed by a ), (optionally after another punctuation) goes to array mode, that is there is no subfield indicator stripped from the values - if followed by multiple word characters, (including '-' and '_', optionally after an initial punctuation) searches for subfields starting with that sequence followed by '=' or ':' Examples: - '^)' splits at TABs - '%)' splits at TABs with quote removal - '%a' selects a sequence following a TAB and 'a' - '%,)' splits a line of comma separated values - '%;*' selects the primary value of a MIME property - '%;charset' selects the charset attribute of a MIME property * occurence selector By default, all occurences of fields, childs and subfields are used. One or multiple occurences can be selected explicitly following a tag, child selector or subfield selector using brackets [] (counting from 1) or parentheses (counting from 0) like (i) or (i..j). - If i is ommited, it defaults to the first (1 or 0, resp.). - If j is ommited, it defaults to last. * literals Each tag, child or subfield selector may be followed by one or more literals. Every literal but the / extends to the next occurence of the same special character by which it is introduced. This special character may be escaped using a backslash. A literal backslash may be escaped as two (but need not, except at the end). The special character governs when and where the literal is output: - " before the first occurence (of the entity in question; i.e. field, child or subfield) - ' before each - ` after each - | inbetween (after each but the last) - + after the last - / this single-character literal starts a new output field after each occurence - ~ this literal is used if the given entitity does NOT occur Literals are not subject to the string modifiers. * substring selector Introduced by a colon ':', it has the form :l or :o.l, where o and l are integers denoting an offset and length to cut from the currently selected value. * extension exits An exit is a C-function (i.e., using C calling convention) in a dynamic library. TODO: describe interface. * script evaluation If a scripting environment like Tcl is available, a {} block may contain a script to be evaluated. TODO: describe interface. --- $Id: Views.txt,v 1.2 2003/05/05 11:00:44 kripke Exp $ doc/WritePerf.txt0100664000113000011300000000527207654006331012764 0ustar kripkripwriting performance test: copying the 58970 records of unesb (~ 500 byte each) with simple full index (-idxall) making one entry for every field (total of 585334 fields) system: $ cat /proc/cpuinfo ... model name : Pentium III (Coppermine) cpu MHz : 801.832 cache size : 256 KB free total used free shared buffers cached Mem: 256252 150428 105824 0 640 70560 -/+ buffers/cache: 79228 177024 Swap: 265032 7668 257364 $ initial write to non existing db db/test/t, creating a 30MB .mst and 18MB .oxi $ time ./openisis -db db/unesb/unesb -append db/test/t -fmt mfn -idxall >/dev/null real 0m20.304s user 0m15.680s sys 0m3.600s ./openisis -db db/test/t -fmt mfn | wc -l 58970 ./openisis -db db/test/t -ifdump | wc -l 585334 $ that is about 2,900 rows + 29,000 index entries per second. (without indexing, it writes nearly 15,000 rows a second). successive append are substantially slower, as IO cache is exhausted. * the never reading system On a 650 MHz system with *lots* of cache memory (it actually *never* reads a block), a test of five successive appends yields pretty much constants times around 29 seconds per pass (~ 2,000 rows + 20,000 entries / sec), although the .mst stacks up to 294850 records and the index then has 2926670 entries. That same system does on that fat db one million random reads in 10 seconds (~ rows 100,000 / sec) and an ifchk (252465 keys in 50596 blks of 1K depth 4) in 5.18 secs, accessing 48,738 per second. * dual CPU use Using both CPUs of this box, four parallel writers (copying unesb to four different targets) take a total of 60 seconds, giving a combined throughput of more than 3,900 rows + 39,000 index entries / sec. * comparision While these numbers are all derived with single threaded debugging code, they yet outperform the (more or less) "relational" databases by one or even two orders of magnitude. For example, MySQL tested 350768 inserts without key in in 619*0.62 seconds, yielding 914 rows/sec as compared 15,000 with OpenIsis. Insert with key (one, probably) did less than 35 rows per second, compared to around 3,000 (with ten keys each). Select (probably reading all rows) did 20,000 in 39*0.54 seconds, that is 950 per second -- compared to 100,000 with OpenIsis. I couldn't figure out what kind of system was used for these > http://www.mysql.com/information/benchmark-results/result-db2,informix,ms-sql,mysql,solid,sybase-relative.html MySQL benchmarks, however, it probably was not much less than our 800 MHz Pentium. Most other databases are even slower, and that, by the way, is also our experience when working with Oracle and the like. doc/XReunion.txt0100664000113000011300000000355607654006331012627 0ustar kripkripOpenIsis joined the X. regional meeting of CDS/ISIS distributors and experts in Latin America and the Caribbean, Feb. 3.-5. in la Havana, Cuba. * Overview Among the more than 50 participants where Davide Storti, chief CDS/ISIS manager in UNESCO HQ Paris, Adalberto Tardelli, senior developer at BIREME, and known expert Ernesto Spinak from Uruguay. Davide announced that there are new developments underway to build a much advanced successor of WinISIS, which shall finally be released under some open source license. OpenIsis was welcomed in the "CDS/ISIS family" and possibly may make some contributions to further developments. * Presentations Just to pick a few among a couple of interesting speeches: - Adalberto explained how BIREME is focussing their next software generation on XML storage to enhance, for example, the > http://www.scielo.org/ SCIELO platform. - Ernesto introduced IsisMarc, a windows app for data entry with strong support for any MARC standard, including the highly complex MARC21. - Elysio Mira Soares from Brasil developed the library automation system > http://www.elysio.com.br/eng_index.htm PHL , which is under GPL, but users are asked to contribute their contents to a pool. - A group of Cuban developers from INFOMED are working on CLABEL (Catálogo en linea para automatizar bibliotecas electrónicas), which shall be a KOHA replacement based on CDS/ISIS (under GPL, of course), while their colleagues are developing > http://webservices.sld.cu/test/isisearch/ XML-RPC interfaces for ISIS. - Paulo Cattelan from Porto Alegre presented the ISIS information and helpdesk system > http://www.oraculo.inf.br/ ORACULO. - From Argentina we saw the LGPLd > http://www.bl.fcen.uba.ar/openopac.php OpenOPAC. OpenIsis prepared bilingual (english/espanol) papers on > Cuba state and roadmap of our development and > CubaPHP the OpenIsis PHP extension by Braulio. doc/btree.txt0100644000113000011300000003261107654006331012151 0ustar kripkripOpenIsis deploys a B-Tree structure similar to that of CDS/ISIS. There are several differences to the original CDS/ISIS file layout: - The complete index is kept in one file, including control information (.CNT), inner nodes (.N0x) and leave nodes (.L0x) of any key length and "inverted file pointers" (.IFP, see below) - The maximum key length is configurable up to 255 bytes (CDS/ISIS: 30) - The maximum mfn is configurable up to 6 bytes (CDS/ISIS: 3) - The keys are not blank padded. - The blocks have no fixed spread factor (CDS/ISIS: 10), but a fixed size of mK (where m is 1,2,4 or 8), allowing for about 200 keys of 30 bytes. - A custom string comparison function may be used, allowing for proper unicode collation. - To enhance safe high concurrency, it is a B-Link-Tree as used in postgres (see Lehmann/Yao 1982), i.e. each block has a link to it's right sibling. Advantages: - The number of necessary I/O operations is dramatically reduced, typically to 1. Since the invention of B-Tree's by Bayer, the ratio of available main memory to disk storage is ~ 1:100. Since the ratio of the total size of the inner nodes to the total index size is aproximately equal to the spread factor, it is crucial that this is at least 100, allowing all the inner nodes to be in RAM. - The price paid for this, is that, since the inner nodes locate the leaf wanted less precise (up to a block of 200 entries instead of 10), there is more unwanted data read from disk. However, all I/O is in mK blocks on mK boundaries, matching the "page size" of most modern operating systems. This is the basic idea of B-Trees in the first place. - The maximum database size is much raised with regard to the mfn. This allows for an index spanning several databases, where the higher mfn bytes indicate the relevant master file. Limits: The theoretical maximum number of terms is also somewhat raised from 20G (2G leaf blocks per 10 terms) to about 400G (4G blocks per 200 entries). Without support for large files (> 2G), however, the limit of the one-file design may actually even be lower, since the 2G file limits the total size of terms + IFPs. * Basic operation: The B-L-Tree relates keys to "inverted file pointers" (IFP). The key is a byte sequence of a length between 0 and 255. The IFPs are 8 to 11 bytes long, with 3 to 6 bytes for an mfn (rowid), and 5 bytes describing where the key occurs within the given record: 2 field tag, 1 occ(repetition of field), 2 word position. These numbers are in big-endian byte order for easy sorting. The mfn/value length is fixed for a given B-Tree. File structure: The index file consists of consecutive mK blocks, with block number n at file offset n*mK. Block 0 is the root, Block 1 the leftmost leaf. Block structure: The inner structure of a block is similar to an ISIS record: some header fields followed by a dictionary of entries, which is an array of positions (offset into block) and lengths. The dictionary is sorted according to the entry's keys. The actual entry data is located at the end of the block, growing towards the dictionary as entries are added. Entry structure: Each entry starts with len bytes key as stated in the dictionary. For an inner node, it is optionally followed by an IFP, as indicated by a dictionary flag, then by a 4 byte block number. For a leaf node, it is followed by a sorted array of IFPs. Searching: Since it is not uncommon for a key to relate to more distinct IFPs than would fit within one block, we actually consider the pair of key and IFP as the key when searching the tree. When looking for all IFPs for a key, we start with empty (nulled) IFP. Upon modification (insert/delete), we use the given IFP as second key segment. Where the IFPs for one key span several leaf nodes, the separating inner node entry will have it's key augmented with the IFP, so we may quickly locate the proper leaf node. Deleting: The entry is searched and deleted, no restructuring. (Background cleanup may be implemented some day). Inserting: If a new entry fit's within the target leaf, everything is fine. Else either the new entry, or one or more other entries may have to be shifted to the right sibling node. If the right sibling is non-existent or too full (or we didn't yet implement checking it), a new right sibling is created. The new separating key, which is lower than the one we stopped at while locating our node, is then updated or inserted in the parent. Multi-Process Concurrency: There *must not* be more than 1 process having the index open for writing. Readers, however, should work correctly, even in the presence of a writer. Multi-Thread Concurrency: Within a multi-threaded process, all access to internal structures is interlocked within a monitor. The monitor is released during I/O, with the block marked as being read or written from/to disk. A thread wishing to access a block being read, or wishes to write to a block while it is written to disk, must wait on the monitor. Lehmann/Yao Concurrency: While this design should make sure that each block is always in a consistent state, the interrelation between blocks may be disturbed by node splitting or shifting: The key may be found to be no longer in the expected block. Thus, while the search key is greater than the greatest key in the current block, one has to try the right sibling. * Concurrency Details: Within a multi-threaded process, all access to internal structures is interlocked within a monitor. The monitor is released during I/O, with the block marked as being read or written from/to disk. This design will give the best possible over-all utilization on a single-CPU system: letting a given thread run exclusively while it can utilize the CPU. (I.e. unless it's waiting for I/O, swapping shouldn't be an issue, if you're seriously running a db server). If you have a multi-CPU system for nothing but OpenIsis, run secondary read-only servers, each bound to one CPU. This will be by far more efficient than sharing multiple CPUs by one processe's threads. Actually, we deploy one mutex (LOCK/UNLOCK) and one condition (WAIT/WAKE) bound to it. Optionally a set of conditions bound to the same mutex may be used. All access to btree blocks is via a common cache, so that there is never more than one copy of a block in memory. During I/O, the cache block is flagged as being read or written, resp. The read flag is held only temporarily during the I/O; it marks the block contents as being asynchronously changed (by the read call) and thus completely useless. Any thread wishing to access the block has to wait on it. After the read call returns, the thread initiating the read will re-acquire the monitor, clear the flag and wake any waiting threads. The write flag, on the other hand, acts as a WRITE LOCK on the block. Threads wishing read-only access to the block completely ignore it, since the block content is valid and will NOT asynchronously change: it is changed by the writing thread only while it has the monitor, thus no other thread will ever observe a change. A thread wishing to modify the block, however, has to wait, since a modification during an ongoing write could corrupt data. Moreover, the lock need not be released immediatly after the write call returns, but can be held until a second block is also written. * The block reading sequence is as follows: $ r0 find the block in the cache. if it is not found, set it reading, release the monitor, read. else, if it is being read, increase waiter count, wait (releases the monitor). else use it (skip next state). [r1] SUSPENDED wait for either our own read to return or to be waked by another reader. if we return from read, re-acquire the monitor. (if we are awaked, the monitor is re-aqcuired implicitly). r2 if we return from read, clear the reading flag, wake waiting threads. if we are awaked, decrease waiter count. (Since we might have less different conditions than blocks, we have to check the reading flag to see whether we were meant and possibly wait again). we have the block wanted. on searching, if it's not the leaf we wanted, start over. $ NOTE: - the thread starting the read on a block is always the first to get it, since waiters can proceed only after this thread cleared the flag. - for a block that already is in cache, the reader proceeds without releasing the monitor. This especially holds for blocks being written. These properties can be used for a save background cleanup. - a reader does not need to hold a block in the cache while it is suspended in I/O: all processing is immediately after reading. For a secondary reader on a database being written, at least leaf blocks should be invalidated asap. (Otherwise it will obviously miss some updates). * The overall search algorithm: $ s1 starting from the root, locate the first entry with key greater than our key, read the block it pointed to by the previous entry. if there is no such entry found, and the block has a right sibling, check it for possible split or shift (Lehmann/Yao). s2 when reaching a matching entry in a leaf, read subsequent right siblings according to the search range criteria and buffer. $ NOTE: - due to the right sibling checking, it does not lead to inaccurate search results, if inner nodes have missing or too high keys for some childs, it only has a performance penalty (dropping the tree altogether means linear search). * The general block writing sequence: $ w0 reading "for update" as r0, but with additional condition: if it has the WRITE LOCK flag set, wait, else use it (skip next state). [w1] SUSPENDED wait for read or to be waked by other writer. on return from read, re-acquire the monitor. w2 if we are awaked and there's a WRITE LOCK, wait again. (the original reader or some other waiting thread might have set it). after returning from read, set the WRITE LOCK flag, wake waiting readers. if we need to write another block, start over. process block(s). release the monitor, start write call on first block to write. [w3] SUSPENDED wait for the write call to return. repeat while there are more blocks to be written. re-acquire the monitor. w4 clear the WRITE LOCK flag on some blocks, wake waiting threads. if there are more blocks to be processed, start over. $ * The insert algorithm: $ i0 locate the leaf L where to insert the entry. if it has enough room, process it and we're done. i1 if it has a right sibling S, read it for update. if this hasn't enough room to take some entries from L, release the write lock and use a new block S as L's right sibling instead. shift entries to S, insert the new entry into L (or S). i2 write first S, then L to disk. release the write lock on L. i3 read the parent P of S for update, i.e. the parent of L or some block to the right. if we used old S, delete it's old entry. insert the new lower bound on the keys of S. if the parent hasn't enough room, start over at i1 (with P for L). $ when unfolded to synchronized/unsynchronized steps, a simplified version of this might look like: $ read your way down to L until it's read for update... u0 not enough space in L, read S for update [u1] wait for S (L locked) v2 modify both blocks, initiate read of P for update [v3] wait for P, write S then L (L,S locked) v4 release lock on L and S, update P or go to u1 (P locked) $ NOTE: - this is deadlock-free, since locking is ordered left to right, bottom to top. - by always keeping one lock, we avoid conflicts with other writers for the same entries, since those have to follow the same lock path, and thus work always in a serializable manner. - if we are the original reader for P, other threads wanting to read P have to wait for three I/Os. The detailled variant below avoids this. - reading the sibling and therefore the [u1] suspend is optional. depending on the key distribution, it might not hurt much or even be more efficient to always create a new block. - the only inconsistency a reader thread might notice is during v3 (or a later suspend in order to lock P's sibling): there are new versions of L and S, but P is in the old state. The key of S (or L's former sibling) is too high and thus an entry now in S may be searched for in L. In this situation, the reader has to follow L's right link to find the key. There is a slightly more complex variant of insert, wich allows for somewhat enhanced concurreny but with higher synchronization effort: $ u2 modify both blocks (L,S locked) [u3] write S then L u4 release lock on L (S locked), read P for update [u5] wait for P u6 release lock on S, update P or go to u1 (P locked) $ * Multi-Process NOTEs: - we have to assume, that block writes * are atomic (seen by others completely or not at all) and * are ordered (if a later write is seen, an earlier is also). This does NOT hold for the file state on disk, but should hold for the operating system's file cache. - a reader in another process might also read the new version of S but the old version of L and thus has to deal with duplicate entries when searching ranges. - if a reader is caching, and thus write order is not in effect, it might miss entries completely. This does still give correct results, if * no leaves are cached * for all cached blocks, the parent is older This is ok for one-shot readers, or if caching is strictly from top with forced re-read after cache miss on search path. Deleting an entry is trivial, since it affects one block only (if we're lazy). In a single-process multi-threaded environment, however, we can take advantage of certain properties to provide relatively simple and save cleanup synchronously or in background. doc/charsets.html0100644000113000011300000001465407655441447013034 0ustar kripkrip ISIS charsets and unicode

some notes on the use of charsets with ISIS

what are charsets?

Since computers can store nothing but numbers, but we want them to store characters, there has to a table telling which character is stored as which number, or, vice versa, which number is to display and print as which character. such tables are called charsets.
Since the smallest unit of number storage is a byte, which can hold 256 different numbers from 0 to 255, many charsets are based on one byte and thus can hold up to 256 characters. such charsets are called one-byte-charsets .
For many scripts, like the various versions of latin, greek, cyrillic, hebrew and arabic, 256 characters are more than enough.
For others, namely chinese, japanese and korean ( CJK ) scripts with several thousand characters, it's not enough. The modern vietnamese script is based on latin letters but needs a vast amount of accented letters, so 256 isn't enough. Those scripts don't get by with one byte per character, so they need multi-byte-charsets, where two or more bytes are needed to encode one character.

what is UNICODE

UNICODE is a big multi-byte-charset designed to include all characters needed in the world (over 40.000 by now), even for some ancient languages. The problems having several charsets are a) you have to know which charset is used in a given text, b) computer systems need to be aware of all possible charsets and c) it's not possible to have a text or database contain characters which are encoded in different charsets. Having all text in unicode solves those problems. Check out this sample page - with a 21st century browser like Mozilla 5 (Netscape 6) you will see most or all of the letters.

ASCII-compatible charsets and encodings

Many charsets use the numbers 0 to 127 in the same way: to represent the basic set of latin characters defined by ASCII . Whenever there's a byte with a number in that range, this byte has the meaning of the corresponding ASCII-character. For example, the number 43 always is a plus sign +, which is important if a query expression is scanned for such characters.
All ISO-8859-x charsets are ASCII-compatible. Older Cyrillic charsets are NOT compatible with ASCII. Some of the eastern multi-byte-charsets are, some are not.
Some of the multi-byte-charsets have different encodings, that is, there is only one table mapping numbers to letters, but distinct ways to use multiple bytes to express such a number, some of which use the numbers in the ASCII-range only for ASCII characters, others don't. UNICODE has two widely used encodings, UTF-8 and UTF-16 (UCS-2). UTF-8 is ASCII-compatible, UTF-16 is not.

so what about ISIS

  • the ISIS database format itself is capable of storing anything and thus can store text in any charset/encoding.
    tools like biremes mx may store and retrieve (by MFN) text in nearly any encoding (but depending on how the programming is done, UTF-16 may not work because it may use bytes with value 0).
  • the ISIS query and formatting language depends on special ASCII-characters having special meaning and therefore will require an ASCII-compatible encoding. All the ISO-8859-x charsets will do as will UTF-8 encoded unicode (although some care must be taken when multiple bytes representing one character are cut off in the midth). At least in theory, mx and wwwisis are able to search for records in any ASCII-compatible encoding including UTF-8 unicode (given carefull web-programming).
  • winisis doesn't know about the possibility of one character having multiple bytes. It will work with any ASCII-compatible one-byte-charset , as long as it doesn't have to know what it does. That is, if your computer has some preferred charset installed, you will see all characters displayed according to that charset, and a character possibly entered as the german ä could show up as greek delta :). No support for multi-byte-charsets, especially not unicode.
  • Like any Java software, JavaISIS is - in theory - able to handle unicode characters and even to do the transformation between unicode and most of the other charsets. Some limitations may result from the underlying wwwisis. In practice, version 3.5 claims to give "Multi-language encoding support", but unfortunately it's in beta since March 2001 (sources made available in Feb 2002).
  • openisis supports any charset and with it's Java-binding, especially unicode and all the conversions. openisis alone can do it on the web, and in combination with JavaISIS (once new sources are available) also with a winisis-like interface.

some other resources on unicode

To see all those characters, you need fonts to tell your display or printer how they look like. Here's a very fine page on how to acquire and install those fonts (and some more advice). James Kass has a long list of high quality links related to Unicode. If you for some reason have to waste your time with M$ products, you may want to check out this page . Especially there's the one-size(23 MB)-fits-all fat font Arial Unicode MS (TM, (c), ... expect the worst) containing nearly all unicode glyphs, which is also included with newer Windoze and/or Ophice versions. doc/encoding.txt0100644000113000011300000001224207654006331012634 0ustar kripkrip95 single-byte and 45 multi-byte encodings as supported by Java rt+i18n.jar $ ASCII American Standard Code for Information Interchange Cp037 USA, Canada (Bilingual, French), Netherlands, Portugal, Brazil, Australia Cp273 IBM Austria, Germany Cp277 IBM Denmark, Norway Cp278 IBM Finland, Sweden Cp280 IBM Italy Cp284 IBM Catalan/Spain, Spanish Latin America Cp285 IBM United Kingdom, Ireland Cp297 IBM France Cp420 IBM Arabic Cp424 IBM Hebrew Cp437 MS-DOS United States, Australia, New Zealand, South Africa Cp500 EBCDIC 500V1 Cp737 PC Greek Cp775 PC Baltic Cp838 IBM Thailand extended SBCS Cp850 MS-DOS Latin-1 Cp852 MS-DOS Latin-2 Cp855 IBM Cyrillic Cp856 IBM Hebrew Cp857 IBM Turkish Cp858 Variant of Cp850 with Euro character Cp860 MS-DOS Portuguese Cp861 MS-DOS Icelandic Cp862 PC Hebrew Cp863 MS-DOS Canadian French Cp864 PC Arabic Cp865 MS-DOS Nordic Cp866 MS-DOS Russian Cp868 MS-DOS Pakistan Cp869 IBM Modern Greek Cp870 IBM Multilingual Latin-2 Cp871 IBM Iceland Cp874 IBM Thai Cp875 IBM Greek Cp918 IBM Pakistan (Urdu) Cp921 IBM Latvia, Lithuania (AIX, DOS) Cp922 IBM Estonia (AIX, DOS) Cp1006 IBM AIX Pakistan (Urdu) Cp1025 IBM Multilingual Cyrillic: Bulgaria, Bosnia, Herzegovinia, Macedonia (FYR) Cp1026 IBM Latin-5, Turkey Cp1046 IBM Arabic - Windows Cp1097 IBM Iran (Farsi)/Persian Cp1098 IBM Iran (Farsi)/Persian (PC) Cp1112 IBM Latvia, Lithuania Cp1122 IBM Estonia Cp1123 IBM Ukraine Cp1124 IBM AIX Ukraine Cp1140 Variant of Cp037 with Euro character Cp1141 Variant of Cp273 with Euro character Cp1142 Variant of Cp277 with Euro character Cp1143 Variant of Cp278 with Euro character Cp1144 Variant of Cp280 with Euro character Cp1145 Variant of Cp284 with Euro character Cp1146 Variant of Cp285 with Euro character Cp1147 Variant of Cp297 with Euro character Cp1148 Variant of Cp500 with Euro character Cp1149 Variant of Cp871 with Euro character Cp1250 Windows Eastern European Cp1251 Windows Cyrillic Cp1252 Windows Latin-1 Cp1253 Windows Greek Cp1254 Windows Turkish Cp1255 Windows Hebrew Cp1256 Windows Arabic Cp1257 Windows Baltic Cp1258 Windows Vietnamese ISO8859_1 ISO 8859-1, Latin alphabet No. 1 ISO8859_2 ISO 8859-2, Latin alphabet No. 2 ISO8859_3 ISO 8859-3, Latin alphabet No. 3 ISO8859_4 ISO 8859-4, Latin alphabet No. 4 ISO8859_5 ISO 8859-5, Latin/Cyrillic alphabet ISO8859_6 ISO 8859-6, Latin/Arabic alphabet ISO8859_7 ISO 8859-7, Latin/Greek alphabet ISO8859_8 ISO 8859-8, Latin/Hebrew alphabet ISO8859_9 ISO 8859-9, Latin alphabet No. 5 ISO8859_13 ISO 8859-13, Latin alphabet No. 7 ISO8859_15_FDIS ISO 8859-15, Latin alphabet No. 9 KOI8_R KOI8-R, Russian MacArabic Macintosh Arabic MacCentralEurope Macintosh Latin-2 MacCroatian Macintosh Croatian MacCyrillic Macintosh Cyrillic MacDingbat Macintosh Dingbat MacGreek Macintosh Greek MacHebrew Macintosh Hebrew MacIceland Macintosh Iceland MacRoman Macintosh Roman MacRomania Macintosh Romania MacSymbol Macintosh Symbol MacThai Macintosh Thai MacTurkish Macintosh Turkish MacUkraine Macintosh Ukraine MS874 Windows Thai TIS620 TIS620, Thai Big5 Big5, Traditional Chinese Cp930 Japanese Katakana-Kanji mixed with 4370 UDC, superset of 5026 Cp933 Korean Mixed with 1880 UDC, superset of 5029 Cp935 Simplified Chinese Host mixed with 1880 UDC, superset of 5031 Cp937 Traditional Chinese Host miexed with 6204 UDC, superset of 5033 Cp939 Japanese Latin Kanji mixed with 4370 UDC, superset of 5035 Cp942 IBM OS/2 Japanese, superset of Cp932 Cp942C Variant of Cp942 Cp943 IBM OS/2 Japanese, superset of Cp932 and Shift-JIS Cp943C Variant of Cp943 Cp948 OS/2 Chinese (Taiwan) superset of 938 Cp949 PC Korean Cp949C Variant of Cp949 Cp950 PC Chinese (Hong Kong, Taiwan) Cp964 AIX Chinese (Taiwan) Cp970 AIX Korean Cp1381 IBM OS/2, DOS People's Republic of China (PRC) Cp1383 IBM AIX People's Republic of China (PRC) Cp33722 IBM-eucJP - Japanese (superset of 5050) EUC_CN GB2312, EUC encoding, Simplified Chinese EUC_JP JIS X 0201, 0208, 0212, EUC encoding, Japanese EUC_KR KS C 5601, EUC encoding, Korean EUC_TW CNS11643 (Plane 1-3), EUC encoding, Traditional Chinese GBK GBK, Simplified Chinese ISO2022CN ISO 2022 CN, Chinese (conversion to Unicode only) ISO2022CN_CNS CNS 11643 in ISO 2022 CN form, Traditional Chinese (conversion from Unicode only) ISO2022CN_GB GB 2312 in ISO 2022 CN form, Simplified Chinese (conversion from Unicode only) ISO2022JP JIS X 0201, 0208 in ISO 2022 form, Japanese ISO2022KR ISO 2022 KR, Korean JIS0201 JIS X 0201, Japanese JIS0208 JIS X 0208, Japanese JIS0212 JIS X 0212, Japanese JISAutoDetect Detects and converts from Shift-JIS, EUC-JP, ISO 2022 JP (conversion to Unicode only) Johab Johab, Korean MS932 Windows Japanese MS936 Windows Simplified Chinese MS949 Windows Korean MS950 Windows Traditional Chinese SJIS Shift-JIS, Japanese UnicodeBig Sixteen-bit Unicode Transformation Format, big-endian byte order, with byte-order mark UnicodeBigUnmarked Sixteen-bit Unicode Transformation Format, big-endian byte order UnicodeLittle Sixteen-bit Unicode Transformation Format, little-endian byte order, with byte-order mark UnicodeLittleUnmarked Sixteen-bit Unicode Transformation Format, little-endian byte order UTF8 Eight-bit Unicode Transformation Format UTF-16 Sixteen-bit Unicode Transformation Format, byte order specified by a mandatory initial byte-order mark $ doc/formatting.txt0100644000113000011300000002364407654006331013230 0ustar kripkrip* supported features As of version 0.8.6, openisis supports only a *very* limited subset of the formatting language. It should however be sufficient for the most important purposes: building indexes and basic preprocessing. - all kinds of literals (',",|) - all modes (P,D,H) - Vn[]^c[] field selector including repeated subfields - implicit and explicit loops The rest of this text describes sort of a merger of formatting features from both WinISIS and CISIS/wwwisis, which openisis may attempt to support one day. * formatting basics Formatting in openisis is separated into two tasks, that used to be mixed up in traditional ISIS software. - record processing In openisis, execution of a ("print"-)format actually transforms one or more records into one new ("output") record. It loops and selects fields, applies Mxx modes, REFs other records, but screen formatting directives are simply added as special fields. In terms of relational databases, a format defines a view. - screen rendering It is then the task of separate rendering engines to turn those printformat fields into well-indented plaintext or HTML or Postscript or TeX or Windows GDI commands or you name it. * elements of formatting A formatting expression is a series of literals and functions. Funtions use zero, one or more data items as parameters. Functions that expect one of the parameters to the left are called operators. * input types The expected type of a function parameter can be one of: - s string (auto-concatenated auto-stringified any) - n numeric expression (including string, boolean) - o output (auto-stringified, but not concatenated) - v variable or iterator - r row iterator (list of rowids) Immediatly after the function name, the following can be used: - c a single character - a alphanumeric bareword (identifier in the C language) - i integer literal - x anything * output types The output of an expression is zero, one or more fields, which can be string, numeric or value. The type of the last field determines which operators are recognized. Field tags are positive for values (i.e. the output of value iterators), zero for literals and negative in the range -1 .. -999 for printformats. Other negative tags are only temporary: value iterators are finally evaluated to emit values, conditional literals and numbers are later eliminated or changed to zero, resulting in a literal. * context during processing, we have the following context: o output record (not changed) r input record/s, db and loop; changed by REF, LR and loops f format, mode and variables; changed by @, Mcc x frame: function, signature, parameter type and position new frames are opened for functions, operators and blocks * show stoppers: ( ) , .. ] During processing the format string is scanned left to right and fields are appended to the ouput record as encountered. Function "calls" can be explicit (parameter list is enclosed in parentheses), or implicit, taking only one parameter (operators and literals like V24, X3). An opening parentheses following an operator or literal function makes the frame explicit. A parameter type of i or a is changed to n or s, resp. Otherwise a ( starts - in numeric context: an anonymous (n) arithmetic parentheses - in string context: an S(s_) function - in output context inside a loop: an indentation function - otherwise starts an explicit loop A closing parenthes closes whatever was opened by the last opening one. A comma ',', or range '..' closes a parameter, which closes an operator or saturated implicit loop. If there was no parameter to close, an empty string or 0 number is emitted, thus [..] is equivalent to [0..0]. * functions and type coercion Every function has a signature which denotes the types of the expected parameters. Functions and switches are put on the stack, the type expected from the following expressions is set according to their signature. Whenever a comma or the closing parentheses is encountered, the fields added for this parameter are converted to the expected type. This is not much work for numeric or boolean types, since expressions are evaluated while parsing. Where a string is expected, we might have a record (i.e. multiple strings), which is then collapsed as with the S function: all printformats are discarded and other strings are concatenated. When a function sees it's closing ')', it replaces the parameters pushed on the output record by the function value calculated from them. The return value is tagged with the first data tag encountered, even from records evaluating to empty. The signature is denoted by a string, where - the 1st char gives the left hand operand ('_' for none) - the 2nd char is a digit of required params or '_' for specials - the following chars give types of each parameter in turn - an trailing '_' denotes, that the last type may be repeated * record loops: "x" |x| + (s_) WHILE n (s_) CONTINUE BREAK OCC IOCC A loop is either started explicitly by a '(' in record context, or implicitly by a conditional literal or a V value iterator. An explicit loop is closed by the matching ')', while an implicit one ends when saturated (after first iterator), with a comma ',' or any other V iterator. The loop content is then repeatedly executed while incrementing the OCC counter from 1 (and possibly subject to a preceding WHILE). On each turn there is a "last" flag (initially true), which is cleared by any iterator that expects to have more fields (this may even be true if there was no OCCth field, e.g. if the OCCth field didn't have some subfield), consequently, if there are no iterators, the first turn is the last. Iterators also set or clear a "had" state (initially unknown). During each execution of the loop, we're then in string type context: - a "" conditional emits it's contents only on OCC=1 if before the first iterator, else on last=true - a || conditional emits it's contents on had!=false - a + undoes an immediatly preceding conditional on OCC==1 and sets had to false if last is true - an iterator adds the OCCth of (the selected) occurences (see below) - if an iterator has no OCCth occurence, an immediatly preceding || is undone and had is set to false - other tokens are processed normally * field iterator and operators: Vi Di Ni [n_] ^c A field clause is always evaluated in a loop over OCC. The iterator can be modified by the range and subfield operators. What are the occurences in question, depends on the selected options: field occ range, subfield selector and subfield range. The integer list n_ may be given as x..y, where a missing x is 1 and missing y or the keyword LAST means up to last occurence. If no range is specified, that means all like [..]. If a subfield is specified, that subfield is selected. If a subfield range is given, those occurences of the subfield are used: if no range is specified for the field, subfield occs are relative to the record, else relative to each individual field occ, so use V71[..]^a[1] to get the first occ of a in each occ of v71. With 71=^afoo^ax, 71=^abar^ay and 71=^abaz^az, V71^a[1..3] will give the first three total occurences of subfield a, i.e. foo, x and bar, whereas in CISIS it would give foo, bar and baz. * other values and operators on values: Ei Si := * operators on string: *n .n (n) (n,n) The * and . string operators modify the top (the last pushed field) by removing the first n chars or all but the first n chars, resp. An () indentation operator is a somewhat late indentation printformat, which exchanges itself with the previous field. * arithmetic operators on numbers: * / + - * relational operators on numbers: = <> < <= > >= * relational operators on strings: = <> < <= > >= : * boolean operators on numbers: AND NOT OR * literals: 123 123.45 'x' "x" |x| /*x*/ !cxc * switches: IF n THEN o ELSE o FI SELECT s CASE s: o ELSECASE o ENDSEL * format: @a MPL MPU MHL MHU MDL MDU * db: DB MSTNAME MFN[i] L(s) NPST(s) NPOST(s) LR(s) LR(s[,n,n]) REF(r,o) * xternal db: L->a(s) LR->a(s) NPOST->a(s) REF->a(r,o) * printformats # / % { } !cxc () QC QJ B I UL Ci Xi Fi FSi CLi NEWLINE(s) LW(n) PICT(s) M(n[,n]) TAB[i] BOX[i] NP[i] NC[i] BPICT(s[,n]) FONTS(a_) COLS(s_) LINK(s_) * other functions &a(s_) CAT(s) GETENV(s) PUTENV(s) SYSTEM(s) DATE[i] PROC(s) DATETIME DATEONLY VAL(s) RMAX(n_) RMIN(n_) RSUM(n_) RAVR(n_) LEFT(s,n) RIGHT(s,n) SS(n,n,s) MID(s,n,n) REPLACE(s,s,s) INSTR(s,s) SIZE(s) F(n) F(n,n) F(n,n,n) TYPE(s) TYPE(s,s) S(s_) LAST NOCC(s_) P(s_) A(s_) * unsupported syntax L([s]s) NPOST([s]s) REF([s]j,...) -- hopeless, use winisis notation * list of tokens by syntax - empty tokens stopper: , .. ) ] THEN ELSE FI CASE ELSECASE ENDSEL CONTINUE BREAK state: MPL MPU MHL MHU MDL MDU values: DB MSTNAME OCC IOCC # / % { } QC QJ B I UL DATETIME DATEONLY LAST - immediate literal i MFN[i] DATE[i] TAB[i] BOX[i] NP[i] NC[i] Ci Xi Fi FSi CLi Vi Di Ni Ei Si @a L->a(s) LR->a(s) NPOST->a(s) REF->a(r,o) ^c 'x' "x" |x| /*x*/ !cxc - syntax blocks (jump & run) o ( o ) WHILE n ( o ) REF( r, o ) IF n THEN o ELSE o FI SELECT s CASE s: o ELSECASE o ENDSEL - operators [n..n] ^c (n[,n]) := . * / + - = <> < <= > >= : AND OR NOT - all others are properly braced functions - ambiguous tokens S F might be Si, Fi or S(), F() / + * might be arithmetic or string operators or a newline = <> < <= > >= might compare strings or numbers := assigns number to Ei, string else ( opens one or the other frame ... * tokenizing & processing * read a token and literal - get (longest matching) token - if token accepts a literal, get the literal - if token accepts an opening (, get it - resolve S/F ambiguity syntactically depending on presence of i literal * process - if token is possibly an operator of higher precedence, check operator ambiguities, coerce field according to operators wishes and go opening the operator frame - else coerce field according to frame context - if we had a field in o-context or token is a stopper, close the parameter - if the frame is implicit and saturated or token is a frame closer, close the frame and start over processing - add the token or literal doc/fst.txt0100644000113000011300000000302207654006331011636 0ustar kripkripfst lines read "tag tec mode,fmt", where - tag is the field number used in the index posting - tec is the indexing technique (see below) - mode is a Mmc formatting mode or, for tec 5..8, a prefixing format - fmt is some format expression, typically a single field like v24 * formatting modes indicated by Mmc, where m is a character indicating some translations, and c is a character indicating the handling of case; default MPL m - P proof mode: no changes applied - H heading mode: angle brackets are removed, ^x is replaced as ';' for x=a, ',' for x=b..i, '.' for others - D data mode: like heading mode plus '. ' after each field c - L lower case: no changes applied - U upper case: (ASCII-)characters are converted to "uppercase" as listed in the file ISISUC.TAB (a textfile containing 256 replacement byte values like ... '096 065 066' ... in 16 lines of 16 numbers each) * indexing techniques - 0 one entry per line (format should produce appropriate newlines with /) - 1 one entry per line or subfield (use proof mode MPx to retain ^x) - 2 one entry per each term in (use proof mode) - 3 one entry per each term /between slashes/ - 4 one term per "word": a word is a sequence of 'alphabetical characters' as listed in the file ISISAC.TAB (a textfile containing numbers like ... '098 099 100 101' ...). "stopwords" (listed in the db's .stw-file) are ignored - 5..8 like 1..4, but instead of applying a mode Mmc, every entry with delimiter 'd' is given a prefix 'p' like 1 8 '/TI=/',v24 (prefix //-entries with TI=) doc/gizmo.txt0100664000113000011300000000421707654006331012200 0ustar kripkripfrom the Winisis "reference" manual: $ i. Gizmo conversion table This field may contain the name of a Gizmo conversion table to be used for import. This facility is provided to facilitate interchange of text containing diacritical marks between different hardware and/or software platforms. A gizmo conversion table is an ASCII text file, which you may create using any familiar text editor, and is defined as follows. The first character of the first line contains the diacritical prefix (normally the @ sign for CDS/ISIS). Each subsequent line has the following format: dxyxyxyxyxy....xy where : d is the diacritical code (e.g. 7 for acute accent), x is the character to be accented, and y is the accented character. A sample gizmo conversion table is given below: @ 6aàeè 7aáeé (([ ))] CDS/ISIS uses this table to provide encoded diacriticals on export (e.g. é will be converted to @7e), and to decode encoded diacriticals on import (e.g. @6a will be converted to à). When a Gizmo conversion table is used for import, CDS/ISIS will simply strip the diacritical for all those characters not listed in the table. For example, using the table given above, the input string @7e. will be converted to é, whereas the string @7y will be converted to y, because y is not listed in line 7. Note: this feature is not implemented in this version. $ at some (?) other source i found: $ a gizmo is a masterfile containing text replacements. gizmo records have fields with tags 1 and 2. every occurence of the contents of field 1 is replaced by the contents of the associated field 2. gizmo translation can be used - to perform charset translation, e.g. Cp850 to ISO-8859-1 - to translate special chars like <> with something nicer like / - to htmlify output like with < to < - to replace certain keywords that might be used in field contents if a gizmo record has a field 11, and it contains the three characters: - asc then field 1 is used as decimal number giving a char value - hex then field 1 is used as hex number (e.g. d7) giving a char value in the same way field 21 may control interpretation of field 2. $ doc/limits.txt0100644000113000011300000000107107654006331012345 0ustar kripkrip$ lengths in bits record length: 15 (32K) mfrl signed short in MST DOS version has 8K mfn: 24 bit (16M) 3 byte in IFP mfn long in MST db size(record position): 29 bit (512M) xrmfb/xrmfp 20+9 in XRF IFP occurence: 8 bit IFP word count: 16 bit #fields in FDT: 200 #lines in FST: 600 #stopwords: 799 #characters PFT: 10000 #characters PFT result: 64000 proposed limits: record length: 24 bits (16M) db size: 40 bit (1T) index size: 32+13 bit (4G*8K blocks = 32T) mfn: 32+8bit (256 databases per 4G records) this needs 64 bit xrf (40pos+24len) and 80 bit ifp $ doc/longidx.txt0100644000113000011300000000522507654006331012515 0ustar kripkrip With the emerging use of unicode in ISIS databases, the 30 byte maximum on keys might become a somewhat narrow limit for languages using 2 or 3 UTF-8 bytes per character while still having lots of characters per word. Where UTF-16/UCS-2 is used, all languages would be equally affected by an effective maximum key of 15 characters. Therefore, there is a clear need for a new index structure supporting longer keys for some environments. This would in itself not raise any compatibility issues, since an index is always redundant and can be reconstructed by any plattform in a suitable format for this plattform (of course according to the plattform's builtin limits). While it would be straightforward to just create some n03/l03 pair of files or change 01/02 parameters, and some tools might even care for the control-file contents and work without modification using such indexes, some other tools wouldn't. In this situation, one may want to spend some time thinking about how an index should be organised. For programmers having coped with ISAM and databases in theory and practice, the reasons for some design decisions of traditional ISIS index structures are less than obvious. * blank padding a typical index-block lookup routine does not benefit in speed from blank-padding. while the fixed-array organisation allows for very easy Pascal implementation (where you can't cast some part of a block to a struct), C code doesn't care. instead, the additional space accounts for additional IO and therefore slows down index access. considering that many DBs even allow for compressed index files, where every key is stored as a number of leading bytes it shares with it's predecessor followed by the trailing difference, one should expect some performance improvement from at least eliminating all those blanks. * block alignment the node and leaves files are made up of blocks not aligned at any page boundaries. Every B*tree in textbooks as well as in real-world databases uses page-aligned blocks for best IO performance. potential reader/writer concurrency problems (readers reading partially updated blocks) are also reduced for most systems when IO occurs at filesystem page boundaries. * multiple files The separation of one index structures nodes and leaves is due to the fact that they do have different block sizes. The obvious advantage of having two index file structures is to reduce the space wasted by blank padding. Using page-aligned blocks of fix size with unpadded keys would allow the complete index to go into one single file. Moreover, the postings could be stored together with the leaves, further reducing IO. --- $Id: longidx.txt,v 1.2 2002/11/13 18:52:55 kripke Exp $ doc/mx.txt0100664000113000011300000000552407654006331011501 0ustar kripkripmx overview running mx with no params yields: ========= CISIS Interface v3.33/GC/M/30000/10/30/I - Utility MX Copyright (c) BIREME/PAHO 1997. All rights reserved. mx [cipar=] [{mfrl|fmtl}=] {|{in|cgi|seq|iso[=]}=|tmp} [
this is the textbody
of the td node
Isi
at sourceforge

switchboard
downloads
CVS tree
OpenIsis
By now, OpenIsis is mainly used for web publishing of Isis databases. Web developers use the Java, Perl or PHP binding for this.

There also is a command line version which serves as a demonstration and test of the various features and can be used as a utility for data import/export and similar tasks.

A graphical, standalone user interface similar to UNESCO's WinIsis will be developed during 2003, after the Tcl/Tk binding is finished.
developing OpenIsis

For more info, please visit openisis.org
doc/threading.txt0100644000113000011300000000700707654006331013016 0ustar kripkrip the benefit of multithreading depends heavily on two factors: - whether the db is in the filesystem cache or needs real disk I/O - whether you use Java or pure C if there is real disk I/O, then multithreading gives a big performance enhancement: 8 threads outperform a single one by a factor of 4.2 (C) or 3.3 (Java) respectively, and C outperforms Java only by a factor of 1.3 (single-threaded) or 1.6 (8 threads). if there is sufficient memory available to cache the whole db, so that the process is mostly CPU bound, then 8 threads give only a 9% throughput advantage over one in C, and an 18% performance degradation (!) in Java. C outperforms Java by a factor of 50 (yes, 5000% percent !) with 8 threads or 38 single-threaded. ====== conclusion: For a mostly I/O-bound server, you might well use threaded Java, it is only about 40% slower than multithreaded C, but still beats single-threaded C by a factor of 2.5 . If, on the other hand, you want to go for real performance, get the memory needed and use C. As long as you are not concerned too much with fairness and there is no risk of some too lengthy operations, it may as well be single threaded. ====== the following numbers were derived without any thread interlocking (mutex) on a single processor some numbers from the C multithreaded crashtest (make crash) on the unesb db for prefix "Z" running on an 800MHz PIII with 1/4G RAM under Linux 2.4.13 409 terms 4205 postings max mfn 58880 * with lots of memory available for caching: $ sequential read 58880 rows in 0.618 seconds 95275 rows per sec 8 threads 82384 rows per sec 2 threads 79600 rows per sec 1 threads 75544 rows per sec 4 threads 81916 rows per sec real 0m3.040s user 0m1.860s sys 0m0.920s $ * with nearly all memory locked down by root: $ sequential read 58880 rows in 6.720 seconds 8761 rows per sec 8 threads 1199 rows per sec 2 threads 467 rows per sec 1 threads 288 rows per sec 4 threads 624 rows per sec real 5m6.422s user 0m1.940s sys 0m2.230s $ some numbers from the Java multithreaded crashtest (make jcrash) on the same db and machine 409 terms 4205 postings max mfn 58880 * with lots of memory available for caching: $ sequential read 58880 rows in 30.508 seconds 1929 rows per sec 8 threads 1645 rows per sec 2 threads 1836 rows per sec 1 threads 2013 rows per sec 4 threads 1785 rows per sec 129.00user 7.40system 2:20.80elapsed 96%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (4978major+7237minor)pagefaults 0swaps $ * with nearly all memory locked down by root: $ sequential read 58880 rows in 58.827 seconds 1000 rows per sec 8 threads 738 rows per sec 2 threads 380 rows per sec 1 threads 225 rows per sec 4 threads 558 rows per sec 123.68user 3.87system 8:16.98elapsed 25%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (5286major+7366minor)pagefaults 0swaps $ * on a 360MHz Mob. PII with 40MB and Linux 2.2.12, numbers differ even more: without disk activity, a single thread gives about 50% more throughput than 8 with lots of disk I/O on the other hand, 8 threads nearly quadruple the single thread throughput * the mobile now with an 2.4.18 kernel: $ enough memory: 8 threads 54558 rows per sec 4 threads 49822 rows per sec 2 threads 43575 rows per sec 1 threads 37656 rows per sec with mutex: 8 threads 49525 rows per sec 4 threads 45174 rows per sec 2 threads 41859 rows per sec 1 threads 36262 rows per sec no memory (mutex): 8 threads 696 rows per sec 4 threads 367 rows per sec 2 threads 258 rows per sec 1 threads 150 rows per sec $ doc/trad-fmt.txt0100664000113000011300000001400607654006331012566 0ustar kripkripthe status quo: =============== notations: UPPERCASE letters are verbatim a,b,x string literals c,d,e single characters f,g,h formats producing a single string value i,j,k,l,m,n,o numbers (formats processed in integer context) ll list of numbers (pushed by an expression) * mode Mcd c = P, H or D, d = U or L P proof mode: no changes applied H heading mode: ^x is replaced as ';' for x=a, ',' for x=b..i, '.' for others angle brackets are removed (>< replaced by '; '): gives b in index, a else gives nothing in index, a else D data mode: like heading mode plus '. ' after each field not followed by a suffix. no '.', if field ends in "punctuation" character. L lower case: no changes applied U upper case: characters are converted to "uppercase" as listed in the file ISISUC.TAB * field selector s Vn^c[i..j] Vn[i..j]^c select field n (0 = all), optional subfield c (* = first), optional range of occ i to j (j may be ommited or LAST), Dn^c select presence (optionally of subfield c) Nn^c select absence (optionally of subfield c) NOCC(s) total number of occurences of field s (NUMERIC) P(s) true, if s is present (BOOLEAN, see IF) A(s) true, if s is absent (BOOLEAN, see IF) ["x" a ["x" a ...]][|x|[+]]s[*o][.l][(k[,m])][[+]|x|]["x"] here [] are not literal, but indicate optional parts. x are strings not containing their respective delimiter a arbitrary formatting or mode commands conditionally executed repeated prefix or suffix in || output for any repetition, with + first prefix / last suffix is ommited, optional string offset o (0=start) and length l, optional indentation k,m (see formatting) * structural elements they do not themselves generate output, but rather control execution of formats u,v,w , separator (end some construct) /*...*/ comment 'a' unconditional literal @a, include format file a &a(f) call external function a (u) repeat group (over IOCC, only in record or string mode) BREAK break a repeat group or REF CONTINUE continue a repeat group SELECT f CASE g: u CASE h: v ... ENDSEL SELECT f CASE g: u CASE h: v ... ELSECASE w ENDSEL REF(m,u) apply format to row number m REF->a(ll,u) REF([f]ll,u) apply format to row number m in db En:=m assignment to numerical variable Sn:=(f) assignment to string variable IF t THEN u FI IF t THEN u ELSE v FI WHILE t (u) t a boolean expression using: fxg comparision, where x is one of = <> < <= > >= AND NOT OR () logical combination f:g true iff f contains g case insensitiv P(s) A(s) (see field selector) f format not empty * calculation (NUMERIC) calculation is performed on longs until a decimal point is seen or RAVR is used. statistical functions run over format giving a list of numbers, separated by ; or blank or any other non-numerical characters +-*/() arithmetic 123 integer literal 123.45 float literal VAL(f) numeric value of string RMAX(ll) maximum RMIN(ll) minimum RSUM(ll) sum RAVR(ll) average (always float) En numerical variable n * database functions OCC IOCC current occurence (1..) in repeat MFN current mfn (with 6 digits) (NUMERIC) MFN(n) current mfn (with n digits) DB MSTNAME current masterfile (possibly changed by REF) L(f) mfn of 1st posting for key f (NUMERIC) L->a(f) L([g]f) mfn of 1st posting for key f in db (NUMERIC) LR((f)) LR((f),m,n) list of mfns (postings m through n) NPOST(f) NPST(f) NPOST([g]f) number of postings for f (in db) (NUMERIC) * system functions CAT(f) contents of file f DATE date as yyyymmdd hhmmss w nnn (w day of week, nnn julian) DATE(DATETIME) dd/mm/yy hh:mm:ss DATE(DATEONLY) dd/mm/yy DATE(n) mm-dd-yy (n=2), hh:mm:ss (n=3) or both (n=1) GETENV(f) value of env variable f PUTENV(f) read f as x=y, set env var x to y SYSTEM(f) execute system command * string functions LEFT(f,n) substring (first n chars of f) RIGHT(f,n) substring (last n chars of f) SS(n,m,f) MID(f,n,m) substring (m chars starting at n) REPLACE(f,g,h) replace g in f with h F(f) F(f,n) F(f,n,m) number to string S(f)*n.m stringify format f, apply optional offset n, length m INSTR(f,g) pos of g in f (0 = none) (NUMERIC) SIZE(f) length of string f (NUMERIC) TYPE(f) A (if alphabetic), N (if numeric) or X (else) TYPE(?a?,f) 0 or 1, if f matches pattern .... ???? TYPE(n,f) 0 or 1, if f is alphanumeric (n=1), alpha(n=2), digits(n=3), integer, incl. sign(n=4), number, incl. E-notation(n=5) Sn string variable n * changing records PROC(f) apply changes resulting from f: D* delete all fields Dn delete all fields n Dn/m delete occ m of field n Ancac add string a (delim by any c) as new field n Hn m a add string a of length m as new field n * formatting # newline unconditional / newline if not on line start % reset blank lines { } open/close formatting scope Vn..(m) or Vn..(m,o) indentation in columns for one field selector NEWLINE(f) change newline to emit f (\r,\n are recognized in f) Cm mv to column m Xm m blanks M(n) M(n,m) indentation in twips (twentieth point = 1/1440 inch) TAB mv to some (?) position TAB(m) mv to position in twips LW(n) linewidth QC center QJ justify BOX BOX(n) box (with color n) NP NP(n) newpage (if less than n lines left) NC NC(n) new column (if less than n lines left) FONTS((a,b),...) setup fonts table, each font with family a and name b COLS((i,j,k),...) setup color table, each color with RGB-values i,j,k Fn select font n FSn select fontsize n B bold I italic UL underline CLn select color n !cac print ESC a PICT(f) picture BPICT(f) background picture BPICT((f),m) background picture in mode m LINK((f),g,...) link text f to command g, where g is: * link commands TEXTBOX f open box with text f TEXTBOXCHILD f same, as child of DB window TEXTBOXRCHILD f same, as child of record window TEXTBOX[[R]CHILD]LOAD f[,a,b] display format f, optionally changing a to b TEXTBOX[[R]CHILD]IMG f display img file f TEXTBOX[i]... with literal [] identify window number i OPENFILE f open according to filetype CMD f run external prog f GOTO m display record with mfn m LGOTO f display record with key f GOBACK ... :) FORMAT f[,a,b] switch format to f, optionally changing a to b BROWSE a[,m[,f]] browse DB a, mfn m, format f \,{ and } "must be escaped with a \" (???) doc/unirec.txt0100644000113000011300000000740607654006331012341 0ustar kripkrip The universal ISIS record Users of CDS/ISIS are accustomed to stuffing not only bibliographic, but all sorts of data into ISIS records. The 1994 edition of ANSI/NISO Z39.2 ("Information Interchange Format" alias ISO2709), after which ISIS records are modelled (more or less), contained a "reduction of references to 'bibliographic' data, because the standard is used for many other types". For example, CDS/ISIS uses a ISIS database to hold the various texts for language specific versions. Various control files like the syspar, FDT and FST are well suited for storage in ISIS records. Probably some implementations of CDS/ISIS internally use ISIS records to hold that data, but none seems to be able to read/write those to/from ISIS master files, ISO2709 or a choice of textformats, including XML. * nesting But then, there is more. I already mentioned that e-mail and many simple XML structures are conveniently stored in ISIS records. But even data which does not all that obviously follow a tag-value-list scheme can be fit easily into such a record. Consider what is happening when ASN.1/BER-encoded structures are sent down the wire to a Z39.50 server, for example structured Type-1 queries to locate bibliographic information: They are turned into a series of tags and values. "Twain OR Clemens" is sent as an "OR" field followed by two term fields valued Twain and Clemens. We can do that same serialization trick, of course: embedding one structure (i.e. record i.e. tag-value-list) into another by simple inserting the fields. That way we can achieve nesting of arbitrary depth no less than with XML. One problem that comes to mind: how do we tell the boundaries? - for structures with a fixed number of fields, like the "OR" node having two childs, boundaries are implicit. - the length (number of childs) may be given with the opening field. This may be inconvenient and/or error-prone, if not computed automatically. - a closing item like "" may be used, e.g. a reserved field tag. These approaches are now > Struct discussed in more detail. Based on such a schema, not only queries can be expressed and stored as records, but also formats, with proper nesting of IFs, loops and so on. This approach has a couple of advantages: - formats may be specified in any of a couple of external representations including XML - the variants of the CDS/ISIS formatting language with different names for the same functions can be supported using input filters - formats can be stored, retrieved and exchanged using standard means On the other hand, the formatting language could be augmented to support substructures. A straightforward and relatively easy to use and implement extension would be a PASCAL-style WITH r DO. The current OpenIsis bindings, especially Tcl as preferred formatting language, contain such support. * external representation Besides CDS/ISIS master files and ISO2709 files, there are a couple of text based formats suitable to store or exchange ISIS records. Most follow a name=value style and are using separators like '=', ':' and linebreaks, with different quoting rules. Among these are - RFC 822 emails - Java properties - Windows-style .INI files - character or tabulator separated values (tsv/csv) (think of the TAB as subfield delimiter) Then there are XML/HTML/SGML and finally freestyle languages, like the query or formatting language, where item boundaries are determined depending on context. Conversion would typically be based on one or more FDTs, mapping between names and numbers. Such a mapping, when used with formats, could also enable the use of symbolic field names like author instead of v24. The "plain" representation as preferrably used by OpenIsis is described in the papaer on > Serialized records. ----- $Id: unirec.txt,v 1.6 2003/04/07 13:12:43 kripke Exp $ doc/whatabout.txt0100644000113000011300000003232107654006331013044 0ustar kripkrip* what makes ISIS ISIS ? Andrew Giles-Peters raised the important question "What is it about ISIS that makes it ISIS?" So here are some thougts on this topic from the OpenIsis team: - As a database used for bibliographic data (among other), ISIS must be able to store and retrieve records as exchanged via ISO2709 efficiently and with no or minimal loss of information. - Besides the ability to retrieve records by number, ISIS must support an indexing mechanism which is essentially "function based", that is, index entries are not the immediate field values, but rather the values of a "view" derived by some computation are indexed. - ISIS must efficiently support typical query elements commonly used on bibliographic databases, like looking up a value without regard for the field or in several fields at once and specifying a distance within search terms should occur. Since these are minimal requirements, they would not stop anybody from adding tons of features on top. For example, it's relatively easy to store ISO2709 data in a relational database like Sybase (used by OCLC/Pica), each record covering several rows (mfn, field number, field occ, value), then compute a second similar table for the index and so on. However, there is the word "efficiently", which practically turns out to put some restrictions on the feature-load, especially when combined with: - ISIS must be widely usable even in the face of *very* low budgets. Therefore, not only the software itself must be available for at most a nominal fee, but it also must not require very new, very powerful or otherwise expensive hardware and system. Even very large catalogs should get by with moderate system costs. The OCLC/Pica system for example requires one to spend hundreds of thousands of dollars for powerful Sun machines. * end of story ? Still, it would be very nice if more areas of application could be explored for ISIS, both for the librarians in order to be able to use their favourite DB (i.e. ISIS) for a broader range of tasks and also to expand the user community, possibly leading to more support for everybody. One important question is whether ISIS needs some fundamental changes deep in it's guts, or whether it already has everything that's needed to build a broad range of sophisticated solutions on top of it. As you might expect, we are pretty well convinced of the latter. * file formats Just like it doesn't harm a database much to be exported to and imported from ISO2709, there is not much of a problem with different file formats, as long as there do exist conversion tools. As you know, CISIS/Unix-DBs are incompatible to WinIsis/DOS-DBs, but may be converted via ISO files. As long as the basic data structures are the same, lossless conversion is just a matter of tools. It's even less of a problem if the software itself can read several file formats (like openisis does). You won't care much whether your wordprocessor is reading a .doc or .rtf file, would you? We did an interesting and very successful study implementing an ISIS-like DB in pure Java using a plaintext masterfile very similar to the Mbox mailfolder format (hope to be able to release the code soon). Likewise there is no reason why one should not be able to read directly from an ISO2709 file. Besides convertible masterfile formats, one might well use other formats for xref and index, which always can be reconstructed as needed. There are several reasons like improved performance or robustness to do so. So I don't think ISIS is defined in terms of detailled file formats, but rather in terms of the basic data structures. One problem that might come to mind when talking about file formats are the limits. While the maximum number of records per DB as well as the maximum total file sizes are bypassed relatively easy by logically joining several databases, the maximum record size of about 32K is a limit which might be unacceptable for some applications. (Although it can partly be resolved by deploying external files like OCLC/Pica does to circumvent Sybase's varchar limits). Raising this limit would clearly restrict lossless conversion to one way, from small to large DB. Where a large DB model is needed, all parties developing ISIS software should agree on one format to allow for as-painless-as-possible interoperability. * so what kind of database is ISIS ? Classical database theory basically distinguishes ISAM, network, hierarchical and relational database systems. ISIS is strongly related to ISAM DBs, however it's flexible indexing is rarely paralleled by any of these systems and it's non-flat data model is targeted by hierarchical DBs only (in greater generality and with much higher costs). - Although direct joins by MFN shouldn't be too costly, ISIS is not the database of choice when several records typically need to be combined in queries or transactions. However, in many application cases, only one ISIS record is needed as opposed to several relational table rows. In such situations, ISIS is even an excellent and efficient transaction (OLTP) database (since save writing of an ISIS record is much simpler than other DB's undo/redo logs). - ISIS is not the database of choice when records are updated by the hour. However, where only about 10% of records are changed between two (monthly, weekly or daily) runs of backup and compactification, the space overhead is not a big problem. Where old versions of data need to be retained anyway (as often needed and supported, for example, by postgres history), you would hardly find a more efficient solution. - ISIS is not the database of choice when it comes to high volume online analytical processing (querying statistics on several dimensions, OLAP). However, after reading some database books and Oracle manuals, one learns that OLAP requires a well designed ("star schema") database separate from the transactional one, anyway. - ISIS does not, in itself, provide any concurrency control (actual implementations do, to some extend). This doesn't hurt when running a read-only multi-user catalogue, a stand-alone application and in some insert-only situations. For distributed multi-client update, there are mechanisms based on timestamps or stored procedures that need to be supported by some ISIS server to come. While these data models are strongly tied to the logical nature and physical organisation of the data, newer notions like that of an 'object oriented' or 'XML' database rather describe a way to use and access a database. Actually OO or XML DBs are usually based on one of the above mentioned systems (mostly relational ones). For the most part, using a DB as OO or XML storage does require nothing but some libraries and optionally precompilers for C++ or Java -- these can be build on top of existing ISIS without changing it, and ISIS will be an excellent choice for many applications. Some aspects of increased functionality and performance will require sort of "stored procedures" running inside the database. In the case of a XML DB they are used for example to decomposite structures, in the OO case they might need some sort of "magic switch" (method overriding) to perform differently for some records than for others. We believe that all this magic can be achieved based on ISIS. The concepts of an ISIS database server and a scripting language as an alternative to formatting exits are to be discussed elsewhere ... First we want to shed some more light on the great flexibility the ISIS database system has by it's very nature. * ISIS is a mail database Looking at http://www.faqs.org/rfcs/rfc822.html (or its updates) one will find many similarities between ISO2709 records and internet mails, which are, after all, essentially a series of header names and values. After assigning numbers to the 100 or 200 most commonly used headers and some sort of subfield encoding (e.g. "^nname^vvalue", "namevalue" or simply "name: value") to store other header lines with a special field number, mails are easily and very efficiently stored in an ISIS database. Given the enormous number of communication, groupware and workflow systems that are nowadays built upon standard plain internet mails (typically using a set of special mail headers), this is a very large area to be served by ISIS databases. The above mentioned Mbox-style implementation of ISIS tends towards that direction, building upon the javax.mail standard. IMAP mail servers could greatly benefit from the powerful indexing and retrieval system of ISIS databases. If also the mail sending application allows to select special headers from an entry form prepared by a skilled librarian with thesauri and systematics, an institution or company could really come to a new way of using mail as a system of qualified, living information. * ISIS is a multimedia database After all the mail not only has got headers, but also a body. A plaintext body of reasonable length (some KB, like sent by nice people), fits without problem in a field whose number means "body". A multipart body is easily decomposed to a series of body fields. Wether larger or non-plaintext bodies are stored within or outside the masterfile is a matter of the actual implementation and doesn't need to be discussed here, both approaches have their pros and cons. Anyway, the MIME standard, up and running since 1982, allows for storage and transmission of anything that uses bytes, and is easily integrated with ISIS databases (we partly did it, code to be released). * ISIS is a XML database Likewise XML, which basically is text, can be stored in an ISIS database (with respect to the implementation's maximum record length). Add some formatting exits to address the XML node content via a DOM-style a.b.c notation as used in javascript, use them in your FST and you will for sure have one of the world's best indexed and fastest XML database -- most others are using a relational DB as basis. So indexing, retrieving and displaying XML data is more or less simply a matter of some formatting functions. However, when thinking about data entry forms, for example, the dark side of the force shows up: Even with a very sophisticated database system with the ability to make sense out of XML DTDs, it is anyway potentially much more complicated. XML was meant to provide arbitrary complexity in the first place. And when it comes to DTDs like that of XHTML, which will carry just about the same content as any HTML page, one easily understands that reasonable automatic processing becomes nearly impossible -- that's the reason why HTML pages are largely beefed up with headers (Dublin Core and others). If you really desperately need it, it's good to have it, but else using it might be looking for trouble. When having to work with XML structures for one or the other reason, typically because they should be imported or exported, one should think of a mapping between XML and ISIS structures. In many situations XML structures are shallow and can be ISIfied by simply mapping the first level of sister nodes to ISIS fields and the second level to subfields (may require repeated subfield support). In other situations a closer look at the data structure may reveal that it is not well designed with regard to Ockham's razor but contains totally unnecessary depth which may be collapsed to the first case. Actually, during several years of work with XML structures as suggested by several "standards", I rarely found a reasonable structure which can not be mapped to a field-subfield-schema. But even if you really need XML structures "as is", they can be stored very > Struct efficiently in ISIS, with all the benefits of the flexible index (c.f. > unirec the universal ISIS record) . Anyway, Dublin Core metadata or other RDF (resource description framework) headers are conveniently stored in ISIS just like mail headers. Maybe, as this schema was created to suit the needs of the very old science of bibliographic knowledge management, much of that experience was built into it. On the other hand, XML's ancestor SGML was conceived for a document's body, not the head, and I guess there still is it's place in spite of programming industry's hype. The use of XML for structuring documents that are ment to be read by humans rather than machines of course is perfectly reasonable. Transparent access to file based data associated with a record and a XML add-on to the formatting language could aid in converting extracts of document contents to metadata accessible in the ISIS database and/or it's index. To wrap it up, I'd suggest to look at XML as an optional add-on to ISIS rather than an integral part. ISIS already has all the functionality needed to support any reasonable use of XML. ISIS data can much more efficiently contain XML structures than the other way round. * ISIS is a database for document/content management systems It follows that ISIS may very well support the needs of systems for XML documents or website content in XML or HTML. With increasing experience with such systems, people tend to understand that content metadata should be organized according to bibliographic principles. (Not that surprising, is it)? In cooperation with the oc4science.org there are projects at german universities to integrate publishing, document management and website CMS, based on an (Open)ISIS DB and directed by the librarian. ----------- $Id: whatabout.txt,v 1.8 2003/02/14 17:30:33 kripke Exp $ doc/writing.txt0100644000113000011300000000450507654006331012534 0ustar kripkripSince version 0.8.7 (October 2002) openisis is able to store new or updated records in the masterfile. Programmers may call the function openIsisWrite, see openisis.c for sample usage. From the command line, the openisis executable accepts a new -write option which specifies a database, where all retrieved records are written to. This works with all record selections like dump, search, scan, mfnlist and the -stream option for reading records from text files. examples: $ # dump (copy) the whole unesb database to cds openisis -db db/unesb/unesb -write db/cds/cds # copy only records found for azteca openisis -db db/unesb/unesb -search azteca -write db/cds/cds # add the syspar.par as new record to the cds db openisis -stream -write db/cds/cds $BASE/boot.log $ doc/xmlisis.txt0100644000113000011300000000631707654006331012544 0ustar kripkrip* some notes on the relation of XML and ISIS XML is in widespread use as a lingua franca for glueing software components together. Several tools for this can be found at xml.apache.org. What is missing here is an efficient, easy to use way of storing XML data. Only the most trivial cases are easily mapped onto the relational data model, which uses flat records, consisting of a fixed number of fields. The data structures modelled in XML typically have a variable number of childs. Hierarchical databases like ADABAS C are well suited and actually used by SoftwareAG in their Tamino XML DB, but aren't widely and freely available. * ISIS to XML ISIS records can be easily and canonically converted to XML. Anything up to the first subfield delimiter is the body, subfields are attributes (strictly XML-ish this is ok only for non-repeated subfields). Other special subdivisions of field content like the typical <key word> may split to real child nodes. The result (as generated by make pdemo) may look like: $ Educational Psychology universities Kenya Okatcha, F.M.M.O. Personal statement Symposium on the Psychological Bases of Programmed Learning $ Instead of tag numbers and subfield characters, symbolic names from the FDT may be used. * XML to ISIS XML data structures can be easily and efficiently mapped to the data model of ISO2709. The general conversion (based on a SAX parser) works as follows: - when encountering an opening tag, look up it's name in the FDT. If there is no FDT provided, create one on the fly. If the FDT does not contain the tag name, create a new entry using tag number max(100,1+highest tag in FDT). Create a field using the tag number found and field value '+'. - when encountering an attribute, look up it's name in the FDT (using either > Meta extended FDT entries or the FDT21 as of ISISMARC). Create a new subfield entry if needed using code 'a' or 1+highest code used (for this tag). Append a subfield using the code found. - When encountering a self-closer (the current field ends with />), change the starting '+' to '-'. - When encountering a text node, add a field using tag number 0 with the node's body as value. - When enceountering a closing tag, look up it's name as for opening tags, add a field with an empty value. For example look at RDF ( > http://www.w3.org/RDF , > http://archive.dstc.edu.au/RDU/reports/RDF-Idiot ). A structure like $ Dr Jacky J Crystal Director jacky@dstc.com.au Researcher $ canonically maps to $ 100 +^aResource 101 + 0 Dr Jacky J Crystal 101 102 + ... $ . If they had made an attribute what can be an attribute (not substructered, not repeatable) instead of a child, it would read (with explicitly assigned subfield codes) much more efficiently like $ 100 ^pResource^fDr Jacky J Crystal^tDirector^ejacky@dstc.com.au^rResearcher $ Also see > unirec and > Struct --- $Id: xmlisis.txt,v 1.6 2003/04/07 15:08:27 kripke Exp $ foo.oip0100755000113000011300000000051507635353345011050 0ustar kripkrip#!/opt/openisis/oip <%! int i; static void foo () { /* as an extension over common JSP-behaviour, a global %! block *ending* with !% introduces literal-printing code to be added outside of main */ !%> lala <%! } %> <% for ( i=0; i<3; i++ ) { %> bla <%=i%> blub <%="%f",i/3.%> <% } foo(); %> <%! int j; %> fulltext0100755000113000011300000000154707626500561011346 0ustar kripkrip#!/usr/bin/perl # read files for fulltext index # filelist is on stdin (e.g. by find) # index entries go to stdout # usage: # find /foo -name \*.html | ./fulltext >/tmp/idx 2>/tmp/mst # find /usr/share/doc -type f -a \! -name \*.htm\* | ./fulltext >/tmp/idx 2>/tmp/mst # sort -o /tmp/idx /tmp/idx # time ./openisis -write db/test/ft -stream -fmt mfn ) { chomp; $f = $_; next unless open( F, $f ); $fn++; print STDERR "100\t$f\n"; $line = 0; while ( ) { next if /^\s*$/; last if 255 < ++$line; chomp; $w = 0; for $word (split /\W+/) { next unless $word; printf "%s\t%d\t%d\t%d\t%d\n", uc($word), $fn, 800, $line, ++$w; } } print STDERR "\f\n"; } iAPI.txt0100755000113000011300000012055107355467174011107 0ustar kripkrip/* this file composed by Klaus Ripke from the iAPI files by Robert Janusz. this is the test app "ix.c" with all the includes slurped in, see // #include "isomething.c" you may run this with: cp iAPI.txt iAPI.c gcc iAPI.c ./a.out */ /* ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ iAPI software: The Example of ISIS (UNIX v.3) Database Handle Copyright (C) 2000 by Robert Janusz E-mail: rj@jezuici.krakow.pl Address: Robert Janusz, ul. Kopernika 26, 31-501 Krakow, Poland tel: (0048-12) 4294416/432 ; fax: (0048-12) 4295003. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */ // Linux gcc Compiler conventions ! #include #include #include #include // Global variables +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ static char DBpath [71]; char UpCase [256]; #define NFIND 3 // let's keep it for future #define PACKED __attribute__((packed)) // it shoud work on Linux gcc ;-) // #include "idef.c" /* ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ iAPI software: The Type Definitions for Unix-ISIS databases Copyright (C) 2000 by Robert Janusz E-mail: rj@jezuici.krakow.pl Address: Robert Janusz, ul. Kopernika 26, 31-501 Krakow, Poland tel: (0048-12) 4294416/432 ; fax: (0048-12) 4295003. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */ // gcc Linux Compiler conventions ! // *** ONLY UNIX ISIS = the difference between DOS/UNIX structures // *************************** TYPE DEFINITIONS ***************************** // Cross Reference File XRF structure +++++++++++++++++++++++++++++++++++++++ typedef /* Packed */ struct // /* Packed */ and order are necessary for block reading { long int xpos PACKED; // Block nr: 1, 2...; < 0: last block long int xrec [127] PACKED; // References to Master File records... // xrec [...] = 2048 *xrmfb [21 bit] & xrmfp [11 bit] // rec. deleted log. < 0 & > 0 (restore: -xrmfb) // rec. deleted phis. =-1 & = 0 (ref@MST.ctrl) // rec. does not exist = 0 & = 0 // new rec. to invert bit 1024 = 1 // rec. edited to inv. (add & del post.) bit 512 = 1 } xrfBlock; // Master File (MST) record structures... +++++++++++++++++++++++++++++++++++ typedef /* Packed */ struct // The first record in Master File { long int ctlmfn PACKED // Always 0 , nxtmfn PACKED // Next Master File Number , nxtmfb PACKED; // Next block in Master File, the 1st has 1 short int nxtmfp PACKED // Next position in the last block , mftype PACKED; // 0 for user's bases; 1 for messages long int reccnt PACKED // Reserved for future , mfcxx1 PACKED // Reserved for future , mfcxx2 PACKED // LAN, > 0: inversion, update not possible , mfcxx3 PACKED; // LAN, = 1: exists exclusive user } mstControl; typedef /* Packed */ struct // Fixed part of record header that must be... { // included in the same block; // the header can begin only on 0, 2, 4...498 long int mfn PACKED; // MFN number short int mfrl PACKED; // Length of record (allways eaven)... // LAN: < 0 ==> record blocked long int mfbwb PACKED; // Block Pointer | If the record waits for inv... short int mfbwp PACKED; // Position Pointer | ...ref. to old location // mfbwb, mfbwp: = 0 at creation; // after modif. = ref to the old record. // xrmfp bit 1024 = 1 at creation // (= waits to be inverted); // xrmfp bit 512 = 1 at edit // (=waits to be inv. & delete old references) } mstHdrFix; typedef /* Packed */ struct { short int base PACKED // Position of data fields in the record , nvf PACKED // Number of fields in the record , status PACKED; // 0 = active, 1 = deleted logicaly (- xrfb < 0) } mstHdrTxt; typedef /* Packed */ struct { mstHdrFix mhf PACKED; // This header is allways in MST file block mstHdrTxt mht PACKED; } RecHeaders; typedef /* Packed */ struct // Index for every field in the record { // dimension = 6 *nvf; base = 18 + 6 *nvf short int tag PACKED // Label of the field , pos PACKED // Position of the first char in data part: 0, 1... , len PACKED; // Length of the field } mstIndex; // Inverted File structures +++++++++++++++++++++++++++++++++++++++++++++++++ // Control File (CNT) structure.............................................. typedef /* Packed */ struct { short int idtype PACKED // B*tree type; 1 = N01/L01, 2 = N02/L02 , ordn PACKED // = 5; a half of size of key-record in N0x , ordf PACKED // = 5; a half of size of key-record in L0x , n PACKED // = 15; number of buffers for nodes , k PACKED // = 5; number of buffers for 1st index level (k < n) , lev PACKED; // Current number of levels in index; -1: no N0x long int posrx PACKED // Reference to the root in N0x , nmaxpos PACKED // Next record in N0x , fmaxpos PACKED; // Next record in L0x short int abnormal PACKED; // 0: N0x has only the root } cntBlock; // Index of Dictionary (N0x) structures...................................... typedef /* Packed */ struct { long int pos PACKED; // Number of record short int ock PACKED // Number of active keys in the record: 1, 2, 3...2*ordn , it PACKED; // B*tree type } nodeHeader; typedef /* Packed */ struct { char key [10] PACKED; long int ref PACKED; // > 0: reference to next node; // < 0: ref. to leaf: idx [1st].key = key // = 0: inactive } nodeIndex1; typedef /* Packed */ struct { char key [30] PACKED; long int ref PACKED; } nodeIndex2; typedef /* Packed */ struct { nodeHeader hdr PACKED; nodeIndex1 idx [10] PACKED; // Vector of 2*ordn elements } nodeBlock1; typedef /* Packed */ struct { nodeHeader hdr PACKED; nodeIndex2 idx [10] PACKED; // Vector of 2*ordn elements } nodeBlock2; typedef /* Packed */ union { nodeBlock1 nb1 PACKED; nodeBlock2 nb2 PACKED; } nodeBlock; // Leaf File (L0x) structures (all index terms).............................. typedef /* Packed */ struct { long int pos PACKED; // Number of leaf record: 1, 2... short int ock PACKED // Number of active keys in the record: 1..2*ordf , it PACKED; // = 1 for L01; = 2 for L02 long int ps PACKED; // Next record in the order: key [ock] < ps^ idx [1st].key } leafHeader; typedef /* Packed */ struct { char key [10] PACKED; short int xxx PACKED; // *** ONLY UNIX ISIS; long int infb PACKED // Reference to IFP segment *** != ISIS.MANUAL *** , infp PACKED; } leafKey1; typedef /* Packed */ struct { char key [30] PACKED; short int xxx PACKED; // *** ONLY UNIX ISIS; long int infb PACKED // Reference to IFP segment *** != ISIS.MANUAL *** , infp PACKED; } leafKey2; typedef /* Packed */ struct { leafHeader hdr PACKED; leafKey1 idx [10] PACKED; // 2*ordf keys } leafBlock1; typedef /* Packed */ struct { leafHeader hdr PACKED; leafKey2 idx [10] PACKED; // 2*ordf keys } leafBlock2; typedef /* Packed */ union { leafBlock1 lb1 PACKED; leafBlock2 lb2 PACKED; } leafBlock; // Inverted File Posting structures.......................................... // ordained postings organized in segments: Index --> Master File typedef /* Packed */ struct { long int ifblk PACKED // Number of block , ifrec [127] PACKED; // Vector of long int *** != ISIS.MANUAL *** // In 1st block ifrec [1st/2nd] = next pos. in IFP } ifpBlock; typedef /* Packed */ struct // This header and 1st posting - always in IFP block { long int ifpnxtb PACKED // Pointer to the next segment: 1, 2... | = 0: last seg. , ifpnxtp PACKED // 0, 1... | = 0: last seg. , ifptotp PACKED // Tot. number of postings (o.k. only in the 1st segment) // = Sum (segment=1, nseg; ifpsegp) , ifpsegp PACKED // Number of postings in this segment , ifpsegc PACKED; // Max. number of postings in this segment (<= 32768) } ifpHeader; // ifpHeader & 1st ifpPosting always in the same ifpBlock // Postings in Segment are always in chain of IFP blocks typedef /* Packed */ struct // Stored left-->right + 0 --> 8 char { long int pmfn PACKED; // First 3 bytes --> Master File short int ptag PACKED; // Field ident - according to File Selection Table (FST) char pocc PACKED; // Number of occurence of the field short int pcnt PACKED; // Counter of the termin inside the field } ifpPosting; typedef struct { int opened; // Which files are opened (bit spec.)... FILE *mst; // 1 FILE *xrf; // 2 FILE *ifp; // 4 FILE *n01; // 8 FILE *l01; // 16 FILE *n02; // 32 FILE *l02; // 64 mstControl mctrl; // MST ctrl record cntBlock cnt1, cnt2; // Inv. Files ctrl records long int ibm, ipm; // IFP new Segment Bock/Pos char name [72]; // Database name } BaseCtrl; typedef struct // This structer must be allways corelated with BaseCtrl { char st [31]; // Max. string to find short int v; // Field to find; 0: all fields short int it; // Type of B*Tree <== strlen (st) leafBlock lbl; // Block L0x with index terms short int ilk; // Position of term in lbl...idx char *term; // The term in lbl...idx long int ib, ip; // Number of block/position for IFP ifpBlock ibl; // Block IFP with posting/s ifpHeader ihd; // Header for Segment of postings long int iiseg; // Number of posting in current Segment (index) long int iitot; // Number of posting in total (index) long int iall; // Number of posting in all Segments char *ipst; // Ref. to last posting string [8 chars] cntBlock *cntr; // BaseCtrl.cntr1/2 <== it FILE *n0x; // N01 or N02 <== it, = BaseCtrl.n0x FILE *l0x; // L01 or L02 <== it, = BaseCtrl.l0x } TableQuery; #define RECLENGTH 8192 typedef union { char buf [RECLENGTH]; RecHeaders hdrs; } RecCtrl; // *** MAIN DATABASE FILE RELATIONS (MST <--> XRF) *** // 1 2 3 // MST |-mblk-|-mblk-|-mblk-|-mblk-|-mblk-|-mblk-|-mblk-|-mblk-|... // ctrl |----record----| // |--->| // 1 2 3 | pointer: B/P (P in [0, 2, 4... 498]) // XRF |-xblk-|-xblk-|-xblk-|... // MFN= 1... 128... 255... // Record |-->-pos->--| // |-mhf-mht-|-idx[1]...idx[nvf]-|---text...---| // |-------------- record buffer --------------| // *** INVERTED FILE RELATIONS (N0x <--> L0x <-->IFP) *** // N0x |-nbl-|-nbl-|-nbl-|-nbl-|-nbl-|-nbl-|... // -->-root->-| |-->--next-->--| |->-->-| r<0 // | // L0x |--lbl--|--lbl--|--lbl--|--lbl--|--lbl--|--lbl--|... // | // |-<--------------<-------------<-| r<0 // *** POSTING *** | // IFP |---ibl---|---ibl---|---ibl---|---ibl---|---ibl---|---ibl---|... // |next |-ihd-|---Segment---| |-ihd-|---Segment---| // |Seg. |-->------>------>-next->--| |||... // MFN... // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ // #include "idef.c" // Let's specify non int header ;-) just to keep order void ShowRec (RecCtrl *rc); void setQuery (BaseCtrl *bc, TableQuery *tq, char *sx, short int fld); void isis (); // #include "iutil.c" /* ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ iAPI software: The UTIL procedures for Unix-ISIS databases Copyright (C) 2000 by Robert Janusz E-mail: rj@jezuici.krakow.pl Address: Robert Janusz, ul. Kopernika 26, 31-501 Krakow, Poland tel: (0048-12) 4294416/432 ; fax: (0048-12) 4295003. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */ // gcc Linux Compiler conventions ! // *** ONLY UNIX ISIS = the difference between DOS/UNIX structures // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int ReadMem (void *ptr, FILE *f, long int nr, long int size) { // Read "nr"'s block of size "size" from file "f" into "ptr"... if (fseek (f, nr * size, SEEK_SET) != 0) return (-1); if (fread (ptr, size, 1, f) == 0) return (-2); return (0); // O.K. } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void doUpcase (char *s, char *st, int slen) { int i; for (i = 0; i < slen; i++) { if (i < strlen (st)) s [i] = UpCase [(int) st [i]]; else s [i] = ' '; } s [slen] = 0; } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int sComp (char *s1, char *s2, int len) { int i; for (i = 0; i < len; i++) { if (s1 [i] > s2 [i]) return (1); // .gt. if (s1 [i] < s2 [i]) return (-1); // .lt. } return (0); // .eq. } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int bfOpen (FILE **f, BaseCtrl *bc, char *ext, char *opt) { char st [81]; if (strlen (bc->name) + strlen (ext) > 80) return (-2); strcpy (st, bc->name); strcat (st, ext); if ((*f = fopen (st, opt)) == NULL) return (-1); return (0); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int InitBase (BaseCtrl *bc) // Error return: -1 base file/s; -2 inv. file/s; -3 mst. ctrl rec/s; // -4 cnt param/s; -5 ifp param/s { ifpBlock ib; bc->opened = 0; if (bfOpen (&bc->mst, bc, ".MST", "rb") != 0) return (-1); bc->opened += 1; // MST opened if (bfOpen (&bc->xrf, bc, ".XRF", "rb") != 0) return (-1); bc->opened += 2; // XRF opened if (bfOpen (&bc->ifp, bc, ".CNT", "rb") != 0) return (-2); if (fread (&bc->cnt1, sizeof (cntBlock), 1, bc->ifp) == 0 || fread (&bc->cnt2, sizeof (cntBlock), 1, bc->ifp) == 0) return (-2); if (fclose (bc->ifp) != 0) return (-3); if (bfOpen (&bc->ifp, bc, ".IFP", "rb") != 0) return (-2); bc->opened += 4; // IFP opened if (bfOpen (&bc->n01, bc, ".N01", "rb") != 0) return (-2); bc->opened += 8; // N01 opened if (bfOpen (&bc->l01, bc, ".L01", "rb") != 0) return (-2); bc->opened += 16; // L01 opened if (bfOpen (&bc->n02, bc, ".N02", "rb") != 0) return (-2); bc->opened += 32; // N02 opened if (bfOpen (&bc->l02, bc, ".L02", "rb") != 0) return (-2); bc->opened += 64; // L02 opened if (fread (&bc->mctrl, sizeof (mstControl), 1, bc->mst) == 0) return (-3); if (bc->cnt1.ordn != 5 || bc->cnt1.ordf != 5 || bc->cnt1.n != 15 || bc->cnt1.k != 5 || bc->cnt2.ordn != 5 || bc->cnt2.ordf != 5 || bc->cnt2.n != 15 || bc->cnt2.k != 5) return (-4); if (fread (&ib, sizeof (ifpBlock), 1, bc->ifp) == 0) return (-5); if (ib.ifblk != 1) return (-5); bc->ibm = ib.ifrec [0]; bc->ipm = ib.ifrec [1]; return (0); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void Done (BaseCtrl *bc) // Close the database files { if (bc->opened && 1) fclose (bc->mst); if (bc->opened && 2) fclose (bc->xrf); if (bc->opened && 4) fclose (bc->ifp); if (bc->opened && 8) fclose (bc->n01); if (bc->opened && 16) fclose (bc->l01); if (bc->opened && 32) fclose (bc->n02); if (bc->opened && 64) fclose (bc->l02); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ char UpMe (char c) // ^A = ^a etc. { if (c > 96) c = c - 32; return (c); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void outS (char *st) { char i; if (st == NULL) return; for (i = 0; i < strlen (st); i++) // c = tblconvert [st [i]]; if (st [i] == (char) 145) printf ("<"); // This is HTML < else if (st [i] == (char) 146) printf (">"); // and that is > else printf ("%c", st [i]); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void outC (char ch) { char st [2]; st [0] = ch; st [1] = 0; outS (st); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void outD (double x, int n) { char s [80]; gcvt (x, n, s); outS (s); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void outN (long int x) // long int -> sting @ base=10 { // char st [33]; // ltoa (x, st, 10); outS (st); printf ("%li", x); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ // #include "imasterf.c" /* ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ iAPI software: The Master File access to Unix-ISIS databases Copyright (C) 2000 by Robert Janusz E-mail: rj@jezuici.krakow.pl Address: Robert Janusz, ul. Kopernika 26, 31-501 Krakow, Poland tel: (0048-12) 4294416/432 ; fax: (0048-12) 4295003. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */ // gcc Linux Compiler conventions ! // *** ONLY UNIX ISIS = the difference between DOS/UNIX structures // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int GetRec (BaseCtrl *bc, long int mfnr, RecCtrl *rc, int bufvol) { char mblk [512]; // Master File block; long int xmfb, xmfp; // Pointer to Master File (last read) record // xMfB < 0 : record deleted (status = 1), -xMfB --> block; // xMfP = 0: deleted phisicaly // = 0 & = 0: does not exist // 1st block: xMfB = 1, xMfP = 0; pointer = xMfB * 2048 + xMfP // ==> mst < 500 MBy; long int xn, xb; int ret, nch; xrfBlock xblk; mstHdrFix *hf = &rc->hdrs.mhf; mstHdrTxt *ht = &rc->hdrs.mht; char *b = rc->buf; if ((mfnr < 1) || (mfnr >= bc->mctrl.nxtmfn)) return (-2); // No record // Find the address in XRF... xn = mfnr - 1; xb = xn / 127; if (ReadMem (&xblk, bc->xrf, xb, sizeof (xrfBlock)) != 0) return (-1); xn = xblk.xrec [xn % 127]; // = address in MST xmfb = xn / 2048; // Master File Block xmfp = xn % 512; // Master File Position // Find the record in MST... if (xmfb < 1) return (1); // Record deleted ret = 0; // Record o.k. if (ReadMem (&mblk, bc->mst, xmfb - 1, sizeof (mblk)) != 0) return (-1); // Copy headers into record buffer... memcpy (b, &mblk [xmfp], sizeof (mstHdrFix)); if (hf->mfn != mfnr) return (-2); // Database corrupted if (hf->mfrl >= bufvol) return (-3); // Buffer too small // Get the rest of the record xmfp += sizeof (mstHdrFix); // Posinion in block xb = sizeof (mstHdrFix); // Position in buf xn = labs (hf->mfrl) - xb; // Number of chars to read if (xn <= 0) return (4); // No fields ?? for ( ; ; ) // We have some chars { // Find number of chars to copy from this block if (xn <= (nch = sizeof (mblk) - xmfp)) nch = xn; memcpy (b + xb, &mblk [xmfp], nch); xb += nch; xn -= nch; xmfp = 0; if (xn <= 0) break; // Nothing more to read if (fread (&mblk, sizeof (mblk), 1, bc->mst) == 0) return (-1); } if (ht->status != 0) ret = 1; // Record deleted logicaly if (hf->mfbwb != 0 || hf->mfbwp != 0) ret = 2; // Record not inv. if (hf->mfrl < 0) ret = 3; // Record blocked return (ret); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int IsF (RecCtrl *rc, short int v, char ch, char *st) { short int n, k, kk, flag; char *b = rc->buf; mstHdrTxt *mht = &rc->hdrs.mht; mstIndex *ind = (mstIndex *) &rc->buf [sizeof (RecHeaders)]; flag = 0; for (n = 0; n < mht->nvf; n++) { if (ind->tag == v) // there is a tag { if (ch == ' ') // field found { flag = 1; break; } else // there is sub-field { flag = 1; kk = 0; k = mht->base + ind->pos; while (b [k] != '^' || b [k + 1] != ch) // convention ;-) { k += 1; kk += 1; if (kk >= ind->len) // out of field return (0); } break; } } ind += 1; // next index ref. } if (flag == 0) return (0); // false if (st == NULL) return (1); // true else if (sComp (&b [mht->base + ind->pos], st, ind->len) == 0) return (1); else return (0); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void RepF (RecCtrl *rc, char nf, ...) // Repeated "nf" fields in record "rc", the left label appears only if: // - begins with "\n", - there was another field and there is defined // filed; the right label appears only if ther was defined field. { va_list pinfo; // for ... parameter list int i, c, flag, n, k, kk, v; int hit = 0; char *sl, *sr; char *b = rc->buf; mstHdrTxt *mht = &rc->hdrs.mht; mstIndex *ind = (mstIndex *) &rc->buf [sizeof (RecHeaders)]; for (n = 0; n < mht->nvf; n++) { va_start (pinfo, nf); for (i = 0; i < nf; i++) { kk = 0; k = mht->base + ind->pos; sl = va_arg (pinfo, char *); // left v = va_arg (pinfo, short int); // field c = va_arg (pinfo, char); // sub-field c = UpMe (c); sr = va_arg (pinfo, char *); // right if (v != ind->tag) // another Tag break; if (c != ' ') // sub-field { flag = 1; while (b [k] != '^' || UpMe (b [k + 1]) != c) // convention ;-) { k += 1; kk += 1; if (kk >= ind->len) { flag = 0; break; } } if (flag) { k += 2; kk += 2; } } else flag = 1; if (flag) { // Reference to NULL = NW error !!! if (sl == NULL) flag = 0; else if (*sl == '\n') flag = 1; else flag = 0; if (hit != 0 || flag) outS (sl); hit += 1; if (c == ' ') while (kk < ind->len) { if (b [k] != '<' && b [k] != '>') // HTML special ;-) outC (b [k]); k += 1; kk += 1; } else while (kk < ind->len && b [k] != '^') // convention ;-) { if (b [k] != '<' && b [k] != '>') // HTML special ;-) outC (b [k]); k += 1; kk += 1; } outS (sr); } } va_end (pinfo); ind += 1; } } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ // #include "iinvterm.c" /* ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ iAPI software: The Inv. File Terms access for Unix-ISIS databases Copyright (C) 2000 by Robert Janusz E-mail: rj@jezuici.krakow.pl Address: Robert Janusz, ul. Kopernika 26, 31-501 Krakow, Poland tel: (0048-12) 4294416/432 ; fax: (0048-12) 4295003. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */ // gcc Linux Compiler conventions ! // *** ONLY UNIX ISIS = the difference between DOS/UNIX structures // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int FindFirst (TableQuery *tq) { short int i, j // Indexes for idx [] , srec, klen, sidx; // Size (rec), key length, size (idx) long int r; // Reference (record number); < 0: --> L0x --> IFP char *ref; leafHeader *lhdr = &tq->lbl.lb1.hdr; // = ...lb2.hdr nodeBlock nbl; nodeHeader *nhdr = &nbl.nb1.hdr; // = ...nb2.hdr if (tq->it == 1) { klen = 10; srec = sizeof (nodeBlock1); sidx = sizeof (nodeIndex1); } else { klen = 30; srec = sizeof (nodeBlock2); sidx = sizeof (nodeIndex2); } r = tq->cntr->posrx; // Root in N0x while (r > 0L) { if (ReadMem (&nbl, tq->n0x, r - 1, srec) != 0) return (-1); j = nhdr->ock; if (nhdr->pos != r || nhdr->it != tq->it || j < 1 || j > 10) return (-3); // 10 = 2*ordn ref = (char *) &nbl.nb1.idx; // Reference to first elem. of index ref += sidx; // The first key is allways <= st, skip it for (i = 1; i < j; i++) if (sComp (ref, tq->st, klen) > 0) break; else ref += sidx; ref -= sidx; // Correct value r = *(long int *) (ref + klen); // We have the "ref" value if (r > tq->cntr->nmaxpos) return (-3); // Out of N0x } // Now find the term in L0x index if (tq->it == 1) { srec = sizeof (leafBlock1); sidx = sizeof(leafKey1); } else { srec = sizeof(leafBlock2); sidx = sizeof(leafKey2); } r = - r; if (ReadMem (&tq->lbl, tq->l0x, r - 1, srec) != 0) return (-1); j = lhdr->ock; if (lhdr->pos != r || lhdr->it != tq->it || j < 1 || j > 10) return (-2); // 10 = 2*ordf tq->ilk = 0; ref = (char *) &tq->lbl.lb1.idx; // Reference to first elem. of index ref += sidx; // The first key is allways <= s, skip it for (i = 1; i < j; i++) if (sComp (ref, tq->st, klen) > 0) break; else { ref += sidx; tq->ilk += 1; } ref -= sidx; tq->term = ref; // We have the term address if (tq->it == 1) ref += 12; // += 10; *** NOT UNIX else ref += 32; // += 30; *** NOT UNIX tq->ib = *(long int *) ref; // = infb ref += sizeof (long int); tq->ip = *(long int *) ref; // = infp return (0); // O.K. } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int FindNext (TableQuery *tq) { int srec; // Block size long int r; // Reference (record number); < 0: --> L0x --> IFP leafHeader *lhdr = &tq->lbl.lb1.hdr; // = ...lb2.hdr if (tq->it == 1) srec = sizeof (leafBlock1); else srec = sizeof (leafBlock2); tq->ilk += 1; // Next key in the index if (tq->ilk >= lhdr->ock) // Out of index { r = lhdr->ps; // Next L0x record in order if (r <= 0) { tq->ilk -= 1; return (1); } // Nothing more, end of index if (ReadMem (&tq->lbl, tq->l0x, r - 1, srec) != 0) return (-1); if (lhdr->pos != r || lhdr->it != tq->it || lhdr->ock < 1 || lhdr->ock > 10) return (-2); // 10 = 2*ordf tq->ilk = 0; } srec = tq->ilk; if (tq->it == 1) { tq->term = tq->lbl.lb1.idx [srec].key; tq->ib = tq->lbl.lb1.idx [srec].infb; tq->ip = tq->lbl.lb1.idx [srec].infp; } else { tq->term = tq->lbl.lb2.idx [srec].key; tq->ib = tq->lbl.lb2.idx [srec].infb; tq->ip = tq->lbl.lb2.idx [srec].infp; } return (0); // We have the new key } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int Find (BaseCtrl *bc, TableQuery tabq [], int ifind, long int *npg , RecCtrl *rc, long int recvol) { int i; if (ifind <= 0 || ifind > NFIND) return (-1); for (i = 0; i < ifind; i++) if (FindFirst (&tabq [i]) < 0) return (-1); i = Postings (bc, tabq, ifind, npg, rc, recvol); return (i); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int IndexPrep (BaseCtrl *bc, TableQuery *tq1, TableQuery *tq2, char *st) { char sx [31]; // tqx->v contains the status of L0x index int i; if (st == NULL) for (i = 0; i < 30; i++) sx [i] = ' '; else for (i = 0; i < 30; i++) if (i < strlen (st)) sx [i] = st [i]; else sx [i] = ' '; sx [30] = 0; setQuery (bc, tq2, sx, 0); // We have a long term in tq2; v := 0 if (FindFirst (tq2) < 0) return (-1); sx [10] = 0; setQuery (bc, tq1, sx, 0); // We have a short term in tq1; v := 0 if (FindFirst (tq1) < 0) return (-1); while (sComp (tq1->term, tq1->st, 10) < 0) if ((tq1->v = FindNext (tq1)) != 0) break; // At the end or error if (tq1->v < 0) return (-1); while (sComp (tq2->term, tq2->st, 30) < 0) if ((tq2->v = FindNext (tq2)) != 0) break; // At the end or error if (tq2->v < 0) return (-1); return (IndexTest (tq1, tq2)); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int IndexTest (TableQuery *tq1, TableQuery *tq2) { if (tq1->v == 1) // At the end of 1st if (tq2->v == 1) // At the end of 2nd return (0); // All done else return (2); // We have only 2nd else if (tq2->v == 1) return (1); // We have only 1st else if (sComp (tq1->term, tq2->term, 10) <= 0) return (1); else return (2); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int IndexTerm (int i, TableQuery *tq1, TableQuery *tq2) { if (i == 1) tq1->v = FindNext (tq1); // The term was in 1st else if (i == 2) tq2->v = FindNext (tq2); // The term was in 2nd else return (-1); return (IndexTest (tq1, tq2)); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void setQuery (BaseCtrl *bc, TableQuery *tq, char *sx, short int fld) { // Must be activated AFTER the initialization of database tq->v = fld; if (strlen (sx) <= 10) { tq->it = 1; doUpcase (tq->st, sx, 10); tq->cntr = &bc->cnt1; tq->n0x = bc->n01; tq->l0x = bc->l01; } else { tq->it = 2; doUpcase (tq->st, sx, 30); tq->cntr = &bc->cnt2; tq->n0x = bc->n02; tq->l0x = bc->l02; } } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ // #include "ipost.c" /* ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ iAPI software: The Inv. File Postings access for Unix-ISIS databases Copyright (C) 2000 by Robert Janusz E-mail: rj@jezuici.krakow.pl Address: Robert Janusz, ul. Kopernika 26, 31-501 Krakow, Poland tel: (0048-12) 4294416/432 ; fax: (0048-12) 4295003. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */ // gcc Linux Compiler conventions ! // *** ONLY UNIX ISIS = the difference between DOS/UNIX structures // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int PostSeg (BaseCtrl *bc, TableQuery *tq) { if (ReadMem (&tq->ibl, bc->ifp, tq->ib - 1, sizeof (ifpBlock)) != 0) return (-1); if (labs (tq->ibl.ifblk) - tq->ib != 0) return (-2); memcpy (&tq->ihd, &tq->ibl.ifrec [tq->ip], sizeof (ifpHeader)); tq->ip += 3; // = long int.s / (header - posting) ***!= ISIS.MANUAL *** tq->iiseg = 0; return (0); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int PostNext (BaseCtrl *bc, TableQuery *tq) { if (tq->iitot >= tq->iall) return (1); // No more postings if (tq->iiseg >= tq->ihd.ifpsegp) // No more postings in this seg. { // Get next Segment... tq->ib = tq->ihd.ifpnxtb; tq->ip = tq->ihd.ifpnxtp; if (PostSeg (bc, tq) != 0) return (-2); } tq->iiseg += 1; tq->iitot += 1; tq->ip += 2; // = Number of long int / posting ***!= ISIS.MANUAL *** if (tq->ip > 125) // ***!= ISIS.MANUAL *** { // Get next block for this segment; we HAVE TO find the right block! tq->ib += 1; tq->ip = 0; if (ReadMem (&tq->ibl, bc->ifp, tq->ib - 1, sizeof (ifpBlock)) != 0) return (-1); if (labs (tq->ibl.ifblk) != tq->ib) return (-3); } tq->ipst = (char *) &tq->ibl.ifrec [tq->ip]; return (0); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void PostMake (long int mfn, short int tag, char occ, short int pcnt , char *s, char *slen) { ifpPosting rnp; char *p = (char *) &rnp; rnp.pmfn = mfn; rnp.ptag = tag; rnp.pocc = occ; rnp.pcnt = pcnt; if (tag == 0) *slen = 3; // Only MFN else *slen = 5; // MFN and TAG s [2] = p [0]; s [1] = p [1]; s [0] = p [2]; // = pmfn s [4] = p [4]; s [3] = p [5]; s [5] = p [6]; s [7] = p [7]; s [6] = p [8]; } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void PostInv (ifpPosting *ifpp, void *ps) { char *p = (char *) ifpp; char *s = (char *) ps; // Posting string: |0 a |1 b |2 c |3 d |4 e |5 f |6 g |7 h | ps *s // ifpPosting: |0 c|1 b |2 a |3 0 |4 e |5 d |6 f |7 h |8 g | ifpp *p // |pmfn |ptag |pocc|pcnt | p [0] = s [2]; p [1] = s [1]; p [2] = s [0]; p [3] = 0; // = pmfn p [4] = s [4]; p [5] = s [3]; p [6] = s [5]; p [7] = s [7]; p [8] = s [6]; } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ int main (int argc, char **argv) // char *argv [] { char c; printf ("ix \n"); printf (" ::= UNIX-ISIS (v.3) MF without .MST\n"); printf ("ix ver. 0.3, Copyright (C) 2000 by Robert Janusz\n"); printf ("ix comes with ABSOLUTELY NO WARRANTY\n"); printf ("License: GPL; see http://www.gnu.org/ for details\n"); for (c = 'A'; c <= 'z'; c++) // Let's prepare it, we use API, not an example UpCase [c] = UpMe (c); if (argv [1] != NULL) { strcpy (DBpath, argv [1]); // Database; no test ! ;-) it's only an example isis (); } return (0); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void isis () { int j, v, k; long int nrc; BaseCtrl bc; RecCtrl rc; char st [31]; TableQuery tabq [NFIND]; strcpy (bc.name, DBpath); InitBase (&bc); // ShowRec but not all... printf ("\n*** ShowRec ***\n\n"); //for (nrc = 1L; nrc < bc.mctrl.nxtmfn; nrc++) { for (nrc = 10L; nrc < 20L; nrc++) { if ((j = GetRec (&bc, nrc, &rc, RECLENGTH)) == 0) { ShowRec (&rc); } } // ShowIndex printf ("\n*** ShowIndex: %s ***\n\n", st); strcpy (st, "TEST"); // just an Inv. File term v = 0; // term counter k = IndexPrep (&bc, &tabq [0], &tabq [1], st); // prepare indexes while (k > 0) { j = 20 * k - 10; // We have the term length ;-) memcpy (&st, &tabq [k - 1].term [0], j); // keep it in st st [j] = 0; if ((v += 1) > 12) // we want 12 terms in a list { printf ("->%s\n", st); // the next term break; } printf ("%s\n", st); k = IndexTerm (k, &tabq [0], &tabq [1]); // Next term } // Find a term & show records printf ("\n*** Search for: %s ***\n\n", st); // Let's use the last term setQuery (&bc, &tabq [0], st, 0); // String to search in a filed 0 = any j = Find (&bc, tabq, 1, &nrc, &rc, RECLENGTH); // 1 term to search for // Clean all printf ("\nO.K.\n"); Done (&bc); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ void ShowRec (RecCtrl *rc) { long int nr; short int n, k, kk; char st [256]; char *b = rc->buf; mstHdrTxt *mht = &rc->hdrs.mht; mstIndex *ind = (mstIndex *) &rc->buf [sizeof (RecHeaders)]; nr = rc->hdrs.mhf.mfn; // Record header printf ("=== %li\n", nr); for (n = 0; n < mht->nvf; n++) { // Show fileds without formating printf ("*** [%li] ", (long int) ind->tag); k = mht->base + ind->pos; for (kk = 0; kk < ind->len; kk++) printf ("%c", b [k++]); printf ("\n"); ind += 1; // Next index } printf ("___\n"); } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ // This procedure is prepared to do a simple AND operator after some // extensions, so it can be localy a little strange ;-) int Postings (BaseCtrl *bc, TableQuery tabq [], int ifind, long int *npg , RecCtrl *rc, long int recvol) { int j, state; long int ipg, rn; ifpPosting ip; TableQuery *tq; if (ifind < 1) goto finish; // nothing to do ipg = 1L; // Index for page break { tq = &tabq [0]; // Let's think only about one term to handle (example) if (PostSeg (bc, tq) != 0) return (-1); tq->iall = tq->ihd.ifptotp; // Only 1st Segment has tot. post. number tq->iitot = 0; if (PostNext (bc, tq) != 0) return (-1); PostInv (&ip, tq->ipst); if (sComp (tq->st, tq->term, 20 * tq->it - 10) != 0) ipg = 0L; // the term not found } if (ipg != 1L) goto finish; for ( ; ; ) // Global loop... { // Is there a field { tq = &tabq [0]; if (tq->v != 0) while (ip.ptag != tq->v) { if (PostNext (bc, tq) > 0) goto finish; // No more postings PostInv (&ip, tq->ipst); } } rn = ip.pmfn; state = 1; if (ip.pmfn != rn) state = 0; if (state) // We have a record to show { if ((j = GetRec (bc, rn, rc, recvol)) == 0) ShowRec (rc); else printf ("Record state = %li\n", j); do // Shift all postings { tq = &tabq [0]; if (PostNext (bc, tq) > 0) goto finish; // No more postings PostInv (&ip, tq->ipst); } while (ip.pmfn <= rn); // MFN equals } else // state == 0, Shift all postings while (ip.pmfn < rn) // The last is to examin { tq = &tabq [0]; if (PostNext (bc, tq) > 0) goto finish; // No more postings PostInv (&ip, tq->ipst); } } finish: return (0); // The output is finished } // ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ idm.jsp0100644000113000011300000005141207576450563011046 0ustar kripkrip<%-- openisis - an open implementation of the ISIS database Version 0.8.x (microversion see file Version) Copyright (C) 2001-2002 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA This software is dedicated to the memory of Eckart Dietrich. This software is inspired by (but contains no code of) the iAPI Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. See iAPI.txt for what it contains. EOH --%><%@page import="org.openisis.*,java.io.IOException,java.util.*,java.net.URLEncoder" contentType="text/html; charset=iso-8859-1" %><%! static final String ID = "$Id: idm.jsp,v 1.4 2002/12/13 21:30:27 kripke Exp $"; static final String EMPTY = ""; static Db db; static String dberr = EMPTY; static String me = "idm"; static { try { String[] args = new String[] { "-v", "info", "-encoding", "Cp850" }; db = Db.open( "/var/db/idm/doku", args ); } catch (IOException e) { e.printStackTrace(); dberr = e.toString(); } }; static String NV ( String x ) { return null==x ? EMPTY : x; } static int ROWS = 10; // number of rows to show at once static int TERMS = 50; // number of terms to show at once static int[] tag = { 100, 101, 200, 201, 301, 302, 430, 435, 28, 450, 800, 18 }; static String[] t = { "Autor(en)", "author", // 100 "weitere Personen", "other contributors", // 101 "Körperschaft als Hrsg.", "corporate body as editor", // 200 "weitere Körperschaften", "other corporate bodies", // 201 "Titel ohne Autorennennung", "title (without indiv. author)", // 301 "Titel mit Autor", "title (with author)", // 302 "Erscheinungsvermerk", "publisher: name and place", // 430 "Erscheinungsjahr", "year of publication", // 435 "Erscheinungsland", "country of publication", // 28 "Ges.Titel/Ser.Titel/Bandnr.", "series title", // 450 "Schlagworte", "keywords", // 800 "Dokument Nr.", "no of document", // 18 // unused fields: "Universitätseinrichtung (10):", EMPTY, "Projekt (12):", EMPTY, "Erfassungsdatum (16):", EMPTY, "Dokumententyp (22):", EMPTY, "Physikalische Form (24):", EMPTY, "Sprache des Dokumentes (26):", EMPTY, "Physische Form (440):", EMPTY, "Quelle (Zeitschrift) (480):", EMPTY, "Quelle (Monographie) (490):", EMPTY, "ISBN (600):", EMPTY, "Abstract (900):", EMPTY, EMPTY, EMPTY, EMPTY, EMPTY, // 50: other texts "Dokumentations-Datenbank des IDM der Universität Bielefeld", "documentation database of the IDM at the university of Bielefeld", "Suchen", "search", "English", "Deutsch", //54 "?", "?", "weiter", "more", "allen Feldern", "all fields", //60 // the ops "und","and", "oder","or", "ohne","without", "zurücksetzen","reset", "Name","Surname",//70 "Vorname","Given name", "Institut","Institution", "Strasse","Street", "Ort","City", "Land","Country", // 80 "Postleitzahl","Zip code", "Bestellen", "order", "Abschicken", "Submit", "Einträge", "rows", "bis", "to", // 90 "von", "of", "für", "for", "Indexeinträge beginnend mit", "terms starting with ", "neue Suche", "new Query", // 100: hilfe "

Willkommen bei der Dokumentations-Datenbank des\n" +" IDM der Universität Bielefeld

\n" +" Sie können bis zu drei Suchbegriffe verknüpfen.\n" +" Jeder Suchbegriff kann auf ein Feld eingeschränkt werden.
\n" +" Die ?-Funktion zeigt Ihnen alle Suchbegriffe,\n" +" die mit dem Feldinhalt beginnen, ggf. auf mehreren Seiten.\n" +" Diese Werte können per Klick übernommen werden.

\n" +" Mit 'Suchen' wird die Anfrage ausgelöst.\n" +" Es werden jeweils "+ROWS+" Sätze angezeigt.

\n" +" Kopien von Titeln können gegen Kostenerstattung bestellt werden.
\n" +"

Suchtipps

\n" +" Für Prefixsuche beenden Sie den Begriff mit * oder $.
\n" +" Die Suche wird auf jeweils 1000 Treffer eingeschränkt.\n" +" Bei sehr häufigen Suchbegriffen, etwa 'a$', gibt es in jedem Satz\n" +" mehrere Treffer, sodass weniger als 1000 Sätze gefunden werden.
\n" +" \n" +" Eine engere Suche, etwa durch Einschränkung auf ein Feld,\n" +" kann in einem solchen Fall u.U. mehr Sätze finden!
\n" +" Verwenden Sie bei mehreren Begriffen den präzisesten zuerst.
\n" +" \n" +"

Cookies

\n" +" Die Funktion ist stark beeinträchtigt,\n" +" falls Sie das Session-Cookie ablehnen.
\n" +" Ein Session-Cookie wird nicht auf Ihrer Festplatte gespeichert,\n" +" sondern verschwindet beim Schließen des Browsers.\n" +" \n" , // 101: help "

welcome to the documentation database of the" +" IDM at the university of Bielefeld

" +" Up to three search terms may be combined.\n" +" Each term can be restricted to a specific database field.
\n" +" Use the ? beneath a field to display all entries in the index\n" +" starting with the currently entered term.\n" +" Click them for use in your query.

\n" +" The 'search' button starts the query.\n" +" Up to "+ROWS+" records are displayed at once.

\n" +" You can order copies of each article for a nominal fee.
\n" +"

search tips

\n" +" Use a trailing * or $ to search by prefix.
\n" +" The query will be restricted to 1000 hits.\n" +" Very frequent terms like 'a$' will have several hits within\n" +" each record, giving less than 1000 records total.
\n" +" Restricting the query, e.g. by choosing a database field,\n" +" may thus find more records!
\n" +" When combining multiple terms, use the most specific one first\n" +" for best results.
\n" +" \n" +"

Cookies

\n" +" This page will not work properly,\n" +" if you do not accept the session cookie.
\n" +" This is not stored on your hard disc,\n" +" but deleted after you quit your browser.\n" , "entfernen", "remove", "E-Mail", "e mail", "Telefon", "phone", "bitte füllen Sie die markierten Felder aus", "please fill in all marked fields", // 110 "Ihre Bestellung wurde abgeschickt. Vielen Dank", "Your order has been submitted. Thank you", "Ihre Bestellung konnte leider nicht abgeschickt werden." +" Bitte versuchen Sie es später", "Your order could not be submitted. Please try again later", }; static final String[] ORDER = { "Xlastname", "firstname", "institution", "Xstreet", "Xcity", "country", "Xzip", "mail", "phone", }; static final int[] ORDERT = { 70,72,74,76, 78,80,82,104,106 }; // check button value independet of language used static boolean chkbutton ( int n, String param ) { return t[n].equals( param ) || t[n+1].equals( param ); } static int chkint ( int min, int max, int def, String param ) { if ( null != param ) try { int j = Integer.parseInt(param); if ( min <= j && j <= max ) return j; } catch (Exception e) { } return def; } static final int OP_AND = 0; static final int OP_OR = 1; static final int OP_NOT = 2; static final char[] OP = { '*', '+', '^' }; static final String ORDERMAIL = "idmdoku@openisis.org"; static final String[] MAILHEADERS = { Mail.TO, ORDERMAIL, Mail.FROM, "Online-Katalog ", Mail.SUBJECT, "online Bestellung", }; static final int LINES = 3; static class Ses { int l; // lang, 0 or 1 int off = 0; int rows = ROWS; int view; // last seen view int idx; // field to save index into int msec; // String qry; int[] set; int len; // form content: String[] t = new String[LINES]; String[] o = new String[ORDER.length]; // order data int[] f = new int[LINES]; // 0 or 1+index of selected tag int[] op = new int[LINES]; Iterator i; String term = EMPTY; ArrayList orders; public Ses () { for ( int i=0; i<% Ses s; try { s = (Ses) session.getAttribute( me ); } catch ( Exception e ) { s = null; } if ( null == s ) session.setAttribute( me, s = new Ses() ); int task = DO_HELP; String order = null; // figure out, what to do String wants = request.getParameter( "do" ); if ( chkbutton( 52, wants ) ) task = DO_QUERY; else if ( chkbutton( 54, wants ) ) task = DO_LANG; else if ( "order".equals( wants ) ) task = DO_ORDER; else if ( chkbutton( 86, wants ) ) task = DO_ORDERS; else if ( "more".equals( wants ) ) task = DO_INDEXN; else if ( "new".equals( wants ) ) task = DO_BLANK; else if ( null != request.getParameter( "off" ) ) task = DO_QUERYN; else // check "index" buttons i0, i1 ... for ( int i=0; i" ) .append(t[112+s.l]) .append( "\n\n" ) .toString(); task = DO_ORDER; } } s.view = VW_ORDER; } int maxrow = s.off + s.rows; if ( maxrow >= s.len ) maxrow = s.len; %> <%=t[50+s.l]%>
service Universität Bielefeld
=================================================
 <%=t[50+s.l]%>
<% if ( DO_ORDER == task ) { %> <% for ( int i=0; i <% } for(int i=0;i <% } %>
<%=lbl%> <%=req%>
<%=Field.html(o[0])%> <%=Field.html(o[1])%>  (<%=t[102+s.l]%>)
-------------------------------------------------
<% } else { %> <% for ( int l=0; l <% } // for lines %>
<% if ( 0 < l ) for ( int o=0; o value="<%=o%>"/> <%=t[62+2*o+s.l] %> <% } // for ops %> in
-------------------------------------------------
<%-- rows at once --%> <% } // query form if ( VW_QUERY == s.view && 0 < s.len ) { if ( 0 < s.off ) { int boff = s.off-s.rows; if ( boff < 0 ) boff = 0; %>    <<<%=t[88+s.l]%> <%=boff+1%> <%=t[90+s.l]%> <%=boff+s.rows%> <% } if ( maxrow < s.len ) { int nmax = maxrow+s.rows; if ( nmax > s.len ) nmax = s.len; %>    <%=t[88+s.l]%> <%=maxrow+1%> <%=t[90+s.l]%> <%=nmax%>>> <% } } %>   <%=t[98+s.l]%>
<% switch ( s.view ) { default: case VW_HELP: %> <%=t[100+s.l]%> <% break; case VW_QUERY: if ( 0 == s.len ) { %> 0 <%=t[88+s.l]%> <%=t[94+s.l]%> '<%=s.qry%>' <% } else { %>

<%=t[88+s.l]%> <%=s.off+1%> <%=t[90+s.l]%> <%=maxrow%> <%=t[92+s.l]%> <%=s.len%> <%=t[94+s.l]%> '<%=s.qry%>'

<%-- start output --%> <% for ( int i=s.off; i <% for ( int j=0; j") ); } String ti = r.v("",301,Db.MDL, 0, 0,""); if ( null == ti || 0 == ti.length() ) ti = r.v("",302,Db.MDL, 0, 0,""); String link = me+"?do=order" + "&id=" + URLEncoder.encode( r.v("",18,Db.MDL, 0, 0,"") ) + "&ti=" + URLEncoder.encode( ti ); %> <% } %>
"+t[2*j+s.l] +"", null, // cond,repeated prefix tag[j], Db.MDL|Db.PS|Db.HTI, 0, 0, // tag and mode, len, off ";
","
>> <%=t[84+s.l]%>
<% } break; // case VW_QUERY case VW_INDEX: %>

<%=t[96+s.l]%> '<%=Field.html(s.term)%>'

<% if ( null == s.i || ! s.i.hasNext() ) { } else { int i = TERMS; while ( 0 < i-- && s.i.hasNext() ) { String ix = (String)s.i.next(); if ( 0 == ix.length() ) continue; String u = URLEncoder.encode( ix ); %> <%=Field.html(ix)%>
<% } if ( 0 > i ) { // more %>

>> <%=t[58+s.l]%>

<% } else s.i = null; } break; // case VW_INDEX case VW_ORDER: %> 
<%= order %>
<% } // switch %> img/Makefile0100644000113000011300000000221507644412745011764 0ustar kripkrip# #/* # openisis - an open implementation of the CDS/ISIS database # Version 0.8.x (patchlevel see file Version) # Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # see README for more information #EOH */ # # # $Id: Makefile,v 1.2 2003/04/08 00:20:53 kripke Exp $ # img Makefile # all: logo.gif isi.gif favicon.ico favicon.ico: isi.pgm ppmtowinicon $< >$@ %.gif: %.pgm ppmtogif $< >$@ %.ico: %.pgm ppmtowinicon $< >$@ img/isi.cursor0100664000113000011300000000050107615301370012332 0ustar kripkrip#define isi_width 16 #define isi_height 16 #define isi_x_hot 1 #define isi_y_hot 5 static unsigned char isi_bits[] = { 0x00, 0x00, 0x00, 0x18, 0x80, 0x25, 0xc0, 0x27, 0xe0, 0x47, 0xfe, 0x27, 0xe4, 0x27, 0xc4, 0x23, 0x08, 0x10, 0x10, 0x08, 0xe0, 0x07, 0xc0, 0x03, 0xc0, 0x03, 0xc0, 0x03, 0xe0, 0x07, 0x00, 0x00}; img/isi.pgm0100664000113000011300000000167407615275612011625 0ustar kripkripP2 16 16 255 255 255 255 255 255 255 255 255 255 255 255 237 151 255 255 255 255 255 255 255 255 255 255 255 255 255 221 110 76 159 255 255 255 255 255 255 255 255 169 58 69 204 103 203 245 84 242 255 255 255 255 255 255 178 7 0 0 30 111 203 255 104 200 255 255 255 255 138 255 75 0 0 0 0 88 154 255 134 181 255 255 255 132 183 106 73 0 0 0 11 109 170 255 115 191 255 255 133 133 255 213 105 68 14 19 106 109 230 252 86 229 255 255 240 84 221 255 178 101 90 91 106 195 255 154 118 255 255 255 255 181 100 240 255 237 184 187 248 255 203 96 233 255 255 255 255 255 176 84 185 255 255 255 252 146 85 220 255 255 255 255 255 255 255 210 55 43 79 72 30 98 240 255 255 255 255 255 255 255 255 255 187 0 0 0 0 188 255 255 255 255 255 255 255 255 255 255 187 0 0 0 2 178 255 255 255 255 255 255 255 255 255 255 187 0 0 0 2 187 255 255 255 255 255 255 255 255 255 236 0 0 0 0 0 0 252 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 img/logo.pgm0100664000113000011300000010765107615303427011777 0ustar kripkripP2 100 100 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 112 158 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 158 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 0 0 158 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 0 0 0 15 207 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 15 191 46 0 0 46 239 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 15 207 255 223 15 0 0 112 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 15 207 255 255 255 191 0 0 0 207 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 15 207 255 255 255 255 255 112 0 0 64 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 15 207 255 255 255 255 255 255 239 15 0 0 191 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 15 207 255 255 255 255 255 255 255 255 158 0 0 64 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 15 207 255 255 255 255 255 255 255 255 255 255 32 0 0 207 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 239 158 112 64 64 64 64 112 174 255 255 255 255 255 112 0 0 0 207 255 255 255 255 255 255 255 255 255 255 255 158 0 0 112 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 239 127 15 0 0 0 0 0 0 0 0 32 127 255 255 255 158 0 0 0 146 255 255 255 255 255 255 255 255 255 255 255 239 15 0 15 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 207 32 0 0 0 0 0 0 0 0 0 0 0 0 46 207 255 255 158 0 0 0 223 255 255 255 255 255 255 255 255 255 255 255 96 0 0 191 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 158 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 207 255 255 127 0 0 64 255 255 255 255 255 255 255 255 255 255 255 174 0 0 112 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 207 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 223 255 255 32 0 0 191 255 255 255 255 255 255 255 255 255 255 255 0 0 46 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 239 32 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 64 255 255 174 0 0 96 255 255 255 255 255 255 255 255 255 255 255 64 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 127 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 191 255 255 32 0 0 239 255 255 255 255 255 255 255 255 255 255 127 0 0 191 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 32 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 64 255 255 127 0 0 174 255 255 255 255 255 255 255 255 255 255 158 0 0 158 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 191 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 239 255 191 0 0 127 255 255 255 255 255 255 255 255 255 255 191 0 0 127 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 127 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 174 255 255 0 0 64 255 255 255 255 255 255 255 255 255 255 223 0 0 127 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 127 255 255 32 0 64 255 255 255 255 255 255 255 255 255 255 255 0 0 84 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 158 127 127 127 127 127 127 127 127 127 127 127 127 127 127 127 158 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 127 255 255 64 0 64 255 255 255 255 255 255 255 255 255 255 255 0 0 64 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 64 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 127 255 255 64 0 64 255 255 255 255 255 255 255 255 255 255 255 0 0 64 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 64 255 255 127 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 146 255 255 0 0 64 255 255 255 255 255 255 255 255 255 255 255 0 0 112 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 255 255 255 255 255 255 255 255 255 255 255 64 0 15 255 255 158 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 207 255 223 0 0 96 255 255 255 255 255 255 255 255 255 255 191 0 0 127 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 112 0 0 223 255 255 255 255 255 255 255 255 255 255 112 0 0 207 255 223 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 255 255 158 0 0 146 255 255 255 255 255 255 255 255 255 255 191 0 0 127 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 127 0 0 191 255 255 255 255 255 255 255 255 255 255 174 0 0 112 255 255 84 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 127 255 255 84 0 0 207 255 255 255 255 255 255 255 255 255 255 127 0 0 191 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 191 0 0 127 255 255 255 255 255 255 255 255 255 255 239 15 0 32 255 255 191 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 239 255 223 0 0 32 255 255 255 255 255 255 255 255 255 255 255 84 0 0 223 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 239 0 0 84 255 255 255 255 255 255 255 255 255 255 255 96 0 0 146 255 255 112 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 158 255 255 96 0 0 146 255 255 255 255 255 255 255 255 255 255 255 32 0 15 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 32 0 0 255 255 255 255 255 255 255 255 255 255 255 223 0 0 15 239 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 96 255 255 207 0 0 15 239 255 255 255 255 255 255 255 255 255 255 223 0 0 84 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 112 0 0 174 255 255 255 255 255 255 255 255 255 255 255 112 0 0 64 255 255 239 84 0 0 0 0 0 0 0 0 0 0 0 0 0 0 112 255 255 239 32 0 0 146 255 255 255 255 255 255 255 255 255 255 255 146 0 0 158 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 174 0 0 96 255 255 255 255 255 255 255 255 255 255 255 239 32 0 0 96 255 255 255 146 15 0 0 0 0 0 0 0 0 0 0 32 174 255 255 239 46 0 0 64 255 255 255 255 255 255 255 255 255 255 255 255 32 0 0 223 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 32 0 0 239 255 255 255 255 255 255 255 255 255 255 255 207 15 0 0 96 239 255 255 239 146 46 0 0 0 0 0 0 46 158 255 255 255 239 46 0 0 15 239 255 255 255 255 255 255 255 255 255 255 255 191 0 0 64 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 127 0 0 127 255 255 255 255 255 255 255 255 255 255 255 255 174 0 0 0 46 207 255 255 255 255 239 191 191 191 191 255 255 255 255 255 191 32 0 0 15 207 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 158 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 223 0 0 15 239 255 255 255 255 255 255 255 255 255 255 255 255 207 15 0 0 0 96 223 255 255 255 255 255 255 255 255 255 255 191 64 0 0 0 32 207 255 255 255 255 255 255 255 255 255 255 255 255 223 0 0 32 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 127 255 255 255 255 255 255 255 255 255 255 255 255 255 223 64 0 0 0 0 46 127 191 223 255 255 207 191 112 32 0 0 0 0 84 239 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 158 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 223 15 0 0 223 255 255 255 255 255 255 255 255 255 255 255 255 255 255 158 32 0 0 0 0 0 0 0 0 0 0 0 0 0 0 32 174 255 255 255 255 255 255 255 255 255 255 255 255 255 255 174 0 0 32 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 127 0 0 64 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 158 84 0 0 0 0 0 0 0 0 0 0 84 174 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 239 15 0 0 174 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 46 0 0 127 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 239 191 146 127 127 127 127 146 191 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 84 0 0 64 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 207 0 0 0 174 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 127 0 0 15 239 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 146 0 0 15 207 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 158 0 0 0 174 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 96 0 0 15 207 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 158 0 0 0 146 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 15 207 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 158 0 0 0 96 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 239 46 0 0 0 158 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 112 0 0 0 96 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 239 96 0 0 0 96 239 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 223 64 0 0 0 96 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 112 0 0 0 32 174 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 146 15 0 0 0 158 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 174 15 0 0 0 64 191 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 158 32 0 0 0 32 207 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 223 84 0 0 0 0 64 158 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 239 158 32 0 0 0 0 112 239 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 191 32 0 0 0 0 15 96 174 239 255 255 255 255 255 255 255 255 255 255 255 255 223 158 84 0 0 0 0 0 64 191 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 158 46 0 0 0 0 0 0 46 84 127 127 191 191 191 174 127 127 64 32 0 0 0 0 0 0 64 191 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 223 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 64 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 191 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 64 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 191 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 64 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 191 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 64 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 223 119 84 46 68 101 195 255 255 255 255 255 135 101 101 101 101 101 101 178 195 255 255 255 255 96 101 101 101 101 101 101 101 101 101 101 255 211 96 101 101 223 255 255 255 255 146 101 101 158 255 255 223 135 135 195 255 255 255 255 223 146 75 75 75 75 146 223 255 255 255 255 255 135 135 165 255 255 255 255 239 146 84 75 75 75 135 223 255 255 255 255 255 255 255 255 255 255 255 84 68 174 207 255 223 191 68 46 239 255 255 255 46 211 207 207 207 207 191 137 84 46 239 255 255 55 211 207 207 207 207 207 207 207 211 55 255 174 112 211 174 46 255 255 255 255 46 211 211 101 255 255 195 15 15 135 255 255 255 165 32 15 15 15 15 15 15 32 165 255 255 255 255 15 15 75 255 255 255 211 32 15 15 15 15 15 15 15 165 255 255 255 255 255 255 255 255 255 55 158 255 255 223 207 207 255 255 191 46 239 255 255 46 255 255 255 255 255 255 255 255 174 46 255 255 68 255 255 255 255 255 255 255 255 255 75 255 178 137 255 255 101 158 255 255 255 46 255 207 101 255 255 195 15 15 135 255 255 223 32 15 15 105 135 135 75 15 15 15 223 255 255 255 15 15 75 255 255 239 32 15 15 84 135 135 84 15 15 15 195 255 255 255 255 255 255 255 119 119 255 255 68 55 101 84 15 207 255 191 55 255 255 46 255 255 15 46 46 32 64 255 255 105 202 255 64 255 202 32 46 46 46 46 46 46 46 255 178 135 255 255 255 32 239 255 255 46 255 207 101 255 255 195 15 15 135 255 255 135 15 15 178 255 255 255 255 135 15 15 119 255 255 255 15 15 75 255 255 165 15 15 146 255 255 255 255 146 15 15 105 255 255 255 255 255 255 255 32 255 255 46 178 255 255 255 239 32 239 255 68 223 255 46 255 255 46 255 255 255 119 146 255 174 119 255 64 255 202 105 255 255 255 255 255 255 255 255 178 135 255 255 255 158 84 255 255 46 255 207 101 255 255 195 15 15 135 255 255 135 15 15 223 255 255 255 255 239 75 84 165 255 255 255 15 15 75 255 255 135 15 15 178 255 255 255 255 255 75 84 165 255 255 255 255 255 255 195 105 255 185 75 255 255 255 255 255 158 119 255 174 119 255 46 255 255 46 255 255 255 185 135 255 207 101 255 64 255 202 105 255 255 255 255 255 255 255 255 178 135 255 195 211 255 68 202 255 46 255 207 101 255 255 195 15 15 135 255 255 146 15 15 46 165 239 255 255 255 255 255 255 255 255 255 15 15 75 255 255 165 15 15 32 165 223 255 255 255 255 255 255 255 255 255 255 255 255 135 158 255 135 158 255 255 255 255 255 255 64 255 223 75 255 46 255 255 32 185 178 158 32 202 255 119 158 255 64 255 211 64 68 68 68 68 68 46 255 255 178 135 255 127 55 255 223 32 255 46 255 207 101 255 255 195 15 15 135 255 255 239 32 15 15 15 15 64 119 195 255 255 255 255 255 255 15 15 75 255 255 239 46 15 15 15 15 64 119 178 255 255 255 255 255 255 255 255 255 101 207 255 135 178 255 255 255 255 255 255 68 255 255 46 255 46 255 255 127 135 137 135 223 255 255 32 239 255 64 255 255 255 255 255 255 255 255 46 255 255 178 135 255 135 32 158 255 119 127 46 255 207 101 255 255 195 15 15 135 255 255 255 211 46 15 15 15 15 15 15 46 135 255 255 255 255 15 15 75 255 255 255 211 75 15 15 15 15 15 15 32 135 255 255 255 255 255 255 255 101 207 255 135 178 255 255 255 255 255 255 68 255 255 46 255 46 255 255 255 255 255 255 255 191 46 158 255 255 64 255 211 64 64 68 68 68 68 46 255 255 178 135 255 135 165 15 255 255 46 32 255 207 101 255 255 195 15 15 135 255 255 255 255 255 211 119 64 15 15 15 15 15 119 255 255 255 15 15 75 255 255 255 255 255 211 119 64 15 15 15 15 15 84 255 255 255 255 255 255 158 137 255 158 137 255 255 255 255 255 223 84 255 207 84 255 46 255 255 64 64 68 32 46 68 223 255 255 255 64 255 202 105 255 255 255 255 255 255 255 255 178 135 255 135 185 158 84 255 185 0 255 207 101 255 255 195 15 15 135 255 255 255 255 255 255 255 255 223 165 75 15 15 15 223 255 255 15 15 75 255 255 255 255 255 255 255 255 239 165 84 15 15 15 211 255 255 255 255 255 223 84 255 239 46 255 255 255 255 255 101 165 255 158 137 255 46 255 255 46 255 255 255 255 255 255 255 255 255 64 255 202 101 255 255 255 255 255 255 255 255 178 135 255 137 178 255 46 207 255 158 255 202 101 255 255 195 15 15 135 255 255 135 105 119 255 255 255 255 255 255 84 15 15 195 255 255 15 15 75 255 255 135 105 105 255 255 255 255 255 255 119 15 15 195 255 255 255 255 255 255 46 239 255 119 75 239 255 255 119 46 255 255 46 239 255 46 255 255 46 255 255 255 255 255 255 255 255 255 64 255 202 105 255 255 255 255 255 255 255 255 185 135 255 137 178 255 223 46 255 255 255 202 101 255 255 195 15 15 135 255 255 46 15 15 223 255 255 255 255 255 105 15 15 195 255 255 15 15 75 255 255 75 15 15 211 255 255 255 255 255 119 15 15 195 255 255 255 255 255 255 174 55 255 255 174 46 46 46 135 255 255 105 101 255 255 46 255 255 46 255 255 255 255 255 255 255 255 255 64 255 223 64 68 68 68 68 68 68 32 255 185 135 255 137 178 255 255 101 135 255 255 202 101 255 255 195 15 15 135 255 255 135 15 15 64 223 255 255 255 146 15 15 46 255 255 255 15 15 75 255 255 165 15 15 64 195 255 255 255 195 32 15 32 239 255 255 255 255 255 255 255 135 55 239 255 255 255 255 255 255 105 68 255 255 255 46 255 255 46 255 255 255 255 255 255 255 255 255 68 255 255 255 255 255 255 255 255 255 135 255 185 137 255 146 174 255 255 255 15 239 255 207 101 255 255 195 15 15 135 255 255 239 64 15 15 15 46 75 32 15 15 15 178 255 255 255 15 15 75 255 255 255 64 15 15 15 46 75 32 15 15 15 165 255 255 255 255 255 255 255 255 255 174 46 68 137 137 137 84 32 135 255 255 255 255 46 75 75 46 255 255 255 255 255 255 255 255 255 15 68 68 64 68 68 68 68 68 68 32 255 178 32 75 46 174 255 255 255 174 32 68 55 96 255 255 195 15 15 135 255 255 255 239 105 15 15 15 15 15 15 46 195 255 255 255 255 15 15 75 255 255 255 239 105 15 15 15 15 15 15 46 178 255 255 255 255 255 255 255 255 255 255 255 255 239 174 174 174 223 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 239 195 195 223 255 255 255 255 255 223 178 135 135 165 211 255 255 255 255 255 255 195 195 211 255 255 255 255 255 239 195 135 135 146 211 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 index-es.html0100644000113000011300000005750307534721352012157 0ustar kripkrip OpenIsis 0.8.x versión actual isis openisis manual práctico acerca de enlaces in english

bienvenido a OpenIsis.org

versión actual

DEMOS:
normal o trate de buscar caracteres unicode ( índice unicode)

Disponible para "descarga":
Código fuente Binarios
  • win32 (0.8.5) ( mingw32 cross-build) debería trabajar en todas las plataformas win32 desde win5b.
    Nota: Si usted quiere utilizar java va a necesitar el JDK1.3.x
    La versíón actual de windows aún no es "thread-safe", por lo tanto utilizarla con una máquina servlet de Java bajo ambiente windows requiere algún cuidado.
  • solaris (0.8.4) ( compilado en solaris 5.8 )
    Nota: el JDK que viene con solaris 5.7 / 5.8 ( java 1.2 ) debería funcionar.
Características:
  • Julio 2002
    trabajo de artículo ¿Qué de ISIS hace a ISIS?
  • 0.8.6 Junio 2002
    Esta versión implementa formateo básico. Mientras que la mayoría de características de WinISIS o CISIS, especialmente las características gráficas, aún no están implementadas (de todas maneras no son típicamente utilizadas en ambiente web), hay apoyo para subcampos repetidos como se declara en MARC para varios campos. Vea notas de formateo para más detalles. La extensión Perl ahora implementa formateo (véase test.pl), y versiones mejoradas para los otros lenguajes seguirán pronto.
  • PHP Junio 2002
    Braulio José Solano Rojas de Costa Rica creó una extensión PHP, que puede ser vista en acción en galileo. Disponible como descarga o desde sourceforge modulo php-openisis. Además, el Institut Teknologi Bandung de Indonesia cambió su índice web a una versión basada en PHP/OpenIsis.
  • 0.8.5.2 Marzo 2002
    Extensión PERL bastante mejorada. Vea el archivo OpenIsis.pm incluido en el código fuente.
  • 0.8.5.1 March 2002
    Java ahora implementa apoyo para modos de formateo básico como MHL, varios modos HTML-seguros (como escapar todos los caracteres no ASCII), un método Vn-field-selector-style y varias utilerías interesantes. La indentación no está propiamente manejada, dado que no hay una solución fácil y común en HTML. Se harán herramientas para una serie de estrategias estándares luego...
  • 0.8.5 Marzo 2002
    Finalmente una implementación del lenguaje de consulta (demostración). Todos los operadores están ahí (incluyendo /(etiqueta) excepto /(t1,t2...)). Se hace lo posible para limitar el costo potencial de consultas extremadamente estúpidas como "$"^"$", de tal manera que resultados no históricos (#n) o resultados intermedios (para precedencia) son guardados. Las consultas son procesadas estrictamente de izquierda a derecha en un almacenamiento de 1000 coincidencias.
  • 0.8.4 Enero 2002
    Reescritura casi completa del código de búsqueda para apoyar condiciones NEAR. Arreglado los problemas de alineamiento en IFP, ahora trabaja con la bd unesb (-format aligned) y la bd cds tal como se distribuyen con winisis.
    La bd cds que tenímos en versiones previas (de una distribución vieja de CDS) tiene un formato mezclado: archivos hoja alineados, otros desalineados. Aunque el apoyo a un formato mixto se logra fácilmente con openisis, no será incluido a menos que alguien lo necesite (envíen un correo).
    El demo JSP ahora apoya la búsqueda en la bd unesb con más de 70.000 filas (noten que el servidor anfitrión es un Celeron de 500MHZ).
    La búsqueda está limitada a 1000 "postings", usualmente resultando en un número de filas de alguna manera menor (donde las filas tienen múltiples "postings" coincidentes). El menor número de fila (MFN) que fue cortado es grabado, y es posible (aún no en el JSP) repetir la búsqueda empezando desde esa fila.
  • 0.8.3 Octubre 2001
    Primer lanzamiento realmente utilizable, dado que tenemos búsqueda verdadera por índice (como prefijo o palabra completa). La búsqueda da una lista (arreglo) de MFNs ordenados; la aritmética lógica (y, o, negación) en estas listas es directa. La demostración JSP muestra como una consulta es refinada (más angosta) iterativamente haciendo un "y" lógico con otra consulta. Gracias a Verónica Lencinas y colegas, tenemos una primer versión en español de este documento originalmente en inglés.
  • Septiembre 2001
    No hay un lanzamiento nuevo aún, pero hay mantenimiento y pruebas. Nuevo mecanismo de bitacoreo de estructuras. El código fuente está disponible via CVS en sourceforge. La versión para Windows openisis.exe corre.
  • 0.8.2 Agosto 2001
    openisis se encuentra ahora bajo la LGPL, no queda código legado. La conversión de estructuras de archivo se maneja ahora mediante una descripción abstracto y dinámica más que estructuras-C. De esta manera podemos dar soporte a diferentes configuraciones de archivos, bases de datos más grandes, grandes procesadores endian y más aún. Disponibilidad de búsqueda simple en texto libre mediante C-Lib y Java. Obteniendo unos 30.000 registros/seg en una lectura aleatoria en una lenta Notebook de 300 MHz Notebook parece ser de utilidad en la práctica. Una demo jsp demo está en desarrollo ;).
  • 0.8.1 Junio 2001
    Versión con interfaz nativa Java. Package Java org.openisis incluye Db, Rec, Field, Test. NativeDb implementada en libopenjsis.so. Extracción de subcampos y traducción de entidades html en java puro.
  • 0.8.0 Mayo 2001
    Extracción de subcampos y traducción de entidades html. Acceso general desde Perl también mediante un xsub. El registro aparece como hash, manuable pero sin campos repetibles. Makefile.PL, test.pl etc.
  • 0.7.9 Mayo 2001
    Primer versión: librería-C estática (C-Library) libopenisis.a para leer registros a partir de la clave de fila (Mfn). Ejecutable "openisis" hace test. Logeado, lectura de argumentos, Makefile, demo etc.

qué es isis

Isis es un sistema de bases de datos sencillo pero potente, que cuenta con una importante base de usuarios desde los 80. Debido a su buena adaptabilidad a datos bibliográficos, se usa generalmente en bibliotecas y por su bajo costo, especialmente en aquellas con escaso presupuesto.

introdución a bases de datos isis

Una base de datos isis consiste en filas de estructura no específica, identificada con un número único, la clave de fila (mfn). Cada fila contiene una lista de campos, con un número de campo (tag) y un valor de cadena. Dentro de una fila puede haber cero, uno o más campos con un número de campo determinado. El valor del campo generalmente es la representación textual de datos en una u otra codificación de caracteres (generalmente una de las páginas de código IBM/DOS), pero también puede contener bytes arbitrarios.

subcampos
Hay una convención que consiste en codificar campos múltiples dentro de un solo campo, separándolos con un ' ^ ' seguido por un caracter identificador del subcampo. De esta manera el valor del campo ' ^afoo^bbar^bbaz' representa un campo con un subcampo 'a' con valor 'foo' y dos subcampos 'b': 'bar' y 'baz' respectivamente.
formateo
Hay un lenguaje de formateo con variables de texto, campos y subcampos, estructuras de control if-else (según existencia de campos) y búcles for (sobre repetición de campos) (dicho de manera sencilla).
indización

Se construye un índice, convirtiendo una fila en una lista de palabras (aplicando formatos opcionales) y ubicándolas según su posición de su ocurrencia en la fila, en un árbol B (que se extiendo sobre seis archivos). Se puebe buscar una palabra o un prefijo indicando - o no - la posición (campo). Debido a que todos los campos pueden ser combinados en un índice, no es necesario (pero posible) generar múltiples índices.

consultas

Un lenguaje de consultas permite la combinación de palabras usando los operadores "y" (and), "o" (or) y "no" (not).

uso

Mientras que isis carece de la mayoría de las características de un sistema de gestión de bases de datos relacionales (RDBMS), como relaciones complejas entre diferentes entidades, su flexibilidad resulta práctico para catálogos y directorios con registros altamente variables y un solo nivel de subestructura, los que actualmente se vierten en documentos XML más que en filas. El mecanismo flexible de indización combina lo mejor de búsqueda en texto completo con recuperación estructurada.

resúmen de software isis

La madre de todo el software isis es una versión DOS de "MicroISIS", un sistema integrado con interfaz de texto. Hay una versión BSD de "CDS/ISIS" que también corre bajo linux hasta kernels 2.2.x (los kernels 2.4 actuales no permiten el módulo iBCS para binarios COFF). Además hay varias versiones de "WinISIS" (sólo para M$-Windows).

Existe una librería compartida de funciones de acceso a bases de datos isis "isis.dll" que, a pesar de su nombre, también se distribuye en una versión para linux "isilux". Sin embargo se necesitan algunas librerías muy especiales para usarla. Un conjunto de utilitarios de línea de comando ("cisis") realizan tareas tales como la importación de archivos de intercambio de bases de datos bibliográficas - ISO2709, construcción de archivos invertido, etc. Lo más cercano a un servidor de bases de datos isis es "wwwisis", que corre como un CGI o bien desde la línea de comandos y realiza la mayoría de tareas de isis (versiones win y lin). Sin embargo, corre sólo por encargo, no como un servidor propiamente dicho y no puede proveer control de procesos simultáneos.

Este software "oficial" de isis, es mantenido por Unesco y/o Bireme, y ha sido complementado por aportes independientes, algunos de dominio público. Javaisis es una interfaz gráfica basada en AWT con su correspondiente servidor, que usa wwwisis. Es un software bueno y abierto que se concentra sobre un frontend multiplataforma, pero sin acceso directo a la base de datos. Robert Janusz escribió una librería C de cero (iAPI), que fué el punto de partida para el software openisis.

qué es openisis

¿Y entonces, porqué estamos haciendo openisis? Porque Isis no es un software de fuente abierta, ni siquiera es gratuito, y esto conduce a una serie de problemas.

problemas resultantes de software cerrado

  1. Disponibilidad (en teoría)
    Existen versiones para algunos sistemas operativos, determinadas versiones de librerías e idiomas. Para otros entornos, no hay versiones disponibles, y nada se puede hacer al respecto.
  2. Disponibilidad (en la práctica)
    Puede bajar la mayoría del software pero está protegido, en parte, con contraseñas que se deben solicitarse a algún distribuidor nacional. Existen algunos aranceles y/o la condición de exponer las razones de la necesidad del software, hecho lo cual hay un tiempo de espera hasta conseguirlo, por ejemplo en Alemania más de un mes.
  3. Disponibilidad (en términos legales)
    Algunas partes del software incluyen información sobre las licencias bajo las cuales se distribuyen, otras no. Los términos de las mismas difieren según el país. No es fácil establecer exactamente cuál puede ser el uso permitido.
  4. Disponibilidad (de documentación)
    Alguna documentación está disponible en inglés, otra solamente en portugués, castellano o italiano. Sólo una pequeña parte se puede "bajar", la mayoría consta en papel.
  5. Control de fallas (bugs)
    No hay forma de corregir una falla y tampoco se puede hay mucho que se pueda hacer para que alguien la corrija.
  6. Extensión
    La única forma de escribir una asociación (Binding) para perl o Java sería mediante isis.dll. Hay problemas respecto la necesidad de librerías adicionales (especialmente en C++), no hay información respecto a seguridad de procesos, compatibilidad con unicode, etc. En consecuencia es practicamente imposible escribir una aplicación web para bases de datos isis con las tecnologías corrientes.
  7. Mejoramiento
    Con la práctica, muchos usuarios desarrollan ideas útiles para el mejoramiento del software. Su experiencia se pierde cuando no son capaces de transferirla a un software mejorado.
  8. Habilitación
    Mientras que software de fuente abierta permite a personas de todo el mundo a adaptar sus herramientas a sus necesidades, el software cerrado los mantiene dependientes.

beneficios de software de fuentes abiertas

Para encarar estos problemas, percibimos la necesidad de una implementación abierta de isis. Sin duda lo mejor sería de contar con el código existente para isis bajo una u otra forma de licencia abierta (GPL, LGPL, artística o similar, según sea apropiado).

Por otro lado, una implementación secundaria e independiente, tiene sus ventajas. Permitiría un enfoque diferente, desarrollando determinadas potencialidades, mientras que en otras situaciones el abordaje sería distinto. Por ejemplo, openisis tendrá algún soporte para procesos múltiples y unicode, pagando un cierto costo. Realizar una recodificación por un desarrollador con un bagaje de conocimientos diferentes, puede aportar nuevas ideas que, luego de haber sido probadas, pueden ayudar a mejorar el estándar.

OpenIsis como software para acceder bases de datos isis está y estará disponible gratuitamente para cualquiera, con código fuente completo, sin aranceles ni restricciones.

desarrollando openisis

En términos generales, no está planeado reimplementar cada línea de código escrita para isis. Para ser útil, OpenIsis debe mantener la compatibilidad con el formato de los archivos de la base de datos. De esta manera, se puede usar winisis o cualquier otro script existente para crear y mantener la base de datos, empleando la interfaz perl de OpenIsis para un motor de consultas potentes y la interfaz nativa de Java para aplicaciones web basada en servlets.

OpenIsis privilegiará la provisión de herramientas más que aplicaciones. Por ejemplo, no intentará recrear la funcionalidad de winisis hasta que el conjunto de herramientas GUI haya sido completado. Para alcanzar este objetivo, OpenIsis provee acceso a los lenguajes de programación más importantes: Java para web (HECHO), perl para scripts (HECHO) y, algún día, Tcl/Tk para interfaces gráficas independientes de la plataforma (parcialmente hecho). Todos los demás (lenguajes) pueden por cierto, utilizar enlaces a la librería.

Proximos pasos:

  • hacer un diseño de archivos configurable para permitir bases de datos más grandes y de implementación exclusiva en java (HECHO)
  • implementar la búsqueda (búsqueda en texto libre HECHO)
  • implementar búsqueda en base a archivo invertido
  • mayor rendimiento: probar std (HECHO, rinde mal) y homegrown io buffering, además de acelerar bucles en la función de conversión de ldb's
  • empezar a trabajar en una implementación pura de java "thread-save" (voluntarios ?)
  • preparar lanzamientos binarios para windows (.exe y .dll para java) (HECHO)
  • implementar el lenguaje de consulta (mayormente HECHO)
  • implementar formateo (va en camino)
  • implementar la escritura (mayormente planeada)
  • ... los voluntarios son bienvenidos !

manual práctico open isis

Se comienza bajando el software. Hasta ahora es todo fuente, sin binarios. Crearemos distribuciones binarias cuando se logre mayor estabilidad. Se desempaca todo en algún directorio cualquiera. Para las pruebas, se necesitará también alguna base de datos isis que deberá ser ubicada como archivo db/cds/cds.*. Pruebe ésta. Controlar que los nombres de archivo estén en minúsculas.

Si se está en Windows, deberá contar con un entorno cygwin con herramientas como gmake y gcc o portar el código y escribir el archivo make para su make y compilador. Erik armará una versión Windows empleando Linux gcc como compilador cruzado. Si se está en Linux, está todo bien. Con un MAC ... ni idea.

Tipee "make" y disfrute de los mensajes del compilador. Tipee "make demo" y disfrute de su primer registro openisis. (Ya instaló una base db/cds/cds.*, no es así? Tiene 42 registros?) Tipee "make run" y observe el contenido de su base de datos. Ttipee "make test", no debería haber diferencia entre testout.txt y testres.txt como se entregan (usando esta base de datos cds). Tipee "make time" para medir el rendimiento, pruebas subsiguientes generalmente son mucho más rápidas. Mi 800 MHz P3 logra en lectura aleatoria más de 80.000 registros por segundo, una vez que los archivos estén en el cache del sistema.

$ make time
time ./openisis -v 3 -perf 100000 -db db/cds/cds
0.63user 0.61system 0:06.20elapsed 19%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (95major+11607minor)pagefaults 0swaps
$ make time
time ./openisis -v 3 -perf 100000 -db db/cds/cds
0.71user 0.53system 0:01.23elapsed 100%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (95major+11607minor)pagefaults 0swaps

Tipee "make perl" para armar el material para perl. Se sobreentiende que previamente deberá haberse instalado algún tipo de perl 5.*. Tipee "make java" o, si no le alcanza, tipee "make jdump", para verlo aparecer en su nuevo y brillante JDK 1.3 Java VM. Algún 1.2.* JDK debería hacerlo también, pero indíquele al Makefile que no busque en /usr/java/jdk1.3, seteándolo a JAVAHOME.

instalando openisis

libopenisis puede enlazarse con tu código; no es necesario instalarlo. Si desea instalar los binarios 'openisis' en alguna ubicación de su path, simplemente cópielo, no hay registros de configuración mágicos.

Para instalar openisis para su disponibilidad general en su /usr/lib/perl5 o donde esté instalado, cambie al subdirectorio perl (después de "make perl") e ingrese "make install" (como root o legitimado de otra manera). Después pruebe "perldoc OpenIsis" y el script de demostración demo.pl.

Java, like perl, needs to dynamically slurp both some stuff in the own language and a native shared object. The former is openisis.jar, set your CLASSPATH to include it, or specify when invoking java like in the Makefile. The latter is libopenjsis.so on linux (yes, it's jsis). The system dynamic linker must be able to find it; see NativeDb.java for details.

sobre openisis.org

OpenIsis.org es patrocinado por allmaxx.de, a service of merconic, Berlín, Alemania. Como un sitio destinado a estudiantes, allmaxx apoya software de fuente abierto dirigido a la educación y el manejo de conocimientos. Véase también la open community for science.

Actualmente el sitio es mantenido por Erik y Paul. Los voluntarios son bienvenidos. El código fuente de Openisis está disponible en SourceForge lado a lado con el proyecto PHP isis de Franck Martin's PHP. Gracias Franck!

links

openisis y PHP isis en sourceforge
sitios núcleo isis:
Unesco
Bireme
documentación:
EL LIBRO manual de referencia CDS/ISIS incl. formato de datos (en español)
gente y proyectos:
Robert Janusz' iAPI
Kafkas Caprazli's TODO acerda de CDS/ISIS
software de código abierto para bibliotecas
javaisis
Institut Teknologi Bandung IsisOnline en Indonesia
grupos de usuarios:
Netherlands / international
UK (ISIS PLUS)
Germany (isisnetz)
personal:
Erik Grziwotz
Klaus "Paul" Ripke
Braulio José Solano Rojas

documentación

ISIS, conjuntos de caracteres y unicode
$Revision: 1.2 $ last changed $Date: 2002/09/02 09:30:45 $ by $Author: kripke $
(this page intentionally left blank :) index.html0100644000113000011300000005717307554752720011562 0ustar kripkrip OpenIsis 0.8.x current version isis openisis howto about links en espanol
Isi

welcome to OpenIsis.org

current version

DEMOS:
standard or try searching for unicode characters ( unicode index)

Available for download:
Sources Binaries
  • win32 (0.8.6) ( mingw32 cross-build) should work on all win32 platforms from win95b on.
    Note: If you want to use java you will need the JDK1.3.x
    The current windows version is not yet thread-safe, so using it with a Java servlet engine under windows requires some care.
  • solaris (0.8.4) ( build on solaris 5.8 ) Please mail us if you need a more up-to-date solaris binary.
    Note: the JDK shipped with solaris 5.7 / 5.8 ( java 1.2 ) should work.
News:
  • October 19th 2002
    The OpenIsis society ("Verein") has been founded with 15 members. Chairman is Erik Grziwotz, other board members are Gabi Rohmann, Ingo Struck and Thomas Sonnemann.
  • 0.8.7 October 2002
    Version 0.8.7 supports writing of DOS/WinIsis masterfiles and xrf. This currently works fine for a single process on Linux. See writing.txt for details. Current TODOs: interlock multiprocess writing (PHP, Perl CGI) and fix the windows and solaris versions. Besides writing support, there is a new streaming record reader, which groks various formats like the SYSPAR.PAR, email headers and property files, so you can fill your db from such textfiles. Next step: new indexing engine.
  • August 2002
    No new software yet. Still very busy doing metawork. We are preparing to set up an organisation to support OpenIsis development with much more momentum and a company for professional services like help on large scale ISIS installations. paperwork on the universal ISIS record
  • July 2002
    some paperwork on What is it about ISIS that makes it ISIS?
  • 0.8.6 June 2002
    This version supports basic formatting. While most, especially graphical, features of WinISIS or CISIS formatting are not yet supported (which are typically not used in a web environment anyway), there is support for repeated subfields as declared by MARC for many fields. See formatting notes for details. The perl binding supports formatting (see the test.pl), enhanced versions for all those languages are to follow soon.
  • PHP June 2002
    Braulio José Solano Rojas from Costa Rica created a PHP binding, which can be seen in action at galileo. Available as download or from sourceforge module php-openisis. Also, the Institut Teknologi Bandung of Indonesia switched it's web index to a PHP/OpenIsis based version.
  • 0.8.5.2 March 2002
    Much enhanced PERL binding. See the OpenIsis.pm included in the sources.
  • 0.8.5.1 March 2002
    Java now has support for basic formatting modes like MHL, various HTML-safety modes (like escaping all non-ASCIIs), a Vn-field-selector-style method and several nice utils. Indentation is not properly handled, since there is no easy common solution in HTML. Will build tools for a choice of standard strategies later ...
  • 0.8.5 March 2002
    Finally some implementation of the query language (demo). All the operators are there (including /(tag), but not /(t1,t2...)). Every attempt is made to limit the potential costs of even extremely stupid queries like "$"^"$", so no historical (#n) or intermediate results (for precedence) are stored. Queries are processed strictly left to right on a buffer of 1000 hits.
  • 0.8.4 January 2002
    Nearly complete rewrite of search code with support for NEAR conditions. Fixed alignment problems in IFP, now works with the unesb db (-format aligned) and the cds db as distributed with winisis.
    The cds db we had with previous versions (from an old CDS distribution) has a mixed format: aligned leaf files, others unaligned. Although support for a mixed format is easily achieved with openisis, it will not be included unless somebody has a need for it (send us mail).
    The JSP demo now supports searching the unesb db with over 58.000 rows (note the hosting server is a 500MHz Celeron).
    Searching is limited to 1000 postings, usually resulting in a somewhat lower number of rows (where rows have multiple matching postings). The lowest row number (MFN) that was cut off is recorded, and it is possible (not yet in the JSP) to repeat the search starting from that rowid.
  • 0.8.3 October 2001
    First truly usable release, since we have true search by index now (as prefix or complete word). Search gives a list (array) of sorted MFNs; arithmetic on those lists (and, or, not) is straightforward. The JSP demo shows how a query is refined (narrowed) iteratively by ANDing it with a second query. And thanks to Veronica Lencinas and colleagues, we have a version of this document en espanol.
  • September 2001
    Not a new release yet, but maintenance and testing. New structure logging mechanism. Sources are available via CVS at sourceforge. Windows version openisis.exe running.
  • 0.8.2 August 2001
    openisis now under LGPL, no legacy code. Conversion from file structures governed by abstract dynamic description rather than C-structs, so we can support different file layouts, much larger databases, big endian processors and more. Simple full-scan search available in C-Lib and Java. Given a random read throughput of about 30.000 Records/sec on a lame 300 MHz Notebook this seems to be of practic use. jsp demo under development ;).
  • 0.8.1 June 2001
    Java native interface version available. Java package org.openisis has Db, Rec, Field, Test. NativeDb implemented in libopenjsis.so. Subfield splitting and htmlifying in pure java.
  • 0.8.0 May 2001
    Subfield splitting and htmlifying. Everything also available from perl as an xsub. Record shows up as hash, handy but no repeated fields. Makefile.PL, test.pl etc.
  • 0.7.9 May 2001
    First version: static C-Library libopenisis.a for reading records by rowid(Mfn). Executable "openisis" does test. Logging, argumentparsing, Makefile, demo etc.

what is isis

Isis is a simple, yet powerful database system with a large installed base since the 80s. Since it's well suited for bibliographic data, it's commonly used in libraries, and since it's very low cost, especially in those running on a low budget.

introduction to the isis db

An isis DB is a list of rows of unspecified structure, each identified by a unique number, the rowid (a.k.a. mfn). Each row is a list of fields, and each field has number (tag) and a string value. Within a row there may be zero, one or more fields with a given tag. While the field's value usually is a textual representation of data in one or the other character encoding (commonly one of the IBM/DOS code pages), it may actually contain arbitrary bytes. This is closely modelled after ISO2709 "Information Interchange Format" (IIF, a.k.a. ANSI/NISO Z39.2)
subfields
There is a convention to encode multiple fields in one by separating them with a '^' followed by one character tagging the subfield. So the field value '^afoo^bbar^bbaz' represents a field having one 'a' subfield with value 'foo' and two 'b' subfields 'bar' and 'baz'. An other separator char may be used, e.g. ASCII character 31 ("Unit Separator") is used in the MARC standard.
formatting
There is a formatting language, with literal text, field and subfield variables, if-else branches (on field existance) and for loops (over field repetitions) (roughly speaking).
indexing
An index is build by converting a row into a list of words (optionally applying formats) and stuffing every word, qualified by the position of it's occurence in the row, into a B+-Tree (which is actually spread to six files). Searching for a word or word prefix is possible with or without qualifying the position (field). Since all fields can be combined into one index, it is usually not necessary (but possible) to set up multiple indexes.
queries
A query language allows for combination of word lookups using and, or and not(without) operators. This is very similar to the "Type-1" query of Z39.50.
usage
While isis lacks most features of RDBMS like complex relations between different entities, it's flexibility comes in handy for many catalogues and directories with highly varying records and one single level of substructure, which today are usually modelled in XML documents rather than table rows. In other words, isis is an ideal storage for many XML applications. The flexible indexing mechanism combines the best of full text searching and structured retrieval.

overview of isis software

The mother of all isis software is a DOS version of "MicroISIS" as an integrated system with textual user interface. There is a BSD version of "CDS/ISIS" which also runs under linux up to some 2.2.x kernels (current 2.4 kernels do not support the iBCS module for COFF binaries). Then there are several versions of "WinISIS" (M$-Windows only, but runs under linux/wine).

A shared library version "isis.dll" of functions to access an isis db from your code exists, despite it's name, also in a linux version ("isilux"); however, you need some pretty special libs to make it run. A set of command line tools ("cisis") performs tasks like importing ISO2709 bibliographic databases, inverting (index building) etc. The thing next to an isis database server is "wwwisis", which runs as CGI or from the command line and performs most isis tasks (win and lin versions). However, it actually runs per request, not as a server itself, and thus cannot provide concurrency control.

This "official" isis software, which is maintained by Unesco and/or Bireme, is accompanied by a couple of independent developments, some of which are in the public domain. Javaisis is an AWT-based GUI (3.5 uses SWING) and a corresponding server, which in turn uses wwwisis. Robert Janusz wrote a C-lib (iAPI) from scratch, which was the starting point for the openisis software.

what is openisis

So why are we writing the openisis software? Because Isis is not open source software, it's not even free software, and that leads to a whole bunch of problems.

problems resulting from closed software

  1. Availability (in theory)
    Versions of the software exist for some operating systems, library versions and languages. For other environments, there is no version of the software, and there is not much one can do about it.
  2. Availability (in practice)
    You may download most the software, but it's partly protected with passwords, which you have to order at some national distributor. You have to pay some fee and/or declare some good reasons, why you want to use the software. Then you have to wait. In germany, for example, it didn't work at all for some time, until the newly founded Isisnetz remedied the situation.
  3. Availability (in legal terms)
    Some parts of the software are accompanied by different documents stating some license terms, others are not. Terms seem to be pretty different between countries. One can not easily figure out, what exactly might be allowed usage.
  4. Availability (of documentation)
    Some documentation is available in english, some only in portugese, espanol or italiano. Only a small part is downloadable at all, most is paperware.
  5. Bugfixing
    There is no way one can fix a bug, and not much one can do about having somebody fix it.
  6. Extending
    The only way one could write a Binding for perl or Java would be using the isis.dll. There are problems with regard to required additional libraries (especially some C++ stuff), there are no statements about thread safety, unicode compatibility and so on. As a consequence, it's practically impossible to write a state-of-the-art web application based on an isis db.
  7. Improving
    Many users develop useful ideas for improvements from practice. Their expertise is lost as they are not able to turn it into improved software.
  8. Enabling
    While open source software enables people all over the world to shape their tools themselves, closed software lets them in dependency.

benefits of open software

To address these problems we feel the need for an open source implementation of isis. Of course it would be best to have all of the existing isis code under one or the other form of open license (GPL, LGPL, artistic or similar as appropriate).

On the other hand, an independent secondary implementation has advantages in it's own right. It may have a different focus and develop strengths in one aspect while another aproach performs better in other situations. For example, openisis will have some support for multithreading and unicode, which is paid for by a certain overhead. A rewrite by developers with a different background might introduce new ideas which finally, after having had their indepent test bed, help improve the standard.

OpenIsis as a software to access isis databases is and will be freely available for everybody with full sourcecode, no fee, no restrictions.

developing openisis

In general, there are no plans to reimplement every piece of code ever written for isis. To be of practical value, OpenIsis has to maintain compatibility in the format of the database files anyway. So, one may use winisis or whatever existing import scripts to create and maintain the database, yet deploy OpenIsis' perl interface to run powerful reports and the Java Native Interface to allow queries from a Servlet based web application.

OpenIsis will focus on providing tools rather than applications. For example, there will be no attempt to mirror functionality of winisis unless the GUI toolkit is done. To achive this, OpenIsis provides access from the most important programming languages: Java and PHP for the web (DONE), perl for the scripts (DONE) and Tcl/Tk for platform independent GUIs (partly DONE). All others can, of course, link the lib.

Next steps:

  • make file layout configurable to allow for larger db (DONE)
  • implement searching (full-scan searching DONE)
  • implement index-based searching (DONE)
  • more performance: try std (DONE, performs badly) and homegrown io buffering, further accelerate loop in ldb's convert function (DONE)
  • start work on thread-save pure-java implementation (cancelled due to lack of demand)
  • prepare binary releases for windows (.exe and .dll for java) (DONE)
  • implement query language (simple version DONE)
  • implement formatting (simple version DONE)
  • implement writing data (masterfile DONE, index underway)
  • finish Tcl binding, create GUI version using TK (-> Erik)
  • implement server version
  • ... volunteers are welcome !

howto open isis

Start by downloading the Software. Unpack everything in some arbitrary directory. For the tests, you will also need some isis database, which must be located as files db/cds/cds.*. Try this one. Make sure filenames are lowercase.

If you are on Windows, you should either get yourself the cygwin environment with tools like gmake and gcc or volunteer as a porter and start writing the Makefile for your make and compiler. Erik has build a Windows version using mingw and Linux gcc as crosscompiler. If you are on Linux, everything is fine. Ports to MAC OS X and other UNIXes should be no problem.

Type "make" and enjoy the compiler messages. (If your make complains, e.g. on BSD, try "gmake"). Type "make demo" and enjoy your first open isis record. (You installed a db/cds/cds.*, didn't you? It has 42 rows?) Type "make run" and watch the guts of your db passing by. Type "make test", there should be no difference between the testout.txt and the testres.txt as provided (using this cds database from winisis and this 15 MB 58.000+ row unesb.zip db). Type "make time" to measure performance, subsequent tries are usually much faster. My 800 MHz P3 random-reads more than 179.000 records a second, once the files are in the system cache. Typical values: 

time ./openisis -perf 1000000 -db db/cds/cds >/dev/null

real    0m5.655s
user    0m3.650s
sys     0m2.000s

time ./openisis -perf 100000 -db db/unesb/unesb >/dev/null

real    0m0.991s
user    0m0.670s
sys     0m0.320s

time ./openisis -fmt mfnf -search 'k$' -db db/unesb/unesb >/dev/null
860     rows for        k

real    0m0.044s
user    0m0.040s
sys     0m0.000s
Type "make perl" to build the perl stuff; some perl 5.* must be installed beforehand. Type "make java" or, if you just can't get enough, "make jdump", to see it all happen in your shiny new JDK1.3 Java VM. Some 1.2.* JDKs should do, but tell the Makefile to not look in /usr/java/jdk1.3 by setting JAVAHOME.

installing openisis

libopenisis.a can be linked with your code; no installation necessary. You may wan't to install the 'openisis' binary somewhere in your path for the fun of it; go ahead, just copy, no magic registry entries.

To install the perl stuff for general availability in your /usr/lib/perl5 or whatever, cd to the perl subdir (after "make perl") and issue "make install" (as root or otherwise legitimated). After that, try "perldoc OpenIsis" and the demo.pl script.

Java, like perl, needs to dynamically slurp both some stuff in the own language and a native shared object. The former is openisis.jar, set your CLASSPATH to include it, or specify when invoking java like in the Makefile. The latter is libopenjsis.so on linux (yes, it's jsis). The system dynamic linker must be able to find it; see NativeDb.java for details.

about openisis.org

OpenIsis.org is sponsored by allmaxx.de, a service of merconic, Berlin, Germany. As a student's site, allmaxx supports open software with a focus on education and knowledge management. See also the open community for science.

Currently the site is maintained by Erik and Paul. Volunteers are very welcome. Openisis sources are available at SourceForge side by side with Franck Martin's PHP isis project. Thanks, Franck!

links

openisis and PHP isis at sourceforge
isis core sites:
Unesco
Bireme
documentation:
THE BOOK CDS/ISIS reference manual incl. data formats (en espanol)
standards:
ISO2709 "Information Interchange Format", a.k.a. ANSI/NISO Z39.2
(US) MARC 21, overview
Z39.50, overview at OCLC|Pica,
links at indexdata, makers of excellent free Z39.50 software.
people and projects:
Robert Janusz' iAPI
Kafkas Caprazli's EVERYTHING about CDS/ISIS
open source software for libraries
javaisis
Institut Teknologi Bandung IsisOnline in Indonesia
user groups:
Netherlands / international
UK (ISIS PLUS)
Germany (isisnetz)
staff:
Erik Grziwotz
Klaus "Paul" Ripke
Braulio José Solano Rojas

documentation

ISIS, charsets and unicode
What is it about ISIS that makes it ISIS?
the universal ISIS record
record writing implementation
multi-threading performance
$Revision: 1.32 $ last changed $Date: 2002/10/21 10:24:16 $ by $Author: kripke $
(this page intentionally left blank :) ioperf.c0100664000113000011300000000163107644411041011167 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ isocvt0100775000113000011300000000036207427563367011016 0ustar kripkrip#!/usr/bin/perl # tools for ISO 2709 format # # field sep is ^^ (char val 30) # record sep is ^]^J (char val 29 and newline) $/ = "\035\n"; while () { print; last if $. > 20; # chomp; # y/\n\t\036/ \t/; # print $_, "\n"; } jdocstyle.css0100664000113000011300000000233107445455333012262 0ustar kripkrip/* Javadoc style sheet */ /* Define colors, fonts and other style attributes here to override the defaults */ /* Page background color */ body { background-color: #FFFFFF } /* Table colors */ .TableHeadingColor { background: #00CC00 } /* Dark Green */ .TableSubHeadingColor { background: #AAEEAA } /* Light Green */ .TableRowColor { background: #FFFFFF } /* White */ /* Font used in left-hand frame lists */ .FrameTitleFont { font-size: normal; font-family: normal } .FrameHeadingFont { font-size: normal; font-family: normal } .FrameItemFont { font-size: normal; font-family: normal } /* Example of smaller, sans-serif font in frames */ /* .FrameItemFont { font-size: 10pt; font-family: Helvetica, Arial, sans-serif } */ /* Navigation bar fonts and colors */ .NavBarCell1 { background-color:#AAEEAA;}/* Light Green */ .NavBarCell1Rev { background-color:#00CC00;}/* Dark Green */ .NavBarFont1 { font-family: Arial, Helvetica, sans-serif; color:#000000;} .NavBarFont1Rev { font-family: Arial, Helvetica, sans-serif; color:#FFFFFF;} .NavBarCell2 { font-family: Arial, Helvetica, sans-serif; background-color:#FFFFFF;} .NavBarCell3 { font-family: Arial, Helvetica, sans-serif; background-color:#FFFFFF;} jtest.bat0100664000113000011300000000367507533161161011374 0ustar kripkrip@echo off rem openisis - an open implementation of the ISIS database rem Version 0.8.x (microversion see file Version) rem Copyright (C) 2001-2002 by Erik Grziwotz, erik@openisis.org rem rem This library is free software; you can redistribute it and/or rem modify it under the terms of the GNU Lesser General Public rem License as published by the Free Software Foundation; either rem version 2.1 of the License, or (at your option) any later version. rem rem This library is distributed in the hope that it will be useful, rem but WITHOUT ANY WARRANTY; without even the implied warranty of rem MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU rem Lesser General Public License for more details. rem rem You should have received a copy of the GNU Lesser General Public rem License along with this library; if not, write to the Free Software rem Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA rem rem This software is dedicated to the memory of Eckart Dietrich. rem rem This software is inspired by (but contains no code of) the iAPI rem Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. rem See iAPI.txt for what it contains. rem rem rem $Id: jtest.bat,v 1.1 2002/08/28 14:47:13 erikisis Exp $ rem Testscript for openisis java classes on win32 platforms rem set CLASSPATH=. if not exist testout.txt goto bug del testout.txt :bug java org.openisis.Test -db db/cds/cds -encoding Cp850 -row 42 >> testout.txt java org.openisis.Test -db db/cds/cds -encoding Cp850 -search development >> testout.txt java org.openisis.Test -db db/cds/cds -encoding Cp850 -terms "" >> testout.txt java org.openisis.Test -db db/cds/cds -encoding Cp850 -mpu -row 130 >> testout.txt java org.openisis.Test -db db/cds/cds -encoding Cp850 -mhl -hta -row 130 >> testout.txt java org.openisis.Test -db db/cds/cds -encoding Cp850 -m1 -row 130 >> testout.txt fc testout.txt jtest.txt jtest.txt0100644000113000011300000005570607445370772011461 0ustar kripkripgot row 42 <24><The> Meteorological problems of the deltaic flood-plains of East Pakistan <26>^c1966 <26c>1966 <30>^ap. 123-133^billus., maps <30a>p. 123-133 <30b>illus., maps <44>Scientific problems of the humid tropical zone deltas and their implications: proceedings of the Dacca Symposium <50>Incl. bibl. <69>Paper on: <meteorology><deltas><floods><sea level><climatic data><rain><Bangladesh> <70>Naqvi, Sibte Nabi got row 100 <24>Micropolitan development: theory and practice of greater-rural economic development <26>^aAmes^bIowa State University Press^c1976 <26a>Ames <26b>Iowa State University Press <26c>1976 <30>^a456 p.^bmap <30a>456 p. <30b>map <50>Incl. bibl. <69><regional planning><rural development><USA> - <human resources development><employment><economics of education><poverty><community development><industrialization><land use><economic policy><statistical data> <70>Tweeten, Luther G. <70>Brinkman, George Lorris got row 114 <24>Educational development; a new discipline for self-renewal <26>^aEugene, Or.^bCenter for the Advanced Study of Educational Administration^c1973 <26a>Eugene, Or. <26b>Center for the Advanced Study of Educational Administration <26c>1973 <30>^a343 p.^billus. <30a>343 p. <30b>illus. <50>(CASEA monograph, no.24). Incl. bibl. <69><educational development><public education><educational planning><educational research><USA> - <educational programmes><curriculum development><teacher training><evaluation of education><information systems><educational management><educational statistics> <70>Hemphill, John Knox <70>Rosenau, Fred S. <71>Far West Laboratory for Educational Research and Development (USA) <71>University of Oregon (USA). Center for the Advanced Study of Educational Administration got row 130 <12>Joint FAO/WHO Expert Committee on Nutrition^n9th^pRome^d1974 <24>Food and nutrition strategies in national development <26>^aRome^bFAO ^c1976 <26a>Rome <26b>FAO <26c>1976 <30>^a64 p. <30a>64 p. <50>(FAO nutrition meetings report series, no.56). WHO technical report series, no.584 <69><agricultural development><food production><nutrition><national development> - <economic and social development><food resources><research priorities><developing countries> <71>FAO <71>WHO <76>Stratégies de l'alimentation et de la nutrition dans le développement national^zfre got row 131 <24><The> Measurement of scientific and technical activities - Proposed standard practice for surveys of research and experimental development <26>^aParis^bOECD^c1976 <26a>Paris <26b>OECD <26c>1976 <30>^a138 p. <30a>138 p. <69><technology assessment><science and technology><research and development><science statistics><statistical analysis><survey analysis><international standards><guides> <71>OECD <74>Proposed standard practice for surveys of research and experimental development <76>La Mesure des activités scientifiques et techniques^zfre got row 141 <24>Energy for rural development: renewable resources and alternative technologies for developing countries <26>^aWashington^bNational Academy of Sciences^c1976 <26a>Washington <26b>National Academy of Sciences <26c>1976 <30>^a306 p.^billus. <30a>306 p. <30b>illus. <50>(PB-260-606). Incl. bibl. <69><renewable resources><solar energy><energy resources><rural development><technology transfer><appropriate technology><developing countries> - <energy conversion><geothermal energy><electrical engineering><wind power><biomass><manufacturing industries><research and development><research centres><directories> <71>National Academy of Sciences (USA). Ad Hoc Panel on Renewable Energy Resources <71>National Research Council (USA) <71>National Academy of Sciences (USA). Board on Science and Technology for International Development got row 149 <12>Symposium on the Psychological Bases of Programmed Learning ^pTbilisi, USSR^d1976 <24>For a training project based on programmed learning in the developing countries: objectives, means and resources, organization and management, development <26>^c14 June 1976 <26c>14 June 1976 <30>^a18 p. <30a>18 p. <69><educational assistance><programmed instruction><developing countries> - <educational innovations><educational technology><educational needs><case studies><teaching methods><educational experiments><technical and vocational education><adult education><working population><production><programmed materials> <70>Pocztar, Jerry got row 150 <24>Micropolitan development: theory and practice of greater-rural economic development <26>^aAmes^bIowa State University Press^c1976 <26a>Ames <26b>Iowa State University Press <26c>1976 <30>^a456 p.^bmap <30a>456 p. <30b>map <50>Incl. bibl. <69><regional planning><rural development><USA> - <human resources development><employment><economics of education><poverty><community development><industrialization><land use><economic policy><statistical data> <70>Tweeten, Luther G. <70>Brinkman, George Lorris ABBAS, B.M. ABEYWICKRAMA, B.A. ABOUT ABSENCE ABSORPTION ACCOMPANYING ACCOUNT ACCOUNTING ACHIEVEMENTS ACID ACTION ACTIVITIES ACTUAL ADDRESSES ADMINISTRATIVE REFORM ADMISSION ADMISSION CRITERIA ADULT ADULT EDUCATION AERIAL AERIAL PHOTOGRAPHY AESTHETICS AFFECTIVE AFGHANISTAN AFRICA AFRICAN LANGUAGES AGE AGE GROUPS AGENT AGENTS AGRICULTURAL DEVELOPMENT AGRICULTURAL TRAINING AGRICULTURE AGRONOMY AHMAD, MAJEED AHMAD, NAZIR ALLUVIAUX ALLUVIONS ALTERNATIVE AMAZON AMERICA ANALYSIS ANALYTICAL ANAYA, RICARDO ANCIENT CIVILIZATIONS ANIMAL ANNA ANNOTATED ANNUAL ANTI ANTIQUITIES APERCU APPARATUS APPLICATION APPROACH APPROPRIATE TECHNOLOGY APRIL APTER, DAVID ERNEST AQUIFERS ARAB COUNTRIES ARABES ARCHITECTURE AREA AREAS ARID ARMAMENTS ARMS CONTROL ARMS SALES ART BOOKS ART HISTORY ARTISTIC CREATION ARTS ASIA ASPECTS ASPIRATION ASSIMILATION ASSURANCES AT ATANGANA, ENGELBERT ATTITUDE AUDIT AUSTRALIA AUSTRALIAN AUTOMATIC AVAILABLE AWAY AYALA, FRANCISCO JOSE BAEZ, ALBERT V. BAHOKEN, J.C. BALANCE BANGLADESH BARRS, H.D. BARS BASED BASEL BASIC BASIC EDUCATION BASIS BAUMGARTNER, ALBERT BE BEHAVIOUR BELOW BENGAL BETWEEN BIBLIOGRAPHIE BIBLIOGRAPHIES BIBLIOGRAPHY BIOLOGICAL EVOLUTION BIOLOGICAL RESEARCH BIOLOGY BIOMASS BITSIOS, DIMITRI S. BJELICA, MIHALO BLACKS BLOOM, JOSEPH BOARDING SCHOOLS BOESCH, H. BOLIVIA BONDURANT, JOAN V. BOOK BORDIA, ANIL BORICHANSKY, L.S. BORING BOROBUDUR BOSIAN, G. BOURDEAU, PHILIPPE F. BOWRA, SIR CECIL MAURICE BRACKISH BRACKISH WATER BRAIN DRAIN BRASIL BRAZIL BRINKMAN, GEORGE LORRIS BRIX, H. BROADCAST BROADCASTING BROTHERS BUDDHISM BUDEL, J. BUDGET BUILDING BURMA BUSINESS MANAGEMENT C CALCULATION CAMEROON CAMEROON UR CANADA CARACTERISTIQUES CARBON CARBON DIOXIDE CARBONIQUE CARIBBEAN AREA CARLIER, G. CARMEL CARTE CARTOGRAPHY CASE STUDIES CASE, JOHN H. CASPERS, H. CAT CATALOGUES CATSKY, J. CENTRE CEREALS CEREMONY CEYLON CHAMBER CHANCELLOR CHANDI CHANGE CHANGES CHANGING CHANGING SOCIETY CHAO CHAPMAN, VALENTINE J. CHEMICAL ANALYSIS CHEMICAL PROPERTIES CHEMICAL/BIOLOGICAL WARFARE CHI CHILDREN CHILDREN'S BOOKS CHIMIQUE CHO CHOWDHURY, M.I. CHURCH CHURCHILL, STACY CIVILIZATION CLASSICAL CLASSIFICATION CLAYS CLIMATE CLIMATIC CLIMATIC DATA CLUBS CO COASTAL COASTAL ZONES COLOMBIA COLORIMETRIC COMMUNICATION COMMUNICATION HISTORY COMMUNICATION IMPACT COMMUNICATION LEGISLATION COMMUNICATION PERSONNEL TRAINI COMMUNICATION PLANNING COMMUNICATION POLICY COMMUNICATION PROCESS COMMUNICATION RESEARCH COMMUNICATION STATISTICS COMMUNITIES COMMUNITY COMMUNITY DEVELOPMENT COMMUNITY EDUCATION COMPARATIVE COMPARATIVE ANALYSIS COMPARISON COMPULSORY EDUCATION COMPUTER APPLICATIONS COMPUTER HARDWARE COMPUTER SOFTWARE COMUNICACAO CONCENTRATION CONCLUSIONS CONDITIONS CONFERENCE CONFERENCES CONFLICT CONFLICT RESOLUTION CONFORM CONQUEST CONSULTANCY CONTEMPORARY SOCIETY CONTRIBUTES CONTROL CONTROLLED CORPORATION COUGHLAN, MARGARET N. COUNTRIES COUNTRY REPORTS COURSES COVER COVERS CRAIN, N. CRITICISM CROISSANCE CROP CROPPED CROSS CUBA CULTIVATION CULTURAL CULTURAL ACTIVITIES CULTURAL DEVELOPMENT CULTURAL FACILITIES CULTURAL FACTORS CULTURAL HISTORY CULTURAL IDENTITY CULTURAL INSTITUTIONS CULTURAL PLANNING AND ADMINIST CULTURAL POLICY CULTURAL PROGRAMMES CULTURE CURRENT CURRENTS CURRICULUM DEVELOPMENT CURRICULUM PLANNING CYPRUS D DACCA DANCE, F. ROBERT DANS DATA COLLECTION DATA PROCESSING DAVE DAVE, RAVINDRAKUMAR H. DE DECENTRALIZATION DECISION DECISION MAKING DEFICIT DEFINIES DEFINITIONS DEGRE DEGREES DEGUCHI, K. DELIVERED DELTA DELTAIC DELTAS DEMAND DEMOCRATIC KAMPUCHEA DEMOCRATIZATION OF EDUCATION DES DESARROLLO DESCRIPTION DESIGN DETERMINATION DETERMINING DEVELOPED COUNTRIES DEVELOPING DEVELOPING COUNTRIES DEVELOPMENT DEVELOPMENT PLANNING DEVELOPMENT STRATEGIES DEVELOPPEMENT DEW DIFFERENT DIFFUSION DIOXIDE DIPLOMAS DIRECTOR DIRECTORIES DISARMAMENT DISC DISCHARGE OF RIVERS DISCIPLINE DISCUSSED DISPOSITION DISSEMINATION OF INFORMATION DOBZHANSKY, THEODOSIUS GRIGORI DOCUMENTATION DOCUMENTATION CENTRES DOCUMENTOS DOTY, M.S. DOW, PHILIP S. DRAINAGE DRAPER, JAMES A. DROUGHT DRY DU DUE DUKE, CHRISTOPHER E EAST EASTERN EUROPE EAUX ECOLOGY ECONOMIC ECONOMIC AND SOCIAL DEVELOPMEN ECONOMIC BEHAVIOUR ECONOMIC CONCENTRATION ECONOMIC CONDITIONS ECONOMIC GROWTH ECONOMIC POLICY ECONOMICS OF EDUCATION ECONOMY ECOSYSTEMS EDUCATION EDUCATIVES EFFECT EFFECTIFS EFFECTS EGOROV, V.V. ELECTRIC ELECTRICAL ENGINEERING ELEMENT ELKANA, YEHUDA EMISSION EMPLOYERS EMPLOYMENT EN ENERGY ENERGY BALANCE ENERGY CONVERSION ENERGY RESOURCES ENROLMENT ENROLMENT PROJECTIONS ENROLMENT RATIOS ENSEIGNEMENT ENVIRONMENTAL ENVIRONMENTAL CHANGES ENVIRONMENTAL CONTROL ENVIRONMENTAL EDUCATION ENVIRONMENTAL QUALITY EQUATION ESCAP ESTABLISHMENT ESTUARIES ESTUARINE ESTUARY ET ETHICS ETUDE EUROPE EVALUATING EVALUATION METHOD EVALUATION OF EDUCATION EVAPORATION EVAPOTRANSPIRATION EVERYDAY LIFE EXAMINATIONS EXCHANGE EXISTENCE EXPECTATIONS EXPERIENCE EXPERIMENT EXPERIMENTAL EXPERIMENTS EXPERTS EXTRACTS FACTORS FAJARDO, JOSE CARLOS FAMILY LIFE FAUNA FAURE, EDGAR FEATURES FELLOWSHIPS FERTILIZERS FIELD FIELDS FILMS FINAL FIRST FISH FISHERIES FLOOD FLOOD CONTROL FLOODING FLOODS FLORA FLORAS FLOW MEASUREMENT AND INSTRUMEN FOLKLORE FOOD FOOD PRODUCTION FOOD RESOURCES FORCE FOREIGN INVESTMENT FOREIGN TRADE FOREST FORESTS FORMATION FORMERLY FOSBERG, F. RAYMOND FRADIER, GEORGES FRANCE FRANCO, C.M. FRANEKER FREEDOM FRESH FRESHWATER FULLY FUTURE GALE, J. GANDHIAN GANGES GASEOUS GAUGES GAZ GENERAL GENESIS GEODYNAMICS GEOLOGICAL GEOLOGY GEOMORPHOLOGIE GEOMORPHOLOGY GEOTHERMAL ENERGY GERMANY FR GHANI, Q. GOODMAN, LOUIS WOLF GOOR, G.A.W. VAN DE GRADUAL GRADUATES GRASS GREAT GREATER GREECE GRIEVE, B.J. GROOT, A.J. DE GROUND GROUNDWATER GROUP DISCUSSIONS GROWTH GUIDANCE SERVICE GUIDE GUIDES GUTTMANN, LUDWIG HALLAK, JACQUES HANDBOOK HAVELOCK, MARY C. HAVELOCK, RONALD G. HEALTH HEAT HEAT TRANSFER HEEMST, H.D.J. VAN HEMPHILL, JOHN KNOX HERVIEU, J. HIGH HIGHER HIGHER EDUCATION HIGHER TECHNICAL PERSONNEL HISTOIRE HISTORIC MONUMENTS HISTORICAL ANALYSIS HISTORY HOLDINGS HOLLANDER, JACK M. HOLLERWOGER, F. HOLT, MICHAEL HOME HOOK, SIDNEY HORLAMUS, SEPP HOUSE HOUSING HUDSON, J.P. HUISMAN, L. HUMAN HUMAN ACTIVITIES EFFECTS HUMAN ECOLOGY HUMAN RESOURCES DEVELOPMENT HUMAN RIGHTS HUMAN SETTLEMENT HUMID HUMID ZONES HUMIDITE HUMIDITY HUNGARY HYDRATURE HYDRAULIC HYDRAULIC ENGINEERING HYDRODYNAMICS HYDROGEOCHEMISTRY HYDROGEOLOGY HYDROGRAPHY HYDROLOGICAL HYDROLOGY HYDROMETEOROLOGICAL HYDROMETEOROLOGY HYGROMETER HYGROMETERS IBE IDEAL IMPLICATIONS IMPORTANT IMPROVEMENT IN-SERVICE TEACHER TRAINING INCLUDING INCOME INCOME DISTRIBUTION INCOMUNICACAO INDEX INDIA INDIAN INDIRECT INDIVIDUAL INDONESIA INDUSTRIAL PRODUCTS INDUSTRIALIZATION INDUSTRIES INFLUENCE INFORMATION PROCESSING AUTOMAT INFORMATION SERVICES INFORMATION SOURCES INFORMATION SYSTEMS INFORMATION USERS INFORMATION/LIBRARY FINANCING INFORMATIQUE INFRA INNOVATION INNOVATION BEHAVIOUR INSTRUCTIONAL INTEGRATED INTEGRATED CURRICULUM INTERACTION INTERDISCIPLINARY APPROACH INTERGOVERNMENTAL ORGANIZATION INTERNAL INTERNATIONAL INTERNATIONAL CONFLICT INTERNATIONAL COOPERATION INTERNATIONAL LAW INTERNATIONAL ORGANIZATIONS INTERNATIONAL RELATIONS INTERNATIONAL STANDARDS INTERNATIONAL TRADE INTERNATIONALE INVERTEBRATA INVOLVED IRAN (ISLAMIC REPUBLIC) IRRAWADDY IRRIGATION ISLAM, M. AMIRUL ISRAEL IT ITALY JACKMAN, MARY ELIZABETH JAVA JOBORU, MAGDA JOURNALISM JOURNALISTS JUSQU KARAMAZOV KARENINA KARIM, A.Q.M.B. KENYA KHAN, F. KARIM KHUDA, M. QUDRAT KIDD, JAMES ROBBINS KLAUSING, O. KLEINFELD, JUDITH S. KRAMER, P.J. KREEB, K. L LA LABORATORY LABOUR LABOUR MARKET LABOUR MIGRATION LABOUR POLICY LABOUR RELATIONS LAND LAND RECLAMATION LAND USE LANGE, O.L. LANGUAGE BEHAVIOUR LANGUAGE OF INSTRUCTION LATIN LATIN AMERICA LAW LE LEADERSHIP LEAF LEARNING LEARNING PROCESSES LEAST DEVELOPED COUNTRIES LEAVES LEBLOND, C. LECTURES LEGISLATION LEIDEN LEKOVIC, ZDRAVKO LEROY, GEORGES LES LEVITT, J. LIBERATION MOVEMENTS LIETH, H. LIFE LIFE-LONG EDUCATION LIFELONG LIKELY LIST LITERACY PROGRAMMES LITERARY CRITICISM LITERATURE LIVING CONDITIONS LONG LOSS LUMIERE LYSIMETER M MADAGASCAR MAGALHAES, A.C. MAHEU MAIN MAKING MAKKINK, G.F. MAN MANAGEMENT MANAGEMENT INFORMATION SYSTEMS MANAGEMENT OPERATIONS MANGANESE MANGOKY MANGROVE MANGROVES MANKIND MANPOWER NEEDS MANUEL MANUFACTURING INDUSTRIES MAQUIS MARJORAM, D.T.E. MARKET MARRAKECH MARSHES MASS MASS CULTURE MASS MEDIA MASTER MATHEMATICAL MATHEMATICAL LOGIC MATHEMATICAL MODELS MATHEMATICS MATTER MAY, WILLIAM H. MCQUAIL, DENIS MEANING MEANS MEASUREMENT MEASUREMENT AND INSTRUMENTS MEASUREMENTS MEASURING MEDIA MEDICINAL PLANTS MEETING MEETINGS MELO, JOSE MARQUES DE MEMORIAL MESURE METABOLISM METEOROLOGICAL METEOROLOGY METHOD METHODOLOGIE METHODOLOGY METHODS MICHAELI, A. MICHIGAN MICROBIOLOGY MICROCLIMATE MICROCLIMATIC MICROPOLITAN MIDDLEMARCH MIKHAILOV, V.N. MILITARY ENGINEERING MINERAL DEPOSITS MINERAL RESOURCES MINERALOGIQUES MIRON, MURRAY S. MISE MODEL MODELS MODIFICATIONS MOISTURE MOLECULAR BIOLOGY MOLLUSCA MONGOLIA MONOLITHS MONUMENT MOORMANN, F.R. MORPHOLOGY MORPHOLOGY (LINGUISTICS) MOUNTAINS MOUTH MR MUCHEKE MUD MUELLER-DOMBOIS, D. MULTINATIONAL MULTIRACIAL SOCIETIES NAGARAJA, V.N. NAMIBIA NAQVI, SIBTE NABI NATIONAL NATIONAL DEVELOPMENT NATIONAL PLANNING NATIONAL POLICY NATIONALISM NATIONS NATURAL NEED NEEDS NETHERLANDS NEW NEWLY-INDEPENDENT STATES NEWS AGENCIES NEWSPAPERS NIGER NO NON-VIOLENCE NORTH AMERICA NOTES NOVEL NOVELS NUCLEAR NUCLEAR WEAPONS NUTRIENT NUTRIENTS NUTRITION OBJECTIVES OBLIGATOIRE OCCUPANT OCCUPATIONAL QUALIFICATIONS OCCURENCE OCEAN CURRENTS OCEANIA OHYA, MASAHIKO OIAPOQUE OKATCHA, F.M.M.O. ONE ONTARIO OPEN OPENING OPERATIONS ORGANES ORGANIC MATTER ORGANIZATION ORGANIZATIONAL OSGOOD, CHARLES EGERTON OTHER OTRAS OUEST PACIFIC ISLANDS PAKISTAN PAR PARASITES PART PART-TIME COURSES PATEL, A.M. PAYS PECS PERDU PERFORMANCE PERIODICALS PERRY, WALTER PERSONAL PERSONNEL MANAGEMENT PERSPECTIVES PERU PERUVIAN PESTS PETTIGREW, ANDREW M. PHENOMENOLOGY PHILIPPINES PHILOSOPHICAL PHILOSOPHY PHOTOGRAPHS PHOTOSYNTHESIS PHYA PHYSICAL EDUCATION PHYSICAL PROPERTIES PHYSIOLOGY PIED PILOT PILOT PROJECTS PINE PIRES, J. MURCA PLACE PLAINS PLAN PLANKTON PLANNING PLANT PLANT ECOLOGY PLANT EVAPOTRANSPIRATION PLANT PHOTOSYNTHESIS PLANT PHYSIOLOGY PLANT TRANSPIRATION PLANTS POCZTAR, JERRY POINT POLICIES POLICY POLITICAL BEHAVIOUR POLITICAL CONFLICT POLITICAL ECONOMY POLITICAL PARTICIPATION POLITICAL PARTIES POLITICAL SYSTEMS POLITICAL THOUGHT POLITICS POLITIQUES POLJAKOFF-MAYBER, A. POLLUTION POLSTER, H. PONENCIAS POPULATION POPULATION DENSITY POPULATION DYNAMICS POROMETER PORTUGAL POST-GRADUATE COURSES POTENTIAL POUR POVERTY POWER INDUSTRY PRACTICE PRATIQUE PRESENTED PRESERVATION OF CULTURAL PROPE PRESERVATION OF MONUMENTS PRESS PREVAILING PRIMAIRE PRIMARY EDUCATION PRIMARY SCHOOLS PROBLEM PROBLEMES PROBLEMS PROCEEDINGS PRODUCT PRODUCTION PRODUCTIVITY PROFESSIONAL ASSOCIATIONS PROFESSIONAL TRAINING PROGRAMMED PROGRAMMED INSTRUCTION PROGRAMMED MATERIALS PROGRAMMES PROGRAMS PROGRESS PROJECT PROJECTIONS PROJECTS PROPOSED PROWSE, G.A. PSYCHOANALYSIS PSYCHOLINGUISTICS PUBLIC PUBLIC EDUCATION PUBLIC HEALTH PUBLIC OPINION QUALITY QUALITY OF LIFE QUANTITATIF QUANTITATIFS QUESTIONNAIRES QURESHI, M. RAHIMULLAH R RACE RACE RELATIONS RACIAL SEGREGATION RADIATION MEASUREMENT AND INST RADIO RAIN RAINFALL RATES READINGS RECENT RECENTES RECENTS RECHERCHE RECOMMENDATIONS RED REDUCTION REFERENCE REFORM REFRACTIVE REGARD REGIME REGIME OF WATERS REGIONAL COOPERATION REGIONAL PLANNING REGIONS REGULATION RELATED RELATING RELATION RELATIVE RELEVANCE RENE RENEWABLE RENEWABLE RESOURCES RENEWAL REPORT REPUBLIC RESEARCH RESEARCH AND DEVELOPMENT RESEARCH CENTRES RESEARCH PRIORITIES RESEARCH WORKERS RESISTANCE RESOLUTION RESOLUTIONS RESOURCE CONSERVATION RESOURCES RESPONSE RESTORATION REVIEW RHODESIA RHODESIAN RICE RIEPE, DALE MAURICE RIGHT TO INFORMATION RIGHTS RIVER RIVER BASINS RIVER CONTROL RIVERS ROLE ROLE OF EDUCATION ROONWAL, M.L. ROSENAU, FRED S. RUN-OFF RURAL RURAL AREAS RURAL DEVELOPMENT RURAL POPULATION RURAL YOUTH S SA, ARTUR MOREIRA DE SAFETY SAINT ANTONIN, CLAUDE SALINE SALINE WATER SALINITY SAMMAK, A. SANCTIONS SANTEMA, P. SANYAL, BIKAS C. SATELLITE SATELLITE BROADCASTING SATURATION SCALE SCHMID, MAURICE SCHOELLER, MARC SCHOLARSHIPS SCHOOL LABORATORIES SCHOOL MAPPING SCHOOL-AGE POPULATION SCHOOLS SCIENCE SCIENCE AND DEVELOPMENT SCIENCE AND TECHNOLOGY SCIENCE EDUCATION SCIENCE FINANCING SCIENCE POLICY SCIENCE STATISTICS SCIENTIFIC SCIENTIFIC ACTIVITIES SCIENTIFIC DEVELOPMENT SCIENTIFIC INSTITUTIONS SCIENTIFIC PERSONNEL SCIENTIFIC PROGRAMMES SCIENTIFIC PROGRESS SCIENTIFIC RESEARCH SCOLAIRE SCOLAIRES SCOLARISATION SCULPTURE SEA SEA LEVEL SEA WATER SECOND SECONDARY EDUCATION SECONDARY SCHOOL CURRICULUM SECONDARY SCHOOLS SEDIMENT SEDIMENT TRANSPORT SEDIMENTATION SEDIMENTS SELECTED SELF SELF-TEACHING METHODS SEMANTICS SEMI SENEGAL SEPTEMBER SERVICES SEX SHACHORI, A.Y. SHIFTING SHIPWORMS SHUI SIGNIFICANCE SIGNIFICANT SIMMONS, MELVIN K. SINGH, B.N. SIOLI, HAROLD SKILLED WORKERS SLATYER, R.O. SLAVIK, B. SOCIAL SOCIAL CHANGE SOCIAL ENVIRONMENT SOCIAL FACTORS SOCIAL INTEGRATION SOCIAL LIFE SOCIAL PARTICIPATION SOCIAL PLANNING SOCIAL REFORM SOCIAL SCIENCE RESEARCH SOCIAL SECURITY SOCIAL STATUS SOCIAL STRATIFICATION SOCIAL SYSTEMS SOCIALES SOCIALISM SOCIO-ECONOMIC FACTORS SOCIOLOGY SOCIOLOGY OF COMMUNICATION SOCIOLOGY OF SCIENCE SOEKMONO, R. SOIL SOIL EROSION SOIL FERTILITY SOIL GENESIS SOIL MECHANICS SOIL MOISTURE SOIL SCIENCE SOIL SURVEYS SOIL TYPES SOILS SOLAR ENERGY SOLAR RADIATION SOLS SOME SOUTERRAINES SOUTH EAST ASIA SOUTHERN AFRICA SPAIN SPECIAL SPORT SRI LANKA STAKE, ROBERT E. STANDARD STANDARDIZATION STANHILL, G. STATEMENT STATEMENTS STATES STATISTICAL ANALYSIS STATISTICAL DATA STATISTICAL METHODOLOGY STATISTICS STATUS STELLING-MICHAUD, SVEN STOMATAL STONE, E.C. STORM STORMS STRATEGIES STRATEGY STRESS STRUCTURAL STUDENT ATTITUDE STUDENT BEHAVIOUR STUDENT SELECTION STUDENT SOCIOLOGY STUDIES STUDY SUCTION SUD SULPHATE SUNDERBANS SUR SURFACE SURGES SURVEY ANALYSIS SURVEYS SURVIVAL SWITZERLAND SYMPOSIUM SYSTEMS ANALYSIS TAIWAN TEACHER BEHAVIOUR TEACHER EDUCATORS TEACHER GUIDES TEACHER ROLE TEACHER TRAINING TEACHING METHODS TEACHING SKILLS TECHNICAL TECHNICAL AND VOCATIONAL EDUCA TECHNICAL ASSISTANCE TECHNICAL PERSONNEL TECHNICAL TRAINING TECHNIQUE TECHNIQUES TECHNOLOGIES TECHNOLOGY TECHNOLOGY ASSESSMENT TECHNOLOGY TRANSFER TELEVISION TELEVISION PROGRAMMES TEMPERATE TEMPERATURE TEMPLES TEMPS TENDANCES TENTATIVE TEREDINID TERMS TERRITORY TEXTURALES THAILAND THEBAUD, SCHILLER THEIR THEORY THERMAL PROPERTIES OF SUBSTANC THREE TIDAL TIDES TIME TISON, LEON J. TOOL TOPICS TOPOGRAPHY TOPOLOGY TOWNS TRACTS TRADITIONAL TRAINING TRAINING ABROAD TRAINING CENTRES TRAINING COURSES TRANSNATIONAL CORPORATIONS TRANSPIRANTS TRANSPIRATION TRANSPORT TRICART, J. TROPICAL TROPICAL ZONES TROPICAUX TROPICS TURGIDITY TURKEY TWEETEN, LUTHER G. TYPE TYPES TYPOLOGY U UNDER UNDERDEVELOPMENT UNESCO UNITED UNIVERSALS UNIVERSIDAD UNIVERSITES UNIVERSITIES UNIVERSITY UNIVERSITY COOPERATION UNIVERSITY COURSES UNIVERSITY CURRICULUM UNIVERSITY EXTENSION URBAN AREAS URBAN DEVELOPMENT URBAN LIFE URBAN PLANNING URBAN POPULATION URBAN RENEWAL URUGUAY USA USE USING USSR V VALEUR VAPOUR VARIABILITY VASAK, KAREL VEGETATION VEGETAUX VELOCITY VERSLUYS, J.D.N. VERSTAPPEN, H.TH. VICE VIET NAM SR VILLAGE VILLAGES VIOLENCE VISSER, W.C. VITALITY VOLKER, ADRIAAN VULNERABLE WANIEWICZ, IGNACY WAR WATER WATER BALANCE WATER YIELD WAVES WAY WEINRICH, A.K.H. WENT, F.W. WEST WESTERN WESTERN EUROPE WETLANDS WIDE WILTING WIND WIND POWER WOMEN WOOD WOODWELL, GEORGE M. WORKERS EDUCATION WORKING POPULATION WORKS WORLD WORLDWIDE WYNTER, HECTOR Y YIELDS YOSUFZAI, H.K. YOUTH YOUTH ORGANIZATIONS YUGOSLAVIA ZACKLIN, RALPH ZAMBIA ZIMBABWE ZONE got row 130 <12>JOINT FAO/WHO EXPERT COMMITTEE ON NUTRITION^N9TH^PROME^D1974 <24>FOOD AND NUTRITION STRATEGIES IN NATIONAL DEVELOPMENT <26>^AROME^BFAO ^C1976 <26a>ROME <26b>FAO <26c>1976 <30>^A64 P. <30a>64 P. <50>(FAO NUTRITION MEETINGS REPORT SERIES, NO.56). WHO TECHNICAL REPORT SERIES, NO.584 <69> - <71>FAO <71>WHO <76>STRATÉGIES DE L'ALIMENTATION ET DE LA NUTRITION DANS LE DÉVELOPPEMENT NATIONAL^ZFRE got row 130 <12>Joint FAO/WHO Expert Committee on Nutrition. 9th. Rome, 1974 <24>Food and nutrition strategies in national development <26>Rome, FAO , 1976 <30>64 p. <50>(FAO nutrition meetings report series, no.56). WHO technical report series, no.584 <69>agricultural development; food production; nutrition; national development - economic and social development; food resources; research priorities; developing countries <71>FAO <71>WHO <76>Stratégies de l'alimentation et de la nutrition dans le développement national. fre MFN : 0130 CONFERENCE : Joint FAO/WHO Expert Committee on Nutrition. 9th. Rome, 1974. TITLE : Food and nutrition strategies in national development. IMPRINT : Rome, FAO , 1976. COLLATION : 64 p. NOTES : (FAO nutrition meetings report series, no.56). WHO technical report series, no.584. CORP. BODY : FAO. WHO. OTHER LANG : Stratégies de l'alimentation et de la nutrition dans le développement national. KEYWORDS : agricultural development; food production; nutrition; national development - economic and social development; food resources; research priorities; developing countries. lbt.c0100644000113000011300000017532307665150673010512 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lbt.c,v 1.35 2003/05/27 11:03:30 kripke Exp $ the btree. */ #include #include /* special #if defined( __GNUC__ ) && defined ( alloca ) #include #endif */ #if defined( sparc ) || defined( __ppc__ ) # define LDB_BIG_ENDIAN #endif #include "lio.h" #include "luti.h" #include "lbt.h" /* probably all relevant cases can be handled by the more efficient approach of storing recoded key a la strxfrm */ #define LBT_USECMP 0 /* ************************************************************ private types */ enum { /* little endian magic text */ LBT_OPEN = 'O' | 'P'<<8 | 'E'<<16 | 'N'<<24, LBT_ISIS = 'I' | 'S'<<8 | 'I'<<16 | 'S'<<24 }; enum { LBT_LEVELS = 16, /* max depth */ LBT_POSBITS = 13, /* bits for position within block */ LBT_POSMASK = 0x1fff, LBT_POSMAX = LBT_POSMASK /* obviously :) */ }; typedef struct Dict { /* dictionary of block entries */ unsigned short pos; /* 13 bits position of entry, 2 bits val for leaves */ unsigned char vln; /* length of entry's value */ /* for inner nodes, this is the length of qualifying second key segment (either 0 or bt->vsz <= 23). for leaves, this, together with higher bits of pos, is the number of values of bt->vsz each. */ unsigned char kln; /* length of entry's key */ /* the following access macros are needed for leaves only */ #define LBT_POS( d ) (LBT_POSMASK & (d)->pos) #define LBT_VLN( d ) ((d)->vln | (~LBT_POSMASK & (d)->pos)>>(LBT_POSBITS-8)) #define LBT_POSADD( d, a ) do { \ (d)->pos = (~LBT_POSMASK & (d)->pos) | (LBT_POSMASK & ((d)->pos + (a))); \ } while (0) #define LBT_VLNADD( d, a ) do { \ unsigned _vln = LBT_VLN( d ) + (a); \ (d)->pos = (LBT_POSMASK & (d)->pos) \ | (~LBT_POSMASK & ((_vln)<<(LBT_POSBITS-8))); \ (d)->vln = (unsigned char)(_vln); \ } while (0) #define LBT_PV( d, p, v ) do { \ (d)->pos = (LBT_POSMASK & (p)) | (~LBT_POSMASK & ((v)<<(LBT_POSBITS-8))); \ (d)->vln = (unsigned char)(v); \ } while (0) #define LBT_PVK( d, p, v, k ) do { \ LBT_PV( d, p, v ); \ (d)->kln = (unsigned char)(k); \ } while (0) } Dict; /* guard space around block */ #ifndef NDEBUG # define LBT_GUARD 64 /* template for guard space -- blank or pattern */ static unsigned char guard[LBT_GUARD]; #else # define LBT_GUARD 0 #endif /* TODO: check the possible performance gain in findkey from separating disk block and administrative data and aligning disk blocks on PAGE boundaries in memory */ typedef struct Block { #if LBT_GUARD unsigned char grd[LBT_GUARD]; #endif unsigned char ird; /* is reading */ unsigned char wrd; /* #threads waiting for read */ unsigned char lck; /* is write locked */ unsigned char wlk; /* #threads waiting for write lock */ #define LBT_BUSY( b ) (*(int*)&(b)->ird) /* any of 1st 4 bytes != 0 */ struct Block *nhs; /* next in hash chain */ struct Block *nru; /* next (more) recently used */ /* TODO: possibly double-link lru list, so we can unlink on every access ? */ /* the disk block -- from here on bsz bytes are on disk */ #define LBT_BOFF ((char *)&((Block*)0)->num - (char*)0) #define LBT_BASE( b ) ((unsigned char *)&(b)->num) unsigned num; /* number of this block, 0 is root */ unsigned nxt; /* right sibling, "OPEN" or "ISIS" for root */ /* typ,key,col are redundant in non-root blocks, maybe abused ... */ unsigned char typ; /* type: bsz, vsz, flags */ unsigned char key; /* max key length */ unsigned char col; /* collation */ unsigned char lev; /* level over bottom, 0 for leaves */ unsigned short ent; /* 10 bits # entries */ #define LBT_ENT( b ) (b)->ent /* currently other bits arent abused */ unsigned short stk; /* 13 bits offset of keyspace stack */ /* offset of dict RELATIVE to base */ #define LBT_DOFF ((char *)&((Block*)0)->dct - (char*)&((Block*)0)->num) Dict dct[1]; } Block; /* LAYOUT: dct is an array of dictionary entries in increasing key order stk is growing downwards from block end towards dict position of entry i is b->stk <= LBT_POS(b->dct+i) < bt->bsz decreasing on i (i.e. stack space is in reverse ordering) at LBT_POS(b->dct+i), there are b->dct[i].kln bytes key for a leaf entry, the key is followed by LBT_VLN(b->dct+i) (between 0..1023 incl.) values of bt->vsz (between 8..23 incl) each for an inner entry, the key is followed by 4 bytes block number and optionally bt->vsz qualifying value SIZES: block size >= 1024 block base overhead LBT_DOFF is 16 bytes avail size >= 1008 Dict entry is 4 bytes max. used stack space per entry is: inner node: kln+4+bt->vsz <= 255+4+23 = 282 new leaf entry (one value): kln+bt->vsz <= 278 max total space for a new entry incl. dict is 286 < avail size/3 */ typedef struct Chunk { struct Chunk *nxt; } Chunk; enum { /* number of levels for lru list */ LBT_LRULEV = sizeof(((Idx*)0)->lru)/sizeof(((Idx*)0)->lru[0]), LBT_ENTBITS = 10, /* bits for # enries */ LBT_ENTMASK = 0x03ff, LBT_ENTMAX = LBT_ENTMASK, LBT_MATCH = 0x8000, /* entry matched */ LBT_ENTSIZE = 8 /* minimum cost of entry: 4 byte dict, 4 byte val */ }; typedef struct Batch { Block *cur; unsigned got; /* current for this key */ unsigned tot; /* total for this key */ unsigned let; /* stats */ unsigned keys; unsigned keyl; unsigned maxl; unsigned vals; unsigned maxv; unsigned span; unsigned free; /* buffer */ Key key; Val val[LBT_ENTMAX]; } Batch; typedef enum { LBT_RD, /* read -- no lock */ LBT_WR, /* write */ LBT_EX /* exclusive */ } lbt_get; typedef struct { /* structure filled by findkey */ Idx *bt; /* the tree */ Key *key; /* the key searched */ Block *b[LBT_LEVELS]; /* path to leaf, b[0] is leaf */ unsigned n[LBT_LEVELS]; /* block numbers of path */ unsigned short e[LBT_LEVELS]; /* entries in path */ } FindKey; typedef struct { unsigned get; unsigned miss; unsigned cmp; unsigned key; } Stat; /* buffer to print a value to. lowest 11 bytes are printed as 4 shorts + one int < 4*5+10 = 30 digits+5sep higher 12 bytes are printed hex = 24 bytes. */ typedef char PrVal[64]; /* ************************************************************ private data */ /* stat access is not guarded by memory barrier. may be wrong when running on multiple processors */ static Stat stat; /* ************************************************************ private functions */ #if defined( sparc ) static unsigned GETINT ( void *m ) /* bus error safe version of *(int*) */ { unsigned l; memcpy( &l, m, 4 ); return l; } static void SETINT ( void *m, unsigned i ) { memcpy( m, &i, 4 ); } #else #define GETINT( m ) (*(unsigned*)(m)) #define SETINT( m, i ) (*(unsigned*)(m) = (i)) #endif static void mkval ( Val *v, Hit *ifp ) { unsigned char *c = v->byt; switch ( v->len ) { default: memset( c, 0, v->len-11 ); c += v->len-11; case 11: *c++ = (unsigned char) (ifp->dbn >> 8); case 10: *c++ = (unsigned char) ifp->dbn; case 9: *c++ = (unsigned char) (ifp->mfn >> 24); case 8: *c++ = (unsigned char) (ifp->mfn >> 16); case 7: *c++ = (unsigned char) (ifp->mfn >> 8); case 6: *c++ = (unsigned char) ifp->mfn; case 5: *c++ = (unsigned char) (ifp->tag >> 8); case 4: *c++ = (unsigned char) ifp->tag; case 3: *c++ = (unsigned char) ifp->occ; case 2: *c++ = (unsigned char) (ifp->pos >> 8); case 1: *c++ = (unsigned char) ifp->pos; case 0: ; } } /* mkval */ static Hit *rdval ( Hit *ifp, unsigned char *c, unsigned vsz ) { ifp->mfn = ifp->tag = ifp->occ = ifp->pos = ifp->dbn = 0; switch ( vsz ) { default: c += vsz-11; case 11: ifp->dbn = (unsigned short)*c++ << 8; case 10: ifp->dbn |= *c++; case 9: ifp->mfn = (unsigned int)*c++ << 24; case 8: ifp->mfn |= (unsigned int)*c++ << 16; case 7: ifp->mfn |= (unsigned int)*c++ << 8; case 6: ifp->mfn |= *c++; case 5: ifp->tag = (unsigned short)*c++ << 8; case 4: ifp->tag |= *c++; case 3: ifp->occ = *c++; case 2: ifp->pos = (unsigned short)*c++ << 8; case 1: ifp->pos |= *c++; case 0: ; } return ifp; } /* rdval */ static char *prval ( PrVal buf, unsigned char *c, unsigned vsz ) { char *d = buf; Hit h; if ( ! vsz ) { buf[0] = '#'; buf[1] = 0; return buf; } if ( 23 < vsz ) vsz = 23; if ( 11 < vsz ) { while ( 11 < vsz-- ) { *d++ = luti_hex[ *c >> 4 ]; *d++ = luti_hex[ *c++ & 0xf ]; } *d++ = '.'; } rdval( &h, c, vsz ); d += u2a( d, h.dbn ); *d++ = '.'; d += u2a( d, h.mfn ); *d++ = '.'; d += u2a( d, h.tag ); *d++ = '.'; d += u2a( d, h.occ ); *d++ = '.'; d += u2a( d, h.pos ); *d++ = '.'; *d = 0; return buf; } /* prval */ /* for PRVAL f->key->val */ #define PRVAL( buf, v ) prval( buf, v.byt, v.len ) /* ************************************************************ block I/O & cache mgmt */ static void iniblk ( Block *b, Idx *bt, unsigned num, unsigned char lev ) { memset( b, 0, LBT_BOFF + bt->bsz + LBT_GUARD ); b->num = num; b->typ = bt->typ; b->key = bt->key; b->col = bt->col; b->lev = lev; b->stk = bt->bsz; } /* iniblk */ static int chkblk ( Block *b, Idx *bt ) { #if 0 (void)b; (void)bt; #else int i; unsigned short eod = LBT_DOFF + b->ent*sizeof(Dict), siz, pos, vln, key; unsigned end = bt->bsz, tot = 0; unsigned char *base = LBT_BASE(b); #if LBT_GUARD if ( memcmp( guard, b->grd, LBT_GUARD ) ) return log_msg( ERR_TRASH, "OOPS! initial guard frobbed on blk %d", b->num ); if ( memcmp( guard, LBT_BASE(b)+bt->bsz, LBT_GUARD ) ) return log_msg( ERR_TRASH, "OOPS! initial guard frobbed on blk %d", b->num ); #endif if ( b->num > bt->len || (b->num && b->nxt > bt->len) ) return log_msg( ERR_TRASH, "bad num %d / nxt %d on lev %d", b->num, b->nxt, b->lev ); if ( b->typ != bt->typ || b->key != bt->key || b->col != bt->col ) return log_msg( ERR_TRASH, "bad sig 0x%02x,%d,%d in block %d want 0x%02x,%d,%d", b->typ, b->key, b->col, b->num, bt->typ, bt->key, bt->col ); if ( b->ent > bt->bsz / LBT_ENTSIZE || b->stk > bt->bsz || b->stk < eod ) return log_msg( ERR_TRASH, "bad ent/stk %d,%d in block %d eod %d sz %d", b->ent, b->stk, b->num, eod, bt->bsz ); for ( i=0; ient; i++ ) { Dict *d = &b->dct[i]; pos = LBT_POS( d ); vln = LBT_VLN( d ); key = d->kln; siz = key + (b->lev ? vln+4u : vln * bt->vsz); if ( pos < b->stk || end != (unsigned)pos + siz ) return log_msg( ERR_TRASH, "bad pos/vln/key %d/%d/%d in entry %d of block %d" " stack %d sz %d end %d expected %d", pos, vln, key, i, b->num, b->stk, siz, pos+siz, end ); if ( b->lev ) { unsigned lnk; if ( vln && vln != +bt->vsz ) return log_msg( ERR_TRASH, "bad vln %d in entry %d of block %d vsz %d", vln, i, b->num, bt->vsz ); lnk = GETINT( base + pos + key + vln ); if ( lnk >= bt->len ) return log_msg( ERR_TRASH, "bad link %u in entry %d of block %d", lnk, i, b->num ); } #if 1 && ! defined( NDEBUG ) if ( i ) { unsigned short op = LBT_POS( d-1 ), okl = d[-1].kln; int mc = memcmp( base+pos, base+op, keylev && vln ) { /* compare values */ unsigned ovln = d[-1].vln; if ( ! ovln || 0 < (j = memcmp( base+pos+key, base+op+okl, vln )) ) goto ok; } return log_msg( ERR_TRASH, "bad key order entry %d of block %d lev %d" " mc %d j %d key %d okl %d vln %d", i, b->num, b->lev, mc, j, key, okl, vln ); } ok: #endif tot += siz; end = pos; } if ( tot > bt->bsz - b->stk ) return log_msg( ERR_TRASH, "bad total entry size %d in block %d stack %d", tot, b->num, b->stk ); #ifndef NDEBUG if ( LOG_DO( LOG_TRACE ) ) { log_msg( LOG_TRACE, "block %d lev %d sig 0x%02x,%d,%d ent %d eod %d stk %d", b->num, b->lev, b->typ, b->key, b->col, b->ent, eod, b->stk ); for ( i=0; i<2; i++ ) { int j = i ? b->ent-1 : 0; Dict *d = &b->dct[j]; if ( i >= b->ent ) /* if b->ent is 0 or 1, log no or one key */ break; pos = LBT_POS( d ); vln = LBT_VLN( d ); key = d->kln; end = pos + b->lev ? vln+4u : vln * bt->vsz; /* start of value */ if ( b->lev ) log_msg( LOG_TRACE, "key %d '%.*s'/%d -> %d", j, key, base+pos, vln, GETINT(base+pos+key+vln) ); else { Hit h; rdval( &h, base+pos+key, bt->vsz ); log_msg( LOG_TRACE, "key %d '%.*s'[%d] = %d[%d,%d,%d]", j, key, base+pos, vln, h.mfn, h.tag, h.occ, h.pos ); } } } #endif #endif return 0; } /* chkblk */ static Block *addchunk ( Idx *bt, unsigned cnt ) { unsigned ibsz = LBT_BOFF + bt->bsz + LBT_GUARD; /* internal block size */ /* increase cnt so it's > 0 and add one extra to return */ char *mem = (char *)mAlloc( sizeof(Chunk) + (1 + ++cnt)*ibsz ); Chunk *chk = (Chunk *)mem; if ( ! mem ) return 0; /* chain the chunk */ chk->nxt = bt->mem; bt->mem = chk; mem += sizeof(Chunk); bt->clen += cnt; /* add cnt blocks to level 0 lru list */ if ( ! bt->mru[0] ) /* level 0 cache is empty */ bt->mru[0] = (Block *)mem; for ( ; cnt--; mem += ibsz ) { Block *b = (Block *)mem; iniblk( b, bt, 0 , 0 ); /* initialise empty block */ b->nru = bt->lru[0]; bt->lru[0] = b; } iniblk( (Block *)mem, bt, 0 , 0 ); return (Block *)mem; } /* addchunk */ static int readblk ( Block *b, Idx *bt, unsigned num ) { int got = lio_pread( &bt->fd, LBT_BASE(b), bt->bsz, num*bt->bsz ); log_msg( LOG_VERBOSE, "reading block %u got %u", num, got ); if ( bt->bsz != (unsigned)got ) return log_msg( LOG_SYSERR, "\twhen reading bt block %d", num ); if ( b->num != num ) return log_msg( ERR_TRASH, "bad num %d in block %d", b->num, num ); return chkblk( b, bt ); } /* readblk */ static void unlock ( Block *b ) { b->lck = 0; if ( b->wlk ) (void)LIO_WAKE( b->num ); } #define ULK( b ) do { \ if ( (b)->lck ) { (b)->lck = 0; if ( b->wlk ) unlock( b ); } \ } while (0) static int writeblk ( Block *b, Idx *bt, int ulk ) { int got; if ( chkblk( b, bt ) ) /* don't write no shize */ return log_msg( ERR_TRASH, "\tbefore writing bt block %d", b->num ); got = lio_pwrite( &bt->fd, LBT_BASE(b), bt->bsz, b->num*bt->bsz ); log_msg( LOG_VERBOSE, "writing block %u got %u", b->num, got ); if ( ulk ) ULK( b ); if ( bt->bsz != (unsigned)got ) return log_msg( LOG_SYSERR, "\twhen writing bt block %d", b->num ); return 0; } /* writeblk */ #define WRULK( b, bt ) writeblk( b, bt, 1 ) #define NEWBLK( bt, lev, lock ) getblk( bt, bt->len, lev, lock ) /** get block #num. 1) locate block #num in cache. 2) if it is found: if it is READING, or we want to lock and it is locked, incr # waiters, wait for it. 3) if it is not found: locate a free cache entry, extending the cache if necessary, set the num, set the READING flag, go read it. after reading, if there are waiters, wake them up. since we are the first to get a hold on the block, we may write lock it, if we want. 4) return the block. */ static Block *getblk ( Idx *bt, unsigned num, unsigned char lev, lbt_get lock ) { Block *b; stat.get++; /* dirty */ if ( !(stat.get & 0x3fff) ) { /* every 16K */ static unsigned lmiss; log_msg( LOG_INFO, "searched %i keys in %u blks cache %d %d%%" " depth %u cmp/get/miss %u/%u/%u %d%% (%d %d%%)", stat.key, bt->len, bt->clen, bt->clen*100 / bt->len, bt->root->lev, stat.cmp, stat.get, stat.miss, stat.miss*100 / stat.get, (stat.miss - lmiss), (stat.miss - lmiss)*100 / 0x4000 ); lmiss = stat.miss; } startover: /* only used on absurd congestion */ b = 0; LOG_DBG( LOG_DEBUG, "get bt block %u %u %d", num, lev, lock ); if ( ! num ) b = bt->root; else if ( num < bt->len ) { b = bt->hash[ num % bt->hlen ]; while ( b && num != b->num ) b = b->nhs; } if ( b ) while ( b->ird || (lock && b->lck) ) { int wrd = b->ird; int wlk = lock && b->lck; LOG_DBG( LOG_VERBOSE, "lck on block %u %u r%d l%d", num, lev, b->ird, b->lck ); if ( LBT_EX <= lock ) { log_msg( LOG_ERROR, "concurrent mod on EX lock" ); return 0; } if ( ! lio_lock ) { log_msg( LOG_FATAL, "bad MT environment: concurrent mod, no lock" ); return 0; } if ( wrd && 0xff != b->wrd ) b->wrd++; if ( wlk && 0xff != b->wlk ) b->wlk++; (void)LIO_WAIT( num ); /* zzzzz .... sleep tight */ /* some hours later */ if ( num != b->num ) { log_msg( LOG_WARN, "heavy congestion: block %d has gone", num ); goto startover; } if ( wrd && b->wrd ) b->wrd--; if ( wlk && b->wlk ) b->wlk--; } else { /* not found or new block */ /* find free block */ int i = 0, llev, err = 0; stat.miss++; /* dirty */ LOG_DBG( LOG_DEBUG, "bt block %u %u not found", num, lev ); for ( ; i < (int)LBT_LRULEV; i++ ) { /* check lru list for level i */ Block **bp = &bt->lru[i]; while ( (b = *bp) && LBT_BUSY(b) ) bp = &b->nru; if ( ! b ) /* try next level */ continue; *bp = b->nru; /* dequeue */ if ( bt->mru[i] == b ) /* was the tail */ bt->mru[i] = 0; break; } if ( ! b ) { log_msg( LOG_WARN, "no free blocks in cache" ); if ( !(b = addchunk( bt, 64 )) ) return 0; /* TODO: should we also extend the hash ? */ } else if ( b->num ) { /* got block -- should be hashed somewhere */ Block **bp = &bt->hash[ b->num % bt->hlen ]; while ( *bp && *bp != b ) bp = &(*bp)->nhs; if ( *bp == b ) *bp = b->nhs; /* dehash */ /* else ... hmmm should have found it */ } assert( ! LBT_BUSY( b ) ); iniblk( b, bt, num, lev ); /* hash it, so waiters can find it */ b->nhs = bt->hash[ b->num % bt->hlen ]; bt->hash[ b->num % bt->hlen ] = b; if ( num == bt->len ) { /* extend */ b->lev = lev; bt->len++; /* TODO: should be configurable */ if ( bt->len > bt->clen << 5 && bt->clen < 0x4000 ) addchunk( bt, bt->clen / 4 ); } else { b->ird = 1; (void)LIO_RELE(); err = readblk( b, bt, num ); (void)LIO_LOCK(); b->ird = 0; if ( ! err && lev != b->lev ) err = log_msg( ERR_TRASH, "bad lev %d in block %d want %d", b->lev, b->num, lev ); if ( err ) { Block **bp = &bt->hash[ num % bt->hlen ]; while ( *bp && *bp != b ) bp = &(*bp)->nhs; if ( *bp == b ) *bp = b->nhs; /* dehash */ b->lev = 0; /* add to level 0 lru list */ b->num = 0; /* tell others that it isn't the expected block */ } } /* add to it's levels lru list */ llev = LBT_LRULEV > b->lev ? b->lev : LBT_LRULEV-1; if ( bt->mru[ llev ] ) bt->mru[ llev ]->nru = b; else bt->lru[ llev ] = b; (bt->mru[ llev ] = b)->nru = 0; if ( b->wrd ) (void)LIO_WAKE( num ); if ( err ) { b->wrd = b->wlk = 0; /* mark block unused */ return 0; } } /* got free block */ if ( LBT_WR == lock ) { if ( b->lck ) { log_msg( LOG_FATAL, "bt consi: unexpected lock on %u", num ); return 0; } b->lck = 1; } return b; } /* getblk */ /** binsearch position for key. position is either the index of the matching entry (0..b->ent-1) or the index (0..b->ent) where key should be inserted. @return position for key (0..b->ent), |ifyed with LBT_MATCH, if found */ static unsigned short keypos ( #if LBT_USECMP Idx *bt, #define KEYPOS( bt, b, key ) keypos( bt, b, key ) #else #define KEYPOS( bt, b, key ) keypos( b, key ) #endif Block *b, Key *key ) { unsigned char *base = LBT_BASE( b ); unsigned char *kbt = key->byt; unsigned char kln = key->len; unsigned short l = 0, r = b->ent; Dict *d = b->dct; /* binsearch invariant: key is < all entries i >= r (we do know relation at r) key is > all entries i < l (don't know relation at l) */ while ( l < r ) { /* invariant: l <= keypos < r */ unsigned short i = (l + r) >> 1; /* l <= i < r */ unsigned short pos = LBT_POS(&d[i]); unsigned char cln = d[i].kln; int cmp = #if LBT_USECMP bt->cmp ? bt->cmp( kbt, base+pos, klnlev /* ignore on leaves -- no duplicate keys here */ || (!d[i].vln && !key->val.len) /* no value */ ) return LBT_MATCH | i; /* so we're on an inner node */ if ( ! d[i].vln ) /* common case */ cmp = 1; else if ( ! key->val.len ) cmp = -1; else { unsigned char vln = LBT_VLN(&d[i]); int c = key->val.len < d[i].vln ? key->val.len : vln; if ( !(cmp = memcmp( key->val.byt, base+pos+cln, c )) && !(cmp = (int) key->val.len - vln) ) return LBT_MATCH | i; } } if ( cmp > 0 ) l = i+1; else r = i; } return l; /* == r */ } /* keypos */ /** find position for (leaf) val. search v in array b of n values of size sz @return position for value (0..n), |ifyed with LBT_MATCH, if found */ static unsigned short valpos ( unsigned char *b, unsigned short n, unsigned char sz, unsigned char *v ) { unsigned short l = 0, r = n; /* binsearch invariant: v is < all values i >= r (we do know relation at r) v is > all values i < l (don't know relation at l) */ while ( l < r ) { /* invariant: l <= valpos < r */ unsigned short i = (l + r) >> 1; /* l <= i < r */ int cmp = memcmp( v, b+i*sz, sz ); if ( ! cmp ) return LBT_MATCH | i; if ( cmp > 0 ) l = i+1; else r = i; } return l; /* == r */ } /* valpos */ #ifndef NDEBUG static void TRACEKEY ( unsigned lev, FindKey *f ) { int i; unsigned n = 0; Hit h; PrVal pv; if ( LOG_DONT( lev ) ) return; log_msg( lev, "key '%.*s' vln %d val %s", f->key->len, f->key->byt, f->key->val.len, PRVAL( pv, f->key->val ) ); if ( f->key->val.len ) { rdval( &h, f->key->val.byt, f->key->val.len ); log_msg( lev, "\tdb %d mfn %d tag %d occ %d pos %d", h.dbn, h.mfn, h.tag, h.occ, h.pos ); } for ( i=LBT_LEVELS; i--; ) { Block *b = f->b[i]; unsigned on = n, e; Dict *d; if ( ! b ) continue; d = b->dct + (e = (LBT_ENTMASK & f->e[i])); if ( i ) n = GETINT( LBT_BASE(b)+d->pos+d->kln+d->vln ); log_msg( lev, "lev %d n %d e 0x%x = %d %c '%.*s' val %d %s -> %d%s", i, f->n[i], f->e[i], e, (LBT_MATCH & f->e[i]) ? 'M' : '-', d->kln, LBT_BASE(b)+d->pos, d->vln, prval( pv, LBT_BASE(b)+d->pos+d->kln, d->vln ), n, on == b->num ? "" : " LY!" ); if ( d->vln ) { rdval( &h, LBT_BASE(b)+d->pos+d->kln, d->vln ); log_msg( lev, "\tdb %d mfn %d tag %d occ %d pos %d", h.dbn, h.mfn, h.tag, h.occ, h.pos ); } if ( e ) { d--; log_msg( lev, "\tleft '%.*s' val %d %s", d->kln, LBT_BASE(b)+d->pos, d->vln, prval( pv, LBT_BASE(b)+d->pos+d->kln, d->vln ) ); if ( d->vln ) { rdval( &h, LBT_BASE(b)+d->pos+d->kln, d->vln ); log_msg( lev, "\tdb %d mfn %d tag %d occ %d pos %d", h.dbn, h.mfn, h.tag, h.occ, h.pos ); } } } } /* TRACEKEY */ #else #define TRACEKEY( l, f ) (void)(f) #endif static int findkey ( FindKey *f, lbt_get lock ) { int i; Block *b = f->bt->root; unsigned char *pkey = 0, pkln = 0, pvln = 0; if ( b->lev >= LBT_LEVELS ) return log_msg( ERR_TRASH, "tree is too deep: %u", b->lev ); stat.key++; for ( i=LBT_LEVELS; i--; ) { /* paranoia */ f->b[i] = 0; f->n[i] = f->e[i] = 0; } for ( i=b->lev;; ) { unsigned short e; unsigned n; Dict *d; #ifndef NDEBUG chkblk( b, f->bt ); #endif f->n[i] = (f->b[i] = b)->num; f->e[i] = e = KEYPOS( f->bt, b, f->key ); LOG_DBG( LOG_DEBUG, "keypos %x", e ); #ifndef NDEBUG if ( pkey && b->ent ) { /* compare parent key against our first key pkey must be equal for inner node and not greater for leave */ int cmp; /* this - pkey >= 0 */ unsigned char *ent, ln; d = b->dct; ent = LBT_BASE(b)+LBT_POS(d); ln = pkln < d->kln ? pkln : d->kln; cmp = ln ? memcmp( ent, pkey, ln ) : 0; if ( ! cmp ) cmp = (int)d->kln - (int)pkln; if ( ! cmp ) { if ( ! d->vln || ! pvln ) cmp = (int)d->vln - (int)pvln; else cmp = memcmp( ent+d->kln, pkey+pkln, f->bt->vsz ); } if ( cmp && (i || 0 > cmp) ) { Hit h; PrVal pv; TRACEKEY( ERR_TRASH, f ); if ( d->vln ) { rdval( &h, ent+d->kln, d->vln ); log_msg( ERR_TRASH, "\tfirst val db %d mfn %d tag %d occ %d pos %d", h.dbn, h.mfn, h.tag, h.occ, h.pos ); } return log_msg( ERR_TRASH, "\tbad first '%.*s' val %d %s", d->kln, ent, d->vln, prval( pv, ent+d->kln, d->vln ) ); } } #endif /* here goes the Lehmann/Yao race detection: If we ran on right boundary, follow right link */ while ( !(LBT_MATCH & e) && b->ent == e && b->nxt && b->num ) { Block *s; unsigned se; b->lck = LBT_WR; /* hold b while checking s */ s = getblk( f->bt, b->nxt, i, i ? LBT_RD : lock ); if ( i || ! lock ) ULK( b ); else b->lck = lock; se = KEYPOS( f->bt, s, f->key ); if ( !se ) { /* lower than first */ ULK( s ); break; } LOG_DBG( LOG_DEBUG, "LYao %d -> %d lev %d for '%.*s' got 0x%x", b->num, s->num, i, f->key->len, f->key->byt, se ); /* trade b for s */ ULK( b ); f->n[i] = (f->b[i] = b = s)->num; e = se; } f->e[i] = e; if ( ! i ) return 0; if ( !(LBT_MATCH & e) ) { if ( e ) f->e[i] = --e; else { TRACEKEY( ERR_TRASH, f ); return log_msg( ERR_IDIOT, "ran on left bound of block %u ent %u", b->num, b->ent ); } } d = b->dct + (LBT_ENTMASK & e); /* direct access ok in inner node */ pkey = LBT_BASE(b)+d->pos; pkln = d->kln; pvln = d->vln; n = GETINT( pkey+pkln+pvln ); i--; b = getblk( f->bt, n, i, i ? LBT_RD : lock ); } return -1; /* go away */ } /* findkey */ static int mktree ( Idx *bt ) { Batch *ba = bt->bat; unsigned char lev = 0; /* the level to be indexed */ unsigned left = 1; /* block on child level */ /* stats */ unsigned size = bt->len * bt->bsz; unsigned sval = ba->vals * bt->vsz; unsigned sove = bt->len * LBT_DOFF + sizeof(Dict)*(ba->keys+ba->span); unsigned iblk = 0; unsigned ikey = 0; unsigned ikeyl = 0; unsigned ipfx = 0; unsigned ispn = 0; log_msg( LOG_INFO, "end batch:\n" "\t#blks %u total len %u free %u (%.1f%%) overhead %u (%.1f%%)\n" "\t#keys %u total len %u (%.1f%%) avg %.1f max %u\n" "\t#vals %u total len %u (%.1f%%) avg %.1f max %u spans %u", bt->len, size, ba->free, ba->free*100./size, sove, sove*100./size, ba->keys, ba->keyl, ba->keyl*100./size, (float)ba->keyl/ba->keys, ba->maxl, ba->vals, sval, sval*100./size, (float)ba->vals/ba->keys, ba->maxv, ba->span ); /* build tree */ for (;;) { /* levels */ unsigned len = 0; /* current is n'th block on level */ Block *b = 0; unsigned have = 0; /* room in current block b */ unsigned char*base = 0; /* of current block b */ Dict *d = 0; /* of current block b */ unsigned next = left; /* child */ left = bt->len; /* get child c and create child pointer in block b */ while ( next ) { unsigned char vln = 0; unsigned char *key; unsigned char kln; unsigned need; Block *c = getblk( bt, next, lev, LBT_EX ); if ( ! c ) return log_msg( ERR_NOMEM, "no child block" ); next = c->nxt; if ( ! c->ent ) /* OOPS empty node */ continue; key = LBT_BASE(c)+LBT_POS(c->dct); kln = c->dct[0].kln; if ( ! b ) /* the leftmost child on this level */ kln = 0; /* always 0 key on left */ else if ( lev ) vln = LBT_VLN(c->dct); else if ( ba->key.len == kln /* compare key to last of prev. */ && !memcmp( ba->key.byt, key, kln ) /* which was saved in ba->key */ ) { /* key spans blocks */ LOG_DBG( LOG_DEBUG, "found span key '%.*s' child %u", kln, key, c->num ); vln = bt->vsz; /* qualify with value */ } else { #if 1 unsigned char x = 0; /* find prefix key */ if ( kln > ba->key.len ) kln = ba->key.len; while ( x < kln && key[x] == ba->key.byt[x] ) x++; if ( x < kln ) /* they differ at [x] */ kln = x+1; else /* equal on the shorter length */ kln++; if ( kln > c->dct[0].kln ) return log_msg( ERR_TRASH, "bad key '%.*s' on level %u matching previous of length %u", c->dct[0].kln, key, lev, ba->key.len ); LOG_DBG( LOG_DEBUG, "prefix on key '%.*s' child %u save %u", kln, key, c->num, c->dct[0].kln - kln ); ipfx += c->dct[0].kln - kln; #endif } if ( vln ) ispn++; ikeyl += kln; need = sizeof(Dict) + kln + vln + 4; if ( need > have ) { if ( b ) { log_msg( LOG_VERBOSE, "level %u block %u is full %u ent stack %u", lev, b->num, b->ent, b->stk ); b->nxt = bt->len; /* link */ b->lck = 0; if ( WRULK( b, bt ) ) goto nowrite; len++; } c->lck = LBT_EX; /* paranoia */ b = NEWBLK( bt, lev+1, LBT_EX ); c->lck = 0; if ( ! b ) return log_msg( ERR_NOMEM, "no free block" ); if ( b->ent || b->stk != bt->bsz ) return log_msg( ERR_TRASH, "new block %u is not empty: ent %u stk %u", b->num, b->ent, b->stk ); base = LBT_BASE( b ); b->lck = LBT_EX; /* lock */ have = bt->bsz - LBT_DOFF; /* - ba->let; TODO: should we ??? */ d = b->dct; len++; iblk++; } SETINT( base+(b->stk -= 4), c->num ); d->kln = kln; d->vln = vln; if ( kln += vln ) /* key + consecutive qualifying value */ memcpy( base+(b->stk -= kln), key, kln ); d->pos = b->stk; d++; b->ent++; ikey++; have -= need; /* save the last dance for me */ memcpy( ba->key.byt, LBT_BASE(c)+LBT_POS(&c->dct[c->ent-1]), ba->key.len = c->dct[c->ent-1].kln ); } if ( !len ) return log_msg( ERR_TRASH, "level %u was empty", lev ); /* close level */ log_msg( LOG_INFO, "closing level %u with %u blocks %u..%u", lev, len, left, b->num ); b->lck = 0; if ( 1 == len ) { memcpy( bt->root, b, LBT_BOFF + bt->bsz ); b = bt->root; b->num = 0; } if ( WRULK( b, bt ) ) goto nowrite; lev++; if ( 2 > len ) break; } log_msg( LOG_INFO, "\tused %u inner blocks %u keys %u bytes avg %.1f\n" "\t%u #spans (%.1f%%) %u pfx (- %.1f%%)", iblk, ikey, ikeyl, (float)ikeyl/ikey, ispn, ispn*100.*bt->vsz/ikeyl, ipfx, ipfx*100./(ipfx+ikeyl) ); mFree( bt->bat ); bt->bat = 0; return 0; nowrite: return log_msg( LOG_SYSERR, "\twhen writing batch block" ); } /* mktree */ /* ************************************************************ package data */ /* ************************************************************ package functions */ int lbt_batch ( Idx *bt, unsigned char pctfree ) { bt->bat = (Batch*)mAlloc( sizeof(Batch) ); if ( ! bt->bat ) return log_msg( ERR_NOMEM, "\twhen getting batch" ); memset( bt->bat, 0, sizeof(Batch) ); if ( pctfree > 50 ) pctfree = 50; bt->bat->let = pctfree * bt->bsz / 100; lio_trunc( &bt->fd, 0 ); iniblk( bt->root, bt, 0, 0 ); /* TODO: kill hash */ bt->len = 1; bt->bat->cur = NEWBLK( bt, 0, LBT_EX ); return 0; } /* lbt_batch */ int lbt_batchval ( Idx *bt, Key *key ) { int cmp = 0; Batch *ba = bt->bat; if ( key != &ba->key ) { int l = key->len < ba->key.len ? key->len : ba->key.len; /* min */ if ( ! (cmp = memcmp(key->byt, ba->key.byt, l)) ) cmp = (int)key->len - (int)ba->key.len; if ( 0 > cmp ) return log_msg( LOG_ERROR, "batch key '%.*s' < last '%.*s'", key->len, key->byt, ba->key.len, ba->key.byt ); } /* flush ? */ while ( ! key->val.len /* end */ || LBT_ENTMAX == ba->got /* full */ || cmp /* next key */ ) { Block *b = ba->cur; unsigned have = b->stk - LBT_DOFF - b->ent*sizeof(Dict) - ba->let; unsigned need = sizeof(Dict) + ba->key.len; unsigned want = need + ba->got * bt->vsz; unsigned take = ba->got; if ( have < want ) { /* you can't always get ... */ if ( have <= need ) take = 0; else { take = (have - need) / bt->vsz; if ( 256 > take * bt->vsz ) take = 0; } } #if 0 /* prefix key turbo ??? -- doesn't seem to work all too well */ if ( have < (bt->bsz >> 1) ) { /* ready to waste half a block ? */ Dict *d = &b->dct[ b->ent-1 ]; unsigned char *last = LBT_BASE(b)+LBT_POS(d); if ( ba->key.byt[0] != last[0] || ba->key.byt[1] != last[1] || ba->key.byt[2] != last[2] ) take = 0; } #endif LOG_DBG( LOG_DEBUG, "batch '%.*s' got %u need %u want %u have %u take %u", ba->key.len, ba->key.byt, ba->got, need, want, have, take ); /* we have enough room in b for take entries */ if ( take ) { unsigned char *base = LBT_BASE( b ); unsigned char *s = base + b->stk; int i = take; Dict *d = &b->dct[ b->ent++ ]; while ( i-- ) memcpy( s -= bt->vsz, ba->val[i].byt, bt->vsz ); memcpy( s -= ba->key.len, ba->key.byt, ba->key.len ); b->stk = (unsigned short) (s - base); LBT_PV( d, b->stk, take ); d->kln = ba->key.len; if ( ba->got > take ) memcpy( ba->val, ba->val+take, sizeof(ba->val[0])*(ba->got - take) ); ba->got -= take; } if ( ! ba->got ) break; if ( take ) /* key spans blocks */ ba->span++; ba->free += b->stk - LBT_DOFF - b->ent*sizeof(Dict); /* wasted */ b->nxt = bt->len; /* link */ LOG_DBG( LOG_DEBUG, "batch: %u entries left, getting block %u", ba->got, bt->len ); if ( WRULK( b, bt ) ) return log_msg( LOG_SYSERR, "\twhen writing batch block" ); /* so couldn't flush all, thus we need another block */ b = ba->cur = NEWBLK( bt, 0, LBT_EX ); /* will continue on end or keychange */ if ( b->ent || b->stk < bt->bsz ) return log_msg( ERR_TRASH, "new block %u is not empty: ent %u stk %u", b->num, b->ent, b->stk ); LOG_DBG( LOG_VERBOSE, "new block %u has %u ent %u stk, will keep %u", b->num, b->ent, b->stk, ba->let ); } if ( cmp || ! key->val.len ) /* done with key: stats */ if ( ba->maxv < ba->tot ) ba->maxv = ba->tot; if ( key->val.len ) { if ( cmp ) { /* set new key */ if ( ba->got ) return log_msg( ERR_IDIOT, "batch not empty on new key" ); ba->tot = 0; memcpy( ba->key.byt, key->byt, ba->key.len = key->len ); ba->keys++; ba->keyl += key->len; if ( ba->maxl < key->len ) ba->maxl = key->len; } /* add new val */ /* LOG_HEX( key->val.byt, bt->vsz ); */ memcpy( ba->val[ ba->got++ ].byt, key->val.byt, bt->vsz ); ba->vals++; ba->tot++; } else { WRULK( ba->cur, bt ); mktree( bt ); } return 0; } /* lbt_batchval */ /* TODO: this function is an awful mess and needs complete rewrite, KR */ int lbt_add ( Idx *bt, Key *key ) { int ret = 0; unsigned short e, i = 0; Block *b, *s = 0; /* mainblock, sibling */ Dict *d; unsigned char *base, *sbase, *src, *v, *stk; FindKey f; int hadkey, split = 0, insibling = 0; unsigned need; PrVal pv; f.bt = bt; f.key = key; if ( bt->key && key->len > bt->key ) key->len = bt->key; if ( findkey( &f, LBT_WR ) ) { if ( f.b[0] ) ULK( f.b[0] ); b = 0; LOG_OTO( done, ( ERR_TRASH, "could not find key" ) ); } e = (LBT_ENTMASK & f.e[0]); b = f.b[0]; base = LBT_BASE(b); d = b->dct + e; if ( ! (hadkey = LBT_MATCH & f.e[0]) ) /* new key */ need = key->len + sizeof(Dict) + bt->vsz; else { unsigned vln = LBT_VLN(d); v = base+LBT_POS(d)+d->kln; if ( 1 == vln ) { /* check stopword -- value entirely zeroed */ int l = bt->vsz; unsigned char *C = v; while ( l-- && !*C++ ) /* huh huh, he said "C++" */ ; if ( -1 == l ) { LOG_DBG( LOG_VERBOSE, "key '%.*s' is a stopword", key->len, key->byt ); goto done; } } i = valpos( v, vln, bt->vsz, key->val.byt ); if ( LBT_MATCH & i ) { LOG_DBG( LOG_DEBUG, "adding key '%.*s'/%s: already there pos %u", key->len, key->byt, PRVAL( pv, key->val ), i & ~LBT_MATCH ); goto done; /* value already there */ } need = bt->vsz; } LOG_DBG( LOG_DEBUG, "adding key '%.*s'/%s in block %d had %d ent %d stk %d need %d", key->len, key->byt, PRVAL( pv, key->val ), b->num, hadkey, b->ent, b->stk, need ); if ( need > b->stk - LBT_DOFF - b->ent*sizeof(Dict) ) { /* now that's bad: not enough space create new block s, move last entries there starting from the end, entries (including the new one) are collected, until their combined size reaches more than half a block. If the combined size then is more than 2/3 a block, and the last entry alone takes more than 1/3 a block, it is split. TODO: check existing sibling for free space */ unsigned half = (bt->bsz - LBT_DOFF) >> 1; unsigned tot = 0; unsigned mv = b->ent; LOG_DBG( LOG_DEBUG, "splitting block %d need %d have %d (stk %d ent %d)", b->num, need, b->stk - LBT_DOFF - b->ent*sizeof(Dict), b->stk, b->ent ); if ( need > (bt->bsz - LBT_DOFF) / 3 ) LOG_OTO( done, (ERR_IDIOT, "OOPS! excessive need for %d bytes", need) ); /* figure out split position */ if ( !hadkey && b->ent == e ) { tot = need; insibling = !0; } d = b->dct+mv; while ( d--, mv-- ) { unsigned sz = sizeof(Dict)+d->kln+LBT_VLN(d)*bt->vsz; if ( half > (tot += sz) ) { /* sz fits */ if ( e == mv && !hadkey ) { /* check size of new entry. In case of hadkey, we don't bother for the bt->vsz bytes. */ if ( half < (tot + need) ) { if ( (half - tot) > need/2 ) { /* more balanced if we move */ tot += need; insibling = !0; } break; } tot += need; insibling = !0; } if ( mv ) continue; /* else croak on ! mv below */ } /* exceeded half -- maybe we move entry mv anyway */ if ( tot > (bt->bsz - LBT_DOFF)*2/3 ) { if ( tot-sz < (bt->bsz - LBT_DOFF)/3 ) split = !0; else { /* undo -- don't move this entry */ tot -= sz; d++; mv++; } } else if ( ! mv ) LOG_OTO( done, (ERR_IDIOT, "OOPS! didn't find end tot %d bytes", tot) ); break; } /* while mv-- */ if ( mv < e || (mv == e && hadkey && !split) ) insibling = !0; /* now mv is the first entry to move */ s = NEWBLK( bt, 0, LBT_WR ); LOG_DBG( LOG_DEBUG, "split: move %d of %d to new blk %d e %d tot %d insibling %d split %d", mv, b->ent, s->num, e, tot, insibling, split ); s->nxt = b->nxt; b->nxt = s->num; sbase = LBT_BASE(s); if ( split ) { /* now total includes full size of entry mv. if we included a new entry, it might even exceed bsz. calculate sizes for both blocks w/o mv's values. */ unsigned esz = sizeof(Dict)+d->kln; /* entry's base cost */ unsigned vln = LBT_VLN(d); unsigned keep, diff, mvval, kpval, mvsiz, pos = LBT_POS(d); /* from currently used space, substract tot (including all values and possibly new key "need"), but add entry base cost. that should be the size remaining if we where moving all of the values but keep an entry and key. maybe slightly biased if need is bt->vsz (the hadkey case). */ keep = bt->bsz - b->stk + b->ent*sizeof(Dict) + need - tot + esz; tot -= vln * bt->vsz; /* undo entries */ diff = (keep > tot ? keep - tot : tot - keep) / bt->vsz; if ( diff > vln ) LOG_OTO( done, (ERR_IDIOT, "OOPS! got diff %d vln %d entry %d tot %d keep %d", diff, vln, mv, tot, keep) ); mvval = (vln - diff)/2; if ( keep > tot ) mvval += diff; kpval = vln - mvval; if ( ! mvval || ! kpval ) { LOG_DBG( LOG_WARN, "bad mv/kpval %d/%d keep %d tot %d vln %d diff %d", mvval, kpval, keep, tot, vln, diff ); if ( ! mvval ) kpval = vln - (mvval = 1); else mvval = vln - (kpval = 1); } tot += mvsiz = mvval*bt->vsz; keep += kpval*bt->vsz; src = base+pos; LOG_DBG( LOG_DEBUG, "split: entry %d '%.*s' mv %d of %d values, i %d keep %d tot %d", mv, d->kln, src, mvval, vln, i, keep, tot ); if ( keep > (bt->bsz-LBT_DOFF) || tot > (bt->bsz-LBT_DOFF) ) LOG_OTO( done, (ERR_IDIOT, "OOPS! got diff %d vln %d mvval %d entry %d tot %d keep %d", diff, vln, mvval, mv, tot, keep) ); /* SIGH -- do it */ memcpy( sbase + (s->stk -= mvsiz), src+d->kln+kpval*bt->vsz, mvsiz ); if ( e == mv && hadkey && i > kpval ) { insibling = !0; s->stk -= bt->vsz; i -= kpval; memmove( sbase+s->stk, sbase+s->stk+bt->vsz, i*bt->vsz ); memcpy( sbase+s->stk+i*bt->vsz, key->val.byt, bt->vsz ); mvval++; } memcpy( sbase + (s->stk -= d->kln), src, d->kln ); LBT_PVK( s->dct, s->stk, mvval, d->kln ); s->ent = 1; memmove( src + mvsiz, src, d->kln+kpval*bt->vsz ); pos += mvsiz; LBT_PV( d, pos, kpval ); d++; mv++; } /* now entries mv..b->ent are moved, mv > 0 */ b->stk = LBT_POS(b->dct+mv-1); /* pre-adjust stack */ if ( mv < e ) { /* move entries mv..e-1 */ /* used stack space is from pos of entry e-1 to end of entry mv (might differ from pos of mv-1 due to split) */ unsigned from = LBT_POS(b->dct+e-1), nent = e-mv, sz = LBT_POS(b->dct+mv)+b->dct[mv].kln +LBT_VLN(b->dct+mv)*bt->vsz - from; int adj = (s->stk -= sz) - from; /* entry position adjustment */ memcpy( sbase + s->stk, base+from, sz ); memcpy( s->dct+s->ent, b->dct+mv, nent*sizeof(Dict) ); d = s->dct + s->ent; s->ent += nent; for ( ; nent--; d++ ) LBT_POSADD( d, adj ); } if ( insibling && mv <= e ) { /* now mv or insert e, unless done on split */ d = b->dct + e; if ( hadkey ) { /* mv existing entry e >= mv, add new val */ /* i is valpos */ unsigned vln = LBT_VLN(d); src = base+LBT_POS(d)+d->kln; if ( i < vln ) { unsigned sz = (vln - i)*bt->vsz; memcpy( sbase + (s->stk -= sz), src + i*bt->vsz, sz ); } memcpy( sbase + (s->stk -= bt->vsz), key->val.byt, bt->vsz ); if ( i ) { unsigned sz = i*bt->vsz; memcpy( sbase + (s->stk -= sz), src, sz ); } memcpy( sbase + (s->stk -= d->kln), base+LBT_POS(d), d->kln ); LBT_PVK( s->dct + s->ent, s->stk, vln+1, d->kln ); e++; /* e is moved */ } else { memcpy( sbase + (s->stk -= bt->vsz), key->val.byt, bt->vsz ); memcpy( sbase + (s->stk -= key->len), key->byt, key->len ); LBT_PVK( s->dct + s->ent, s->stk, 1, key->len ); } s->ent++; } /* mv or insert e */ if ( e < b->ent && mv < b->ent ) { /* move entries max(e,mv) to end */ unsigned fst = mv < e ? e : mv; /* used stack space is from pos of entry b->ent-1 (the original b->stk) to end of entry fst (might differ from pos of fst-1 due to split) */ unsigned from = LBT_POS(b->dct+b->ent-1), nent = b->ent-fst, sz = LBT_POS(b->dct+fst)+b->dct[fst].kln +LBT_VLN(b->dct+fst)*bt->vsz - from; int adj = (s->stk -= sz) - from; /* entry position adjustment */ LOG_DBG( LOG_DEBUG, "split: entries %d..%d sz %d from %d to %d", fst, b->ent-1, sz, from, s->stk ); memcpy( sbase + s->stk, base+from, sz ); memcpy( s->dct+s->ent, b->dct+fst, nent*sizeof(Dict) ); d = s->dct + s->ent; s->ent += nent; for ( ; nent--; d++ ) { #if !defined(NDEBUG) unsigned p = LBT_POS(d); #endif LBT_POSADD( d, adj ); #if !defined(NDEBUG) LOG_DBG( LOG_DEBUG, "new ent %d p/v/k %d/%d/%d o %d", d - s->dct, LBT_POS(d), LBT_VLN(d), d->kln, p ); #endif } } b->ent = mv; /* kill entries */ LOG_DBG( LOG_DEBUG, "split: blks %d/%d with %d/%d entries, s starts '%.*s'", b->num, s->num, b->ent, s->ent, s->dct[0].kln, sbase+LBT_POS(s->dct) ); } /* new sibling */ if ( ! insibling ) { /* now we do have enough space */ d = b->dct + e; stk = base + b->stk; /* bottom of data stack */ if ( hadkey ) { /* insert value at i */ unsigned char *val = base + LBT_POS(d)+d->kln+i*bt->vsz; need = bt->vsz; #if 0 LOG_DBG( LOG_DEBUG, "in entry %d '%.*s' val %d: memmove( %p %p %d) val %p need %d", e, d->kln, base+LBT_POS(d), i, stk-need, stk, (val-stk)+need, val, need ); #endif memmove( stk - need, stk, (val - stk)+need ); memcpy( val - need, key->val.byt, bt->vsz ); LBT_VLNADD( d, 1 ); } else { /* insert key, value and dict entry */ /* take dict and data space from entry e */ unsigned char *nu; need = key->len + bt->vsz; if ( e == b->ent ) nu = base + b->stk - need; else { /* at end of next entry, i.e. end of block or beg of previous */ nu = base + LBT_POS(d) + d->kln + LBT_VLN(d)*bt->vsz; memmove( stk - need, stk, nu - stk ); nu -= need; memmove( d+1, d, (b->ent - e)*sizeof(Dict) ); } memcpy( nu, key->byt, key->len ); memcpy( nu + key->len, key->val.byt, bt->vsz ); LBT_PV( d, nu-base, 1 ); d->kln = key->len; d++; e++; b->ent++; } /* need is set to the needed (i.e. additionally used) data space, e is the first dict entry to redirect need bytes lower */ b->stk -= need; for ( ; e < b->ent; e++, d++ ) LBT_POSADD( d, -need ); } /* ! insibling */ while ( s ) { /* created sibling */ unsigned num = s->num; unsigned char *e0 = LBT_BASE(s) + LBT_POS(s->dct); unsigned l = s->lev + 1; Block *ins; Key k; memcpy( k.byt, e0, k.len = s->dct[0].kln ); k.val.len = 0; if ( s->lev ? s->dct[0].vln : split ) memcpy( k.val.byt, e0+k.len, k.val.len = bt->vsz ); LOG_DBG( LOG_DEBUG, "split: had %d/%d on lev %d key '%.*s' vln %d %s", b->num, num, s->lev, k.len, k.byt, k.val.len, PRVAL( pv, k.val ) ); /* get parent */ if ( f.b[l]->num == f.n[l] && ! LBT_BUSY(f.b[l]) ) f.b[l]->lck = LBT_WR; else f.b[l] = getblk( bt, f.n[l], l, LBT_WR ); LOG_DBG( LOG_DEBUG, "split: got parent %d lev %d", f.b[l]->num, f.b[l]->lev ); /* MT TODO: again L/Y link hunt */ /* ok, locked the parent -- release s and b */ if ( (ret = WRULK( s, bt )) || (ret = WRULK( b, bt )) ) break; s = 0; ins = b = f.b[l]; base = LBT_BASE(b); e = KEYPOS( bt, b, &k ); LOG_DBG( LOG_DEBUG, "split: got parent pos %d", e ); if ( LBT_MATCH & e ) /* !??? */ LOG_OTO( done, (ERR_TRASH, "OOPS! found key in parent" ) ); if ( ! e ) { TRACEKEY( ERR_TRASH, &f ); LOG_OTO( done, (ERR_TRASH, "OOPS! parent insert at 0 lev %d", l ) ); } need = sizeof(Dict) + k.len + 4 + k.val.len; if ( need > b->stk - LBT_DOFF - b->ent*sizeof(Dict) ) { unsigned half = ((bt->bsz - b->stk)+b->ent*sizeof(Dict)+need) / 2; unsigned mv = b->ent, tot = 0, nent, sz; int adj; if ( e == b->ent ) { tot = need; ins = 0; } for ( d = b->dct+mv; d--, mv--; ) { sz = sizeof(Dict) + d->kln + 4 + d->vln; if ( half > (tot += sz) ) { if ( e == mv ) { if ( half > tot+need || tot+need-half < half-tot ) { tot += need; ins = 0; /* new link goes to sibling */ } else break; if ( half <= tot ) break; } if ( mv ) continue; } if ( tot - half > sz/2 ) { /* don't move this */ tot -= sz; d++; mv++; } else if ( ! mv ) /* !??? */ mv = 1; break; } /* split here a little bit simpler */ s = NEWBLK( bt, l, LBT_WR ); s->nxt = b->nxt; b->nxt = s->num; /* straight move entries mv..b->ent-1 to sibling */ sbase = LBT_BASE(s); d = b->dct + mv; sz = b->dct[mv-1].pos - b->stk; nent = b->ent - mv; LOG_DBG( LOG_DEBUG, "split: parent lev %d split %d to new %d mv %d" " '%.*s'/%s nent %d sz %d", l, b->num, s->num, mv, d->kln, base+d->pos, prval( pv, base+d->pos+d->kln, d->vln ), nent, sz ); memcpy( sbase + (s->stk -= sz), base + b->stk, sz ); adj = s->stk - b->stk; b->stk += sz; memcpy( s->dct, d, nent*sizeof(Dict) ); b->ent = mv; s->ent = nent; for ( d = s->dct; nent--; d++ ) d->pos += adj; if ( ! ins ) { ins = s; e -= mv; } if ( (ret = chkblk( s, bt )) ) break; } /* now insert link in block ins at pos e */ { unsigned isz = k.len + 4 + k.val.len, nent = ins->ent - e; unsigned char *ibase = LBT_BASE(ins); unsigned end = e ? ins->dct[e-1].pos : bt->bsz; LOG_DBG( LOG_DEBUG, "split: ins parent %d lev %d pos %d key '%.*s' -> %d vln %d %s", ins->num, ins->lev, e, k.len, k.byt, num, k.val.len, PRVAL( pv, k.val ) ); memmove( ibase + ins->stk - isz, ibase + ins->stk, end - ins->stk ); ins->stk -= isz; memcpy( ibase + (end -= 4), &num, 4 ); if ( k.val.len ) memcpy( ibase + (end -= k.val.len), k.val.byt, k.val.len ); memcpy( ibase + (end -= k.len), k.byt, k.len ); for ( d = ins->dct+ins->ent; nent--; d-- ) { *d = d[-1]; d->pos -= isz; } d->pos = end; d->vln = k.val.len; d->kln = k.len; ins->ent++; if ( (ret = chkblk( ins, bt )) ) break; } if ( ! s ) /* write b below or in next turn */ break; if ( ! b->num ) { /* root split */ Block *bb = NEWBLK( bt, l, LBT_WR ); LOG_DBG( LOG_DEBUG, "root split!" ); /* cp all but the num */ memcpy( LBT_BASE(bb)+4, LBT_BASE(b)+4, bt->bsz-4 ); b->nxt = s->nxt; /* restore root's special nxt */ s->nxt = 0; b->stk = bt->bsz; /* empty the root */ /* add two childs bb and s */ SETINT( base + (b->stk -= 4), bb->num ); LBT_PVK( b->dct, b->stk, 0, 0 ); SETINT( base + (b->stk -= 4), s->num ); e0 = LBT_BASE(s) + LBT_POS(s->dct); if ( s->dct[0].vln ) memcpy( base + (b->stk -= bt->vsz), e0+s->dct[0].kln, bt->vsz ); memcpy( base + (b->stk -= s->dct[0].kln), e0, s->dct[0].kln ); LBT_PVK( b->dct+1, b->stk, s->dct[0].vln, s->dct[0].kln ); b->ent = 2; b->lev++; /* increase tree depth */ if ( (ret = WRULK( s, bt )) || (ret = WRULK( bb, bt )) ) break; s = 0; break; } } if ( b && !ret ) ret = WRULK( b, bt ); done: if ( s ) ULK( s ); if ( b ) ULK( b ); if ( !ret && b->lev && findkey( &f, LBT_RD ) ) ret = log_msg( ERR_TRASH, "could not find key after add" ); return ret; } /* lbt_add */ int lbt_del ( Idx *bt, Key *key ) { int ret = 0; unsigned short e, n, i, mv; Block *b; Dict *d; unsigned char *base, *v; FindKey f; PrVal pv; f.bt = bt; f.key = key; LOG_DBG( LOG_DEBUG, "deleting key '%.*s'/%s", key->len, key->byt, PRVAL( pv, key->val ) ); if ( findkey( &f, LBT_WR ) ) LOG_OTO( done, ( ERR_TRASH, "could not find key" ) ); if ( !(LBT_MATCH & f.e[0]) ) /* key not found - ok */ goto done; LOG_DBG( LOG_DEBUG, "key found" ); e = (LBT_ENTMASK & f.e[0]); b = f.b[0]; d = b->dct + e; base = LBT_BASE(b); v = base+LBT_POS(d)+d->kln; /* start of values */ n = LBT_VLN(d); if ( ! n ) goto done; i = valpos( v, n, bt->vsz, key->val.byt ); LOG_DBG( LOG_DEBUG, "val %sfound pos %u of %u", (LBT_MATCH & i) ? "" : "not ", i & ~LBT_MATCH, n ); if ( !(LBT_MATCH & i) ) /* val not found - ok */ goto done; i &= ~LBT_MATCH; LOG_DBG( LOG_DEBUG, "del blk %d e %d i/n %d/%d", b->num, e, i, n ); if ( 1 < n ) { v += i*bt->vsz; /* end of stack area to move: start of this value */ mv = bt->vsz; /* amount to move: one value size */ LBT_VLNADD( d, -1 ); } else { /* sole value -- rm whole entry */ v = base+LBT_POS(d); /* start of key */ mv = d->kln+bt->vsz; for ( i=b->ent - e - 1; i--; d++ ) /* move the i entries after d down */ *d = d[1]; d = b->dct + e; /* reset */ b->ent--; } memmove( base+b->stk+mv, base+b->stk, v-(base+b->stk) ); b->stk += mv; for ( ; e++ < b->ent; d++ ) LBT_POSADD( d, mv ); WRULK( b, bt ); done: if ( f.b[0] ) ULK( f.b[0] ); return ret; } /* lbt_del */ int lbt_loop ( Idx *bt, DXLoop *l ) { int ret = 0; unsigned i = 0; int mode = IDXMODE & l->flg; Block *b = 0; Key *cmp = &l->key; if ( ! cmp->len ) /* start at 1st leaf */ b = getblk( bt, 1, 0, LBT_RD ); else { /* locate key */ FindKey f; f.bt = bt; f.key = cmp; if ( findkey( &f, LBT_RD ) ) LOG_OTO( done, ( ERR_TRASH, "could not find key" ) ); b = f.b[0]; i = (unsigned)(LBT_ENTMASK & f.e[0]); if ( !(LBT_MATCH & f.e[0]) && IDXEQ == mode ) goto done; } if ( IDXUPTO <= mode ) cmp = &l->to; for ( ;; i=0 ) { /* blocks */ unsigned char *base = LBT_BASE( b ); Dict *d; if ( ! b || b->lev ) return -ERR_TRASH; b->lck = LBT_EX; /* lock */ d = b->dct + i; for ( ; ient; i++, d++ ) { unsigned short pos = LBT_POS( d ); unsigned short vln = LBT_VLN( d ); unsigned short kln = d->kln; unsigned short j; unsigned char *v = base+pos+kln; Key key; Hit hit; int diff = memcmp( cmp->byt, base+pos, kln <= cmp->len ? kln : cmp->len ); switch ( mode ) { case IDXPF: if ( diff || kln < cmp->len ) goto moan; break; case IDXEQ: if ( diff || kln != cmp->len ) goto done; break; case IDXUPTO: if ( 0 >= diff ) goto done; break; case IDXINCL: if ( 0 > diff || (!diff && kln > cmp->len) ) goto done; break; default: /* unused */ moan: log_msg( LOG_INFO, "loop ends on key '%.*s'(%d) %d %d", kln, base+pos, kln, diff, cmp->len ); goto done; } memcpy( key.byt, base+pos, key.len = kln ); if ( l->cb ) for ( j=0; jvsz ) { rdval( &hit, v, bt->vsz ); ret = l->cb( l->me, &key, &hit ); if ( ret ) goto done; } else { /* internal check */ FindKey f; f.bt = bt; f.key = &key; memcpy( key.val.byt, v, key.val.len = bt->vsz ); if ( findkey( &f, LBT_RD ) ) LOG_OTO( done, ( ERR_TRASH, "could not find key" ) ); if ( !(LBT_MATCH & f.e[0]) || b != f.b[0] || i != (unsigned)(LBT_ENTMASK & f.e[0]) ) { log_msg( ERR_TRASH, "OOPS! want %u(%p).%u got %u(%p).%u (match %x)", b->num, b, i, f.b[0]->num, f.b[0], LBT_ENTMASK & f.e[0], f.e[0] ); rdval( &hit, v, bt->vsz ); log_msg( ERR_TRASH, "\twhen looking for '%.*s'/%u.%u.%u.%u", key.len, key.byt, hit.mfn, hit.tag, hit.occ, hit.pos ); if ( f.b[0]->ent <= (LBT_ENTMASK & f.e[0]) ) log_msg( ERR_TRASH, "\tI found nothing" ); else { Dict *dd = f.b[0]->dct + (LBT_ENTMASK & f.e[0]); unsigned char *bb = LBT_BASE(f.b[0])+LBT_POS(dd); rdval( &hit, bb+dd->kln, bt->vsz ); log_msg( ERR_TRASH, "\tI found '%.*s'/%u.%u.%u.%u", dd->kln, bb, hit.mfn, hit.tag, hit.occ, hit.pos ); } } else log_msg( LOG_TRACE, "key ok" ); } } ULK( b ); if ( ! b->nxt || !(b = getblk(bt, b->nxt, 0, LBT_RD)) ) break; } done: if ( b ) ULK( b ); if ( l->cb ) l->cb( l->me, 0, 0 ); else log_msg( LOG_INFO, "checked %i keys in %u blks of %dK depth %u cmp/get/miss %u/%u/%u", stat.key, bt->len, bt->bsz>>10, bt->root->lev, stat.cmp, stat.get, stat.miss ); return ret; } /* lbt_loop */ /* mimik the plain old ldb_search */ int lbt_search ( Idx *bt, Key *key, LdbPost *post, Rec *rec ) { int ret = 0; unsigned i = 0; int cont = rec && rec->len; /* continuing previous terms search */ char *stk = rec ? ((char*)rec + rec->bytes) : 0; /* term stack */ int prefix; Block *b = 0; Key start; LdbP * const p = post->p; /* really just an alias */ /* check for prefix match */ if ( post ) { if ( 8 != bt->vsz ) return log_msg( ERR_INVAL, "bad legacy search on vsz %d", bt->vsz ); prefix = LDB_PFX & post->mode; } else { if ( ! rec ) return log_msg( ERR_INVAL, "lbt_search needs rec or post" ); if ( (prefix = key->len && '$' == key->byt[key->len - 1]) ) key->len--; } if ( ! cont ) start = *key; else { memset( &start, 0, sizeof(start) ); start.len = 255 < rec->field[rec->len-1].len ? 255 : rec->field[rec->len-1].len; memcpy( start.byt, rec->field[rec->len-1].val, start.len ); rec->len = 0; } if ( ! start.len ) /* start at 1st leaf */ b = getblk( bt, 1, 0, LBT_RD ); else { /* locate key */ FindKey f; f.bt = bt; f.key = &start; if ( findkey( &f, LBT_RD ) ) LOG_OTO( ouch, ( ERR_TRASH, "could not find key" ) ); b = f.b[0]; i = (unsigned)(LBT_ENTMASK & f.e[0]); if ( !(LBT_MATCH & f.e[0]) && ! prefix ) goto done; } for ( ;; i=0 ) { /* blocks */ unsigned char *base = LBT_BASE( b ); Dict *d; if ( ! b || b->lev ) return -ERR_TRASH; b->lck = LBT_EX; /* lock */ d = b->dct + i; for ( ; ient; i++, d++ ) { /* entries */ unsigned char *ent = base + LBT_POS( d ); /* start of key */ unsigned short kln = d->kln; if ( memcmp( key->byt, ent, kln <= key->len ? kln : key->len ) || (prefix ? kln < key->len : kln != key->len ) ) goto done; if ( rec ) { /* record term */ Field *f = rec->field + rec->len; /* field to assign */ /** append term to record, unless it's the same as previous */ if ( rec->len ) { if ( kln == f[-1].len && !memcmp( f[-1].val, ent, kln ) ) continue; } else if ( cont ) { if ( kln == start.len && !memcmp( start.byt, ent, kln ) ) continue; } stk -= kln; if ( stk < (char*)(f+1) ) /* stack reached field dict */ goto done; memcpy( stk, ent, kln ); f->tag = 0; f->val = stk; f->len = kln; rec->len++; continue; } { /* more locals */ LdbP merge[1024]; unsigned short vln, k, m=0; unsigned char *c; int f = post->fil-1; /* highest pos to consider in given postings */ if ( !(vln = LBT_VLN( d )) ) continue; ent += kln; /* set to start of values */ c = ent + 8*(vln-1); /* last value */ /* collect postings */ for ( k=vln; k--; c-=8 ) { /* loop backwards (for the fun of it) postings in segment */ int prow, ptag, ppos; LdbP e; /* the entry */ LdbP samerow; /* highest possible entry w/ same row as e */ #if defined( LDB_BIG_ENDIAN ) /* the 8 bytes of a posting are BIG ENDIAN ! */ memcpy(e.bytes,c,8); #else e.bytes[0] = c[7]; e.bytes[1] = c[6]; e.bytes[2] = c[5]; e.bytes[3] = c[4]; e.bytes[4] = c[3]; e.bytes[5] = c[2]; e.bytes[6] = c[1]; e.bytes[7] = c[0]; #endif prow = LDBP_ROW( &e ); ptag = LDBP_TAG( &e ); ppos = LDBP_POS( &e ); LOG_DBG( LOG_VERBOSE, "post %d.%hd pos %06x", prow, ptag, ppos ); if ( !prow ) continue; if ( (post->cut && prow >= post->cut) || (post->tag && post->tag != ptag) ) continue; if ( prow < post->skp ) /* quickly bail out on skip condition */ break; LDBP_SETROWTOP( &samerow, &e ); /* for mfn comparison */ /* sweep down to postings for the same row as e ... */ while ( f >= 0 && LDBP_GT( p+f, &samerow ) ) f--; if ( LDB_AND & post->mode ) { int l; /* loop postings for same row, mark all (that are near enough) */ LDBP_SETROWBOT( &samerow, &e ); /* for mfn comparison */ /* NOTE: postings for row are GT than bottom even if marked */ for ( l = f; l>=0 && LDBP_GT( p+l, &samerow ); l-- ) { if ( post->near ) { int dist; if ( ptag != LDBP_TAG( p+l ) ) continue; if ( LDB_NEAR_G != post->near ) { dist = LDBP_POS( p+l ) - LDBP_POS( &e ); if ( dist < 0 ) dist = -dist; if ( 0 < post->near ? post->near < dist : -post->near != dist /* exact $$$$ */ ) continue; } } LDBP_SETMARK( p+l ); } } else { /* OR mode */ int add; if ( ! post->near ) /* add if row not found: ignore details */ add = 0 > f || prow > LDBP_ROW( p+f ); else { /* add if no exact match */ int l; /* NOTE: we don't use mark bit in OR mode, do we ? */ for ( l = f; l>=0 && LDBP_GT( p+l, &e ); l-- ) ; add = 0 > l || LDBP_GT( &e, p+l ); } if ( add ) merge[ m++ ] = e; } } /* for postings in segment */ if ( m ) { /* merge in the merge buffer */ LdbP *mm = merge; for ( k = post->fil += m; m && k--; ) { LdbP src; if ( k < m || LDBP_GT( mm, &p[k-m] ) ) { src = *mm++; m--; LOG_DBG( LOG_DEBUG, "merging %d at %d", LDBP_ROW(&src), k ); } else src = p[k-m]; if ( k < post->len ) p[k] = src; else { /* set cut */ int row = LDBP_ROW( &src ); if ( row < post->cut || !post->cut ) post->cut = row; } } if ( post->fil > post->len ) post->fil = post->len; if ( post->cut ) /* postings for cut row are unreliable */ while ( post->fil && post->cut <= LDBP_ROW(p+post->fil-1) ) post->fil--; } /* if ( m ) */ } /* more locals */ } /* for entries in block */ ULK( b ); if ( ! b->nxt || !(b = getblk(bt, b->nxt, 0, LBT_RD)) ) break; } done: if ( rec ) ret = rec->len; ouch: if ( b ) ULK( b ); return ret; } /* lbt_search */ int lbt_init ( Idx *bt ) { int ret = 0, blkbits, i; Block *l; int size = lio_size( bt->fd ); bt->root = 0; bt->hash = 0; bt->mem = 0; for ( i=LBT_LRULEV; i--; ) bt->lru[i] = bt->mru[i] = 0; if ( 0x3ff & size ) /* not on 1K boundary */ return log_msg( LOG_ERROR, "bt has bad size %dK + %d", size >> 10, (int)size & 0x3ff ); if ( size ) { /* read root header */ Block b; unsigned want = sizeof(b) - LBT_BOFF; int got = lio_pread( &bt->fd, LBT_BASE(&b), want, 0 ); if ( want != (unsigned)got ) return log_msg( LOG_SYSERR, "could not get root head" ); if ( b.num ) return log_msg( ERR_TRASH, "root has num %d", b.num ); /*if ( LBT_ISIS != b.nxt ) log_msg( LOG_WARN, "magic is 0x%08x", b.nxt );*/ if ( bt->typ && bt->typ != b.typ ) log_msg( LOG_WARN, "typ want 0x02%x got 0x02%x", bt->typ, b.typ ); bt->typ = b.typ; if ( bt->key && bt->key != b.key ) log_msg( LOG_WARN, "key want %d got %d", bt->key, b.key ); bt->key = b.key; if ( bt->col && bt->col != b.col ) log_msg( LOG_WARN, "col want %d got %d", bt->col, b.col ); bt->col = b.col; bt->dpt = b.lev; /* b.ent b.stk */ } else { /* new index */ bt->dpt = 1; if ( ! bt->key ) bt->key = 30; } bt->vsz = 8+(0xf & bt->typ); blkbits = 10 + (3 & (bt->typ >> 4)); bt->bsz = 1 << blkbits; if ( size % bt->bsz ) return log_msg( ERR_TRASH, "size 0x%08x % blksize", size, bt->bsz ); bt->len = size >> blkbits; if ( 2 > bt->len && !(LBT_WRITE & bt->flg) ) return log_msg( LOG_SYSERR, "attempt to read empty index" ); bt->hlen = 16 + bt->len / 32; /* initial cache size */ if ( bt->hlen < 1000 ) bt->hlen = 1000; if ( !(bt->root = addchunk( bt, bt->hlen )) ) return log_msg( LOG_SYSERR, "\twhen getting bt cache" ); if ( bt->hlen < 256 ) /* the hash */ bt->hlen = 256; if ( !(bt->hash = (Block**)mAlloc( bt->hlen * sizeof(Block*) )) ) return log_msg( LOG_SYSERR, "\twhen getting bt hash" ); memset( bt->hash, 0, bt->hlen * sizeof(Block*) ); if ( size && readblk( bt->root, bt, 0 ) ) return log_msg( ERR_TRASH, "could not read root" ); if ( ! bt->len ) { /* init: link root to leaf */ SETINT( LBT_BASE(bt->root)+(bt->root->stk -= 4), 1 ); LBT_PVK( bt->root->dct, bt->root->stk, 0, 0 ); bt->root->ent = 1; bt->len = 1; } if ( !(l = getblk( bt, 1, 0, LBT_EX )) ) return log_msg( ERR_TRASH, "could not read 1st leaf" ); if ( ! size ) { log_msg( LOG_INFO, "creating index bsz %u ksz %u vsz %u", bt->bsz, bt->key, bt->vsz ); bt->root->lev = 1; if ( WRULK( bt->root, bt ) || WRULK( l, bt ) ) return log_msg( LOG_SYSERR, "\twhen writing base" ); } return ret; } /* lbt_init */ void lbt_close ( Idx *bt ) { Chunk *c, *n = bt->mem; while ( (c = n) ) { n = c->nxt; mFree( c ); } mFree( bt->hash ); lio_close( &bt->fd, LIO_INOUT ); memset( bt, 0, sizeof(*bt) ); } /* lbt_close */ /* ************************************************************ public functions */ void cXMkVal ( Idx *bt, Val *val, Hit *hit ) { if ( hit ) { val->len = bt->vsz; mkval( val, hit ); } else val->len = 0; } int cXAdd ( Idx *bt, Key *key, Hit *hit ) { if ( bt->bat ) { if ( ! key ) key = &bt->bat->key; cXMkVal( bt, &key->val, hit ); return lbt_batchval( bt, key ); } else if ( !key ) return 0; else if ( hit ) { cXMkVal( bt, &key->val, hit ); if ( hit->dbn == 0xffff ) return lbt_del( bt, key ); } return lbt_add( bt, key ); } /* cXAdd */ lbt.h0100644000113000011300000000664107664643002010503 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LBT_H /* $Id: lbt.h,v 1.12 2003/05/27 11:03:30 kripke Exp $ package interface of the btree. see Lehmann/Yao or the GiST for how it works. */ #include "ldb.h" /* LdbPost */ /** key comparision function. however, we still assume that two keys have to be identical in length and bytes to be equal. */ typedef int lbt_comp ( const unsigned char *a, const unsigned char *b, unsigned int l ); /** variables: 4 bits: #bytes for value 8+(0..15) 2 bits: blocksize 1024<<(0..3) max key length (4..255) collation (comparision function) 0..255 */ typedef struct Idx { /* actually it's a B-L-Tree ;) */ int fd; /* the file */ int flg; /* flags: writeable, batch */ unsigned char typ; /* type: bsz, ifp, flags */ unsigned char key; /* max key length */ unsigned char col; /* collation */ unsigned char dpt; /* depth (level of root over bottom > 0) */ lbt_comp *cmp; /* comparision function */ /* following members (and the depth above) are set automatically. they are going to stay and you may check them, if you're interested. */ unsigned vsz; /* ifp size computed from type */ unsigned bsz; /* block size computed from type */ unsigned len; /* # blocks in index */ /* following members are considered internal. if you import lbt and rely on them, don't blame me if they change. */ unsigned hlen; /* hash length */ unsigned clen; /* cache length */ struct Block *root; /* the root */ struct Block **hash; /* hash array */ struct Block *lru[4]; /* least recently used list for lowest levels */ struct Block *mru[4]; /* tail of lru list (most recently used) */ struct Chunk *mem; struct Batch *bat; } Idx; enum { /* btree flags */ LBT_WRITE = 0x01 /* open for writing */ }; enum { /* btree type */ LBT_BLK1K = 0x00, /* 1K blocks */ LBT_BLK2K = 0x10, /* 2K blocks */ LBT_BLK4K = 0x20, /* 4K blocks */ LBT_BLK8K = 0x30, /* 8K blocks */ LBT_CMPRS = 0x80 /* compressed keys */ }; /** initialise from an already open fd. */ extern int lbt_init ( Idx *bt ); /** flush and release any cache, close fd. */ extern void lbt_close ( Idx *bt ); extern int lbt_batch ( Idx *bt, unsigned char pctfree ); extern int lbt_batchval ( Idx *bt, Key *key ); extern int lbt_add ( Idx *bt, Key *key ); extern int lbt_del ( Idx *bt, Key *key ); extern int lbt_loop ( Idx *bt, DXLoop *l ); extern int lbt_search ( Idx *bt, Key *key, LdbPost *post, Rec *rec ); /* half public ... ??? */ extern void cXMkVal ( Idx *bt, Val *val, Hit *hit ); #define LBT_H #endif /* LBT_H */ lcli.c0100644000113000011300000002153007662161237010635 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lcli.c,v 1.13 2003/05/15 19:17:17 mawag Exp $ OpenIsis client side of communication */ #ifndef WIN32 #include #include /* socket etc */ #include /* IPPROTO_TCP, htons */ #include /* gethostbyname */ #endif #include #include #include #include "openisis.h" #include "loi.h" #include "lio.h" #include "lcli.h" /* gcc -D_MAIN_ -I. -Lsto lcli.c -lopenisis && ./a.out 111 `find .` gcc -D_MAIN_ -DWIN32 -I. lcli.c win/libopenisis.a /mingw/lib/libwsock32.a */ #define LF 10 /* LineFeed a.k.a. newline - '\n' isn't really well defined */ #define TAB 9 /* horizontal, that is */ #define VT 11 /* vertical, used as newline replacement */ #define STAT_CONT 0x01 /* not at beginning of line */ #define STAT_BIN 0x10 /* binary mode */ /* plain protocol * @return done or -err */ int cliPlain (Ios *s, int *stat, Rec **rec) { Field *f; int l = s->b.fill - s->b.done; unsigned char *b = s->b.c + s->b.done; unsigned char *end = b+l, *v, *p; if ( ! l ) { /* EOF: done */ if ( ! (STAT_CONT & *stat) ) /* ok */ return 1; /* last field wasn't closed by LF */ return sMsg (ERR_INVAL, "cliPlain(%d): no EOL", LIO_FD & s->file); } if ( STAT_CONT & *stat ) RSPACE( *rec, l, !0 ); /* add text lines */ while ( blen ) { RADD( *rec, 0,0,end-b, !0 ); } else { /* binary mode continuation line */ conti = 1; if ( !(STAT_BIN & *stat)) { sMsg( LOG_INFO, "cliPlain(%d): detected binary mode", LIO_FD & s->file); *stat |= STAT_BIN; } RSPACE( *rec, end-b, !0 ); } if ( ! *rec ) return -ERR_NOMEM; *stat |= STAT_CONT; case STAT_CONT: /* add to last field */ f = (*rec)->field + (*rec)->len-1; v = (unsigned char*)f->val; p = v + f->len; if ( conti ) { *p++ = LF; b++; } if ( STAT_BIN & *stat ) { for ( ; bused += (p - v) - f->len; f->len = p - v; if ( LF == b[-1] ) { int ret = a2il( f->val, f->len, &f->tag ); if ( ret ) { if ( ret < f->len && TAB == v[ret] ) ret++; if ( ret < f->len ) memmove( v, v+ret, f->len - ret ); f->len -= ret; } *stat &= ~STAT_CONT; } sMsg (LOG_VERBOSE, "cliPlain(%d): fld[%2d] %3d = '%.*s'", LIO_FD & s->file, (*rec)->len-1, f->tag, f->len, f->val); } } return 0; } Rec* cliRead (CliChnl *ch) { Ios str; Rec *rec; int rt, stat; if (!ch || 0 > ch->sd) { return 0; } memset (&str, 0, sizeof (Ios)); str.file = ch->sd | LIO_IN; rec = 0; stat = 0; ch->err = errno = 0; while (1) { str.b.done = str.b.fill; rt = ioStdio (&str, LIO_SFILL); if (0 > rt) { ch->err = errno; if (rec) { mFree (rec); } if (!(LIO_IN & str.file)) { ch->sd = -1; } return 0; } rt = cliPlain (&str, &stat, &rec); if (rt) { if (0 > rt) { if (rec) { mFree (rec); } return 0; } log_rec (LOG_VERBOSE, rec, "cliRead(%d): ", 0); sMsg (LOG_INFO, "cliRead(%d): read #flds %d", ch->sd, (rec ? rec->len : -1)); return rec; } } } int cliWrite (CliChnl *ch, Rec *rec) { char buf [4096]; char *b; int fd, len, sl, rl, rt; if (!ch || 0 > ch->sd) { return -ERR_BADF; } ch->err = 0; if (!rec || !rec->len) { return 0; } len = sizeof (buf); b = rSerA (rec, buf, &len); if (!b) { return sMsg (ERR_NOMEM, "cliWrite(%d): write (%d)", ch->sd, rec->used); } fd = ch->sd | LIO_OUT; for (sl = rt = errno = 0, rl = len; rl; sl += rt, rl -= rt) { rt = lio_write (&fd, b + sl, (unsigned)rl); ch->err = errno; if (0 > rt) { if (!(LIO_OUT & fd)) { ch->sd = -1; } break; } } if (b != buf) { mFree (b); } if (0 > rt) { return sMsg (ERR_IO, "cliWrite(%d): write (%d) = %d(%d)", len, rt, ch->err); } sMsg (LOG_INFO, "cliWrite(%d): wrote %d,%d", ch->sd, rec->len, len); return 0; } int cliConnect (CliChnl *ch, const char *hname, int port) { #ifdef WIN32 return sMsg (ERR_TRASH, "cliConnect: operation not supported"); #else struct sockaddr_in addr; struct hostent *hostp; int rt; if (!ch) { return sMsg (ERR_IDIOT, "cliConnect: null channel"); } memset (ch, 0, sizeof (CliChnl)); rt = socket (PF_INET, SOCK_STREAM, IPPROTO_TCP); if (0 > rt) { ch->err = errno; return sMsg (ERR_IO, "cliConnect: cannot create socket: %d", errno); } ch->sd = rt; memset (&addr, 0, sizeof (addr)); if (hname) { hostp = gethostbyname (hname); if (! hostp) { ch->err = errno; return sMsg (ERR_INVAL, "cliConnect: unknown host <%s>: %d", hname, errno); } addr.sin_family = hostp->h_addrtype; /* already network-byte-order */ addr.sin_addr.s_addr = *(long*)(*(hostp->h_addr_list)); /* dto */ } else { addr.sin_family = AF_INET; addr.sin_addr.s_addr = htonl (INADDR_LOOPBACK); } addr.sin_port = htons (port); rt = connect (ch->sd, &addr, sizeof (addr)); if (0 > rt) { ch->err = errno; return sMsg (ERR_IO, "cliConnect: cannot connect to %s:%d: %d", (hname ? hname : "localhost"), port, errno); } return 0; #endif } void cliClose (CliChnl *ch) { if (ch && 0 <= ch->sd) { close (ch->sd); ch->sd = -1; } } /* ----------------------------------------------------------------------- * test */ #ifdef _MAIN_ #include /* printf */ #include /* atoi */ #include /* strerror */ #include "luti.h" /* log_rec */ void cmp (Rec *in, Rec *out) { Field *F1, *F2; int l1 = in->len; int l2 = out->len; int j, k; for (j = k = 0, F1 = in->field, F2 = out->field; l2 > k; ++k, ++F2) { if (0 > F2->tag) { continue; } if (l1 <= j) { printf ("ERR too many answers %d %d\n", j, k); log_rec (0, in, "ERR IN ", 0); log_rec (0, out, "ERR OUT ", 0); exit (1); } if (F1->tag != F2->tag) { printf ("ERR tag %d mismatch %d %d\n", j, F1->tag, F2->tag); log_rec (0, in, "ERR IN ", 0); log_rec (0, out, "ERR OUT ", 0); exit (1); } if (F1->len != F2->len) { printf ("ERR len %d mismatch %d %d\n", j, F1->tag, F2->tag); log_rec (0, in, "ERR IN ", 0); log_rec (0, out, "ERR OUT ", 0); exit (1); } if (memcmp (F1->val, F2->val, F1->len)) { printf ("ERR val %d mismatch\n", j); log_rec (0, in, "ERR IN ", 0); log_rec (0, out, "ERR OUT ", 0); exit (1); } ++j; ++F1; } } int loop (CliChnl *ch, Rec *demo, int fail) { Rec *rsp; int rt; rt = cliWrite (ch, demo); if (0 > rt) { printf ("%s write %d %s\n", (fail ? "ERR":"WARN"), ch->err, strerror (ch->err)); if (!fail) { return -1; } exit (1); } rsp = cliRead (ch); if (! rsp) { printf ("%s read %d %s\n", (fail ? "ERR":"WARN"), ch->err, strerror (ch->err)); if (!fail) { return -1; } exit (1); } cmp (demo, rsp); mFree (rsp); return 0; } int main (int argc, char **argv) { CliChnl ch; char *hname = 0; int port = 0; Rec *demo = 0; int num = 0; int j; for (j = 1; argc > j; ++j) { if (*argv[j] == '-') { if (argv[j][1] == 'h') { hname = argv[j] + 2; continue; } if (argv[j][1] == 'p') { port = atoi (argv[j] + 2); continue; } } num = atoi (argv[j]); break; } if (0 >= num) { num = 1; } for (j = argc; 0 <= --j; ) { RADDS (demo, j, argv[j], !0); } if (0 >= port) { port = 8080; } if (cliConnect (&ch, hname, port)) { printf ("ERR connect %d %s\n", ch.err, strerror (ch.err)); exit (1); } j = num; while (1) { if (loop (&ch, demo, 0)) { sleep (15); if (cliConnect (&ch, hname, port)) { printf ("ERR reconnect %d %s\n", ch.err, strerror (ch.err)); exit (1); } loop (&ch, demo, !0); } if (! --j) { break; } if (! (j % 10)) { printf ("loop %d ...\n", j); } } cliClose (&ch); printf ("ok.\n"); return 0; } #endif /* _MAIN_ */ lcli.h0100664000113000011300000000357207656021241010643 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LCLI_H #define LCLI_H /* $Id: lcli.h,v 1.8 2003/05/06 21:01:21 mawag Exp $ OpenIsis client side of communication */ typedef struct CliChnl { int sd; /* socket */ int err; /* errno */ } CliChnl; /** open channel and connect to server. @return 0 success, <0 error code */ extern int cliConnect (CliChnl *ch, const char *hostname, int port); /** close channel. */ extern void cliClose (CliChnl *ch); /** serialize and write record. @return number of bytes written, <0 error code */ extern int cliWrite (CliChnl *ch, Rec *rec); /** deserialize record from stream using plain protocol. @param str input stream @param stat internal state, must be initialized with 0 before first call @param rec address of record to fill @return 0 - more input needed, 1 - done, <0 error code */ extern int cliPlain (Stream str, int *stat, Rec **rec); /** read and deserialize record. @return deserialized record, 0 on error */ extern Rec* cliRead (CliChnl *ch); #endif /* LCLI_H */ lcs.c0100644000113000011300000001464607644412744010507 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lcs.c,v 1.8 2003/04/08 00:20:52 kripke Exp $ charset support */ #include /* memset */ #include "luti.h" #include "lio.h" #include "lcs.h" /* ************************************************************ private types */ /* ************************************************************ private data */ /* ************************************************************ private functions */ /* ************************************************************ package functions */ unsigned char lcs_latin1_uc[256] = { 0,' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ', ' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ', ' ','!','"','#','$','%','&','\'','(',')','*','+',',','-','.','/', '0','1','2','3','4','5','6','7','8','9',':',';','<','=','>','?', '@','A','B','C','D','E','F','G','H','I','J','K','L','M','N','O', 'P','Q','R','S','T','U','V','W','X','Y','Z','[','\\',']','^','_', '`','A','B','C','D','E','F','G','H','I','J','K','L','M','N','O', 'P','Q','R','S','T','U','V','W','X','Y','Z','{','|','}','~',' ', ' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ', ' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ', ' ',161,162,163,164,165,166,167,168,169,170,171,172,173,174,175, 176,177,178,179,180,181,182,183,184,185,186,187,188,189,190,191, 'A','A','A','A','A','A','A','C','E','E','E','E','I','I','I','I', 'T','N','O','O','O','O','O',215,'O','U','U','U','U','Y','T','S', 'A','A','A','A','A','A','A','C','E','E','E','E','I','I','I','I', 'T','N','O','O','O','O','O',247,'O','U','U','U','U','Y','T','Y' }; unsigned char lcs_latin1_ct[256] = { /* NUL STX SOT ETX EOT ENQ ACK BEL */ LCS_F,LCS_C,LCS_C,LCS_C,LCS_C,LCS_C,LCS_C,LCS_C, /* BS HT LF VT FF CR SO SI */ LCS_C,LCS_C,LCS_F,LCS_C,LCS_R,LCS_F,LCS_C,LCS_C, /* DLE DC1 DC2 DC3 DC4 NAK SYN ETB */ LCS_C,LCS_C,LCS_C,LCS_C,LCS_C,LCS_C,LCS_C,LCS_C, /* CAN EM SUB ESC FS GS RS US */ LCS_C,LCS_C,LCS_C,LCS_C,LCS_C,LCS_R,LCS_F,LCS_C, /*blank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` a b c d e f g */ LCS_P,LCS_L,LCS_A,LCS_L,LCS_A,LCS_A,LCS_A,LCS_A, LCS_A,LCS_L,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A, LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A, /* x y z { | } ~ DEL */ LCS_L,LCS_A,LCS_A,LCS_P,LCS_P,LCS_P,LCS_P,LCS_C, /* 32 C0 controls */ LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P, LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P, LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P, LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P,LCS_P, /* 32 mostly symbols */ LCS_S,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y, LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y, LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y, LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y,LCS_Y, /* 64 Latin alphas including 2 symbols */ LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A, LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A, LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_Y, LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A, LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A, LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A, LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_Y, LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A,LCS_A }; /* ************************************************************ package functions */ int lcs_mktab ( LcsTab *dst, char *n, int l, int bits ) { int ret = 0; unsigned char *c = dst->c; unsigned short *u = dst->u; int line = 1; int i = 0; memset( dst, 0, sizeof(*dst) ); for (;; i++ ) { unsigned short in = 0; for ( ; l && (*n < '0' || '9' < *n); n++, l-- ) /* skip nondigs */ if ( '\n' == *n ) line++; if ( ! l ) break; for ( ; l && '0' <= *n && *n <= '9'; n++, l--) in = 10*in + *n - '0'; if ( bits < 0 ) { if ( i > 127 ) { log_msg( LOG_WARN, "too many unicode values in input" ); return 0; } if ( in <= 0xffff ) u[i] = in; else log_msg( ERR_INVAL, "unicode %d too big %dth value line %d", in, i, line ); continue; } if ( in > 0xff ) log_msg( ERR_INVAL, "byte %d too big %dth value line %d", in, i, line ); else if ( bits ) c[in] |= bits; else if ( i > 255 ) { log_msg( LOG_WARN, "too many values in input" ); return 0; } else c[i] = (unsigned char)in; } return ret; } /* lcs_mktab */ int lcs_mkrecod ( unsigned char *dst, unsigned short *from, unsigned short *to ) { int i = 0, j; for ( i=0; i<128; i++ ) dst[i] = (unsigned char)i; /* identity */ for ( ; i<256; i++ ) { unsigned short s = from[ i - 128 ]; dst[i] = ' '; if ( to ) for ( j=128; j--; ) { if ( s == to[j] ) { dst[i] = (unsigned char)128+j; break; } } else if ( 256 > s ) /* map to latin 1 */ dst[i] = (unsigned char)s; } return 0; } /* lcs_mkrecod */ /* ************************************************************ public functions */ lcs.h0100644000113000011300000001001507645531147010476 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LCS_H /* $Id: lcs.h,v 1.3 2003/04/08 00:20:52 kripke Exp $ charset support */ typedef enum { LCS_SINGLE, /* some ASCII-compliant single byte charset */ LCS_UTF8, /* unicode in 8bit transfer encoding */ LCS_UCS2 /* the unicode base multilingual plane (BMP, first 64K chars) in 2byte encoding, native (platform) byte order */ } lcs_type; typedef enum { LCS_LATIN1, /* the ISO-8859-1 (Latin 1) charset */ LCS_CP850, /* MS-DOS Latin 1, known as the "OEM" charset */ LCS__SBCS /* # of single byte charsets */ } lcs_single; typedef enum { LCS_CTYPE, /* flags see below */ LCS_UNICO, /* 128 unicodes */ LCS_UCASE, /* each characters uppercase code */ LCS_INDEX, /* single byte recoding for index ("strxfrm") */ LCS_XEDNI, /* inverse */ LCS_RECOD, /* single byte recoding to extern (typically Latin1) */ LCS_DOCER, /* inverse */ LCS__TABS } lcs_tabid; /* some conversion table on a single byte charset c may hold an actab (!0 for any character deemed "alpha"), uctab ("uppercase" replacement code, typically removing diacriticals), code-to-code conversion or other. u may hold unicode values for the upper half of an ASCII compatible single byte charset. */ typedef union { unsigned char c[256]; unsigned short u[128]; } LcsTab; enum { /* controls (0..31,127) */ LCS_R = 0x00, /* record separator (FF,GS), */ LCS_F = 0x10, /* field separator (NUL,CR,LF,RS) */ LCS_C = 0x28, /* other control */ /* other separators */ LCS_S = 0x48, /* space(blank,nbsp), */ LCS_T = 0x68, /* other token separator (,:;=), */ /* other punctuation assumed "word" characters */ LCS_P = 0x80, LCS_Y = 0xb0, /* symbol */ /* identifiers */ LCS_I = 0xc0, /* C-ident (_). */ LCS_D = 0xd0, /* digits ('0'..'9') */ LCS_A = 0xe0, /* ASCII alpha */ LCS_L = 0xe4, /* ... among those formatting literals a,c,i,x */ LCS_N = 0xf0 /* other alpha ("national"/non-ASCII) */ }; #define LCS_ISCONTR( t ) (!(0xc0 & (t))) #define LCS_ISSPACE( t ) (LCS_S >= (t)) #define LCS_ISPRINT( t ) (0xc0 & (t)) /* (LCS_S <= (t)) */ #define LCS_ISIDENT( t ) (0xc0 == (0xc0 & (t))) #define LCS_ISALPHA( t ) (0xe0 == (0xe0 & (t))) /* for record parsing */ #define LCS_ISSEP( t ) (!(0x80 & (t))) #define LCS_ISWORD( t ) (0x80 & (t)) #define LCS_ISFR( t ) (!(0xe0 & (t))) /* field or record separator */ #define LCS_ISCST( t ) (0x08 & (t)) /* other separator */ enum { LCS_MKUNI = -1 /* as "bits" value, have mktab create unicode table */ }; /** create the table dst from string containing a free style sequence of decimal numbers (ignoring any non-digit). if bits is 0, numbers are assigned sequentially to dst->c, else if bits < 0, they are assigned to dst->u, else for every number, bits are set in the corresponding array element else something strange may happen. */ extern int lcs_mktab ( LcsTab *dst, char *numbers, int len, int bits ); /** create single byte conversion table from two unicode tables. if to is NULL, the trivial (Latin1) table is used. */ extern int lcs_mkrecod ( unsigned char *dst, unsigned short *from, unsigned short *to ); extern unsigned char lcs_latin1_uc[256]; extern unsigned char lcs_latin1_ct[256]; #define LCS_H #endif /* LCS_H */ ldb.c0100644000113000011300000023437507665647540010501 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: ldb.c,v 1.92 2003/05/30 11:28:15 kripke Exp $ implementation of general db access functions. */ #include #include #include /* PATH_MAX */ #include /* special */ #if defined( __GNUC__ ) && defined ( alloca ) #include #endif /* gcc defines always a cpu type - this we use for byteorder checking */ #if defined( sparc ) || defined( __ppc__ ) # define LDB_BIG_ENDIAN /* TODO: figure out fastest "htonl" on those boxes that usually don't swap */ static int rvi ( int i ) { int r; ((char*)&r)[0] = ((char*)&i)[3]; ((char*)&r)[1] = ((char*)&i)[2]; ((char*)&r)[2] = ((char*)&i)[1]; ((char*)&r)[3] = ((char*)&i)[0]; return r; } static short rvs ( short i ) { short r; ((char*)&r)[0] = ((char*)&i)[1]; ((char*)&r)[1] = ((char*)&i)[0]; return r; } #define SWI( i ) i = rvi( i ) #define SWS( s ) s = rvs( s ) #else # define rvi(i) i # define rvs(s) s #define SWI( i ) #define SWS( s ) #endif #if defined( sparc ) # define LDB_NEEDALIGN #endif #if defined( LDB_NEEDALIGN ) static unsigned GETINT ( const void *m ) { unsigned l; memcpy( &l, m, 4 ); return l; } static unsigned short GETSHORT ( const void *m ) { unsigned short s; memcpy( &s, m, 2 ); return s; } #else #define GETINT( m ) (*(unsigned*)(m)) #define GETSHORT( m ) (*(unsigned short*)(m)) #endif #include "lstr.h" #include "lio.h" #include "lbt.h" #include "lcs.h" #include "ldb.h" #include "lfdt.h" #include "luti.h" #ifdef WIN32 #define IsAbsPath(p) \ ((p) && *(p) && ( \ '/' == *(p) || '\\' == *(p) || ( \ ':' == (p)[1] && ( \ '/' == (p)[2] || '\\' == (p)[2] \ )))) #else #define IsAbsPath(p) \ ((p) && '/' == *(p)) #endif #define LF 10 /* LineFeed a.k.a. newline - '\n' isn't really well defined */ #define TAB 9 /* horizontal, that is */ #define VT 11 /* vertical, used as newline replacement */ /* ************************************************************ private types */ /** extension of master file proper. */ static const char EXT_MST_MST[] = ".mst"; /** extension of master file xref. */ static const char EXT_MST_XRF[] = ".xrf"; typedef enum { MST_MST, MST_XRF, MST_FILES } mst_file; static const char * const EXT_MST[MST_FILES] = { EXT_MST_MST, EXT_MST_XRF }; /** extension of inverted file short term nodes. */ static const char EXT_INV_N01[] = ".n01"; /** extension of inverted file short term leaves. */ static const char EXT_INV_L01[] = ".l01"; /** extension of inverted file long term nodes. */ static const char EXT_INV_N02[] = ".n02"; /** extension of inverted file long term leaves. */ static const char EXT_INV_L02[] = ".l02"; /** extension of inverted file control records. */ static const char EXT_INV_CNT[] = ".cnt"; /** extension of inverted file postings. */ static const char EXT_INV_IFP[] = ".ifp"; typedef enum { INV_N01, INV_L01, INV_N02, INV_L02, INV_CNT, INV_IFP, INV_FILES } inv_file; static const char * const EXT_INV[INV_FILES] = { EXT_INV_N01, EXT_INV_L01, EXT_INV_N02, EXT_INV_L02, EXT_INV_CNT, EXT_INV_IFP }; /** extension of lbt B-Link-Tree. It's named oxi because that is nicer than oix for OpenIsis indeX. however, see http://www.oxicenter.com.br/ */ static const char EXT_LBT_OXI[] = ".oxi"; static const char * const EXT_LBT[] = { EXT_LBT_OXI }; /** plaintext master file */ static const char EXT_TXT_TXT[] = ".txt"; static const char EXT_TXT_PTR[] = ".ptr"; static const char EXT_TXT_OPT[] = ".opt"; typedef enum { TXT_TXT, TXT_PTR, TXT_FILES } txt_file; static const char * const EXT_TXT[] = { EXT_TXT_TXT, EXT_TXT_PTR }; static const char ISIX[] = "ISIX"; /* ptr magic */ /** extension of supporting file alpha character table. */ static const char EXT_SUP_ACT[] = ".act"; /** extension of supporting file uppercase table. */ static const char EXT_SUP_UCT[] = ".uct"; typedef enum { SUP_ACT, SUP_UCT, SUP_FILES } sup_file; static const char * const EXT_SUP[SUP_FILES] = { EXT_SUP_ACT, EXT_SUP_UCT }; typedef int lblk[128]; typedef struct { Db head; int flags; const char *path; int mst[MST_FILES]; /* master file */ int inv[INV_FILES]; /* primary inverted file */ int mfc[LMFC__FL]; /* master file control record */ unsigned mflen; /* master file length */ int xrf[129]; /* last used xrf block : THREAD THREAT */ int xrlen; /* length of xrf (in blocks) */ unsigned short ptr; /* type of pointer file (new style xrf) */ unsigned short ptrl; /* pointer bytes, 512 for old xrf */ char *mmap; /* memory map of xrf/ptr */ int mmlen; /* length of map (in ptrl) */ int cnt[LDB_INDEXES][LCNT__FL]; /* two cnt records */ short tlen[LDB_INDEXES]; /* max term length for each index */ LcsTab ctab[LCS__TABS]; Idx oxi; } LDb; typedef union { lll bar; char r[16]; } Ptr; /* db flags */ #define DB_OPEN 0x010000 #define DB_INVOPEN 0x020000 #define DB_LBTOPEN 0x040000 #define DB_TXTOPEN 0x080000 #define DB_WRITABLE 0x100000 #define DB_MODIFIED 0x200000 #define DB_TXTMODE 0x20 #define DB_MMAP 0x10 #define DB_VARI 0xf /* mask for variant */ /* get xstr for record rec in set */ #define DB_XSTR( db, set, rec ) \ lstrlib[ set ].desc[ DB_VARI & (db)->flags ][ rec ] /* get record names for record rec in set */ #define DB_RNAM( db, set, rec ) \ lstrlib[ set ].name[ rec ] /** packed little endian masterfile control structure. */ typedef struct Mfc { int ctlm; int nmfn; int nmfb; short nmfp; short type; int rcnt; int mfx1; int mfx2; int mfx3; } Mfc; /** packed little endian masterfile record. */ typedef struct Dict { short tag; short pos; short len; } Dict; /** packed little endian masterfile record. */ typedef struct Mfr { int mfn; short recl; /* a.k.a. mfrl */ short bwbl; /* low part of int */ short bwbh; /* high part of int */ short bwp; /* it is believed, that this first five fields up to here (12 bytes packed) are to be in one 512-byte block; the manual mentiones even 14 bytes ... ??? */ short base; short nvf; short stat; Dict dict[1]; } Mfr; /* ************************************************************ private data */ static LDb defdbspace[32]; /* array of open dbs. should expand dynamically. */ static LDb *dbs = defdbspace; static int dbs_len = sizeof(defdbspace)/sizeof(defdbspace[0]); static int init; /* ************************************************************ private functions */ static LDb *getDb ( int id ) { if ( 0 <= id && id < dbs_len && dbs[id].flags ) { return &dbs[id]; } log_msg( LOG_ERROR, "attempt to access bad db id %d", id ); return 0; } /* getDb */ /* ************************************************************ start of io section */ enum { /* additional flags in the LIO_FD range */ OPEN_TRY = 1, /* try writable, open readonly else */ OPEN_UC = 2, /* use uppercase ext */ /* commonly used combinations */ /* 1) open as is, do not complain about any failure, do not create */ OPEN_ASIS = LIO_SEEK|LIO_RDWR|OPEN_TRY, /* 2) open readonly, do not complain about any failure, do not create */ OPEN_RDIF = LIO_SEEK|LIO_RD|LIO_TRY, /* 3) open readonly, complain about any failure */ OPEN_RD = LIO_SEEK|LIO_RD, /* 4) open or create writable, complain on failure */ OPEN_NEW = LIO_SEEK|LIO_RDWR|LIO_CREAT, OPEN_BLANK = LIO_SEEK|LIO_RDWR|LIO_CREAT|LIO_TRUNC }; /* figure out wether to use uppercase extension on path. if last path component (everything after the last / and \) does contain an uppercase ascii and does not contain a lowercase ascii, return OPEN_UC, else 0. */ static int autocase ( const char *path ) { int ret = 0; const char *e = path + strlen( path ); while ( e-- > path ) if ( 'A'<=*e && *e<= 'Z' ) ret = OPEN_UC; else if ( 'a'<=*e && *e<= 'z' ) return 0; else if ( '/'==*e || '\\' == *e ) break; return ret; } /* set extension. fname MUST already end with .xxx. if how has OPEN_UC set, use uppercase extension */ static char *setext ( char *fname, const char *ext, int how ) { int l = strlen( fname ) - 4; memcpy( fname+l, ext, 4 ); if ( OPEN_UC & how ) { char *p = fname+l; for ( ;*p; p++ ) /* use uppercase extensions */ if ( 'a' <= *p && *p <= 'z' ) *p -= 'a'-'A'; } return fname; } /** try to open all files according to how. ldb is only interested in seekable readable true files. @return 1 if all files could be opened writable 0 if all files could be opened readonly, and that was requested by a RD mode or try write something negative else */ static int openfiles ( int *fid, char *path, const char * const *ext, int nfiles, int how ) { int i; int wr = LIO_WR&how ? 1 : 0, mode = LIO_WANT & how; for ( i=0; i got ) return got; #ifndef NDEBUG if ( LOG_DO( LOG_ALL ) ) LOG_HEX( dst, got ); #endif if ( siz == (int)got ) return 0; log_msg( LOG_WARN, "got %u bytes wanted %d at %d in 0x%x", got, siz, where, fid ); return 1+(int)got; } /* readblk */ /* ************************************************************ end of io section */ static int *nrec ( int *xstr ) { int *dst = (int*)mAlloc( xstr[LSTR_ILEN] ); if ( dst ) *dst = *xstr; return dst; } /* nrec */ typedef struct b8 { char x[8]; } b8; typedef struct b4 { char x[4]; } b4; typedef struct b2 { char x[2]; } b2; static int convert ( int *dst, char *src, int *xstr ) { static int pow2[] = { 1, 2, 4, 8 }; int occ = -1; int *xmbrs = xstr+LSTR_XMBR; int nmbrs = LSTRFIX(*xstr); int *mbr; char *buf, *part = src, *srcend = src + xstr[LSTR_XLEN]; if ( ! dst ) return log_msg( ERR_NOMEM, "out of memory (no dst) in convert" ); if ( LSTRLEN(*xstr) > xstr[LSTR_ILEN] ) return log_msg( ERR_INVAL, "bad ilen %d need %d", xstr[LSTR_ILEN], (int)LSTRLEN(*xstr) ); if ( *dst != *xstr ) return log_msg( ERR_INVAL, "bad *dst 0x%08x need 0x%08x", *dst, *xstr ); /* clean and re-init */ memset( dst, 0, xstr[LSTR_ILEN] ); *dst = *xstr; mbr = dst+1; buf = ((char*)dst) + LSTRLEN(*dst); /* cvt the fixed part (occ==-1) and each occurrence of repeated part. */ for ( ;/* occ < LSTROCC(*dst) */; ) { /* cvt one part */ int i; for ( i=0; i sbytes ) return log_msg( ERR_INVAL, "srcbuf too short %d have %d need %d occ %d mbr %d", xstr[LSTR_XLEN], sbytes, len, occ, i ); if ( !NEEDSWAP( xmbr ) ) /* for ( j = len; j--; ) num.buf[j] = s[j]; */ switch ( ld ) { case 3: num.x8 = *(b8*)s; break; case 2: num.x4 = *(b4*)s; break; case 1: num.x2 = *(b2*)s; break; case 0: num.buf[0] = *s; break; } else /* swap bytes */ for ( j = len; j--; ) num.buf[j] = s[len - 1 - j]; switch ( len ) { case 8: *mbr = num.ll; break; /* TODO: defines for 16 and 64 bit compilers */ case 4: *mbr = num.i; break; case 2: *mbr = num.s; break; case 1: *mbr = num.buf[0]; break; } if ( LMBRISBITS( xmbr ) ) { /* apply bit shift and mask */ *mbr >>= LMBRBITOFF( xmbr ); *mbr &= ~(-1L << LMBRBITLEN( xmbr )); } continue; } /* else raw data -- that's easy :) */ { int offset = buf - (char*)dst; int need; len = LONG2LEN(xmbr); need = offset + len; if ( need > xstr[LSTR_ILEN] ) return log_msg( ERR_INVAL, "bad buflen %d need %d+%d in occ %d mbr %d", xstr[LSTR_ILEN], offset, len, occ, i ); if ( len > sbytes ) return log_msg( ERR_INVAL, "srcbuf too short %d have %d need %d occ %d mbr %d", xstr[LSTR_XLEN], sbytes, len, occ, i ); memcpy( buf, s, len ); *mbr = buf - (char*)dst; buf += len; } } /* for mbrs */ if ( ++occ >= LSTROCC(*dst) ) break; if ( occ ) part += (short)xstr[LSTR_XRLO]; /* adv. rep. part len */ else { /* was the fixed part, setup for repeated */ nmbrs = LSTRREP(*xstr); part += (short)(xstr[LSTR_XRLO]>>16); /* adv. rep. part off */ xmbrs += i; } } return 0; } /* convert */ static int readrec ( int *dst, int fid, int where, int *xstr ) { char *buf = (char *) #ifdef alloca alloca( xstr[LSTR_XLEN] ) #else mAlloc( xstr[LSTR_XLEN] ) #endif ; int got = 0; int ret = 0; if ( ! buf ) return -ERR_NOMEM; if ( 0 > where ) where = -where * xstr[LSTR_XLEN]; got = readblk( buf, xstr[LSTR_XLEN], fid, where ); ret = got ? got : convert( dst, buf, xstr ); #ifndef alloca mFree( buf ); #endif return ret; } /* readrec */ /* read and log */ static int readlog ( int *dst, int fid, int where, LDb *db, int set, int rec ) { int ret = readrec( dst, fid, where, DB_XSTR( db, set, rec ) ); if ( !ret && LOG_DO( LOG_TRACE ) ) LOG_STR( dst, lstrlib[ set ].name[ rec ] ); return ret; } /* readlog */ /* ************************************************************ access functions for the record types */ static int getOff ( LDb *db, int rowid, int xr ) { int rowix = rowid - 1; /* mfns count from 1 */ int xrf_block = rowix / 127; int blkix = xrf_block + 1; /* ... so do xrf blocks */ int off; if ( xrf_block < db->mmlen ) { if ( xr ) ((lblk*)db->mmap)[xrf_block][1+(int)(rowix % 127)] = rvi( xr ); else xr = rvi( ((lblk*)db->mmap)[xrf_block][1+(int)(rowix % 127)] ); } else { int err = 0; /* if ( LIO_LOCK() ) return -3; */ if ( xr ) { /* write */ if ( blkix <= db->xrlen ) { SWI( xr ); if ( 4 != lio_pwrite( &db->mst[MST_XRF], &xr, 4, xrf_block*512 + 4*(1 + (rowix % 127))) ) return 0; } else { /* extent */ lblk extend; memset( extend, 0, sizeof(extend) ); while ( db->xrlen < blkix ) { /* extend */ extend[0] = db->xrlen + 1; /* set blk id */ if ( blkix == extend[0] ) extend[1+(int)(rowix % 127)] = rvi( xr ); SWI( extend[0] ); if ( 512 != lio_pwrite( &db->mst[MST_XRF], extend, 512, db->xrlen*512) ) return 0; db->xrlen++; } } if ( blkix == db->xrf[LXRF_XPOS] ) db->xrf[ LXRF_XREC + (int)(rowix % 127) ] = xr; /* and go on read it back, just to check ... */ } if ( blkix != db->xrf[LXRF_XPOS] ) { int ret; LOG_DBG( LOG_VERBOSE, "fetching xrf block %d had %d", blkix, ! db->xrf ? -1 : db->xrf[LXRF_XPOS] ); ret = readlog( db->xrf, db->mst[MST_XRF], -xrf_block, db, LSET_MST, LSTR_XRF ); if ( ret ) { log_msg( LOG_ERROR, "\twhen fetching xrf block %d", blkix ); err = -1; } else if ( blkix == -db->xrf[LXRF_XPOS] ) { LOG_DBG( LOG_DEBUG, "hmmm ... negative" ); db->xrf[LXRF_XPOS] = blkix; } else if ( blkix != db->xrf[LXRF_XPOS] ) { log_msg( LOG_WARN, "bad xrf %d wanted %d", db->xrf[LXRF_XPOS], blkix ); err = -2; } } xr = db->xrf[ LXRF_XREC + (int)(rowix % 127) ]; /* LIO_RELE(); */ if ( err ) return err; } /* 21 bits (<<11) signed for the (512 byte) block ("xrmfb") 1 for the first block (offset 0) 0 means, never had such a record -1 and xrmfp=0: record removed from MST (there is no record at pos 0 in 1st block, since there resides the MST header) other negative value -x or pos!=0: record logically deleted, was at +x 1 bit (1<<10): this record is new and not yet inverted 1 bit (1<<9): this record is changed and not yet re-inverted 9 bits for the block-relative position ("xrmfp") */ off = (((xr & 0xfffff800) >> 2) - 0x200) | (0x1ff & xr); if ( 0 < (xr & ~0x600) ) { LOG_DBG( LOG_DEBUG, "offset for rowid %d is %d (blk %d pos %d) flg 0x%08x at %d[%d]", rowid, off, (xr>>11)&0xfffff, xr&0x1ff, xr&0x80000600, blkix, rowix%127 ); return off; } /* deleted */ log_msg( LOG_INFO, "offset for rowid %d is %d (blk %d pos %d) flg 0x%08x at %d[%d]", rowid, off, (xr>>11)&0xfffff, xr&0x1ff, xr&0x80000600, blkix, rowix%127 ); return 0; } /* getOff */ static int* getMfr ( LDb *db, int off, int *nxtoff ) { struct mfcxstr { int xstr[LSTR_LONGS(7+3)]; } my = *(struct mfcxstr *)DB_XSTR( db, LSET_MST, LSTR_MFR ); int head[1+7+3]; int len = 0, base; int *rec = 0; char *buf = 0; #ifdef alloca int notalloca = 0; #endif LOG_DBG( LOG_VERBOSE, "getting MFR at off %d", off ); if ( 498 < off % 512 ) log_msg( LOG_WARN, "blk pos > 498 in offset 0x%08x", off ); *head = *my.xstr; if ( readrec( head, db->mst[MST_MST], off, my.xstr ) ) { log_msg( LOG_ERROR, "\twhen reading MFR head at %d", off ); return 0; } /* log_str( LOG_VERBOSE, head, lstrlib[LSET_MST].name[LSTR_MFR] ); */ len = head[LMFR_RECL]; LOG_DBG( LOG_VERBOSE, "got MFR %d reclen %d", head[LMFR_MFN], len ); if ( nxtoff ) { *nxtoff = off + (0 < len ? len : -len); /* an odditiy */ if ( 1 & *nxtoff ) /* an odditiy */ (*nxtoff)++; /* round up to even */ if ( 498 < *nxtoff % 512 ) { *nxtoff += 512; *nxtoff &= ~0x1ff; } if ( *nxtoff > db->mfc[LMFC_NMFB]*512 + db->mfc[LMFC_NMFP] ) { LOG_DBG( LOG_VERBOSE, "at end of db: %d > %d*512+%hd", *nxtoff, db->mfc[LMFC_NMFB], db->mfc[LMFC_NMFP] ); *nxtoff = -1; } } if ( len < 0 ) { log_msg( LOG_INFO, "found deleted rec len %hd at offset %d", len, off ); return 0; } /* check external base length */ base = LONG2OFF(my.xstr[LSTR_XRLO]) + head[LMFR_NVF]*LONG2LEN(my.xstr[LSTR_XRLO]); if ( 0 > head[LMFR_BASE] || 0 > head[LMFR_NVF] || 0x8fff < head[LMFR_NVF] || len < head[LMFR_BASE] || len < base || (head[LMFR_NVF] && head[LMFR_BASE] < base) ) { log_msg( LOG_ERROR, "bad len %d base %d nvf %d need base %d at offset %d", len, head[LMFR_BASE], head[LMFR_NVF], base, off ); /* check alignment problem */ base = LONG2OFF(my.xstr[LSTR_XRLO]) + head[LMFR_STAT]*LONG2LEN(my.xstr[LSTR_XRLO]); if ( 0 > head[LMFR_NVF] || 0 > head[LMFR_STAT] || 0x8fff < head[LMFR_STAT] || len < head[LMFR_NVF] || len < base || (head[LMFR_STAT] && head[LMFR_NVF] < base) ) ; else log_msg( LOG_ERROR, "probably alignment problem, try -format aligned" ); goto cleanup; } /* set up external structure for this rec */ my.xstr[LSTR_SIZE] |= head[LMFR_NVF] << 16; /* occ of rep. part */ my.xstr[LSTR_XLEN] = len; /* internal base length */ base = LSTRLEN( *my.xstr ); /* internal len adjusted for the slightly longer base */ my.xstr[LSTR_ILEN] = len + base - head[LMFR_BASE]; /* the buffer */ rec = nrec( my.xstr ); if ( ! rec ) { log_msg( LOG_SYSERR, "could not alloc MFR of len %hd", my.xstr[LSTR_ILEN] ); goto cleanup; } buf = (char *) #ifdef alloca alloca( len ); /* first try faster alloca, but stack may be too limited for large records */ notalloca = ! buf; if ( notalloca ) buf = #endif mAlloc( len ); if ( ! buf ) { log_msg( LOG_SYSERR, "could not alloc MFR of len %hd", len ); goto cleanup; } if ( readblk( buf, len, db->mst[MST_MST], off ) ) { log_msg( LOG_ERROR, "\twhen reading MFR" ); goto cleanup; } #ifndef LDB_BIG_ENDIAN if ( LVAR_PAC != (DB_VARI & db->flags) ) { #endif if ( convert( rec, buf, my.xstr ) ) { log_msg( LOG_ERROR, "\twhen converting MFR" ); goto cleanup; } #ifndef LDB_BIG_ENDIAN } else { /* 10% faster */ Mfr *mfr = (Mfr*)buf; short *s = &mfr->dict->tag; int *f = rec + LMFR__FL; int *e = f + 3*head[LMFR_NVF]; rec[LMFR_MFN] = mfr->mfn; rec[LMFR_RECL] = mfr->recl; /* rec[LMFR_BWB] = mfr->bwbh<<16 | mfr->bwbl; rec[LMFR_BWP] = mfr->bwp; */ rec[LMFR_BASE] = mfr->base; rec[LMFR_NVF] = mfr->nvf; rec[LMFR_STAT] = mfr->stat; while ( f < e ) { *f++ = *s++; *f++ = *s++; *f++ = *s++; } } #endif if ( rec[LMFR_STAT] ) { log_msg( LOG_WARN, "found status %hd", rec[LMFR_STAT] ); goto cleanok; } /* do a consistency check */ if ( rec[LMFR_NVF] < 0 || rec[LMFR_BASE] < 0 ) { log_msg( LOG_ERROR, "found neg. field nvf %hd base %hd", rec[LMFR_NVF], rec[LMFR_BASE] ); goto cleanup; } /* now care for the field values */ { char *valsrc = buf+rec[LMFR_BASE]; char *recsta = ((char*)rec); char *valdst = recsta + base; int xbufl = rec[LMFR_RECL] - rec[LMFR_BASE]; int sumlens = 0; int i; for ( i=0; i < rec[LMFR_NVF]; i++ ) { int *d = &rec[LMFR__FL + i*LMFR__RL]; if ( d[LMFR_POS] < 0 || d[LMFR_LEN] < 0 ) { log_msg( LOG_ERROR, "bad field %d at off %d: negativ pos %hd or len %hd", i, off, d[LMFR_POS], d[LMFR_LEN] ); goto cleanup; } if ( d[LMFR_POS] + d[LMFR_LEN] > xbufl ) { log_msg( LOG_ERROR, "bad field %d at off %d: pos %hd + len %hd > buf %d", i, off, d[LMFR_POS], d[LMFR_LEN], xbufl ); goto cleanup; } sumlens += d[LMFR_LEN]; if ( sumlens > xbufl ) { log_msg( LOG_ERROR, "bad fields at off %d: sum of lengths %d > buf %d", off, sumlens, xbufl ); goto cleanup; } memcpy( valdst, valsrc+d[LMFR_POS], d[LMFR_LEN] ); d[LMFR_POS] = valdst - recsta; valdst += d[LMFR_LEN]; } } /* consistency check */ rec[LMFR_BWB] = /* "used" bytes */ rec[LMFR_RECL] = my.xstr[LSTR_ILEN]; rec[LMFR_BWP] = rec[LMFR_NVF]; /* avail fields = used fields */ rec[LMFR_BASE] = base; if ( LOG_TRACE <= log_lev ) LOG_STR( rec, lstrlib[LSET_MST].name[LSTR_MFR] ); goto done; cleanup: if ( nxtoff ) *nxtoff = -1; cleanok: if ( rec ) { mFree( rec ); rec = 0; } done: if ( buf #ifdef alloca && notalloca #endif ) mFree( buf ); if ( rec ) *rec = db->head.dbid; return rec; } /* getMfr */ /** write the record. If it doesn't yet have a mfn, assign one. NOTE: on a BIG_ENDIAN, anything but the mfn and recl will be frobbed after this call */ static int putMfr ( LDb *db, Mfr *mfr ) { int oldpos, newpos; int ret; if ( !(db->flags & DB_WRITABLE) ) { log_msg( LOG_ERROR, "db is not writable" ); return -1; } /* minimalist sanity check */ if ( mfr->nvf < 0 || mfr->base != 18 + 6*mfr->nvf || mfr->base > mfr->recl ) { log_msg( LOG_ERROR, "bad nvf/base/recl %d/%d/%d ", mfr->nvf, mfr->base, mfr->recl ); return -2; } db->flags |= DB_MODIFIED; if ( mfr->mfn ) { int block; if ( db->mfc[LMFC_NMFN] > mfr->mfn ) oldpos = getOff( db, mfr->mfn, 0 ); else { db->mfc[LMFC_NMFN] = mfr->mfn+1; oldpos = 0; } mfr->bwp = 511 & oldpos; block = 1 + (oldpos >> 9); /* blockno counting from 1 */ mfr->bwbl = 0xffff & block; mfr->bwbh = block >> 16; if ( db->mfc[LMFC_NMFN] <= mfr->mfn ) db->mfc[LMFC_NMFN] = mfr->mfn+1; } else { mfr->mfn = db->mfc[LMFC_NMFN]++; /* assign new mfn */ oldpos = 0; mfr->bwbl = mfr->bwbh = mfr->bwp = 0; } mfr->stat = 0; newpos = db->mflen; if ( 498 < (newpos & 511) ) /* round up to next block boundary */ newpos = ~511 & (newpos + 14); if ( 1 & newpos ) newpos++; db->mflen = newpos + mfr->recl; #ifdef LDB_BIG_ENDIAN { /* swap swap swap */ Dict *d = mfr->dict; short nvf = mfr->nvf; SWI( mfr->mfn ); SWS( mfr->recl ); SWS( mfr->bwbl ); SWS( mfr->bwbh ); SWS( mfr->bwp ); SWS( mfr->base ); SWS( mfr->nvf ); SWS( mfr->stat ); for ( ; nvf--; d++ ) { SWS( d->tag ); SWS( d->pos ); SWS( d->len ); } } #endif ret = lio_pwrite( &db->mst[MST_MST], (char*)mfr, rvs(mfr->recl), newpos ); #ifdef LDB_BIG_ENDIAN /* restore mnf, recl */ SWI( mfr->mfn ); SWS( mfr->recl ); #endif if ( ret != mfr->recl ) return log_msg( ERR_TRASH, "could not write Mfr %d bytes got %d", mfr->recl, ret ); getOff( db, mfr->mfn, (1 << (oldpos ? 9 : 10)) | (((newpos & 0xfffffe00) + 0x200) << 2) | (0x1ff & newpos) ); return 0; } /* putMfr */ static int putRec ( LDb *db, Rec *rec ) { int ret = 0, i; Mfr *mfr = 0; int buflen = 0; int reclen = 0; int contig = 1; #ifdef alloca int notalloca = 0; #endif const char *rbase = ((char *)rec) + rec->base; Dict *d; Field *f = rec->field; /* TODO: if not rec->len, delete ? */ for ( i = rec->len; i--; f++ ) { if ( ! f->len ) continue; if ( ! f->val ) LOG_OTO( cleanup, ( ERR_FAULT, "bad rec NULL val" ) ); contig = contig && (f->val == rbase + buflen); buflen += f->len; } reclen = 18 + 6*rec->len + buflen; if ( 1 & reclen ) reclen++; mfr = (Mfr*) #ifdef alloca alloca( reclen ); notalloca = ! mfr; if ( notalloca ) mfr = (Mfr*) #endif mAlloc( reclen ); if ( ! mfr ) LOG_OTO( cleanup, ( ERR_NOMEM, "could not alloc MFR of len %hd", reclen ) ); mfr->mfn = rec->rowid; mfr->recl = reclen; mfr->bwbl = mfr->bwbh = mfr->bwp = 0; mfr->base = 18 + 6*rec->len; mfr->nvf = rec->len; mfr->stat = 0; d = mfr->dict; f = rec->field; buflen = 0; for ( i = rec->len; i--; d++, f++ ) { d->tag = f->tag; d->pos = buflen; buflen += (d->len = f->len); } if ( contig ) memcpy( ((char*)mfr)+mfr->base, rbase, buflen ); else { char *mbase = ((char*)mfr)+mfr->base; d = mfr->dict; f = rec->field; for ( i = rec->len; i--; d++, f++ ) if ( d->len ) memcpy( mbase + d->pos, f->val, d->len ); } ret = putMfr( db, mfr ); if ( !ret && !rec->rowid ) rec->rowid = mfr->mfn; cleanup: if ( mfr #ifdef alloca && notalloca #endif ) mFree( mfr ); return ret; } /* putRec */ /* ************************************************************ access functions for plaintext db */ /* create a pointer from the least significant bytes of pos, len, fld buf must have db->ptrl bytes (up to 16 = 8+4+4) and the most strict alignment (i.e. 4 or 8) possible for db->ptrl returns buf */ static char *mkptr ( char *buf, LDb *db, unsigned pos, unsigned len, unsigned fld ) { switch ( db->ptr ) { case 0x0134: /* '4' */ ((unsigned*)buf)[0] = pos; if ( ~0xff & fld ) fld = 0; #ifdef LDB_BIG_ENDIAN /* the first = high order 3 bytes are len */ ((unsigned*)buf)[1] = (0xff&fld) | len<<8; #else /* the first = low order 3 bytes are len */ ((unsigned*)buf)[1] = (0xffffff&len) | fld<<24; #endif return buf; case 0x0044: /* 'D' */ ((unsigned*)buf)[0] = pos; ((unsigned*)buf)[1] = len; return buf; case 0x0035: /* '5' */ #ifdef LDB_BIG_ENDIAN /* the first = high order 5 bytes are pos */ *(lll*)buf = (0xffffff&len) | ((lll)pos)<<24; #else /* the first = low order 5 bytes are pos */ *(lll*)buf = pos | ((lll)len)<<40; #endif return buf; } /* TODO */ (void)fld; assert( 0 ); return 0; } /* read pointer, return len if 0x0f00 & db->ptr, fld must not be 0, else *fld is untouched */ static unsigned rdptr ( unsigned *pos, unsigned *fld, LDb *db, char *buf ) { switch ( db->ptr ) { case 0x0134: *pos = *(unsigned*)buf; *fld = ((unsigned char *)buf)[7]; #ifdef LDB_BIG_ENDIAN return ((unsigned*)buf)[1] >> 8; #else return 0xffffff & ((unsigned*)buf)[1]; #endif case 0x0044: *pos = *(unsigned*)buf; return ((unsigned*)buf)[1]; case 0x0035: #ifdef LDB_BIG_ENDIAN /* *pos = (*(unsigned*)(buf+1)); would bus error on sparc */ *pos = (unsigned) (*(lll*)buf >> 24); return 0xffffff & (unsigned)*(lll*)buf; /* last 3 bytes */ #else /* guess there is no little endian that needs alignment ? */ *pos = *(unsigned*)buf; /* use low order = first 4 of first 5 bytes */ return 0xffffff & (*(unsigned*)(buf+5)); #endif } /* TODO */ (void)fld; assert( 0 ); return 0; } static void setPtr ( LDb *db, int mfn, unsigned pos, unsigned len, unsigned fld ) { Ptr pt; if ( mfn < db->mmlen ) { mkptr( db->mmap + mfn*db->ptrl, db, pos, len, fld ); return; } lio_pwrite( &db->mst[MST_XRF], mkptr( pt.r, db, pos, len, fld), db->ptrl, mfn*db->ptrl ); } /* setPtr */ static unsigned getPtr ( unsigned *pos, unsigned *fld, LDb *db, int mfn ) { Ptr pt; return mfn < db->mmlen ? rdptr( pos, fld, db, db->mmap + mfn*db->ptrl ) : db->ptrl == lio_pread( &db->mst[MST_XRF], pt.r, db->ptrl, mfn*db->ptrl ) ? rdptr( pos, fld, db, pt.r ) : 0; } /* getPtr */ #if 0 static int putPlain ( LDb *db, Rec *rec ) { return 0; } /* putPlain */ #endif /** get text the original text is read contigously at base. the record is then cooked as requested: 0 well done: do full fixup, apply conversions and create fields. 1 english: do not create fields (rec->fields is 0), apply no conversions, but set rec->len to actual number of fields (counting if necessary). 2 raw: set len only if it's known from the pointer */ static Rec *dText ( LDb *db, int mfn, int raw ) { unsigned base, sz, pos, len, fld; /* #fields actually used */ unsigned n = 0; /* known #fields */ Rec *r, *x; Field *f, *fe; char *p, *q, *e; len = getPtr( &pos, &n, db, mfn ); LOG_DBG( LOG_TRACE, "dText %d pos %d len %d fld %d", mfn, pos, len, n ); if ( !len ) return 0; if ( raw ) fld = 0; else if ( !(fld = n) ) { fld = len / 36; /* assume one (costing 12 bytes) per 36 bytes data */ if ( fld < 8 ) /* small record is likely to have some short fields */ fld = 8; } base = BASESZ(fld); sz = base + len; r = (Rec*)mAlloc( sz ); p = ((char*)r) + base; if ( (int)len != lio_pread( &db->mst[MST_MST], p, len, pos ) ) { mFree( r ); return 0; } LOG_DBG( LOG_TRACE, "'%.*s'", len, p ); r->dbid = db->head.dbid; r->rowid = mfn; r->used = r->bytes = sz; r->fields = fld; r->base = base; r->len = n; if ( raw && (n || 1 != raw) ) return r; e = p + len; f = r->field; /* next field to assign */ fe = f + fld; /* end of assignable fields */ /* loop through buffer lines to count a/o assign count them in n while f < fe, also fix and assign them */ for ( n=0;;) { /* possibly 2 passes needed */ for ( ;p < e; p = q+1 ) { if ( !(q = memchr( p, LF, e-p )) ) q = e; /* > p */ if ( TAB != *p || !n ) { if ( f < fe ) { p += a2il( p, q-p, &f->tag ); if ( p < q && TAB == *p ) p++; f->len = q - (f->val = p); f++; /* f == r->field+n, as long as we don't hit fe */ } n++; continue; } /* continuation line */ if ( f != r->field+n ) continue; /* we ARE assigning & didn't loose sync at fe */ /* append to previous */ { char *dest = (char*)f[-1].val + f[-1].len; int dist = p - dest, l = q-p; *p = LF; memmove( dest, p, l ); memset( q-dist, ' ', dist ); /* cleanup */ f[-1].len += l; } } /* now n != 0, since initially p < e, since len != 0 */ if ( r->len && r->len != (int)n ) { log_msg( LOG_WARN, "rec %d len %d != ptr %d", mfn, n, r->len ); break; } if ( raw || (int)n <= r->fields ) /* all counted/assigned */ break; /* extend the record to n fields */ log_msg( LOG_INFO, "extending rec %d %d -> %d fields", mfn, fld, n ); fld = n; base = BASESZ(fld); sz = base + len; x = (Rec*)mAlloc( sz ); x->dbid = db->head.dbid; x->rowid = mfn; x->used = x->bytes = sz; x->fields = fld; x->base = base; x->len = n; p = ((char*)x) + base; e = p + len; q = ((char*)r) + r->base; memcpy( p, q, len ); memcpy( x->field, r->field, r->fields*sizeof(Field) ); for ( f=x->field, n=r->fields; n--; ) (f++)->val += p-q; n = r->fields; mFree( r ); r = x; f = r->field + n; fe = r->field + fld; p = (char*)f[-1].val + f[-1].len; /* seek behind the LF that delimited the last field */ while (LF != *p++) ; } r->len = n; if ( !raw && (DB_TXTMODE & db->flags) ) for ( f = r->field, fe = r->field + r->fields; f < fe; f++ ) for ( p = (char*)f->val, e = p+f->len; (p = memchr(p,VT,e-p)); ) *p++ = LF; return r; } /* dText */ static int pText ( LDb *db, Rec *r, const char *mark ) { char buf[128 + 65536]; unsigned pos = 0, len = 0, fld, off; char *p, *b; int ret; if ( !(db->flags & DB_WRITABLE) ) { log_msg( LOG_ERROR, "db is not writable" ); return -1; } if ( ! r->rowid ) r->rowid = db->mfc[LMFC_NMFN]++; /* assign new mfn */ else if ( db->mfc[LMFC_NMFN] <= r->rowid ) db->mfc[LMFC_NMFN] = r->rowid + 1; else { fld = 0; len = getPtr( &pos, &fld, db, r->rowid ); } p = b = 32768 >= r->used ? buf : mAlloc(128+2*r->used); *p++ = LF; *p++ = 'W'; *p++ = TAB; p += u2a( p, r->rowid ); *p++ = TAB; if ( pos ) { p += u2a( p, pos ); *p++ = '.'; p += u2a( p, len ); if ( fld ) { *p++ = '.'; p += u2a( p, fld ); } } *p++ = TAB; if ( mark ) { int l = strlen(mark); if ( l > 31 ) { log_msg( LOG_WARN, "mark '%.48s'%s has length %d", mark, l<48 ? "" : "...", l ); l = 31; } memcpy( p, mark, l ); p += l; } else { timeGtfm( p, 0 ); p += 17; } *p++ = LF; off = p - b; p += len = rSerB( p, r ) - 1; /* discard trailing blank line */ if ( len ) len--; /* discard last newline */ db->flags |= DB_MODIFIED; pos = db->mflen; db->mflen += p - b; ret = lio_pwrite( &db->mst[MST_MST], b, p - b, pos ); if ( ret == p - b ) { setPtr( db, r->rowid, pos+off, len, r->len ); ret = 0; } if ( buf != b ) mFree( b ); return ret; } /* pText */ /* ************************************************************ utilities */ static int search ( LDb *db, const char *key, LdbPost *post, Rec *rec, DXLoop *lp ) { int i, j, prefix, idx, ret, ock; int pos; int *leaf, *entry; char *term; int *xstr; struct { /* terms cursor */ char key[LDB_MAX_KEYLEN+1]; /* key or key prefix */ short klen; /* key length to compare */ char imin; /* minimum index to search */ char imax; /* maximum index to search */ int leaf[LDB_INDEXES][LDB_TERMBUF]; /* one leaf buffer per index */ short lpos[LDB_INDEXES]; /* next position in leaf, -1 if done */ } crs; short klen; /* length for initial locate */ int block[128]; /* buffer to read one block */ int blockpos = 0; LdbP *p = 0; if ( ! key ) key = "$"; /* prepare cursor struct */ memset( &crs, 0, sizeof(crs) ); /* tabula rasa */ crs.klen = strlen( key ); /* check for prefix match */ if ( post ) prefix = LDB_PFX & post->mode; else if ( (prefix = crs.klen && '$' == key[crs.klen - 1]) ) crs.klen--; /* check out minimum index to search */ for ( crs.imin=0; crs.klen > db->tlen[(int)crs.imin]; ) if ( LDB_INDEXES == ++(crs.imin) ) return log_msg( ERR_INVAL, "bad keylen %d key '%.64s'", crs.klen, key ); /* prepare key */ memset( crs.key, ' ', sizeof(crs.key)-1 ); { unsigned char *uc = (unsigned char*)crs.key; unsigned char *uk = (unsigned char*)key; for ( i=crs.klen; i--; ) uc[i] = db->ctab[LCS_UCASE].c[ uk[i] ]; } if ( prefix ) crs.imax = LDB_INDEXES-1; else { crs.imax = crs.imin; crs.klen = db->tlen[(int)crs.imin]; } log_msg( LOG_INFO, "search for '%.*s'%c", crs.klen, crs.key, prefix?'$':' ' ); key = crs.key; klen = crs.klen; if ( rec && rec->len ) { /* use last key from record to locate starting position */ key = rec->field[rec->len-1].val; klen = rec->field[rec->len-1].len; rec->len = 0; } for ( i=crs.imin; i<=crs.imax; i++ ) { /* find leaf positions */ int nFile = INV_N01 + 2*i; /* node file index */ int nStr = LSTR_N01 + 2*i; /* node struct index */ int *nstr = DB_XSTR( db, LSET_INV, nStr ); int lvl; short cmplen = klen <= db->tlen[i] ? klen : db->tlen[i]; pos = db->cnt[i][LCNT_POSR]; /* pos of root record */ j = 0; for ( lvl = 0; 0= nstr[LSTR_ILEN] ); *node = *nstr; ret = readlog( node, db->inv[nFile], 1-pos, db, LSET_INV, nStr ); if ( pos != node[LN0X_POS] /* wrong address */ || i+1 != node[LN0X_TYPE] /* wrong type */ || node[LN0X_OCK] < 1 /* no keys */ || 2*db->cnt[i][LCNT_ORDN] < node[LN0X_OCK] /* too many keys */ ) return log_msg( ERR_TRASH, "bad node pos %d type %d keys %d", node[LN0X_POS], node[LN0X_TYPE], node[LN0X_OCK] ); ock = node[LN0X_OCK]; for ( j=1; jp; xstr = DB_XSTR( db, LSET_INV, LSTR_IFP ); for (;;) { /* loop terms in prefix mode */ /* vars for postings: */ int infb, infp; /* block and pos where to read postings */ int added; /* postings added or marked per term */ int blkno; /* postings block number */ int remain = 0; /* postings to fetch from next block of segment */ int ifp[LIFP__FL]; /* postings header */ idx = -1; /* index to use */ /* compare index terms, load leafes if needed */ for ( i = crs.imin; i <= crs.imax; i++ ) { short cmplen = klen <= db->tlen[i] ? klen : db->tlen[i]; leaf = crs.leaf[i]; if ( leaf[LL0X_OCK] <= crs.lpos[i] ) { /* load */ int lFile = INV_L01 + 2*i; /* leaf file index */ int lStr = LSTR_L01 + 2*i; /* leaf struct index */ int *lstr = DB_XSTR( db, LSET_INV, lStr ); crs.lpos[i] = -1; reread: if ( ! (pos = leaf[LL0X_PS]) ) continue; LOG_DBG( LOG_DEBUG, "leaf %d", pos ); assert( (int)sizeof(crs.leaf[i]) >= lstr[LSTR_ILEN] ); *leaf = *lstr; ret = readlog( leaf, db->inv[lFile], 1-pos, db, LSET_INV, lStr ); if ( pos != leaf[LL0X_POS] /* wrong address */ || i+1 != leaf[LL0X_TYPE] /* wrong type */ || leaf[LL0X_OCK] < 1 /* no keys */ || 2*db->cnt[i][LCNT_ORDN] < leaf[LL0X_OCK] /* too many keys */ ) return log_msg( ERR_TRASH, "bad leaf pos %d type %d keys %d", leaf[LL0X_POS], leaf[LL0X_TYPE], leaf[LL0X_OCK] ); ock = leaf[LL0X_OCK]; /* advance to first term which is not too small (should be needed only for first leaf of an index) */ for ( j=0; j crs.lpos[i] ) continue; if ( 0 > idx ) { idx = i; continue; } /* compare this index next term to that of index idx */ /* assume that index w/ lower number has shorter keys */ ret = memcmp( ((char*)leaf)+leaf[LL0X__FL + crs.lpos[i]*LL0X__RL + LL0X_KEY], ((char*)crs.leaf[idx])+ crs.leaf[idx][LL0X__FL + crs.lpos[idx]*LL0X__RL + LL0X_KEY], db->tlen[idx] ); if ( 0 > ret ) idx = i; } if ( 0 > idx ) goto done; j = crs.lpos[idx]; leaf = crs.leaf[idx]; entry = leaf + LL0X__FL + j*LL0X__RL; term = ((char*)leaf) + entry[LL0X_KEY]; if ( memcmp( crs.key, term, crs.klen ) ) goto done; crs.lpos[idx]++; if ( rec ) { /* record the term */ /* field to assign */ Field *f = rec->field + rec->len; short tlen = db->tlen[idx]; /* end of available buffer */ char *b = rec->len ? (char*)f[-1].val /* before previously assigned field */ : ((char*)rec + rec->bytes); /* end of record */ while ( tlen && ' ' == term[tlen-1] ) tlen--; b -= tlen; if ( b < (char*)(f+1) ) /* no space left on device */ goto done; /* probably we're nuking the locator now: */ memcpy( b, term, tlen ); f->tag = 0; f->val = b; f->len = tlen; rec->len++; /* reset key from locator to prefix */ key = crs.key; klen = crs.klen; } if ( ! post && ! lp ) continue; /* collect postings */ infb = entry[LL0X_INFB]; infp = entry[LL0X_INFP]; /* the IFP file is organized in blocks of 128 longs. 1st int is block number followed by 127 data. postings are organized in chained segments so that each segment fits within one such block. a segment has five longs header, giving number of postings and pointer to next segment. */ added = 0; for ( blkno=0; infb; blkno++ ) { /* segments */ LdbP merge[127/2]; /* buffer to collect new postings */ int *base; /* start of data */ int *b; /* start of postings */ int n; /* max postings in this seg's 1st block */ int xlen; /* external length to read */ int f = post ? post->fil - 1 : 0; /* highest pos to consider in given postings */ int m = 0; /* fill merge buffer */ int k; /* loop segment */ if ( infp > 127-2-5 ) { return log_msg( ERR_TRASH, "found bad IFP pos %d blk %d for %.*s", infp, blkno, klen, key ); } if ( remain ) { /* consecutive block of same segment */ n = remain; if ( n > 127/2 ) n = 127/2; xlen = 8*n; } else { n = (127 - 5 - infp)/2; xlen = 20 + 8*n; } pos = (infb - 1) * 512 + (infp + 1) * 4; if ( blockpos && !((pos-blockpos) >> 9) /* 0 <= (pos-blockpos) < 512 */ && pos+xlen <= blockpos+ 1 + (0x1ff & ~blockpos) ) base = block + (pos - blockpos)/sizeof(int); else { int blklen = 1 + (0x1ff & ~pos); assert( xlen <= blklen ); assert( blklen <= (int)sizeof(block) ); assert( 0 == (0x1ff & (pos + blklen)) ); ret = readblk( block, blklen, db->inv[INV_IFP], pos ); if ( ret ) return log_msg( ERR_IO, "\twhen reading IFP" ); blockpos = pos; base = block; } if ( remain ) { /* no header to convert */ remain -= n; b = base; /* no header */ } else { assert( (int)sizeof(ifp) >= xstr[LSTR_ILEN] ); *ifp = *xstr; ret = convert( ifp, (char *)base, xstr ); if ( ret ) return log_msg( ERR_TRASH, "\twhen converting IFP header" ); if ( n > ifp[LIFP_SEGP] ) n = ifp[LIFP_SEGP]; remain = ifp[LIFP_SEGP] - n; b = base+5; /* after header */ } LOG_DBG( LOG_VERBOSE, "key %d.%d '%.*s' blk %d post %d/%d r %d xlen %d at b/p %d.%d=%d", leaf[LL0X_PS], j, db->tlen[idx], term, blkno, n, ifp[LIFP_TOTP], remain, xlen, infb, infp, pos ); if ( LOG_DO( LOG_TRACE ) ) LOG_STR( ifp, lstrlib[ LSET_INV ].name[ LSTR_IFP ] ); assert( (size_t)n <= sizeof(merge)/sizeof(merge[0]) ); if ( lp ) { Key kbf; Hit hit; unsigned char tlen = (unsigned char) db->tlen[idx]; while ( tlen && ' ' == term[tlen-1] ) tlen--; memcpy( kbf.byt, term, kbf.len = tlen ); for ( k=0; k> 16); hit.pos = (unsigned short)ppos; if ( lp->cb( lp->me, &kbf, &hit ) ) goto done; } } if ( post ) for ( k=n; k--; ) { /* loop backwards (for the fun of it) postings in segment */ int prow, ptag, ppos; unsigned char *c = (unsigned char *)&b[k*2]; LdbP e; /* the entry */ LdbP samerow; /* highest possible entry w/ same row as e */ #ifdef LDB_BIG_ENDIAN /* the 8 bytes of a posting are BIG ENDIAN ! */ memcpy(e.bytes,c,8); #else e.bytes[0] = c[7]; e.bytes[1] = c[6]; e.bytes[2] = c[5]; e.bytes[3] = c[4]; e.bytes[4] = c[3]; e.bytes[5] = c[2]; e.bytes[6] = c[1]; e.bytes[7] = c[0]; #endif prow = LDBP_ROW( &e ); ptag = LDBP_TAG( &e ); ppos = LDBP_POS( &e ); LOG_DBG( LOG_VERBOSE, "post %d.%hd pos %06x key '%.*s'", prow, ptag, ppos, db->tlen[idx], term ); if ( 0 >= ptag /* bad tag */ || !prow || prow >= db->mfc[LMFC_NMFN] /* bad mfn */ ) continue; if ( ! post || (post->cut && prow >= post->cut) || (post->tag && post->tag != ptag) ) continue; if ( prow < post->skp ) /* quickly bail out on skip condition */ break; LDBP_SETROWTOP( &samerow, &e ); /* for mfn comparison */ /* sweep down to postings for the same row as e ... */ while ( f >= 0 && LDBP_GT( p+f, &samerow ) ) f--; if ( LDB_AND & post->mode ) { int l; /* loop postings for same row, mark all (that are near enough) */ LDBP_SETROWBOT( &samerow, &e ); /* for mfn comparison */ /* NOTE: postings for row are GT than bottom even if marked */ for ( l = f; l>=0 && LDBP_GT( p+l, &samerow ); l-- ) { if ( post->near ) { int dist; if ( ptag != LDBP_TAG( p+l ) ) continue; if ( LDB_NEAR_G != post->near ) { dist = LDBP_POS( p+l ) - LDBP_POS( &e ); if ( dist < 0 ) dist = -dist; if ( 0 < post->near ? post->near < dist : -post->near != dist /* exact $$$$ */ ) continue; } } LDBP_SETMARK( p+l ); added++; } } else { /* OR mode */ int add; if ( ! post->near ) /* add if row not found: ignore details */ add = 0 > f || prow > LDBP_ROW( p+f ); else { /* add if no exact match */ int l; /* NOTE: we don't use mark bit in OR mode, do we ? */ for ( l = f; l>=0 && LDBP_GT( p+l, &e ); l-- ) ; add = 0 > l || LDBP_GT( &e, p+l ); } if ( add ) merge[ m++ ] = e; } } /* for postings in segment */ if ( m ) { /* merge in the merge buffer */ LdbP *mm = merge; added += m; for ( k = post->fil += m; m && k--; ) { LdbP src; if ( k < m || LDBP_GT( mm, &p[k-m] ) ) { src = *mm++; m--; LOG_DBG( LOG_DEBUG, "merging %d at %d", LDBP_ROW(&src), k ); } else src = p[k-m]; if ( k < post->len ) p[k] = src; else { /* set cut */ int row = LDBP_ROW( &src ); if ( row < post->cut || !post->cut ) post->cut = row; } } if ( post->fil > post->len ) post->fil = post->len; if ( post->cut ) /* postings for cut row are unreliable */ while ( post->fil && post->cut <= LDBP_ROW(p+post->fil-1) ) post->fil--; } if ( remain ) { /* advance to start of next block */ infb++; infp = 0; } else { infb = ifp[LIFP_NXTB]; infp = ifp[LIFP_NXTP]; } } /* for segments */ LOG_DBG( LOG_VERBOSE, "added %d postings for key '%.*s'", added, db->tlen[idx], term ); } /* for terms in prefix/postings mode */ done: if ( post /* fixup */ && LDB_AND & post->mode && !(LDB_KEEPMARKS & post->mode) ) { int mark = LDB_NOT & post->mode ? 0 : 0x8000; j=0; for ( i=0; ifil; i++ ) if ( mark == LDBP_MARK(p+i) ) { LDBP_CLRMARK(p+i); p[j++] = p[i]; } post->fil = j; } return ! rec ? 0 : rec->len; } /* search */ static int ldb_last_path_sep (const char *path) { char *p2; int i2; #ifdef WIN32 char *p3; int i3; #endif if (! path) { return -1; } p2 = strrchr (path, '/'); i2 = p2 ? p2 - path : -1; #ifdef WIN32 p3 = strrchr (path, '\\'); i3 = p3 ? p3 - path : -1; if (i3 > i2) { i2 = i3; } #endif return i2; } /* ldb_last_path_sep */ static int ldb_open (const char *dbname, Rec *dbpar, Rec *syspar, Fdt *fdt) { LDb ndb, *db; int i, plen, sz, dbid, uc, lck = LIO_TLOCK; /* WLOCK only on special demand */ int ret = 0, invret = -1, lbtret = 0, autoformat = 1, writable = -1; int gotopt = 0, txtfd = 0, copyidx = 0; char *autoenc = 0; char *p, *q; char buf[65536+1]; /* need 64K buf for copying DO NOT SHRINK !!! */ char path[ PATH_MAX ]; memset( &ndb, 0, sizeof(ndb) ); /* these should be 0 by memsetting to 0 anyway ... */ ndb.path = 0; ndb.mmap = 0; ndb.flags |= DB_MMAP; /* it mean's: we'll try */ /* loglevel */ if ( 0 <= (i = rInt2(dbpar, syspar, OPENISIS_SLOGV, -1)) ) cLog( i, 0 ); /* prepare name ... */ if (! dbname) { if (! dbpar) return log_msg( ERR_FAULT, "ldb_open: dbname not given"); dbname = rString (dbpar, OPENISIS_DNAME, 0, buf, sizeof(buf)); if (! dbname) return log_msg( ERR_FAULT, "ldb_open: no dbname parameter"); } plen = strlen (dbname); if (0 >= plen) return log_msg( ERR_FAULT, "ldb_open: empty dbname"); if ( sizeof(buf) <= (unsigned)plen || sizeof(path) <= (unsigned)(plen + 4 + 1) ) return log_msg( ERR_FAULT, "ldb_open: dbname too long '%s'", dbname); if ('/' == dbname[plen - 1] #ifdef WIN32 || '\\' == dbname[plen - 1] #endif ) return log_msg( ERR_FAULT, "ldb_open: must not specify directory as dbname '%s'", dbname); if (DBNLEN > plen) strcpy(ndb.head.name, dbname); else { int i1 = 1 + plen - DBNLEN ; int i2 = ldb_last_path_sep (dbname); if (0 <= i2 && plen > ++i2 && i2 > i1) { i1 = i2; } strncpy(ndb.head.name, dbname + i1, DBNLEN - 1) [DBNLEN - 1] = 0; log_msg( LOG_WARN, "ldb_open: truncating dbname '%s' to '%s'", dbname, ndb.head.name); } /* ... and path */ strcpy(path, dbname); if (! IsAbsPath (path)) { int plen2; if ( (dbpar || syspar) && (p = rString2 (dbpar, syspar, OPENISIS_DPATH, buf, sizeof(buf))) ) { plen2 = strlen (p); if (sizeof(path) <= (unsigned)(plen + plen2 + 4 + 1 + 1)) return log_msg( ERR_FAULT, "ldb_open: dbname or dbpath too long: %d %d '%s'", plen, plen2, path); memmove (path + 1 + plen2, path, 1 + plen); path[plen2] = '/'; memcpy (path, p, plen2); plen += 1 + plen2; } if ( !IsAbsPath(path) && syspar && (p = rString(syspar, OPENISIS_SPATH, 0, buf, sizeof(buf))) ) { plen2 = strlen(p); if (sizeof(path) <= (unsigned)(plen + plen2 + 4 + 1 + 1)) return log_msg( ERR_FAULT, "ldb_open: dbname or syspath too long: %d %d '%s'", plen, plen2, path); memmove(path + 1 + plen2, path, 1 + plen); path[plen2] = '/'; memcpy(path, p, plen2); plen += 1 + plen2; } } /* name and path */ /* more init AFTER honoring verbosity */ if ( ! init ) { lstr_auto(0); init = !0; } for ( dbid=0; dbidhead.dbid = dbid; /* preset record sizes */ db->mfc[0] = *DB_XSTR( db, LSET_MST, LSTR_MFC ); db->xrf[0] = *DB_XSTR( db, LSET_MST, LSTR_XRF ); db->cnt[0][0] = db->cnt[1][0] = *DB_XSTR( db, LSET_INV, LSTR_CNT ); /* isis-1 index term lengths */ db->tlen[0] = 10; db->tlen[1] = 30; /* only the packed little endian ("DOS") format is writable test later ... if ( LVAR_PAC != (DB_VARI & db->flags) ) writable = 0; */ db->path = mDup( path, plen+1 ); /* save path */ memcpy( path+plen, ".???", 5 ); if ( dbpar ) dbpar = rDup(dbpar, 0, 0); /* check options file and extension case */ if ( 0 > (i = lio_open( setext(path,EXT_TXT_OPT,uc=0), OPEN_RDIF )) && 0 > (i = lio_open( setext(path,EXT_TXT_OPT,uc=OPEN_UC), OPEN_RDIF )) ) uc = autocase( db->path ); else { if ( 0 < (sz = lio_size(i)) ) { p = sz < (int)sizeof(buf) ? buf : mAlloc(sz); if ( (gotopt = (sz == lio_read( &i, p, sz ))) ) rDeser( &dbpar, p, sz, 0 ); log_msg( LOG_INFO, "reading %d bytes options from '%s' %s", sz, path, gotopt ? "ok" : "nok" ); if ( buf != p ) mFree( p ); } lio_close( &i, LIO_INOUT ); } lck |= uc; if ( (dbpar || syspar) && 0 <= (i = rInt2(dbpar, syspar, OPENISIS_DRO, -1))) writable = !i; /* explicit 0/1 */ /* open files */ /* trad. index is never openend writable. */ invret = openfiles( db->inv, path, EXT_INV, INV_FILES, uc|OPEN_RDIF ); if (dbpar || syspar) { char fmtstr[32]; if (rString2 (dbpar, syspar, OPENISIS_DTYPE, fmtstr, sizeof(fmtstr))) { if (! strcmp ("aligned", fmtstr)) { db->flags |= LVAR_ALI; autoformat = 0; } else if (! strcmp ("naligned", fmtstr)) autoformat = 0; } } if ( autoformat ) { if ( invret ) log_msg( LOG_WARN, "cannot guess format -- no inverted file" ); else { unsigned len = lio_size( db->inv[INV_CNT] ); if ( 56L == len ) { db->flags |= LVAR_ALI; autoenc = "iso8859-1"; /* writable = 0; we do not write aligned format */ } else if ( 52L == len ) autoenc = "cp850"; else log_msg( LOG_WARN, "cannot guess format -- bad .cnt len %d", len ); log_msg( LOG_INFO, "using autoformat %saligned for .cnt len %d", (db->flags & LVAR_ALI) ? "":"un", len ); } } /* data */ #ifdef NOTXTDB if ( !(ret = openfiles( db->mst, path, EXT_MST, MST_FILES, lck|OPEN_ASIS|LIO_CREAT )) ) writable = 0; else if (0 > ret) #else if ( 0 <= (ret = openfiles( &txtfd, path, EXT_TXT, 1, lck|LIO_SYNC|(writable?OPEN_ASIS:OPEN_RDIF) )) ) { /* .txt exists: use it */ if ( ret ) writable = 1; else if (1 == writable) { log_msg( LOG_ERROR, "file '%s' is readonly", path ); goto cleanup; } else writable = 0; } else if ( 0 <= (ret = openfiles( db->mst, path, EXT_MST, MST_FILES, lck|((writable && !(db->flags & LVAR_ALI))?OPEN_ASIS:OPEN_RDIF) )) && (ret || 1!=writable) ) { /* trad. files are ok */ if ( !ret ) writable = 0; } else if ( 1 != (ret = openfiles( &txtfd, path, EXT_TXT, 1, lck|(ret ? LIO_SYNC : 0)|OPEN_NEW )) ) /* don't sync on autoconv */ #endif goto cleanup; /* MW: creation mode? KR: ugo+rw & ~umask */ if ( 1 == (lbtret = openfiles( &db->oxi.fd, path, EXT_LBT, 1, lck|(writable?OPEN_ASIS:OPEN_RDIF) )) ) lbtret = 0; else if ( !writable ) ;/* no problem */ else if ( !lbtret ) { /* exists ro */ log_msg( LOG_ERROR, "file '%s' is readonly", path ); goto cleanup; } else { /* create and copy to oxi */ if ( 1 != openfiles( &db->oxi.fd, path, EXT_LBT, 1, lck|OPEN_NEW ) ) goto cleanup; lbtret = 0; copyidx = 1; } if ( db->mst[MST_MST] ) { /* care for the traditionals */ if ( (ret = readlog( db->mfc, db->mst[MST_MST], 0, db, LSET_MST, LSTR_MFC )) ) { /* NEW goto cleanup; */ memset( db->mfc, 0, sizeof(db->mfc) ); db->mfc[LMFC_NMFN] = 1; db->mfc[LMFC_NMFB] = 1; db->mfc[LMFC_NMFP] = 64; db->mflen = 64; } else { /* int lastblock = (db->mflen = lio_size( db->mst[MST_MST] ))/512; if ( 511 & db->mflen ) lastblock++; counting from 1 the next record's block should be either the last one we have or the next one to follow if ( db->mfc[LMFC_NMFB] != lastblock && db->mfc[LMFC_NMFB] != lastblock+1 ) log_msg( LOG_VERBOSE, "NMFB mismatch: NMFB %d ~ %d", db->mfc[LMFC_NMFB], lastblock ); */ /* set LOGICAL mf length */ db->mflen = (db->mfc[LMFC_NMFB]-1)*512 + db->mfc[LMFC_NMFP]; } db->ptrl = 512; db->xrlen = lio_size( db->mst[MST_XRF] ) / 512; if ( (DB_MMAP & db->flags) && db->xrlen && db->xrlen*512 == lio_mmap( &db->mst[MST_XRF], (void**)&db->mmap, db->xrlen*512 ) ) db->mmlen = db->xrlen; } if ( txtfd ) { int remake = 0; /* TODO: make on-demand preparation even faster using buffered IO */ if ( db->mst[MST_MST] ) { /* copy to new empty txt */ int end = db->mfc[LMFC_NMFN]; /* max recsize for traditionals is 32K. field values may double, if consisting entirely of newlines. rec->used may be more than 32K, since we 12 bytes per field. However, we know there are only sign+5digits+tab+newline used per tag, fitting within 2* the original 6 bytes per field. */ if ( gotopt && 0 < (i = lio_open( setext(path,EXT_TXT_OPT,uc), LIO_RD )) ) { /* copy the options file */ log_msg( LOG_INFO, "copying %d bytes options", lio_size(i) ); while ( 0 < (sz = lio_read( &i, buf, sizeof(buf)-1 )) ) lio_write( &txtfd, buf, sz ); if ( LIO_INOUT & i ) { /* is supposed to autoclose */ log_msg( LOG_WARN, "tss tss tss ..." ); lio_close( &i, LIO_INOUT ); } } log_msg( LOG_INFO, "copying traditional data" ); db->flags |= DB_OPEN; /* pretend */ buf[0] = LF; for ( i=1; iflags &= ~DB_OPEN; /* pret end */ remake = 1; if ( db->mmap ) lio_mmap( 0, (void**)&db->mmap, db->mmlen*512 ); db->mmlen = 0; closefiles( db->mst, MST_FILES ); } /* copying */ db->mst[MST_MST] = txtfd; db->mflen = lio_size( db->mst[MST_MST] ); db->ptr = 0x0134; /* should be config opt */ if ( !remake ) { /* other reasons why we should remake */ unsigned short ptr; unsigned isix = GETINT(ISIX); unsigned magic; remake = 1; if ( 0 > (db->mst[MST_XRF] = lio_open( setext(path,EXT_TXT_PTR,uc), LIO_SEEK|(writable?LIO_RDWR:LIO_RD) )) ) log_msg( LOG_INFO, "'%s' not found", path ); else if ( 6 != lio_read(&db->mst[MST_XRF],buf,6) ) log_msg( LOG_WARN, "'%s' too short", path ); else if ( isix != (magic = GETINT(buf)) ) /* FOO! */ log_msg( LOG_WARN, "'%s' has black magic 0x%08x", path, magic ); /* TODO: save that foo if it doesn't read ISIX ? */ else if ( 0xf000 & (ptr = GETSHORT(buf+4)) ) /* bad endianess */ log_msg( LOG_WARN, "'%s' has bad endianess type 0x%04x", path, ptr ); else if ( (db->ptr && db->ptr != ptr) ) /* other type configured */ log_msg( LOG_WARN, "'%s' type 0x%04x != cfg 0x%04x", path, ptr, db->ptr ); else if ( lio_time(db->mst[MST_XRF]) < lio_time(db->mst[MST_MST]) ) log_msg( LOG_WARN, "'%s' older than data", path ); else { db->ptr = ptr; remake = 0; } } if ( ! db->ptr ) { db->ptr = 0x0134; /* m*256 + l*16 + k, doc/Serialized */ /* BTW: 0x34 is ASCII digit '4', so it's ISIX4^A on little endian */ db->ptrl = 8; } else { /* fix unsupported type */ unsigned m = 0xf&(db->ptr>>8); unsigned l = 0xf&(db->ptr>>4); unsigned k = 0xf&db->ptr; int mod = 0; if ( m > 4 ) { m = 4; mod = 1; } if ( l > 4 ) { l = 4; mod = 1; } if ( k > 4 ) { k = 4; mod = 1; } /* TODO: allow 8 with large files */ /* total ptr bytes = sum(nibbles) <= 45, but won't use more than 8+4+4 */ if ( mod ) { log_msg( LOG_WARN, "fixing unsupported ptr type 0x%04x", db->ptr ); db->ptr = (unsigned short)(m<<8 | l<<4 | k); remake = 1; } db->ptrl = k+l+m; } if ( remake ) { Ptr pt; unsigned base = 0; /* of current block */ unsigned pos = 0; /* of last record */ unsigned fld = 0; /* of last record */ unsigned nmfn = 0; /* next mfn = maxmfn+1 */ unsigned xmfn = 0; /* explicitly given */ char op = 0; int more; /* buf not empty flag */ char *last; /* of current block */ lio_close( &db->mst[MST_XRF], LIO_INOUT ); if ( 0 > (db->mst[MST_XRF] = lio_open( setext(path,EXT_TXT_PTR,uc), OPEN_BLANK )) ) goto cleanup; /* write signature */ memcpy( pt.r, "ISIX", 4 ); memcpy( pt.r+4, &db->ptr, 2 ); memcpy( pt.r+6, ":)", 2 ); if ( 8 < db->ptrl ) memset( pt.r+8, ')', db->ptrl - 8 ); lio_pwrite( &db->mst[MST_XRF], pt.r, db->ptrl, 0 ); /* loop the masterfile */ lio_seek( &db->mst[MST_MST], 0 ); last = (p = buf) + lio_read( &db->mst[MST_MST], buf, 8192 ) - 1; more = last > buf; /* one byte is no byte ;) */ if ( more && LF == *p ) { /* no options: no \n\n */ nmfn = pos = 1; p++; } for (;;) { /* records */ unsigned len, mfn; for (;;) { /* lines and stuff to end of record */ if ( p < last ) { /* have one lookahead */ if ( LF != *p++ ) continue; /* the tight loop ... or use memchr ? */ if ( LF != *p ) { /* now p <= last */ if ( fld || !(0xc0 & *p) ) { /* < '@', 'A', ... */ if ( TAB != *p ) /* no continuation */ fld++; continue; } fld++; /* count field, unless we really recognize a opline */ if ( 'Z' < *p ) continue; /* now we have '@'...'Z' at start of 1st line */ sz = last - p; /* avail after p */ if ( sz && TAB != p[1] ) /* no opline */ continue; switch (*p) { case 'D': case 'I': case 'W': break; /* give it a try */ default: log_msg( LOG_WARN, "unknown opline %c at mfn %d", *p, nmfn ); continue; } if ( sz > 127 ) /* longer -> no opline */ sz = 127; if ( ! sz || ! (q = memchr(p+1, LF, sz)) ) { if ( sz >= 127 || ! more ) continue; /* too long or undelimited last */ p--; /* back to \n, so we come here again */ goto gimmemore; } if ( q < p+3 || p[2] < '0' || '9' < p[2] ) continue; /* TODO: take a closer look at whether the whole line makes sense */ if ( op ) { /* yeah, two metas in sequence! weird stuff! */ p--; /* step back to newline */ pos = base+(p-buf); /* fake pos as if we had no line at all */ break; /* go handle the PREVIOUS opline */ } op = *p; xmfn = a2i( p+2, q-p-2 ); fld--; /* uncount this line */ pos = base + (q-buf) + 1; /* start after q */ continue; } break; } gimmemore: LOG_DBG( LOG_DEBUG, "MORE %d at pos %d base %d p +%d last +%d", more, pos, base, p-buf, last-buf ); if ( !more ) goto schicht; /* german: done */ base += p - buf; /* shift out bytes before p */ len = last-p; /* bytes to keep after p; < 128 */ if ( len ) /* we're probing for more lookahead */ memmove( buf, p, 1+last-p ); else /* typically */ *buf = *p; /* but save the last dance */ p = buf; last = buf + len; /* reload */ if ( 0 < (sz = lio_read( &db->mst[MST_MST], buf+1+len, 8192 )) ) { last += sz; continue; } more = 0; /* but yet, finish this up */ /* since *buf = *last was the files last character, we'd expect a newline */ if ( last == p ) p = buf+(LF==*buf ? 1 : 2); /* pretend buf started \n */ if ( ! len ) break; /* else try again opline */ } /* lines and stuff */ /* now p is on a delimiting blank lines \n -- or such ... */ len = base + (p-buf) - pos; /* >= 0 */ mfn = xmfn ? xmfn : nmfn; log_msg( LOG_INFO, "ptr %c %d(%d/%d) pos %d len %d", op?op:'>', mfn, xmfn, nmfn, pos, len ); if ( base + (p-buf) < pos ) /* FOO !!! */ len = 0; if ( len ) /* could have been completely empty */ len--; /* mute last \n */ if ( 'D' == op && len ) /* FOO !!! */ len = 0; if ( mfn && (len || op) ) lio_pwrite( &db->mst[MST_XRF], mkptr( pt.r, db, pos, len, fld), db->ptrl, mfn*db->ptrl ); pos = base + (p-buf) + 1; /* next starts after p */ if ( 'D' != op ) { /* 'D'elete does not lead to implicit reuse */ if ( op && nmfn < xmfn ) nmfn = xmfn; nmfn++; /* continue after this */ } xmfn = fld = op = 0; } schicht: ; } /* remake */ db->mfc[LMFC_NMFN] = db->xrlen = lio_size( db->mst[MST_XRF] ) / db->ptrl; if ( (DB_MMAP & db->flags) && db->xrlen && db->xrlen*db->ptrl == lio_mmap( &db->mst[MST_XRF], (void**)&db->mmap, db->xrlen*db->ptrl ) ) db->mmlen = db->xrlen; log_msg( LOG_INFO, "mapped %d*%d = %d", db->xrlen, db->ptrl, db->xrlen*db->ptrl ); db->flags |= DB_TXTOPEN; db->flags &= ~DB_VARI; /* clear alignment and such */ } /* if ( txtfd ) */ /* supporting files, ctables */ p = buf; if ( 0 >= (sz = lio_slurp( &p, sizeof(buf), setext(path,EXT_SUP_ACT,uc), 1 )) || lcs_mktab( db->ctab+LCS_CTYPE, p, sz, LCS_A ) ) memcpy( db->ctab+LCS_CTYPE, lcs_latin1_ct, sizeof(db->ctab[0]) ); if ( 0 >= (sz = lio_slurp( &p, sizeof(buf), setext(path,EXT_SUP_UCT,uc), 1 )) || lcs_mktab( db->ctab+LCS_UCASE, p, sz, 0 ) ) memcpy( db->ctab+LCS_UCASE, lcs_latin1_uc, sizeof(db->ctab[0]) ); /* fill header */ if (! fdt) { if ( (p = rString (dbpar, OPENISIS_DFDT, 0, buf, sizeof(buf))) ) { Rec *recfdt = 0; Db *dbfdt = nDbByName (openisis_stub0, p); if ( dbfdt) recfdt = dRead (dbfdt->dbid, 1); else { int idfdt = ldb_open (p, 0, syspar, 0); if (0 <= idfdt) { recfdt = dRead (idfdt, 1); /*MMM*/ cDClose (idfdt); } } if (recfdt) fdt = fRec2Fdt (recfdt); } else if ( gotopt ) fdt = fRec2Fdt(dbpar); if (! fdt) fdt = fFromFile (path); } db->head.fdt = fdt; if (fdt) log_msg( LOG_INFO, "have %d fdt entries for %s", fdt->len, db->head.name); else log_msg( LOG_INFO, "have no fdt for %s", db->head.name); db->head.tms = timeUpd(0); /* what watch? */ log_msg( LOG_INFO, "tms %d for %s", db->head.tms, db->head.name); /* set path and name */ if (0 <= (i = ldb_last_path_sep (db->path))) { if (i) strncpy(path, db->path, i)[i] = 0; else strcpy (path, "/"); dbpar = rSet (dbpar, RCHG | RDIS, OPENISIS_DPATH, path, 0); } dbpar = rSet (dbpar, RCHG | RDIS, OPENISIS_DNAME, db->head.name, 0); /* set encoding */ if (!(p = rString (dbpar, OPENISIS_DENC, 0, buf, sizeof(buf)))) if ( (syspar && (p = rString (syspar, OPENISIS_DENC, 0, buf, sizeof(buf)))) || (p = autoenc) ) dbpar = rSet(dbpar, RDIS, OPENISIS_DENC, p, 0); if ( p ) log_msg( LOG_INFO, "using encoding %s for %s", p, db->head.name); db->head.cfg = dbpar; /* done */ db->flags |= DB_OPEN; if ( writable && LVAR_PAC == (DB_VARI & db->flags) ) db->flags |= DB_WRITABLE; /* if ( (dbpar || syspar) && 0 < rInt2(dbpar, syspar, OPENISIS_DDUMP, -1) ) { int off = 0; int *r; do { if ( (r = ldb_readRecAtOff(dbid,off,&off)) ) mFree( r ); } while ( 0 < off ); exit(0); } */ /* init oxi */ if ( writable ) db->oxi.flg |= LBT_WRITE; if ( (p = getenv("OXITYP")) && 0 < (i = atoi(p)) && 4 > i ) db->oxi.typ = i << 4; if ( !lbtret && !lbt_init( &db->oxi ) ) db->flags |= DB_LBTOPEN; if ( ! invret && ! (ret = readlog( db->cnt[0], db->inv[INV_CNT], 0, db, LSET_INV, LSTR_CNT )) && ! (ret = readlog( db->cnt[1], db->inv[INV_CNT], -1, db, LSET_INV, LSTR_CNT )) ) { if ( lbtret ) db->flags |= DB_INVOPEN; else { if ( copyidx ) { DXLoop l; log_msg( LOG_INFO, "copying traditional index" ); lbtret = 0; memset( &l, 0, sizeof(l) ); l.me = & db->oxi; l.cb = (DXCb*)cXAdd; lbt_batch( & db->oxi, 5 ); search( db, 0, 0, 0, &l ); cXAdd( & db->oxi, 0, 0 ); } closefiles( db->inv, INV_FILES ); } } return dbid; cleanup: /* cleanup ... */ db->flags = 0; closefiles( &db->oxi.fd, 1 ); closefiles( &txtfd, 1 ); closefiles( db->inv, INV_FILES ); closefiles( db->mst, MST_FILES ); return 0 > ret ? ret : ret ? -ret : -1; } /* ldb_open */ /* ************************************************************ package data */ /* ************************************************************ package functions */ int *ldb_readRecAtOff ( int dbid, lxref off, int *nxtoff ) { int *rec; LDb *db = getDb( dbid ); if ( ! db ) { log_msg( LOG_ERROR, "\tat ldb_readRecAtOff" ); return 0; } if ( 0 == off ) off = 64; rec = getMfr( db, off, nxtoff ); if ( ! rec ) return 0; LOG_DBG( LOG_VERBOSE, "db %d off %d: got %hd bytes", dbid, off, !rec ? -1 : rec[LMFR_RECL] ); return rec; } /* ldb_readRecAtOff */ int ldb_search ( int dbid, const char *key, LdbPost *post, Rec *rec ) { LDb *db = getDb( dbid ); Key k; if ( ! db ) return -ERR_BADF; if ( post ) { /* prepare for postings */ if ( ! post->len ) post->len = sizeof(post->p)/sizeof(post->p[0]); /* standard length */ if ( LDB_NOT & post->mode ) post->mode |= LDB_AND; } if ( DB_INVOPEN & db->flags ) return search( db, key, post, rec, 0 ); if ( !(DB_LBTOPEN & db->flags) ) return -ERR_BADF; if ( db->oxi.bat ) return -ERR_BUSY; memset( &k, 0, sizeof(k) ); if ( ! key ) { k.byt[0] = '$'; k.len = 1; } else { unsigned char *uk = (unsigned char*)key; int l = strlen( key ); if ( l > 255 ) l = 255; k.len = (unsigned char)l; while ( l-- ) k.byt[l] = db->ctab[LCS_UCASE].c[ uk[l] ]; } return lbt_search( &db->oxi, &k, post, rec ); } /* ldb_search */ int ldb_p2s ( Set *set, LdbPost *post ) { int *s = set->id; int last=0, max = set->len; int i; set->len = 0; if ( ! max ) max = OPENISIS_SETLEN; max--; if ( !post->fil ) return 0L; s[0] = LDBP_ROW(post->p); for ( i=1; ifil && last < max; i++ ) { int row = LDBP_ROW(post->p+i); if ( s[last] != row ) s[++last] = row; } return set->len = last+1; } /* ldb_p2s */ #if 0 LcsTab *ldb_tabs( int dbid ) { LDb *db = getDb( dbid ); return ! db ? 0 : db->ctab; } /* ldb_tabs */ #endif Db *ldb_getdb (int dbid) { LDb *db = getDb (dbid); return db ? &db->head : 0; } /* ************************************************************ public functions */ int dMaxId ( int dbid ) { LDb *db = getDb( dbid ); if ( ! db ) return -ERR_BADF; return db->mfc[LMFC_NMFN] - 1; } /* dMaxId */ Raw *dRaw ( int dbid, int rowid ) { int off; int *rec = 0; LDb *db; if ( LIO_LOCK() ) return 0; db = getDb( dbid ); if ( ! db ) { log_msg( LOG_ERROR, "\tat openIsisReadRaw %d", rowid ); goto done; } off = getOff( db, rowid, 0 ); log_msg( LOG_INFO, "found xref 0x%08x for %d", off, rowid ); if ( 0 >= off ) { log_msg( LOG_INFO, "found deleted xref 0x%08x for %d", off, rowid ); goto done; } rec = getMfr( db, off, 0 ); if ( ! rec ) { log_msg( LOG_WARN, "\tno record at %d rowid %d", off, rowid ); goto done; } LOG_DBG( LOG_VERBOSE, "db %d row %d: got %hd bytes", dbid, rowid, !rec ? -1 : rec[LMFR_RECL] ); if ( rec[LMFR_MFN] != rowid ) { log_msg( LOG_ERROR, "got mfn %d expected %d", rec[LMFR_MFN], rowid ); mFree( rec ); rec = 0; goto done; } done: (void)LIO_RELE(); return (Raw*)rec; } /* dRaw */ Rec *dRead ( int dbid, int rowid ) { LDb *db = getDb( dbid ); Rec *r; if ( DB_TXTOPEN & db->flags ) return dText( db, rowid, 0 ); if ( (r = (Rec *) dRaw( dbid, rowid )) ) { char * base = (char*)r; Field *f = r->field; int i = r->len; for ( ; i--; f++ ) f->val = base + (int)f->val; assert( RECOK( r ) ); } return r; } /* dRead */ int dWritex ( int dbid, Rec *rec, Rec *idx ) { LDb *db = getDb( dbid ); int ret = 0; if ( ! db ) return -ERR_BADF; if ( !(DB_WRITABLE & db->flags) ) return log_msg( ERR_INVAL, "db %d not writable", dbid ); if ( rec && (ret = DB_TXTOPEN & db->flags ? pText( db, rec, 0 ) : putRec( db, rec ) ) ) return ret; if ( idx ) { const unsigned char *const uc = db->ctab[LCS_UCASE].c; int delmode = 0; int tag = -1; int mode = 'f'; /* 'w', 's' */ int occ = 0; int pos = 0; int cut = 30; int mfn = rec ? rec->rowid : 0; Hit h; Key k; Field *f = idx->field, *last = f + idx->len - 1; for ( ; f <= last; f++ ) { const char *val = f->val; int len = f->len; int del = delmode; k.val.len = 0; switch ( f->tag ) { case XCTL: { /* index cmd [opt] */ const char *cmd = val, *e = val + len; int cmdlen, opt = 0, haveopt; while ( val < e && 64 < *val ) /* eat ASCII letters */ val++; cmdlen = val - cmd; if ( val < e && (TAB == *val || ' ' == *val) ) val++; haveopt = val < e && a2il( val, e-val, &opt ); if ( ! cmdlen ) { cut = haveopt ? opt : 30; continue; } switch (*cmd) { case 'f': /* fields */ mode = 'f'; occ = opt; pos = 0; continue; case 'w': /* words */ mode = 'w'; pos = opt; continue; case 's': /* split */ mode = 's'; pos = opt; continue; case 'a': /* add */ delmode = 0; occ = pos = 0; continue; case 'd': /* del */ delmode = 1; occ = pos = 0; continue; case 'm': /* mfn */ mfn = opt; occ = pos = 0; continue; } return log_msg( ERR_INVAL, "bad index control '%.*s'", cmdlen, cmd ); } case XHIT: { int i = 0, v[5], *pv = v; const char *e = val + len; if ( len ) switch (*val) { case '+': del = 0; val++; break; case '-': del = 1; val++; break; } for ( ; val < e && i<5; i++ ) { int dig = a2il( val, e-val, v+i ); val += dig; if ( val >= e || TAB == *val ) break; if ( '.' != *val ) return log_msg( ERR_INVAL, "bad HIT '%.*s' after %d", e-val, val, v[i] ); val++; } h.dbn = 0; h.mfn = mfn; h.pos = pos; h.occ = occ; h.tag = tag; switch ( i ) { case 5: h.dbn = (unsigned short)*pv++; case 4: h.mfn = (unsigned)*pv++; case 3: h.pos = (unsigned short)pv[2]; case 2: h.occ = (unsigned short)pv[1]; case 1: h.tag = (unsigned short)pv[0]; /* case 0: ! f->len */ } if ( val < e && TAB == *val ) val++; len = e - val; } break; /* case XHIT */ case XFST: return log_msg( ERR_IDIOT, "sorry, XFST not implemented" ); #if 0 case XADD: /* binary key */ /* if ( f->len < db->oxi.vsz ) memset( k.val.byt, 0, db->oxi.vsz - f->len ); */ memcpy( k.val.byt + (f->len < (int)db->oxi.vsz ? (int)db->oxi.vsz - f->len : 0), f->val, f->len > (int)db->oxi.vsz ? (int)db->oxi.vsz : f->len ); k.val.len = db->oxi.vsz; break; #endif default: if ( 0 > f->tag ) return log_msg( ERR_INVAL, "bad index control tag %d", f->tag ); switch ( mode ) { /* check for tag change */ case 'f': if ( tag == f->tag ) occ++; else occ = 0; break; case 'w': if ( tag == f->tag ) pos++; else pos = 0; break; } tag = f->tag; h.dbn = 0; h.mfn = mfn; h.pos = pos; h.occ = occ; h.tag = f->tag; } if ( ! k.val.len ) { /* not ADD/DEL: use hit, val */ unsigned char *dst = k.byt; const unsigned char *src = (const unsigned char *)val; if ( cut < len ) len = cut; k.len = (unsigned char)len; while ( len-- ) *dst++ = uc[ *src++ ]; cXMkVal( &db->oxi, &k.val, &h ); LOG_DBG( LOG_DEBUG, "#%d %c key '%.*s' hit %d.%d.%d.%d.%d", f - idx->field, del ? '-' : '+', k.len, k.byt, h.dbn, h.mfn, h.tag, h.occ, h.pos ); } ret = del ? lbt_del( &db->oxi, &k ) : lbt_add( &db->oxi, &k ); } } return ret; } /* dWritex */ int dWrite ( int dbid, Rec *rec ) { /* TODO: use FST lines as idx */ return dWritex( dbid, rec, 0 ); } /* dWrite */ Rec* dTerm ( Rec *rec, int dbid, const char *key ) { return 0 > ldb_search( dbid, key, 0, rec ) ? 0 : rec; } /* dTerm */ int dXLoop ( int dbid, DXLoop *l ) { LDb *db = getDb( dbid ); if ( !db ) return -ERR_BADF; if ( OPENISIS_IDXTRAD & l->flg ) { if ( !(db->flags & DB_INVOPEN) ) return -ERR_BUSY; return search( db, 0, 0, 0, l ); } if ( !(db->flags & DB_LBTOPEN) || db->oxi.bat ) return -ERR_BUSY; return lbt_loop( & db->oxi, l ); } /* dXLoop */ int cInit ( int argc, const char **argv, CLockFunc lockfunc ) { (void)argc; (void)argv; cOpen( 0 ); if ( lockfunc ) lio_lock = lockfunc; return 0; } Db* cDOpen (const char *dbname, Rec *dbpar, Rec *syspar, Fdt *fdt) { int dbid; if ( ! init ) cOpen( 0 ); dbid = ldb_open (dbname, dbpar, syspar, fdt); if (0 <= dbid) { return &dbs[dbid].head; } return 0; } int cDOpenv ( const char *dbname, const char **argv, int argc ) { Rec *dbpar = 0; int rt; if ( ! init ) cOpen( 0 ); if (argc) { dbpar = rSet (0, RARGV | RFDT | RNOC | RIGN | argc, openIsisFdtDbpar, argv); } rt = ldb_open (dbname, dbpar, 0, 0); if (dbpar) { mFree (dbpar); } return rt; } int cDClose ( int dbid ) { LDb *db = getDb( dbid ); if ( ! db ) return -ERR_BADF; if ( LIO_LOCK() ) return -ERR_BUSY; if ( DB_MODIFIED == ((DB_MODIFIED|DB_TXTOPEN) & db->flags) ) { /* write back the MF control */ Mfc mfc; /* if ( 498 < (db->mflen & 511) ) db->mflen = ~511 & (db->mflen + 14); */ mfc.ctlm = rvi( db->mfc[LMFC_CTLM] ); mfc.nmfn = rvi( db->mfc[LMFC_NMFN] ); mfc.nmfb = rvi( 1 + (db->mflen >> 9) ); mfc.nmfp = rvs( 511 & db->mflen ); mfc.type = rvs( db->mfc[LMFC_TYPE] ); mfc.rcnt = rvi( db->mfc[LMFC_RCNT] ); mfc.mfx1 = rvi( db->mfc[LMFC_MFX1] ); mfc.mfx2 = rvi( db->mfc[LMFC_MFX2] ); mfc.mfx3 = rvi( db->mfc[LMFC_MFX3] ); if ( sizeof(mfc) != lio_pwrite( &db->mst[MST_MST], &mfc, sizeof(mfc), 0) ) log_msg( ERR_TRASH, "could not write MST header" ); } if ( db->mmap ) { if ( (DB_MODIFIED|DB_TXTOPEN) == ((DB_MODIFIED|DB_TXTOPEN) & db->flags) ) memcpy( db->mmap, ISIX, 4 ); /* force newer mtime on proper close */ lio_mmap( 0, (void**)&db->mmap, db->mmlen*db->ptrl ); } db->mmlen = 0; closefiles( db->mst, MST_FILES ); if ( DB_INVOPEN & db->flags ) closefiles( db->inv, INV_FILES ); if ( DB_LBTOPEN & db->flags ) lbt_close( &db->oxi ); db->flags = 0L; if ( db->path ) mFree( (char*)db->path ); if (db->head.cfg) mFree (db->head.cfg); if (db->head.fdt) fFree (db->head.fdt); memset( db, 0, sizeof(db) ); (void)LIO_RELE(); return 0; } /* cDClose */ int cDCheck ( int dbid, int flags ) { static char dot = '.'; int *r; LDb *db = getDb( dbid ); int nxtoff = 64, off; if ( ! db ) return -ERR_BADF; (void)flags; do { lio_write( &lio_out, &dot, 1 ); if ( (r = ldb_readRecAtOff(dbid,off=nxtoff,&nxtoff)) ) { int o = getOff( db, r[LMFR_MFN], 0 ); if ( o != off ) { log_msg( LOG_WARN, "mfn %d xrf %d != real %d\n", r[LMFR_MFN], o, off ); } mFree( r ); } } while ( 0 < nxtoff ); return 0; } /* cDCheck */ OpenIsisIdx *cXOpen ( int dbid, int mode ) { LDb *db = getDb( dbid ); if ( !db || !(db->flags & DB_LBTOPEN) || !(db->oxi.flg & LBT_WRITE) /* may be writable if db is not */ || db->oxi.bat /* preliminary undocumented feature: mode -1 gives direct access in non-batch mode */ || (0 <= mode && lbt_batch( & db->oxi, (unsigned char)mode )) ) return 0; return & db->oxi; } /* cXOpen */ ldb.h0100644000113000011300000001277007644412744010470 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LDB_H /* $Id: ldb.h,v 1.19 2003/04/08 00:20:52 kripke Exp $ package interface of the general db access functions. */ #include "luti.h" #include "lll.h" /* #define LDB_MAGIC 0x0FE91515 read OPENISIS */ enum { /* MFR members */ LMFR_MFN = 1, /* the rowid */ LMFR_RECL, /* total external length (even), negative for locked record */ LMFR_BWB, /* block of rec's previous version */ LMFR_BWP, /* pos of " */ LMFR_BASE, /* offset of contents area */ LMFR_NVF, /* number of fields */ LMFR_STAT, /* state, if != 0, about to be deleted */ LMFR__FL, /* offset of repeated part */ LMFR_TAG = 0, LMFR_POS, LMFR_LEN, LMFR__RL /* length of repeated part */ }; /** read a raw isis record by rowid. the memory must be freed after usage. */ extern int *ldb_readRec ( int db, int rowid ); /** numerical type for a records file position. To support large DBs, this may be lll. */ typedef int lxref; /** read a raw isis record by xref. the memory must be freed after usage. */ extern int *ldb_readRecAtOff ( int dbid, lxref off, int *nxtoff ); /* ** ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ** detailed search structures and functions ** ** ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */ enum { LDB_MAX_KEYLEN = 31, /* actually, it's 30 for isis-1 */ LDB_TERMBUF = 128, /* ints */ LDB_INDEXES = 2 }; /** the isis-1 posting is an 64bit number. In the file it's in big endian order, to allow ordering using memcmp. In the LdbP we have native order to use native 64bit. Until we need more bits for postings, we may use the 64bit type provided by gcc and some other compilers. the 8 bytes are: highest 3 for the rowid, then 2 tag, then 3 pos. the 3 pos bytes in turn are 1 byte for the occurence and 2 bytes for the word count. (e.g. value 2<<16 | 7 if term was found in the 7th word of the 2nd occurence of field 'tag'). we usually don't care, since for a near condition we always want a small distance w/in the same occurence which may be conveniently checked by the difference of 3 byte pos values. BUT NOTE: ... unfortunately we DO NEED one more bit for marking postings during AND operations. we abuse the highest bit of the word counter, since it's for sure useless for any reasonable distance check. *** BE CAREFUL WHEN COMPARING POSTINGS DURING "AND" OP *** use the LDBP_xxM macros below. */ typedef union { char bytes[8]; lll val; } LdbP; /* all macros operate on pointers */ /* access parts */ #define LDBP_ROW( p ) (0xffffffL & (int)((p)->val >> 40)) #define LDBP_TAG( p ) ( 0xffffL & (int)((p)->val >> 24)) #define LDBP_POS( p ) (0xff7fffL & (int)((p)->val)) #define LDBP_MARK( p ) ( 0x8000 & (int)((p)->val)) #define LDBP_SETMARK( p ) ((p)->val |= LLL( 0x8000)) #define LDBP_CLRMARK( p ) ((p)->val &= LLL(0xffffffffffff7fff)) /* value w/o the mark bit */ #define LDBP_IGNMARK( p ) ((p)->val & LLL(0xffffffffffff7fff)) /* compare according to rowid, tag, pos */ #define LDBP_EQ( a, b ) ((a)->val == (b)->val) #define LDBP_GT( a, b ) ((a)->val > (b)->val) /* compare IGNORING MARK according to rowid, tag, pos */ #define LDBP_EQM( a, b ) (LDBP_IGNMARK(a) == LDBP_IGNMARK(b)) #define LDBP_GTM( a, b ) (LDBP_IGNMARK(a) > LDBP_IGNMARK(b)) /* manipulation: set bottom / top for row of given posting */ #define LDBP_SETROWBOT( d, s ) ((d)->val = (s)->val & LLL(0xffffff0000000000)) #define LDBP_SETROWTOP( d, s ) ((d)->val = (s)->val | LLL( 0xffffffffff)) /** postings come in arrays with some header. The standard structure gives a 8k buffer, but member len may give other actual length. */ typedef struct { short mode; /* in: merge flags */ short near; /* in: near distance; in OR mode: collect pos info */ int tag; /* in: tag, to which postings are restricted */ int skp; /* in: ignore mfns < skp */ int len; /* in: length (# of postings) of buffer (if 0 : default length) */ int fil; /* io: number of postings actually used */ int cut; /* io: min mfn ignored due to buffer length */ LdbP p[OPENISIS_SETLEN]; /** 8 bytes as in IFP file: mfn[3],tag[2],occ[1],cnt[2] highest bit of cnt (1LL<<15) is used as mark */ } LdbPost; enum { LDB_OR, LDB_AND, LDB_NOT, /* like AND, but keep unmarked postings */ LDB_PFX = 4, /* prefix match */ LDB_KEEPMARKS = 8 /* do not compact after AND/NOT */ }; enum { LDB_NEAR_F = 0x7fff, /* the (F): same occurence of field */ LDB_NEAR_G = -0x8000 /* the (G): same field */ }; /** */ extern int ldb_search ( int db, const char *key, LdbPost *post, OpenIsisRec *rec ); extern int ldb_p2s ( OpenIsisSet *set, LdbPost *post ); extern Db* ldb_getdb (int dbid); #define LDB_H #endif /* LDB_H */ ldsp.c0100644000113000011300000003311007662161240010643 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: ldsp.c,v 1.31 2003/05/16 12:32:06 mawag Exp $ server side request processing. */ #include #include /* PATH_MAX */ #include #include /* vsnprintf */ #include #include "openisis.h" #include "loi.h" #include "ldb.h" #include "ldsp.h" #include "lfdt.h" /* #include "lsv.h" */ #include "luti.h" #ifdef WIN32 #define vsnprintf _vsnprintf #endif #define ISID 0 #define ISER 1 #define ITMS 2 #define SSTS 3 #define RSPVAR \ Db *dbh = 0; \ Rec *rsp = 0; \ int sdbid; #define RSPADDR \ if (! rspa) { rspa = &rsp; } #define RSPHEAD \ *rspa = rAddI (*rspa, OPENISIS_COM_SID, sst[ISID], !0); \ *rspa = rAddI (*rspa, OPENISIS_COM_SER, sst[ISER], !0); #define RSPNOERR \ *rspa = rAddI (*rspa, OPENISIS_RSP_ERR, 0, !0); #define RSPDB \ RADDS (*rspa, OPENISIS_COM_DBN, dbn, !0); #define RSPFREE mFree (rsp); #define GETDB(rname) \ sdbid = ldspGetDb (&dbh, sst, dbn); \ if (0 > sdbid) { \ int rt = _RqsError (sst, ERR_INVAL, rspa, cb, \ cld, rname, sst[ITMS] ? \ "db %s remounted" : "db %s not open", dbn); \ return rt; \ } #define RSPFDT \ if (dbh->fdt) { \ *rspa = openIsisFFdt2Rec (dbh->fdt, *rspa, 0); \ } \ if (dbh->cfg) { \ *rspa = luti_wrap (*rspa, dbh->cfg, OPENISIS_COM_CFG); \ } #define DELIVER \ if (*rspa) { \ int rt = 0; \ if (cb) { \ rt = (*cb) (rspa, cld); \ mFree (*rspa); \ } \ return rt; \ } \ return sMsg (ERR_NOMEM, "ldsp: deliver: rsp == 0"); #define UTF8REC(rec,inv,rname) \ if (openIsisEnc2Utf8 && dbh->cfg) { \ char ebuf[128]; \ char *enc = rString (dbh->cfg, OPENISIS_DENC, 0, ebuf, sizeof(ebuf)); \ if (enc && *enc && strcmp (enc, "utf-8")) { \ ldspUtf8Rec (&rec, enc, inv); \ } \ } /* ************************************************************ */ static int ldspGetDb (Db **db, int sst[], const char *dbn) { Db *dbh = nDbByName (stub0, dbn); if (! dbh) { sst[ITMS] = 0; return -1; } if (sst[ITMS]) { if (-1 == sst[ITMS] || /* dont want default of rInt call */ (dbh->tms && dbh->tms != sst[ITMS]) ) { sst[ITMS] = dbh->tms; return -1; } } if (db) { *db = dbh; } return dbh->dbid; } static int ldspUtf8Rec (Rec **rec, const char *enc, int invert) { char buf[4096]; const char *dst; Rec *res = 0; Field *F, *E; if (*rec && (*rec)->len) { E = (*rec)->field + (*rec)->len; for (F = (*rec)->field; E > F; ++F) { if (0 == F->len || (1 == F->len && ! *(F->val))) { RADDS (res, F->tag, "", !0); } else { dst = (*openIsisEnc2Utf8) (enc, F->val, F->len, buf, sizeof(buf), invert); if (! dst) { if (res) { mFree (res); } return sMsg (ERR_TRASH, "ldsp: cannot convert %s %s utf8: %.*s", enc, (invert ? "from" : "to"), F->len, F->val); } RADDS (res, F->tag, dst, !0); if (dst != buf && dst != F->val) { (*openIsisEnc2Utf8) (0, dst, 0, 0, 0, 0); } } if (! res) { return sMsg (ERR_NOMEM, "ldsp: convert %s %s utf8: res == 0", enc, (invert ? "from" : "to")); } } res->rowid = (*rec)->rowid; res->dbid = (*rec)->dbid; mFree (*rec); *rec = res; } return 0; } static int _RqsError ( int sst[], int err, Rec **rspa, LdspDlvCb *cb, void *cld, const char *rname, const char *fmt, ... ) { Rec *rsp = 0; char msg[OPENISIS_ERRMSGLEN]; va_list ap; RSPADDR RSPHEAD sMsg (LOG_VERBOSE, "ldsp: _RqsError %s %x from %d/%d", rname, err, sst[ISID], sst[ISER]); if (sst[ITMS] && -1 != sst[ITMS]) { *rspa = rAddI (*rspa, OPENISIS_COM_TMS, sst[ITMS], !0); } *rspa = rAddI (*rspa, OPENISIS_RSP_ERR, err, !0); if (0 != fmt && 0 != *fmt) { va_start (ap, fmt); vsnprintf (msg, OPENISIS_ERRMSGLEN - 1, fmt, ap); msg[OPENISIS_ERRMSGLEN - 1] = 0; va_end (ap); RADDS (*rspa, OPENISIS_RSP_MSG, msg, !0); sMsg (LOG_WARN, "ldsp: err on rqs %s: %d(%x) - %s", rname, err, err, msg); } else { sMsg (LOG_WARN, "ldsp: err on rqs %s: %d(%x)", rname, err, err); } DELIVER } static int _RqsLsDb ( int sst[], Rec **rspa, LdspDlvCb *cb, void *cld ) { Rec *rsp = 0; Schema *sc; char *n; int j; sc = nSchema (stub0); if (! sc) { j = _RqsError (sst, ERR_BADF, rspa, cb, cld, "lsdb", "local schema not initialized"); return j; } RSPADDR RSPHEAD RSPNOERR sMsg (LOG_VERBOSE, "ldsp: _RqsLsDb from %d/%d", sst[ISID], sst[ISER]); for (j = 0; sc->ndbs > j; ++j) { if (*(n = sc->dbs[j]->name)) { RADDS (*rspa, OPENISIS_COM_DBN, n, !0); } } DELIVER } static int _RqsOpen ( int sst[], const char *dbn, Rec *dbpar, Fdt *fdt, Rec **rspa, LdspDlvCb *cb, void *cld ) { #define STRERRNOLEN 128 RSPVAR Schema *sc; char msg[STRERRNOLEN]; sst[ITMS] = 0; sc = nSchema (stub0); if (! sc) { int rt = _RqsError (sst, ERR_BADF, rspa, cb, cld, "open", "%s: local schema not initialized", dbn); mFree (dbpar); return rt; } sMsg (LOG_VERBOSE, "ldsp: _RqsOpen %s from %d/%d", dbn, sst[ISID], sst[ISER]); errno = 0; dbh = cDOpen (dbn, dbpar, sc->cfg, fdt); if (! dbh) { int rt; strncpy (msg, strerror (errno), STRERRNOLEN - 1) [STRERRNOLEN - 1] = 0; rt = _RqsError (sst, ERR_BADF, rspa, cb, cld, "open", "%s: %s", dbn, msg); mFree (dbpar); return rt; } sdbid = dbh->dbid; dbn = dbh->name; RSPADDR RSPHEAD RSPNOERR RSPDB *rspa = rAddI (*rspa, OPENISIS_RSP_DBID, sdbid, !0); *rspa = rAddI (*rspa, OPENISIS_COM_TMS, dbh->tms, !0); mFree (dbpar); DELIVER } static int _RqsClose ( int sst[], const char *dbn, Rec **rspa, LdspDlvCb *cb, void *cld ) { RSPVAR int err; GETDB("close") sMsg (LOG_VERBOSE, "ldsp: _RqsClose %s from %d/%d", dbn, sst[ISID], sst[ISER]); err = cDClose (sdbid); RSPADDR RSPHEAD *rspa = rAddI (*rspa, OPENISIS_RSP_DBID, sdbid, !0); *rspa = rAddI (*rspa, OPENISIS_RSP_ERR, err, !0); RSPDB DELIVER } static int _RqsMount ( int sst[], const char *dbn, Rec **rspa, LdspDlvCb *cb, void *cld ) { RSPVAR GETDB("mount") RSPADDR RSPHEAD sMsg (LOG_VERBOSE, "ldsp: _RqsMount %s from %d/%d", dbn, sst[ISID], sst[ISER]); RSPNOERR RSPDB RSPFDT DELIVER } static int _RqsMaxrow ( int sst[], const char *dbn, int flags, Rec **rspa, LdspDlvCb *cb, void *cld ) { RSPVAR int maxrow; GETDB("maxrow") maxrow = dMaxId (sdbid); if (0 > maxrow) { int rt = _RqsError (sst, ERR_TRASH, rspa, cb, cld, "maxrow", "no maxrow for %s", dbn); RSPFREE return rt; } RSPADDR RSPHEAD sMsg (LOG_VERBOSE, "ldsp: _RqsMaxrow %s from %d/%d", dbn, sst[ISID], sst[ISER]); RSPNOERR RSPDB if (OPENISIS_RQSF_MNT & flags) { RSPFDT } *rspa = rAddI (*rspa, OPENISIS_COM_ROW, maxrow, !0); DELIVER } static int _RqsQuery ( int sst[], const char *dbn, const char *key, int mode, int skip, int size, int flags, Rec **rspa, LdspDlvCb *cb, void *cld ) { RSPVAR Set set; Rec *rec; int lrt; int j; int rdr = OPENISIS_RQSF_QRR & flags; GETDB("query") sMsg (LOG_VERBOSE, "ldsp: _RqsQuery %s from %d/%d", dbn, sst[ISID], sst[ISER]); set.len = 0; /* input */ lrt = dQuery (&set, sdbid, key, mode, skip); if (0 > lrt) { int rt = _RqsError (sst, ERR_TRASH, rspa, cb, cld, "query", "%d,%s = %d", sdbid, key, lrt); return rt; } RSPADDR RSPHEAD RSPNOERR RSPDB if (OPENISIS_RQSF_MNT & flags) { RSPFDT } *rspa = rAddI (*rspa, OPENISIS_RSP_NUMT, lrt, !0); if (0 < size && lrt > size) { lrt = size; } *rspa = rAddI (*rspa, OPENISIS_RSP_NUMR, lrt, !0); for (j = 0; lrt > j; ++j) { *rspa = rAddI (*rspa, OPENISIS_COM_ROW, set.id[j], !0); } if (rdr) { for (j = 0; *rspa && lrt > j; ++j) { rec = dRead (sdbid, set.id[j]); if (rec) { UTF8REC (rec, 0, "query") } else { sMsg (ERR_TRASH | LOG_WARN, "db %s: non-existent mfn %d for key %s", dbn, set.id[j], key); rec = rDup (0, 0, 0); } *rspa = luti_wrap (*rspa, rec, OPENISIS_COM_REC); mFree (rec); } } DELIVER } static int _RqsRead ( int sst[], const char *dbn, int row, int flags, Rec **rspa, LdspDlvCb *cb, void *cld ) { RSPVAR Rec *rec; GETDB("read") RSPADDR RSPHEAD sMsg (LOG_VERBOSE, "ldsp: _RqsRead %s %d from %d/%d", dbn, row, sst[ISID], sst[ISER]); RSPNOERR RSPDB if (OPENISIS_RQSF_MNT & flags) { RSPFDT } rec = dRead (sdbid, row); if (rec) { *rspa = rAddI (*rspa, OPENISIS_RSP_NUMT, 1, !0); *rspa = rAddI (*rspa, OPENISIS_RSP_NUMR, 1, !0); *rspa = rAddI (*rspa, OPENISIS_COM_ROW, row, !0); UTF8REC (rec, 0, "read") *rspa = luti_wrap (*rspa, rec, OPENISIS_COM_REC); mFree (rec); } else { *rspa = rAddI (*rspa, OPENISIS_RSP_NUMT, 0, !0); *rspa = rAddI (*rspa, OPENISIS_RSP_NUMR, 0, !0); } DELIVER } static int _RqsUpdate ( int sst[], const char *dbn, int row, Rec *rec, Rec *idx, int flags, Rec **rspa, LdspDlvCb *cb, void *cld ) { RSPVAR int rtw; GETDB("update") sMsg (LOG_VERBOSE, "ldsp: _RqsUpdate %s %d from %d/%d", dbn, row, sst[ISID], sst[ISER]); if (rec) { rec->rowid = row; rec->dbid = sdbid; UTF8REC (rec, !0, "update") } rtw = dWritex (sdbid, rec, idx); if (rtw) { rtw = _RqsError (sst, ERR_TRASH, rspa, cb, cld, "update", "write error %d", rtw); mFree (rec); mFree (idx); return rtw; } RSPADDR RSPHEAD RSPNOERR RSPDB if (OPENISIS_RQSF_MNT & flags) { RSPFDT } if (rec) { row = rec->rowid; mFree (rec); } mFree (idx); *rspa = rAddI (*rspa, OPENISIS_COM_ROW, row, !0); DELIVER } static int _RqsDelete ( int sst[], const char *dbn, int row, int flags, Rec **rspa, LdspDlvCb *cb, void *cld ) { (void)flags; sMsg (LOG_VERBOSE, "ldsp: _RqsDelete %s %d from %d/%d", dbn, row, sst[ISID], sst[ISER]); return _RqsError (sst, ERR_TRASH, rspa, cb, cld, "delete", "request not implemented yet"); } /* ************************************************************ package functions */ int ldspProcess (Rec *rqs, Rec **rspa, LdspDlvCb *cb, void *cld) { char buf[PATH_MAX]; char *dbn = 0; int sst[SSTS]; int rtyp; int flags = 0; if (0 == rqs) { return sMsg (ERR_IDIOT, "ldspProcess: no request"); } if (0 == rspa && 0 == cb) { return sMsg (ERR_IDIOT, "ldspProcess: no callback"); } rtyp = rInt (rqs, OPENISIS_RQS_TYPE, -1, 0); sst[ISID] = rInt (rqs, OPENISIS_COM_SID, -1, 0); sst[ISER] = rInt (rqs, OPENISIS_COM_SER, -1, 0); sst[ITMS] = rInt (rqs, OPENISIS_COM_TMS, -1, 0); if (OPENISIS_RQST_LSDB != rtyp) { dbn = rString (rqs, OPENISIS_COM_DBN, 0, buf, sizeof(buf)); if (!dbn) { i2a (buf, rtyp); return _RqsError (sst, ERR_INVAL, rspa, cb, cld, buf, "no db name given"); } flags = rInt (rqs, OPENISIS_RQS_FLG, 0, 0); } switch (rtyp) { case OPENISIS_RQST_LSDB: return _RqsLsDb (sst, rspa, cb, cld); case OPENISIS_RQST_OPEN: { Rec *dbpar = luti_unwrap (rqs, 0, OPENISIS_COM_CFG, -1); Fdt *fdt = fRec2Fdt (rqs); return _RqsOpen (sst, dbn, dbpar, fdt, rspa, cb, cld); } case OPENISIS_RQST_CLOS: return _RqsClose (sst, dbn, rspa, cb, cld); case OPENISIS_RQST_MNT: return _RqsMount (sst, dbn, rspa, cb, cld); case OPENISIS_RQST_MROW: return _RqsMaxrow (sst, dbn, flags, rspa, cb, cld); case OPENISIS_RQST_QRY: { char kybuf[OPENISIS_QRY_KEYLEN]; char *key; int mode, skip, size; key = rString (rqs, OPENISIS_RQS_KEY, 0, kybuf, OPENISIS_QRY_KEYLEN); mode = rInt (rqs, OPENISIS_RQS_QMOD, -1, 0); skip = rInt (rqs, OPENISIS_RQS_SKIP, 0, 0); size = rInt (rqs, OPENISIS_RQS_SIZE, -1, 0); if (0 > mode || 0 == key) { return _RqsError (sst, ERR_INVAL, rspa, cb, cld, "query", "missing mode or key"); } return _RqsQuery (sst, dbn, key, mode, skip, size, flags, rspa, cb, cld); } case OPENISIS_RQST_READ: { int rowid; rowid = rInt (rqs, OPENISIS_COM_ROW, -1, 0); if (0 > rowid) { return _RqsError (sst, ERR_INVAL, rspa, cb, cld, "read", "missing rowid"); } return _RqsRead (sst, dbn, rowid, flags, rspa, cb, cld); } case OPENISIS_RQST_INS: case OPENISIS_RQST_UPD: { Rec *rec, *idx; int rowid = 0; int ins = rtyp == OPENISIS_RQST_INS; rec = luti_unwrap (rqs, 0, OPENISIS_COM_REC, -1); idx = luti_unwrap (rqs, 0, OPENISIS_RQS_IDX, -1); if (!(rec || idx)) { return _RqsError (sst, ERR_INVAL, rspa, cb, cld, "update", "missing record or index"); } rowid = rInt (rqs, OPENISIS_COM_ROW, 0, 0); if (! ins && 0 >= rowid) { return _RqsError (sst, ERR_INVAL, rspa, cb, cld, "update", "missing rowid"); } return _RqsUpdate (sst, dbn, rowid, rec, idx, flags, rspa, cb, cld); } case OPENISIS_RQST_DEL: { int rowid; rowid = rInt (rqs, OPENISIS_COM_ROW, -1, 0); if (0 > rowid) { return _RqsError (sst, ERR_INVAL, rspa, cb, cld, "delete", "missing rowid"); } return _RqsDelete (sst, dbn, rowid, flags, rspa, cb, cld); } } /* switch */ i2a (buf, rtyp); return _RqsError (sst, ERR_INVAL, rspa, cb, cld, buf, "unrecognized request type"); } /* ldspProcess */ ldsp.h0100644000113000011300000000414407662161240010655 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LDSP_H #define LDSP_H /* $Id: ldsp.h,v 1.3 2003/05/16 12:32:07 mawag Exp $ openisis internal request dispatcher interface. */ /** server side method for transformation of record fields to utf-8 encoding. @param enc 0-terminated encoding of src; if 0, free src after usage @param src source string @param slen length of src @param dstbuf optional output buffer @param dlen length of dstbuf @param invert if != 0, convert src from utf-8 to enc @return utf-8 representation of src, 0 on error, returned pointer must be freed if it not equals dstbuf or src */ typedef const char* OpenIsisEnc2Utf8Func (const char *enc, const char *src, int slen, char *dstbuf, int dlen, int invert); /** pointer to transformation function provided by external environment. if environment defines no or null transformation function, no transformation will happen. */ extern OpenIsisEnc2Utf8Func *openIsisEnc2Utf8; /** callback for delivering response record. */ typedef int LdspDlvCb (OpenIsisRec **rsp, void *cld); /** server side request processing. * @param rsp optional address where to store response */ extern int ldspProcess ( OpenIsisRec *rqs, OpenIsisRec **rsp, LdspDlvCb *cb, void *cld); #endif /* LDSP_H */ lenc.c0100664000113000011300000000210407653462065010634 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lenc.c,v 1.1 2003/04/29 11:38:29 mawag Exp $ dummy encoding function. */ #include "openisis.h" #include "ldsp.h" OpenIsisEnc2Utf8Func *openIsisEnc2Utf8 = 0; lfdt.c0100644000113000011300000006043107662161240010640 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lfdt.c,v 1.23 2003/05/15 16:33:33 mawag Exp $ implementation of FDT calls and static FDTs. */ #include /* CHAR_MAX,PATH_MAX */ #include #include #include "openisis.h" #include "loi.h" #include "lfdt.h" #include "luti.h" #define FDF_EXT ".fdt" #define FDF_NAMLEN 30 #define FDF_PATLEN 20 /* format of *.fdt */ #define FDF_TOOLEN 6 #if FDF_NAMLEN >= FD_NAMELEN fix me #endif /* ************************************************************ private types */ /* ************************************************************ private data */ #define _FD_DFLTDB \ { OPENISIS_SC_DFLTDB, 0, FTX, 0, 0, DBNLEN, "defaultdb", \ "Name of default db", 0, 0, 0, 0 } #define _FD_DPATH \ { OPENISIS_DPATH, 0, FTX, 0, 0, PATH_MAX, "dbpath", \ "Database Path", 0, 0, 0, 0 } #define _FD_DENC \ { OPENISIS_DENC, 0, FTX, 0, 0, 32, "encoding", \ "Encoding", 0, 0, 0, 0 } static Fd _fdsys[] = { /* 700 ... OpenIsis system parameters */ { OPENISIS_SPATH, 0, FTX, 0, 0, PATH_MAX, "syspath", "Global Database Path", 0, 0, 0, 0 }, { OPENISIS_SLOGF, 0, FTX, 0, 0, PATH_MAX, "logfile", "Logfile Name", 0, 0, 0, 0 }, { OPENISIS_SLOGV, 0, FTE, 0, 0, 16, "v", "Verbosity of Logging", 0, 0, 0, 0 }, { OPENISIS_SLOGV, 0, FTV, 0, 0, 0, "off", "don't log anything", 0, 0, 0, 0 }, { OPENISIS_SLOGV, 0, FTV, 0, 0, 1, "fatal", "", 0, 0, 0, 0 }, { OPENISIS_SLOGV, 0, FTV, 0, 0, 3, "syserr", "", 0, 0, 0, 0 }, { OPENISIS_SLOGV, 0, FTV, 0, 0, 4, "error", "", 0, 0, 0, 0 }, { OPENISIS_SLOGV, 0, FTV, 0, 0, 5, "warn", "", 0, 0, 0, 0 }, { OPENISIS_SLOGV, 0, FTV, 0, 0, 6, "info", "", 0, 0, 0, 0 }, { OPENISIS_SLOGV, 0, FTV, 0, 0, 7, "verbose", "", 0, 0, 0, 0 }, { OPENISIS_SLOGV, 0, FTV, 0, 0, 8, "trace", "", 0, 0, 0, 0 }, { OPENISIS_SLOGV, 0, FTV, 0, 0, 9, "debug", "", 0, 0, 0, 0 }, { OPENISIS_SLOGV, 0, FTV, 0, 0, 10, "all", "", 0, 0, 0, 0 }, _FD_DFLTDB, _FD_DPATH, _FD_DENC }; static Fdt _fdtsys = { sizeof (_fdsys) / sizeof (_fdsys[0]), _fdsys, 0 }; static Fd _fdsch[] = { { OPENISIS_SC_NAME, 0, FTX, 0, 0, SCNLEN, "name", "Identification of remote scheme", 0, 0, 0, 0 }, { OPENISIS_SC_HOST, 0, FTX, 0, 0, 64, "host", "Hostname of remote scheme", 0, 0, 0, 0 }, { OPENISIS_SC_PORT, 0, FTN, 0, 0, 5, "port", "Port of remote scheme", 0, 0, 0, 0 }, _FD_DFLTDB, _FD_DPATH, _FD_DENC }; static Fdt _fdtsch = { sizeof (_fdsch) / sizeof (_fdsch[0]), _fdsch, 0 }; static Fd _fddb[] = { { OPENISIS_DNAME, 0, FTX, 0, 0, DBNLEN, "db", "Identification of database", 0, 0, 0, 0 }, { OPENISIS_DTYPE, 0, FTE, 0, 0, 256, "format", "Database Format", 0, 0, 0, 0 }, { OPENISIS_DTYPE, 0, FTV, 0, 0, 0, "autoformat", "Database Format", 0, 0, 0, 0 }, { OPENISIS_DTYPE, 0, FTV, 0, 0, 1, "naligned", "Database Format", 0, 0, 0, 0 }, { OPENISIS_DTYPE, 0, FTV, 0, 0, 2, "aligned", "Database Format", 0, 0, 0, 0 }, { OPENISIS_DRO , 0, FTB, 0, 0, 1, "ro", "Readonly Flag", 0, 0, 0, 0 }, _FD_DPATH, { OPENISIS_DDUMP, 0, FTB, 0, 0, 1, "internaldump", "Internal Dump Flag", 0, 0, 0, 0 }, _FD_DENC, { OPENISIS_DFDT, 0, FTX, 0, 0, 256, "fdt", "Path to fdt", 0, 0, 0, 0 } }; static Fdt _fdtdb = { sizeof (_fddb) / sizeof (_fddb[0]), _fddb, 0 }; static Fd _fdfd[] = { { OPENISIS_FDID, 0, FTN, 0, 0, 10, "tag", "Tag number of field", 0, 0, 0, 0 }, { OPENISIS_FDSUB, 0, FTX, 0, 0, 1, "subfield", "Subfield", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTE, 0, 0, 2, "type", "Field type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTX, "alphanum", "Field type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTA, "alpha", "Field type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTN, "numeric", "Field type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTP, "pattern", "Field type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTI, "iso", "Field type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTE, "enum", "Field type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTB, "boolean", "Field type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTT, "table", "Field type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTS, "structure", "Field type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTF | FTX, "subalphanum", "Subfield type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTF | FTA, "subalpha", "Subfield type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTF | FTN, "subnumeric", "Subfield type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTF | FTP, "subpattern", "Subfield type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTF | FTI, "subiso", "Subfield type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTF | FTE, "subenum", "Subfield type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTF | FTB, "subbool", "Subfield type", 0, 0, 0, 0 }, { OPENISIS_FDTYPE, 0, FTV, 0, 0, FTV, "enum value", "Enumeration value", 0, 0, 0, 0 }, { OPENISIS_FDREP, 0, FTB, 0, 0, 1, "repeatable", "Repeatable flag", 0, 0, 0, 0 }, { OPENISIS_FDNUMC, 0, FTN, 0, 0, 2, "numchilds", "Number of subfield childs", 0, 0, 0, 0 }, { OPENISIS_FDLEN, 0, FTN, 0, 0, 10, "length", "Field length or enum value", 0, 0, 0, 0 }, { OPENISIS_FDNAME, 0, FTX, 0, 0, 30, "name", "Field name", 0, 0, 0, 0 }, { OPENISIS_FDDESC, 0, FTX, 0, 0, 31, "description", "Description", 0, 0, 0, 0 }, { OPENISIS_FDPAT, 0, FTX, 0, 0, 128, "pattern", "Pattern", 0, 0, 0, 0 }, { OPENISIS_FDINFO, 0, FTS, 0, 0, 1, "info", "Embedded info record", 0, 0, 0, 0 }, { OPENISIS_FDCHLD, 0, FTX, 1, 0, 1, "children", "Subfield childs", 0, 0, 0, 0 } }; static Fdt _fdtfd = { sizeof (_fdfd) / sizeof (_fdfd[0]), _fdfd, 0 }; #define _FD_FDT \ { OPENISIS_FDT_LEN, 0, FTN, 0, 0, 3, "flen", \ "Length of fdt", 0, 0, 0, 0 }, \ { OPENISIS_FDT_FD, 0, FTS, 1, 0, 1, "fd", \ "Field description", 0, 0, 0, 0 }, \ { OPENISIS_FDT_REC, 0, FTS, 0, 0, 1, "frec", \ "Embedded info record", 0, 0, 0, 0 } static Fd _fdfdt[] = { _FD_FDT }; static Fdt _fdtfdt = { sizeof (_fdfdt) / sizeof (_fdfdt[0]), _fdfdt, 0 }; #define _FD_COM \ { OPENISIS_COM_SID, 0, FTN, 0, 0, 2, "sid", \ "Client Session Id", 0, 0, 0, 0 }, \ { OPENISIS_COM_SER, 0, FTN, 0, 0, 5, "ser", \ "Request Serial No.", 0, 0, 0, 0 }, \ { OPENISIS_COM_DBN, 0, FTX, 0, 0, DBNLEN, "db", \ "DB Identification", 0, 0, 0, 0 }, \ { OPENISIS_COM_TMS, 0, FTN, 0, 0, 10, "tms", \ "Server Db Timestamp", 0, 0, 0, 0 }, \ { OPENISIS_COM_ROW, 0, FTN, 0, 0, 10, "rowid", \ "RowId", 0, 0, 0, 0 }, \ _FD_FDT, \ { OPENISIS_COM_CFG, 0, FTS, 0, 0, 0, "config", \ "Config", 0, 0, 0, 0 }, \ { OPENISIS_COM_REC, 0, FTS, 1, 0, 0, "rec", \ "Data", 0, 0, 0, 0 } static Fd _fdrqs[] = { _FD_COM, { OPENISIS_RQS_TYPE, 0, FTE, 0, 0, 32, "type", "Request type", 0, 0, 0, 0 }, { OPENISIS_RQS_TYPE, 0, FTV, 0, 0, OPENISIS_RQST_OPEN, "open", "open db", 0, 0, 0, 0 }, { OPENISIS_RQS_TYPE, 0, FTV, 0, 0, OPENISIS_RQST_CLOS, "close", "close db", 0, 0, 0, 0 }, { OPENISIS_RQS_TYPE, 0, FTV, 0, 0, OPENISIS_RQST_MNT, "mount", "mount db", 0, 0, 0, 0 }, { OPENISIS_RQS_TYPE, 0, FTV, 0, 0, OPENISIS_RQST_LSDB, "ls", "list dbs", 0, 0, 0, 0 }, { OPENISIS_RQS_TYPE, 0, FTV, 0, 0, OPENISIS_RQST_MROW, "maxrow", "get maxrowid", 0, 0, 0, 0 }, { OPENISIS_RQS_TYPE, 0, FTV, 0, 0, OPENISIS_RQST_QRY, "query", "exec query", 0, 0, 0, 0 }, { OPENISIS_RQS_TYPE, 0, FTV, 0, 0, OPENISIS_RQST_READ, "read", "fetch row", 0, 0, 0, 0 }, { OPENISIS_RQS_TYPE, 0, FTV, 0, 0, OPENISIS_RQST_INS, "insert", "insert rec", 0, 0, 0, 0 }, { OPENISIS_RQS_TYPE, 0, FTV, 0, 0, OPENISIS_RQST_UPD, "update", "update rec", 0, 0, 0, 0 }, { OPENISIS_RQS_TYPE, 0, FTV, 0, 0, OPENISIS_RQST_DEL, "delete", "delete row", 0, 0, 0, 0 }, { OPENISIS_RQS_FLG, 0, FTN, 0, 0, 4, "flags", "Request flags", 0, 0, 0, 0 }, { OPENISIS_RQS_QMOD, 0, FTN, 0, 0, 5, "mode", "Query Mode", 0, 0, 0, 0 }, { OPENISIS_RQS_SKIP, 0, FTN, 0, 0, 5, "skip", "Query Skip", 0, 0, 0, 0 }, { OPENISIS_RQS_SIZE, 0, FTN, 0, 0, 5, "size", "Result Length", 0, 0, 0, 0 }, { OPENISIS_RQS_KEY, 0, FTX, 0, 0, OPENISIS_QRY_KEYLEN, "key", "Query Key", 0, 0, 0, 0 }, { OPENISIS_RQS_IDX, 0, FTS, 0, 0, 0, "idx", "Index to Update", 0, 0, 0, 0 } }; static Fdt _fdtrqs = { sizeof (_fdrqs) / sizeof (_fdrqs[0]), _fdrqs, 0 }; static Fd _fdrsp[] = { _FD_COM, { OPENISIS_RSP_DBID, 0, FTN, 0, 0, 2, "dbid", "Id of local db", 0, 0, 0, 0 }, { OPENISIS_RSP_ERR, 0, FTN, 0, 0, 5, "error", "Error Code", 0, 0, 0, 0 }, { OPENISIS_RSP_MSG, 0, FTX, 0, 0, OPENISIS_ERRMSGLEN, "msg", "Error Message", 0, 0, 0, 0 }, { OPENISIS_RSP_NUMT, 0, FTN, 0, 0, 5, "total", "Total No. of Records", 0, 0, 0, 0 }, { OPENISIS_RSP_NUMR, 0, FTN, 0, 0, 5, "size", "Number of Records", 0, 0, 0, 0 }, { OPENISIS_RSP_CERR, 0, FTN, 0, 0, 5, "error2", "Client Side Error", 0, 0, 0, 0 } }; static Fdt _fdtrsp = { sizeof (_fdrsp) / sizeof (_fdrsp[0]), _fdrsp, 0 }; /* ************************************************************ public data */ const Fdt *openIsisFdtSyspar = &_fdtsys; const Fdt *openIsisFdtScheme = &_fdtsch; const Fdt *openIsisFdtDbpar = &_fdtdb; const Fdt *openIsisFdtFd = &_fdtfd; const Fdt *openIsisFdtFdt = &_fdtfdt; const Fdt *openIsisFdtRqs = &_fdtrqs; const Fdt *openIsisFdtRsp = &_fdtrsp; /* ************************************************************ private functions */ static char *fFileGets (FILE *fp) { static char buf[4096]; char *res; int len; while (1) { nxtl: if (0 == (res = fgets (buf, sizeof(buf) - 1, fp))) { break; } buf[sizeof(buf) - 1] = 0; if (!(len = strlen (res))) { continue; } --len; while (0 <= len) { if ('\n' != res[len] && '\r' != res[len]) { break; } res[len] = 0; if (0 > --len) { goto nxtl; } } break; } return res; } static void strrtrim (char *str) { char *p = str + strlen (str) - 1; while (p >= str && (' ' == *p || '\t' == *p)) { *p-- = 0; } } static void fDesc2Name (char *d, char *n) { while (1) { if (!(*n = *d)) { return; } if ('A' <= *n && 'Z' >= *n) { *n += 'a' - 'A'; ++n; ++d; continue; } if ('_' == *n || ('a' <= *n && 'z' >= *n) || ('0' <= *n && '9' >= *n)) { ++n; ++d; continue; } while (1) { ++d; if (!*d) { *n = 0; return; } if (('A' <= *d && 'Z' >= *d) || ('a' <= *d && 'z' >= *d) || ('0' <= *d && '9' >= *d)) { break; } } *n++ = '_'; } } static int fLine2Fd (char *line, Fd **fd, int *num) { char name[FD_NAMELEN]; char pat[1 + FDF_PATLEN]; Fd buff, bufs; char *P; char *L = line; int tag, len, typ, rep, idx; if (FDF_NAMLEN + FDF_PATLEN >= strlen (L)) { return sMsg (ERR_INVAL, "fFileFd: illegal line <%s>", line); } strncpy (name, L, FDF_NAMLEN) [FDF_NAMLEN] = 0; strrtrim (name); if (! *name) { return sMsg (ERR_INVAL, "fFileFd: no descr in line <%s>", line); } L += FDF_NAMLEN; strncpy (pat, L, FDF_PATLEN) [FDF_PATLEN] = 0; strrtrim (pat); L += FDF_PATLEN; if (4 != sscanf (L, "%d %d %d %d", &tag, &len, &typ, &rep)) { return sMsg (ERR_INVAL, "fFileFd: no type in line <%s>", line); } switch (typ) { case 0: typ = FTX; break; case 1: typ = FTA; break; case 2: typ = FTN; break; case 3: typ = FTP; break; default: return sMsg (ERR_INVAL, "fFileFd: unrecognized type %d in line <%s>", typ, line); } memset (&buff, 0, sizeof (Fd)); buff.id = tag; buff.type = typ; buff.rep = 0 != rep; buff.len = len; strcpy (buff.desc, name); fDesc2Name (name, buff.name); if (! *buff.name) { return sMsg (ERR_INVAL, "fFileFd: illegal name in line <%s>", line); } if (FTP == typ) { if (! *pat) { return sMsg (ERR_INVAL, "fFileFd: illegal pat in line <%s>", line); } buff.pat = mDup (pat, -1); if (! buff.pat) { return sMsg (ERR_NOMEM, "fFileFd: cannot allocate pat"); } } idx = luti_ptrincr (fd, num, 1, sizeof(Fd), -1); if (0 > idx) { return sMsg (ERR_NOMEM, "fFileFd: cannot extend fd array %d", *num); } memcpy (*fd + idx, &buff, sizeof(Fd)); if (FTP != typ) { for (P = pat; *P; ++P) { memcpy (&bufs, &buff, sizeof(Fd)); bufs.subf = *P; bufs.type = FTX; idx = luti_ptrincr (fd, num, 1, sizeof(Fd), -1); if (0 > idx) { return sMsg (ERR_NOMEM, "fFileFd: cannot extend fd array %d", *num); } memcpy (*fd + idx, &bufs, sizeof(Fd)); } } return 0; } static int fLine2Tool (char *line, Rec **rec) { char tool[1 + FDF_TOOLEN]; char *L = line; int len, typ; switch (*L) { case 'w': case 'W': typ = OPENISIS_DFMT; break; case 'f': case 'F': typ = OPENISIS_DPFT; break; case 's': case 'S': typ = OPENISIS_DFST; break; default: return sMsg (ERR_INVAL, "fLine2Tool: illegal line <%s>", line); } L += 2; len = strlen (L); while (0 < len) { strncpy (tool, L, FDF_TOOLEN) [FDF_TOOLEN] = 0; strrtrim (tool); RADDS (*rec, typ, tool, !0); if (! *rec) { return sMsg (ERR_NOMEM, "fLine2Tool: cannot extend rec"); } L += FDF_TOOLEN; len -= FDF_TOOLEN; } return 0; } static int fResolveChilds (Fd *fd, int len) { char msg[256] = { 0 }; Fd *buf[CHAR_MAX]; Fd **C; Fd *F, *G; int numc; int err = 0; for (F = fd + len; --F >= fd; ) { if (! F->subf) { numc = 0; C = buf; for (G = fd + len; --G >= fd; ) { if (G->subf && (! G->id || G->id == F->id)) { if (CHAR_MAX == numc) { err = ERR_INVAL; sprintf (msg, "fResolveChilds: too many childs for %s", F->name); break; } *C++ = G; ++numc; } } if (numc) { C = mAlloc (numc * sizeof(Fd)); if (! C) { return sMsg (ERR_NOMEM, "fResolveChilds"); } F->subs = (Fd**) memcpy (C, buf, numc * sizeof(Fd)); F->slen = numc; } } } if (err) { return sMsg (err, msg); } return 0; } static void fFreeFd (Fd *fd) { if (fd) { if (fd->pat) { mFree (fd->pat); } if (fd->dflt) { mFree (fd->dflt); } if (fd->info) { mFree (fd->info); } if (fd->subs) { mFree (fd->subs); } } } static Fdt* fCleanupArr (Fdt *fdt, Fd *arr, int len) { Fd *F; for (F = arr + len; --F >= arr; ) { fFreeFd (F); } mFree (arr); mFree (fdt); return 0; } /* ************************************************************ package functions */ /* ************************************************************ public functions */ Fd* fById ( const Fdt *fdt, int id, int subf ) { Fd *f, *e; if (! fdt) { return 0; } for (e = (f = fdt->fd) + fdt->len; --e >= f; ) { if (id == e->id && subf == e->subf && ! (0xf0 & e->type)) { return e; } } return 0; } Fd* fByName ( const Fdt *fdt, const char *name ) { Fd *f, *e, *g; const char *p; int l, cnt; if (! fdt || ! name) { return 0; } f = fdt->fd; e = f + fdt->len; if ( '-' == *name ) /* ignore leading dash */ name++; if (! *name) { return 0; } if ( '0' <= *name && *name <= '9' ) return fById( fdt, a2i( name, -1 ), 0 ); p = name; while ( 'a' <= *p ? *p <= 'z' : '9' >= *p ? *p >= '0' : '_' == *p ) p++; l = p - name; if ( ! l || l > FD_NAMELEN - 1 ) return 0; for ( cnt = 0, g = 0; f < e; f++ ) { if ( *name == *f->name && !(0xf0 & f->type) /* is field */ && !memcmp( name, f->name, l ) ) { if (!f->name[l]) { return f; } g = f; ++cnt; } } return 1 == cnt ? g : 0; } /** lookup enum: if name is numeric, the value is returned, if legal considered are value entries with given id or id 0 - if there is an exact match with same id, this is used - if there is an exact match with id 0, this is used - if name is a unique prefix on given id, this is used - if name is no prefix on given id, but a unique prefix on 0, this is used */ int fEnum ( Fdt *fdt, int id, const char *name ) { Fd *f = fdt->fd, *e = f + fdt->len, *x0 = 0, *pi = 0, *p0 = 0; int ui = 1, u0 = 1; /* unique */ int l = strlen( name ); if ( ! l || l > FD_NAMELEN - 1 ) return NOENUM; if ( ('0' <= *name && *name <= '9') || ('-' == *name && '0' <= name[1] && name[1] <= '9') ) { int v = a2i( name, l ); for ( ; f < e; f++ ) { if ( FTV == f->type && v == f->len && (f->id == id || !f->id) ) return v; } return NOENUM; } for ( ; f < e; f++ ) { if ( FTV != f->type || *name != *f->name || memcmp( name, f->name, l ) || (f->id && f->id != id) ) continue; if ( !f->name[l] ) { /* exact match */ if ( f->id == id ) return f->len; x0 = f; /* f->id is 0 */ continue; } /* prefix match */ if ( f->id == id ) { if ( pi ) ui = 0; else pi = f; } else if ( p0 ) u0 = 0; else p0 = f; } return x0 ? x0->len : pi ? (ui ? pi->len : NOENUM) : p0 && u0 ? p0->len : NOENUM; } Fdt* fFree ( Fdt *fdt ) { Fd *F, *E; if (fdt) { if (fdt->fd) { for (E = (F = fdt->fd) + fdt->len; --E >= F; ) { fFreeFd (E); } mFree (fdt->fd); } if (fdt->rec) { mFree (fdt->rec); } mFree (fdt); } return 0; } Fdt* fFromFile (char *path) { FILE *fp; char *line; Fdt *res; Fd *fd; Rec *rec; int len, err; /* ldb::setext */ len = strlen (path) - 4; memcpy (path + len, FDF_EXT, 4); if ( 'A'<=path[len-1] && path[len-1]<= 'Z' ) { char *p = path + len; for ( ;*p; p++ ) /* use uppercase extensions */ if ( 'a' <= *p && *p <= 'z' ) *p -= 'a'-'A'; } fp = fopen (path, "r"); if (! fp) { sMsg (LOG_INFO | ERR_BADF, "no such fdt: %s", path); return 0; } res = mAlloc (sizeof(Fdt)); if (! res) { fclose (fp); sMsg (ERR_NOMEM, "fFromFile"); return 0; } sMsg (LOG_VERBOSE, "> reading fdt: %s", path); fd = 0; rec = 0; err = len = 0; while ((line = fFileGets (fp))) { if ('*' == *line) { continue; } if (':' == line[1]) { if (-ERR_NOMEM == (err = fLine2Tool (line, &rec))) { break; } continue; } if (-ERR_NOMEM == (err = fLine2Fd (line, &fd, &len))) { break; } } fclose (fp); res->fd = fd; res->rec = rec; res->len = len; if (-ERR_NOMEM == err) { return fFree (res); } sMsg (LOG_VERBOSE, "< %d entries in fdt", len); return res; } #define ADDFDS(tag,val) \ RADDS (rec, tag, val, !0); \ if (! rec) { return 0; } Rec* fFd2Rec (const Fd *fd, Rec *rec, int embchld) { char buf[16]; Fd **E, **F; Rec *child; if (! fd) { return rec; } i2a (buf, fd->id); ADDFDS (OPENISIS_FDID, buf); if ((*buf = fd->subf)) { buf[1] = 0; ADDFDS (OPENISIS_FDSUB, buf); } i2a (buf, fd->type); ADDFDS (OPENISIS_FDTYPE, buf); i2a (buf, fd->rep); ADDFDS (OPENISIS_FDREP, buf); i2a (buf, fd->len); ADDFDS (OPENISIS_FDLEN, buf); ADDFDS (OPENISIS_FDNAME, fd->name); ADDFDS (OPENISIS_FDDESC, fd->desc); if (fd->pat) { ADDFDS (OPENISIS_FDPAT, fd->pat); } if (fd->dflt) { ADDFDS (OPENISIS_FDDFLT, fd->dflt); } if (fd->info) { rec = luti_wrap (rec, fd->info, OPENISIS_FDINFO); if (! rec) { return 0; } } if (embchld && fd->slen) { for (E = (F = fd->subs) + fd->slen; rec && F < E; ++F) { child = fFd2Rec (*F, 0, 0); if (! child) { return rec; } rec = luti_wrap (rec, child, OPENISIS_FDCHLD); } if (rec) { rec = rAddI (rec, OPENISIS_FDNUMC, fd->slen, !0); } } return rec; } Rec* fFdt2Rec (const Fdt *fdt, Rec *rec, int embchld) { Fd *E, *F; Rec *R; int len; if (! fdt) { return rec; } if (fdt->rec) { rec = luti_wrap (rec, fdt->rec, OPENISIS_FDT_REC); if (! rec) { return 0; } } for (E = (F = fdt->fd) + fdt->len, len = 0; F < E; ++F) { if (! F->subf || ! embchld) { R = fFd2Rec (F, 0, embchld); if (! R) { return rec; } rec = luti_wrap (rec, R, OPENISIS_FDT_FD); mFree (R); if (! rec) { return 0; } ++len; } } rec = rAddI (rec, OPENISIS_FDT_LEN, len, !0); return rec; } Fd *fRec2Fd (Rec *rec, Fd *buf) { char name[FD_NAMELEN]; Field *F, *E; Fd *fd; int got = 0; if (! rec) { return 0; } if (! (fd = buf)) { fd = mAlloc (sizeof(Fd)); if (! fd) { return 0; } } *name = 0; for (E = (F = rec->field) + rec->len; F < E; ++F) { switch (F->tag) { case OPENISIS_FDID: got |= 0x01; fd->id = a2id (F->val, F->len, 0); break; case OPENISIS_FDSUB: if (! (1 == F->len || (2 == F->len && ! F->val[1]))) { sMsg (ERR_INVAL, "fRec2Fd: ignoring illegal subfield id (%s)", name); } else { fd->subf = F->val[0]; } break; case OPENISIS_FDTYPE: got |= 0x02; fd->type = (char) a2id (F->val, F->len, 0); break; case OPENISIS_FDREP: if (! (1 == F->len || (2 == F->len && ! F->val[1]))) { sMsg (ERR_INVAL, "fRec2Fd: ignoring illegal repeatable flag (%s)", name); } else { fd->rep = F->val[0] && '0' != F->val[0]; } break; case OPENISIS_FDLEN: fd->len = (short) a2id (F->val, F->len, 0); break; case OPENISIS_FDNAME: got |= 0x04; if (FD_NAMELEN <= F->len) { memcpy (name, F->val, FD_NAMELEN - 1); name[FD_NAMELEN - 1] = 0; sMsg (ERR_INVAL, "fRec2Fd: name too long (%d) - truncated to %s", F->len, name); } else { memcpy (name, F->val, F->len); name[F->len] = 0; } if (! *fd->desc) { strcpy (fd->desc, name); } fDesc2Name (name, fd->name); if (!*(fd->name)) { sMsg (ERR_INVAL, "fRec2Fd: illegal name (%s)", name); got &= ~0x04; } break; case OPENISIS_FDDESC: if (FD_NAMELEN <= F->len) { memcpy (fd->desc, F->val, FD_NAMELEN - 1); fd->desc[FD_NAMELEN - 1] = 0; sMsg (ERR_INVAL, "fRec2Fd: descr too long (%d) - truncated to %s (%s)", F->len, fd->desc, name); } else { memcpy (fd->desc, F->val, F->len); fd->desc[F->len] = 0; } break; case OPENISIS_FDPAT: if (fd->pat) { sMsg (ERR_INVAL, "fRec2Fd: ignoring multiple occurences of pattern (%s)", name); } else { fd->pat = (char*) mAlloc (1 + F->len); if (! fd->pat) { goto err; } memcpy (fd->pat, F->val, F->len); fd->pat[F->len] = 0; } break; case OPENISIS_FDDFLT: if (fd->dflt) { sMsg (ERR_INVAL, "fRec2Fd: ignoring multiple occurences of dflt (%s)", name); } else { fd->dflt = (char*) mAlloc (1 + F->len); if (! fd->dflt) { goto err; } memcpy (fd->dflt, F->val, F->len); fd->dflt[F->len] = 0; } break; case OPENISIS_FDINFO: if (fd->info) { sMsg (ERR_INVAL, "fRec2Fd: ignoring multiple occurences of info (%s)", name); } else { int pos = F - rec->field; fd->info = luti_unwrap (rec, &pos, OPENISIS_FDINFO, -1); if (! fd->info) { goto err; } F = rec->field + pos - 1; } break; default: sMsg (ERR_INVAL, "fRec2Fd: ignoring unexpected tag %d (%s)", F->tag, name); } } if (0x07 != got) { sMsg (ERR_TRASH, "fRec2Fd: incomplete field description [%x] (%s)", got, name); err: if (fd->pat) { mFree (fd->pat); } if (fd->dflt) { mFree (fd->dflt); } if (fd->info) { mFree (fd->info); } if (fd != buf) { mFree (fd); } return 0; } return fd; } Fdt *fRec2Fdt (Rec *rec) { Rec *R, *cfg; Fdt *fdt; Fd fdbuf; Fd *F, *arr; int err, num, len, pos; if (!rec) { return 0; } num = rInt (rec, OPENISIS_FDT_LEN, 0, 0); if (0 >= num) { return 0; } arr = (Fd*) mAlloc (num * sizeof (Fd)); if (!arr) { return 0; } fdt = (Fdt*) mAlloc (sizeof(Fdt)); if (! fdt) { mFree (arr); return 0; } for (F = arr, len = pos = 0; num; --num) { R = luti_unwrap (rec, &pos, OPENISIS_FDT_FD, -1); if (!R) { return fCleanupArr (fdt, arr, len); } memset (&fdbuf, 0, sizeof(Fd)); if (fRec2Fd (R, &fdbuf)) { memcpy (F, &fdbuf, sizeof(Fd)); ++len; ++F; } mFree (R); } if (len != num) { num = len * sizeof(Fd); F = (Fd*) mAlloc (num); if (! F) { return fCleanupArr (fdt, arr, len); } memcpy (F, arr, num); mFree (arr); arr = F; } cfg = luti_unwrap (rec, 0, OPENISIS_FDT_REC, -1); fdt->len = len; fdt->fd = arr; fdt->rec = cfg; err = fResolveChilds (arr, len); if (-ERR_NOMEM == err) { return fFree (fdt); } return fdt; } lfdt.h0100644000113000011300000000316107656453761010661 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LFDT_H /* $Id: lfdt.h,v 1.10 2003/05/07 16:32:36 mawag Exp $ FDT support */ #include "loi.h" /** build fd from rec. @param rec source @param buf optional target buffer @return buf or allocated fd or null on error */ extern Fd* fRec2Fd (Rec *rec, Fd *buf); /** convert fd to record. @param fd source @param rec target where to append fd, may be null @param embchld boolean indicating whether subfield childs should be embedded in target record @return target or null if no memory available */ extern Rec* fFd2Rec (const Fd *fd, Rec *rec, int embchld); /** read fdt from file named path extended by '.fdt'. * @return fdt or null */ extern Fdt* fFromFile (char *path); #define LFDT_H #endif /* LFDT_H */ lfmt.c0100644000113000011300000010545407644412744010666 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lfmt.c,v 1.14 2003/04/08 00:20:52 kripke Exp $ implementation of record cooking. */ #include /* memset et al */ #include "ldb.h" #define LFMT_DBG LOG_WARN /* level of debugging */ /* ************************************************************ private types */ typedef struct { /* function definition */ const char *name; /* hmmm ... fixed char[] probably faster */ const char *sign; int tag; } LfmtToken; typedef enum { LFMT_MP, /* proof */ LFMT_MH, /* heading */ LFMT_MD /* data */ } lfmtMode; typedef struct LfmtIter_ { /* value iterator */ const unsigned char *def; /* position in format where iter was opened. */ const unsigned char *end; /* position in format where iter was closed. */ int pos; /* next field pos */ int off; /* offset to start search for next subfield */ int occ; /* next occ */ /* filter */ int tag; char sub; char submode; /* primary loop (counted by occ) is over subfields. */ int from; /* primary range. */ int to; int sfrom; /* additional subfield range. */ int sto; } LfmtIter; #define LFMT_NITER 8 typedef struct LfmtRec_ { /* record context -- pushed by the r in REF */ struct LfmtRec_ *prev; Rec *r; int frame; /* stack frame associated with this record */ const unsigned char *loop; /* position in format where loop was opened. */ LfmtIter iter[LFMT_NITER]; /* iterators in loop */ int occ; /* loop occ */ unsigned char i; /* iterator count */ char more; /* more flag */ char had; /* had flag */ char sub; /* database's subfield separator, typically ^ or $ */ } LfmtRec; typedef struct LfmtFmt_ { /* format context -- pushed by @ */ struct LfmtFmt_ *prev; const unsigned char *f; const unsigned char *e; /* end -- just for paranoia check */ lfmtMode mode; int upcase; /* uppercase flag */ } LfmtFmt; typedef struct { /* other type */ const LfmtToken *tok; /* recognized token */ int data; /* start of params in output */ int used; /* used len of output up to our params */ char pos; /* position in signature */ char want; /* currently expected type */ char expl; /* explicit flag */ char rec; /* record flag */ } LfmtFrame; /* list of special token tags after the format tags 1 .. 999. unlike formatting tokens, which just emit a field with their negative tag, the literal's tag is reset to 0, and all other tokens emit fields with tags based on their parameters. tokens are organized in groups of (up to) 16. the group of an operator determines it's precedence. */ #define LFMT_DEFGROUP(i) ((i) << 4) /* * 16 */ #define LFMT_GROUP(i) ((i) >> 4) /* token groups */ enum { LFMT_LITERAL = 63, /* literals group starting from 63*16 = 1008 */ /* syntactical groups starting from 64*16 = 0x400 */ LFMT_FUNCTIONS, LFMT_VALUES = 96, /* starting from 96*16 = 0x600 */ LFMT_ITERATORS, LFMT_OPITER, /* iterator operators */ LFMT_OPSTR, /* string operators */ LFMT_OPMUL, /* multiplicative operators */ LFMT_OPADD, /* additive operators */ LFMT_OPREL, /* relational operators */ LFMT_OPBOOL, /* boolean operators */ LFMT_OPASSIGN, /* assignment */ LFMT_STOPPER = 128, /* starting from 128*16 = 0x800 */ LFMT_SYNTAX }; enum { /* literals */ LFMT_LIT = LFMT_DEFGROUP( LFMT_LITERAL ), LFMT_COND, LFMT_REP, LFMT_COMM, LFMT_BARE, /* field of bareword C&A literal */ LFMT_NUMBER, /* standalone numeric literal */ /* functions */ LFMT_OPEN = LFMT_DEFGROUP( LFMT_FUNCTIONS ), LFMT_AMP, LFMT_AT, LFMT_XREF, LFMT_S, LFMT_Z, LFMT_F, /* */ LFMT_MDL, LFMT_MDU, LFMT_MHL, LFMT_MHU, LFMT_MPL, LFMT_MPU, /* values */ LFMT_SI = LFMT_DEFGROUP( LFMT_VALUES ), LFMT_EI, /* iterators */ LFMT_V = LFMT_DEFGROUP( LFMT_ITERATORS ), LFMT_D, LFMT_N, /* operator groups by decreasing precedence */ LFMT_INDEX = LFMT_DEFGROUP( LFMT_OPITER ), LFMT_SUBFIELD, LFMT_STAR = LFMT_DEFGROUP( LFMT_OPSTR ), LFMT_DOT, LFMT_INDENT, LFMT_MUL = LFMT_DEFGROUP( LFMT_OPMUL ), LFMT_DIV, LFMT_PLUS = LFMT_DEFGROUP( LFMT_OPADD ), LFMT_MINUS, LFMT_EQ = LFMT_DEFGROUP( LFMT_OPREL ), LFMT_NE, LFMT_LT, LFMT_LE, LFMT_GT, LFMT_GE, LFMT_CT, LFMT_AND = LFMT_DEFGROUP( LFMT_OPBOOL ), LFMT_OR, LFMT_NOT, LFMT_ASSIGN = LFMT_DEFGROUP( LFMT_OPASSIGN ), /* stoppers */ LFMT_COMMA = LFMT_DEFGROUP( LFMT_STOPPER ), LFMT_RANGE, LFMT_CLOSE, /* syntax */ LFMT_BLANK = LFMT_DEFGROUP( LFMT_SYNTAX ), LFMT_LOOP, LFMT_WHILE, LFMT_REF, LFMT_IF, LFMT_THEN, LFMT_ELSE, LFMT_FI, LFMT_SELECT, LFMT_CASE, LFMT_ELSECASE, LFMT_ENDSEL, LFMT__LAST }; /* while fields emitted by formatting and literal tokens have tags between -1 and -1023 (0xffffffff to 0xfffffc01) with all high bits set, numerics have lowest negative values ranging from 0x80000000 to 0xbfffffff (-2147483648 to -1073741825). That means the highest two bits are 1 and 0. All numeric fields have a val of NULL and use len to hold a number. Besides pure int or fixed-point numbers, there are other numerical types: next two bits may indicate a range or field selector. The lower 3 bytes may give fractional part or range end. */ #define LFMT_NMASK 0xf0000000L /* highest bits 1 and 0 mark a number */ #define LFMT_NUM 0x80000000L /* basic number */ #define LFMT_ITR 0xa0000000L /* with iterator bit */ #define LFMT_VAL 0x00ffffffL /* mask for 3 bytes additional info */ /* is any numeric */ #define LFMT_ISNUM(t) (LFMT_NUM == (LFMT_NMASK & (t))) /* is pure int */ #define LFMT_ISINT(t) (LFMT_NUM == (t)) /* is iterator */ #define LFMT_ISITR(t) (LFMT_ITR == (LFMT_NMASK & (t))) /* ************************************************************ private data */ static const LfmtToken number = { "", "_1i", LFMT_NUMBER }; /* while a numeric *field* has large LFMT_NUM tag, the anonymous number *token* has a literal tag, so we don't take it as high precedence operator. */ static const LfmtToken blank = { "_", "_0o_", LFMT_BLANK }; static const LfmtToken loop = { "loop", "_0o_", LFMT_LOOP }; /* 32 = 33 = ! 34 = " 35 = # 36 = $ 37 = % 38 = & 39 = ' 40 = ( 41 = ) 42 = * 43 = + 44 = , 45 = - 46 = . 47 = / 48 = 0 49 = 1 50 = 2 51 = 3 52 = 4 53 = 5 54 = 6 55 = 7 56 = 8 57 = 9 58 = : 59 = ; 60 = < 61 = = 62 = > 63 = ? 64 = @ 65 = A 66 = B 67 = C 68 = D 69 = E 70 = F 71 = G 72 = H 73 = I 74 = J 75 = K 76 = L 77 = M 78 = N 79 = O 80 = P 81 = Q 82 = R 83 = S 84 = T 85 = U 86 = V 87 = W 88 = X 89 = Y 90 = Z 91 = [ 92 = \ 93 = ] 94 = ^ 95 = _ 96 = ` 97 = a 98 = b 99 = c 100 = d 101 = e 102 = f 103 = g 104 = h 105 = i 106 = j 107 = k 108 = l 109 = m 110 = n 111 = o 112 = p 113 = q 114 = r 115 = s 116 = t 117 = u 118 = v 119 = w 120 = x 121 = y 122 = z 123 = { 124 = | 125 = } 126 = ~ */ /* stops dictionary search */ static const char sentinel[] = "\x7f"; #define ENDDICT { sentinel, NULL, 0 } static const LfmtToken dictSpecial[] = { { "", "_0", 0 }, /* eof */ { "!", "_1x", OPENISIS_FMT_ESC }, { "\"", "_1x", LFMT_COND }, { "#", "_0", OPENISIS_FMT_SHARP }, /* { "$", "_0", LFMT_LIT }, */ { "%", "_0", OPENISIS_FMT_PERCENT }, { "&", "_1as_", LFMT_AMP }, { "'", "_1x", LFMT_LIT }, { "(", "_0", LFMT_OPEN }, { ")", "_0", LFMT_CLOSE }, { "*", "n1n", LFMT_MUL }, { "+", "n1n", LFMT_PLUS }, { ",", "_0", LFMT_COMMA }, { "-", "n1n", LFMT_MINUS }, { ".", "s1i", LFMT_DOT }, { "..","_0", LFMT_RANGE }, { "/", "n1n", LFMT_DIV }, /* alias { "/", "_0", OPENISIS_FMT_SLASH }, */ { "/*","_1x", LFMT_COMM }, /* 0 - 9 */ { ":", "s1s", LFMT_CT }, { ":=","v1f", LFMT_ASSIGN }, /* { ";", "b_", LFMT_LIT }, */ { "<", "n1n", LFMT_LT }, { "<=","n1n", LFMT_LE }, { "<>","n1n", LFMT_NE }, { "=", "n1n", LFMT_EQ }, { ">", "n1n", LFMT_GT }, { ">=","n1n", LFMT_GE }, /* { "?", "b_", LFMT_LIT }, */ { "@", "_1a", LFMT_AT }, /* A - Z */ { "[", "v1nn", LFMT_INDEX }, /* { "\\", "_0", LFMT_LIT }, */ { "]", "_0", LFMT_CLOSE }, { "^", "v1c", LFMT_SUBFIELD }, /* { "_", "_0", LFMT_LIT }, */ /* { "`", "_0", LFMT_LIT }, */ /* a - z */ { "{", "_0", OPENISIS_FMT_OPEN }, { "|", "_1x", LFMT_REP }, { "}", "_0", OPENISIS_FMT_CLOSE }, { "~", "_1n", LFMT_NOT }, ENDDICT }; static const LfmtToken dictA[] = { ENDDICT }; static const LfmtToken dictB[] = { { "B", "_0", OPENISIS_FMT_B }, ENDDICT }; static const LfmtToken dictC[] = { { "C", "_1i", OPENISIS_FMT_C }, ENDDICT }; static const LfmtToken dictD[] = { { "D", "_1i", LFMT_D }, ENDDICT }; static const LfmtToken dictE[] = { { "E", "_1i", LFMT_EI }, { "ELSE", "_1s_", LFMT_ELSE }, ENDDICT }; static const LfmtToken dictF[] = { /* { "F", "_1nii", LFMT_F }, */ { "F", "_1i", OPENISIS_FMT_F }, ENDDICT }; static const LfmtToken dictG[] = { ENDDICT }; static const LfmtToken dictH[] = { ENDDICT }; static const LfmtToken dictI[] = { { "I", "_0", OPENISIS_FMT_I }, { "IF", "_1n", LFMT_IF }, ENDDICT }; static const LfmtToken dictJ[] = { ENDDICT }; static const LfmtToken dictK[] = { ENDDICT }; static const LfmtToken dictL[] = { { "LINK", "_1s_", OPENISIS_FMT_LINK }, ENDDICT }; static const LfmtToken dictM[] = { { "M", "_1ii", OPENISIS_FMT_M }, { "MDL", "_0", LFMT_MDL }, { "MDU", "_0", LFMT_MDU }, { "MHL", "_0", LFMT_MHL }, { "MHU", "_0", LFMT_MHU }, { "MPL", "_0", LFMT_MPL }, { "MPU", "_0", LFMT_MPU }, ENDDICT }; static const LfmtToken dictN[] = { { "N", "_1i", LFMT_N }, { "NC", "_0i", OPENISIS_FMT_NC }, ENDDICT }; static const LfmtToken dictO[] = { ENDDICT }; static const LfmtToken dictP[] = { { "PICT", "_1s", OPENISIS_FMT_PICT }, ENDDICT }; static const LfmtToken dictQ[] = { { "QC", "_0", OPENISIS_FMT_QC }, ENDDICT }; static const LfmtToken dictR[] = { { "REF", "_2rs_", LFMT_REF }, ENDDICT }; static const LfmtToken dictS[] = { { "S", "_1s_", LFMT_S }, ENDDICT }; static const LfmtToken dictT[] = { { "TAB", "_0i", OPENISIS_FMT_TAB }, ENDDICT }; static const LfmtToken dictU[] = { { "UL", "_0", OPENISIS_FMT_UL }, ENDDICT }; static const LfmtToken dictV[] = { { "V", "_1i", LFMT_V }, ENDDICT }; static const LfmtToken dictW[] = { ENDDICT }; static const LfmtToken dictX[] = { { "X", "_0i", OPENISIS_FMT_X }, ENDDICT }; static const LfmtToken dictY[] = { ENDDICT }; static const LfmtToken dictZ[] = { { "Z", "_1is", LFMT_Z }, ENDDICT }; static const LfmtToken * const dictAZ[26] = { dictA, dictB, dictC, dictD, dictE, dictF, dictG, dictH, dictI, dictJ, dictK, dictL, dictM, dictN, dictO, dictP, dictQ, dictR, dictS, dictT, dictU, dictV, dictW, dictX, dictY, dictZ }; /* character conversion */ static const unsigned char upcase[256] = { 0,' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ', ' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ',' ', ' ','!','"','#','$','%','&','\'', '(',')','*','+',',','-','.','/', '0','1','2','3','4','5','6','7', '8','9',':',';','<','=','>','?', '@','A','B','C','D','E','F','G','H','I','J','K','L','M','N','O', 'P','Q','R','S','T','U','V','W','X','Y','Z','[','\\',']','^','_', '`','A','B','C','D','E','F','G','H','I','J','K','L','M','N','O', 'P','Q','R','S','T','U','V','W','X','Y','Z','{','|','}','~',' ', '?','?','?','?','?','?','?','?','?','?','?','?','?','?','?','?', '?','?','?','?','?','?','?','?','?','?','?','?','?','?','?','?', '?','?','?','?','?','?','?','?','?','?','?','?','?','?','?','?', '?','?','?','?','?','?','?','?','?','?','?','?','?','?','?','?', '?','?','?','?','?','?','?','?','?','?','?','?','?','?','?','?', '?','?','?','?','?','?','?','?','?','?','?','?','?','?','?','?', '?','?','?','?','?','?','?','?','?','?','?','?','?','?','?','?', '?','?','?','?','?','?','?','?','?','?','?','?','?','?','?','?' }; enum { DIG=0x01, /* digit */ ALP=0x02, /* ascii alphabetic */ IDE=0x04, /* other identifier character */ LIT=0x10, /* as signature type, eat literal (a,c,i,x) */ ALT=ALP|LIT }; /* character type */ static const unsigned char ctype[256] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, DIG,DIG,DIG,DIG,DIG,DIG,DIG,DIG,DIG,DIG, 0, 0, 0, 0, 0, 0, /*'@','A','B','C','D','E','F','G','H','I','J','K','L','M','N','O',*/ 0,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP, /*'P','Q','R','S','T','U','V','W','X','Y','Z','[','\\',']','^','_',*/ ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP, 0, 0, 0, 0,IDE, 0,ALT,ALP,ALT,ALP,ALP,ALP,ALP,ALP,ALT,ALP,ALP,ALP,ALP,ALP,ALP, ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALP,ALT,ALP,ALP, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }; /* ************************************************************ private functions */ /* don't rely on braindead ctype.h, it may use locale */ #define ISALPHA(u) ( ALP & ctype[(unsigned char)(u)]) #define ISDIGIT(u) ( DIG & ctype[(unsigned char)(u)]) #define ISALNUM(u) ((ALP|DIG) & ctype[(unsigned char)(u)]) #define ISIDENT(u) ((ALP|DIG|IDE) & ctype[(unsigned char)(u)]) #define ISLITER(u) ( LIT & ctype[(unsigned char)(u)]) /* ************************************************************ package functions */ /* ************************************************************ public functions */ Rec* rFmt ( Rec *buf, const char *fmt, Rec *irec ) { /* context */ Rec *o = buf; #define ADD( ntag, s, n ) do { \ RADD( o, ntag, s, n, o!=buf ); \ if ( NULL == o ) goto outofmem; \ } while(0) #define CAT( s, n ) do { \ RCAT( o, s, n, o!=buf ); \ if ( NULL == o ) goto outofmem; \ } while(0) /* field add */ #define FADD( f ) ADD( f.tag, f.val, f.len ) /* string add */ #define SADD( ntag, s ) ADD( ntag, s, strlen(s) ) #define SCAT( s ) CAT( s, strlen(s) ) #define SERR( err ) do { \ o = rMsg( o, o!=buf, -LOG_ERROR, \ "@%d: %s", fp - f->f, err ); \ if ( NULL == o ) goto outofmem; \ } while(0) #define SERR1( fmt, a ) do { \ o = rMsg( o, o!=buf, -LOG_ERROR, \ "@%d: " fmt, fp - f->f, a ); \ if ( NULL == o ) goto outofmem; \ } while(0) #define SERR2( fmt, a, b ) do { \ o = rMsg( o, o!=buf, -LOG_ERROR, \ "@%d: " fmt, fp - f->f, a, b ); \ if ( NULL == o ) goto outofmem; \ } while(0) /* number add */ #define NADD( i, d ) do { \ ADD( LFMT_NUM | (d), NULL, 0 ); \ o->field[o->len - 1].val = NULL; \ o->field[o->len - 1].len = i; \ } while(0) LfmtFrame stack[128], *s = stack; /* current frame (always == stack+sp) */ int sp = -1; /* push new frame for token t */ #define PUSHFRAME( t ) do { \ if ( ++sp == sizeof(stack)/sizeof(stack[0]) ) goto overflow; \ s = stack+sp; \ memset( s, 0, sizeof(*s) ); \ s->data = o->len; \ s->used = o->used; \ s->tok = (t); \ s->want = (t)->sign[2]; \ } while(0) LfmtFmt fmt0, *f = NULL; /* default and current format */ const unsigned char *fp = NULL; /* format pointer */ #define PUSHFORMAT( pfmt, fstr ) do { \ memset( pfmt, 0, sizeof(*(pfmt)) ); \ (pfmt)->prev = f; \ f = (pfmt); \ fp = f->f = (const unsigned char *)fstr; \ f->e = fp + strlen(fstr); \ } while(0) LfmtRec rec0, *r = NULL; /* default and current record */ #define PUSHRECORD( prec, nrec ) do { \ memset( prec, 0, sizeof(*(prec)) ); \ (prec)->prev = r; \ r = (prec); \ r->frame = sp; \ r->r = nrec; \ r->sub = '^'; \ } while(0) int tlen = 0; /* token length */ PUSHFRAME( &blank ); s->rec = 1; PUSHFORMAT( &fmt0, fmt ); PUSHRECORD( &rec0, irec ); o->dbid = irec->dbid ; o->rowid = irec->rowid; for ( ;; fp += tlen ) { const LfmtToken *tok = NULL; Field lit; unsigned char u; int start = s->data + s->pos; /* start of current param in output */ int flds = o->len - start; /* # fields for current param */ int vpos = 0; /* pos in dict where we had match of tlen chars */ int dpos = 0; /* pos in dict */ int expl = 0; unsigned char eatlit = 0; const LfmtToken *dict; enum { CLOSE_OP, /* don't close anything, operator wants field */ CLOSE_VAR, /* close variables only */ CLOSE_FIELD, /* coerce field only */ CLOSE_PARAM, /* compact param only (eat comma) */ CLOSE_EXPL, /* close frame explicitly (eat closing token) */ CLOSE_IMPL /* close frame implicitly, continue on token */ } close; static const char *closename[] = { "", "var", "field", "param", "expl", "impl" }; lit.tag = 0; tlen = 0; /* verified length */ /* LOG_DBG( LFMT_DBG, "frame %s%c%s id %d pos %d got %d want %c%s", s->tok->name, s->expl ? '(' : ' ', s->tok->sign, s->tok->tag, s->pos, o->len - s->data, s->want, flds ? " (have)" : "" ); */ /* lookup token */ while ( ' ' == (u = upcase[*fp]) ) /* eat whitespace */ fp++; if ( ISALPHA(u) ) { dict = dictAZ[u-'A']; if ( sentinel != dict->name ) u = upcase[ fp[tlen = 1] ]; } else if ( ISDIGIT( *fp ) ) { tok = &number; /* read numeric literal */ goto countnumber; } else dict = dictSpecial; /* tok is the last candidate that matched up to IT'S length */ for ( ;; dpos++ ) { unsigned char test; LOG_DBG( LOG_TRACE, "trying token '%s'", dict[dpos].name ); if ( tlen && vpos != dpos && strncmp( dict[dpos].name, dict[vpos].name, tlen ) ) break; /* ran out of verified length */ if ( u == dict[dpos].name[tlen] ) vpos = dpos; while ( (test = dict[dpos].name[tlen]) && u == test ) /* test this entry -- same pos, next char */ u = upcase[ fp[++tlen] ]; if ( !test ) /* dict entry ends here -- a hit so far */ tok = &dict[dpos]; /* LOG_DBG( LFMT_DBG, "\thit on token %s(%s) len %d", tok->name, tok->sign, tlen ); */ if ( u < test || !u ) /* too large - bail out*/ break; } if ( NULL != tok ) { /* had some match */ if ( tok < &dict[vpos] ) /* had later == longer match */ tok = NULL; else if ( dict != dictSpecial ) { /* alpha-token must not be followed by alpha */ assert( tlen == (int)strlen(tok->name) ); if ( ISALPHA( u ) && ISALPHA( tok->name[tlen-1] ) ) tok = NULL; } } if ( NULL == tok /* no hit in alphadict */ || (dictSpecial == tok && *fp) /* false eof hit in special dict */ ) { SERR2( "unrecognized token '%.*s'", tlen+1, fp ); goto broken; } if ( ISLITER( tok->sign[2] ) ) /* eat literals */ switch ( tok->sign[2] ) { /* c, a, i, x */ case 'c': if ( (lit.len = (eatlit = fp[tlen]) ? 1 : 0 ) ) { lit.tag = -LFMT_BARE; lit.val = (char *)fp + tlen; tlen++; } break; case 'a': if ( (eatlit = ISIDENT( fp[tlen] )) ) { lit.tag = -LFMT_BARE; lit.val = (char *)fp + tlen; lit.len = 0; while ( ISIDENT( fp[tlen] ) ) { lit.len++; tlen++; } } break; case 'i': countnumber: lit.val = NULL; lit.len = 0; if ( (eatlit = ISDIGIT( u = fp[tlen] )) ) { lit.tag = LFMT_NUM; while ( u && 10 > (u -= '0') ) { lit.len = 10*lit.len + u; u = fp[++tlen]; } } /* decimal digits on standalone numeric literal */ if ( &number == tok && (unsigned char)('.'-'0') == u && '.' != fp[tlen+1] /* avoid eating 1..2 */ ) { int dec = 0; int v = 0; while ( (u = fp[++tlen]) && 10 > (u -= '0') ) if ( 6 > dec ) { v = 10*v + u; dec++; } while ( 6 < dec++ ) v *= 10; lit.tag = LFMT_NUM | (v & LFMT_VAL); } break; case 'x': switch ( tok->tag ) { case LFMT_COMM: eatlit='*'; break; case OPENISIS_FMT_ESC: eatlit=fp[tlen++]; break; default: eatlit=tok->name[0]; break; } lit.tag = - tok->tag; lit.val = (char *)fp+1; while ( (u = fp[tlen++]) && (eatlit != u || (LFMT_COMM==tok->tag && '/' != fp[tlen])) ) ; if ( !u ) { SERR1( "unterminated %s-literal", tok->name ); goto broken; } lit.len = tlen - 2; if ( LFMT_COMM == tok->tag ) { tlen++; continue; } } if ( tok->sign[2] ) { /* token wants params */ if ( (!eatlit || tok->sign[3]) && (expl = '(' == fp[tlen]) ) tlen++; /* eat ( */ } /* got token */ LOG_DBG( LFMT_DBG, "at %d: '%.*s' %s(%s) id %d (%x) group %d pos %d%c of %s%c%s", fp - f->f, tlen > 10 ? 10 : tlen, fp, tok->name, tok->sign, tok->tag, tok->tag, LFMT_GROUP(tok->tag), s->pos, s->want, s->tok->name, s->expl?'(':' ', s->tok->sign ); /* close what needs to be closed */ do { int group = LFMT_GROUP( tok->tag ); int dflt, wantnum, gotnum; close = CLOSE_FIELD; if ( '_' != tok->sign[0] /* token is operator ... */ && (s->expl /* ... within explicit frame or ... */ || group < LFMT_GROUP( s->tok->tag )) /* ... of higher prec */ ) /* we take it -- even as a variable ? */ close = 'v' == tok->sign[0] ? CLOSE_OP : CLOSE_VAR; else switch ( tok->tag ) { case LFMT_COMMA: case LFMT_RANGE: close = CLOSE_PARAM; break; case LFMT_CLOSE: close = CLOSE_EXPL; break; } if ( CLOSE_VAR > close ) /* leave variable to operator */ break; /* close loop */ if ( flds && LFMT_ISITR( o->field[ o->len-1 ].tag ) ) { /* dereference iterator variable */ LfmtIter * iter = r->iter + r->i; if ( ! iter->occ ) { if ( ! iter->from ) iter->from = 1; /* field defaults to 1..0 (all) */ if ( ! iter->sfrom ) iter->sfrom = iter->sto = 1; /* subfield defaults to 1..1 (1st) */ } LOG_DBG( LFMT_DBG, "iterator %d/%d V%d[%d..%d]^%c[%d..%d]%s", o->field[ o->len-1 ].len, r->occ, iter->tag, iter->from, iter->to, iter->sub ? iter->sub : ' ', iter->sfrom, iter->sto, iter->submode ? " subfield mode" : "" ); o->len--; /* kill the variable. */ /* move to next legal position. In standard mode, where we advance one field occurence at a time, this is just the occ >= from. A legal position may still emit no field, if a selected subfield is not available in the current field occ. In subfield mode, we may have to advance 0, 1 or several field occurences to find next occ of subfield. */ r->had = 0; do { int socc; if ( 0 > iter->occ ) /* we were already done */ break; /* if have legal occurence, ADD it */ if ( iter->occ >= iter->from ) for ( socc=1;; socc++ ) { /* not initialization pass. */ Field *v = r->r->field + iter->pos; const char *src = v->val; int len = v->len; if ( iter->sub ) { /* find subfield */ const char *p = src + iter->off, *e = src+len; while ( p < e && ( r->sub != *p++ || e == p || (iter->sub != *p++ && iter->sub != '*') ) ) ; iter->off = p - src; if ( p >= e ) break; if ( iter->sfrom && socc < iter->sfrom ) continue; if ( iter->sto && socc > iter->sto ) break; src = p; while ( p < e && r->sub != *p ) p++; iter->off += len = p - src; } /* make sure there's enough room, +3 for DATA mode */ ADD( v->tag, NULL, len+3 ); if ( ! f->mode ) CAT( src, len ); else { const char *e = src + len; char *dst = (char *)o->field[ o->len - 1 ].val; for ( ; src < e; ) { if ( '<' == *src ) { for ( src++; src < e; *dst++ = *src++ ) if ( '=' == *src || '>' == *src ) { while ( '>' != *src++ && src < e ) ; if ( src < e && '<' == *src ) { /* have >< */ *dst++ = ';'; *dst++ = ' '; } break; } } else if ( r->sub == *src ) { if ( ++src == e ) break; if ( dst == o->field[ o->len - 1 ].val ) ; /* no spearator */ else if ( 'a' == *src ) *dst++ = ';'; else if ( 'j' > *src ) *dst++ = ','; else *dst++ = '.'; src++; } else *dst++ = *src++; } o->used += o->field[ o->len - 1 ].len = dst - o->field[ o->len - 1 ].val; if ( LFMT_MD == f->mode ) CAT( ". ", 3 ); } /* f->mode */ if ( f->upcase ) { char *p = (char *)o->field[ o->len - 1 ].val; char *e = p + o->field[ o->len - 1 ].len; for ( ; p < e; p++ ) *p = upcase[(unsigned char)*p]; } r->had = 1; if ( iter->submode || ! iter->sub ) break; } if ( iter->submode ) { if ( iter->occ && iter->occ < iter->from ) /* was skipped */ iter->off++; for ( ; iter->pos < r->r->len; iter->pos++, iter->off = 0 ) if ( iter->tag == r->r->field[ iter->pos ].tag || ! iter->tag ) { Field *v = r->r->field + iter->pos; const char *p = v->val + iter->off, *e = v->val + v->len; while ( p < e && ( r->sub != *p++ || e == p || (iter->sub != *p++ && iter->sub != '*') ) ) ; if ( p < e ) { /* hit */ iter->off = (p - v->val) - 2; break; } } } else { if ( iter->occ ) iter->pos++; while ( iter->pos < r->r->len && iter->tag && iter->tag != r->r->field[ iter->pos ].tag ) iter->pos++; iter->off = 0; } iter->occ++; if ( iter->pos >= r->r->len /* end of record */ || (iter->to && iter->to < iter->occ) /* end of selected occ */ ) { iter->occ = -1; break; } } while ( iter->occ <= iter->from ); if ( 0 < iter->occ ) /* we have a next occurence */ r->more = 1; if ( ! r->had && flds && -LFMT_REP == o->field[ o->len - 1 ].tag ) { flds--; o->len--; o->used -= o->field[o->len].len; } r->i++; /* advance to next iterator */ } /* dereference iterator variable */ if ( CLOSE_FIELD > close ) /* leave field to operator */ break; /* close loop */ if ( ! flds ) { /* had no field */ if ( CLOSE_PARAM > close ) /* nothing to do */ break; /* close loop */ } else if ( ! s->expl /* frame is implicit ... */ && ( LFMT_LOOP != s->tok->tag /* ... and not a loop ... */ || CLOSE_EXPL == close /* ... or we saw a hard closer ... */ || (r->i && CLOSE_PARAM == close) /* ... or a loop after 1st iterator on param closer */ ) ) /** upgrade the closing mode to IMPL */ close = CLOSE_IMPL; /* close the field, i.e. coerce it */ dflt = !flds /* param was not given ... */ && (s->pos /* but was started explicitly by a comma like F( 3, ) */ || CLOSE_PARAM == close /* or is closed explicitly like F( ,3 ) */ ); wantnum = 'i' == s->want || 'n' == s->want; gotnum = flds && LFMT_ISNUM( o->field[ o->len-1 ].tag ); LOG_DBG( LFMT_DBG, "\tclose %s %d%c pos %d%c of %s%c%s id %x", closename[close], flds, flds ? gotnum ? 'n' : 's' : dflt ? 'd' : '-', s->pos, s->want, s->tok->name, s->expl?'(':' ', s->tok->sign, s->tok->tag ); /* close and fix last field */ if ( gotnum && !wantnum ) { /* coerce number to string, so it may get concatenated */ char b[32]; int ll = lprint( b, o->field[ o->len-1 ].len ); b[ll++] = ' '; o->len--; ADD( 0, b, ll ); gotnum = 0; } if ( CLOSE_PARAM > close ) break; /* close loop */ /* close and fix param */ if ( wantnum ) { if ( dflt ) { /* default 0 */ NADD( 0, 0 ); flds = 1; } else if ( 1 < flds ) { SERR( "multiple fields for numerical param" ); goto broken; } else if ( flds && !gotnum ) { SERR( "expected number" ); goto broken; } if ( 'i' == s->want ) /* kill decimal */ o->field[s->data].tag &= ~LFMT_VAL; } else { if ( dflt ) { SADD( 0, "" ); flds = 1; } else if ( gotnum ) { SERR( "expected string" ); /* shouldn't happen !? */ goto broken; } else if ( 1 < flds /* concat strings ... */ && ! s->rec /* ... unless in record mode frame (blank,loop,ref) */ ) { int i; for ( i = start+1; i < o->len; i++ ) if ( o->field[i].val ) o->field[ start ].len += o->field[ i ].len; o->len = start + 1; } } /* done fixing param */ if ( CLOSE_IMPL > close ) { tok = NULL; /* eat token */ if ( CLOSE_EXPL > close ) { /* i.e. CLOSE_PARAM */ if ( flds ) { /* increment param pos, set next wanted type */ int sl = strlen( s->tok->sign ); /* frame takes sl-2 params */ if ( ++(s->pos) < sl-2 ) { if ( '_' != s->tok->sign[ 2 + s->pos ] ) /* repeated param */ s->want = s->tok->sign[ 2 + s->pos ]; /* else keep last val */ } else if ( '_' != s->tok->sign[sl-1] ) { /* too much */ SERR2( "expected at most %d params for '%s'", sl-2, s->tok->name ); goto broken; } } break; /* close loop */ } } /* done with token -- close and fix frame */ #define KILLARGS() do { \ o->len = s->data; \ o->used = s->used; \ } while(0) #define KILLOP() do { \ o->len = s->data - 1; \ o->used = s->used - (o->field[o->len].val ? o->field[o->len].len : 0); \ } while(0) switch ( s->tok->tag ) { case LFMT_PLUS: /* arithmetic */ case LFMT_MINUS: case LFMT_MUL: case LFMT_DIV: { lll a = o->field[s->data-1].len; lll b = o->field[s->data].len; /* LOG_DBG( LFMT_DBG, "\tarith on %Ld %Ld", a, b ); */ switch ( s->tok->tag ) { case LFMT_PLUS: a += b; break; case LFMT_MINUS: a -= b; break; case LFMT_MUL: a *= b; break; case LFMT_DIV: if ( 0 != b ) a /= b; break; } KILLOP(); NADD( (int)a, 0 ); } break; case LFMT_F: { char b[32]; int ll = lprint( b, o->field[ s->data ].len ); KILLARGS(); ADD( 0, b, ll ); } break; case LFMT_V: r->iter[ r->i ].tag = o->field[ s->data ].len; /* set iterator field */ o->field[ s->data ].tag = LFMT_ITR | r->iter[ r->i ].tag; o->field[ s->data ].len = r->i; break; case LFMT_INDEX: { LfmtIter * iter = r->iter + r->i; if ( 1 != r->occ ) { SERR( "OOPS! index not within loop" ); goto internalerr; } if ( ! iter->sub || (iter->submode = !iter->from) ) { /* primary loop */ if ( !(iter->from = (int) o->field[ s->data ].len) ) iter->from = 1; iter->to = s->pos ? (int) o->field[ s->data+1 ].len : iter->from; } else { /* additional subfield loop */ if ( !(iter->sfrom = (int) o->field[ s->data ].len) ) iter->sfrom = 1; iter->sto = s->pos ? (int) o->field[ s->data+1 ].len : iter->sfrom; } iter->end = fp + tlen; KILLARGS(); } break; /* case LFMT_INDEX: */ case LFMT_SUBFIELD: if ( 1 != r->occ ) { SERR( "OOPS! subfield not within loop" ); goto internalerr; } r->iter[ r->i ].sub = o->field[ s->data ].val[0]; r->iter[ r->i ].end = fp + tlen; KILLARGS(); break; case LFMT_LOOP: if ( r->more ) { fp = r->loop; /* start over */ tlen = 0; /* care for += tlen */ tok = NULL; r->occ++; r->more = 0; r->i = 0; goto doneclosing; } r->loop = 0; break; } /* close frame, unless it's the outmost */ if ( ! sp ) break; /* TODO: close loop ? */ s = &stack[ --sp ]; start = s->data + s->pos; /* start of current param in output */ flds = o->len - start; /* # fields for current param */ } while ( CLOSE_IMPL == close ); doneclosing: if ( NULL == tok ) continue; if ( ! tok->tag ) goto done; /* add new token */ if ( NULL == r->loop ) /* open loop ? */ switch ( LFMT_GROUP(tok->tag) ) { case LFMT_FUNCTIONS: if ( LFMT_OPEN != tok->tag ) break; goto openloop; /* else should do, but gcc doesn't grok it right */ case LFMT_LITERAL: if ( LFMT_COND != tok->tag && LFMT_REP != tok->tag ) break; case LFMT_ITERATORS: openloop: LOG_DBG( LFMT_DBG, "\topening loop on %s", tok->name ); PUSHFRAME( &loop ); s->rec = stack[sp-1].rec; /* inherit record mode */ r->loop = fp; r->occ = 1; r->more = 0; r->had = 0; r->i = 0; if ( (s->expl = (LFMT_OPEN == tok->tag)) ) { r->loop += tlen; continue; } } if ( tok->sign[2] && LFMT_LITERAL != LFMT_GROUP(tok->tag) ) { /* token opens standard frame */ if ( LFMT_ITERATORS == LFMT_GROUP(tok->tag) ) { if ( 1 != r->occ ) { /* not first run */ if ( r->iter[r->i].def != fp ) { SERR1( "internal error at %d. iterator", r->i + 1 ); goto internalerr; } NADD( r->i, LFMT_ITR | r->iter[r->i].tag ); /* push iterator */ fp = r->iter[r->i].end; /* skip parsing */ tlen = 0; /* care for += tlen */ continue; /* next token */ } if ( LFMT_NITER-1 == r->i ) { SERR1( "maximum number of iterators %d exceeded", LFMT_NITER ); goto overflow; } memset( r->iter + r->i, 0, sizeof(r->iter[0]) ); r->iter[r->i].def = fp; r->iter[r->i].end = fp + tlen; } PUSHFRAME( tok ); switch ( tok->tag ) { /* special treatment */ case LFMT_INDEX: s->expl = 1; break; default: if ( expl ) s->expl = 1; } LOG_DBG( LFMT_DBG, "\topen %s%c%s id %x", s->tok->name, s->expl?'(':' ', s->tok->sign, s->tok->tag ); } if ( lit.tag ) { /* a literal */ if ( NULL != lit.val ) { /* string literal */ switch ( lit.tag ) { case -LFMT_COND: if ( r->i ? !r->more : r->occ > 1 ) continue; break; case -LFMT_REP: if ( r->i && !r->had ) continue; break; } ADD( lit.tag, lit.val, lit.len ); LOG_DBG( LFMT_DBG, "\tlit %s%.*s id %x", tok->name, lit.len, lit.val, tok->tag ); } else { /* numeric literal */ NADD( lit.len, lit.tag ); LOG_DBG( LFMT_DBG, "\tlit %d", lit.len ); } /* new implicit frame immediatly closed by literal ? problem are implicit loops if ( o->len == s->data + 1 && ! s->expl ) { tok = NULL; goto closeframe; } */ } else switch ( tok->tag ) { /* other special tokens */ case LFMT_MDL: f->mode = LFMT_MD; f->upcase = 0; break; case LFMT_MDU: f->mode = LFMT_MD; f->upcase = 1; break; case LFMT_MHL: f->mode = LFMT_MH; f->upcase = 0; break; case LFMT_MHU: f->mode = LFMT_MH; f->upcase = 1; break; case LFMT_MPL: f->mode = LFMT_MP; f->upcase = 0; break; case LFMT_MPU: f->mode = LFMT_MP; f->upcase = 1; break; } } /* for token */ broken: internalerr: overflow: outofmem: done: if ( NULL != o ) { int i = o->len; while ( i-- ) if ( NULL == o->field[i].val ) { ; } else if ( -LFMT_LIT >= o->field[i].tag ) o->field[i].tag = 0; } return o; } /* rFmt */ lio.c0100644000113000011300000006136207665640211010501 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lio.c,v 1.42 2003/05/30 11:29:13 kripke Exp $ I/O support for the openisis library. */ #include #include /* vsnprintf */ #include #include #include #include /* PATH_MAX */ #include #ifdef WIN32 #define WIN32_LEAN_AND_MEAN #define NONAMELESSUNION # include # include # define SHUT_RD SD_RECEIVE # define SHUT_WR SD_SEND # define SHUT_RDWR SD_BOTH # define snprintf _snprintf # define vsnprintf _vsnprintf # include #else /* have operating system */ # define __USE_UNIX98 /* # define _POSIX_MAPPED_FILES # define _POSIX_SYNCHRONIZED_IO */ # include /* or try unixio */ # ifndef O_SYNC # if defined( __FreeBSD__ ) # define O_SYNC O_FSYNC # else # define O_SYNC 0 # endif # endif # include # include /* flock */ # include # include # include #endif /* WIN32 */ #include /* gettimeofday */ #include /* localtime */ #include "luti.h" /* logging */ #include "lses.h" #include "lstr.h" /* log_str */ #include "lio.h" /* ************************************************************ private data */ static int init; #ifdef WIN32 /* it's not thread-safe, anyway */ static HANDLE hand[128]; #endif /* ************************************************************ private functions */ #ifndef WIN32 static int lerrno ( int errno_ ) { switch ( errno_ ) { case EINTR: case EAGAIN: return ERR_OK; case EFAULT: /* structurally bad address */ case ENAMETOOLONG: case ELOOP: return ERR_FAULT; case EBADF: /* file "does not exist" */ case ENOENT: case ENOTDIR: return ERR_BADF; case EIO: /* file "is not accessible" */ case EEXIST: case EISDIR: case EACCES: case ENXIO: case ENODEV: case EROFS: case ETXTBSY: case ENOSPC: case EPIPE: case ESPIPE: return ERR_IO; case ENOMEM: case EMFILE: case ENFILE: return ERR_NOMEM; case EBUSY: return ERR_BUSY; } return ERR_INVAL; } /* lerrno */ /** check unixio errno. if harmless, return != 0 (is ok). else return 0, possibly closing the file */ int unixio_ok ( int *file, int io ) { switch ( errno ) { case EINTR: case EAGAIN: return !0; case EFAULT: case ESPIPE: break; default: log_msg( LOG_IOERR, "closing file %x for %x on %d", *file, io, errno ); lio_close( file, io ); } return 0; } /* unixio_err */ #endif /** return wether printing should be skipped */ static int seterrlev ( int *err, int *level, int code ) { *level = LOG_MASK & code; *err = ERR_MASK & code; if ( ! code ) return 0; if ( ! *err ) switch ( *level ) { case LOG_FATAL: *err = ERR_IDIOT; break; case LOG_ERROR: *err = ERR_INVAL; break; case LOG_IOERR: #ifndef WIN32 if ( EAGAIN == errno || EINTR == errno ) /* no error */ return !0; case LOG_SYSERR: *err = lerrno( errno ); #else case LOG_SYSERR: *err = ERR_INVAL; #endif break; default: *err = 0; } else if ( ! *level ) { if ( ERR_TRASH <= *err ) *level = LOG_FATAL; else if ( ERR_IO <= *err ) *level = LOG_SYSERR; else if ( ERR_FAULT <= *err ) *level = LOG_ERROR; else *level = LOG_VERBOSE; } return *level > (int)log_lev; } /* seterrlev */ /* ************************************************************ package data */ CLockFunc *lio_lock; LogLevel log_lev = LOG_ERROR; #ifndef WIN32 # define LIO_STATICFD( nm, flg ) int nm = flg #else # define LIO_STATICFD( nm, flg ) int nm #endif LIO_STATICFD( lio_in , LIO_IN | 0 ); LIO_STATICFD( lio_out, LIO_OUT | 1 ); LIO_STATICFD( lio_err, LIO_OUT | 2 ); /* ************************************************************ package functions */ static void lio_init () { #ifndef WIN32 /* static */ #else lio_in = lio_open( "CONIN$", LIO_RD ); lio_out = lio_open( "CONOUT$", LIO_WR ); lio_err = lio_open( "oisiserr.txt", LIO_WR|LIO_CREAT ); #endif } /* lio_init */ /* has "official" name, since visible to linker use ld -init openIsisInit for shared object */ void openIsisInit () { static const char inimsg[] = "openIsisInit\n"; const char *ll = getenv("OPENISIS_LOGLEVEL"); if ( init ) return; init = 42; lio_init(); if ( ll ) { cLog( ll[0], 0 ); #ifndef NDEBUG if ( LOG_DEBUG <= log_lev ) lio_write( &lio_err, inimsg, sizeof(inimsg)-1 ); #endif } lses_init(); } /* openIsisInit */ /* use ld -fini openIsisFini for shared object */ void openIsisFini () { static const char finimsg[] = "openIsisFini\n"; lses_fini(); #ifndef NDEBUG if ( LOG_DEBUG <= log_lev ) lio_write( &lio_err, finimsg, sizeof(finimsg)-1 ); #endif } /* openIsisFini */ void linit () /* called by cOpen() */ { if ( init ) /* we're an .so -- implicitly initialized */ return; /* we're statically linked */ openIsisInit(); atexit( openIsisFini ); } int lio_open ( const char *name, int flags ) { int fd = -1; int faillvl = (LIO_TRY == ((LIO_TRY|LIO_WR)&flags)) ? (LOG_VERBOSE|ERR_BADF) : LOG_INFO|ERR_BADF /* LOG_IOERR too annoying ... */; if ( '&' == name[0] ) { int i = 1; fd = 0; while ( '0' <= name[i] && name[i] <= '9' ) fd = 10*fd + name[i++] - '0'; } #ifndef WIN32 if ( 0 > fd ) { int f = !(LIO_WR&flags) ? O_RDONLY : ((LIO_RD&flags) ? O_RDWR : O_WRONLY); if ( LIO_CREAT & flags && LIO_WR & flags ) f |= O_CREAT; if ( LIO_NBLK & flags ) f |= O_NONBLOCK; if ( LIO_SYNC & flags ) f |= O_SYNC; if ( LIO_TRUNC & flags ) f |= O_TRUNC; if ( ! (LIO_SEEK & flags) ) f |= O_APPEND; fd = open( name, f, 00664 ); /* let umask finetune */ if ( LIO_FD < fd ) { log_msg( LOG_ERROR, "got big fd %d", fd ); fd = -1; } if ( 0 > fd ) return log_msg( faillvl, "could not open '%s' %x", name, flags ); if ( LIO_FLOCK & flags ) { const char *lck = LIO_WR&flags ? "exclusive" : "shared"; /* we want an advisory lock, so that tail -f can read changes. SYSV/POSIX fcntl might be mandatory, depending on files mode bits (and filesystem). Therefore prefer BSD style flock, if available. under linux, flock is never mandatory, while lockf/fcntl may be. /usr/src/linux/Documentation/mandatory.txt moreover fcntl attempts to lock over NFS, which is a very feeble idea. */ #ifndef LOCK_SH /* on solaris, LOCK_SH is declared only with the /usr/ucb/cc includes. their flock on fcntl emulation "must not be used in MT environments". */ struct flock fl; memset( &fl, 0, sizeof(fl) ); fl.l_type = LIO_WR&flags ? F_WRLCK : F_RDLCK; fl.l_whence = SEEK_SET; fl.l_start = 0; fl.l_len = 1; /* 0 locks the whole file -- could be mandatory :( */ #endif log_msg( LOG_VERBOSE, "attempting %s lock on '%s'", lck, name ); if ( #ifndef LOCK_SH /* hmm .. at least this is "POSIX" */ fcntl( fd, LIO_WLOCK&flags ? F_SETLKW : F_SETLK, &fl ) #else flock( fd, (LIO_WR&flags ? LOCK_EX : LOCK_SH) | (LIO_WLOCK&flags ? 0 : LOCK_NB) ) #endif ) { close( fd ); return log_msg( ERR_BADF, "could not get %s lock on '%s'", lck, name ); } } } else if ( -1 == fcntl( fd, F_GETFL ) ) /* check open */ return -ERR_BADF; #else if ( 0 > fd ) { HANDLE h; int acc = 0; int shr = FILE_SHARE_READ | (LIO_FLOCK&flags ? 0 : FILE_SHARE_WRITE); int cre = (LIO_CREAT & flags && LIO_WR & flags) ? (LIO_TRUNC & flags) ? CREATE_ALWAYS : OPEN_ALWAYS : (LIO_TRUNC & flags) ? TRUNCATE_EXISTING : OPEN_EXISTING; int f = (LIO_WR&flags) ? FILE_ATTRIBUTE_ARCHIVE : FILE_ATTRIBUTE_NORMAL; if ( LIO_RD & flags ) acc |= GENERIC_READ; /* FILE_READ_DATA; */ if ( LIO_WR & flags ) acc |= GENERIC_WRITE; /* some routines like GetFileSize are "documented" to require GENERIC_READ/GENERIC_WRITE, so it may not be sufficient to use: (LIO_SEEK & flags) ? FILE_WRITE_DATA : FILE_APPEND_DATA */; /* if ( LIO_NBLK & flags ) f |= FILE_FLAG_OVERLAPPED; */ if ( LIO_SYNC & flags ) f |= FILE_FLAG_WRITE_THROUGH; f |= (LIO_SEEK & flags) ? FILE_FLAG_RANDOM_ACCESS : FILE_FLAG_SEQUENTIAL_SCAN; for ( fd=0; hand[fd]; ) if ( sizeof(hand)/sizeof(hand[0]) == ++fd ) return -EMFILE; h = CreateFile( name, acc, shr, 0, cre, f, 0 ); if ( INVALID_HANDLE_VALUE == h ) return log_msg( faillvl, "could not open '%s' %x", name, flags ); hand[fd] = h; if ( (LIO_WR & flags) && ! (LIO_SEEK & flags) ) SetFilePointer( hand[fd], 0, 0, FILE_END ); } else if ( (int)(sizeof(hand)/sizeof(hand[0])) <= fd || ! hand[fd] ) return -ERR_BADF; #endif fd |= flags & LIO_WANT; /* add status flags */ if ( LIO_RD & flags ) fd |= LIO_IN; if ( LIO_WR & flags ) fd |= LIO_OUT; return fd; } /* lio_open */ int lio_close ( int *file, int flags ) { int fd = 0xffff & *file; int op = LIO_INOUT & *file; int cl = LIO_INOUT & flags; if ( ! op || ! cl ) return 0; if ( 0 > *file ) { /* probably failed open */ *file = 0; return 0; } *file &= ~cl; if ( op & ~cl ) { /* remains partly open */ int how = (LIO_IN & cl) ? SHUT_RD : SHUT_WR; if ( !(LIO_SOCK & fd) ) return 0; #ifndef WIN32 if ( ! shutdown( fd, how ) ) return 0; #else /* TODO */ #endif log_msg( LOG_SYSERR, "could not shutdown sock %d %d", fd, how ); return errno ? -errno : -1; } #ifndef WIN32 for ( fsync( fd ); close( fd ); errno = 0 ) if ( EINTR != errno ) { log_msg( LOG_SYSERR, "could not close file %d", fd ); return errno ? -errno : -1; } #else if ( (int)(sizeof(hand)/sizeof(hand[0])) <= fd ) return -ERR_BADF; CloseHandle( hand[fd] ); hand[fd] = 0; #endif return 0; } /* lio_close */ int lio_size ( int file ) { if ( !((LIO_IN|LIO_OUT) & file) ) return log_msg( ERR_BADF, "*file 0x%x not open for stat", file ); { int fd = LIO_FD & file; #ifndef WIN32 struct stat s; return fstat( fd, &s ) ? 0 : s.st_size; #else return GetFileSize( hand[fd], 0 ); #endif } } /* lio_size */ unsigned lio_time ( int file ) { if ( !((LIO_IN|LIO_OUT) & file) ) return log_msg( ERR_BADF, "*file 0x%x not open for stat", file ); { int fd = LIO_FD & file; #ifndef WIN32 struct stat s; return fstat( fd, &s ) ? 0 : (unsigned)s.st_mtime; #else FILETIME foo; /* time since 160101011200 UTC in hundred nanoseconds !!! */ ull bar; if ( ! GetFileTime( hand[fd], 0, 0, &foo ) ) return 0; bar = ((ull)foo.dwHighDateTime<<32 | (ull)foo.dwLowDateTime) / ULL(10000000); return (unsigned)(bar - ULL(11644473600)); #endif } } /* lio_size */ int lio_read ( int *file, void *buf, unsigned count ) { if ( !(LIO_IN & *file) ) return log_msg( ERR_BADF, "*file 0x%x not open for reading", *file ); { int fd = LIO_FD & *file; #ifndef WIN32 int got = read( fd, buf, count ); /* log_msg( LOG_ERROR, "read 0x%x got %d", *file, got ); */ if ( got ) return 0 < got ? got : unixio_ok( file, LIO_IN ) ? 0 : log_msg( LOG_IOERR, "could not read %d bytes from %d", count, fd ); else if ( LIO_SEEK & *file ) /* special EOF treatment */ return 0; else { lio_close( file, LIO_IN ); return -ERR_EOF; } #else DWORD got = 0; int ok = ReadFile( hand[fd], buf, count, &got, 0 ); return ok ? got : log_msg( LOG_IOERR, "could not read %d bytes from %d", count, fd ); #endif } } /* lio_read */ int lio_write ( int *file, const void *buf, unsigned count ) { if ( !(LIO_OUT & *file) ) return log_msg( ERR_BADF, "*file 0x%x not open for writing", *file ); { int fd = LIO_FD & *file; #ifndef WIN32 int got = write( fd, buf, count ); return 0 <= got ? got : unixio_ok( file, LIO_OUT ) ? 0 : log_msg( LOG_IOERR, "could not read %d bytes from %d", count, fd ); #else DWORD got = 0; int ok = WriteFile( hand[fd], buf, count, &got, 0 ); return ok ? got : log_msg( LOG_IOERR, "could not read %d bytes from %d", count, fd ); #endif } } /* lio_write */ int lio_seek ( int *file, int offset ) { if ( !(LIO_INOUT & *file) ) return log_msg( ERR_BADF, "*file 0x%x not open", *file ); { int fd = LIO_FD & *file; #ifndef WIN32 int got = (int)lseek( fd, offset, SEEK_SET ); return offset == got ? 0 : log_msg( LOG_IOERR, "could not seek to %d", offset ); #else return 0xffffffff /* "INVALID_SET_FILE_POINTER" */ == SetFilePointer( hand[fd], offset, 0, FILE_BEGIN ) ? -1 : 0; #endif } } /* lio_seek */ int lio_pread ( int *file, void *buf, unsigned count, int offset ) { if ( !(LIO_IN & *file) ) return log_msg( ERR_BADF, "*file 0x%x not open for reading", *file ); { int fd = LIO_FD & *file; #ifndef WIN32 int got = pread( fd, buf, count, offset ); return 0 <= got ? got : unixio_ok( file, LIO_IN ) ? 0 : log_msg( LOG_IOERR, "could not read %d bytes from %d at %d", count, fd, offset ); #else return 0xffffffff /* "INVALID_SET_FILE_POINTER" */ == SetFilePointer( hand[fd], offset, 0, FILE_BEGIN ) ? -1 : lio_read( file, buf, count ); #endif } } /* lio_pread */ int lio_pwrite ( int *file, const void *buf, unsigned count, int offset ) { if ( !(LIO_OUT & *file) ) return log_msg( ERR_BADF, "*file 0x%x not open for writing", *file ); { int fd = LIO_FD & *file; #ifndef WIN32 int got = pwrite( fd, buf, count, offset ); return 0 <= got ? got : unixio_ok( file, LIO_OUT ) ? 0 : log_msg( LOG_IOERR, "could not read %d bytes from %d at %d", count, fd, offset ); #else return 0xffffffff /* "INVALID_SET_FILE_POINTER" */ == SetFilePointer( hand[fd], offset, 0, FILE_BEGIN ) ? -1 : lio_write( file, buf, count ); #endif } } /* lio_pwrite */ int lio_trunc ( int *file, int offset ) { int fd = LIO_FD & *file; #ifndef WIN32 ftruncate( fd, offset ); #else /* SetFileValidData( hand[fd], offset ); XP only */ return 0xffffffff /* "INVALID_SET_FILE_POINTER" */ == SetFilePointer( hand[fd], offset, 0, FILE_BEGIN ) ? -1 : SetEndOfFile( hand[fd] ) ? 0 : log_msg( LOG_IOERR, "could not truncate %d to %d", fd, offset ); #endif return 0; } /* lio_trunc */ int lio_mmap ( int *file, void **map, int length ) { if ( ! map ) return log_msg( ERR_INVAL, "no map" ); if ( 0 >= length && 0 >= (length = lio_size( *file )) ) { *map = 0; return length; } #ifndef WIN32 #if 0 /* solaris states it will round up itself, and linux does */ { static int ps = 0; if ( !ps ) ps = sysconf(_SC_PAGESIZE); /* getpagesize(); */ } #endif if ( file ) { if ( *map ) return msync( *map, length, (LIO_SYNC & *file) ? (MS_SYNC|MS_INVALIDATE) : MS_ASYNC ) ? log_msg( LOG_IOERR, "msync" ) : 0; *map = mmap( 0, length, PROT_READ | ((LIO_OUT&*file) ? PROT_WRITE : 0), MAP_SHARED, LIO_FD&*file, 0 ); if ( MAP_FAILED != *map ) return length; *map = 0; return log_msg( ERR_NOMEM, "mmap failed on fd %08x", *file ); } if ( ! *map ) return log_msg( ERR_INVAL, "no *map" ); munmap( *map, length ); *map = 0; return 0; #else /* awfully complicated here -- need CreateFileMapping, MapViewOfFile stupid piece of shrott true mapping supported in NT family only, 9x copies to swap */ *map = 0; return 0; #endif } /* lio_mmap */ int lio_slurp ( char **buf, int sz, const char *name, int opt ) { int file = lio_open( name, LIO_RD|LIO_SEEK|(opt ? LIO_TRY : 0) ); int size; int ret = -ERR_NOMEM; char *p = *buf; /* log_msg( LOG_IOERR, "open '%s' = %d", name, file ); */ if ( 0 > file ) return file; size = lio_size( file ); if ( 0 >= size ) return size; if ( size > sz ) LOG_OTO( done, ( ERR_INVAL, "file '%.30s' too big: %d > %d", name, size, sz ) ); if ( ! p && !(p = mAlloc(size)) ) goto done; ret = lio_read( &file, p, size ); if ( size == ret ) *buf = p; else { if ( ret >= 0 ) ret = log_msg( ERR_IO, "OOPS! got %d of %d bytes from '%.30s'", ret, size, name ); if ( ! *buf ) free( p ); } done: lio_close( &file, LIO_INOUT ); return ret; } /* lio_slurp */ int log_msg ( int code, const char *fmt, ... ) { static const char toolong[] = ": message too long !!!\n"; int err, level; va_list ap; va_start( ap, fmt ); #if 0 ret = sMsg( MSG_VA|code, fmt, ap ); make log_msg unbuffered, use sMsg for buffered logging #else if ( ! seterrlev( &err, &level, code ) && (LIO_OUT & lio_err) ) { char buf[4096]; int len = vsnprintf( (char*)buf, sizeof(buf), fmt, ap ); if ( 0 < len && len < (int)sizeof(buf) ) { buf[len++] = '\n'; lio_write( &lio_err, buf, len ); } else { lio_write( &lio_err, fmt, strlen(fmt) ); lio_write( &lio_err, toolong, sizeof(toolong)-1 ); } } #endif va_end( ap ); return -err; } /* log_msg */ void log_str ( LogLevel level, int *rec, const char **desc ) { int occ = -1; int nmbrs = LSTRFIX(*rec); int *mbr = rec+1; char *base = (char*)rec; if ( level > log_lev || ! desc ) return; sMsg( 2, "record %.20s\n", *desc++ ); /* dump the fixed part (occ==-1) and each occurrence of repeated part. */ for ( ;/* occ < LSTROCC(*dst) */; ) { /* dump one part */ int i; for ( i=0; i= LSTROCC(*rec) ) break; if ( ! occ ) { /* was the fixed part, setup for repeated */ nmbrs = LSTRREP(*rec); desc += i; } } } /* log_str */ void log_hex ( LogLevel level, const void *mem, int len ) { const char *p = (const char *)mem; char buf[82]; int i = 0; if ( level > log_lev ) return; for ( ; i 16 ) left = 16; for ( ; j < left; j++ ) { sprintf( buf+pos, "%02x", p[j] ); pos += 2; if ( 3 == j%4 ) { buf[pos++] = ' '; buf[pos++] = ' '; } } for ( ; j < 16; j++ ) { buf[pos++] = ' '; buf[pos++] = ' '; if ( 3 == j%4 ) { buf[pos++] = ' '; buf[pos++] = ' '; } } /* got 50 = 10 + 16*2 + 4*2 */ for ( j=0; j < left; j++ ) /* add up to 16 */ buf[pos++] = (0x60 & p[j]) ? p[j] : '.'; buf[pos++] = '\n'; buf[pos] = 0; sMsg( 2, "%.*s", pos, buf ); } } /* log_hex */ /* ************************************************************ public functions */ int timeUpd ( Tm *tm ) { Tm o; #ifdef WIN32 struct _timeb tb; _ftime( &tb ); if ( !tm ) return tb.time; o = *tm; tm->millis = tb.time*LLL(1000) + tb.millitm; #else struct timeval tv; gettimeofday( &tv, 0 ); if ( !tm ) return tv.tv_sec; o = *tm; tm->millis = tv.tv_sec*LLL(1000) + tv.tv_usec/1000; #endif return (int)(tm->millis - o.millis); } /* timeUpd */ static int timeLoc ( struct tm *t, Tm *tm ) { Tm x; time_t tt; if ( !tm || !tm->millis ) timeUpd( tm ? tm : (tm = &x) ); tt = (time_t)(tm->millis / 1000); #ifdef WIN32 /* did I mention it's not threadsafe ? */ *t = *localtime( &tt ); #else localtime_r( &tt, t ); #endif return (int)(tm->millis % 1000); } /* timeLoc */ char *timeGtf ( char *buf, Tm *tm ) { struct tm t; timeLoc( &t, tm ); snprintf( buf, 15, "%04u%02u%02u%02u%02u%02u", 1900+t.tm_year, 1+t.tm_mon, t.tm_mday, t.tm_hour, t.tm_min, t.tm_sec ); buf[14] = 0; return buf; } /* timeGtf */ char *timeGtfm ( char *buf, Tm *tm ) { struct tm t; int millis = timeLoc( &t, tm ); snprintf( buf, 19, "%04u%02u%02u%02u%02u%02u%03u", 1900+t.tm_year, 1+t.tm_mon, t.tm_mday, t.tm_hour, t.tm_min, t.tm_sec, millis ); buf[18] = 0; return buf; } /* timeGtfm */ void timeSleep ( Tm *tm ) { #ifdef WIN32 Sleep( tm->millis ); #else struct timespec ts; ts.tv_sec = tm->millis / 1000; ts.tv_nsec = 1000000 * (int)(tm->millis % 1000); nanosleep( &ts, 0 ); #endif } /* timeSleep */ int ioStream ( Ios *s, int op ) { Buf *b; switch ( op ) { case LIO_SSIZE: return sizeof(Ios); case LIO_SOPEN: s->pos = s->b.fill = s->b.done = 0; case LIO_SCLOSE: s->file &= ~LIO_INOUT; return 0; case LIO_SPURGE: if ( !(LIO_IN & s->file) ) return -ERR_BADF; /* try to get 4k chunk */ b = /* s->alt ? s->alt : */ &s->b; if ( LIO_BUFSIZ/2 < b->fill && b->done ) { /* move contents downwards */ if ( b->done < b->fill ) memmove( b->c, b->c + b->done, b->fill - b->done ); s->pos += b->done; b->fill -= b->done; b->done = 0; } return LIO_BUFSIZ <= b->fill ? 0 : LIO_BUFSIZ/2 < b->fill ? LIO_BUFSIZ - b->fill : LIO_BUFSIZ/2; case LIO_SFILL: case LIO_SFLUSH: s->file &= ~LIO_INOUT; } return -ERR_EOF; } /* ioStream */ int ioStdio ( Ios *s, int op ) { Buf *b; int ret; switch ( op ) { case LIO_SOPEN: ioStream( s, op ); if ( 0 < (ret = lio_open( s->name, s->file )) ) s->file = ret; return ret; case LIO_SCLOSE: if ( LIO_OUT & s->file ) s->func( s, LIO_SFLUSH ); /* don't close borrowed fd */ return '&' == *s->name ? 0 : lio_close( &s->file, LIO_INOUT ); case LIO_SFILL: if ( !(LIO_IN & s->file) ) return -ERR_BADF; if ( 0 < (ret = ioStream( s, LIO_SPURGE )) && 0 <= (ret = lio_read( &s->file, s->b.c + s->b.fill, ret )) ) s->b.fill += ret; log_msg( LOG_ALL, "LIO_FILL: got %d now %d", ret, s->b.fill ); return ret; case LIO_SFLUSH: if ( !(LIO_OUT & s->file) ) return -ERR_BADF; b = /* s->alt ? s->alt : */ &s->b; if ( b->fill <= b->done ) return 0; if ( 0 < (ret = lio_write( &s->file, b->c + b->done, b->fill - b->done )) ) b->done += ret; if ( b->fill == b->done ) { s->pos += b->done; b->fill = b->done = 0; } else if ( b->done && LIO_BUFSIZ/2 < b->fill ) ioStream( s, LIO_SPURGE ); return ret; } return ioStream( s, op ); } /* ioStdio */ int sMsg ( int to, const char *fmt, ... ) { SESDECL int code = (LOG_MASK|ERR_MASK) & to, err, level; int fd = LIO_FD & to; /* stream id uses same mask as system fd */ Ios *o = ses->io[ fd ? fd : code ? 2 : 1 ]; va_list ap; int space; int ret; if ( LSES_FILE_MAX <= fd || ! o ) return -1; if ( seterrlev( &err, &level, code ) ) /* logging */ return -err; /* core fprintf */ if ( LIO_BUFSIZ/2 < o->b.fill ) LIO_FLUSH( o ); space = LIO_BUFSIZ - o->b.fill; va_start( ap, fmt ); ret = vsnprintf( (char*)o->b.c + o->b.fill, space, fmt, (MSG_VA & to) ? va_arg( ap, va_list ) : ap ); va_end( ap ); if ( ret < 0 || space < ret ) /* outta space */ ret = space; o->b.fill += ret; if ( ! code ) return ret; /* logging afterburner */ if ( LOG_SYSERR == level || LOG_IOERR == level ) { int len; char *syserr = #ifndef WIN32 strerror(errno); #else 0; char buf[256] = ""; if ( FormatMessage( FORMAT_MESSAGE_FROM_SYSTEM|FORMAT_MESSAGE_IGNORE_INSERTS, 0, GetLastError(), 0, buf, sizeof(buf)-1, 0 ) ) syserr = buf; buf[ sizeof(buf)-1 ] = 0; #endif if ( syserr ) { len = strlen( syserr ); if ( o->b.fill+4+len < LIO_BUFSIZ ) { memcpy( o->b.c + o->b.fill, "\n\t: ", 4 ); memcpy( o->b.c + o->b.fill+4, syserr, len ); o->b.fill += 4+len; } } } if ( o->b.fill < LIO_BUFSIZ ) o->b.c[ o->b.fill++ ] = '\n'; LIO_FLUSH( o ); return -err; } /* sMsg */ int sGets ( int fd, char **ptr, char delim ) { SESDECL Ios *s = ses->io[ LIO_FD & fd ]; if ( s->b.fill > s->b.done || 0 < LIO_FILL( s ) ) for (;;) { char *f = memchr( s->b.c + s->b.done, delim, s->b.fill - s->b.done ); int len; log_msg( LOG_ALL, "sGets: done %d fill %d", s->b.done, s->b.fill ); f = memchr( s->b.c + s->b.done, delim, s->b.fill - s->b.done ); if ( ! f ) { if ( (s->b.fill < LIO_BUFSIZ || s->b.done) /* buffer avail */ && (LIO_IN & s->file) /* file open */ && 0 < LIO_FILL( s ) ) continue; else if ( s->b.done >= s->b.fill ) break; f = (char*)s->b.c + s->b.fill; } len = f - (*ptr = (char*)s->b.c + s->b.done); s->b.done += len + 1; if ( len && '\n' == delim && '\r' == (*ptr)[len-1] ) len--; return len; } *ptr = 0; return -1; } /* sGets */ void cLog ( int level, const char *filename ) { if ( 0 > level ) /* no change */ ; else if ( LOG_LEVELS > level ) /* by basic number */ level <<= LOG_SHIFT; else if ( 'z' >= level ) { /* by ascii value */ switch ( level ) { case '-': level = LOG_NOCHANGE; break; case 'o': level = LOG_OFF; break; case 'f': level = LOG_FATAL; break; case 's': level = LOG_IOERR /*LOG_SYSERR*/; break; default: if ( '0' <= level && level <= '9' ) { level = (level - '0')<= 0 ) log_lev = (LogLevel)level; (void)filename; /* TODO if ( filename ) sOpen( cOpen(0), "2>"filename, 0, ioStdio ) */ } /* cLog */ lio.h0100644000113000011300000002142107665600575010507 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LIO_H /* $Id: lio.h,v 1.22 2003/05/29 18:02:54 kripke Exp $ I/O support for the openisis library. Provides non-blocking, look-ahead network streams and other nice stuff. */ #include "luti.h" #include "lll.h" /* ************************************************************ time */ /* struct wrapper just for struct Tm *Time */ typedef struct Tm { lll millis; } Tm; /** set Tm to current time, return difference in millis */ extern int timeUpd ( Tm * ); /** print generalized time yyyyMMddHHmmss + 0 byte to buffer if Tm is 0, current time is used if Tm.millis is 0, Tm is updated return buffer, which must have 15 bytes */ extern char *timeGtf ( char *buf, Tm * ); /** like timeAsc, but with additional 3 digits millis return buffer, which must have 18 bytes */ extern char *timeGtfm ( char *buf, Tm * ); /** nanosl */ extern void timeSleep ( Tm * ); /* ************************************************************ system files */ /** system file is a non-negative 32bit int including status bits. main status bits are LIO_IN and LIO_OUT, which are set by open and cleared by close. negative values are reserved for use by non-system based streams. */ #define LIO_WANT 0xfff00000 /* what we want: open flags */ #define LIO_FD 0x0000ffff /* what we get: system fd */ #define LIO_STAT 0x000f0000 /* how it's going: it's status */ /* status flags */ #define LIO_IN 0x00010000 /* support input */ #define LIO_OUT 0x00020000 /* support output */ #define LIO_INOUT 0x00030000 /* support both */ #define LIO_ISOPEN( file ) (LIO_INOUT & (file)) /* common requirements that are also used by non-system streams. */ #define LIO_RD 0x00100000 /* shall be opened for input */ #define LIO_WR 0x00200000 /* shall be opened for output */ #define LIO_RDWR 0x00300000 /* shall be opened for both */ #define LIO_SYNC 0x00400000 /* with WR: syncing output */ #define LIO_NBLK 0x00800000 /* non blocking IO (not used) */ /* system files only. */ #define LIO_CREAT 0x01000000 /* with WR: shall be created */ #define LIO_TRY 0x01000000 /* w/o WR: do not complain if open fails */ #define LIO_TRUNC 0x02000000 /* with WR: shall be truncated */ #define LIO_SEEK 0x04000000 /* random access (with WR: else append mode) */ #define LIO_SOCK 0x08000000 /* is a socket (support shutdown) */ /* locking */ #define LIO_TLOCK 0x10000000 /* try locking (EX with WR) */ #define LIO_WLOCK 0x20000000 /* lock waiting (EX with WR) */ #define LIO_FLOCK 0x30000000 /* any locking is set */ /** create a new fid based on name and flags. name may be &i to use fd i, esp. for i=0,1 or 2. @return non-negative file or some error code */ extern int lio_open ( const char *name, int flags ); extern int lio_close ( int *file, int flags ); extern int lio_size ( int file ); extern unsigned lio_time ( int file ); /* mtime sec */ extern int lio_in; extern int lio_out; extern int lio_err; /* Like the syscalls, this returns the number of bytes on success. Unlike the syscalls, it returns 0 rather than an error when no bytes are available after interrupt (EINTR) or on non-blocking IO (EAGAIN); i.e. you may try later. On error, a negative value is returned. On errors that render the file unusable, it is closed for input or output, resp. These are most errors but EFAULT (bad buf) and ESPIPE (which clears the SEEK bit). For files that do not have the SEEK bit set, EOF is considered such an error and -LERR_EOF is returned. If you want to stream a disk file like 'tail -f', set the SEEK bit. */ extern int lio_read ( int *file, void *buf, unsigned count ); extern int lio_write ( int *file, const void *buf, unsigned count ); extern int lio_pread ( int *file, void *buf, unsigned count, int offset ); extern int lio_pwrite ( int *file, const void *buf, unsigned count, int offset ); extern int lio_seek ( int *file, int offset ); extern int lio_trunc ( int *file, int offset ); /** open, sync or close a memory mapping @param file pointer to open file handle to map. if file is NULL, *map is unmapped (and set to NULL) else if *map is NULL, length bytes are mapped else *map is synced (flushed) @param length number of bytes to map; if 0, size will be used @return mapped length; if <= 0, *map is set to 0, else to memory region */ extern int lio_mmap ( int *file, void **map, int length ); /** slurp in a whole file at once. @param buf points to buffer of size sz. lio_slurp will allocate one, if *buf is NULL @param sz maximum number of bytes to read @param name of file to slurp @param opt if != 0, do not complain on failure @return number of bytes read or negative on error */ extern int lio_slurp ( char **buf, int sz, const char *name, int opt ); /* ************************************************************ streams */ /** an I/O buffered stream similar to stdio's FILE. Unlike FILE with it's stupid ungetc, it supports efficient scanning of the whole input buffer without actually consuming it. Also unlike at least Linux' glibc's stdio, access is not interlocked. The buffer is typically used for input OR output. However, for internal pipes, the same buffer is used from both ends. High level access functions like printf operate directly on the buffer and call the stream function on buffer underflow or overflow (or explicit flush). */ typedef int lio_sfunc ( struct Ios *s, int op ); enum { LIO_BUFSIZ = 8192 }; typedef struct Buf { /* struct Buf *nxt; chain -- not used */ short fill; /* total #bytes in the buffer */ short done; /* #bytes done */ unsigned char c[LIO_BUFSIZ]; /* actual buffer may be longer */ } Buf; /* fill - done is the number of available bytes */ #define LIO_BAVAIL( b ) ((b)->fill - (b)->done) #define LIO_BINIT( b ) do { \ /*(b)->nxt = 0;*/ (b)->fill = 0; (b)->done = 0; \ } while (0) #define LIO_BINITIALIZER { 0,0,"" } typedef struct Ios { lio_sfunc *func; const char *name; /* stream's name */ int file; /* file "id" and flags as above */ int pos; /* logical position of buffer start within stream */ Buf b; } Ios; #define LIO_SISOPEN( s ) LIO_ISOPEN( (s)->file ) #define LIO_SAVAIL( s ) LIO_BAVAIL( &(s)->b ) #define LIO_SINIT( s, fun, nam, fil ) do { \ (s)->func = fun; (s)->name = nam; (s)->file = fil; \ (s)->pos = 0; \ LIO_BINIT( &(s)->b ); \ } while (0) #define LIO_STDINIT( s, nam, fil ) LIO_SINIT( s, lio_stdio, nam, fil ) #define LIO_SINITIALIZER( fun, nam, fil ) { fun, nam, fil, 0, LIO_BINITIALIZER } #define LIO_STDINITIALIZER( nam, fil ) LIO_SINITIALIZER( lio_stdio, nam, fil ) /** stream operations. */ enum { LIO_SSIZE, /* tell the size of the stream structure */ LIO_SOPEN, /* initialize the structure */ LIO_SCLOSE, /* close */ LIO_SPURGE, /* kill the done bytes. return fill buffer size */ LIO_SFILL, /* try to refill the buffer (poll input). return # new bytes */ LIO_SFLUSH, /* try to flush the buffer. return # bytes written */ LIO_SPUSH /* push input stream: munge prefilled input */ }; /** "abstract" base implementation of stream. not very useful in itself, but used by derived impl. */ extern int ioStream ( Ios *s, int op ); /** stream func based on system file to mimic stdio. on LIO_SOPEN, the fid is set by lio_open based on name. other ops used the correponding lio system call. */ extern int ioStdio ( Ios *s, int op ); #define LIO_OPEN( s ) (s)->func( (s), LIO_SOPEN ) #define LIO_CLOSE( s ) (s)->func( (s), LIO_SCLOSE ) #define LIO_FILL( s ) (s)->func( (s), LIO_SFILL ) #define LIO_FLUSH( s ) (s)->func( (s), LIO_SFLUSH ) /* ************************************************************ inter process communication */ extern CLockFunc *lio_lock; #define LIO_LOCK() (NULL != lio_lock && lio_lock(OPENISIS_LOCK)) #define LIO_RELE() (NULL != lio_lock && lio_lock(OPENISIS_RELE)) #define LIO_WAIT( c ) (NULL != lio_lock && lio_lock(OPENISIS_WAIT \ | (OPENISIS_COND & (c)))) #define LIO_WAKE( c ) (NULL != lio_lock && lio_lock(OPENISIS_WAKE \ | (OPENISIS_COND & (c)))) #define LIO_H #endif /* LIO_H */ lll.h0100644000113000011300000000304007665640164010502 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LLL_H /* $Id: lll.h,v 1.3 2003/05/30 11:28:52 kripke Exp $ 64 bit integers */ #ifdef __GNUC__ # define LLL_BUILTIN long long # define LLL( v ) v##LL # define ULL( v ) v##ULL #elif defined( _MSC_VER ) || defined( __BORLANDC__ ) # define LLL_BUILTIN __int64 # define LLL( v ) v /* is there some special suffix ??? */ # define ULL( v ) v #elif 0xFFFFFFFFL != ~0L # define LLL_BUILTIN long # define LLL( v ) v##L # define ULL( v ) v##UL #endif #ifndef LLL_BUILTIN # error "please use a compiler providing a 64 bit integer type. thanks." #endif typedef LLL_BUILTIN lll; typedef unsigned LLL_BUILTIN ull; #define LLL_H #endif /* LLL_H */ loi.h0100644000113000011300000001740607664636753010525 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LOI_H /* $Id: loi.h,v 1.30 2003/05/27 10:28:59 kripke Exp $ lib openisis: provides the "openisis namespace", i.e. short names for most openIsis names, so we don't get to dislike those letters ... */ #include #define OPENISIS_NOPRE09 #include "openisis.h" #define Fld OpenIsisFld #define Field OpenIsisField #define Raw OpenIsisRaw #define Rec OpenIsisRec #define BASESZ OPENISIS_BASESZ #define INITREC OPENISIS_INITREC #define INITSTD OPENISIS_INITSTD #define INITBUF OPENISIS_INITBUF #define CLRREC OPENISIS_CLRREC #define RECOK OPENISIS_RECOK #define rGet openIsisRGet #define rOccurence openIsisROccurence #define rInt openIsisRInt #define rInt2 openIsisRInt2 #define rEnum openIsisREnum #define rString openIsisRString #define rString2 openIsisRString2 #define a2i openIsisA2i #define a2il openIsisA2il #define a2id openIsisA2id #define i2a openIsisI2a #define u2a openIsisU2a #define rSer openIsisRSerialize #define rSerA openIsisRSerializeAlloc #define rSerB openIsisRSerializeBin #define rDeser openIsisRDeserialize #define FTX OPENISIS_FTX #define FTA OPENISIS_FTA #define FTN OPENISIS_FTN #define FTP OPENISIS_FTP #define FTI OPENISIS_FTI #define FTE OPENISIS_FTE #define FTB OPENISIS_FTB #define FTT OPENISIS_FTT #define FTS OPENISIS_FTS #define FTF OPENISIS_FTF #define FTV OPENISIS_FTV #define NOENUM OPENISIS_NOENUM #define FD_NAMELEN OPENISIS_FD_NAMELEN #define Fd OpenIsisFd #define Fdt OpenIsisFdt #define fById openIsisFById #define fByName openIsisFByName #define fEnum openIsisFEnum #define fFree openIsisFFree #define fFdt2Rec openIsisFFdt2Rec #define fRec2Fdt openIsisFRec2Fdt #define Ses OpenIsisSes #define Session OpenIsisSession /* Ses* */ #define mAlloc openIsisMAlloc #define mFree openIsisMFree #define mDup openIsisMDup #define toHtml openIsisToHtml #define utf8Chk openIsisUtf8Chk #define Stb OpenIsisStb #define Stub OpenIsisStub /* Stb* */ #define Db OpenIsisDb #define Schema OpenIsisSchema #define DBNLEN OPENISIS_DB_NAMELEN #define SCNLEN OPENISIS_SC_NAMELEN #define DBIMSK OPENISIS_DBIDMSK #define SCIMSK OPENISIS_SCIDMSK #define nInit openIsisNInit #define nDeinit openIsisNDeinit #define nOpen openIsisNOpen #define nClose openIsisNClose #define nSchema openIsisNSchema #define nSend openIsisNSend #define nRecv openIsisNRecv #define nClean openIsisNClean #define nResult openIsisNGetResult #define nNtfRsp openIsisNNotifyResponse #define nNtfErr openIsisNNotifyError #define nChannel openIsisNGetChannel #define nDbByName openIsisNDbByName #define nDbById openIsisNDbById #define Ios OpenIsisIos #define Stream OpenIsisStream /* Ios* */ #define SFunc OpenIsisSFunc #define ERR_OK OPENISIS_ERR_OK #define ERR_EOF OPENISIS_ERR_EOF #define ERR_FAULT OPENISIS_ERR_FAULT #define ERR_INVAL OPENISIS_ERR_INVAL #define ERR_BADF OPENISIS_ERR_BADF #define ERR_IO OPENISIS_ERR_IO #define ERR_NOMEM OPENISIS_ERR_NOMEM #define ERR_BUSY OPENISIS_ERR_BUSY #define ERR_TRASH OPENISIS_ERR_TRASH #define ERR_IDIOT OPENISIS_ERR_IDIOT #define ERR_ERRORS OPENISIS_ERR_ERRORS #define ERR_MASK OPENISIS_ERR_MASK #define ERR_SHIFT OPENISIS_ERR_SHIFT #define LOG_OFF OPENISIS_LOG_OFF #define LOG_FATAL OPENISIS_LOG_FATAL #define LOG_SYSERR OPENISIS_LOG_SYSERR #define LOG_IOERR OPENISIS_LOG_IOERR #define LOG_ERROR OPENISIS_LOG_ERROR #define LOG_WARN OPENISIS_LOG_WARN #define LOG_INFO OPENISIS_LOG_INFO #define LOG_VERBOSE OPENISIS_LOG_VERBOSE #define LOG_TRACE OPENISIS_LOG_TRACE #define LOG_DEBUG OPENISIS_LOG_DEBUG #define LOG_ALL OPENISIS_LOG_ALL #define LOG_NOCHANGE OPENISIS_LOG_NOCHANGE #define LOG_LEVELS OPENISIS_LOG_LEVELS #define LOG_MASK OPENISIS_LOG_MASK #define LOG_SHIFT OPENISIS_LOG_SHIFT #define MSG_VA OPENISIS_MSG_VA #define LogLevel OpenIsisLogLevel #define sMsg openIsisSMsg #define sOpen openIsisSOpen #define sGets openIsisSGets #define sReadln openIsisSReadln #define sGetr openIsisSGetr #define sSet openIsisSSet #define sGet openIsisSGet #define rDup openIsisRDup #define rMsg openIsisRMsg #define rFmt openIsisRFmt #define rSplitf openIsisRSplitf #define rSet openIsisRSet #define rAddI openIsisRAddI #define RSPACE OPENISIS_RSPACE #define RADD OPENISIS_RADD #define RCAT OPENISIS_RCAT #define RADDS OPENISIS_RADDS #define RADDF OPENISIS_RADDF #define RCATS OPENISIS_RCATS #define RARGC OPENISIS_RARGC #define RARGV OPENISIS_RARGV #define RFDT OPENISIS_RFDT #define RDIS OPENISIS_RDIS #define RNOC OPENISIS_RNOC #define RIGN OPENISIS_RIGN #define RCHG OPENISIS_RCHG #define RDEL OPENISIS_RDEL #define RDFLT OPENISIS_RDFLT #define ROP OPENISIS_ROP #define Set OpenIsisSet #define Val OpenIsisVal #define Key OpenIsisKey #define Hit OpenIsisHit #define XCTL OPENISIS_XCTL #define XFST OPENISIS_XFST #define XHIT OPENISIS_XHIT #define XADD OPENISIS_XADD #define XDEL OPENISIS_XDEL #define dRead openIsisDRead #define dRaw openIsisDRaw #define dFmt openIsisDFmt #define dMaxId openIsisDMaxId #define dWrite openIsisDWrite #define dWritex openIsisDWritex #define dScan openIsisDScan #define dTerm openIsisDTerm #define dQuery openIsisDQuery #define DXCb OpenIsisDXCb #define DXLoop OpenIsisDXLoop #define IDXEQ OPENISIS_IDXEQ #define IDXPF OPENISIS_IDXPF #define IDXUPTO OPENISIS_IDXUPTO #define IDXINCL OPENISIS_IDXINCL #define IDXMODE OPENISIS_IDXMODE #define dXLoop openIsisDXLoop #define Idx OpenIsisIdx #define Index OpenIsisIndex /* Idx* */ #define CLockFunc OpenIsisCLockFunc #define cOpen openIsisCOpen #define cInit openIsisCInit #define cLog openIsisCLog #define cSession openIsisCSession #define cDOpen openIsisCDOpen #define cDOpenv openIsisCDOpenv #define cDCheck openIsisCDCheck #define cDClose openIsisCDClose #define cXOpen openIsisCXOpen #define cXMkVal openIsisCXMkVal #define cXAdd openIsisCXAdd #define cXClose openIsisCXClose #define stub0 openisis_stub0 #define ses openisis_ses /* the following are not yet blessed by an official OpenIsis prefix */ /* lio */ #define Buf LioBuf #define ioStream lio_stream #define ioStdio lio_stdio /* lsv */ #define Que LsvQue #define Pool LsvPool #define Wrk LsvWrk #define Srv LsvSrv #define Con LsvCon #define svCur lsv_cur #define svPlain lsv_plain #define svEcho lsv_echo #define svRun lsv_run #define LOI_H #endif /* LOI_H */ lqry.c0100644000113000011300000001326207644412744010706 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lqry.c,v 1.7 2003/04/08 00:20:52 kripke Exp $ implementation of record cooking. */ #include /* memset et al */ #include "ldb.h" /* ************************************************************ private types */ /* ************************************************************ private data */ static unsigned char op[256]; /* ************************************************************ private functions */ static int lqry_scan ( OpenIsisSet *set, int db, const char *key, int mode, int skip ) { (void)set;(void)db;(void)key;(void)mode;(void)skip; log_msg( LOG_ERROR, "scan mode not implemented yet, sorry!" ); return -1; } /* lqry_scan */ /* ************************************************************ package functions */ /* ************************************************************ public functions */ int dQuery ( Set *set, int db, const char *key, int mode, int skip ) { LdbPost p; int ret = 0; memset( &p, 0, sizeof(p) ); p.skp = skip; p.tag = mode >> 16; mode &= 0xff; if ( OPENISIS_QRY_SIMPLE > mode ) { char *tmp = (char*)key; if ( OPENISIS_QRY_SCANE <= mode ) return lqry_scan( set, db, key, mode, skip ); switch ( (int)mode ) { case OPENISIS_QRY_KEYEQ: break; case OPENISIS_QRY_KEYAT: { size_t l = strlen(key); if ( !l || '$' != key[l-1] ) break; tmp = mAlloc( l ); if ( ! tmp ) return -ERR_NOMEM; memcpy( tmp, key, l-1 ); tmp[l-1] = 0; } case OPENISIS_QRY_KEYPF: p.mode = LDB_PFX; } ret = ldb_search( db, tmp, &p, 0 ); if ( tmp != key ) mFree( tmp ); /* } else if ( OPENISIS_QRY_SIMPLE < mode ) { */ } else { /* handling query expression (simple) */ size_t klen = strlen(key); unsigned char *buf, *uc, *term, *end, utmp; if ( ! op['.'] ) /* operators */ op['.'] = op['$'] = op['('] = op['*'] = op['+'] = op['^'] = op['/'] = 1; buf = (unsigned char*)mAlloc( klen+1 ); if ( ! buf ) return -ERR_NOMEM; /* startover: */ memcpy( buf, key, klen+1 ); /* reload, might have frobed the buf on \ */ uc = buf; p.mode = LDB_OR; nextterm: while ( *uc && ' ' >= *uc ) uc++; /* skip white */ if ( !*uc ) goto done; /* take as term, whatever it is ... */ term = uc; if ( '"' == *term ) { /* anything up to next unescaped " */ int bs = 0; /* count of backslashes seen so far */ term++; while ( *++uc && '"' != *uc ) { if ( '\\' == *uc ) { if ( ! *++uc ) break; /* trailing \ */ bs++; } if ( bs ) uc[-bs] = *uc; } end = uc++ - bs; if ( '$' == end[-1] ) { p.mode |= LDB_PFX; end--; } } else { while ( *++uc && !op[*uc] && (' '<*uc || (uc==term+3 && !memcmp("ANY",term,3))) ); /* find white or op */ end = uc; if ( '$' == *uc ) { p.mode |= LDB_PFX; uc++; } } /* now end is on 1st char after term, possibly on a $ or ". uc is on next pos to scan. */ /* check for tag */ p.tag = 0; if ( '/' == *uc ) { int parenths = '(' == *++uc; if ( parenths ) uc++; /* accept tag w o w/o () */ /* eat arbitrary large numbers :)) */ while ( '0'<=*uc && *uc<='9' ) p.tag = 10*p.tag + *uc++ - '0'; if ( parenths ) while ( *uc && ')'!=*uc++ ); /* mv behind ) */ } /* do it */ utmp = *end; *end = 0; /* terminate term */ if ( !(p.mode & (LDB_AND|LDB_NOT)) ) p.near = -1; /* collect all pos */ else if ( !p.fil ) /* nothing to AND */ goto operator; if ( 0 > (ret = ldb_search( db, (const char*)term, &p, 0 )) ) goto done; sMsg( LOG_INFO, "src %d '%.30s'@%d %hd(%hd): fill %d cut %d", p.skp, term, p.tag, p.mode, p.near, p.fil, p.cut ); operator: *end = utmp; /* restore */ /* look for next operator */ p.mode = LDB_AND; p.near = 0; while ( *uc && ' ' >= *uc ) uc++; /* skip white */ if ( !*uc ) goto done; switch ( *uc ) { case '*': uc++; break; case '^': p.mode = LDB_NOT; uc++; break; case '+': p.mode = LDB_OR; uc++; break; case '.': while ( '.' == *uc++ ) p.near++; break; case '$': while ( '$' == *uc++ ) p.near--; break; case '(': if ( ! uc[1] ) goto done; if ( ')' != uc[2] ) break; if ( 'F' == uc[1] ) p.near = LDB_NEAR_F; if ( 'G' == uc[1] ) p.near = LDB_NEAR_G; if ( '0'= uc[2] ) { p.mode = LDB_OR; uc += 3; } break; case 'A': if ( 'N' == uc[1] && 'D' == uc[2] && ' ' >= uc[3] ) uc += 4; break; case 'N': if ( 'O' == uc[1] && 'T' == uc[2] && ' ' >= uc[3] ) { p.mode = LDB_NOT; uc += 4; } break; } goto nextterm; done: /* somewhat screws the idea of limiting costs ... if ( !p.fil && p.cut > p.skp ) { sMsg( LOG_INFO, "OUCH! full cut at %d for '%.50s' >= %d", p.cut, key, p.skp ); p.skp = p.cut; p.cut = 0; goto startover; } */ mFree( buf ); } return 0 > ret ? ret : ldb_p2s( set, &p ); } /* openIsisQuery */ lrec.c0100644000113000011300000005401207665600575010646 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lrec.c,v 1.53 2003/05/29 18:03:35 kripke Exp $ implementation of record cooking. */ #include #include #include /* vsnprintf */ #include /* free */ #include /* memset et al */ #ifdef WIN32 # define vsnprintf _vsnprintf #endif /* WIN32 */ #include "ldb.h" #include "lio.h" #include "lcs.h" /* ************************************************************ private types */ typedef struct { int len; /* length of tagf */ int tagf[2]; /* tag and modification flags */ } LrecMF; #define MF_SET 0x01 #define MF_DEL 0x02 /* ************************************************************ private data */ /* ************************************************************ private functions */ static LrecMF* mfCtor (Rec *rec) { int len = (int)(2 * rec->len); int *M; Field *F; LrecMF *lmf = (LrecMF*) mAlloc ((1 + len) * sizeof (int)); if (0 == lmf) { return 0; } memset (lmf->tagf, 0, len * sizeof (int)); lmf->len = len; for (M = lmf->tagf, F = rec->field; len > M - lmf->tagf; ++F, M += 2) { *M = (int) F->tag; } return lmf; } static int rCompact (Rec *rec) { int rt = 0; if (0 != rec) { char *V = (char*)rec + rec->base; Field *F = rec->field; Field *L = F + rec->len; int used = rec->base; while (F < L) { if (V != F->val) { rt = !0; break; } V += F->len; used += F->len; ++F; } if (rt) { char *buf = (char*) mAlloc (rec->used - (V - (char*)rec)); char *B = buf; int nb = 0; if (0 == buf) { return 0; } while (F < L) { memcpy (B, F->val, F->len); F->val = V; B += F->len; V += F->len; nb += F->len; ++F; } memcpy ((char*)rec + used, buf, nb); rec->used = used + nb; mFree (buf); } } return rt; } /** @return -2 - malloc failed, -1 - tag not present, 0 - compactification not neccessary, 1 - rec data buffer has gaps */ static int rReplace ( Rec **rec, LrecMF *lmf, int tag, int occr, int lmode, const char *val, int disc ) { Rec *nr; Field *F; int *M; int vlen; int fnd = -1; int ocnt = -1; for (M = lmf->tagf; lmf->len > M - lmf->tagf; M += 2) { if (tag == M[0]) { if (0 <= occr) { if (++ocnt == occr) { M[1] = lmode; fnd = (M - lmf->tagf) / 2; break; } } else { if (0 == (MF_SET & M[1])) { M[1] = lmode; fnd = (M - lmf->tagf) / 2; if (lmode) { for (M += 2; lmf->len > M - lmf->tagf; M += 2) { if (tag == M[0]) { M[1] = MF_DEL; } } } break; } } } } if (0 > fnd) { return -1; } if (MF_SET != lmode) { return 1; } F = (*rec)->field + fnd; vlen = strlen (val); if (F->len == vlen) { memcpy ((char*)F->val, val, vlen); return 0; } if (F->len > vlen) { memcpy ((char*)F->val, val, vlen); F->len = vlen; return 1; } if ((*rec)->bytes < (*rec)->used + vlen) { nr = rDup (*rec, vlen, disc); if (0 == nr) { return -2; } F = (*rec = nr)->field + fnd; } F->val = (char*)(*rec) + (*rec)->used; F->len = vlen; memcpy ((char*)F->val, val, vlen); (*rec)->used += vlen; return 1; } static int rDel (Rec *rec, LrecMF *lmf) { Field *F; int *M; int idx; int dcnt = 0; int done = 0; for (M = lmf->tagf; lmf->len > M - lmf->tagf; M += 2) { if (0 != (MF_DEL & M[1])) { idx = (M - lmf->tagf) / 2 - dcnt; if (idx < rec->len - 1) { F = rec->field + idx; memmove (F, F + 1, (rec->len - idx - 1) * sizeof (Field)); } --(rec->len); ++dcnt; done = !0; } } return done; } /* ************************************************************ package functions */ /* ************************************************************ public functions */ Field *rGet ( Rec *r, int tag, int *pos ) { if (r) { int p = pos ? *pos : 0; for ( ; p < r->len; p++ ) if ( tag == r->field[p].tag ) { if ( pos ) *pos = p+1; return &r->field[p]; } if ( pos ) *pos = r->len; } return 0; } /* rGet */ Field *rOccurence ( Rec *r, int tag, int occ ) { Field *F; int pos = 0; do { F = rGet (r, tag, &pos); } while (F && 0 <= --occ); return F; } char *rString (Rec *rec, int tag, int *pos, char *buf, int len) { Field *F = rGet (rec, tag, pos); if (!F) { return 0; } if (len > 1 + F->len) { len = 1 + (int) F->len; } if (0 >= --len) { *buf = 0; return buf; } strncpy (buf, F->val, len) [len] = 0; return buf; } char *rString2 (Rec *rec, Rec *dflt, int tag, char *buf, int len) { char *rt = rString (rec, tag, 0, buf, len); if (rt) { return rt; } return rString (dflt, tag, 0, buf, len); } int rInt ( Rec *r, int tag, int def, int *pos ) { if (r) { int p = pos ? *pos : 0; int rt; for ( ; p < r->len; p++ ) if ( tag == r->field[p].tag ) { if ( pos ) *pos = p+1; rt = a2i( r->field[p].val, r->field[p].len ); if (rt) { return rt; } if (0 < luti_true (r->field[p].val, r->field[p].len)) { return 1; } return 0; } if ( pos ) *pos = r->len; } return def; } /* rInt */ int rInt2 (Rec *rec, Rec *dflt, int tag, int def) { int rt1 = rInt (rec, tag, def, 0); if (def == rt1) { return rInt (dflt, tag, def, 0); } return rt1; } int rEnum (Fdt *fdt, Rec *rec, int tag, int def, int *pos) { char buf[OPENISIS_FD_NAMELEN]; Field *F; Fd *D; int len, ev; if (! fdt || ! rec) { return def; } D = fById (fdt, tag, 0); if (! D) { return def; } F = rGet (rec, tag, pos); if (! F) { return def; } len = (int) F->len; if (OPENISIS_FD_NAMELEN <= len) { len = OPENISIS_FD_NAMELEN - 1; } strncpy (buf, F->val, len) [len] = 0; ev = fEnum (fdt, tag, buf); if (NOENUM == ev) { return def; } return ev; } Rec* rDup ( Rec *r, int room, int discard ) { Rec *nr = 0; int nfields, nbytes, hadfields, hadcontent; if ( ! r ) { nfields = 80; /* for a 1K base */ nbytes = 8*1024; if ( nbytes < BASESZ(80) + room*3/2 ) nbytes = BASESZ(80) + room*3/2; hadfields = 0; hadcontent = 0; } else { rCompact (r); hadfields = r->len; hadcontent = r->used - r->base; if ( 0 > room ) { /* shrink to fit */ nfields = r->len; nbytes = BASESZ(nfields) + hadcontent; } else { nfields = 54 > r->len ? 80 : (r->len * 3 / 2); /* add 50% */ if ( nfields < r->fields ) nfields = r->fields; nbytes = 6*1024 > r->bytes ? 8*1024 : (r->bytes *3 / 2); if ( nbytes < r->used + room ) nbytes = r->used + room*3/2; nbytes += (nfields - r->fields)*sizeof(Field); } } LOG_DBG( LOG_DEBUG, "extending rec size %d (cont %d) to %d bytes, %d -> %d fields", !r ? -1 : r->bytes, hadcontent, nbytes, hadfields, nfields ); assert( nbytes >= BASESZ( nfields ) + hadcontent + room ); if ( ! (nr = mAlloc( nbytes )) ) return 0; memset( nr, 0, nbytes ); /* paranoia */ nr->bytes = nbytes; nr->fields = nfields; nr->base = BASESZ( nfields ); nr->used = nr->base + hadcontent; nr->len = hadfields; if ( ! r ) { nr->dbid = -1; /* no valid dbid if new rec */ } else { char *obuf = ((char*)r) + r->base; char *nbuf = ((char*)nr) + nr->base; nr->dbid = r->dbid; /* for ! r, the following are 0 by memset */ nr->rowid = r->rowid; nr->state = r->state; if ( hadcontent ) memcpy( nbuf, obuf, hadcontent ); if ( hadfields ) { int i; int displace = nbuf - obuf; /* ptrdiff_t */ Field *f = nr->field; memcpy( nr->field, r->field, hadfields*sizeof(Field) ); for ( i=hadfields; i--; ) if ( f[i].val ) f[i].val += displace; #ifndef NDEBUG { char *end = ((char*)nr) + nr->used; for ( i=hadfields; i--; ) if ( f[i].val && (f[i].val < nbuf || f[i].val + f[i].len > end) ) { int wasok = r->field[i].val >= obuf && r->field[i].val + r->field[i].len <= obuf + hadcontent; sMsg( LOG_ERROR, "OOPS! nuked field %d which previously was %s", i, wasok ? "ok" : "already broken" ); return 0; /* no cleanup, we're nearly dead anyway */ } } #endif } if ( discard ) free( r ); } assert( RECOK(nr) ); return nr; } /* rDup */ Rec* rMsg ( Rec *r, int discard, int tag, const char *fmt, ... ) { char buf[1024]; int l; va_list ap; va_start( ap, fmt ); l = vsnprintf( buf, sizeof(buf), fmt, ap ); if ( 0 > l ) /* older versions return -1 */ l = sizeof(buf); va_end( ap ); RADD( r, tag, buf, l, discard ); return r; } /* rMsg */ /* #define openIsisPrintf( r, d, t, f, a... ) \ openIsisRMsg( OPENISIS_SES0(), r, d, t, f, ## a ) requires gcc ... */ Rec* openIsisPrintf ( Rec *r, int discard, int tag, const char *fmt, ... ) { char buf[1024]; int l; va_list ap; va_start( ap, fmt ); l = vsnprintf( buf, sizeof(buf), fmt, ap ); if ( 0 > l ) /* older versions return -1 */ l = sizeof(buf); va_end( ap ); RADD( r, tag, buf, l, discard ); return r; } /* openIsisPrintf */ Rec *rAddI (Rec *rec, int tag, int val, int discard) { char buf[32]; int len = i2a (buf, (int)val); RADD (rec, tag, buf, len, discard); return rec; } Rec* rSet ( Rec *oldr, int mode, ... ) { const char **argv = 0; LrecMF *lmf = 0; Fdt *fdt = 0; Rec *newr = oldr; int argc = 0xffff & mode; int dis = RDIS & mode; int op = ROP & mode; int lmode = MF_SET; int compact = 0; va_list ap; va_start( ap, mode ); if (RFDT & mode) { fdt = va_arg( ap, Fdt* ); } if (RARGV & mode) { argv = va_arg( ap, const char** ); } if (! fdt) { fdt = luti_fdt_from_rec (oldr); } if (op) { switch (op) { case RDEL: lmode = MF_DEL; break; case RCHG: break; case RDFLT: lmode = 0; break; default: sMsg (ERR_IDIOT, "rSet: illegal mode %x", op); return newr; } if (oldr) { if (0 == (lmf = mfCtor (oldr))) { goto done; } } else if (RDEL == op) { goto done; } } for ( ;; ) { char ib[16]; Rec *tmpr; const char *v = 0; int id = -1; int occr = -1; Fd *fd = 0; int rpl; if ( argv ) { if ( 1 > argc ) break; v = luti_parse_path (*argv, fdt, &fd, &id, &occr); if (!v || *v) { fd = 0; id = -1; } v = 0; ++argv; --argc; } else { id = va_arg( ap, int ); if ( ! id ) break; fd = fdt ? fById( fdt, id, 0 ) : 0; if (RDEL != op) { v = va_arg( ap, char* ); } } if ( fd ) { if (RDEL != op) { int e; if (argv) { if ( 1 > argc ) { if (FTB != fd->type) { break; } } else { v = *argv++; --argc; } } switch ( fd->type ) { case FTE: e = fEnum( fdt, fd->id, v ); if ( NOENUM == e ) { if ( RIGN & mode ) continue; sMsg( LOG_ERROR, "bad enum value '%s' for id %d", v, fd->id ); goto error; } i2a( ib, e ); v = ib; break; case FTB: e = v ? luti_true (v, -1) : 1; if (0 > e) { /* non-given values default to true */ ++argc; --argv; v = "1"; } else { v = e ? "1" : "0"; } break; } } /* RDEL != op */ } /* fd */ else { if (argv) { if (RDEL != op) { if ( 1 > argc ) { break; } v = *argv++; --argc; } } if (fdt /*|| 0 >= id*/) { if ( RIGN & mode ) continue; if ( argv ) sMsg( LOG_ERROR, "unknown field name '%s'", argv[RDEL == op ? -1 : -2] ); else sMsg( LOG_ERROR, "unknown field id %d", id ); break; } } if (!v) { v = ""; } if (op) { rpl = lmf ? rReplace ( &newr, lmf, id, occr, lmode, v, dis || newr != oldr) : -1; if (RDFLT == op) { if (-1 != rpl) { continue; } } else { if (-2 == rpl) { goto done; } if (-1 != rpl) { if (1 == rpl) { compact = !0; } continue; } if (RDEL == op) { continue; } } /* fall thru */ } tmpr = newr; RADDS( newr, id, v, dis || newr != oldr ); if (0 == newr) { goto done; } if (newr != tmpr) { compact = 0; } sMsg( LOG_VERBOSE, "added v '%s' id %d as %dth", v, id, newr->len ); } error: if (lmf) { if (rDel (newr, lmf)) { compact = !0; } } if (compact && 0 == (RNOC & mode)) { rCompact (newr); } done: if (lmf) { mFree (lmf); } va_end( ap ); return newr; } /* rSet */ Rec *dFmt ( Rec *buf, const char *fmt, int db, int rowid ) { Rec *r = dRead( db, rowid ); Rec *q; if ( ! r ) return 0; q = rFmt( buf, fmt, r ); free( r ); return q; } /* dFmt */ Rec *dScan ( int db, int rowid, int tag, const char *txt ) { int max = dMaxId( db ); int tlen = strlen( txt ); const char f = *txt; for ( ; rowid <= max; rowid++ ) { int i, found = 0; Rec *r = dRead( db, rowid ); if ( ! r ) continue; for ( i=0; i < r->len; i++ ) { const char *c, *e; if ( found || (tag > 0 && tag != r->field[i].tag) || tlen > r->field[i].len ) continue; c = r->field[i].val; e = c + r->field[i].len - tlen; for ( ; c<=e; c++ ) if ( f == *c && !memcmp( c, txt, tlen ) ) { found = !0; break; } } if ( found ) return r; } return 0; } /* dScan */ Rec *rSplitf ( Rec *r, const Field* field ) { int i; int nfields = 0; /* number of subfields */ int size; /* byte length of field list */ const char *p, *e; if ( ! field || 2 > field->len ) return 0; if ( '^' != field->val[0] ) /* initial anonymous subfield */ nfields = 1; /* go counting hats ... */ for ( e = (p=field->val) + field->len - 1; plen, field->val ); if ( r ) { if ( 1 > r->len ) return 0; size = sizeof(Rec) + (r->len - 1) * sizeof(Field); if ( nfields > r->len ) nfields = r->len; } else { /* first field already counted in sizeof(Rec) */ size = sizeof(Rec) + (nfields - 1) * sizeof(Field); r = (Rec *)mAlloc( size ); if ( ! r ) return 0; } memset( r, 0, size ); r->rowid = 0; r->len = nfields; e = (p = field->val) + field->len; if ( '^' == *p ) /* else initial anonymous subfield */ p++; for ( i=0; i= e ) { sMsg( LOG_ERROR, "confused cats hound counting hats at %d in '%.*s'", i, field->len, field->val ); r->len = i; break; } /* p is after a hat */ r->field[ i ].tag = p==field->val ? 0 : (unsigned char)*p++; r->field[ i ].val = p; /* advance in buf */ while ( pfield[ i ].len = p - r->field[ i ].val; /* log_msg( LOG_ERROR, "subf %d of %d tag %c off %d", i, nfields, r->field[ i ].tag, r->field[ i ].val - field->val ); */ } return r; } /* rSplitf */ enum { /* stream reading state */ RS_REC, /* not in any record */ RS_TAG, /* in tag: done is at beginning of line */ RS_SEP, /* in sep: done is at beginning of sep */ RS_VAL, /* done is somewhere in the value ... */ RS_EOL /* not really a state, but a flag */ }; /** read a record from a stream. This may either
  1. read a record and return 1
  2. not get enough bytes on a non-blocking stream and return 0
  3. have some error and return negative
*/ int sGetr ( OpenIsisRecStream *s ) { Ios *lio = s->in; int state = 3 & s->flg; int emptyline = 0; int ret = 0; if ( ! s->rec && ! (s->rec = s->buf) /* have buffer ? */ && ! (s->rec = rDup( 0, 0, 0 )) /* allocate and prepare an 8k standard record */ ) { ret = sMsg( LOG_ERROR, "could not get record" ); goto fatal; } if ( RS_REC == state ) { /* initialize */ s->rec->rowid = 0; s->rec->len = 0; /* nuke all fields */ s->rec->used = s->rec->base; /* nuke value buffer */ /* let base/fields untouched, it hopefully has a reasonable value */ state = RS_TAG; } if ( ! s->rec->len ) state = RS_TAG; for (;;) { unsigned char *b, *p, *e; /* begin, pointer, end */ if ( lio->b.done >= lio->b.fill ) { reload: if ( ! LIO_FILL( lio ) ) goto done; /* EAGAIN -- see ya later */ /* so we have bytes or EOF */ if ( lio->b.done >= lio->b.fill ) { if ( LIO_IN & lio->file ) { sMsg( LOG_ERROR, "confused in instream: no bytes and no EOF" ); lio_close( &lio->file, LIO_IN ); } break; } } /* so we have bytes */ p = b = lio->b.c + lio->b.done; e = lio->b.c + lio->b.fill; /* try to eat one line */ switch ( state ) { case RS_TAG: { int tag = -1; int recognized = 0; if ( LCS_ISFR(lcs_latin1_ct[*p]) ) { /* line starts with line/record sep */ if ( '\n' == *p && (RS_EOL & s->flg) ) { /* TODO */ s->flg &= ~RS_EOL; continue; } emptyline = !0; break; } for (;;p++) { if ( p < e ) { if ( LCS_ISWORD(lcs_latin1_ct[*p]) ) continue; /* tight loop */ break; } if ( 4096 > p - b ) goto reload; sMsg( ERR_TRASH, "tag too long" ); goto fatal; } /* identify tag between s and p */ if ( p == b ) recognized = !0;/* empty tag */ else if ( '0' <= *b && *b <= '9' ) { unsigned char *d = b; /* digit ? */ tag = 0; while ( d < p && '0' <= *d && *d <= '9' ) tag = 10*tag + *d++ - '0'; if ( !(recognized = d == p) ) /* all digits */ tag = -1; } if ( ! recognized && s->dict ) { /* dict lookup */ int i=s->dict->len, l = p-b; Field *f = s->dict->field; for ( ; i--; f++ ) if ( l == f->len && memcmp( b, f->val, l ) ) { tag = f->tag; break; } } /* create new field unless it's a continuation line */ if ( p == b && s->rec->len ) ; else if ( recognized ) RADD( s->rec, tag, 0, 0, s->rec != s->buf ); else RADD( s->rec, tag, b, p-b, s->rec != s->buf ); if ( ! s->rec ) goto outamem; /* did that */ lio->b.done = (b = p) - lio->b.c; if ( LCS_ISFR(lcs_latin1_ct[*p]) ) break; state = RS_SEP; } case RS_SEP: /* p = b is at beginning of separator */ while ( p < e && ' ' == *p ) /* skip leading blanks */ p++; if ( p < e && LCS_ISCST(lcs_latin1_ct[*p]) && '\t' != *p++ ) while ( p < e && ' ' == *p ) /* skip trailing blanks */ p++; if ( e == p ) { if ( 4096 > p - b ) goto reload; sMsg( ERR_TRASH, "sep too long" ); goto fatal; } assert( s->rec->len ); if ( s->rec->field[s->rec->len - 1].len ) RCAT( s->rec, b, p-b, s->rec != s->buf ); if ( ! s->rec ) goto outamem; lio->b.done = (b = p) - lio->b.c; if ( LCS_ISFR(lcs_latin1_ct[*p]) ) break; state = RS_VAL; case RS_VAL: default: /* ??? */ while ( p < e && !LCS_ISFR(lcs_latin1_ct[*p]) ) p++; assert( s->rec->len ); if ( p > b ) RCAT( s->rec, b, p-b, s->rec != s->buf ); if ( ! s->rec ) goto outamem; lio->b.done = (b = p) - lio->b.c; if ( p == e ) goto reload; /* else we found line or record separator */ } /* switch state */ if ( '\r' == *p ) { /* do that ole' ugly CR/NL quatsch */ if ( e-p < 2 && 0 < LIO_FILL(lio) ) /* extend end */ e = lio->b.c + lio->b.fill; if ( e-p < 2 ) s->flg |= RS_EOL; /* possibly check later */ else if ( '\n' == p[1] ) p++; } lio->b.done = 1 + p - lio->b.c; /* consider p eaten */ if ( LCS_R == lcs_latin1_ct[*p] || (emptyline && (OPENISIS_STOPONEMPTY & s->flg)) ) break; state = RS_TAG; } /* got record */ state = RS_REC; if ( s->rec && s->rec->len ) { if ( (OPENISIS_AUTOCLONE & s->flg) && s->rec == s->buf ) s->rec = rDup( s->buf, 0, 0 ); ret = 1; } goto done; outamem: fatal: if ( LIO_SISOPEN( s->in ) ) LIO_CLOSE( s->in ); done: s->flg = state | (~3 & s->flg); /* save state */ /* sMsg( LOG_ERROR, "fill %d done %d", lio->b.fill, lio->b.done ); */ return ret; } /* openIsisReadStream */ int rSer ( char *buf, OpenIsisRec *rec ) { Field *f = rec->field; int i = rec->len; char *p = buf, *e; for ( ; i--; f++ ) { p += i2a( p, f->tag ); *p++ = 9; /* TAB */ memcpy( p, f->val, f->len ); for ( e = p + f->len; p= 2*rec->used */ int rSerB ( char *buf, OpenIsisRec *rec ) { Field *f = rec->field; int i = rec->len; char *p = buf; const char *s, *e; for ( ; i--; f++ ) { p += i2a( p, f->tag ); *p++ = 9; /* TAB */ for ( e = (s = f->val) + f->len; sused) { buf = (char*) mAlloc (1 + rec->used); if (!buf) { sMsg (ERR_NOMEM, "rSerA: cannot malloc %d", (1 + rec->used)); return 0; } } *len = rSer (buf, rec); return buf; } int rDeser ( OpenIsisRec **rp, const char *buf, int len, int flg ) { Rec *rec = *rp; Field *f; const char *p = buf, *e = buf+len; char *q; int dis = RDIS & flg; int tag; /* non-transparent mode - no continuation lines */ for ( ; e > p; p++ ) { if ( 10 == *p ) { /* blank line */ if ( OPENISIS_STOPONEMPTY & flg ) break; continue; } p += a2il( p, e-p, &tag ); if ( e > p && 9 == *p ) /* TAB */ p++; RADD( rec, tag, 0, e-p, dis || rec!=*rp ); if ( !rec ) break; f = rec->field + rec->len-1; for ( q=(char*)f->val; e>p && 10 != *p; ) if ( 11 == (*q++ = *p++) ) q[-1] = 10; rec->used += f->len = q - f->val; } *rp = rec; return p-buf; } /* rDeser */ lrectest.c0100644000113000011300000002023407654477732011552 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lrectest.c,v 1.12 2003/05/01 10:05:41 mawag Exp $ tests for lrec methods. */ #include #include #include #include #include #include "openisis.h" #include "loi.h" #include "lfdt.h" #include "luti.h" #ifdef WIN32 #define snprintf _snprintf #define vsnprintf _vsnprintf #endif static const char* _TheSummary = "Meeting on the Implementation of the Recommendations of the Third Regional Conference of Ministers of Education and Ministers Responsible for Economic Planning in the Arab States"; #define SUMLEN (strlen (_TheSummary)) static Fd _TheFds[] = { { 20, 0, FTX, 0, 0, 32, "author", "", 0, 0, 0, 0 }, { 21, 0, FTX, 1, 0, 32, "names", "", 0, 0, 0, 0 }, { 22, 0, FTX, 1, 0, 32, "year", "", 0, 0, 0, 0 }, { 42, 0, FTX, 1, 0, 32, "title", "", 0, 0, 0, 0 }, { 49, 0, FTX, 1, 0, 32, "summary", "", 0, 0, 0, 0 }, { 70, 0, FTX, 1, 0, 32, "country", "", 0, 0, 0, 0 }, { 80, 0, FTX, 1, 0, 32, "conference", "", 0, 0, 0, 0 }, { 81, 0, FTE, 0, 0, 32, "format", "", 0, 0, 0, 0 }, { 81, 0, FTV, 0, 0, 1, "autoformat", "", 0, 0, 0, 0 }, { 81, 0, FTV, 0, 0, 2, "naligned", "", 0, 0, 0, 0 }, { 81, 0, FTV, 0, 0, 73, "aligned", "", 0, 0, 0, 0 }, { 82, 0, FTB, 0, 0, 10, "soldout", "", 0, 0, 0, 0 } }; static Fdt _TheFdt = { sizeof(_TheFds) / sizeof(_TheFds[0]), _TheFds, 0 }; static void check ( char *msg, Rec *rec, int *tag, const char **val, int len, int used ) { Field *F; int idx, vlen, ok; if (0 == rec) { fprintf (stderr, "ERR %s: 0 == rec\n", msg); exit (1); } if (! OPENISIS_RECOK (rec)) { fprintf (stderr, "ERR %s: ! RECOK: b %d u %d/%d f %d/%d\n", msg, rec->base, rec->used, rec->bytes, rec->len, rec->fields); exit (1); } if (len != rec->len) { fprintf (stderr, "ERR %s: len %d != %d\n", msg, rec->len, len); log_rec (LOG_ERROR, rec, 0, 0); exit (1); } if (0 <= used && used != rec->used - rec->base) { fprintf (stderr, "ERR %s: used %d (%d-%d) != %d\n", msg, (rec->used - rec->base), rec->used, rec->base, used); log_rec (LOG_ERROR, rec, 0, 0); exit (1); } ok = !0; for (idx = 0; len > idx; ++idx) { vlen = strlen (val[idx]); F = rec->field + idx; if (F->tag != tag[idx]) { fprintf (stderr, "ERR %s: fld[%d] tag %d != %d\n", msg, idx, F->tag, tag[idx]); ok = 0; } if (F->len != vlen || strncmp (F->val, val[idx], vlen)) { fprintf (stderr, "ERR %s: fld[%d] val <%.*s> != <%s>\n", msg, idx, (int)F->len, F->val, val[idx]); ok = 0; } } if (! ok) { exit (1); } } int main (int argc, char **argv) { const char *args[32]; const char *vals[32]; int tags[32]; Rec *rec = 0; /* call lio_init to prepare stderr on windows */ openIsisCOpen (0); if (1 < argc) { #define EXPFLD (1332 + (1000 - 888)) int bytes = 0, used = 0, num = atoi (argv[1]); printf ("entering loops (%d) ...\n", num); while (num) { int j; rec = 0; for (j = 1000; 1 <= --j; ) { RADD (rec, j, _TheSummary, SUMLEN, !0); } for (j = 888; 666 <= --j; ) { rec = rSet (rec, RDEL | RNOC, j, 0); } for (j = 666; 444 <= --j; ) { rec = rSet (rec, RDEL, j, 0); } for (j = 444; 222 <= --j; ) { rec = rSet (rec, RDEL | RNOC, j, 0); } for (j = 1333; 222 <= --j; ) { RADD (rec, j, _TheSummary, SUMLEN, !0); } if (EXPFLD != rec->len) { printf ("%d: ERR: #len = %d, expected %d\n", num, rec->len, EXPFLD); exit (1); } if (0 == bytes) { bytes = rec->bytes; used = rec->used; printf ("mem %d use %d ...\n", rec->bytes, rec->used); } else { if (bytes != rec->bytes || used != rec->used) { printf ("%d: ERR: mem = %d use = %d\n", num, rec->bytes, rec->used); exit (1); } if (0 == (num % 100)) { printf ("%d ...\n", num); } } mFree (rec); --num; } printf ("ok.\n"); exit (0); } args[0] = "20"; args[1] = "Mary"; rec = rSet (rec, RARGV | 2, args); tags[0] = 20; vals[0] = args[1]; check ("A", rec, tags, vals, 1, 4); args[0] = "22"; args[1] = "1953"; args[2] = "70"; args[3] = "Italy"; args[4] = "98"; args[5] = "illegal"; args[6] = "20"; args[7] = "John"; rec = rSet (rec, RARGV | RFDT | 8, &_TheFdt, args); tags[0] = 20; vals[0] = "Mary"; tags[1] = 22; vals[1] = "1953"; tags[2] = 70; vals[2] = "Italy"; tags[3] = 20; vals[3] = "John"; check ("B", rec, tags, vals, 3, 13); args[0] = "98"; args[1] = "illegal"; args[2] = "20"; args[3] = "John"; rec = rSet (rec, RARGV | RFDT | RIGN | 4, &_TheFdt, args); check ("C", rec, tags, vals, 4, 17); rec = rSet (rec, RDEL, 20, 0); tags[0] = 22; vals[0] = "1953"; tags[1] = 70; vals[1] = "Italy"; check ("D", rec, tags, vals, 2, 9); rec = rSet (rec, RDFLT | RFDT, &_TheFdt, 81, "naligned", 22, "already there", 0); tags[2] = 81; vals[2] = "2"; check ("D1", rec, tags, vals, 3, 10); tags[3] = 234; vals[3] = "foo"; rec = rSet (rec, RDFLT, 234, "foo", 0); check ("D2", rec, tags, vals, 4, 13); rec = rSet (rec, RDEL, 234, 81, 0); check ("D3", rec, tags, vals, 2, 13); rec = rSet (rec, 0, 20, "Mary", 20, "James", 90, "Alabama", 0); tags[2] = 20; vals[2] = "Mary"; tags[3] = 20; vals[3] = "James"; tags[4] = 90; vals[4] = "Alabama"; check ("E", rec, tags, vals, 5, 29); rec = rSet (rec, RFDT | RCHG, &_TheFdt, 20, "Lary", 70, "Greece", 70, "Cuba", 42, "Folklore from Africa", 81, "invalid", 0); tags[0] = 22; vals[0] = "1953"; tags[1] = 70; vals[1] = "Greece"; tags[2] = 20; vals[2] = "Lary"; tags[3] = 90; vals[3] = "Alabama"; tags[4] = 70; vals[4] = "Cuba"; tags[5] = 42; vals[5] = "Folklore from Africa"; check ("F", rec, tags, vals, 6, 45); rec = rSet (rec, 0, 20, "Peter", 20, "Gary", 70, "Egypt", 0); tags[6] = 20; vals[6] = "Peter"; tags[7] = 20; vals[7] = "Gary"; tags[8] = 70; vals[8] = "Egypt"; check ("G", rec, tags, vals, 9, 59); args[0] = "author[1]"; args[1] = "81"; args[2] = "70"; rec = rSet (rec, RFDT | RARGV | RDEL | 3, &_TheFdt, args); tags[0] = 22; vals[0] = "1953"; tags[1] = 20; vals[1] = "Lary"; tags[2] = 90; vals[2] = "Alabama"; tags[3] = 42; vals[3] = "Folklore from Africa"; tags[4] = 20; vals[4] = "Gary"; check ("H", rec, tags, vals, 5, 39); args[0] = "author[0]"; args[1] = "May"; args[2] = "summary"; args[3] = _TheSummary; args[4] = "soldout"; args[5] = "format"; args[6] = "aligned"; rec = rSet (rec, RFDT | RARGV | RCHG | RNOC | 7, &_TheFdt, args); vals[1] = "May"; tags[5] = 49; vals[5] = _TheSummary; tags[6] = 82; vals[6] = "1"; tags[7] = 81; vals[7] = "73"; check ("I", rec, tags, vals, 8, 42 + SUMLEN); args[0] = "20[0]"; args[1] = "Marilyn"; args[2] = "49[1]"; args[3] = _TheSummary; rec = rSet (rec, RARGV | RCHG | 4, args); vals[1] = "Marilyn"; tags[8] = 49; vals[8] = _TheSummary; check ("J", rec, tags, vals, 9, 45 + 2 * SUMLEN); rec = rSet (rec, RDEL | 2, 82, 22, 0); tags[0] = 20; vals[0] = "Marilyn"; tags[1] = 90; vals[1] = "Alabama"; tags[2] = 42; vals[2] = "Folklore from Africa"; tags[3] = 20; vals[3] = "Gary"; tags[4] = 49; vals[4] = _TheSummary; tags[5] = 81; vals[5] = "73"; tags[6] = 49; vals[6] = _TheSummary; check ("K", rec, tags, vals, 7, 40 + 2 * SUMLEN); rec = rSet (rec, RFDT | RCHG, &_TheFdt, 82, "false", 81, "2", 49, "overview", 0); vals[4] = "overview"; vals[5] = "2"; tags[6] = 82; vals[6] = "0"; check ("L", rec, tags, vals, 7, 48); mFree (rec); printf ("ok.\n"); return 0; } lscvs0100755000113000011300000000146707576377762010655 0ustar kripkrip#!/usr/bin/perl # list files under CVS control # $Id: lscvs,v 1.2 2002/12/13 15:42:42 kripke Exp $ sub dodir { my $dir = shift; print STDERR "$dir\n"; local *FH; open( FH, $dir.'CVS/Entries' ); while () { # print STDERR $_; chomp; my ($type,$name,$version,$date,$opt,$tagdate) = split( /\// ); if ( '' eq $type ) { if ( $l ) { print join("\t",$dir.$name,$version,$date,$x),"\n"; } else { print $dir.$name,"\n"; } } elsif ( 'D' eq $type ) { dodir( $dir.$name.'/' ) if $r && $name; # "`D' on a line all by itself indicates that the program # which wrote the `Entries' file does record subdirectories" } else { print STDERR "uknown type '$type' in $dir ", join("\t",$name,$version,$date,$x),"\n"; } } close FH; } $l = $ARGV[0] =~ /l/; $r = $ARGV[0] =~ /r/; dodir; lses.c0100644000113000011300000001757007664371301010666 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lses.c,v 1.18 2003/05/26 10:55:10 kripke Exp $ implementation of general db access functions. */ #include #include /* memset */ #ifdef _REENTRANT #ifdef WIN32 # define WIN32_LEAN_AND_MEAN # define NONAMELESSUNION # include /* they call it "thread local storage" */ # define pthread_key_t int # define pthread_key_create( pkey, func ) *(pkey) = TlsAlloc() # define pthread_setspecific( key, val ) TlsSetValue( key, val ) # define pthread_getspecific( key ) TlsGetValue( key ) #else # include #endif #endif #include "lses.h" /* ************************************************************ private types */ typedef struct { Ses h; /* int brk; */ } SesI; enum { LSES_MAX = 0x1000, LSES_MSK = 0x0fff }; /* ************************************************************ package data */ /* ************************************************************ private data */ static Ios stdio[3] = { LIO_STDINITIALIZER( "&0", LIO_IN ), LIO_STDINITIALIZER( "&1", LIO_OUT|1 ), LIO_STDINITIALIZER( "&2", LIO_OUT|2 ) }; static SesI cses = { /* the default session */ { 0, "", { stdio, stdio+1, stdio+2 }, -1, 0, 0, /* nxt, accnt */ {0},{0},{0}, /* times */ 0,0,0,0 /* state */ } }; static SesI *tab[ LSES_MAX ] = { &cses }; static Ses *htab[ LSES_MAX ]; static int maxid; static const char oops[] = "out of memory\n"; #ifndef _REENTRANT Session ses = &cses.h; #else int openisis_threaded = 42; static pthread_key_t seskey; #endif /* ************************************************************ private functions */ static void *memordie ( int size ) { void *p = malloc( size ); if ( p ) return p; /* TODO: ask btree to release some cache */ lio_write( &lio_err, oops, sizeof(oops)-1 ); exit( 71/*EX_OSERR*/ ); /* exit handlers MUST NOT alloc memory */ return 0; } /* ************************************************************ package functions */ void lses_init () /* called from openIsisInit */ { #ifdef _REENTRANT pthread_key_create( &seskey, 0 ); /* TODO: use ses destr func */ pthread_setspecific( seskey, &cses.h ); #endif } /* lses_init */ void lses_fini () /* called from openIsisFini */ { int i, s = LSES_MAX; /* lio_write( &lio_out, "fini time\n", 9 ); */ while ( s-- ) if ( tab[s] ) for ( i=0; i<3; i++ ) if ( tab[s]->h.io[i] ) LIO_CLOSE( tab[s]->h.io[i] ); } /* lses_fini */ int lses_open ( Session s, const char *name, int flags, SFunc *type ) { int fd = 0; const char *p = name; lio_sfunc *func = type ? (lio_sfunc*)type : lio_stdio; /* check for i<> */ while ( '0' <= *p && *p <= '9' ) fd = 10*fd + *p++ - '0'; if ( '<' != *p && '>' != *p ) p = name; /* forget it */ else { if ( name == p && '>' == *p ) /* no fd given, > defaults to 1 */ fd = 1; flags |= '<' == *p ? LIO_RD : LIO_WR; p++; } if ( p == name ) { /* no fd given, find one */ for ( fd=0; s->io[fd]; ) if ( LSES_FILE_MAX == ++fd ) return -ERR_NOMEM; } else if ( s->io[fd] ) { /* close existing */ func( s->io[fd], LIO_SCLOSE ); /* don't kill the statics :) */ if ( s->io[fd] < stdio || s->io[fd] > stdio+2 ) mFree( s->io[fd] ); } s->io[fd] = (Ios*)mAlloc( func( 0, LIO_SSIZE ) ); if ( ! s->io[fd] ) return -ERR_NOMEM; s->io[fd]->func = func; s->io[fd]->name = mDup( p, -1 ); s->io[fd]->file = flags; return func( s->io[fd], LIO_SOPEN ); } /* lses_open */ /* ************************************************************ public functions */ extern Ses *sGet () { #ifdef _REENTRANT Ses *ses; if ( ! (ses = (Ses*)pthread_getspecific( seskey )) ) pthread_setspecific( seskey, ses = cSession( 0 ) ); #endif return ses; } extern void sSet ( Ses *s ) { #ifndef _REENTRANT ses = s; #else pthread_setspecific( seskey, s ); #endif } Ses *cOpen ( Rec *args ) { extern void linit(); linit(); if ( args ) { /* add args */ /* TODO: merge args */ cses.h.prop = rDup( args, -1, 0 ); } return &cses.h; } /* cOpen */ Ses *cSession ( Rec *args ) { SesI *sesi = 0; Ses *s = 0; int id = 1; /** XXX unless we actually free sessions, we should start checking with maxid or check for expiry */ while ( tab[id] ) if ( LSES_MAX == ++id ) return 0; if ( maxid < id ) maxid = id; sesi = tab[id] = (SesI*)malloc( sizeof(SesI) ); memset( sesi, 0, sizeof(SesI) ); s = &sesi->h; s->id = id; s->hash = -1; /* not hashed (hash is non-negative) */ s->prop = args ? rDup( args, -1, 0 ) : 0; /* check props */ /* open files */ lses_open( s, "2>&2", 0, lio_stdio ); return s; } /* cSession */ void *mAlloc ( int size ) { void *p = memordie( size ); memset(p, 0, size); return p; } /* mAlloc */ void mFree ( void *mem ) { if ( mem ) free( mem ); } /* mFree */ void *mDup ( const void *str, int sz ) { void *m = memordie( 0<=sz ? sz : (sz = strlen( str ) + 1) ); memcpy( m, str, sz ); return m; } /* mDup */ int sOpen ( const char *name, int flags, SFunc *type ) { SESDECL return lses_open( ses, name, flags, type ); } /* sOpen */ Ses *openIsisId2Session (int sid) { if (0 > sid || LSES_MAX <= sid || 0 == tab[sid]) { return (Ses*)0; } return &(tab[sid]->h); } /* openIsisId2Session */ int openIsisSession2Id (Ses *s) { if (0 == s) { return -1; } return s->id; } /* openIsisSession2Id */ Ses *cSesByName ( char *name, int nlen, Tm *now, Tm *expire ) { char tbuf[20]; int h,i; Ses *s; if ( nlen < 0 ) nlen = strlen(name); if ( nlen > 63 ) nlen = 63; for ( i=nlen; i--; ) if ( !(0xe0 & name[i]) ) name[i] = '_'; h = lhash( name, nlen ); for ( s = htab[h & LSES_MSK]; s; s = s->nxt ) if ( h == s->hash && !memcmp( name, s->name, nlen ) && !s->name[nlen] ) { /* match */ if ( !expire || s->atime.millis >= expire->millis ) goto gotit; /* valid */ if ( s->cur || s->que ) { log_msg( LOG_WARN, "expired session %d %s still busy!", s->id, s->name ); goto gotit; /* valid anyway */ } goto refresh; } /* didn't find by name */ if ( ! expire ) /* bad thing -- we're going to run out of sessions ... */ i = maxid+1; else /* check other expired */ for ( i=1; i<=maxid; i++ ) { s = &tab[i]->h; /* sind wir heute defensiv ... */ if ( s && s->atime.millis < expire->millis && s->atime.millis && 0 <= s->hash /* else is none of our biz */ && !s->cur && !s->que /* else is busy !? */ ) break; } if ( i > maxid ) { s = cSession( 0 ); if ( ! s ) return 0; } else /* got one */ log_msg( LOG_INFO, "reusing session %d '%s' atime %s", s->id, s->name, s->atime.millis ? timeGtf(tbuf,&s->atime) : "-" ); if ( 0 <= s->hash ) { /* was hashed -- unlink */ Ses **spp = &htab[s->hash & LSES_MSK]; for ( ; *spp; spp = &(*spp)->nxt ) if ( s == *spp ) { *spp = s->nxt; break; } } memset( s->name, 0, sizeof(s->name) ); memcpy( s->name, name, nlen ); s->hash = h; s->nxt = htab[h & LSES_MSK]; htab[h & LSES_MSK] = s; refresh: s->ctime.millis = s->mtime.millis = now ? now->millis : 0; s->accnt = 0; if ( s->prop ) CLRREC( s->prop ); if ( s->res ) CLRREC( s->res ); gotit: s->atime.millis = now ? now->millis : 0; return s; } /* cSesByName */ lses.h0100644000113000011300000000441507645272420010666 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LSES_H /* $Id: lses.h,v 1.8 2003/04/10 13:43:44 kripke Exp $ session */ #include "loi.h" #include "lio.h" enum { LSES_FILE_MAX = 16 }; typedef struct Ses { int id; char name[64]; Ios *io[LSES_FILE_MAX]; int hash; /* -value of name */ struct Ses *nxt; /* chain in hash */ int accnt; /* # accesses */ Tm ctime; Tm mtime; Tm atime; Rec *prop; /* session state */ Rec *res; /* a reusable result */ struct Con *cur; /* current connection */ struct Con *que; /* connection queue */ /* actually this is followed by internal members ... */ } Ses; /** get the Session for the current thread. if it doesn't have one, one is created. */ extern Ses *sGet (); extern void sSet ( Ses *ses ); #ifdef _REENTRANT #define SESDECL Session ses = sGet(); #define SESGET() sGet() #define SESSET(s) sSet(s) #else extern Ses *ses; #define SESDECL #define SESGET() ses #define SESSET(s) do { ses = (s); } while (0) #endif extern void lses_init (); extern void lses_fini (); /** get or create session by name (of len nlen). name will be cleaned of control chars (<32). touch atime (and ctime on new) with now (if not 0). clean and reuse sessions with atime < expire (if not 0) */ extern Ses *cSesByName ( char *name, int nlen, Tm *now, Tm *expire ); #define LSES_H #endif /* LSES_H */ lstb.c0100644000113000011300000005225707662161240010662 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lstb.c,v 1.40 2003/05/16 12:32:07 mawag Exp $ implementation of client side session and communication functions. */ #include #include /* PATH_MAX */ #include #include #include #include "openisis.h" #include "loi.h" #include "lcli.h" #include "ldb.h" #include "ldsp.h" #include "lfdt.h" #include "luti.h" #ifdef WIN32 #define snprintf _snprintf #define vsnprintf _vsnprintf #endif #define LSTB_STHOSTLEN 32 #define LSTB_MAXDBS 255 /* according to OPENISIS_DBIDMSK */ #define LSTB_MAXSTUBS 255 /* according to OPENISIS_SCIDMSK */ #define LSTB_INCRDB 8 #define LSTB_INCRSTUB 8 /* ************************************************************ * meta */ /* stub id 0 has idx -1 */ /* stub id 0x0100 has idx 0 */ #define SC_X2I(s) ( ( (1 + (s)) << 8 ) & OPENISIS_SCIDMSK ) #define SC_I2X(s) ( ( ((s) & OPENISIS_SCIDMSK) >> 8 ) - 1 ) #define DB_X2I(d,s) ( ((s) & OPENISIS_SCIDMSK) | ((d) & OPENISIS_DBIDMSK) ) #define DB_I2X(d) ( (d) & OPENISIS_DBIDMSK ) #define DB_IDX_OK(st,idx) \ (0 <= (idx) && (st)->head.ndbs > (idx) && * (st)->head.dbs[idx]->name) /* ************************************************************ * client view of db */ typedef struct { Db head; } CDb; static void cdbCtor (CDb *that, const char *name, int dbid, int tms, Fdt *fdt, Rec *cfg ) { strncpy (that->head.name, name, DBNLEN - 1) [DBNLEN - 1] = 0; that->head.dbid = dbid; that->head.tms = tms; that->head.fdt = fdt; that->head.cfg = cfg; } static void cdbDtor (CDb *that) { if (that->head.fdt) { fFree (that->head.fdt); } if (that->head.cfg) { mFree (that->head.cfg); } memset (that, 0, sizeof (CDb)); } /* ************************************************************ * client session */ typedef struct OpenIsisStb { Schema head; char host[LSTB_STHOSTLEN]; int port; int srq; /* rqs serial no */ Rec *rqs; /* last request */ Rec *rsp; /* last response */ Db *rdb; /* db of embedded rsp recs */ CDb *dbase; /* db array of non-local schema */ LutiLT hsh; CliChnl chn; OpenIsisStbRqsProc *rqsp; OpenIsisStbDelProc *delp; void *cld; /* client_data of rqsp */ /*MMM deprecated*/ OpenIsisRspCb *actcb; void *actcld; OpenIsisStubCbData cbdta; } OpenIsisStb; Stub stub0 = 0; static Stub RmtStubs = 0; static int NumStubs = 0; static LutiLT StubLT = 0; #define STBSIZE (sizeof(Stb)) static int stbLocalRqs (Stub that, Rec *rqs); static int stbSyncRqs (Stub that, Rec *rqs); static void stbClear (Stub that) { memset (that, 0, STBSIZE); that->chn.sd = -1; } static int stbCtor (Stub that, const char *name, int scid, Rec *cfg, OpenIsisStbRqsProc *rqsp, OpenIsisStbDelProc *delp, void *cld, OpenIsisStubCbData *cbd ) { strncpy (that->head.name, name, SCNLEN - 1) [SCNLEN - 1] = 0; that->head.scid = scid; if (scid) { if (! rString (cfg, OPENISIS_SC_HOST, 0, that->host, LSTB_STHOSTLEN)) { return 0; } that->port = rInt (cfg, OPENISIS_SC_PORT, 0, 0); if (0 >= that->port) { return 0; } if (rqsp) { that->rqsp = rqsp; } else { that->rqsp = &stbSyncRqs; } } else { that->rqsp = &stbLocalRqs; } that->head.cfg = rSet (cfg, RCHG | RDIS, OPENISIS_SC_NAME, that->head.name, 0); that->delp = delp; that->cld = cld; that->hsh = luti_ltnew (); if (cbd) { memcpy (&that->cbdta, cbd, sizeof (OpenIsisStubCbData)); } luti_ltadd (StubLT, that->head.name, scid); return !0; } /* return idx or -err */ static int cdbNewIdx (Stub that) { int ndb = that->head.ndbs; int rt, j; for (j = 0; ndb > j; ++j) { if (! * that->dbase[j].head.name) { return j; } } rt = luti_ptrincr (&that->dbase, &ndb, LSTB_INCRDB, sizeof (CDb), LSTB_MAXDBS); if (0 > rt) { return sMsg (ERR_NOMEM, "%s: cdbNewIdx: out of memory", that->head.name); } rt = luti_ptrincr (&that->head.dbs, &that->head.ndbs, LSTB_INCRDB, sizeof (CDb*), LSTB_MAXDBS); if (0 > rt) { return sMsg (ERR_NOMEM, "%s: cdbNewIdx: out of memory", that->head.name); } for (j = that->head.ndbs; 0 <= --j; ) { that->head.dbs[j] = (Db*) (that->dbase + j); } return rt; } /* return bool newly allocated or -err */ static int cdbNewIdx0 (Stub that, int dbid) { int rt; if (dbid < that->head.ndbs) { if (that->head.dbs[dbid]) { return 0; } return 1; } rt = luti_ptrincr (&that->head.dbs, &that->head.ndbs, 1 + dbid - that->head.ndbs, sizeof (CDb*), -1); if (0 > rt) { return sMsg (ERR_NOMEM, "%s: cdbNewIdx0(%d): out of memory", that->head.name, dbid); } return 1; } static int stbNewIdx () { int j; for (j = 0; NumStubs > j; ++j) { if (! * RmtStubs[j].head.name) { return j; } } j = luti_ptrincr (&RmtStubs, &NumStubs, LSTB_INCRSTUB, STBSIZE, LSTB_MAXSTUBS); /* stbClear (RmtStubs + j) done in nOpen */ return j; } static Stub stbById (int sid) { int idx; if (0 > sid || ! stub0) { return 0; } if (0 == sid) { return stub0; } idx = SC_I2X (sid); if (0 > idx) { log_msg (LOG_ERROR, "stbById: illegal id %x", sid); return 0; } if (NumStubs <= idx) { return 0; } if (! * RmtStubs[idx].head.name) { return 0; } return RmtStubs + idx; } static void stbDtor (Stub that) { luti_ltrmv (StubLT, that->head.name); if (that->cbdta.delcb) { if (that->actcld) { (*that->cbdta.delcb) ( that->cbdta.delcld, that, that->actcld); } if (that->cbdta.dfltcld) { (*that->cbdta.delcb) ( that->cbdta.delcld, that, that->cbdta.dfltcld); } (*that->cbdta.delcb) (that->cbdta.delcld, that, 0); } if (that->dbase) { int j; for (j = that->head.ndbs; 0 <= --j; ) { if (* that->dbase[j].head.name) { cdbDtor (that->dbase + j); } } mFree (that->dbase); } if (that->head.dbs) { mFree (that->head.dbs); } if (that->head.cfg) { mFree (that->head.cfg); } luti_ltdel (that->hsh); nClean (that); stbClear (that); } static int stbDbidFromName (Stub that, const char *dbn) { int id; if (! dbn) { return -1; } id = luti_ltget (that->hsh, dbn); return id; } static int stbDbIdxFromName (Stub that, const char *dbn) { int idx = stbDbidFromName (that, dbn); if (0 > idx) { return -1; } idx = DB_I2X (idx); if (! DB_IDX_OK (that, idx)) { return -1; } return idx; } static int stbOpenDb (Stub that) { char dbn[DBNLEN]; Fdt *fdt; Rec *cfg; int idx, dbid, tms; int newdb = 0; if (! rString (that->rsp, OPENISIS_COM_DBN, 0, dbn, DBNLEN)) { return sMsg (ERR_IDIOT, "%s: stbOpenDb: missing db name in rsp", that->head.name); } idx = stbDbIdxFromName (that, dbn); if (0 > idx) { idx = cdbNewIdx (that); if (0 > idx) { return idx; } newdb = !0; } else { cdbDtor (that->dbase + idx); } fdt = fRec2Fdt (that->rsp); cfg = luti_unwrap (that->rsp, 0, OPENISIS_COM_CFG, -1); tms = rInt (that->rsp, OPENISIS_COM_TMS, 0, 0); dbid = DB_X2I (idx, that->head.scid); cdbCtor (that->dbase + idx, dbn, dbid, tms, fdt, cfg); if (newdb) { luti_ltadd (that->hsh, dbn, dbid); } return 0; } static int stbOpenDb0 (Stub that) { Db *dbh; int newdb = 0; int dbid = rInt (that->rsp, OPENISIS_RSP_DBID, -1, 0); if (0 > dbid) { return sMsg (ERR_IDIOT, "stbOpenDb0: missing dbid in rsp"); } newdb = cdbNewIdx0 (that, dbid); if (0 > newdb) { return newdb; } dbh = ldb_getdb (dbid); if (! dbh) { return sMsg (ERR_TRASH, "stbOpenDb0: no such db %d", dbid); } log_msg (LOG_INFO, "stbOpenDb0 %d %s", dbid, dbh->name); that->head.dbs[dbid] = dbh; if (newdb) { luti_ltadd (that->hsh, dbh->name, dbid); } return 0; } static int stbCloseDb (Stub that) { char dbn[DBNLEN]; CDb *db; int dbid; if (! rString (that->rsp, OPENISIS_COM_DBN, 0, dbn, sizeof (dbn))) { return sMsg (ERR_IDIOT, "%s: stbCloseDb: missing db name in rsp", that->head.name); } dbid = stbDbIdxFromName (that, dbn); if (0 > dbid) { return sMsg (ERR_TRASH, "%s: stbCloseDb: illegal db name <%s> in rsp", that->head.name, dbn); } db = that->dbase + dbid; luti_ltrmv (that->hsh, db->head.name); cdbDtor (db); return 0; } static int stbCloseDb0 (Stub that) { char dbn[DBNLEN]; int dbid = rInt (that->rsp, OPENISIS_RSP_DBID, -1, 0); if (0 > dbid) { return sMsg (ERR_IDIOT, "stbCloseDb0: missing dbid in rsp"); } if (that->head.ndbs <= dbid) { return sMsg (ERR_TRASH, "stbCloseDb0: illegal dbid %d(%d) in rsp", dbid, that->head.ndbs); } if (! rString (that->rsp, OPENISIS_COM_DBN, 0, dbn, DBNLEN)) { return sMsg (ERR_IDIOT, "stbCloseDb0: missing db name in rsp\n"); } log_msg (LOG_INFO, "stbCloseDb0 %d %s", dbid, dbn); luti_ltrmv (that->hsh, dbn); that->head.dbs[dbid] = 0; return 0; } /* (local) response callback */ static int stbRspCb (Rec **rsp, void *cld) { Stub that; Rec *rsp2; int sid, ser, err, err2, err3; (void)cld; sid = rInt (*rsp, OPENISIS_COM_SID, -1, 0); ser = rInt (*rsp, OPENISIS_COM_SER, -1, 0); err = rInt (*rsp, OPENISIS_RSP_ERR, -1, 0); that = stbById (sid); if (! that) { return sMsg (ERR_TRASH, "illegal stub id %d in stbRspCb\n", sid); } if (ser != that->srq) { return sMsg (ERR_TRASH, "%s: rqs serial mismatch %d != %d", that->head.name, ser, that->srq); } if (! that->rqs) { return sMsg (ERR_TRASH, "%s/%d: response without request", that->head.name, sid); } if (that->rsp) { sMsg (ERR_TRASH, "%s: multiple responses", that->head.name); mFree (that->rsp); } that->rdb = 0; that->rsp = *rsp; *rsp = 0; err2 = err3 = 0; if (0 == err) { int rtyp = rInt (that->rqs, OPENISIS_RQS_TYPE, -1, 0); switch (rtyp) { case OPENISIS_RQST_LSDB: break; case OPENISIS_RQST_CLOS: if (sid) { err2 = sMsg (ERR_IDIOT, "%s: db close on remote stub", that->head.name); } else { err2 = stbCloseDb0 (that); } break; case OPENISIS_RQST_OPEN: if (sid) { err2 = sMsg (ERR_IDIOT, "%s: db open on remote stub", that->head.name); break; } err2 = stbOpenDb0 (that); goto setdb; case OPENISIS_RQST_MNT: if (0 == sid) { err2 = sMsg (ERR_IDIOT, "db mount on local stub"); break; } err2 = stbOpenDb (that); goto setdb; default: setdb: { char dbn[DBNLEN]; int rdid, rx; if (! rString (that->rsp, OPENISIS_COM_DBN, 0, dbn, sizeof(dbn))) { if (sid) { err2 = sMsg (ERR_IDIOT, "%s: stbRspCb: missing db name in rsp", that->head.name); } else { err2 = sMsg (ERR_IDIOT, "stbRspCb: missing db name in rsp"); } } else { rdid = stbDbidFromName (that, dbn); if (0 > rdid) { if (sid) { err2 = sMsg (ERR_TRASH, "%s: stbRspCb: illegal db name <%s> in rsp", that->head.name, dbn); } else { err2 = sMsg (ERR_TRASH, "stbRspCb: illegal db name <%s> in rsp", dbn); } } else { /* see nDbById */ rx = DB_I2X (rdid); if (! DB_IDX_OK (that, rx)) { if (sid) { err2 = sMsg (ERR_TRASH, "%s: stbRspCb: illegal db id %d %d <%s>", that->head.name, rdid, rx, dbn); } else { err2 = sMsg (ERR_TRASH, "stbRspCb: illegal db id %d %d <%s>", rdid, rx, dbn); } } else { that->rdb = that->head.dbs[rx]; } } } } } /* switch */ } else { /* err */ int tms; tms = rInt (that->rsp, OPENISIS_COM_TMS, 0, 0); if (tms && sid) { /* remounted */ err2 = stbCloseDb (that); } } rsp2 = rAddI (that->rsp, OPENISIS_RSP_CERR, err2, !0); if (rsp2) { that->rsp = rsp2; } if (that->actcb) { err3 = (*that->actcb) (that->actcld, that, that->rsp, that->rdb); if (that->actcld && that->cbdta.delcb) { (*that->cbdta.delcb) ( that->cbdta.delcld, that, that->actcld); } that->actcb = 0; that->actcld = 0; } else if (that->cbdta.dfltcb) { err3 = (*that->cbdta.dfltcb) (that->cbdta.dfltcld, that, that->rsp, that->rdb); } sMsg (LOG_VERBOSE, "stbRspCb(%s): %d/%d err = %x,%x,%x", that->head.name, sid, ser, err, err2, err3); return err3; } static int stbLocalRqs (Stub that, Rec *rqs) { (void)that; return ldspProcess (rqs, 0, &stbRspCb, 0); } static int stbSyncRqs (Stub that, Rec *rqs) { Rec *rsp; int rt; if (0 > that->chn.sd) { rt = cliConnect (&that->chn, that->host, that->port); if (0 > rt) { return rt; } } rt = cliWrite (&that->chn, rqs); if (0 > rt) { return rt; } rsp = cliRead (&that->chn); if (! rsp) { return sMsg (ERR_IO, "stbSyncRqs: read error %d - %s", that->chn.err, strerror (that->chn.err)); } rt = stbRspCb (&rsp, 0); if (rsp) { mFree (rsp); } return rt; } /* ************************************************************ package functions */ /* ************************************************************ public functions */ Stub nInit ( int argc, const char **argv, OpenIsisStubCbData *dta ) { Rec *cfg = 0; if (stub0) { return stub0; } StubLT = luti_ltnew (); stub0 = mAlloc (STBSIZE); stbClear (stub0); if (0 < argc) { char buf[PATH_MAX]; char *logn; int logl; cfg = rSet (0, RFDT | RARGV | RIGN | argc, openIsisFdtSyspar, argv); logn = rString (cfg, OPENISIS_SLOGF, 0, buf, sizeof(buf)); logl = rInt (cfg, OPENISIS_SLOGV, -1, 0); if (logn || 0 <= logl) { cLog (logl, logn); } } stbCtor (stub0, "", 0, cfg, 0, 0, 0, dta); return stub0; } Stub nOpen ( const char *name, int argc, const char *argv[], OpenIsisStubCbData *dta ) { OpenIsisStb buf; Stub that; Rec *cfg = 0; int idx, sid; if (! stub0 || ! name || ! *name) { return 0; } idx = luti_ltget (StubLT, name); if (0 <= idx) { idx = SC_I2X (idx); if (0 > idx || NumStubs <= idx || ! * RmtStubs[idx].head.name) { log_msg (LOG_ERROR, "nOpen: illegal idx %d(%d) for %s", idx, NumStubs, name); return 0; } return RmtStubs + idx; } idx = stbNewIdx (); if (0 > idx) { return 0; } if (0 < argc) { cfg = rSet (0, RFDT | RARGV | RIGN | argc, openIsisFdtScheme, argv); } sid = SC_X2I (idx); stbClear (&buf); if (! stbCtor (&buf, name, sid, cfg, 0, 0, 0, dta)) { if (cfg) { mFree (cfg); } return 0; } that = RmtStubs + idx; memcpy (that, &buf, STBSIZE); return that; } void nClose (Stub that) { if (that && *that->head.name) { stbDtor (that); } } void nDeinit () { if (stub0) { int j; if (RmtStubs) { for (j = NumStubs; 0 <= --j; ) { if (* RmtStubs[j].head.name) { stbDtor (RmtStubs + j); } } mFree (RmtStubs); RmtStubs = 0; NumStubs = 0; } for (j = stub0->head.ndbs; 0 <= --j; ) { if (stub0->head.dbs[j]) { cDClose (stub0->head.dbs[j]->dbid); } } stbDtor (stub0); mFree (stub0); stub0 = 0; luti_ltdel (StubLT); StubLT = 0; } } Schema* nSchema (Stub that) { int idx; if (that && ( stub0 == that || ( *that->head.name && 0 <= (idx = SC_I2X (that->head.scid)) && NumStubs > idx ))) { return &that->head; } return 0; } void nClean (Stub that) { if (that->rqs) { mFree (that->rqs); that->rqs = 0; } if (that->rsp) { mFree (that->rsp); that->rsp = 0; } that->rdb = 0; } int nSend (Stub that, Rec *rqs, OpenIsisRspCb *actcb, void *actcld, int dup ) { char buf[DBNLEN]; char sid[16]; char ser[16]; char tms[16]; char *dbl, *dbn; int dbid, rtyp, setn; if (!(that && rqs)) { return sMsg (ERR_IDIOT, "nSend: null pointer %x %x", (int)that, (int)rqs); } if (stub0 != that && ! *that->head.name) { return sMsg (ERR_IDIOT, "nSend: invalid stub"); } setn = !0; rtyp = rInt (rqs, OPENISIS_RQS_TYPE, -1, 0); switch (rtyp) { case -1: return sMsg (ERR_IDIOT, "nSend(%s): request without type", that->head.name); case OPENISIS_RQST_LSDB: setn = 0; break; case OPENISIS_RQST_OPEN: case OPENISIS_RQST_CLOS: if (stub0 != that) { return sMsg (ERR_IDIOT, "nSend(%s): request %d denied", that->head.name, rtyp); } break; case OPENISIS_RQST_MNT: if (stub0 == that) { return sMsg (ERR_IDIOT, "nSend: local mount denied"); } rqs = rSet (rqs, RFDT | RDIS | RCHG, openIsisFdtRqs, OPENISIS_COM_TMS, "0", 0); if (! rqs) { return sMsg (ERR_NOMEM, "nSend(%s): out of memory", that->head.name); } break; } dbn = 0; dbl = that->rdb ? that->rdb->name : 0; nClean (that); i2a (sid, that->head.scid); if (0 >= ++that->srq) { that->srq = 1; log_msg (LOG_INFO, "stub %s: srq loop over 0", that->head.name); } i2a (ser, that->srq); if (setn) { if ((dbn = rString (rqs, OPENISIS_COM_DBN, 0, buf, DBNLEN))) { setn = 0; } else { if (dbl) { dbn = dbl; } else { dbn = rString (that->head.cfg, OPENISIS_SC_DFLTDB, 0, buf, DBNLEN); } } if (! dbn) { return sMsg (ERR_IDIOT, "nSend(%s): request without db name", that->head.name); } } dbid = -1; if (0 > rInt (rqs, OPENISIS_COM_TMS, -1, 0)) { dbid = stbDbIdxFromName (that, dbn); if (0 <= dbid) { i2a (tms, that->head.dbs[dbid]->tms); } } if (dup) { rqs = rDup (rqs, 0, 0); if (! rqs) { return -1; } } rqs = rSet (rqs, RFDT | RDIS | RCHG, openIsisFdtRqs, OPENISIS_COM_SID, sid, OPENISIS_COM_SER, ser, 0 <= dbid ? OPENISIS_COM_TMS : 0, tms, 0); if (! rqs) { return sMsg (ERR_NOMEM, "nSend(%s): out of memory", that->head.name); } if (setn) { rqs = rSet (rqs, RFDT | RDIS | RCHG, openIsisFdtRqs, OPENISIS_COM_DBN, dbn, 0); if (! rqs) { return sMsg (ERR_NOMEM, "nSend(%s): out of memory", that->head.name); } } that->rqs = rqs; if (that->actcld && that->cbdta.delcb) { /* ooups */ (*that->cbdta.delcb) ( that->cbdta.delcld, that, that->actcld); } if (actcb) { that->actcb = actcb; that->actcld = actcld; } else { that->actcb = 0; that->actcld = 0; } return (*that->rqsp) (that, that->rqs); } Rec* nRecv (Stub that, Db **db) { if (that && ( stub0 == that || *that->head.name )) { if (db) { *db = that->rdb; } return that->rsp; } return 0; } Db* nDbByName (Stub that, const char *dbname) { if (that && ( stub0 == that || *that->head.name )) { int dbid = stbDbIdxFromName (that, dbname); if (0 <= dbid) { return that->head.dbs[dbid]; } } return 0; } Db* nDbById (int id) { Stub that = stbById (id & OPENISIS_SCIDMSK); if (! that) { return 0; } id = DB_I2X (id); if (! DB_IDX_OK (that, id)) { return 0; } return that->head.dbs[id]; } #define FREEPTR(p) if (p) { mFree (p); p = 0; } #define FREEROW if (rowids) { FREEPTR(*rowids) } int nResult ( Stub that, int **rowids, Rec ***recs, Db **db, int *tot ) { Field *F; Rec **R; int *I; int numt, numr, err, pos, j; if (rowids) { *rowids = 0; } if (recs) { *recs = 0; } if (db) { *db = 0; } if (tot) { *tot = 0; } if (! that) { return sMsg (ERR_IDIOT, "nGetResult: null pointer"); } if (! that->rsp) { return sMsg (ERR_IDIOT, "nGetResult(%s): no response", that->head.name); } if (! that->rdb) { return sMsg (ERR_IDIOT, "nGetResult(%s): no db in response", that->head.name); } err = rInt (that->rsp, OPENISIS_RSP_ERR, 0, 0); if (err) { return sMsg (ERR_IDIOT, "nGetResult(%s): response error %d", that->head.name, err); } numr = rInt (that->rsp, OPENISIS_RSP_NUMR, -1, 0); numt = rInt (that->rsp, OPENISIS_RSP_NUMT, -1, 0); if (0 > numr || 0 > numt) { return sMsg (ERR_IDIOT, "nGetResult(%s): no record count in response", that->head.name); } if (numr) { if (rowids) { *rowids = (int*) mAlloc (numr * sizeof (int)); if (! *rowids) { return sMsg (ERR_NOMEM, "nGetResult(%s): out of memory", that->head.name); } for (I = *rowids, j = pos = 0; numr > j; ++I, ++j) { *I = rInt (that->rsp, OPENISIS_COM_ROW, -1, &pos); if (0 > *I) { FREEROW; return sMsg (ERR_TRASH, "nGetResult(%s): missing rowid %d(%d)", that->head.name, j, numr); } } } if (recs) { *recs = (Rec**) mAlloc (numr * sizeof (Rec*)); if (! *recs) { FREEROW; return sMsg (ERR_NOMEM, "nGetResult(%s): out of memory", that->head.name); } memset (*recs, 0, numr * sizeof (Rec*)); R = *recs; I = rowids ? *rowids : 0; j = pos = 0; F = rGet (that->rsp, OPENISIS_COM_REC, &pos); if (! F) { sMsg (LOG_WARN, "nGetResult(%s): no records in response", that->head.name); FREEPTR (*recs); } else { --pos; while (numr > j) { *R = luti_unwrap (that->rsp, &pos, OPENISIS_COM_REC, that->rdb->dbid); if (! *R) { FREEROW; luti_free ((void**)*recs, j); *recs = 0; return sMsg (ERR_NOMEM, "nGetResult(%s): out of memory", that->head.name); } if (I) { (*R)->rowid = *I; ++I; } ++R; ++j; } } } /* recs */ } /* numr */ if (db) { *db = that->rdb; } if (tot) { *tot = numt; } return numr; } lstbtest.c0100644000113000011300000004145107656730161011563 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lstbtest.c,v 1.21 2003/05/08 18:00:49 mawag Exp $ tests for lstb methods. */ #include #include #include #include #include #include /* unlink */ #include "openisis.h" #include "loi.h" #include "ldb.h" /* ldb_getdb */ #include "luti.h" /* luti_unwrap */ static int _RspCld[1] = { 0 }; static int _DelCld[1] = { 0 }; static OpenIsisStubCbData cbdta; static Stub st0, st1, st2, st3; static Schema *sc0, *sc1, *sc2, *sc3; static Db *db; static Rec *rqs, *rsp; static char *args[64]; static int alen; /* ************************************************************ utils */ static void _lstb_assert (int cond, const char *where, char *fmt, ...) { if (! cond) { va_list ap; va_start (ap, fmt); fprintf (stderr, "ERR %s: ", where ? where : ""); vfprintf (stderr, fmt, ap); fprintf (stderr, "\n"); va_end (ap); exit (1); } } static int _rsp_cb (void *cld, Stub stb, Rec *rsprec, Db *x) { (void)stb; (void)rsprec; (void)x; if (cld) { ++(*((int*)cld)); } return 0; } static void _del_cb (void *cld, Stub stb, void *cbd) { (void)stb; ++(*((int*)cld)); if (cbd) { _lstb_assert (cbd == (void*)_RspCld, "DELCB", "%d: %p != %p", *((int*)cld), cbd, _RspCld); --(*((int*)cbd)); } } static void _cmp_rec (const char *msg, Rec *r1, Rec *r2) { char buf1[32], buf2[32]; Field *f1, *f2; int j; _lstb_assert (0 != r1, msg, "chkr: r1 == 0"); _lstb_assert (0 != r2, msg, "chkr: r2 == 0"); _lstb_assert (r1->len == r2->len, msg, "chkr: reclen = %d/%d", r1->len, r2->len); buf1[31] = buf2[31] = 0; for (f1 = r1->field, f2 = r2->field, j = 0; r1->len > j; ++f1, ++f2, ++j) { _lstb_assert (f1->tag == f2->tag, msg, "chkr: tag[%d] = %d/%d", j, f1->tag, f2->tag); _lstb_assert (f1->len == f2->len, msg, "chkr: len[%d] = %d/%d", j, f1->len, f2->len); strncpy (buf1, f1->val, 31 < f1->len ? 31 : f1->len); strncpy (buf2, f2->val, 31 < f1->len ? 31 : f1->len); _lstb_assert (0 == memcmp (f1->val, f2->val, f1->len), msg, "field[%d] = %s/%s", j, buf1, buf2); } } static void _chk_rec (const char *msg, Rec *rec, int dbid, int rowid) { Rec *r2 = dRead (dbid, rowid); _cmp_rec (msg, rec, r2); mFree (r2); } static Rec* _buildrqs (int argc, char **argv) { return rSet (0, RARGV | RFDT | argc, openIsisFdtRqs, argv); } static int _set_tms (const char *msg, char *dbn, int delta) { static char tbuf[32]; int j; _lstb_assert (0 != sc0, msg, "(tms) sc0 == 0"); _lstb_assert (0 < sc0->ndbs, msg, "(tms) sc0->ndbs = %d", sc0->ndbs); for (j = 0, db = 0; sc0->ndbs > j; ++j) { if (0 == strcmp (sc0->dbs[j]->name, dbn)) { db = sc0->dbs[j]; break; } } _lstb_assert (0 != db, msg, "(tms) no such db: %s", dbn); _lstb_assert (0 != db->tms, msg, "(tms) tms == 0"); sprintf (tbuf, "%d", (db->tms + delta)); args[alen++] = "tms"; args[alen++] = tbuf; return db->tms; } /* ************************************************************ tests */ static void _open_srv (const char *msg, int argc, const char **argv) { const char *scargs[] = { "host", "tonno", "port", "4242" }; int scnum = sizeof (scargs) / sizeof (scargs[0]); st0 = openIsisNInit (argc, argv, &cbdta); _lstb_assert (0 != st0, msg, "st0 == 0"); sc0 = openIsisNSchema (st0); _lstb_assert (0 != sc0, msg, "sc0 == 0"); _lstb_assert (0 == sc0->scid, msg, "sc0.scid = %x", sc0->scid); st1 = openIsisNOpen ("nonno", scnum, scargs, &cbdta); _lstb_assert (0 != st1, msg, "st1 == 0"); _lstb_assert (st0 != st1, msg, "st1 == st0"); sc1 = openIsisNSchema (st1); _lstb_assert (0 != sc1, msg, "sc1 == 0"); _lstb_assert (0x0100 == sc1->scid, msg, "sc1.scid = %x", sc1->scid); st2 = openIsisNInit (0, 0, 0); _lstb_assert (st2 == st0, msg, "st2 %p != %p", st2, st0); sc2 = openIsisNSchema (st2); _lstb_assert (sc2 == sc0, msg, "sc2 %p != %p", sc2, sc0); st3 = openIsisNOpen ("nonno", 0, 0, 0); _lstb_assert (st3 == st1, msg, "st3 %p != %p", st3, st1); sc3 = openIsisNSchema (st1); _lstb_assert (sc3 == sc1, msg, "sc3 %p != %p", sc3, sc1); } static void _open_db (const char *msg, char *dbn) { int rt; alen = 0; args[alen++] = "type"; args[alen++] = "open"; args[alen++] = "db"; args[alen++] = dbn; rqs = _buildrqs (alen, args); _lstb_assert (0 != rqs, msg, "rqs == 0"); rt = openIsisNSend (st0, rqs, &_rsp_cb, _RspCld, 0); _lstb_assert (0 == rt, msg, "rt = %d", rt); rsp = openIsisNRecv (st0, 0); _lstb_assert (0 != rsp, msg, "rsp == 0"); rt = rInt (rsp, OPENISIS_RSP_ERR, -1, 0); _lstb_assert (0 == rt, msg, "err = %d", rt); db = ldb_getdb (0); _lstb_assert (0 != db, msg, "db == 0"); _lstb_assert (0 == strcmp (db->name, "cds"), msg, "db0: %s", db->name); } static void _close_db (const char *msg, char *dbn, int dbid) { int rt; alen = 0; args[alen++] = "type"; args[alen++] = "close"; args[alen++] = "db"; args[alen++] = dbn; rqs = _buildrqs (alen, args); _lstb_assert (0 != rqs, msg, "rqs == 0"); rt = openIsisNSend (st0, rqs, &_rsp_cb, _RspCld, 0); _lstb_assert (0 == rt, msg, "rt = %d", rt); rsp = openIsisNRecv (st0, 0); _lstb_assert (0 != rsp, msg, "rsp == 0"); rt = rInt (rsp, OPENISIS_RSP_ERR, -1, 0); _lstb_assert (0 == rt, msg, "err = %d", rt); if (0 <= dbid) { db = ldb_getdb (dbid); _lstb_assert (0 == db, msg, "db != 0"); } } static void _lsdb (const char *msg) { char dbn[OPENISIS_DB_NAMELEN]; char *n1; int rt, pos; alen = 0; args[alen++] = "type"; args[alen++] = "ls"; rqs = _buildrqs (alen, args); _lstb_assert (0 != rqs, msg, "rqs == 0"); rt = openIsisNSend (st0, rqs, 0, 0, 0); _lstb_assert (0 == rt, msg, "rt = %d", rt); rsp = openIsisNRecv (st0, 0); _lstb_assert (0 != rsp, msg, "rsp == 0"); pos = 0; n1 = rString (rsp, OPENISIS_COM_DBN, &pos, dbn, OPENISIS_DB_NAMELEN); _lstb_assert (0 != n1, msg, "db(0) == 0"); _lstb_assert (0 == strcmp (n1, "cds"), msg, "db(0) = %s", n1); n1 = rString (rsp, OPENISIS_COM_DBN, &pos, dbn, OPENISIS_DB_NAMELEN); _lstb_assert (0 != n1, msg, "db(1) == 0"); _lstb_assert (0 == strcmp (n1, "lstbtest"), msg, "db(1) = %s", n1); n1 = rString (rsp, OPENISIS_COM_DBN, &pos, dbn, OPENISIS_DB_NAMELEN); _lstb_assert (0 == n1, msg, "db(2) = %s", n1); } static void _maxrow (const char *msg, char *dbn, int exp) { int rt; alen = 0; args[alen++] = "type"; args[alen++] = "maxrow"; args[alen++] = "db"; args[alen++] = dbn; rqs = _buildrqs (alen, args); _lstb_assert (0 != rqs, msg, "rqs == 0"); rt = openIsisNSend (st0, rqs, &_rsp_cb, _RspCld, 0); _lstb_assert (0 == rt, msg, "rt = %d", rt); rsp = openIsisNRecv (st0, 0); _lstb_assert (0 != rsp, msg, "rsp == 0"); rt = rInt (rsp, OPENISIS_RSP_ERR, -1, 0); _lstb_assert (0 == rt, msg, "err = %d", rt); rt = rInt (rsp, OPENISIS_RSP_NUMT, -1, 0); _lstb_assert (-1 == rt, msg, "numt = %d", rt); rt = rInt (rsp, OPENISIS_COM_ROW, -1, 0); if (0 <= exp) { _lstb_assert (exp == rt, msg, "rowid = %d(%d)", rt, exp); } else { _lstb_assert (0 <= rt, msg, "rowid = %d", rt); } } static void _maxrowerr (const char *msg, char *dbn) { int rt, tms; alen = 0; args[alen++] = "type"; args[alen++] = "maxrow"; args[alen++] = "db"; args[alen++] = dbn; tms = _set_tms (msg, dbn, -2); rqs = _buildrqs (alen, args); _lstb_assert (0 != rqs, msg, "rqs == 0"); rt = openIsisNSend (st0, rqs, &_rsp_cb, _RspCld, 0); _lstb_assert (0 == rt, msg, "rt = %d x%x", rt, rt); rsp = openIsisNRecv (st0, 0); _lstb_assert (0 != rsp, msg, "rsp == 0"); rt = rInt (rsp, OPENISIS_RSP_ERR, -1, 0); _lstb_assert (ERR_INVAL == rt, msg, "err = %d", rt); rt = rInt (rsp, OPENISIS_COM_TMS, -1, 0); _lstb_assert (tms == rt, msg, "tms = %d(%d)", rt, tms); } static void _lstb_read (const char *msg, int rid, int rlen) { /* no _rsp_cb */ Rec *rec; Db *rdb; char rbuf[32]; int rt; alen = 0; args[alen++] = "type"; args[alen++] = "read"; args[alen++] = "db"; args[alen++] = "cds"; _set_tms (msg, "cds", 0); sprintf (rbuf, "%d", rid); args[alen++] = "rowid"; args[alen++] = rbuf; rqs = _buildrqs (alen, args); _lstb_assert (0 != rqs, msg, "rqs == 0"); rt = openIsisNSend (st0, rqs, 0, _RspCld, 0); _lstb_assert (0 == rt, msg, "rt = %d", rt); rsp = openIsisNRecv (st0, &rdb); _lstb_assert (0 != rsp, msg, "rsp == 0"); _lstb_assert (0 != rdb, msg, "rdb == 0"); rt = rInt (rsp, OPENISIS_RSP_ERR, -1, 0); _lstb_assert (0 == rt, msg, "err = %d", rt); rt = rInt (rsp, OPENISIS_RSP_NUMT, -1, 0); _lstb_assert (1 == rt, msg, "numt = %d", rt); rt = rInt (rsp, OPENISIS_COM_ROW, -1, 0); _lstb_assert (rid == rt, msg, "rowid = %d(%d)", rt, rid); rec = luti_unwrap (rsp, 0, OPENISIS_COM_REC, rdb->dbid); _lstb_assert (0 != rec, msg, "rec == 0"); if (0 < rlen) { _lstb_assert (rlen == rec->len, msg, "rec->len = %d(%d)", rec->len, rlen); } else { _lstb_assert (0 < rec->len, msg, "rec->len = %d", rec->len); } _lstb_assert (rdb->dbid == rec->dbid, msg, "rec->dbid = %d(%d)", rec->dbid, rdb->dbid); _chk_rec (msg, rec, 0, rid); mFree (rec); } static int _query ( /* no _rsp_cb */ const char *msg, int flg, int mode, int skip, int size, int exp ) { Rec **recs; Db *rdb; char fbuf[32], mbuf[32], kbuf[32], sbuf[32]; int *rowids; int rt, rr, j; alen = 0; args[alen++] = "type"; args[alen++] = "query"; args[alen++] = "db"; args[alen++] = "cds"; sprintf (mbuf, "%d", mode); args[alen++] = "mode"; args[alen++] = mbuf; if (flg) { sprintf (fbuf, "%d", flg); args[alen++] = "flags"; args[alen++] = fbuf; } if (skip) { sprintf (kbuf, "%d", skip); args[alen++] = "skip"; args[alen++] = kbuf; } if (size) { sprintf (sbuf, "%d", size); args[alen++] = "size"; args[alen++] = sbuf; } args[alen++] = "key"; args[alen++] = "Africa$"; rqs = _buildrqs (alen, args); _lstb_assert (0 != rqs, msg, "rqs == 0"); rt = openIsisNSend (st0, rqs, 0, _RspCld, 0); _lstb_assert (0 == rt, msg, "rt = %d", rt); rsp = openIsisNRecv (st0, &rdb); _lstb_assert (0 != rsp, msg, "rsp == 0"); _lstb_assert (0 != rdb, msg, "rdb == 0"); rt = rInt (rsp, OPENISIS_RSP_ERR, -1, 0); _lstb_assert (0 == rt, msg, "err = %d", rt); rr = openIsisNGetResult (st0, &rowids, &recs, 0, &rt); if (0 <= exp) { _lstb_assert (exp == rr, msg, "numr = %d(%d)", rr, exp); } else { _lstb_assert (0 <= rr, msg, "numr = %d", rr); } _lstb_assert (rt >= rr, msg, "numt = %d/%d", rt, rr); if (recs) { for (j = 0; rr > j; ++j) { _chk_rec (msg, recs[j], 0, rowids[j]); mFree (recs[j]); } mFree (recs); } rt = rr ? rowids[rr - 1] : -1; mFree (rowids); return rt; } static int _update (const char *msg, int rid, int rtag, char *rval, int itag, char *ival, int exp ) { Rec *rec; char rbuf[32]; int rt; alen = 0; args[alen++] = "type"; args[alen++] = 0 <= rid ? "update" : "insert"; args[alen++] = "db"; args[alen++] = "lstbtest"; _set_tms (msg, "lstbtest", 0); if (0 <= rid) { sprintf (rbuf, "%d", rid); args[alen++] = "rowid"; args[alen++] = rbuf; } rqs = _buildrqs (alen, args); rec = rSet (0, 0, rtag, rval, 88, "summary of the conference in gondor", 0); rqs = luti_wrap (rqs, rec, OPENISIS_COM_REC); mFree (rec); if (itag) { rec = rSet (0, 0, itag, ival, 0); rqs = luti_wrap (rqs, rec, OPENISIS_COM_REC); mFree (rec); } rt = openIsisNSend (st0, rqs, &_rsp_cb, _RspCld, 0); _lstb_assert (0 == rt, msg, "rt = %d", rt); rsp = openIsisNRecv (st0, 0); _lstb_assert (0 != rsp, msg, "rsp == 0"); rt = rInt (rsp, OPENISIS_RSP_ERR, -1, 0); _lstb_assert (0 == rt, msg, "err = %d", rt); rt = rInt (rsp, OPENISIS_COM_ROW, -1, 0); if (0 < exp) { _lstb_assert (exp == rt, msg, "rowid = %d(%d)", rt, exp); } else { _lstb_assert (0 < rt, msg, "rowid = %d", rt); } return rt; } static void _rmtstdb () { char path[1024]; char *p; int err; if (! sc0) { fprintf (stderr, "rmtstdb: st0 == 0\n"); return; } if (! sc0->cfg) { fprintf (stderr, "rmtstdb: st0->cfg == 0\n"); return; } if (! rString (sc0->cfg, OPENISIS_SPATH, 0, path, 990)) { fprintf (stderr, "rmtstdb: path not given\n"); return; } p = path + strlen (path); *p++ = '/'; if (rString (sc0->cfg, OPENISIS_DPATH, 0, p, 10)) { p += strlen (p); *p++ = '/'; } err = 0; strcpy (p, "lstbtest.mst"); errno = 0; if (unlink (path)) { fprintf (stderr, "rmtstdb: unlink mst = %d\n", errno); err = 1; } strcpy (p, "lstbtest.oxi"); errno = 0; if (unlink (path)) { fprintf (stderr, "rmtstdb: unlink oxi = %d\n", errno); err = 1; } strcpy (p, "lstbtest.xrf"); errno = 0; if (unlink (path)) { fprintf (stderr, "rmtstdb: unlink xrf = %d\n", errno); err = 1; } if (! err) { log_msg (LOG_INFO, "db lstbtest removed."); } } static void _tstUnwrap () { Rec *r1, *r2, *r31, *r32, *r4; r31 = rSet (0, 0, 11, "r3111", 12, "r3112", 13, "r3113", 0); r32 = rSet (0, 0, 11, "r3211", 12, "r3212", 13, "r3213", 0); r2 = rSet (0, 0, 21, "r221", 22, "r222", 0); r2 = luti_wrap (r2, r31, 89); r2 = rSet (r2, 0, 23, "r223", 0); r2 = luti_wrap (r2, r32, 89); r1 = 0; r1 = luti_wrap (r1, r2, 79); r1 = rSet (r1, 0, 111, "r1111", 112, "r1112", 0); r4 = luti_getembed (r1, ".79.89[1]", 0); _cmp_rec ("unwrap", r4, r32); mFree (r1); mFree (r2); mFree (r31); mFree (r32); mFree (r4); } int main (int argc, char **argv) { char *argd[64]; int lastid, lvl, j; lvl = LOG_ERROR; switch (argc) { case 2: j = atoi (argv[1]); goto dfltargs; case 0: case 1: j = 1; dfltargs: argc = 0; argd[argc++] = "syspath"; argd[argc++] = "/opt/openisis/db"; argd[argc++] = "dbpath"; argd[argc++] = "cds"; argv = argd; break; default: if ((j = atoi (argv[1]))) { --argc; ++argv; if (0 < (lastid = atoi (argv[2]))) { lvl = lastid; --argc; ++argv; } } else { j = 1; } --argc; ++argv; } cLog (lvl, 0); do { _RspCld[0] = _DelCld[0] = 0; st0 = st1 = st2 = st3 = 0; sc0 = sc1 = sc2 = sc3 = 0; db = 0; rqs = rsp = 0; openIsisNInit (0, 0, 0); openIsisNDeinit (); memset (&cbdta, 0, sizeof (OpenIsisStubCbData)); cbdta.delcb = &_del_cb; cbdta.delcld = _DelCld; _open_srv ("A", argc, (const char**)argv); openIsisNDeinit (); _lstb_assert (0 == stub0, "B", "stub0 != 0"); _lstb_assert (2 == _DelCld[0], "B", "_DelCld = %d", _DelCld[0]); sc1 = openIsisNSchema (st1); _lstb_assert (sc1 == 0, "B", "sc1 != 0"); _open_srv ("C", argc, (const char**)argv); openIsisNClose (st3); _lstb_assert (3 == _DelCld[0], "D", "_DelCld = %d", _DelCld[0]); sc1 = openIsisNSchema (st1); _lstb_assert (sc1 == 0, "D", "sc1 != 0"); _open_db ("E", "cds"); _close_db ("F", "cds", 0); _open_db ("G", "cds"); _close_db ("H", "cds", 0); _open_db ("I", "cds"); _lstb_assert (0 == _RspCld[0], "I", "_RspCld = %d", _RspCld[0]); _lstb_assert (8 == _DelCld[0], "I", "_DelCld = %d", _DelCld[0]); _maxrow ("J", "cds", 154); _lstb_read ("K", 94, 7); _lstb_assert (0 == _RspCld[0], "K", "_RspCld = %d", _RspCld[0]); _lstb_assert (9 == _DelCld[0], "K", "_DelCld = %d", _DelCld[0]); _query ("L", 0, OPENISIS_QRY_KEYAT, 0, 0, 5); lastid = _query ("M", 0, OPENISIS_QRY_KEYAT, 0, 4, 4); _query ("N", 0, OPENISIS_QRY_KEYAT, 1 + lastid, 4, 1); _query ("P", OPENISIS_RQSF_QRR, OPENISIS_QRY_KEYAT, 0, 0, 5); lastid = _query ("Q", OPENISIS_RQSF_QRR, OPENISIS_QRY_KEYAT, 0, 4, 4); _query ("R", OPENISIS_RQSF_QRR, OPENISIS_QRY_KEYAT, 1 + lastid, 4, 1); _lstb_assert (0 == _RspCld[0], "R", "_RspCld = %d", _RspCld[0]); _lstb_assert (9 == _DelCld[0], "R", "_DelCld = %d", _DelCld[0]); _open_db ("S", "lstbtest"); lastid = _update ("T", -1, 22, "Gandalf", 22, "Frodo", 1); _update ("U", -1, 23, "Larry", 0, 0, 2); _update ("V", lastid, 22, "Harry", 0, 0, 1); _maxrow ("W", "lstbtest", 2); _lsdb ("X"); _close_db ("Y", "lstbtest", 1); _lstb_assert (0 == _RspCld[0], "Y", "_RspCld = %d", _RspCld[0]); _lstb_assert (15 == _DelCld[0], "Y", "_DelCld = %d", _DelCld[0]); _rmtstdb (); _maxrowerr ("Z", "cds"); _lstb_read ("a", 94, 7); _maxrow ("b", "cds", 154); _lstb_read ("c", 94, 7); _tstUnwrap (); openIsisNDeinit (); _lstb_assert (0 == _RspCld[0], "z", "_RspCld = %d", _RspCld[0]); _lstb_assert (18 == _DelCld[0], "z", "_DelCld = %d", _DelCld[0]); if (0 == (j % 100)) { printf ("%d ...\n", j); } } while (--j); printf ("ok.\n"); return 0; } lstr.c0100644000113000011300000002255307644412745010707 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lstr.c,v 1.13 2003/04/08 00:20:53 kripke Exp $ implementation of record cooking. */ #include #include "lstr.h" #include "luti.h" /* ************************************************************ private types */ /* ************************************************************ private data */ static const char * nameMfc[] = { "Mfc", "CTLM", "NMFN", "NMFB", "NMFP", "TYPE", "RCNT", "MFX1", "MFX2", "MFX3" }; static int descMfc[] = { LSTRSIZE( 9, 0, 0 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, LMBRINT, LMBRINT, LMBRSHORT, LMBRSHORT, LMBRINT, LMBRINT, LMBRINT, LMBRINT }; /* descMfc */ static const char * nameMfr[] = { "Mfr", "MFN", "RECL", "BWB", "BWP", "BASE", "NVF", "STAT", "TAG", "'POS", "LEN" }; static int descMfr[] = { LSTRSIZE( 7, 3, 0 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, LMBRSHORT, LMBRINT, LMBRSHORT, LMBRSHORT, LMBRSHORT, LMBRSHORT, LMBRSHORT, LMBRSHORT, LMBRSHORT }; /* descMfr */ /* aligned version */ static int descMfrA[] = { LSTRSIZE( 7, 3, 0 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, LMBRSHORT, LSTRLOFF( LMBRINT, 8 ), LMBRSHORT, LMBRSHORT, LMBRSHORT, LMBRSHORT, LMBRSHORT, LMBRSHORT, LMBRSHORT }; /* descMfr */ static const char * nameXrf[] = { "Xrf", "XPOS", "XREC" }; static int descXrf[] = { LSTRSIZE( 1, 1, 127 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, LMBRINT }; /* descXrf */ static const char ** nameMst[LSTR_MST] = { nameMfc, nameMfr, nameXrf }; static int* descMst[LSTR_MST] = { descMfc, /* MST head */ descMfr, /* MST record */ descXrf /* XRF record */ }; static int* descMstA[LSTR_MST] = { descMfc, /* MST head */ descMfrA, /* MST record */ descXrf /* XRF record */ }; static const char * nameCnt[] = { "Cnt", "TYPE", "ORDN", "ORDF", "N", "K", "LEV", "POSR", "NMAX", "FMAX", "ABNO" }; static int descCnt[] = { LSTRSIZE( 10, 0, 0 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRSHORT, /* TYPE */ LMBRSHORT, /* ORDN */ LMBRSHORT, /* ORDF */ LMBRSHORT, /* N */ LMBRSHORT, /* K */ LMBRSHORT, /* LEV */ LMBRINT, /* POSR */ LMBRINT, /* NMAX */ LMBRINT, /* FMAX */ LMBRSHORT /* ABNO */ }; /* descCnt */ static int descCntA[] = { LSTRSIZE( 10, 0, 0 ), /* sizes (fix, rep, occ) */ 0,28,44, LMBRSHORT, /* TYPE */ LMBRSHORT, /* ORDN */ LMBRSHORT, /* ORDF */ LMBRSHORT, /* N */ LMBRSHORT, /* K */ LMBRSHORT, /* LEV */ LMBRINT, /* POSR */ LMBRINT, /* NMAX */ LMBRINT, /* FMAX */ LMBRSHORT /* ABNO */ }; /* descCntA */ static const char * nameN0[] = { "N0", "POS", "OCK", "TYPE", "'KEY", "REF" }; static int descN01[] = { LSTRSIZE( 3, 2, 10 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, /* POS */ LMBRSHORT, /* OCK */ LMBRSHORT, /* TYPE */ 10, /* KEY */ LMBRINT /* REF */ }; static int descN01A[] = { LSTRSIZE( 3, 2, 10 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, /* POS */ LMBRSHORT, /* OCK */ LMBRSHORT, /* TYPE */ 10, /* KEY */ LSTRLOFF( LMBRINT, 12 ) /* REF */ }; static int descN02[] = { LSTRSIZE( 3, 2, 10 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, /* POS */ LMBRSHORT, /* OCK */ LMBRSHORT, /* TYPE */ 30, /* KEY */ LMBRINT /* REF */ }; static int descN02A[] = { LSTRSIZE( 3, 2, 10 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, /* POS */ LMBRSHORT, /* OCK */ LMBRSHORT, /* TYPE */ 30, /* KEY */ LSTRLOFF( LMBRINT, 32 ) /* REF */ }; static const char * nameL0[] = { "L0", "POS", "OCK", "TYPE", "PS", "'KEY", "INFB", "INFP" }; static int descL01[] = { LSTRSIZE( 4, 3, 10 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, /* POS */ LMBRSHORT, /* OCK */ LMBRSHORT, /* TYPE */ LMBRINT, /* PS */ 10, /* KEY */ LMBRINT, /* INFB */ LMBRINT /* INFP */ }; static int descL01A[] = { LSTRSIZE( 4, 3, 10 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, /* POS */ LMBRSHORT, /* OCK */ LMBRSHORT, /* TYPE */ LMBRINT, /* PS */ 10, /* KEY */ LSTRLOFF( LMBRINT, 12 ), /* INFB */ LMBRINT /* INFP */ }; static int descL02[] = { LSTRSIZE( 4, 3, 10 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, /* POS */ LMBRSHORT, /* OCK */ LMBRSHORT, /* TYPE */ LMBRINT, /* PS */ 30, /* KEY */ LMBRINT, /* INFB */ LMBRINT /* INFP */ }; static int descL02A[] = { LSTRSIZE( 4, 3, 10 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, /* POS */ LMBRSHORT, /* OCK */ LMBRSHORT, /* TYPE */ LMBRINT, /* PS */ 30, /* KEY */ LSTRLOFF( LMBRINT, 32 ), /* INFB */ LMBRINT /* INFP */ }; static const char * nameIfp[] = { "IfH", "NXTB", "NXTP", "TOTP", "SEGP", "SEGC", "MFN", "INFO" }; static int descIfp[] = { #ifdef CONVERT_POSTINGS /* to properly convert postings, the repeated part needs to be a big-endian 64bit int however, converting can't be done in a straightforward fashion anyway, since posting segments have holes of one or two longs at block boundaries */ LSTRSIZE( 5, 2, 1 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, LMBRINT, /* */ LMBRINT, /* */ #else LSTRSIZE( 5, 0, 0 ), /* sizes (fix, rep, occ) */ LSTR_AUTOLENGTHS, #endif LMBRINT, /* */ LMBRINT, /* */ LMBRINT, /* */ LMBRINT, /* */ LMBRINT /* */ }; static const char ** nameInv[LSTR_INV] = { nameCnt, nameN0, nameL0, nameN0, nameL0, nameIfp }; static int* descInv[LSTR_INV] = { descCnt, /* CNT record */ descN01, /* N01 record */ descL01, /* L01 record */ descN02, /* N02 record */ descL02, /* L02 record */ descIfp /* Ifp record */ }; static int* descInvA[LSTR_INV] = { descCntA, /* CNT record */ descN01A, /* N01 record */ descL01A, /* L01 record */ descN02A, /* N02 record */ descL02A, /* L02 record */ descIfp /* Ifp record */ }; /* ************************************************************ private functions */ /* ************************************************************ package data */ const LstrSet lstrlib[LSET_SETS] = { { /* Isis 1 MST */ nameMst, { descMst, descMstA } }, { /* Isis 1 INV */ nameInv, { descInv, descInvA } }, }; /* lstrlib */ /* ************************************************************ package functions */ /* ************************************************************ public functions */ int lstr_auto ( int *str ) { static int pow2[] = { 1, 2, 4, 8 }; int ret = 0; int fix,rep,occ,i; int *xmbrs, txlen, tilen; int nmbrs, off = 0, xlen[2] = {0,0}, blen[2] = {0,0}, x=0; if ( NULL == str ) { #define DOALL( x ) \ for ( i=0; i<(int)(sizeof(x)/sizeof(x[0])); i++ ) \ if ( NULL != x[i] ) ret |= lstr_auto( x[i] ) DOALL( descMst ); DOALL( descInv ); DOALL( descMstA ); DOALL( descInvA ); return ret; } fix = LSTRFIX(*str); rep = LSTRREP(*str); occ = LSTROCC(*str); xmbrs = str + LSTR_XMBR; nmbrs = fix; for ( x=0; ; x++ ) { /* fix and rep */ for ( i=0; i> 16) ) str[LSTR_XRLO] |= xlen[0] << 16; else if ( xlen[0] > (short)(str[LSTR_XRLO] >> 16) ) log_msg( LOG_WARN, "short xlen[0] %hd < %d", (short)(str[LSTR_XRLO] >> 16), xlen[0] ); if ( !(short)str[LSTR_XRLO] ) str[LSTR_XRLO] |= 0xffff & xlen[1]; else if ( xlen[1] > (short)str[LSTR_XRLO] ) log_msg( LOG_WARN, "short xlen[1] %hd < %d", (short)str[LSTR_XRLO], xlen[1] ); /* build totals applying occ */ txlen = (short)(str[LSTR_XRLO] >> 16) + (int)occ*(short)str[LSTR_XRLO]; if ( !str[LSTR_XLEN] ) str[LSTR_XLEN] = txlen; else if ( txlen > str[LSTR_XLEN] ) log_msg( LOG_WARN, "short xlen %d < %d", str[LSTR_XLEN], txlen ); tilen = LSTRLEN(*str) + blen[0] + (int)occ*blen[1]; if ( !str[LSTR_ILEN] ) str[LSTR_ILEN] = tilen; else if ( tilen > str[LSTR_ILEN] ) log_msg( LOG_WARN, "short ilen %d < %d", str[LSTR_ILEN], tilen ); LOG_DBG( LOG_ALL, "%08x xrlo %08x xl %d il %d", str[0], str[1], str[2], str[3] ); return ret; } /* lstr_auto */ lstr.h0100644000113000011300000002233407644412745010711 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LSTR_H /* $Id: lstr.h,v 1.15 2003/04/08 00:20:53 kripke Exp $ definition of ISIS internal structures for openisis lib implementation. This header defines the common internal data structures and their members and external structure and member definition types. A internal data structure is the common in-memory representation of ISIS data; it may be backed by different external (in-file) structures. An in-memory data structure is a contigous int* chunk of memory, consisting of the sizes (one int), an array of members (int) and some buffer (array of char). A data structures sizes consists of the lengths fix and rep of the fixed and repeated part, resp., and the number of occurences of the repeated part. It is used in both an abstract description of a data type and each instance of a data structure. If the number of occurences of the repeated part is not fixed, it is 0 in the abstract case. The lengths are suitable as offsets into the lmbr array. This structure is designed to fit within 32 bits. The sizing is the only common meta-data description of a logical data type. For readability of code, however, the member indexes are enum-erated. When actually used, the type has to match the code's assumptions anyway. The numbers fix and rep are fixed for a given data structure. Although the logical data type has more properties, especially the type of each member (numeric, C-string or raw mem), a representation of this data is needed only when converting external data and thus is stored (somewhat redundantly) with the external properties. There are only two types of members: numeric and memory (mostly some character data). both are represented by a int, which, for memory, denotes an offset (counted in BYTES, not ints) from the records start (NOT offset into the data buffer, as with the isis mst record). The member array consists of a fixed number fix of initial members and a number occ of occurences of member arrays of fixed length rep. While occ is usually fixed, it varies for masterfile records. Note that this layout somewhat resembles the isis record itself, which supports a single level of nesting and repetition. */ /* id of a structure */ /* construction */ #define LSTRID( set, no, vari ) \ (((int)(vari))<<16 | (set)<<8 | (no) ) /* access */ #define LSTRNO( l ) (char)(l) /* no of record type in set */ #define LSTRSET( l ) (char)((l)>>8) /* record set number */ #define LSTRVAR( l ) (short)((l)>>16) /* variant */ /* size of a structure */ /* construction */ #define LSTRSIZE( fix, rep, occ ) \ (((int)(occ))<<16 | (rep)<<8 | (fix)) /* access */ #define LSTRFIX( l ) (char)(l) /* length of fixed part */ #define LSTRREP( l ) (char)((l)>>8) /* length of repeated part */ #define LSTROCC( l ) (short)((l)>>16) /* the members */ /* actual number of members */ #define LSTRMEMS( l ) (LSTRFIX(l) + LSTRREP(l)*LSTROCC(l)) /* in-memory size of members (w/o buffer). */ #define LSTRLEN( t ) ((int)sizeof(int)*(1 + LSTRMEMS(t))) /** Description of a external members layout is encoded as a short "len" and a short offset "off" into the data or repeated part, resp. offset and len are combined into one int. If offsets are 0, they are computed based on the assumption, that members are layouted packed in the declared ordering. For character data, len just gives the length of the member in bytes (up to 2**15), with the highest bit cleared. A length of 0 denotes character data of variable size. Numerical data is marked with the highest bit set. For numbers, the lowest two bits give the ld of the length in bytes (i.e. 0=1,1=2,2=4,3=8 bytes). The offset and length of the actually used bits within these bytes are each encoded in six bits, which is enough for 0..63. A length of 0 (as resulting from 6-bit overflow on '64') means no masking. The 15th bit, if set, denotes high endian data. Note that the only reason for stuffing all that into a short is, that it is sufficient. The amount of space saved is minimal and depending on the actual processor there may be a slight performance penalty for the bit operations. */ #define LONG2LEN( l ) ((short)(l)) #define LONG2OFF( l ) ((short)((l) >> 16)) /* constants */ #define LMBR_FNUM 0x8000 /* flag numeric data */ #define LMBR_FHIE 0x4000 /* flag high endian data */ /* construction */ #define LMBRNUM( ld ) (LMBR_FNUM | ld) /* standard nueric types: the world according to Java :) */ #define LMBRBYTE LMBRNUM(0) #define LMBRSHORT LMBRNUM(1) #define LMBRINT LMBRNUM(2) #define LMBRLONG LMBRNUM(3) /* bit mask construction for numerical data */ #define LMBRBITS( off, len, ld ) (LMBR_FNUM | (off)<<8 | (len)<<2 | (ld)) #define LMBRBITS4( off, len ) (LMBR_FNUM | (off)<<8 | (len)<<2 | 2) /* if you have to specify the offset explicitly */ #define LSTRLOFF( len, off ) ((len) | (off)<<16) /* access */ #define LMBRISNUM( f ) (LMBR_FNUM & (f)) #define LMBRLD( f ) (0x3 & (f)) /* bit access for numerical data */ #define LMBRISBITS( f ) (0x3ffc & (f)) #define LMBRBITLEN( f ) (0x003f & ((f) >> 2)) #define LMBRBITOFF( f ) (0x003f & ((f) >> 8)) /** Description of an external structure layout. This is usually fixed for a given (implementation of a) type; for the isis mst record, however, the occurences of the repeated part and the total external and internal lengths vary for each record. */ enum { LSTR_SIZE, /* sizes */ LSTR_XRLO, /* repeated part byte length and offset <<16; auto */ LSTR_XLEN, /* ext. total length of data; auto */ LSTR_ILEN, /* int. total length of data; auto */ LSTR_XMBR /* members */ }; #define LSTR_AUTOLENGTHS 0,0,0 #define LSTR_LONGS( totmbrs ) (4+(totmbrs)) /* technical variants of a record */ typedef enum { LVAR_PAC, /* packed little endian base variant */ LVAR_ALI, /* aligned variant */ LVAR_BIG, /* big endian aligned */ LVAR_VARI } lstr_variant; /* known record sets */ enum { LSET_MST, /* isis 1 master file */ LSET_INV, /* isis 1 inverted file */ LSET_SETS }; typedef struct { const char ***name; /* an array of names per record type */ int **desc[LVAR_VARI]; /* an array of descs per variant */ } LstrSet; /* data types of ISIS 1 master file */ typedef enum { LSTR_MFC, /* MST head */ LSTR_MFR, /* MST record */ LSTR_XRF, /* XRF record */ LSTR_MST /* count */ } lstr_mst; enum { /* MFC members */ LMFC_CTLM = 1, LMFC_NMFN, /* next available MFN */ LMFC_NMFB, /* next free block in MF counting from 1 */ LMFC_NMFP, /* next free pos in MF */ LMFC_TYPE, /* "0 for user's bases; 1 for messages" */ LMFC_RCNT, /* record count ? not used by CISIS */ LMFC_MFX1, LMFC_MFX2, /* LAN lock */ LMFC_MFX3, /* LAN lock */ LMFC__FL }; enum { /* XRF members */ LXRF_XPOS = 1, LXRF_XREC }; /* data types of ISIS 1 inverted file */ typedef enum { LSTR_CNT, /* CNT record */ LSTR_N01, /* N01 record */ LSTR_L01, /* L01 record */ LSTR_N02, /* N02 record */ LSTR_L02, /* L02 record */ LSTR_IFP, /* IFP record */ LSTR_INV /* count */ } lstr_inv; enum { /* CNT members */ LCNT_TYPE = 1, /* 1 or 2: N0x */ LCNT_ORDN, /* fix 5 */ LCNT_ORDF, /* fix 5 */ LCNT_N, /* fix 15 */ LCNT_K, /* fix 5 */ LCNT_LEV, /* depth of index */ LCNT_POSR, /* position of root */ LCNT_NMAX, LCNT_FMAX, LCNT_ABNO, LCNT__FL }; enum { /* N0x members */ LN0X_POS = 1, /* record number */ LN0X_OCK, /* number of keys */ LN0X_TYPE, /* type 1 or 2 */ LN0X__FL, /* offset of repeated part */ LN0X_KEY = 0, /* key chars */ LN0X_REF, /* ref to next node (>0) or leaf (<0) */ LN0X__RL /* length of repeated part */ }; enum { /* L0x members */ LL0X_POS = 1, /* record number */ LL0X_OCK, /* number of keys */ LL0X_TYPE, /* type 1 or 2 */ LL0X_PS, /* type 1 or 2 */ LL0X__FL, /* offset of repeated part */ LL0X_KEY = 0, /* key chars */ LL0X_INFB, /* ifp block */ LL0X_INFP, /* ifp pos */ LL0X__RL /* length of repeated part */ }; enum { /* IFP members */ LIFP_NXTB = 1, /* block of next segment */ LIFP_NXTP, /* pos of next segment */ LIFP_TOTP, /* total postings */ LIFP_SEGP, /* postings in this segment */ LIFP_SEGC, /* postings in this segment */ LIFP__FL, /* offset of repeated part */ LIFP_MFN = 0, /* mfn */ LIFP_ADD, /* additional qualifyers */ LIFP__RL /* length of repeated part */ }; extern const LstrSet lstrlib[LSET_SETS]; /** auto-initialise the given structure or all builtins, iff NULL. */ extern int lstr_auto ( int *str ); #define LSTR_H #endif /* LSTR_H */ lsv.c0100644000113000011300000005772707645531473010543 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: lsv.c,v 1.12 2003/04/11 12:21:15 kripke Exp $ OpenIsis server */ #include #include #include #include #include #include #include #include /* memcpy et al */ #include #include #include "lsv.h" static Con *con[FD_SETSIZE]; /* usually 1024, but up to 1<<16 */ static pthread_mutex_t lsv_mutex_init = PTHREAD_MUTEX_INITIALIZER; static pthread_cond_t lsv_cond_init = PTHREAD_COND_INITIALIZER; static pthread_key_t lsv_key_wrk; static int lsv_term; static siginfo_t lsv_sig; static const char * pdg[] = { "---", "NEW", "RED", "EOF" }; static const char * stg[] = { "NEW", "CON", "INP", "SES", "WRK", "RUN", "COM", "DON", "RET" }; static void lsv_sighand ( int sig, siginfo_t *info, void *ucontext ) { if ( SIGINT == sig || SIGTERM == sig ) lsv_term = !0; if ( info ) lsv_sig = *info; else { /* ??? */ lsv_sig.si_signo = sig; lsv_sig.si_pid = 0; } (void)ucontext; } /* lsv_sighand */ static int lsv_enq ( Que *q, Con *c ) { c->nxt = 0; q->len++; if ( q->tail ) { q->tail->nxt = c; q->tail = c; return 1; } q->head = q->tail = c; return 0; } /* lsv_enq */ static void lsv_penq ( Srv *srv, Con *c ) { Pool *work = c->grp ? &srv->pool : &srv->main; srv->jobs++; c->stg = LSV_STGWRK; if ( c->ses ) c->ses->cur = c; lsv_enq( &work->que, c ); sMsg( LOG_DEBUG, "enq req %d from %s qlen %d", LIO_FD&c->ios.file, c->nam, work->que.len ); if ( work->wait ) { Wrk *w = work->wait; work->wait = w->nxt; w->nxt = 0; srv->plen--; pthread_cond_signal( &w->todo ); sMsg( LOG_DEBUG, "woke %d of %d", w->id, srv->plen ); } } /* lsv_penq */ static Con *lsv_deq ( Que *q ) { Con *c = q->head; if ( c ) { q->len--; q->head = c->nxt; c->nxt = 0; if ( q->tail == c ) q->tail = 0; } return c; } /* lsv_deq */ Wrk *svCur () { return (Wrk*)pthread_getspecific( lsv_key_wrk ); } static void *lsv_wrk ( void *arg ) { Wrk *self = (Wrk*)arg; Srv *srv = self->srv; Pool *work = self->id ? &srv->pool : &srv->main; Con *job = 0; sigset_t blk; /* block signals, they should go to the receiver */ sigemptyset( &blk ); sigaddset( &blk, SIGHUP ); sigaddset( &blk, SIGINT ); sigaddset( &blk, SIGTERM ); sigaddset( &blk, SIGPIPE ); pthread_sigmask( SIG_BLOCK, &blk, 0 ); pthread_setspecific( lsv_key_wrk, arg ); if ( self->id ) srv->app( 0, LSV_APPINI ); self->ses = sGet(); /* allocate session for thread */ for (;/*ever*/;) { int ret; /* **************************************** critical section enter */ pthread_mutex_lock( &srv->mut ); if ( job ) { /* from last turn */ sMsg( LOG_DEBUG, "thr %d enq job %s pdg %s", self->id, job->nam, pdg[job->pdg] ); job->stg = LSV_STGRET; lsv_enq( &srv->recv, job ); if ( job->pdg ) /* wakeup receiver */ write( srv->pip[1], &job->pdg, 1 ); if ( job->ses->que ) { /* immediatly reschedule queued job */ Con *q = job->ses->que; job->ses->que = q->nxt; sMsg( LOG_VERBOSE, "dequeing con %s for ses '%s'", q->nam, job->ses->name ); lsv_penq( srv, q ); } } self->cur = 0; while ( !(job = lsv_deq( &work->que )) && !(LSV_SHUTDN & srv->flg) ) { self->nxt = work->wait; work->wait = self; /* pls pls lemme wrk */ srv->plen++; self->waits++; sMsg( LOG_DEBUG, "thr %d waiting", self->id ); /* wait releases the mutex */ pthread_cond_wait( &self->todo, &srv->mut ); } self->cur = job; if ( job ) job->stg = LSV_STGRUN; pthread_mutex_unlock( &srv->mut ); /* **************************************** critical section leave */ /* end of monday morning meeting blah blah -- start working */ if ( ! job || (LSV_SHUTDN & srv->flg) ) break; self->jobs++; if ( job->ses ) sSet( job->ses ); else job->ses = self->ses; job->ses->io[1] = &job->ios; if ( job->ses->res ) CLRREC( job->ses->res ); sMsg( LOG_DEBUG, "thr %d run job %s ses '%s'.%d", self->id, job->nam, job->ses->name, job->ses->accnt ); ret = srv->app( job, LSV_APPRUN ); sMsg( LOG_DEBUG, "thr %d got %d buf %d", self->id, ret, LIO_SAVAIL( &job->ios ) ); /* lock only in order to create SMP memory barrier. on a single processor system, the byte should be safely readable by receiver without this lock. TOD: check alternatives: since receiver is checking the stage of active connections only in rare cases, it might be more efficient to guard either this set or the processing up to here with a per thread mutex. */ pthread_mutex_lock( &srv->mut ); job->stg = LSV_STGDON; job->ses->cur = 0; pthread_mutex_unlock( &srv->mut ); ret = srv->prt( &job->ios, LIO_SCLOSE ); job->ses->io[1] = 0; sMsg( LOG_DEBUG, "thr %d done job %s", self->id, job->nam ); if ( job->ses != self->ses ) sSet( self->ses ); /* reset session */ } if ( self->id ) srv->app( 0, LSV_APPFIN ); return self; } /* lsv_wrk */ static void *lsv_rcv ( void *arg ) { Srv *srv = (Srv*)arg; fd_set fd; int fdlen = 0; int inlock = 0; int pending = 0; int ret = 0, sel, i; FD_ZERO( &fd ); FD_SET( srv->lsn, &fd ); FD_SET( srv->pip[0], &fd ); fdlen = srv->pip[0]+1; if ( fdlen < srv->lsn+1 ) fdlen = srv->lsn+1; for (;/*ever*/;) { struct sockaddr_in npeer; /* of new connection */ int nsock = -1; /* new socket */ Con *c; fd_set ready = fd; Tm expire; /* select on read only; writing is done in workers and exceptions (i.e. telnet OOB data) are not used */ if ( 0 > (sel = select( fdlen, &ready, 0, 0, 0 )) ) { if ( EINTR == errno ) { sMsg( LOG_DEBUG, "select woken by signal" ); } else { ret = sMsg( LOG_SYSERR, "select" ); goto done; } } if ( lsv_sig.si_signo ) { sMsg( LOG_DEBUG, "SIG%s (%d) from pid %d", SIGHUP==lsv_sig.si_signo ? "HUP" : SIGINT==lsv_sig.si_signo ? "INT" : SIGTERM==lsv_sig.si_signo ? "TERM" : SIGPIPE==lsv_sig.si_signo ? "PIPE" : "?", lsv_sig.si_signo, lsv_sig.si_pid ); if ( 1 /*SIGHUP==lsv_sig.si_signo*/ ) { for ( i=FD_SETSIZE; i--; ) if ( con[i] ) { c = con[i]; sMsg( LOG_INFO, "con %3d peer %s stg %s prt %d app %d ios %d %s (0x%x)", i, c->nam, stg[c->stg], c->prt, c->app, c->ios.pos + c->ios.b.done, LIO_SISOPEN(&c->ios) ? "opn" : "clo", c->ios.file ); } } if ( lsv_term ) /* even if it wasn't the last signal delivered */ goto done; lsv_sig.si_signo = 0; /* must NOT access the set after interrupt, probably all fds are set */ continue; } timeUpd( &srv->tim ); timeGtfm( srv->gtm, &srv->tim ); expire.millis = srv->tim.millis - srv->sto*LLL(1000); pthread_mutex_lock( &srv->mut ); /* back in locked land */ inlock = !0; /* -- we're the acceptor */ if ( FD_ISSET( srv->lsn, &ready ) ) { /* new connection */ unsigned /*socklen_t is broken*/ alen = sizeof(npeer); nsock = accept( srv->lsn, (struct sockaddr*)&npeer, &alen ); if ( 0 > nsock ) { ret = sMsg( LOG_SYSERR, "accept" ); goto done; } if ( FD_SETSIZE <= nsock ) { ret = sMsg( ERR_BADF, "socket %d >= FD_SETSIZE %d", nsock, FD_SETSIZE ); goto done; } if ( sizeof(npeer) != alen ) { ret = sMsg( ERR_INVAL, "bad peer len %d", alen ); goto done; } /* ok */ /* setsockopt SO_SNDTIMEO fixed value on linux SO_OOBINLINE ... hmm, don't care SO_LINGER off */ /* prepare con */ if ( !(c = con[nsock]) ) { c = con[nsock] = mAlloc( sizeof(*c) ); c->ios.file = nsock; /* preliminary -- further setup below */ } c->pdg = LSV_PDGNEW; switch ( c->stg ) { case LSV_STGCON: /* still in control of receiver */ case LSV_STGINP: sMsg( ERR_IDIOT, "bad recv stage %d on new con", c->stg ); c->stg = LSV_STGNEW; case LSV_STGNEW: /* clean idle socket -- pass to receiver */ lsv_enq( &srv->recv, c ); break; case LSV_STGSES: /* still in control of receiver */ case LSV_STGWRK: sMsg( ERR_IDIOT, "bad worker stage %d on new con", c->stg ); case LSV_STGRUN: /* worker will queue it */ /* hmmm ... should have COM ... */ case LSV_STGCOM: case LSV_STGDON: case LSV_STGRET: /* worker has queued it */ pending++; break; } srv->conn++; FD_CLR( srv->lsn, &ready ); /* skip in fd check */ } /* new connection */ if ( FD_ISSET( srv->pip[0], &ready ) ) { /* pending events */ char trash[LSV_NUMWRK]; read( srv->pip[0], trash, sizeof(trash) ); FD_CLR( srv->pip[0], &ready ); } /* switching roles -- now we're the receiver */ while ( (c = lsv_deq( &srv->recv )) ) { /* care for ready connections */ int sock = LIO_FD & c->ios.file; sMsg( LOG_DEBUG, "deq %d from %s pdg %s", sock, c->nam, pdg[c->pdg] ); if ( c->pdg && c->stg ) pending--; if ( !(LIO_IN & c->ios.file) && (LIO_RDWR & c->ios.file) ) { /* cleanup closed socks */ if ( c != con[sock] ) { ret = sMsg( ERR_IDIOT, "deq bad sock %d (0x%x)", sock, c->ios.file ); goto done; /* PANIC time */ } sMsg( LOG_DEBUG, "clos%s sock %d", (LIO_OUT & c->ios.file) ? "ing" : "ed", sock ); if ( (LIO_OUT & c->ios.file) ) lio_close( &c->ios.file, LIO_INOUT ); c->ios.file &= ~LIO_RDWR; /* clear also, so we're not closing again */ FD_CLR( sock, &fd ); FD_CLR( sock, &ready ); /* probably an EOF to read */ if ( fdlen == sock+1 ) while ( fdlen /* && !FD_ISSET( fdlen-1, &fd ) risk of muting tmp. disabled */ && (!con[fdlen-1] || !LIO_SISOPEN(&con[fdlen-1]->ios)) ) fdlen--; } if ( LSV_PDGNEW == c->pdg ) { /* new connection */ struct sockaddr_in peer; /* of new connection */ if ( nsock == sock ) /* the one accepted just above */ peer = npeer; else { unsigned alen = sizeof(peer); if ( getpeername( sock, (struct sockaddr*)&peer, &alen ) || sizeof(peer) != alen ) { ret = sMsg( LOG_SYSERR, "getpeername alen %d", alen ); goto done; } } c->srv = srv; c->ses = 0; { unsigned host = (unsigned)ntohl( peer.sin_addr.s_addr ); unsigned short port = (unsigned short)ntohs( peer.sin_port ); char *p = c->nam; c->host = host; memset( c->nam, 0, sizeof(c->nam) ); p += u2a( p, 0xff & (host >> 24) ); *p++ = '.'; p += u2a( p, 0xff & (host >> 16) ); *p++ = '.'; p += u2a( p, 0xff & (host >> 8) ); *p++ = '.'; p += u2a( p, 0xff & (host ) ); *p++ = ':'; p += u2a( p, port ); /* *p++ = '@'; p += u2a( p, sock ); */ } c->con++; /* never mind overflow ... */ sMsg( LOG_INFO, "connect %d on %d from %s", c->con+1, sock, c->nam ); LIO_SINIT( &c->ios, srv->prt, c->nam, sock | LIO_INOUT|LIO_RDWR|LIO_SOCK ); LIO_BINIT( &c->flt ); FD_SET( sock, &fd ); if ( fdlen < sock+1 ) fdlen = sock+1; c->bin = 0; /* do not reset on every request */ } if ( ! (LIO_IN & c->ios.file) ) c->stg = LSV_STGNEW; else { c->stg = LSV_STGCON; if ( ! FD_ISSET( sock, &fd ) ) { sMsg( LOG_DEBUG, "reenabling %s", c->nam ); FD_SET( sock, &fd ); if ( fdlen < sock+1 ) fdlen = sock+1; } } c->grp = c->pdg = c->prt = c->app = 0; /* prepare for new request */ c->ios.pos = 0; c->ios.b.done = c->ios.b.fill = 0; if ( c->req ) CLRREC( c->req ); else { /* this is done in the receiver rather than the acceptor since in multi-receiver model, receiver owns the req */ c->req = mAlloc( LSV_BUFSIZ ); OPENISIS_INITREC( c->req, LSV_BUFSIZ, 64 ); } if ( !(LSV_CONSES & srv->flg) ) c->ses = 0; } /* prepare connections */ for ( i=fdlen; sel && i--; ) { /* read from the ready */ Field *sid; int got; if ( ! FD_ISSET( i, &ready ) ) continue; sel--; c = con[i]; if ( ! c ) { ret = sMsg( ERR_IDIOT, "ran on empty con %d", i ); goto done; } if ( LSV_STGINP < c->stg ) { sMsg( LOG_DEBUG, "con %d busy stage %d!", i, c->stg ); c->pdg = LSV_PDGRED; pending++; FD_CLR( i, &fd ); /* mute this */ continue; } c->ios.b.done = 0; got = lio_read( &c->ios.file, c->ios.b.c, sizeof(c->ios.b.c) ); if ( 0 >= got ) { if ( -ERR_EOF != got ) goto err; sMsg( LOG_DEBUG, "EOF on 0x%x line %d from %s", c->ios.file, c->req->len, c->nam ); FD_CLR( i, &fd ); if ( ! c->req->len ) /* initial eof */ goto eof; /* close immediatly */ /* lio_close( &c->ios.file, LIO_IN ); */ c->ios.file &= ~LIO_IN; /* don't shut */ c->pdg = LSV_PDGEOF; /* so it will be dequed and closed */ pending++; got = 0; /* tell proto to close ... */ } if ( ! c->req->len ) { /* stage should be CON */ RADDS( c->req, -LSV_PEER, c->nam, 1 ); sMsg( LOG_DEBUG, "req on %d from %s", LIO_FD&c->ios.file, c->nam ); } c->stg = LSV_STGINP; c->ios.b.fill = got; got = srv->prt( &c->ios, LIO_SPUSH ); sMsg( LOG_DEBUG, "got 0x%x from %s", got, c->nam ); if ( ! got && (LIO_IN & c->ios.file) ) /* ok: more next time */ continue; if ( 0 > got ) /* nok */ goto err; got = c->srv->app( c, LSV_APPGOT ); if ( 0 > got ) /* nok */ goto err; if ( !(LSV_CONSES & srv->flg) && c->req && (sid = rGet( c->req, -LSV_SID, 0 )) ) { c->ses = cSesByName( (char*)sid->val, sid->len, &srv->tim, &expire ); if ( ! c->ses ) { /* sad sad sad */ sMsg( LOG_ERROR, "out of sessions" ); goto err; /* drop connection */ } if ( ! c->ses->accnt && (! c->ses->prop || ! c->ses->prop->len) /* overflow paranoia */ ) got = c->srv->app( c, LSV_APPSES ); if ( 0 > got ) /* nok */ goto err; } c->tim = srv->tim; RADDS( c->req, -LSV_TIME, srv->gtm, 1 ); if ( c->ses ) { sMsg( LOG_DEBUG, "req %s for ses '%s' %s", c->nam, c->ses->name, c->ses->cur ? "busy" : "idle" ); c->ses->accnt++; if ( c->ses->cur || c->ses->que ) { /* oh - busy !? */ Con **cpp = &c->ses->que; int l = 0; for ( ; *cpp; cpp = &(*cpp)->nxt ) if ( 4 < ++l ) { sMsg( LOG_WARN, "too many con queing on ses '%s' - dropping %s", c->ses->name, c->nam ); goto err; } sMsg( LOG_VERBOSE, "queing con %s for ses '%s'", c->nam, c->ses->name ); c->nxt = 0; *cpp = c; c->stg = LSV_STGSES; continue; } } /* yo - got it */ lsv_penq( srv, c ); continue; err: sMsg( ERR_BADF, "err 0x%x from %s", got, c->nam ); eof: lio_close( &c->ios.file, LIO_INOUT ); c->stg = LSV_STGNEW; FD_CLR( i, &fd ); } /* read from the ready */ srv->turn++; srv->wlen += srv->pool.que.len; srv->busy += srv->nwr - srv->plen; sMsg( LOG_INFO, "\n==== %s\nturn %d: %d jobs for %d idle workers %d pending", srv->gtm, srv->turn, srv->pool.que.len, srv->plen, pending ); for ( i=srv->nwr; i--; ) if ( srv->wrk[i].cur ) sMsg( LOG_INFO, "thr %d serving %d from %s", i, LIO_FD&srv->wrk[i].cur->ios.file, srv->wrk[i].cur->nam ); inlock = 0; pthread_mutex_unlock( &srv->mut ); } /* for(ever) */ done: if ( inlock ) pthread_mutex_unlock( &srv->mut ); srv->flg |= LSV_SHUTDN; for ( i=srv->nwr; i--; ) pthread_cond_signal( &srv->wrk[i].todo ); return srv; } /* lsv_rcv */ int svRun ( Srv *srv, const char *addr ) { static const int yes = !0; struct sigaction siga; struct sockaddr_in sa; /* server address */ int ret = 0, i; /* assume we're the main-thread -- try to make sure we get the controlling session */ sGet(); srv->mut = lsv_mutex_init; /* set defaults */ if ( ! srv->sto ) srv->sto = 30*60; /* 30 minutes session timeout */ if ( ! srv->nwr ) srv->nwr = LSV_NUMWRK/2; /* half the max */ else if ( LSV_NUMWRK < srv->nwr ) srv->nwr = LSV_NUMWRK; /* max */ /* prepare signals */ memset( &siga, 0, sizeof(siga) ); siga.sa_sigaction = lsv_sighand; sigemptyset( &siga.sa_mask ); sigaddset( &siga.sa_mask, SIGHUP ); sigaddset( &siga.sa_mask, SIGINT ); sigaddset( &siga.sa_mask, SIGTERM ); sigaddset( &siga.sa_mask, SIGPIPE ); siga.sa_flags = SA_SIGINFO; sigaction( SIGHUP, &siga, 0 ); sigaction( SIGINT, &siga, 0 ); sigaction( SIGTERM, &siga, 0 ); sigaction( SIGPIPE, &siga, 0 ); /* prepare socket */ srv->lsn = socket( PF_INET, SOCK_STREAM, IPPROTO_TCP/*6*/ ); /* fcntl( srv->lsn, O_NONBLOCK ); not needed */ sa.sin_family = AF_INET; sa.sin_port = htons( addr ? a2i( addr, -1 ) : 8080 ); sa.sin_addr.s_addr = #ifdef LSV_LOCALHOST_ONLY htonl(INADDR_LOOPBACK); #else INADDR_ANY; #endif if ( setsockopt( srv->lsn, SOL_SOCKET, SO_REUSEADDR, &yes, sizeof(yes) ) ) { ret = sMsg( LOG_SYSERR, "setsockopt" ); goto inierr; } if ( bind( srv->lsn, (struct sockaddr*)&sa, sizeof(sa) ) ) { ret = sMsg( LOG_SYSERR, "bind on port %u", htons(sa.sin_port) ); goto inierr; } if ( listen( srv->lsn, 64/*backlog*/ ) ) { ret = sMsg( LOG_SYSERR, "listen" ); goto inierr; } /* prepare pending event pipe */ if ( pipe( srv->pip ) ) { ret = sMsg( LOG_SYSERR, "pipe" ); goto inierr; } if ( pthread_key_create( &lsv_key_wrk, 0 ) ) { ret = sMsg( LOG_SYSERR, "key_create" ); goto inierr; } /* setup workers */ memset( srv->wrk, 0, sizeof(srv->wrk) ); for ( i=srv->nwr; i--; ) { srv->wrk[i].id = i; srv->wrk[i].srv = srv; srv->wrk[i].nxt = 0; srv->wrk[i].ses = cSession(0); memcpy( srv->wrk[i].ses->name, "wrk", 3 ); u2a( srv->wrk[i].ses->name+3, i ); srv->wrk[i].todo = lsv_cond_init; if ( i ) pthread_create( &srv->wrk[i].thr, 0, lsv_wrk, srv->wrk+i ); } /* start the server thread */ pthread_create( &srv->rcv, 0, lsv_rcv, srv ); /* run main worker */ srv->wrk[0].thr = pthread_self(); lsv_wrk( srv->wrk ); /* cleanup */ pthread_join( srv->rcv, 0 ); for ( i=srv->nwr; i--; ) pthread_join( srv->wrk[i].thr, 0 ); for ( i=srv->nwr; i--; ) sMsg( LOG_INFO, "worker %d had %d jobs %d waits", i, srv->wrk[i].jobs, srv->wrk[i].waits ); sMsg( LOG_INFO, "server had %d jobs on %d connections %d turns" " avg queue len %.02f workers %.02f", srv->jobs, srv->conn, srv->turn, srv->wlen/srv->turn, srv->busy/srv->turn ); inierr: close( srv->lsn ); return ret; } /* svRun */ /** helper, probably should go to lio. cat l bytes from c to buf b flushing as needed. if ! c, empty buf b. */ int lio_sout ( Ios *s, Buf *b, const char *c, unsigned l ) { if ( ! b ) b = &s->b; do { if ( LIO_BUFSIZ/2 < b->fill && b->done ) { /* purge */ if ( b->done < b->fill ) memmove( b->c, b->c + b->done, b->fill - b->done ); s->pos += b->done; b->fill -= b->done; b->done = 0; } if ( c && l && b->fill < LIO_BUFSIZ ) { unsigned cp = LIO_BUFSIZ - b->fill; if ( cp > l ) cp = l; memcpy( b->c + b->fill, c, cp ); b->fill += cp; c += cp; l -= cp; } if ( b->fill > b->done ) { int wr = lio_write( &s->file, b->c + b->done, b->fill - b->done ); if ( 0 > wr ) return wr; if ( ! wr ) return -ERR_BUSY; b->done += wr; if ( b->done > b->fill ) { b->done = b->fill = 0; return -ERR_IDIOT; } } if ( b->done >= b->fill ) { s->pos += b->done; b->done = b->fill = 0; } } while ( c ? l : (unsigned)(b->fill > b->done) ); return 0; } /* lio_sout */ #define LF 10 /* LineFeed a.k.a. newline - '\n' isn't really well defined */ #define TAB 9 /* horizontal, that is */ #define VT 11 /* vertical, used as newline replacement */ static const char lf = LF; static const char tab = TAB; static const char vt = VT; /* the plain protocol */ int svPlain ( Ios *s, int op ) { Field *f; Con *c = (Con*)s; switch ( op ) { case LIO_SPUSH: { int l = s->b.fill - s->b.done; unsigned char *b = s->b.c + s->b.done; unsigned char *end = b+l, *v, *p; if ( ! l ) { /* EOF: done */ if ( ! c->prt ) /* ok */ return 1; /* last field wasn't closed by LF */ return sMsg( ERR_INVAL, "no EOL from %s", c->nam ); } if ( c->prt ) RSPACE( c->req, l, !0 ); /* add text lines */ while ( bprt ) { case 0: /* at beginning of line -- start new field */ if ( LF == *b ) /* empty line */ return 1; if ( TAB != *b || !c->req->len ) RADD( c->req, 0,0,end-b, !0 ); else { /* binary mode continuation line */ conti = 1; if ( ! c->bin ) { sMsg( LOG_INFO, "detected binary mode on con %s", c->nam ); c->bin = 1; } RSPACE( c->req, end-b, !0 ); } if ( ! c->req ) return -ERR_NOMEM; c->prt = 1; case 1: /* add to last field */ f = c->req->field + c->req->len-1; v = (unsigned char*)f->val; p = v + f->len; if ( conti ) { *p++ = LF; b++; } if ( c->bin ) { for ( ; breq->used += (p - v) - f->len; f->len = p - v; if ( LF == b[-1] ) { int ret = a2il( f->val, f->len, &f->tag ); if ( ret ) { if ( ret < f->len && TAB == v[ret] ) ret++; if ( ret < f->len ) memmove( v, v+ret, f->len - ret ); f->len -= ret; } ret = c->srv->app( c, LSV_APPARG ); c->prt = 0; if ( ret ) return ret; } sMsg( LOG_INFO, "prs from %s: [%2d] %3d = '%.*s'", c->nam, c->req->len-1, f->tag, f->len, f->val ); } } return 0; } /* case LIO_SPUSH */ case LIO_SFLUSH: if ( s->b.fill <= s->b.done ) return 0; case LIO_SCLOSE: if ( !(LIO_OUT & s->file) ) return -ERR_INVAL; /* this switch is not protected, since receiver doesn't care */ if ( c->stg < LSV_STGCOM ) c->stg = LSV_STGCOM; if ( ! s->pos+s->b.done && c->ses->res ) { char num[32]; int n = c->ses->res->len; sMsg( LOG_INFO, "com %d fields to %s", c->ses->res->len, c->nam ); for ( f = c->ses->res->field; n--; f++ ) { const char *v = f->val, *nl; int vl = f->len; int nlen = i2a( num, f->tag ); num[nlen++] = TAB; lio_sout( s, &c->flt, num, nlen ); while ( vl && (nl = memchr(v,'\n',vl)) ) { int ll = 1+(nl-v); if ( c->bin ) { lio_sout( s, &c->flt, v, ll ); lio_sout( s, &c->flt, &tab, 1 ); } else { lio_sout( s, &c->flt, v, ll-1 ); lio_sout( s, &c->flt, &vt, 1 ); } v += ll; vl -= ll; } if ( vl ) lio_sout( s, &c->flt, v, vl ); lio_sout( s, &c->flt, &lf, 1 ); } } if ( s->b.fill ) s->b.done = s->b.fill = 0; /* { lio_sout( s, &c->flt, 0, 0 ); lio_stdio( s, LIO_SFLUSH ); } */ if ( LIO_SCLOSE == op ) { lio_sout( s, &c->flt, &lf, 1 ); lio_sout( s, &c->flt, 0, 0 ); /* finally flush */ } return 0; } return lio_stream( s, op ); /* hmm -- SPURGE ? */ } /* svPlain */ /* the echo application */ int svEcho ( Con *c, int task ) { static Tm tm = { LLL(50) }; Field *f; int i; if ( task ) { /* preparing: we don't care */ if ( LSV_APPGOT == task ) c->grp = 1; /* don't require mainthead */ return 0; } /* run it */ sMsg( LOG_INFO, "echo %d: %d fields to %s", svCur()->id, c->req->len, c->nam ); #if 0 timeSleep( &tm ); #endif for ( i=c->req->len, f = c->req->field; i--; f++ ) { RADDF( c->ses->res, f, !0 ); /* plain protocol actually does not support plain output ... sMsg( c->ses, 0, "%d = '%.*s'\n", f->tag, f->len, f->val ); */ } return 0; } /* svEcho */ #if 0 int main ( int argc, const char **argv ) { Srv echo; memset( &echo, 0, sizeof(echo) ); echo.prt = lsv_plain; echo.app = lsv_echo; cOpen(0); cLog( LOG_DEBUG, 0 ); return svRun( &echo, 1 #include "loi.h" #include "lio.h" #include "lses.h" /** network service RECV: loop in receiver thread - loop selecting on all connections - dequeue connections from recv queue, clear BUSY - for every available connection: - if request is new (no input so far), prepare req with peer header, preset default flags - read input (which may clear the LIO_IN flag on EOF) - call proto to extract input proto calls check (if any) to preprocess each piece of input - if done, set BUSY, pass to dispatcher (via queue or in same thread) DISP: the one and only session master (in separate thread, if multiple recv) - no input after this point, the req is considered read-only any data on connection is considered an error (i.e. no pipelining) - if request contains a SID header, find or create session. if session is busy, enqueue to session's wait queue - enqueue to work queue WORK: (in worker thread) - once a worker thread dequeued a request, it calls process function this may add fields to response and/or write to output stream - before output buffer is flushed first time, it is set aside and proto is called to output the response - after process is done, the request is marked in flush phase: the req is not accessed after this point, the output is flushed and the response freed - enqueue to recv queue events in connection lifetime: - accept: if the socket hasn't been used so far, a new connection is created. an existing connection is reused. The connection object may still be in use by a worker. Since the socket file id is reused, the former socket with the same fid must have been closed, and the LIO_INOUT flags must have been cleared, so lio_write on that old lio file id will fail. The connection is marked as new and, if not in use by a worker, entered into the receive queue. - dequeue from receive queue: if the connection is marked new, it is prepared and added to the select set. else, if it was closed, it is removed from the select set. - read events: if the connection is in use, that is, a previous request hasn't yet entered the flush phase, there is an attempt to pipeline. while this is considered an error, for robustness it's handled: the connection is marked pipeline and the socket is cleared from the select set, to be reenabled later. Once the connection is idle, input is read until the protocol decides on a complete request. Excess data (immediate pipelining) is considered an error. */ /* common header tags -- actually negative tags are used */ enum { /* nonstandard */ LSV_TEXT, /* +- 0: a small, inlined untyped or text/plain body */ LSV_RID, /* row id, a.k.a MFN */ LSV_SID, /* session id */ LSV_PEER, /* network peer ("remote host") */ LSV_TIME, /* GTF + millis */ LSV_ATTACH, /* wrapped-up attachment, TBD */ /* MIME basics */ LSV_CTYPE, /* HTTP: CGI/servlet */ LSV_URI, LSV_QUERY }; enum { LSV_NUMWRK = 8, LSV_BUFSIZ = 0x04000, /* 16384 a.k.a. 16K */ LSV_BUFMSK = 0x03fff /* mask buf lengths */ }; typedef struct Que { struct Con *head; struct Con *tail; int len; } Que; /* queue with pool of workers waiting */ typedef struct Pool { Que que; struct Wrk *wait; /* stack */ } Pool; typedef struct Wrk { pthread_t thr; unsigned id; struct Srv *srv; struct Wrk *nxt; Session ses; pthread_cond_t todo; /* work condition waited for by workers */ /* stats */ struct Con *cur; /* current job */ int jobs; int waits; } Wrk; typedef struct Srv { /* the protocol stream method - with LIO_SPUSH, input is pushed to request (len 0 indicates EOF) return 0=more, >0=done, <0=error - on LIO_SFLUSH, prt must convert output buffer to flt - on LIO_SCLOSE, prt must flush and close answer */ int (*prt) ( Ios*, int op ); /* the application - with task APPARG (called by prt for each field): check (and typically modify) last field extracted by prt return 0=more, >0=done, <0=error - with task APPGOT (after request is complete): check request, add -LSV_SID field, if needed (e.g. to extract from an URL or cookie) - with task APPRUN (0): process request */ int (*app) ( struct Con*, int task ); Fdt *fdt; /* field def */ int flg; /* flags */ unsigned sto; /* config: session timeout (seconds) */ unsigned nwr; /* config: #workers (<= LSV_NUMWRK) */ Session ses; /* usually the default session */ Tm tim; /* tim of last turn */ char gtm[20]; /* generalized time (actually 18 digits) */ Wrk wrk[LSV_NUMWRK]; /* single receiver model */ pthread_t rcv; /* the receiver thread */ pthread_mutex_t mut; /* single mutex held by recv/disp while not in select */ int lsn; /* listening socket */ int pip[2]; /* pipe for pending events */ Que recv; /* enq by worker, deq by receiver */ Pool main; /* enq by dispatcher, deq by workers */ Pool pool; /* enq by dispatcher, deq by workers */ /* stats */ int plen; /* length of pool */ int jobs; /* total #jobs */ int conn; /* total #conn */ int turn; /* total #turn */ float wlen; /* cumulated length of work queue (on each turn) */ float busy; /* cumulated # of active workers (on each turn) */ } Srv; enum { /* server flags */ LSV_CONSES = 0x01, /* session bound to connection */ LSV_WRKSES = 0x02, /* session bound to worker */ LSV_SHUTDN = 0x100 /* server is shutting down */ }; enum { /* called in worker thread: */ LSV_APPRUN, /* app called to process request */ LSV_APPINI, /* app called to initialise non-main thread (no con) */ LSV_APPFIN, /* app called to finalize non-main thread (no con) */ /* called in receiver thread: */ LSV_APPARG, /* app called to check parameter */ LSV_APPGOT, /* app called after all input */ LSV_APPSES /* app called on new session */ }; enum { /* connection pending state */ LSV_PDGNON, /* new connection pending */ LSV_PDGNEW, /* new connection pending */ LSV_PDGRED, /* ready to read pending */ LSV_PDGEOF /* EOF pending */ /* when adding, keep array of names in sync */ }; enum { /* connection processing stage */ LSV_STGNEW, /* connection needs to be setup */ LSV_STGCON, /* connected, no request on the way */ LSV_STGINP, /* reading input */ LSV_STGSES, /* waiting for session */ LSV_STGWRK, /* queued for work */ /* set by worker: */ LSV_STGRUN, /* dequeued */ LSV_STGCOM, /* commited (output has started) */ LSV_STGDON, /* done processing, flushing output */ LSV_STGRET /* returning (queued to recv) */ /* when adding, keep array of names in sync */ }; typedef struct Con { Ios ios; /* raw output */ Buf flt; /* protocol filtered output */ const Srv *srv; struct Con *nxt; unsigned char stg; /* processing stage (guarded by mutex) */ unsigned char pdg; /* pending flags set by receiver only */ unsigned char grp; /* thread group (0=main) set by receiver app calls */ unsigned char bin; /* client requested binary mode */ int prt; /* for free use by prt, 0 init */ int app; /* for free use by app, 0 init */ int host; char nam[64]; /* 123.123.123.123:123456 for IPv4 */ Session ses; Rec *req; /* request -- owned by srv->ses */ Tm tim; /* time when the request was scheduled */ /* Rec *res; response -- owned by ses */ int con; /* connection counter */ } Con; extern Wrk *svCur (); extern int svRun ( Srv *srv, const char *addr ); /* the plain protocol */ extern int svPlain ( Ios *s, int op ); extern int svEcho ( Con *c, int task ); #define LSV_H #endif /* LSV_H */ luti.c0100644000113000011300000002776207655441446010711 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: luti.c,v 1.47 2003/05/03 13:16:06 mawag Exp $ utilities */ #include /* strlen */ #include "openisis.h" #include "luti.h" /* ************************************************************ private types */ typedef struct { char *name; int id; } LutiLTEnt; typedef struct OpenIsisLT { LutiLTEnt *arr; int siz; /* malloc'ed */ int num; /* used */ } LutiLTNod; #define LT_NUM 27 #define LT_SIZE (LT_NUM * sizeof (LutiLTNod)) #define LTIDX(name) ( \ ('a' <= *(name) && 'z' >= *(name)) ? *(name) - 'a' : ( \ ('A' <= *(name) && 'Z' >= *(name)) ? *(name) - 'A' : 26 \ )) #define LT_INCR 2 /* ************************************************************ private data */ /* ************************************************************ private functions */ /* ************************************************************ package data */ const char luti_hex[16] = { '0','1','2','3','4','5','6','7','8','9','a','b','c','d','e','f' }; /* ************************************************************ package functions */ void log_rec ( int level, Rec *rec, const char *msg, const char *delim ) { Field *F; int ch = 2; if (0 > level) { level = -level; ch = 1; } if ((unsigned)level > log_lev) { return; } if (msg && *msg) { sMsg (ch, msg); } sMsg (ch, "%p", rec); if (rec) { if (! delim) { delim = "; "; } sMsg (ch, " #f=%d", (int)rec->len); for (F = rec->field; rec->len > F - rec->field; ++F) { sMsg (ch, delim); sMsg (ch, "%d", (int)F->tag); sMsg (ch, " %.*s", (int)F->len, F->val); } } sMsg (ch, "\n"); } int lprint ( void *buf, int i ) { char *c = buf; if ( 0 > i ) { *c++ = '-'; i = -i; } if ( ! i ) *c++ = '0'; else { int dig = 0, j=i; char *p; for ( ; j; j /= 10 ) dig++; p = c += dig; for ( ; i; i /= 10 ) *--p = '0' + i % 10; } *c = 0; return c - (char*)buf; } /* lprint */ int a2il ( const char *p, int l, int *res ) { const char *s = p; const char *e = p + (0<=l ? l : (int)strlen(p)); int ret = 0, min = 0; while ( p < e && ' ' >= (unsigned char)*p ) p++; if ( p < e && '0' > *p ) { if ( '+' == *p ) p++; else if ( '-' == *p ) { min = 1; p++; } } if ( p < e && '0' == *p && ++p < e && 'x' == *p ) { while ( ++p < e ) { int v = *p; if ( '0' <= v && v <= '9' ) v -= '0'; else if ( 'a' <= v && v <= 'f' ) v -= 'a'-10; else if ( 'A' <= v && v <= 'F' ) v -= 'A'-10; else break; ret = (ret<<4) + v; } } else { while ( p < e && '0' <= *p && *p <= '9' ) ret = 10*ret + *p++ - '0'; } if (0 != res) { *res = min ? -ret : ret; } return p - s; } /* a2il */ int a2i ( const char *p, int l ) { int res; a2il (p, l, &res); return res; } int a2id ( const char *p, int l, int dflt ) { int res; if (0 > l) { l = (int) strlen (p); } if (l != a2il (p, l, &res)) { return dflt; } return res; } int i2a ( char *p, int i ) { if ( i ) { int min, dig = 0, j; if ( 0 < i ) min = 0; else { min = 1; *p++ = '-'; i = -i; } for ( j=i; j; j/=10 ) dig++; for ( *(p+=dig)=0; i; i/=10 ) *--p = '0' + (i % 10); return min+dig; } *p++ = '0'; *p = 0; return 1; } /* i2a */ int u2a ( char *p, unsigned u ) { if ( u ) { int dig = 0, j; for ( j=u; j; j/=10 ) dig++; for ( *(p+=dig)=0; u; u/=10 ) *--p = '0' + (u % 10); return dig; } *p++ = '0'; *p = 0; return 1; } /* u2a */ LutiLT luti_ltnew () { LutiLT res = (LutiLT) mAlloc (LT_SIZE); return res; } void luti_ltdel (LutiLT lt) { LutiLTNod *N; LutiLTEnt *E; if (lt) { for (N = lt + LT_NUM; --N >= lt; ) { if (N->arr) { for (E = N->arr + N->num; --E >= N->arr; ) { if (E->name) { mFree (E->name); } } mFree (N->arr); } } mFree (lt); } } void luti_ltadd (LutiLT lt, const char *name, int id) { if (0 != lt) { LutiLTNod *N; LutiLTEnt *E; int idx; if (!name) { name = ""; } idx = LTIDX (name); N = lt + idx; if (N->siz == N->num) { E = mAlloc ( (LT_INCR + N->siz) * sizeof (LutiLTEnt)); if (N->siz) { memcpy (E, N->arr, N->siz * sizeof (LutiLTEnt)); mFree (N->arr); } N->arr = E; N->siz += LT_INCR; } N->arr[N->num].name = mDup (name, -1); N->arr[N->num].id = id; ++N->num; } } int luti_ltget (LutiLT lt, const char *name) { if (0 != lt) { LutiLTNod *N; LutiLTEnt *E; int idx; if (!name) { name = ""; } idx = LTIDX (name); N = lt + idx; if (N->num) { for (E = N->arr + N->num; --E >= N->arr; ) { if (! strcmp (E->name, name)) { return E->id; } } } } return -1; } void luti_ltrmv (LutiLT lt, const char *name) { if (0 != lt) { LutiLTNod *N; LutiLTEnt *E; int idx; if (!name) { name = ""; } idx = LTIDX (name); N = lt + idx; if (N->num) { for (E = N->arr + N->num; --E >= N->arr; ) { if (! strcmp (E->name, name)) { int len = N->num - (E - N->arr) - 1; mFree (E->name); if (len) { memmove (E, E + 1, len * sizeof (LutiLTEnt)); } --N->num; return; } } } log_msg (LOG_WARN, "luti_ltrmv: no such entry <%s>", name); } } static Rec *_unwrap ( Rec *env, int *pos, int tag, int dbid, int rdonly ) { Field *F; Rec *rec = 0; int rlen, np; if (! env ) { return 0; } if (! pos) { np = 0; pos = &np; } F = rGet (env, tag, pos); if (!F) { return 0; } rlen = a2id (F->val, (int)F->len, -1); if (0 > rlen) { sMsg (LOG_ERROR, "luti_unwrap: illegal reclen %d, pos %d, tag %d", rlen, *pos, F->tag); return 0; } ++F; if (env->len < *pos + rlen) { sMsg (LOG_ERROR, "luti_unwrap: illegal reclen %d, pos %d, tag %d, env %d", rlen, *pos, F->tag, env->len); return 0; } if (! rlen && ! rdonly) { OPENISIS_RSPACE (rec, 0, 0); if (0 == rec) { return 0; } } while (0 <= --rlen) { if (! rdonly) { RADD (rec, F->tag, F->val, F->len, !0); if (0 == rec) { return 0; } } ++F; ++(*pos); } if (rdonly) { return (Rec*)1; } rec->dbid = dbid; return rec; } Rec *luti_unwrap ( Rec *env, int *pos, int tag, int dbid) { return _unwrap (env, pos, tag, dbid, 0); } Rec* luti_append (Rec *tgt, Rec *src) { Field *F, *E; if (!src) { return tgt; } E = src->field + src->len; for (F = src->field; tgt && E > F; ++F) { RADD (tgt, F->tag, F->val, F->len, !0); } return tgt; } Rec *luti_wrap (Rec *env, Rec *rec, int tag) { if (!rec) { return env; } env = rAddI (env, tag, rec->len, !0); return luti_append (env, rec); } int luti_ptrincr ( void *start, int *num, int incr, int siz, int maxn ) { char **base = (char**)start; char *arr; int oldn = *num; if (0 < maxn && oldn >= maxn) { return -1; } arr = mAlloc ((oldn + incr) * siz); if (!arr) { return -1; } if (oldn) { memcpy (arr, *base, oldn * siz); mFree (*base); } memset (arr + oldn * siz, 0, incr * siz); *num += incr; *base = arr; return oldn; } void luti_free (void **arr, int num) { if (arr) { while (0 <= --num) { if (arr[num]) { mFree (arr[num]); } } mFree (arr); } } int luti_true (const char *str, int len) { if (! str || ! *str) { return -1; } if (0 > len) { len = strlen (str); } if (1 == len) { if ('0' == *str || 'f' == *str || 'F' == *str || 'n' == *str || 'N' == *str ) { return 0; } if ('1' == *str || 'y' == *str || 'Y' == *str || 't' == *str || 'T' == *str ) { return 1; } return -1; } if (! strncmp ("false", str, len) || ! strncmp ("no", str, len)) { return 0; } if (! strncmp ("true", str, len) || ! strncmp ("yes", str, len)) { return 1; } return -1; } /* ------------------------------------------------------------------------ * unwrap by path */ Fdt* luti_fdt_from_rec (Rec *rec) { Db *dbh; if (rec && 0 <= rec->dbid && (dbh = nDbById (rec->dbid)) ) { return dbh->fdt; } return 0; } const char* luti_parse_path ( const char *path, const Fdt *fdt, Fd **fd, int *tag, int *occ ) { Fd *_f = 0; int _t = -1; int _o = -1; int num, qsub; if (fd) { *fd = 0; } if (tag) { *tag = -1; } if (occ) { *occ = -1; } if (! path) { return 0; } if ('.' == *path || ('-' == *path && (path[1]<'0' || '9' OPENISIS_FD_NAMELEN - 1) { return 0; } strncpy (name, path, num) [num] = 0; _f = fByName ( fdt, name ); if (! _f) { return 0; } _t = _f->id; } path += num; if (tag) { *tag = _t; } if (fd) { *fd = _f; } if ('[' != *path && '(' != *path) { return path; } qsub = *path; ++path; num = a2il (path, -1, &_o); if (! num || 0 > _o) { return 0; } path += num; if ((']' != *path && '[' == qsub) || (')' != *path && '(' == qsub)) { return 0; } if (occ) { *occ = _o; } return path + 1; } Rec* luti_getembed ( Rec *env, const char *path, const Fdt *fdt) { Rec *res; int tag, occ, idx, pos; if ( ! env || ! path || ! *path) { return 0; } if (! fdt) { fdt = luti_fdt_from_rec (env); } path = luti_parse_path (path, fdt, 0, &tag, &occ); if (! path) { return 0; } res = 0; idx = pos = 0; while (1) { res = _unwrap (env, &pos, tag, -1, idx < occ); if (! res) { return 0; } if (idx >= occ) { break; } ++idx; } if (*path) { Rec *r2 = luti_getembed (res, path, 0); mFree (res); return r2; } return res; } int lhash ( const char *str, int len ) { int h = 0; while ( len-- ) h = 31*h + *str++; return 0<=h ? h : -h; } /* ************************************************************ public functions */ char *toHtml ( const char *str, int len ) { int l; char *p, *e, *ret; if ( ! str ) return 0; if ( 0 > len ) len = strlen(str); l = len + 1; for ( e = (p=(char*)str) + len; p': l += 3; break; case '&': l += 4; break; case '"': l += 5; break; /* do not replace ' w/ ' -- many clients don't grok it */ } ret = mAlloc( l ); if ( ! ret ) return 0; for ( p=ret; str': *p='&'; p[1]='g'; p[2]='t'; p[3]=';'; p+=4; break; case '&': *p='&'; p[1]='a'; p[2]='m'; p[3]='p'; p[4]=';'; p+=5; break; case '"': *p='&';p[1]='q';p[2]='u';p[3]='o';p[4]='t';p[5]=';';p+=6;break; default: *p++ = *str; } *p = 0; return ret; } /* toHtml */ int utf8Chk ( void *mem, int len, int *tof ) { unsigned char *c = (unsigned char *)mem; int f = 0; /* expected followers */ int l = len; if ( tof ) f = *tof; for ( ; l--; c++ ) { unsigned char u = *c; int tofollow = u < 128 ? 0 : u < 192 ? -1 : u < 224 ? 1 : 2; if ( f ) { if ( 0 > tofollow ) { f--; continue; } break; } if ( 0 <= tofollow ) { f = tofollow; continue; } goto croak; } if ( tof ) *tof = f; if ( ! f ) return 0; if ( 0 > l && tof ) /* exhausted buf, want follower */ return 0; sMsg( LOG_ERROR, "expected follower at %d got %x", len-l-1, (int)*c ); return len-l; croak: sMsg( LOG_ERROR, "expected no follower at %d got %x", len-l-1, (int)*c ); return len-l; } /* utf8Chk */ luti.h0100644000113000011300000001144707653266132010702 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef LUTI_H /* $Id: luti.h,v 1.30 2003/04/17 13:14:41 mawag Exp $ common utilities for openisis lib implementation. */ #include #include "loi.h" extern LogLevel log_lev; #define LOG_DO( lev ) ((lev) <= log_lev) #define LOG_DONT( lev ) ((lev) > log_lev) /** NOTE: unlike sMsg, this directly and immediatly lio_writes to file 2. So this is more expensive, but threadsafe. print some log message to stderr or some configured logdevice. for errors, use one of the LERR_xxx codes or a loglevel of SYSERR or IOERR to determine the error from errno. loglevels FATAL and ERROR, when used w/o error code, assum IDIOT and INVAL, resp. other logging levels do not assume an error (i.e. 0 == -LERR_OK). any error or loglevel code implies stream 2. @param code errorcode (-LERR_xxx) or loglevel as described above @return a non-positive code as described above */ extern int log_msg ( int code, const char *fmt, ... ); #ifndef NDEBUG # define LOG_DBG log_msg #else # ifdef __GNUC__ # define LOG_DBG( args... ) # else # define LOG_DBG (void) /* compiler should dispose statement off */ # endif #endif #define LOG_OTO( to, args ) do { ret = log_msg args; goto to; } while (0) /* content logging is usually on level LOG_TRACE */ #define LOG_STR( rec, desc ) log_str( LOG_TRACE, rec, desc ) #define LOG_HEX( mem, len ) log_hex( LOG_TRACE, mem, len ) /** dump a record @param level loglevel necessary for output to occurr. */ extern void log_str ( LogLevel level, int *rec, const char **desc ); /* * print fields of record to stderr * @param level if negative, print msg to stdout * @param delim field delimiter (default '; ') */ extern void log_rec (int level, Rec *rec, const char *msg, const char *delim); /** hexdump memory @param level loglevel necessary for output to occurr. */ extern void log_hex ( LogLevel level, const void *mem, int len ); /** do a malloc and print error if out of memory. */ extern void *lmalloc ( int size ); /* more specific log functions in log.h */ /* printint: optional sign, up to 10 digits, 0 byte. buf must be at least 12 bytes long. return length w/o the 0 byte. */ extern int lprint ( void *buf, int i ); /* Lookup of id from name */ typedef struct OpenIsisLT *LutiLT; /* create lookup table */ extern LutiLT luti_ltnew (); /* destroy lookup table */ extern void luti_ltdel (LutiLT lt); /* add entry */ extern void luti_ltadd (LutiLT lt, const char *name, int id); /* get id for name */ extern int luti_ltget (LutiLT lt, const char *name); /* remove entry */ extern void luti_ltrmv (LutiLT lt, const char *name); /* get fdt from rec */ extern Fdt* luti_fdt_from_rec (Rec *rec); /* get fd, tag, occurence from path */ extern const char* luti_parse_path ( const char *path, const Fdt *fdt, Fd **fd, int *tag, int *occ); /* extract embedded record from envelope */ extern Rec* luti_unwrap (Rec *env, int *pos, int tag, int dbid); /* extract embedded record from envelope */ extern Rec* luti_getembed ( Rec *env, const char *path, const Fdt *fdt); /* embed record in envelope */ extern Rec* luti_wrap (Rec *env, Rec *rec, int tag); /* append src fields to tgt */ extern Rec* luti_append (Rec *tgt, Rec *src); /** incr base ptr by incr*siz, if 0 < maxnum && *num < maxnum new allocated range is memsetted 0 NOTE: base is address of base for reallocation (whatever_type**) @return index of first new allocated address in array or -1 if out of mem */ extern int luti_ptrincr ( void *base, int *num, int incr, int siz, int maxnum); /** free pointer array */ extern void luti_free (void **arr, int num); /** check whether string represents a boolean value @return 0 str is false, 1 str is true, -1 str is something else */ extern int luti_true (const char *str, int len); /** s[0]*31^(n-1) + s[1]*31^(n-2) + ... + s[n-1] forced to non-negative */ extern int lhash ( const char *str, int len ); extern const char luti_hex[16]; /* '0'..'9','a'..'f' */ #define LUTI_H #endif /* LUTI_H */ make.mingw0100664000113000011300000000022207571422633011522 0ustar kripkripPATH:=/opt/cross-tools/i386-mingw32msvc/bin/:/opt/cross-tools/bin/:$(PATH) JNIINC=-I/opt/cross-tools/include/jni/win32 -I/usr/java/jdk1.3/include mlock.c0100664000113000011300000000212507644411042011010 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #include int main ( int argc, char **argv ) { size_t s = (size_t)1024*1024*atoi(argv[1]); char *p = (char *)malloc( s ); if ( ! mlock( p, s ) ) sleep( 10000 ); return 0; } msgdb/key.mst0100664000113000011300000000000007660413546012147 0ustar kripkripmsgdb/key.oxi0100664000113000011300000000400007662161243012143 0ustar kripkripüü¹î Þм ¦’  l VD -ùëÛ˹ MEN/SEARCH º*MEN/HELP ½*MEN/EDIT »*MEN/DB ¹*MEN/CFG ¼*LANG*CMD/SEARCH/RESULT Ï*CMD/SEARCH/HIST Î*CMD/SEARCH Í*CMD/HELP/INDEX ë*CMD/DB/OPEN Ã*CMD/DB/EXIT Å*CMD/DB/CLOSE Ä*CMD/CFG/TOGGLE â*CMD/CFG/LANG á*BTN/OK*BTN/HELP*BTN/CANCEL*msgdb/key.xrf0100664000113000011300000000000007660413546012143 0ustar kripkripoip.h0100644000113000011300000000237007644412745010512 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef OIP_H /* $Id: oip.h,v 1.2 2003/04/08 00:20:53 kripke Exp $ OpenIsis page */ /* standalone stdio version */ #include #define OIPRUN int main ( int argc, const char **argv ) { #define OIPEND return 0; } #define OIPP fprintf #define OIPF stdout #define OIPS( s ) fputs( s, stdout ) #define OIP_H #endif /* OIP_H */ oipc0100755000113000011300000000653307644412745010437 0ustar kripkrip#!/usr/bin/perl -w #/* # openisis - an open implementation of the CDS/ISIS database # Version 0.8.x (patchlevel see file Version) # Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # see README for more information #EOH */ # $Id: oipc,v 1.2 2003/04/08 00:20:53 kripke Exp $ # # usage # ./oipc foo.oip gcc %o && ./a.out # ./oipc foo.oip gcc %o -shared -o foo.oip.so && ./oipl $PWD/foo.oip.so sub mk { my $ifile = shift; my $ofile = $ifile . '.c'; my $buf = ''; # collected strings from output mode my @code = ('',''); # main and global code my $code = 0, $state = -1; # -1: file header # 0: stuff to output literally # run code: # 1: in <% # 2: in <%= # global code: # 3: in <%! # 4: in <%@ die "could not read $ifile" unless open I, $ifile; # transform line by line LIN: while () { if ( -1 == $state ) { # skipping initial # lines next if /^#/; $state = 0; # switch to output state } while (1) { if ( ! $state ) { # output up to <% my ($o,$s,$m,$x) = /^(.*?)(<%([=!@])?(.*))?$/s; if ( length $o ) { $o =~ s/\\/\\\\/g; $o =~ s/"/\\"/g; if ( defined $s ) { $buf .= '"' . $o . '"'; } else { $buf .= '"' . $o . '\n"' . "\n"; } } next LIN unless defined $s; $code[$code] .= "\tOIPS(\n".$buf."\t);\n" if length $buf; $buf = ''; ($m ||= 0) =~ y/=!@/123/; $state = 1+$m; $code = 2 < $state ? 1 : 0; $_ = $x; $code[$code] .= "\n#line ".$.."\n"; next LIN unless length $x; } my ($x,$s,$m,$o) = /^(.*?)((!)?%>(.*))?$/s; $buf .= $x; if ( defined $s ) { if ( 2 == $state ) { # <%= if ( 1 == length $buf ) { if ( $buf =~ /[i-n]/ ) { # integer shorthand $buf = '"%d",'.$buf; } elsif ( $buf =~ /[u-w]/ ) { # unsigned shorthand $buf = '"%u",'.$buf; } elsif ( $buf =~ /[f-h]/ ) { # OpenIsis Field shorthand $buf = '"%.*s",'.$buf.'->len,'.$buf.'->val'; } } if ( $buf =~ /^\s*"/ ) { $buf = 'OIPP(OIPF,'.$buf.');'; } else { $buf = 'OIPS('.$buf.');'; } } $code[$code] .= $buf; # don't reset print code destination if global ends with !%> $code = 0 unless $code && defined $m; $state = 0; if ( length $o ) { $buf = ''; $_ = $o; next; } $buf = '"\n"'; } else { $buf .= "\n"; } last; } } close I; $code[0] .= "OIPS(\n".$buf.");\n" if length $buf; die "could not write $ofile" unless open O, '>', $ofile; print O '#include "oip.h"',"\n"; print O $code[1]; print O "OIPRUN\n"; print O $code[0]; print O "OIPEND\n"; close O; map s/%o/$ofile/, @_; die 'exec '.join(' ',@_)." failed\n" unless exec @_; } mk @ARGV; oipl.c0100644000113000011300000000306307644412745010661 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: oipl.c,v 1.2 2003/04/08 00:20:53 kripke Exp $ OpenIsis page loader */ #include /* dlopen et al. */ #include int main ( int argc, const char **argv ) { int (*m)( int, const char ** ); const char *err = 0; const char *dlfile = argv[1]; void *dl = dlopen( dlfile, RTLD_LAZY ); int ret; if ( ! dl ) { fprintf( stderr, "could not open dlfile '%s': %s\n", dlfile, (err = dlerror()) ? err : "?" ); return 1; } m = dlsym( dl, "main" ); if ( ! m ) { fprintf( stderr, "could not link main in '%s': %s\n", dlfile, (err = dlerror()) ? err : "?" ); return 2; } ret = m( --argc, ++argv ); dlclose( dl ); return ret; } ois.c0100644000113000011300000000340207645521151010476 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: ois.c,v 1.8 2003/04/08 00:20:53 kripke Exp $ OpenIsis page server */ #include /* mem* */ #include "lsv.h" #if 0 /* the echo application */ static int echo ( Con *c, int task ) { Field *f; int i; if ( task ) /* preparing: we don't care */ return LSV_APPGOT==task; /* run it */ sMsg( c->ses, LOG_INFO, "echo %d: %d fields to %s", svCur()->id, c->req->len, c->nam ); for ( i=c->req->len, f = c->req->field; i--; f++ ) { RADDF( c->ses, c->ses->res, f, !0 ); /* plain protocol actually does not support plain output ... sMsg( c->ses, 0, "%d = '%.*s'\n", f->tag, f->len, f->val ); */ } return 0; } /* echo */ #endif int main ( int argc, const char **argv ) { Srv echo; memset( &echo, 0, sizeof(echo) ); echo.prt = svPlain; echo.app = svEcho; cOpen(0); cLog( LOG_DEBUG, 0 ); return svRun( &echo, 1/dev/tcp/$host/$port local i # running in subshell anyway ... declare -i i i=$requests ses=ses$[id / 4] while [ 0 != $i ]; do if [ 0 -eq $[i % 7] ]; then # add to session echo "-2 $ses $i $id $i foo $i $id $i bar " >&3 else # fetch session info only echo "-2 $ses " >&3 fi while read line; do # echo $line [ -z "$line" ] && break; done <&3 i=i-1 done echo "$id: `date`" } if [ x-trash == "x$1" ]; then requests=${2:-500} clients=${3:-8} port=${4:-8080} host=${5:-localhost} declare -i j echo "starting $clients trashers for $requests requests at `date`" j=$clients; while [ 0 != $j ]; do time trasher $j & j=j-1; done echo "started $clients trashers on $host:$port at `date`" j=$clients; while [ 0 != $j ]; do wait; j=j-1; done echo "collected $clients trashers at `date`" exit fi # default: run manually # exec 3<>/dev/tcp/${2:-localhost}/${1:-8080} cat <&3 & cp=$! cat >&3 kill $cp # copy kill ;/ openisis.c0100644000113000011300000005316407665653760011565 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: openisis.c,v 1.66 2003/05/30 13:09:04 kripke Exp $ main file of openisis executable. */ #ifdef HAVE_PTHREAD #define HAVE_THREADS #endif #include #include /* free */ #include /* strcmp */ #ifdef WIN32 # include # include # define timeval _timeb #else # include /* gettimeofday */ #endif #include /* gettimeofday */ #ifdef HAVE_PTHREAD #include /* threaded crashtest */ #endif #include "openisis.h" /* temporary library includes for testing of functions that should later be accessible via openisis.h */ #include "lio.h" /* ************************************************************ private types */ typedef enum { /* what to do */ DO_DUMP, DO_SCAN, /* simple full scan */ DO_SEARCH, /* basic index based searching */ DO_TERMS, /* list terms */ DO_CHECK, /* check db */ DO_PERF, /* do random reads for performance checking */ DO_CRASH, /* do multi-threaded crashtest */ DO_SPLIT, /* split a field value */ DO_STREAM, /* stream in records */ DO_MFNLIST, /* fetch records by mfn list */ DO_IFLOAD, /* read a lk2-style file from stdin */ DO_SWLOAD, /* load stopwords */ DO_IFDUMP, /* dump a lk2-style file to stdout */ DO_VUTF /* validate UTF-8 input */ } todo; typedef enum { FMT_MFN, /* rowid only */ FMT_MFNF, /* rowid, 1st field */ FMT_PROP, /* property style */ FMT_TXT, /* plaintext masterfile style */ FMT_TXTW /* plaintext masterfile style with W lines */ } format; typedef enum { IFM_DUMP, IFM_TAB, IFM_OLD, /* dump old index */ IFM_COPY, /* copy old index */ IFM_CHK /* check (new oxi) index */ } ifmode; /* ************************************************************ private data */ static const char *pft; static const char **term, **val; static OpenIsisSet *post; int nterm, nval; static int db = -1, wdb = -1, append = 0, idxall = 0; /* ************************************************************ private functions */ static int argchk ( const char *param, const char *n, const char *v ) { return strcmp( param, n ) ? 0 : v ? 1 : (openIsisSMsg( OPENISIS_ERR_INVAL, "no value for param '%s'", param ), 0); } /* argchk */ static int print ( OpenIsisRec *r, int freeit, format f ) { union { OpenIsisRec r; char buf[10000]; } x; int i, ret; if ( ! r ) { if ( FMT_TXT == f ) openIsisSMsg( 1, "\n" ); return -1; } ret = r->rowid; if ( pft ) { OpenIsisRec *q; OPENISIS_INITBUF(x); q = openIsisFmt( &x.r, pft, r ); if ( freeit ) free( r ); freeit = q != &x.r; r = q; } if ( FMT_MFN == f ) openIsisSMsg( 1, "%d\n", r->rowid ); else if ( FMT_MFNF == f ) openIsisSMsg( 1, "%d %.*s\n", r->rowid, 0 == r->len ? 1 : (int)r->field[0].len, 0 == r->len ? "-" : r->field[0].val ); else if ( FMT_TXT <= f ) { openIsisSMsg( 1, "\n" ); /* blank line */ if ( FMT_TXTW == f ) openIsisSMsg( 1, "W\t%d\n", r->rowid ); for ( i=0; ilen; i++ ) { if ( r->field[i].val ) openIsisSMsg( 1, "%d\t%.*s\n", r->field[i].tag, (int)r->field[i].len, r->field[i].val ); else openIsisSMsg( 1, "%d\t%d\n", r->field[i].tag, r->field[i].len ); } } else for ( i=0; ilen; i++ ) { if ( ! r->field[i].val ) { /* shouldn't happen -- numeric ? */ openIsisSMsg( 1, "%d.?=%d\n", r->rowid, r->field[i].len ); continue; } openIsisSMsg( 1, "%d.%d=%.*s\n", r->rowid, r->field[i].tag, (int)r->field[i].len, r->field[i].val ); if ( r->field[i].len && '^' == *r->field[i].val ) { /* split subfields */ OpenIsisRec *rf = openIsisReadField( 0, r->field+i ); if ( rf ) { int j; for ( j=0; jlen; j++ ) openIsisSMsg( 1, "%d.%d.%c=%.*s\n", r->rowid, r->field[i].tag, (0x60 & (int)rf->field[j].tag ) ? (int)rf->field[j].tag : ' ', (int)rf->field[j].len, rf->field[j].val ); free( rf ); } } } if ( 0 <= wdb ) { OpenIsisRec *q = 0; int ok; if ( append ) r->rowid = 0; if ( idxall && r != &x.r ) { /* add index entries for all fields */ OPENISIS_INITBUF(x); q = &x.r; for ( i=0; ilen; i++ ) { char hit[64]; OpenIsisField *fld = r->field + i; sprintf( hit, "%d.%d.%d.1 ", r->rowid, fld->tag, i ); OPENISIS_RADDS( q, OPENISIS_XHIT, hit, q != &x.r ); OPENISIS_RCAT( q, fld->val, fld->len, q != &x.r ); } } ok = openIsisDWritex( wdb, r, q ); openIsisSMsg( 1, "wrote mfn %d (%d)\n", r->rowid, ok ); } if ( freeit ) free( r ); return ret; } /* print */ static void printid ( int id, format f ) { if ( FMT_MFN == f && 0 > wdb ) openIsisSMsg( 1, "%d\n", id ); else print( openIsisReadRow( db, id ), !0, f ); } /* printid */ static int printlk2 ( void *me, OpenIsisKey *key, OpenIsisHit *hit ) { (void)me; if ( key && hit ) /* 30 key BLANK 7 mfn BLANK 5 tag BLANK 4 occ BLANK 4 pos*/ openIsisSMsg( 1, "%-30.*s %7u %5u %4u %4u\n", key->len, key->byt, hit->mfn, hit->tag, hit->occ, hit->pos ); return 0; } /* printlk2 */ static int printtab ( void *me, OpenIsisKey *key, OpenIsisHit *hit ) { (void)me; if ( key && hit ) /* 30 key BLANK 7 mfn BLANK 5 tag BLANK 4 occ BLANK 4 pos*/ openIsisSMsg( 1, "%.*s\t%u\t%u\t%u\t%u\n", key->len, key->byt, hit->mfn, hit->tag, hit->occ, hit->pos ); return 0; } /* printtab */ /* timing utility. set the timeval, return milliseconds since last call. */ #ifdef WIN32 static int millis ( struct _timeb *tb ) { struct _timeb otb = *tb; _ftime( tb ); return (tb->time - otb.time)*1000 + (tb->millitm - otb.millitm); } /* millis */ #else static int millis ( struct timeval *tv ) { struct timeval otv = *tv; gettimeofday( tv, 0 ); return (tv->tv_sec - otv.tv_sec)*1000 + (tv->tv_usec - otv.tv_usec)/1000; } /* millis */ #endif #ifdef HAVE_PTHREAD int myOpenIsisLockFunc ( int lock ) { static pthread_mutex_t mut = PTHREAD_MUTEX_INITIALIZER; /* the "fast" kind */ static pthread_cond_t cond = PTHREAD_COND_INITIALIZER; /* LOG_DBG( LOG_ERROR, "thread %d op 0x%08x", (int)pthread_self(), lock ); */ switch ( OPENISIS_WAIT & lock ) { case OPENISIS_RELE: return pthread_mutex_unlock( &mut ); case OPENISIS_LOCK: return pthread_mutex_lock( &mut ); case OPENISIS_WAKE: return pthread_cond_broadcast( &cond ); case OPENISIS_WAIT: return pthread_cond_wait( &cond, &mut ); } return -1; } #else # define myOpenIsisLockFunc 0 #endif typedef struct { int start; } threadarg; /* (p)thread routine */ static void * run ( void *arg ) { struct timeval tv; int m; threadarg *my = (threadarg *)arg; int start = my->start; int i; millis( &tv ); for ( i=start+1; i++!=start; ) { int j; if ( i >= nterm ) i = 0; for ( j=0; j= r->len ) openIsisSMsg( 2, "no fields for %d\n", mfn ); else if ( r->field[0].len != (int)strlen(val[mfn]) || memcmp( r->field[0].val, val[mfn], r->field[0].len ) ) openIsisSMsg( 2, "mismatch on %d\n", mfn ); free( r ); } } m = millis(&tv); openIsisSMsg( 1, "thread %d@%d terminated after %.3f seconds\n", #ifdef HAVE_THREADS (int)pthread_self() #else 0L #endif , start, m/1000. ); return (void *)m; } /* run */ /* multithreaded crashtest */ static int crash ( const char *pre ) { struct timeval tv; union { OpenIsisRec r; char buf[10000]; } x; int l; int p = 0; int i, pass; int fd = 2; threadarg def; millis( &tv ); x.r.len = 0; x.r.bytes = sizeof(x); nterm = 0; while ( openIsisTerm( &x.r, db, pre ) && x.r.len ) nterm += x.r.len; openIsisSMsg( fd, "%d terms\n", nterm ); openIsisSMsg( fd, "%.3f sec\n", millis(&tv)/1000. ); term = (const char **)malloc( nterm * sizeof(*term) ); post = (OpenIsisSet *)malloc( nterm * sizeof(*post) ); nterm = 0; nval = 0; while ( openIsisTerm( &x.r, db, pre ) && x.r.len ) { for ( i=0; ilen = 0; openIsisQuery( set, db, c, OPENISIS_QRY_KEYEQ, 0 ); if ( 0 >= set->len ) { openIsisSMsg( 2, "no results for '%s'\n", c ); return 1; } for ( cnt = 0; cnt < set->len; cnt++ ) if ( nval < set->id[cnt] ) nval = set->id[cnt]; p += set->len; } } openIsisSMsg( fd, "%d postings max mfn %d\n", p, nval ); openIsisSMsg( fd, "%.3f sec\n", millis(&tv)/1000. ); val = (const char **)malloc( (1+nval) * sizeof(*val) ); for ( i=1; i<=nval; i++ ) { OpenIsisRec *r = openIsisReadRow( db, i ); val[i] = 0; if ( ! r ) { openIsisSMsg( 2, "no rec %d\n", i ); continue; } if ( 0 < r->len ) { char *c = malloc( r->field[0].len + 1 ); memcpy( c, r->field[0].val, r->field[0].len ); c[ r->field[0].len ] = 0; val[i] = c; } free( r ); } l = millis(&tv); openIsisSMsg( fd, "sequential read %d rows in %.3f seconds %d rows per sec\n", nval, l/1000., nval*1000/(l?l:1) ); def.start = 0; i = 0; run( &def ); l = millis(&tv); run( &def ); l = millis(&tv); run( &def ); l = millis(&tv); openIsisSMsg( fd, "in-thread run in %.3f seconds %d rows per sec\n", l/1000., p*1000/(l?l:1) ); #ifdef HAVE_THREADS #define passes 3 { static int nThreads[] = { 8, 2, 1, 4 }; int res[ sizeof(nThreads)/sizeof(nThreads[0]) ]; pthread_t th[ 8 /* max. nThreads */ ]; threadarg arg[ sizeof(th)/sizeof(th[0]) ]; int j; for ( i=0; i<(int)(sizeof(nThreads)/sizeof(nThreads[0])); i++ ) res[i] = 0; for ( j=0; j<(int)(sizeof(th)/sizeof(th[0])); j++ ) #if 0 if ( ! (arg[j].ses = openIsisSesGet( -1, 0 )) ) { openIsisSMsg( fd, "could not get %dth session\n", j ); return 1; } #endif for ( pass=0; passvalue\n" ); ret = 1; goto bye; } if ( dowrite && 0 > (wdb = openIsisOpen( dowrite, 0, 0 )) ) { openIsisSMsg( 2, "could not open write target db '%s'\n", dowrite ); ret = 2; goto bye; } switch ( what ) { case DO_DUMP: { int max = openIsisMaxRowid( db ); int rowid; for ( rowid = 1; rowid <= max; rowid++ ) printid( rowid, fmt ); } break; /* DO_DUMP */ case DO_MFNLIST: { char *buf = 0; int l; while ( 0 <= (l = openIsisSReadln( &buf )) ) { int id = 0; while ( l-- ) id = 10*id + *buf++ - '0'; if ( id ) printid( id, fmt ); } } break; /* DO_MFNLIST */ case DO_SCAN: { int max = openIsisMaxRowid( db ); int rowid; for ( rowid = 1; 0 < rowid && rowid <= max; rowid++ ) rowid = print( openIsisScan( db, rowid, 0, search ), !0, fmt ); } break; /* DO_SCAN */ case DO_SEARCH: { int cnt; OpenIsisSet set; set.len = 0; openIsisQuery( &set, db, search, searchmode, 0 ); if ( 0 >= set.len ) { openIsisSMsg( 2, "no results for '%s'\n", search ); ret = 1; goto bye; } /* openIsisSMsg( 2, "%d\trows for\t%s\n", set.len, search ); */ for ( cnt = 0; cnt < set.len; cnt++ ) printid( set.id[cnt], fmt ); } break; /* DO_SEARCH */ case DO_TERMS: { union { OpenIsisRec r; char buf[10000]; } x; x.r.len = 0; x.r.bytes = sizeof(x); while ( openIsisTerm( &x.r, db, search ) && x.r.len ) { /* openIsisSMsg( 1, "%d terms\n", x.r.len ); */ for ( i=0; ilen; i++ ) openIsisSMsg( 1, "%c=%.*s\n", (int)r->field[i].tag, (int)r->field[i].len, r->field[i].val ); } break; case DO_STREAM: { OpenIsisIos ios; OpenIsisRecStream rs = { 0, OPENISIS_STOPONEMPTY, 0, 0, 0 }; LIO_SINIT( &ios, lio_stdio, "stdin", LIO_IN ); rs.in = &ios; /* some gcc versions need it this way */ while ( 0 < (i = openIsisSGetr( &rs )) ) print( rs.rec, 0, fmt ); } break; case DO_IFLOAD: case DO_SWLOAD: { OpenIsisKey key; OpenIsisHit hit; OpenIsisIndex idx = openIsisIdxOpen( db, intarg ); char *buf = 0; int l, lines = 0; memset( &hit, 0, sizeof(hit) ); hit.dbn = (-2 == intarg) ? 0xffff : 0; /* secret key for ifdel */ while ( 0 <= (l = openIsisSReadln( &buf )) && buf ) { char *t = memchr( buf, '\t', l ); if ( DO_SWLOAD == what ) { memcpy( key.byt, buf, key.len = (unsigned char)l ); } else if ( t ) { /* tab delimited */ key.len = (unsigned char)(t - buf); memcpy( key.byt, buf, key.len ); if ( 0 >= (l -= t-buf+1) || !(t = memchr( buf=t+1, '\t', l )) ) continue; hit.mfn = (unsigned) openIsisA2i( buf, t-buf ); if ( 0 >= (l -= t-buf+1) || !(t = memchr( buf=t+1, '\t', l )) ) continue; hit.tag = (unsigned short) openIsisA2i( buf, t-buf ); if ( 0 >= (l -= t-buf+1) || !(t = memchr( buf=t+1, '\t', l )) ) continue; hit.occ = (unsigned short) openIsisA2i( buf, t-buf ); if ( 0 >= (l -= t-buf+1) ) continue; hit.pos = (unsigned short) openIsisA2i( t+1, l ); } else { /* 10/30 key BLANK 7 mfn BLANK 5 tag BLANK 4 occ BLANK 4 pos*/ int eok = l - 24; /* pos of blank after key, 10 or 30 */ if ( 54 != l && 34 != l ) { openIsisSMsg( OPENISIS_ERR_INVAL, "bad ifload input len %d, want 34 or54 bytes + newline\n", l ); break; } for ( i=eok-1; ' ' == buf[i] && i--; ) ; key.len = (unsigned char) (++i); memcpy( key.byt, buf, key.len ); log_msg( LOG_VERBOSE, "'%.*s'", 7, buf+eok+1 ); hit.mfn = (unsigned) openIsisA2i( buf+eok+1, 7 ); hit.tag = (unsigned short)openIsisA2i( buf+eok+9, 5 ); hit.occ = (unsigned short)openIsisA2i( buf+eok+15, 4 ); hit.pos = (unsigned short)openIsisA2i( buf+eok+20, 4 ); } log_msg( LOG_VERBOSE, "'%.*s' %d %d %d %d", key.len, key.byt, hit.mfn, hit.tag, hit.occ, hit.pos ); if ( openIsisIdxAdd( idx, &key, &hit ) ) break; if ( !(++lines & 0x3ff) ) log_msg( LOG_INFO, "%dK lines", lines >> 10 ); } openIsisIdxDone( idx ); } break; case DO_IFDUMP: { OpenIsisIdxLoop l; memset( &l, 0, sizeof(l) ); l.flg = idxmode; switch ( ifm ) { case IFM_OLD: l.flg |= OPENISIS_IDXTRAD; case IFM_DUMP: l.cb = (OpenIsisIdxCb*)printlk2; break; case IFM_TAB: l.cb = (OpenIsisIdxCb*)printtab; break; case IFM_COPY: l.flg |= OPENISIS_IDXTRAD; l.me = openIsisIdxOpen( 0 <= wdb ? wdb : db, 0 ); l.cb = (OpenIsisIdxCb*)openIsisIdxAdd; break; case IFM_CHK: /* nuttin */ break; } if ( search ) memcpy( l.key.byt, search, l.key.len = (unsigned char)strlen( search ) ); if ( idxto ) memcpy( l.to.byt, idxto, l.to.len = (unsigned char)strlen( idxto ) ); openIsisIdxLoop( db, &l ); if ( IFM_COPY == ifm ) openIsisIdxDone( (OpenIsisIndex)l.me ); } break; case DO_VUTF: { char buf[1024]; int t = 0, f = 0, g; while ( 0 < (g = lio_read( &lio_in, buf, sizeof(buf) )) ) { int l = openIsisValidUTF8( buf, g, &f ); if ( l ) { openIsisSMsg( OPENISIS_ERR_INVAL, "at total %d = %d+%d\n", l-1+t, l-1, t ); ret = 1; goto bye; } t += g; } } } /* switch ( what ) */ bye: if ( 0 <= db ) openIsisClose( db ); if ( 0 <= wdb ) openIsisClose( wdb ); /* at least with WINE, atexit cleanup is not performed unless we explicitly call exit :( */ exit( ret ); return ret; } /* openisis main */ openisis.h0100644000113000011300000012564007665654533011570 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ #ifndef OPENISIS_H /* $Id: openisis.h,v 1.89 2003/05/30 13:15:07 kripke Exp $ public interface of the openisis lib. usage samples of most calls are found in openisis.c */ #define OPENISIS_VERSION "0.8.9" #ifndef __STDC__ # ifndef const # define const # endif #endif #ifdef __cplusplus extern "C" { #endif /* ************************************************************ record structure */ /** cooked version of an ISIS field. * values are *NOT* terminated by a 0 byte. */ typedef struct OpenIsisField { int tag; const char *val; int len; } OpenIsisField; /** cooked version of an ISIS Record. */ typedef struct OpenIsisRec { int dbid; int rowid; int bytes; /* total avail bytes */ int used; /* total used bytes (originally BWB) */ int fields; /* avail number of fields (originally BWP) */ int base; /* byte offset of contents area (after fields) */ int len; /* used number of fields */ int state; OpenIsisField field[1]; } OpenIsisRec; /* compute base from fields (for both raw and cooked record) */ #define OPENISIS_BASESZ( nfields ) ((int)sizeof(int)*(8 + 3*(nfields))) /* initialise record of f fields, n bytes */ #define OPENISIS_INITREC( r, n, f ) \ do { \ memset( r, 0, n ); \ (r)->bytes = (n); \ (r)->fields = (f); \ (r)->used = (r)->base = OPENISIS_BASESZ( (r)->fields ); \ } while (0) /* init with std 1 field per 100 bytes */ #define OPENISIS_INITSTD( r, n ) OPENISIS_INITREC( r, n, (n)/100 ) /* init a buffer structure which has a member r */ #define OPENISIS_INITBUF( x ) OPENISIS_INITSTD( &x.r, sizeof(x) ) #define OPENISIS_CLRREC( r ) \ do { \ (r)->len = (r)->state = 0; \ (r)->used = (r)->base = OPENISIS_BASESZ( (r)->fields ); \ } while (0) /* check some invariants (for both raw and cooked record) */ #define OPENISIS_RECOK( r ) \ ( (r)->base <= (r)->used \ && (r)->used <= (r)->bytes \ && (r)->len <= (r)->fields \ && (r)->base == OPENISIS_BASESZ( (r)->fields ) ) /** field description. */ enum { /* field types */ OPENISIS_FTX = 0, /* alphanum */ OPENISIS_FTA = 1, /* STRICTLY alpha */ OPENISIS_FTN = 2, /* numeric */ OPENISIS_FTP = 3, /* pattern */ OPENISIS_FTI = 10, /* ISO */ OPENISIS_FTE = 12, /* enum */ OPENISIS_FTB = 13, /* bool */ OPENISIS_FTT = 14, /* table */ OPENISIS_FTS = 15, /* structure */ OPENISIS_FTF = 16, /* subfield */ OPENISIS_FTV = 31 /* enum value */ }; #define OPENISIS_FD_NAMELEN 32 typedef struct OpenIsisFd { int id; /* tag */ char subf; /* subfield char or 0 */ char type; /* type */ char rep; /* !0 for repeated */ char slen; /* length of childs */ short len; /* max length or enum value */ char name[OPENISIS_FD_NAMELEN]; /* name, 0 terminated, max 30 */ char desc[OPENISIS_FD_NAMELEN]; /* 0 terminated description */ char *pat; /* 0 terminated pattern */ char *dflt; /* 0 terminated default value */ OpenIsisRec *info; /* additional application dependent attributes */ struct OpenIsisFd **subs; /* subfield childs */ } OpenIsisFd; /** field description table. */ typedef struct { int len; OpenIsisFd *fd; OpenIsisRec *rec; /* additional fields like formats, worksheets ... */ /* additional internal data like a hash might follow */ } OpenIsisFdt; typedef struct { /* where a key has a hit */ unsigned mfn; unsigned short tag; unsigned short occ; /* hit is in occ'th occurence of field tag in row mfn */ unsigned short pos; /* ... as the pos'th word */ unsigned short dbn; /* for multi-db index; ignored by now */ } OpenIsisHit; typedef struct { unsigned char len; unsigned char byt[23]; } OpenIsisVal; typedef struct { OpenIsisVal val; unsigned char len; unsigned char byt[255]; } OpenIsisKey; /** Writex tags, see doc/Inverted. Non-negative tags add (or delete) themselves. Negative control indexing. */ enum { OPENISIS_XCTL = -1, /* the "index" control field */ OPENISIS_XFST = -2, /* fst view line */ OPENISIS_XHIT = -3, /* dotted decimal hit plus val */ OPENISIS_XADD = -4, /* (prefix of) binary OpenIsisKey to add */ OPENISIS_XDEL = -5 /* (prefix of) binary OpenIsisKey to del */ }; /** get occurence of field. @param pos if given, the first occ starting at pos is searched and pos is set to one after the found position (or after end). may be used to loop all on an int var initialized to 0. @return the field or 0 */ extern OpenIsisField *openIsisRGet ( OpenIsisRec *r, int tag, int *pos ); extern OpenIsisField *openIsisROccurence ( OpenIsisRec *r, int tag, int occ ); /** similar to atoi/strtol, but - string needs not be 0 terminated (unless l < 0) - cares for hex 0x, but not octal 0. */ extern int openIsisA2i ( const char *p, int l ); /** similar to openIsisA2i but - return number of parsed characters - put result in *res */ extern int openIsisA2il ( const char *p, int l, int *res ); /** similar to openIsisA2i but - return dflt if less than l characters of p has been successfully parsed */ extern int openIsisA2id ( const char *p, int l, int dflt ); /** print 0 terminated int. p must have 12 bytes space @return number of chars (up to 10 digits + minus sign) */ extern int openIsisI2a ( char *p, int i ); extern int openIsisU2a ( char *p, unsigned u ); /** get field value as int as of A2i. boolean values are recognized as 0 and 1. if not found, return def. */ extern int openIsisRInt ( OpenIsisRec *r, int tag, int def, int *pos ); /** * get field value as string from rec or default rec. * note: def should be a non-legal value, since a def result in the * first record causes a lookup in dflt. */ extern int openIsisRInt2 (OpenIsisRec *rec, OpenIsisRec *dflt, int tag, int def); /** get field value as enum. if not found, return def. */ extern int openIsisREnum ( OpenIsisFdt *fdt, OpenIsisRec *r, int tag, int def, int *pos ); /** get field value as string of length len (including 0-byte). if not found, return 0. */ extern char* openIsisRString (OpenIsisRec *r, int tag, int *pos, char *buf, int len); /** * get field value as string from rec or default rec. */ extern char* openIsisRString2 (OpenIsisRec *rec, OpenIsisRec *dflt, int tag, char *buf, int len); /** flatten (serialize) text record to tag\tval lines ended by a blank line. Text means the record is assumed to not contain vertical tabs of any importance, so newlines in field values are replaced with vertical tabs (^K). buf must be of size rec->used, wich has enough room for all values + 24bytes header + 12 bytes per field (ok for sign+10digits+tab). @return # of bytes written */ extern int openIsisRSerialize ( char *buf, OpenIsisRec *rec ); /** like Serialize, but newlines are turned into newline-tab. buf must be of size 2*rec->used (unless you know you don't have newlines) */ extern int openIsisRSerializeBin ( char *buf, OpenIsisRec *rec ); /** alloc buffer with appropriate length, if necessary. @param *len input: length of buf @param *len output: length of serialized record written to returned buffer @return buffer, must be freed after usage if reallocated (i.e. != buf), null on error */ extern char* openIsisRSerializeAlloc (OpenIsisRec *rec, char *buf, int *len); /** flatten fully transparently, replacing newlines with newline-tab. extern int openIsisRBinSerialize ( char *buf, int len, OpenIsisRec *rec ); */ /* ************************************************************ fdt */ /** get field description by id (tag). */ extern OpenIsisFd* openIsisFById ( const OpenIsisFdt *fdt, int id, int subf ); /** get field description by name. */ extern OpenIsisFd* openIsisFByName (const OpenIsisFdt *fdt, const char *name); enum { OPENISIS_NOENUM = 0x10000 /* enum NOENUM :) -- not a short */ }; /** lookup enum value for field id. name must be exact match or unique prefix. @return the enum value or NOENUM */ extern int openIsisFEnum ( OpenIsisFdt *fdt, int id, const char *name ); /** free fdt and fd structures. @return 0 */ extern OpenIsisFdt* openIsisFFree (OpenIsisFdt *fdt); /** convert fdt to record. @param fdt source @param rec target where to append fd, may be null @param embchld boolean indicating whether subfield childs should be embedded in fd part of target rather than in fdt part @return target or null if no memory available */ extern OpenIsisRec* openIsisFFdt2Rec ( const OpenIsisFdt *fdt, OpenIsisRec *rec, int embchld); /** build fdt from rec. @param rec source @return fdt or null on error */ extern OpenIsisFdt* openIsisFRec2Fdt (OpenIsisRec *rec); /* ************************************************************ session type and utilities */ typedef struct OpenIsisSes *OpenIsisSession; extern void *openIsisMAlloc ( int size ); extern void openIsisMFree ( void *mem ); extern void *openIsisMDup ( const void *str, int size ); /** convert string to html by replacing the four characters <,>,& and ". free the returned pointer after usage. XML has a fifth standard entity apos, but many HTML clients don't grok it. */ extern char *openIsisToHtml ( const char *str, int len ); extern int openIsisUtf8Chk ( void *mem, int len, int *tof ); /* ************************************************************ stream IO */ typedef struct OpenIsisIos *OpenIsisStream; typedef int OpenIsisSFunc ( OpenIsisStream s, int op ); extern int openIsisSOpen ( const char *name, int flags, OpenIsisSFunc *type ); /** error and loglevel codes one nibble each. */ enum { /* our very own errno */ OPENISIS_ERR_OK = 0, /* 0 is no error, also read(2)'s EINTR, EAGAIN */ /* errors logged at level VERBOSE */ OPENISIS_ERR_EOF = 0x10000, /* end of file */ /* errors logged at level ERROR */ OPENISIS_ERR_FAULT = 0x20000, /* NULL pointer or bad sized buffer given */ OPENISIS_ERR_INVAL = 0x30000, /* general invalid parameters, any EINVAL errno */ OPENISIS_ERR_BADF = 0x40000, /* bad file, also read(2)'s EINVAL, some of open(2) */ /* errors logged at level SYSERR */ OPENISIS_ERR_IO = 0x50000, /* real IO error, also write(2)'s ENOSPC, EPIPE */ OPENISIS_ERR_NOMEM = 0x60000, /* out of memory, also open(2)'s EMFILE, ENFILE */ OPENISIS_ERR_BUSY = 0x70000, /* object is busy */ /* errors logged at level FATAL */ OPENISIS_ERR_TRASH = 0x80000, /* database internal consistency */ OPENISIS_ERR_IDIOT = 0x90000 /* caught programming error */ }; typedef enum { OPENISIS_LOG_OFF = 0, /* don't log anything */ OPENISIS_LOG_FATAL = 0x100000, /* fatal internal errors: we can't go on */ OPENISIS_LOG_SYSERR = 0x200000, /* problem with system ressources: bad file, no mem */ OPENISIS_LOG_IOERR = 0x300000, /* problem on IO */ OPENISIS_LOG_ERROR = 0x400000, /* unusable input, database or query */ OPENISIS_LOG_WARN = 0x500000, /* bad input */ OPENISIS_LOG_INFO = 0x600000, /* some major event like opening a db */ OPENISIS_LOG_VERBOSE = 0x700000, /* any event like reading a record */ OPENISIS_LOG_TRACE = 0x800000, /* database content (log_str) */ /* highest levels only with debug built */ OPENISIS_LOG_DEBUG = 0x900000, /* lots of processing details */ OPENISIS_LOG_ALL = 0xa00000 /* just everything, even built-in content */ } OpenIsisLogLevel; enum { OPENISIS_ERR_ERRORS = 10, OPENISIS_ERR_MASK = 0xf0000, /* mask for filtering error code */ OPENISIS_ERR_SHIFT = 16, /* shift to make codes 0,1...ERR_ERRORS */ OPENISIS_LOG_NOCHANGE = -1, /* leave as is when initialising log */ OPENISIS_LOG_LEVELS = 11, OPENISIS_LOG_MASK = 0xf00000, /* mask for filtering log levels */ OPENISIS_LOG_SHIFT = 20, /* shift to make codes 0,1...LOG_LEVELS */ /* the following ids are provided just in case you don't feel comfortable assuming that in,out,err are always 0,1,2 ;) typically, you don't need any of these, 0 will work ... */ OPENISIS_SIN = 0, /* stream id of stdin */ OPENISIS_SOUT = 1, /* stdout */ OPENISIS_SERR = 2, /* stderr */ OPENISIS_SMASK = 0xffff, /* mask of stream id part of to */ OPENISIS_MSG_VA = 0x1000000 /* 1st parameter after fmt IS a va_list */ }; /** printf to session's stream given by to. The streams 0,1 and 2 of the default session are usually bound to the system file descriptors 0,1 and 2 under unixes. On windows, 2 is redirected to file oisiserr.txt. 0 and 1 are bound to CONIN$ and CONOUT$ for console apps, to NULL and oisisout.txt for GUI apps. For server sessions, 0 and 1 are connected to the client and 2 is redirected to some per-session logging. @param to the lower bytes as given by SMASK specify the stream to use. the higher bytes may specify loglevel and or error code. if either loglevel or error code is given, the other will be derived accordingly and stream defaults to 2 (if stream id is 0). otherwise, stream defaults to 1, so you never can print to 0. @return the negative error code */ extern int openIsisSMsg ( int to, const char *fmt, ... ); /** get a string delimited by char delim (typically '\n') on stream number fd. on success, ptr is set to the string and it's length is returned, not including the delimiting char. Note that there need not be a delimiting char near end of file. If there is no string available, ptr is set to 0 (NULL). If input might be available later, 0 is returned, else some negative error code. */ extern int openIsisSGets ( int fd, char **ptr, char delim ); #define openIsisSReadln( p ) openIsisSGets( 0, p, '\n' ) typedef struct OpenIsisRecStream { OpenIsisStream in; /* input stream to read from */ int flg; /* flags */ OpenIsisRec *buf; /* original buffer */ OpenIsisRec *rec; /* record read */ OpenIsisRec *dict; /* dictionary of field names */ } OpenIsisRecStream; /* stream in records in some free-style format suitable for mails, properties, syspar ... fields are delimited by CR or NL or CR-NL within a field, tag stops at blank,TAB,= or : after a TAB, val starts immediately after a = or :, val starts at 1st character not a blank after a blank, val starts at 1st char not one of the above separators processing - a field or record separator is recognized anywhere - within each field, the tag is read up to the first separator (any "special" char, all controls, whitespace and most punctuation). upon reaching a separator, the tag is identified. if it is empty, and the record has a field, now new field is created. else, if the tag cannot be identified, a tag of -1 is used and the tag string appended as value. - the separator consists of any number of leading blanks, some separator char, and, unless the sepchar is TAB, trailing blanks. if the current field is not empty, the separator is appended as value. - the rest of the field is appended as value - a field with empty tag and value (blank line) terminates the record in most modes - a field with empty tag (continuation line) has it's separator AND value appended to the previous field - if either a tag or a separator extend over more than 4k, the stream is considered over limit. */ enum { OPENISIS_STOPONEMPTY = 0x10, /* stop at empty line */ OPENISIS_AUTOCLONE = 0x20, /* clone the record even if not needed */ OPENISIS_TRANSPARENT = 0x40, /* do not convert vtabs to newlines */ OPENISIS_MAIL = 0x100, /* process mail body */ OPENISIS_MBOX = 0x200, /* input stream is mbox format */ OPENISIS_DETACH = 0x400, /* detach body */ OPENISIS_CONVERT = 0x800 /* fix body parts */ }; extern int openIsisSGetr ( OpenIsisRecStream *stream ); /* ************************************************************ record utilities requiring a session */ /** clone record or create empty record. with room >= 0 and rec != NULL, make sure that there is at least one field and room bytes available. else, it works in shrink-to-fit mode: exactly the required bytes are used. @param rec to extend, may be NULL @param room number of required free bytes (<0 for shrink-to-fit) @param discard wether the original record should be freed after extension @return a rec with suitable room */ extern OpenIsisRec* openIsisRDup ( OpenIsisRec *r, int room, int discard ); /** printf a new field of up to 1024 bytes at the end of record. */ extern OpenIsisRec* openIsisRMsg ( OpenIsisRec *r, int discard, int tag, const char *fmt, ... ); /** check for free headspace in a record. make sure that there is at least one field and room bytes available. NOTE: this may change recp @see openIsisRDup */ #define OPENISIS_RSPACE( recp, room, disc ) \ do { if ( ! (recp) \ || (recp)->len >= (recp)->fields \ || (recp)->bytes < (recp)->used + (room) \ ) (recp) = openIsisRDup( (recp), (room), (disc) ); } while (0) /** add a field to a record. a new field entry and n bytes space are provided. if the pointer s is not NULL, n bytes are copied from it to the field. NOTE: this may change recp @see openIsisRDup */ #define OPENISIS_RADD( recp, ntag, s, n, disc ) \ do { \ int _ll = (n); \ OPENISIS_RSPACE( recp, _ll, disc ); \ if ( (recp) ) { \ OpenIsisField *_ff = (recp)->field + (recp)->len++; \ _ff->tag = (ntag); \ _ff->val = ((char*)recp)+(recp)->used; \ if ( ! s ) \ _ff->len = 0; \ else \ memcpy( (char*)_ff->val, s, _ff->len = _ll ); \ (recp)->used += _ff->len; \ } \ } while(0) /** append n bytes to the last field of a record. NOTE: this may change recp NOTE: this requires the last field to reside at the end of the buffer @see openIsisRDup */ #define OPENISIS_RCAT( recp, s, n, disc ) \ do { \ OpenIsisField *_ff; \ int _ll = (n); \ assert ( (recp) ); \ assert ( (recp)->len ); \ if ( (recp)->bytes < (recp)->used + _ll ) \ (recp) = openIsisRDup( (recp), _ll, (disc) ); \ _ff = (recp)->field + (recp)->len - 1; \ memcpy( (char*)(_ff->val+_ff->len), s, _ll ); \ _ff->len += _ll; \ (recp)->used += _ll; \ } while(0) /* like OPENISIS_RADD, but based on strlen */ #define OPENISIS_RADDS( recp, ntag, s, disc ) \ OPENISIS_RADD( recp, ntag, s, strlen(s), disc ) /* like OPENISIS_RADD, but based on field */ #define OPENISIS_RADDF( recp, f, disc ) \ OPENISIS_RADD( recp, (f)->tag, (f)->val, (f)->len, disc ) /* like OPENISIS_RCAT, but based on strlen */ #define OPENISIS_RCATS( recp, s, disc ) \ OPENISIS_RCAT( recp, s, strlen(s), disc ) /** create a record from a field, i.e. split subfields. if a rec buffer is given, it is filled up to it's rec->len and returned. else an appropriately allocated rec is returned, which must be freed after usage. NOTE: USE WITH CARE, IT SOMEWHAT BREAKS THE GENERAL CONTRACT OF REC! DO NOT TRY TO EXTEND THE RETURNED RECORD !!! The rec does NOT have it's own data buffer, but instead field values are pointers into the input field's value. The returned rec will therefore become invalid, if the inpur field is freed. @param rec a rec buffer to fill or NULL @param field some string to be split up at '^' characters @return an OpenIsisRec or NULL, if splitting failed. */ extern OpenIsisRec *openIsisRSplitf ( OpenIsisRec *rec, const OpenIsisField* field ); enum { /* RSet flags */ OPENISIS_RARGC = 0xffff, /* argc mask */ OPENISIS_RFDT = 0x0010000, /* first vararg is the Fdt to use */ OPENISIS_RARGV= 0x0020000, /* next vararg is a char **argv */ OPENISIS_RDIS = 0x0040000, /* discard original record on change */ OPENISIS_RNOC = 0x0080000, /* do not recompact after CHG/DEL op */ OPENISIS_RIGN = 0x0100000, /* ignore unknown fields */ /* default op is to append fields at the end */ OPENISIS_RCHG = 0x1000000, /* change mode: overwrite previous field values */ OPENISIS_RDEL = 0x2000000, /* args is a list of tags to delete */ OPENISIS_RDFLT= 0x4000000, /* default mode: set field value only if not already present */ OPENISIS_ROP = 0xf000000 /* op mask */ }; /** set fields in a record. this is going to be the real field setting one-stop-shop. This is called EITHER with real varargs tag, val, tag, val ... 0, stopped by a 0 tag, OR with an char **argv as vararg. In most modes, the varargs are like tag, val, tag, val ... Tags may be either numeric (int) or by field name. Field names are translated to numerical ids by the Fdt, which may be given explicitly or derived from the rec's dbid. @param rec record to modify or 0 to create new one @param mode bitwise OR of flags and the number of fields (2 byte) if the number of fields is 0, args must be zero terminated @return the same record or a copy, depending on space needs */ extern OpenIsisRec *openIsisRSet ( OpenIsisRec *rec, int mode, ... ); /** add serialized "text" to rec supports the RDIS, STOPONEMPTY and TRANSPARENT flags @return # of bytes read */ extern int openIsisRDeserialize ( OpenIsisRec **rec, const char *buf, int len, int flg ); /** append tag with integer value to rec */ extern OpenIsisRec *openIsisRAddI ( OpenIsisRec *rec, int tag, int value, int discard); enum { OPENISIS_FMT_SHARP = 32, /* # */ OPENISIS_FMT_SLASH, /* / */ OPENISIS_FMT_PERCENT, /* % */ OPENISIS_FMT_OPEN, /* { */ OPENISIS_FMT_CLOSE, /* } */ OPENISIS_FMT_ESC, /* ! */ OPENISIS_FMT_B, /* bold */ OPENISIS_FMT_BOX, /* draw a box (0i) */ OPENISIS_FMT_BPICT, /* draw background pict (1sn) */ OPENISIS_FMT_C, /* column (1i) */ OPENISIS_FMT_CL, /* colour (1i) */ OPENISIS_FMT_COLS, /* colour table (s_) - emits a string of hex vals */ OPENISIS_FMT_FONTS, /* font table (s_) */ OPENISIS_FMT_FS, /* font size (1i) */ OPENISIS_FMT_F, /* font (1i) */ OPENISIS_FMT_I, /* italic */ OPENISIS_FMT_LINK, /* link (2s_) */ OPENISIS_FMT_LW, /* line width (1i) */ OPENISIS_FMT_M, /* indentation in twips(1in) */ OPENISIS_FMT_NC, /* new column (0i) */ OPENISIS_FMT_NEWLINE, /* set newline (1s) */ OPENISIS_FMT_NP, /* newpage (0i) */ OPENISIS_FMT_PICT, /* draw pict (1s) */ OPENISIS_FMT_QC, /* center */ OPENISIS_FMT_QJ, /* justify */ OPENISIS_FMT_TAB, /* tab (0i) */ OPENISIS_FMT_UL, /* underline */ OPENISIS_FMT_X /* blanks (1i) */ }; extern OpenIsisRec* openIsisRFmt ( OpenIsisRec *buf, const char *fmt, OpenIsisRec *r ); /* ************************************************************ db access */ /** get highest rowid in db @param dbid id returned by openIsisOpen @return positive rowid on success, 0 if unknown, negative on error */ extern int openIsisDMaxId ( int dbid ); /** read a cooked row. free the returned pointer after usage. @param db id returned by openIsisOpen @param rowid the mfn @return an OpenIsisRec or NULL, if row not available. */ extern OpenIsisRec *openIsisDRead ( int db, int rowid ); /** for convenience: read and format in one. */ extern OpenIsisRec *openIsisDFmt ( OpenIsisRec *buf, const char *fmt, int db, int rowid ); /** write a record and/or index entries. If the rowid is 0, the record will be modified by assigning a new rowid. @param db id returned by openIsisOpen @param rec the OpenIsisRec to write @param idx record containing index entries @return 0 or an error code */ extern int openIsisDWritex ( int dbid, OpenIsisRec *rec, OpenIsisRec *idx ); /** write a record as with DWritex. index entries should be derived automatically. */ extern int openIsisDWrite ( int dbid, OpenIsisRec *rec ); #define OPENISIS_DWRITE( rec ) openIsisDWrite( (rec)->dbid, rec ) /** simple full scan search. read the first row with rowid >= given rowid and containing txt in subfield tag (any, if < 0). free the returned pointer after usage. @param db id returned by openIsisOpen @param rowid the mfn @param tag field in which to look for key, 0 for any @return an OpenIsisRec or NULL, if row not available. */ extern OpenIsisRec *openIsisDScan ( int db, int rowid, int tag, const char *key ); /** get a term for key w/ or w/o the postings. prefix mode is determined by a trailing '$' in key. for prefix mode, the last preset field in the record gives a starting point, i.e. terms up to and including it are ignored. thus, if there are more matching terms than fit in the record, the search can be continued by passing the same record. @return an OpenIsisRec or NULL */ extern OpenIsisRec* openIsisDTerm ( OpenIsisRec *rec, int db, const char *key ); #ifndef OPENISIS_SETLEN #define OPENISIS_SETLEN 1000 #endif /** record set type as returned from openIsisSearch. first entry is the set length n, followed by rowids 1..n. actual array length is thus n+1. */ typedef struct { int len; int id[OPENISIS_SETLEN]; } OpenIsisSet; /* query mode */ enum { /* search given string using index */ OPENISIS_QRY_KEYEQ, /* index scan key equal */ OPENISIS_QRY_KEYPF, /* index scan key prefix */ OPENISIS_QRY_KEYAT, /* index scan key auto (checks for '$') */ /* fulltext scan for given string */ OPENISIS_QRY_SCANE=64, /* fulltext scan for equal */ OPENISIS_QRY_SCANC, /* fulltext scan for contains */ /* parse string as query expression */ OPENISIS_QRY_SIMPLE=128, /* simple left-to-right binding */ OPENISIS_QRY_PROPER /* proper binding using precedence, () */ }; /** query the db. @param set set buffer to fill. set->len must contain the actual buffer length (# of ids). 0 means default length OPENISIS_SETLEN. @param db id returned by openIsisOpen @param key keyword or query expression to search for @param mode query mode, optionally |ed with a field tag<<16. a field tag affects key and scan modes. @param skip mfns < skip are ignored, useful to continue a query @return the number of rowids in set (= set->len) on success, <0 on error. */ extern int openIsisDQuery ( OpenIsisSet *set, int db, const char *key, int mode, int skip ); /** callback for index loop */ typedef int OpenIsisDXCb ( void *me, OpenIsisKey *key, OpenIsisHit *hit ); typedef struct { int flg; void *me; /* commonly the session */ OpenIsisDXCb *cb; OpenIsisKey key; OpenIsisKey to; } OpenIsisDXLoop; enum { /* flags */ OPENISIS_IDXPF = 0, /* loop where from is prefix of key */ OPENISIS_IDXEQ = 1, /* loop from == key */ OPENISIS_IDXUPTO = 2, /* loop from <= key < to */ OPENISIS_IDXINCL = 3, /* loop from <= key <= to */ OPENISIS_IDXMODE = 0x03, /* mask for match mode */ OPENISIS_IDXTRAD = 0x10000 /* use traditional index */ }; /** */ extern int openIsisDXLoop ( int dbid, OpenIsisDXLoop *l ); /* ************************************************************ system control -- session 0 only */ enum { /* system (schema 0) parameters */ OPENISIS_SPATH = 5, /* db dir */ OPENISIS_SLOGF = 700, /* logfile */ OPENISIS_SLOGV = 701, /* verbosity */ /* schema parameters */ OPENISIS_SC_NAME = 710, /* schema n name */ OPENISIS_SC_HOST = 711, /* schema n host */ OPENISIS_SC_PORT = 712, /* schema n port */ OPENISIS_SC_DFLTDB = 721, /* schema default db */ /* database parameters */ OPENISIS_DNAME = 800, OPENISIS_DTYPE = 801, OPENISIS_DRO = 802, OPENISIS_DPATH = 803, OPENISIS_DDUMP = 804, OPENISIS_DENC = 810, /* encoding */ OPENISIS_DFDT = 811, /* path to serialized fdt */ OPENISIS_DFMT = 850, /* worksheets */ OPENISIS_DPFT = 851, /* printformats */ OPENISIS_DFST = 852, /* field selections */ /* fdt entries */ OPENISIS_FDID = 860, /* tag */ OPENISIS_FDSUB = 861, /* subfield char */ OPENISIS_FDTYPE = 862, /* field type */ OPENISIS_FDREP = 863, /* repeatable flag */ OPENISIS_FDNUMC = 864, OPENISIS_FDLEN = 865, /* length: 0 variable, >0 maxlen, <0 fixed len */ OPENISIS_FDNAME = 866, /* field name */ OPENISIS_FDDESC = 867, /* field description */ OPENISIS_FDPAT = 868, /* pattern */ OPENISIS_FDDFLT = 869, /* default value */ OPENISIS_FDINFO = 870, /* application specific addons */ OPENISIS_FDCHLD = 871, OPENISIS_FDT_LEN = 880, /* fdt: number of fd's */ OPENISIS_FDT_FD = 881, /* fdt: embedded fd */ OPENISIS_FDT_REC = 882, /* fdt: application specific addons */ /* communication parameters */ OPENISIS_COM_SID = 900, /* session id */ OPENISIS_COM_SER = 901, /* serial no of request */ OPENISIS_COM_DBN = 902, /* db name */ OPENISIS_COM_TMS = 903, /* server timestamp */ OPENISIS_COM_ROW = 904, /* rowid */ OPENISIS_COM_CFG = 907, /* embedded cfg */ OPENISIS_COM_REC = 908, /* embedded data record */ /* communication parameters: request */ OPENISIS_RQS_TYPE = 920, /* request type */ OPENISIS_RQS_FLG = 921, /* flags */ OPENISIS_RQS_QMOD = 922, /* mode (query) */ OPENISIS_RQS_SKIP = 923, /* skip (query) */ OPENISIS_RQS_SIZE = 924, /* output size (query) */ OPENISIS_RQS_KEY = 925, /* key value (query) */ OPENISIS_RQS_IDX = 926, /* index record (insert, update) */ /** communication parameters: response */ OPENISIS_RSP_DBID = 940, /* local db id */ OPENISIS_RSP_ERR = 941, /* error code */ OPENISIS_RSP_MSG = 942, /* error message */ OPENISIS_RSP_NUMT = 943, /* total number of records (query, read) */ OPENISIS_RSP_NUMR = 944, /* number in this response (query, read) */ OPENISIS_RSP_CERR = 945 /* client side error */ }; extern const OpenIsisFdt *openIsisFdtSyspar; /* schema 0 */ extern const OpenIsisFdt *openIsisFdtScheme; /* schema n */ extern const OpenIsisFdt *openIsisFdtDbpar; /* db cfg */ extern const OpenIsisFdt *openIsisFdtFd; /* fd rec */ extern const OpenIsisFdt *openIsisFdtFdt; /* fdt rec */ extern const OpenIsisFdt *openIsisFdtRqs; /* request rec */ extern const OpenIsisFdt *openIsisFdtRsp; /* response rec */ /** request types. */ enum { OPENISIS_RQST_OPEN = 1, /* open db */ OPENISIS_RQST_CLOS = 3, /* close db */ OPENISIS_RQST_MNT = 4, /* mount db */ OPENISIS_RQST_LSDB = 6, /* list available dbs */ OPENISIS_RQST_MROW = 10, /* get maxrowid of db */ OPENISIS_RQST_QRY = 11, /* exec query on db */ OPENISIS_RQST_READ = 12, /* get rec for rowid */ OPENISIS_RQST_INS = 20, /* insert rec */ OPENISIS_RQST_UPD = 21, /* update rec */ OPENISIS_RQST_DEL = 22 /* delete row */ }; /** request flags. */ enum { OPENISIS_RQSF_MNT = 0x0001, /* mount db */ OPENISIS_RQSF_QRR = 0x0002 /* fetch records on query */ }; /** max length of key in query request. */ #define OPENISIS_QRY_KEYLEN 1024 /** max length of error messages. */ #define OPENISIS_ERRMSGLEN 1024 enum { OPENISIS_RELE = 0x0000000, /* release, unlock */ OPENISIS_LOCK = 0x1000000, OPENISIS_WAKE = 0x2000000, /* broadcast signal, notifyall */ OPENISIS_WAIT = 0x3000000, OPENISIS_COND = 0x0ffffff /* condition mask */ }; /** a function to provide a global lock in a multi threaded process. @param lock coded action. the high byte is one of release, lock, wake and wait. the lower bytes indicate a condition for wake and wait. the value may be hashed to a small set of conditions, typically eight or one (ignore). @return 0 on success sample with a single condition: int myOpenIsisLockFunc ( int lock ) { static pthread_mutex_t mut = PTHREAD_MUTEX_INITIALIZER; static pthread_cond_t cond = PTHREAD_COND_INITIALIZER; switch ( OPENISIS_WAIT & lock ) { case OPENISIS_RELE: return pthread_mutex_unlock( &mut ); case OPENISIS_LOCK: return pthread_mutex_lock( &mut ); case OPENISIS_WAKE: return pthread_cond_broadcast( &cond ); case OPENISIS_WAIT: return pthread_cond_wait( &cond, &mut ); } return -1; } */ typedef int OpenIsisCLockFunc ( int lock ); /** initialize system with some common parameters and a lock func. @param lock a function to provide a global lock in a multi threaded process. */ extern int openIsisCInit ( int argc, const char **argv, OpenIsisCLockFunc lock ); /** set logging level and device. @param level the logging level may be given numerical or as ascii char in -ofsewivtda for nochange(-1), off (0), fatal (1), syserr, error, warn, info, verbose, trace, debug, all(9) default is set to error @param filename specifies, if not NULL, a file to append log to */ extern void openIsisCLog ( int level, const char *filename ); /** get the default session, possibly adding args. only one (session master) thread in a process may call this. @param args initialisation properties (cloned to prop) */ extern OpenIsisSession openIsisCOpen ( OpenIsisRec *args ); /** get a new session. only one (session master) thread in a process may call this. @param args initialisation properties (cloned to prop) */ extern OpenIsisSession openIsisCSession ( OpenIsisRec *args ); /* ************************************************************ schema */ #define OPENISIS_SC_NAMELEN 32 /* max length of schema identification */ #define OPENISIS_DB_NAMELEN 32 /* max length of db id relative to schema */ #define OPENISIS_SCIDMSK 0x0ff00 /* schema part of id */ #define OPENISIS_DBIDMSK 0x000ff /* db part of id */ typedef struct { char name[OPENISIS_DB_NAMELEN]; int dbid; int mntcnt; /* number of sessions currently using this */ int tms; /* last (re)opening timestamp */ OpenIsisRec *cfg; OpenIsisFdt *fdt; } OpenIsisDb; typedef struct { char name[OPENISIS_SC_NAMELEN]; int scid; int mntcnt; OpenIsisRec *cfg; int ndbs; OpenIsisDb **dbs; } OpenIsisSchema; /* ************************************************************ client side */ typedef struct OpenIsisStb *OpenIsisStub; /** function that sends request record rqs to remote address of stb */ typedef int OpenIsisStbRqsProc (OpenIsisStub stb, OpenIsisRec *rqs); /** callback for freeing external resources when a stub is deleted. @param client_data client_data registered in openIsisNOpen */ typedef void OpenIsisStbDelProc (OpenIsisStub stb, void *client_data); /** deprecated */ typedef int OpenIsisRspCb (void *client_data, OpenIsisStub stub, OpenIsisRec *response, OpenIsisDb *db); typedef struct { /* function for request sending. if 0, defaults to internal blocking sender */ int (*sendproc) (OpenIsisStub stub, OpenIsisRec *rqs); /* default response callback */ OpenIsisRspCb *dfltcb; /* delete callback: cleanup client data cbd that is no longer used. if 0==cbd, session is going to be deleted. */ void (*delcb) (void *cld, OpenIsisStub session, void *cbd); /* client data for dfltcb */ void *dfltcld; /* client data for delcb */ void *delcld; } OpenIsisStubCbData; extern OpenIsisStub openisis_stub0; /** create stub for local schema. */ extern OpenIsisStub openIsisNInit ( int argc, const char **argv, OpenIsisStubCbData *dta ); /** destroy local schema. */ extern void openIsisNDeinit (); /** create a new stub. */ extern OpenIsisStub openIsisNOpen ( const char *name, int argc, const char **argv, OpenIsisStubCbData *dta ); /** destroy a stub. */ extern void openIsisNClose (OpenIsisStub stub); /** get schema from stub. */ extern OpenIsisSchema* openIsisNSchema (OpenIsisStub stub); /** send request. SID,SER,TMS are set in this call. Note: rqs (or the internal copy) will be freed at the next call to openIsisNClean or openIsisNSend. @param stub communication partner @param rqs request parameter @param actcb optional async callback for this request @param actcld optional client data for actcb @param rdup make a internal copy of rqs @return 0 on success */ extern int openIsisNSend ( OpenIsisStub stub, OpenIsisRec *rqs, OpenIsisRspCb *actcb, void *actcld, int rdup ); /** get stub's response record. @param dbid optional db id of embedded result records */ extern OpenIsisRec *openIsisNRecv (OpenIsisStub stub, OpenIsisDb **db); /** clean request, response, session used in last openIsisNSend call. */ extern void openIsisNClean (OpenIsisStub stub); /** util: unwrap embbeded response records. */ extern int openIsisNGetResult (OpenIsisStub stub, int **rowids, OpenIsisRec ***recs, OpenIsisDb **db, int *total); /** Notification of a successfully received response record. Interface for applications that handle asynchronous communications with a remote server. @param rsp address of received response record */ extern void openIsisNNotifyResponse (OpenIsisStub stub, OpenIsisRec **rsp); /** Error notification when server communication breaks down. Interface for applications that handle asynchronous communications with a remote server. @param errcode reason of failure */ extern void openIsisNNotifyError (OpenIsisStub stub, int errcode); /** get communication channel from stub. */ extern struct CliChnl* openIsisNGetChannel (OpenIsisStub stub); /** get db by name. */ extern OpenIsisDb *openIsisNDbByName (OpenIsisStub stub, const char *dbname); /** get db by id. */ extern OpenIsisDb *openIsisNDbById (int dbid); /* ************************************************************ db control */ /** open an isis database. The following options are supported as text-style parameters:
  • db <name>
    basename of database. same as param dbname.
  • dbpath <path>
    path to database. usefull when opening secondary indexes or other files whose names are not based on the db basename.
  • v <level>
    set verbosity level
@param dbname identification name of database. the dbpath arg, if given, will be prepended, and the ISIS file extensions appended to build the actual filename. dbname may be NULL, if argv includes a db arg. @param argv array of names and values. May be NULL. argv contains parameter names, which may be prefixed by a dash '-'. See the syspar and dbpar fdt above for supported parameter names. Unrecognized parameter names are ignored. @param argc length of argv. You may supply -1 to indicate that argv is NULL terminated. @param syspar optional syspar record @param dbpar optional dbpar record, a clone will be saved in the db header cfg rec @param fdt optional fdt of this db, the reference will be saved in the db header fdt, so fdt must *NOT* be freed after call @return db structure or 0 on error */ extern OpenIsisDb* openIsisCDOpen ( const char *dbname, OpenIsisRec *dbpar, OpenIsisRec *syspar, OpenIsisFdt *fdt ); /** open by vector @return some non-negative dbid on success, some negative errorcode on error. Errorcodes include -ENOENT, -ENOMEM, -EACCESS and others returned by file access functions. */ extern int openIsisCDOpenv ( const char *dbname, const char **argv, int argc ); enum { OPENISIS_CHK_ONLY = 0x01, /* check only, do not fix */ OPENISIS_CHK_XRF = 0x02, /* check the xrf */ OPENISIS_CHK_FIX = 0xfe, /* check all and fix */ OPENISIS_CHK_ALL = 0xff /* check all, no fix */ }; /** check and optionally fix the database. */ extern int openIsisCDCheck ( int dbid, int flags ); /** close a database. @param dbid id returned by openIsisOpen @return 0 on success, some negative errorcode else. */ extern int openIsisCDClose ( int dbid ); /* ************************************************************ low level index access DO NOT USE unless you know what you're doing !!!!!!!!! */ typedef struct OpenIsisIdx *OpenIsisIndex; /** get the db's index for loading. the index may not be access until the loading session is finished. @param mode lowest byte give percent free for index */ extern OpenIsisIndex openIsisCXOpen ( int dbid, int mode ); /** load a sorted series of keys and hits into index. @param key may be NULL to append to previous key. @param hit a value of NULL terminates the loading session */ extern int openIsisCXAdd ( OpenIsisIndex idx, OpenIsisKey *key, OpenIsisHit *hit ); #define openIsisCXClose( idx ) openIsisCXAdd( idx, 0, 0 ) #ifdef _REENTRANT extern int openisis_threaded; #endif /* ************************************************************ single session and legacy mode pre 0.9 */ /** raw version of an ISIS field. * off gives an offset from the start of the record taken as char*. * values are *NOT* terminated by a 0 byte. */ typedef struct { /* although isis-1 dbs have only 2 bytes for field tags (numbers), we use ints with the lower 3 bytes for the tag. The highest byte may have special usage in some context. */ int tag; int off; int len; } OpenIsisFld; /** raw version of an ISIS Record. rarely used. */ typedef struct { int dbid; int rowid; int bytes; /* total avail bytes */ int used; /* total used bytes (originally BWB) */ int fields; /* avail number of fields (originally BWP) */ int base; /* byte offset of contents area (after fields) */ int len; /* used number of fields */ int state; OpenIsisFld field[1]; } OpenIsisRaw; /** read raw. */ extern OpenIsisRaw *openIsisDRaw ( int db, int rowid ); #ifndef OPENISIS_NOPRE09 /* utilities */ #define openIsis2Html( str, len ) \ openIsisToHtml ( str, len ) #define openIsisValidUTF8( mem, len, tof ) \ openIsisUtf8Chk ( mem, len, tof ) /* record tools */ #define openIsisClone( r, ro, d ) \ openIsisRDup( r, ro, d ) /* GNUC only, and waaarns ... #define openIsisPrintf( r, d, t, f, a... ) \ openIsisRMsg( r, d, t, f, ## a ) */ #define OPENISIS_CHKFREE( recp, room, disc ) \ OPENISIS_RSPACE( recp, room, disc ) #define OPENISIS_ADD( recp, ntag, s, n, disc ) \ OPENISIS_RADD( recp, ntag, s, n, disc ) #define OPENISIS_CAT( recp, s, n, disc ) \ OPENISIS_RCAT( recp, s, n, disc ) #define OPENISIS_SADD( recp, ntag, s, disc ) \ OPENISIS_RADDS( recp, ntag, s, disc ) #define OPENISIS_SCAT( recp, s, disc ) \ OPENISIS_RCATS( recp, s, disc ) #define openIsisReadField( rec, f ) \ openIsisRSplitf( rec, f ) #define openIsisFmt( buf, fmt, r ) \ openIsisRFmt( buf, fmt, r ) #define openIsisReadStream( stream ) \ openIsisSGetr( stream ) /* database */ #define openIsisMaxRowid( db ) \ openIsisDMaxId( db ) #define openIsisReadRow( db, rowid ) \ openIsisDRead( db, rowid ) #define openIsisReadRaw( db, rowid ) \ openIsisDRaw( db, rowid ) #define openIsisRead( buf, fmt, db, rowid ) \ openIsisDFmt( buf, fmt, db, rowid ) #define openIsisWrite( db, r ) \ openIsisDWrite( db, r ) #define openIsisWritex( db, r, i ) \ openIsisDWritex( db, r, i ) #define openIsisScan( db, rowid, tag, key ) \ openIsisDScan( db, rowid, tag, key ) #define openIsisTerm( rec, db, key ) \ openIsisDTerm( rec, db, key ) #define openIsisQuery( set, db, key, mode, skip ) \ openIsisDQuery( set, db, key, mode, skip ) /* database index */ #define OpenIsisIdxCb OpenIsisDXCb #define OpenIsisIdxLoop OpenIsisDXLoop #define openIsisIdxLoop( dbis, cb ) \ openIsisDXLoop( dbis, cb ) /* system control (main session only) */ #define openIsisLog openIsisCLog #define OpenIsisLockFunc OpenIsisCLockFunc #define openIsisInit( argc, argv, lock ) \ openIsisCInit( argc, argv, lock ) #define openIsisSesGet( id, args ) \ (id) ? openIsisCSession( args ) : openIsisCOpen( args ) /* database control (main session only) */ #define openIsisOpen openIsisCDOpenv #define openIsisCheck openIsisCDCheck #define openIsisClose openIsisCDClose /* database index control (main session only) */ #define openIsisIdxOpen openIsisCXOpen #define openIsisIdxAdd openIsisCXAdd #define openIsisIdxDone( idx ) openIsisCXClose( idx ) #endif /* OPENISIS_NOPRE09 */ #ifdef __cplusplus } #endif /* extern "C" */ #define OPENISIS_H #endif /* OPENISIS_H */ openjsis.c0100644000113000011300000002014307644412745011546 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: openjsis.c,v 1.17 2003/04/08 00:20:53 kripke Exp $ main file of openjsis java native lib. */ #include /* free */ #include /* memcpy */ #include "openisis.h" #define index stupidgcc /* java/jdk1.3/include/jni.h:607: warning: declaration of `index' shadows global declaration jobject (JNICALL *GetObjectArrayElement) (JNIEnv *env, jobjectArray array, jsize index); */ #include "org/openisis/NativeDb.h" #undef index /* ************************************************************ private types */ /* ************************************************************ private data */ /* ************************************************************ private functions */ static jobjectArray rec2bytes (JNIEnv* env, OpenIsisRec *r ) { jobjectArray jarr = 0; jintArray ia; jint ji; jbyteArray ba; int i; if ( ! r ) return 0; jarr = (*env)->NewObjectArray( env, 1+r->len, (*env)->FindClass( env, "java/lang/Object" ), 0 ); if ( ! jarr ) return 0; ia = (*env)->NewIntArray( env, 1+r->len ); (*env)->SetObjectArrayElement( env, jarr, 0, ia ); ji = r->rowid; (*env)->SetIntArrayRegion( env, ia, 0, 1, &ji ); for ( i=0; ilen; i++ ) { OpenIsisField *f = &r->field[i]; ba = (*env)->NewByteArray( env, f->len ); if ( ! ba ) break; ji = f->tag; (*env)->SetIntArrayRegion( env, ia, i+1, 1, &ji ); (*env)->SetByteArrayRegion( env, ba, 0, f->len, (jbyte*)f->val ); (*env)->SetObjectArrayElement( env, jarr, i+1, ba ); } return jarr; } /* rec2bytes */ static OpenIsisRec* bytes2rec ( JNIEnv* env, jobjectArray arr ) { jsize len; jsize alen; jintArray ia; jbyteArray ba; OpenIsisRec* rec; int i; jint ji; jint* tags; char *txt; /* initialize record */ rec = 0; OPENISIS_RSPACE( rec, 8*1024, -1 ); len = (*env)->GetArrayLength( env, arr ); if ( 1 > len ) return 0; ia = (*env)->GetObjectArrayElement( env, arr, 0 ); if ( ! ia ) return 0; (*env)->GetIntArrayRegion( env, ia, 0, 1, &ji ); rec->rowid = ji; /* strip rowid / type description from length */ --len; tags = malloc( len * sizeof( jint ) ); (*env)->GetIntArrayRegion( env, ia, 1, len, tags ); for ( i = 0; i < len; i++ ) { ba = (*env)->GetObjectArrayElement( env, arr, i + 1 ); alen = (*env)->GetArrayLength( env, ba ); txt = malloc( alen ); memset( txt, 0, alen ); (*env)->GetByteArrayRegion( env, ba, 0, alen, (jbyte*)txt ); OPENISIS_RADD( rec, tags[ i ], txt, alen, 0 ); if ( txt ) free( txt ); } if ( tags ) free( tags ); return rec; } /* bytes2rec */ /* ************************************************************ package functions */ /* ************************************************************ public functions */ JNIEXPORT jshort JNICALL Java_org_openisis_NativeDb_nopen (JNIEnv* env, jclass cls, jstring jdbname, jobjectArray jargs) { int db; const char *dbname = 0; const char **argv = 0; int argc = 0; jsize i; (void)cls; dbname = ! jdbname ? 0 : (*env)->GetStringUTFChars( env, jdbname, 0 ); if ( jargs ) { argc = (*env)->GetArrayLength( env, jargs ); argv = malloc( argc * sizeof(*argv) ); if ( argv ) for ( i=0; iGetStringUTFChars( env, (*env)->GetObjectArrayElement( env, jargs, i ), 0 ); } db = openIsisOpen( dbname, argv, argc ); if ( dbname ) (*env)->ReleaseStringUTFChars( env, jdbname, dbname ); if ( argv ) for ( i=0; iReleaseStringUTFChars( env, (*env)->GetObjectArrayElement( env, jargs, i ), argv[i] ); return db; } /* Java_org_openisis_Db_open */ JNIEXPORT jobjectArray JNICALL Java_org_openisis_NativeDb_nreadRow (JNIEnv* env, jclass cls, jshort db, jint rowid, jint tag, jbyteArray txt ) { OpenIsisRec *r; jobjectArray jarr = 0; (void)cls; if ( ! txt ) r = openIsisReadRow( db, rowid ); else { jsize l = (*env)->GetArrayLength( env, txt ); char *search = malloc( l+1 ); (*env)->GetByteArrayRegion( env, txt, 0, l, (jbyte*)search ); search[l] = 0; r = openIsisScan( db, rowid, tag, search ); free( search ); } jarr = rec2bytes( env, r ); if ( r ) free( r ); #ifndef NDEBUG fflush( stderr ); #endif return jarr; } /* Java_org_openisis_Db_readRow */ /* * please leave me in for debugging ;) * static int print ( OpenIsisRec *r ) { int i; if ( r ) { for ( i=0; ilen; i++ ) { if ( ! r->field[i].val ) { openIsisSMsg( 1, "%d.?=%d\n", r->rowid, r->field[i].len ); continue; } openIsisSMsg( 1, "%d.%d=%.*s\n", r->rowid, r->field[i].tag, (int)r->field[i].len, r->field[i].val ); if ( r->field[i].len && '^' == *r->field[i].val ) { OpenIsisRec *rf = openIsisReadField( 0, r->field+i ); if ( rf ) { int j; for ( j=0; jlen; j++ ) openIsisSMsg( 1, "%d.%d.%c=%.*s\n", r->rowid, r->field[i].tag, (0x60 & (int)rf->field[j].tag ) ? (int)rf->field[j].tag : ' ', (int)rf->field[j].len, rf->field[j].val ); free( rf ); } } } } return ! r ? -1 : r->rowid; } */ JNIEXPORT jshort JNICALL Java_org_openisis_NativeDb_nwriteRow (JNIEnv* env, jclass cls, jshort db, jobjectArray arr ) { OpenIsisRec *r; jint result; /* no warning */ (void) cls; if ( ! arr ) return -1; r = bytes2rec( env, arr ); if ( ! r ) return -1; result = openIsisWrite( db, r ); if ( r ) free( r ); return result; } JNIEXPORT jshort JNICALL Java_org_openisis_NativeDb_nwriteXRow (JNIEnv* env, jclass cls, jshort db, jobjectArray arr, jobjectArray idx ) { OpenIsisRec *r; OpenIsisRec *i; jint result; /* no warning */ (void) cls; if ( ! arr || ! idx ) return -1; r = bytes2rec( env, arr ); i = bytes2rec( env, idx ); if ( ! r || ! i ) return -1; result = openIsisWritex( db, r, i ); if ( r ) free( r ); if ( i ) free( i ); return result; } JNIEXPORT jobjectArray JNICALL Java_org_openisis_NativeDb_nTerms (JNIEnv* env, jclass cls, jshort db, jbyteArray term, jbyteArray start ) { OpenIsisRec *r; jobjectArray jarr = 0; jsize lterm = (*env)->GetArrayLength( env, term ); char *cterm = malloc( lterm+1 ); union { OpenIsisRec r; char buf[4096]; } x; x.r.len = 0; x.r.bytes = sizeof(x); (void)cls; (*env)->GetByteArrayRegion( env, term, 0, lterm, (jbyte*)cterm ); cterm[lterm] = 0; if ( start ) { char *cstart; jsize lstart = (*env)->GetArrayLength( env, start ); if ( 128 < lstart ) lstart = 128; cstart = x.buf + sizeof(x.r); (*env)->GetByteArrayRegion( env, start, 0, lstart, (jbyte*)cstart ); x.r.field[0].val = cstart; x.r.field[0].len = lstart; x.r.len = 1; } r = openIsisTerm( &x.r, db, cterm ); free( cterm ); jarr = rec2bytes( env, r ); #ifndef NDEBUG fflush( stderr ); #endif return jarr; } /* Java_org_openisis_Db_Terms */ JNIEXPORT jintArray JNICALL Java_org_openisis_NativeDb_nsearch (JNIEnv * env, jclass cls, jshort db, jbyteArray key, jint mode) { jsize l; char *search; OpenIsisSet set; jintArray ia; (void)cls; if ( ! key ) return 0; l = (*env)->GetArrayLength( env, key ); search = malloc( l+1 ); (*env)->GetByteArrayRegion( env, key, 0, l, (jbyte*)search ); search[l] = 0; set.len = 0; openIsisQuery( &set, db, search, mode, 0 ); free( search ); if ( 0 >= set.len ) return 0; ia = (*env)->NewIntArray( env, set.len ); (*env)->SetIntArrayRegion( env, ia, 0, set.len, (jint*)set.id ); return ia; } /* Java_org_openisis_Db_search */ openjsis.def0100644000113000011300000000260507631453445012063 0ustar kripkrip; ; openisis - an open implementation of the ISIS database ; Version 0.8.3 ; Copyright (C) 2001 by Erik Grziwotz, erik@openisis.org ; ; This library is free software; you can redistribute it and/or ; modify it under the terms of the GNU Lesser General Public ; License as published by the Free Software Foundation; either ; version 2.1 of the License, or (at your option) any later version. ; ; This library is distributed in the hope that it will be useful, ; but WITHOUT ANY WARRANTY; without even the implied warranty of ; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ; Lesser General Public License for more details. ; ; You should have received a copy of the GNU Lesser General Public ; License along with this library; if not, write to the Free Software ; Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA ; ; This software is dedicated to the memory of Eckart Dietrich. ; ; This software is inspired by (but contains no code of) the iAPI ; Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. ; See iAPI.txt for what it contains. ;EOH ; $Id: openjsis.def,v 1.4 2003/03/05 19:52:28 istr Exp $ ; export definition file for win32 dll ; EXPORTS Java_org_openisis_NativeDb_nopen Java_org_openisis_NativeDb_nreadRow Java_org_openisis_NativeDb_nwriteRow Java_org_openisis_NativeDb_nwriteXRow Java_org_openisis_NativeDb_nsearch Java_org_openisis_NativeDb_nTerms org/openisis/Charset.java0100644000113000011300000001253707644412745014434 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ package org.openisis; import java.io.*; import java.util.Arrays; /** Charset tests and utils for openisis java binding.

$Id: Charset.java,v 1.3 2003/04/08 00:20:53 kripke Exp $ @version $Revision: 1.3 $ @author $Author: kripke $ */ public class Charset { static final String[] encodings = { "ISO-8859-5", // cyrillic "ISO-8859-7", // greek "ISO-8859-9", // turkish }; // bytes 160(=-96) to 255(=-1) static final byte[] isohigh = new byte[96]; static { for ( int b=0; b<96; b++ ) isohigh[b] = (byte)(b-96); } static void utf8seq () throws IOException { Writer w = new OutputStreamWriter( System.out, "UTF-8" ); for ( int i=0; i r.getLen() || null == r.getField(0) || null == r.getField(1) ) { System.err.println( "bad fields on mfn "+n[0]+" for encoding "+e ); continue; } Field f = r.getField(0); if ( 1 != f.tag || !e.equals( f.val ) ) { System.err.println( "bad field 0 "+f.tag+"="+f.val +" on mfn "+n[0]+" for encoding "+e ); continue; } f = r.getField(1); if ( 2 != f.tag ) { System.err.println( "bad field 1 "+f.tag+"="+f.val +" on mfn "+n[0]+" for encoding "+e ); continue; } if ( ! f.val.equals( new String(isohigh,e) ) ) { System.err.println( "UTF-8 mismatch on mfn "+n[0]+" for encoding "+e ); continue; } byte[] b = f.val.getBytes(e); if ( ! Arrays.equals( b, isohigh ) ) // ok for some encodings System.err.println( "note: no round-trip for encoding "+e ); } } public final String enc; public Charset ( String encoding ) { enc = encoding; } public OutputStreamWriter osw ( OutputStream s ) throws IOException { return new OutputStreamWriter( s, enc ); } public void chartab ( OutputStreamWriter w ) throws IOException { if ( null == w ) w = new OutputStreamWriter( System.out ); w.write( "... 0 1 2 3 4 5 6 7 8 9 A B C D E F\n" ); byte[] b = new byte[32]; for ( int l=0; l<8; l++ ) { for ( int j=0; j<16; j++ ) { b[2*j] = (byte)' '; b[2*j+1] = (byte)(l*16 + j - 128); } String s = new String( b, enc ); w.write( (128+l*16) + s + "\n" ); // char[] c = s.toCharArray[]; } w.flush(); } public void codetab ( OutputStreamWriter w ) throws IOException { if ( null == w ) w = new OutputStreamWriter( System.out ); byte[] b = new byte[16]; for ( int l=0; l<8; l++ ) { for ( int j=0; j<16; j++ ) b[j] = (byte)(l*16 + j - 128); String s = new String( b, enc ); char[] c = s.toCharArray(); for ( int j=0; j\n" + "\n" + "

\n"
		);
		byte[] b = new byte[1];
		for ( b[0] = -96; b[0]<-1; b[0]++ ) {
			w.write( (256+(int)b[0]) + " = "
				+ new String( b, enc ) + "\n" );
		}
		w.write( "
" ); w.flush(); } public static void main ( String[] args ) throws IOException { String enc = "Cp850"; char task = 's'; int a = 0; while ( a $Id: Db.java,v 1.6 2003/04/08 00:20:53 kripke Exp $ @version $Revision: 1.6 $ @author $Author: kripke $ */ public abstract class Db { /** the default encoding. * the constant String value is "ISO-8859-1". */ public static final String DEFAULT_ENCODING = "ISO-8859-1"; /* standard empty String returned by several methods */ public static final String EMPTY = ""; /** index scan key equal */ public static final int QRY_KEYEQ = 0; /** index scan key prefix */ public static final int QRY_KEYPF = 1; public static final int PREFIX = QRY_KEYPF; // legacy alias /** index scan key auto (checks for '$') */ public static final int QRY_KEYAT = 2; /** fulltext scan for equal */ public static final int QRY_SCANE = 64; /** fulltext scan for contains */ public static final int QRY_SCANC = 65; /** query expression with simple left-to-right binding */ public static final int QRY_SIMPLE = 128; /** query expression with proper binding using precedence, () */ public static final int QRY_PROPER = 129; /** field v: heading mode */ public static final int MHL = 0x10000; /** field v: data flag (no effect w/o heading) */ public static final int MD = 0x20000; /** field v: data mode (includes heading flag) */ public static final int MDL = 0x30000; /** field v: uppercase mode flag */ public static final int MXU = 0x40000; /** field v: uppercase heading mode */ public static final int MHU = 0x50000; /** field v: uppercase data mode */ public static final int MDU = 0x70000; /** field v: index formatting flag */ public static final int MI = 0x80000; /** field v: 16bit UNICODE HTML. HTML-entities are escaped. */ public static final int HTU = 0x100000; /** field v: 8bit ISO-8859-1 HTML. HTML-entities and all chars greater than 255 are escaped. */ public static final int HTI = 0x300000; /** field v: 7bit ASCII HTML. HTML-entities and all chars greater than 127 are escaped. */ public static final int HTA = 0x700000; /** rec v: repeated prefix plus operator (not on 1st occ) */ public static final int PP = 0x2000000; /** rec v: repeated suffix plus operator (not on last occ) */ public static final int PS = 0x1000000; /** rec v: dummy presence selector Dn */ public static final int DS = 0x4000000; /** rec v: dummy absence selector Nn */ public static final int NS = 0x8000000; /** rec v: occurence range up to last occ */ public static final int LOCC = 0xffff << 16; static final Map _dbs = new HashMap(); /** open an isis database. The following options are supported as text-style parameters:
  • -db <name>
    basename of database. same as param dbname.
  • -dbpath <path>
    path to database. usefull when opening secondary indexes or other files whose names are not based on the db basename.
  • -v <level>
    set verbosity level
  • -encoding <enc>
    select encoding used in db bytes. default is DEFAULT_ENCODING.
  • -puredb
    select pure java implementation. by default the NativeDb is used.
@param dbname basename of database. the dbhome arg, if given, will be prepended, and the ISIS file extensions appended to build the actual filename. Therefore, depending on the value of dbhome and location of files, this may need to include a path. db may be null, if argv includes a dbname arg. @param argv array of names and values. May be null. argv contains parameter names, which may be prefixed by a dash '-'. Depending on the actual parameter, the next string may or must contain a corresponding parameter value. An optional parameter value may be omitted, if at end of argv or the next name is prefixed with a dash. To avoid ambiguity, values never start with a dash. See above for a list of supported parameter names. Unrecognized parameter names are ignored. @return a new Db (unless some RuntimeException is thrown) @throws IoException especially FileNotFoundExceptiond and UnsupportedEncodingException */ public static synchronized Db open ( String dbname, String[] args ) throws IOException { String enc = DEFAULT_ENCODING; boolean usePureDb = false; Db db = (Db) _dbs.get( dbname ); if ( null != db ) return db; if ( null != args ) for ( int i=0; i $Id: Fdt.java,v 1.5 2003/04/08 00:20:53 kripke Exp $ @version $Revision: 1.5 $ @author $Author: kripke $ */ public class Fdt { /** alphanum field type. */ public static final int FTX = 0; /** strictly alpha field type. */ public static final int FTA = 1; /** numeric field type. */ public static final int FTN = 2; /** pattern field type. */ public static final int FTP = 3; /** iso field type. */ public static final int FTI = 10; /** enum field type. */ public static final int FTE = 12; /** boolean field type. */ public static final int FTB = 13; /** table field type. */ public static final int FTT = 14; /** structure field type. */ public static final int FTS = 15; /** subfield field type. */ public static final int FTF = 16; /** enum value type. */ public static final int FTV = 31; /** field description. */ public static class Fd { /** tag */ public final int id; /** type */ public final int type; /** max length or enum value */ public final int len; /** repeated flag */ public final boolean rep; /** name */ public final String name; /** subfields/pattern */ public final String subf; /** description */ public final String desc; /** @exception IllegalArgumentException if name is null or "" */ public Fd (int id, int type, int len, boolean rep, String name, String subf, String desc) { this.id = id; this.type = type; this.len = len; this.rep = rep; this.name = name; this.subf = subf; this.desc = desc; if (null == name || 0 == name.length ()) { throw new IllegalArgumentException ( "illegal field desc: id " + id + ", name " + name); } } /** shorthand for Fd (id, type, len, false, name, null, null) */ public Fd (int id, int type, int len, String name) { this (id, type, len, false, name, null, null); } public String toString () { return name + ';' + id + ';' + type; } } public final Fd[] _fd; /** @exception IllegalArgumentException if fd is null or has length 0 */ public Fdt (Fd[] fd) { _fd = fd; if (null == fd || 0 == fd.length) { throw new IllegalArgumentException (); } } public Fd fdById (int id) { for (int j = _fd.length; 0 <= --j; ) { if (id == _fd[j].id) { return _fd[j]; } } return null; } public Fd fdByName (String name) { if (null == name || 0 == name.length ()) { return null; } char c = name.charAt (0); if ('0' <= c && '9' >= c) { try { int id = Integer.parseInt (name); return fdById (id); } catch (Exception ex) { } } for (int j = _fd.length; 0 <= --j; ) { if (name.equals (_fd[j].name)) { return _fd[j]; } } return null; } public String dump() { StringBuffer d=new StringBuffer(); for(int i=0;i<_fd.length;i++) { d.append(_fd[i].toString()); if (_fd[i].subf!=null) { d.append(','); d.append(_fd[i].subf); } d.append('\n'); } return d.toString(); } } org/openisis/Field.java0100644000113000011300000001715507644412745014067 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ package org.openisis; import java.util.StringTokenizer; // split /** Field implementation of openisis java binding. A Field is an immutable structure, binding a String value to a tag. It provides methods for formatted access to the value.

$Id: Field.java,v 1.5 2003/04/08 00:20:53 kripke Exp $ @version $Revision: 1.5 $ @author $Author: kripke $ */ public class Field { /** what a single field can contribute to the V operator. @param b a StringBuffer to append to @param mode OR-combination of the field formatting flags defined in Db and an optional subfield character. @param plain a String to format @param len a 16bit length constraint | offset << 16 @return false iff a non-existing subfield was requested, else true */ public static boolean v ( StringBuffer b, int mode, String plain, int len ) { if ( 0 != (Db.MXU & mode) ) // won't change any specials plain = plain.toUpperCase(); final char[] c = plain.toCharArray(); final int cl=c.length; char sub = (char)mode; int o=0, e=cl; if ( 0 != sub ) { if ( '*' == sub ) { // first subfield if ( e > 1 && '^' == c[0] ) o = 2; } else { if ( 0 != (Db.MXU & mode) ) // oops - upcased subfield marks ... sub = Character.toUpperCase(sub); for (;;) { if ( cl == o ) return false; // no subfield if ( '^' != c[o++] ) continue; if ( cl == o ) return false; // no subfield if ( sub == c[o++] ) break; } } // got sub starting at o e = o; while ( cl > e && '^' != c[e] ) e++; } // now the relevant data is at offset o and ends at e if ( 0 != len ) { int off = len >>> 16; len |= 0xffff; o += off; if ( 0 != len && e > o+len ) e = o+len; } if ( e <= o ) return true; // existing but empty boolean mh = Db.MHL == (Db.MHL & mode); boolean mi = Db.MI == (Db.MI & mode); boolean ht = Db.HTU == (Db.HTU & mode); boolean ab = false; // inside angel brackets: stop at = if ( ! (mh || ht) ) { b.append( c, o, e-o ); return true; } final char htlim = // chars > htlim are escaped ( Db.HTA == (Db.HTA & mode) ) ? 127 : ( Db.HTI == (Db.HTI & mode) ) ? 255 : Character.MAX_VALUE; int i=o; char x = 0; // now we run at least one of mh or ht mode replacements LOOP: for (;;i++) { if ( i < e && htlim >= (x = c[i]) ) switch ( x ) { case '"': case '&': if ( ht ) break; continue; case '^': if ( mh ) break; continue; case '=': if ( ab ) break; continue; case '<': case '>': break; // special in any case default: continue; } if ( i > o ) // flush what we had up to current position b.append( c, o, i-o ); if ( i == e ) // done break; o = i+1; // new offset after this if ( htlim < x ) { // write decimal value b.append( "&#" ).append( (int)x ).append( ';' ); continue; } // now we have one of the special chars switch ( x ) { case '"': b.append( """ ); continue; // ht case '&': b.append( "&" ); continue; // ht case '^': // mh if ( i+1 < e ) { x = Character.toUpperCase( c[++i] ); if ( 1 < o ) // not on beginning of field b.append( ('A' == x) ? ';' : ('A'' != c[i] ) if ( e == ++i ) break LOOP; // ran to end o = i+1; // don't break, go on checking for ><-pair case '>': ab = false; if ( ! mh ) b.append( ">" ); else if ( i+1 while ( '>' != c[i] && '=' != c[i] ) if ( e == ++i ) break LOOP; // ran to end o = i+1; break; } } if ( Db.MDL == (Db.MDL & mode) ) // data mode switch ( Character.getType( c[e-1] ) ) { case Character.START_PUNCTUATION: case Character.END_PUNCTUATION: case Character.CONNECTOR_PUNCTUATION: case Character.OTHER_PUNCTUATION: b.append( " " ); break; default: b.append( ". " ); } return true; } // v /** format a String in HTU-mode. @see #v(StringBuffer,int,String,int) */ public static String html ( String plain ) { StringBuffer b = new StringBuffer( plain.length() + 3 ); v( b, Db.HTU, plain, 0 ); // always true w/o subfield spec return b.toString(); } // html // // so far the statics // /** the tag of this field. For primary fields, this is the field number. For subfields, this is actually a char. */ public final int tag; public final String val; public Field ( int tag_, String val_ ) { tag = tag_; val = (val_==null)?"":val_; } public boolean equals (Object that) { if (null == that) { return false; } if (getClass () != that.getClass ()) { return false; } String v1 = val; if (null == v1) v1 = ""; String v2 = ((Field)that).val; if (null == v2) v2 = ""; return tag == ((Field)that).tag && v1.equals (v2); } public String toString () { return tag + ";" + val; } /** split the Field into subfields. */ public Field[] split () { if ( null == val || 0 == val.length() ) return null; String tval=val; int p; if ((p=tval.indexOf('^'))<0) return null; if (p>0) tval=tval.substring(p+1,tval.length()); StringTokenizer st = new StringTokenizer( tval, "^" ); Field[] f = new Field[ st.countTokens() ]; for ( int i=0; i $Id: Log.java,v 1.1 2003/04/12 14:48:21 mawag Exp $ @version $Revision: 1.1 $ @author $Author: mawag $ */ public class Log { /** Error log level */ public static final int ERROR = 1; /** Warning log level */ public static final int WARN = ERROR + 1; /** Info log level */ public static final int INFO = WARN + 1; /** Verbose log level */ public static final int VERBOSE = INFO + 1; private static SimpleDateFormat TheFormat = new SimpleDateFormat ("yyMMddHHmmss"); private static PrintStream TheStream = System.err; private static int TheLevel = VERBOSE; /** Set log level. Default level is VERBOSE. */ public static void setLevel (int lvl) { TheLevel = lvl; } private static void doPrint ( String pref, Object that, String msg, Throwable ex ) { StringBuffer buf = new StringBuffer (128); buf.append (pref). append (' '). append (TheFormat.format (new Date ())). append (' '); if (null != that) { buf.append ('['). append (that). append (']'); } if (null != msg) { if (null != that) { buf.append (": "); } buf.append (msg); } if (null != ex) { if (null != that || null != msg) { buf.append (": "); } } synchronized (TheStream) { TheStream.print (buf.toString ()); if (null != ex) { ex.printStackTrace (TheStream); } else { TheStream.println (); } } } /** Log a verbose message. */ public static void verbose (Object module, String msg, Throwable ex) { if (VERBOSE <= TheLevel) { doPrint (" ", module, msg, ex); } } /** Log an info message. */ public static void info (Object module, String msg, Throwable ex) { if (INFO <= TheLevel) { doPrint ("*** ", module, msg, ex); } } /** Log a warning. */ public static void warn (Object module, String msg, Throwable ex) { if (WARN <= TheLevel) { doPrint ("WARN", module, msg, ex); } } /** Log an error. */ public static void error (Object module, String msg, Throwable ex) { if (ERROR <= TheLevel) { doPrint ("ERR ", module, msg, ex); } } /* Tester public static void main (String argv[]) { error (new Object (), "error", new Error ("fatal")); warn (null, "warning", null); info (null, null, null); verbose (new Exception (), null, new Exception ()); } */ } org/openisis/Mail.java0100644000113000011300000005351207644412745013723 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ package org.openisis; import java.io.*; import java.util.*; import javax.activation.*; import javax.mail.*; import javax.mail.internet.*; /** This class contains a collection of javax.mail based E-Mail utilities.

$Id: Mail.java,v 1.4 2003/04/08 00:20:53 kripke Exp $ @version $Revision: 1.4 $ @author $Author: kripke $ */ public abstract class Mail { public static final String EMPTY = ""; public static final String MESSAGEID = "message-id"; public static final String DATE = "date"; public static final String TO = "to"; public static final String CC = "cc"; public static final String BCC = "bcc"; public static final String FROM = "from"; public static final String SENDER = "sender"; public static final String REPLYTO = "reply-to"; public static final String SUBJECT = "subject"; public static final String REFERENCES = "references"; public static final String INREPLYTO = "in-reply-to"; public static final String KEYWORDS = "keywords"; public static final String COMMENTS = "comments"; public static final String ENCRYPTED = "encrypted"; public static final String PRECEDENCE = "precedence"; // non-standard public static final String XPRIORITY = "x-priority"; // non-standard public static final String RETURNPATH = "return-path"; public static final String DLVRDTO = "delivered-to"; // qmail public static final String RECEIVED = "received"; public static final String XMAILER = "x-mailer"; public static final String CTYPE = "content-type"; public static final String CTRANSENC = "content-transfer-encoding"; public static final String TEXT_PLAIN = "text/plain"; public static final String TEXT_ISO = "text/plain; charset=\"iso-8859-1\""; public static final String TEXT_HTML = "text/html"; public static final String MESS_MIME = "message/rfc822"; public static final String MULT_MIXED = "multipart/mixed"; public static final String MULT_ALT = "multipart/alternative"; public static final String MULT_APPLE = "multipart/appledouble"; public static final String MULT_SIGN = "multipart/signed"; public static final String MULT_ANY = "multipart/*"; public static final String APPL_WORD = "application/msword"; public static final String APPL_PDF = "application/pdf"; public static final String APPL_RTF = "application/rtf"; public static final String APPL_BYTES = "application/octet-stream"; public static final String APPL_APPLE = "application/applefile"; public static final String APPL_ANY = "application/*"; public static final String IMAG_GIF = "image/gif"; public static final String IMAG_JPEG = "image/jpeg"; public static final String IMAG_ANY = "image/*"; public static final String SMTP_HOST = "mail.smtp.host"; public static final Session SES; static { if ( null == System.getProperty( SMTP_HOST ) ) System.setProperty( SMTP_HOST, "localhost" ); SES = Session.getDefaultInstance( System.getProperties(), null ); } public static interface Source { Message[] get () throws Exception; void done ( boolean expunge ) throws Exception; } public static interface Sink { void put ( Message m ) throws Exception; void put ( Message[] m ) throws Exception; } // ////////////////////////////////////////////////// // utilities // public static void del ( Message m ) throws Exception { m.setFlag( Flags.Flag.DELETED, true ); } // del public static void del ( Message[] m ) throws Exception { for ( int i=0; i" ) ); } } public Dir ( String base ) throws Exception { _t = base + File.separator + "tmp"; _n = base + File.separator + "new"; _c = base + File.separator + "cur"; _ndir = new File(_n); File t = new File(_t); File c = new File(_c); if ( ! t.isDirectory() && ! t.mkdirs() ) throw new FileNotFoundException( _t ); if ( ! _ndir.isDirectory() && ! _ndir.mkdirs() ) throw new FileNotFoundException( _n ); if ( ! c.isDirectory() && ! c.mkdirs() ) throw new FileNotFoundException( _c ); } public String toString () { return "Dir "+_ndir.getParentFile().getAbsolutePath(); } public Message[] get () throws Exception { File[] l = _ndir.listFiles(); System.err.println( _ndir+": "+l.length ); int i = l.length; if ( 0 != _lim && i > _lim ) i = _lim; got = new Msg[i]; while ( 0 < i-- ) { File f = new File( _t+File.separator+l[i].getName() ); l[i].renameTo( f ); got[i] = new Msg( f ); } return got; } public void done ( boolean expunge ) throws Exception { if ( expunge && null != got ) for ( int i=0; i" ); System.err.println( _b+" += "+mid ); Date d = m.getSentDate(); if ( null == d ) d = m.getReceivedDate(); if ( null == d ) { System.err.println( "warning: no date found in "+mid ); d = new Date(); } String base = path( d ); String path = base; File f = new File( _b + path ); File dir = f.getParentFile(); if ( ! dir.isDirectory() && ! dir.mkdirs() ) throw new FileNotFoundException( dir.getAbsolutePath() ); for ( int i=0; ! f.createNewFile(); i++ ) f = new File( _b + (path = base + Integer.toString( i, 36 )) ); String name = f.getName(); // got unique path path System.out.println( "new message "+mid+" date "+d+" -> "+path ); MimeMessage mm = new MimeMessage( SES ); Enumeration e = m.getAllHeaders(); while ( e.hasMoreElements() ) { Header h = (Header)e.nextElement(); String n = h.getName().toLowerCase(); if ( OK == OKHDR.get(n) ) mm.addHeader( n, h.getValue() ); } mm.addHeader( CTYPE, TEXT_ISO ); mm.addHeader( CTRANSENC, "8bit" ); StringBuffer b = new StringBuffer( 4096 ); Stack s = new Stack(); s.push( m ); int i = 0; int totlines = 0; List binparts = new ArrayList( 32 ); while ( ! s.isEmpty() ) { Part p = (Part)s.pop(); String ct = oneline( p.getContentType() ); // ct = null == ct ? TEXT_PLAIN : ct.toLowerCase(); if ( p.isMimeType( MULT_ANY ) ) { Object content = p.getContent(); if ( ! (content instanceof Multipart) ) { // ignore System.err.println( "multipart type "+ct +" has class "+content.getClass().getName() ); mm.addHeader( COMMENTS, "dropped bad multipart "+ct ); } else { Multipart mu = (Multipart)content; int cnt = mu.getCount(); if ( cnt > 8 ) { System.err.println( "multipart with too many parts: "+cnt ); cnt = 8; } Part[] pu = new Part[cnt]; for ( int j=0; j 0; ) // push parts backwards s.push( pu[j] ); } continue; } // multi i++; // about to create part i String part = path+'-'+i; // if ( ! p.isMimeType( TEXT_PLAIN ) ) { /* it's so awfully unreliable ... many RTFs go as APPL_WORD ... if ( p.isMimeType( TEXT_HTML ) ) ext = ".html"; else if ( p.isMimeType( APPL_WORD ) ) ext = ".doc"; else ext = ""; */ OutputStream o = new FileOutputStream( _b+part ); p.getDataHandler().writeTo( o ); o.close(); // can we convert it ? String[] args = new String[] { "any2txt", "-v", name+'-'+i }; Process fix = Runtime.getRuntime().exec( args, null, dir ); BufferedReader err = new BufferedReader( new InputStreamReader( fix.getErrorStream() ) ); String line; while ( null != (line = err.readLine()) ) System.err.println( "any2txt:\t"+line ); err.close(); fix.waitFor(); int ret = fix.exitValue(); String ext = EMPTY; boolean havetext = 0 != (ret & 1) || 0 == ret; int type = ret & ~1; switch ( type ) { case 0: ext = ".txt"; break; case 2: ext = ".html"; break; case 4: ext = ".pdf"; break; case 6: ext = ".rtf"; break; case 8: ext = ".doc"; break; case 16: // mail -- reread and start over try { s.push( new MimeMessage( SES, new FileInputStream( _b+part ) ) ); // TODO: delete continue; } catch (Exception ee) { System.err.println( "could not reload mail "+mid+"["+i+"] "+ct +" from "+part ); ee.printStackTrace(); } ext = ".eml"; break; case 32: ext = ".gif"; break; case 34: ext = ".jpg"; break; case 36: ext = ".bmp"; break; case 38: ext = ".png"; break; case 40: ext = ".tiff"; break; } System.err.println( mid+"["+i+"] "+ct+" is "+ext+" ("+ret+")" ); if ( EMPTY != ext && 1 != ret ) // rename (new File(_b+part)).renameTo( new File(_b+part+ext) ); if ( 0 != type ) { binparts.add( name+'-'+i+ext ); // mm.addHeader( COMMENTS, "part "+i+" "+ct+" is "+part ); if ( ! havetext ) { if ( ! _zip ) { if ( 0 < totlines ) b.append( "\r\n---\r\n\r\n" ); b.append( "\r\nget part "+i+" at "+_u+part+ext+"\r\n" ); } totlines++; continue; } } // } // now _b+part.txt contains plaintext File plain = new File(_b+part+".txt"); BufferedReader br = new BufferedReader( new InputStreamReader( new FileInputStream( plain ) ) ); int lines = 0; int max = _maxlines - totlines; if ( max < _minlines ) max = _minlines; // String line; if ( 0 < totlines ) b.append( "\r\n---\r\n\r\n" ); while ( null != (line = br.readLine()) ) { b.append( line ).append( "\r\n" ); if ( max == ++lines ) { mm.addHeader( COMMENTS, "plaintext "+i+" continued at "+part ); b.append( "... read more: "+_u+part+".txt (" +(1 + plain.length()/1024)+" KB)\r\n" ); break; } } if ( 0 != type && !_zip ) b.append( "\r\nget original part "+i+" at "+_u+part+ext+"\r\n" ); totlines += lines; br.close(); } int nparts = binparts.size(); if ( _zip && 0 != nparts ) { String[] args = new String[3+nparts]; int a=0; args[a++] = "zip"; args[a++] = "-m"; // move args[a++] = name+".zip"; for ( int p = 0; pInstallation The shared library "openjsis" must be loadable. On a UNIX System, make sure that the file libopenjsis.so resides in a directory listed by the environment variable LD_LIBRARY_PATH or in /etc/ld.so.conf. If you just drop it in /usr/lib or similar dir, you may have to run a command like ldconfig.

On Windows, put openjsis.dll somewhere in the PATH.

$Id: NativeDb.java,v 1.5 2003/04/08 00:20:53 kripke Exp $ @version $Revision: 1.5 $ @author $Author: kripke $ */ public class NativeDb extends Db { private static synchronized native short nopen ( String dbname, String[] args ); /** read a record by rowid. * The native implementation gives character data as byte[], * since we need to employ java encoding mechanisms anyway. * The tag is packed as four bytes little endian. * @param db a db number as returned by open * @param rowid the number (aka mfn) of the row * @param tag in which to look for txt, negative for any * @param txt to grep from record, null for no grep * @return an array of Objects: * the first is an array of ints filled with rowid an tags * the others are byte[] with field character data in encoding * @see #encoding */ private static synchronized native Object[] nreadRow ( short db, int rowid, int tag, byte[] txt ); /** search for terms */ private static synchronized native Object[] nTerms ( short db, byte[] term, byte[] start ); /** get matching rowids */ private static synchronized native int[] nsearch ( short db, byte[] key, int mode ); /** plain write */ private static synchronized native short nwriteRow ( short db, Object[] data ); /** * Indexed write * @param db database handle returned by open * @param data an Object array with an int[] on idx 0 containing * rowid and tags and with byte[] containing appropriate data for * the tags; thus a simple sanity check for a properly structured * data is: data.length == data[ 0 ].length * @param idx same as data but contains a record that holds index * information */ private static synchronized native short nwriteXRow ( short db, Object[] data, Object[] idx ); static { // TODO: try some explicit paths like our jar's location with System.load // in order to not have to export LD_LIBRARY_PATH or so System.loadLibrary( "openjsis" ); } public byte[] cvt ( String txt ) { try { return txt.getBytes(encoding); } catch ( UnsupportedEncodingException e ) { // shouldn't happen -- our fault throw new Error( "failed to check for support of encoding "+encoding ); } } public String cvt ( byte[] b ) { try { // System.err.println( "got "+b.length+" bytes "+b ); for ( int l=b.length; 0 0 ? "0" : // 0... self-contained b[i] < -64 ? "f" : // 10... six follower bits b[i] < -32 ? "1" : // 110... five bits, one follower b[i] < -16 ? "2" : // 1110... four bits, two followers "b" // bad ) + "("+b[i]+")" ); return ""; } return ""; } catch ( UnsupportedEncodingException e ) { // shouldn't happen -- our fault throw new Error( "failed to check for support of encoding "+encoding ); } } /** Number of this db. */ public final short db; public NativeDb ( String enc, String dbname, String[] args ) throws IOException { super( enc ); db = nopen( dbname, args ); if ( 0 > db ) throw new FileNotFoundException(); } Rec mkRow ( Object[] buf ) { if ( null == buf ) return null; int[] ids = (int[])buf[0]; int rowid = ids[0]; Field[] f = new Field[ buf.length-1 ]; for ( int i=1; i= buf.length ) { buf = null; return false; } nxt = 1; return true; } public Object next () { if ( ! hasNext() ) throw new NoSuchElementException(); return cvt( (byte[])buf[nxt++] ); } public void remove () { throw new UnsupportedOperationException(); } } public Iterator terms ( String prefix ) { return new TermIterator( prefix ); } } // NativeDb org/openisis/PlainSerializer.java0100644000113000011300000003173607655441447016145 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ package org.openisis; import java.io.*; import java.net.Socket; import java.util.ArrayList; /** Serialization of fields by plain protocol.

$Id: PlainSerializer.java,v 1.8 2003/05/02 18:41:55 mawag Exp $ @version $Revision: 1.8 $ @author $Author: mawag $ */ public class PlainSerializer { private static class ProtocolViolation extends RuntimeException { ProtocolViolation () { super (); } ProtocolViolation (String msg) { super (msg); } ProtocolViolation (byte arr[]) { super ("got <" + new String (arr) + ">"); } } private static class ByteBuf { byte _b[]; int _cap, _pos; ByteBuf (int cap) { _b = new byte [_cap = cap]; _pos = 0; } void append (byte b[], int off, int len) { if (0 < len) { int np = _pos + len; if (np > _cap) { int nc = _cap; while (np > (nc *= 2)) { ; } byte nb[] = new byte [nc]; System.arraycopy (_b, 0, nb, 0, _pos); _b = nb; _cap = nc; } System.arraycopy (b, off, _b, _pos, len); _pos = np; } } } private static final byte[] NL = new byte [] { (byte)'\n' }; private static final byte[] HT = new byte [] { (byte)'\t' }; private static final byte[] VT = new byte [] { (byte)11 }; private InputStream _in; private OutputStream _out; private String _enc = null; private boolean _dbg = false; private boolean _bin = false; private final void _write (byte[] arr, int off, int len) throws Exception { if (_dbg) { Log.verbose (this, "w <" + new String (arr, off, len) + ">", null); } _out.write (arr, off, len); } private final void _write (byte[] arr) throws Exception { if (_dbg) { Log.verbose (this, "w <" + new String (arr, 0, arr.length) + ">", null); } _out.write (arr); } private final int _read (byte[] arr) throws Exception { int rd = _in.read (arr); if (_dbg) { Log.verbose (this, "r <" + new String (arr, 0, rd) + ">", null); } return rd; } /** write end of record to output stream. @return success */ public boolean flush () { try { _write (NL); return true; } catch (Exception ex) { Log.warn (this, "flush", ex); } return false; } /** serialize one field. @return success */ public boolean write (Field fld) { int j, lst; byte[] bv; if (null == fld) { return false; } try { bv = Integer.toString (fld.tag).getBytes (); _write (bv); _write (HT); if (null != fld.val && 0 < fld.val.length ()) { bv = null != _enc ? fld.val.getBytes (_enc) : fld.val.getBytes (); lst = j = 0; while (bv.length > j) { if (NL[0] != bv[j]) { ++j; continue; } if (_bin) { ++j; _write (bv, lst, j - lst); _write (HT); } else { if (j > lst) { _write (bv, lst, j - lst); } _write (VT); ++j; } lst = j; } if (bv.length > lst) { _write (bv, lst, bv.length - lst); } } _write (NL); return true; } catch (Exception ex) { Log.warn (this, "serializing " + fld, ex); } return false; } /** serialize fields. @return success */ public boolean write (Field[] flds) { if (null == flds) { return false; } for (int j = 0; flds.length > j; ++j) { if (! write (flds[j])) { return false; } } return flush (); } /** deserialize fields. @return array of fields read from input stream, null on error */ public Field[] read () { byte[] rbuf = new byte [1024]; ByteBuf sbuf = new ByteBuf (1024); ArrayList lst = new ArrayList (); Field fld; int rd, pos, tag, vs, val; boolean rdt; // reading tag boolean lnl; // last byte was NL boolean neg; // tag negative boolean bin = _bin; try { tag = 0; rdt = true; neg = false; lnl = false; all: while (true) { rd = _read (rbuf); if (0 >= rd) { throw new ProtocolViolation ("no more bytes available"); } vs = pos = 0; while (pos < rd) { if (rdt) { if (HT[0] == rbuf[pos]) { rdt = false; vs = ++pos; continue; } if (0 == tag && '-' == rbuf[pos]) { neg = true; ++pos; continue; } val = rbuf[pos] - '0'; if (0 > val || 9 < val) { if (0 == pos && NL[0] == rbuf[0] && 0 == lst.size ()) { Log.warn (this, null, new ProtocolViolation ("empty response")); break; } throw new ProtocolViolation (rbuf); } tag = 10 * tag + val; ++pos; continue; } if (lnl) { lnl = false; if (HT[0] == rbuf[pos]) { bin = true; sbuf.append (NL, 0, 1); vs = ++pos; continue; } fld = new Field (neg ? - tag : tag, null == _enc ? new String (sbuf._b, 0, sbuf._pos) : new String (sbuf._b, 0, sbuf._pos, _enc) ); lst.add (fld); if (NL[0] == rbuf[pos]) { if (pos != rd - 1) { throw new ProtocolViolation ("trailing bytes"); } break all; } sbuf._pos = 0; rdt = true; neg = false; tag = 0; continue; } if (VT[0] == rbuf[pos]) { if (! bin) { rbuf[pos] = NL[0]; } } else if (NL[0] == rbuf[pos]) { if (pos > vs) { sbuf.append (rbuf, vs, pos - vs); } lnl = true; } ++pos; } if (! rdt && rd > vs) { sbuf.append (rbuf, vs, rd - vs); } } return (Field[]) lst.toArray (new Field [lst.size ()]); } catch (Exception ex) { Log.warn (this, "read", ex); } return null; } public void setBinaryMode (boolean bin) { _bin = bin; } /** Set character encoding to use in subsequent calls to read and write. @param enc if null, the platforms default encoding will be used */ public void setEncoding (String enc) { if (null != enc) { try { new String (NL, enc); } catch (UnsupportedEncodingException ex) { Log.warn (this, "setEncoding(" + enc + ")", ex); return; } } _enc = enc; } /** Constructor with read and write stream. @param enc character encoding @param bin binary mode */ public PlainSerializer ( InputStream in, OutputStream out, String enc, boolean bin ) { _in = in; _out = out; setEncoding (enc); setBinaryMode (bin); Log.info (this, "ctor(" + enc + "," + bin + ")", null); } /** Constructor using text mode and platforms default encoding. */ public PlainSerializer (InputStream in, OutputStream out) { this (in, out, null, false); } /** Constructor for a socket connection to port@host. @param enc character encoding @param bin binary mode @exception all exceptions a socket might throw */ public PlainSerializer ( String hostname, int port, String enc, boolean bin ) throws Exception { try { Socket sock = new Socket (hostname, port); _in = sock.getInputStream (); _out = sock.getOutputStream (); setEncoding (enc); setBinaryMode (bin); Log.info (this, "ctor(" + hostname + ',' + port + "," + enc + "," + bin + ")", null); } catch (Exception ex) { Log.warn (this, "ctor(" + hostname + ',' + port + ")", ex); throw ex; } } /** Constructor using text mode and platforms default encoding. */ public PlainSerializer (String hostname, int port) throws Exception { this (hostname, port, null, false); } /** Close communication channels. */ public void close () { if (null != _in) { try { _in.close (); } catch (Exception ex) { } _in = null; } if (null != _out) { try { _out.close (); } catch (Exception ex) { } _out = null; } } // ////////////////////////////////////////////////////////////////////// private class PSWriter extends Thread { Field[] _rqs; int _numr; PSWriter (Field[] rqs, int numr) { _rqs = rqs; _numr = numr; } public void run () { for (int j = 0; _numr > j; ++j) { if (! write (_rqs)) { Log.warn (null, "error writing rqs #" + j, null); System.exit (1); } try { synchronized (PSWriter.this) { PSWriter.this.wait (); } } catch (Exception ex) { Log.warn (PSWriter.this, "wait", ex); } } } } /** Tester:

		java org.openisis.PlainSerializer -p
		java org.openisis.PlainSerializer -b -p
		java org.openisis.PlainSerializer -e UTF-8 -p
		java org.openisis.PlainSerializer -e UTF-8 -b -p
		java org.openisis.PlainSerializer -e UTF-16 -p
		java org.openisis.PlainSerializer -e UTF-16 -b -p
		java org.openisis.PlainSerializer -p 55
		java org.openisis.PlainSerializer -b -p 55
		java org.openisis.PlainSerializer -e UTF-16 -p 55
		java org.openisis.PlainSerializer -e UTF-16 -b -p 55
*/ public static void main (String argv[]) { try { String novel = "blah"; String host = "localhost"; String enc = null; int port = 8080; int numr = 1; boolean pipe = false; boolean dbg = false; boolean bin = false; int j = 0; int len = argv.length; for (int jj = 0; argv.length > jj; ++jj) { if (argv[jj].startsWith ("-d")) { dbg = true; ++j; --len; continue; } if (argv[jj].startsWith ("-b")) { bin = true; ++j; --len; continue; } if (argv[jj].startsWith ("-e")) { enc = argv[++jj]; j += 2; len -= 2; continue; } break; } switch (len) { case 3: numr = Integer.parseInt (argv[2 + j]); case 2: if (argv[j].startsWith ("-p")) { pipe = true; numr = Integer.parseInt (argv[1 + j]); break; } port = Integer.parseInt (argv[1 + j]); case 1: if (argv[j].startsWith ("-p")) { pipe = true; } else { host = argv[j]; } case 0: break; default: throw new IllegalArgumentException (); } PlainSerializer ps; if (pipe) { PipedInputStream pi = new PipedInputStream (); PipedOutputStream po = new PipedOutputStream (pi); ps = new PlainSerializer (pi, po, enc, bin); } else { ps = new PlainSerializer (host, port, enc, bin); } ps._dbg = dbg; for (j = 0; 8 > j; ++j) { novel = novel + novel; } String[] vals = new String [] { "localhorst", "nothing", "one", null, "", novel, "\nfive\nfire\t", "\t\n\t\n" + (bin ? new String (VT) : ""), "seven", "\n" }; Field[] rqs = new Field [] { new Field (-2, vals[0]), new Field (0, vals[1]), new Field (1, vals[2]), new Field (2, vals[3]), new Field (3, vals[4]), new Field (4, vals[5]), new Field (5, vals[6]), new Field (6, vals[7]), new Field (7, vals[8]), new Field (8, vals[9]) }; int k, kk; Field[] rsp; PSWriter writer = ps.new PSWriter (rqs, numr); writer.start (); for (j = 0; numr > j; ++j) { rsp = ps.read (); if (null == rsp) { Log.warn (null, "error reading rsp #" + j, null); System.exit (1); } for (k = kk = 0; rsp.length > k; ++k) { if (0 > rsp[k].tag) { if (kk >= rqs.length || ! rsp[k].equals (rqs[kk])) { // skip additional server fields if (dbg) { Log.verbose (null, "skipping server field " + rsp[k] + " ...", null); } continue; } } if (rqs.length <= kk) { for (kk = 0; rsp.length > kk; ++kk) { Log.verbose (new Integer (kk), rsp[kk].toString (), null); } Log.warn (null, "length mismatch rsp #" + j + ": " + k + "/" + kk + "/" + rsp.length + ">=" + rqs.length, null); System.exit (1); } if (! rsp[k].equals (rqs[kk])) { Log.warn (null, "field #" + k + " rqs #" + j + " differs: <" + rsp[k] + "> != <" + rqs[kk] + ">", null); System.exit (1); } ++kk; } if (0 != j && 0 == (j % 100)) { System.out.println (j + " ..."); } synchronized (writer) { writer.notify (); } } ps.close (); System.out.println ("ok."); } catch (Exception ex) { ex.printStackTrace (System.err); System.err.println ("usage: " + PlainSerializer.class.getName () + " [-b(inary)] [-d(ebug)] [-e(ncoding) ]" + " [ [ []]] |" + " -p(ipe) []"); System.exit (1); } } } org/openisis/PureDb.java0100644000113000011300000000431307644412745014215 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ package org.openisis; import java.io.IOException; import java.io.FileNotFoundException; import java.util.Iterator; /** This class is the pure java implementation of class Db. Unfortunately, it's still empty :(

$Id: PureDb.java,v 1.4 2003/04/08 00:20:53 kripke Exp $ @version $Revision: 1.4 $ @author $Author: kripke $ */ public class PureDb extends Db { public PureDb ( String enc, String dbname, String[] args ) throws IOException { super( enc ); throw new UnsupportedOperationException( "sorry, not implemented" ); } public Rec readRow ( int rowid ) { throw new UnsupportedOperationException( "sorry, not implemented" ); } // readrow public short writeRow ( Rec row ) { throw new UnsupportedOperationException( "sorry, not implemented" ); } // writerow public short writeXRow ( Rec row, Rec idx ) { throw new UnsupportedOperationException( "sorry, not implemented" ); } // writerow public Rec scanRow ( int rowid, int tag, String txt ) { throw new UnsupportedOperationException( "sorry, not implemented" ); } // scanrow public int[] search ( String key, int mode ) { throw new UnsupportedOperationException( "sorry, not implemented" ); } // search public Iterator terms ( String prefix ) { throw new UnsupportedOperationException( "sorry, not implemented" ); } } // PureDb org/openisis/Rec.java0100644000113000011300000001564007644412745013552 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ package org.openisis; /** Record implementation of openisis java binding.

$Id: Rec.java,v 1.4 2003/04/08 00:20:53 kripke Exp $ @version $Revision: 1.4 $ @author $Author: kripke $ */ public class Rec { private static final int APPEND_LENGTH=10; private int myRowId; private Field[] myFields; private int myLength; public Rec ( int rowid, Field[] fields ) { myRowId = rowid; myFields = fields; myLength=fields.length; } public Rec (int rowid) { myRowId=rowid; myFields=new Field[0]; myLength=0; } public int getRowId () { return myRowId; } public int getLen () { return myLength; } public Field getField ( int i ) { return myFields[i]; } public Field[] getFields () { int len = getLen(); Field[] clone = new Field[ len ]; System.arraycopy( myFields, 0, clone, 0, len ); return clone; } public Field[] getFields(int offset,int len) { Field[] clone=new Field[len]; System.arraycopy(myFields,offset,clone,0,len); return clone; } public boolean append(int tag,String value) { return append(new Field(tag,value)); } public boolean append(Field field) { if (field==null) { System.err.println("FIELD==null"); field=new Field(0,""); } try { if (myFields.length<=myLength) { if (myFields.length0) { for(int i=0;ipos) myFields[pos]=f; } protected Field get(int pos) { if (myLength>pos) return myFields[pos]; else return null; } /** the field selector. format one or more occurences of a field to a StringBuffer, mimicking the ISIS Vn-operator. @param b the StringBuffer to fill @param cpre conditional prefix: printed first, if anything is printed @param rpre repeated prefix: printed before every occ (but not the first, if mode include PP) @param tag the field number (0 = all) @param mode several mode bits as defined in Db + optional subfield char @param occ a 16bit occ | end << 16. if occ is 0, all occurences are printed. if the end (higher bytes) are 0, only this occ is printed, else all up to incl. end (use Db.LOCC for all). @param len a 16bit length constraint | offset << 16 this is passed to Field.v and applied for each occ. @param rsuf repeated suffix: printed after every occ (but not the last, if mode include PS) @param csuf conditional suffix: printed last, if anything was printed @return the number of formatted occurences @see Field#v(StringBuffer,int,String,int) */ public int v ( StringBuffer b, String cpre, String rpre, int tag, int mode, int occ, int len, String rsuf, String csuf ) { int end = occ >>> 16; occ &= 0xffff; int o = 0; // occurences seen int did = 0; // occurences done (after range and subfield selection) int lrsuf = -1; // last pos of repeated suffix // data mode ending should not be used with suffix ? // if ( null != rsuf || null != csuf ) mode &= ~Db.MD; // loop all fields int flen=getLen(); for ( int i=0; i occ && o > end ) break; System.err.println( "in range"); // seems we are in range int l = b.length(); if ( 0 == did && null != cpre ) b.append( cpre ); if ( null != rpre && (0 myFields[i].val.indexOf( new String(sf) ); if ( 0 != (Db.NS & mode) ) // the N absence dummy ok = !ok; } if ( ! ok ) { // subfield not found -- however, it counts as an field occ ... // backup to where we were b.delete( l, Integer.MAX_VALUE ); continue; } did++; // gotcha if ( null == rsuf ) continue; lrsuf = b.length(); b.append( rsuf ); } if ( 0 <= lrsuf && 0 != (Db.PS & mode) ) // backup over lrsuf b.delete( lrsuf, Integer.MAX_VALUE ); if ( 0 != did && null != csuf ) b.append( csuf ); return did; } // v /** format fields to a new StringBuffer, return it's content as String. @return a new String or the String Db.EMPTY, iff no fields where done. @see #v(StringBuffer,String,String,int,int,int,int,String,String) */ public String v ( String cpre, String rpre, int tag, int mode, int occ, int len, String rsuf, String csuf ) { StringBuffer b = new StringBuffer(); return 0 < v( b, cpre, rpre, tag, mode, occ, len, rsuf, csuf ) ? b.toString() : Db.EMPTY; } // v /** format fields to a new StringBuffer, return it's content as String. the conditional pre and suf are null here. @return a new String or the String Db.EMPTY, iff no fields where done. @see #v(StringBuffer,String,String,int,int,int,int,String,String) */ public String v ( String rpre, int tag, int mode, int occ, int len, String rsuf ) { return v( null, rpre, tag, mode, occ, len, rsuf, null ); } // v } // Rec org/openisis/RecToXml.java0100664000113000011300000000466707644411042014534 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ package org.openisis; /** convert OpenIsisRecord to Xml */ import java.io.*; import java.util.*; public class RecToXml { private Fdt myFdt; public RecToXml(Fdt fdt) { myFdt=fdt; } public String toXML(Rec rec,String add) { return toXML(rec,"",add); } public String toXML(Rec rec,String pre,String add) { StringBuffer xml=new StringBuffer(); Field f; Rec subRec; int subLen; int i=0; String lastTag=null; String h; while(i"); lastTag=null; } } else { if (lastTag!=null) { xml.append(""); lastTag=null; } if (f!=null && myFdt.fdById(f.tag)!=null) { Field sf[]=f.getSubFields(); String value=f.getValue(); String atts=""; if (sf!=null && sf.length>0) { atts=atts; for(int j=0;j"); if (value.indexOf("&")>=0 || value.indexOf("<")>=0) { value=""; } xml.append(value); } //xml.append("\n"); } i++; } if (lastTag!=null) xml.append(pre+""); return xml.toString(); } } org/openisis/Test.java0100644000113000011300000002126607644412745013761 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ package org.openisis; import java.io.IOException; import java.util.Date; import java.util.Iterator; import java.text.DecimalFormat; /** Test of openisis java binding.

$Id: Test.java,v 1.4 2003/04/08 00:20:53 kripke Exp $ @version $Revision: 1.4 $ @author $Author: kripke $ */ class Test extends Thread { static DecimalFormat mfn4 = new DecimalFormat( "0000" ); public static boolean dumpRow ( Db db, Rec r, int m ) { if ( null == r ) { System.err.println( "could not get row " ); return false; } if ( 1 == m ) { // cds1 test special m = Db.MDL | Db.PS; String r13 = "\n ", s = "\n"; System.out.println( "MFN : "+mfn4.format(r.getRowId()) ); System.out.print( r.v("CONFERENCE : ",null,12,m,0,0,r13,s) ); System.out.print( r.v("TITLE : ",null,24,m,0,0,r13,s) ); System.out.print( r.v("AUTHOR(S) : ",null,14,m,0,0,"; "+r13+" ",s) ); System.out.print( r.v("EDITION : ",null,25,m,0,0,r13,s) ); System.out.print( r.v("IMPRINT : ",null,26,m,0,0,r13,s) ); System.out.print( r.v("COLLATION : ",null,30,m,0,0,r13,s) ); System.out.print( r.v("SERIES : ",null,44,m,0,0,r13,s) ); System.out.print( r.v("NOTES : ",null,50,m,0,0,r13,s) ); System.out.print( r.v("CORP. BODY : ",null,71,m,0,0,r13,s) ); System.out.print( r.v("ADDED CONF : ",null,72,m,0,0,r13,s) ); System.out.print( r.v("OTHER TITLE: ",null,74,m,0,0,r13,s) ); System.out.print( r.v("OTHER LANG : ",null,76,m|'*',0,0,r13,s) ); System.out.print( r.v("KEYWORDS : ",null,69,m,0,0,r13,s) ); System.out.print( "\n\n" ); return true; } System.out.println( "got row "+r.getRowId() ); int l = r.getLen(); for ( int i=0; i"+f.v(m)+"" ); if ( 0 == (Db.MHL & m) && '^' == f.val.charAt(0) ) { Field[] s = f.split(); if ( null == s ) continue; for ( int j=0; j" +s[j].v(m) +"" ); } } return true; } // dumpRow public static boolean dumpRow ( Db db, Rec r ) { return dumpRow( db, r, Db.HTU ); } // dumpRow static String[] term, val; static Db db; static int[][] post; int start; Test ( int s ) { start = s; } /** implementation of runnable for crashtest. */ public void run () { long then = (new Date()).getTime(); for ( int i=start+1; i++!=start; ) { // read recs for term i if ( i >= term.length ) i = 0; // System.err.println( "post "+i ); for ( int j=0; j args.length ) break; if ( "-terms".equals(args[a]) ) for ( Iterator i=db.terms(args[a+1]); i.hasNext(); ) System.out.println( i.next() ); else if ( "-row".equals(args[a]) ) dumpRow( db, db.readRow( Integer.parseInt(args[a+1]) ), m ); else if ( "-scan".equals(args[a]) ) for ( int i=1; null != (r = db.scanRow(i,-42,args[a+1])); i++ ) { dumpRow( db, r, m ); i = r.getRowId(); } else if ( "-search".equals(args[a]) || "-query".equals(args[a]) ) { String src = args[a+1]; int mode = 0; int[] set = db.search(src, "-search".equals(args[a]) ? Db.QRY_KEYAT : Db.QRY_SIMPLE ); System.err.println( set.length + " rows" ); for ( int i=0; i $Id: Utils.java,v 1.2 2003/04/08 00:20:53 kripke Exp $ @version $Revision: 1.2 $ @author $Author: kripke $ */ public abstract class Utils { private Utils () {} /** Build field array from map. Map keys must be valid field description names (if fdt is given) or objects whose string representation build tag numbers. @param ign if false, method will return null on an occurence of a map entry that cannot be transformed in a field @return field array or null, if ign is false and map contains illegal entries */ public static Field[] ToFields (Map map, Fdt fdt, boolean ign) { ArrayList lst = new ArrayList (); Iterator it; Fdt.Fd fd; Map.Entry me; String str; Object key, val; int tag; if (null != map) { it = map.entrySet ().iterator (); while (it.hasNext ()) { me = (Map.Entry) it.next (); key = me.getKey (); if (null == key) { if (ign) { continue; } return null; } if (key instanceof Number) { tag = ((Number)key).intValue (); } else { str = key.toString (); try { tag = Integer.parseInt (str); } catch (Exception ex) { fd = null; if (null != fdt) { fd = fdt.fdByName (str); } if (null == fd) { if (ign) { continue; } return null; } tag = fd.id; } } val = me.getValue (); str = null == val ? "" : val.toString (); lst.add (new Field (tag, str)); } } return (Field[]) lst.toArray (new Field [lst.size ()]); } /** Fill tgt with field tag-val-entries.
If fdt is given, only fields having a field decription are put in tgt with string keys. If fdt is null, entry keys are set as integers.
If tgt is null, a new HashMap will be allocated. */ public static Map ToMap (Field[] flds, Fdt fdt, Map tgt) { Fdt.Fd fd; if (null != flds) { if (null == tgt) { tgt = new HashMap (flds.length); } for (int j = flds.length; 0 <= --j; ) { if (null != fdt) { if (null != (fd = fdt.fdById (flds[j].tag))) { tgt.put (fd.name, flds[j].val); } } else { tgt.put (new Integer (flds[j].tag), flds[j].val); } } } return tgt; } } org/openisis/XmlTest.java0100664000113000011300000000444307644411042014427 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ package org.openisis; /** * some tests */ import java.io.*; import org.xml.sax.*; import org.xml.sax.helpers.*; public class XmlTest { public static void main(String argv[]) { try { XMLReader reader=XMLReaderFactory.createXMLReader(); XmlToRec x2r=new XmlToRec(reader); Rec r=x2r.toRec(new FileReader(argv[0])); StringBuffer b=new StringBuffer(); r.v(b,"","",0,0,0,1000,"\n",""); System.out.println("R:"+r.getLen()+",\n"+b.toString()); System.out.println("\nFTD:"+x2r.getFdt().dump()); RecToXml r2x=new RecToXml(x2r.getFdt()); String xml1=r2x.toXML(r,""); x2r=new XmlToRec(reader); Rec r2=x2r.toRec(new StringReader(xml1)); String xml2=r2x.toXML(r2,""); Rec r3=x2r.toRec(new StringReader(xml2)); String xml3=r2x.toXML(r3,""); System.out.println("COMPARE:"+xml3.compareTo(xml2)); if (xml3.compareTo(xml2)!=0) { boolean diff=false; for(int i=0;i0) { Iterator it=myTags.entrySet().iterator(); Map.Entry e; ArrayList fdl=new ArrayList(); List att; String atts; while(it.hasNext()) { e=(Map.Entry)it.next(); att=(List)myAttributes.get(e.getKey()); if (att!=null && att.size()>0) { atts=""; for(int i=0;i0) { //node.attributes=new HashMap(); for(int i=0;i0) { Object tO=mySubRec.get(mySubRec.size()-1); if (tO instanceof Field) { mySubRec.remove(mySubRec.size()-1); Rec r=new Rec(0); r.append((Field)tO); mySubRec.add(r); } }*/ Rec r=new Rec(0); r.append(new Field(getTag(myPath),value.length()==0?null:value.toString())); mySubRec.add(r); super.startElement(uri,localName,qName,attributes); } public void endElement(java.lang.String uri, java.lang.String localName, java.lang.String qName) throws SAXException { if (mySubRec.size()>=2) { Object tO=mySubRec.get(mySubRec.size()-1); Rec r=(Rec)mySubRec.get(mySubRec.size()-2); if (tO instanceof Field) { r.append((Field)tO); } else { //Rec r2=new Rec(r.getRowId()); //r2.wrap((Rec)tO,Integer.MAX_VALUE); r.wrap((Rec)tO,Integer.MAX_VALUE); } Rec r2=new Rec(r.getRowId()); r2.append(r.getFields(0,1)); r2.wrap(new Rec(0,r.getFields(1,r.getLen())),Integer.MAX_VALUE); //r2.wrap(r,Integer.MAX_VALUE); mySubRec.set(mySubRec.size()-2,r2); } else { if (mySubRec.size()==1) { myRec=(Rec)mySubRec.get(mySubRec.size()-1); //myRec=new Rec(r.getRowId()); //myRec.wrap(r,Integer.MAX_VALUE); //myRec=(Rec)mySubRec.get(mySubRec.size()-1); } } myPath.remove(myPath.size()-1); //myNodes.remove(myNodes.size()-1); mySubRec.remove(mySubRec.size()-1); super.endElement(uri,localName,qName); } public void characters(char[] ch, int start, int length) throws SAXException { //OINode node=(OINode)myNodes.get(myPath.size()-1); Object tO=mySubRec.get(mySubRec.size()-1); String value=null; Field f=null; Rec r=null; if (tO instanceof Field) { f=(Field)tO; value=f.val; } else { r=(Rec)tO; f=r.get(0); value=f.val; } if (value==null || value.length()==0) { value=new String(ch,start,length); } else { value=new String(ch,start,length)+value; } if (r==null) { mySubRec.set(mySubRec.size()-1,new Field(f.tag,value)); } else { r.set(0,new Field(f.tag,value)); } } public void endDocument() throws SAXException { //myRec=null; super.endDocument(); } } } perl/Makefile.PL0100644000113000011300000000311407645531473012464 0ustar kripkrip# #/* # openisis - an open implementation of the CDS/ISIS database # Version 0.8.x (patchlevel see file Version) # Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # see README for more information #EOH */ # # $Id: Makefile.PL,v 1.4 2003/04/11 12:21:15 kripke Exp $ # the MakeMaker Makefile for the openisis perl XSUB # use ExtUtils::MakeMaker; # See lib/ExtUtils/MakeMaker.pm for details of how to influence # the contents of the Makefile that is written. WriteMakefile( 'NAME' => 'OpenIsis', 'VERSION_FROM' => 'OpenIsis.pm', # finds $VERSION 'LDFROM' => '$(OBJECT) ../sto/libopenisis.a', # force to use static archive 'DEFINE' => '', # e.g., '-DHAVE_SOMETHING' 'INC' => '-I..', # e.g., '-I/usr/include/other' 'CC' => 'gcc', 'LD' => 'gcc', 'dynamic_lib' => { OTHERLDFLAGS => '-Wl,-init,openIsisInit,-fini,openIsisFini' } ); perl/OpenIsis.pm0100644000113000011300000001352107644412745012603 0ustar kripkrip# #/* # openisis - an open implementation of the CDS/ISIS database # Version 0.8.x (patchlevel see file Version) # Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # see README for more information #EOH */ # # $Id: OpenIsis.pm,v 1.3 2003/04/08 00:20:53 kripke Exp $ # perl module wrapper for openisis perl XSUB # package OpenIsis; use strict; use vars qw($VERSION @ISA @EXPORT @EXPORT_OK); require Exporter; require DynaLoader; # require AutoLoader; @ISA = qw(Exporter DynaLoader); # Items to export into callers namespace by default. Note: do not export # names by default without a very good reason. Use EXPORT_OK instead. # Do not simply export all your public functions/methods/constants. @EXPORT = qw( ); $VERSION = '0.874'; bootstrap OpenIsis $VERSION; # query mode constants *QRY_KEYEQ = \0; *QRY_KEYPF = \1; *QRY_KEYAT = \2; *QRY_SCANE = \64; *QRY_SCANC = \65; *QRY_SIMPLE = \128; *QRY_PROPER = \129; # error constants *ERR_EOF = \-1; *ERR_FAULT = \-2; *ERR_INVAL = \-3; *ERR_BADF = \-4; *ERR_IO = \-5; *ERR_NOMEM = \-6; *ERR_BUSY = \-7; *ERR_TRASH = \-8; *ERR_IDIOT = \-9; # loglevel constants *LOG_OFF = \0; *LOG_FATAL = \1; *LOG_SYSERR = \2; *LOG_IOERR = \3; *LOG_ERROR = \4; *LOG_WARN = \5; *LOG_INFO = \6; *LOG_VERBOSE = \7; *LOG_TRACE = \8; *LOG_DEBUG = \9; *LOG_ALL = \10; # Preloaded methods go here. sub MHL { for ( @_ ) { s/(<[^=>]*)=[^>]+>/\1>/g; # dump substitutions s/>< pairs s/[><]//g; # nuke other >< s/^\^.//; # kill initial subfield spec s/\^a/; /; # ^a -> ; s/\^[b-i]/, /g; # ^[b-i] -> , s/\^./. /g; # others -> . } } # Autoload methods go after =cut, and are processed by the autosplit program. 1; __END__ # Below is the stub of documentation for the OpenIsis module. =head1 NAME OpenIsis - Perl extension for accessing ISIS databases =head1 SYNOPSIS use OpenIsis; $db = OpenIsis::open( 'basename', '-dbpath', '/home/me/mydb' ); $maxmfn = OpenIsis::maxRowid( $db ); for (OpenIsis::query( $db, 'plant * water' )) { my $row = OpenIsis::read( $db, $_ ); # get hashref $row = OpenIsis::read( $db, $_, 'MHL V24,V26,V71' ); # read formatted my $v24 = $row->{'24'}; # field tags map to arrayrefs OpenIsis::MHL( @$v24 ); # apply heading mode to array print $row->{'mfn'}, "TITLE :", join("\n ", @$v24 ), "\n"; my $v30 = $row->{'30'}->[0]; # contents is like ^ap. 233-238 ^billus. my $subfields = OpenIsis::subfields( $v30 ); # get subfield values print "a = ", $subfields->{'a'}, " b = ", $subfields->{'b'}, "\n"; } # print terms starting with A print join('; ',OpenIsis::terms( $db, 'a$' )), "\n"; =head1 DESCRIPTION The OpenIsis perl extension provides access to ISIS databases based on the openisis C library. =over 4 =item OpenIsis::open( name [, args ...] ) Open a CDS/Isis data base. name: basename of a CDS/Isis database. If given, will be prepended, and the Isis file extensions appended to build the actual filename. Depending on the value of dbname and location of files, this may need to include a path. Dbname may be NULL, if argv includes a dbname argument. args: array of names and values; may be empty. Contains parameters names, which may be prefixed by a dash ('-'). Depending on the actual parameter, the next string may or must contain a corresponding parameter value. An optional parameter value may be omitted, if at end of argv or the next name is prefixed with a dash. To avoid ambiguity, values never start with a dash. Unrecognized parameter names are ignored. Possible values: -db: basename of database; same as param name; -dbpath: path to database; usefull when opening secondary indexes or other files whose names are not based on the db basename; -v: set verbosity level. return value: non-negative number (dbid)Success OpenIsis::ERR_NOENT OpenIsis::ERR_BADF examples: $db = OpenIsis::open( 'cds', '-dbpath', '/var/db/cds/' ); $db = OpenIsis::open( '/var/db/cds/cds' ); $db = OpenIsis::open( '/var/db/cds/cds', '-v', -1 ); if ( 0 <= $db ) { print "ok"; } elsif ( OpenIsis::ERR_BADF == $db ) { print STDERR "/var/db/cds/cds does not exist\n"; } elsif ( OpenIsis::ERR_IO == $db ) { print STDERR "/var/db/cds/cds is not accessible\n"; } else { print STDERR "/var/db/cds/cds had strange error $db\n"; } see also:Query; Terms; Maxrowid type:Database =item OpenIsis::maxRowid( $db ) Get highest rowid (a.k.a. MFN) in data base. dbid: Data base identifier returned by open return value: positive number correspondent highest rowid ZERO unknown value negative number Error examples: $db = OpenIsis::open( $base ); $maxmfn = OpenIsis::maxRowid( $db ); see also:Open type:Database =item OpenIsis::query( $db, $qry ) =item OpenIsis::terms( $db, $key ) =item OpenIsis::read( $db, $rowid [, pft] ) =item OpenIsis::2html( $str ) =item OpenIsis::subfields( $str ) =item OpenIsis::log( $level, $filename ) set logging level to a number between 0 (no logging) and 10 (lots of log). A filename for logging output is currently ignored. =item OpenIsis::MHL( ... ) =back =head1 AUTHOR Erik Grziwotz, erik@openisis.org Documentation added by Paulo Cattelan, cattelan@control.com.br =head1 SEE ALSO L. =cut perl/OpenIsis.xs0100644000113000011300000001213207644412745012616 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: OpenIsis.xs,v 1.5 2003/04/08 00:20:53 kripke Exp $ the openisis XSUB */ #include "EXTERN.h" #include "perl.h" #include "XSUB.h" #ifdef assert #undef assert /* perl.h assert doesn't really assert */ #endif #include /* XSUB.h: #if (defined(PERL_CAPI) || defined(PERL_IMPLICIT_SYS)) && !defined(PERL_CORE) # define malloc PerlMem_malloc # define free PerlMem_free but we need to free memory that was acquired by real malloc I don't now wether these should be defined or not, they are on some Serhij's windows box and aren't on my linux !?? However, we will have to move to OpenIsisMFree, anyway ... */ #ifdef free #define PERLFREE PerlMem_free #undef free #else #define PERLFREE free #endif #define OPENISIS_NOPRE09 #include "openisis.h" /* turn negative error codes in nice little numbers */ static int sherr ( int x ) { return -(1<> OPENISIS_ERR_SHIFT); } MODULE = OpenIsis PACKAGE = OpenIsis PROTOTYPES: ENABLE VERSIONCHECK: ENABLE void log( level, file ) int level char* file CODE: openIsisCLog( level, *file ? file : 0 ); int open( name, ... ) char* name CODE: { const char **argv = 0; int argc = 0; if ( 1 < items ) { STRLEN n_a; int i; argv = (const char **)malloc( (items-1) * sizeof(argv[0]) ); for ( i=1; irowid ), 0 ); if ( items > 2 ) { #ifdef SvPV_nolen /* 5.6.x has this. see man perlguts for why it's useful. */ char *fmt = (char *)SvPV_nolen( ST(2) ); #else STRLEN unused; char *fmt = (char *)SvPV( ST(2), unused ); #endif OpenIsisRec *q; OPENISIS_INITBUF(x); q = openIsisRFmt( &x.r, fmt, r ); free( r ); r = q; } for ( i=0; i < r->len; i++ ) { char buf[7]; AV *ar; SV *ref, **entry; SV *val; val = newSVpv( (char*)r->field[i].val, r->field[i].len ); snprintf( buf, sizeof(buf), "%hu", r->field[i].tag ); buf[sizeof(buf)-1] = 0; entry = hv_fetch( hv, buf, strlen(buf), 1 ); assert ( entry ); /* out of memory */ if ( ! *entry || ! SvROK(*entry) ) { ar = newAV(); *entry = newRV_noinc((SV*)ar); } else { ar = (AV*)SvRV(*entry); assert( SVt_PVAV == SvTYPE(ar) ); } av_push( ar, val ); } if ( r != &x.r ) free( r ); } RETVAL = newRV_noinc( (SV*)hv ); } OUTPUT: RETVAL SV* subfields( field ) char *field CODE: { OpenIsisField f; OpenIsisRec *r; HV *hv = newHV(); f.tag = 0; f.val = field; f.len = strlen(field); r = openIsisRSplitf( 0, &f ); if ( r ) { int i; char buf[2]; buf[sizeof(buf)-1] = 0; for ( i=0; i < r->len; i++ ) { buf[0] = (char)r->field[i].tag; hv_store( hv, buf, buf[0] ? 1 : 0, newSVpv( (char*)r->field[i].val, r->field[i].len ), 0 ); } free( r ); } RETVAL = newRV_noinc( (SV*)hv ); } OUTPUT: RETVAL void query( db, key, ... ) int db char *key PPCODE: { int mode = OPENISIS_QRY_SIMPLE; int skip = 0; int got; OpenIsisSet set; set.len = 0; got = (int)openIsisDQuery( &set, db, key, mode, skip ); if ( 0 < got ) { int i; EXTEND(SP,got); for ( i=0; i{'mfn'}; print '', "\n"; for $t (keys(%$rec)) { my $flds = $rec->{$t}; # all fields for tag for $f (@$flds) { # encoding conversion generated by encoding.pl # convert cp850 to Latin1 $f =~ y/\x80-\xff/\xc7\xfc\xe9\xe2\xe4\xe0\xe5\xe7\xea\xeb\xe8\xef\xee\xec\xc4\xc5\xc9\xe6\xc6\xf4\xf6\xf2\xfb\xf9\xff\xd6\xdc\xf8\xa3\xd8\xd7?\xe1\xed\xf3\xfa\xf1\xd1\xaa\xba\xbf\xae\xac\xbd\xbc\xa1\xab\xbb?????\xc1\xc2\xc0\xa9????\xa2\xa5???????\xe3\xc3???????\xa4\xf0\xd0\xca\xcb\xc8?\xcd\xce\xcf????\xa6\xcc?\xd3\xdf\xd4\xd2\xf5\xd5\xb5\xfe\xde\xda\xdb\xd9\xfd\xdd\xaf\xb4\xad\xb1?\xbe\xb6\xa7\xf7\xb8\xb0\xa8\xb7\xb9\xb3\xb2?\xa0/; print "\t{$s}; } } if ( '' eq $ind ) { print "/>\n"; } else { print ">$ind\n"; } } elsif ( $f =~ /^<.*>$/ ) { # probably print ">\n"; for $kw (split(/>[^<]*", $kw, "\n"; } print "\t\n"; } else { OpenIsis::MHL( $f ); print '>', $f, '\n"; } } } print '', "\n"; } # pr # OpenIsis::log(0,''); # turn off logging $db = OpenIsis::open( $ARGV[0] || 'db/cds/cds' ); if ( 'dump' eq $mode ) { for (1..OpenIsis::maxRowid( $db )) { pr OpenIsis::read( $db, $_ ); } } perl/encoding.pl0100775000113000011300000000373707627123452012647 0ustar kripkrip#!/usr/bin/perl # perl/encoding.pl encodings # # read a Tcl encoding file (see man Tcl_GetEncoding) # spit out a perl translator $enc = 'unknown'; $_ = ; if ( /^#/ ) { ($enc) = /: ([^,]+),/; $_ = ; } /^S/ || die "not a single byte encoding: got '$_'"; $_ = ; ($fallback,$symbol,$pages) = split /\s+/; 1 == $pages || die "not a single byte encoding: got $fallback,$symbol,$pages"; $fallback = hex $fallback; $fbchr = 31<$fallback && $fallback<127 ? chr($fallback) : sprintf('\x%02x', $fallback); print STDERR "fallback character is $fallback ($fbchr)\n"; # read page $_ = ; hex && die "base is not 00 but '$_'"; for $n (0..7) { # check for ASCII @v = unpack( 'n*', pack( 'H64', scalar() ) ); 15 == $#v || die "length is ".$#v; # print STDERR join(',',@v), "\n"; for (0..15) { $v[$_] == 16*$n + $_ || die "bad val $v[$_] at row $n pos $_"; } } $#rv = 127; for (0..127) { $rv[$_] = $fbchr; } for $n (8..15) { @v = unpack( 'n*', pack( 'H64', scalar() ) ); 15 == $#v || die "length is ".$#v; print STDERR join(',',@v), "\n"; for (0..15) { if ( 255 < $v[$_] ) { $tr .= '?'; } else { $tr .= sprintf('\x%02x',$v[$_]); $rv[$v[$_]-128] = sprintf('\x%02x',16*$n + $_) if 127 < $v[$_]; } } } print "# convert $enc to Latin1\n"; print 'y/\x80-\xff/',$tr,"/;\n"; $rv = join('',@rv); print "# convert Latin1 to $enc\n"; print 'y/\x80-\xff/',$rv,"/;\n"; # Latin1 roundtrip test $_ = $test = pack( 'C*', (160..255) ); eval 'y/\x80-\xff/'.$rv.'/'; print STDERR "Latin1 160..255 in encoding $enc\n"; for $n (0..5) { print STDERR 160+16*$n,': ',join(',',unpack('C*',substr($_,16*$n,16))), "\n"; } print STDERR 'test: ',$test,"\n"; eval 'y/\x80-\xff/'.$tr.'/'; print STDERR "trip: ",$_,"\n"; print STDERR "Latin1 160..255 after roundtrip through $enc\n"; for $n (0..5) { print STDERR 160+16*$n,': ',join(',',unpack('C*',substr($_,16*$n,16))), "\n"; } perl/encodings0100664000113000011300000010320507627123504012402 0ustar kripkrip# convert cp1250 to Latin1 y/\x80-\xff/?\x81?\x83????\x88???????\x90???????\x98???????\xa0???\xa4?\xa6\xa7\xa8\xa9?\xab\xac\xad\xae?\xb0\xb1??\xb4\xb5\xb6\xb7\xb8??\xbb?????\xc1\xc2?\xc4??\xc7?\xc9?\xcb?\xcd\xce????\xd3\xd4?\xd6\xd7??\xda?\xdc\xdd?\xdf?\xe1\xe2?\xe4??\xe7?\xe9?\xeb?\xed\xee????\xf3\xf4?\xf6\xf7??\xfa?\xfc\xfd??/; # convert Latin1 to cp1250 y/\x80-\xff/?\x81?\x83????\x88???????\x90???????\x98???????\xa0???\xa4?\xa6\xa7\xa8\xa9?\xab\xac\xad\xae?\xb0\xb1??\xb4\xb5\xb6\xb7\xb8??\xbb?????\xc1\xc2?\xc4??\xc7?\xc9?\xcb?\xcd\xce????\xd3\xd4?\xd6\xd7??\xda?\xdc\xdd?\xdf?\xe1\xe2?\xe4??\xe7?\xe9?\xeb?\xed\xee????\xf3\xf4?\xf6\xf7??\xfa?\xfc\xfd??/; # convert cp1251 to Latin1 y/\x80-\xff/????????????????????????\x98???????\xa0???\xa4?\xa6\xa7?\xa9?\xab\xac\xad\xae?\xb0\xb1???\xb5\xb6\xb7???\xbb????????????????????????????????????????????????????????????????????/; # convert Latin1 to cp1251 y/\x80-\xff/????????????????????????\x98???????\xa0???\xa4?\xa6\xa7?\xa9?\xab\xac\xad\xae?\xb0\xb1???\xb5\xb6\xb7???\xbb????????????????????????????????????????????????????????????????????/; # convert cp1252 to Latin1 y/\x80-\xff/?\x81???????????\x8d?\x8f\x90????????????\x9d??\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff/; # convert Latin1 to cp1252 y/\x80-\xff/?\x81???????????\x8d?\x8f\x90????????????\x9d??\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff/; # convert cp1253 to Latin1 y/\x80-\xff/?\x81??????\x88?\x8a?\x8c\x8d\x8e\x8f\x90???????\x98?\x9a?\x9c\x9d\x9e\x9f\xa0??\xa3\xa4\xa5\xa6\xa7\xa8\xa9\x00\xab\xac\xad\xae?\xb0\xb1\xb2\xb3?\xb5\xb6\xb7???\xbb?\xbd????????????????????\x00????????????????????????????????????????????\x00/; # convert Latin1 to cp1253 y/\x80-\xff/?\x81??????\x88?\x8a?\x8c\x8d\x8e\x8f\x90???????\x98?\x9a?\x9c\x9d\x9e\x9f\xa0??\xa3\xa4\xa5\xa6\xa7\xa8\xa9?\xab\xac\xad\xae?\xb0\xb1\xb2\xb3?\xb5\xb6\xb7???\xbb?\xbd??????????????????????????????????????????????????????????????????/; # convert cp1254 to Latin1 y/\x80-\xff/?\x81???????????\x8d\x8e\x8f\x90????????????\x9d\x9e?\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf?\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc??\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef?\xf1\xf2\xf3\xf4\xf5\xf6\xf7\xf8\xf9\xfa\xfb\xfc??\xff/; # convert Latin1 to cp1254 y/\x80-\xff/?\x81???????????\x8d\x8e\x8f\x90????????????\x9d\x9e?\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf?\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc??\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef?\xf1\xf2\xf3\xf4\xf5\xf6\xf7\xf8\xf9\xfa\xfb\xfc??\xff/; # convert cp1255 to Latin1 y/\x80-\xff/?\x81????????\x8a?\x8c\x8d\x8e\x8f\x90?????????\x9a?\x9c\x9d\x9e\x9f\xa0\xa1\xa2\xa3?\xa5\xa6\xa7\xa8\xa9\xd7\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xf7\xbb\xbc\xbd\xbe\xbf??????????\x00??????????????\x00\x00\x00\x00\x00\x00\x00???????????????????????????\x00\x00??\x00/; # convert Latin1 to cp1255 y/\x80-\xff/?\x81????????\x8a?\x8c\x8d\x8e\x8f\x90?????????\x9a?\x9c\x9d\x9e\x9f\xa0\xa1\xa2\xa3?\xa5\xa6\xa7\xa8\xa9?\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9?\xbb\xbc\xbd\xbe\xbf???????????????????????\xaa???????????????????????????????\xba????????/; # convert cp1256 to Latin1 y/\x80-\xff/????????????????????????????????\xa0?\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9?\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9?\xbb\xbc\xbd\xbe????????????????????????\xd7????????\xe0?\xe2????\xe7\xe8\xe9\xea\xeb??\xee\xef????\xf4??\xf7?\xf9?\xfb\xfc???/; # convert Latin1 to cp1256 y/\x80-\xff/????????????????????????????????\xa0?\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9?\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9?\xbb\xbc\xbd\xbe????????????????????????\xd7????????\xe0?\xe2????\xe7\xe8\xe9\xea\xeb??\xee\xef????\xf4??\xf7?\xf9?\xfb\xfc???/; # convert cp1257 to Latin1 y/\x80-\xff/?\x81?\x83????\x88?\x8a?\x8c\xa8?\xb8\x90???????\x98?\x9a?\x9c\xaf?\x9f\xa0\x00\xa2\xa3\xa4\x00\xa6\xa7\xd8\xa9?\xab\xac\xad\xae\xc6\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xf8\xb9?\xbb\xbc\xbd\xbe\xe6????\xc4\xc5???\xc9?????????\xd3?\xd5\xd6\xd7????\xdc??\xdf????\xe4\xe5???\xe9?????????\xf3?\xf5\xf6\xf7????\xfc???/; # convert Latin1 to cp1257 y/\x80-\xff/?\x81?\x83????\x88?\x8a?\x8c???\x90???????\x98?\x9a?\x9c??\x9f\xa0?\xa2\xa3\xa4?\xa6\xa7\x8d\xa9?\xab\xac\xad\xae\x9d\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\x8f\xb9?\xbb\xbc\xbd\xbe?????\xc4\xc5\xaf??\xc9?????????\xd3?\xd5\xd6\xd7\xa8???\xdc??\xdf????\xe4\xe5\xbf??\xe9?????????\xf3?\xf5\xf6\xf7\xb8???\xfc???/; # convert cp1258 to Latin1 y/\x80-\xff/?\x81????????\x8a??\x8d\x8e\x8f\x90?????????\x9a??\x9d\x9e?\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0\xc1\xc2?\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb?\xcd\xce\xcf?\xd1?\xd3\xd4?\xd6\xd7\xd8\xd9\xda\xdb\xdc??\xdf\xe0\xe1\xe2?\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb?\xed\xee\xef?\xf1?\xf3\xf4?\xf6\xf7\xf8\xf9\xfa\xfb\xfc??\xff/; # convert Latin1 to cp1258 y/\x80-\xff/?\x81????????\x8a??\x8d\x8e\x8f\x90?????????\x9a??\x9d\x9e?\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0\xc1\xc2?\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb?\xcd\xce\xcf?\xd1?\xd3\xd4?\xd6\xd7\xd8\xd9\xda\xdb\xdc??\xdf\xe0\xe1\xe2?\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb?\xed\xee\xef?\xf1?\xf3\xf4?\xf6\xf7\xf8\xf9\xfa\xfb\xfc??\xff/; # convert cp437 to Latin1 y/\x80-\xff/\xc7\xfc\xe9\xe2\xe4\xe0\xe5\xe7\xea\xeb\xe8\xef\xee\xec\xc4\xc5\xc9\xe6\xc6\xf4\xf6\xf2\xfb\xf9\xff\xd6\xdc\xa2\xa3\xa5??\xe1\xed\xf3\xfa\xf1\xd1\xaa\xba\xbf?\xac\xbd\xbc\xa1\xab\xbb?????????????????????????????????????????????????\xdf????\xb5??????????\xb1????\xf7?\xb0?\xb7??\xb2?\xa0/; # convert Latin1 to cp437 y/\x80-\xff/????????????????????????????????\xff\xad\x9b\x9c?\x9d????\xa6\xae\xaa???\xf8\xf1\xfd??\xe6?\xfa??\xa7\xaf\xac\xab?\xa8????\x8e\x8f\x92\x80?\x90???????\xa5????\x99?????\x9a??\xe1\x85\xa0\x83?\x84\x86\x91\x87\x8a\x82\x88\x89\x8d\xa1\x8c\x8b?\xa4\x95\xa2\x93?\x94\xf6?\x97\xa3\x96\x81??\x98/; # convert cp737 to Latin1 y/\x80-\xff/?????????????????????????????????????????????????????????????????????????????????????????????????????????????????\xb1????\xf7?\xb0?\xb7??\xb2?\xa0/; # convert Latin1 to cp737 y/\x80-\xff/????????????????????????????????\xff???????????????\xf8\xf1\xfd????\xfa???????????????????????????????????????????????????????????????\xf6????????/; # convert cp775 to Latin1 y/\x80-\xff/?\xfc\xe9?\xe4?\xe5???????\xc4\xc5\xc9\xe6\xc6?\xf6?\xa2??\xd6\xdc\xf8\xa3\xd8\xd7\xa4??\xf3????\xa6\xa9\xae\xac\xbd\xbc?\xab\xbb????????????????????????????????????????????????\xd3\xdf??\xf5\xd5\xb5?????????\xad\xb1?\xbe\xb6\xa7\xf7?\xb0?\xb7\xb9\xb3\xb2?\xa0/; # convert Latin1 to cp775 y/\x80-\xff/????????????????????????????????\xff?\x96\x9c\x9f?\xa7\xf5?\xa8?\xae\xaa\xf0\xa9?\xf8\xf1\xfd\xfc?\xe6\xf4\xfa?\xfb?\xaf\xac\xab\xf3?????\x8e\x8f\x92??\x90?????????\xe0?\xe5\x99\x9e\x9d???\x9a??\xe1????\x84\x86\x91??\x82?????????\xa2?\xe4\x94\xf6\x9b???\x81???/; # convert cp850 to Latin1 y/\x80-\xff/\xc7\xfc\xe9\xe2\xe4\xe0\xe5\xe7\xea\xeb\xe8\xef\xee\xec\xc4\xc5\xc9\xe6\xc6\xf4\xf6\xf2\xfb\xf9\xff\xd6\xdc\xf8\xa3\xd8\xd7?\xe1\xed\xf3\xfa\xf1\xd1\xaa\xba\xbf\xae\xac\xbd\xbc\xa1\xab\xbb?????\xc1\xc2\xc0\xa9????\xa2\xa5???????\xe3\xc3???????\xa4\xf0\xd0\xca\xcb\xc8?\xcd\xce\xcf????\xa6\xcc?\xd3\xdf\xd4\xd2\xf5\xd5\xb5\xfe\xde\xda\xdb\xd9\xfd\xdd\xaf\xb4\xad\xb1?\xbe\xb6\xa7\xf7\xb8\xb0\xa8\xb7\xb9\xb3\xb2?\xa0/; # convert Latin1 to cp850 y/\x80-\xff/????????????????????????????????\xff\xad\xbd\x9c\xcf\xbe\xdd\xf5\xf9\xb8\xa6\xae\xaa\xf0\xa9\xee\xf8\xf1\xfd\xfc\xef\xe6\xf4\xfa\xf7\xfb\xa7\xaf\xac\xab\xf3\xa8\xb7\xb5\xb6\xc7\x8e\x8f\x92\x80\xd4\x90\xd2\xd3\xde\xd6\xd7\xd8\xd1\xa5\xe3\xe0\xe2\xe5\x99\x9e\x9d\xeb\xe9\xea\x9a\xed\xe8\xe1\x85\xa0\x83\xc6\x84\x86\x91\x87\x8a\x82\x88\x89\x8d\xa1\x8c\x8b\xd0\xa4\x95\xa2\x93\xe4\x94\xf6\x9b\x97\xa3\x96\x81\xec\xe7\x98/; # convert cp852 to Latin1 y/\x80-\xff/\xc7\xfc\xe9\xe2\xe4??\xe7?\xeb??\xee?\xc4?\xc9??\xf4\xf6????\xd6\xdc???\xd7?\xe1\xed\xf3\xfa??????\xac???\xab\xbb?????\xc1\xc2????????????????????????\xa4???\xcb??\xcd\xce????????\xd3\xdf\xd4??????\xda??\xfd\xdd?\xb4\xad????\xa7\xf7\xb8\xb0\xa8?????\xa0/; # convert Latin1 to cp852 y/\x80-\xff/????????????????????????????????\xff???\xcf??\xf5\xf9??\xae\xaa\xf0??\xf8???\xef???\xf7??\xaf?????\xb5\xb6?\x8e??\x80?\x90?\xd3?\xd6\xd7????\xe0\xe2?\x99\x9e??\xe9?\x9a\xed?\xe1?\xa0\x83?\x84??\x87?\x82?\x89?\xa1\x8c????\xa2\x93?\x94\xf6??\xa3?\x81\xec??/; # convert cp855 to Latin1 y/\x80-\xff/??????????????????????????????????????????????\xab\xbb???????????????????????????????\xa4????????????????????????????????\xad????????????\xa7?\xa0/; # convert Latin1 to cp855 y/\x80-\xff/????????????????????????????????\xff???\xcf??\xfd???\xae?\xf0?????????????\xaf????????????????????????????????????????????????????????????????????/; # convert cp857 to Latin1 y/\x80-\xff/\xc7\xfc\xe9\xe2\xe4\xe0\xe5\xe7\xea\xeb\xe8\xef\xee?\xc4\xc5\xc9\xe6\xc6\xf4\xf6\xf2\xfb\xf9?\xd6\xdc\xf8\xa3\xd8??\xe1\xed\xf3\xfa\xf1\xd1??\xbf\xae\xac\xbd\xbc\xa1\xab\xbb?????\xc1\xc2\xc0\xa9????\xa2\xa5???????\xe3\xc3???????\xa4\xba\xaa\xca\xcb\xc8\x00\xcd\xce\xcf????\xa6\xcc?\xd3\xdf\xd4\xd2\xf5\xd5\xb5\x00\xd7\xda\xdb\xd9\xec\xff\xaf\xb4\xad\xb1\x00\xbe\xb6\xa7\xf7\xb8\xb0\xa8\xb7\xb9\xb3\xb2?\xa0/; # convert Latin1 to cp857 y/\x80-\xff/????????????????????????????????\xff\xad\xbd\x9c\xcf\xbe\xdd\xf5\xf9\xb8\xd1\xae\xaa\xf0\xa9\xee\xf8\xf1\xfd\xfc\xef\xe6\xf4\xfa\xf7\xfb\xd0\xaf\xac\xab\xf3\xa8\xb7\xb5\xb6\xc7\x8e\x8f\x92\x80\xd4\x90\xd2\xd3\xde\xd6\xd7\xd8?\xa5\xe3\xe0\xe2\xe5\x99\xe8\x9d\xeb\xe9\xea\x9a??\xe1\x85\xa0\x83\xc6\x84\x86\x91\x87\x8a\x82\x88\x89\xec\xa1\x8c\x8b?\xa4\x95\xa2\x93\xe4\x94\xf6\x9b\x97\xa3\x96\x81??\xed/; # convert cp860 to Latin1 y/\x80-\xff/\xc7\xfc\xe9\xe2\xe3\xe0\xc1\xe7\xea\xca\xe8\xcd\xd4\xec\xc3\xc2\xc9\xc0\xc8\xf4\xf5\xf2\xda\xf9\xcc\xd5\xdc\xa2\xa3\xd9?\xd3\xe1\xed\xf3\xfa\xf1\xd1\xaa\xba\xbf\xd2\xac\xbd\xbc\xa1\xab\xbb?????????????????????????????????????????????????\xdf????\xb5??????????\xb1????\xf7?\xb0?\xb7??\xb2?\xa0/; # convert Latin1 to cp860 y/\x80-\xff/????????????????????????????????\xff\xad\x9b\x9c??????\xa6\xae\xaa???\xf8\xf1\xfd??\xe6?\xfa??\xa7\xaf\xac\xab?\xa8\x91\x86\x8f\x8e???\x80\x92\x90\x89?\x98\x8b???\xa5\xa9\x9f\x8c\x99???\x9d\x96?\x9a??\xe1\x85\xa0\x83\x84???\x87\x8a\x82\x88?\x8d\xa1???\xa4\x95\xa2\x93\x94?\xf6?\x97\xa3?\x81???/; # convert cp861 to Latin1 y/\x80-\xff/\xc7\xfc\xe9\xe2\xe4\xe0\xe5\xe7\xea\xeb\xe8\xd0\xf0\xde\xc4\xc5\xc9\xe6\xc6\xf4\xf6\xfe\xfb\xdd\xfd\xd6\xdc\xf8\xa3\xd8??\xe1\xed\xf3\xfa\xc1\xcd\xd3\xda\xbf?\xac\xbd\xbc\xa1\xab\xbb?????????????????????????????????????????????????\xdf????\xb5??????????\xb1????\xf7?\xb0?\xb7??\xb2?\xa0/; # convert Latin1 to cp861 y/\x80-\xff/????????????????????????????????\xff\xad?\x9c???????\xae\xaa???\xf8\xf1\xfd??\xe6?\xfa???\xaf\xac\xab?\xa8?\xa4??\x8e\x8f\x92\x80?\x90???\xa5??\x8b??\xa6??\x99?\x9d?\xa7?\x9a\x97\x8d\xe1\x85\xa0\x83?\x84\x86\x91\x87\x8a\x82\x88\x89?\xa1??\x8c??\xa2\x93?\x94\xf6\x9b?\xa3\x96\x81\x98\x95?/; # convert cp862 to Latin1 y/\x80-\xff/???????????????????????????\xa2\xa3\xa5??\xe1\xed\xf3\xfa\xf1\xd1\xaa\xba\xbf?\xac\xbd\xbc\xa1\xab\xbb?????????????????????????????????????????????????\xdf????\xb5??????????\xb1????\xf7?\xb0?\xb7??\xb2?\xa0/; # convert Latin1 to cp862 y/\x80-\xff/????????????????????????????????\xff\xad\x9b\x9c?\x9d????\xa6\xae\xaa???\xf8\xf1\xfd??\xe6?\xfa??\xa7\xaf\xac\xab?\xa8?????????????????\xa5?????????????\xe1?\xa0???????????\xa1???\xa4?\xa2???\xf6??\xa3?????/; # convert cp863 to Latin1 y/\x80-\xff/\xc7\xfc\xe9\xe2\xc2\xe0\xb6\xe7\xea\xeb\xe8\xef\xee?\xc0\xa7\xc9\xc8\xca\xf4\xcb\xcf\xfb\xf9\xa4\xd4\xdc\xa2\xa3\xd9\xdb?\xa6\xb4\xf3\xfa\xa8\xb8\xb3\xaf\xce?\xac\xbd\xbc\xbe\xab\xbb?????????????????????????????????????????????????\xdf????\xb5??????????\xb1????\xf7?\xb0?\xb7??\xb2?\xa0/; # convert Latin1 to cp863 y/\x80-\xff/????????????????????????????????\xff?\x9b\x9c\x98?\xa0\x8f\xa4??\xae\xaa??\xa7\xf8\xf1\xfd\xa6\xa1\xe6\x86\xfa\xa5??\xaf\xac\xab\xad?\x8e?\x84????\x80\x91\x90\x92\x94??\xa8\x95????\x99????\x9d?\x9e\x9a??\xe1\x85?\x83????\x87\x8a\x82\x88\x89??\x8c\x8b???\xa2\x93??\xf6?\x97\xa3\x96\x81???/; # convert cp865 to Latin1 y/\x80-\xff/\xc7\xfc\xe9\xe2\xe4\xe0\xe5\xe7\xea\xeb\xe8\xef\xee\xec\xc4\xc5\xc9\xe6\xc6\xf4\xf6\xf2\xfb\xf9\xff\xd6\xdc\xf8\xa3\xd8??\xe1\xed\xf3\xfa\xf1\xd1\xaa\xba\xbf?\xac\xbd\xbc\xa1\xab\xa4?????????????????????????????????????????????????\xdf????\xb5??????????\xb1????\xf7?\xb0?\xb7??\xb2?\xa0/; # convert Latin1 to cp865 y/\x80-\xff/????????????????????????????????\xff\xad?\x9c\xaf?????\xa6\xae\xaa???\xf8\xf1\xfd??\xe6?\xfa??\xa7?\xac\xab?\xa8????\x8e\x8f\x92\x80?\x90???????\xa5????\x99?\x9d???\x9a??\xe1\x85\xa0\x83?\x84\x86\x91\x87\x8a\x82\x88\x89\x8d\xa1\x8c\x8b?\xa4\x95\xa2\x93?\x94\xf6\x9b\x97\xa3\x96\x81??\x98/; # convert cp866 to Latin1 y/\x80-\xff/????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????\xb0?\xb7??\xa4?\xa0/; # convert Latin1 to cp866 y/\x80-\xff/????????????????????????????????\xff???\xfd???????????\xf8??????\xfa????????????????????????????????????????????????????????????????????????/; # convert cp869 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85?\x87\xb7\xac\xa6????????\x93\x94??\xa9?\xb2\xb3?\xa3??????????????\xbd??\xab\xbb????????????????????????????????????????????????????????????????\xad\xb1???\xa7??\xb0\xa8?????\xa0/; # convert Latin1 to cp869 y/\x80-\xff/\x80\x81\x82\x83\x84\x85?\x87???????????\x93\x94???????????\xff??\x9c??\x8a\xf5\xf9\x97?\xae\x89\xf0??\xf8\xf1\x99\x9a???\x88???\xaf?\xab??????????????????????????????????????????????????????????????????/; # convert cp874 to Latin1 y/\x80-\xff/?\x81\x82\x83\x84?\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90???????\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??????????????????????????????????????????????????????????\x00\x00\x00\x00?????????????????????????????\x00\x00\x00\x00/; # convert Latin1 to cp874 y/\x80-\xff/?\x81\x82\x83\x84?\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90???????\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0???????????????????????????????????????????????????????????????????????????????????????????????/; # convert iso8859-1 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff/; # convert Latin1 to iso8859-1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff/; # convert iso8859-10 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??????\xa7?????\xad??\xb0??????\xb7?????????\xc1\xc2\xc3\xc4\xc5\xc6??\xc9?\xcb?\xcd\xce\xcf\xd0??\xd3\xd4\xd5\xd6?\xd8?\xda\xdb\xdc\xdd\xde\xdf?\xe1\xe2\xe3\xe4\xe5\xe6??\xe9?\xeb?\xed\xee\xef\xf0??\xf3\xf4\xf5\xf6?\xf8?\xfa\xfb\xfc\xfd\xfe?/; # convert Latin1 to iso8859-10 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??????\xa7?????\xad??\xb0??????\xb7?????????\xc1\xc2\xc3\xc4\xc5\xc6??\xc9?\xcb?\xcd\xce\xcf\xd0??\xd3\xd4\xd5\xd6?\xd8?\xda\xdb\xdc\xdd\xde\xdf?\xe1\xe2\xe3\xe4\xe5\xe6??\xe9?\xeb?\xed\xee\xef\xf0??\xf3\xf4\xf5\xf6?\xf8?\xfa\xfb\xfc\xfd\xfe?/; # convert iso8859-13 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0?\xa2\xa3\xa4?\xa6\xa7\xd8\xa9?\xab\xac\xad\xae\xc6\xb0\xb1\xb2\xb3?\xb5\xb6\xb7\xf8\xb9?\xbb\xbc\xbd\xbe\xe6????\xc4\xc5???\xc9?????????\xd3?\xd5\xd6\xd7????\xdc??\xdf????\xe4\xe5???\xe9?????????\xf3?\xf5\xf6\xf7????\xfc???/; # convert Latin1 to iso8859-13 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0?\xa2\xa3\xa4?\xa6\xa7?\xa9?\xab\xac\xad\xae?\xb0\xb1\xb2\xb3?\xb5\xb6\xb7?\xb9?\xbb\xbc\xbd\xbe?????\xc4\xc5\xaf??\xc9?????????\xd3?\xd5\xd6\xd7\xa8???\xdc??\xdf????\xe4\xe5\xbf??\xe9?????????\xf3?\xf5\xf6\xf7\xb8???\xfc???/; # convert iso8859-14 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??\xa3???\xa7?\xa9???\xad\xae???????\xb6?????????\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf?\xd1\xd2\xd3\xd4\xd5\xd6?\xd8\xd9\xda\xdb\xdc\xdd?\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef?\xf1\xf2\xf3\xf4\xf5\xf6?\xf8\xf9\xfa\xfb\xfc\xfd?\xff/; # convert Latin1 to iso8859-14 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??\xa3???\xa7?\xa9???\xad\xae???????\xb6?????????\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf?\xd1\xd2\xd3\xd4\xd5\xd6?\xd8\xd9\xda\xdb\xdc\xdd?\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef?\xf1\xf2\xf3\xf4\xf5\xf6?\xf8\xf9\xfa\xfb\xfc\xfd?\xff/; # convert iso8859-15 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0\xa1\xa2\xa3?\xa5?\xa7?\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3?\xb5\xb6\xb7?\xb9\xba\xbb???\xbf\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff/; # convert Latin1 to iso8859-15 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0\xa1\xa2\xa3?\xa5?\xa7?\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3?\xb5\xb6\xb7?\xb9\xba\xbb???\xbf\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff/; # convert iso8859-16 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??????\xa7?\xa9?\xab?\xad??\xb0\xb1????\xb6\xb7???\xbb????\xc0\xc1\xc2?\xc4?\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf??\xd2\xd3\xd4?\xd6??\xd9\xda\xdb\xdc??\xdf\xe0\xe1\xe2?\xe4?\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef??\xf2\xf3\xf4?\xf6??\xf9\xfa\xfb\xfc??\xff/; # convert Latin1 to iso8859-16 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??????\xa7?\xa9?\xab?\xad??\xb0\xb1????\xb6\xb7???\xbb????\xc0\xc1\xc2?\xc4?\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf??\xd2\xd3\xd4?\xd6??\xd9\xda\xdb\xdc??\xdf\xe0\xe1\xe2?\xe4?\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef??\xf2\xf3\xf4?\xf6??\xf9\xfa\xfb\xfc??\xff/; # convert iso8859-2 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0???\xa4??\xa7\xa8????\xad??\xb0???\xb4???\xb8????????\xc1\xc2?\xc4??\xc7?\xc9?\xcb?\xcd\xce????\xd3\xd4?\xd6\xd7??\xda?\xdc\xdd?\xdf?\xe1\xe2?\xe4??\xe7?\xe9?\xeb?\xed\xee????\xf3\xf4?\xf6\xf7??\xfa?\xfc\xfd??/; # convert Latin1 to iso8859-2 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0???\xa4??\xa7\xa8????\xad??\xb0???\xb4???\xb8????????\xc1\xc2?\xc4??\xc7?\xc9?\xcb?\xcd\xce????\xd3\xd4?\xd6\xd7??\xda?\xdc\xdd?\xdf?\xe1\xe2?\xe4??\xe7?\xe9?\xeb?\xed\xee????\xf3\xf4?\xf6\xf7??\xfa?\xfc\xfd??/; # convert iso8859-3 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??\xa3\xa4\x00?\xa7\xa8????\xad\x00?\xb0?\xb2\xb3\xb4\xb5?\xb7\xb8????\xbd\x00?\xc0\xc1\xc2\x00\xc4??\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\x00\xd1\xd2\xd3\xd4?\xd6\xd7?\xd9\xda\xdb\xdc??\xdf\xe0\xe1\xe2\x00\xe4??\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef\x00\xf1\xf2\xf3\xf4?\xf6\xf7?\xf9\xfa\xfb\xfc???/; # convert Latin1 to iso8859-3 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??\xa3\xa4??\xa7\xa8????\xad??\xb0?\xb2\xb3\xb4\xb5?\xb7\xb8????\xbd??\xc0\xc1\xc2?\xc4??\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf?\xd1\xd2\xd3\xd4?\xd6\xd7?\xd9\xda\xdb\xdc??\xdf\xe0\xe1\xe2?\xe4??\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef?\xf1\xf2\xf3\xf4?\xf6\xf7?\xf9\xfa\xfb\xfc???/; # convert iso8859-4 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0???\xa4??\xa7\xa8????\xad?\xaf\xb0???\xb4???\xb8????????\xc1\xc2\xc3\xc4\xc5\xc6??\xc9?\xcb?\xcd\xce?????\xd4\xd5\xd6\xd7\xd8?\xda\xdb\xdc??\xdf?\xe1\xe2\xe3\xe4\xe5\xe6??\xe9?\xeb?\xed\xee?????\xf4\xf5\xf6\xf7\xf8?\xfa\xfb\xfc???/; # convert Latin1 to iso8859-4 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0???\xa4??\xa7\xa8????\xad?\xaf\xb0???\xb4???\xb8????????\xc1\xc2\xc3\xc4\xc5\xc6??\xc9?\xcb?\xcd\xce?????\xd4\xd5\xd6\xd7\xd8?\xda\xdb\xdc??\xdf?\xe1\xe2\xe3\xe4\xe5\xe6??\xe9?\xeb?\xed\xee?????\xf4\xf5\xf6\xf7\xf8?\xfa\xfb\xfc???/; # convert iso8859-5 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0????????????\xad???????????????????????????????????????????????????????????????????????????????\xa7??/; # convert Latin1 to iso8859-5 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??????\xfd?????\xad??????????????????????????????????????????????????????????????????????????????????/; # convert iso8859-6 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0\x00\x00\x00\xa4\x00\x00\x00\x00\x00\x00\x00?\xad\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00?\x00\x00\x00?\x00??????????????????????????\x00\x00\x00\x00\x00???????????????????\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00/; # convert Latin1 to iso8859-6 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0???\xa4????????\xad??????????????????????????????????????????????????????????????????????????????????/; # convert iso8859-7 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??\xa3\x00\x00\xa6\xa7\xa8\xa9\x00\xab\xac\xad\x00?\xb0\xb1\xb2\xb3???\xb7???\xbb?\xbd????????????????????\x00????????????????????????????????????????????\x00/; # convert Latin1 to iso8859-7 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0??\xa3??\xa6\xa7\xa8\xa9?\xab\xac\xad??\xb0\xb1\xb2\xb3???\xb7???\xbb?\xbd??????????????????????????????????????????????????????????????????/; # convert iso8859-8 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0\x00\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9\xd7\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xf7\xbb\xbc\xbd\xbe\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00????????????????????????????\x00\x00??\x00/; # convert Latin1 to iso8859-8 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0?\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9?\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9?\xbb\xbc\xbd\xbe????????????????????????\xaa???????????????????????????????\xba????????/; # convert iso8859-9 to Latin1 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf?\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc??\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef?\xf1\xf2\xf3\xf4\xf5\xf6\xf7\xf8\xf9\xfa\xfb\xfc??\xff/; # convert Latin1 to iso8859-9 y/\x80-\xff/\x80\x81\x82\x83\x84\x85\x86\x87\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90\x91\x92\x93\x94\x95\x96\x97\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8\xa9\xaa\xab\xac\xad\xae\xaf\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf?\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc??\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef?\xf1\xf2\xf3\xf4\xf5\xf6\xf7\xf8\xf9\xfa\xfb\xfc??\xff/; # convert koi8-r to Latin1 y/\x80-\xff/??????????????????????????\xa0?\xb0\xb2\xb7\xf7???????????????????????????????\xa9????????????????????????????????????????????????????????????????/; # convert Latin1 to koi8-r y/\x80-\xff/????????????????????????????????\x9a????????\xbf??????\x9c?\x9d????\x9e???????????????????????????????????????????????????????????????\x9f????????/; # convert koi8-u to Latin1 y/\x80-\xff/??????????????????????????\xa0?\xb0\xb2\xb7\xf7???????????????????????????????\xa9????????????????????????????????????????????????????????????????/; # convert Latin1 to koi8-u y/\x80-\xff/????????????????????????????????\x9a????????\xbf??????\x9c?\x9d????\x9e???????????????????????????????????????????????????????????????\x9f????????/; # convert macCentEuro to Latin1 y/\x80-\xff/\xc4??\xc9?\xd6\xdc\xe1??\xe4???\xe9???\xed????\xf3?\xf4\xf6\xf5\xfa??\xfc?\xb0?\xa3\xa7?\xb6\xdf\xae\xa9??\xa8?????????????????????\xac????\xab\xbb?\xa0??\xd5????????\xf7????????????????\xc1??\xcd???\xd3\xd4??\xda?????\xdd\xfd??????/; # convert Latin1 to macCentEuro y/\x80-\xff/????????????????????????????????\xca??\xa3???\xa4\xac\xa9?\xc7\xc2?\xa8?\xa1?????\xa6????\xc8?????\xe7??\x80????\x83???\xea?????\xee\xef\xcd\x85???\xf2?\x86\xf8?\xa7?\x87??\x8a????\x8e???\x92?????\x97\x99\x9b\x9a\xd6??\x9c?\x9f\xf9??/; # convert macCroatian to Latin1 y/\x80-\xff/\xc4\xc5\xc7\xc9\xd1\xd6\xdc\xe1\xe0\xe2\xe4\xe3\xe5\xe7\xe9\xe8\xea\xeb\xed\xec\xee\xef\xf1\xf3\xf2\xf4\xf6\xf5\xfa\xf9\xfb\xfc?\xb0\xa2\xa3\xa7?\xb6\xdf\xae??\xb4\xa8??\xd8?\xb1???\xb5?????\xaa\xba??\xf8\xbf\xa1\xac????\xab??\xa0\xc0\xc3\xd5????????\xf7??\xa9????\xc6\xbb?\xb7???\xc2?\xc1?\xc8\xcd\xce\xcf\xcc\xd3\xd4?\xd2\xda\xdb\xd9???\xaf?\xcb?\xb8\xca\xe6?/; # convert Latin1 to macCroatian y/\x80-\xff/????????????????????????????????\xca\xc1\xa2\xa3???\xa4\xac\xd9\xbb\xc7\xc2?\xa8\xf8\xa1\xb1??\xab\xb5\xa6\xe1\xfc?\xbc\xdf???\xc0\xcb\xe7\xe5\xcc\x80\x81\xde\x82\xe9\x83\xfd\xfa\xed\xea\xeb\xec?\x84\xf1\xee\xef\xcd\x85?\xaf\xf4\xf2\xf3\x86??\xa7\x88\x87\x89\x8b\x8a\x8c\xfe\x8d\x8f\x8e\x90\x91\x93\x92\x94\x95?\x96\x98\x97\x99\x9b\x9a\xd6\xbf\x9d\x9c\x9e\x9f???/; # convert macCyrillic to Latin1 y/\x80-\xff/?????????????????????????????????\xb0?\xa3\xa7?\xb6?\xae\xa9???????\xb1???\xb5????????????\xac????\xab\xbb?\xa0???????????\xf7?????????????????????????????????????????/; # convert Latin1 to macCyrillic y/\x80-\xff/????????????????????????????????\xca??\xa3???\xa4?\xa9?\xc7\xc2?\xa8?\xa1\xb1???\xb5\xa6????\xc8???????????????????????????????????????????????????????????\xd6????????/; # convert macGreek to Latin1 y/\x80-\xff/\xc4\xb9\xb2\xc9\xb3\xd6\xdc?\xe0\xe2\xe4?\xa8\xe7\xe9\xe8\xea\xeb\xa3?\xee\xef?\xbd?\xf4\xf6\xa6\xad\xf9\xfb\xfc???????\xdf\xae\xa9??\xa7?\xb0\xb7?\xb1??\xa5?????????????\xac????\xab\xbb?\xa0???????????\xf7?????????????????????????????????????????/; # convert Latin1 to macGreek y/\x80-\xff/????????????????????????????????\xca??\x92?\xb4\x9b\xac\x8c\xa9?\xc7\xc2\x9c\xa8?\xae\xb1\x82\x84???\xaf?\x81?\xc8?\x97??????\x80????\x83????????????\x85?????\x86??\xa7\x88?\x89?\x8a??\x8d\x8f\x8e\x90\x91??\x94\x95????\x99?\x9a\xd6?\x9d?\x9e\x9f???/; # convert macIceland to Latin1 y/\x80-\xff/\xc4\xc5\xc7\xc9\xd1\xd6\xdc\xe1\xe0\xe2\xe4\xe3\xe5\xe7\xe9\xe8\xea\xeb\xed\xec\xee\xef\xf1\xf3\xf2\xf4\xf6\xf5\xfa\xf9\xfb\xfc\xdd\xb0\xa2\xa3\xa7?\xb6\xdf\xae\xa9?\xb4\xa8?\xc6\xd8?\xb1??\xa5\xb5?????\xaa\xba?\xe6\xf8\xbf\xa1\xac????\xab\xbb?\xa0\xc0\xc3\xd5????????\xf7?\xff???\xd0\xf0\xde\xfe\xfd\xb7???\xc2\xca\xc1\xcb\xc8\xcd\xce\xcf\xcc\xd3\xd4?\xd2\xda\xdb\xd9???\xaf???\xb8???/; # convert Latin1 to macIceland y/\x80-\xff/????????????????????????????????\xca\xc1\xa2\xa3?\xb4?\xa4\xac\xa9\xbb\xc7\xc2?\xa8\xf8\xa1\xb1??\xab\xb5\xa6\xe1\xfc?\xbc\xc8???\xc0\xcb\xe7\xe5\xcc\x80\x81\xae\x82\xe9\x83\xe6\xe8\xed\xea\xeb\xec\xdc\x84\xf1\xee\xef\xcd\x85?\xaf\xf4\xf2\xf3\x86\xa0\xde\xa7\x88\x87\x89\x8b\x8a\x8c\xbe\x8d\x8f\x8e\x90\x91\x93\x92\x94\x95\xdd\x96\x98\x97\x99\x9b\x9a\xd6\xbf\x9d\x9c\x9e\x9f\xe0\xdf\xd8/; # convert macRoman to Latin1 y/\x80-\xff/\xc4\xc5\xc7\xc9\xd1\xd6\xdc\xe1\xe0\xe2\xe4\xe3\xe5\xe7\xe9\xe8\xea\xeb\xed\xec\xee\xef\xf1\xf3\xf2\xf4\xf6\xf5\xfa\xf9\xfb\xfc?\xb0\xa2\xa3\xa7?\xb6\xdf\xae\xa9?\xb4\xa8?\xc6\xd8?\xb1??\xa5\xb5?????\xaa\xba?\xe6\xf8\xbf\xa1\xac????\xab\xbb?\xa0\xc0\xc3\xd5????????\xf7?\xff????????\xb7???\xc2\xca\xc1\xcb\xc8\xcd\xce\xcf\xcc\xd3\xd4?\xd2\xda\xdb\xd9???\xaf???\xb8???/; # convert Latin1 to macRoman y/\x80-\xff/????????????????????????????????\xca\xc1\xa2\xa3?\xb4?\xa4\xac\xa9\xbb\xc7\xc2?\xa8\xf8\xa1\xb1??\xab\xb5\xa6\xe1\xfc?\xbc\xc8???\xc0\xcb\xe7\xe5\xcc\x80\x81\xae\x82\xe9\x83\xe6\xe8\xed\xea\xeb\xec?\x84\xf1\xee\xef\xcd\x85?\xaf\xf4\xf2\xf3\x86??\xa7\x88\x87\x89\x8b\x8a\x8c\xbe\x8d\x8f\x8e\x90\x91\x93\x92\x94\x95?\x96\x98\x97\x99\x9b\x9a\xd6\xbf\x9d\x9c\x9e\x9f??\xd8/; # convert macRomania to Latin1 y/\x80-\xff/\xc4\xc5\xc7\xc9\xd1\xd6\xdc\xe1\xe0\xe2\xe4\xe3\xe5\xe7\xe9\xe8\xea\xeb\xed\xec\xee\xef\xf1\xf3\xf2\xf4\xf6\xf5\xfa\xf9\xfb\xfc?\xb0\xa2\xa3\xa7?\xb6\xdf\xae\xa9?\xb4\xa8????\xb1??\xa5\xb5?????\xaa\xba???\xbf\xa1\xac????\xab\xbb?\xa0\xc0\xc3\xd5????????\xf7?\xff??\xa4?????\xb7???\xc2\xca\xc1\xcb\xc8\xcd\xce\xcf\xcc\xd3\xd4?\xd2\xda\xdb\xd9???\xaf???\xb8???/; # convert Latin1 to macRomania y/\x80-\xff/????????????????????????????????\xca\xc1\xa2\xa3\xdb\xb4?\xa4\xac\xa9\xbb\xc7\xc2?\xa8\xf8\xa1\xb1??\xab\xb5\xa6\xe1\xfc?\xbc\xc8???\xc0\xcb\xe7\xe5\xcc\x80\x81?\x82\xe9\x83\xe6\xe8\xed\xea\xeb\xec?\x84\xf1\xee\xef\xcd\x85??\xf4\xf2\xf3\x86??\xa7\x88\x87\x89\x8b\x8a\x8c?\x8d\x8f\x8e\x90\x91\x93\x92\x94\x95?\x96\x98\x97\x99\x9b\x9a\xd6?\x9d\x9c\x9e\x9f??\xd8/; # convert macThai to Latin1 y/\x80-\xff/\xab\xbb??????????????????????????????\xa0?????????????????????????????????????????????????????????????????????????????????????????\xae\xa9????/; # convert Latin1 to macThai y/\x80-\xff/????????????????????????????????\xa0????????\xfb?\x80??\xfa????????????\x81????????????????????????????????????????????????????????????????????/; # convert macTurkish to Latin1 y/\x80-\xff/\xc4\xc5\xc7\xc9\xd1\xd6\xdc\xe1\xe0\xe2\xe4\xe3\xe5\xe7\xe9\xe8\xea\xeb\xed\xec\xee\xef\xf1\xf3\xf2\xf4\xf6\xf5\xfa\xf9\xfb\xfc?\xb0\xa2\xa3\xa7?\xb6\xdf\xae\xa9?\xb4\xa8?\xc6\xd8?\xb1??\xa5\xb5?????\xaa\xba?\xe6\xf8\xbf\xa1\xac????\xab\xbb?\xa0\xc0\xc3\xd5????????\xf7?\xff????????\xb7???\xc2\xca\xc1\xcb\xc8\xcd\xce\xcf\xcc\xd3\xd4?\xd2\xda\xdb\xd9???\xaf???\xb8???/; # convert Latin1 to macTurkish y/\x80-\xff/????????????????????????????????\xca\xc1\xa2\xa3?\xb4?\xa4\xac\xa9\xbb\xc7\xc2?\xa8\xf8\xa1\xb1??\xab\xb5\xa6\xe1\xfc?\xbc\xc8???\xc0\xcb\xe7\xe5\xcc\x80\x81\xae\x82\xe9\x83\xe6\xe8\xed\xea\xeb\xec?\x84\xf1\xee\xef\xcd\x85?\xaf\xf4\xf2\xf3\x86??\xa7\x88\x87\x89\x8b\x8a\x8c\xbe\x8d\x8f\x8e\x90\x91\x93\x92\x94\x95?\x96\x98\x97\x99\x9b\x9a\xd6\xbf\x9d\x9c\x9e\x9f??\xd8/; # convert macUkraine to Latin1 y/\x80-\xff/?????????????????????????????????\xb0?\xa3\xa7?\xb6?\xae\xa9???????\xb1???\xb5????????????\xac????\xab\xbb?\xa0???????????\xf7????????????????????????????????????????\xa4/; # convert Latin1 to macUkraine y/\x80-\xff/????????????????????????????????\xca??\xa3\xff??\xa4?\xa9?\xc7\xc2?\xa8?\xa1\xb1???\xb5\xa6????\xc8???????????????????????????????????????????????????????????\xd6????????/; perl/test.pl0100664000113000011300000000640007644412444012024 0ustar kripkrip# #/* # openisis - an open implementation of the CDS/ISIS database # Version 0.8.x (patchlevel see file Version) # Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # see README for more information #EOH */ # # Before `make install' is performed this script should be runnable with # `make test'. After `make install' it should work as `perl test.pl' ######################### We start with some black magic to print on failure. # Change 1..1 below to 1..last_test_to_print . # (It may become useful if the test is moved to ./t subdirectory.) BEGIN { $| = 1; print "1..6\n"; } END {print "not ok 1\n" unless $loaded;} use OpenIsis; $loaded = 1; print "ok 1\n"; ######################### End of black magic. # Insert your test code below (better if it prints "ok 13" # (correspondingly "not ok 13") depending on the success of chunk 13 # of the test code): sub pok { print $_[1] ? '' : 'not ', 'ok ', $_[0], ' (', $_[2], ") \n"; } OpenIsis::log(0,''); pok( 2, 1, 'unfortunately the effect of logging cannot be tested here' ); $db = OpenIsis::open( "cds", "-dbpath", "../db/cds" ); pok( 3, $db >= 0, $db ); $maxrow = OpenIsis::maxRowid( $db ); pok( 4, $maxrow > 0, $maxrow ); $row = OpenIsis::read( $db, 28 ); for (keys(%$row)) { next if 'mfn' eq $_; $a = $row->{$_}; OpenIsis::MHL( @$a ); # print STDERR "\t$_=",join( "\n\t", @$a ),"\n"; } pok( 5, "HASH" eq ref($row), $row ); $row = OpenIsis::subfields( '^ap. 233-238 ^billus.' ); # for (keys(%$row)) { print STDERR "\t$_=$row->{$_}\n"; } pok( 6, "HASH" eq ref($row) && 'p. 233-238 ' eq $row->{'a'}, $row->{'a'} ); @res = OpenIsis::query( $db, 'plant * water' ); # print STDERR "mfns for 'plant * water': ", join(',',@res), "\n"; pok( 7, 2 == @res, scalar(@res) ); @res = OpenIsis::query( $db, 'devel$' ); # print STDERR "mfns for 'devel$': ", join(',',@res), "\n"; pok( 8, 18 == @res, scalar(@res) ); # test error code $nodb = OpenIsis::open( "foo", "-dbpath", "/no/such/database" ); pok( 9, $nodb == $OpenIsis::ERR_BADF, $nodb ); # not really tests for (@res) { $row = OpenIsis::read( $db, $_ ); $title = $row->{'24'}; OpenIsis::MHL( @$title ); # print STDERR "mfn ", $row->{'mfn'}, " title: ", $title->[0], "\n"; } @res = OpenIsis::terms( $db, 'a$' ); # print STDERR 'terms for "a$": ', join('; ',@res), "\n"; $str = OpenIsis::2html( 'lt: < gt: > amp: & quot: "' ); # print STDERR "\t$str\n"; @res = OpenIsis::query( $db, 'Dacca' ); for (@res) { $row = OpenIsis::read( $db, $_, 'V24, V26, V26^*[..], V71,' ); for (sort(keys(%$row))) { $vals = $row->{$_}; # for $v (@$vals) { print STDERR "\t$_=$v\n"; } } } php/Isis.php0100644000113000011300000000225607663441444011763 0ustar kripkrip php/Isis/Db.php0100644000113000011300000000351507663453771012315 0ustar kripkrip number hash (where name should be an "identifier") may contain some additional entries using non-identifier chars ... */ var $fdt; var $name; var $srv; function Isis_Db ( $fdt, $name = null, $server = null ) { $this->fdt = $fdt; $this->name = $name; $this->srv = $server; } function req ( $type, $req ) { $req->append( 920, $type ); $req->append( 903, 0 ); $req->append( 902, $this->name ); return $this->srv->request( $req ); } function query ( $expr ) { $ret = $this->req( 11, new Isis_Rec( 922, 128, 925, $expr ) ); return $ret->get(904);; } function read ( $mfn ) { $ret = $this->req( 12, new Isis_Rec( 904, $mfn ) ); return $ret->v( 908, 0 ); } } // class Isis_Db ?> php/Isis/File.php0100644000113000011300000001247507664371302012642 0ustar kripkripname = $name; $this->w = $write; $this->mst = fopen($name.'.txt', $write ? 'a+' : 'r'); if ( ! $this->mst ) return; flock($this->mst, $write ? LOCK_EX : LOCK_SH); $s = fstat($this->mst); $mtime = $s['mtime']; if ( file_exists($name.'.ptr') ) // avoid annoying warning $this->xrf = fopen($name.'.ptr', $write ? 'r+' : 'r'); if (!$this->xrf) $remake = 1; else { $s = fstat($this->xrf); $magic = fread($this->xrf, 8); $remake = $s['mtime'] < $mtime || ISIS_FILE_MAGIC != $magic; if ( !$remake ) $this->len = $s['size'] / 8 - 1; if ( $remake && !$write ) { fclose($this->xrf); $this->xrf = 0; } } if ( !$this->xrf && !($this->xrf = fopen($name.'.ptr','w+')) ) { fclose($this->mst); return; } if ( $remake ) $this->mkxref(); } function mkxref () { fseek($this->mst, 0); ftruncate($this->xrf, 0); fseek($this->xrf, 0); fwrite( $this->xrf, ISIS_FILE_MAGIC, 8 ); $mfn = 0; $lines = 0; for (;;) { $line = fgets($this->mst); $end = !is_string($line) || '' == $line; if ( $end || "\n" == $line{0} ) { $now = ftell($this->mst); $len = $now - $pos - 1; if ( $len && !$end ) $len--; if ($mfn) { $xrf = pack('VV', $pos, $len); // echo "mfn $mfn pos $pos len $len\n"; if ( 8 != ($foo = fwrite($this->xrf, $xrf)) ) echo "FOO! $foo\n"; if ($this->len < $mfn) $this->len = $mfn; } if ($end) break; $mfn++; $pos = $now; $lines = 0; continue; } if ( ! $lines && ! (int)$line && 2 == sscanf($line, "%c %d", &$ctrl, &$num) && 'W' == $ctrl ) { echo "W $num\n"; $pos = ftell($this->mst); $mfn = $num; fseek($this->xrf, 8*$mfn); continue; } $lines++; } fflush( $this->xrf ); } // mkxref /** @return string number $mfn or null */ function read ( $mfn ) { fseek($this->xrf, 8*$mfn); $xrf = fread( $this->xrf, 8 ); if ( 8 != strlen($xrf) ) { // echo "no xrf for $mfn\n"; return null; } $xrf = unpack('Vpos/Vlen', $xrf); // echo "mfn $mfn pos ${xrf['pos']} len ${xrf['len']}\n"; if ( !$xrf['pos'] || !$xrf['len'] ) return null; fseek($this->mst, $xrf['pos']); $rec = fread($this->mst, $xrf['len']); return $xrf['len'] == strlen($rec) ? $rec : null; } /** write $data as $mfn or ++$this->len @return mfn */ function write ( $data, $mfn = 0 ) { if ( !$this->w ) trigger_error( $this->name.' not writable', E_USER_ERROR ); if ( !$mfn ) $mfn = ++$this->len; elseif ($this->len < $mfn) $this->len = $mfn; if ( !is_string($data) ) $len = 0; elseif ( $len = strlen($data) ) { if ("\n" == $data{$len-1}) $len--; else $data .= "\n"; } fseek($this->mst, 0, SEEK_END); // prbly not needed by mode a+ fwrite($this->mst, "\nW\t$mfn\n"); fflush($this->mst); $pos = ftell($this->mst); if ( $len ) fwrite($this->mst, $data); fseek($this->xrf, 8*$mfn); // echo "write $mfn len $len\n"; fwrite($this->xrf, pack('VV', $pos, $len)); fflush($this->xrf); return $mfn; } function compact () { if ( !$this->w ) trigger_error( $this->name.' not writable', E_USER_ERROR ); $c = fopen($this->name.'.new', 'w+'); flock($c, LOCK_EX); // so we have exclusive locks on both for ( $n=1; $n<=$this->len; $n++ ) fwrite($c, is_null($data = $this->read($n)) ? "\n" : "\n".$data."\n" ); fflush($c); if ( file_exists($this->name.'.old') ) unlink($this->name.'.old'); rename($this->name.'.txt', $this->name.'.old'); rename($this->name.'.new', $this->name.'.txt'); fclose($this->mst); $this->mst = $c; $this->mkxref(); // frob the xref in place } } // class Isis_File ?> php/Isis/Http.php0100644000113000011300000001047207663152500012671 0ustar kripkriplen()) for ( $res = '', $t = reset($rec->tag), $v = reset($rec->val); ; $t = next($rec->tag), $v = next($rec->val) ) { if ( $nuts ) // use v42[]: the [] so we can send multiple values, // and the v cause numerical named params as [] // tend to crash PHP; GET always, POST sometimes works ... // (it's such a feeble ugly pile of shirt) $res .= 'v' . $t . '%5B%5D=' . urlencode($v); else $res .= $t . '=' . urlencode($v); if ( !--$i ) break; $res .= '&'; } return $res; } /** static we ARE on the nutty side -- try to turn http request into something usable see toUrl nuts above for the format UNFORTUNATELY we have no way to tell the original order of parameters. @param mixed $db_or_fdt a Isis_Db object or just it's fdt array to look up names. */ function fromReq ( $db_or_fdt = null ) { if ( is_object($db_or_fdt) ) { $req = new Isis_Rec( '-db', $db_or_fdt ); $fdt = $db->fdt; } else { $req = new Isis_Rec; $fdt = $db_or_fdt; // array or null } foreach (array($_GET,$_POST) as $in) foreach ($in as $k => $v) { if ( is_numeric($k) ) $k = (int)$k; elseif ( sscanf( $k, 'v%d', &$i ) ) $k = (int)$i; elseif ( $fdt && is_int($fdt[$k]) ) $k = $fdt[$k]; else continue; if (is_string($v)) $req->add($k, $v); else foreach ($v as $w) // Wolfsburg $req->add($k, $w); } return $req; } function Isis_Http ( $host, $url, $php = 0, $port = 80 ) { $this->host = $host; $this->port = $port; $this->head = 'POST '.$url." HTTP/1.1\r\n" . 'Host: '.$host."\r\n" . "User-Agent: Isis\r\n" // try keep-alive. fnord does not support that with cgi . "Connection: keep-alive\r\n" . "Content-Type: application/x-www-form-urlencoded\r\n" . 'Content-Length: '; $this->nuts = $php; $this->open(); } function request ( $req ) { if ( !$this->sock && !$this->open() ) return null; $url = Isis_Http::toUrl( $req, $this->nuts ); fwrite($this->sock, $this->head . strlen($url) ."\r\n\r\n" . $url ); // as opposed to folklore (http://www.php.net/manual/en/ref.stream.php), // there is NO trailing CRFL (RTFM err ... RFC 2068 4.1 Page 30) fflush($this->sock); $ok = ($line = fgets($this->sock)) && sscanf( $line, "HTTP/1.%d 200 OK", &$foo ); $ka = 0; while ( ($line = fgets($this->sock)) && trim($line) ) if ( !strncasecmp('connection: keep-alive', $line, 22) ) $ka = $ok; if ( !$ok ) $res = null; else { $res = new Isis_Rec; // HERE we rely on our other end being cooperative, // nicely ending response with blank line while ( ($line = fgets($this->sock)) && trim($line) ) $res->parse($line, ISIS_REC_TEXT); } if ( !$ka ) { fclose( $this->sock ); $this->sock = 0; } return $res; } } // class Isis_Http ?> php/Isis/Rec.php0100644000113000011300000004300207663453771012474 0ustar kripkrip]*)=[^>]+>/', # dump substitutions '/>< pairs '/[><]/', # nuke other >< '/^\^./', # kill initial subfield spec '/\^a/', # ^a -> ; '/\^[b-i]/', # ^[b-i] -> , '/\^./' # others -> . ); $GLOBALS['ISIS_REC_MHL_REPL'] = array( '\1>', '; ', '', '', '; ', ', ', '. ' ); /** An ISIS(/IIF/Z39.2/ISO2709)-style record in pure PHP. First of all: Use Tcl, Java, Perl or plain C, but please don't use PHP -- it's a mess. This is only loosely connected to an Isis Database, most functions can be used without having a DB. Note: Most getter methods work regardless of the type of tags. However, the setters are strongly biased towards numeric tags. @version $Revision: 1.8 $ @license LGPL @package Isis */ class Isis_Rec { /** The associated database, if any. Better be an Isis_Db. @var object $db */ var $db = 0; /** The records MFN ("master file number", a.k.a. rowid). A value of 0 means the record did not yet have a number in it's db, a write will be handled as insert. @var int $mfn */ var $mfn = 0; /** Array of tags for the record. Keys and values better be integers (others might be ignored or yield unexpected results in some contexts). The count of the tag array is taken as length of the record. The tag array may have holes (unassigned ints) if unset is used. @var array $tag */ var $tag; /** Array of values for the record. Values better be strings (i.e. will be forced to strings by some operations). @var array $tag */ var $val; /** last tag accessed using function v */ var $v = null; /** number of holes in record */ var $holy = 0; // //////////////////////////////////////////////////////////// // static // /** static function to format a value. maybe used as Isis_Rec::fmt. @param string $val a value to be formatted @param string $fmt a format specification, defaults to 'MHL' - if $fmt is null, the value is returned unchanged - if $fmt starts with '&' or '%', that character is stripped and htmlspecialchars or urlencode, resp., is applied to the (each) value as last step - if $fmt starts with 'MHL' (the constant ISIS_REC_MHL), that is stripped and the classical MHL ISIS formatting applied to values (before & or % mangling, rarely needed with subfields) - if $fmt is (now) empty, the complete value is used - else we're going for subfields: - if $fmt starts with a hat, the hat is stripped an used as subfield delimiter (TAB otherwise) - if $fmt matches /^([^(]*)\((\d*)(\.\.(\d*))?/, the part from the first '(' on is stripped as occurence selector (note that an optional closing ')' and additional chars are ignored) - the (remaining) characters in $fmt are subfield names, '*' selects any subfield (including the initial unnamed, even if it's empty !), '' any without stripping subfield names - if there is a occurence or range selected, for every character in the $fmt the specified occurences are used (counted from 0). If either bound is empty, 0 is used. If a range is specified (.. given), an upper bound of 0 means up to end. By default, only the first occurence (0) is used. @return if there is only a single character (remaining) in $fmt, and only a single occurence selectedi (no ..), a string is returned. Else you've been asking for an array, and so an array is returned even if it contains only a single value. If more than one subfield name was specified or the '*', the names are used ('' for the initial). If a range is selected, the index is used (in addition). Fields are added by first looping over subfield names, then occurences. PHP may or may not loop the array in that order. Example: '^ab' gives keys 'a' and 'b', 'b(1..' gives keys '1','2'...m '^cab(..', gives 'c0', 'c1', ... 'a0', ... */ function fmt ( $val, $fmt = ISIS_REC_MHL ) { global $ISIS_REC_MHL_PAT, $ISIS_REC_MHL_REPL; if ( null == $fmt ) return $val; if ( '&' != $fmt{0} && '%' != $fmt{0} ) $mode = ''; else { $mode = '&' == $fmt{0} ? 'htmlspecialchars' : 'urlencode'; $fmt = substr($fmt,1); } if ($mhl = ('M' == $fmt{0} && ISIS_REC_MHL == substr($fmt,0,3)) ) $fmt = substr($fmt,3); if ( '' == $fmt ) { if ( $mhl ) $val = preg_replace($ISIS_REC_MHL_PAT, $ISIS_REC_MHL_REPL, $val); return $mode ? $mode($val) : $val; } if ( '^' != $fmt{0} ) $sep = "\t"; else { $sep = '^'; $fmt = substr($fmt,1); } $sub = explode($sep,$val); // this really is not for performance ... $n = count($sub); if ( false !== strpos($fmt,'(') // quick check && preg_match( '/^([^(]*)\((\d*)(\.\.(\d*))?/', $fmt, $m ) // actually matches everything with a ( in it ) { $fmt = $m[1]; // initial part if (!($fst = $m[2]) ) $fst = 0; if (!($range = $m[3])) $lst = $fst; elseif (!($lst = $m[4])) $lst = $n; } else $fst = $range = 0; $strip = $l = strlen($fmt); if (!$strip) { // plain mode $fmt = '*'; $l = 1; } if ( !($single = 1 == $l && !$range) ) $ret = array(); // echo "sub '$fmt'($fst..$lst) range $range single $single strip $strip n $n\n"; for ($i = 0; $i < $l; $i++ ) { $s = $fmt{$i}; $sel = '*' == $s ? 0 : 1; $o = -1; for ($j = $sel; $j < $n; $j++ ) { // echo "test '$s'($fst..$lst) $o\n"; if ( ($sel && $s != $sub[$j]{0}) || ++$o < $fst ) continue; if ( $lst < $o ) break; // ok, it's in range $v = ($strip && $j) ? substr($sub[$j],1) : $sub[$j]; if ( $mhl ) $v = preg_replace($ISIS_REC_MHL_PAT, $ISIS_REC_MHL_REPL, $v); if ( $mode ) $v = $mode($v); if ( $single ) return $v; $key = $sel ? (1 == $l ? '' : $s) : $strip && $j ? $sub[$j]{0} : ''; if ($range) $key .= $o; $ret[$key] = $v; if ( !$range ) break; } } return $ret; } // //////////////////////////////////////////////////////////// // ctor // /** create an ISIS record. @return object Isis_Rec a new ISIS record */ function Isis_Rec () { $this->tag = array(); $this->val = array(); if ( func_num_args() ) $this->add( func_get_args() ); } // //////////////////////////////////////////////////////////// // getters // /** @return the number of fields */ function len () { return count($this->tag); } /** try to look up non-numeric tags in the fdt */ function fdt ( &$tag ) { if ( !is_int($tag) ) { if (is_numeric($tag)) $tag = (int)$tag; else if ($this->db && $this->db->fdt && is_int($this->db->fdt[$tag])) $tag = $this->db->fdt[$tag]; } } /** reset tag and val array pointers */ function res () { reset($this->tag); reset($this->val); } /** get all values for tag as array */ function get ( $tag ) { $ret = array(); reset($this->tag); while (list($p,$t) = each($this->tag)) if ( $tag == $t ) $ret[] = $this->val[$p]; return $ret; } /** v is for value -- get the value of the next occurence of tag in the record. The position is reset by res or when using v with a different tag. @param mixed $fmt format to apply - if $fmt is null (or the value is null), the value is returned unmodified - if $fmt is a string, the value is formatted by Isis_Rec::fmt (returns string or array). - if $fmt is 0, a new Isis_Rec is used as $fmt - if $fmt is an object, it is assumed to be a record, into which a subrecord (as of embed) is to be extracted. This also advances the loop position to after the child fields. The record is returned. - other values of $fmt are reserved for future extensions and currently return null. */ function v ( $tag = null, $fmt = null ) { if ( !is_int($tag) ) $this->fdt($tag); if ( $this->v != $tag ) { reset($this->tag); // no each on val needed $this->v = $tag; } /* This not working in 4.3.1 was due to an earlier print_r on tag, which completely corrupts the data (prob one of the 282 print_r bugs). So don't print_r. */ $v = null; if ( is_null($tag) ) { if (list($p,$t) = each($this->tag)) $v = $this->val[$p]; } else while (list($p,$t) = each($this->tag)) if ( $tag == $t ) { $v = $this->val[$p]; break; } if (is_null($v) || is_null($fmt)) return $v; if ( is_string($fmt) ) return Isis_Rec::fmt($v,$fmt); if ( 0 === $fmt ) $fmt = new Isis_Rec(); if ( is_object($fmt) ) { for ( $i = (int)$v; $i-- && list($p,$t) = each($this->tag); ) { $fmt->tag[] = $t; $fmt->val[] = $this->val[$p]; } return $fmt; } return null; } /** same as v($tag,'&'.$fmt) */ function h ( $tag = null, $fmt = '' ) { return $this->v($tag,'&'.$fmt); } /** same as v($tag,'&MHL') */ function mhl ( $tag = null ) { return $this->v($tag,'&MHL'); } // //////////////////////////////////////////////////////////// // setters // /** append a new field (tag-value-pair) to the end of the record. @param int $tag tag to use in the field. It is not enforced that tag is an integer. @param string $val the new fields value. The string type is not enforced here. @return the new value */ function append ( $tag, $val ) { $this->tag[] = $tag; return $this->val[] = $val; } /** add an array to the record. @param mixed $argv an array either as a single parameter or as a variable number of arguments. The array is processed as follows: - if an item is an int, it is appended, with the following item as value. - if an item is an array, add is called recursively on this array - if an item is '-db' or '-mfn', the corresponding properties are set - if we have an fdt which maps the item to an int, the following item as value is added with the tag given by the fdt. - if an item is ISIS_REC_TEXT, the following item is parsed in text mode. - else the item is parsed in standard mode. */ function add ( $argv ) { $added = 0; $fdt = $this->db ? $this->db->fdt : null; if (1 < func_num_args() || !is_array($argv)) $argv = func_get_args(); // standardized to having an array for ( $i = reset($argv); $i || !is_bool($i); $i = next($argv) ) { if ( is_int($i) ) { $this->append($i, next($argv)); $added++; } elseif ( is_array($i) ) $added += $this->add($i); elseif ( '-mfn' == $i ) $this->mfn = next($argv); elseif ( '-db' == $i ) { // they don't grok $fdt = ($this->db = next($argv))->fdt; $this->db = next($argv); $fdt = $this->db->fdt; } elseif ($fdt && is_int($fdt[$i])) { $this->append($fdt[$i], next($argv)); $added++; } elseif ( ISIS_REC_TEXT == $i ) $added += $this->parse( next($argv), ISIS_REC_TEXT ); else $added += $this->parse( $i ); } } /** recompact after unsetting */ function pack () { $this->tag = array_values( $this->tag ); $this->val = array_values( $this->val ); $this->holy = 0; } /** remove a field at given pos, poking a hole in the field list. */ function rm ( $pos, $pack = FALSE ) { unset( $this->tag[$pos] ); unset( $this->val[$pos] ); if ( $pack ) $this->pack(); else $this->holy++; } /** remove all fields or all with a given tag. */ function del ( $tag = null, $pack = FALSE ) { if ( is_null($tag) ) { $this->tag = array(); $this->val = array(); } else { if ( !is_int($tag) ) $this->fdt($tag); foreach ($this->tag as $p => $t) if ($tag == $t) $this->rm($p); if ( $pack ) $this->pack(); } } /** set fields with tag to values. set( 42, 'foo', 'bar', 'baz' ) will change the first three occurences of 42 to 'foo', 'bar' and 'baz', resp. - if there are less than three occurences, the remaining values are appended - if there are more than three occurences, the remaining occurences are deleted - if a value is the integer 0, processing stops (i.e. remaining occurences are left unchanged) - if a value is a positive integer n, processing skips n occurences (letting them unchanged) - if a value is an array, it's elements are used (it is flattened out non-recursively) @param mixed $tag tag by int or name @param mixed values... variable number of values */ function set ( $tag ) { if ( !is_int($tag) ) $this->fdt($tag); $argv = func_get_args(); reset($argv); each($argv); // eat $tag reset($this->tag); $ary = null; $more = 1; for (;;) { if ( $ary ) { if ( !(list($k,$arg) = each($ary)) ) { $ary = null; continue; } } else { if ( !(list($k,$arg) = each($argv)) ) break; if ( is_array($arg) ) { $ary = $arg; reset($ary); continue; } } echo "setting '$arg'\n"; if (is_int($arg)) { if (!$arg) return; while ($more && 0 < $arg--) while (($more = (list($k,$v) = each($this->tag))) && $tag != $v) ; continue; } // now arg is a value to set/add if ($more) { while (($more = (list($k,$v) = each($this->tag))) && $tag != $v) ; if ($more) { $this->val[$k] = $arg; continue; } } $this->append($tag,$arg); } if ( $more ) while (list($k,$v) = each($this->tag)) if ($tag == $v) $this->rm($k); } /* transparently embed a record. */ function embed ( $tag, $that ) { $this->add( $tag, $i = $that->len() ); for ( $t = reset($that->tag), $v = reset($that->val); $i--; $t = next($that->tag), $v = next($that->val) ) { $this->tag[] = $t; $this->val[] = $v; } } // //////////////////////////////////////////////////////////// // conversion to/from other representations // /** serialize record to a string. After each field, including the last one, a newline is added. @param string $mode replacement value for newlines. suggested is one of the predefined constants. defaults to ISIS_REC_BIN. @return the string representation of the record */ function toString ( $mode = ISIS_REC_BIN ) { $s = ''; // PHP is so braindead in almost every possible aspect, // they even managed to get the iterator wrong ... // why in heaven is next a pre-increment ??? $t = reset($this->tag); $v = reset($this->val); for ( $i = count($this->tag); $i--; $t = next($this->tag), $v = next($this->val) ) $s .= $t . "\t" . str_replace("\n",$mode,$v) . "\n"; return $s; } /** parse text as record fields to add. For each non-empty line, initial digits are used as tag (empty == 0), an optional following tab is skipped, and the rest used as value, after replacing $repl, if given, with newlines. If the line starts with a tab and the record is not empty, a newline and the value are appended to the last field, else a new field is appended to the record. @param string $repl string to be converted back to newlines. use ISIS_REC_TEXT, if you know text is from toString(ISIS_REC_TEXT) @return number of fields added */ function parse ( $text, $repl = null ) { // need compact array in order to reliably know last index $conv = 0; if ( $this->holy ) $this->pack(); foreach (explode("\n",$text) as $line) { if ('' == $line) // blank line or trailing newline continue; $dig = strspn($line,'0123456789'); $t = $dig ? (int)substr($line,0,$dig) : 0; $v = substr( $line, $dig + ("\t" == $line{$dig} ? 1 : 0) ); if ( $repl ) $v = str_replace($repl,"\n",$v); if ("\t" == $line{0}) { // continuation line // need compact array in order to reliably know last index if ( $this->holy ) $this->pack(); if ($c = count($this->val)) { $this->val[$c-1] .= "\n".$v; continue; } } $this->tag[] = $t; $this->val[] = $v; $conv++; } return $conv; } // parse /** return the "data fork" of this record by mapping a function to tags and values in parallel. @param function $func defaults to null, resulting in an array of fields, each an array [0] => $tag, [1] => $val. @return a new array as of array_map */ function map ($func = null) { return array_map($func,$this->tag,$this->val); } } // class Isis_Rec ?> php/Isis/Server.php0100644000113000011300000000432407664425364013233 0ustar kripkriphost = $host; $this->port = $port; $this->open(); } function open () { // pfsockopen will reuse an existing socket on same host:port // should we use stream_socket_client ??? $this->sock = pfsockopen( $this->host, $this->port // , $errno, $errstr ); // if ( false == $this->sock ) // echo "could not connect to $host:$port: $errno '$errstr'"; // stupid PHP prints warnings itself %< return $this->sock; } function request ( $req ) { if ( !$this->sock && !$this->open() ) return null; // echo "SEND\n", $req->toString(ISIS_REC_TEXT); fwrite($this->sock, $req->toString(ISIS_REC_TEXT) . "\n"); fflush($this->sock); $res = new Isis_Rec; while ( ($line = fgets($this->sock)) && trim($line) ) { // echo "RETR ", $line; $res->parse($line, ISIS_REC_TEXT); } return $res; } } // class Isis_Server ?> php/echo0100664000113000011300000000014007662711675011201 0ustar kripkriptoString(), "\n"; ?> php/file0100664000113000011300000000102707663440605011200 0ustar kripkrip 70 ) ); $f = new Isis_File( '/tmp/test', 1 ); if ( $r->len() ) { $n = $f->write( $r->toString(), $_GET['W'] ); echo "wrote mfn $n\n"; } if ($_GET['C']) $f->compact(); for ( $n=0; ++$n <= $f->len; ) { $foo = $f->read($n); // if ( 1000 < $n ) break; echo "\nW\t$n\n$foo\n"; // echo "\nW $n\n"; // pretty print // for ( $r = new Isis_Rec('-mfn', $n, $foo); $v = $r->mhl(); ) echo "$v\n"; } ?> php/foo0100644000113000011300000001153307663441444011047 0ustar kripkrip Demo for package Isis 22, 'b' => 42 ) ); // create a db with field list ("fdt") $db = new Isis_Db( array( 'title' => 24, 'author' => 70, 'keywords' => 69 ), 'cds' ); // a special value using lotsa subfields with the traditional sep ^ $subs = 'initial^aParis^bUnesco^b^c-1965'; // create record using various add parameters $r = new Isis_Rec( // set the db '-db', $db, // first some lines from CDS, some using field names, some plain int tags 'keywords', 'Paper on: ', 'author', 'Magalhaes, A.C.', 24, ' Controlled climate in the plant chamber', 76, 'Les Politiques de la communication en Yougoslavie^zfre', 'author', 'Franco, C.M.', 26, $subs, // a field to test del 77, 'ave Caesar', // a field using tab as subfield separator 42, "foo\tbar\tbaz", // a field containing newline 99, "two\nlines", // a serialized record (as of toString) as parameter "70\tyet another author\n99two more\n\tlines\na 0 field\n42\tthe\tanswer" ); // dump the stuff ?>

dump of record

have len()?> fields 
--
toString())?>--

examples of field selector

tag); // print_r corrupts data with 4.3.1 // print_r($r->map()); $r->del( 77 ); // ... morituri te salutant // various examples how to access fields foreach ($db->fdt as $name => $tag) while ($v = $r->mhl($tag)) echo "$name = $v
\n"; ?> v26^a = h(26,'^a')?>
res(); // more convenient to get them all at once // the same as Isis_rec::fmt($r->v(26),'&^abc'), foreach ($r->h(26,'^abc') as $k => $v) echo "v26^$k = $v
\n"; ?> v69 = mhl('keywords')?>
v76 = mhl(76)?>
v42(1) = h(42,'(1)')?> formats returning arrays on '$subs'\n"; foreach (array( '&^cab(..', '&^*(..', '&^(..', '&^bca(1..', '&MHL^cab(..' ) as $fmt) { echo "\n\n

Format "$fmt"

\n"; foreach (Isis_rec::fmt($subs,$fmt) as $k => $v) echo "v26^$k = $v
\n"; } ?>

embedding and TEXT mode

embed( 22, $r ); $s = $q->toString(ISIS_REC_TEXT); echo "

record embedded as field 22

\n
---\n$s---\n
\n"; $q->del(); // restore from the string $q->parse( $s, ISIS_REC_TEXT ); // extract the original r $r = $q->v( 22, 0 ); $r->db = $db; echo "

record restored from field 22

\n", "
---\n", $r->toString(), "---\n
\n"; ?>

set operator

set( 'title', 'new title', 'second new title' ); // change 99s, removing one $r->set( 99, 'now a oneliner' ); // change author[1] only, also test array parameter $r->set( 'author', array(1, 'Blanco', 0) ); echo "
---\n", $r->toString(), "---\n
\n"; ?>

talking to echo server

Unfortunately, PHP GET/POST based echo server can not preserve field order. request( $r ); if ( $res ) echo "\n
---\n", $res->toString(), "---
\n"; else echo "failed\n"; ?> php/php.cgi0100775000113000011300000000103207661762723011616 0ustar kripkrip#!/bin/sh # echo -e 'Content-Type: text/plain\n'; env # on /php.cgi/foo, fnord will give us # REQUEST_URI=/php.cgi/foo # SCRIPT_NAME=/php.cgi # PATH_INFO=/foo # (+ QUERY_STRING, CONTENT_TYPE/LENGTH, if any, + some) # php wants at least SCRIPT_FILENAME, fix two others, too # could need a little bit refinement, if part of PATH_INFO is really PATH_INFO SCRIPT_NAME=$SCRIPT_NAME$PATH_INFO \ PATH_INFO= \ SCRIPT_FILENAME=/opt/openisis/php$PATH_INFO \ PHP_FCGI_CHILDREN=3 exec /opt/bin/cgi-fcgi -connect /tmp/php.fcgi /opt/PHP432RC3/bin/php php/query0100664000113000011300000000043207663453654011435 0ustar kripkripquery('plant * water'); foreach ($mfns as $n) { echo "---$n\n"; $r = $db->read($n); if ( $r ) while ( $v = $r->mhl() ) echo "$v\n"; } ?> searchtest.txt0100644000113000011300000001327207415453163012456 0ustar kripkrip2 Methodology of plant eco-physiology: proceedings of the Montpellier Symposium 3 Control of conditions in the plant chamber: fully automatic regulation of wind velocity, temperature and relative humidity to conform to microclimatic field conditions 4 Methodology of plant eco-physiology: proceedings of the Montpellier Symposium 5 Methodology of plant eco-physiology: proceedings of the Montpellier Symposium 6 Methodology of plant eco-physiology: proceedings of the Montpellier Symposium 17 Measurement of drought resistance in crop plants 22 Methodology of plant eco-physiology: proceedings of the Montpellier Symposium 24 Calculation model for the actual evapotranspiration from cropped areas and other terms of the water balance equation 29 Sur les caract‚ristiques texturales et min‚ralogiques des s‚diments r‚cents et des sols alluviaux dans le delta du Mangoky, sud-ouest de Madagascar 30 Apercu sur la v‚g‚tation occupant les alluvions r‚centes 31 Place de la g‚omorphologie dans l'‚tude de la mise en valeur des deltas tropicaux 34 Use of aerial photographs in delta studies 37 Mud transport studies using manganese as an accompanying element under temperate and tropical climatic conditions: Western Europe, Amazon area and Chao-Phya river area in Thailand 39 Soils in the estuary of the Amazon 41 Morphology, genesis and occurence of acid sulphate soils, cat-clays in deltas of the humid tropics 43 Interaction of river and sea water in the absence of tides 46 Methodology in area analysis 47 Hydrological achievements and social problems 50 Influence of man's traditional activities on soils and vegetation 55 Vegetation as a geological agent in tropical deltas 57 Ground water in deltas 58 Surface hydrology of deltaic areas 59 Control of floods in East Pakistan 60 General features of East Pakistan vegetation 61 Estuarine vegetation of Ceylon 62 Estuaries of the Amazon and Oiapoque rivers and their floras 65 Plankton productivity in relation to tropical fish culture 67 Fish and fisheries of the Sunderbans 69 Humid tropics and their vegetation 70 Some important animal pests and parasites of East Pakistan 71 Agriculture with special reference to rice cultivation in humid tropical zone deltas 73 Need for land-use studies in East Pakistan 74 Deltaic area of the Irrawaddy river in Burma 75 General features of the delta of the Amazon 79 La Politique culturelle en R‚publique unie du Cameroun^zfre 80 Hydrometeorological and tidal problems of the deltaic areas in India 82 Cross-cultural universals of affective meaning 84 Adult education in India; a book of readings 85 Training for change agents; a guide to the design of training programs in education and other fields 86 Higher education and the labour market in Zambia: expectations and performance 88 Open university; a personal account by the first Vice-Chancellor 89 Classical Greece 90 Long way from home; effects of public high schools on village children away from home 92 Peruvian model of innovation: the reform of basic education 94 Chandi Borobudur; a monument of mankind 95 Mucheke: race, status and politics in a Rhodesian community 97 Human rights studies in universities 99 Studies in the philosophy of biology: reduction and related problems 101 Manuel pratique des assurances sociales 102 Unesco, an ideal in action 103 About the quality of life 105 Les Statistiques de la science et de la technologie en Am‚rique latine ^zfre 109 Annual review of energy - v.1, 1976 110 Multinational product strategy; a typology for analysis of worldwide product innovation and diffusion 111 Multinational corporation and social change 115 Folklore from Africa to the United States; an annotated bibliography 117 Demand for part-time learning in Ontario 119 Meeting on the Implementation of the Recommendations of the Third Regional Conference of Ministers of Education and Ministers Responsible for Economic Planning in the Arab States^pCairo^d1976 120 Meeting on the Implementation of the Recommendations of the Third Regional Conference of Ministers of Education and Ministers Responsible for Economic Planning in the Arab States^pCairo^d1976 122 Meeting of Experts on the Content of Education in the Context of Life-long Education^pParis^d1975 123 Meeting on the Implementation of the Recommendations of the Third Regional Conference of Ministers of Education and Ministers Responsible for Economic Planning in the Arab States^pCairo^d1976 124 Meeting on the Implementation of the Recommendations of the Third Regional Conference of Ministers of Education and Ministers Responsible for Economic Planning in the Arab States^pCairo^d1976 125 Regional Seminar on Training^n2nd^pDakar^d1975 128 - 129 Incl. bibl. 130 Joint FAO/WHO Expert Committee on Nutrition^n9th^pRome^d1974 131 Measurement of scientific and technical activities - Proposed standard practice for surveys of research and experimental development 133 Consultancy services available in developing ESCAP countries 134 Training courses available in developing ESCAP countries 137 Audit informatique 138 Addresses delivered at a memorial ceremony for Ren‚ Maheu formerly Director-General of Unesco (1961-1974), Unesco House, 26 April 1976 140 Armaments and disarmament in the nuclear age; a handbook 141 Energy for rural development: renewable resources and alternative technologies for developing countries 145 Phenomenology and natural existence 146 Australian perspectives on lifelong education 149 Symposium on the Psychological Bases of Programmed Learning ^pTbilisi, USSR^d1976 151 Manuel pratique des assurances sociales tcl/Makefile0100644000113000011300000001536307665623725012007 0ustar kripkrip# #/* # openisis - an open implementation of the CDS/ISIS database # Version 0.8.x (patchlevel see file Version) # Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # see README for more information #EOH */ # # # $Id: Makefile,v 1.20 2003/05/30 09:43:49 kripke Exp $ # the Makefile # # make # make ORATCL=3.3 # make ORATCL=4.1 # make PGLIB=/opt/postgres/lib # make.tcltk is basically a compilation of the t*Config.sh scripts # it's created using # ./configure # see maketcltk for hints on building the tcl/tk # # ./configure /opt/TclTk/8.4.2.t.so # ./configure /opt/TclTk/8.3.5.t.so # # for the binary distribution, we build statics like: # ./configure /opt/TclTk/8.3.5.a/lib # # actually used are: # TCL_SHARED_BUILD 0 or 1 # TCL_DL_LIBS e.g. -ldl # TCL_LD_FLAGS e.g. -rdynamic to enable explit loading of extensions # TCL_DL_LIBS e.g. -ldl for the same reason # TCL_LIB_SPEC e.g. -L/opt/TclTk/8.3.5.t.so/lib -ltcl8.3 # TCL_LD_SEARCH_FLAGS e.g. -Wl,-rpath,${LIB_RUNTIME_DIR} # TCL_LIBS e.g. -lpthread -lieee -lm # TCL_DEFS lots of -Defines like -D_REENTRANT=1 or -DSTATIC_BUILD=1 # TK_LIB_SPEC -L/opt/TclTk/8.3.5.t.so/lib -ltk8.3 # TK_LIBS -L/usr/X11R6/lib -lX11 -ldl -lpthread -lieee -lm ifeq (,$(WIN32)) include make.tcltk else # ./configure -mingw /opt/TclTk/8.3.5.w.a include make.tcltk.mingw endif # mingw build leaves many options unset (newer configure cares) ifeq (,$(TCL_LIB_SPEC)) TCL_LIB_SPEC=$(TCL_PREFIX)/lib/$(TCL_LIB_FILE) endif ifeq (,$(TK_LIB_SPEC)) TK_LIB_SPEC=$(TK_PREFIX)/lib/$(TK_LIB_FILE) endif # 8.4 only ? ifeq ($(TCL_INCLUDE_SPEC),) TCL_INCLUDE_SPEC=-I$(TCL_PREFIX)/include endif MT=$(findstring TCL_THREADS,$(TCL_DEFS)) ifeq (,$(MT)) O=sto else O=mto OITS=bin/oits endif BIN=bin EXE= RUN= GUI= ifneq ($(WIN32),) include ../make.mingw O=win BIN=win EXE=.exe RUN=wine -- GUI=-mwindows tclsubdir=win endif ifneq ($(TK_VERSION),) WISISH=$(BIN)/wisish$(EXE) else WISISH= endif TARGETS=$(BIN)/isish$(EXE) lib/pkgIndex.tcl $(BIN)/tspr$(EXE) $(OITS) $(WISISH) OBJS=$(BIN)/openisistcl.o # standard make defines CC=gcc RANLIB=ranlib ifeq (solaris,$(OSTYPE)) # for nanosleep; others are in TCL_LIBS SYSLIBS=-lrt endif ifneq ($(TKHTML),NO) # ... wisish needs it TKHDEF=-DTKHTML=20 ifeq (1,$(TCL_SHARED_BUILD)) TKHLINK=-ltkhtml else TKHLINK=$(TCL_PREFIX)/lib/libtkhtml.a endif endif ifneq ($(TIX),) TIXDEF=-DTIX=$(TIX) TIXLINK=-ltix$(TIX) endif ifneq ($(ORATCL),) ORADEF=-DORATCL=$(ORATCL) ORALINK=-loratcl$(ORATCL) -L$(ORACLE_HOME)/lib ORAPATH=:$(ORACLE_HOME)/lib endif ifneq ($(PGLIB),) PGDEF=-DPGTCL PGLINK=-lpgtcl -L$(PGLIB) PGPATH=:$(PGLIB) endif # referenced by TCL_LD_SEARCH_FLAGS in make.tcl LIB_RUNTIME_DIR=$(TCL_PREFIX)/lib$(ORAPATH)$(PGPATH) # DBG= ifeq ($(NDEBUG),) DBG=-g else DBG=-DNDEBUG endif # note that -O is necessary for -Wuninitialized warnings # java needs long long # -fstack-check ??? CFLAGS=-O2 -Wall -W -Wshadow -pedantic -Wno-long-long \ $(DBG) $(TCL_DEFS) $(TKHDEF) $(TIXDEF) $(ORADEF) $(PGDEF) \ -I.. $(TCL_INCLUDE_SPEC) # DB to use -- define DB for db/DB/DB, or DBDIR for other path ifeq ($(DB),) DB=cds endif ifeq ($(DBDIR),) DBDIR=db/$(DB)/$(DB) endif # standard default target all: $(TARGETS) .PHONY: all clean test # do not delete intermediates .PRECIOUS: clean: rm -rf *.o *.a *.so *.dll *.tmp *.exe \ bin win Makefile.old testout.txt lib/pkgIndex.tcl: openIsis.tcl @- [ -d lib ] || mkdir lib @- [ -h lib/openIsis.tcl ] || ln -s ../openIsis.tcl lib/. echo "pkg_mkIndex lib" | tclsh ifeq (,$(MT)) $(BIN)/oits$(EXE): : please use a threaded build of Tcl exit 1 else $(BIN)/oits$(EXE): $(BIN)/oits.o $(OBJS) ../mto/libopenisis.a ../mto/liboisrv.a @- [ -d $(BIN) ] || mkdir $(BIN) $(CC) $(TCL_LD_FLAGS) -o $@ $^ \ $(TCL_LIB_SPEC) $(TCL_DL_LIBS) \ $(ORALINK) $(PGLINK) \ $(TCL_LD_SEARCH_FLAGS) \ $(TCL_LIBS) $(SYSLIBS) endif $(BIN)/isish$(EXE): $(BIN)/isish.o $(OBJS) ../$(O)/libopenisis.a @- [ -d $(BIN) ] || mkdir $(BIN) $(CC) $(TCL_LD_FLAGS) -o $@ $^ \ $(TCL_LIB_SPEC) $(TCL_DL_LIBS) \ $(ORALINK) $(PGLINK) \ $(TCL_LD_SEARCH_FLAGS) \ $(TCL_LIBS) $(SYSLIBS) $(BIN)/tspr$(EXE): $(BIN)/tspr.o $(OBJS) ../$(O)/libopenisis.a @- [ -d $(BIN) ] || mkdir $(BIN) $(CC) $(TCL_LD_FLAGS) -o $@ $^ \ $(TCL_LIB_SPEC) $(TCL_DL_LIBS) \ $(ORALINK) $(PGLINK) \ $(TCL_LD_SEARCH_FLAGS) \ $(TCL_LIBS) $(SYSLIBS) $(BIN)/wisish$(EXE): $(BIN)/wisish.o $(OBJS) ../$(O)/libopenisis.a @- [ -d $(BIN) ] || mkdir $(BIN) $(CC) $(TCL_LD_FLAGS) $(GUI) -o $@ $^ \ $(TCL_LIB_SPEC) $(TCL_DL_LIBS) $(TK_LIB_SPEC) $(TKHLINK) \ $(ORALINK) $(PGLINK) \ $(TIXLINK) \ $(TCL_LD_SEARCH_FLAGS) \ $(TK_LIBS) $(SYSLIBS) test: $(BIN)/isish$(EXE) $(RUN) $(BIN)/isish$(EXE) test/test1.tcl $(RUN) $(BIN)/isish$(EXE) test/test2.tcl $(RUN) $(BIN)/isish$(EXE) test/testrec.tcl $(RUN) $(BIN)/isish$(EXE) test/teststb.tcl run: $(BIN)/wisish$(EXE) $(RUN) $(BIN)/wisish$(EXE) tisis ../db/cds/cds install: all strip $(BIN)/* cp $(BIN)/* $(TCL_PREFIX)/bin [ -d $(TCL_PREFIX)/lib/openIsis ] || mkdir $(TCL_PREFIX)/lib/openIsis cp openIsis.tcl lib/pkgIndex.tcl $(TCL_PREFIX)/lib/openIsis cp tisis detail.tsp list.tsp $(TCL_PREFIX) ../mto/libopenisis.a: @ $(MAKE) -C .. mto/libopenisis.a ../mto/liboisrv.a: @ $(MAKE) -C .. mto/liboisrv.a ../win/libopenisis.a: @ $(MAKE) -C .. WIN32=1 win/libopenisis.a $(BIN)/%.o: %.c @- [ -d $(BIN) ] || mkdir $(BIN) $(CC) -c $(CPPFLAGS) $(MTFLAGS) $(CFLAGS) -o $@ $< # please do not try to make deps w/o sed, gcc and mv ... deps: sed -n -e '1,/^# generated deps/p' Makefile >Makefile.tmp gcc -I. -I.. -MM `../lscvs |grep '\.c$$' |sort` | \ sed -e '/^[^ ]/s/^/$$(BIN)\//' >>Makefile.tmp mv -f Makefile Makefile.old mv Makefile.tmp Makefile -diff Makefile Makefile.old # generated deps -- do not edit below this line $(BIN)/isish.o: isish.c ../openisis.h openisistcl.h $(BIN)/oits.o: oits.c ../lsv.h ../loi.h ../openisis.h ../lio.h ../luti.h \ ../lll.h ../lses.h ../ldsp.h openisistcl.h $(BIN)/openisistcl.o: openisistcl.c ../openisis.h openisistcl.h ../luti.h \ ../loi.h ../ldsp.h $(BIN)/tspr.o: tspr.c ../openisis.h openisistcl.h $(BIN)/wisish.o: wisish.c ../openisis.h openisistcl.h tcl/README0100644000113000011300000000366407654002045011210 0ustar kripkrip# openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2001-2002 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: README,v 1.2 2003/04/30 17:13:09 kripke Exp $ # tcl/tk binding # To compile the tcl/tk binding, you need version 8.3.5 of tcl and tk, though it may work with other versions too. The standard GUI application uses D. Richard Hipp's TkHTML 2.0. It is recommended, but not necessary, that you create a fresh compile of Tcl/Tk and extensions using the maketcltk script. Then use configure to create the file make.tcltk, which will contain various settings as appropriate for your installation (e.g. wether it is a shared and/or threaded build). Then switch to the tkhtml subdirectory to build tkhtml using make TKHTML= install to install the TkHTML libraries. If the environment variable WIN32 is set to something non-empty, maketcltk, configure and the Makefiles will attempt a mingw build. tcl/addmsgkey.tcl0100755000113000011300000000244307663442007013007 0ustar kripkrip#!./bin/isish # adds a key -> msgID mapping to the key index package require openIsis 0.8 proc usage {} { puts "usage: ./addmsgkey {msgid} {key}" exit } # check for args if { 2 != $argc } usage set key [lindex $argv 1 ] set mfn [lindex $argv 0 ] incr mfn 1 set stb [ openIsis ] $stb add \ syspath /opt/openisis/db set kdb "[ $stb get syspath ]/msg/key" $stb req type open db $kdb openIsis::prnres "open result: " $stb set db [$stb .res get db ] $stb req type query key $key mode 1 set size [$stb .res get size] if { $size } { puts "removing all $key -> xxx mappings" set mfns [$stb .res get {rowid {}}] puts "mfns: $mfns" set i $size while { $i } { set m [lindex $mfns [expr $i-1]] puts "removing mfn $m" set idx [ $stb .req .idx ] $idx add -1 "del" $idx add -1 "mfn\t$m" $idx add -1 "fields" $idx add 42 $key $stb .req wrap -tag 926 $idx $stb req type delete incr i -1 } } puts "try to add $key -> $mfn" set idx [ $stb .req .idx ] $idx add -1 "add" $idx add -1 "mfn\t$mfn" $idx add -1 "fields" $idx add 42 $key $stb .req wrap -tag 926 $idx $stb req type insert openIsis::prnres "result: " $stb puts "now try to get $key" $stb req type query key $key mode 1 set size [$stb .res get size] if { 1 == $size } { puts "ok." } else { puts "err. please inspect $kdb.oxi" } tcl/configure0100755000113000011300000000634307646060557012251 0ustar kripkrip#!/bin/bash # $Id: configure,v 1.5 2003/04/12 18:56:47 kripke Exp $ # ./configure /opt/TclTk/8.3.5.t.so # ./configure -mingw /opt/TclTk/8.3.5.w.a mingw=false case "$1" in -mingw) mingw=true; shift;; *) [ -n "$WIN32" ] && mingw=true;; esac cfg=${1:-/usr} dir=$cfg tgt=make.tcltk $mingw && tgt=$tgt.mingw for ext in "" /lib /unix; do dir=$cfg$ext [ -f $dir/tclConfig.sh ] && break done if [ ! -f $dir/tclConfig.sh ]; then echo " no file tclConfig.sh under $cfg usage $0 dir-containing-tclConfig.sh " >&2 exit 1 fi echo using cfg $dir date="`date`" cat <$tgt # configured $date by # $0 $cfg # # $dir/tclConfig.sh: EOF . $dir/tclConfig.sh # mingw build leaves many options unset # note it's NOT -ltcl$TCL_VERSION, but libtcl83s.a for mingw :( # specify -L$TCL_PREFIX/lib anyway for other libs ... [ -z "$TCL_LIB_SPEC" ] && TCL_LIB_SPEC="-L$TCL_PREFIX/lib $TCL_PREFIX/lib/$TCL_LIB_FILE" while read line; do case "$line" in TCL*=*) var=${line%%=*} eval echo $var=\$$var eval echo $var=\$$var >&2 esac done <"$dir"/tclConfig.sh >>$tgt # include_spec 8.4 only, it seems [ -z "$TCL_INCLUDE_SPEC" ] && echo TCL_INCLUDE_SPEC=-I$TCL_PREFIX/include >>$tgt if [ -f "$dir"/tclxConfig.sh ]; then cat <>$tgt # # $dir/tclxConfig.sh: # EOF . $dir/tclxConfig.sh while read line; do case "$line" in TCLX*=*) var=${line%%=*} eval echo $var=\$$var eval echo $var=\$$var >&2 esac done <"$dir"/tclxConfig.sh >>$tgt fi # tcldom if [ -f "$dir"/TcldomConfig.sh ]; then cat <>$tgt # # $dir/TcldomConfig.sh: # EOF . $dir/TcldomConfig.sh while read line; do case "$line" in Tcldom*=*) var=${line%%=*} eval echo $var=\$$var eval echo $var=\$$var >&2 esac done <"$dir"/TcldomConfig.sh >>$tgt fi # tclxml if [ -f "$dir"/TclxmlConfig.sh ]; then cat <>$tgt # # $dir/TclxmlConfig.sh: # EOF . $dir/TclxmlConfig.sh while read line; do case "$line" in Tclxml*=*) var=${line%%=*} eval echo $var=\$$var eval echo $var=\$$var >&2 esac done <"$dir"/TclxmlConfig.sh >>$tgt fi ### Tk stuff # if [ ! -f "$dir"/tkConfig.sh ]; then echo "no tkConfig.sh in $dir" >&2 echo "# (no $dir/tkConfig.sh)" >>$tgt # skip other Tk extensions exit fi cat <>$tgt # # $dir/tkConfig.sh: # EOF . $dir/tkConfig.sh [ -z "$TK_LIB_SPEC" ] && TK_LIB_SPEC="-L$TK_PREFIX/lib $TK_PREFIX/lib/$TK_LIB_FILE" while read line; do case "$line" in TK*=*) var=${line%%=*} eval echo $var=\$$var eval echo $var=\$$var >&2 esac done <"$dir"/tkConfig.sh >>$tgt ### tkx if [ -f "$dir"/tkxConfig.sh ]; then cat <>$tgt # # $dir/tkxConfig.sh: # EOF . $dir/tkxConfig.sh while read line; do case "$line" in TKX*=*) var=${line%%=*} eval echo $var=\$$var eval echo $var=\$$var >&2 esac done <"$dir"/tkxConfig.sh >>$tgt fi ### tix if [ -f "$dir"/tixConfig.sh ]; then cat <>$tgt # # $dir/tixConfig.sh: # EOF . $dir/tixConfig.sh while read line; do case "$line" in TIX*=*) var=${line%%=*} eval echo $var=\$$var eval echo $var=\$$var >&2 esac done <"$dir"/tixConfig.sh >>$tgt fi ### blt if [ -f "$dir/libBLT24.a" ]; then cat <>$tgt # # $dir/libBLT24.a: # BLT_STATIC=$dir/libBLT24.a EOF fi if [ -f "$dir/libBLT24.so" ]; then cat <>$tgt # # $dir/libBLT24.so: # BLT_SHARED=-lBLT24 EOF fi tcl/demo.cgi0100775000113000011300000000465607653545772011771 0ustar kripkrip#!/opt/isis/OpenIsis/tcl/bin/tspr <%-- test Tcl server page $Id: demo.cgi,v 1.1 2003/04/29 19:01:14 kripke Exp $ --%><%! # package require mime; # namespace import ::mime::*; # a procedure to make live with requests easier: # it uses the current value of the $isis server variable to send a request # so we can switch servers by setting isis, yet use a single command proc ::isis {request args} { global isis $isis .req delete $isis .req set -type $request # eval so that the single list of args becomes multiple parameters eval $isis req $args ::puts stderr "###req\n[$isis .req serialize]###res\n[$isis .res serialize]###" # we don't clone the result rec, but provide a procedure # if {"read" == $request} { # clone result data as v; throws exception if no such record # $isis .res .rec clone v # } # return the size (number of records or mfns) $isis .res get size } # convenient access to the result proc ::res {args} { global isis eval $isis .res $args } # wrapper for current result record # add the "rowid" command, which the record is lacking for some reason # add a "get first", if args is a single integer proc ::v {args} { global isis if {"rowid" == $args} { return [$isis .res get rowid] } if {[string is integer $args]} { # v 24 return [lindex [$isis .res .rec get $args] 0] } eval $isis .res .rec $args } %><% # request standard tags # -10 SCRIPT_NAME # -11 PATH_INFO # -12 REMOTE_USER # global env set script [lindex [request get -10] 0] set path [lindex [request get -11] 0] set query [lindex [request get 1] 0] set db cds set ::isis [openIsis] ;# contact local server $::isis add -syspath /opt/isis/db -dbpath $db -defaultdb $db isis open # query max row isis maxrow set max [res get rowid] %> max mfn in <%=db%> is <%=max%>
enter query or mfn:

<% if [string is integer $query] { set mfns $query } else { # query isis query -mode 128 -key $query set mfns [res get rowid] %> found mfns <% foreach mfn $mfns { %> <%=mfn%> <% } } %> <% foreach mfn $mfns { isis read -rowid $mfn %> <% ::openIsis::Tspr detail.tsp } %>
MFN<%=mfn%>
tcl/detail.tsp0100644000113000011300000000150107656443645012327 0ustar kripkrip<%! # define a procedure to be called proc pr {tag name} { # list of all occurences of $tag set occ [v $tag] if {"" == $occ} {return 0} ;# empty set occ [MHLl $occ] # this section ends with a ! before the %, # so the HTML becomes part of this procedure !%> <%=name%><%=join $occ
%> <%! return 1 } %> <% if [pr 24 TITLE] { # looks like CDS ;) pr 70 AUTHOR(S) pr 12 CONFERENCE pr 25 EDITION pr 26 IMPRINT pr 30 COLLATION pr 44 SERIES pr 50 NOTES pr 71 "CORP. BODY" pr 72 "ADDED CONF" pr 74 "OTHER TITLE" pr 76 "OTHER LANG" pr 69 KEYWORDS } else { # default to "ALL" listing foreach {tag val} [v get -tags] { %> <% } } %>
<%=tag%><%=MHL $val%>
 tcl/dexml0100775000113000011300000000526607653234152011376 0ustar kripkrip#!/opt/isis/OpenIsis/tcl/bin/isish # let editor know it's Tcl vim:syntax=tcl package require openIsis; package require xml; # map element names to tags array set tag {} # highest ag used; set to 99 so we start from 100 set ltag 99 # map attribute names to subfields # map keys are "tag\tatt" array set subf {} openIsisRec rec proc cdata {data} { # text # puts "0\t[string map {\n " "} $data]" rec add 0 [string map {\n " "} $data] } proc comment {data} { # puts "#\t[string map {\n " "} $data]" } proc doctype {data} { puts "#\t$data" } proc estart {name attlist args} { global tag ltag subf lsubf set start + foreach {arg val} $args { switch -- $arg -empty { if $val { set start - } } } if [catch {set t $tag($name)}] { set t [set tag($name) [incr ltag]] } set line $start foreach {att val} $attlist { set attkey "$t\t$att" if [catch {set s $subf($attkey)}] { # use $subf($t) as counter for last subfield if [catch {set c [incr subf($t)]}] { set c [set subf($t) 97] } set s [set subf($attkey) [format %c $c]] } append line "^$s$val" } # puts "$t\t$line" rec add $t $line } proc eend {name args} { global tag foreach {arg val} $args { switch -- $arg -empty { if $val { return } } } # puts $tag($name) rec add $tag($name) "" } set xp [xml::parser \ -ignorewhitespace yes \ -reportempty yes \ -characterdatacommand cdata \ -commentcommand comment \ -doctypecommand doctype \ -elementstartcommand estart \ -elementendcommand eend \ -defaultcommand error ] # doesn't like that: # -endcdatasectioncommand ecdata # -startcdatasectioncommand scdata $xp parse [read stdin] # invert the FDT tag,subf to name,attr foreach {n v} [array get tag] { set name($v) $n} foreach {n v} [array get subf] { if {2 == [scan $n "%u\t%s" t a]} { set attr($t\t$v) $a } } foreach t [lsort [array names name]] { # puts stderr "$t\t$name($t)" ;# verbose entry set line $name($t) foreach s [lsort [array names attr "$t\t*"]] { # puts stderr "$s\t$attr($s)" ;# verbose entry scan $s "%u\t%s" - c append line "^$c$attr($s)" } puts stderr "$t\t$line" ;# the one-line DTD ;) } # use the inverted FDT to create the XML text puts "" set depth 0 rec do t v { if !$t { # text puts [string repeat \t $depth]$v continue } if {"" == $v} { # close incr depth -1 puts [string repeat \t $depth] continue } set line "<$name($t)" set alist [split $v ^] set hasChilds [expr {"+" == [lindex $alist 0]}] foreach a [lrange $alist 1 end] { append line " $attr($t\t[string index $a 0])=\"[string range $a 1 end]\"" } if $hasChilds { append line > } else { append line /> } puts [string repeat \t $depth]$line if $hasChilds { incr depth } } tcl/fdt21.tcl0100755000113000011300000001220407656730114011754 0ustar kripkrip#!/opt/openisis/tcl/bin/isish # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: fdt21.tcl,v 1.3 2003/05/08 18:00:48 mawag Exp $ # convert fdt21 to openIsis fdt package require mlcm::util set syspath "" # usage: fdt21.tcl ?fdt21path fdtpath? ?...? if !$argc { # test set syspath /opt/openisis/db/mlcm/ADB/ISISMARC set argv [list fdt21 fdttest] set argc 2 } # comparisation of two fd recs proc cmpfd {f1 f2} { set res [expr [$f1 get 860] - [$f2 get 860]] if $res { return $res } string compare [$f1 get {861 ""}] [$f2 get {861 ""}] } # op on db and check response for error proc sendrqs {msg db type args} { global stb0 if [catch { eval $stb0 req type $type db $db $args $stb0 .req delete }] { global errorInfo puts $errorInfo puts "$msg: $stb0 sendrqs $db $type $args" exit 1 } set err 0 set err2 0 if [catch { set err [$stb0 .res get error] set err2 [$stb0 .res get error2] }] { global errorInfo puts $errorInfo puts "$msg: $stb0 checkerr $db $type $args" exit 1 } if {$err || $err2} { puts "ERR = $err $err2" puts "$msg: $stb0 response $db $type $args" exit 1 } } # set stb0 [openIsis syspath $syspath v 9] set stb0 [openIsis syspath $syspath] set ok 1 # loop over db's foreach {db out} $argv { sendrqs openIn $db open sendrqs maxIn $db maxrow set maxin [$stb0 .res get rowid] sendrqs openOut $out open sendrqs maxOut $out maxrow set maxout [$stb0 .res get rowid] puts "db $db maxrow $maxin $maxout" # our fdt rec and list of fd's set fdt [openIsisRec] set fdlst [list] # read fdt21 db for {set r 1} {$maxin >= $r} {incr r} { set rlst "" if [catch { sendrqs "read($r)" $db read rowid $r # mainfield setup set rec [$stb0 .res .rec] set rlst [$rec get -tags] set fd [openIsisRec] set info [openIsisRec] set tag [$rec get 1] set len [$rec get {20 0}] $fd add \ 862 0 \ 860 $tag \ 866 [lindex [$rec get 2] 0] \ 863 [$rec get {5 0}] if $len { $fd add 865 -$len } { $fd add 865 0 } # loop over subfields for {set pos 0} 1 {incr pos} { if [catch { set subrec [$rec get 100\[$pos\]] }] { break } set issub 1 set sublst [mlcm::util::splitfld $subrec] foreach {c val} $sublst { if [string equal c $c] { set issub [string length $val] break } } # subfield entry if $issub { set fdsub [openIsisRec] set infosub [openIsisRec] foreach {c val} $sublst { switch $c { c {$fdsub add 861 $val} n {$fdsub add 866 $val} r {$fdsub add 863 $val} l {$fdsub add 865 [expr $val ? -$val : 0]} d {$fdsub add 869 $val} w {$infosub add 10 $val} # codes for data entry proc x {$infosub add 11 $val} # db with coded data k {$infosub add 12 $val} # prefix in coded data db a {$infosub add 20 $val} # edition attributes } } # MMM extract type from w field? $fdsub add 860 $tag 862 0 if [llength [$infosub get]] { $fdsub wrap -done -tag 870 $infosub } { $infosub done } lappend fdlst $fdsub # additional mainfield attributes } { foreach {c val} $sublst { switch $c { d {$fd add 869 $val} w {$info add 10 $val} x {$info add 11 $val} k {$info add 12 $val} a {$info add 20 $val} } } } } # mainfield entry if [llength [$info get]] { $fd wrap -done -tag 870 $info } { $info done } lappend fdlst $fd }] { puts "ERR on rec $r@$db: $rlst: $errorInfo" set ok 0 } };# for r set flen [llength $fdlst] if !$flen { puts "WARN empty fdt in $db" set ok 0 } { set fdlst [lsort -command cmpfd $fdlst] $fdt add 880 $flen for {set f 0} {$flen > $f} {incr f} { $fdt wrap -done -tag 881 [lindex $fdlst $f] } # puts [$fdt get -tags] set orec [$stb0 .req .rec] foreach {t v} [$fdt get -tag] { $orec add $t $v } if $maxout { sendrqs update $out update rowid 1 } { sendrqs insert $out insert } } $fdt done sendrqs closeIn $db close sendrqs closeOut $out close };# for db $stb0 done if !$ok { puts "terminated with warnings." exit 1 } puts ok. tcl/fullinv0100755000113000011300000000273007656454236011745 0ustar kripkrip#!/opt/isis/OpenIsis/tcl/bin/isish # vim:syntax=tcl # fully invert db using fsp # $Id: fullinv,v 1.2 2003/05/08 13:16:14 kripke Exp $ # package require openIsis 0.8 # TODO: use cmd line args set db test set v 0 set n 5 foreach arg {db v n} { catch {set $arg $env($arg)} } set isis [openIsis] ;# contact local server $isis add -syspath ../db/test -defaultdb $db $isis req -type open $isis req -type maxrow set max [$isis .res get rowid] set tot 0 $isis new ::openIsis::idx set fsp [source ../db/test/$db.fsp] proc ix {i {del 0}} { global isis db tot fsp $isis .req delete $isis req -type read -rowid $i $isis .res .rec clone rec ::openIsis::idx delete ::openIsis::idx add -1 "mfn\t$i" if $del { ::openIsis::idx add -1 del ;# delete mode } if [catch { $fsp incr tot [expr [::openIsis::idx length] / 2] $isis .req delete $isis .req wrap -tag 926 ::openIsis::idx if $::v { ::puts stderr [$isis .req serialize] } else { puts -nonewline "\r$i\t$tot"; flush stdout } $isis req -type insert }] { puts stderr "$i: $::errorInfo" } } switch -- [lindex $argv 0] \ del { # delete every nth for {set i 1} {$i <= $max} {incr i $n} { ix $i 1 } } \ add { # add every nth for {set i 1} {$i <= $max} {incr i $n} { ix $i } } \ test { # delete and reindex every nth for {set i 1} {$i <= $max} {incr i $n} { ix $i 1 } for {set i 1} {$i <= $max} {incr i $n} { ix $i } } \ default { # full inv for {set i 1} {$i <= $max} {incr i} { ix $i } } # puts "\n$tot entries" tcl/isish.c0100644000113000011300000000303107664637050011612 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: isish.c,v 1.5 2003/05/26 19:17:48 mawag Exp $ OpenIsis Tcl shell */ #include "openisis.h" #include "openisistcl.h" #ifdef ORATCL extern int Oratcl_Init ( Tcl_Interp *ip ); #endif #ifdef PGTCL extern int Pgtcl_Init ( Tcl_Interp *ip ); #endif static int appinit ( Tcl_Interp *interp ) { if ( TCL_OK != Tcl_Init( interp ) || TCL_OK != openIsisTclInit ( interp ) #ifdef ORATCL || TCL_OK != Oratcl_Init( interp ) #endif #ifdef PGTCL || TCL_OK != Pgtcl_Init( interp ) #endif ) return TCL_ERROR; return TCL_OK; } int main ( int argc, char **argv ) { openIsisCOpen(0); Tcl_Main( argc, argv, appinit ); return 0; } tcl/lib/mlcm.db.tcl0100644000113000011300000000336307664370557013133 0ustar kripkrip package provide mlcm::db 0.1 namespace eval mlcm::db { set _body { set len [llength $args] # puts "MMM $this CMD [lindex $args 0]" # puts "MMM $this REC $myrec" # puts "MMM $this REC [$myrec get -tags]" switch -glob -- [lindex $args 0] { all { set max [$myrec get maxrow] # puts "MMM $this all $max" $myrec set num $max $myrec set act 0 $myrec delete mfn for {set j 1} {$max >= $j} {incr j} { $myrec add mfn $j } return $max } hasNext { if {2 <= $len} { set incr [lindex $args 1] } { set incr 1 } set idx [$myrec get act] set max [$myrec get num] set nidx [expr $idx + $incr] # puts "MMM $this hasNext $len $incr $idx $max" return [expr 0 < $nidx && $nidx <= $max] } name { return [$myrec get name] } next { if {2 <= $len} { set incr [lindex $args 1] } { set incr 1 } set idx [$myrec get act] set max [$myrec get num] incr idx $incr if {$idx < 1} { set idx 1 } elseif {$idx > $max} { set idx $max } $myrec set act $idx # puts "MMM $this next -> $idx" return [$myrec get mfn\[[expr $idx - 1]\]] } rec { return $myrec } stub { return [$myrec get stub] } * { return [eval $myrec $args] } } } proc new {dbn stb} { variable _body set rec [openIsisRec -fdt $mlcm::meta::fdtDb] $rec set name $dbn stub $stb $rec set config [$stb db $dbn] if [catch { $stb .req delete $stb req type maxrow db $dbn set maxrow [$stb .res get row] # puts "MMM db $dbn maxrow $maxrow" $rec set maxrow $maxrow }] { global errorInfo puts $errorInfo } proc $dbn args " set myrec $rec set this $dbn $_body " return mlcm::db::$dbn } } tcl/lib/mlcm.gui.tcl0100644000113000011300000001200007664370557013316 0ustar kripkrip package provide mlcm::gui 0.1 namespace eval mlcm::gui { variable _gid 0 proc _newwindow {} { variable _gid return ".top[incr _gid]" } proc _activate btn { set bg1 [$btn cget -background] set bg2 [$btn cget -activebackground] if ![string equal $bg1 $bg2] { $btn configure -background $bg2 after 200 $btn configure -background $bg1 $btn invoke } } proc _button args { set res [eval button $args] if [catch { set under [$res cget -underline] set key [$res cget -text] if {0 <= $under} { set key [string index $key $under] bind . "mlcm::gui::_activate $res" puts "bound $key to $res" } }] { global errorInfo puts $errorInfo } return $res } proc _header {that frm db} { set maxrow [$db get maxrow] set scrw $frm.dbhead tixScrolledWindow $scrw pack $scrw -side left -expand yes -fill both set panel [$scrw subwidget window] set lbl [label $panel.maxrow -text "[mlcm::lang::msg maxrow]: $maxrow"] set bf [_button $panel.first -command "mlcm::gui::_dbfirst $db" \ -text [mlcm::lang::msg first] -underline 0] set bp [_button $panel.prev -command "mlcm::gui::_dbprev $db" \ -text [mlcm::lang::msg prev] -underline 0] set bn [_button $panel.next -command "mlcm::gui::_dbnext $db" \ -text [mlcm::lang::msg next] -underline 0] set bl [_button $panel.last -command "mlcm::gui::_dblast $db" \ -text [mlcm::lang::msg last] -underline 0] pack $panel.maxrow $panel.first $panel.prev $panel.next \ $panel.last -side left } proc _dbfirst {db} { puts "first [$db name]" } proc _dbprev {db} { puts "prev [$db name]" } proc _dbnext {db} { puts "next [$db name]" } proc _dblast {db} { puts "last [$db name]" } proc _workarea {that frm db} { set pnw $frm.pnw tixPanedWindow $pnw -orientation horizontal pack $pnw -side top -expand yes -fill both $pnw add pane1 -size 444 $pnw add pane2 # left area: result set list set htmlw $pnw.pane1.text set scb $pnw.pane1.scby html $htmlw scrollbar $scb -command "$htmlw yview" $htmlw configure -yscrollcommand "$scb set" pack $scb -side right -expand no -fill y pack $htmlw -side right -expand yes -fill both $that add html $htmlw # right area: TBD label $pnw.pane2.lbl -text "edit me" pack $pnw.pane2.lbl -side left -expand yes -fill both } # create new gui hierarchy for db # @param parent parent window where to put the gui into # if empty, a new toplevel window is created # @param profile actual profile to use # @param db actual database to use # @return gui record proc new {parent profile db} { set that [openIsisRec -fdt $mlcm::meta::fdtGui] set top 0 set path $parent if ![string length $parent] { set parent [_newwindow] toplevel $parent set path $parent. } elseif [string equal . $parent] { set top 1 } { set path $parent. set top 2 } $that add top $parent db $db stub [$db stub] set main ${path}main if $top { destroy $main } frame $main pack $main -side top -fill both -expand yes set actmenu [$profile menu] set mbar [frame $main.menubar -bd 5] pack $mbar -side top -fill x mlcm::menu::showmenu $actmenu $mbar set fhead [frame $main.frhead] pack $fhead -side top -expand no -fill x _header $that $fhead $db set fwork [frame $main.workarea] pack $fwork -side top -expand yes -fill both _workarea $that $fwork $db if $top { wm title $parent "MLCM - [$db name]" if {2 == $top} { wm geom $parent 500x400+80+80 } } return $that } # setup global/openIsis variables for active window proc activate that { set stb [$that get stub] set openIsis::srv $stb set htmlw [$that get html] proc ::openIsis::puts msg " $htmlw parse \$msg " } # show all db entries proc all that { set db [$that get db] set htmlw [$that get html] $htmlw clear catch { set tsp [tsp list $db] set dbn [$db name] for {$db all} {[$db hasNext]} {} { set mfn [$db next] if ![openIsis::isis read -rowid $mfn -db $dbn] { puts "no record for mfn $mfn" continue } if [catch {openIsis::Tspr $tsp}] { global errorInfo puts "error in $tsp:\n$errorInfo" break } } } } proc _chktsp {name dbn path res} { upvar $res ret set ret "$path/$dbn.$name" if [file readable $ret] { return 1 } set ret "$path/$name" file readable $ret } # find tsp file # first, look for /..tsp # second, look for /.tsp # third, scan $::auto_path for ..tsp # fourth, scan $::auto_path for .tsp # fifth, raise an error # @param tsp name of tsp file without extension # @param db actual databse object # @return path to tsp file proc tsp {tsp db} { global auto_path set tsp "$tsp.tsp" set dbn [$db name] set path "$auto_path" catch { set cfg [$db get 4002] set dbp [$cfg get {803 {}}] if [string length $dbp] { set path "$dbp $auto_path" } } set res "" foreach p "$path" { if [_chktsp $tsp $dbn $p res] { return $res } } error "cannot find $tsp in $path" } } tcl/lib/mlcm.lang.tcl0100644000113000011300000000626407662161245013460 0ustar kripkrip package provide mlcm::lang 0.1 namespace eval mlcm::lang { array set _keys {} array set _dbs {} # default lang: en set _alang "" set _keydb "" set _stb "" # loads some defaults proc init {} { variable _stb set _stb [ openIsis ] $_stb add syspath /opt/openisis/db set lang "en" loadmap "" load $lang "" activate $lang } # loads a key -> value mapping for messages # @param stb db stub to be used # @param path db holding the (reverse) key value mappings # the mfn stands for the message id, tag 0 for the key proc loadmap { path } { variable _stb variable _keydb if { "" == $path } { set syspath [$_stb get .syspath ] set path "$syspath/msg/key" } $_stb req type open db $path set err [openIsis::errno $_stb] if { $err } { puts "WARNING: could not open key map db @ $path" return -1 } else { puts "opened key db on path $path" } set _keydb [ $_stb .res get db ] } # loads a message set for a specific language # @param stb db stub to be used # @param lang the language key to access the message set; # should be a two-letter language code # @param path db holding the message set; the message db # is compatible to the WinISIS style # @return 0 on success else the openisis error code proc load { lang path } { variable _stb variable _dbs # no db path given; revert to default if { "" == $path } { set syspath [$_stb get .syspath ] set path "$syspath/msg/ms$lang" } puts "using msgdb path: $path for lang $lang" $_stb req type open db $path set err [openIsis::errno $_stb] if { $err } { return $err } openIsis::prnres "open $path" $_stb set _dbs($lang) $path } # activates a given lang if previously loaded proc activate { lang } { variable _alang variable _dbs set db "" if { [ catch { set db $_dbs($lang) } ] } { puts "Language $lang not loaded" } else { set _alang "$lang" puts "Language $lang now active" } } # returns a raw msg proc rawMsg { key } { variable _stb variable _alang variable _dbs set id [ _lookup $key ] if { $id } { incr id -1 set msg [ _xlate $id ] if { "" != $msg } { return $msg } } return $key } # returns a msg w/o any appended command proc strippedMsg { key } { return [_cleanCmd [rawMsg $key ]] } # returns a completely cleaned msg proc msg { key } { return [_cleanAmp [strippedMsg $key]] } # looks up a key from the proc _lookup { key } { variable _stb variable _dbs variable _keydb $_stb req type query mode 0 key $key db $_keydb set size [$_stb .res get size] if { 1 != $size } { return 0 } set id [$_stb .res get rowid] return $id } # translates a message id to the given lang proc _xlate { id } { variable _stb variable _alang variable _dbs set mfn [ expr {$id/10+1} ] set tag [ expr {[expr {$id+1}]%10} ] $_stb req type read rowid $mfn db $_dbs($_alang) set size [$_stb .res get size] if { ! $size } { return "" } set msg [lindex [$_stb .res .rec get $tag ] 0 ] return $msg } # cleans msg from appended command proc _cleanCmd msg { regsub {\[.*\]} $msg {} msg return $msg } # cleans msg from ampersands proc _cleanAmp msg { regsub -all {\&} $msg {} msg return $msg } } tcl/lib/mlcm.menu.tcl0100644000113000011300000000655507662161245013506 0ustar kripkrip package provide mlcm::menu 0.1 namespace eval mlcm::menu { array set _all {} array set _men {} set _act 0 proc addentry {name rec} { variable _all if [catch {set menus $_all($name)}] { set menus [list] } set len [llength $menus] set ent [$rec get 3000] set idx [expr $ent / 100] for {set j $len} {$idx >= $j} {incr j} { lappend menus [list] } set lst [lindex $menus $idx] set cfg [openIsisRec {} -fdt $mlcm::meta::fdtMenu] eval $cfg add -ignore [$rec get -tags] $rec done lappend lst $cfg lset menus $idx $lst set _all($name) $menus } proc _sortentries {e1 e2} { set m1 [$e1 get 3000] set m2 [$e2 get 3000] expr $m1 - $m2 } proc doneentries {} { variable _all foreach {key lst} [array get _all] { set len [llength $lst] set slst [list] for {set j 0} {$len > $j} {incr j} { set ent [lindex $lst $j] if [llength $ent] { set ent [lsort -command _sortentries $ent] } lappend slst "$ent" } set _all($key) $slst } } proc _name2lbl {name underl} { upvar $underl idx set idx [string first & $name] if {0 <= $idx} { set res [string range $name [expr $idx + 1] end] if $idx { set n1 [string range $name 0 [expr $idx - 1]] set res $n1$res } set idx "-underline $idx" return $res } set idx "" return $name } proc _rec2btn rec { set name [_name2lbl [mlcm::lang::strippedMsg [$rec get name\[0\]]] underl] return "-text $name $underl" } proc _rec2cmd rec { set name [_name2lbl [mlcm::lang::strippedMsg [$rec get name\[0\]]] underl] set cmd [$rec get {command\[0\] {}}] if [string length $cmd] { set cmd "-command \{$cmd\}" } set act [$rec get {active\[0\] 1}] if [eval expr $act] { set act "" } { set act "-state disabled" } return "-label \"$name\" $underl $cmd $act" } proc _test {} { variable _act set _act [ expr ! $_act ] } proc showmenu {name parent} { variable _all variable _men set j 0 foreach lst $_all($name) { set len [llength $lst] if $len { set rec [lindex $lst 0] set bttn $parent.$j set subm $bttn.sub set opts [_rec2btn $rec] eval menubutton $bttn $opts \ -menu $subm -bd 2 -relief flat pack $bttn -side left menu $subm -tearoff 0 -borderwidth 2 for {set s 1} {$len > $s} {incr s} { set rec [lindex $lst $s] set opts [_rec2cmd $rec] eval $subm add command $opts } } incr j } if $j { set _men("$parent\ $name") $j } } proc notify {name parent} { variable _all variable _men if {0 == [string length $name]} { set name * } if {0 == [string length $parent]} { set parent * } foreach {key maxm} [array get _men "\"$parent\ $name\""] { set parent [eval lindex $key 0] set name [eval lindex $key 1] if ![llength [info command $parent]] { puts "mlcm::menu: removing $parent" array unset _men $key continue } set j 0 foreach lst $_all($name) { set len [llength $lst] if $len { set subm $parent.$j.sub for {set s 1} {$len > $s} {incr s} { set rec [lindex $lst $s] set act [$rec get {active\[0\] 1}] if [eval expr $act] { set stn normal } { set stn disabled } set s1 [expr $s - 1] set sto [$subm entrycget $s1 -state] if ![string equal $sto $stn] { $subm entryconfigure $s1 -state $stn } } } incr j } } } } tcl/lib/mlcm.meta.tcl0100644000113000011300000001341407664370557013472 0ustar kripkrip package require mlcm::menu package require mlcm::profile package require mlcm::util package provide mlcm::meta 0.1 namespace eval mlcm::meta { # fdt records # OPENISIS_FDID = 860, /* tag */ # OPENISIS_FDSUB = 861, /* subfield char */ # OPENISIS_FDTYPE = 862, /* field type */ # OPENISIS_FDREP = 863, /* repeatable flag */ # OPENISIS_FDLEN = 865, /* length: 0 variable, >0 maxlen, <0 fixed len */ # OPENISIS_FDNAME = 866, /* field name */ # OPENISIS_FDDESC = 867, /* field description */ # OPENISIS_FDPAT = 868, /* pattern */ # OPENISIS_FDDFLT = 869, /* default value */ # OPENISIS_FDINFO = 870, /* application specific addons */ # meta record set fdtMeta [mlcm::util::buildfdt { {860 1000 862 12 863 0 865 0 866 type 867 {type of entry}} {860 1000 862 31 863 0 865 0 866 profile 867 {profile entry}} {860 1000 862 31 863 0 865 1 866 stub 867 {stub entry}} {860 1000 862 31 863 0 865 2 866 db 867 {db entry}} {860 1000 862 31 863 0 865 3 866 lang 867 {language entry}} {860 1000 862 31 863 0 865 4 866 menu 867 {menu entry}} {860 1001 862 0 863 0 865 0 866 name 867 {id of entry}} {860 1010 862 15 863 1 865 0 866 config 867 {embedded config rec}} {860 1011 862 15 863 0 865 0 866 gui 867 {global gui opts}} {860 1020 862 0 863 0 865 0 866 stub 867 {profile: known stub}} {860 1021 862 0 863 0 865 0 866 db 867 {profile: known db}} {860 1022 862 0 863 0 865 0 866 lang 867 {profile: known lang}} {860 1023 862 0 863 0 865 0 866 menu 867 {profile: known menu}} {860 1030 862 0 863 0 865 0 866 dbname 867 {db,lang: dbid}} {860 1031 862 0 863 0 865 0 866 dbstub 867 {db,lang: stub to use}} {860 1032 862 13 863 0 865 0 866 dbopen 867 {db: open/mount flag}} }] # profile data set fdtProf [mlcm::util::buildfdt { {860 1011 862 15 863 0 865 0 866 gui 867 {global gui opts}} {860 1020 862 0 863 1 865 0 866 stub 867 {known stubs}} {860 1021 862 0 863 1 865 0 866 db 867 {known dbs}} {860 1022 862 0 863 1 865 0 866 lang 867 {known langs}} {860 1023 862 0 863 1 865 0 866 menu 867 {known menus}} {860 2020 862 0 863 0 865 0 866 astub 867 {currently active stub}} {860 2021 862 0 863 0 865 0 866 adb 867 {currently active db}} {860 2022 862 0 863 0 865 0 866 alang 867 {currently active lang}} {860 2023 862 0 863 0 865 0 866 amenu 867 {currently active menu}} }] # menu data set fdtMenu [mlcm::util::buildfdt { {860 3000 862 2 863 0 865 4 866 key 867 {menu key mmss}} {860 3001 862 0 863 0 865 0 866 name 867 {key in lang db}} {860 3002 862 0 863 0 865 0 866 command 867 {menubuttons command}} {860 3003 862 0 863 0 865 0 866 active 867 {active condition}} }] # database set fdtDb [mlcm::util::buildfdt { {860 4000 862 0 863 0 865 32 866 name 867 {db identification}} {860 4001 862 0 863 0 865 32 866 stub 867 {stub to use}} {860 4002 862 0 863 0 865 0 866 config 867 {rec cmd of db metas}} {860 4003 862 15 863 0 865 0 866 gui 867 {db specific gui opts}} {860 4010 862 2 863 0 865 32 866 maxrow 867 {maxrow}} {860 4011 862 2 863 0 865 32 866 num 867 {length of result set}} {860 4012 862 2 863 0 865 32 866 act 867 {act index in result set}} {860 4013 862 2 863 1 865 32 866 mfn 867 {result set}} }] # gui layout data set fdtGui [mlcm::util::buildfdt { {860 5000 862 0 863 0 865 0 866 top 867 {frame/toplevel of gui}} {860 5001 862 0 863 0 865 0 866 db 867 {actual db}} {860 5002 862 0 863 0 865 0 866 stub 867 {stub to use}} {860 5003 862 0 863 0 865 0 866 html 867 {html widget for list.tsp}} }] # default config proc defaults {} { variable fdtMeta return [mlcm::util::buildlst { {type profile name "" menu dfltmenu lang en lang de gui { 5000 24 } } {type menu name dfltmenu config { 3000 0 3001 men/&db }} {type menu name dfltmenu config { 3000 1 3001 cmd/db/&open 3002 { puts "cmd open ..." catch {::open_db} } }} {type menu name dfltmenu config { 3000 2 3001 cmd/db/&close 3002 {puts "cmd close ..."} 3003 0 }} {type menu name dfltmenu config { 3000 3 3001 cmd/db/&exit 3002 exit }} {type menu name dfltmenu config { 3000 100 3001 men/search }} {type menu name dfltmenu config { 3000 102 3001 cmd/search 3002 {puts "cmd search ..."} 3003 { $mlcm::menu::_act } }} {type menu name dfltmenu config { 3000 103 3001 cmd/search/hist 3002 {puts "cmd search hist ..."} 3003 { $mlcm::menu::_act } }} {type menu name dfltmenu config { 3000 104 3001 cmd/search/result 3002 {puts "cmd search result ..."} 3003 { $mlcm::menu::_act } }} {type menu name dfltmenu config { 3000 200 3001 men/edit }} {type menu name dfltmenu config { 3000 201 3001 cmd/edit 3002 {puts "cmd edit ..."} }} {type menu name dfltmenu config { 3000 300 3001 men/cfg }} {type menu name dfltmenu config { 3000 301 3001 cmd/cfg/lang 3002 {puts "cmd lang ..."} }} {type menu name dfltmenu config { 3000 302 3001 cmd/cfg/toggle 3002 { puts "toggle ..." mlcm::menu::_test mlcm::menu::notify "" "" } }} {type menu name dfltmenu config { 3000 900 3001 men/help }} {type menu name dfltmenu config { 3000 901 3001 cmd/help/index 3002 {puts "cmd lang ..."} }} } {config gui} $fdtMeta] } # add config rec proc addconfig rec { switch -- [$rec get type] { 0 { puts "add prof [$rec get -tagn]" mlcm::profile::addentry [lindex [$rec get {name ""}] 0] $rec } 1 { puts "add stub [$rec get -tagn]" } 2 { puts "add db [$rec get -tagn]" } 3 { puts "add lang [$rec get -tagn]" } 4 { puts "add menu [$rec get -tagn]" mlcm::menu::addentry [lindex [$rec get {name ""}] 0] \ [$rec .config] } } } } tcl/lib/mlcm.prof.tcl0100644000113000011300000000270507663442011013473 0ustar kripkrip package provide mlcm::profile 0.1 namespace eval mlcm::profile { array set _all {} # gets an "active" variable proc _getA { rec var } { if [catch {set val [$rec get "a$var"]}] { set val [$rec get $var] if [llength $val] { set val [lindex $val 0] $rec set "a$var" $val return $val } error "have no $var" } return $val } # sets an "active" variable proc _setA { rec var val } { if {0 > [lsearch [$rec get $var] "a$var"]} { error "unknown $var $val" } $rec set "a$var" $val return $val } set _body { set len [llength $args] switch -regexp -- [lindex $args 0] { db { if {2 > $len} { return [$rec get {adb {}}] } set adb [lindex $args 1] if {0 > [lsearch [$rec get db] $adb]} { error "unknown db $adb" } $rec set adb $adb # set astub return $adb } stub { return [$rec get {astub {}}] } menu|lang { if {2 > $len} { return [_getA $rec [lindex $args 0]] } return [_setA $rec [lindex $args 0] [lindex $args 1]] } .* { return [eval $rec $args] } } } proc addentry {name rec} { variable _all if [catch {set cfg $_all($name)}] { set _all($name) [set cfg [ openIsisRec {} -fdt $mlcm::meta::fdtProf]] } mlcm::util::deepcopy $rec $cfg return $cfg } proc activate name { variable _all variable _body set rec $_all($name) set cmd p0 proc $cmd args " variable _all set rec $rec $_body " return mlcm::profile::$cmd } } tcl/lib/mlcm.util.tcl0100644000113000011300000000571507663442011013506 0ustar kripkrip package provide mlcm::util 0.1 namespace eval mlcm::util { # construct fdt from list proc buildfdt lst { set fdt [openIsisRec] set num 0 foreach ent "$lst" { set fd [eval openIsisRec \{\} set $ent] $fdt wrap -tag 881 -done $fd incr num } $fdt set 880 $num return $fdt } # build list of records proc buildlst {lst {emb {}} {fdt {}}} { if [string length $fdt] { set fdt "-fdt $fdt" } set res [list] foreach ent "$lst" { set rec [eval openIsisRec \{\} $fdt] foreach {tag val} "$ent" { if {0 <= [lsearch $emb $tag]} { set rec2 [eval openIsisRec \{\} add $val] $rec wrap -done $tag $rec2 } { $rec add -ignore $tag $val } } lappend res $rec } return $res } # split recfield into list of subfield's char and value proc splitfld fld { set lst [list] set c "" while {0 <= [set e [string first ^ $fld]]} { set val [string range $fld 0 [expr $e - 1]] set nc [string index $fld [incr e]] if ![string length $nc] { break } if {[string length $c] || [string length $val]} { lappend lst "$c" "$val" } set c $nc set fld [string range $fld [incr e] end] } lappend lst "$c" "$fld" return $lst } proc _rec2subs rec { set fdt [$rec fdt] set num [$fdt get 880] set res {} for {set j 0} {$num > $j} {incr j} { set fd [$fdt .881\[$j\]] if {15 == [$fd get 862]} { lappend res [$fd get 860] } $fd done } $fdt done return $res } proc _tgttag {tag rel} { foreach {s t} $rel { if {$s == $tag} { return $t } } return -1 } # make a deep record copy # @param src command of source record # @param tgt target record, optional # @param fdt fdt of target record, optional # @param rel relation between source and target record tags # if empty, source tags are taken over unmodified # @return target record proc deepcopy {src {tgt {}} {fdt {}} {rel {}}} { if [catch {$tgt row}] { if [string length $fdt] { set tgt [openIsisRec -fdt $fdt] } { set tgt [openIsisRec] } } set hasr [llength $rel] set hast 0 if [catch { set subrecs [_rec2subs $src] set subend -1 foreach {t v} [$src get -tags] { if {0 <= $subend} { if $hast { $emb add $t $v } if {$subend == [incr subcnt]} { if $hast { $tgt wrap -tag $subtag -done $emb } set subend -1 } continue } if {0 <= [lsearch $subrecs $t]} { set subend $v set subcnt 0 if $hasr { set subtag [_tgttag $t "$rel"] } { set subtag $t } set hast [expr 0 <= $subtag] if $hast { set emb [openIsisRec] } continue } if $hasr { set tt [_tgttag $t "$rel"] } { set tt $t } if {0 <= $tt} { $tgt add -ignore $tt $v } } }] { global errorInfo puts $errorInfo } if $hast { if ![catch { $emb done }] { set lst "" catch { set lst [$src get -tags] } puts "deepcopy: corrupted source: $lst" } } return $tgt } } tcl/list.tsp0100644000113000011300000000034407656442536012042 0ustar kripkrip<% global idx set mfn [v rowid] ;# get the mfn # use first field, if we don't have CDS' famous 24 if [catch {v get 24} list] { set list [v get] } %> <%=mfn%>: <%=MHL [lindex $list 0]%>
tcl/lp0100755000113000011300000001227407624143521010670 0ustar kripkrip#!/opt/openisis/tcl/tkbuild/wish # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: lp,v 1.6 2003/02/17 11:27:23 kripke Exp $ # # lowerpoint :) # # usage: lp [db [encoding [ptSz]]] # default encoding is cp850 ("DOS Latin1"), you may want iso8859-1 or utf8 # default font point size is 240, try 180 # # consult documentation of your window manager # on how to tell it to put no frames on this window # openIsisInit set dbname [expr {$argc ? [lindex $argv 0] : {db/cds/cds}}] set enc [expr {1 < $argc ? [lindex $argv 1] : {cp850}}] set pt [expr {2 < $argc ? [lindex $argv 2] : {240}}] set db [openIsisOpen $dbname] set max [$db maxrowid] set cur 0 # puts "$db $dbname $max" # configure look set sw [winfo screenwidth .] set sh [winfo screenheight .] set tw [expr $sw - 400] set lh [expr $pt / 6] ;# line height set fn "-*-new century schoolbook-medium-r-*-*-*-$pt-75-75-*-*-iso10646-1" set dot img/isi.gif set logo img/logo.gif set cursor {@img/isi.cursor black} proc show n { global db cur max lh tw fn enc switch -- $n \ + { if {$max == $cur} return; incr cur } - { if {1 == $cur} return; incr cur -1 } default { if {$cur == $n} return; set cur $n } .c delete t set y 100 set rec [$db read $cur] if {"utf-8" != $enc} {set rec [encoding convertfrom $enc $rec]} set f [lindex $rec 0] regexp {([^\t]*)(\t(.*))?} [lindex $f 1] - text - next set it [.c create text 150 $y \ -anchor nw -font $fn -width $tw -tags t -text $text] if {"" != $next} { set ny [expr [lindex [.c bbox $it] 3] + $lh/4] set it [.c create text 210 $ny \ -anchor nw -font $fn -width $tw -tags t -fill darkgreen -text $next] } set ny [lindex [.c bbox t] 3] .c lower [.c create rectangle 120 [expr $y-2] [expr 150+$tw] $ny \ -tags t -outline lightgrey -fill lightgrey] set y [expr $ny + 2*$lh] foreach f [lrange $rec 1 end] { regexp {([^\t]*)(\t(.*))?} [lindex $f 1] - text - next if [string match *.gif $text] { regexp {(.*/)?(.*)} $text - - text if ![catch {image create photo $text -file img/$text}] { set it [.c create image 150 $y -anchor nw -tags t -image $text] set y [expr [lindex [.c bbox $it] 3] + $lh] continue } } .c create image 120 $y -anchor nw -tags t -image dot set it [.c create text 150 $y \ -anchor nw -font $fn -width $tw -tags t -text $text] if {"" != $next} { set y [expr [lindex [.c bbox $it] 3] + $lh/4] set it [.c create text 210 $y \ -anchor nw -font $fn -width $tw -tags t -fill darkgreen -text $next] } set y [expr [lindex [.c bbox $it] 3] + $lh] } .c itemconfigure counter -text "http://openisis.org/ $cur / $max" };# show proc print {} { # A4dj 595 842 24 50 571 818 # interior on Deskjet landscape is 768*547 # to keep 4x3 ratio use 728x546 .c postscript -rotate 1 -pagewidth 728 -file /tmp/lp.ps };# print proc printall {} { global cur max set f [open /tmp/lp.ps w] set o $cur for {set i 1} {$i <= $max} {incr i} { show $i .c postscript -rotate 1 -pagewidth 728 -channel $f } close $f show $o };# printall # create widgets canvas .c -width $sw -height $sh -bg white catch {.c configure -cursor $cursor} bind .c exit bind .c q exit bind .c - {show -} bind .c b {show -} bind .c p print bind .c P printall bind .c {show -} bind .c {show -} bind .c {show 1} bind .c "show $max" bind .c {show +} bind .c {show +} if [catch {image create photo logo -file $logo} err] { puts stderr "error loading logo '$logo':\n\t$err" } else { .c create image [expr $sw - 150] 150 -tags logo -image logo } if [catch {image create photo dot -file $dot} err] { # builtin image create bitmap dot -data { #define isi_width 16 #define isi_height 16 #define isi_x_hot 1 #define isi_y_hot 5 static unsigned char isi_bits[] = { 0x00, 0x00, 0x00, 0x18, 0x80, 0x25, 0xc0, 0x27, 0xe0, 0x47, 0xfe, 0x27, 0xe4, 0x27, 0xc4, 0x23, 0x08, 0x10, 0x10, 0x08, 0xe0, 0x07, 0xc0, 0x03, 0xc0, 0x03, 0xc0, 0x03, 0xe0, 0x07, 0x00, 0x00}; } } .c create text [expr $sw - 100] [expr $sh - 50] -tags counter -text 0 show 1 pack .c -expand 1 -fill both focus .c wm geometry . ${sw}x${sh}+0+0 # if [wm frame .] ... # puts stderr "for real fullscreen experience tell your wm to not decorate me" tcl/lp2db0100755000113000011300000000106707624143521011256 0ustar kripkrip#!/usr/bin/perl # convert a bilingual MikiDoc text # to a database suitable for lowerpoint[tm] presentation # using tcl/lp # # this is actually a little bit useful ... :) # usage: # cat doc/Cuba*.txt | tcl/lp2db | ./openisis -write db/pres/Cuba -stream $c = 0; $odd = 0; while (<>) { chomp; /^\*\t(.*)/ && do { $odd = !$odd; if ( $odd ) { push @a, $1; } else { print "\f\n" if $c++; print "100\t$1\t", shift(@a), "\n"; } next; }; /^-\t(.*)/ && do { if ( $odd ) { push @a, $1; } else { print "200\t$1\t", shift(@a), "\n"; } }; } tcl/lpdb0100775000113000011300000000062507615315677011212 0ustar kripkrip#!/usr/bin/perl # convert a MikiDoc text to a databse suitable for lowerpoint[tm] presentation # using tcl/lp # # this is not actually useful, but serves as a demo # usage: # tcl/lpdb ) { chomp; /^\*\t(.*)/ && do { print "\f\n" if $c++; print "100\t$1\n"; next; }; /^-\t(.*)/ && do { print "200\t$1\n"; }; } tcl/maketcltk0100755000113000011300000001166507653266141012245 0ustar kripkrip#!/bin/bash # $Id: maketcltk,v 1.14 2003/04/28 16:03:43 kripke Exp $ # # sample how to build Tcl/Tk # we can build shared or static versions, # however, --disable-threads seems to break some tests ... # # recommended wish build: # maketcltk -thr 8.3.5 tclx8.3.5 tix-8.1.4 blt2.4z tcllib-1.3 tclxml-2.6 # maketcltk -mingw 8.3.5 # make -C tkhtml # other examples # maketcltk -thr 8.4.2 oratcl-4-1-branch # maketcltk -thr 8.3.5 oratcl3.3 # maketcltk 8.3.4 oratcl3.3 # # for dir in tcl tclx tix blt incrtcl oratcl; do # http://belnet.dl.sourceforge.net/sourceforge/$dir # done unset TCLLIBPATH CC=gcc; export CC calledas="$0 $*" base=`pwd`/ case "$1" in -base) base="$2"; shift 2;; esac mingw=false static=false case "$1" in -mingw) mingw=true; shift;; -a) static=true; shift;; *) [ -n "$WIN32" ] && mingw=true;; esac tclsubdir=unix other=win w= CFG=--enable-gcc if $mingw; then static=true tclsubdir=win other=unix w=.w if [ -d $HOME/mingw/bin ]; then PATH=$HOME/mingw/bin:$HOME/mingw/i386-mingw32msvc/bin:$PATH elif [ -d /opt/cross-tools/bin ]; then PATH=/opt/cross-tools/bin:/opt/cross-tools/i386-mingw32msvc/bin:$PATH else echo "could not find mingw installation. suggest you link $HOME/mingw to it">&2 exit 1 fi WIN32=yes export PATH WIN32 CFG="$CFG --cache-file=config.cross.cache --host=i386-mingw32msvc" CFG="$CFG --target=i386-mingw32msvc --build=i386-linux" fi case "$remake" in true);; *) remake=false;; esac thr=--disable-threads t= case "$1" in -thr) thr=--enable-threads; t=.t; shift;; esac ver=${1:-8.3.5} for arg; do case "$arg" in tclxml*) tclxml="$arg";; tclx*) tclx="$arg";; oratcl*) oratcl="$arg";; tix*) tix="$arg";; blt*) blt="$arg";; tcllib*) tcllib="$arg";; esac done makepkg () { echo "### makepkg $*" pfx=$1 pkg=$2 cfg=$3 shift 3 ( $remake && ( cd $pkg/$other; make clean distclean; ) cd $pkg/$tclsubdir $remake && rm config.{cache,log,status} $remake && make distclean echo CC=gcc $cfg $CFG --prefix=$pfx $thr "$@" [ -f config.status ] || $cfg $CFG --prefix=$pfx $thr "$@" $remake && make clean make install );# popd } maketix () { echo "### maketix $*" pfx=$1 pkg=$2 shift 2 with="--with-tclconfig=$dir/lib --with-tkconfig=$dir/lib" # should be able to figure it out # with="$with --with-tclinclude=$dir/include --with-tkinclude=$dir/include" ( $remake && ( cd $pkg/$other; make clean distclean; ) cd $pkg/$tclsubdir $remake && make distclean echo CC=gcc ./configure --prefix=$pfx $thr $with "$@" [ -f config.status ] || ./configure --prefix=$pfx $thr $with "$@" # figure out subdir case "$pfx" in *8.4*) subdir=tk8.4;; *) subdir=tk8.3;; esac echo "# tix subdir $subdir" cd $subdir $remake && make distclean echo CC=gcc ./configure $CFG --prefix=$pfx $thr $with "$@" ./configure $CFG --prefix=$pfx $thr $with "$@" make install cd .. make install );# popd } makeblt () { echo "### makeblt $*" pfx=$1 pkg=$2 shift 2 ( cd $pkg $remake && rm config.{cache,log,status} # make distclean echo CC=gcc ./configure $CFG --prefix=$pfx $thr "$@" [ -f config.status ] || ./configure $CFG --prefix=$pfx $thr "$@" $remake && make clean make ;# important to make first, else bltsh is missing! make install );# popd } makexml () { echo "### makexml $*" pfx=$1 pkg=$2 shift 2 ( cd $pkg $remake && rm config.{cache,log,status} $remake && make distclean echo CC=gcc ./configure $CFG --prefix=$pfx $thr "$@" [ -f config.status ] || ./configure $CFG --prefix=$pfx $thr "$@" $remake && make clean make install );# popd } makeall () { sfx=$1 shr=$2 dir=$base$ver$sfx mkdir $dir echo "### $calledas" if [ -d "tk$ver" ]; then mktk=true usetk="--enable-tk --with-tk=$dir/lib" else mktk=false usetk="--disable-tk --without-tk" fi echo tail -f $dir/makeall.log $remake && [ -f $dir/makeall.log ] && mv $dir/makeall.log $dir/makeall.log.old exec >>$dir/makeall.log 2>&1 echo "### $calledas" makepkg $dir tcl$ver ./configure $shr $mktk && makepkg $dir tk$ver ./configure $shr --with-tcl=$dir/lib [ -n "$tclx" ] && makepkg $dir $tclx ./configure $shr \ --with-tcl=$dir/lib $usetk [ -n "$oratcl" ] && makepkg $dir $oratcl ../configure $shr --with-tcl=$dir/lib if $mktk; then [ -n "$tkhtml" ] && maketkhtml $dir $tkhtml [ -n "$tix" ] && maketix $dir "$tix" $shr [ -n "$blt" ] && makeblt $dir "$blt" $shr \ --with-tcl=$dir/lib --with-tk=$dir/lib fi if [ -n "$tcllib" ]; then echo "### tcllib $tcllib" (cd $tcllib && ./install.sh $dir) fi if [ -n "$tclxml" ]; then ver=${tclxml#tclxml} makexml $dir "$tclxml" $shr --with-tcl=$dir/lib [ -d "tcldom-$ver" ] \ && makexml $dir "tcldom-$ver" $shr --with-tcl=$dir/lib if false; then # requires gnome libxslt, libxml2 cat $dir/lib/TcldomConfig.sh $dir/lib/TclxmlConfig.sh \ >$dir/lib/tcldomxmlConfig.sh [ -d "tclxslt-$ver" ] \ && makexml $dir "tclxslt-$ver" $shr --with-tcl=$dir/lib fi fi } if $static; then # only used for building binary dist ... makeall $w$t.a --disable-shared else makeall $w$t.so --enable-shared fi tcl/md0100775000113000011300000001230207650525036010653 0ustar kripkrip#!/opt/bin/tspr <%-- test Tcl server page $Id: md,v 1.2 2003/04/20 14:09:34 kripke Exp $ --%><%! package require mime; namespace import ::mime::*; # return list file from date subject size content-type proc mscan file { array set a {from {} date {} subject {} content-type {}} set fid [open $file] # for quick scan, ignore continuation lines and repeated headers while {[gets $fid line]} { if [regexp {^([\w\-]+):\s*(.*)} $line - key val] { set a([string tolower $key]) $val } } close $fid list $file $a(from) $a(date) $a(subject) [file size $file] $a(content-type) } # return list file from date subject size encoding content params parts msg proc mparse file { set l [list $file] set msg [initialize -file $file] # mime want's whole messages # headers foreach key {From Date Subject} { if [catch {getheader $msg $key} value] { lappend l {} } else { lappend l $value } } # properties (parsed content-* headers) foreach key {size encoding content params parts} { if [catch {getproperty $msg $key} value] { lappend l {} } else { lappend l $value } } lappend l $msg return $l } %><% # request standard tags # -10 SCRIPT_NAME # -11 PATH_INFO # -12 REMOTE_USER # # request parameters # 1 move destination (checkbox) # 2 the ok button # 3 write to # 4 write subject # 5 write body # global env # foreach e [array names env] { ::puts stderr "$e\t$env($e)" } ::puts stderr [request serialize] set script [lindex [request get -10] 0] set path [lindex [request get -11] 0] set user [lindex [request get -12] 0] set body [lindex [request get 5] 0] if {"" == $user} return set l [string length $user] set fid [open /var/qmail/users/popasswd] while {[gets $fid line]} { if ![string match "$user:*" $line] continue set upignh [split $line :] set name [lindex $upignh 4] set base [lindex $upignh 5]/Maildir break } close $fid cd $base ::puts stderr [pwd] if ![regexp {^/(\w+)(/\d+[\w.:,]*)?$} $path - dir file] { set dir new set file "" set rel $script/ } elseif {"" == $file} { set rel "" } else { set rel ../ } if ![file isdirectory $dir] return # scan for dirs set dirs {} ;# dirs to list set mvdirs {} ;# dirs to move to foreach d [glob ???] { if ![file isdirectory $d] continue switch -- $d tmp {} default { lappend dirs $d } switch -- $d new - tmp - uns - out - $dir {} default { lappend mvdirs $d } } if {"" != $body && [catch { # send set to [lindex [request get 3] 0] set subject [lindex [request get 4] 0] if {"" == $to} { # get from message set msg [initialize -file $dir$file] if [catch {getheader $msg Reply-To} to] { set to [getheader $msg From] } set to [lindex $to 0] if {"" == $subject} { set subject "Re: [lindex [getheader $msg Subject] 0]" } finalize $msg } set tim [clock seconds] set gtf [clock format $tim -format %Y%m%d%H%M%S -gmt 1] set rfc [clock format $tim -format {%a, %d %b %Y %H:%M%S %Z}] set fid [open uns/$gtf.[pid] w] set msg [initialize -canonical text/plain \ -header [list Return-Path $name] \ -header [list From $name] \ -header [list To $to] \ -header [list Date $rfc] \ -header [list Subject $subject] \ -encoding quoted-printable \ -string $body ] copymessage $msg $fid close $fid }]} { global errorInfo ::puts stderr $errorInfo } %> <%"name%>/<%=dir%> list: <% foreach d $dirs { %> <%=d%> <% } %> <%-- main display --%>
<% if {"" == $file} { %> <% foreach mv [request get 1] { ::puts stderr $mv if [regexp {^(\w\w\w)/(\S*)$} $mv - dest file] { file rename $dir/$file $dest %> <% } } cd $dir foreach f [glob *] { # file from date subject size content-type set l [mscan $f] %> <% } } else { # file view set msg [initialize -file $dir$file] foreach h [getheader $msg -names] { foreach v [getheader $msg $h] { %> <% } } set tpart "" if {"text/plain" == [getproperty $msg content]} { set tpart $msg } elseif {![catch {set parts [getproperty $msg parts]}]} { foreach p $parts { set ct [getproperty $p content] %> <% if {"text/plain" == $ct} { set tpart $p } } } if {"" != $tpart} { %> <% } } %>
<% foreach d $mvdirs { %><%=d%> <% } %> From Date Size Subject
moved <%"file%> to <%"dest%>
<% foreach d $mvdirs { %> <% } %> <%"lindex $l 1%> <%"lindex $l 2%> <%"lindex $l 4%> <%"lindex $l 3%>
<%"h%> <%"v%>
part <%"ct%> 
	<%"getbody $tpart%>
<%-- send form --%>
to
subject
tcl/mf0100775000113000011300000002007307660461712010662 0ustar kripkrip#!/opt/isis/tcl/bin/tcl # vim:syntax=tcl # mail fix # split a message into parts # create a new message from selected headers and textified parts # keep original non-text parts in files # qualify cleaned message by dbacl # in mime.tcl, BE SURE to remove the comment from the line # error "termination string missing in $state(content)" # cf. BUG 631314 "infinite loop on bad data" # # unterminated multipart is very frequent in SPAM, # be VERY rare in non-automated mail # when not erroring out, mime.tcl may hang package require mime; namespace import ::mime::*; proc mail2txt file { exec sed -e {1,/^$/d} <$file >$file.txt } proc html2txt file { exec lynx -dump -force_html -nolog $file >$file.txt } proc winl2txt file { # stupid M$ "Cp1252" exec tr {\200-\237} {E?,f".+#^%S<Ö?Z??''''""*--"Ts>ö?zY} <$file \ | fmt -s >$file.txt } proc win2txt file { exec tr {\200-\237} {E?,f".+#^%S<Ö?Z??''''""*--"Ts>ö?zY} <$file >$file.txt } proc long2txt file { exec fmt -s <$file >$file.txt } proc txt2txt file { file rename $file $file.txt } proc pdf2txt file { exec pdftotext -raw $file - >$file.txt } proc rtf2txt file { exec unrtf --nopict --html $file 2>/dev/null \ | lynx -dump -force_html -nolist /proc/self/fd/0 >$file.txt } proc doc2txt file { exec antiword $file >$file.txt } # fix a part in file # - figure out real file type # - figure out an extension so that webserver will deliver with proper ctype # - in any case, rename or convert file to that extension # - for some "text" types, also create additional plain text version proc fixpart {file {ctype ""}} { # figure out file type set ftype "" # shouldn't harm to believe image switch -glob -- $ctype \ image/*gif { set ftype "GIF image" } \ image/*jpeg { set ftype "JPEG image" } \ image/*png { set ftype "PNG image" } if {"" == $ftype} { set ftype [exec file -b $file] } # figure out extension and fix from file type set ext txt set fix "" set nofix 0 switch -glob -- $ftype \ "*mail text*" { set fix [set ext mail]2txt } \ *HTML* { # maybe ASCII or ISO-8859 HTML document text, so check before ASCII set fix [set ext html]2txt } \ "*Non-ISO extended-ASCII*long lines" { # assume windows stuff set fix winl2txt } \ "*Non-ISO extended-ASCII*" { set fix win2txt } \ "ASCII*text*long lines" - "ISO-8859*text*long lines" { set fix long2txt } \ ASCII*text* - ISO-8859*text* {} \ PDF* { set fix [set ext pdf]2txt } \ "Rich Text Format*" { set fix [set ext rtf]2txt } \ "Microsoft Office Document*" { set fix [set ext doc]2txt } \ "*GIF image*" { set ext gif } \ "*JPEG image*" { set ext jpg } \ "*PC bitmap*" { set ext bmp } \ "*PNG image*" { set ext png } \ "*TIFF image*" { set ext tiff } \ Zip* { set ext zip } \ gzip* { set ext gz } \ *image* { set ext img } \ default { set ext bin } ::puts stderr "file $file; c=\"$ctype\"; f=\"$ftype;\" e=$ext; fix=$fix" # if we don't have no fix, # or can not successfully apply a fix, we stick with the original if {"" == $fix || [catch {$fix $file}]} { if {"" != $fix} { ::puts stderr "FIXPART $file:\t$::errorInfo" catch {file delete $file.txt} } file rename $file $file.$ext return $ext } # if the original is represented by a non-destructive txt fix, # delete it and keep only the fix if {"txt" == $ext} { file delete $file return txt } # we have a successfull non-equivalence fix -- keep both file rename $file $file.$ext list txt $ext } ;# fixpart # operate on array env # file => base filename # attc => saved attachments (list of extensions and original ctypes) # part => list of plaintext part tokens proc saveparts {envname msg} { upvar $envname env if [catch {getproperty $msg content} ctype] { set ctype "text/plain" } # multiparts recurse if ![catch {getproperty $msg parts} parts] { if {"multipart/alternative" == $ctype} { # select one set select [lindex $parts 0] foreach p $parts { if {![catch {getproperty $msg ctype} pc] && "text/plain" == $pc } { set select $p break } } set parts [list $select] } # TODO: also care for mp/apple-double foreach p $parts { saveparts env $p } return } # a leave is not plain or too long: save as $file.i, i=0,1.. set file $env(file).[expr [llength $env(attc)]/2] set fid [open $file w] fconfigure $fid -translation binary puts -nonewline $fid [getbody $msg] close $fid # check and fix the file, get primary and secondary extension foreach {ext sec} [fixpart $file $ctype] break # optionally create new part if {"txt" == $ext} { # create new text/plain part set fid [open $file.txt] fconfigure $fid -translation binary set body [read $fid 8000] close $fid set complete [expr 8000 >= [file size $file.txt]] set charset ISO-8859-1 if {"text/plain" == $ctype} { foreach {key val} [getproperty $msg params] { if {"charset" == $key} { set charset $val } } } lappend env(part) [initialize -canonical text/plain \ -param [list charset $charset] \ -encoding quoted-printable -string $body] if {$complete || "" != $sec} { file delete $file.txt } if {"" != $sec} { # use that as attachment set ext $sec } elseif {$complete} { # no attachment return } } lappend env(attc) $ext $ctype } ;# saveparts proc fixmsg {file dir link} { set msg [initialize -file $dir$file] array set env [list file $dir$file part "" attc ""] saveparts env $msg ::puts stderr "$dir$file\t$env(attc)\t$env(part)" if {"" != $env(attc)} { # create referring part set i -1 if ![file isdirectory .attc] { file mkdir .attc } set body "" foreach {ext ctype} $env(attc) { set f $file.[incr i].$ext file rename $dir$f .attc append body "$link$f\t$ctype\n" } lappend env(part) [initialize -canonical text/plain \ -param {charset US-ASCII} -encoding 7bit -string $body] } if {1 == [llength $env(part)]} { set nmsg [lindex $env(part) 0] } else { set nmsg [initialize -canonical multipart/mixed -parts $env(part)] } foreach {key vallist} [getheader $msg] { foreach val $vallist { # ::puts stderr "$key $val" switch -- [string tolower $key] \ subject - date - from - to - cc - delivered-to - \ message-id - references - in-reply-to { setheader $nmsg $key $val -mode append } } } if ![file isdirectory .orig] { file mkdir .orig } file rename $dir$file .orig finalize $msg set fid [open $dir$file w] copymessage $nmsg $fid close $fid finalize $nmsg } ;# fixmsg proc main {argv} { foreach {file dir base link} $argv break if {"" != $base} { cd $base } if {"" != $dir} { set dir $dir/ } elseif {[file isdirectory tmp]} { set dir tmp/ } if {"-" == $file || "" == $file} { set file [clock format [clock seconds] -format %Y%m%d%H%M%S -gmt 1] set try 0 set base $file while {[catch {open $dir$file {WRONLY CREAT EXCL}} fid]} { if {99 == $try} { error "too many tries on $dir$file" } set file $base-[incr try] } while {![eof stdin]} { puts -nonewline $fid [read stdin 8192] } close $fid } set spam "" if [catch {fixmsg $file $dir $link} err] { # some errors have 99.99% SPAM probability switch -glob -- $err \ "termination string missing in multipart*" - \ "multiple Content-Type fields*" { set spam "FIXMSG $err" } ::puts stderr "FIXMSG $file:\t$::errorInfo" } if ![file isdirectory .spam] { file mkdir .spam } if {"" == $spam} { catch {exec dbacl -c .cat/ok -c .cat/spam $dir$file} if {2 != [scan $::errorCode "CHILDSTATUS %d %d" pid code]} { ::puts stderr "OOPS! could not scan '$::errorCode'" } elseif {2 == $code} { set spam "DBACL" } } if {"" != $spam} { ::puts stderr "FIXSPAM\t$file\t$spam" file rename $dir$file .spam # collect some headers even if message parsing failed # for quick scan, ignore continuation lines set headers "$file\t$spam\n" set fid [open .spam/$file] while {[gets $fid line]} { if [regexp {^([\w\-]+):\s*(.*)} $line - key val] { switch -- [string tolower $key] \ from - to - subject - delivered-to { append headers $line "\n" } } } close $fid set fid [open .spam/.log {WRONLY APPEND CREAT}] puts $fid $headers ;# let puts append a newline close $fid return } if [file isdirectory new] { file rename $dir$file new } } main $argv tcl/mlcm0100755000113000011300000000415307664370557011221 0ustar kripkrip#!/opt/openisis/tcl/bin/wisish package require Tcl 8.3 package require openIsis 0.8 # package require mlcm::menu package require mlcm::meta # package require mlcm::profile # package require mlcm::util package require mlcm::lang package require mlcm::db package require mlcm::gui set cfglst [mlcm::meta::defaults] foreach cfg $cfglst { mlcm::meta::addconfig $cfg } mlcm::menu::doneentries set actprof [mlcm::profile::activate {}] mlcm::lang::init mlcm::lang::load [$actprof lang] "" mlcm::lang::activate [$actprof lang] set lang [mlcm::lang::msg "lang" ] set stb0 [openIsis syspath /opt/openisis/db] proc open_file {fname} { global stb0 actprof global db0 ;# MMM puts "open_file: $fname" set idx [string last / $fname] $stb0 .req delete if {0 < $idx} { set dbpath [string range $fname 0 [expr $idx - 1]] $stb0 .req .config add dbpath $dbpath set fname [string range $fname [expr $idx + 1] end] } set idx [string last . $fname] if {0 < $idx} { set dbn [string range $fname 0 [expr $idx - 1]] } { set dbn $fname } $stb0 req type open db $dbn set dbn [$stb0 .res get db] set db0 [mlcm::db::new $dbn $stb0] set gui [mlcm::gui::new . $actprof $db0] mlcm::gui::activate $gui mlcm::gui::all $gui } proc open_db {} { global stb0 set dialog [tix filedialog tixFileSelectDialog] $dialog config -command "open_file" set fsbox [$dialog subwidget fsbox] $fsbox configure -directory [$stb0 get syspath] -pattern *.mst $dialog popup } open_file /opt/openisis/db/cds/cds.mst if 0 { set xxx [openIsisRec -fdt $mlcm::meta::fdtDb] set ggg [openIsisRec] $ggg add 111 eins 222 zwei $xxx add 4000 myname $xxx wrap -tag 4003 -done $ggg $xxx add 4013 4646 set yyy [mlcm::util::deepcopy $xxx {} $mlcm::meta::fdtProf {4003 1011 4013 2023}] puts "MMM >>> [$yyy get -tags]" puts "MMM >>> [$yyy .1011 get -tags]" set yyy [mlcm::util::deepcopy $xxx {} $mlcm::meta::fdtDb] set yyy [mlcm::util::deepcopy $xxx $yyy] puts "MMM >>> [$yyy get -tags]" puts "MMM >>> [$yyy .4003 get -tags]" puts "MMM [$actprof get -tags]" set guirec [$actprof .1011] puts "MMM [$guirec get -tags]" $guirec done puts "MMM list -> [mlcm::gui::tsp list $db0]" } tcl/oits.c0100644000113000011300000001246407664637050011463 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: oits.c,v 1.11 2003/05/26 19:17:48 mawag Exp $ OpenIsis Tcl server */ #include "lsv.h" #include "ldsp.h" #include "openisistcl.h" /* include this after the Tcl stuff for the benefit of those who use the 150% braindead gcc 2.96 which barfs on declaration of `index' shadows global declaration in generic/tclDecls.h cause string.h declares char *index(const char *s, int c) */ #include /* mem* */ static Tcl_Interp *interp; /* the 1 + only */ static int oitsid; static int oitrid[3]; /* the db srv application */ static int oitdbs (Con *con, int task) { int rt; switch ( task ) { case LSV_APPGOT: con->grp = 1; default: return 0; case LSV_APPRUN: break; } sMsg( LOG_INFO, "dbs %d: %d fields from %s", svCur()->id, con->req->len, con->nam); rt = ldspProcess (con->req, &(con->ses->res), 0, 0); sMsg( LOG_INFO, "dbs %d: %d fields to %s (%d)", svCur()->id, con->ses->res->len, con->nam, rt); return rt; } /* the echo application */ static int oits ( Con *c, int task ) { Field *f; int i; switch ( task ) { case LSV_APPINI: /* init thread */ case LSV_APPFIN: /* fini thread */ case LSV_APPARG: /* new request param arrived */ default: return 0; case LSV_APPGOT: /* decide wether request may go to background thread */ c->grp = 1; /* non Tcl */ for ( i=c->req->len, f = c->req->field; i--; f++ ) /* by convention, request with fields >99 are handled by Tcl */ if ( 99 < f->tag ) { c->grp = 0; break; } return 0; case LSV_APPRUN: break; /* main code below */ } /* run it */ sMsg( LOG_INFO, "echo %d: %d fields to %s", svCur()->id, c->req->len, c->nam ); /** we have records: req -- the request, allocated by the srv->ses and readonly here ses -- session props, allocated and managed by ses, read/write res -- the responses, allocated and managed by ses, read/write (the identified) ses may be a different session on each call, however, calls are strictly serialized (all in main thread) alternatively, we may use the one main session bound to the main thread as allocator and use the identified session only to keep the session props -- given they are never modified in non-main-thread calls */ switch ( c->grp ) { case 0: { /* main thread Tcl request */ Rec *rec[3]; rec[0] = c->req; rec[1] = c->ses->res; rec[2] = c->ses->prop; if ( openIsisTclEval( oitsid, 3, oitrid, rec, "oitscb" ) ) { const char *errorInfo = Tcl_GetVar( interp, "errorInfo", TCL_GLOBAL_ONLY ); sMsg( LOG_ERROR, "error in callback\n%s", errorInfo ? errorInfo : "-" ); } c->req = rec[0]; c->ses->res = rec[1]; c->ses->prop = rec[2]; /* don't break; */ } /* case 0 */ default: /* cp session to response */ if ( c->ses->prop ) for ( i=c->ses->prop->len, f = c->ses->prop->field; i--; f++ ) { RADDF( c->ses->res, f, !0 ); } break; } return 0; } /* oits */ #ifdef ORATCL extern int Oratcl_Init ( Tcl_Interp *ip ); #endif #ifdef PGTCL extern int Pgtcl_Init ( Tcl_Interp *ip ); #endif int main ( int argc, const char **argv ) { char *file; int ret, db; Srv echo; cOpen(0); Tcl_FindExecutable (argv[0]); interp = Tcl_CreateInterp(); ret = Tcl_Init( interp ); sMsg( LOG_VERBOSE, "appinit %d", ret ); db = 0; if (1 < argc && 0 == strcmp ("-db", argv[1])) { db = 1; ret = openIsisTclInit (interp); sMsg( LOG_VERBOSE, "openisisinit %d", ret); --argc; ++argv; } #ifdef ORATCL ret = Oratcl_Init( interp ); sMsg( LOG_VERBOSE, "orainit %d", ret ); #endif #ifdef PGTCL ret = Pgtcl_Init( interp ); sMsg( LOG_VERBOSE, "pginit %d", ret ); #endif oitsid = openIsisTclNewSession( interp ); sMsg( LOG_VERBOSE, "oitsid %d", oitsid ); oitrid[0] = openIsisTclCreateRecCmd( oitsid, "request", 0 ); sMsg( LOG_VERBOSE, "request %d", oitrid[0] ); oitrid[1] = openIsisTclCreateRecCmd( oitsid, "response", 0 ); sMsg( LOG_VERBOSE, "response %d", oitrid[1] ); oitrid[2] = openIsisTclCreateRecCmd( oitsid, "session", 0 ); sMsg( LOG_VERBOSE, "session %d", oitrid[2] ); file = 2= $i};# condition for having a literal while {"#" == [string index $tsp $i]} { set i [string first "\n" $tsp $i] if {0 > $i} { error "unterminated line" } incr i } while 1 { # we're in literal (outside <% %>) mode if {0 <= [set j [string first <% $tsp $i]]} { incr j -1 } else { set j end set lit {$i <= [string length $tsp]} } if $lit { append $sect "puts " [list [string range $tsp $i $j]] "\n" } if {"end" == $j} break # have <% code, j was last char before incr j 3 if {0 > [set i [string first %> $tsp $j]]} { error "unbalanced <% at pos $j '[string range $tsp $j [expr $i+80]]'..." } incr i -1 switch -- [set type [string index $tsp $j]] \ - { # comment -- actually should start -- # extend to end of comment, skipping all intervening %> if {0 > [set i [string first --%> $tsp $j]]} { error "unbalanced <%-- at pos $j ..." } incr i 4 ;# skip --%> continue } = - "\"" - ? { incr j set expr [string range $tsp $j $i] # apply some magic guessing if [regexp {^[.\-\d]} $expr] { # looks like v get args append $sect put$put($type) { [lindex [v get } $expr "] 0]\n" } elseif [regexp {^[\w\d:]+($|\()} $expr] { # looks like variable append $sect "put$put($type) \"\$" $expr "\"\n" } elseif [regexp {^\w} $expr] { # looks like command append $sect put$put($type) { [} $expr "]\n" } else { # use as is append $sect "put$put($type) \"" $expr "\"\n" } set type = } @ { error "@-section is meaningless with TSPs" } ! { set sect global if [set globlit [expr {"!" == [string index $tsp $i]}]] { incr i -1 } incr j } default { set type "" set sect main } if {"=" != $type} { append $sect [string range $tsp $j $i] "\n" } if {"!" == $type} { if $globlit { incr i 1 } else { # reset set sect main } } incr i 3 } return "$intro $global proc main {} { $main } "; } ;# openIsis::Tspc proc openIsis::puts {str} { ::puts -nonewline $str } proc openIsis::2html {str} { # substitution of very important entities string map {& & < < > > \" "} $str } proc openIsis::puth {str} { puts [2html $str] } proc openIsis::2url {str} { variable umap regsub -all {[\0-\37]} [ string map $umap $str ] {+} str set str } proc openIsis::putu {str} { puts [2url $str] } # run tspfile, compiling if necessary # proc openIsis::Tspr {tspfile} { variable ver set compile 1 set mtime [file mtime $tspfile] catch { if {[set ::tsp::${tspfile}::mtime] == $mtime} { set compile 0 } } if $compile { catch {namespace delete ::tsp::$tspfile} set f [open $tspfile] namespace eval ::tsp::$tspfile [Tspc [read $f] " variable mtime $mtime namespace import ::openIsis::* "] close $f if $ver { ::puts stderr "compiled mtime [set ::tsp::${tspfile}::mtime] [info body ::tsp::${tspfile}::main]" } } ::tsp::${tspfile}::main } ;# openIsis::Tspr proc openIsis::MHL {str} { global env if [info exists env(ENCODING)] { set str [encoding convertfrom $env(ENCODING) $str] } # ::puts stderr "MHLing '$str'" regsub -all {(<[^=>]*)=[^>]+>} $str {\1>} str; # dump substitutions regsub -all {><} $str {; } str; # replace >< pairs regsub -all {[><]} $str {} str; # nuke other >< regsub {^\^.} $str {} str; # kill initial subfield spec regsub -all {\^a} $str {; } str; # ^a -> ; regsub -all {\^[b-i]} $str {, } str; # ^[b-i] -> , regsub -all {\^.} $str {. } str; # others -> . return $str } proc openIsis::MHLl {list} { set ret {} foreach str $list { lappend ret [MHL $str] } return $ret } # a procedure to make live with requests easier: # it uses the current value of the $srv server variable to send a request # so we can switch servers by setting srv, yet use a single command proc openIsis::isis {request args} { variable srv variable ver $srv .req delete $srv .req set -type $request # eval so that the single list of args becomes multiple parameters eval $srv req $args if $ver { ::puts stderr "###req [$srv .req serialize]###res [$srv .res serialize]###" } # return the size (number of records or mfns) $srv .res get {size 0} } # convenient access to the result proc openIsis::res {args} { variable srv eval $srv .res $args } # wrapper for current result record # add a "get first", if args is a single integer proc openIsis::v {args} { variable srv # if {"rowid" == $args} { return [$srv .res get {rowid 0}] } if {[string is integer $args]} { # v 24 return [$srv .res .rec get "$args {}"] } eval $srv .res .rec $args } # write record rec as mfn to server srv proc openIsis::save {{rec rec} {mfn 0}} { variable srv variable fsp set dbpath [$srv get 5]/[$srv get 721] ;# syspath/db if {$fsp(db) != $dbpath} { # check for fsp set fsp(db) $dbpath if {[catch {source $dbpath.fsp} fsp(proc)] || "" == [info procs $fsp(proc)]} { ::puts stderr "fsp $fsp(proc) : $::errorInfo" set fsp(proc) "" } ::puts stderr "using fsp $fsp(proc)" } $srv .req delete if {"" != $fsp(proc)} { $srv new ::openIsis::idx idx delete if $mfn { # delete entries for old version isis read -rowid $mfn idx add -1 del ;# set delete mode $fsp(proc) ::openIsis::v ::openIsis::idx idx add -1 add ;# back to add mode } # add new index entries $fsp(proc) $rec ::openIsis::idx ::puts stderr "index entries:\n[idx serialize]" $srv .req wrap -tag 926 ::openIsis::idx } $srv .req set 904 $mfn # wrap the record into request $srv .req wrap -tag 908 $rec ::puts stderr "write:\n[$srv .req serialize]" if $mfn { $srv req -type update -rowid $mfn } else { $srv req -type insert } res get rowid } ;# save proc openIsis::prnres {msg stb} { if [$stb .res] { set res [$stb .res get sid ser error error2] set tms "" set dta "" catch { set tms [$stb .res get tms] } catch { set dta [$stb .res get rec] } ::puts stderr "\[$stb/[lindex $res 0]] $msg:\ ser [lindex $res 1],\ err [lindex $res 2]/[lindex $res 3],\ tms $tms,\ dta $dta" } else { ::puts stderr "\[$stb] $msg: no response available" } } proc openIsis::errno { stb } { if { ![ $stb .res ] } { return 0 } set err [$stb .res get error] if { $err } { return $err } set err [$stb .res get error2] return $err } proc openIsis::maxrow {stb db wrn} { $stb .req delete $stb req type maxrow db $db prnres "maxrow $db" $stb set row 0 if [catch { set row [$stb .res get rowid] }] { if $wrn { global errorInfo ::puts stderr $errorInfo } } { # puts stderr "maxrow = $row" } return $row } tcl/openisistcl.087.c0100644000113000011300000004105607644412746013356 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: openisistcl.087.c,v 1.2 2003/04/08 00:20:54 kripke Exp $ tcl/tk binding */ #include #include #include #include #include #include #include #ifdef WIN32 #define snprintf _snprintf #endif /* ************************************************************ private functions */ #define SESPRE "oises" #define SPLEN 5 #define MAXVALLEN 4096 /* <= INT_MAX */ #define BUFLEN 32 #define PRNBUF(fmt,arg) \ memset (buf, 0, BUFLEN); \ snprintf (buf, BUFLEN - 1, fmt, arg); #define PRNBUF2(fmt,a1,a2) \ memset (buf, 0, BUFLEN); \ snprintf (buf, BUFLEN - 1, fmt, a1, a2); typedef struct { OpenIsisSession ses; long idx; int dbid; OpenIsisSet ois; } OIT_Set; static int OIT_CmdSess _ANSI_ARGS_ ( ( OpenIsisSession ses, Tcl_Interp *ip, int argc, const char **argv ) ); static int OIT_CmdDb _ANSI_ARGS_ ( ( int *dbid, Tcl_Interp *ip, int argc, const char **argv ) ); static int OIT_FreeDb _ANSI_ARGS_ ( ( int *dbid, Tcl_Interp *ip, int argc, const char **argv ) ); static int OIT_CmdSet _ANSI_ARGS_ ( ( OIT_Set *set, Tcl_Interp *ip, int argc, const char **argv ) ); static int OIT_FreeSet _ANSI_ARGS_ ( ( OIT_Set *set, Tcl_Interp *ip, int argc, const char **argv ) ); static int scanLong (const char *str, long *res) { char *endp; if (0 == str) { return 0; } *res = strtol (str, &endp, 0); if (LONG_MAX == *res || LONG_MIN == *res) { return 0; } if (endp == str) { return 0; } return 1; } static int scanInt (const char *str, int *res) { long lval; if (! scanLong (str, &lval)) { return 0; } if (INT_MAX <= lval || INT_MIN >= lval) { return 0; } *res = (int) lval; return 1; } static OpenIsisSession scanSes (const char *str) { int sid; if (strncmp (SESPRE, str, SPLEN)) { return (OpenIsisSession)0; } if (! scanInt (str + SPLEN, &sid)) { return (OpenIsisSession)0; } return openIsisId2Session (sid); } static int resultSes (Tcl_Interp *ip, OpenIsisSession ses) { char buf[BUFLEN]; int rt; rt = openIsisSession2Id (ses); if (0 > rt) { PRNBUF ("%d", rt); Tcl_SetResult (ip, buf, TCL_VOLATILE); return TCL_ERROR; } PRNBUF2 ("%s%d", SESPRE, rt); Tcl_CreateCommand (ip, buf, (Tcl_CmdProc*)OIT_CmdSess, (ClientData)ses, (Tcl_CmdDeleteProc*)0); Tcl_SetResult (ip, buf, TCL_VOLATILE); return TCL_OK; } static int _SetCnt = 0; /* ??? */ static int NewSetCounterFor (OIT_Set *set) { return _SetCnt++; } static char *Rec2String (OpenIsisRec *rec, Tcl_DString *ds) { Tcl_DStringInit (ds); if (0 != rec) { OpenIsisField *fld; char val[MAXVALLEN]; char buf[BUFLEN]; unsigned len; for (fld = rec->field; rec->len > fld - rec->field; ++fld) { Tcl_DStringStartSublist (ds); PRNBUF ("%ld", fld->tag); Tcl_DStringAppendElement (ds, buf); if (0 != fld->len) { if (0 == (len = (unsigned) fld->len) || MAXVALLEN <= len) { fprintf (stderr, "truncation of %ld %ld %ld\n", rec->rowid, fld->tag, fld->len); len = MAXVALLEN - 1; } strncpy (val, fld->val, len) [len] = 0; } else { *val = 0; } Tcl_DStringAppendElement (ds, val); Tcl_DStringEndSublist (ds); } } return Tcl_DStringValue (ds); } static int OIT_ReadRec ( Tcl_Interp *ip, OpenIsisSession ses, int dbid, int row ) { Tcl_DString ds; OpenIsisRec *rec = openIsisDRead (ses, dbid, row); Tcl_SetResult (ip, Rec2String (rec, &ds), TCL_VOLATILE); if (0 != rec) { free (rec); } Tcl_DStringFree (&ds); return TCL_OK; } static int Append2Rec ( OpenIsisSession ses, OpenIsisRec **recp, const char *tag, const char *val ) { int ntag; if (0 == val || ! scanInt (tag, &ntag) || 0 >= ntag) { return 0; } OPENISIS_RADDS (ses, *recp, ntag, val, !0); return 1; } /* ============================ general ================================ */ /* openIsisCLog ( int level, const char *filename ) */ static int OIT_Log ( ClientData cd, Tcl_Interp *ip, int argc, const char **argv ) { const char *fname = (char*)0; int lvl; switch (argc) { case 3: fname = argv[2]; case 2: if (scanInt (argv[1], &lvl) && -1 <= lvl && 9 >= lvl) { break; } if (1 == strlen (argv[1])) { lvl = *argv[1]; break; } default: Tcl_AppendResult (ip, "usage: ", argv[0], " []", (char*)0); return TCL_ERROR; } openIsisCLog (lvl, fname); /* fname currently ignored */ return TCL_OK; } /* openIsisCInit(int argc, const char **argv,OpenIsisCLockFunc lock) incl. openIsisCOpen ( OpenIsisRec *args ) */ static int OIT_Init ( ClientData cd, Tcl_Interp *ip, int argc, const char **argv ) { char buf[BUFLEN]; int rt; rt = openIsisCInit (argc - 1, argv + 1, 0); if (0 != rt) { PRNBUF ("%d", rt); Tcl_SetResult (ip, buf, TCL_VOLATILE); return TCL_ERROR; } return resultSes (ip, OPENISIS_SES0 ()); } /* cDOpenv ( const char *dbname, const char **argv, int argc ) */ static int OIT_Open ( ClientData cd, Tcl_Interp *ip, int argc, const char **argv ) { Tcl_CmdInfo cinfo; char buf[BUFLEN]; int *dbid; int rt; if (2 > argc) { Tcl_AppendResult (ip, "usage: ", argv[0], " [args...]", (char*)0); return TCL_ERROR; } errno = 0; rt = openIsisCDOpenv (argv[1], argv + 2, argc - 2); if (0 > rt) { if (0 != errno) { Tcl_SetResult (ip, strerror (errno), TCL_VOLATILE); } else { PRNBUF ("%d", rt); Tcl_SetResult (ip, buf, TCL_VOLATILE); } return TCL_ERROR; } PRNBUF ("oidb%d", rt); *(dbid = (int*) Tcl_Alloc (sizeof (int))) = rt; if (! Tcl_GetCommandInfo (ip, buf, &cinfo)) { Tcl_CreateCommand (ip, buf, (Tcl_CmdProc*)OIT_CmdDb, (ClientData)dbid, (Tcl_CmdDeleteProc*)OIT_FreeDb); } Tcl_SetResult (ip, buf, TCL_VOLATILE); return TCL_OK; } /* ============================ session ================================ */ static int OIT_CmdSess ( OpenIsisSession ses, Tcl_Interp *ip, int argc, const char **argv ) { if (2 != argc || strncmp ("new", argv[1], strlen (argv[1]))) { Tcl_AppendResult (ip, "usage: ", argv[0], " new", (char*)0); return TCL_ERROR; } return resultSes (ip, openIsisCSession ((OpenIsisRec*)0)); } /* ============================ db ================================ */ static int OIT_CmdDb ( int *dbid, Tcl_Interp *ip, int argc, const char **argv ) { OpenIsisSession ses = (OpenIsisSession)0; const char *sname = ""; const char *rname = ""; char buf[BUFLEN]; unsigned len; int rt; if (2 > argc || 0 == (len = strlen (argv[1]))) { goto usage; } switch (*argv[1]) { case 'a': if (0 == strncmp ("append", argv[1], len)) { Tcl_DString ds; OpenIsisRec *rec; long rid; int fidx = 3; if (5 > argc || 0 == (argc & 1)) { goto usage; } if (! scanLong (argv[2], &rid)) { goto usage; } if (0 == strncmp ("-session", argv[3], strlen (argv[3]))) { if (0 == (ses = scanSes (sname = argv[4]))) { goto nosess; } if ((fidx = 5) == argc) { goto usage; } } if (0 == ses) { ses = OPENISIS_SES0 (); } if (0 == (rec = openIsisDRead (ses, *dbid, rid))) { rname = argv[2]; goto norec; } while (fidx < argc) { if (! Append2Rec (ses, &rec, argv[fidx], argv[1 + fidx])) { Tcl_AppendResult (ip, argv[0], ": error appending field ", argv[fidx], (char*)0); return TCL_ERROR; } fidx += 2; } Tcl_SetResult (ip, Rec2String (rec, &ds), TCL_VOLATILE); if (0 != rec) { free (rec); } Tcl_DStringFree (&ds); return TCL_OK; } /* 'a' */ goto usage; case 'c': if (1 != len) { if (0 == strncmp ("close", argv[1], len)) { if (2 != argc) { goto usage; } Tcl_DeleteCommand (ip, (char*)argv[0]); return TCL_OK; } if (0 == strncmp ("check", argv[1], len)) { int flags = 0; if (2 != argc) { if (3 != argc) { goto usage; } if (! scanInt (argv[2], &flags)) { goto usage; } } rt = openIsisCDCheck (*dbid, flags); if (0 == rt) { return TCL_OK; } PRNBUF ("%d", rt); Tcl_SetResult (ip, buf, TCL_VOLATILE); return TCL_ERROR; } } goto usage; case 'd': /* $db delete */ goto usage; case 'i': /* $db insert {rec} */ goto usage; case 'm': if (0 == strncmp ("maxrowid", argv[1], len)) { if (2 != argc) { if (4 != argc || strncmp ("-session", argv[2], strlen (argv[2]))) { goto usage; } if (0 == (ses = scanSes (sname = argv[3]))) { goto nosess; } } if (0 == ses) { ses = OPENISIS_SES0 (); } rt = openIsisDMaxId (ses, *dbid); PRNBUF ("%d", rt); if (0 > rt) { Tcl_AppendResult (ip, "could not determine max rowid: ", buf, (char*)0); return TCL_ERROR; } Tcl_SetResult (ip, buf, TCL_VOLATILE); return TCL_OK; } goto usage; case 'q': { OIT_Set *set; const char *key; long mode = OPENISIS_QRY_KEYEQ; long skip = 0; long lrt; int j; if (strncmp ("query", argv[1], len)) { goto usage; } if (3 > argc) { goto usage; } key = argv[2]; for (j = 3; argc > j; ++j) { if (argv[j][0] != '-') { goto usage; } switch (argv[j][1]) { case 'm': if (strncmp ("-mode", argv[j], strlen (argv[j]))) { goto usage; } if (argc <= ++j) { goto usage; } if (! scanLong (argv[j], &mode)) { goto usage; } continue; case 's': { unsigned ll = strlen (argv[j]); if (3 > ll) { goto usage; } if (0 == strncmp ("-session", argv[j], ll)) { if (argc <= ++j) { goto usage; } if (0 == (ses = scanSes (sname = argv[j]))) { goto nosess; } continue; } if (0 == strncmp ("-skip", argv[j], ll)) { if (argc <= ++j) { goto usage; } if (! scanLong (argv[j], &skip)) { goto usage; } continue; } } /* fall thru */ default: goto usage; } } /* for args */ if (0 == ses) { ses = OPENISIS_SES0 (); } set = (OIT_Set*) Tcl_Alloc (sizeof (OIT_Set)); set->ois.len = set->idx = 0; set->ses = ses; set->dbid = *dbid; lrt = openIsisDQuery (ses, &(set->ois), *dbid, key, mode, skip); if (0 > lrt) { Tcl_Free ((ClientData)set); PRNBUF ("%ld", lrt); Tcl_SetResult (ip, buf, TCL_VOLATILE); return TCL_ERROR; } rt = NewSetCounterFor (set); PRNBUF ("oiset%d", rt); Tcl_CreateCommand (ip, buf, (Tcl_CmdProc*)OIT_CmdSet, (ClientData)set, (Tcl_CmdDeleteProc*)OIT_FreeSet); Tcl_SetResult (ip, buf, TCL_VOLATILE); return TCL_OK; } /* 'q' */ goto usage; case 'r': { long rid; if (strncmp ("read", argv[1], len)) { goto usage; } switch (argc) { case 5: if (strncmp ("-session", argv[3], strlen (argv[3]))) { goto usage; } if (0 == (ses = scanSes (sname = argv[4]))) { goto nosess; } case 3: if (scanLong (argv[2], &rid) && 0 <= rid) { break; } default: goto usage; } if (0 == ses) { ses = OPENISIS_SES0 (); } return OIT_ReadRec (ip, ses, *dbid, rid); } /* 'r' */ goto usage; case 'u': if (0 == strncmp ("update", argv[1], len)) { Tcl_DString ds; OpenIsisRec *rec; long rid; int fidx = 3; if (5 > argc || 0 == (argc & 1)) { goto usage; } if (! scanLong (argv[2], &rid)) { goto usage; } if (0 == strncmp ("-session", argv[3], strlen (argv[3]))) { if (0 == (ses = scanSes (sname = argv[4]))) { goto nosess; } if ((fidx = 5) == argc) { goto usage; } } if (0 == ses) { ses = OPENISIS_SES0 (); } if (0 == (rec = openIsisRDup (ses, (OpenIsisRec*)0, 1024, !0))) { Tcl_AppendResult (ip, argv[0], ": cannot allocate record", (char*)0); return TCL_ERROR; } while (fidx < argc) { if (! Append2Rec (ses, &rec, argv[fidx], argv[1 + fidx])) { Tcl_AppendResult (ip, argv[0], ": error appending field ", argv[fidx], (char*)0); return TCL_ERROR; } fidx += 2; } rec->rowid = rid; rt = openIsisDWrite (ses, *dbid, rec); if (0 != rt) { PRNBUF ("%d", rt); Tcl_SetResult (ip, buf, TCL_VOLATILE); return TCL_ERROR; } Tcl_SetResult (ip, Rec2String (rec, &ds), TCL_VOLATILE); if (0 != rec) { free (rec); } Tcl_DStringFree (&ds); return TCL_OK; } /* 'u' */ goto usage; } /* switch (*argv[1]) */ usage: Tcl_AppendResult (ip, "ops on db:\n ", argv[0], " check [] |", " close |", " maxrowid [-session ] |", " query [-session ] [-mode ] [-skip ] |", " read [-session ]\n", /* "ops on recs:\n ", argv[0], " append [-session ] |", " delete [-session ] |", " insert [-session ] |", " update [-session ] ", */ (char*)0); return TCL_ERROR; nosess: Tcl_AppendResult (ip, argv[0], ": no such session: ", sname, (char*)0); return TCL_ERROR; norec: Tcl_AppendResult (ip, argv[0], ": no such record: ", rname, (char*)0); return TCL_ERROR; } static int OIT_FreeDb ( int *dbid, Tcl_Interp *ip, int argc, const char **argv ) { char buf[BUFLEN]; int rt; rt = openIsisCDClose (*dbid); if (0 == rt) { Tcl_Free ((ClientData)dbid); return TCL_OK; } PRNBUF2 ("%d: %d", *dbid, rt); Tcl_Free ((ClientData)dbid); Tcl_AppendResult (ip, "could not close database ", buf, (char*)0); return TCL_ERROR; } /* ============================ set ================================ */ static int OIT_CmdSet ( OIT_Set *set, Tcl_Interp *ip, int argc, const char **argv ) { unsigned len; switch (argc) { case 3: { int idx; if (0 == (len = strlen (argv[1])) || 'g' != *argv[1]) { goto usage; } if (strncmp ("get", argv[1], len)) { goto usage; } if (! scanInt (argv[2], &idx)) { goto usage; } if (0 > idx || set->ois.len <= idx) { char buf[BUFLEN]; PRNBUF2 ("%d / %ld", idx, set->ois.len); Tcl_AppendResult (ip, "index out of bounds: ", buf, (char*)0); return TCL_ERROR; } return OIT_ReadRec (ip, set->ses, set->dbid, set->ois.id[idx]); } /* 3 */ case 2: if (0 != (len = strlen (argv[1])) && 'g' != *argv[1]) { break; } default: goto usage; } /* switch (argc) */ switch (*argv[1]) { case 'c': if (0 == strncmp ("close", argv[1], len)) { Tcl_DeleteCommand (ip, (char*)argv[0]); return TCL_OK; } break; case 'l': if (0 == strncmp ("list", argv[1], len)) { char buf[BUFLEN]; long *rid; int j; for (j = 0, rid = set->ois.id; set->ois.len > j; ++j, ++rid) { PRNBUF ("%ld", *rid); Tcl_AppendElement (ip, buf); } return TCL_OK; } break; case 'n': if (0 == strncmp ("next", argv[1], len)) { if (set->ois.len > set->idx) { char buf[BUFLEN]; PRNBUF ("%ld", set->ois.id[set->idx++]); Tcl_AppendElement (ip, buf); } return TCL_OK; } break; case 'r': if (0 == strncmp ("rewind", argv[1], len)) { set->idx = 0; return TCL_OK; } break; case 's': if (0 == strncmp ("size", argv[1], len)) { char buf[BUFLEN]; PRNBUF ("%ld", set->ois.len); Tcl_SetResult (ip, buf, TCL_VOLATILE); return TCL_OK; } break; } /* switch (*argv[1]) */ usage: Tcl_AppendResult (ip, "usage: ", argv[0], " close | get | list | next | rewind | size", (char*)0); return TCL_ERROR; } static int OIT_FreeSet ( OIT_Set *set, Tcl_Interp *ip, int argc, const char **argv ) { Tcl_Free ((ClientData)set); return TCL_OK; } /* ************************************************************ public functions */ int OpenisisTcl_Init ( Tcl_Interp *ip ) { Tcl_CreateCommand (ip, "openIsisInit", (Tcl_CmdProc*)OIT_Init, (ClientData)0, (Tcl_CmdDeleteProc*)0); Tcl_CreateCommand (ip, "openIsisLog", (Tcl_CmdProc*)OIT_Log, (ClientData)0, (Tcl_CmdDeleteProc*)0); Tcl_CreateCommand (ip, "openIsisOpen", (Tcl_CmdProc*)OIT_Open, (ClientData)0, (Tcl_CmdDeleteProc*)0); return TCL_OK; } int OpenisisTcl_AppInit ( Tcl_Interp *interp ) { if ( TCL_ERROR == OpenisisTcl_Init( interp ) ) return TCL_ERROR; return Tcl_AppInit( interp ); } tcl/openisistcl.c0100644000113000011300000015063407663442011013031 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: openisistcl.c,v 1.66 2003/05/23 14:40:48 mawag Exp $ tcl/tk binding */ #include "openisis.h" #include "openisistcl.h" #include "luti.h" /* luti_getembed, * luti_ptrincr, * luti_parse_path, * luti_free, * luti_append */ #include "ldsp.h" /* openIsisEnc2Utf8 */ /* include this after the Tcl stuff for the benefit of those who use the 150% braindead gcc 2.96 which barfs on declaration of `index' shadows global declaration in generic/tclDecls.h cause string.h declares char *index(const char *s, int c) */ #include #include #include #include #include #ifdef WIN32 #define snprintf _snprintf #endif #ifdef _REENTRANT extern int openisis_threaded; static int *link_dummy = &openisis_threaded; /* force correct linkage */ #endif /* ============================ types ================================== */ /* name of local schema */ #define OIT_STB0 "openIsisRoot" typedef enum { /* record commands */ RC_ADD, RC_CLON, RC_DB, /* stub, too */ RC_DEL, RC_DESER, RC_DO, RC_DONE, /* stub, too */ RC_FDT, /* stub, too */ RC_FMT, RC_GET, RC_ROW, RC_LEN, RC_SERI, RC_SET, RC_WRAP, RC_LAST = RC_WRAP, /* used for checking last rec cmd */ /* stub commands */ SC_NEW, SC_RECV, SC_RQS, SC_ARQS, SC_ARSP } OITCmd; static const char *OITOpts[] = { /* record commands */ "add", "clone", "db", "delete", "deserialize", "do", "done", "fdt", "format", "get", "rowid", "length", "serialize", "set", "wrap", /* stub commands */ "new", "recv", "request", ".req", ".res", 0 }; #define OIT_RS_SZM 0x000FF /* size mask */ #define OIT_RS_RQS 0x00100 /* schemas request rec */ #define OIT_RS_RSP 0x00200 /* schemas response rec */ #define OIT_RS_STC 0x00400 /* schemas config rec */ #define OIT_RS_DBC 0x00800 /* db config rec */ #define OIT_RS_DBF 0x01000 /* db fdt rec */ #define OIT_RS_USED 0x10000 /* rec is in use */ #define OIT_RS_FRE 0x20000 /* free associated memory */ #define OIT_RS_OWNF 0x40000 /* record has own fdt that must be freed */ #define RecType(r) (0x01F00 & (r)->stat) #define NonWritable(r) (0x01A00 & (r)->stat) #define NonDeletable(r) (0x00700 & (r)->stat) typedef struct OITSess OITSess; typedef struct { OpenIsisRec *rec; OpenIsisDb *db; /* own db; tmp set to target db in rqs */ const OpenIsisFdt *fdt; /* own fdt */ const char *cmd; /* associated tcl ip cmd */ int sid; /* allocator */ int stat; } OITRec; #define RecSess(r) \ (0 <= ((OITRec*)(r))->sid && NumSessions > ((OITRec*)(r))->sid ? \ Sessions + ((OITRec*)(r))->sid : 0) typedef struct { OITRec env; OpenIsisStub stb; OITRec **recs; /* embedded recs */ int numr; } OITCont; #define OIT_ST_ROOT 0x0001 #define OIT_ST_TCL 0x0002 #define OIT_ST_OINIT 0x0004 typedef struct { OpenIsisStub stb; OITRec cfg; /* direct schema copy allocated by ses0 */ OITCont *rqs; OITCont *rsp; const char *cmd; /* associated tcl ip cmd */ Tcl_Obj *dfltproc; Tcl_Obj *actproc; int ases; /* act session of rqs and rsp */ int stat; } OITStub; #define StbSess(s) (0 <= (s)->ases && NumSessions > (s)->ases ? \ Sessions + (s)->ases : 0) struct OITSess { Tcl_Interp *ip; OITRec **recs; int numr; int six; }; static OITSess *Sessions = 0; static int NumSessions = 0; /* ============================ records ================================ */ #define OIT_RECINCR 32 #define OIT_MAXRECS 65535 #define OIT_SESSINCR 1 #define OIT_MAXSESS 255 static void CtorRec (OITRec *that, int sid, int siz) { memset (that, 0, (unsigned)siz); that->stat = siz; that->sid = sid; } static int AllcRec (OITSess *ses, int siz, int type) { int j; if (siz > OIT_RS_SZM) { return openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] AllcRec: unexpected size %d", siz); } for (j = ses->numr; 0 <= --j; ) { if (! ses->recs[j]) { goto allcj; } if (! (OIT_RS_USED & ses->recs[j]->stat) && siz == (OIT_RS_SZM & ses->recs[j]->stat) /* may be <= */ ) { goto done; } } j = luti_ptrincr ( &ses->recs, &ses->numr, OIT_RECINCR, sizeof (OITRec*), OIT_MAXRECS); if (0 > j) { return openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] AllcRec: out of memory"); } allcj: ses->recs[j] = (OITRec*) openIsisMAlloc (siz); if (! ses->recs[j]) { return openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] AllcRec: out of memory"); } CtorRec (ses->recs[j], ses->six, siz); done: ses->recs[j]->stat |= type | OIT_RS_USED; return j; } static int NewRec ( OITSess *ses, OpenIsisDb *db, const OpenIsisFdt *fdt, int type ) { int j = AllcRec (ses, sizeof (OITRec), type); if (0 <= j) { ses->recs[j]->db = db; ses->recs[j]->fdt = fdt; } return j; } static int NewCont (OITSess *ses, const OpenIsisFdt *fdt, int type) { int j = AllcRec (ses, sizeof (OITCont), type); if (0 <= j) { ses->recs[j]->fdt = fdt; } return j; } static void DtorRecs (OITRec **recs, int numr, int frmem); static void DtorRec (OITRec *that, int frmem) { if (that) { OITCont *con; OITSess *ses; int siz, type; ses = RecSess (that); if (! ses) { openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] DtorRec: illegal sid %d(%d)", that->sid, NumSessions); return; } if (that->cmd) { if (frmem) { that->stat |= OIT_RS_FRE; } Tcl_DeleteCommand (ses->ip, (char*)that->cmd); return; } type = RecType (that); siz = OIT_RS_SZM & that->stat; switch (type) { case 0: case OIT_RS_DBF: if (that->rec) { openIsisMFree (that->rec); } break; case OIT_RS_DBC: /* that->rec = 0; readonly cfg handled by db */ break; case OIT_RS_RQS: if (that->rec) { openIsisMFree (that->rec); } /* fall thru */ case OIT_RS_RSP: /* that->rec = 0; response record handled by stub */ con = (OITCont*)that; DtorRecs (con->recs, con->numr, 0); break; default: /* OIT_RS_STC embedded in OITStub and handled by stub */ openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] DtorRec: unexpected type %x", type); return; } if ((OIT_RS_OWNF & that->stat)) { openIsisFFree ((OpenIsisFdt*)that->fdt); } if (frmem || (OIT_RS_FRE & that->stat)) { openIsisMFree (that); } else { CtorRec (that, ses->six, siz); } } } static void DtorRecs (OITRec **recs, int numr, int frmem) { if (recs) { while (0 <= --numr) { DtorRec (recs[numr], frmem); } openIsisMFree (recs); } } static int CtorSess (Tcl_Interp *ip) { int j = luti_ptrincr ( &Sessions, &NumSessions, OIT_SESSINCR, sizeof (OITSess), OIT_MAXSESS); if (0 > j) { return -1; } Sessions[j].ip = ip; Sessions[j].six = j; return j; } static void DtorSess (OITSess *that) { DtorRecs (that->recs, that->numr, !0); } static void ExitSess () { int j; if (NumSessions) { for (j = NumSessions; 0 <= --j; ) { DtorSess (Sessions + j); } openIsisMFree (Sessions); Sessions = 0; NumSessions = 0; } } /* --------------------------------------------------------------------- */ static unsigned _RecId = 0; static char* NewRecId (char *buf) { sprintf (buf, "openIsisRec%u", ++_RecId); return buf; } static void TclDelRec (ClientData cld) { OITRec *that = (OITRec*)cld; if (that->cmd) { openIsisMFree ((void*)that->cmd); that->cmd = 0; } DtorRec ((OITRec*)cld, 0); } static int CmdRec ( ClientData rid, Tcl_Interp *ip, int argc, Tcl_Obj* const argv[] ); static int CrtRecCmd (OITSess *ses, const char *name, OITRec *rec, int srst) { char buf[64]; if (! name || ! *name) { name = NewRecId (buf); } rec->cmd = (const char*) openIsisMDup (name, -1); if (! rec->cmd) { Tcl_AppendResult (ses->ip, "CrtRecCmd: out of memory", 0); return TCL_ERROR; } Tcl_CreateObjCommand (ses->ip, (char*)name, &CmdRec, rec, &TclDelRec); if (srst) { Tcl_SetResult (ses->ip, (char*)name, TCL_VOLATILE); } return TCL_OK; } static char **ToArgv ( Tcl_Obj* const objv[], int objc, char* buf, int siz ) { char *str, *nb; char *res = buf; int posp = 0; int posv = objc * sizeof (char*); int j, len, nsz; for (j = 0; objc > j; ++j) { str = Tcl_GetStringFromObj (objv[j], &len); ++len; nsz = posv + len; if (siz < nsz) { nsz *= 2; nb = (char*) openIsisMAlloc (nsz); if (! nb) { openIsisSMsg (OPENISIS_ERR_NOMEM, "[openisistcl] ToArgv: out of memory"); return 0; } memcpy (nb, res, (unsigned) posv); if (res != buf) { openIsisMFree (res); } siz = nsz; res = nb; } *(char**)(res + posp) = res + posv; strcpy (res + posv, str); posp += sizeof (char*); posv += len; } return (char**)res; } static int BuildEmbRecs ( OITCont *that, OpenIsisRec **recs, int numr, int frr ) { OITSess *ses; OITRec **oirs; int buf[1000]; int *idx = buf; int j; ses = RecSess (that); if (! ses) { return openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] BuildEmbRecs: illegal sid %d(%d)", that->env.sid, NumSessions); } oirs = (OITRec**) openIsisMAlloc ( (int) (numr * sizeof (OITRec*))); if (! oirs) { if (frr) { luti_free ((void**)recs, numr); } return openIsisSMsg (OPENISIS_ERR_NOMEM, "[openisistcl] BuildEmbRecs(%d): out of memory", numr); } if (1000 < numr) { idx = (int*) openIsisMAlloc ( (int) (numr * sizeof (int))); if (! idx) { if (frr) { luti_free ((void**)recs, numr); } openIsisMFree (oirs); return openIsisSMsg (OPENISIS_ERR_NOMEM, "[openisistcl] BuildEmbRecs(%d): out of memory", numr); } } for (j = numr; 0 <= --j; ) { idx[j] = NewRec (ses, 0, 0, 0); if (0 > idx[j]) { if (frr) { luti_free ((void**)recs, numr); } openIsisMFree (oirs); if (idx != buf) { openIsisMFree (idx); } return openIsisSMsg (OPENISIS_ERR_NOMEM, "[openisistcl] BuildEmbRecs(%d): out of memory", numr); } } for (j = numr; 0 <= --j; ) { oirs[j] = ses->recs[idx[j]]; oirs[j]->rec = recs[j]; } that->recs = oirs; that->numr = numr; if (idx != buf) { openIsisMFree (idx); } if (frr) { openIsisMFree (recs); } return numr; } static int BuildRqsRecs (OITCont *that) { OpenIsisRec *recs[4] = { 0, 0, 0, 0 }; /* REC, IDX, CFG, fdt */ int numr; numr = BuildEmbRecs (that, recs, 4, 0); if (4 == numr) { if ((that->recs[0]->db = that->env.db)) { that->recs[0]->fdt = that->env.db->fdt; } that->recs[2]->fdt = openIsisFdtDbpar; that->recs[3]->fdt = openIsisFdtFdt; } return numr; } static int BuildRspRecs (OITCont *that, Tcl_Interp *ip, const char *arg0) { OpenIsisRec **recs; OpenIsisDb *db; int *rows; /* save rowid in recs */ int numr, j; numr = openIsisNGetResult (that->stb, &rows, &recs, &db, 0); if (0 > numr) { Tcl_AppendResult (ip, arg0, ": child allocation failure", 0); return numr; } if (rows) { openIsisMFree (rows); } if (0 == numr || ! recs) { return 0; } j = BuildEmbRecs (that, recs, numr, !0); if (j != numr) { Tcl_AppendResult (ip, arg0, ": child allocation failure", 0); return j; } if (db) { for (j = numr; 0 <= --j; ) { that->recs[j]->db = db; that->recs[j]->fdt = db->fdt; } } return numr; } static int UsageRec (Tcl_Interp *ip, const char *arg0) { if (! arg0) { arg0 = ""; } Tcl_AppendResult (ip, "usage: ", arg0, " add field value ?field value ...? |", " clone ?options? newname ?field value ...? |", " db ?options? |", " delete ?field ...? |", " deserialize line |", " do ?tagvar? valvar body |", " done |", " fdt ?options? |", " format ?options? format |", " get ?-tags | -tagnames | field ...? |", " rowid |", " serialize |", " set field ?value field value ...? |", " wrap ?options? recname |", " .path ?option arg ...?", 0); return TCL_ERROR; } static int OpPath ( OITRec *that, Tcl_Interp *ip, const char *arg0, Tcl_Obj *arg1, int argc, Tcl_Obj* const argv[] ); static int OpDb (Tcl_Interp *ip, OITSess *ses, OpenIsisDb *db, int argc, Tcl_Obj* const argv[] ) { OITRec *oir; int rt; rt = NewRec (ses, 0, openIsisFdtDbpar, OIT_RS_DBC); if (0 > rt) { Tcl_AppendResult (ip, "openIsisDb: out of memory", 0); return TCL_ERROR; } oir = ses->recs[rt]; oir->rec = db->cfg; if (argc) { rt = OpPath (oir, ip, "openIsisDb", argv[0], argc - 1, argv + 1); DtorRec (oir, 0); return rt; } return CrtRecCmd (ses, 0, oir, !0); } static int OpFdt (Tcl_Interp *ip, OITSess *ses, const OpenIsisFdt *fdt, int argc, Tcl_Obj* const argv[] ) { OITRec *oir; int rt; rt = NewRec (ses, 0, openIsisFdtFdt, OIT_RS_DBF); if (0 > rt) { Tcl_AppendResult (ip, "openIsisFdt: out of memory", 0); return TCL_ERROR; } oir = ses->recs[rt]; oir->rec = openIsisFFdt2Rec (fdt, 0, 0); if (argc) { rt = OpPath (oir, ip, "openIsisFdt", argv[0], argc - 1, argv + 1); DtorRec (oir, 0); return rt; } return CrtRecCmd (ses, 0, oir, !0); } #define FldObj( f ) Tcl_NewStringObj( (f)->val, (f)->len ) static Tcl_Obj* FdObj (OpenIsisField *fld, const OpenIsisFdt *fdt) { char buf[32]; OpenIsisFd *fd = fdt ? openIsisFById (fdt, fld->tag, 0) : 0; if (fd) { return Tcl_NewStringObj (fd->name, -1); } openIsisI2a (buf, fld->tag); return Tcl_NewStringObj (buf, -1); } static int OpRec ( OITRec *that, Tcl_Interp *ip, const char *arg0, const OITCmd cmd, int argc, Tcl_Obj* const argv[] ) { OITSess *ses = RecSess (that); if (! ses) { Tcl_AppendResult (ip, arg0, ": record corrupted", 0); return TCL_ERROR; } switch (cmd) { case RC_SET: case RC_ADD: { char buf[2048]; OpenIsisRec *oldrec; char **argp = 0; char **args = 0; int setf = 0; int len; if (0 == argc) { return UsageRec (ip, arg0); } if (RC_SET == cmd) { if (1 == argc) { goto op_get; } if ( that->rec ) { setf = OPENISIS_RCHG; } } if (NonWritable (that)) { Tcl_AppendResult (ip, arg0, ": readonly record", 0); return TCL_ERROR; } args = argp = ToArgv (argv, argc, buf, sizeof (buf)); if (! argp) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); return TCL_ERROR; } while (argc) { if ('-' != args[0][0]) { break; } len = strlen (args[0]); if (1 == len || '-' == args[0][1]) { break; } if (0 == strncmp ("-ignore", args[0], len)) { setf |= OPENISIS_RIGN; ++args; --argc; continue; } if (RC_SET != cmd) { break; } if (0 == strncmp ("-default", args[0], len)) { setf = OPENISIS_RDFLT | (OPENISIS_RIGN & setf); ++args; --argc; continue; } break; } oldrec = that->rec; that->rec = openIsisRSet (oldrec, OPENISIS_RARGV | OPENISIS_RFDT | OPENISIS_RDIS | setf | argc, that->fdt, args); if (argp != (char**)buf) { openIsisMFree (argp); } if (! that->rec && (oldrec || (argc && !(OPENISIS_RIGN & setf)))) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); return TCL_ERROR; } return TCL_OK; } /* RC_SET, RC_ADD */ case RC_CLON: { char buf[64]; OITRec *nrec; char *opt = 0; char *name = 0; int empty = 0; int nn = 0; int j, len, rt, id; if ((OIT_RS_RQS | OIT_RS_RSP) & that->stat) { Tcl_AppendResult (ip, arg0, ": container cloning not allowed", 0); return TCL_ERROR; } for (j = 0; argc > j; ++j) { opt = Tcl_GetStringFromObj (argv[j], &len); if (0 == len) { return UsageRec (ip, arg0); } if ('-' != *opt) { break; } if (0 == opt[1]) { nn = !0; ++j; break; } if (0 == strncmp ("-empty", opt, len)) { empty = !0; continue; } return UsageRec (ip, arg0); } if (argc > j && ! nn) { name = Tcl_GetStringFromObj (argv[j], &len); if (0 == len) { return UsageRec (ip, arg0); } ++j; } if (! name) { name = NewRecId (buf); } id = NewRec (ses, that->db, that->fdt, 0); if (0 > id) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); return TCL_ERROR; } nrec = ses->recs[id]; if (! empty) { nrec->rec = openIsisRDup (that->rec, 0, 0); } if (argc > j) { rt = OpRec (nrec, ip, name, RC_SET, argc - j, argv + j); if (TCL_OK != rt) { return rt; } } rt = CrtRecCmd (ses, name, ses->recs[id], !0); return rt; } /* RC_CLON */ case RC_DB: if (! that->db) { Tcl_AppendResult (ip, arg0, ": no db", 0); return TCL_ERROR; } return OpDb (ip, ses, that->db, argc, argv); case RC_DEL: { if (NonWritable (that)) { Tcl_AppendResult (ip, arg0, ": readonly record", 0); return TCL_ERROR; } if (argc) { char buf[2048]; char **argp = ToArgv (argv, argc, buf, sizeof (buf)); if (! argp) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); return TCL_ERROR; } that->rec = openIsisRSet (that->rec, OPENISIS_RARGV | OPENISIS_RFDT | OPENISIS_RDEL | argc, that->fdt, argp); if (argp != (char**)buf) { openIsisMFree (argp); } return TCL_OK; } if (that->rec) { OPENISIS_CLRREC( that->rec ); } if (OIT_RS_RQS == RecType (that)) { OITCont *con = (OITCont*)that; if (con->numr) { DtorRecs (con->recs, con->numr, 0); con->numr = 0; con->recs = 0; } } return TCL_OK; } /* RC_DEL */ case RC_DO: { Tcl_Obj *tagvar = 0; Tcl_Obj *valvar; Tcl_Obj *tag = 0; Tcl_Obj *val; Tcl_Obj *body; OpenIsisField *f; int i, rt; if ( 2 > argc ) return UsageRec (ip, arg0); body = argv[--argc]; valvar = argv[--argc]; if ( 1 == argc ) tagvar = argv[0]; #ifndef DONTREUSE /* prepare object for var */ if ( ! Tcl_ObjSetVar2( ip, valvar, 0, val = Tcl_NewObj(), TCL_LEAVE_ERR_MSG ) ) return TCL_ERROR; if ( tagvar && ! Tcl_ObjSetVar2( ip, tagvar, 0, tag = Tcl_NewIntObj(0), TCL_LEAVE_ERR_MSG ) ) return TCL_ERROR; #endif /* go loop */ rt = TCL_OK; for ( i = that->rec->len, f = that->rec->field; i--; f++ ) { #ifndef DONTREUSE /* if somebody shares the object, we have to set a new one (or Tcl_SetStringObj will panic). if it's not owned by the var, it's either owned by someone else or nobody, i.e. deleted */ if ( val == Tcl_ObjGetVar2( ip, valvar, 0, TCL_LEAVE_ERR_MSG ) && !Tcl_IsShared( val ) ) Tcl_SetStringObj( val, (char*)f->val, f->len ); else #endif if ( !Tcl_ObjSetVar2( ip, valvar, 0, val = FldObj( f ), TCL_LEAVE_ERR_MSG ) ) return TCL_ERROR; if ( tagvar ) { #ifndef DONTREUSE if ( tag == Tcl_ObjGetVar2( ip, tagvar, 0, TCL_LEAVE_ERR_MSG ) && !Tcl_IsShared( tag ) ) Tcl_SetIntObj( tag, f->tag ); else #endif if ( !Tcl_ObjSetVar2( ip, tagvar, 0, tag = Tcl_NewIntObj( f->tag ), TCL_LEAVE_ERR_MSG ) ) return TCL_ERROR; } /* tagvar */ switch (rt = Tcl_EvalObjEx( ip, body, 0 )) { case TCL_CONTINUE: rt = TCL_OK; case TCL_OK: continue; case TCL_BREAK: rt = TCL_OK; case TCL_RETURN: case TCL_ERROR: default: return rt; } } return rt; } /* RC_DO */ case RC_DONE: { if (! that->cmd) { Tcl_AppendResult (ip, arg0, ": no command bound to rec", 0); return TCL_ERROR; } if (strcmp (that->cmd, arg0)) { Tcl_AppendResult (ip, arg0, ": command mismatch: ", that->cmd, 0); return TCL_ERROR; } if (NonDeletable (that)) { Tcl_AppendResult (ip, arg0, ": record not deletable", 0); return TCL_ERROR; } Tcl_DeleteCommand (ip, (char*)arg0); return TCL_OK; } /* RC_DONE */ case RC_FDT: if (! that->fdt) { Tcl_AppendResult (ip, arg0, ": no fdt", 0); return TCL_ERROR; } return OpFdt (ip, ses, that->fdt, argc, argv); case RC_FMT: { Tcl_AppendResult (ip, arg0, ": sorry, format not implemented yet", 0); return TCL_ERROR; } /* RC_FMT */ case RC_GET: op_get: { OpenIsisField *fld; const char *path, *rem; int tag, occ, i, j, objc, len; int witht = 0; int usedf = !0; Tcl_Obj *list, *dflt, *val; Tcl_Obj **objv; Tcl_ResetResult (ip); if (! that->rec) { return TCL_OK; } list = val = 0; if (1 == argc) { path = Tcl_GetStringFromObj (argv[0], &i); if (3 < i && *path == '-') { if (0 == strncmp ("-tags", path, i)) { --argc; witht = 1; } else if (0 == strncmp ("-tagnames", path, i)) { --argc; witht = 2; } } } if (! argc) { list = Tcl_NewListObj( 0, 0 ); for (i = that->rec->len, fld = that->rec->field; i--; fld++) { if (witht) { if (TCL_OK != Tcl_ListObjAppendElement (ip, list, FdObj (fld, 2 == witht ? that->fdt : 0))) { goto geterr; } } if (TCL_OK != Tcl_ListObjAppendElement (ip, list, FldObj (fld))) { goto geterr; } } Tcl_SetObjResult( ip, list ); return TCL_OK; } for (j = 0; argc > j; ) { path = Tcl_GetStringFromObj (argv[j], &len); if (3 <= len && '-' == *path && 0 == strncmp ("-nodefaults", path, len)) { usedf = 0; ++j; continue; } val = dflt = 0; if (usedf && TCL_OK == Tcl_ListObjGetElements (0, argv[j], &objc, &objv) && 2 == objc ) { path = Tcl_GetStringFromObj (objv[0], 0); dflt = objv[1]; } rem = luti_parse_path (path, that->fdt, 0, &tag, &occ); if (! rem || *rem) { Tcl_ResetResult (ip); Tcl_AppendResult (ip, arg0, ": no such path: ", path, 0); goto geterr; } if ( 0 > occ ) { for ( i = that->rec->len, fld = that->rec->field; i--; fld++ ) { if ( tag == fld->tag ) { if (! val) { val = Tcl_NewListObj (0, 0); } if (TCL_OK != Tcl_ListObjAppendElement ( ip, val, FldObj(fld))) { goto geterr; } } } } else { /* specific occ wanted */ fld = openIsisROccurence (that->rec, tag, occ); if (fld) { val = FldObj(fld); } } if (! val) { if (! dflt) { Tcl_ResetResult (ip); Tcl_AppendResult (ip, arg0, ": no such field: ", path, 0); goto geterr; } /* don't force a list even for empty default -- take the default as list. A literal default foo IS the list containing foo. If you REALLY wan't a list with one empty element as default, you can explicitly specify one. if (0 > occ) { Tcl_Obj *tmp[1]; tmp[0] = dflt; val = Tcl_NewListObj (1, tmp); } else */ val = dflt; } if (++j >= argc) { if (! list) { Tcl_SetObjResult( ip, val ); return TCL_OK; } if (TCL_OK != Tcl_ListObjAppendElement(ip, list, val)) { goto geterr; } break; } if (! list) { list = Tcl_NewListObj( 0, 0 ); } if (TCL_OK != Tcl_ListObjAppendElement(ip, list, val)) { goto geterr; } } if (! list) { return UsageRec (ip, arg0); } Tcl_SetObjResult( ip, list ); return TCL_OK; geterr: /* free */ if (val) { Tcl_DecrRefCount (val); } if (list) { Tcl_DecrRefCount( list ); } return TCL_ERROR; } /* RC_GET */ case RC_ROW: Tcl_SetObjResult( ip, Tcl_NewIntObj( (that->rec && 0rec->rowid) ? that->rec->rowid : 0 ) ); return TCL_OK; case RC_LEN: Tcl_SetObjResult( ip, Tcl_NewIntObj( (that->rec && 0rec->len) ? that->rec->len : 0 ) ); return TCL_OK; case RC_SERI: { char buf[2048]; char *b; int len; if (! that->rec) { Tcl_ResetResult (ip); return TCL_OK; } len = sizeof (buf); b = openIsisRSerializeAlloc (that->rec, buf, &len); if (! b) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); return TCL_ERROR; } /* do NOT include the final blankline */ Tcl_SetObjResult( ip, Tcl_NewStringObj( b, len-1 ) ); /* faster, but not allowed :( Tcl_Obj *ret = Tcl_NewStringObj(0,0); Tcl_SetObjLength( ret, that->rec->used ); Tcl_SetObjLength( ret, openIsisRSerialize( ret->bytes, that->rec ) ); */ if ( buf != b ) mFree( b ); return TCL_OK; } /* RC_SERI */ case RC_DESER: { char *b; int len; if ( 1 != argc ) return UsageRec (ip, arg0); if (NonWritable (that)) { Tcl_AppendResult (ip, arg0, ": readonly record", 0); return TCL_ERROR; } b = Tcl_GetStringFromObj( argv[0], &len ); if ( ! b ) Tcl_ResetResult (ip); else { int ret = openIsisRDeserialize( &that->rec, b, len, OPENISIS_RDIS|OPENISIS_STOPONEMPTY ); Tcl_SetObjResult( ip, Tcl_NewIntObj( ret ) ); } return TCL_OK; } /* RC_DESER */ case RC_WRAP: { Tcl_CmdInfo info; char buf[2048]; OITRec *emb = 0; char **argp = (char**)buf; char *name = 0; char *tgnm = 0; int tag = -1; int num = -1; int del = 0; int len, j; if (NonWritable (that)) { Tcl_AppendResult (ip, arg0, ": readonly record", 0); return TCL_ERROR; } argp = ToArgv (argv, argc, buf, sizeof(buf)); if (! argp) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); return TCL_ERROR; } for (j = 0; argc > j; ++j) { if (! *argp[j]) { goto wuserr; } if ('-' == *argp[j]) { len = strlen (argp[j]); if (2 > len) { goto wuserr; } if (0 == strncmp ("-done", argp[j], len)) { del = !0; continue; } if (0 == strncmp ("-number", argp[j], len)) { if (argc <= ++j) { goto wuserr; } num = openIsisA2id (argp[j], -1, -1); if (0 > num) { goto wuserr; } continue; } if (0 == strncmp ("-tag", argp[j], len)) { if (argc <= ++j) { goto wuserr; } tag = openIsisA2id (argp[j], -1, 0); if (0 >= tag) { goto wuserr; } continue; } goto wuserr; } if (0 > tag && 0 == tgnm) { tgnm = argp[j]; continue; } if (name) { goto wuserr; } name = argp[j]; if (! Tcl_GetCommandInfo (ip, name, &info)) { Tcl_AppendResult (ip, arg0, ": no such record: ", name, 0); goto wrperr; } if (info.objProc != &CmdRec) { Tcl_AppendResult (ip, arg0, ": ", name, " is not a record", 0); goto wrperr; } emb = (OITRec*) info.objClientData; if (! emb) { Tcl_AppendResult (ip, arg0, ": internal record error", 0); goto wrperr; } } /* for argc */ if (tgnm) { OpenIsisFd *fd = openIsisFByName (that->fdt, tgnm); if (! fd) { Tcl_AppendResult (ip, arg0, ": no such field description: ", tgnm, 0); goto wrperr; } tag = fd->id; } if (0 > tag) { goto wuserr; } if (num && ! emb) { Tcl_AppendResult (ip, arg0, ": record to embed not given", 0); goto wrperr; } if (0 <= num) { that->rec = openIsisRAddI (that->rec, tag, num, !0); if (num) { if (! emb->rec || ! (len = emb->rec->len)) { Tcl_AppendResult (ip, arg0, ": record to embed is empty", 0); goto wrperr; } that->rec = luti_append (that->rec, emb->rec); } if (! that->rec) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); goto wrperr; } } else { len = that->rec || emb->rec; that->rec = luti_wrap (that->rec, emb->rec, tag); if (len && ! that->rec) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); goto wrperr; } } if (del && emb) { len = OpRec (emb, ip, name, RC_DONE, 0, 0); if (TCL_OK != len) { goto wrperr; } } if (argp != (char**)buf) { openIsisMFree (argp); } return TCL_OK; wuserr: Tcl_AppendResult (ip, "usage: " , arg0, " wrap [-done] ", "[-number ] {-tag | } recname", 0); wrperr: if (argp != (char**)buf) { openIsisMFree (argp); } return TCL_ERROR; } /* RC_WRAP */ default: { char buf[654]; openIsisI2a (buf, cmd); Tcl_AppendResult (ip, arg0, ": unrecognized command ", buf, 0); return TCL_ERROR; } /* default */ } /* switch */ } static int OpPath ( OITRec *that, Tcl_Interp *ip, const char *arg0, Tcl_Obj *arg1, int argc, Tcl_Obj* const argv[] ) { char buf[128]; OITSess *ses; const char *path; int cmd, rt; if (! arg1) { return UsageRec (ip, arg0); } if (! arg0) { arg0 = ""; } if (! that || ! (ses = RecSess (that))) { Tcl_AppendResult (ip, arg0, ": record corrupted", 0); return TCL_ERROR; } if (ip != ses->ip) { Tcl_AppendResult (ip, arg0, ": session corrupted", 0); return TCL_ERROR; } path = Tcl_GetStringFromObj (arg1, 0); /* path to embedded rec */ if ('.' == *path) { int type = RecType (that); switch (type) { case OIT_RS_RQS: { OITCont *con; OITRec *rec; const char *p2; int tag, occ; con = (OITCont*) that; if (strncmp (".fdt", path, 4)) { p2 = luti_parse_path (path, openIsisFdtRqs, 0, &tag, &occ); if (0 == p2 || 0 < occ) { Tcl_AppendResult (ip, arg0, ": no such child: ", path, 0); return TCL_ERROR; } } else { p2 = path + 4; tag = -42; occ = 0; } if (! con->numr) { rt = BuildRqsRecs (con); if (0 > rt) { Tcl_AppendResult (ip, arg0, ": child allocation failure", 0); return TCL_ERROR; } } switch (tag) { case OPENISIS_COM_REC: rec = con->recs[0]; break; case OPENISIS_RQS_IDX: rec = con->recs[1]; break; case OPENISIS_COM_CFG: rec = con->recs[2]; break; case -42: rec = con->recs[3]; break; default: Tcl_AppendResult (ip, arg0, ": no such child: ", path, 0); return TCL_ERROR; } NewRecId (buf); if (*p2) { Tcl_Obj *obj = Tcl_NewStringObj (p2, (int) strlen (p2)); if (obj) { Tcl_IncrRefCount (obj); rt = OpPath (rec, ip, buf, obj, argc, argv); Tcl_DecrRefCount (obj); return rt; } return TCL_ERROR; } if (! argc) { return CrtRecCmd (ses, buf, rec, !0); } return OpPath (rec, ip, buf, argv[0], argc - 1, argv + 1); } case OIT_RS_RSP: { OITCont *con; OITRec *rec; const char *p2; int tag, occ; con = (OITCont*) that; p2 = luti_parse_path (path, openIsisFdtRsp, 0, &tag, &occ); if (0 == p2 || OPENISIS_COM_REC != tag ) { Tcl_AppendResult (ip, arg0, ": no such child: ", path, 0); return TCL_ERROR; } if (!(rt = con->numr)) { rt = BuildRspRecs (con, ip, arg0); if (0 > rt) { return TCL_ERROR; } } if (0 > occ) { occ = 0; } if (rt <= occ) { sprintf (buf, "%d", rt); Tcl_AppendResult (ip, arg0, ": no such child: ", path, ", have ", buf, " childs", 0); return TCL_ERROR; } rec = con->recs[occ]; NewRecId (buf); if (*p2) { Tcl_Obj *obj = Tcl_NewStringObj (p2, (int) strlen (p2)); if (obj) { Tcl_IncrRefCount (obj); rt = OpPath (rec, ip, buf, obj, argc, argv); Tcl_DecrRefCount (obj); return rt; } return TCL_ERROR; } if (! argc) { return CrtRecCmd (ses, buf, rec, !0); } return OpPath (rec, ip, buf, argv[0], argc - 1, argv + 1); } default: { OpenIsisRec *rec; OITRec *oir; rec = luti_getembed (that->rec, path, that->fdt); if (! rec) { Tcl_AppendResult (ip, arg0, ": no such child: ", path, 0); return TCL_ERROR; } rt = NewRec (ses, 0, OIT_RS_DBF == RecType(that) ? openIsisFdtFd : 0, 0); if (0 > rt) { Tcl_AppendResult (ip, "OpPath: out of memory", 0); return TCL_ERROR; } oir = ses->recs[rt]; oir->rec = rec; NewRecId (buf); if (! argc) { return CrtRecCmd (ses, buf, oir, !0); } rt = OpPath (oir, ip, buf, argv[0], argc - 1, argv + 1); DtorRec (oir, 0); return rt; } } } /* path */ rt = Tcl_GetIndexFromObj (ip, arg1, OITOpts, "option", 0, &cmd); if (TCL_OK != rt) { return TCL_ERROR; } if (RC_LAST < cmd) { return UsageRec (ip, arg0); } return OpRec (that, ip, arg0, (OITCmd)cmd, argc, argv); } static int CmdRec ( ClientData cld, Tcl_Interp *ip, int argc, Tcl_Obj* const argv[] ) { const char *arg0 = 0 >= argc ? 0 : Tcl_GetStringFromObj (argv[0], 0); if (1 >= argc) { return UsageRec (ip, arg0); } return OpPath ((OITRec*)cld, ip, arg0, argv[1], argc - 2, argv + 2); } /* ============================== stubs ================================ */ static int RspCb ( void *cld, OpenIsisStub stb, OpenIsisRec *rsp, OpenIsisDb *db ) { OITStub *that = (OITStub*) cld; OITSess *ses; const char *res; int rt; (void)db; if (! that) { return openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] RspCb: response without stub"); } if (stb != that->stb) { return openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] RspCb: stub changed"); } ses = StbSess (that); if (! ses) { return openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] RspCb: response without session"); } if (! that->rqs) { return openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] RspCb: response without request"); } if (that->rsp) { if (that->rsp->env.rec) { openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] RspCb: multiple responses"); DtorRec ((OITRec*)that->rsp, 0); } } rt = NewCont (ses, openIsisFdtRsp, OIT_RS_RSP); if (0 > rt) { return rt; } that->rsp = (OITCont*) ses->recs[rt]; that->rsp->env.rec = rsp; that->rsp->stb = that->stb; if (! that->actproc) { if (! that->dfltproc) { return 0; } rt = Tcl_EvalObj (ses->ip, that->dfltproc); } else { rt = Tcl_EvalObj (ses->ip, that->actproc); } res = Tcl_GetStringResult (ses->ip); if (! res) { res = ""; } if (TCL_ERROR == rt) { return openIsisSMsg (OPENISIS_ERR_IDIOT, "[openisistcl] callback eval: %s", res); } if (TCL_OK != rt) { openIsisSMsg (OPENISIS_LOG_WARN, "[openisistcl] callback eval = %d, %s", rt, res); return 0; } openIsisSMsg (OPENISIS_LOG_INFO, "[openisistcl] callback eval : %s", res); return 0; } static void StbDelCb (void *cld, OpenIsisStub stb, void *cbd); static void TclDelStb (ClientData cld); static int CmdStub ( ClientData cld, Tcl_Interp *ip, int argc, Tcl_Obj* const argv[] ); static OITStub* CtorStub (Tcl_Interp *ip, const char *name, int argc, const char **argv, const char *proc ) { OpenIsisStubCbData scd; OpenIsisSchema *sch; OITStub *that; int stat = 0; if (! name) { name = OIT_STB0; stat = OIT_ST_ROOT; } if (! NumSessions) { openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] CtorStub(%s): no root session", name); return 0; } if (ip != Sessions->ip) { openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] CtorStub(%s): must not create stub in derived session", name); return 0; } that = (OITStub*) openIsisMAlloc (sizeof (OITStub)); if (! that) { openIsisSMsg (OPENISIS_ERR_NOMEM, "[openisistcl] CtorStub(%s): out of memory", name); return 0; } that->cmd = openIsisMDup (name, -1); if (! that->cmd) { openIsisMFree (that); openIsisSMsg (OPENISIS_ERR_NOMEM, "[openisistcl] CtorStub(%s): out of memory", name); return 0; } that->ases = -1; memset (&scd, 0, sizeof (OpenIsisStubCbData)); scd.dfltcb = &RspCb; scd.delcb = &StbDelCb; scd.dfltcld = scd.delcld = that; if (OIT_ST_ROOT & stat) { that->cfg.fdt = openIsisFdtSyspar; that->stb = openIsisNInit (argc, argv, &scd); } else { that->cfg.fdt = openIsisFdtScheme; that->stb = openIsisNOpen (name, argc, argv, &scd); } if (! that->stb) { openIsisMFree ((void*)that->cmd); openIsisMFree (that); openIsisSMsg (OPENISIS_ERR_IDIOT, "[openisistcl] CtorStub(%s): deficient configuration", name); return 0; } sch = openIsisNSchema (that->stb); that->cfg.rec = sch->cfg; that->cfg.sid = 0; that->cfg.stat = OIT_RS_USED | OIT_RS_STC; if (proc) { that->dfltproc = Tcl_NewStringObj (proc, (int) strlen (proc)); if (that->dfltproc) { Tcl_IncrRefCount (that->dfltproc); } } that->stat = stat | OIT_ST_TCL | OIT_ST_OINIT; Tcl_CreateObjCommand (ip, (char*)that->cmd, &CmdStub, that, &TclDelStb); Tcl_SetResult (ip, (char*)that->cmd, TCL_VOLATILE); return that; } static void DtorStub (OITStub *that, int where) { const char *cmd = that->cmd; if (! cmd) { cmd = "<>"; } if (! NumSessions) { openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] DtorStub(%s): no root session", cmd); return; } that->stat &= ~where; if (OIT_ST_TCL & that->stat) { if (! that->cmd) { openIsisSMsg (OPENISIS_ERR_TRASH, "[openisistcl] DtorStub: no command"); return; } Tcl_DeleteCommand (Sessions->ip, (char*)that->cmd); return; } if (OIT_ST_OINIT & that->stat) { if (OIT_ST_ROOT & that->stat) { openIsisNDeinit (); } else { openIsisNClose (that->stb); } return; } if (that->dfltproc) { Tcl_DecrRefCount (that->dfltproc); } if (that->actproc) { Tcl_DecrRefCount (that->actproc); } if (that->rqs) { DtorRec ((OITRec*)that->rqs, 0); } if (that->rsp) { DtorRec ((OITRec*)that->rsp, 0); } /* that->cfg.rec holds the same ref as stub->cfg, which is freed in openIsisNClose */ if (that->cmd) { openIsisMFree ((void*)that->cmd); } openIsisMFree (that); } static void TclDelStb (ClientData cld) { DtorStub ((OITStub*)cld, OIT_ST_TCL); } static void StbDelCb (void *cld, OpenIsisStub stb, void *cbd) { (void) stb; if (cbd) { if (((OITStub*)cbd)->actproc) { Tcl_DecrRefCount (((OITStub*)cbd)->actproc); ((OITStub*)cbd)->actproc = 0; } return; } DtorStub ((OITStub*)cld, OIT_ST_OINIT); } static int BuildRqsCont (OITStub *that, Tcl_Interp *ip, const char *arg0, OITSess *ses, int *argc, Tcl_Obj* const **argv ) { OpenIsisDb *db; char *dbn; int rt; if (! that->rqs) { rt = NewCont (ses, openIsisFdtRqs, OIT_RS_RQS); if (0 > rt) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); return 0; } that->rqs = (OITCont*) ses->recs[rt]; } if (1 < *argc && 0 == strcmp ("-db", Tcl_GetStringFromObj ((*argv)[0], 0))) { dbn = Tcl_GetStringFromObj ((*argv)[1], 0); db = openIsisNDbByName (that->stb, dbn); if (! db) { Tcl_AppendResult (ip, arg0, ": no such db: ", dbn, 0); return 0; } that->rqs->env.db = db; *argc -= 2; *argv += 2; } return !0; } static const OpenIsisFdt* SysFdtFromName (const char *dbn, int len) { switch (dbn[1]) { case 'd': if (0 == strncmp (dbn, "-dbpar", len)) { return openIsisFdtDbpar; } return 0; case 'f': if ('d' == dbn[2]) { if (0 == dbn[3]) { return openIsisFdtFd; } if ('t' == dbn[3] && 0 == dbn[4]) { return openIsisFdtFdt; } } return 0; case 'r': if (3 < len) { if (0 == strncmp (dbn, "-request", len)) { return openIsisFdtRqs; } if (0 == strncmp (dbn, "-response", len)) { return openIsisFdtRsp; } } return 0; case 's': if (2 < len) { if (0 == strncmp (dbn, "-syspar", len)) { return openIsisFdtSyspar; } if (0 == strncmp (dbn, "-scheme", len)) { return openIsisFdtScheme; } } return 0; } return 0; } static int UsageStub (Tcl_Interp *ip, const char *argv0) { Tcl_AppendResult (ip, "usage: ", (argv0 ? argv0 : ""), " db db ?option ...? |", " fdt db ?option ...? |", " new -schema name ?-cfg val ...? |", " new ?-db db? ?name? |", " recv |", " request ?-db db? ?-param val? |", " .req ?-db db? ?option ...? |", " .res ?option ...?", 0); return TCL_ERROR; } static int CmdStub ( ClientData cld, Tcl_Interp *ip, int argc, Tcl_Obj* const argv[] ) { OITStub *that = (OITStub*) cld; OITSess *ses; const char *arg0; OITCmd cmd; int rt; if (1 > argc) { return TCL_ERROR; } arg0 = Tcl_GetStringFromObj (argv[0], 0); if (2 > argc) { return UsageStub (ip, arg0); } ses = StbSess (that); if (! ses) { if (! NumSessions) { Tcl_AppendResult (ip, arg0, ": no session", 0); return TCL_ERROR; } ses = Sessions; that->ases = 0; } if (ip != ses->ip) { Tcl_AppendResult (ip, arg0, ": session changed", 0); return TCL_ERROR; } rt = Tcl_GetIndexFromObj (ip, argv[1], OITOpts, "option", 0, (int*)&cmd); if (TCL_OK != rt) { return UsageStub (ip, arg0); } argc -= 2; argv += 2; switch (cmd) { case RC_DB: { OpenIsisDb *db; char *dbn; if (! argc) { return UsageStub (ip, arg0); } dbn = Tcl_GetStringFromObj (argv[0], 0); db = openIsisNDbByName (that->stb, dbn); if (! db) { Tcl_AppendResult (ip, arg0, ": no such db <", dbn, ">", 0); return TCL_ERROR; } return OpDb (ip, ses, db, argc - 1, argv + 1); } /* RC_DB */ case RC_DONE: Tcl_DeleteCommand (ip, (char*)arg0); return TCL_OK; case RC_FDT: { const OpenIsisFdt *fdt = 0; OpenIsisDb *db; char *dbn; int len; if (! argc) { return UsageStub (ip, arg0); } dbn = Tcl_GetStringFromObj (argv[0], &len); if (1 < len && '-' == *dbn) { fdt = SysFdtFromName (dbn, len); } if (! fdt) { db = openIsisNDbByName (that->stb, dbn); if (! db) { Tcl_AppendResult (ip, arg0, ": no such db <", dbn, ">", 0); return TCL_ERROR; } fdt = db->fdt; if (! fdt) { Tcl_AppendResult (ip, arg0, ": ", dbn, " has no fdt", 0); return TCL_ERROR; } } return OpFdt (ip, ses, fdt, argc - 1, argv + 1); } /* RC_FDT */ case SC_NEW: { char buf[2048]; OITStub *nstb = 0; const OpenIsisFdt *fdt = 0; OpenIsisDb *db = 0; const char *arg2 = 0; const char *dbn = 0; const char *proc = 0; char *name = 0; char **argp = 0; int len = 0; int dbl = 0; switch (argc) { /* new record */ case 0: goto newrec; case 1: name = Tcl_GetStringFromObj (argv[0], 0); goto newrec; case 3: arg2 = Tcl_GetStringFromObj (argv[0], &len); if (2 > len || 0 != strncmp ("-db", arg2, len)) { return UsageStub (ip, arg0); } dbn = Tcl_GetStringFromObj (argv[1], &dbl); name = Tcl_GetStringFromObj (argv[2], 0); newrec: if (1 < dbl && '-' == *dbn) { fdt = SysFdtFromName (dbn, dbl); } if (! fdt) { if (! dbn) { dbn = openIsisRString (that->cfg.rec, OPENISIS_SC_DFLTDB, 0, buf, sizeof(buf)); if (! dbn) { Tcl_AppendResult (ip, arg0, ": no db specified", 0); return TCL_ERROR; } } db = openIsisNDbByName (that->stb, dbn); if (! db) { Tcl_AppendResult (ip, arg0, ": no such db <", dbn, ">", 0); return TCL_ERROR; } fdt = db->fdt; } rt = NewRec (ses, db, fdt, 0); if (0 > rt) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); return TCL_ERROR; } rt = CrtRecCmd (ses, name, ses->recs[rt], !0); return rt; /* new stub */ default: arg2 = Tcl_GetStringFromObj (argv[0], &len); if (2 > len) { return UsageStub (ip, arg0); } if (2 == argc && 0 == strncmp ("-db", arg2, len)) { dbn = Tcl_GetStringFromObj (argv[1], &dbl); goto newrec; } if (strncmp ("-schema", arg2, len) && strncmp ("schema", arg2, len)) { return UsageStub (ip, arg0); } name = Tcl_GetStringFromObj (argv[1], 0); argc -= 2; argv += 2; if (0 < argc) { int j; argp = ToArgv (argv, argc, buf, sizeof (buf)); if (! argp) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); return TCL_ERROR; } for (j = 0; argc > j; ++j) { if (0 == strcmp ("-async", argp[j]) && argc > ++j) { proc = argp[j]; } } } nstb = CtorStub (ip, name, argc, (const char**)argp, proc); if (argp && argp != (char**)buf) { openIsisMFree (argp); } if (! nstb) { Tcl_AppendResult (ip, arg0, ": deficient configuration for ", name, " or out of memory", 0); return TCL_ERROR; } return TCL_OK; } /* switch (argc) */ } /* SC_NEW */ case SC_RECV: { if (that->rsp && that->rsp->env.rec) { return TCL_OK; } Tcl_AppendResult (ip, arg0, ": waiting for response in async mode not implemented yet", 0); return TCL_ERROR; } /* SC_RECV */ case SC_RQS: { OpenIsisRec *rqs; if (that->rsp) { DtorRec ((OITRec*) that->rsp, 0); } if (! BuildRqsCont (that, ip, arg0, ses, &argc, &argv)) { return TCL_ERROR; } if (0 < argc) { Tcl_Obj *setcmd = Tcl_NewStringObj ("set", 3); if (! setcmd) { Tcl_AppendResult (ip, arg0, ": out of memory", 0); return TCL_ERROR; } Tcl_IncrRefCount (setcmd); rt = OpPath ( (OITRec*)that->rqs, ip, 0, setcmd, argc, argv); Tcl_DecrRefCount (setcmd); if (TCL_OK != rt) { return rt; } } rqs = that->rqs->env.rec; if (that->rqs->numr) { OpenIsisRec *rec; if ((rec = that->rqs->recs[0]->rec)) { rqs = luti_wrap (rqs, rec, OPENISIS_COM_REC); } if ((rec = that->rqs->recs[1]->rec)) { rqs = luti_wrap (rqs, rec, OPENISIS_RQS_IDX); } if ((rec = that->rqs->recs[2]->rec)) { rqs = luti_wrap (rqs, rec, OPENISIS_COM_CFG); } if ((rec = that->rqs->recs[3]->rec)) { rqs = luti_append (rqs, rec); } } if (! openIsisRGet (rqs, OPENISIS_COM_DBN, 0) && that->rqs->env.db) { OPENISIS_RADDS (rqs, OPENISIS_COM_DBN, that->rqs->env.db->name, !0); } rt = openIsisNSend (that->stb, that->rqs->env.rec = rqs, 0, 0, !0); if (0 != rt) { char buf[64]; sprintf (buf, "%x", rt); Tcl_AppendResult (ip, arg0, ": error ", buf, " sending request", 0); return TCL_ERROR; } return TCL_OK; } /* SC_RQS */ case SC_ARQS: { if (! BuildRqsCont (that, ip, arg0, ses, &argc, &argv)) { return TCL_ERROR; } if (1 > argc) { return UsageStub (ip, arg0); } rt = OpPath ( (OITRec*)that->rqs, ip, 0, argv[0], argc - 1, argv + 1); return rt; } /* SC_ARQS */ case SC_ARSP: { int hasrsp = that->rsp && that->rsp->env.rec; if (1 > argc) { Tcl_SetObjResult (ip, Tcl_NewBooleanObj (hasrsp)); return TCL_OK; } if (! hasrsp) { Tcl_AppendResult (ip, arg0, ": no response available", 0); return TCL_ERROR; } rt = OpPath ( (OITRec*)that->rsp, ip, 0, argv[0], argc - 1, argv + 1); return rt; } /* SC_ARSP */ default: { /* record commands */ OpenIsisSchema *sch; rt = OpRec (&that->cfg, ip, arg0, cmd, argc, argv); sch = openIsisNSchema (that->stb); /* sch->cfg is a reference to our cfg at every time, OpRec changes with RDIS, so we dont free the old rec here */ sch->cfg = that->cfg.rec; return rt; } /* default */ } /* switch cmd */ } static int CmdInit ( ClientData cld, Tcl_Interp *ip, int argc, const char *argv[] ) { OITStub *news; const char *proc = 0; int j, len; (void)cld; if (openisis_stub0) { Tcl_CmdInfo info; if (Tcl_GetCommandInfo (ip, OIT_STB0, &info)) { Tcl_SetResult (ip, OIT_STB0, TCL_STATIC); return TCL_OK; } } for (j = 1; argc > j; ++j) { if (! argv[j] || ! (len = strlen (argv[j]))) { goto usage; } if ('-' == *argv[j]) { if (! strncmp ("-async", argv[j], (unsigned) len)) { if (argc <= ++j) { goto usage; } proc = argv[j]; continue; } } break; } /* openIsisNInit can be called multiple times */ news = CtorStub (ip, 0, argc - j, argv + j, proc); if (! news) { Tcl_AppendResult (ip, OIT_STB0, ": out of memory", 0); return TCL_ERROR; } return TCL_OK; usage: Tcl_AppendResult (ip, "usage: ", argv[0], " ?-async ? ?options?", 0); return TCL_ERROR; } static int CmdOIR ( ClientData cld, Tcl_Interp *ip, int argc, Tcl_Obj* const argv[] ) { Tcl_CmdInfo info; OITSess *ses; OITRec *rec, *rfdt; OpenIsisFdt *fdt; char *fname; const char *rname; char buf[64]; int j, rt, len; (void)cld; if (! NumSessions) { Tcl_AppendResult (ip, "session not initialized", 0); return TCL_ERROR; } ses = Sessions; rec = 0; rname = 0; fdt = 0; for (j = 1; argc > j; ++j) { fname = Tcl_GetStringFromObj (argv[j], &len); if (! fdt && 2 <= len && 0 == strncmp ("-fdt", fname, len)) { if (argc <= ++j) { Tcl_AppendResult (ip, "usage: openIsisRec ?-fdt name? ?options...?", 0); return TCL_ERROR; } fname = Tcl_GetStringFromObj (argv[j], 0); if (! Tcl_GetCommandInfo (ip, fname, &info)) { Tcl_AppendResult (ip, "no such fdt: ", fname, 0); return TCL_ERROR; } if (info.objProc != &CmdRec) { Tcl_AppendResult (ip, fname, " is not a fdt", 0); return TCL_ERROR; } rfdt = (OITRec*) info.objClientData; fdt = openIsisFRec2Fdt (rfdt->rec); if (! fdt) { Tcl_AppendResult (ip, fname, " is an illegal fdt", 0); return TCL_ERROR; } if (rname) { ++j; break; } continue; } if (rname) { break; } rname = fname; } if (! rname || ! *rname) { rname = NewRecId (buf); } rt = NewRec (ses, 0, fdt, fdt ? OIT_RS_OWNF : 0); if (0 > rt) { Tcl_AppendResult (ip, Tcl_GetStringFromObj (argv[0], 0), ": out of memory", 0); goto error; } rec = ses->recs[rt]; if (j < argc - 1) { rt = OpPath (rec, ip, rname, argv[j], argc - (j+1), argv + (j+1)); if (TCL_OK != rt) { goto error; } } rt = CrtRecCmd (ses, rname, rec, !0); if (TCL_OK == rt) { return TCL_OK; } error: if (rec) { DtorRec (rec, 0); } else if (fdt) { openIsisFFree (fdt); } return TCL_ERROR; } static void FreeEnc (); static void AtExit (ClientData cld) { (void)cld; openIsisNDeinit (); ExitSess (); FreeEnc (); } static void AddCmds (Tcl_Interp *ip, int root) { Tcl_CreateCommand (ip, "openIsis", (Tcl_CmdProc*)CmdInit, 0, root ? &AtExit : 0); Tcl_CreateObjCommand (ip, "openIsisRec", &CmdOIR, 0, 0); } /* =========================== encoding ================================ */ static Tcl_HashTable Encodings; static int InitEnc = 0; static Tcl_Encoding GetEnc (Tcl_Interp *ip, const char *name, int *frs) { Tcl_HashEntry *he; Tcl_Encoding enc; int nw; if (! InitEnc) { Tcl_InitHashTable (&Encodings, TCL_STRING_KEYS); InitEnc = !0; } he = Tcl_FindHashEntry (&Encodings, name); if (he) { return (Tcl_Encoding) Tcl_GetHashValue (he); } enc = Tcl_GetEncoding (ip, name); he = Tcl_CreateHashEntry (&Encodings, name, &nw); Tcl_SetHashValue (he, enc); if (frs && ! enc) { *frs = !0; } return enc; } static void FreeEnc () { if (InitEnc) { Tcl_Encoding enc; Tcl_HashSearch hs; Tcl_HashEntry *he; for (he = Tcl_FirstHashEntry (&Encodings, &hs); he; he = Tcl_NextHashEntry (&hs) ) { enc = (Tcl_Encoding) Tcl_GetHashValue (he); if (enc) { Tcl_FreeEncoding (enc); } } Tcl_DeleteHashTable (&Encodings); InitEnc = 0; } } static const char* TrfEnc (const char *ename, const char *src, int slen, char *dst, int dlen, int invert ) { Tcl_Encoding enc; Tcl_DString str; char *tgt; int tlen, frs; if (! ename) { if (src) { openIsisMFree ((void*)src); } return 0; } if (! src || 0 >= slen) { return src; } /* tclEncoding.c says that a null interp is ok */ frs = 0; enc = GetEnc (0, ename, &frs); if (frs) { openIsisSMsg (OPENISIS_ERR_INVAL, "[openisistcl] TrfEnc: no such encoding <%s>", ename); return src; } Tcl_DStringInit (&str); if (invert) { tgt = Tcl_UtfToExternalDString (enc, src, slen, &str); } else { tgt = Tcl_ExternalToUtfDString (enc, src, slen, &str); } tlen = Tcl_DStringLength (&str); if (! dst || tlen >= dlen) { dst = (char*) openIsisMAlloc (1 + tlen); if (! dst) { return 0; } } memcpy (dst, tgt, tlen); dst[tlen] = 0; Tcl_DStringFree (&str); return dst; } OpenIsisEnc2Utf8Func *openIsisEnc2Utf8 = &TrfEnc; /* ************************************************************ public functions */ int openIsisTclNewSession (Tcl_Interp *ip) { int id; for (id = 0; NumSessions > id; ++id) { if (ip == Sessions[id].ip) { return id; } } id = CtorSess (ip); if (0 <= id) { AddCmds (ip, 0 == id); } return id; } int openIsisTclGetSession (int sid, Tcl_Interp **ip) { if (0 > sid || NumSessions <= sid) { return 0; } if (ip) { *ip = Sessions[sid].ip; } return !0; } /* void openIsisTclDelSession (int sid) { if (! sid) { ExitSess (); return; } if (0 < sid && NumSessions > sid) { DtorSess (Sessions + sid); } } */ int openIsisTclCreateRecCmd ( int sid, char *nam, int flg ) { OITSess *ses = Sessions+sid; int rid; (void)flg; /* TODO: set readonly */ rid = NewRec( ses, 0, 0, 0 ); if ( 0 <= rid ) CrtRecCmd( ses, nam, ses->recs[rid], 0 ); return rid; } /* openIsisTclCreateRecCmd */ int openIsisTclEval ( int sid, int cnt, const int *ids, OpenIsisRec **recs, char *script ) { OITSess *ses = Sessions+sid; int ret, i; if ( recs ) for ( i=cnt; i--; ) ses->recs[ ids[i] ]->rec = recs[i]; ret = Tcl_Eval( ses->ip, script ); if ( recs ) for ( i=cnt; i--; ) recs[i] = ses->recs[ ids[i] ]->rec; return ret; } /* openIsisTclEval */ int openIsisTclInit (Tcl_Interp *ip) { if (! NumSessions) { CtorSess (ip); } AddCmds (ip, !0); Tcl_CreateExitHandler (&AtExit, 0); return TCL_OK; } tcl/openisistcl.h0100644000113000011300000000513107653266142013035 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: openisistcl.h,v 1.10 2003/04/14 21:15:25 mawag Exp $ tcl/tk binding */ #ifndef OPENISISTCL #define OPENISISTCL #include EXTERN int openIsisTclInit _ANSI_ARGS_ ( ( Tcl_Interp *interp ) ); /** @return session id on success, -1 on error */ EXTERN int openIsisTclNewSession _ANSI_ARGS_ ( ( Tcl_Interp *ip ) ); /** @return boolean ok */ EXTERN int openIsisTclGetSession _ANSI_ARGS_ ( (int sessid, Tcl_Interp **ip ) ); /* interface needed by "main-threaded" server: - create commands for records "req", "res", "ses" - bind Recs to those commands and set "the" session. all recs are dynamic and can be assumed to belong to "the" session. should be discarded by "the" session when extended. the "req" record must not be modified (since it's real owner session might be working in parallel). modification would require "copy-on-write". - retrieve Recs future extension: the same game with multiple openIsisTclSessions, each with one interp (constant) */ /** create record command slot on a clean session, ids are guaranteed to increment 0,1 ... return slot id >=0 or err <0 */ EXTERN int openIsisTclCreateRecCmd ( int sid, char *nam, int flg ); /** set/get cnt records to/from slots specified by ids return cnt or err <0 EXTERN int openIsisTclSetRecs ( int oitses, int cnt, const int *ids, OpenIsisRec **recs ); EXTERN int openIsisTclGetRecs ( int oitses, int cnt, const int *ids, OpenIsisRec **recs ); */ /** set session to ses and eval command in oitses' interp if recs is not 0, set cnt recs as specified by ids (as created by openIsisTclCreateRecCmd) */ EXTERN int openIsisTclEval ( int sid, int cnt, const int *ids, OpenIsisRec **recs, char *script ); #endif /* OPENISISTCL */ tcl/sample.fsp0100755000113000011300000000224107656443745012336 0ustar kripkrip# vim:syntax=tcl # create the index record # if this file is found as $db.fsp, it is used by openIsis::save # we use the doc/Inverting controls to set word/field mode # # cds.fst is: # 70 0 MHU,(V70/) # 24 4 MHU,V24 # 69 2 V69 namespace eval test {} ;# create namespace proc test::fsp {{rec rec} {idx ::openIsis::idx}} { foreach {v24 v69 v70} [$rec get {24 {}} {69 {}} {70 {}}] break ;# lassign # method 4: add one for each word of 24 (don't count occ; it's not repeatable) set v [lindex $v24 0] set len [string length $v] $idx add -1 words ;# set word mode for {set i 0} {$i < $len} {set i $j} { set j [string wordend $v $i] set e [expr $j - 1] if {$e > $i} { # ignore single character words or non-words set w [string range $v $i $e] $idx add 24 $w } } # method 2: grep from 69 foreach {- kw} [regexp -all -inline {<([^>]+)>} [lindex $v69 0]] { $idx add 69 $kw } # method 0: add one for each occurence of 70 $idx add -1 fields ;# set field mode foreach f $v70 { regsub -all {<[^=>]*=?([^>]*)>} $f {\1} f; # dump substitutions $idx add 70 $f } return $idx } # must return the name of the FSP as result of source return test::fsp tcl/test/readall.tcl0100755000113000011300000000330107656453721013423 0ustar kripkrip#!/opt/openisis/tcl/bin/isish # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: readall.tcl,v 1.2 2003/05/07 14:15:53 mawag Exp $ # print all records from given db if {1 > $argc || 2 < $argc} { puts "usage: $argv0 []" exit 1 } set stb0 [openIsis] $stb0 add defaultdb [lindex $argv 0] if {2 == $argc} { $stb0 add syspath [lindex $argv 1] } $stb0 req type open $stb0 .req delete $stb0 req type maxrow set maxrow [ $stb0 .res get {rowid 0} ] for {set r 1} {$maxrow >= $r} {incr r} { $stb0 .req delete $stb0 req type read rowid $r catch {puts "$r: [$stb0 .res .rec get -tags]"} } $stb0 .req delete $stb0 req type close $stb0 done tcl/test/test.tsp0100755000113000011300000000110207653266143013014 0ustar kripkrip#!/opt/openisis/tcl/bin/tspr <%-- test Tcl server page $Id: test.tsp,v 1.2 2003/04/28 16:05:23 kripke Exp $ --%><%! proc foo {bla} { # proc with embedded literals !%> foo: <%=bla%> <%! };#foo %><% # preset record openIsisRec v v add 24 Foo 25 bar^abaz %> hello ISIS! this goes <% for {set i 1} {$i < 4} {incr i} { %> <%=i%> <% } %> HOME is <%=::env(HOME)%> <%-- variable --%> 2+2 is <%=expr 2+2%> <%-- command --%> v24 is <%=24%> <%-- field --%> <% foo bar %> <%-- procedure --%> <%=request serialize%> tcl/test/test1.087.tcl0100775000113000011300000000350307631670353013375 0ustar kripkrip#!/opt/openisis/tcl/tclbuild/tclsh # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: test1.087.tcl,v 1.1 2003/03/06 16:01:47 mawag Exp $ # test script for tcl-binding of openisis # openIsisLog 9 /opt/openisis/tcl/test/test1.log # openIsisLog v set ses0 [openIsisInit] puts $ses0 set db1 [openIsisOpen /opt/openisis/db/cds/cds] puts $db1 # $db1 check set maxrow [$db1 maxrowid -s $ses0] # set maxrow [$db1 maxrowid] puts $maxrow if 0 { set i 0 while {$maxrow > $i} { set rec [$db1 r $i -s $ses0] puts "$i -> $rec" incr i } } # set rs1 [$db1 query Micropolitan] set rs1 [$db1 query Africa] set rs1s [$rs1 size] puts "$rs1 total = $rs1s" set i 0 while {$rs1s > $i} { set rid [$rs1 next] set rec [$rs1 get $i] puts "found $i -> $rid -> $rec" incr i } $rs1 cl $db1 cl tcl/test/test1.tcl0100755000113000011300000001142007665600605013053 0ustar kripkrip#!/opt/openisis/tcl/bin/isish # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: test1.tcl,v 1.21 2003/05/29 18:06:32 kripke Exp $ # test script for (re)opening db, maxrow, read, query requests to # cds demo db proc putt msg { global prnout if $prnout { puts $msg } } proc prnres {msg stb {ex 1}} { if [$stb .res] { set res [$stb .res get sid ser error error2] set tms "" set dta "" catch { set tms [$stb .res get tms] } catch { set dta [$stb .res get rec] } set err [lindex $res 2] set err2 [lindex $res 3] if {$ex && ($err || $err2)} { puts "ERR \[$stb/[lindex $res 0]] $msg:\ ser [lindex $res 1],\ err $err/$err2,\ tms $tms,\ dta $dta" exit 1 } putt "\[$stb/[lindex $res 0]] $msg:\ ser [lindex $res 1],\ err $err/$err2,\ tms $tms,\ dta $dta" } else { puts "\[$stb] $msg: no response available" exit 1 } } proc maxrow {stb db wrn} { $stb .req delete $stb req type maxrow db $db prnres "maxrow $db" $stb $wrn set row 0 if [catch { set row [$stb .res get rowid] }] { if $wrn { global errorInfo putt $errorInfo } } { putt "maxrow = $row" } if {$wrn && 154 != $row && 151 != $row} { puts "ERR maxrow = $row" exit 1 } return $row } set nn 3 set mm 1 set prnout 0 if {$argc} { set nn [lindex $argv 0] if {1 < $argc} { set mm [lindex $argv 1] if {2 < $argc} { set prnout [lindex $argv 2] } } } for {set n $nn} {0 != $n} {incr n -1} { set stb0 [openIsis] putt "$stb0 created." switch [expr $n % 3] { 0 { $stb0 add \ syspath /opt/openisis/db \ dbpath cds \ defaultdb cds } 1 { $stb0 add \ dbpath /opt/openisis/db/cds } } set flds [ $stb0 get ] putt "$stb0 fields = $flds" if [catch { set stb1 [$stb0 new schema nonno -host tonno -port 2929] putt $stb1 $stb1 done if [catch $stb1] { putt "$stb1 done" } { puts "ERROR: $stb1 still alive" exit 1 } }] { global errorInfo putt $errorInfo } set stb1 "" set dbn cds $stb0 .req set type open switch [expr $n % 3] { 0 { $stb0 req } 1 { $stb0 req db $dbn } 2 { set dbn /opt/openisis/db/cds/ for {set k 0} {75 > $k} {incr k} { # PATH_MAX >= 1024 ? set dbn "$dbn../../db/cds/" } set dbn "${dbn}cds" $stb0 req db $dbn } } prnres "open cds" $stb0 maxrow $stb0 cds 1 $stb0 .req delete $stb0 req type close db cds prnres "close cds" $stb0 maxrow $stb0 cds 0 $stb0 .req delete $stb0 req type open db $dbn prnres "open cds" $stb0 maxrow $stb0 cds 1 $stb0 .req delete $stb0 req type close db cds prnres "close cds" $stb0 $stb0 .req delete $stb0 req type open db $dbn prnres "open cds" $stb0 for {set m $mm} {0 != $m} {incr m -1} { set row [maxrow $stb0 cds 1] for {set j 1} {$j <= $row} {incr j} { $stb0 .req delete $stb0 req type read db cds rowid $j set ttl "" catch { set ttl [$stb0 .res .rec get 24] } if {60 < [string length $ttl]} { set ttl "[string range $ttl 0 57] ..." } putt "row [$stb0 .res get "rowid $j"] $ttl" } for {set k 0} {2 > $k} {incr k} { $stb0 .req delete if $k { $stb0 .req set flags 2 } $stb0 req typ query db cds key Africa\$ mode 2 prnres "query cds" $stb0 set numr [$stb0 .res get size] for {set j 0} {$j < $numr} {incr j} { set ttl "" catch { set ttl [$stb0 .res .rec\[$j\] get 24] } if {60 < [string length $ttl]} { set ttl "[string range $ttl 0 57] ..." } putt "rec $j [$stb0 .res get rowid\[$j\]] $ttl" } } # set flds [ $stb0 .res get ] # puts $flds if {! $prnout && 0 == [expr $m % 50]} { puts "$n $m ..." } } ;# for m $stb0 .req delete $stb0 req type close db cds prnres "close cds" $stb0 $stb0 dele putt "$stb0 deleted." if {! $prnout && 0 == [expr $n % 50]} { puts "$n ..." } } ;# for n puts "ok." tcl/test/test2.tcl0100755000113000011300000000437307653266143013067 0ustar kripkrip#!/opt/openisis/tcl/bin/isish # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: test2.tcl,v 1.4 2003/04/19 12:25:24 mawag Exp $ # test script proc putt msg { global prnout if $prnout { puts "$msg" } } set nn 1 set nrecs 1 set nfields 200 set prnout 0 switch $argc { 0 { } 1 { set nn [lindex $argv 0] } 2 { set nn [lindex $argv 0] set nrecs [lindex $argv 1] } 3 { set nn [lindex $argv 0] set nrecs [lindex $argv 1] set nfields [lindex $argv 2] } default { set nn [lindex $argv 0] set nrecs [lindex $argv 1] set nfields [lindex $argv 2] set prnout [lindex $argv 3] } } for {set n $nn} {0 != $n} {incr n -1} { set lst "" for {set j $nrecs} {0 <= [incr j -1]} {} { set cmd [openIsisRec {} set 23232 aa 45454 bb 23232 cc] lappend lst $cmd set val jdkjgjkdsdsfdffgfdggggggfdfkglfdsjödafdafdsaö $cmd set 1 $val $cmd set 2 $val $cmd set 3 $val $cmd set 476 $val $cmd set 477 $val $cmd set 478 $val for {set f 1} {$nfields >= $f} {incr f} { $cmd set $f njjhrehjghtgthgrthtbrbrthbrtibtrbhjtiobhghbhbkgbt } $cmd delete 2 {23232[0]} 477 $cmd add 23232 dd putt [$cmd get] } foreach cmd "$lst" { $cmd done } if {! $prnout && ! [expr $n % 50]} { puts "$n ..." } } puts "ok." tcl/test/test4.087.tk0100775000113000011300000001153107631670353013234 0ustar kripkrip#!/opt/openisis/tcl/tkbuild/wish # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: test4.087.tk,v 1.1 2003/03/06 16:01:47 mawag Exp $ # test script for tk-binding of openisis entry .fsele .fsele insert 0 /opt/openisis/db/cds/cds frame .bbar button .bbar.open -text Open -command {dbOpen [.fsele get]} button .bbar.exit -text Exit -command dbExit pack configure .bbar.open .bbar.exit -side left pack .fsele .bbar wm title . TkIsis # globals: # Session0 # MaxIndex # Db$index(session,dbid,dbname,title) openIsisLog v set Session0 [openIsisInit] set MaxIndex -1 # openIsisOpen # ######################################################################## proc dbOpen fname { global ses0 errorInfo set i1 [string last / $fname] set i2 [string last . $fname] if {0 <= $i2 && $i1 < $i2} { set fname [string range $fname 0 [expr $i2 - 1]] } if [catch {dbNew $fname}] { # set i3 [string first "while executing" $errorInfo] set i3 -1 if {0 <= $i3} { set msg [string range $errorInfo 0 [expr $i3 - 1]] } { set msg $errorInfo } toplevel .erropen message .erropen.msg -text "$msg\n$fname" button .erropen.ok -text Ok -command {destroy .erropen} pack .erropen.msg .erropen.ok } } proc dbExit {} { global MaxIndex for {set j 0} {$MaxIndex >= $j} {incr j} { catch { upvar #0 Db$j db $db(dbid) close puts "$db(dbid) closed" } } exit 0 } # db # ######################################################################## proc dbNew dbname { global Session0 MaxIndex set sid [$Session0 new] set dbid [openIsisOpen $dbname] set idx [string range $sid 5 end] puts "dbNew $idx $sid $dbid" if {$MaxIndex < $idx} { set MaxIndex $idx } set i1 [string last / $dbname] if {0 <= $i1} { set title [string range $dbname [incr i1] end] } { set title $dbname } upvar #0 Db$idx db set db(dbid) $dbid set db(session) $sid set db(dbname) $dbname set db(title) $title toplevel .top$idx wm title .top$idx $title frame .top$idx.menu button .top$idx.menu.search -text Search -command "dbSearchCb $idx" button .top$idx.menu.close -text Close -command "dbCloseCb $idx" label .top$idx.lbl1 -text $dbname label .top$idx.lbl2 -text "maxrow = [$dbid maxrowid]" pack .top$idx.menu.search .top$idx.menu.close -side left pack .top$idx.menu .top$idx.lbl1 .top$idx.lbl2 -anchor w } proc dbCloseCb idx { upvar #0 Db$idx db $db(dbid) close unset db destroy .top$idx } proc dbSearchCb idx { if [catch { upvar #0 Db$idx db toplevel .ts$idx wm title .ts$idx Search entry .ts$idx.key button .ts$idx.ok -text Ok -command "dbDoTheQuery $idx" button .ts$idx.cancel -text Cancel -command "destroy .ts$idx" pack .ts$idx.key .ts$idx.ok .ts$idx.cancel -side left }] { global errorInfo puts $errorInfo } } proc dbDoTheQuery idx { if [catch { upvar #0 Db$idx db set key [.ts$idx.key get] set res [$db(dbid) query $key -session $db(session)] toplevel .tr$idx$key wm title .tr$idx$key "$db(title): $key" frame .tr$idx$key.frm listbox .tr$idx$key.frm.lst -yscrollcommand ".tr$idx$key.frm.sy set" \ -xscrollcommand ".tr$idx$key.sx set" scrollbar .tr$idx$key.frm.sy -command ".tr$idx$key.frm.lst yview" scrollbar .tr$idx$key.sx -command ".tr$idx$key.frm.lst xview" \ -orient horizontal button .tr$idx$key.close -text Close -command "destroy .tr$idx$key" set rids [$res list] set num [llength $rids] for {set j 0} {$num > $j} {incr j} { set rec [$res get $j] .tr$idx$key.frm.lst insert end \ [concat \ [lindex $rids $j] "| " \ [lindex $rec 0] "| " \ [lindex $rec 1] ] } $res close destroy .ts$idx pack .tr$idx$key.frm.lst -side left -fill both -expand 1 pack .tr$idx$key.frm.sy -side right -fill y pack .tr$idx$key.frm -fill both -expand 1 pack .tr$idx$key.sx -fill x pack .tr$idx$key.close }] { global errorInfo puts $errorInfo } } tcl/test/test5.087.tcl0100775000113000011300000000541507631670353013405 0ustar kripkrip#!/opt/openisis/tcl/tclbuild/tclsh # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: test5.087.tcl,v 1.1 2003/03/06 16:01:47 mawag Exp $ # test script for tcl-binding of openisis # openIsisLog v set ses0 [openIsisInit] set db [openIsisOpen /opt/openisis/db/cds/cds] set res [$db query africa -session $ses0] puts ---------------------------------------------------------------- puts "africa 0 -> [$res list]" $res close set res [$db query africa -mode 1 -session $ses0] puts ---------------------------------------------------------------- puts "africa 1 -> [$res list]" $res close set res [$db query africa$ -mode 2 -session $ses0] puts ---------------------------------------------------------------- puts "africa$ 2 -> [$res list]" $res close # -mode 64 not implemented yet set res [$db query africa -mode 128 -session $ses0] puts ---------------------------------------------------------------- puts "africa 128 -> [$res list]" $res close # -mode 129 no effect set res [$db query universities -mode 0 -session $ses0] puts ---------------------------------------------------------------- puts "universities 0 -> [$res list]" $res close set res [$db query {universities + africa} -mode 128 -session $ses0] puts ---------------------------------------------------------------- puts "universities+africa 128 -> [$res list]" $res close set res [$db query {universities ^ africa} -mode 128 -session $ses0] puts ---------------------------------------------------------------- puts "universities^africa 128 -> [$res list]" $res close set res [$db query {universities africa} -mode 128 -session $ses0] puts ---------------------------------------------------------------- puts "universities africa 128 -> [$res list]" $res close $db close tcl/test/testcli.tcl0100755000113000011300000001117507663442012013463 0ustar kripkrip#!/opt/openisis/tcl/bin/isish # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: testcli.tcl,v 1.3 2003/05/22 15:14:10 mawag Exp $ # test script for db servers client proc checklst {msg exp got} { if [catch { set len1 [llength $exp] set len2 [llength $got] if {$len1 != $len2} { puts "ERR $msg: length != $len1" puts "REC = $got" exit 1 } for {set j 0} {$len1 > $j} {incr j} { set gg [lindex $got $j] set ee [lindex $exp $j] if {! [string equal $gg $ee]} { puts "ERR $msg: got\[$j] = $gg, exp = $ee" puts "REC = $got" exit 1 } } }] { global errorInfo puts $errorInfo puts "REC = $got" exit 1 } } proc checkrec {msg cmd tag val {complete 0}} { set lst "" set witht "" if [catch { set lst [eval $cmd get] set witht [eval $cmd get -tags] set len [llength $val] if $complete { if {[llength $lst] != $len} { puts "ERR $msg: got len = [llength $lst], exp = $len" puts "REC = $witht" exit 1 } } for {set j 0} {$len > $j} {incr j} { set exp [lindex $val $j] if $complete { set got [lindex $lst $j] if {! [string equal $got $exp]} { puts "ERR $msg: rec\[$j] = $got, exp = $exp" puts "REC = $witht" exit 1 } } set fld [lindex $tag $j] set got [eval $cmd get $fld] set got [lindex $got 0] if {! [string equal $got $exp]} { puts "ERR $msg: tag $fld = $got, exp = $exp" puts "REC = $witht" exit 1 } } }] { global errorInfo puts $errorInfo puts "$msg: REC = $witht" exit 1 } } proc checkrsp {msg stb fld val} { checkrec "$msg" "$stb .res" "$fld" "$val" } proc checknoerr {msg stb fld val} { lappend fld sid error error2 lappend val [expr 1 << 8] 0 0 checkrsp "$msg" $stb "$fld" "$val" } proc sendrqs {msg stb type db args} { if [string length $db] { set db "db $db" } if [catch { eval $stb req type $type $db $args $stb .req delete }] { global errorInfo puts $errorInfo puts "$msg: $stb sendrqs($type)" exit 1 } } # ----------------------------------------------------------------- if [catch { set nn [lindex $argv 0] incr nn 0 }] { set nn 1 } if [catch { set dbg "v [lindex $argv 1]" }] { set dbg "" } set verbose [expr 1 == $nn && [string length $dbg]] set stb0 [eval openIsis $dbg] set stb1 [$stb0 new -schema tstsrv -host localhost -port 8080] for {set n $nn} {0 != $n} {incr n -1} { sendrqs RQS_LS $stb1 ls "" checknoerr RSP_LS $stb1 "" "" set dbs [$stb1 .res get db] foreach db $dbs { if {$verbose} { puts "tstsrv has $db" } } set db cds if {0 > [lsearch -exact $dbs $db]} { puts "ERR expected cds db for testing" exit 1 } sendrqs RQS_MOUNT $stb1 mount $db checknoerr RSP_MOUNT $stb1 "" "" set fdt [$stb1 fdt $db] checkrec FDT $fdt {881 {866\[0\]}} {6 conference_main_entry} if {$verbose} { puts [$fdt seri] } $fdt done set maxrow 154 sendrqs RQS_MAXROW $stb1 maxrow $db checknoerr RSP_MAXROW $stb1 904 $maxrow for {set r 1} {$maxrow >= $r} {incr r} { sendrqs RQS_READ($r) $stb1 read $db rowid $r if {23 == $r || 152 <= $r} { set exp "0 0" set fld "total size" } { set exp "1 1 $r" set fld "total size rowid" } checknoerr RSP_READ($r) $stb1 $fld $exp } sendrqs RQS_QUERY $stb1 query $db key africa\$ mode 2 flags 2 checknoerr RSP_QUERY $stb1 {total size} {5 5} if $verbose { for {set i 0} {5 > $i} {incr i} { set ttl "" catch { set ttl [$stb1 .res .rec\[$i\] get 24] } if {60 < [string length $ttl]} { set ttl "[string range $ttl 0 57] ..." } puts "rec($i) [$stb1 .res get row\[$i\]]: $ttl" } } if {0 == [expr $n % 20]} { puts "$n ..." } } $stb1 done puts ok. exit 0 tcl/test/testfdt.tcl0100775000113000011300000000321007654770417013477 0ustar kripkrip#!/opt/openisis/tcl/bin/isish # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: testfdt.tcl,v 1.1 2003/05/03 16:40:47 mawag Exp $ # test script for reading fdt from metadb set syspath /opt/openisis/db/mlcm/ADB set db adb set df ISISMARC/fdttest set stb0 [openIsis syspath $syspath] set cfg [$stb0 .req .config] $cfg add fdt $df $stb0 req type open db $db $stb0 .req delete set fdt [$stb0 fdt $db] set flen [$fdt get flen] for {set j 0} {$flen > $j} {incr j} { set fd [$fdt .fd\[$j\]] puts [$fd get -tag] $fd done } $fdt done $stb0 req type close db $db $stb0 .req delete $stb0 done tcl/test/testrec.tcl0100755000113000011300000001662707660011234013467 0ustar kripkrip#!/opt/openisis/tcl/bin/isish # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: testrec.tcl,v 1.7 2003/05/12 15:41:04 mawag Exp $ # test script for record commands proc checklst {msg exp got} { if [catch { set len1 [llength $exp] set len2 [llength $got] if {$len1 != $len2} { puts "ERR $msg: length != $len1" puts "REC = $got" exit 1 } for {set j 0} {$len1 > $j} {incr j} { set gg [lindex $got $j] set ee [lindex $exp $j] if {! [string equal $gg $ee]} { puts "ERR $msg: got\[$j] = $gg, exp = $ee" puts "REC = $got" exit 1 } } }] { global errorInfo puts $errorInfo puts "REC = $got" exit 1 } } proc checkrec {msg cmd tag val} { set lst "" set tlst "" if [catch { set lst [$cmd get] set tlst [$cmd get -tags] set len [llength $val] if {[llength $lst] != $len} { puts "ERR $msg: got len = [llength $lst], exp = $len" puts "REC = $tlst" exit 1 } for {set j 0} {$len > $j} {incr j} { set got [lindex $lst $j] set exp [lindex $val $j] if {! [string equal $got $exp]} { puts "ERR $msg: rec\[$j] = $got, exp = $exp" puts "REC = $tlst" exit 1 } set fld [lindex $tag $j] set got [$cmd get $fld] if {! [string equal $got $exp]} { puts "ERR $msg: tag $fld = $got, exp = $exp" puts "REC = $tlst" exit 1 } } }] { global errorInfo puts $errorInfo puts "REC = $tlst" exit 1 } } set nn 1 set mm 1 switch $argc { 0 {} 1 { set nn [lindex $argv 0] } default { set nn [lindex $argv 0]; set mm [lindex $argv 1] } } for {set n $nn} {0 != $n} {incr n -1} { # -fdt option set fd [openIsisRec {} add 860 1 862 0 866 name 867 desc] set fdt [openIsisRec {} wrap -done -tag 881 $fd] $fdt add 880 1 set cmd [openIsisRec -fdt $fdt] set len [$cmd fdt get 880] if {1 != $len} { puts "ERR fdt len = $len" puts "FDT = [$cmd fdt get -tags]" exit 1 } set fd2 [$cmd fdt .881] checkrec fd $fd2 {860 862 863 865 866 867} {1 0 0 0 name desc} $fd2 done $fdt done $cmd done set cmd [openIsisRec] checkrec "ctor1" $cmd {} {} for {set m $mm} {0 != $m} {incr m -1} { # add field value ?field value ...? $cmd add 11 22 checkrec add1 $cmd 11 22 $cmd add 33 44 55 66 77 88 checkrec add2 $cmd {11 33 55 77} {22 44 66 88} $cmd add 33 333 77 777 33 3333 checkrec add3 $cmd {11 33\[0] 55 77\[0] 33\[1] 77\[1] 33\[2]} \ {22 44 66 88 333 777 3333} # clone ?options? newname ?field value ...? set clone [$cmd clone] checkrec clone1 $clone {11 33\[0] 55 77\[0] 33\[1] 77\[1] 33\[2]} \ {22 44 66 88 333 777 3333} $clone done set clone [$cmd clone -empty] checkrec clone2 $clone {} {} $clone done set clone [$cmd clone - 99 999 77\[1\] 7171] if [string compare -length 11 $clone openIsisRec] { puts "ERR clone3: recname = $clone" exit 1 } checkrec clone3 $clone {11 33\[0] 55 77\[0] 33\[1] 77\[1] 33\[2] 99} \ {22 44 66 88 333 7171 3333 999} $clone done set clone [$cmd clone -empty bla 99 999 77\[0\] 7171] if ![string equal $clone bla] { puts "ERR clone4: recname = $clone" exit 1 } checkrec clone4 $clone {99 77} {999 7171} $clone done # db ?options? # fdt ?options? # see teststb.tcl # delete ?field ...? $cmd dele 77\[0] 33\[1] checkrec del1 $cmd {11 33\[0] 55 77 33\[1]} \ {22 44 66 777 3333} $cmd delete checkrec del2 $cmd {} {} $cmd add 11 22 33 44 55 66 33 333 33 3333 11 111 checkrec add4 $cmd {11\[0] 33\[0] 55 33\[1] 33\[2] 11\[1]} \ {22 44 66 333 3333 111} # deserialize line, serialize set ser [$cmd seri] $cmd dele checkrec del3 $cmd {} {} $cmd deser $ser checkrec deser $cmd {11\[0] 33\[0] 55 33\[1] 33\[2] 11\[1]} \ {22 44 66 333 3333 111} # do ?tagvar? valvar body $cmd do t v { set t1 [string range $t 0 1] if {11 == $t1 || 55 == $t1} { $cmd add $t1 $v$v } } checkrec do $cmd {11\[0] 33\[0] 55\[0] 33\[1] 33\[2] 11\[1] 11\[2] 55\[1] 11\[3]} \ {22 44 66 333 3333 111 2222 6666 111111} # get ?field ...? set third [$cmd get 33] if ![string equal {44 333 3333} "$third"] { puts "ERR get1: rec\[33] = $third" exit 1 } set got [$cmd get {888 8181}] if ![string equal 8181 $got] { puts "ERR getdflt: rec\[888] = $got" exit 1 } set got [$cmd get] checklst get2 {22 44 66 333 3333 111 2222 6666 111111} $got set got [$cmd get -tag] checklst get3 { 11 22 33 44 55 66 33 333 33 3333 11 111 11 2222 55 6666 11 111111 } $got # set field ?value field value ...? $cmd set 55 5555 77 88 33\[1] 3131 checkrec set $cmd {11\[0] 33\[0] 55 33\[1] 33\[2] 11\[1] 11\[2] 11\[3] 77} \ {22 44 5555 3131 3333 111 2222 111111 88} # wrap ?options? recname set emb [openIsisRec emb set 111 aaa 112 bbb 111 ccc] if ![string equal $emb emb] { puts "ERR emb: recname = $emb" exit 1 } checkrec ctor2 $emb {111\[0] 112 111\[1]} {aaa bbb ccc} $cmd wrap -tag 999 $emb checkrec wrap1 $cmd {11\[0] 33\[0] 55 33\[1] 33\[2] 11\[1] 11\[2] 11\[3] 77 999 111\[0] 112 111\[1]} \ {22 44 5555 3131 3333 111 2222 111111 88 3 aaa bbb ccc} $cmd wrap -tag 888 -num 42 $emb checkrec wrap2 $cmd {11\[0] 33\[0] 55 33\[1] 33\[2] 11\[1] 11\[2] 11\[3] 77 999 111\[0] 112\[0] 111\[1] 888 111\[2] 112\[1] 111\[3]} {22 44 5555 3131 3333 111 2222 111111 88 3 aaa bbb ccc 42 aaa bbb ccc} $emb set 112 ddd $cmd wrap -done -tag 999 $emb checkrec wrap3 $cmd {11\[0] 33\[0] 55 33\[1] 33\[2] 11\[1] 11\[2] 11\[3] 77 999\[0] 111\[0] 112\[0] 111\[1] 888 111\[2] 112\[1] 111\[3] 999\[1] 111\[4] 112\[2] 111\[5]} { 22 44 5555 3131 3333 111 2222 111111 88 3 aaa bbb ccc 42 aaa bbb ccc 3 aaa ddd ccc} if [catch {$emb get}] {} { puts "ERR done1" exit 1 } # .path ?option arg ...? set emb [$cmd .999] checkrec path1 $emb {111\[0] 112 111\[1]} {aaa bbb ccc} $emb done set emb [$cmd .999\[1\]] checkrec path2 $emb {111\[0] 112 111\[1]} {aaa ddd ccc} $emb done set emb [$cmd .999\[1\] get 112] if ![string equal ddd $emb] { puts "ERR path2: got = $emb" exit 1 } $cmd .999\[1\] do t v { set x $v } $cmd .999\[1\] add 42 43 44 45 set emb [$cmd .999\[1\] clone emb] if ![string equal $emb emb] { puts "ERR path3: recname = $emb" exit 1 } checkrec path3 $emb {111\[0] 112 111\[1]} {aaa ddd ccc} $emb done if {! [expr $m % 1000]} { puts "$n $m ..." } $cmd dele } # done $cmd done if [catch {$cmd get}] {} { puts "ERR done2" exit 1 } if {1000 > $mm && ! [expr $n % 1000]} { puts "$n ..." } } puts "ok." tcl/test/testsrv.tcl0100775000113000011300000000274407660761610013537 0ustar kripkrip#!/opt/openisis/tcl/bin/isish # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: testsrv.tcl,v 1.3 2003/05/15 18:59:20 mawag Exp $ # oitsrc for db srv # bin/oits -db 8080 test/testsrv.tcl set syspath /opt/openisis/db/cds # set stb0 [openIsis syspath $syspath v verbose] set stb0 [openIsis syspath $syspath] $stb0 req type open db cds puts "testsrv: db cds opened with err \ [$stb0 .res get error] / [$stb0 .res get error2]" tcl/test/teststb.tcl0100755000113000011300000003120707656730116013511 0ustar kripkrip#!/opt/openisis/tcl/bin/isish # # openisis - an open implementation of the ISIS database # Version 0.8.x (microversion see file Version) # Copyright (C) 2003 by Erik Grziwotz, erik@openisis.org # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # This software is dedicated to the memory of Eckart Dietrich. # # This software is inspired by (but contains no code of) the iAPI # Copyright (C) 2000 by Robert Janusz, rj@jezuici.krakow.pl. # See iAPI.txt for what it contains. # # $Id: teststb.tcl,v 1.16 2003/05/08 18:26:22 mawag Exp $ # test script for stub commands proc checklst {msg exp got} { if [catch { set len1 [llength $exp] set len2 [llength $got] if {$len1 != $len2} { puts "ERR $msg: length != $len1" puts "REC = $got" exit 1 } for {set j 0} {$len1 > $j} {incr j} { set gg [lindex $got $j] set ee [lindex $exp $j] if {! [string equal $gg $ee]} { puts "ERR $msg: got\[$j] = $gg, exp = $ee" puts "REC = $got" exit 1 } } }] { global errorInfo puts $errorInfo puts "REC = $got" exit 1 } } proc checkrec {msg cmd tag val complete} { set lst "" set witht "" if [catch { set lst [eval $cmd get] set witht [eval $cmd get -tags] set len [llength $val] if $complete { if {[llength $lst] != $len} { puts "ERR $msg: got len = [llength $lst], exp = $len" puts "REC = $witht" exit 1 } } for {set j 0} {$len > $j} {incr j} { set exp [lindex $val $j] if $complete { set got [lindex $lst $j] if {! [string equal $got $exp]} { puts "ERR $msg: rec\[$j] = $got, exp = $exp" puts "REC = $witht" exit 1 } } set fld [lindex $tag $j] set got [eval $cmd get $fld] set got [lindex $got 0] if {! [string equal $got $exp]} { puts "ERR $msg: tag $fld = $got, exp = $exp" puts "REC = $witht" exit 1 } } }] { global errorInfo puts $errorInfo puts "$msg: REC = $witht" exit 1 } } proc checkrsp {msg stb fld val} { checkrec "$msg" "$stb .res" "$fld" "$val" 0 } proc checknoerr {msg stb fld val} { lappend fld sid error error2 lappend val 0 0 0 checkrsp "$msg" $stb "$fld" "$val" } proc sendrqs {msg stb type args} { global db if [catch { eval $stb req type $type db $db $args $stb .req delete }] { global errorInfo puts $errorInfo puts "$msg: $stb sendrqs($type)" exit 1 } } proc puterrline {} { global errorInfo set idx [string first \n $errorInfo] if {0 < $idx} { puts [string range $errorInfo 0 [incr idx -1]] } { puts $errorInfo } } proc clntestdb dbn { if [catch { file delete $dbn.mst }] { puterrline } if [catch { file delete $dbn.oxi }] { puterrline } if [catch { file delete $dbn.xrf }] { puterrline } } proc opentestdb {stb db {msg open}} { set fd1 [openIsisRec] $fd1 add 860 10 862 0 863 1 865 256 866 name 867 "" set fd2 [openIsisRec] $fd2 add 860 11 862 0 863 1 865 256 866 phone 867 "" set fd3 [openIsisRec] $fd3 add 860 12 862 0 863 1 865 256 866 title 867 "" set fd4 [openIsisRec] $fd4 add 860 13 862 0 863 1 865 256 866 street 867 "" set fd5 [openIsisRec] $fd5 add 860 20 862 0 863 1 865 256 866 city 867 "" set fd6 [openIsisRec] $fd6 add 860 88 862 0 863 1 865 256 866 summary 867 "" set fd61 [openIsisRec] $fd61 add 860 88 861 q 862 0 863 0 865 256 866 overview 867 "" set fdt [$stb .req .fdt] $fdt add flen 7 $fdt wrap -done fd $fd1 $fdt wrap -done fd $fd2 $fdt wrap -done fd $fd3 $fdt wrap -done fd $fd4 $fdt wrap -done fd $fd5 $fdt wrap -done fd $fd6 $fdt wrap -done fd $fd61 sendrqs $msg $stb open checknoerr $msg $stb {db dbid} "$db 0" } if 0 { proc reopendb {stb db msg} { sendrqs close$msg $stb close checknoerr close$msg $stb db $db opentestdb $stb $db open$msg } } set db testdb set syspath /opt/openisis/db/cds set nn 1 set mm 1 set debug -1 if {$argc} { set x [lindex $argv 0] if [string length $x] { set syspath $x } unset x if {1 < $argc} { set nn [lindex $argv 1] if {2 < $argc} { set mm [lindex $argv 2] if {3 < $argc} { set debug [lindex $argv 3] } } } } for {set n $nn} 1 {} { set args "syspath $syspath" if ![string equal $debug -1] { set args "$args v $debug" } set stb [eval openIsis $args] for {set m $mm; set hasdb 0} {0 != $m} {incr m -1; set hasdb 1} { # rqs open if !$hasdb { puts "using $syspath/$db" clntestdb $syspath/$db opentestdb $stb $db } # rqs ls sendrqs ls $stb ls checknoerr ls $stb db $db # fdt db ?option ...? set fdt [$stb fdt $db] checkrec fdt1 $fdt flen 7 0 checkrec fd11 "$fdt .fd\\\[0]" {tag type rep len name descr} \ {10 0 1 256 name {}} 1 checkrec fd12 "$fdt .fd\\\[1]" {tag type rep len name descr} \ {11 0 1 256 phone {}} 1 checkrec fd13 "$fdt .fd\\\[2]" {tag type rep len name descr} \ {12 0 1 256 title {}} 1 checkrec fd14 "$fdt .fd\\\[3]" {tag type rep len name descr} \ {13 0 1 256 street {}} 1 checkrec fd15 "$fdt .fd\\\[4]" {tag type rep len name descr} \ {20 0 1 256 city {}} 1 checkrec fd16 "$fdt .fd\\\[5]" {tag type rep len name descr} \ {88 0 1 256 summary {}} 1 checkrec fd161 "$fdt .fd\\\[6]" {tag sub type rep len name descr} \ {88 q 0 0 256 overview {}} 1 $fdt done checkrec fd21 "$stb fdt $db .fd\\\[0]" \ {tag type rep len name descr} \ {10 0 1 256 name ""} 1 checkrec fd22 "$stb fdt $db .fd\\\[1]" \ {tag type rep len name descr} \ {11 0 1 256 phone ""} 1 checkrec fd23 "$stb fdt $db .fd\\\[2]" \ {tag type rep len name descr} \ {12 0 1 256 title ""} 1 checkrec fd24 "$stb fdt $db .fd\\\[3]" \ {tag type rep len name descr} \ {13 0 1 256 street ""} 1 checkrec fd25 "$stb fdt $db .fd\\\[4]" \ {tag type rep len name descr} \ {20 0 1 256 city ""} 1 checkrec fd26 "$stb fdt $db .fd\\\[5]" \ {tag type rep len name descr} \ {88 0 1 256 summary ""} 1 checkrec fd261 "$stb fdt $db .fd\\\[6]" \ {tag sub type rep len name descr} \ {88 q 0 0 256 overview ""} 1 checkrec fdsys "$stb fdt -sys .fd\\\[0]" \ {tag name} {5 syspath} 0 checkrec fdsch "$stb fdt -sche .fd\\\[0]" \ {tag name} {710 name} 0 checkrec fdsch "$stb fdt -sche .fd\\\[1]" \ {tag name} {711 host} 0 checkrec fddb "$stb fdt -db .fd\\\[0]" \ {tag name} {800 db} 0 checkrec fdfd "$stb fdt -fd .fd\\\[0]" \ {tag name} {860 tag} 0 checkrec fdfd "$stb fdt -fdt .fd\\\[0]" \ {tag name} {880 flen} 0 checkrec fdrqs0 "$stb fdt -req .fd\\\[0]" \ {tag name} {900 sid} 0 checkrec fdrqs10 "$stb fdt -req .fd\\\[10]" \ {tag name} {920 type} 0 checkrec fdrsp0 "$stb fdt -res .fd\\\[0]" \ {tag name} {900 sid} 0 checkrec fdrsp10 "$stb fdt -res .fd\\\[10]" \ {tag name} {940 dbid} 0 # db db ?option ...? set rec [$stb db $db] checkrec meta $rec {800 803} "$db $syspath" 0 rename $rec "" if [catch {$rec get}] {} { puts "ERR meta: $rec still present" exit 1 } # new -schema name ?-cfg val ...? set rec [$stb new -sche testsrv -host localhost -port 3434] checkrec schema $rec {711 712} {localhost 3434} 0 rename $rec "" set rec [$stb new -sche testsrv -host localhost -port 3434] checkrec schema $rec {711 712} {localhost 3434} 0 $rec done # new ?-db db? ?name? # $rec db # $rec fdt if [catch {set rec [$stb new]}] {} { puts "ERR newrec1 = $rec" exit 1 } set rec [$stb new -db $db] checkrec recfdt1 "$rec fdt .fd\\\[1]" \ {tag name} {11 phone} 0 if [catch { set dbn [$rec db get db] if ![string equal $dbn $db] { puts "ERR recdb = $dbn" exit 1 } }] { global errorInfo puts $errorInfo puts "recdb failed" exit 1 } set rec [$stb new -db -sys $rec] checkrec recfdt2 "$rec fdt .fd\\\[1]" \ {tag name} {700 logfile} 0 set rec [$rec clone $rec] checkrec recfdt3 "$rec fdt .fd\\\[1]" \ {tag name} {700 logfile} 0 set dbn "" if [catch {set dbn [$rec db get]}] {} { puts "ERR recdb = $dbn" exit 1 } $stb add defaultdb $db if [catch {set rec [$stb new $rec]}] { global errorInfo puts $errorInfo puts "newrec2 failed" exit 1 } checkrec recfdt4 "$rec fdt .fd\\\[1]" \ {tag name} {11 phone} 0 $stb del defaultdb $rec done set rec [$stb clone] checkrec recfdt5 $rec syspath $syspath 0 checkrec recfdt6 "$rec fdt .fd\\\[0]" \ {tag name} {5 syspath} 0 $rec done # rqs insert if !$hasdb { set rec [$stb .req -db $db .rec] set idx [$stb .req .idx] $rec add name harry phone 4711 city montevideo # XDOT [[dbn.]mfn.]tag.occ.pos or tag[.occ] $idx add 10 harry 10 hurry sendrqs insert1 $stb insert checknoerr insert1 $stb {db rowid} "$db 1" set row 1 } if {1000 > $row} { set rec [$stb .req -db $db .rec] set idx [$stb .req .idx] set phone "+4711" $rec add name mary phone $phone summary {Meeting Implementation} $idx add 11 $phone 11 marry sendrqs insert2 $stb insert checknoerr insert2 $stb {} {} set row2 [$stb .res get rowid] if {[incr row] != $row2} { puts "ERR insert2: rowid == $row2, exp = $row" exit 1 } } # rqs maxrow sendrqs maxrow1 $stb maxrow checknoerr maxrow1 $stb rowid $row # .req ?-db db? ?option ...? # rqs update set phone "+49-30-$row" set rec [$stb .req .rec] $rec add name lary phone $phone street chausseestr $stb .req -db $db set rowid $row sendrqs update1 $stb update rowid $row checknoerr update1 $stb rowid $row set idx [$stb .req .idx] $idx add -3 "-11.1\tmarry" sendrqs update2 $stb update rowid $row checknoerr update2 $stb rowid $row set idx [$stb .req .idx] $idx add -3 "11\tlary" sendrqs update3 $stb update rowid $row checknoerr update3 $stb rowid $row # rqs maxrow sendrqs maxrow2 $stb maxrow checknoerr maxrow2 $stb rowid $row # recv if [catch { $stb recv if ![$stb .res] { puts "ERR: .res == 0" exit 1 } }] { global errorInfo puts $errorInfo puts "recv failed" exit 1 } # rqs read (row 1) $stb .req set rowid 1 sendrqs read1 $stb read checknoerr read1 $stb {total size rowid} "1 1 1" # checknoerr read1 $stb "" "" set rec [$stb .res .rec] checkrec read1/fdt "$rec fdt .fd\\\[1]" \ {tag name} {11 phone} 0 if [catch { set dbn [$rec db get db] if ![string equal $dbn $db] { puts "ERR read1/recdb = $dbn" exit 1 } }] { global errorInfo puts $errorInfo puts "read1/recdb failed" exit 1 } checkrec read1 $rec {name phone city} {harry 4711 montevideo} 1 $rec done # rqs read (row n) for {set r 2} {$row >= $r} {incr r} { set msg "read($r)" $stb .req set rowid $r sendrqs $msg $stb read checknoerr $msg $stb {total size rowid} "1 1 $r" set rec [$stb .res .rec] checkrec $msg/fdt "$rec fdt .fd\\\[1]" \ {tag name} {11 phone} 0 if [catch { set dbn [$rec db get db] if ![string equal $dbn $db] { puts "ERR $msg/recdb = $dbn" exit 1 } }] { global errorInfo puts $errorInfo puts "$msg/recdb failed" exit 1 } set phone "+49-30-$r" checkrec $msg $rec {name phone street} "lary $phone chausseestr" 1 set rec2 [$rec clone] set rero [$rec2 row] $rec2 done if {$rero != $r} { puts "ERR $msg: rowid = $rero, exp: $r" exit 1 } if {2 == $r} { set got [$rec get -tag] checklst gettags \ "10 lary 11 $phone 13 chausseestr" $got set got [$rec get -tagn] checklst gettagn \ "name lary phone $phone street chausseestr" $got } } # rqs query set tot [expr $row - 1] set siz 100 if {100 >= $row} { set siz $tot } $stb .req set typ query key marr size 100 flags 2 $stb req mode 1 checknoerr query $stb {total size} "$tot $siz" for {set j 0} {$siz > $j} {incr j} { set r [$stb .res get rowid\[$j\]] set phone "+49-30-$r" set rec [$stb .res .rec\[$j\]] checkrec "query\[$j]" $rec {name phone street} \ "lary $phone chausseestr" 1 set rero [$rec row] if {$rero != $r} { puts "ERR query\[$j]: rowid = $rero, exp: $r" exit 1 } } if ![expr $m % 100] { puts "$n $m ($row) ..." } } if {100 > $mm && ! [expr $n % 100]} { puts "$n ..." } # rqs close if ![incr n -1] { sendrqs close $stb close checknoerr close $stb db $db clntestdb $syspath/$db break } # done $stb done if [catch {$stb get}] {} { puts "ERR done" exit 1 } } puts "ok." tcl/tisis0100755000113000011300000002526007656444377011431 0ustar kripkrip#!/opt/isis/OpenIsis/tcl/bin/wisish # if wisish is in the PATH, but we don't know where, delete this first 2 lines #!/bin/sh # let editor know it's Tcl vim:syntax=tcl # the old Tcl trick of calling wisish regardless where it's installed # sh sees 1 comment line and execs wisish; Tcl joins lines, ignoring exec \ exec wisish "$@" # load the openIsis.tcl code # if we are NOT the wisish, but a standard wish, # this should also load the dynamic OpenIsis/Tcl lib package require openIsis 0.8 namespace import openIsis::isis openIsis::v openIsis::res # ############################################################ # global variables # set isis noserver ;# will give error set mfns {} ;# empty set idx 0 set max 0 set mfn 0 ;# this is a workaround; record should tell it's mfn # ############################################################ # first define procedures # # set the global max idx proc max {} { global max mfns # if the list is empty, get max idx if {0 == [set max [llength $mfns]]} { isis maxrow set max [res get {rowid 0}] } } # check wether an increment of incr (default 1) would give a legal index proc haveNext {{incr 1}} { global idx max set nidx [expr $idx + $incr] expr 0 < $nidx && $nidx <= $max } # get mfn for index i (or current idx, if not given) # if we have a set of mfns, idx is index into this proc mfn {{i ""}} { global idx max mfns if {"" == $i} { set i $idx } if {"" == $mfns} { # direct mfn runs 1..max return $i } else { # list index is 0..max-1 return [lindex $mfns [expr $i - 1]] } } # get next mfn based on global variable idx (after incrementing by idx) # if we have a set of mfns, idx is index into this proc next {{incr 1}} { global idx max mfns incr idx $incr if {$idx < 1} { set idx 1 } elseif {$idx > $max} { set idx $max } mfn } # render all mfns proc all {{from 0}} { global idx max .main.text clear max ;# determine maximum set limit 100 for {set idx $from} {[haveNext]} {} { if {! [isis read -rowid [next]]} { puts "no record for idx $idx:" continue } if {[catch {openIsis::Tspr list.tsp}]} { global errorInfo puts stderr "error in list.tsp:\n$errorInfo" break } if [incr limit -1] continue openIsis::puts "
more" break } } ;# all # show detail for index i proc detail {i} { global idx max set idx $i isis read -rowid [mfn] .main.text clear .main.text configure -base /idx/$idx # print application header openIsis::puts "

detail for MFN [v rowid] ($idx of $max)

" if {[haveNext -1]} { set prev [expr $idx-1] openIsis::puts "prev ($prev) " } openIsis::puts {all edit } if {[haveNext]} { set next [expr $idx+1] openIsis::puts "next ($next)
" } openIsis::puts {
} if {[catch {openIsis::Tspr detail.tsp}]} { global errorInfo puts stderr "error in detail.tsp:\n$errorInfo" } } # run a query, display results proc runQuery {qry} { global mfns puts "query $qry" isis query -mode 128 -key $qry set mfns [res get {rowid {}}] puts "mfns for '$qry': $mfns" all } proc showText {} { pack .main.text -side left -expand yes -fill both pack .main.sbar -side right -fill y ;# cuts needed width from right } # create the edit dialog toplevel .main.edlg # to edit existing (if mfn != 0) or new record proc editDialog {{mfn 0}} { pack forget .main.text .main.sbar destroy .main.edlg if {$mfn} { if {! [isis read -rowid $mfn]} { tk_messageBox -icon error \ -message "could not read $mfn" -type ok return } } frame .main.edlg if {$mfn} { v do tag val { set row [lindex [grid size .main.edlg] 1] label .main.edlg.tag$row -text $tag entry .main.edlg.val$row -width 80 .main.edlg.val$row insert 0 $val grid .main.edlg.tag$row .main.edlg.val$row -sticky we } } # create a frame for button bar frame .main.edlg.bbar button .main.edlg.bbar.atag -text "add field tag:" -command addNewField entry .main.edlg.bbar.ntag -width 3 button .main.edlg.bbar.save -text "save record as MFN:" -command saveRecord entry .main.edlg.bbar.mfn -width 6 .main.edlg.bbar.mfn insert 0 $mfn button .main.edlg.bbar.close -text "close" -command closeEdit pack .main.edlg.bbar.atag .main.edlg.bbar.ntag \ .main.edlg.bbar.save .main.edlg.bbar.mfn .main.edlg.bbar.close \ -side left -fill x -expand yes grid .main.edlg.bbar -columnspan 2 -sticky we # make the first column not stretching, give all space to 2nd grid columnconfigure .main.edlg 0 -weight 0 grid columnconfigure .main.edlg 1 -weight 1 pack .main.edlg -side left -expand yes -fill both } proc addNewField {} { set tag [.main.edlg.bbar.ntag get] if {![string is integer -strict $tag]} { tk_messageBox -icon error \ -message "please specify integer tag" -type ok return } set row [lindex [grid size .main.edlg] 1] ;# number of rows in dialog # mv the bar downwards grid configure .main.edlg.bbar -row $row incr row -1 label .main.edlg.tag$row -text $tag entry .main.edlg.val$row -width 80 grid .main.edlg.tag$row .main.edlg.val$row -row $row -sticky we } proc saveRecord {} { # get the mfn from the entry in the button bar set mfn [.main.edlg.bbar.mfn get] # prepare new record $openIsis::srv new rec set rows [lindex [grid size .main.edlg] 1] ;# number of rows in dialog incr rows -1 ;# don't count the button bar # get tags and values for {set row 0} {$row < $rows} {incr row} { set tag [.main.edlg.tag$row cget -text] set val [.main.edlg.val$row get] rec add $tag $val } set mfn [openIsis::save rec $mfn] tk_messageBox -icon info -message "saved as MFN $mfn" -type ok closeEdit } proc closeEdit {} { destroy .main.edlg showText } # ############################################################ # procedures called from menu and other interface events # # the tsps use openIsis::puts -- have them render to then text widget proc openIsis::puts msg { .main.text parse $msg } # a click in the .main.text widget proc htmlclick {x y} { set ref [.main.text href $x $y] if {"" == $ref} return puts "hyperlink: $ref" if {"/all" == $ref} { all } elseif [scan $ref "/all/%d" i] { all $i } elseif [scan $ref "/idx/%d" i] { detail $i } elseif {"/edit" == $ref} { command_edit [mfn] } } proc command_open {} { puts "command_open" } proc command_close {} { puts "hasta la vista" isis close exit } proc command_edit {{mfn 0}} { puts "command_edit" editDialog $mfn } proc command_search {} { puts "command_search" # open a query dialog in a new toplevel window named .qdlg # TODO for Tk experts: build a guided search dialog with AND,OR,NOT choice destroy .qdlg ;# destroy the toplevel, if existed toplevel .qdlg entry .qdlg.e .qdlg.e insert 0 water button .qdlg.b -text query -command runQueryFromDialog pack .qdlg.e .qdlg.b -fill both focus .qdlg.e } # called by the button in the query dialog proc runQueryFromDialog {} { runQuery [.qdlg.e get] ;# what's in the entry widget } # called by the button in the menu bar proc runQueryFromMBar {} { runQuery [.mbar.e get] ;# what's in the entry widget } # ############################################################ # configuration # # TODO: check the syspar # openIsisRec syspar # set file [open syspar.par] # syspar deserialize [read $file] # close $file # set defaultdb [syspar get 1001] # defaults set syspath ../db set encoding "" set defaultdb cds/cds # parameters # can't use -v for verbose; TkMain will take it as "-visual" foreach {opt val} $argv { switch -glob -- $opt \ -syspath { set syspath $val } \ -verbose - -ver { set openIsis::ver 1 } \ -encoding { set encoding $val } \ -defaultdb - -db { set defaultdb $val } \ -* { puts stderr "unknown option '$opt'" } \ default { if {"" == $val} { # probably empty last param ? set defaultdb $opt } } } # a record to hold menu labels # TODO: we should try to read this from a "message" database openIsisRec menulabels # set fallback values in english # 1: labels for menubar buttons file ... set menubar {file edit} menulabels add 1 File 1 Edit # 2: labels for File menu open close ... set menu(file) {open close} menulabels add 2 Open 2 Close set menu(edit) {edit search} menulabels add 3 Edit 3 Search if {[info exists env(LANG)] && "es_AR" == $env(LANG)} { # we're en Argentina # overwrite main menubar labels: # menulabels set 1 Archivo 1 Edit } puts "menu language:\n[menulabels serialize]" # ############################################################ # now we have all procedures defined and the config -- create the windows # # create a frame to hold the menubuttons # remove ;# to give it a raised relief frame .mbar ;# -relief raised -bd 2 # create the menubar, looping texts over fields with tag 1 set texttag 1 foreach button $menubar text [menulabels get $texttag] { puts "creating menu $button labeled $text" # underline the first character as "hotkey" menubutton .mbar.$button -text $text -underline 0 -menu .mbar.$button.menu pack .mbar.$button -side left # create the menu belonging to this button menu .mbar.$button.menu incr texttag ;# next number foreach entry $menu($button) text [menulabels get $texttag] { .mbar.$button.menu add command -label $text -command command_$entry } } # add a quick expert search entry .mbar.e -width 20 .mbar.e insert 0 water button .mbar.b -text query -command runQueryFromMBar pack .mbar.b -side right pack .mbar.e -side right -expand yes -fill x frame .main ;# yeah, it runs on a mainframe !;) # create the html widget html .main.text ;# -relief raised -bd 2 # add an imagecommand to deliver images for tags # see the TkHtml doku and the tisis example how this works # .main.text configure -imagecommand imgcb # give it a scrollbar to talk with scrollbar .main.sbar -command ".main.text yview" .main.text configure -yscrollcommand ".main.sbar set" showText # pack the main frames in our toplevel window pack .mbar -side top -fill x ;# cuts needed height from top pack .main -side left -expand yes -fill both ;# expands to all remaining space # non-standard event bindings: # search on s # bind . command_search # activate hyperlinks: # bind a mouse button 1 click on the htmlwidget('s inner window .x) # to call procedure htmlclick with the coordinates bind .main.text.x <1> {htmlclick %x %y} # ############################################################ # we have the windows -- startup # # connect to the "local server", i.e. the builtin request dispatcher # actually the value (command name) will always be openIsisRoot, # but we use a variable so we could switch between servers set openIsis::srv [openIsis] # open the default db $openIsis::srv add -syspath $syspath -defaultdb $defaultdb if {"" == $encoding && [string equal cds/cds $defaultdb]} { set encoding cp850 } if {"" != $encoding} { $openIsis::srv add -encoding $encoding } # Open Isis isis open; all tcl/tkhtml/Makefile0100664000113000011300000000660707646054635013312 0ustar kripkrip# manually generated by KriP # Do Edit! ifeq (,$(TKHTML)) TKHTML=/opt/TclTk/htmlwidget endif S=$(TKHTML)/src ifeq (,$(WIN32)) include ../make.tcltk OSDEF=-DOS_UNIX=1 -DOS_WIN=0 PIC=-fPIC else include ../../make.mingw include ../make.tcltk.mingw OSDEF=-DOS_UNIX=0 -DOS_WIN=1 endif CC=gcc RANLIB=ranlib ifeq ($(NDEBUG),) DBG=-g else DBG=-DNDEBUG endif CPPFLAGS=$(TCL_DEFS) $(OSDEF) -I. $(TCL_INCLUDE_SPEC) $(DBG) CFLAGS=-O2 -Wall -W -Wshadow -pedantic -Wno-long-long ifeq (,$(TCL_SHLIB_LD)) all: headers libtkhtml.a else all: headers libtkhtml.a libtkhtml.so pkgIndex.tcl endif # use native compiler makeheaders: $(TKHTML)/tools/makeheaders.c : this may fail if cc uses the cross compilers as : try to run it manually /usr/bin/cc -o makeheaders $(TKHTML)/tools/makeheaders.c htmltokens.c: $S/tokenlist.txt $(TKHTML)/tools/maketokens.tcl tclsh $(TKHTML)/tools/maketokens.tcl $S/tokenlist.txt >htmltokens.c pkgIndex.tcl: libtkhtml.so echo 'pkg_mkIndex . libtkhtml.so' | tclsh # variants of the builtin rules %.o: $S/%.c $(CC) -c $(CPPFLAGS) $(CFLAGS) -o $@ $< sh%.o: $S/%.c $(CC) -c $(CPPFLAGS) $(TCL_SHLIB_CFLAGS) $(CFLAGS) -o $@ $< sh%.o: %.c $(CC) -c $(CPPFLAGS) $(TCL_SHLIB_CFLAGS) $(CFLAGS) -o $@ $< # deps htmlcmd.o: $S/htmlcmd.c htmlcmd.h htmldraw.o: $S/htmldraw.c htmldraw.h htmlform.o: $S/htmlform.c htmlform.h htmlimage.o: $S/htmlimage.c htmlimage.h htmlindex.o: $S/htmlindex.c htmlindex.h htmllayout.o: $S/htmllayout.c htmllayout.h htmlparse.o: $S/htmlparse.c htmlparse.h htmlsizer.o: $S/htmlsizer.c htmlsizer.h htmltable.o: $S/htmltable.c htmltable.h htmltest.o: $S/htmltest.c htmltest.h htmltokens.o: htmltokens.c htmltokens.h htmlurl.o: $S/htmlurl.c htmlurl.h htmlwidget.o: $S/htmlwidget.c htmlwidget.h shhtmlcmd.o: $S/htmlcmd.c htmlcmd.h shhtmldraw.o: $S/htmldraw.c htmldraw.h shhtmlform.o: $S/htmlform.c htmlform.h shhtmlimage.o: $S/htmlimage.c htmlimage.h shhtmlindex.o: $S/htmlindex.c htmlindex.h shhtmllayout.o: $S/htmllayout.c htmllayout.h shhtmlparse.o: $S/htmlparse.c htmlparse.h shhtmlsizer.o: $S/htmlsizer.c htmlsizer.h shhtmltable.o: $S/htmltable.c htmltable.h shhtmltest.o: $S/htmltest.c htmltest.h shhtmltokens.o: htmltokens.c htmltokens.h shhtmlurl.o: $S/htmlurl.c htmlurl.h shhtmlwidget.o: $S/htmlwidget.c htmlwidget.h libtkhtml.a: htmlcmd.o htmldraw.o htmlform.o htmlimage.o htmlindex.o htmllayout.o htmlparse.o htmlsizer.o htmltable.o htmltest.o htmlurl.o htmlwidget.o htmltokens.o ar cr $@ $^ ranlib $@ libtkhtml.so: shhtmlcmd.o shhtmldraw.o shhtmlform.o shhtmlimage.o shhtmlindex.o shhtmllayout.o shhtmlparse.o shhtmlsizer.o shhtmltable.o shhtmltest.o shhtmlurl.o shhtmlwidget.o shhtmltokens.o strip $^ $(TCL_SHLIB_LD) -o $@ $^ \ $(TCL_LIB_SPEC) $(TK_LIB_SPEC) $(TCL_LD_SEARCH_FLAGS) $(TK_LIBS) headers: makeheaders htmltokens.c ./makeheaders $S/htmlcmd.c:htmlcmd.h $S/htmldraw.c:htmldraw.h $S/htmlform.c:htmlform.h $S/htmlimage.c:htmlimage.h $S/htmlindex.c:htmlindex.h $S/htmllayout.c:htmllayout.h $S/htmlparse.c:htmlparse.h $S/htmlsizer.c:htmlsizer.h $S/htmltable.c:htmltable.h $S/htmltest.c:htmltest.h $S/htmlurl.c:htmlurl.h $S/htmlwidget.c:htmlwidget.h htmltokens.c $S/html.h clean: rm -f makeheaders *.o *.h *.a *.so htmltokens.c pkgIndex.tcl rm -rf srcdir install: all cp libtkhtml.a $(TK_PREFIX)/lib ifneq (,$(TCL_SHLIB_LD)) [ -d $(TK_PREFIX)/lib/Tkhtml ] || mkdir $(TK_PREFIX)/lib/Tkhtml cp libtkhtml.so $(TK_PREFIX)/lib/Tkhtml cp pkgIndex.tcl $(TK_PREFIX)/lib/Tkhtml endif tcl/tspr.c0100644000113000011300000001066007664637050011471 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: tspr.c,v 1.5 2003/05/26 19:17:48 mawag Exp $ OpenIsis TSP runner must be binary in order to use with #! lines */ #include /* getenv */ #include /* read/write */ #include "openisis.h" #include "openisistcl.h" #define CL_MAX 512*1024 static int urldec ( OpenIsisRec **preq, const char *q, int l ) { OpenIsisRec *req = *preq; const char *e = q + (0 <= l ? l : (l = strlen(q))); while ( *q ) { char *v, *p; int tag; /* assume numeric tags */ q += openIsisA2il( q, e-q, &tag ); if ( '=' != *q || 0 >= tag ) { openIsisSMsg( OPENISIS_LOG_ERROR, "bad param %d '%.32s'", tag, q ); return 1; } q++; OPENISIS_RADD( req, tag, 0, e-q, !0 ); if ( ! req ) return 3; p = v = (char*)req->field[req->len-1].val; for (;;p++,q++) { switch (*q) { case '+': *p = ' '; continue; default: *p = *q; continue; case '%': /* dehex 2 */ *p = 0; q++; if ( '0' <= *q && *q <= '9' ) *p += *q - '0'; else if ( 'A' <= *q && *q <= 'F' ) *p += *q - 'A' + 10; else if ( 'a' <= *q && *q <= 'f' ) *p += *q - 'a' + 10; else return 2; /* carp bad hex */ *p <<= 4; q++; if ( '0' <= *q && *q <= '9' ) *p += *q - '0'; else if ( 'A' <= *q && *q <= 'F' ) *p += *q - 'A' + 10; else if ( 'a' <= *q && *q <= 'f' ) *p += *q - 'a' + 10; else return 2; continue; case '&': q++; case 0: break; } /* hit proper end */ req->used += req->field[req->len-1].len = p - v; break; } } /* while q */ *preq = req; return 0; } #ifdef ORATCL extern int Oratcl_Init ( Tcl_Interp *ip ); #endif #ifdef PGTCL extern int Pgtcl_Init ( Tcl_Interp *ip ); #endif static int appinit ( Tcl_Interp *interp ) { if ( TCL_OK != Tcl_Init( interp ) /* done in new session || TCL_OK != OpenisisTcl_Init( interp ) */ #ifdef ORATCL || TCL_OK != Oratcl_Init( interp ) #endif #ifdef PGTCL || TCL_OK != Pgtcl_Init( interp ) #endif ) return TCL_ERROR; return TCL_OK; } static char prelude[] = "Content-Type: text/html\r\n\r\n"; static char code[] = "package require openIsis 0.8\nopenIsis::Tspr $tsp"; int main ( int argc, const char **argv ) { OpenIsisRec *req = 0; int sid, rid, ret; const char *q; Tcl_Interp *interp; if ( 2 != argc ) return 42; openIsisCOpen(0); /* Tcl_Main( argc, argv, appinit ); */ Tcl_FindExecutable(argv[0]); /* adds ../lib to auto_path */ interp = Tcl_CreateInterp(); appinit( interp ); if ( (q = getenv("SCRIPT_NAME")) ) OPENISIS_RADDS( req, -10, q, !0 ); if ( (q = getenv("PATH_INFO")) ) OPENISIS_RADDS( req, -11, q, !0 ); if ( (q = getenv("REMOTE_USER")) ) OPENISIS_RADDS( req, -12, q, !0 ); /* URL-decode env(QUERY_STRING) into request record */ if ( (q = getenv("QUERY_STRING")) && (ret = urldec( &req, q, -1 )) ) return ret; if ( (q = getenv("CONTENT_LENGTH")) && *q ) { char buf[CL_MAX]; int l = openIsisA2i( q, -1 ); if ( l > CL_MAX ) return 4; if ( !(q = getenv("CONTENT_TYPE")) || strcmp("application/x-www-form-urlencoded", q) ) return 5; if ( l != read( 0, buf, l ) ) return 6; if ( (ret = urldec( &req, buf, l )) ) return ret; } sid = openIsisTclNewSession( interp ); rid = openIsisTclCreateRecCmd( sid, "request", 0 ); Tcl_SetVar( interp, "tsp", (char*)argv[1], 0 ); write( 1, prelude, sizeof(prelude)-1 ); if ( openIsisTclEval( sid, 1, &rid, &req, code ) ) { const char *errorInfo = Tcl_GetVar( interp, "errorInfo", TCL_GLOBAL_ONLY ); openIsisSMsg( OPENISIS_LOG_ERROR, "error in tsp\n%s", errorInfo ? errorInfo : "-" ); return 3; } return 0; } tcl/wisish.c0100644000113000011300000000355607664637050012015 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: wisish.c,v 1.7 2003/05/26 19:17:48 mawag Exp $ OpenIsis Tcl shell */ #include "openisis.h" #include "openisistcl.h" #include #ifdef TKHTML extern int Tkhtml_Init ( Tcl_Interp *ip ); #endif #ifdef TIX extern int Tix_Init ( Tcl_Interp *ip ); #endif #ifdef ORATCL extern int Oratcl_Init ( Tcl_Interp *ip ); #endif #ifdef PGTCL extern int Pgtcl_Init ( Tcl_Interp *ip ); #endif static int appinit ( Tcl_Interp *interp ) { if ( TCL_OK != Tcl_Init( interp ) || TCL_OK != Tk_Init( interp ) || TCL_OK != openIsisTclInit ( interp ) #ifdef TKHTML || TCL_OK != Tkhtml_Init( interp ) #endif #ifdef TIX || TCL_OK != Tix_Init( interp ) #endif #ifdef ORATCL || TCL_OK != Oratcl_Init( interp ) #endif #ifdef PGTCL || TCL_OK != Pgtcl_Init( interp ) #endif ) return TCL_ERROR; Tcl_StaticPackage( interp, "Tk", Tk_Init, Tk_SafeInit ); return TCL_OK; } int main ( int argc, char **argv ) { openIsisCOpen(0); Tk_Main( argc, argv, appinit ); return 0; } test.pft0100664000113000011300000000023707505600362011234 0ustar kripkripMHL "V26" V26, "a" V71 'b', "V71[2]:" V71[2], ( "V71:" V71, |V26^*[2..]| V26^*[2..] |b| ) "V69" V69, /* arithmetic */ '123.4 + 7 / 2 + 5 = ' 123.4 + 7 / 2 + 5 testres.txt0100644000113000011300000010733207665600575012013 0ustar kripkrip47.24=Hydrological achievements and social problems 47.26=^c1966 47.26.c=1966 47.30=^ap. 357-359 47.30.a=p. 357-359 47.44=Scientific problems of the humid tropical zone deltas and their implications: proceedings of the Dacca Symposium 47.69=Paper on: 47.70=Versluys, J.D.N. 50.24=Influence of man's traditional activities on soils and vegetation 50.26=^c1966 50.26.c=1966 50.30=^ap. 301-304 50.30.a=p. 301-304 50.44=Scientific problems of the humid tropical zone deltas and their implications: proceedings of the Dacca Symposium 50.69=Paper on: 50.70=Boesch, H. 51.24=Deltas: a basis of culture and civilization 51.26=^c1966 51.26.c=1966 51.30=^ap. 295-300 51.30.a=p. 295-300 51.44=Scientific problems of the humid tropical zone deltas and their implications: proceedings of the Dacca Symposium 51.69=Paper on: 51.70=B<=ue>del, J. 71.24=Agriculture with special reference to rice cultivation in humid tropical zone deltas 71.26=^c1966 71.26.c=1966 71.30=^ap. 305-316 71.30.a=p. 305-316 71.44=Scientific problems of the Dacca Symposium 71.50=Incl. bibl. 71.69=Paper on: 71.70=Goor, G.A.W. van de 73.24=Need for land-use studies in East Pakistan 73.26=^c1966 73.26.c=1966 73.30=^ap. 329-332 73.30.a=p. 329-332 73.44=Scientific problems of the humid tropical zone deltas and their implications: proceedings of the Dacca Symposium 73.50=Incl. bibl. 73.69=Paper on: 73.70=Ahmad, Nazir 73.70=Khan, F. Karim 79.76=La Politique culturelle en R‚publique unie du Cameroun^zfre 79.44=Studies and documents on cultural policies 79.50=Incl. bibl. 79.69= - 79.24=Cultural policy in the United Republic of Cameroon 79.26=^aParis ^bUnesco Press^c1976 79.26.a=Paris 79.26.b=Unesco Press 79.26.c=1976 79.30=^a91 p.^billus.^c24 cm. 79.30.a=91 p. 79.30.b=illus. 79.30.c=24 cm. 79.70=Bahoken, J.C. 79.70=Atangana, Engelbert 86.24=Higher education and the labour market in Zambia: expectations and performance 86.26=^aParis^bUnesco Press^bIIEP^aLusaka^bUniversity of Zambia 86.26.a=Paris 86.26.b=Unesco Press 86.26.b=IIEP 86.26.a=Lusaka 86.26.b=University of Zambia 86.30=^a373 p.^billus., map^c24 cm. 86.30.a=373 p. 86.30.b=illus., map 86.30.c=24 cm. 86.50=Incl. bibl. 86.69= - 86.70=Sanyal, Bikas C. 86.70=Case, John H. 86.70=Dow, Philip S. 86.70=Jackman, Mary Elizabeth 86.71=University of Zambia 86.71=IIEP 87.24=Innovation in science education - world-wide 87.26=^aParis^bUnesco Press^c1976 87.26.a=Paris 87.26.b=Unesco Press 87.26.c=1976 87.30=^a249 p.^billus.^c24 cm. 87.30.a=249 p. 87.30.b=illus. 87.30.c=24 cm. 87.70=Baez, Albert V. 87.50=Incl. bibl. 87.69= - 87.76=L'Innovation dans l'enseignement des sciences: synthŠse mondiale^zfre 93.24=Broadcast training techniques 93.26=^aParis^bUnesco Press^c1976 93.26.a=Paris 93.26.b=Unesco Press 93.26.c=1976 93.30=^a122 p.^billus.^c22 cm. 93.30.a=122 p. 93.30.b=illus. 93.30.c=22 cm. 93.50=Incl. bibl. 93.69= - 93.70=Dance, F. Robert 93.76=T‚cnicas de formaci¢n del personal de comunicaci¢n^zspa 100.24=Micropolitan development: theory and practice of greater-rural economic development 100.26=^aAmes^bIowa State University Press^c1976 100.26.a=Ames 100.26.b=Iowa State University Press 100.26.c=1976 100.30=^a456 p.^bmap 100.30.a=456 p. 100.30.b=map 100.50=Incl. bibl. 100.69= - 100.70=Tweeten, Luther G. 100.70=Brinkman, George Lorris 105.76=Les Statistiques de la science et de la technologie en Am‚rique latine ^zfre 105.76=Las Estad¡sticas de la ciencia y la tecnolog¡a en Am‚rica Latina^zspa 105.24=Statistics on science and technology in Latin America; experience with Unesco pilot projects, 1972-1974 105.26=^aParis^bUnesco Press^c1976 105.26.a=Paris 105.26.b=Unesco Press 105.26.c=1976 105.30=^a76 p.^c27 cm. 105.30.a=76 p. 105.30.c=27 cm. 105.70=Th‚baud, Schiller 105.44=Statistical reports and studies^v20 105.50=Incl. bibl. 105.69= - 105.72=Coloquio sobre los Aspectos Nacionales e Internacionales de los Sistemas de Datos para las Estad¡sticas relativas a la Ciencia y la Tecnolog¡a en Am‚rica Latina^pCaracas^d1974 108.24=Comunica‡ao/incomunica‡ao no Brasil 108.26=^aSao Paulo ^bEdi‡oes Loyola^c1976 108.26.a=Sao Paulo 108.26.b=Edi‡oes Loyola 108.26.c=1976 108.30=^a207 p.^billus. 108.30.a=207 p. 108.30.b=illus. 108.50=(Biblioteca Universit ria de Comunica‡ao, vol.1). Incl. bibl. 108.69= - 108.70=Melo, Jos‚ Marques de 108.71=Uniao Crista Brasileira de Comunica‡ao Social 111.24=Multinational corporation and social change 111.26=^aNew York ^bPraeger Publishers^c1976 111.26.a=New York 111.26.b=Praeger Publishers 111.26.c=1976 111.30=^a234 p. 111.30.a=234 p. 111.50=Incl. bibl. 111.69= - 111.70=Apter, David Ernest 111.70=Goodman, Louis Wolf 114.24=Educational development; a new discipline for self-renewal 114.26=^aEugene, Or.^bCenter for the Advanced Study of Educational Administration^c1973 114.26.a=Eugene, Or. 114.26.b=Center for the Advanced Study of Educational Administration 114.26.c=1973 114.30=^a343 p.^billus. 114.30.a=343 p. 114.30.b=illus. 114.50=(CASEA monograph, no.24). Incl. bibl. 114.69= - 114.70=Hemphill, John Knox 114.70=Rosenau, Fred S. 114.71=Far West Laboratory for Educational Research and Development (USA) 114.71=University of Oregon (USA). Center for the Advanced Study of Educational Administration 118.12=Seminario Iberoamericano de Pol¡tica Cient¡fica y Tecnol¢gica^pAlcal  de Henares and Madrid^d1974 118.24=(Ponencias y otras documentos) 118.26=^aMadrid^bServicio de Publicaciones del Ministerio de Educac¡on y Ciencia^c1976 118.26.a=Madrid 118.26.b=Servicio de Publicaciones del Ministerio de Educac¡on y Ciencia 118.26.c=1976 118.30=^a244 p. 118.30.a=244 p. 118.69= - 118.71=OAS 119.12=Meeting on the Implementation of the Recommendations of the Third Regional Conference of Ministers of Education and Ministers Responsible for Economic Planning in the Arab States^pCairo^d1976 119.24=Selected resolutions and recommendations of recent conferences and meetings of relevance to topics likely to be discussed at the meeting 119.26=^c17 Sept. 1976 119.26.c=17 Sept. 1976 119.30=^a25 p. 119.30.a=25 p. 119.69= 120.12=Meeting on the Implementation of the Recommendations of the Third Regional Conference of Ministers of Education and Ministers Responsible for Economic Planning in the Arab States^pCairo^d1976 120.24= Croissance des effectifs scolaires dans les pays arabes; tendances et projections jusqu'en 1985 120.26=^cJuly 1976 120.26.c=July 1976 120.30=^a96 p. ^billus. 120.30.a=96 p. 120.30.b=illus. 120.69= - 121.12=Meeting of Experts on the Content of Education in the Context of Life-long Education^pParis^d1975 121.24=Final report 121.26=^c6 Feb. 1976 121.26.c=6 Feb. 1976 121.30=^a(18 p. in various pagings) 121.30.a=(18 p. in various pagings) 121.69= - 123.12=Meeting on the Implementation of the Recommendations of the Third Regional Conference of Ministers of Education and Ministers Responsible for Economic Planning in the Arab States^pCairo^d1976 123.24=Aspects quantitatifs de la scolarisation du second degr‚ dans les pays arabes 123.26=^c27 Aug. 1976 123.26.c=27 Aug. 1976 123.30=^a25 p. 123.30.a=25 p. 123.69= - 123.70=Sammak, A. 123.71=Unesco Regional Office for Education in the Arab States (Lebanon) 124.12=Meeting on the Implementation of the Recommendations of the Third Regional Conference of Ministers of Education and Ministers Responsible for Economic Planning in the Arab States^pCairo^d1976 124.24=D‚veloppement quantitatif de l'enseignement primaire dans les pays arabes … la lumiŠre de la r‚solution de la Conf‚rence de Marrakech sur l'enseignement obligatoire 124.26=^c27 Aug. 1976 124.26.c=27 Aug. 1976 124.30=^a(44 p. in various pagings)^billus. 124.30.a=(44 p. in various pagings) 124.30.b=illus. 124.69= - 124.71=Unesco Regional Office for Education in the Arab States (Lebanon) 125.12=Regional Seminar on Training^n2nd^pDakar^d1975 125.24=Final report 125.26=^c25 June 1976 125.26.c=25 June 1976 125.30=^a(17 p. in various pagings) 125.30.a=(17 p. in various pagings) 125.69= - 125.71=Unesco Regional Office for Education in Africa (Senegal) 127.69= - 127.71=IBE 127.12=Meeting of National Directors and Chief Technical Advisers of the Projects for the Training of Educational Personnel^n7th^pParis^d1975 127.24=Report 127.26=^c15 June 1976 127.26.c=15 June 1976 127.30=^a(88 p. in various pagings)^billus. 127.30.a=(88 p. in various pagings) 127.30.b=illus. 128.69= - 128.24=Desarrollo y universidad 128.26=^aLa Paz^bLibrer¡a Editorial "Juventud"^c1975 128.26.a=La Paz 128.26.b=Librer¡a Editorial "Juventud" 128.26.c=1975 128.30=^a206 p. 128.30.a=206 p. 128.70=Anaya, Ricardo 130.12=Joint FAO/WHO Expert Committee on Nutrition^n9th^pRome^d1974 130.24=Food and nutrition strategies in national development 130.26=^aRome^bFAO ^c1976 130.26.a=Rome 130.26.b=FAO 130.26.c=1976 130.30=^a64 p. 130.30.a=64 p. 130.50=(FAO nutrition meetings report series, no.56). WHO technical report series, no.584 130.69= - 130.71=FAO 130.71=WHO 130.76=Strat‚gies de l'alimentation et de la nutrition dans le d‚veloppement national^zfre 131.24= Measurement of scientific and technical activities - Proposed standard practice for surveys of research and experimental development 131.26=^aParis^bOECD^c1976 131.26.a=Paris 131.26.b=OECD 131.26.c=1976 131.30=^a138 p. 131.30.a=138 p. 131.69= 131.71=OECD 131.74=Proposed standard practice for surveys of research and experimental development 131.76=La Mesure des activit‚s scientifiques et techniques^zfre 132.12=UNDP/ESCAP Regional Meeting on Technical Cooperation among Developing Countries^pBangkok^d1976 132.24=Conclusions and recommendations and extracts from statements 132.26=^aBangkok^bESCAP^c1976 132.26.a=Bangkok 132.26.b=ESCAP 132.26.c=1976 132.30=^a76 p. 132.30.a=76 p. 132.69= 132.71=ESCAP 141.24=Energy for rural development: renewable resources and alternative technologies for developing countries 141.26=^aWashington^bNational Academy of Sciences^c1976 141.26.a=Washington 141.26.b=National Academy of Sciences 141.26.c=1976 141.30=^a306 p.^billus. 141.30.a=306 p. 141.30.b=illus. 141.50=(PB-260-606). Incl. bibl. 141.69= - 141.71=National Academy of Sciences (USA). Ad Hoc Panel on Renewable Energy Resources 141.71=National Research Council (USA) 141.71=National Academy of Sciences (USA). Board on Science and Technology for International Development 147.24=Employment, growth and basic needs; a one-world problem 147.26=^aGeneva^bILO^c1976 147.26.a=Geneva 147.26.b=ILO 147.26.c=1976 147.30=^a177 p. 147.30.a=177 p. 147.50="Report of the Director-General." Incl. bibl. 147.69= - 147.71=ILO 147.72=Tripartite World Conference on Employment, Income Distribution and Social Progress and the International Division of Labour^pGeneva^d1976 147.76=L'Emploi, la croissance et les besoins essentiels; problŠme mondial^zfre 147.76=Empleo, crecimiento y necesidades esenciales; problema mundial^zspa 149.12=Symposium on the Psychological Bases of Programmed Learning ^pTbilisi, USSR^d1976 149.24=For a training project based on programmed learning in the developing countries: objectives, means and resources, organization and management, development 149.26=^c14 June 1976 149.26.c=14 June 1976 149.30=^a18 p. 149.30.a=18 p. 149.69= - 149.70=Pocztar, Jerry 150.24=Micropolitan development: theory and practice of greater-rural economic development 150.26=^aAmes^bIowa State University Press^c1976 150.26.a=Ames 150.26.b=Iowa State University Press 150.26.c=1976 150.30=^a456 p.^bmap 150.30.a=456 p. 150.30.b=map 150.50=Incl. bibl. 150.69= - 150.70=Tweeten, Luther G. 150.70=Brinkman, George Lorris ABBAS, B.M. ABEYWICKRAMA, B.A. ABOUT ABSENCE ABSORPTION ACCOMPANYING ACCOUNT ACCOUNTING ACHIEVEMENTS ACID ACTION ACTIVITIES ACTUAL ADDRESSES ADMINISTRATIVE REFORM ADMISSION ADMISSION CRITERIA ADULT ADULT EDUCATION AERIAL AERIAL PHOTOGRAPHY AESTHETICS AFFECTIVE AFGHANISTAN AFRICA AFRICAN LANGUAGES AGE AGE GROUPS AGENT AGENTS AGRICULTURAL DEVELOPMENT AGRICULTURAL TRAINING AGRICULTURE AGRONOMY AHMAD, MAJEED AHMAD, NAZIR ALLUVIAUX ALLUVIONS ALTERNATIVE AMAZON AMERICA ANALYSIS ANALYTICAL ANAYA, RICARDO ANCIENT CIVILIZATIONS ANIMAL ANNA ANNOTATED ANNUAL ANTI ANTIQUITIES APERCU APPARATUS APPROPRIATE TECHNOLOGY APRIL APTER, DAVID ERNEST AQUIFERS ARAB COUNTRIES ARABES ARCHITECTURE AREA AREAS ARID ARMAMENTS ARMS CONTROL ARMS SALES ART BOOKS ART HISTORY ARTISTIC CREATION ARTS ASIA ASPECTS ASPIRATION ASSIMILATION ASSURANCES AT ATANGANA, ENGELBERT ATTITUDE AUDIT AUSTRALIA AUSTRALIAN AUTOMATIC AVAILABLE AWAY AYALA, FRANCISCO JOSE BAEZ, ALBERT V. BAHOKEN, J.C. BALANCE BANGLADESH BARRS, H.D. BARS BASED BASEL BASIC BASIC EDUCATION BASIS BAUMGARTNER, ALBERT BE BEHAVIOUR BELOW BENGAL BETWEEN BIBLIOGRAPHIE BIBLIOGRAPHIES BIBLIOGRAPHY BIOLOGICAL EVOLUTION BIOLOGICAL RESEARCH BIOLOGY BIOMASS BITSIOS, DIMITRI S. BJELICA, MIHALO BLACKS BLOOM, JOSEPH BOARDING SCHOOLS BOESCH, H. BOLIVIA BONDURANT, JOAN V. BOOK BORDIA, ANIL BORICHANSKY, L.S. BORING BOROBUDUR BOSIAN, G. BOURDEAU, PHILIPPE F. BOWRA, SIR CECIL MAURICE BRACKISH BRACKISH WATER BRAIN DRAIN BRASIL BRAZIL BRINKMAN, GEORGE LORRIS BRIX, H. BROADCAST BROADCASTING BROTHERS BUDDHISM BUDEL, J. BUDGET BUILDING BURMA BUSINESS MANAGEMENT C CALCULATION CAMEROON CAMEROON UR CANADA CARACTERISTIQUES CARBON CARBON DIOXIDE CARBONIQUE CARIBBEAN AREA CARLIER, G. CARTE CARTOGRAPHY CASE STUDIES CASE, JOHN H. CASPERS, H. CAT CATALOGUES CATSKY, J. CENTRE CEREALS CEREMONY CEYLON CHAMBER CHANCELLOR CHANDI CHANGE CHANGES CHANGING CHANGING SOCIETY CHAO CHAPMAN, VALENTINE J. CHEMICAL ANALYSIS CHEMICAL PROPERTIES CHEMICAL/BIOLOGICAL WARFARE CHI CHILDREN CHILDREN'S BOOKS CHIMIQUE CHO CHOWDHURY, M.I. CHURCH CHURCHILL, STACY CIVILIZATION CLASSICAL CLASSIFICATION CLAYS CLIMATE CLIMATIC CLIMATIC DATA CLUBS CO COASTAL COASTAL ZONES COLOMBIA COLORIMETRIC COMMUNICATION COMMUNICATION HISTORY COMMUNICATION IMPACT COMMUNICATION LEGISLATION COMMUNICATION PERSONNEL TRAINI COMMUNICATION PLANNING COMMUNICATION POLICY COMMUNICATION PROCESS COMMUNICATION RESEARCH COMMUNICATION STATISTICS COMMUNITIES COMMUNITY COMMUNITY DEVELOPMENT COMMUNITY EDUCATION COMPARATIVE COMPARATIVE ANALYSIS COMPARISON COMPULSORY EDUCATION COMPUTER APPLICATIONS COMPUTER HARDWARE COMPUTER SOFTWARE COMUNICACAO CONCENTRATION CONCLUSIONS CONDITIONS CONFERENCE CONFERENCES CONFLICT CONFLICT RESOLUTION CONFORM CONQUEST CONSULTANCY CONTEMPORARY SOCIETY CONTRIBUTES CONTROL CONTROLLED CORPORATION COUGHLAN, MARGARET N. COUNTRIES COUNTRY REPORTS COURSES COVER COVERS CRAIN, N. CRITICISM CROISSANCE CROP CROPPED CROSS CUBA CULTIVATION CULTURAL CULTURAL ACTIVITIES CULTURAL DEVELOPMENT CULTURAL FACILITIES CULTURAL FACTORS CULTURAL HISTORY CULTURAL IDENTITY CULTURAL INSTITUTIONS CULTURAL PLANNING AND ADMINIST CULTURAL POLICY CULTURAL PROGRAMMES CULTURE CURRENT CURRENTS CURRICULUM DEVELOPMENT CURRICULUM PLANNING CYPRUS D DACCA DANCE, F. ROBERT DANS DATA COLLECTION DATA PROCESSING DAVE DAVE, RAVINDRAKUMAR H. DE DECENTRALIZATION DECISION DECISION MAKING DEFICIT DEFINIES DEFINITIONS DEGRE DEGREES DEGUCHI, K. DELIVERED DELTA DELTAIC DELTAS DEMAND DEMOCRATIC KAMPUCHEA DEMOCRATIZATION OF EDUCATION DES DESARROLLO DESCRIPTION DESIGN DETERMINATION DETERMINING DEVELOPED COUNTRIES DEVELOPING DEVELOPING COUNTRIES DEVELOPMENT DEVELOPMENT PLANNING DEVELOPMENT STRATEGIES DEVELOPPEMENT DEW DIFFERENT DIFFUSION DIOXIDE DIPLOMAS DIRECTOR DIRECTORIES DISARMAMENT DISC DISCHARGE OF RIVERS DISCIPLINE DISCUSSED DISSEMINATION OF INFORMATION DOBZHANSKY, THEODOSIUS GRIGORI DOCUMENTATION DOCUMENTATION CENTRES DOCUMENTOS DOTY, M.S. DOW, PHILIP S. DRAINAGE DRAPER, JAMES A. DROUGHT DRY DU DUE DUKE, CHRISTOPHER E EAST EASTERN EUROPE EAUX ECOLOGY ECONOMIC ECONOMIC AND SOCIAL DEVELOPMEN ECONOMIC BEHAVIOUR ECONOMIC CONCENTRATION ECONOMIC CONDITIONS ECONOMIC GROWTH ECONOMIC POLICY ECONOMICS OF EDUCATION ECONOMY ECOSYSTEMS EDUCATION EDUCATION AND DEVELOPMENT EDUCATION AND EMPLOYMENT EDUCATIONAL EDUCATIONAL ADMINISTRATION EDUCATIONAL ADMINISTRATIVE STR EDUCATIONAL ADMINISTRATORS EDUCATIONAL ADVISERS EDUCATIONAL ASSISTANCE EDUCATIONAL COSTS EDUCATIONAL DEMAND EDUCATIONAL DEVELOPMENT EDUCATIONAL ENVIRONMENT EDUCATIONAL EQUIPMENT EDUCATIONAL EXPENDITURE EDUCATIONAL EXPERIMENTS EDUCATIONAL GOALS EDUCATIONAL HISTORY EDUCATIONAL INNOVATIONS EDUCATIONAL INSTITUTIONS EDUCATIONAL LEVELS EDUCATIONAL MANAGEMENT EDUCATIONAL NEEDS EDUCATIONAL OUTPUT EDUCATIONAL PERSONNEL TRAINING EDUCATIONAL PLANNING EDUCATIONAL POLICY EDUCATIONAL PROGRAMMES EDUCATIONAL PSYCHOLOGY EDUCATIONAL RADIO EDUCATIONAL REFORM EDUCATIONAL RESEARCH EDUCATIONAL STATISTICS EDUCATIONAL STRATEGIES EDUCATIONAL SUPERVISION EDUCATIONAL SYSTEMS EDUCATIONAL TECHNOLOGY EDUCATIONAL TELEVISION EDUCATIONAL TRENDS EDUCATIVES EFFECT EFFECTIFS EFFECTS EGOROV, V.V. ELECTRIC ELECTRICAL ENGINEERING ELEMENT ELKANA, YEHUDA EMISSION EMPLOYERS EMPLOYMENT EN ENERGY ENERGY BALANCE ENERGY CONVERSION ENERGY RESOURCES ENROLMENT ENROLMENT PROJECTIONS ENROLMENT RATIOS ENSEIGNEMENT ENVIRONMENTAL ENVIRONMENTAL CHANGES ENVIRONMENTAL CONTROL ENVIRONMENTAL EDUCATION ENVIRONMENTAL QUALITY EQUATION ESCAP ESTABLISHMENT ESTUARIES ESTUARINE ESTUARY ET ETHICS ETUDE EUROPE EVALUATING EVALUATION METHOD EVALUATION OF EDUCATION EVAPORATION EVAPOTRANSPIRATION EVERYDAY LIFE EXAMINATIONS EXCHANGE EXISTENCE EXPECTATIONS EXPERIENCE EXPERIMENT EXPERIMENTAL EXPERIMENTS EXPERTS EXTRACTS FACTORS FAJARDO, JOSE CARLOS FAMILY LIFE FAUNA FAURE, EDGAR FEATURES FELLOWSHIPS FERTILIZERS FIELD FIELDS FILMS FINAL FIRST FISH FISHERIES FLOOD FLOOD CONTROL FLOODING FLOODS FLORA FLORAS FLOW MEASUREMENT AND INSTRUMEN FOLKLORE FOOD FOOD PRODUCTION FOOD RESOURCES FORCE FOREIGN INVESTMENT FOREIGN TRADE FOREST FORESTS FORMATION FORMERLY FOSBERG, F. RAYMOND FRADIER, GEORGES FRANCE FRANCO, C.M. FRANEKER FREEDOM FRESH FRESHWATER FULLY FUTURE GALE, J. GANDHIAN GANGES GASEOUS GAUGES GAZ GENERAL GENESIS GEODYNAMICS GEOLOGICAL GEOLOGY GEOMORPHOLOGIE GEOMORPHOLOGY GEOTHERMAL ENERGY GERMANY FR GHANI, Q. GOODMAN, LOUIS WOLF GOOR, G.A.W. VAN DE GRADUAL GRADUATES GRASS GREAT GREATER GREECE GRIEVE, B.J. GROOT, A.J. DE GROUND GROUNDWATER GROUP DISCUSSIONS GROWTH GUIDANCE SERVICE GUIDE GUIDES GUTTMANN, LUDWIG HALLAK, JACQUES HANDBOOK HAVELOCK, MARY C. HAVELOCK, RONALD G. HEALTH HEAT HEAT TRANSFER HEEMST, H.D.J. VAN HEMPHILL, JOHN KNOX HERVIEU, J. HIGH HIGHER HIGHER EDUCATION HIGHER TECHNICAL PERSONNEL HISTOIRE HISTORIC MONUMENTS HISTORICAL ANALYSIS HISTORY HOLDINGS HOLLANDER, JACK M. HOLLERWOGER, F. HOLT, MICHAEL HOME HOOK, SIDNEY HORLAMUS, SEPP HOUSE HOUSING HUDSON, J.P. HUISMAN, L. HUMAN HUMAN ACTIVITIES EFFECTS HUMAN ECOLOGY HUMAN RESOURCES DEVELOPMENT HUMAN RIGHTS HUMAN SETTLEMENT HUMID HUMID ZONES HUMIDITE HUMIDITY HUNGARY HYDRATURE HYDRAULIC HYDRAULIC ENGINEERING HYDRODYNAMICS HYDROGEOCHEMISTRY HYDROGEOLOGY HYDROGRAPHY HYDROLOGICAL HYDROLOGICAL CYCLE HYDROLOGY HYDROMETEOROLOGICAL HYDROMETEOROLOGY HYGROMETER HYGROMETERS IBE IDEAL IMPLICATIONS IMPORTANT IMPROVEMENT IN-SERVICE TEACHER TRAINING INCLUDING INCOME INCOME DISTRIBUTION INCOMUNICACAO INDEX INDIA INDIAN INDIRECT INDIVIDUAL INDONESIA INDUSTRIAL PRODUCTS INDUSTRIALIZATION INDUSTRIES INFLUENCE INFORMATION PROCESSING AUTOMAT INFORMATION SERVICES INFORMATION SOURCES INFORMATION SYSTEMS INFORMATION USERS INFORMATION/LIBRARY FINANCING INFORMATIQUE INFRA INNOVATION INNOVATION BEHAVIOUR INSTRUCTIONAL INTEGRATED CURRICULUM INTERACTION INTERDISCIPLINARY APPROACH INTERGOVERNMENTAL ORGANIZATION INTERNAL INTERNATIONAL INTERNATIONAL CONFLICT INTERNATIONAL COOPERATION INTERNATIONAL LAW INTERNATIONAL ORGANIZATIONS INTERNATIONAL RELATIONS INTERNATIONAL STANDARDS INTERNATIONAL TRADE INTERNATIONALE INVERTEBRATA INVOLVED IRAN (ISLAMIC REPUBLIC) IRRAWADDY IRRIGATION ISLAM, M. AMIRUL ISRAEL IT ITALY JACKMAN, MARY ELIZABETH JAVA JOBORU, MAGDA JOURNALISM JOURNALISTS JUSQU KARAMAZOV KARENINA KARIM, A.Q.M.B. KENYA KHAN, F. KARIM KHUDA, M. QUDRAT KIDD, JAMES ROBBINS KLAUSING, O. KLEINFELD, JUDITH S. KRAMER, P.J. KREEB, K. L LA LABORATORY LABOUR LABOUR MARKET LABOUR MIGRATION LABOUR POLICY LABOUR RELATIONS LAND LAND RECLAMATION LAND USE LANGE, O.L. LANGUAGE BEHAVIOUR LANGUAGE OF INSTRUCTION LATIN LATIN AMERICA LAW LE LEADERSHIP LEAF LEARNING LEARNING PROCESSES LEAST DEVELOPED COUNTRIES LEAVES LEBLOND, C. LECTURES LEGISLATION LEIDEN LEKOVIC, ZDRAVKO LEROY, GEORGES LES LEVITT, J. LIBERATION MOVEMENTS LIETH, H. LIFE LIFE-LONG EDUCATION LIFELONG LIKELY LIST LITERACY PROGRAMMES LITERARY CRITICISM LITERATURE LIVING CONDITIONS LONG LOSS LUMIERE LYSIMETER M MADAGASCAR MAGALHAES, A.C. MAHEU MAIN MAKING MAKKINK, G.F. MAN MANAGEMENT MANAGEMENT INFORMATION SYSTEMS MANAGEMENT OPERATIONS MANGANESE MANGOKY MANGROVE MANGROVES MANKIND MANPOWER NEEDS MANUEL MANUFACTURING INDUSTRIES MAQUIS MARJORAM, D.T.E. MARKET MARRAKECH MARSHES MASS MASS CULTURE MASS MEDIA MASTER MATHEMATICAL MATHEMATICAL LOGIC MATHEMATICAL MODELS MATHEMATICS MATTER MAY, WILLIAM H. MCQUAIL, DENIS MEANING MEANS MEASUREMENT MEASUREMENT AND INSTRUMENTS MEASUREMENTS MEASURING MEDIA MEDICINAL PLANTS MEETING MEETINGS MELO, JOSE MARQUES DE MEMORIAL MESURE METABOLISM METEOROLOGICAL METEOROLOGY METHOD METHODOLOGIE METHODOLOGY METHODS MICHAELI, A. MICHIGAN MICROBIOLOGY MICROCLIMATE MICROCLIMATIC MICROPOLITAN MIDDLEMARCH MIKHAILOV, V.N. MILITARY ENGINEERING MINERAL DEPOSITS MINERAL RESOURCES MINERALOGIQUES MIRON, MURRAY S. MISE MODEL MODELS MODIFICATIONS MOISTURE MOLECULAR BIOLOGY MOLLUSCA MONGOLIA MONOLITHS MONUMENT MOORMANN, F.R. MORPHOLOGY MORPHOLOGY (LINGUISTICS) MOUTH MR MUCHEKE MUD MUELLER-DOMBOIS, D. MULTINATIONAL MULTIRACIAL SOCIETIES NAGARAJA, V.N. NAMIBIA NAQVI, SIBTE NABI NATIONAL NATIONAL DEVELOPMENT NATIONAL PLANNING NATIONAL POLICY NATIONALISM NATIONS NATURAL NEED NEEDS NETHERLANDS NEW NEWLY-INDEPENDENT STATES NEWS AGENCIES NEWSPAPERS NIGER NO NON-VIOLENCE NORTH AMERICA NOTES NOVEL NOVELS NUCLEAR NUCLEAR WEAPONS NUTRIENT NUTRIENTS NUTRITION OBJECTIVES OBLIGATOIRE OCCUPANT OCCUPATIONAL QUALIFICATIONS OCCURENCE OCEAN CURRENTS OCEANIA OHYA, MASAHIKO OIAPOQUE OKATCHA, F.M.M.O. ONE ONTARIO OPEN OPENING OPERATIONS ORGANES ORGANIC MATTER ORGANIZATION ORGANIZATIONAL OSGOOD, CHARLES EGERTON OTHER OTRAS OUEST PACIFIC ISLANDS PAKISTAN PAR PARASITES PART PART-TIME COURSES PATEL, A.M. PAYS PECS PERDU PERFORMANCE PERIODICALS PERRY, WALTER PERSONAL PERSONNEL MANAGEMENT PERSPECTIVES PERU PERUVIAN PESTS PETTIGREW, ANDREW M. PHENOMENOLOGY PHILIPPINES PHILOSOPHICAL PHILOSOPHY PHOTOGRAPHS PHOTOSYNTHESIS PHYA PHYSICAL EDUCATION PHYSICAL PROPERTIES PHYSIOLOGY PIED PILOT PILOT PROJECTS PINE PIRES, J. MURCA PLACE PLAINS PLAN PLANKTON PLANNING PLANT PLANT ECOLOGY PLANT EVAPOTRANSPIRATION PLANT PHOTOSYNTHESIS PLANT PHYSIOLOGY PLANT TRANSPIRATION PLANTS POCZTAR, JERRY POINT POLICIES POLICY POLITICAL BEHAVIOUR POLITICAL CONFLICT POLITICAL ECONOMY POLITICAL PARTICIPATION POLITICAL PARTIES POLITICAL SYSTEMS POLITICAL THOUGHT POLITICS POLITIQUES POLJAKOFF-MAYBER, A. POLLUTION POLSTER, H. PONENCIAS POPULATION POPULATION DENSITY POPULATION DYNAMICS POROMETER PORTUGAL POST-GRADUATE COURSES POTENTIAL POUR POVERTY POWER INDUSTRY PRACTICE PRATIQUE PRESENTED PRESERVATION OF CULTURAL PROPE PRESERVATION OF MONUMENTS PRESS PRESSURE PREVAILING PRIMAIRE PRIMARY EDUCATION PRIMARY SCHOOLS PROBLEM PROBLEMES PROBLEMS PROCEEDINGS PRODUCT PRODUCTION PRODUCTIVITY PROFESSIONAL ASSOCIATIONS PROFESSIONAL TRAINING PROGRAMMED PROGRAMMED INSTRUCTION PROGRAMMED MATERIALS PROGRAMMES PROGRAMS PROGRESS PROJECT PROJECTIONS PROJECTS PROPOSED PROWSE, G.A. PSYCHOANALYSIS PSYCHOLINGUISTICS PUBLIC PUBLIC EDUCATION PUBLIC HEALTH PUBLIC OPINION QUALITY QUALITY OF LIFE QUANTITATIF QUANTITATIFS QUESTIONNAIRES QURESHI, M. RAHIMULLAH R RACE RACE RELATIONS RACIAL SEGREGATION RADIO RAIN RATES READINGS RECENT RECENTES RECENTS RECHERCHE RECOMMENDATIONS RED REDUCTION REFERENCE REFORM REFRACTIVE REGARD REGIME REGIME OF WATERS REGIONAL COOPERATION REGIONAL PLANNING REGIONS REGULATION RELATED RELATING RELATION RELATIVE RELEVANCE RENE RENEWABLE RENEWABLE RESOURCES RENEWAL REPORT REPUBLIC RESEARCH RESEARCH AND DEVELOPMENT RESEARCH CENTRES RESEARCH PRIORITIES RESEARCH WORKERS RESISTANCE RESOLUTION RESOLUTIONS RESOURCE CONSERVATION RESOURCES RESPONSE RESTORATION REVIEW RHODESIA RHODESIAN RICE RIEPE, DALE MAURICE RIGHT TO INFORMATION RIGHTS RIVER RIVER BASINS RIVER CONTROL RIVERS ROLE ROLE OF EDUCATION ROONWAL, M.L. ROSENAU, FRED S. RUN-OFF RURAL RURAL AREAS RURAL DEVELOPMENT RURAL POPULATION RURAL YOUTH S SA, ARTUR MOREIRA DE SAFETY SAINT ANTONIN, CLAUDE SALINE SALINE WATER SALINITY SAMMAK, A. SANCTIONS SANTEMA, P. SANYAL, BIKAS C. SATELLITE SATELLITE BROADCASTING SATURATION SCALE SCHMID, MAURICE SCHOELLER, MARC SCHOLARSHIPS SCHOOL LABORATORIES SCHOOL MAPPING SCHOOL-AGE POPULATION SCHOOLS SCIENCE SCIENCE AND DEVELOPMENT SCIENCE AND TECHNOLOGY SCIENCE EDUCATION SCIENCE FINANCING SCIENCE POLICY SCIENCE STATISTICS SCIENTIFIC SCIENTIFIC ACTIVITIES SCIENTIFIC DEVELOPMENT SCIENTIFIC INSTITUTIONS SCIENTIFIC PERSONNEL SCIENTIFIC PROGRAMMES SCIENTIFIC PROGRESS SCIENTIFIC RESEARCH SCOLAIRE SCOLAIRES SCOLARISATION SCULPTURE SEA SEA LEVEL SEA WATER SECOND SECONDARY EDUCATION SECONDARY SCHOOL CURRICULUM SECONDARY SCHOOLS SEDIMENT SEDIMENT TRANSPORT SEDIMENTATION SEDIMENTS SELECTED SELF SELF-TEACHING METHODS SEMANTICS SEMI SENEGAL SEPTEMBER SERVICES SEX SHACHORI, A.Y. SHIFTING SHIPWORMS SHUI SIGNIFICANCE SIGNIFICANT SIMMONS, MELVIN K. SINGH, B.N. SIOLI, HAROLD SKILLED WORKERS SLATYER, R.O. SLAVIK, B. SOCIAL SOCIAL CHANGE SOCIAL ENVIRONMENT SOCIAL FACTORS SOCIAL INTEGRATION SOCIAL LIFE SOCIAL PARTICIPATION SOCIAL PLANNING SOCIAL REFORM SOCIAL SCIENCE RESEARCH SOCIAL SECURITY SOCIAL STATUS SOCIAL STRATIFICATION SOCIAL SYSTEMS SOCIALES SOCIALISM SOCIO-ECONOMIC FACTORS SOCIOLOGY SOCIOLOGY OF COMMUNICATION SOCIOLOGY OF SCIENCE SOEKMONO, R. SOIL SOIL EROSION SOIL FERTILITY SOIL GENESIS SOIL MECHANICS SOIL MOISTURE SOIL SCIENCE SOIL SURVEYS SOIL TYPES SOILS SOLAR ENERGY SOLS SOME SOUTERRAINES SOUTH EAST ASIA SOUTHERN AFRICA SPAIN SPECIAL SPORT SRI LANKA STAKE, ROBERT E. STANDARD STANDARDIZATION STATEMENT STATEMENTS STATES STATISTICAL ANALYSIS STATISTICAL DATA STATISTICAL METHODOLOGY STATISTICS STATUS STELLING-MICHAUD, SVEN STOMATAL STONE, E.C. STORM STORMS STRATEGIES STRATEGY STRESS STRUCTURAL STUDENT ATTITUDE STUDENT BEHAVIOUR STUDENT SELECTION STUDENT SOCIOLOGY STUDIES STUDY SUCTION SUD SULPHATE SUNDERBANS SUR SURFACE SURGES SURVEY ANALYSIS SURVEYS SURVIVAL SWITZERLAND SYMPOSIUM SYSTEMS ANALYSIS TAIWAN TEACHER BEHAVIOUR TEACHER EDUCATORS TEACHER GUIDES TEACHER ROLE TEACHER TRAINING TEACHING METHODS TEACHING SKILLS TECHNICAL TECHNICAL AND VOCATIONAL EDUCA TECHNICAL ASSISTANCE TECHNICAL PERSONNEL TECHNICAL TRAINING TECHNIQUE TECHNIQUES TECHNOLOGIES TECHNOLOGY TECHNOLOGY ASSESSMENT TECHNOLOGY TRANSFER TELEVISION TELEVISION PROGRAMMES TEMPERATE TEMPERATURE TEMPLES TEMPS TENDANCES TENTATIVE TEREDINID TERMS TERRITORY TEXTURALES THAILAND THEBAUD, SCHILLER THEIR THEORY THERMAL PROPERTIES OF SUBSTANC THREE TIDAL TIDES TIME TISON, LEON J. TOOL TOPICS TOPOGRAPHY TOPOLOGY TOWNS TRACTS TRADITIONAL TRAINING TRAINING ABROAD TRAINING CENTRES TRAINING COURSES TRANSNATIONAL CORPORATIONS TRANSPIRANTS TRANSPIRATION TRANSPORT TRICART, J. TROPICAL TROPICAL ZONES TROPICAUX TROPICS TURGIDITY TURKEY TWEETEN, LUTHER G. TYPE TYPOLOGY U UNDER UNDERDEVELOPMENT UNESCO UNITED UNIVERSALS UNIVERSIDAD UNIVERSITES UNIVERSITIES UNIVERSITY UNIVERSITY COOPERATION UNIVERSITY COURSES UNIVERSITY CURRICULUM UNIVERSITY EXTENSION URBAN AREAS URBAN DEVELOPMENT URBAN LIFE URBAN PLANNING URBAN POPULATION URBAN RENEWAL URUGUAY USA USE USING USSR V VALEUR VAPOUR VARIABILITY VASAK, KAREL VEGETATION VEGETAUX VELOCITY VERSLUYS, J.D.N. VERSTAPPEN, H.TH. VICE VIET NAM SR VILLAGE VILLAGES VIOLENCE VISSER, W.C. VITALITY VOLKER, ADRIAAN VULNERABLE WANIEWICZ, IGNACY WAR WATER WATER BALANCE WATER YIELD WAVES WAY WEINRICH, A.K.H. WENT, F.W. WEST WESTERN WESTERN EUROPE WETLANDS WIDE WILTING WIND WIND POWER WOMEN WOOD WOODWELL, GEORGE M. WORKERS EDUCATION WORKING POPULATION WORKS WORLD WORLDWIDE WYNTER, HECTOR Y YIELDS YOSUFZAI, H.K. YOUTH YOUTH ORGANIZATIONS YUGOSLAVIA ZACKLIN, RALPH ZAMBIA ZIMBABWE ZONE 86 Higher education and the labour market in Zambia: expectations and performance 87 Innovation in science education - world-wide 100 Micropolitan development: theory and practice of greater-rural economic development 111 Multinational corporation and social change 114 Educational development; a new discipline for self-renewal 118 Seminario Iberoamericano de Pol¡tica Cient¡fica y Tecnol¢gica^pAlcal  de Henares and Madrid^d1974 124 Meeting on the Implementation of the Recommendations of the Third Regional Conference of Ministers of Education and Ministers Responsible for Economic Planning in the Arab States^pCairo^d1976 127 - 128 - 130 Joint FAO/WHO Expert Committee on Nutrition^n9th^pRome^d1974 131 Measurement of scientific and technical activities - Proposed standard practice for surveys of research and experimental development 132 UNDP/ESCAP Regional Meeting on Technical Cooperation among Developing Countries^pBangkok^d1976 133 Consultancy services available in developing ESCAP countries 134 Training courses available in developing ESCAP countries 141 Energy for rural development: renewable resources and alternative technologies for developing countries 147 Employment, growth and basic needs; a one-world problem 149 Symposium on the Psychological Bases of Programmed Learning ^pTbilisi, USSR^d1976 150 Micropolitan development: theory and practice of greater-rural economic development tomcat0100775000113000011300000000213707445415554010772 0ustar kripkrip#!/bin/sh : ${OI:=$HOME/openisis} : ${JAVA_HOME:=/usr/java/jdk1.3} if [ -z "$CATALINA_HOME" ]; then for d in /opt/*tomcat* $HOME/*tomcat*; do [ -d "$d" ] && CATALINA_HOME="$d" # do not break, use the last one found (highest version ?) done fi export JAVA_HOME CATALINA_HOME export LD_LIBRARY_PATH=$OI:$LD_LIBRARY_PATH cmd () { case "$1" in setup) [ -d $OI/WEB-INF ] \ || mkdir $OI/WEB-INF [ -e $OI/WEB-INF/web.xml ] \ || cat >$OI/WEB-INF/web.xml < OpenIsis EOF [ -d $OI/WEB-INF/lib ] \ || mkdir $OI/WEB-INF/lib [ -e $OI/openisis.jar -a -e $OI/openisis.jar ] \ || make -C $OI java [ -e $OI/WEB-INF/lib/openisis.jar ] \ || ln -s ../../openisis.jar $OI/WEB-INF/lib/. [ -e $CATALINA_HOME/webapps/openisis ] \ || ln -s $OI $CATALINA_HOME/webapps/. ;; start) cmd setup; $CATALINA_HOME/bin/startup.sh ;; stop) $CATALINA_HOME/bin/shutdown.sh ;; *) cmd stop; sleep 5; cmd start;; esac } cmd "$1" tpl.c0100644000113000011300000000272007644412745010514 0ustar kripkrip/* openisis - an open implementation of the CDS/ISIS database Version 0.8.x (patchlevel see file Version) Copyright (C) 2001-2003 by Erik Grziwotz, erik@openisis.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA see README for more information EOH */ /* $Id: tpl.c,v 1.9 2003/04/08 00:20:53 kripke Exp $ implementation of general db access functions. */ #include #include "ldb.h" /* ************************************************************ private types */ /* ************************************************************ private data */ /* ************************************************************ private functions */ /* ************************************************************ package functions */ /* ************************************************************ public functions */ unicode.jsp0100644000113000011300000001056007440777755011727 0ustar kripkrip<%@page import="org.openisis.*,java.io.IOException,java.util.Date,java.util.Iterator,java.net.URLEncoder" contentType="text/html;charset=UTF-8"%><%! static final String EMPTY = ""; static Db db; static String dberr = EMPTY; static { try { String[] args = new String[] { "-v", "9", "-encoding", "UTF-8", "-format", "aligned" }; db = Db.open( "/var/db/come2uni/come2uni", args ); if ( db instanceof NativeDb ) System.err.println( "db.db = "+((NativeDb)db).db ); } catch (IOException e) { dberr = e.toString(); } }; static String NV ( String x ) { return null==x ? EMPTY : x; } static int ROWS = 30; // number of rows to show at once static String vif ( String pre, Rec r, int t, String post ) { StringBuffer ret = null; int l = r.getLen(); for ( int i=0; i<% String search = NV( request.getParameter( "search" ) ); // brute force ... search = new String(search.getBytes("ISO-8859-1"), "UTF-8"); int off = 0; int rows = ROWS; try { off = Integer.parseInt( request.getParameter( "off" ) ); } catch ( Exception e ) {} try { rows = Integer.parseInt( request.getParameter( "rows" ) ); } catch ( Exception e ) {} int[] set = (int[])session.getAttribute( "set" ); %> OpenIsis unicode demo <%= dberr %>
<%-- method="POST" enctype="multipart/form-data" buys us nothing -- the ;charset=UTF-8 is parsed by newer tomcat, but not transmitted by Mozilla, so we have to do conversion manually anyway --%> as prefix rows at once
show index terms, not records
text excerpts are from the invitation to the Tenth International Unicode Conference.

<% if ( EMPTY != search ) { if ( null != request.getParameter( "terms" ) ) { Iterator t = db.terms( search ); while ( t.hasNext() ) { String s = (String)t.next(); if ( 0 == s.length() ) continue; String u = URLEncoder.encode( new String(s.getBytes("UTF-8"), "ISO-8859-1") ); %> <%=s%>
<% } %> <% return; } // list terms long then = (new Date()).getTime(); int mode = null==request.getParameter( "prefix" ) ? 0 : Db.QRY_KEYPF; int[] nset = db.search( search, mode ); long now = (new Date()).getTime(); if ( null == nset ) { %> sorry, no results for <%=search%> <% session.removeAttribute( "set" ); return; } else { %> got <%=nset.length%> rows for <%=search%> in <%= 0==mode ? "word" : "prefix" %> mode in <%=(now - then)/1000.%> seconds.
<% set = nset; session.setAttribute( "set", set ); } } // EMPTY != search if ( null == set ) return; if ( 0 < off ) { int boff = off-rows; if ( boff < 0 ) boff = 0; %> <<rows <%=boff+1%> to <%=boff+rows%> <% } int end = off + rows; if ( end >= set.length ) end = set.length; else { %> rows <%=end+1%> to <%=end+rows%>>> <% } %>

rows <%=off+1%> to <%=end%>

<%-- start output --%> <% for ( int i=off; i <% } // for rows %>
<%=name%>
lang=<%=lang%>><%=head%>
lang=<%=lang%>> <%=body%>
unistat0100775000113000011300000000670107654512267011174 0ustar kripkrip#!/usr/bin/perl print < http://unicode.org/Public/UNIDATA/ EOF $dir = $ARGV[0] || '.'; open UDATA, $dir.'/UnicodeData.txt'; open BLOCKS, $dir.'/Blocks.txt'; $end = -1; $gap = 0; # get next block sub nblock { while () { last unless /^#/; } chomp; $lend = $end; ($beg,$end,$block) = /([0-9A-Z]*)..([0-9A-Z]*); (.*)/; $h = {}; $h->{beg} = $beg; $h->{end} = $end; $beg = hex $beg; $end = hex $end; $gap += $beg - 1 - $lend if 65536 > $beg; $h->{len} = $end - $beg + 1; $h->{nam} = $block; # print "$beg..$end($h->{len}): $block\n"; push @blocks, $h; } nblock; # UnicodeData.txt: # 0 number (hex) # 1 name S # 2 general category E # 3 canonical combining class N # 4 bidi class E # 5 decomposition mapping S # 6-8 numeric type and value E/N # 9 bidi_mirrored B # 10 old name S # 11 comment S # 12-14 upper/lower/titlecase mapping S while () { @l = split /;/; # last if '10000' eq $l[0]; # BMP only $num = hex $l[0]; while ($num > $end) { nblock; } die "$_ not in any block!" if $num < $beg; # if ( $l[1] =~ /, First>$/ ) { $n = ; ($e) = split /;/,$n; $e = hex($e); # print stderr "$l[1]..Last: $num .. $e\n"; die "end $e of $l[1] ($num) not in block $h->{nam}" if $e > $end; $h->{$l[2]} += $e - $num +1; next; } # gen cat $h->{$l[2]}++; # decomp if ($l[5]) { ($type) = ($l[5] =~ /<(.*)>/); $deco{$type||'(canonical)'}++; if ( ! $type ) { # decomp Canon $h->{'dC'}++; } } # has Upper if ($l[12]) { $h->{'uC'}++; } } print < http://unicode.org/Public/UNIDATA/UCD.html#Character_Decomposition_Mappings \$ EOF for (sort keys %deco) { print join("\t",$_,$deco{$_}),"\n"; } print "\$\n\n\n"; # 30 general categories + 2 specials @cat = ( 'Cn','Lu','Ll','Lt','Lm','Lo','Mn','Me','Mc','Nd','Nl','No','Zs','Zl','Zp', 'Cc','Cf','Co','Cs','Pc','Pd','Ps','Pe','Pi','Pf','Po','Sm','Sc','Sk','So' ); @add = ( 'uC','dC' ); # table major categories to blocks print "\n*\tmajor category/block table\n"; print <{$_}; $tot{$_} += $h->{$_}; $maj{substr($_,0,1)} += $h->{$_}; } $maj{'C'} += $h->{'Cn'} = $h->{len} - $ass; # unassigned $Cn += $h->{'Cn'} if 65536 > hex($h->{beg}); # in BMP for (@mcat) { $tot{$_} += $maj{$_}; print $maj{$_},"\t"; } for (@add) { $tot{$_} += $h->{$_}; print $h->{$_}||'0',"\t"; } print join("\t",$h->{beg},$h->{end},$h->{len},$h->{nam}),"\n"; } for (@mcat,@add) { print $tot{$_},"\t"; } print "\n"; print "\$\n"; print "BMP: nonblock $gap unassigned $Cn\n\n\n"; # list blocks to categories print < http://unicode.org/Public/UNIDATA/UCD.html#General_Category_Values \$ EOF for $h (@blocks) { print join("\t",$h->{nam},'b'.$h->{beg},'l'.$h->{len}),"\t"; for (@cat) { if ($h->{$_}) { print $_,$h->{$_},"\t"; } } print "\n"; } print "\$\n\n\n";