1 <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
2 "http://www.w3.org/TR/html4/loose.dtd">
5 <title>The XML C library for Gnome</title>
6 <meta name="GENERATOR" content="amaya 5.1">
7 <meta http-equiv="Content-Type" content="text/html">
10 <body bgcolor="#ffffff">
11 <h1 align="center">The XML C library for Gnome</h1>
13 <h1>Note: this is the flat content of the <a href="index.html">web
16 <h1 style="text-align: center">libxml, a.k.a. gnome-xml</h1>
20 <p>Libxml is the XML C library developed for the Gnome project. XML itself
21 is a metalanguage to design markup languages, i.e. text language where
22 semantic and structure are added to the content using extra "markup"
23 information enclosed between angle brackets. HTML is the most well-known
24 markup language. Though the library is written in C <a href="python.html">a
25 variety of language bindings</a> make it available in other environments.</p>
27 <p>Libxml2 implements a number of existing standards related to markup
30 <li>the XML standard: <a
31 href="http://www.w3.org/TR/REC-xml">http://www.w3.org/TR/REC-xml</a></li>
32 <li>Namespaces in XML: <a
33 href="http://www.w3.org/TR/REC-xml-names/">http://www.w3.org/TR/REC-xml-names/</a></li>
35 href="http://www.w3.org/TR/xmlbase/">http://www.w3.org/TR/xmlbase/</a></li>
36 <li><a href="http://www.cis.ohio-state.edu/rfc/rfc2396.txt">RFC 2396</a> :
37 Uniform Resource Identifiers <a
38 href="http://www.ietf.org/rfc/rfc2396.txt">http://www.ietf.org/rfc/rfc2396.txt</a></li>
39 <li>XML Path Language (XPath) 1.0: <a
40 href="http://www.w3.org/TR/xpath">http://www.w3.org/TR/xpath</a></li>
42 href="http://www.w3.org/TR/html401/">http://www.w3.org/TR/html401/</a></li>
43 <li>most of XML Pointer Language (XPointer) Version 1.0: <a
44 href="http://www.w3.org/TR/xptr">http://www.w3.org/TR/xptr</a></li>
45 <li>XML Inclusions (XInclude) Version 1.0: <a
46 href="http://www.w3.org/TR/xinclude/">http://www.w3.org/TR/xinclude/</a></li>
48 href="http://www.cis.ohio-state.edu/rfc/rfc2044.txt">rfc2044</a> [UTF-8]
49 and <a href="http://www.cis.ohio-state.edu/rfc/rfc2781.txt">rfc2781</a>
50 [UTF-16] core encodings</li>
51 <li>part of SGML Open Technical Resolution TR9401:1997</li>
52 <li>XML Catalogs Working Draft 06 August 2001: <a
53 href="http://www.oasis-open.org/committees/entity/spec-2001-08-06.html">http://www.oasis-open.org/committees/entity/spec-2001-08-06.html</a></li>
54 <li>Canonical XML Version 1.0: <a
55 href="http://www.w3.org/TR/xml-c14n">http://www.w3.org/TR/xml-c14n</a>
56 and the Exclusive XML Canonicalization CR draft <a
57 href="http://www.w3.org/TR/xml-exc-c14n">http://www.w3.org/TR/xml-exc-c14n</a></li>
58 <li>Relax NG Committee Specification 3 December 2001 <a
59 href="http://www.oasis-open.org/committees/relax-ng/spec-20011203.html">http://www.oasis-open.org/committees/relax-ng/spec-20011203.html</a></li>
62 <p>In most cases libxml tries to implement the specifications in a relatively
63 strictly compliant way. As of release 2.4.16, libxml2 passes all 1800+ tests
65 href="http://www.oasis-open.org/committees/xml-conformance/">OASIS XML Tests
68 <p>To some extent libxml2 provides support for the following additional
69 specifications but doesn't claim to implement them completely:</p>
71 <li>Document Object Model (DOM) <a
72 href="http://www.w3.org/TR/DOM-Level-2-Core/">http://www.w3.org/TR/DOM-Level-2-Core/</a>
73 it doesn't implement the API itself, gdome2 does this on top of
75 <li><a href="http://www.cis.ohio-state.edu/rfc/rfc959.txt">RFC 959</a> :
76 libxml implements a basic FTP client code</li>
77 <li><a href="http://www.cis.ohio-state.edu/rfc/rfc1945.txt">RFC 1945</a> :
78 HTTP/1.0, again a basic HTTP client code</li>
79 <li>SAX: a minimal SAX implementation compatible with early expat
81 <li>DocBook SGML v4: libxml2 includes a hackish parser to transition to
85 <p>A partial implementation of XML Schemas is being worked on but it would be
86 far too early to make any conformance statement about it at the moment.</p>
88 <p>Libxml2 is known to be very portable, the library should build and work
89 without serious troubles on a variety of systems (Linux, Unix, Windows,
90 CygWin, MacOS, MacOS X, RISC Os, OS/2, VMS, QNX, MVS, ...)</p>
92 <p>Separate documents:</p>
94 <li><a href="http://xmlsoft.org/XSLT/">the libxslt page</a> providing an
95 implementation of XSLT 1.0 and common extensions like EXSLT for
97 <li><a href="http://www.cs.unibo.it/~casarini/gdome2/">the gdome2 page</a>
98 : a standard DOM2 implementation for libxml2</li>
99 <li><a href="http://www.aleksey.com/xmlsec/">the XMLSec page</a>: an
100 implementation of <a href="http://www.w3.org/TR/xmldsig-core/">W3C XML
101 Digital Signature</a> for libxml2</li>
102 <li>also check the related links section below for more related and active
106 <p>Logo designed by <a href="mailto:liyanage@access.ch">Marc Liyanage</a>.</p>
108 <h2><a name="Introducti">Introduction</a></h2>
110 <p>This document describes libxml, the <a
111 href="http://www.w3.org/XML/">XML</a> C library developed for the <a
112 href="http://www.gnome.org/">Gnome</a> project. <a
113 href="http://www.w3.org/XML/">XML is a standard</a> for building tag-based
114 structured documents/data.</p>
116 <p>Here are some key points about libxml:</p>
118 <li>Libxml exports Push (progressive) and Pull (blocking) type parser
119 interfaces for both XML and HTML.</li>
120 <li>Libxml can do DTD validation at parse time, using a parsed document
121 instance, or with an arbitrary DTD.</li>
122 <li>Libxml includes complete <a
123 href="http://www.w3.org/TR/xpath">XPath</a>, <a
124 href="http://www.w3.org/TR/xptr">XPointer</a> and <a
125 href="http://www.w3.org/TR/xinclude">XInclude</a> implementations.</li>
126 <li>It is written in plain C, making as few assumptions as possible, and
127 sticking closely to ANSI C/POSIX for easy embedding. Works on
128 Linux/Unix/Windows, ported to a number of other platforms.</li>
129 <li>Basic support for HTTP and FTP client allowing applications to fetch
130 remote resources.</li>
131 <li>The design is modular, most of the extensions can be compiled out.</li>
132 <li>The internal document representation is as close as possible to the <a
133 href="http://www.w3.org/DOM/">DOM</a> interfaces.</li>
134 <li>Libxml also has a <a href="http://www.megginson.com/SAX/index.html">SAX
135 like interface</a>; the interface is designed to be compatible with <a
136 href="http://www.jclark.com/xml/expat.html">Expat</a>.</li>
137 <li>This library is released under the <a
138 href="http://www.opensource.org/licenses/mit-license.html">MIT
139 License</a>. See the Copyright file in the distribution for the precise
143 <p>Warning: unless you are forced to because your application links with a
144 Gnome-1.X library requiring it, <strong><span
145 style="background-color: #FF0000">Do Not Use libxml1</span></strong>, use
148 <h2><a name="FAQ">FAQ</a></h2>
150 <p>Table of Contents:</p>
152 <li><a href="FAQ.html#License">License(s)</a></li>
153 <li><a href="FAQ.html#Installati">Installation</a></li>
154 <li><a href="FAQ.html#Compilatio">Compilation</a></li>
155 <li><a href="FAQ.html#Developer">Developer corner</a></li>
158 <h3><a name="License">License</a>(s)</h3>
160 <li><em>Licensing Terms for libxml</em>
161 <p>libxml is released under the <a
162 href="http://www.opensource.org/licenses/mit-license.html">MIT
163 License</a>; see the file Copyright in the distribution for the precise
166 <li><em>Can I embed libxml in a proprietary application ?</em>
167 <p>Yes. The MIT License allows you to keep proprietary the changes you
168 made to libxml, but it would be graceful to send-back bug fixes and
169 improvements as patches for possible incorporation in the main
170 development tree.</p>
174 <h3><a name="Installati">Installation</a></h3>
176 <li>Unless you are forced to because your application links with a Gnome
177 library requiring it, <strong><span style="background-color: #FF0000">Do
178 Not Use libxml1</span></strong>, use libxml2</li>
179 <li><em>Where can I get libxml</em> ?
180 <p>The original distribution comes from <a
181 href="ftp://rpmfind.net/pub/libxml/">rpmfind.net</a> or <a
182 href="ftp://ftp.gnome.org/pub/GNOME/sources/libxml2/2.5/">gnome.org</a></p>
183 <p>Most Linux and BSD distributions include libxml, this is probably the
184 safer way for end-users to use libxml.</p>
185 <p>David Doolin provides precompiled Windows versions at <a
186 href="http://www.ce.berkeley.edu/~doolin/code/libxmlwin32/ ">http://www.ce.berkeley.edu/~doolin/code/libxmlwin32/</a></p>
188 <li><em>I see libxml and libxml2 releases, which one should I install ?</em>
190 <li>If you are not constrained by backward compatibility issues with
191 existing applications, install libxml2 only</li>
192 <li>If you are not doing development, you can safely install both.
193 Usually the packages <a
194 href="http://rpmfind.net/linux/RPM/libxml.html">libxml</a> and <a
195 href="http://rpmfind.net/linux/RPM/libxml2.html">libxml2</a> are
196 compatible (this is not the case for development packages).</li>
197 <li>If you are a developer and your system provides separate packaging
198 for shared libraries and the development components, it is possible
199 to install libxml and libxml2, and also <a
200 href="http://rpmfind.net/linux/RPM/libxml-devel.html">libxml-devel</a>
202 href="http://rpmfind.net/linux/RPM/libxml2-devel.html">libxml2-devel</a>
203 too for libxml2 >= 2.3.0</li>
204 <li>If you are developing a new application, please develop against
208 <li><em>I can't install the libxml package, it conflicts with libxml0</em>
209 <p>You probably have an old libxml0 package used to provide the shared
210 library for libxml.so.0, you can probably safely remove it. The libxml
211 packages provided on <a
212 href="ftp://rpmfind.net/pub/libxml/">rpmfind.net</a> provide
215 <li><em>I can't install the libxml(2) RPM package due to failed
217 <p>The most generic solution is to re-fetch the latest src.rpm , and
218 rebuild it locally with</p>
219 <p><code>rpm --rebuild libxml(2)-xxx.src.rpm</code>.</p>
220 <p>If everything goes well it will generate two binary rpm packages (one
221 providing the shared libs and xmllint, and the other one, the -devel
222 package, providing includes, static libraries and scripts needed to build
223 applications with libxml(2)) that you can install locally.</p>
227 <h3><a name="Compilatio">Compilation</a></h3>
229 <li><em>What is the process to compile libxml ?</em>
230 <p>As most UNIX libraries libxml follows the "standard":</p>
231 <p><code>gunzip -c xxx.tar.gz | tar xvf -</code></p>
232 <p><code>cd libxml-xxxx</code></p>
233 <p><code>./configure --help</code></p>
234 <p>to see the options, then the compilation/installation proper</p>
235 <p><code>./configure [possible options]</code></p>
236 <p><code>make</code></p>
237 <p><code>make install</code></p>
238 <p>At that point you may have to rerun ldconfig or a similar utility to
239 update your list of installed shared libs.</p>
241 <li><em>What other libraries are needed to compile/install libxml ?</em>
242 <p>Libxml does not require any other library, the normal C ANSI API
243 should be sufficient (please report any violation to this rule you may
245 <p>However if found at configuration time libxml will detect and use the
248 <li><a href="http://www.info-zip.org/pub/infozip/zlib/">libz</a> : a
249 highly portable and available widely compression library.</li>
250 <li>iconv: a powerful character encoding conversion library. It is
251 included by default in recent glibc libraries, so it doesn't need to
252 be installed specifically on Linux. It now seems a <a
253 href="http://www.opennc.org/onlinepubs/7908799/xsh/iconv.html">part
254 of the official UNIX</a> specification. Here is one <a
255 href="http://www.gnu.org/software/libiconv/">implementation of the
256 library</a> which source can be found <a
257 href="ftp://ftp.ilog.fr/pub/Users/haible/gnu/">here</a>.</li>
260 <li><em>Make check fails on some platforms</em>
261 <p>Sometimes the regression tests' results don't completely match the
262 value produced by the parser, and the makefile uses diff to print the
263 delta. On some platforms the diff return breaks the compilation process;
264 if the diff is small this is probably not a serious problem.</p>
265 <p>Sometimes (especially on Solaris) make checks fail due to limitations
266 in make. Try using GNU-make instead.</p>
268 <li><em>I use the CVS version and there is no configure script</em>
269 <p>The configure script (and other Makefiles) are generated. Use the
270 autogen.sh script to regenerate the configure script and Makefiles,
272 <p><code>./autogen.sh --prefix=/usr --disable-shared</code></p>
274 <li><em>I have troubles when running make tests with gcc-3.0</em>
275 <p>It seems the initial release of gcc-3.0 has a problem with the
276 optimizer which miscompiles the URI module. Please use another
281 <h3><a name="Developer">Developer</a> corner</h3>
283 <li><em>xmlDocDump() generates output on one line.</em>
284 <p>Libxml will not <strong>invent</strong> spaces in the content of a
285 document since <strong>all spaces in the content of a document are
286 significant</strong>. If you build a tree from the API and want
289 <li>the correct way is to generate those yourself too.</li>
290 <li>the dangerous way is to ask libxml to add those blanks to your
291 content <strong>modifying the content of your document in the
292 process</strong>. The result may not be what you expect. There is
293 <strong>NO</strong> way to guarantee that such a modification won't
294 affect other parts of the content of your document. See <a
295 href="http://xmlsoft.org/html/libxml-parser.html#XMLKEEPBLANKSDEFAULT">xmlKeepBlanksDefault
297 href="http://xmlsoft.org/html/libxml-tree.html#XMLSAVEFORMATFILE">xmlSaveFormatFile
301 <li>Extra nodes in the document:
302 <p><em>For a XML file as below:</em></p>
303 <pre><?xml version="1.0"?>
304 <PLAN xmlns="http://www.argus.ca/autotest/1.0/">
305 <NODE CommFlag="0"/>
306 <NODE CommFlag="1"/>
308 <p><em>after parsing it with the function
309 pxmlDoc=xmlParseFile(...);</em></p>
310 <p><em>I want to the get the content of the first node (node with the
311 CommFlag="0")</em></p>
312 <p><em>so I did it as following;</em></p>
313 <pre>xmlNodePtr pnode;
314 pnode=pxmlDoc->children->children;</pre>
315 <p><em>but it does not work. If I change it to</em></p>
316 <pre>pnode=pxmlDoc->children->children->next;</pre>
317 <p><em>then it works. Can someone explain it to me.</em></p>
319 <p>In XML all characters in the content of the document are significant
320 <strong>including blanks and formatting line breaks</strong>.</p>
321 <p>The extra nodes you are wondering about are just that, text nodes with
322 the formatting spaces which are part of the document but that people tend
323 to forget. There is a function <a
324 href="http://xmlsoft.org/html/libxml-parser.html">xmlKeepBlanksDefault
325 ()</a> to remove those at parse time, but that's an heuristic, and its
326 use should be limited to cases where you are certain there is no
327 mixed-content in the document.</p>
329 <li><em>I get compilation errors of existing code like when accessing
330 <strong>root</strong> or <strong>child fields</strong> of nodes.</em>
331 <p>You are compiling code developed for libxml version 1 and using a
332 libxml2 development environment. Either switch back to libxml v1 devel or
333 even better fix the code to compile with libxml2 (or both) by <a
334 href="upgrade.html">following the instructions</a>.</p>
336 <li><em>I get compilation errors about non existing
337 <strong>xmlRootNode</strong> or <strong>xmlChildrenNode</strong>
339 <p>The source code you are using has been <a
340 href="upgrade.html">upgraded</a> to be able to compile with both libxml
341 and libxml2, but you need to install a more recent version:
342 libxml(-devel) >= 1.8.8 or libxml2(-devel) >= 2.1.0</p>
344 <li><em>XPath implementation looks seriously broken</em>
345 <p>XPath implementation prior to 2.3.0 was really incomplete. Upgrade to
346 a recent version, there are no known bugs in the current version.</p>
348 <li><em>The example provided in the web page does not compile.</em>
349 <p>It's hard to maintain the documentation in sync with the code
350 <grin/> ...</p>
351 <p>Check the previous points 1/ and 2/ raised before, and please send
354 <li><em>Where can I get more examples and information than privoded on the
356 <p>Ideally a libxml book would be nice. I have no such plan ... But you
359 <li>check more deeply the <a href="html/libxml-lib.html">existing
360 generated doc</a></li>
361 <li>look for examples of use for libxml function using the Gnome code.
362 For example the following will query the full Gnome CVS base for the
363 use of the <strong>xmlAddChild()</strong> function:
365 href="http://cvs.gnome.org/lxr/search?string=xmlAddChild">http://cvs.gnome.org/lxr/search?string=xmlAddChild</a></p>
366 <p>This may be slow, a large hardware donation to the gnome project
367 could cure this :-)</p>
370 href="http://cvs.gnome.org/bonsai/rview.cgi?cvsroot=/cvs/gnome&dir=gnome-xml">Browse
371 the libxml source</a> , I try to write code as clean and documented
372 as possible, so looking at it may be helpful. In particular the code
373 of xmllint.c and of the various testXXX.c test programs should
374 provide good examples of how to do things with the library.</li>
378 <p>libxml is written in pure C in order to allow easy reuse on a number
379 of platforms, including embedded systems. I don't intend to convert to
381 <p>There are however a few C++ wrappers which may fulfill your needs:</p>
383 <li>by Ari Johnson <ari@btigate.com>:
385 href="http://lusis.org/~ari/xml++/">http://lusis.org/~ari/xml++/</a></p>
387 href="http://lusis.org/~ari/xml++/libxml++.tar.gz">http://lusis.org/~ari/xml++/libxml++.tar.gz</a></p>
389 <li>by Peter Jones <pjones@pmade.org>
391 href="http://pmade.org/pjones/software/xmlwrapp/">http://pmade.org/pjones/software/xmlwrapp/</a></p>
395 <li>How to validate a document a posteriori ?
396 <p>It is possible to validate documents which had not been validated at
397 initial parsing time or documents which have been built from scratch
398 using the API. Use the <a
399 href="http://xmlsoft.org/html/libxml-valid.html#XMLVALIDATEDTD">xmlValidateDtd()</a>
400 function. It is also possible to simply add a DTD to an existing
402 <pre>xmlDocPtr doc; /* your existing document */
403 xmlDtdPtr dtd = xmlParseDTD(NULL, filename_of_dtd); /* parse the DTD */
405 dtd->name = xmlStrDup((xmlChar*)"root_name"); /* use the given root */
407 doc->intSubset = dtd;
408 if (doc->children == NULL) xmlAddChild((xmlNodePtr)doc, (xmlNodePtr)dtd);
409 else xmlAddPrevSibling(doc->children, (xmlNodePtr)dtd);
412 <li>So what is this funky "xmlChar" used all the time?
413 <p>It is a null terminated sequence of utf-8 characters. And only utf-8!
414 You need to convert strings encoded in different ways to utf-8 before
415 passing them to the API. This can be accomplished with the iconv library
423 <h2><a name="Documentat">Documentation</a></h2>
425 <p>There are several on-line resources related to using libxml:</p>
427 <li>Use the <a href="search.php">search engine</a> to lookup
429 <li>Check the <a href="FAQ.html">FAQ.</a></li>
430 <li>Check the <a href="http://xmlsoft.org/html/libxml-lib.html">extensive
431 documentation</a> automatically extracted from code comments (using <a
432 href="http://cvs.gnome.org/bonsai/rview.cgi?cvsroot=/cvs/gnome&dir=gtk-doc">gtk
434 <li>Look at the documentation about <a href="encoding.html">libxml
435 internationalization support</a>.</li>
436 <li>This page provides a global overview and <a href="example.html">some
437 examples</a> on how to use libxml.</li>
438 <li>John Fleck's libxml tutorial: <a href="tutorial/index.html">html</a> or
439 <a href="tutorial/xmltutorial.pdf">pdf</a>.</li>
440 <li><a href="mailto:james@daa.com.au">James Henstridge</a> wrote <a
441 href="http://www.daa.com.au/~james/gnome/xml-sax/xml-sax.html">some nice
442 documentation</a> explaining how to use the libxml SAX interface.</li>
443 <li>George Lebl wrote <a
444 href="http://www-4.ibm.com/software/developer/library/gnome3/">an article
445 for IBM developerWorks</a> about using libxml.</li>
446 <li>Check <a href="http://cvs.gnome.org/lxr/source/gnome-xml/TODO">the TODO
448 <li>Read the <a href="upgrade.html">1.x to 2.x upgrade path</a>
449 description. If you are starting a new project using libxml you should
450 really use the 2.x version.</li>
451 <li>And don't forget to look at the <a
452 href="http://mail.gnome.org/archives/xml/">mailing-list archive</a>.</li>
455 <h2><a name="Reporting">Reporting bugs and getting help</a></h2>
457 <p>Well, bugs or missing features are always possible, and I will make a
458 point of fixing them in a timely fashion. The best way to report a bug is to
459 use the <a href="http://bugzilla.gnome.org/buglist.cgi?product=libxml">Gnome
460 bug tracking database</a> (make sure to use the "libxml2" module name). I
461 look at reports there regularly and it's good to have a reminder when a bug
462 is still open. Be sure to specify that the bug is for the package libxml2.</p>
464 <p>There is also a mailing-list <a
465 href="mailto:xml@gnome.org">xml@gnome.org</a> for libxml, with an <a
466 href="http://mail.gnome.org/archives/xml/">on-line archive</a> (<a
467 href="http://xmlsoft.org/messages">old</a>). To subscribe to this list,
469 href="http://mail.gnome.org/mailman/listinfo/xml">associated Web</a> page and
470 follow the instructions. <strong>Do not send code, I won't debug it</strong>
471 (but patches are really appreciated!).</p>
473 <p>Check the following <strong><span style="color: #FF0000">before
474 posting</span></strong>:</p>
476 <li>Read the <a href="FAQ.html">FAQ</a> and <a href="search.php">use the
477 search engine</a> to get informations related to your problem.</li>
478 <li>Make sure you are <a href="ftp://xmlsoft.org/">using a recent
479 version</a>, and that the problem still shows up in a recent version.</li>
480 <li>Check the <a href="http://mail.gnome.org/archives/xml/">list
481 archives</a> to see if the problem was reported already. In this case
482 there is probably a fix available, similarly check the <a
483 href="http://bugzilla.gnome.org/buglist.cgi?product=libxml">registered
485 <li>Make sure you can reproduce the bug with xmllint or one of the test
486 programs found in source in the distribution.</li>
487 <li>Please send the command showing the error as well as the input (as an
491 <p>Then send the bug with associated informations to reproduce it to the <a
492 href="mailto:xml@gnome.org">xml@gnome.org</a> list; if it's really libxml
493 related I will approve it. Please do not send mail to me directly, it makes
494 things really hard to track and in some cases I am not the best person to
495 answer a given question, ask on the list.</p>
497 <p>To <span style="color: #E50000">be really clear about support</span>:</p>
499 <li>Support or help <span style="color: #E50000">request MUST be sent to
500 the list or on bugzilla</span> in case of problems, so that the Question
501 and Answers can be shared publicly. Failing to do so carries the implicit
502 message "I want free support but I don't want to share the benefits with
503 others" and is not welcome. I will automatically Carbon-Copy the
504 xml@gnome.org mailing list for any technical reply made about libxml2 or
506 <li>There is <span style="color: #E50000">no garantee for support</span>,
507 if your question remains unanswered after a week, repost it, making sure
508 you gave all the detail needed and the informations requested.</li>
509 <li>Failing to provide informations as requested or double checking first
510 for prior feedback also carries the implicit message "the time of the
511 library maintainers is less valuable than my time" and might not be
515 <p>Of course, bugs reported with a suggested patch for fixing them will
516 probably be processed faster than those without.</p>
518 <p>If you're looking for help, a quick look at <a
519 href="http://mail.gnome.org/archives/xml/">the list archive</a> may actually
520 provide the answer. I usually send source samples when answering libxml usage
521 questions. The <a href="http://xmlsoft.org/html/book1.html">auto-generated
522 documentation</a> is not as polished as I would like (i need to learn more
523 about DocBook), but it's a good starting point.</p>
525 <h2><a name="help">How to help</a></h2>
527 <p>You can help the project in various ways, the best thing to do first is to
528 subscribe to the mailing-list as explained before, check the <a
529 href="http://mail.gnome.org/archives/xml/">archives </a>and the <a
530 href="http://bugzilla.gnome.org/buglist.cgi?product=libxml">Gnome bug
533 <li>Provide patches when you find problems.</li>
534 <li>Provide the diffs when you port libxml to a new platform. They may not
535 be integrated in all cases but help pinpointing portability problems
537 <li>Provide documentation fixes (either as patches to the code comments or
539 <li>Provide new documentations pieces (translations, examples, etc
541 <li>Check the TODO file and try to close one of the items.</li>
542 <li>Take one of the points raised in the archive or the bug database and
543 provide a fix. <a href="mailto:daniel@veillard.com">Get in touch with me
544 </a>before to avoid synchronization problems and check that the suggested
545 fix will fit in nicely :-)</li>
548 <h2><a name="Downloads">Downloads</a></h2>
550 <p>The latest versions of libxml can be found on <a
551 href="ftp://xmlsoft.org/">xmlsoft.org</a> (<a
552 href="ftp://speakeasy.rpmfind.net/pub/libxml/">Seattle</a>, <a
553 href="ftp://fr.rpmfind.net/pub/libxml/">France</a>) or on the <a
554 href="ftp://ftp.gnome.org/pub/GNOME/MIRRORS.html">Gnome FTP server</a> either
555 as a <a href="ftp://ftp.gnome.org/pub/GNOME/sources/libxml2/2.5/">source
556 archive</a><!-- commenting this out because they seem to have disappeared or <a
557 href="ftp://ftp.gnome.org/pub/GNOME/stable/redhat/i386/libxml/">RPM
559 , Antonin Sprinzl also provide <a href="ftp://gd.tuwien.ac.at/pub/libxml/">a
560 mirror in Austria</a>. (NOTE that you need both the <a
561 href="http://rpmfind.net/linux/RPM/libxml2.html">libxml(2)</a> and <a
562 href="http://rpmfind.net/linux/RPM/libxml2-devel.html">libxml(2)-devel</a>
563 packages installed to compile applications using libxml.) <a
564 href="mailto:igor@zlatkovic.com">Igor Zlatkovic</a> is now the maintainer of
566 href="http://www.zlatkovic.com/projects/libxml/index.html">he provides
567 binaries</a>. <a href="mailto:Gary.Pennington@sun.com">Gary Pennington</a>
568 provides <a href="http://garypennington.net/libxml2/">Solaris binaries</a>.
569 <a href="mailto:Steve.Ball@zveno.com">Steve Ball</a> provides <a
570 href="http://www.zveno.com/open_source/libxml2xslt.html">Mac Os X
573 <p><a name="Snapshot">Snapshot:</a></p>
575 <li>Code from the W3C cvs base libxml <a
576 href="ftp://xmlsoft.org/cvs-snapshot.tar.gz">cvs-snapshot.tar.gz</a>.</li>
577 <li>Docs, content of the web site, the list archive included <a
578 href="ftp://xmlsoft.org/libxml-docs.tar.gz">libxml-docs.tar.gz</a>.</li>
581 <p><a name="Contribs">Contributions:</a></p>
583 <p>I do accept external contributions, especially if compiling on another
584 platform, get in touch with me to upload the package, wrappers for various
585 languages have been provided, and can be found in the <a
586 href="contribs.html">contrib section</a></p>
588 <p>Libxml is also available from CVS:</p>
591 href="http://cvs.gnome.org/bonsai/rview.cgi?cvsroot=/cvs/gnome&dir=gnome-xml">Gnome
592 CVS base</a>. Check the <a
593 href="http://developer.gnome.org/tools/cvs.html">Gnome CVS Tools</a>
594 page; the CVS module is <b>gnome-xml</b>.</p>
596 <li>The <strong>libxslt</strong> module is also present there</li>
599 <h2><a name="News">News</a></h2>
601 <h3>CVS only : check the <a
602 href="http://cvs.gnome.org/lxr/source/gnome-xml/ChangeLog">Changelog</a> file
603 for a really accurate description</h3>
605 <p>Items not finished and worked on, get in touch with the list if you want
608 <li>More testing on RelaxNG</li>
609 <li>Finishing up <a href="http://www.w3.org/TR/xmlschema-1/">XML
613 <h3>2.5.4: Feb 20 2003</h3>
615 <li>Conformance testing and lot of fixes on Relax NG and XInclude
617 <li>Implementation of XPointer element() scheme</li>
618 <li>Bug fixes: XML parser, XInclude entities merge, validity checking on
620 <p>2 serialization bugs, node info generation problems, a DTD regexp
621 generation problem.</p>
623 <li>Portability: windows updates and path canonicalization (Igor)</li>
624 <li>A few typo fixes (Kjartan Maraas)</li>
625 <li>Python bindings generator fixes (Stephane Bidoul)</li>
628 <h3>2.5.3: Feb 10 2003</h3>
630 <li>RelaxNG and XML Schemas datatypes improvements, and added a first
631 version of RelaxNG Python bindings</li>
632 <li>Fixes: XLink (Sean Chittenden), XInclude (Sean Chittenden), API fix for
633 serializing namespace nodes, encoding conversion bug, XHTML1
635 <li>Portability fixes: Windows (Igor), AMD 64bits RPM spec file</li>
638 <h3>2.5.2: Feb 5 2003</h3>
640 <li>First implementation of RelaxNG, added --relaxng flag to xmllint</li>
641 <li>Schemas support now compiled in by default.</li>
642 <li>Bug fixes: DTD validation, namespace checking, XInclude and entities,
643 delegateURI in XML Catalogs, HTML parser, XML reader (Stéphane Bidoul),
644 XPath parser and evaluation, UTF8ToUTF8 serialization, XML reader memory
645 consumption, HTML parser, HTML serialization in the presence of
647 <li>added an HTML API to check elements and attributes.</li>
648 <li>Documentation improvement, PDF for the tutorial (John Fleck), doc
649 patches (Stefan Kost)</li>
650 <li>Portability fixes: NetBSD (Julio Merino), Windows (Igor Zlatkovic)</li>
651 <li>Added python bindings for XPointer, contextual error reporting
652 (Stéphane Bidoul)</li>
653 <li>URI/file escaping problems (Stefano Zacchiroli)</li>
656 <h3>2.5.1: Jan 8 2003</h3>
658 <li>Fixes a memory leak and configuration/compilation problems in 2.5.0</li>
659 <li>documentation updates (John)</li>
660 <li>a couple of XmlTextReader fixes</li>
663 <h3>2.5.0: Jan 6 2003</h3>
665 <li>New <a href="xmlreader.html">XmltextReader interface</a> based on C#
666 API (with help of Stéphane Bidoul)</li>
667 <li>Windows: more exports, including the new API (Igor)</li>
668 <li>XInclude fallback fix</li>
669 <li>Python: bindings for the new API, packaging (Stéphane Bidoul),
670 drv_libxml2.py Python xml.sax driver (Stéphane Bidoul), fixes, speedup
671 and iterators for Python-2.2 (Hannu Krosing)</li>
672 <li>Tutorial fixes (john Fleck and Niraj Tolia) xmllint man update
674 <li>Fix an XML parser bug raised by Vyacheslav Pindyura</li>
675 <li>Fix for VMS serialization (Nigel Hall) and config (Craig A. Berry)</li>
676 <li>Entities handling fixes</li>
677 <li>new API to optionally track node creation and deletion (Lukas
679 <li>Added documentation for the XmltextReader interface and some <a
680 href="guidelines.html">XML guidelines</a></li>
683 <h3>2.4.30: Dec 12 2002</h3>
685 <li>2.4.29 broke the python bindings, rereleasing</li>
686 <li>Improvement/fixes of the XML API generator, and couple of minor code
690 <h3>2.4.29: Dec 11 2002</h3>
692 <li>Windows fixes (Igor): Windows CE port, pthread linking, python bindings
693 (Stéphane Bidoul), Mingw (Magnus Henoch), and export list updates</li>
694 <li>Fix for prev in python bindings (ERDI Gergo)</li>
695 <li>Fix for entities handling (Marcus Clarke)</li>
696 <li>Refactored the XML and HTML dumps to a single code path, fixed XHTML1
698 <li>Fix for URI parsing when handling URNs with fragment identifiers</li>
699 <li>Fix for HTTP URL escaping problem</li>
700 <li>added an TextXmlReader (C#) like API (work in progress)</li>
701 <li>Rewrote the API in XML generation script, includes a C parser and saves
702 more informations needed for C# bindings</li>
705 <h3>2.4.28: Nov 22 2002</h3>
707 <li>a couple of python binding fixes</li>
708 <li>2 bug fixes in the XML push parser</li>
709 <li>potential memory leak removed (Martin Stoilov)</li>
710 <li>fix to the configure script for Unix (Dimitri Papadopoulos)</li>
711 <li>added encoding support for XInclude parse="text"</li>
712 <li>autodetection of XHTML1 and specific serialization rules added</li>
713 <li>nasty threading bug fixed (William Brack)</li>
716 <h3>2.4.27: Nov 17 2002</h3>
718 <li>fixes for the Python bindings</li>
719 <li>a number of bug fixes: SGML catalogs, xmlParseBalancedChunkMemory(),
720 HTML parser, Schemas (Charles Bozeman), document fragment support
721 (Christian Glahn), xmlReconciliateNs (Brian Stafford), XPointer,
722 xmlFreeNode(), xmlSAXParseMemory (Peter Jones), xmlGetNodePath (Petr
723 Pajas), entities processing</li>
724 <li>added grep to xmllint --shell</li>
725 <li>VMS update patch from Craig A. Berry</li>
726 <li>cleanup of the Windows build with support for more compilers (Igor),
727 better thread support on Windows</li>
728 <li>cleanup of Unix Makefiles and spec file</li>
729 <li>Improvements to the documentation (John Fleck)</li>
732 <h3>2.4.26: Oct 18 2002</h3>
734 <li>Patches for Windows CE port, improvements on Windows paths handling</li>
735 <li>Fixes to the validation code (DTD and Schemas), xmlNodeGetPath() ,
736 HTML serialization, Namespace compliance, and a number of small
740 <h3>2.4.25: Sep 26 2002</h3>
742 <li>A number of bug fixes: XPath, validation, Python bindings, DOM and
743 tree, xmlI/O, Html</li>
744 <li>Serious rewrite of XInclude</li>
745 <li>Made XML Schemas regexp part of the default build and APIs, small fix
746 and improvement of the regexp core</li>
747 <li>Changed the validation code to reuse XML Schemas regexp APIs</li>
748 <li>Better handling of Windows file paths, improvement of Makefiles (Igor,
749 Daniel Gehriger, Mark Vakoc)</li>
750 <li>Improved the python I/O bindings, the tests, added resolver and regexp
752 <li>New logos from Marc Liyanage</li>
753 <li>Tutorial improvements: John Fleck, Christopher Harris</li>
754 <li>Makefile: Fixes for AMD x86_64 (Mandrake), DESTDIR (Christophe
756 <li>removal of all stderr/perror use for error reporting</li>
757 <li>Better error reporting: XPath and DTD validation</li>
758 <li>update of the trio portability layer (Bjorn Reese)</li>
761 <p><strong>2.4.24: Aug 22 2002</strong></p>
763 <li>XPath fixes (William), xf:escape-uri() (Wesley Terpstra)</li>
764 <li>Python binding fixes: makefiles (William), generator, rpm build, x86-64
766 <li>HTML <style> and boolean attributes serializer fixes</li>
767 <li>C14N improvements by Aleksey</li>
768 <li>doc cleanups: Rick Jones</li>
769 <li>Windows compiler makefile updates: Igor and Elizabeth Barham</li>
770 <li>XInclude: implementation of fallback and xml:base fixup added</li>
773 <h3>2.4.23: July 6 2002</h3>
775 <li>performances patches: Peter Jacobi</li>
776 <li>c14n fixes, testsuite and performances: Aleksey Sanin</li>
777 <li>added xmlDocFormatDump: Chema Celorio</li>
778 <li>new tutorial: John Fleck</li>
779 <li>new hash functions and performances: Sander Vesik, portability fix from
781 <li>a number of bug fixes: XPath (William Brack, Richard Jinks), XML and
782 HTML parsers, ID lookup function</li>
783 <li>removal of all remaining sprintf: Aleksey Sanin</li>
786 <h3>2.4.22: May 27 2002</h3>
788 <li>a number of bug fixes: configure scripts, base handling, parser, memory
789 usage, HTML parser, XPath, documentation (Christian Cornelssen),
790 indentation, URI parsing</li>
791 <li>Optimizations for XMLSec, fixing and making public some of the network
792 protocol handlers (Aleksey)</li>
793 <li>performance patch from Gary Pennington</li>
794 <li>Charles Bozeman provided date and time support for XML Schemas
798 <h3>2.4.21: Apr 29 2002</h3>
800 <p>This release is both a bug fix release and also contains the early XML
801 Schemas <a href="http://www.w3.org/TR/xmlschema-1/">structures</a> and <a
802 href="http://www.w3.org/TR/xmlschema-2/">datatypes</a> code, beware, all
803 interfaces are likely to change, there is huge holes, it is clearly a work in
804 progress and don't even think of putting this code in a production system,
805 it's actually not compiled in by default. The real fixes are:</p>
807 <li>a couple of bugs or limitations introduced in 2.4.20</li>
808 <li>patches for Borland C++ and MSC by Igor</li>
809 <li>some fixes on XPath strings and conformance patches by Richard
811 <li>patch from Aleksey for the ExcC14N specification</li>
812 <li>OSF/1 bug fix by Bjorn</li>
815 <h3>2.4.20: Apr 15 2002</h3>
817 <li>bug fixes: file descriptor leak, XPath, HTML output, DTD validation</li>
818 <li>XPath conformance testing by Richard Jinks</li>
819 <li>Portability fixes: Solaris, MPE/iX, Windows, OSF/1, python bindings,
823 <h3>2.4.19: Mar 25 2002</h3>
825 <li>bug fixes: half a dozen XPath bugs, Validation, ISO-Latin to UTF8
827 <li>portability fixes in the HTTP code</li>
828 <li>memory allocation checks using valgrind, and profiling tests</li>
829 <li>revamp of the Windows build and Makefiles</li>
832 <h3>2.4.18: Mar 18 2002</h3>
834 <li>bug fixes: tree, SAX, canonicalization, validation, portability,
836 <li>removed the --with-buffer option it was becoming unmaintainable</li>
837 <li>serious cleanup of the Python makefiles</li>
838 <li>speedup patch to XPath very effective for DocBook stylesheets</li>
839 <li>Fixes for Windows build, cleanup of the documentation</li>
842 <h3>2.4.17: Mar 8 2002</h3>
844 <li>a lot of bug fixes, including "namespace nodes have no parents in
846 <li>fixed/improved the Python wrappers, added more examples and more
847 regression tests, XPath extension functions can now return node-sets</li>
848 <li>added the XML Canonicalization support from Aleksey Sanin</li>
851 <h3>2.4.16: Feb 20 2002</h3>
853 <li>a lot of bug fixes, most of them were triggered by the XML Testsuite
854 from OASIS and W3C. Compliance has been significantly improved.</li>
855 <li>a couple of portability fixes too.</li>
858 <h3>2.4.15: Feb 11 2002</h3>
860 <li>Fixed the Makefiles, especially the python module ones</li>
861 <li>A few bug fixes and cleanup</li>
862 <li>Includes cleanup</li>
865 <h3>2.4.14: Feb 8 2002</h3>
867 <li>Change of License to the <a
868 href="http://www.opensource.org/licenses/mit-license.html">MIT
869 License</a> basically for integration in XFree86 codebase, and removing
870 confusion around the previous dual-licensing</li>
871 <li>added Python bindings, beta software but should already be quite
873 <li>a large number of fixes and cleanups, especially for all tree
875 <li>cleanup of the headers, generation of a reference API definition in
879 <h3>2.4.13: Jan 14 2002</h3>
881 <li>update of the documentation: John Fleck and Charlie Bozeman</li>
882 <li>cleanup of timing code from Justin Fletcher</li>
883 <li>fixes for Windows and initial thread support on Win32: Igor and Serguei
885 <li>Cygwin patch from Robert Collins</li>
886 <li>added xmlSetEntityReferenceFunc() for Keith Isdale work on xsldbg</li>
889 <h3>2.4.12: Dec 7 2001</h3>
891 <li>a few bug fixes: thread (Gary Pennington), xmllint (Geert Kloosterman),
892 XML parser (Robin Berjon), XPointer (Danny Jamshy), I/O cleanups
894 <li>Eric Lavigne contributed project files for MacOS</li>
895 <li>some makefiles cleanups</li>
898 <h3>2.4.11: Nov 26 2001</h3>
900 <li>fixed a couple of errors in the includes, fixed a few bugs, some code
902 <li>xmllint man pages improvement by Heiko Rupp</li>
903 <li>updated VMS build instructions from John A Fotheringham</li>
904 <li>Windows Makefiles updates from Igor</li>
907 <h3>2.4.10: Nov 10 2001</h3>
909 <li>URI escaping fix (Joel Young)</li>
910 <li>added xmlGetNodePath() (for paths or XPointers generation)</li>
911 <li>Fixes namespace handling problems when using DTD and validation</li>
912 <li>improvements on xmllint: Morus Walter patches for --format and
913 --encode, Stefan Kost and Heiko Rupp improvements on the --shell</li>
914 <li>fixes for xmlcatalog linking pointed by Weiqi Gao</li>
915 <li>fixes to the HTML parser</li>
918 <h3>2.4.9: Nov 6 2001</h3>
920 <li>fixes more catalog bugs</li>
921 <li>avoid a compilation problem, improve xmlGetLineNo()</li>
924 <h3>2.4.8: Nov 4 2001</h3>
926 <li>fixed SGML catalogs broken in previous release, updated xmlcatalog
928 <li>fixed a compile errors and some includes troubles.</li>
931 <h3>2.4.7: Oct 30 2001</h3>
933 <li>exported some debugging interfaces</li>
934 <li>serious rewrite of the catalog code</li>
935 <li>integrated Gary Pennington thread safety patch, added configure option
936 and regression tests</li>
937 <li>removed an HTML parser bug</li>
938 <li>fixed a couple of potentially serious validation bugs</li>
939 <li>integrated the SGML DocBook support in xmllint</li>
940 <li>changed the nanoftp anonymous login passwd</li>
941 <li>some I/O cleanup and a couple of interfaces for Perl wrapper</li>
942 <li>general bug fixes</li>
943 <li>updated xmllint man page by John Fleck</li>
944 <li>some VMS and Windows updates</li>
947 <h3>2.4.6: Oct 10 2001</h3>
949 <li>added an updated man pages by John Fleck</li>
950 <li>portability and configure fixes</li>
951 <li>an infinite loop on the HTML parser was removed (William)</li>
952 <li>Windows makefile patches from Igor</li>
953 <li>fixed half a dozen bugs reported for libxml or libxslt</li>
954 <li>updated xmlcatalog to be able to modify SGML super catalogs</li>
957 <h3>2.4.5: Sep 14 2001</h3>
959 <li>Remove a few annoying bugs in 2.4.4</li>
960 <li>forces the HTML serializer to output decimal charrefs since some
961 version of Netscape can't handle hexadecimal ones</li>
964 <h3>1.8.16: Sep 14 2001</h3>
966 <li>maintenance release of the old libxml1 branch, couple of bug and
967 portability fixes</li>
970 <h3>2.4.4: Sep 12 2001</h3>
972 <li>added --convert to xmlcatalog, bug fixes and cleanups of XML
974 <li>a few bug fixes and some portability changes</li>
975 <li>some documentation cleanups</li>
978 <h3>2.4.3: Aug 23 2001</h3>
980 <li>XML Catalog support see the doc</li>
981 <li>New NaN/Infinity floating point code</li>
982 <li>A few bug fixes</li>
985 <h3>2.4.2: Aug 15 2001</h3>
987 <li>adds xmlLineNumbersDefault() to control line number generation</li>
988 <li>lot of bug fixes</li>
989 <li>the Microsoft MSC projects files should now be up to date</li>
990 <li>inheritance of namespaces from DTD defaulted attributes</li>
991 <li>fixes a serious potential security bug</li>
992 <li>added a --format option to xmllint</li>
995 <h3>2.4.1: July 24 2001</h3>
997 <li>possibility to keep line numbers in the tree</li>
998 <li>some computation NaN fixes</li>
999 <li>extension of the XPath API</li>
1000 <li>cleanup for alpha and ia64 targets</li>
1001 <li>patch to allow saving through HTTP PUT or POST</li>
1004 <h3>2.4.0: July 10 2001</h3>
1006 <li>Fixed a few bugs in XPath, validation, and tree handling.</li>
1007 <li>Fixed XML Base implementation, added a couple of examples to the
1008 regression tests</li>
1009 <li>A bit of cleanup</li>
1012 <h3>2.3.14: July 5 2001</h3>
1014 <li>fixed some entities problems and reduce memory requirement when
1015 substituting them</li>
1016 <li>lots of improvements in the XPath queries interpreter can be
1017 substantially faster</li>
1018 <li>Makefiles and configure cleanups</li>
1019 <li>Fixes to XPath variable eval, and compare on empty node set</li>
1020 <li>HTML tag closing bug fixed</li>
1021 <li>Fixed an URI reference computation problem when validating</li>
1024 <h3>2.3.13: June 28 2001</h3>
1026 <li>2.3.12 configure.in was broken as well as the push mode XML parser</li>
1027 <li>a few more fixes for compilation on Windows MSC by Yon Derek</li>
1030 <h3>1.8.14: June 28 2001</h3>
1032 <li>Zbigniew Chyla gave a patch to use the old XML parser in push mode</li>
1033 <li>Small Makefile fix</li>
1036 <h3>2.3.12: June 26 2001</h3>
1038 <li>lots of cleanup</li>
1039 <li>a couple of validation fix</li>
1040 <li>fixed line number counting</li>
1041 <li>fixed serious problems in the XInclude processing</li>
1042 <li>added support for UTF8 BOM at beginning of entities</li>
1043 <li>fixed a strange gcc optimizer bugs in xpath handling of float, gcc-3.0
1044 miscompile uri.c (William), Thomas Leitner provided a fix for the
1045 optimizer on Tru64</li>
1046 <li>incorporated Yon Derek and Igor Zlatkovic fixes and improvements for
1047 compilation on Windows MSC</li>
1048 <li>update of libxml-doc.el (Felix Natter)</li>
1049 <li>fixed 2 bugs in URI normalization code</li>
1052 <h3>2.3.11: June 17 2001</h3>
1054 <li>updates to trio, Makefiles and configure should fix some portability
1055 problems (alpha)</li>
1056 <li>fixed some HTML serialization problems (pre, script, and block/inline
1057 handling), added encoding aware APIs, cleanup of this code</li>
1058 <li>added xmlHasNsProp()</li>
1059 <li>implemented a specific PI for encoding support in the DocBook SGML
1061 <li>some XPath fixes (-Infinity, / as a function parameter and namespaces
1062 node selection)</li>
1063 <li>fixed a performance problem and an error in the validation code</li>
1064 <li>fixed XInclude routine to implement the recursive behaviour</li>
1065 <li>fixed xmlFreeNode problem when libxml is included statically twice</li>
1066 <li>added --version to xmllint for bug reports</li>
1069 <h3>2.3.10: June 1 2001</h3>
1071 <li>fixed the SGML catalog support</li>
1072 <li>a number of reported bugs got fixed, in XPath, iconv detection,
1073 XInclude processing</li>
1074 <li>XPath string function should now handle unicode correctly</li>
1077 <h3>2.3.9: May 19 2001</h3>
1079 <p>Lots of bugfixes, and added a basic SGML catalog support:</p>
1081 <li>HTML push bugfix #54891 and another patch from Jonas Borgström</li>
1082 <li>some serious speed optimization again</li>
1083 <li>some documentation cleanups</li>
1084 <li>trying to get better linking on Solaris (-R)</li>
1085 <li>XPath API cleanup from Thomas Broyer</li>
1086 <li>Validation bug fixed #54631, added a patch from Gary Pennington, fixed
1087 xmlValidGetValidElements()</li>
1088 <li>Added an INSTALL file</li>
1089 <li>Attribute removal added to API: #54433</li>
1090 <li>added a basic support for SGML catalogs</li>
1091 <li>fixed xmlKeepBlanksDefault(0) API</li>
1092 <li>bugfix in xmlNodeGetLang()</li>
1093 <li>fixed a small configure portability problem</li>
1094 <li>fixed an inversion of SYSTEM and PUBLIC identifier in HTML document</li>
1097 <h3>1.8.13: May 14 2001</h3>
1099 <li>bugfixes release of the old libxml1 branch used by Gnome</li>
1102 <h3>2.3.8: May 3 2001</h3>
1104 <li>Integrated an SGML DocBook parser for the Gnome project</li>
1105 <li>Fixed a few things in the HTML parser</li>
1106 <li>Fixed some XPath bugs raised by XSLT use, tried to fix the floating
1107 point portability issue</li>
1108 <li>Speed improvement (8M/s for SAX, 3M/s for DOM, 1.5M/s for
1109 DOM+validation using the XML REC as input and a 700MHz celeron).</li>
1110 <li>incorporated more Windows cleanup</li>
1111 <li>added xmlSaveFormatFile()</li>
1112 <li>fixed problems in copying nodes with entities references (gdome)</li>
1113 <li>removed some troubles surrounding the new validation module</li>
1116 <h3>2.3.7: April 22 2001</h3>
1118 <li>lots of small bug fixes, corrected XPointer</li>
1119 <li>Non deterministic content model validation support</li>
1120 <li>added xmlDocCopyNode for gdome2</li>
1121 <li>revamped the way the HTML parser handles end of tags</li>
1122 <li>XPath: corrections of namespaces support and number formatting</li>
1123 <li>Windows: Igor Zlatkovic patches for MSC compilation</li>
1124 <li>HTML output fixes from P C Chow and William M. Brack</li>
1125 <li>Improved validation speed sensible for DocBook</li>
1126 <li>fixed a big bug with ID declared in external parsed entities</li>
1127 <li>portability fixes, update of Trio from Bjorn Reese</li>
1130 <h3>2.3.6: April 8 2001</h3>
1132 <li>Code cleanup using extreme gcc compiler warning options, found and
1133 cleared half a dozen potential problem</li>
1134 <li>the Eazel team found an XML parser bug</li>
1135 <li>cleaned up the user of some of the string formatting function. used the
1136 trio library code to provide the one needed when the platform is missing
1138 <li>xpath: removed a memory leak and fixed the predicate evaluation
1139 problem, extended the testsuite and cleaned up the result. XPointer seems
1143 <h3>2.3.5: Mar 23 2001</h3>
1145 <li>Biggest change is separate parsing and evaluation of XPath expressions,
1146 there is some new APIs for this too</li>
1147 <li>included a number of bug fixes(XML push parser, 51876, notations,
1149 <li>Fixed some portability issues</li>
1152 <h3>2.3.4: Mar 10 2001</h3>
1154 <li>Fixed bugs #51860 and #51861</li>
1155 <li>Added a global variable xmlDefaultBufferSize to allow default buffer
1156 size to be application tunable.</li>
1157 <li>Some cleanup in the validation code, still a bug left and this part
1158 should probably be rewritten to support ambiguous content model :-\</li>
1159 <li>Fix a couple of serious bugs introduced or raised by changes in 2.3.3
1161 <li>Fixed another bug in xmlNodeGetContent()</li>
1162 <li>Bjorn fixed XPath node collection and Number formatting</li>
1163 <li>Fixed a loop reported in the HTML parsing</li>
1164 <li>blank space are reported even if the Dtd content model proves that they
1165 are formatting spaces, this is for XML conformance</li>
1168 <h3>2.3.3: Mar 1 2001</h3>
1170 <li>small change in XPath for XSLT</li>
1171 <li>documentation cleanups</li>
1172 <li>fix in validation by Gary Pennington</li>
1173 <li>serious parsing performances improvements</li>
1176 <h3>2.3.2: Feb 24 2001</h3>
1178 <li>chasing XPath bugs, found a bunch, completed some TODO</li>
1179 <li>fixed a Dtd parsing bug</li>
1180 <li>fixed a bug in xmlNodeGetContent</li>
1181 <li>ID/IDREF support partly rewritten by Gary Pennington</li>
1184 <h3>2.3.1: Feb 15 2001</h3>
1186 <li>some XPath and HTML bug fixes for XSLT</li>
1187 <li>small extension of the hash table interfaces for DOM gdome2
1189 <li>A few bug fixes</li>
1192 <h3>2.3.0: Feb 8 2001 (2.2.12 was on 25 Jan but I didn't kept track)</h3>
1194 <li>Lots of XPath bug fixes</li>
1195 <li>Add a mode with Dtd lookup but without validation error reporting for
1197 <li>Add support for text node without escaping (XSLT)</li>
1198 <li>bug fixes for xmlCheckFilename</li>
1199 <li>validation code bug fixes from Gary Pennington</li>
1200 <li>Patch from Paul D. Smith correcting URI path normalization</li>
1201 <li>Patch to allow simultaneous install of libxml-devel and
1203 <li>the example Makefile is now fixed</li>
1204 <li>added HTML to the RPM packages</li>
1205 <li>tree copying bugfixes</li>
1206 <li>updates to Windows makefiles</li>
1207 <li>optimization patch from Bjorn Reese</li>
1210 <h3>2.2.11: Jan 4 2001</h3>
1212 <li>bunch of bug fixes (memory I/O, xpath, ftp/http, ...)</li>
1213 <li>added htmlHandleOmittedElem()</li>
1214 <li>Applied Bjorn Reese's IPV6 first patch</li>
1215 <li>Applied Paul D. Smith patches for validation of XInclude results</li>
1216 <li>added XPointer xmlns() new scheme support</li>
1219 <h3>2.2.10: Nov 25 2000</h3>
1221 <li>Fix the Windows problems of 2.2.8</li>
1222 <li>integrate OpenVMS patches</li>
1223 <li>better handling of some nasty HTML input</li>
1224 <li>Improved the XPointer implementation</li>
1225 <li>integrate a number of provided patches</li>
1228 <h3>2.2.9: Nov 25 2000</h3>
1230 <li>erroneous release :-(</li>
1233 <h3>2.2.8: Nov 13 2000</h3>
1235 <li>First version of <a href="http://www.w3.org/TR/xinclude">XInclude</a>
1237 <li>Patch in conditional section handling</li>
1238 <li>updated MS compiler project</li>
1239 <li>fixed some XPath problems</li>
1240 <li>added an URI escaping function</li>
1241 <li>some other bug fixes</li>
1244 <h3>2.2.7: Oct 31 2000</h3>
1246 <li>added message redirection</li>
1247 <li>XPath improvements (thanks TOM !)</li>
1248 <li>xmlIOParseDTD() added</li>
1249 <li>various small fixes in the HTML, URI, HTTP and XPointer support</li>
1250 <li>some cleanup of the Makefile, autoconf and the distribution content</li>
1253 <h3>2.2.6: Oct 25 2000:</h3>
1255 <li>Added an hash table module, migrated a number of internal structure to
1257 <li>Fixed a posteriori validation problems</li>
1258 <li>HTTP module cleanups</li>
1259 <li>HTML parser improvements (tag errors, script/style handling, attribute
1261 <li>coalescing of adjacent text nodes</li>
1262 <li>couple of XPath bug fixes, exported the internal API</li>
1265 <h3>2.2.5: Oct 15 2000:</h3>
1267 <li>XPointer implementation and testsuite</li>
1268 <li>Lot of XPath fixes, added variable and functions registration, more
1270 <li>Portability fixes, lots of enhancements toward an easy Windows build
1272 <li>Late validation fixes</li>
1273 <li>Integrated a lot of contributed patches</li>
1274 <li>added memory management docs</li>
1275 <li>a performance problem when using large buffer seems fixed</li>
1278 <h3>2.2.4: Oct 1 2000:</h3>
1280 <li>main XPath problem fixed</li>
1281 <li>Integrated portability patches for Windows</li>
1282 <li>Serious bug fixes on the URI and HTML code</li>
1285 <h3>2.2.3: Sep 17 2000</h3>
1288 <li>cleanup of entity handling code</li>
1289 <li>overall review of all loops in the parsers, all sprintf usage has been
1291 <li>Far better handling of larges Dtd. Validating against DocBook XML Dtd
1292 works smoothly now.</li>
1295 <h3>1.8.10: Sep 6 2000</h3>
1297 <li>bug fix release for some Gnome projects</li>
1300 <h3>2.2.2: August 12 2000</h3>
1302 <li>mostly bug fixes</li>
1303 <li>started adding routines to access xml parser context options</li>
1306 <h3>2.2.1: July 21 2000</h3>
1308 <li>a purely bug fixes release</li>
1309 <li>fixed an encoding support problem when parsing from a memory block</li>
1310 <li>fixed a DOCTYPE parsing problem</li>
1311 <li>removed a bug in the function allowing to override the memory
1312 allocation routines</li>
1315 <h3>2.2.0: July 14 2000</h3>
1317 <li>applied a lot of portability fixes</li>
1318 <li>better encoding support/cleanup and saving (content is now always
1319 encoded in UTF-8)</li>
1320 <li>the HTML parser now correctly handles encodings</li>
1321 <li>added xmlHasProp()</li>
1322 <li>fixed a serious problem with &#38;</li>
1323 <li>propagated the fix to FTP client</li>
1324 <li>cleanup, bugfixes, etc ...</li>
1325 <li>Added a page about <a href="encoding.html">libxml Internationalization
1329 <h3>1.8.9: July 9 2000</h3>
1331 <li>fixed the spec the RPMs should be better</li>
1332 <li>fixed a serious bug in the FTP implementation, released 1.8.9 to solve
1333 rpmfind users problem</li>
1336 <h3>2.1.1: July 1 2000</h3>
1338 <li>fixes a couple of bugs in the 2.1.0 packaging</li>
1339 <li>improvements on the HTML parser</li>
1342 <h3>2.1.0 and 1.8.8: June 29 2000</h3>
1344 <li>1.8.8 is mostly a commodity package for upgrading to libxml2 according
1345 to <a href="upgrade.html">new instructions</a>. It fixes a nasty problem
1346 about &#38; charref parsing</li>
1347 <li>2.1.0 also ease the upgrade from libxml v1 to the recent version. it
1348 also contains numerous fixes and enhancements:
1350 <li>added xmlStopParser() to stop parsing</li>
1351 <li>improved a lot parsing speed when there is large CDATA blocs</li>
1352 <li>includes XPath patches provided by Picdar Technology</li>
1353 <li>tried to fix as much as possible DTD validation and namespace
1354 related problems</li>
1355 <li>output to a given encoding has been added/tested</li>
1356 <li>lot of various fixes</li>
1361 <h3>2.0.0: Apr 12 2000</h3>
1363 <li>First public release of libxml2. If you are using libxml, it's a good
1364 idea to check the 1.x to 2.x upgrade instructions. NOTE: while initially
1365 scheduled for Apr 3 the release occurred only on Apr 12 due to massive
1367 <li>The include are now located under $prefix/include/libxml (instead of
1368 $prefix/include/gnome-xml), they also are referenced by
1369 <pre>#include <libxml/xxx.h></pre>
1371 <pre>#include "xxx.h"</pre>
1373 <li>a new URI module for parsing URIs and following strictly RFC 2396</li>
1374 <li>the memory allocation routines used by libxml can now be overloaded
1375 dynamically by using xmlMemSetup()</li>
1376 <li>The previously CVS only tool tester has been renamed
1377 <strong>xmllint</strong> and is now installed as part of the libxml2
1379 <li>The I/O interface has been revamped. There is now ways to plug in
1380 specific I/O modules, either at the URI scheme detection level using
1381 xmlRegisterInputCallbacks() or by passing I/O functions when creating a
1382 parser context using xmlCreateIOParserCtxt()</li>
1383 <li>there is a C preprocessor macro LIBXML_VERSION providing the version
1384 number of the libxml module in use</li>
1385 <li>a number of optional features of libxml can now be excluded at
1386 configure time (FTP/HTTP/HTML/XPath/Debug)</li>
1389 <h3>2.0.0beta: Mar 14 2000</h3>
1391 <li>This is a first Beta release of libxml version 2</li>
1392 <li>It's available only from<a href="ftp://xmlsoft.org/">xmlsoft.org
1393 FTP</a>, it's packaged as libxml2-2.0.0beta and available as tar and
1395 <li>This version is now the head in the Gnome CVS base, the old one is
1396 available under the tag LIB_XML_1_X</li>
1397 <li>This includes a very large set of changes. From a programmatic point
1398 of view applications should not have to be modified too much, check the
1399 <a href="upgrade.html">upgrade page</a></li>
1400 <li>Some interfaces may changes (especially a bit about encoding).</li>
1401 <li>the updates includes:
1403 <li>fix I18N support. ISO-Latin-x/UTF-8/UTF-16 (nearly) seems correctly
1405 <li>Better handling of entities, especially well-formedness checking
1406 and proper PEref extensions in external subsets</li>
1407 <li>DTD conditional sections</li>
1408 <li>Validation now correctly handle entities content</li>
1409 <li><a href="http://rpmfind.net/tools/gdome/messages/0039.html">change
1410 structures to accommodate DOM</a></li>
1413 <li>Serious progress were made toward compliance, <a
1414 href="conf/result.html">here are the result of the test</a> against the
1415 OASIS testsuite (except the Japanese tests since I don't support that
1416 encoding yet). This URL is rebuilt every couple of hours using the CVS
1420 <h3>1.8.7: Mar 6 2000</h3>
1422 <li>This is a bug fix release:</li>
1423 <li>It is possible to disable the ignorable blanks heuristic used by
1424 libxml-1.x, a new function xmlKeepBlanksDefault(0) will allow this. Note
1425 that for adherence to XML spec, this behaviour will be disabled by
1426 default in 2.x . The same function will allow to keep compatibility for
1428 <li>Blanks in <a> </a> constructs are not ignored anymore,
1429 avoiding heuristic is really the Right Way :-\</li>
1430 <li>The unchecked use of snprintf which was breaking libxml-1.8.6
1431 compilation on some platforms has been fixed</li>
1432 <li>nanoftp.c nanohttp.c: Fixed '#' and '?' stripping when processing
1436 <h3>1.8.6: Jan 31 2000</h3>
1438 <li>added a nanoFTP transport module, debugged until the new version of <a
1439 href="http://rpmfind.net/linux/rpm2html/rpmfind.html">rpmfind</a> can use
1440 it without troubles</li>
1443 <h3>1.8.5: Jan 21 2000</h3>
1445 <li>adding APIs to parse a well balanced chunk of XML (production <a
1446 href="http://www.w3.org/TR/REC-xml#NT-content">[43] content</a> of the
1448 <li>fixed a hideous bug in xmlGetProp pointed by Rune.Djurhuus@fast.no</li>
1449 <li>Jody Goldberg <jgoldberg@home.com> provided another patch trying
1450 to solve the zlib checks problems</li>
1451 <li>The current state in gnome CVS base is expected to ship as 1.8.5 with
1455 <h3>1.8.4: Jan 13 2000</h3>
1457 <li>bug fixes, reintroduced xmlNewGlobalNs(), fixed xmlNewNs()</li>
1458 <li>all exit() call should have been removed from libxml</li>
1459 <li>fixed a problem with INCLUDE_WINSOCK on WIN32 platform</li>
1460 <li>added newDocFragment()</li>
1463 <h3>1.8.3: Jan 5 2000</h3>
1465 <li>a Push interface for the XML and HTML parsers</li>
1466 <li>a shell-like interface to the document tree (try tester --shell :-)</li>
1467 <li>lots of bug fixes and improvement added over XMas holidays</li>
1468 <li>fixed the DTD parsing code to work with the xhtml DTD</li>
1469 <li>added xmlRemoveProp(), xmlRemoveID() and xmlRemoveRef()</li>
1470 <li>Fixed bugs in xmlNewNs()</li>
1471 <li>External entity loading code has been revamped, now it uses
1472 xmlLoadExternalEntity(), some fix on entities processing were added</li>
1473 <li>cleaned up WIN32 includes of socket stuff</li>
1476 <h3>1.8.2: Dec 21 1999</h3>
1478 <li>I got another problem with includes and C++, I hope this issue is fixed
1479 for good this time</li>
1480 <li>Added a few tree modification functions: xmlReplaceNode,
1481 xmlAddPrevSibling, xmlAddNextSibling, xmlNodeSetName and
1482 xmlDocSetRootElement</li>
1483 <li>Tried to improve the HTML output with help from <a
1484 href="mailto:clahey@umich.edu">Chris Lahey</a></li>
1487 <h3>1.8.1: Dec 18 1999</h3>
1489 <li>various patches to avoid troubles when using libxml with C++ compilers
1490 the "namespace" keyword and C escaping in include files</li>
1491 <li>a problem in one of the core macros IS_CHAR was corrected</li>
1492 <li>fixed a bug introduced in 1.8.0 breaking default namespace processing,
1493 and more specifically the Dia application</li>
1494 <li>fixed a posteriori validation (validation after parsing, or by using a
1495 Dtd not specified in the original document)</li>
1496 <li>fixed a bug in</li>
1499 <h3>1.8.0: Dec 12 1999</h3>
1501 <li>cleanup, especially memory wise</li>
1502 <li>the parser should be more reliable, especially the HTML one, it should
1503 not crash, whatever the input !</li>
1504 <li>Integrated various patches, especially a speedup improvement for large
1505 dataset from <a href="mailto:cnygard@bellatlantic.net">Carl Nygard</a>,
1506 configure with --with-buffers to enable them.</li>
1507 <li>attribute normalization, oops should have been added long ago !</li>
1508 <li>attributes defaulted from DTDs should be available, xmlSetProp() now
1509 does entities escaping by default.</li>
1512 <h3>1.7.4: Oct 25 1999</h3>
1514 <li>Lots of HTML improvement</li>
1515 <li>Fixed some errors when saving both XML and HTML</li>
1516 <li>More examples, the regression tests should now look clean</li>
1517 <li>Fixed a bug with contiguous charref</li>
1520 <h3>1.7.3: Sep 29 1999</h3>
1522 <li>portability problems fixed</li>
1523 <li>snprintf was used unconditionally, leading to link problems on system
1524 were it's not available, fixed</li>
1527 <h3>1.7.1: Sep 24 1999</h3>
1529 <li>The basic type for strings manipulated by libxml has been renamed in
1530 1.7.1 from <strong>CHAR</strong> to <strong>xmlChar</strong>. The reason
1531 is that CHAR was conflicting with a predefined type on Windows. However
1532 on non WIN32 environment, compatibility is provided by the way of a
1533 <strong>#define </strong>.</li>
1534 <li>Changed another error : the use of a structure field called errno, and
1535 leading to troubles on platforms where it's a macro</li>
1538 <h3>1.7.0: Sep 23 1999</h3>
1540 <li>Added the ability to fetch remote DTD or parsed entities, see the <a
1541 href="html/libxml-nanohttp.html">nanohttp</a> module.</li>
1542 <li>Added an errno to report errors by another mean than a simple printf
1544 <li>Finished ID/IDREF support and checking when validation</li>
1545 <li>Serious memory leaks fixed (there is now a <a
1546 href="html/libxml-xmlmemory.html">memory wrapper</a> module)</li>
1547 <li>Improvement of <a href="http://www.w3.org/TR/xpath">XPath</a>
1549 <li>Added an HTML parser front-end</li>
1552 <h2><a name="XML">XML</a></h2>
1554 <p><a href="http://www.w3.org/TR/REC-xml">XML is a standard</a> for
1555 markup-based structured documents. Here is <a name="example">an example XML
1557 <pre><?xml version="1.0"?>
1558 <EXAMPLE prop1="gnome is great" prop2="&amp; linux too">
1560 <title>Welcome to Gnome</title>
1563 <title>The Linux adventure</title>
1564 <p>bla bla bla ...</p>
1565 <image href="linus.gif"/>
1566 <p>...</p>
1568 </EXAMPLE></pre>
1570 <p>The first line specifies that it is an XML document and gives useful
1571 information about its encoding. Then the rest of the document is a text
1572 format whose structure is specified by tags between brackets. <strong>Each
1573 tag opened has to be closed</strong>. XML is pedantic about this. However, if
1574 a tag is empty (no content), a single tag can serve as both the opening and
1575 closing tag if it ends with <code>/></code> rather than with
1576 <code>></code>. Note that, for example, the image tag has no content (just
1577 an attribute) and is closed by ending the tag with <code>/></code>.</p>
1579 <p>XML can be applied successfully to a wide range of tasks, ranging from
1580 long term structured document maintenance (where it follows the steps of
1581 SGML) to simple data encoding mechanisms like configuration file formatting
1582 (glade), spreadsheets (gnumeric), or even shorter lived documents such as
1583 WebDAV where it is used to encode remote calls between a client and a
1586 <h2><a name="XSLT">XSLT</a></h2>
1588 <p>Check <a href="http://xmlsoft.org/XSLT">the separate libxslt page</a></p>
1590 <p><a href="http://www.w3.org/TR/xslt">XSL Transformations</a>, is a
1591 language for transforming XML documents into other XML documents (or
1592 HTML/textual output).</p>
1594 <p>A separate library called libxslt is being developed on top of libxml2.
1595 This module "libxslt" too can be found in the Gnome CVS base.</p>
1597 <p>You can check the <a
1598 href="http://cvs.gnome.org/lxr/source/libxslt/FEATURES">features</a>
1599 supported and the progresses on the <a
1600 href="http://cvs.gnome.org/lxr/source/libxslt/ChangeLog"
1601 name="Changelog">Changelog</a>.</p>
1603 <h2><a name="Python">Python and bindings</a></h2>
1605 <p>There are a number of language bindings and wrappers available for
1606 libxml2, the list below is not exhaustive. Please contact the <a
1607 href="http://mail.gnome.org/mailman/listinfo/xml-bindings">xml-bindings@gnome.org</a>
1608 (<a href="http://mail.gnome.org/archives/xml-bindings/">archives</a>) in
1609 order to get updates to this list or to discuss the specific topic of libxml2
1610 or libxslt wrappers or bindings:</p>
1612 <li><a href="http://libxmlplusplus.sourceforge.net/">Libxml++</a> seems the
1613 most up-to-date C++ bindings for libxml2, check the <a
1614 href="http://libxmlplusplus.sourceforge.net/reference/html/hierarchy.html">documentation</a>
1616 href="http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/libxmlplusplus/libxml%2b%2b/examples/">examples</a>.</li>
1617 <li>There is another <a href="http://libgdome-cpp.berlios.de/">C++ wrapper
1618 based on the gdome2 bindings</a> maintained by Tobias Peters.</li>
1619 <li>and a third C++ wrapper by Peter Jones <pjones@pmade.org>
1621 href="http://pmade.org/pjones/software/xmlwrapp/">http://pmade.org/pjones/software/xmlwrapp/</a></p>
1624 href="http://mail.gnome.org/archives/xml/2001-March/msg00014.html">Matt
1625 Sergeant</a> developed <a
1626 href="http://axkit.org/download/">XML::LibXSLT</a>, a Perl wrapper for
1627 libxml2/libxslt as part of the <a href="http://axkit.com/">AxKit XML
1628 application server</a>.</li>
1629 <li><a href="mailto:dkuhlman@cutter.rexx.com">Dave Kuhlman</a> provides an
1630 earlier version of the libxml/libxslt <a
1631 href="http://www.rexx.com/~dkuhlman">wrappers for Python</a>.</li>
1632 <li>Gopal.V and Peter Minten develop <a
1633 href="http://savannah.gnu.org/projects/libxmlsharp">libxml#</a>, a set of
1634 C# libxml2 bindings.</li>
1635 <li>Petr Kozelka provides <a
1636 href="http://sourceforge.net/projects/libxml2-pas">Pascal units to glue
1637 libxml2</a> with Kylix, Delphi and other Pascal compilers.</li>
1638 <li>Uwe Fechner also provides <a
1639 href="http://sourceforge.net/projects/idom2-pas/">idom2</a>, a DOM2
1640 implementation for Kylix2/D5/D6 from Borland.</li>
1641 <li>Wai-Sun "Squidster" Chia provides <a
1642 href="http://www.rubycolor.org/arc/redist/">bindings for Ruby</a> and
1643 libxml2 bindings are also available in Ruby through the <a
1644 href="http://libgdome-ruby.berlios.de/">libgdome-ruby</a> module
1645 maintained by Tobias Peters.</li>
1646 <li>Steve Ball and contributors maintains <a
1647 href="http://tclxml.sourceforge.net/">libxml2 and libxslt bindings for
1649 <li>There is support for libxml2 in the DOM module of PHP.</li>
1652 <p>The distribution includes a set of Python bindings, which are guaranteed
1653 to be maintained as part of the library in the future, though the Python
1654 interface have not yet reached the maturity of the C API.</p>
1656 <p><a href="mailto:stephane.bidoul@softwareag.com">Stéphane Bidoul</a>
1657 maintains <a href="http://users.skynet.be/sbi/libxml-python/">a Windows port
1658 of the Python bindings</a>.</p>
1660 <p>Note to people interested in building bindings, the API is formalized as
1661 <a href="libxml2-api.xml">an XML API description file</a> which allows to
1662 automate a large part of the Python bindings, this includes function
1663 descriptions, enums, structures, typedefs, etc... The Python script used to
1664 build the bindings is python/generator.py in the source distribution.</p>
1666 <p>To install the Python bindings there are 2 options:</p>
1668 <li>If you use an RPM based distribution, simply install the <a
1669 href="http://rpmfind.net/linux/rpm2html/search.php?query=libxml2-python">libxml2-python
1670 RPM</a> (and if needed the <a
1671 href="http://rpmfind.net/linux/rpm2html/search.php?query=libxslt-python">libxslt-python
1673 <li>Otherwise use the <a href="ftp://xmlsoft.org/python/">libxml2-python
1674 module distribution</a> corresponding to your installed version of
1675 libxml2 and libxslt. Note that to install it you will need both libxml2
1676 and libxslt installed and run "python setup.py build install" in the
1680 <p>The distribution includes a set of examples and regression tests for the
1681 python bindings in the <code>python/tests</code> directory. Here are some
1682 excerpts from those tests:</p>
1686 <p>This is a basic test of the file interface and DOM navigation:</p>
1689 doc = libxml2.parseFile("tst.xml")
1690 if doc.name != "tst.xml":
1691 print "doc.name failed"
1694 if root.name != "doc":
1695 print "root.name failed"
1697 child = root.children
1698 if child.name != "foo":
1699 print "child.name failed"
1703 <p>The Python module is called libxml2; parseFile is the equivalent of
1704 xmlParseFile (most of the bindings are automatically generated, and the xml
1705 prefix is removed and the casing convention are kept). All node seen at the
1706 binding level share the same subset of accessors:</p>
1708 <li><code>name</code> : returns the node name</li>
1709 <li><code>type</code> : returns a string indicating the node type</li>
1710 <li><code>content</code> : returns the content of the node, it is based on
1711 xmlNodeGetContent() and hence is recursive.</li>
1712 <li><code>parent</code> , <code>children</code>, <code>last</code>,
1713 <code>next</code>, <code>prev</code>, <code>doc</code>,
1714 <code>properties</code>: pointing to the associated element in the tree,
1715 those may return None in case no such link exists.</li>
1718 <p>Also note the need to explicitly deallocate documents with freeDoc() .
1719 Reference counting for libxml2 trees would need quite a lot of work to
1720 function properly, and rather than risk memory leaks if not implemented
1721 correctly it sounds safer to have an explicit function to free a tree. The
1722 wrapper python objects like doc, root or child are them automatically garbage
1725 <h3>validate.py:</h3>
1727 <p>This test check the validation interfaces and redirection of error
1731 #deactivate error messages from the validation
1732 def noerr(ctx, str):
1735 libxml2.registerErrorHandler(noerr, None)
1737 ctxt = libxml2.createFileParserCtxt("invalid.xml")
1739 ctxt.parseDocument()
1741 valid = ctxt.isValid()
1744 print "validity check failed"</pre>
1746 <p>The first thing to notice is the call to registerErrorHandler(), it
1747 defines a new error handler global to the library. It is used to avoid seeing
1748 the error messages when trying to validate the invalid document.</p>
1750 <p>The main interest of that test is the creation of a parser context with
1751 createFileParserCtxt() and how the behaviour can be changed before calling
1752 parseDocument() . Similarly the informations resulting from the parsing phase
1753 are also available using context methods.</p>
1755 <p>Contexts like nodes are defined as class and the libxml2 wrappers maps the
1756 C function interfaces in terms of objects method as much as possible. The
1757 best to get a complete view of what methods are supported is to look at the
1758 libxml2.py module containing all the wrappers.</p>
1762 <p>This test show how to activate the push parser interface:</p>
1765 ctxt = libxml2.createPushParser(None, "<foo", 4, "test.xml")
1766 ctxt.parseChunk("/>", 2, 1)
1771 <p>The context is created with a special call based on the
1772 xmlCreatePushParser() from the C library. The first argument is an optional
1773 SAX callback object, then the initial set of data, the length and the name of
1774 the resource in case URI-References need to be computed by the parser.</p>
1776 <p>Then the data are pushed using the parseChunk() method, the last call
1777 setting the third argument terminate to 1.</p>
1779 <h3>pushSAX.py:</h3>
1781 <p>this test show the use of the event based parsing interfaces. In this case
1782 the parser does not build a document, but provides callback information as
1783 the parser makes progresses analyzing the data being provided:</p>
1788 def startDocument(self):
1790 log = log + "startDocument:"
1792 def endDocument(self):
1794 log = log + "endDocument:"
1796 def startElement(self, tag, attrs):
1798 log = log + "startElement %s %s:" % (tag, attrs)
1800 def endElement(self, tag):
1802 log = log + "endElement %s:" % (tag)
1804 def characters(self, data):
1806 log = log + "characters: %s:" % (data)
1808 def warning(self, msg):
1810 log = log + "warning: %s:" % (msg)
1812 def error(self, msg):
1814 log = log + "error: %s:" % (msg)
1816 def fatalError(self, msg):
1818 log = log + "fatalError: %s:" % (msg)
1820 handler = callback()
1822 ctxt = libxml2.createPushParser(handler, "<foo", 4, "test.xml")
1823 chunk = " url='tst'>b"
1824 ctxt.parseChunk(chunk, len(chunk), 0)
1825 chunk = "ar</foo>"
1826 ctxt.parseChunk(chunk, len(chunk), 1)
1828 reference = "startDocument:startElement foo {'url': 'tst'}:" + \
1829 "characters: bar:endElement foo:endDocument:"
1830 if log != reference:
1831 print "Error got: %s" % log
1832 print "Expected: %s" % reference</pre>
1834 <p>The key object in that test is the handler, it provides a number of entry
1835 points which can be called by the parser as it makes progresses to indicate
1836 the information set obtained. The full set of callback is larger than what
1837 the callback class in that specific example implements (see the SAX
1838 definition for a complete list). The wrapper will only call those supplied by
1839 the object when activated. The startElement receives the names of the element
1840 and a dictionary containing the attributes carried by this element.</p>
1842 <p>Also note that the reference string generated from the callback shows a
1843 single character call even though the string "bar" is passed to the parser
1844 from 2 different call to parseChunk()</p>
1848 <p>This is a basic test of XPath wrappers support</p>
1851 doc = libxml2.parseFile("tst.xml")
1852 ctxt = doc.xpathNewContext()
1853 res = ctxt.xpathEval("//*")
1855 print "xpath query: wrong node set size"
1857 if res[0].name != "doc" or res[1].name != "foo":
1858 print "xpath query: wrong node set value"
1861 ctxt.xpathFreeContext()</pre>
1863 <p>This test parses a file, then create an XPath context to evaluate XPath
1864 expression on it. The xpathEval() method execute an XPath query and returns
1865 the result mapped in a Python way. String and numbers are natively converted,
1866 and node sets are returned as a tuple of libxml2 Python nodes wrappers. Like
1867 the document, the XPath context need to be freed explicitly, also not that
1868 the result of the XPath query may point back to the document tree and hence
1869 the document must be freed after the result of the query is used.</p>
1871 <h3>xpathext.py:</h3>
1873 <p>This test shows how to extend the XPath engine with functions written in
1880 doc = libxml2.parseFile("tst.xml")
1881 ctxt = doc.xpathNewContext()
1882 libxml2.registerXPathFunction(ctxt._o, "foo", None, foo)
1883 res = ctxt.xpathEval("foo(1)")
1885 print "xpath extension failure"
1887 ctxt.xpathFreeContext()</pre>
1889 <p>Note how the extension function is registered with the context (but that
1890 part is not yet finalized, this may change slightly in the future).</p>
1892 <h3>tstxpath.py:</h3>
1894 <p>This test is similar to the previous one but shows how the extension
1895 function can access the XPath evaluation context:</p>
1896 <pre>def foo(ctx, x):
1900 # test that access to the XPath evaluation contexts
1902 pctxt = libxml2.xpathParserContext(_obj=ctx)
1903 ctxt = pctxt.context()
1904 called = ctxt.function()
1907 <p>All the interfaces around the XPath parser(or rather evaluation) context
1908 are not finalized, but it should be sufficient to do contextual work at the
1909 evaluation point.</p>
1911 <h3>Memory debugging:</h3>
1913 <p>last but not least, all tests starts with the following prologue:</p>
1914 <pre>#memory debug specific
1915 libxml2.debugMemory(1)</pre>
1917 <p>and ends with the following epilogue:</p>
1918 <pre>#memory debug specific
1919 libxml2.cleanupParser()
1920 if libxml2.debugMemory(1) == 0:
1923 print "Memory leak %d bytes" % (libxml2.debugMemory(1))
1924 libxml2.dumpMemory()</pre>
1926 <p>Those activate the memory debugging interface of libxml2 where all
1927 allocated block in the library are tracked. The prologue then cleans up the
1928 library state and checks that all allocated memory has been freed. If not it
1929 calls dumpMemory() which saves that list in a <code>.memdump</code> file.</p>
1931 <h2><a name="architecture">libxml architecture</a></h2>
1933 <p>Libxml is made of multiple components; some of them are optional, and most
1934 of the block interfaces are public. The main components are:</p>
1936 <li>an Input/Output layer</li>
1937 <li>FTP and HTTP client layers (optional)</li>
1938 <li>an Internationalization layer managing the encodings support</li>
1939 <li>a URI module</li>
1940 <li>the XML parser and its basic SAX interface</li>
1941 <li>an HTML parser using the same SAX interface (optional)</li>
1942 <li>a SAX tree module to build an in-memory DOM representation</li>
1943 <li>a tree module to manipulate the DOM representation</li>
1944 <li>a validation module using the DOM representation (optional)</li>
1945 <li>an XPath module for global lookup in a DOM representation
1947 <li>a debug module (optional)</li>
1950 <p>Graphically this gives the following:</p>
1952 <p><img src="libxml.gif" alt="a graphical view of the various"></p>
1956 <h2><a name="tree">The tree output</a></h2>
1958 <p>The parser returns a tree built during the document analysis. The value
1959 returned is an <strong>xmlDocPtr</strong> (i.e., a pointer to an
1960 <strong>xmlDoc</strong> structure). This structure contains information such
1961 as the file name, the document type, and a <strong>children</strong> pointer
1962 which is the root of the document (or more exactly the first child under the
1963 root which is the document). The tree is made of <strong>xmlNode</strong>s,
1964 chained in double-linked lists of siblings and with a children<->parent
1965 relationship. An xmlNode can also carry properties (a chain of xmlAttr
1966 structures). An attribute may have a value which is a list of TEXT or
1967 ENTITY_REF nodes.</p>
1969 <p>Here is an example (erroneous with respect to the XML spec since there
1970 should be only one ELEMENT under the root):</p>
1972 <p><img src="structure.gif" alt=" structure.gif "></p>
1974 <p>In the source package there is a small program (not installed by default)
1975 called <strong>xmllint</strong> which parses XML files given as argument and
1976 prints them back as parsed. This is useful for detecting errors both in XML
1977 code and in the XML parser itself. It has an option <strong>--debug</strong>
1978 which prints the actual in-memory structure of the document; here is the
1979 result with the <a href="#example">example</a> given before:</p>
1986 content=gnome is great
1994 content=Welcome to Gnome
1998 content=The Linux adventure
2001 content=bla bla bla ...
2010 <p>This should be useful for learning the internal representation model.</p>
2012 <h2><a name="interface">The SAX interface</a></h2>
2014 <p>Sometimes the DOM tree output is just too large to fit reasonably into
2015 memory. In that case (and if you don't expect to save back the XML document
2016 loaded using libxml), it's better to use the SAX interface of libxml. SAX is
2017 a <strong>callback-based interface</strong> to the parser. Before parsing,
2018 the application layer registers a customized set of callbacks which are
2019 called by the library as it progresses through the XML input.</p>
2021 <p>To get more detailed step-by-step guidance on using the SAX interface of
2023 href="http://www.daa.com.au/~james/gnome/xml-sax/xml-sax.html">nice
2024 documentation</a>.written by <a href="mailto:james@daa.com.au">James
2027 <p>You can debug the SAX behaviour by using the <strong>testSAX</strong>
2028 program located in the gnome-xml module (it's usually not shipped in the
2029 binary packages of libxml, but you can find it in the tar source
2030 distribution). Here is the sequence of callbacks that would be reported by
2031 testSAX when parsing the example XML document shown earlier:</p>
2032 <pre>SAX.setDocumentLocator()
2035 SAX.startElement(EXAMPLE, prop1='gnome is great', prop2='&amp; linux too')
2036 SAX.characters( , 3)
2037 SAX.startElement(head)
2038 SAX.characters( , 4)
2039 SAX.startElement(title)
2040 SAX.characters(Welcome to Gnome, 16)
2041 SAX.endElement(title)
2042 SAX.characters( , 3)
2043 SAX.endElement(head)
2044 SAX.characters( , 3)
2045 SAX.startElement(chapter)
2046 SAX.characters( , 4)
2047 SAX.startElement(title)
2048 SAX.characters(The Linux adventure, 19)
2049 SAX.endElement(title)
2050 SAX.characters( , 4)
2052 SAX.characters(bla bla bla ..., 15)
2054 SAX.characters( , 4)
2055 SAX.startElement(image, href='linus.gif')
2056 SAX.endElement(image)
2057 SAX.characters( , 4)
2059 SAX.characters(..., 3)
2061 SAX.characters( , 3)
2062 SAX.endElement(chapter)
2063 SAX.characters( , 1)
2064 SAX.endElement(EXAMPLE)
2065 SAX.endDocument()</pre>
2067 <p>Most of the other interfaces of libxml are based on the DOM tree-building
2068 facility, so nearly everything up to the end of this document presupposes the
2069 use of the standard DOM tree build. Note that the DOM tree itself is built by
2070 a set of registered default callbacks, without internal specific
2073 <h2><a name="Validation">Validation & DTDs</a></h2>
2075 <p>Table of Content:</p>
2077 <li><a href="#General5">General overview</a></li>
2078 <li><a href="#definition">The definition</a></li>
2079 <li><a href="#Simple">Simple rules</a>
2081 <li><a href="#reference">How to reference a DTD from a document</a></li>
2082 <li><a href="#Declaring">Declaring elements</a></li>
2083 <li><a href="#Declaring1">Declaring attributes</a></li>
2086 <li><a href="#Some">Some examples</a></li>
2087 <li><a href="#validate">How to validate</a></li>
2088 <li><a href="#Other">Other resources</a></li>
2091 <h3><a name="General5">General overview</a></h3>
2093 <p>Well what is validation and what is a DTD ?</p>
2095 <p>DTD is the acronym for Document Type Definition. This is a description of
2096 the content for a family of XML files. This is part of the XML 1.0
2097 specification, and allows one to describe and verify that a given document
2098 instance conforms to the set of rules detailing its structure and content.</p>
2100 <p>Validation is the process of checking a document against a DTD (more
2101 generally against a set of construction rules).</p>
2103 <p>The validation process and building DTDs are the two most difficult parts
2104 of the XML life cycle. Briefly a DTD defines all the possible elements to be
2105 found within your document, what is the formal shape of your document tree
2106 (by defining the allowed content of an element; either text, a regular
2107 expression for the allowed list of children, or mixed content i.e. both text
2108 and children). The DTD also defines the valid attributes for all elements and
2109 the types of those attributes.</p>
2111 <h3><a name="definition1">The definition</a></h3>
2113 <p>The <a href="http://www.w3.org/TR/REC-xml">W3C XML Recommendation</a> (<a
2114 href="http://www.xml.com/axml/axml.html">Tim Bray's annotated version of
2117 <li><a href="http://www.w3.org/TR/REC-xml#elemdecls">Declaring
2119 <li><a href="http://www.w3.org/TR/REC-xml#attdecls">Declaring
2123 <p>(unfortunately) all this is inherited from the SGML world, the syntax is
2126 <h3><a name="Simple1">Simple rules</a></h3>
2128 <p>Writing DTDs can be done in many ways. The rules to build them if you need
2129 something permanent or something which can evolve over time can be radically
2130 different. Really complex DTDs like DocBook ones are flexible but quite
2131 harder to design. I will just focus on DTDs for a formats with a fixed simple
2132 structure. It is just a set of basic rules, and definitely not exhaustive nor
2133 usable for complex DTD design.</p>
2135 <h4><a name="reference1">How to reference a DTD from a document</a>:</h4>
2137 <p>Assuming the top element of the document is <code>spec</code> and the dtd
2138 is placed in the file <code>mydtd</code> in the subdirectory
2139 <code>dtds</code> of the directory from where the document were loaded:</p>
2141 <p><code><!DOCTYPE spec SYSTEM "dtds/mydtd"></code></p>
2145 <li>The system string is actually an URI-Reference (as defined in <a
2146 href="http://www.ietf.org/rfc/rfc2396.txt">RFC 2396</a>) so you can use a
2147 full URL string indicating the location of your DTD on the Web. This is a
2148 really good thing to do if you want others to validate your document.</li>
2149 <li>It is also possible to associate a <code>PUBLIC</code> identifier (a
2150 magic string) so that the DTD is looked up in catalogs on the client side
2151 without having to locate it on the web.</li>
2152 <li>A DTD contains a set of element and attribute declarations, but they
2153 don't define what the root of the document should be. This is explicitly
2154 told to the parser/validator as the first element of the
2155 <code>DOCTYPE</code> declaration.</li>
2158 <h4><a name="Declaring2">Declaring elements</a>:</h4>
2160 <p>The following declares an element <code>spec</code>:</p>
2162 <p><code><!ELEMENT spec (front, body, back?)></code></p>
2164 <p>It also expresses that the spec element contains one <code>front</code>,
2165 one <code>body</code> and one optional <code>back</code> children elements in
2166 this order. The declaration of one element of the structure and its content
2167 are done in a single declaration. Similarly the following declares
2168 <code>div1</code> elements:</p>
2170 <p><code><!ELEMENT div1 (head, (p | list | note)*, div2?)></code></p>
2172 <p>which means div1 contains one <code>head</code> then a series of optional
2173 <code>p</code>, <code>list</code>s and <code>note</code>s and then an
2174 optional <code>div2</code>. And last but not least an element can contain
2177 <p><code><!ELEMENT b (#PCDATA)></code></p>
2179 <p><code>b</code> contains text or being of mixed content (text and elements
2180 in no particular order):</p>
2182 <p><code><!ELEMENT p (#PCDATA|a|ul|b|i|em)*></code></p>
2184 <p><code>p </code>can contain text or <code>a</code>, <code>ul</code>,
2185 <code>b</code>, <code>i </code>or <code>em</code> elements in no particular
2188 <h4><a name="Declaring1">Declaring attributes</a>:</h4>
2190 <p>Again the attributes declaration includes their content definition:</p>
2192 <p><code><!ATTLIST termdef name CDATA #IMPLIED></code></p>
2194 <p>means that the element <code>termdef</code> can have a <code>name</code>
2195 attribute containing text (<code>CDATA</code>) and which is optional
2196 (<code>#IMPLIED</code>). The attribute value can also be defined within a
2199 <p><code><!ATTLIST list type (bullets|ordered|glossary)
2200 "ordered"></code></p>
2202 <p>means <code>list</code> element have a <code>type</code> attribute with 3
2203 allowed values "bullets", "ordered" or "glossary" and which default to
2204 "ordered" if the attribute is not explicitly specified.</p>
2206 <p>The content type of an attribute can be text (<code>CDATA</code>),
2207 anchor/reference/references
2208 (<code>ID</code>/<code>IDREF</code>/<code>IDREFS</code>), entity(ies)
2209 (<code>ENTITY</code>/<code>ENTITIES</code>) or name(s)
2210 (<code>NMTOKEN</code>/<code>NMTOKENS</code>). The following defines that a
2211 <code>chapter</code> element can have an optional <code>id</code> attribute
2212 of type <code>ID</code>, usable for reference from attribute of type
2215 <p><code><!ATTLIST chapter id ID #IMPLIED></code></p>
2217 <p>The last value of an attribute definition can be <code>#REQUIRED
2218 </code>meaning that the attribute has to be given, <code>#IMPLIED</code>
2219 meaning that it is optional, or the default value (possibly prefixed by
2220 <code>#FIXED</code> if it is the only allowed).</p>
2224 <li>Usually the attributes pertaining to a given element are declared in a
2225 single expression, but it is just a convention adopted by a lot of DTD
2227 <pre><!ATTLIST termdef
2229 name CDATA #IMPLIED></pre>
2230 <p>The previous construct defines both <code>id</code> and
2231 <code>name</code> attributes for the element <code>termdef</code>.</p>
2235 <h3><a name="Some1">Some examples</a></h3>
2237 <p>The directory <code>test/valid/dtds/</code> in the libxml distribution
2238 contains some complex DTD examples. The example in the file
2239 <code>test/valid/dia.xml</code> shows an XML file where the simple DTD is
2240 directly included within the document.</p>
2242 <h3><a name="validate1">How to validate</a></h3>
2244 <p>The simplest way is to use the xmllint program included with libxml. The
2245 <code>--valid</code> option turns-on validation of the files given as input.
2246 For example the following validates a copy of the first revision of the XML
2247 1.0 specification:</p>
2249 <p><code>xmllint --valid --noout test/valid/REC-xml-19980210.xml</code></p>
2251 <p>the -- noout is used to disable output of the resulting tree.</p>
2253 <p>The <code>--dtdvalid dtd</code> allows validation of the document(s)
2254 against a given DTD.</p>
2256 <p>Libxml exports an API to handle DTDs and validation, check the <a
2257 href="http://xmlsoft.org/html/libxml-valid.html">associated
2258 description</a>.</p>
2260 <h3><a name="Other1">Other resources</a></h3>
2262 <p>DTDs are as old as SGML. So there may be a number of examples on-line, I
2263 will just list one for now, others pointers welcome:</p>
2265 <li><a href="http://www.xml101.com:8081/dtd/">XML-101 DTD</a></li>
2268 <p>I suggest looking at the examples found under test/valid/dtd and any of
2269 the large number of books available on XML. The dia example in test/valid
2270 should be both simple and complete enough to allow you to build your own.</p>
2274 <h2><a name="Memory">Memory Management</a></h2>
2276 <p>Table of Content:</p>
2278 <li><a href="#General3">General overview</a></li>
2279 <li><a href="#setting">Setting libxml set of memory routines</a></li>
2280 <li><a href="#cleanup">Cleaning up after parsing</a></li>
2281 <li><a href="#Debugging">Debugging routines</a></li>
2282 <li><a href="#General4">General memory requirements</a></li>
2285 <h3><a name="General3">General overview</a></h3>
2287 <p>The module <code><a
2288 href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlmemory.h</a></code>
2289 provides the interfaces to the libxml memory system:</p>
2291 <li>libxml does not use the libc memory allocator directly but xmlFree(),
2292 xmlMalloc() and xmlRealloc()</li>
2293 <li>those routines can be reallocated to a specific set of routine, by
2294 default the libc ones i.e. free(), malloc() and realloc()</li>
2295 <li>the xmlmemory.c module includes a set of debugging routine</li>
2298 <h3><a name="setting">Setting libxml set of memory routines</a></h3>
2300 <p>It is sometimes useful to not use the default memory allocator, either for
2301 debugging, analysis or to implement a specific behaviour on memory management
2302 (like on embedded systems). Two function calls are available to do so:</p>
2304 <li><a href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlMemGet
2305 ()</a> which return the current set of functions in use by the parser</li>
2307 href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlMemSetup()</a>
2308 which allow to set up a new set of memory allocation functions</li>
2311 <p>Of course a call to xmlMemSetup() should probably be done before calling
2312 any other libxml routines (unless you are sure your allocations routines are
2315 <h3><a name="cleanup">Cleaning up after parsing</a></h3>
2317 <p>Libxml is not stateless, there is a few set of memory structures needing
2318 allocation before the parser is fully functional (some encoding structures
2319 for example). This also mean that once parsing is finished there is a tiny
2320 amount of memory (a few hundred bytes) which can be recollected if you don't
2321 reuse the parser immediately:</p>
2323 <li><a href="http://xmlsoft.org/html/libxml-parser.html">xmlCleanupParser
2324 ()</a> is a centralized routine to free the parsing states. Note that it
2325 won't deallocate any produced tree if any (use the xmlFreeDoc() and
2326 related routines for this).</li>
2327 <li><a href="http://xmlsoft.org/html/libxml-parser.html">xmlInitParser
2328 ()</a> is the dual routine allowing to preallocate the parsing state
2329 which can be useful for example to avoid initialization reentrancy
2330 problems when using libxml in multithreaded applications</li>
2333 <p>Generally xmlCleanupParser() is safe, if needed the state will be rebuild
2334 at the next invocation of parser routines, but be careful of the consequences
2335 in multithreaded applications.</p>
2337 <h3><a name="Debugging">Debugging routines</a></h3>
2339 <p>When configured using --with-mem-debug flag (off by default), libxml uses
2340 a set of memory allocation debugging routines keeping track of all allocated
2341 blocks and the location in the code where the routine was called. A couple of
2342 other debugging routines allow to dump the memory allocated infos to a file
2343 or call a specific routine when a given block number is allocated:</p>
2346 href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlMallocLoc()</a>
2348 href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlReallocLoc()</a>
2350 href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlMemStrdupLoc()</a>
2351 are the memory debugging replacement allocation routines</li>
2352 <li><a href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlMemoryDump
2353 ()</a> dumps all the informations about the allocated memory block lefts
2354 in the <code>.memdump</code> file</li>
2357 <p>When developing libxml memory debug is enabled, the tests programs call
2358 xmlMemoryDump () and the "make test" regression tests will check for any
2359 memory leak during the full regression test sequence, this helps a lot
2360 ensuring that libxml does not leak memory and bullet proof memory
2361 allocations use (some libc implementations are known to be far too permissive
2362 resulting in major portability problems!).</p>
2364 <p>If the .memdump reports a leak, it displays the allocation function and
2365 also tries to give some informations about the content and structure of the
2366 allocated blocks left. This is sufficient in most cases to find the culprit,
2367 but not always. Assuming the allocation problem is reproducible, it is
2368 possible to find more easily:</p>
2370 <li>write down the block number xxxx not allocated</li>
2371 <li>export the environment variable XML_MEM_BREAKPOINT=xxxx , the easiest
2372 when using GDB is to simply give the command
2373 <p><code>set environment XML_MEM_BREAKPOINT xxxx</code></p>
2374 <p>before running the program.</p>
2376 <li>run the program under a debugger and set a breakpoint on
2377 xmlMallocBreakpoint() a specific function called when this precise block
2379 <li>when the breakpoint is reached you can then do a fine analysis of the
2380 allocation an step to see the condition resulting in the missing
2384 <p>I used to use a commercial tool to debug libxml memory problems but after
2385 noticing that it was not detecting memory leaks that simple mechanism was
2386 used and proved extremely efficient until now. Lately I have also used <a
2387 href="http://developer.kde.org/~sewardj/">valgrind</a> with quite some
2388 success, it is tied to the i386 architecture since it works by emulating the
2389 processor and instruction set, it is slow but extremely efficient, i.e. it
2390 spot memory usage errors in a very precise way.</p>
2392 <h3><a name="General4">General memory requirements</a></h3>
2394 <p>How much libxml memory require ? It's hard to tell in average it depends
2395 of a number of things:</p>
2397 <li>the parser itself should work in a fixed amount of memory, except for
2398 information maintained about the stacks of names and entities locations.
2399 The I/O and encoding handlers will probably account for a few KBytes.
2400 This is true for both the XML and HTML parser (though the HTML parser
2401 need more state).</li>
2402 <li>If you are generating the DOM tree then memory requirements will grow
2403 nearly linear with the size of the data. In general for a balanced
2404 textual document the internal memory requirement is about 4 times the
2405 size of the UTF8 serialization of this document (example the XML-1.0
2406 recommendation is a bit more of 150KBytes and takes 650KBytes of main
2407 memory when parsed). Validation will add a amount of memory required for
2408 maintaining the external Dtd state which should be linear with the
2409 complexity of the content model defined by the Dtd</li>
2410 <li>If you don't care about the advanced features of libxml like
2411 validation, DOM, XPath or XPointer, but really need to work fixed memory
2412 requirements, then the SAX interface should be used.</li>
2417 <h2><a name="Encodings">Encodings support</a></h2>
2419 <p>Table of Content:</p>
2421 <li><a href="encoding.html#What">What does internationalization support
2423 <li><a href="encoding.html#internal">The internal encoding, how and
2425 <li><a href="encoding.html#implemente">How is it implemented ?</a></li>
2426 <li><a href="encoding.html#Default">Default supported encodings</a></li>
2427 <li><a href="encoding.html#extend">How to extend the existing
2431 <h3><a name="What">What does internationalization support mean ?</a></h3>
2433 <p>XML was designed from the start to allow the support of any character set
2434 by using Unicode. Any conformant XML parser has to support the UTF-8 and
2435 UTF-16 default encodings which can both express the full unicode ranges. UTF8
2436 is a variable length encoding whose greatest points are to reuse the same
2437 encoding for ASCII and to save space for Western encodings, but it is a bit
2438 more complex to handle in practice. UTF-16 use 2 bytes per characters (and
2439 sometimes combines two pairs), it makes implementation easier, but looks a
2440 bit overkill for Western languages encoding. Moreover the XML specification
2441 allows document to be encoded in other encodings at the condition that they
2442 are clearly labeled as such. For example the following is a wellformed XML
2443 document encoded in ISO-8859 1 and using accentuated letter that we French
2444 likes for both markup and content:</p>
2445 <pre><?xml version="1.0" encoding="ISO-8859-1"?>
2446 <très>là</très></pre>
2448 <p>Having internationalization support in libxml means the following:</p>
2450 <li>the document is properly parsed</li>
2451 <li>informations about it's encoding are saved</li>
2452 <li>it can be modified</li>
2453 <li>it can be saved in its original encoding</li>
2454 <li>it can also be saved in another encoding supported by libxml (for
2455 example straight UTF8 or even an ASCII form)</li>
2458 <p>Another very important point is that the whole libxml API, with the
2459 exception of a few routines to read with a specific encoding or save to a
2460 specific encoding, is completely agnostic about the original encoding of the
2463 <p>It should be noted too that the HTML parser embedded in libxml now obey
2464 the same rules too, the following document will be (as of 2.2.2) handled in
2465 an internationalized fashion by libxml too:</p>
2466 <pre><!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"
2467 "http://www.w3.org/TR/REC-html40/loose.dtd">
2468 <html lang="fr">
2470 <META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=ISO-8859-1">
2473 <p>W3C crée des standards pour le Web.</body>
2476 <h3><a name="internal">The internal encoding, how and why</a></h3>
2478 <p>One of the core decision was to force all documents to be converted to a
2479 default internal encoding, and that encoding to be UTF-8, here are the
2480 rationale for those choices:</p>
2482 <li>keeping the native encoding in the internal form would force the libxml
2483 users (or the code associated) to be fully aware of the encoding of the
2484 original document, for examples when adding a text node to a document,
2485 the content would have to be provided in the document encoding, i.e. the
2486 client code would have to check it before hand, make sure it's conformant
2487 to the encoding, etc ... Very hard in practice, though in some specific
2488 cases this may make sense.</li>
2489 <li>the second decision was which encoding. From the XML spec only UTF8 and
2490 UTF16 really makes sense as being the two only encodings for which there
2491 is mandatory support. UCS-4 (32 bits fixed size encoding) could be
2492 considered an intelligent choice too since it's a direct Unicode mapping
2493 support. I selected UTF-8 on the basis of efficiency and compatibility
2494 with surrounding software:
2496 <li>UTF-8 while a bit more complex to convert from/to (i.e. slightly
2497 more costly to import and export CPU wise) is also far more compact
2498 than UTF-16 (and UCS-4) for a majority of the documents I see it used
2499 for right now (RPM RDF catalogs, advogato data, various configuration
2500 file formats, etc.) and the key point for today's computer
2501 architecture is efficient uses of caches. If one nearly double the
2502 memory requirement to store the same amount of data, this will trash
2503 caches (main memory/external caches/internal caches) and my take is
2504 that this harms the system far more than the CPU requirements needed
2505 for the conversion to UTF-8</li>
2506 <li>Most of libxml version 1 users were using it with straight ASCII
2507 most of the time, doing the conversion with an internal encoding
2508 requiring all their code to be rewritten was a serious show-stopper
2509 for using UTF-16 or UCS-4.</li>
2510 <li>UTF-8 is being used as the de-facto internal encoding standard for
2511 related code like the <a href="http://www.pango.org/">pango</a>
2512 upcoming Gnome text widget, and a lot of Unix code (yep another place
2513 where Unix programmer base takes a different approach from Microsoft
2514 - they are using UTF-16)</li>
2519 <p>What does this mean in practice for the libxml user:</p>
2521 <li>xmlChar, the libxml data type is a byte, those bytes must be assembled
2522 as UTF-8 valid strings. The proper way to terminate an xmlChar * string
2523 is simply to append 0 byte, as usual.</li>
2524 <li>One just need to make sure that when using chars outside the ASCII set,
2525 the values has been properly converted to UTF-8</li>
2528 <h3><a name="implemente">How is it implemented ?</a></h3>
2530 <p>Let's describe how all this works within libxml, basically the I18N
2531 (internationalization) support get triggered only during I/O operation, i.e.
2532 when reading a document or saving one. Let's look first at the reading
2535 <li>when a document is processed, we usually don't know the encoding, a
2536 simple heuristic allows to detect UTF-18 and UCS-4 from whose where the
2537 ASCII range (0-0x7F) maps with ASCII</li>
2538 <li>the xml declaration if available is parsed, including the encoding
2539 declaration. At that point, if the autodetected encoding is different
2540 from the one declared a call to xmlSwitchEncoding() is issued.</li>
2541 <li>If there is no encoding declaration, then the input has to be in either
2542 UTF-8 or UTF-16, if it is not then at some point when processing the
2543 input, the converter/checker of UTF-8 form will raise an encoding error.
2544 You may end-up with a garbled document, or no document at all ! Example:
2545 <pre>~/XML -> ./xmllint err.xml
2546 err.xml:1: error: Input is not proper UTF-8, indicate encoding !
2547 <très>là</très>
2549 err.xml:1: error: Bytes: 0xE8 0x73 0x3E 0x6C
2550 <très>là</très>
2553 <li>xmlSwitchEncoding() does an encoding name lookup, canonicalize it, and
2554 then search the default registered encoding converters for that encoding.
2555 If it's not within the default set and iconv() support has been compiled
2556 it, it will ask iconv for such an encoder. If this fails then the parser
2557 will report an error and stops processing:
2558 <pre>~/XML -> ./xmllint err2.xml
2559 err2.xml:1: error: Unsupported encoding UnsupportedEnc
2560 <?xml version="1.0" encoding="UnsupportedEnc"?>
2563 <li>From that point the encoder processes progressively the input (it is
2564 plugged as a front-end to the I/O module) for that entity. It captures
2565 and convert on-the-fly the document to be parsed to UTF-8. The parser
2566 itself just does UTF-8 checking of this input and process it
2567 transparently. The only difference is that the encoding information has
2568 been added to the parsing context (more precisely to the input
2569 corresponding to this entity).</li>
2570 <li>The result (when using DOM) is an internal form completely in UTF-8
2571 with just an encoding information on the document node.</li>
2574 <p>Ok then what happens when saving the document (assuming you
2575 collected/built an xmlDoc DOM like structure) ? It depends on the function
2576 called, xmlSaveFile() will just try to save in the original encoding, while
2577 xmlSaveFileTo() and xmlSaveFileEnc() can optionally save to a given
2580 <li>if no encoding is given, libxml will look for an encoding value
2581 associated to the document and if it exists will try to save to that
2583 <p>otherwise everything is written in the internal form, i.e. UTF-8</p>
2585 <li>so if an encoding was specified, either at the API level or on the
2586 document, libxml will again canonicalize the encoding name, lookup for a
2587 converter in the registered set or through iconv. If not found the
2588 function will return an error code</li>
2589 <li>the converter is placed before the I/O buffer layer, as another kind of
2590 buffer, then libxml will simply push the UTF-8 serialization to through
2591 that buffer, which will then progressively be converted and pushed onto
2593 <li>It is possible that the converter code fails on some input, for example
2594 trying to push an UTF-8 encoded Chinese character through the UTF-8 to
2595 ISO-8859-1 converter won't work. Since the encoders are progressive they
2596 will just report the error and the number of bytes converted, at that
2597 point libxml will decode the offending character, remove it from the
2598 buffer and replace it with the associated charRef encoding &#123; and
2599 resume the conversion. This guarantees that any document will be saved
2600 without losses (except for markup names where this is not legal, this is
2601 a problem in the current version, in practice avoid using non-ascii
2602 characters for tags or attributes names @@). A special "ascii" encoding
2603 name is used to save documents to a pure ascii form can be used when
2604 portability is really crucial</li>
2607 <p>Here is a few examples based on the same test document:</p>
2608 <pre>~/XML -> ./xmllint isolat1
2609 <?xml version="1.0" encoding="ISO-8859-1"?>
2610 <très>là</très>
2611 ~/XML -> ./xmllint --encode UTF-8 isolat1
2612 <?xml version="1.0" encoding="UTF-8"?>
2613 <très>là </très>
2616 <p>The same processing is applied (and reuse most of the code) for HTML I18N
2617 processing. Looking up and modifying the content encoding is a bit more
2618 difficult since it is located in a <meta> tag under the <head>,
2619 so a couple of functions htmlGetMetaEncoding() and htmlSetMetaEncoding() have
2620 been provided. The parser also attempts to switch encoding on the fly when
2621 detecting such a tag on input. Except for that the processing is the same
2622 (and again reuses the same code).</p>
2624 <h3><a name="Default">Default supported encodings</a></h3>
2626 <p>libxml has a set of default converters for the following encodings
2627 (located in encoding.c):</p>
2629 <li>UTF-8 is supported by default (null handlers)</li>
2630 <li>UTF-16, both little and big endian</li>
2631 <li>ISO-Latin-1 (ISO-8859-1) covering most western languages</li>
2632 <li>ASCII, useful mostly for saving</li>
2633 <li>HTML, a specific handler for the conversion of UTF-8 to ASCII with HTML
2634 predefined entities like &copy; for the Copyright sign.</li>
2637 <p>More over when compiled on an Unix platform with iconv support the full
2638 set of encodings supported by iconv can be instantly be used by libxml. On a
2639 linux machine with glibc-2.1 the list of supported encodings and aliases fill
2640 3 full pages, and include UCS-4, the full set of ISO-Latin encodings, and the
2641 various Japanese ones.</p>
2643 <h4>Encoding aliases</h4>
2645 <p>From 2.2.3, libxml has support to register encoding names aliases. The
2646 goal is to be able to parse document whose encoding is supported but where
2647 the name differs (for example from the default set of names accepted by
2648 iconv). The following functions allow to register and handle new aliases for
2649 existing encodings. Once registered libxml will automatically lookup the
2650 aliases when handling a document:</p>
2652 <li>int xmlAddEncodingAlias(const char *name, const char *alias);</li>
2653 <li>int xmlDelEncodingAlias(const char *alias);</li>
2654 <li>const char * xmlGetEncodingAlias(const char *alias);</li>
2655 <li>void xmlCleanupEncodingAliases(void);</li>
2658 <h3><a name="extend">How to extend the existing support</a></h3>
2660 <p>Well adding support for new encoding, or overriding one of the encoders
2661 (assuming it is buggy) should not be hard, just write an input and output
2662 conversion routines to/from UTF-8, and register them using
2663 xmlNewCharEncodingHandler(name, xxxToUTF8, UTF8Toxxx), and they will be
2664 called automatically if the parser(s) encounter such an encoding name
2665 (register it uppercase, this will help). The description of the encoders,
2666 their arguments and expected return values are described in the encoding.h
2669 <p>A quick note on the topic of subverting the parser to use a different
2670 internal encoding than UTF-8, in some case people will absolutely want to
2671 keep the internal encoding different, I think it's still possible (but the
2672 encoding must be compliant with ASCII on the same subrange) though I didn't
2673 tried it. The key is to override the default conversion routines (by
2674 registering null encoders/decoders for your charsets), and bypass the UTF-8
2675 checking of the parser by setting the parser context charset
2676 (ctxt->charset) to something different than XML_CHAR_ENCODING_UTF8, but
2677 there is no guarantee that this will work. You may also have some troubles
2680 <p>Basically proper I18N support is important, this requires at least
2681 libxml-2.0.0, but a lot of features and corrections are really available only
2684 <h2><a name="IO">I/O Interfaces</a></h2>
2686 <p>Table of Content:</p>
2688 <li><a href="#General1">General overview</a></li>
2689 <li><a href="#basic">The basic buffer type</a></li>
2690 <li><a href="#Input">Input I/O handlers</a></li>
2691 <li><a href="#Output">Output I/O handlers</a></li>
2692 <li><a href="#entities">The entities loader</a></li>
2693 <li><a href="#Example2">Example of customized I/O</a></li>
2696 <h3><a name="General1">General overview</a></h3>
2698 <p>The module <code><a
2699 href="http://xmlsoft.org/html/libxml-xmlio.html">xmlIO.h</a></code> provides
2700 the interfaces to the libxml I/O system. This consists of 4 main parts:</p>
2702 <li>Entities loader, this is a routine which tries to fetch the entities
2703 (files) based on their PUBLIC and SYSTEM identifiers. The default loader
2704 don't look at the public identifier since libxml do not maintain a
2705 catalog. You can redefine you own entity loader by using
2706 <code>xmlGetExternalEntityLoader()</code> and
2707 <code>xmlSetExternalEntityLoader()</code>. <a href="#entities">Check the
2709 <li>Input I/O buffers which are a commodity structure used by the parser(s)
2710 input layer to handle fetching the informations to feed the parser. This
2711 provides buffering and is also a placeholder where the encoding
2712 converters to UTF8 are piggy-backed.</li>
2713 <li>Output I/O buffers are similar to the Input ones and fulfill similar
2714 task but when generating a serialization from a tree.</li>
2715 <li>A mechanism to register sets of I/O callbacks and associate them with
2716 specific naming schemes like the protocol part of the URIs.
2717 <p>This affect the default I/O operations and allows to use specific I/O
2718 handlers for certain names.</p>
2722 <p>The general mechanism used when loading http://rpmfind.net/xml.html for
2723 example in the HTML parser is the following:</p>
2725 <li>The default entity loader calls <code>xmlNewInputFromFile()</code> with
2726 the parsing context and the URI string.</li>
2727 <li>the URI string is checked against the existing registered handlers
2728 using their match() callback function, if the HTTP module was compiled
2729 in, it is registered and its match() function will succeeds</li>
2730 <li>the open() function of the handler is called and if successful will
2731 return an I/O Input buffer</li>
2732 <li>the parser will the start reading from this buffer and progressively
2733 fetch information from the resource, calling the read() function of the
2734 handler until the resource is exhausted</li>
2735 <li>if an encoding change is detected it will be installed on the input
2736 buffer, providing buffering and efficient use of the conversion
2738 <li>once the parser has finished, the close() function of the handler is
2739 called once and the Input buffer and associated resources are
2743 <p>The user defined callbacks are checked first to allow overriding of the
2744 default libxml I/O routines.</p>
2746 <h3><a name="basic">The basic buffer type</a></h3>
2748 <p>All the buffer manipulation handling is done using the
2749 <code>xmlBuffer</code> type define in <code><a
2750 href="http://xmlsoft.org/html/libxml-tree.html">tree.h</a> </code>which is a
2751 resizable memory buffer. The buffer allocation strategy can be selected to be
2752 either best-fit or use an exponential doubling one (CPU vs. memory use
2753 trade-off). The values are <code>XML_BUFFER_ALLOC_EXACT</code> and
2754 <code>XML_BUFFER_ALLOC_DOUBLEIT</code>, and can be set individually or on a
2755 system wide basis using <code>xmlBufferSetAllocationScheme()</code>. A number
2756 of functions allows to manipulate buffers with names starting with the
2757 <code>xmlBuffer...</code> prefix.</p>
2759 <h3><a name="Input">Input I/O handlers</a></h3>
2761 <p>An Input I/O handler is a simple structure
2762 <code>xmlParserInputBuffer</code> containing a context associated to the
2763 resource (file descriptor, or pointer to a protocol handler), the read() and
2764 close() callbacks to use and an xmlBuffer. And extra xmlBuffer and a charset
2765 encoding handler are also present to support charset conversion when
2768 <h3><a name="Output">Output I/O handlers</a></h3>
2770 <p>An Output handler <code>xmlOutputBuffer</code> is completely similar to an
2771 Input one except the callbacks are write() and close().</p>
2773 <h3><a name="entities">The entities loader</a></h3>
2775 <p>The entity loader resolves requests for new entities and create inputs for
2776 the parser. Creating an input from a filename or an URI string is done
2777 through the xmlNewInputFromFile() routine. The default entity loader do not
2778 handle the PUBLIC identifier associated with an entity (if any). So it just
2779 calls xmlNewInputFromFile() with the SYSTEM identifier (which is mandatory in
2782 <p>If you want to hook up a catalog mechanism then you simply need to
2783 override the default entity loader, here is an example:</p>
2784 <pre>#include <libxml/xmlIO.h>
2786 xmlExternalEntityLoader defaultLoader = NULL;
2789 xmlMyExternalEntityLoader(const char *URL, const char *ID,
2790 xmlParserCtxtPtr ctxt) {
2791 xmlParserInputPtr ret;
2792 const char *fileID = NULL;
2793 /* lookup for the fileID depending on ID */
2795 ret = xmlNewInputFromFile(ctxt, fileID);
2798 if (defaultLoader != NULL)
2799 ret = defaultLoader(URL, ID, ctxt);
2807 * Install our own entity loader
2809 defaultLoader = xmlGetExternalEntityLoader();
2810 xmlSetExternalEntityLoader(xmlMyExternalEntityLoader);
2815 <h3><a name="Example2">Example of customized I/O</a></h3>
2817 <p>This example come from <a href="http://xmlsoft.org/messages/0708.html">a
2818 real use case</a>, xmlDocDump() closes the FILE * passed by the application
2819 and this was a problem. The <a
2820 href="http://xmlsoft.org/messages/0711.html">solution</a> was to redefine a
2821 new output handler with the closing call deactivated:</p>
2823 <li>First define a new I/O output allocator where the output don't close
2825 <pre>xmlOutputBufferPtr
2826 xmlOutputBufferCreateOwn(FILE *file, xmlCharEncodingHandlerPtr encoder) {
2827 xmlOutputBufferPtr ret;
2829 if (xmlOutputCallbackInitialized == 0)
2830 xmlRegisterDefaultOutputCallbacks();
2832 if (file == NULL) return(NULL);
2833 ret = xmlAllocOutputBuffer(encoder);
2835 ret->context = file;
2836 ret->writecallback = xmlFileWrite;
2837 ret->closecallback = NULL; /* No close callback */
2877 <li>And then use it to save the document:
2879 xmlOutputBufferPtr output;
2886 output = xmlOutputBufferCreateOwn(f, NULL);
2887 res = xmlSaveFileTo(output, doc, NULL);
2892 <h2><a name="Catalog">Catalog support</a></h2>
2894 <p>Table of Content:</p>
2896 <li><a href="General2">General overview</a></li>
2897 <li><a href="#definition">The definition</a></li>
2898 <li><a href="#Simple">Using catalogs</a></li>
2899 <li><a href="#Some">Some examples</a></li>
2900 <li><a href="#reference">How to tune catalog usage</a></li>
2901 <li><a href="#validate">How to debug catalog processing</a></li>
2902 <li><a href="#Declaring">How to create and maintain catalogs</a></li>
2903 <li><a href="#implemento">The implementor corner quick review of the
2905 <li><a href="#Other">Other resources</a></li>
2908 <h3><a name="General2">General overview</a></h3>
2910 <p>What is a catalog? Basically it's a lookup mechanism used when an entity
2911 (a file or a remote resource) references another entity. The catalog lookup
2912 is inserted between the moment the reference is recognized by the software
2913 (XML parser, stylesheet processing, or even images referenced for inclusion
2914 in a rendering) and the time where loading that resource is actually
2917 <p>It is basically used for 3 things:</p>
2919 <li>mapping from "logical" names, the public identifiers and a more
2920 concrete name usable for download (and URI). For example it can associate
2922 <p>"-//OASIS//DTD DocBook XML V4.1.2//EN"</p>
2923 <p>of the DocBook 4.1.2 XML DTD with the actual URL where it can be
2925 <p>http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd</p>
2927 <li>remapping from a given URL to another one, like an HTTP indirection
2929 <p>"http://www.oasis-open.org/committes/tr.xsl"</p>
2930 <p>should really be looked at</p>
2931 <p>"http://www.oasis-open.org/committes/entity/stylesheets/base/tr.xsl"</p>
2933 <li>providing a local cache mechanism allowing to load the entities
2934 associated to public identifiers or remote resources, this is a really
2935 important feature for any significant deployment of XML or SGML since it
2936 allows to avoid the aleas and delays associated to fetching remote
2940 <h3><a name="definition">The definitions</a></h3>
2942 <p>Libxml, as of 2.4.3 implements 2 kind of catalogs:</p>
2944 <li>the older SGML catalogs, the official spec is SGML Open Technical
2945 Resolution TR9401:1997, but is better understood by reading <a
2946 href="http://www.jclark.com/sp/catalog.htm">the SP Catalog page</a> from
2947 James Clark. This is relatively old and not the preferred mode of
2948 operation of libxml.</li>
2949 <li><a href="http://www.oasis-open.org/committees/entity/spec.html">XML
2950 Catalogs</a> is far more flexible, more recent, uses an XML syntax and
2951 should scale quite better. This is the default option of libxml.</li>
2956 <h3><a name="Simple">Using catalog</a></h3>
2958 <p>In a normal environment libxml will by default check the presence of a
2959 catalog in /etc/xml/catalog, and assuming it has been correctly populated,
2960 the processing is completely transparent to the document user. To take a
2961 concrete example, suppose you are authoring a DocBook document, this one
2962 starts with the following DOCTYPE definition:</p>
2963 <pre><?xml version='1.0'?>
2964 <!DOCTYPE book PUBLIC "-//Norman Walsh//DTD DocBk XML V3.1.4//EN"
2965 "http://nwalsh.com/docbook/xml/3.1.4/db3xml.dtd"></pre>
2967 <p>When validating the document with libxml, the catalog will be
2968 automatically consulted to lookup the public identifier "-//Norman Walsh//DTD
2969 DocBk XML V3.1.4//EN" and the system identifier
2970 "http://nwalsh.com/docbook/xml/3.1.4/db3xml.dtd", and if these entities have
2971 been installed on your system and the catalogs actually point to them, libxml
2972 will fetch them from the local disk.</p>
2974 <p style="font-size: 10pt"><strong>Note</strong>: Really don't use this
2975 DOCTYPE example it's a really old version, but is fine as an example.</p>
2977 <p>Libxml will check the catalog each time that it is requested to load an
2978 entity, this includes DTD, external parsed entities, stylesheets, etc ... If
2979 your system is correctly configured all the authoring phase and processing
2980 should use only local files, even if your document stays portable because it
2981 uses the canonical public and system ID, referencing the remote document.</p>
2983 <h3><a name="Some">Some examples:</a></h3>
2985 <p>Here is a couple of fragments from XML Catalogs used in libxml early
2986 regression tests in <code>test/catalogs</code> :</p>
2987 <pre><?xml version="1.0"?>
2988 <!DOCTYPE catalog PUBLIC
2989 "-//OASIS//DTD Entity Resolution XML Catalog V1.0//EN"
2990 "http://www.oasis-open.org/committees/entity/release/1.0/catalog.dtd">
2991 <catalog xmlns="urn:oasis:names:tc:entity:xmlns:xml:catalog">
2992 <public publicId="-//OASIS//DTD DocBook XML V4.1.2//EN"
2993 uri="http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"/>
2996 <p>This is the beginning of a catalog for DocBook 4.1.2, XML Catalogs are
2997 written in XML, there is a specific namespace for catalog elements
2998 "urn:oasis:names:tc:entity:xmlns:xml:catalog". The first entry in this
2999 catalog is a <code>public</code> mapping it allows to associate a Public
3000 Identifier with an URI.</p>
3002 <rewriteSystem systemIdStartString="http://www.oasis-open.org/docbook/"
3003 rewritePrefix="file:///usr/share/xml/docbook/"/>
3006 <p>A <code>rewriteSystem</code> is a very powerful instruction, it says that
3007 any URI starting with a given prefix should be looked at another URI
3008 constructed by replacing the prefix with an new one. In effect this acts like
3009 a cache system for a full area of the Web. In practice it is extremely useful
3010 with a file prefix if you have installed a copy of those resources on your
3013 <delegatePublic publicIdStartString="-//OASIS//DTD XML Catalog //"
3014 catalog="file:///usr/share/xml/docbook.xml"/>
3015 <delegatePublic publicIdStartString="-//OASIS//ENTITIES DocBook XML"
3016 catalog="file:///usr/share/xml/docbook.xml"/>
3017 <delegatePublic publicIdStartString="-//OASIS//DTD DocBook XML"
3018 catalog="file:///usr/share/xml/docbook.xml"/>
3019 <delegateSystem systemIdStartString="http://www.oasis-open.org/docbook/"
3020 catalog="file:///usr/share/xml/docbook.xml"/>
3021 <delegateURI uriStartString="http://www.oasis-open.org/docbook/"
3022 catalog="file:///usr/share/xml/docbook.xml"/>
3025 <p>Delegation is the core features which allows to build a tree of catalogs,
3026 easier to maintain than a single catalog, based on Public Identifier, System
3027 Identifier or URI prefixes it instructs the catalog software to look up
3028 entries in another resource. This feature allow to build hierarchies of
3029 catalogs, the set of entries presented should be sufficient to redirect the
3030 resolution of all DocBook references to the specific catalog in
3031 <code>/usr/share/xml/docbook.xml</code> this one in turn could delegate all
3032 references for DocBook 4.2.1 to a specific catalog installed at the same time
3033 as the DocBook resources on the local machine.</p>
3035 <h3><a name="reference">How to tune catalog usage:</a></h3>
3037 <p>The user can change the default catalog behaviour by redirecting queries
3038 to its own set of catalogs, this can be done by setting the
3039 <code>XML_CATALOG_FILES</code> environment variable to a list of catalogs, an
3040 empty one should deactivate loading the default <code>/etc/xml/catalog</code>
3043 <h3><a name="validate">How to debug catalog processing:</a></h3>
3045 <p>Setting up the <code>XML_DEBUG_CATALOG</code> environment variable will
3046 make libxml output debugging informations for each catalog operations, for
3048 <pre>orchis:~/XML -> xmllint --memory --noout test/ent2
3049 warning: failed to load external entity "title.xml"
3050 orchis:~/XML -> export XML_DEBUG_CATALOG=
3051 orchis:~/XML -> xmllint --memory --noout test/ent2
3052 Failed to parse catalog /etc/xml/catalog
3053 Failed to parse catalog /etc/xml/catalog
3054 warning: failed to load external entity "title.xml"
3056 orchis:~/XML -> </pre>
3058 <p>The test/ent2 references an entity, running the parser from memory makes
3059 the base URI unavailable and the the "title.xml" entity cannot be loaded.
3060 Setting up the debug environment variable allows to detect that an attempt is
3061 made to load the <code>/etc/xml/catalog</code> but since it's not present the
3062 resolution fails.</p>
3064 <p>But the most advanced way to debug XML catalog processing is to use the
3065 <strong>xmlcatalog</strong> command shipped with libxml2, it allows to load
3066 catalogs and make resolution queries to see what is going on. This is also
3067 used for the regression tests:</p>
3068 <pre>orchis:~/XML -> ./xmlcatalog test/catalogs/docbook.xml \
3069 "-//OASIS//DTD DocBook XML V4.1.2//EN"
3070 http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd
3071 orchis:~/XML -> </pre>
3073 <p>For debugging what is going on, adding one -v flags increase the verbosity
3074 level to indicate the processing done (adding a second flag also indicate
3075 what elements are recognized at parsing):</p>
3076 <pre>orchis:~/XML -> ./xmlcatalog -v test/catalogs/docbook.xml \
3077 "-//OASIS//DTD DocBook XML V4.1.2//EN"
3078 Parsing catalog test/catalogs/docbook.xml's content
3079 Found public match -//OASIS//DTD DocBook XML V4.1.2//EN
3080 http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd
3082 orchis:~/XML -> </pre>
3084 <p>A shell interface is also available to debug and process multiple queries
3085 (and for regression tests):</p>
3086 <pre>orchis:~/XML -> ./xmlcatalog -shell test/catalogs/docbook.xml \
3087 "-//OASIS//DTD DocBook XML V4.1.2//EN"
3090 public PublicID: make a PUBLIC identifier lookup
3091 system SystemID: make a SYSTEM identifier lookup
3092 resolve PublicID SystemID: do a full resolver lookup
3093 add 'type' 'orig' 'replace' : add an entry
3094 del 'values' : remove values
3095 dump: print the current catalog state
3096 debug: increase the verbosity level
3097 quiet: decrease the verbosity level
3098 exit: quit the shell
3099 > public "-//OASIS//DTD DocBook XML V4.1.2//EN"
3100 http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd
3102 orchis:~/XML -> </pre>
3104 <p>This should be sufficient for most debugging purpose, this was actually
3105 used heavily to debug the XML Catalog implementation itself.</p>
3107 <h3><a name="Declaring">How to create and maintain</a> catalogs:</h3>
3109 <p>Basically XML Catalogs are XML files, you can either use XML tools to
3110 manage them or use <strong>xmlcatalog</strong> for this. The basic step is
3111 to create a catalog the -create option provide this facility:</p>
3112 <pre>orchis:~/XML -> ./xmlcatalog --create tst.xml
3113 <?xml version="1.0"?>
3114 <!DOCTYPE catalog PUBLIC "-//OASIS//DTD Entity Resolution XML Catalog V1.0//EN"
3115 "http://www.oasis-open.org/committees/entity/release/1.0/catalog.dtd">
3116 <catalog xmlns="urn:oasis:names:tc:entity:xmlns:xml:catalog"/>
3117 orchis:~/XML -> </pre>
3119 <p>By default xmlcatalog does not overwrite the original catalog and save the
3120 result on the standard output, this can be overridden using the -noout
3121 option. The <code>-add</code> command allows to add entries in the
3123 <pre>orchis:~/XML -> ./xmlcatalog --noout --create --add "public" \
3124 "-//OASIS//DTD DocBook XML V4.1.2//EN" \
3125 http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd tst.xml
3126 orchis:~/XML -> cat tst.xml
3127 <?xml version="1.0"?>
3128 <!DOCTYPE catalog PUBLIC "-//OASIS//DTD Entity Resolution XML Catalog V1.0//EN" \
3129 "http://www.oasis-open.org/committees/entity/release/1.0/catalog.dtd">
3130 <catalog xmlns="urn:oasis:names:tc:entity:xmlns:xml:catalog">
3131 <public publicId="-//OASIS//DTD DocBook XML V4.1.2//EN"
3132 uri="http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"/>
3134 orchis:~/XML -> </pre>
3136 <p>The <code>-add</code> option will always take 3 parameters even if some of
3137 the XML Catalog constructs (like nextCatalog) will have only a single
3138 argument, just pass a third empty string, it will be ignored.</p>
3140 <p>Similarly the <code>-del</code> option remove matching entries from the
3142 <pre>orchis:~/XML -> ./xmlcatalog --del \
3143 "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd" tst.xml
3144 <?xml version="1.0"?>
3145 <!DOCTYPE catalog PUBLIC "-//OASIS//DTD Entity Resolution XML Catalog V1.0//EN"
3146 "http://www.oasis-open.org/committees/entity/release/1.0/catalog.dtd">
3147 <catalog xmlns="urn:oasis:names:tc:entity:xmlns:xml:catalog"/>
3148 orchis:~/XML -> </pre>
3150 <p>The catalog is now empty. Note that the matching of <code>-del</code> is
3151 exact and would have worked in a similar fashion with the Public ID
3154 <p>This is rudimentary but should be sufficient to manage a not too complex
3155 catalog tree of resources.</p>
3157 <h3><a name="implemento">The implementor corner quick review of the
3160 <p>First, and like for every other module of libxml, there is an
3161 automatically generated <a href="html/libxml-catalog.html">API page for
3162 catalog support</a>.</p>
3164 <p>The header for the catalog interfaces should be included as:</p>
3165 <pre>#include <libxml/catalog.h></pre>
3167 <p>The API is voluntarily kept very simple. First it is not obvious that
3168 applications really need access to it since it is the default behaviour of
3169 libxml (Note: it is possible to completely override libxml default catalog by
3170 using <a href="html/libxml-parser.html">xmlSetExternalEntityLoader</a> to
3171 plug an application specific resolver).</p>
3173 <p>Basically libxml support 2 catalog lists:</p>
3175 <li>the default one, global shared by all the application</li>
3176 <li>a per-document catalog, this one is built if the document uses the
3177 <code>oasis-xml-catalog</code> PIs to specify its own catalog list, it is
3178 associated to the parser context and destroyed when the parsing context
3182 <p>the document one will be used first if it exists.</p>
3184 <h4>Initialization routines:</h4>
3186 <p>xmlInitializeCatalog(), xmlLoadCatalog() and xmlLoadCatalogs() should be
3187 used at startup to initialize the catalog, if the catalog should be
3188 initialized with specific values xmlLoadCatalog() or xmlLoadCatalogs()
3189 should be called before xmlInitializeCatalog() which would otherwise do a
3190 default initialization first.</p>
3192 <p>The xmlCatalogAddLocal() call is used by the parser to grow the document
3193 own catalog list if needed.</p>
3195 <h4>Preferences setup:</h4>
3197 <p>The XML Catalog spec requires the possibility to select default
3198 preferences between public and system delegation,
3199 xmlCatalogSetDefaultPrefer() allows this, xmlCatalogSetDefaults() and
3200 xmlCatalogGetDefaults() allow to control if XML Catalogs resolution should
3201 be forbidden, allowed for global catalog, for document catalog or both, the
3202 default is to allow both.</p>
3204 <p>And of course xmlCatalogSetDebug() allows to generate debug messages
3205 (through the xmlGenericError() mechanism).</p>
3207 <h4>Querying routines:</h4>
3209 <p>xmlCatalogResolve(), xmlCatalogResolveSystem(), xmlCatalogResolvePublic()
3210 and xmlCatalogResolveURI() are relatively explicit if you read the XML
3211 Catalog specification they correspond to section 7 algorithms, they should
3212 also work if you have loaded an SGML catalog with a simplified semantic.</p>
3214 <p>xmlCatalogLocalResolve() and xmlCatalogLocalResolveURI() are the same but
3215 operate on the document catalog list</p>
3217 <h4>Cleanup and Miscellaneous:</h4>
3219 <p>xmlCatalogCleanup() free-up the global catalog, xmlCatalogFreeLocal() is
3220 the per-document equivalent.</p>
3222 <p>xmlCatalogAdd() and xmlCatalogRemove() are used to dynamically modify the
3223 first catalog in the global list, and xmlCatalogDump() allows to dump a
3224 catalog state, those routines are primarily designed for xmlcatalog, I'm not
3225 sure that exposing more complex interfaces (like navigation ones) would be
3228 <p>The xmlParseCatalogFile() is a function used to load XML Catalog files,
3229 it's similar as xmlParseFile() except it bypass all catalog lookups, it's
3230 provided because this functionality may be useful for client tools.</p>
3232 <h4>threaded environments:</h4>
3234 <p>Since the catalog tree is built progressively, some care has been taken to
3235 try to avoid troubles in multithreaded environments. The code is now thread
3236 safe assuming that the libxml library has been compiled with threads
3241 <h3><a name="Other">Other resources</a></h3>
3243 <p>The XML Catalog specification is relatively recent so there isn't much
3244 literature to point at:</p>
3246 <li>You can find a good rant from Norm Walsh about <a
3247 href="http://www.arbortext.com/Think_Tank/XML_Resources/Issue_Three/issue_three.html">the
3248 need for catalogs</a>, it provides a lot of context informations even if
3249 I don't agree with everything presented. Norm also wrote a more recent
3251 href="http://wwws.sun.com/software/xml/developers/resolver/article/">XML
3252 entities and URI resolvers</a> describing them.</li>
3253 <li>An <a href="http://home.ccil.org/~cowan/XML/XCatalog.html">old XML
3254 catalog proposal</a> from John Cowan</li>
3255 <li>The <a href="http://www.rddl.org/">Resource Directory Description
3256 Language</a> (RDDL) another catalog system but more oriented toward
3257 providing metadata for XML namespaces.</li>
3258 <li>the page from the OASIS Technical <a
3259 href="http://www.oasis-open.org/committees/entity/">Committee on Entity
3260 Resolution</a> who maintains XML Catalog, you will find pointers to the
3261 specification update, some background and pointers to others tools
3262 providing XML Catalog support</li>
3263 <li>Here is a <a href="buildDocBookCatalog">shell script</a> to generate
3264 XML Catalogs for DocBook 4.1.2 . If it can write to the /etc/xml/
3265 directory, it will set-up /etc/xml/catalog and /etc/xml/docbook based on
3266 the resources found on the system. Otherwise it will just create
3267 ~/xmlcatalog and ~/dbkxmlcatalog and doing:
3268 <p><code>export XMLCATALOG=$HOME/xmlcatalog</code></p>
3269 <p>should allow to process DocBook documentations without requiring
3270 network accesses for the DTD or stylesheets</p>
3272 <li>I have uploaded <a href="ftp://xmlsoft.org/test/dbk412catalog.tar.gz">a
3273 small tarball</a> containing XML Catalogs for DocBook 4.1.2 which seems
3274 to work fine for me too</li>
3275 <li>The <a href="http://www.xmlsoft.org/xmlcatalog_man.html">xmlcatalog
3276 manual page</a></li>
3279 <p>If you have suggestions for corrections or additions, simply contact
3282 <h2><a name="library">The parser interfaces</a></h2>
3284 <p>This section is directly intended to help programmers getting bootstrapped
3285 using the XML library from the C language. It is not intended to be
3286 extensive. I hope the automatically generated documents will provide the
3287 completeness required, but as a separate set of documents. The interfaces of
3288 the XML library are by principle low level, there is nearly zero abstraction.
3289 Those interested in a higher level API should <a href="#DOM">look at
3292 <p>The <a href="html/libxml-parser.html">parser interfaces for XML</a> are
3293 separated from the <a href="html/libxml-htmlparser.html">HTML parser
3294 interfaces</a>. Let's have a look at how the XML parser can be called:</p>
3296 <h3><a name="Invoking">Invoking the parser : the pull method</a></h3>
3298 <p>Usually, the first thing to do is to read an XML input. The parser accepts
3299 documents either from in-memory strings or from files. The functions are
3300 defined in "parser.h":</p>
3302 <dt><code>xmlDocPtr xmlParseMemory(char *buffer, int size);</code></dt>
3303 <dd><p>Parse a null-terminated string containing the document.</p>
3307 <dt><code>xmlDocPtr xmlParseFile(const char *filename);</code></dt>
3308 <dd><p>Parse an XML document contained in a (possibly compressed)
3313 <p>The parser returns a pointer to the document structure (or NULL in case of
3316 <h3 id="Invoking1">Invoking the parser: the push method</h3>
3318 <p>In order for the application to keep the control when the document is
3319 being fetched (which is common for GUI based programs) libxml provides a push
3320 interface, too, as of version 1.8.3. Here are the interface functions:</p>
3321 <pre>xmlParserCtxtPtr xmlCreatePushParserCtxt(xmlSAXHandlerPtr sax,
3325 const char *filename);
3326 int xmlParseChunk (xmlParserCtxtPtr ctxt,
3329 int terminate);</pre>
3331 <p>and here is a simple example showing how to use the interface:</p>
3334 f = fopen(filename, "r");
3336 int res, size = 1024;
3338 xmlParserCtxtPtr ctxt;
3340 res = fread(chars, 1, 4, f);
3342 ctxt = xmlCreatePushParserCtxt(NULL, NULL,
3343 chars, res, filename);
3344 while ((res = fread(chars, 1, size, f)) > 0) {
3345 xmlParseChunk(ctxt, chars, res, 0);
3347 xmlParseChunk(ctxt, chars, 0, 1);
3348 doc = ctxt->myDoc;
3349 xmlFreeParserCtxt(ctxt);
3353 <p>The HTML parser embedded into libxml also has a push interface; the
3354 functions are just prefixed by "html" rather than "xml".</p>
3356 <h3 id="Invoking2">Invoking the parser: the SAX interface</h3>
3358 <p>The tree-building interface makes the parser memory-hungry, first loading
3359 the document in memory and then building the tree itself. Reading a document
3360 without building the tree is possible using the SAX interfaces (see SAX.h and
3361 <a href="http://www.daa.com.au/~james/gnome/xml-sax/xml-sax.html">James
3362 Henstridge's documentation</a>). Note also that the push interface can be
3363 limited to SAX: just use the two first arguments of
3364 <code>xmlCreatePushParserCtxt()</code>.</p>
3366 <h3><a name="Building">Building a tree from scratch</a></h3>
3368 <p>The other way to get an XML tree in memory is by building it. Basically
3369 there is a set of functions dedicated to building new elements. (These are
3370 also described in <libxml/tree.h>.) For example, here is a piece of
3371 code that produces the XML document used in the previous examples:</p>
3372 <pre> #include <libxml/tree.h>
3374 xmlNodePtr tree, subtree;
3376 doc = xmlNewDoc("1.0");
3377 doc->children = xmlNewDocNode(doc, NULL, "EXAMPLE", NULL);
3378 xmlSetProp(doc->children, "prop1", "gnome is great");
3379 xmlSetProp(doc->children, "prop2", "& linux too");
3380 tree = xmlNewChild(doc->children, NULL, "head", NULL);
3381 subtree = xmlNewChild(tree, NULL, "title", "Welcome to Gnome");
3382 tree = xmlNewChild(doc->children, NULL, "chapter", NULL);
3383 subtree = xmlNewChild(tree, NULL, "title", "The Linux adventure");
3384 subtree = xmlNewChild(tree, NULL, "p", "bla bla bla ...");
3385 subtree = xmlNewChild(tree, NULL, "image", NULL);
3386 xmlSetProp(subtree, "href", "linus.gif");</pre>
3388 <p>Not really rocket science ...</p>
3390 <h3><a name="Traversing">Traversing the tree</a></h3>
3392 <p>Basically by <a href="html/libxml-tree.html">including "tree.h"</a> your
3393 code has access to the internal structure of all the elements of the tree.
3394 The names should be somewhat simple like <strong>parent</strong>,
3395 <strong>children</strong>, <strong>next</strong>, <strong>prev</strong>,
3396 <strong>properties</strong>, etc... For example, still with the previous
3398 <pre><code>doc->children->children->children</code></pre>
3400 <p>points to the title element,</p>
3401 <pre>doc->children->children->next->children->children</pre>
3403 <p>points to the text node containing the chapter title "The Linux
3406 <p><strong>NOTE</strong>: XML allows <em>PI</em>s and <em>comments</em> to be
3407 present before the document root, so <code>doc->children</code> may point
3408 to an element which is not the document Root Element; a function
3409 <code>xmlDocGetRootElement()</code> was added for this purpose.</p>
3411 <h3><a name="Modifying">Modifying the tree</a></h3>
3413 <p>Functions are provided for reading and writing the document content. Here
3414 is an excerpt from the <a href="html/libxml-tree.html">tree API</a>:</p>
3416 <dt><code>xmlAttrPtr xmlSetProp(xmlNodePtr node, const xmlChar *name, const
3417 xmlChar *value);</code></dt>
3418 <dd><p>This sets (or changes) an attribute carried by an ELEMENT node.
3419 The value can be NULL.</p>
3423 <dt><code>const xmlChar *xmlGetProp(xmlNodePtr node, const xmlChar
3425 <dd><p>This function returns a pointer to new copy of the property
3426 content. Note that the user must deallocate the result.</p>
3430 <p>Two functions are provided for reading and writing the text associated
3433 <dt><code>xmlNodePtr xmlStringGetNodeList(xmlDocPtr doc, const xmlChar
3434 *value);</code></dt>
3435 <dd><p>This function takes an "external" string and converts it to one
3436 text node or possibly to a list of entity and text nodes. All
3437 non-predefined entity references like &Gnome; will be stored
3438 internally as entity nodes, hence the result of the function may not be
3443 <dt><code>xmlChar *xmlNodeListGetString(xmlDocPtr doc, xmlNodePtr list, int
3444 inLine);</code></dt>
3445 <dd><p>This function is the inverse of
3446 <code>xmlStringGetNodeList()</code>. It generates a new string
3447 containing the content of the text and entity nodes. Note the extra
3448 argument inLine. If this argument is set to 1, the function will expand
3449 entity references. For example, instead of returning the &Gnome;
3450 XML encoding in the string, it will substitute it with its value (say,
3451 "GNU Network Object Model Environment").</p>
3455 <h3><a name="Saving">Saving a tree</a></h3>
3457 <p>Basically 3 options are possible:</p>
3459 <dt><code>void xmlDocDumpMemory(xmlDocPtr cur, xmlChar**mem, int
3461 <dd><p>Returns a buffer into which the document has been saved.</p>
3465 <dt><code>extern void xmlDocDump(FILE *f, xmlDocPtr doc);</code></dt>
3466 <dd><p>Dumps a document to an open file descriptor.</p>
3470 <dt><code>int xmlSaveFile(const char *filename, xmlDocPtr cur);</code></dt>
3471 <dd><p>Saves the document to a file. In this case, the compression
3472 interface is triggered if it has been turned on.</p>
3476 <h3><a name="Compressio">Compression</a></h3>
3478 <p>The library transparently handles compression when doing file-based
3479 accesses. The level of compression on saves can be turned on either globally
3480 or individually for one file:</p>
3482 <dt><code>int xmlGetDocCompressMode (xmlDocPtr doc);</code></dt>
3483 <dd><p>Gets the document compression ratio (0-9).</p>
3487 <dt><code>void xmlSetDocCompressMode (xmlDocPtr doc, int mode);</code></dt>
3488 <dd><p>Sets the document compression ratio.</p>
3492 <dt><code>int xmlGetCompressMode(void);</code></dt>
3493 <dd><p>Gets the default compression ratio.</p>
3497 <dt><code>void xmlSetCompressMode(int mode);</code></dt>
3498 <dd><p>Sets the default compression ratio.</p>
3502 <h2><a name="Entities">Entities or no entities</a></h2>
3504 <p>Entities in principle are similar to simple C macros. An entity defines an
3505 abbreviation for a given string that you can reuse many times throughout the
3506 content of your document. Entities are especially useful when a given string
3507 may occur frequently within a document, or to confine the change needed to a
3508 document to a restricted area in the internal subset of the document (at the
3509 beginning). Example:</p>
3510 <pre>1 <?xml version="1.0"?>
3511 2 <!DOCTYPE EXAMPLE SYSTEM "example.dtd" [
3512 3 <!ENTITY xml "Extensible Markup Language">
3516 7 </EXAMPLE></pre>
3518 <p>Line 3 declares the xml entity. Line 6 uses the xml entity, by prefixing
3519 its name with '&' and following it by ';' without any spaces added. There
3520 are 5 predefined entities in libxml allowing you to escape characters with
3521 predefined meaning in some parts of the xml document content:
3522 <strong>&lt;</strong> for the character '<', <strong>&gt;</strong>
3523 for the character '>', <strong>&apos;</strong> for the character ''',
3524 <strong>&quot;</strong> for the character '"', and
3525 <strong>&amp;</strong> for the character '&'.</p>
3527 <p>One of the problems related to entities is that you may want the parser to
3528 substitute an entity's content so that you can see the replacement text in
3529 your application. Or you may prefer to keep entity references as such in the
3530 content to be able to save the document back without losing this usually
3531 precious information (if the user went through the pain of explicitly
3532 defining entities, he may have a a rather negative attitude if you blindly
3533 substitute them as saving time). The <a
3534 href="html/libxml-parser.html#XMLSUBSTITUTEENTITIESDEFAULT">xmlSubstituteEntitiesDefault()</a>
3535 function allows you to check and change the behaviour, which is to not
3536 substitute entities by default.</p>
3538 <p>Here is the DOM tree built by libxml for the previous document in the
3540 <pre>/gnome/src/gnome-xml -> ./xmllint --debug test/ent1
3547 INTERNAL_GENERAL_ENTITY xml
3548 content=Extensible Markup Language
3552 <p>And here is the result when substituting entities:</p>
3553 <pre>/gnome/src/gnome-xml -> ./tester --debug --noent test/ent1
3558 content= Extensible Markup Language</pre>
3560 <p>So, entities or no entities? Basically, it depends on your use case. I
3561 suggest that you keep the non-substituting default behaviour and avoid using
3562 entities in your XML document or data if you are not willing to handle the
3563 entity references elements in the DOM tree.</p>
3565 <p>Note that at save time libxml enforces the conversion of the predefined
3566 entities where necessary to prevent well-formedness problems, and will also
3567 transparently replace those with chars (i.e. it will not generate entity
3568 reference elements in the DOM tree or call the reference() SAX callback when
3569 finding them in the input).</p>
3571 <p><span style="background-color: #FF0000">WARNING</span>: handling entities
3572 on top of the libxml SAX interface is difficult!!! If you plan to use
3573 non-predefined entities in your documents, then the learning curve to handle
3574 then using the SAX API may be long. If you plan to use complex documents, I
3575 strongly suggest you consider using the DOM interface instead and let libxml
3576 deal with the complexity rather than trying to do it yourself.</p>
3578 <h2><a name="Namespaces">Namespaces</a></h2>
3580 <p>The libxml library implements <a
3581 href="http://www.w3.org/TR/REC-xml-names/">XML namespaces</a> support by
3582 recognizing namespace constructs in the input, and does namespace lookup
3583 automatically when building the DOM tree. A namespace declaration is
3584 associated with an in-memory structure and all elements or attributes within
3585 that namespace point to it. Hence testing the namespace is a simple and fast
3586 equality operation at the user level.</p>
3588 <p>I suggest that people using libxml use a namespace, and declare it in the
3589 root element of their document as the default namespace. Then they don't need
3590 to use the prefix in the content but we will have a basis for future semantic
3591 refinement and merging of data from different sources. This doesn't increase
3592 the size of the XML output significantly, but significantly increases its
3593 value in the long-term. Example:</p>
3594 <pre><mydoc xmlns="http://mydoc.example.org/schemas/">
3595 <elem1>...</elem1>
3596 <elem2>...</elem2>
3597 </mydoc></pre>
3599 <p>The namespace value has to be an absolute URL, but the URL doesn't have to
3600 point to any existing resource on the Web. It will bind all the element and
3601 attributes with that URL. I suggest to use an URL within a domain you
3602 control, and that the URL should contain some kind of version information if
3603 possible. For example, <code>"http://www.gnome.org/gnumeric/1.0/"</code> is a
3604 good namespace scheme.</p>
3606 <p>Then when you load a file, make sure that a namespace carrying the
3607 version-independent prefix is installed on the root element of your document,
3608 and if the version information don't match something you know, warn the user
3609 and be liberal in what you accept as the input. Also do *not* try to base
3610 namespace checking on the prefix value. <foo:text> may be exactly the
3611 same as <bar:text> in another document. What really matters is the URI
3612 associated with the element or the attribute, not the prefix string (which is
3613 just a shortcut for the full URI). In libxml, element and attributes have an
3614 <code>ns</code> field pointing to an xmlNs structure detailing the namespace
3615 prefix and its URI.</p>
3617 <p>@@Interfaces@@</p>
3621 <p>Usually people object to using namespaces together with validity checking.
3622 I will try to make sure that using namespaces won't break validity checking,
3623 so even if you plan to use or currently are using validation I strongly
3624 suggest adding namespaces to your document. A default namespace scheme
3625 <code>xmlns="http://...."</code> should not break validity even on less
3626 flexible parsers. Using namespaces to mix and differentiate content coming
3627 from multiple DTDs will certainly break current validation schemes. I will
3628 try to provide ways to do this, but this may not be portable or
3631 <h2><a name="Upgrading">Upgrading 1.x code</a></h2>
3633 <p>Incompatible changes:</p>
3635 <p>Version 2 of libxml is the first version introducing serious backward
3636 incompatible changes. The main goals were:</p>
3638 <li>a general cleanup. A number of mistakes inherited from the very early
3639 versions couldn't be changed due to compatibility constraints. Example
3640 the "childs" element in the nodes.</li>
3641 <li>Uniformization of the various nodes, at least for their header and link
3642 parts (doc, parent, children, prev, next), the goal is a simpler
3643 programming model and simplifying the task of the DOM implementors.</li>
3644 <li>better conformances to the XML specification, for example version 1.x
3645 had an heuristic to try to detect ignorable white spaces. As a result the
3646 SAX event generated were ignorableWhitespace() while the spec requires
3647 character() in that case. This also mean that a number of DOM node
3648 containing blank text may populate the DOM tree which were not present
3652 <h3>How to fix libxml-1.x code:</h3>
3654 <p>So client code of libxml designed to run with version 1.x may have to be
3655 changed to compile against version 2.x of libxml. Here is a list of changes
3656 that I have collected, they may not be sufficient, so in case you find other
3657 change which are required, <a href="mailto:Daniel.Ïeillardw3.org">drop me a
3660 <li>The package name have changed from libxml to libxml2, the library name
3661 is now -lxml2 . There is a new xml2-config script which should be used to
3662 select the right parameters libxml2</li>
3663 <li>Node <strong>childs</strong> field has been renamed
3664 <strong>children</strong> so s/childs/children/g should be applied
3665 (probability of having "childs" anywhere else is close to 0+</li>
3666 <li>The document don't have anymore a <strong>root</strong> element it has
3667 been replaced by <strong>children</strong> and usually you will get a
3668 list of element here. For example a Dtd element for the internal subset
3669 and it's declaration may be found in that list, as well as processing
3670 instructions or comments found before or after the document root element.
3671 Use <strong>xmlDocGetRootElement(doc)</strong> to get the root element of
3672 a document. Alternatively if you are sure to not reference DTDs nor have
3673 PIs or comments before or after the root element
3674 s/->root/->children/g will probably do it.</li>
3675 <li>The white space issue, this one is more complex, unless special case of
3676 validating parsing, the line breaks and spaces usually used for indenting
3677 and formatting the document content becomes significant. So they are
3678 reported by SAX and if your using the DOM tree, corresponding nodes are
3679 generated. Too approach can be taken:
3681 <li>lazy one, use the compatibility call
3682 <strong>xmlKeepBlanksDefault(0)</strong> but be aware that you are
3683 relying on a special (and possibly broken) set of heuristics of
3684 libxml to detect ignorable blanks. Don't complain if it breaks or
3685 make your application not 100% clean w.r.t. to it's input.</li>
3686 <li>the Right Way: change you code to accept possibly insignificant
3687 blanks characters, or have your tree populated with weird blank text
3688 nodes. You can spot them using the commodity function
3689 <strong>xmlIsBlankNode(node)</strong> returning 1 for such blank
3692 <p>Note also that with the new default the output functions don't add any
3693 extra indentation when saving a tree in order to be able to round trip
3694 (read and save) without inflating the document with extra formatting
3697 <li>The include path has changed to $prefix/libxml/ and the includes
3698 themselves uses this new prefix in includes instructions... If you are
3699 using (as expected) the
3700 <pre>xml2-config --cflags</pre>
3701 <p>output to generate you compile commands this will probably work out of
3704 <li>xmlDetectCharEncoding takes an extra argument indicating the length in
3705 byte of the head of the document available for character detection.</li>
3708 <h3>Ensuring both libxml-1.x and libxml-2.x compatibility</h3>
3710 <p>Two new version of libxml (1.8.11) and libxml2 (2.3.4) have been released
3711 to allow smooth upgrade of existing libxml v1code while retaining
3712 compatibility. They offers the following:</p>
3714 <li>similar include naming, one should use
3715 <strong>#include<libxml/...></strong> in both cases.</li>
3716 <li>similar identifiers defined via macros for the child and root fields:
3717 respectively <strong>xmlChildrenNode</strong> and
3718 <strong>xmlRootNode</strong></li>
3719 <li>a new macro <strong>LIBXML_TEST_VERSION</strong> which should be
3720 inserted once in the client code</li>
3723 <p>So the roadmap to upgrade your existing libxml applications is the
3726 <li>install the libxml-1.8.8 (and libxml-devel-1.8.8) packages</li>
3727 <li>find all occurrences where the xmlDoc <strong>root</strong> field is
3728 used and change it to <strong>xmlRootNode</strong></li>
3729 <li>similarly find all occurrences where the xmlNode
3730 <strong>childs</strong> field is used and change it to
3731 <strong>xmlChildrenNode</strong></li>
3732 <li>add a <strong>LIBXML_TEST_VERSION</strong> macro somewhere in your
3733 <strong>main()</strong> or in the library init entry point</li>
3734 <li>Recompile, check compatibility, it should still work</li>
3735 <li>Change your configure script to look first for xml2-config and fall
3736 back using xml-config . Use the --cflags and --libs output of the command
3737 as the Include and Linking parameters needed to use libxml.</li>
3738 <li>install libxml2-2.3.x and libxml2-devel-2.3.x (libxml-1.8.y and
3739 libxml-devel-1.8.y can be kept simultaneously)</li>
3740 <li>remove your config.cache, relaunch your configuration mechanism, and
3741 recompile, if steps 2 and 3 were done right it should compile as-is</li>
3742 <li>Test that your application is still running correctly, if not this may
3743 be due to extra empty nodes due to formating spaces being kept in libxml2
3744 contrary to libxml1, in that case insert xmlKeepBlanksDefault(1) in your
3745 code before calling the parser (next to
3746 <strong>LIBXML_TEST_VERSION</strong> is a fine place).</li>
3749 <p>Following those steps should work. It worked for some of my own code.</p>
3751 <p>Let me put some emphasis on the fact that there is far more changes from
3752 libxml 1.x to 2.x than the ones you may have to patch for. The overall code
3753 has been considerably cleaned up and the conformance to the XML specification
3754 has been drastically improved too. Don't take those changes as an excuse to
3755 not upgrade, it may cost a lot on the long term ...</p>
3757 <h2><a name="Thread">Thread safety</a></h2>
3759 <p>Starting with 2.4.7, libxml makes provisions to ensure that concurrent
3760 threads can safely work in parallel parsing different documents. There is
3761 however a couple of things to do to ensure it:</p>
3763 <li>configure the library accordingly using the --with-threads options</li>
3764 <li>call xmlInitParser() in the "main" thread before using any of the
3765 libxml API (except possibly selecting a different memory allocator)</li>
3768 <p>Note that the thread safety cannot be ensured for multiple threads sharing
3769 the same document, the locking must be done at the application level, libxml
3770 exports a basic mutex and reentrant mutexes API in <libxml/threads.h>.
3771 The parts of the library checked for thread safety are:</p>
3773 <li>concurrent loading</li>
3774 <li>file access resolution</li>
3775 <li>catalog access</li>
3776 <li>catalog building</li>
3777 <li>entities lookup/accesses</li>
3779 <li>global variables per-thread override</li>
3780 <li>memory handling</li>
3783 <p>XPath is supposed to be thread safe now, but this wasn't tested
3786 <h2><a name="DOM"></a><a name="Principles">DOM Principles</a></h2>
3788 <p><a href="http://www.w3.org/DOM/">DOM</a> stands for the <em>Document
3789 Object Model</em>; this is an API for accessing XML or HTML structured
3790 documents. Native support for DOM in Gnome is on the way (module gnome-dom),
3791 and will be based on gnome-xml. This will be a far cleaner interface to
3792 manipulate XML files within Gnome since it won't expose the internal
3795 <p>The current DOM implementation on top of libxml is the <a
3796 href="http://cvs.gnome.org/lxr/source/gdome2/">gdome2 Gnome module</a>, this
3797 is a full DOM interface, thanks to Paolo Casarini, check the <a
3798 href="http://www.cs.unibo.it/~casarini/gdome2/">Gdome2 homepage</a> for more
3801 <h2><a name="Example"></a><a name="real">A real example</a></h2>
3803 <p>Here is a real size example, where the actual content of the application
3804 data is not kept in the DOM tree but uses internal structures. It is based on
3805 a proposal to keep a database of jobs related to Gnome, with an XML based
3806 storage structure. Here is an <a href="gjobs.xml">XML encoded jobs
3808 <pre><?xml version="1.0"?>
3809 <gjob:Helping xmlns:gjob="http://www.gnome.org/some-location">
3813 <gjob:Project ID="3"/>
3814 <gjob:Application>GBackup</gjob:Application>
3815 <gjob:Category>Development</gjob:Category>
3818 <gjob:Status>Open</gjob:Status>
3819 <gjob:Modified>Mon, 07 Jun 1999 20:27:45 -0400 MET DST</gjob:Modified>
3820 <gjob:Salary>USD 0.00</gjob:Salary>
3821 </gjob:Update>
3823 <gjob:Developers>
3824 <gjob:Developer>
3825 </gjob:Developer>
3826 </gjob:Developers>
3828 <gjob:Contact>
3829 <gjob:Person>Nathan Clemons</gjob:Person>
3830 <gjob:Email>nathan@windsofstorm.net</gjob:Email>
3831 <gjob:Company>
3832 </gjob:Company>
3833 <gjob:Organisation>
3834 </gjob:Organisation>
3835 <gjob:Webpage>
3836 </gjob:Webpage>
3837 <gjob:Snailmail>
3838 </gjob:Snailmail>
3841 </gjob:Contact>
3843 <gjob:Requirements>
3844 The program should be released as free software, under the GPL.
3845 </gjob:Requirements>
3848 </gjob:Skills>
3850 <gjob:Details>
3851 A GNOME based system that will allow a superuser to configure
3852 compressed and uncompressed files and/or file systems to be backed
3853 up with a supported media in the system. This should be able to
3854 perform via find commands generating a list of files that are passed
3855 to tar, dd, cpio, cp, gzip, etc., to be directed to the tape machine
3856 or via operations performed on the filesystem itself. Email
3857 notification and GUI status display very important.
3858 </gjob:Details>
3863 </gjob:Helping></pre>
3865 <p>While loading the XML file into an internal DOM tree is a matter of
3866 calling only a couple of functions, browsing the tree to gather the data and
3867 generate the internal structures is harder, and more error prone.</p>
3869 <p>The suggested principle is to be tolerant with respect to the input
3870 structure. For example, the ordering of the attributes is not significant,
3871 the XML specification is clear about it. It's also usually a good idea not to
3872 depend on the order of the children of a given node, unless it really makes
3873 things harder. Here is some code to parse the information for a person:</p>
3877 typedef struct person {
3885 } person, *personPtr;
3888 * And the code needed to parse it
3890 personPtr parsePerson(xmlDocPtr doc, xmlNsPtr ns, xmlNodePtr cur) {
3891 personPtr ret = NULL;
3893 DEBUG("parsePerson\n");
3895 * allocate the struct
3897 ret = (personPtr) malloc(sizeof(person));
3899 fprintf(stderr,"out of memory\n");
3902 memset(ret, 0, sizeof(person));
3904 /* We don't care what the top level element name is */
3905 cur = cur->xmlChildrenNode;
3906 while (cur != NULL) {
3907 if ((!strcmp(cur->name, "Person")) && (cur->ns == ns))
3908 ret->name = xmlNodeListGetString(doc, cur->xmlChildrenNode, 1);
3909 if ((!strcmp(cur->name, "Email")) && (cur->ns == ns))
3910 ret->email = xmlNodeListGetString(doc, cur->xmlChildrenNode, 1);
3917 <p>Here are a couple of things to notice:</p>
3919 <li>Usually a recursive parsing style is the more convenient one: XML data
3920 is by nature subject to repetitive constructs and usually exhibits highly
3921 structured patterns.</li>
3922 <li>The two arguments of type <em>xmlDocPtr</em> and <em>xmlNsPtr</em>,
3923 i.e. the pointer to the global XML document and the namespace reserved to
3924 the application. Document wide information are needed for example to
3925 decode entities and it's a good coding practice to define a namespace for
3926 your application set of data and test that the element and attributes
3927 you're analyzing actually pertains to your application space. This is
3928 done by a simple equality test (cur->ns == ns).</li>
3929 <li>To retrieve text and attributes value, you can use the function
3930 <em>xmlNodeListGetString</em> to gather all the text and entity reference
3931 nodes generated by the DOM output and produce an single text string.</li>
3934 <p>Here is another piece of code used to parse another level of the
3936 <pre>#include <libxml/tree.h>
3938 * a Description for a Job
3940 typedef struct job {
3946 personPtr developers[100]; /* using dynamic alloc is left as an exercise */
3950 * And the code needed to parse it
3952 jobPtr parseJob(xmlDocPtr doc, xmlNsPtr ns, xmlNodePtr cur) {
3955 DEBUG("parseJob\n");
3957 * allocate the struct
3959 ret = (jobPtr) malloc(sizeof(job));
3961 fprintf(stderr,"out of memory\n");
3964 memset(ret, 0, sizeof(job));
3966 /* We don't care what the top level element name is */
3967 cur = cur->xmlChildrenNode;
3968 while (cur != NULL) {
3970 if ((!strcmp(cur->name, "Project")) && (cur->ns == ns)) {
3971 ret->projectID = xmlGetProp(cur, "ID");
3972 if (ret->projectID == NULL) {
3973 fprintf(stderr, "Project has no ID\n");
3976 if ((!strcmp(cur->name, "Application")) && (cur->ns == ns))
3977 ret->application = xmlNodeListGetString(doc, cur->xmlChildrenNode, 1);
3978 if ((!strcmp(cur->name, "Category")) && (cur->ns == ns))
3979 ret->category = xmlNodeListGetString(doc, cur->xmlChildrenNode, 1);
3980 if ((!strcmp(cur->name, "Contact")) && (cur->ns == ns))
3981 ret->contact = parsePerson(doc, ns, cur);
3988 <p>Once you are used to it, writing this kind of code is quite simple, but
3989 boring. Ultimately, it could be possible to write stubbers taking either C
3990 data structure definitions, a set of XML examples or an XML DTD and produce
3991 the code needed to import and export the content between C data and XML
3992 storage. This is left as an exercise to the reader :-)</p>
3994 <p>Feel free to use <a href="example/gjobread.c">the code for the full C
3995 parsing example</a> as a template, it is also available with Makefile in the
3996 Gnome CVS base under gnome-xml/example</p>
3998 <h2><a name="Contributi">Contributions</a></h2>
4000 <li>Bjorn Reese, William Brack and Thomas Broyer have provided a number of
4001 patches, Gary Pennington worked on the validation API, threading support
4002 and Solaris port.</li>
4003 <li>John Fleck helps maintaining the documentation and man pages.</li>
4004 <li><a href="mailto:igor@zlatkovic.com">Igor Zlatkovic</a> is now the
4005 maintainer of the Windows port, <a
4006 href="http://www.zlatkovic.com/projects/libxml/index.html">he provides
4008 <li><a href="mailto:Gary.Pennington@sun.com">Gary Pennington</a> provides
4009 <a href="http://garypennington.net/libxml2/">Solaris binaries</a></li>
4011 href="http://mail.gnome.org/archives/xml/2001-March/msg00014.html">Matt
4012 Sergeant</a> developed <a
4013 href="http://axkit.org/download/">XML::LibXSLT</a>, a Perl wrapper for
4014 libxml2/libxslt as part of the <a href="http://axkit.com/">AxKit XML
4015 application server</a></li>
4016 <li><a href="mailto:fnatter@gmx.net">Felix Natter</a> and <a
4017 href="mailto:geertk@ai.rug.nl">Geert Kloosterman</a> provide <a
4018 href="libxml-doc.el">an emacs module</a> to lookup libxml(2) functions
4020 <li><a href="mailto:sherwin@nlm.nih.gov">Ziying Sherwin</a> provided <a
4021 href="http://xmlsoft.org/messages/0488.html">man pages</a></li>
4022 <li>there is a module for <a
4023 href="http://acs-misc.sourceforge.net/nsxml.html">libxml/libxslt support
4024 in OpenNSD/AOLServer</a></li>
4025 <li><a href="mailto:dkuhlman@cutter.rexx.com">Dave Kuhlman</a> provided the
4026 first version of libxml/libxslt <a
4027 href="http://www.rexx.com/~dkuhlman">wrappers for Python</a></li>
4028 <li>Petr Kozelka provides <a
4029 href="http://sourceforge.net/projects/libxml2-pas">Pascal units to glue
4030 libxml2</a> with Kylix and Delphi and other Pascal compilers</li>
4031 <li><a href="mailto:aleksey@aleksey.com">Aleksey Sanin</a> implemented the
4032 <a href="http://www.w3.org/Signature/">XML Canonicalization and XML
4033 Digital Signature</a> <a
4034 href="http://www.aleksey.com/xmlsec/">implementations for libxml2</a></li>
4035 <li><a href="mailto:Steve.Ball@zveno.com">Steve Ball</a>, <a
4036 href="http://www.zveno.com/">Zveno</a> and contributors maintain <a
4037 href="http://tclxml.sourceforge.net/">tcl bindings for libxml2 and
4038 libxslt</a>, as well as <a
4039 href="http://tclxml.sf.net/tkxmllint.html">tkxmllint</a> a GUI for
4040 xmllint and <a href="http://tclxml.sf.net/tkxsltproc.html">tkxsltproc</a>
4041 a GUI for xsltproc.</li>