Mailman 3 September 2009 - lxml - The Python XML Toolkit

[lxml-dev] problems trying to install lxml 2.2 on Mac OS X
by Chris Withers Sept. 12, 2009

Sept. 12, 2009

Hi All, I'm getting the following error when trying to install: Building lxml version 2.2. NOTE: Trying to build without Cython, pre-generated 'src/lxml/lxml.etree.c' needs to be available. Using build configuration of libxslt 1.1.11 Building against libxml2/libxslt in the following directory: /usr/lib src/lxml/lxml.etree.c:169:31:src/lxml/lxml.etree.c:169:31: error: libxml/schematron.h: No such file or directory error: libxml/schematron.h: No such file or directory <snip> src/lxml/lxml.etree.c:135067: error: dereferencing pointer to incomplete type src/lxml/lxml.etree.c:135068: error: dereferencing pointer to incomplete type src/lxml/lxml.etree.c: At top level: src/lxml/lxml.etree.c:135174: error: invalid application of 'sizeof' to incomplete type 'struct __pyx_obj_4lxml_5etree__ParserSchemaValidationContext' lipo: can't figure out the architecture type of: /var/tmp//ccxExHVU.out error: Setup script exited with error: command 'gcc' failed with exit status 1 What am I doing wrong? cheers, Chris

4 12

[lxml-dev] etree.XMLParser: two possible bugs?
by Yaniv Aknin Sept. 11, 2009

Sept. 11, 2009

Hi, I'm far from being an xml/libxml/lxml guru, so I feel a bit unconfident reporting these two 'peculiarities' I have with lxml. However, I've managed to reduce things to simple enough code that I don't know how to explain in any terms other than lxml/libxml bugs, and I'd appreciate it if the list could confirm my findings before I report them as bugs. Attached is a short piece of my code, called lxml_test.py, defining a class called TerminationDetectingXMLParser. As the name suggests, this is a stream XML parser that 'knows' whether or not it has seen the end of the document it is being fed with. I have two peculiarities with this code: 1. While the code works for: >>> td.feed('<a>') >>> td.feed('<b></b>') >>> td.feed('</a>') ...it doesn't work for: >>> td.feed('<a><b></b></a>') It seems to me that etree.XMLParser doesn't call end() on the treebuilder in the latter case, which, humbly, I think is a bug; etree.XMLParser has everything it needs to call end() on the target. 2. The second peculiarity is a bit weirder. Much to my dismay we have two Python environments built and maintained by different teams. Both environments are 2.6.2 and lxml 2.2.2, though one uses slightly different libxml/libxslt versions. In one environment, peculiarity (1) described above exists, but the code runs well. In the other, an AssertionError is raised from saxparser.pxi during lxml.etree.TreeBuilder.close. This is despite the XML input being really really trivial. I have attached two ASCII tty-screehshots of the two interpreters running against the same code, one with the assertion error, the other showing just peculiarity (1). Am I using the library incorrectly? Are these libxml issues? Should I open bugs for these? What workaround should I use? Thanks in advance, - Yaniv

2 6

[lxml-dev] [Fwd: [xml] Release of libxml2-2.7.4]
by Stefan Behnel Sept. 11, 2009

Sept. 11, 2009

Hi, just wanted to give a note that libxml2 2.7.4 was released, which includes quite a number of bug fixes. Some of those are likely relevant to lxml users, e.g. those for the HTML parser or a couple of XML Schema improvements. I didn't do much testing, but so far, lxml passes its test suite nicely with this release. Kudos to Daniel Veillard for the work he keeps putting into libxml2. Have fun, Stefan -------- Original-Message -------- Subject: [xml] Release of libxml2-2.7.4 Date: Thu, 10 Sep 2009 18:51:18 +0200 From: Daniel Veillard <veillard(a)redhat.com> To: xml(a)gnome.org Better late than never, but an awful lot of pending bug got fixed. Still no major improvement except adding symbol versioning to libxml2 shared libs, which is fairy important for long term maintainance, but not worth jumping to 2.8.0 Tarball and signed rpms available at ftp://xmlsoft.org/libxml2/ There are still a few things which I would have loved to put in the release like per context error handling and the like but I prefer a (nearly) bug fix only release that people can upgrade to without troubles and then work on changing more stuff: Improvements: - Autoregenerate libxml2.syms automated checkings (Daniel Veillard) - Add symbol versioning to libxml2 shared libs (Daniel Veillard) Documentation: - 544910 typo: "renciliateNs" (Leonid Evdokimov) - Add VxWorks to list of OSes (Daniel Veillard) - Regenerate the documentation and update for git (Daniel Veillard) - 560524 – xmlTextReaderLocalName description (Daniel Veillard) - Added sponsoring by AOE media for the server (Daniel Veillard) - updated URLs for GNOME (Vincent Lefevre) - more warnings about xmlCleanupThreads and xmlCleanupParser (Daniel Veillard) Portability: - 593857 try to work around thread pbm MinGW 4.4 (Daniel Veillard) - 594250 rename ATTRIBUTE_ALLOC_SIZE to avoid clashes (Daniel Veillard) - Fix Windows build * relaxng.c: fix windows build (Rob Richards) - Fix the globals.h to use XMLPUBFUN (Paul Smith) - Problem with extern extern in header (Daniel Veillard) - Add -lnetwork for compiling on Haiku (Scott McCreary) - Runtest portability patch for Solaris (Tim Rice) - Small patch to accomodate the Haiku OS (Scott McCreary) - 584605 package VxWorks folder in the distribution (Daniel Veillard) - 574017 Realloc too expensive on most platform (Daniel Veillard) - Fix windows build (Rob Richards) - 545579 doesn't compile without schema support (Daniel Veillard) - xmllint use xmlGetNodePath when not compiled in (Daniel Veillard) - Try to avoid __imp__xmlFree link trouble on msys (Daniel Veillard) - Allow to select the threading system on Windows (LRN) - Fix Solaris binary links, cleanups (Daniel Veillard) - Bug 571059 – MSVC doesn't work with the bakefile (Intron) - fix ATTRIBUTE_PRINTF header clash (Belgabor and Mike Hommey) - fixes for Borland/CodeGear/Embarcadero compilers (Eric Zurcher) Bug fixes: - 594514 memory leaks - duplicate initialization (MOD) - Wrong block opening in htmlNodeDumpOutputInternal (Daniel Veillard) - 492317 Fix Relax-NG validation problems (Daniel Veillard) - 558452 fight with reg test and error report (Daniel Veillard) - 558452 RNG compilation of optional multiple child (Daniel Veillard) - 579746 XSD validation not correct / nilable groups (Daniel Veillard) - 502960 provide namespace stack when parsing entity (Daniel Veillard) - 566012 part 2 fix regresion tests and push mode (Daniel Veillard) - 566012 autodetected encoding and encoding conflict (Daniel Veillard) - 584220 xpointer(/) and xinclude problems (Daniel Veillard) - 587663 Incorrect Attribute-Value Normalization (Daniel Veillard) - 444994 HTML chunked failure for attribute with <> (Daniel Veillard) - Fix end of buffer char being split in XML parser (Daniel Veillard) - Non ASCII character may be split at buffer end (Adiel Mittmann) - 440226 Add xmlXIncludeProcessTreeFlagsData API (Stefan Behnel) - 572129 speed up parsing of large HTML text nodes (Markus Kull) - Fix HTML parsing with 0 character in CDATA (Daniel Veillard) - Fix SetGenericErrorFunc and SetStructured clash (Wang Lam) - 566012 Incomplete EBCDIC parsing support (Martin Kögler) - 541335 HTML avoid creating 2 head or 2 body element (Daniel Veillard) - 541237 error correcting missing end tags in HTML (Daniel Veillard) - 583439 missing line numbers in push mode (Daniel Veillard) - 587867 xmllint --html --xmlout serializing as HTML (Daniel Veillard) - 559501 avoid select and use poll for nanohttp (Raphael Prevost) - 559410 - Regexp bug on (...)? constructs (Daniel Veillard) - Fix a small problem on previous HTML parser patch (Daniel Veillard) - 592430 - HTML parser runs into endless loop (Daniel Veillard) - 447899 potential double free in xmlFreeTextReader (Daniel Veillard) - 446613 small validation bug mixed content with NS (Daniel Veillard) - Fix the problem of revalidating a doc with RNG (Daniel Veillard) - Fix xmlKeepBlanksDefault to not break indent (Nick Wellnhofer) - 512131 refs from externalRef part need to be added (Daniel Veillard) - 512131 crash in xmlRelaxNGValidateFullElement (Daniel Veillard) - 588441 allow '.' in HTML Names even if invalid (Daniel Veillard) - 582913 Fix htmlSetMetaEncoding() to be nicer (Daniel Veillard) - 579317 Try to find the HTML encoding information (Daniel Veillard) - 575875 don't output charset=html (Daniel Veillard) - 571271 fix semantic of xsd:all with minOccurs=0 (Daniel Veillard) - 570702 fix a bug in regexp determinism checking (Daniel Veillard) - 567619 xmlValidateNotationUse missing param test (Daniel Veillard) - 574393 – utf-8 filename magic for compressed files (Hans Breuer) - Fix a couple of problems in the parser (Daniel Veillard) - 585505 – Document ids and refs populated by XSD (Wayne Jensen) - 582906 XSD validating multiple imports of the same schema (Jason Childs) - Bug 582887 – problems validating complex schemas (Jason Childs) - Bug 579729 – fix XSD schemas parsing crash (Miroslav Bajtos) - 576368 – htmlChunkParser with special attributes (Jiri Netolicky) - Bug 565747 – relax anyURI data character checking (Vincent Lefevre) - Preserve attributes of include start on tree copy (Petr Pajas) - Skip silently unrecognized XPointer schemes (Jakub Wilk) - Fix leak on SAX1, xmllint --sax1 option and debug (Daniel Veillard) - potential NULL dereference on non-glibc (Jim Meyering) - Fix an XSD validation crash (Daniel Veillard) - Fix a regression in streaming entities support (Daniel Veillard) - Fix a couple of ABI issues with C14N 1.1 (Aleksey Sanin) - Aleksey Sanin support for c14n 1.1 (Aleksey Sanin) - reader bug fix with entities - use options from current parser context for external entities (Rob Richards) - 581612 use %s to printf strings (Christian Persch) - 584605 change the threading initialization sequence (Igor Novoseltsev) - 580705 keep line numbers in HTML parser (Aaron Patterson) - 581803 broken HTML table attributes init (Roland Steiner) - do not set error code in xmlNsWarn (Rob Richards) - 564217 fix structured error handling problems - reuse options from current parser for entities (Rob Richards) - xmlXPathRegisterNs should not allow enpty prefixes (Daniel Veillard) - add a missing check in xmlAddSibling (Kris Breuker) - avoid leaks on errors (Jinmei Tatuya) Cleanups: - Chasing dead assignments reported by clang-scan (Daniel Veillard) - A few more safety cleanup raised by scan (Daniel Veillard) - Fixing assorted potential problems raised by scan (Daniel Veillard) - Potential uninitialized arguments raised by scan (Daniel Veillard) - Fix a bunch of scan 'dead increments' and cleanup (Daniel Veillard) - Remove a pedantic warning (Daniel Veillard) - 555833 always use rm -f in uninstall-local (Daniel Veillard) - 542394 xmlRegisterOutputCallbacks MAX_INPUT_CALLBACK (Daniel Veillard) - Make xmlRecoverDoc const (Martin Trappel) (Daniel Veillard) - Both args of xmlStrcasestr are const (Daniel Veillard) - hide the nbParse* variables used for debugging (Mike Hommey) - 570806 changed include of config.h (William M. Brack) - cleanups and error reports when xmlTextWriterVSprintf fails (Jinmei Tatuya) Thanks everybody who helped by submitting ideas, bug reports or patches ! Enjoy ! Daniel

1 0

[lxml-dev] copying an element will "reset" is nsmap
by Hervé Cauwelier Sept. 11, 2009

Sept. 11, 2009

Hi, I have another unexpected behaviour. I have a root element with many namespaces declared. When I introspect the first child of the root, the nsmap is full. When I "deepcopy" it, only the the namespace used in the tag name is kept. If the name is not qualified, nsmap turns empty. (Sorry if I don't use the right XML semantic.) I tried to reaffect nsmap but the attribute is read-only. How could I preserve the nsmap attribute of unattached elements? Regards, Hervé -- lxml.etree: (2, 2, 2, 0) libxml used: (2, 7, 3) libxml compiled: (2, 7, 3) libxslt used: (1, 1, 24) libxslt compiled: (1, 1, 24)

2 7

[lxml-dev] current trunk includes static build for libiconv
by Stefan Behnel Sept. 2, 2009

Sept. 2, 2009

Hi, since there were problems also with libiconv on MacOS recently, I added libiconv to the list of libraries that "--static-deps" builds. I'd be happy to get some feedback on this to see if it actually works for those who reported problems. http://codespeak.net/lxml/build.html#subversion http://codespeak.net/lxml/build.html#building-lxml-on-macos-x Note that the trunk builds as lxml version "2.3dev" now. Copying over the buildlibxml.py script to another lxml version should also work. http://codespeak.net/svn/lxml/trunk/buildlibxml.py Stefan

2 2

Re: [lxml-dev] getname() method on 'smart' attribute string values
by Nicholas Dudfield Sept. 2, 2009

Sept. 2, 2009

> You can give it a try on the trunk, if you like. > > https://codespeak.net/viewvc/?view=rev&revision=67010 > > http://codespeak.net/lxml/build.html > > Stefan > I also have need for this functionality and also a bugfix from a revision ahead of the stable version 2.2.2 available for windows. I heard libxml2/lxml is a PITA to build on windows so being inexperienced I'll not bother attempting it before ruling out alternatives. Is there a build bot with dist zips of the latest revisions available anywhere for windows ?

2 1