Revision history for Perl extension XML::LibXML

1.75 Fri Jun 24 19:00:40 IDT 2011

• Correct some typos reported in
  • https://rt.cpan.org/Ticket/Display.html?id=54390
• Fix the handling of XML::LibXML::InputCallbacks at load_xml().
  • https://rt.cpan.org/Ticket/Display.html?id=58190
  • The problem was that the input callbacks were not cloned in _clone().
• Apply the patches from https://rt.cpan.org/Ticket/Display.html?id=56334
  • Convert t/02parse.t to Test::More .
  • Thanks to TODDR .
  • Removed the diag() messages which were annoying.
• Add 'make runtest' and 'make distruntest' targets to run the tests using Test::Run ( http://beta.metacpan.org/module/Test::Run ).
  • Adds colours and stuff like that.
• Add << LICENSE => 'perl' >> to the Makefile.PL for a license meta-data in the META.YML.
• Feature implementation: joining congruent character data together in SAX driver .
  • Apply a somewhat modified patch from:
    • https://rt.cpan.org/Ticket/Display.html?id=52368
• Add t/pod.t .
• Fix https://rt.cpan.org/Ticket/Display.html?id=55000 :
  • Apply modified patch in the bug report.
  • << If an element contains both a default namespace declaration and a second namespace declaration, adding an attribute using the default namespace declaration will cause that attribute to have the other prefix. >>

1.74 Thu Jun 23 16:20:42 IDT 2011

• More work on the t/*.t test scripts.
• Add scripts/Test.pm-to-Test-More.pl to semi-automatically convert a test script from Test.pm to Test::More.
• Change NodeSet to NodeList in the documentation of lib/XML/LibXML/NodeList.pm .
  • Resolved https://rt.cpan.org/Ticket/Display.html?id=60998
• Makefile.PL: now saying we are trying to link against -lm, -lz and -lxml2 . Not only -lxml2:
  • https://rt.cpan.org/Ticket/Display.html?id=51439
• https://rt.cpan.org/Ticket/Display.html?id=61756
  • << $node = XML::LibXML::Comment( $content ); >> is wrong.
• Documentation: moved away from Indirect-object-notation and added some missing "my"s:
  • http://www.modernperlbooks.com/mt/2009/08/the-problems-with-indirect-object-notation.html
• Fix failing t/01basic.t when compiling against libxml2 that comes from git.
  • https://rt.cpan.org/Public/Bug/Display.html?id=54951
  • Thanks to Evan Carroll ( http://www.evancarroll.com/ ) for the report.

1.73 Sat Jun 18 10:53:44 IDT 2011

• Calculating $err->column() properly, so it won't be maxed out at

  80

  • https://rt.cpan.org/Public/Bug/Display.html?id=66642
  • the context still maxes at 80 (to avoid wasting RAM) but we still continue past that to get the accurate verdict.
  • Thanks to SCOP.

• Update the repository in the documentation to point to the bitbucket.org one.
• Revamped Makefile.PL:
  • Got rid of "\t" characters.
  • Add "use strict" and "use warnings".
  • Add resources and keywords to the META_MERGE.
  • Other changes.
• Fix https://rt.cpan.org/Public/Bug/Display.html?id=53632 :
  • << when calling normalize on a node, processing of children nodes will stop when an empty element node is encountered. >>
  • Thanks to Daniel Frett for the patch.
• Apply the patch from Daniel Frett's InputCallbackFix branch.
  • a partial fix to https://rt.cpan.org/Public/Bug/Display.html?id=4263 .
  • Call two $parser->parse_string() in succession.
• Apply the NestedParsing patch.
  • more of https://rt.cpan.org/Public/Bug/Display.html?id=4263
  • Thanks to Daniel Frett for the patch.

        [QUOTE]
        Updated how legacy parser local callbacks are utilized by
        init_callbacks so that the XML::LibXML::InputCallback object doesn't
        have to be temporarily modified during the parsing process.
    
        This change could break code for users that have subclassed
        XML::LibXML::InputCallback and overridden the init_callbacks method
        [/QUOTE]
    - Documentation fixes patch from Daniel Frett on:
        - From https://github.com/frett/perl-libxml .

1.72            Thu Jun 16 19:26:13 IDT 2011
    - Removed a stray file from the MANIFEST
        - http://rt.cpan.org/Ticket/Display.html?id=68865
        - Warned on "kit not complete".
        - Thanks to obrien.jk

1.71            Tue Jun 14 19:43:50 IDT 2011

• turn XML_LIBXML_PARSE_DEFAULTS constant to $XML::LibXML::XML_LIBXML_PARSE_DEFAULTS
• Apply 0001-XML-LibXML-Error-no-need-to-AUTOLOAD-domain.patch from https://rt.cpan.org/Public/Bug/Display.html?id=68575 - no need to AUTOLOAD 'domain' because a method like that exists.

  -- Applied by SHLOMIF. -- Thanks to Aaron Crane.

  • Apply 0002-XML-LibXML-Error-avoid-AUTOLOAD.patch from https://rt.cpan.org/Public/Bug/Display.html?id=68575 - get rid of AUTOLOAD completely. -- Applied by SHLOMIF. -- Thanks to Aaron Crane.
  • Apply 0003-XML-LibXML-Error-make-domain-work-for-unknown-domain.patch from https://rt.cpan.org/Public/Bug/Display.html?id=68575 - handle unknown domains. -- Applied by SHLOMIF. -- Thanks to Aaron Crane.
  • Apply 0004-XML-LibXML-Error-add-domains-from-newer-libxml2.patch from https://rt.cpan.org/Public/Bug/Display.html?id=68575 - add more errors. -- Applied by SHLOMIF. -- Thanks to Aaron Crane.
  • Apply 0005-XML-LibXML-Error-avoid-malformed-UTF-8-warnings.patch from https://rt.cpan.org/Public/Bug/Display.html?id=68575 -- Applied by SHLOMIF. -- Thanks to Aaron Crane.
  • In replaceDataString - use http://perldoc.perl.org/functions/quotemeta.html instead of a long (and incomplete) list of characters to escape.

    -- With test. -- also fix deleteDataString by making it use replaceDataString for help. -- Fixing https://rt.cpan.org/Ticket/Display.html?id=68564 -- Thanks to Daniel Perrett .

1.70

• various fixes and improvements in the documentation
• added (convenient yet non-standard) methods nonBlankChildNodes, firstNonBlankChild, nextNonBlankSibling, prevNonBlankSibling that skip empty or whitespace-only Text and CDATA nodes
• exposed and documented external entity handler
• XPathContext can now be passed to toStringC14N and toStringEC14N (e.g. to provide NS mapping for the XPath expression)
• avoid using libxml2's globals (Nick Wellnhofer)
• added interface to libxml2's regexp implementation: XML::LibXML::RegExp
• added XML::LibXML->load_xml and XML::LibXML->load_html with uniform and cleaner API than the old parse_* family
• cleanup code dealing with parsing flags
• fix bogus validation results if revalidating a modified document
• added 'eq' and 'cmp' overloading on XML::LibXML::Error and set fallback to 1
• lots of bugs fixed

1.69_2

• provide context and more accurate column number in structured errors
• clarify license and copyright
• support for Win32+mingw+ActiveState

1.69_1

• merge with XML::LibXML::Common
• fix compilation on Windows with mingw or msvc
• fix a bug in structured errors preventing the previous errors from being reported
• fix compilation bugs
• fix encoding problem in reader
• added getAttributeHash to the reader interface
• fix segfaults: reconcileNs in domReplaceChild, findnodes with a doc fragment (S. Rezic)

1.69

• fix incorrect output of getAttributeNS and possibly other methods on UTF-8
• added $node_or_xpc->exists($xpath) method
• remove accidental debug output from XML::LibXML::SAX::Builder

1.68

• compilation problem fixes

1.67

• many bugfixes (rt.cpan.org)
• added XML::LibXML::Pattern module and extended pattern support in Reader
• added XML::LibXML::XPathExpression module that can pre-compile an XPath expression
• reimplementation of the thread support (mostly by Tim Brody)
• structured errors XML::LibXML::Error
• memory leak fixes
• documentation fixes
• README - notes for building on Win32 (C.J. Madsen)

1.66

• Perl-thread support contributed by Tim Brody [rt.cpan.org #31945]
• fix [rt.cpan.org #30610] possible segmentation fault when importing nodes from a document to an element created with XML::LibXML::Element->new
• fix [rt.cpan.org #30261] Segmentation fault when extracting elements from an XML chunk
• make Makefile.PL require Perl 5.6.1
• minor fixes and additions to the documentation
• portability patch from [rt.cpan.org #29627]
• give registered Ns declarations precedence over document-specific ones in XML::LibXML::XPathContext; fixes [rt.cpan.org #29650]

1.65

• fix bug in t/40reader.t revealed by a bugfix in Test::More 0.71 (Jonathan Rockway)
• fix possible SIGSEGV when PI's or attrs created with createDocument can get garbage-collected after their owning document (old-standing bug suddenly caught by XML::Compile regression tests)
• skip tests for unsupported features on unsupported versions of Perl/libxml2
• make Reader interface require Perl 5.8 (patches to extend to 5.6 are welcome)

1.64

• fix reconcilation of the "xml" namespace [rt.cpan.org #26450]
• make tests pass with libxml2 2.9.29 - PI regression tests now accept "" as data of an empty PI [rt.cpan.org #27659]
• strip-off UTF8 flag with $node->toString($format,1) for consistent behavior independent on the actual document encoding
• fix in XML::LibXML::Reader::nextSiblingElment
• fix synopsis for XML::LibXML::Reader
• skip tests that require Encode module if not available (perl 5.6)
• finally removed the iterator() method deprecated since 1.54
• set_document_locator support in XML::LibXML::SAX::Parser
• SYNOPSIS sections of the docs now mention which module to use and which other manpage to look into for inherited methods
• XML::LibXML::Namespace API fixed in order to achieve an agreement between the docs and the implementation

1.63

• added no_network parser flag
• added support for exclusive canonicalization (http://www.w3.org/TR/xml-exc-c14n/)
• make XInclude reflect parser flags
• documentation fixes
• better namespace reconciliation implemented by Tim Brody
• $doc->toString always returns octets
• $doc->actualEncoding returns UTF8 if no document encoding is declared (unlike $doc->getEncoding, which returns undef)

1.62

• interface to libxml2's pull-parser XML::LibXML::Reader (initiated by Heiko Klein)
• make error messages intended to the user report the line of the application call rather than that of the internal XS call
• XML::LibXML::Attr->serializeContent added (convenience function)
• fix getAttributeNode etc. w.r.t. #FIXED attributes (as well as some cases with old buggy libxml2 versions)
• warn if runtime libxml2 is older than the one used at the compile time
• if compiled against libxml2 >= 2.6.27, new parse_html_* implementation is used allowing encoding and other options to be passed to the parser
• DOM-compliant nodeNames: #comment, #text, #cdata, #document, #document-fragment
• toString on empty text node returns empty string, not undef
• cloneNode copies attributes on an element as required by the DOM spec

1.61

• get{Elements,Children}By{TagName,TagNameNS,LocalName} now obey wildcards '*', getChildrenByLocalName was added.
• XML::LibXML::XPathContext merged in
• many new tests added
• the module should now be fully compatibile with libxml2 >= 2.6.16 (some older versions compile but have problems with namespaced attributes)
• threads test skipped by default
• documentation updates (namely DOM namespace conformance in XML::LibXML::DOM)
• added setNamespaceDecl{URI,Prefix}
• get/setAttribute(NS)? implementation made xmlns aware
• all sub-modules have the same version as XML::LibXML

1.60

• getElementsById corrected to getElementById and the old name kept as an alias. Also re-implemented without XPath for improved performance
• DOM Level 3 method $attr->isId() added
• make {get,set,has}Attribute(Node)? methods work with full attribute names rather than just localnames. (Although DOM Level 3 is not very clear about the behavior of these methods for an attributes with namespaces, it certainly does not imply that getAttribute('foo') should return value of a bar:foo, which was the old behavior.)
• added publicId and systemId methods to XML::LibXML::Dtd

1.59

• new parser and callback code (Christian Glahn)
• new XML::LibXML::InputCallback class
• many bug fixes (including several memory leaks)
• documentation and regression fixes and enhancements
• Perl wrappers for parse_html_
• make sure parse_ methods are not called on class (bug 11126)
• DOM Layer 3 conformance fixes:
  • lookupNamespaceURI(empty_or_undef) now returns the default NS
• faster getChildrenByTagNameNS implementation
• remove the SGML parser code no longer supported by libxml (Michael Kröll)

1.58

• fixed a pointer initialization in parse_xml_chunk(), fixes random several segmentation faults on document fragments.
• added NSCLEAN feature to the parser interface (bug 4560)
• minor code cleanups
• updated libxml2 blacklist.
• fixed croak while requesting nodeName() of CDATA sections (bug 1694).
• more documentation updates

1.57

• added cloneNode to XML::LibXML::Document
• include Schema/RelaxNG code only with libxml2 >= 2.6.0 (to support old libxml2)
• applied patch to example/cb_example.pl (bug 4262)
• fixed insertBefore/insertAfter on empty elements (bug 3691)
• more DOM conformant XML::LibXML->createDocument API (by Robin Berjon)
• fixed encoding problems with API calls in document encoding
• improved support for importing DTD subsets
• fixed DTD validation error reporting problems with libxml2-2.6.x
• fixed compilation problems with libxml2-2.6.x
• fixed XML::LibXML::Number to support negative numbers
• added XML Schema validation interface (XML::LibXML::Schema)
• added XML RelaxNG validation interface (XML::LibXML::RelaxNG)
• Michael K. Edwards' patch applied with some amendments from Petr Pajas:
  • add debian build files (I added SKIP_SAX_INSTALL flag for Makefile.PL and changed the patch so that it doesn't disable sax parser registration completely by default, and rather made debian build use this flag)
  • general cleanup (use SV_nolen, etc.)
  • SAX parsers cleanup
  • general error reporting code cleanup/rewrite, try preventing possible memory leaks
  • recover(1) now triggers warnings (disable with $SIG{__WARN__}=sub {}) (fixes bug 1968, too)
  • slighlty more strict parse_string behavior (now same as when parsing fh, etc): e.g. parse_string("<foo:bar>"), i.e prefix without NS declaration, raises error unless recover(1) is used
  • documentation fixes/updates
  • slightly updated test set to reflect the new slightly more strict parsing.
• fixed default c14n XPath to include attributes and namespaces (Petr Pajas)
• make libxml2's xmlXPathOrderDocElems available through a new $doc->indexElements method
• added version information of libxml2
• Les Richardson's documentation patch applied.

1.56

• added line number interface (thanks to Peter Haworth)
• patch to make perl 5.8.1 and XML::LibXML work together (thanks to François Pons)
• added getElementById to XML::LibXML::Document (thanks to Robin Berjon)
• fixes symbol problem with versions of libxml2 compiled without thread support (reported by Randal L. Schwartz)
• tiny code clean ups
• corrected tested versions after a local setup problem

1.55

• fixed possible problems with math.h
• added C14N interface "toStringC14N()" (thanks to Chip Turner)
• fixed default namespace bug with libxml2 2.5.8 (by Vaclav Barta)
• fixed a NOOP in the XPath code.
• fixed insertBefore() behaviour to be DOM conform
• fixed a minor problem in Makefile.PL
• improved more documentation
• converted documentation to DocBook

NOTE:
Version 1.54 fixes potentional buffer overflows were possible with earlier versions of the package.

1.54

• fixed some major bugs, works now with libxml2 2.5.x
• fixed problem with empty document fragments
• bad tag and attribute names cannot be created anymore
• Catalog interface is aware about libxml2 configuration
• XML::LibXML should work now on systems without having zlib installed
• cleaned the error handling code, which
  • fixes bad reporting of the validating parser
  • fixes bad reporting in xpath functions
• added getElementsBy*Name() functions for the Document Class
• fixed memory management problem introduced in 1.53 (that fixes a lot strange things)
• interface for raw libxml2 DOM building functions (currently just addChild() and addNewChild(), others will follow)
• fixed namespace handling if nodes are imported to a new DOM.
• fixed segmentation fault during validation under libxml2 2.4.25
• fixed bad CDATA handing in XML::LibXML::SAX::Builder
• fixed namespace handing in XML::LibXML::SAX
• fixed attribute handing in XML::LibXML::SAX
• fixed memory leak in XML::LibXML::SAX
• fixed memory leak in XML::LibXML::Document
• fixed segfault while appending entity ref nodes to documents
• fixed some backward compatibility issues
• fixed cloning with namespaces misbehaviour
• fixed parser problems with libxml2 2.5.3+
• moved iterator classes into a separate package (after realizing some CPAN testers refuse to read their warnings from Makefile.PL)
• improved parser testsuite
• improved M
• more documentation

1.53

Parser

• catalog interface
• enabled SGML parsing
• implemented libxml2 dom recovering
• parsing into GDOME nodes is now possible
• XML::LibXML::SAX is now faster
• made XML::LibXML::SAX parser running without errors in most (all?) cases (DTD handling is still not implemented).

DOM interface

• Node Iterator class
• NodeList Iterator class
• introduced XML::GDOME import and export. (EXPERIMENTAL)
• more security checks

general blur

• removed code shared with XML::GDOME to a separate XML::LibXML::Common module (check CPAN)
• removed some redundand code
• more documentation (and docu fixes) (thanks to Petr Pajas)

major fixes:

• possible buffer overflow with broken XML: This may effect all older versions of XML::LibXML, please upgrade!
• a bug while replacing the document element.
• very stupid encoding bug. all UTF8 strings will now be marked as UTF8 correctly
• namespace functions to work with empty namespaces
• toFH()
• namespace setting in XPath functions: the namespaces of the document element will always be added now
• threaded perl 5.8.0 issues
• calling external entity handlers work again
• XML::LibXML::SAX::Parser will not throw warnings on DTD nodes

1.52

• fixed some typos (thanks to Randy Kobes and Hildo Biersma)
• fixed namespace node handling
• fixed empty Text Node bug
• corrected the parser default values.
• added some documentation

1.51

• fixed parser bug with broken XML declarations
• fixed memory management within documents that have subsets
• fixed some threaded perl issues (special thanks to Andreas Koenig for the patch)
• applied Win32 tests (special thanks to Randy Kobes for the patch)
• fixed findnodes() and find() to return empty arrays in array context if the statement was legal but produced no result.
• fixed namespace handling in xpath functions
• fixed local namespace handling in DOM functions
• pretty formating to all serializing functions NOTE the XML::LibXML::Node::toString interface changed check the XML::LibXML::Node man page
• made xpath functions verbose to perl (one can wrap evals now)
• improved native SAX interface
• improved XML::LibXML::SAX::Builder
• added getNamespaces to the node interface
• better libxml2 version testing
• more documentation

1.50

• fixed major problems with the validating parser
• fixed default behaviour of the generic parser
• fixed attribute setting of the string parser
• fixed external entity loading for entity expansion
• fixed nodeValue() to handle entities and entity refs correctly
• SAX::Parser ignores now hidden XINCLUDE nodes.
• fixed SAX::Builder to recognize namespace declarations correctly
• compatibility fixes
• importNode() bug fix
• fixed library tests and output in Makefile.PL
• added setOwnerDocument() again
• XML::LibXML::Document::process_xincludes reintroduced
• global callbacks reintroduced NOTE: the Interface changed here, read XML::LibXML manpage!
• code cleanings
• push parser interface
• basic native libxml2 SAX interface THIS INTERFACE IS STILL EXPERIMENTAL
• cloneNode clones now within documents
• more documentation

1.49

• memory management has been completely rewritten.

  now the module should not cause that many memory leaks (special thanks to Merijn Broeren and Petr Pajas for providing testcases)

• more libxml2 functions are used
• DOM API is more Level 3 conform
• ownerDocument fixed
• parser validation bug fixed (reported by Erik Ray)
• made parse_xml_chunk() report errors
• fixed the PI interface
• xpath.pl example
• better namespace support
• improved NamedNodeMap support
• restructured the interfaces
• HTML document nodes are recognized as HTML doc nodes instead of plain nodes
• XML::LibXML::SAX::Parser able to handle HTML docs now (patch by D. Hageman [dhageman@dracken.com])
• added serialization flags ($setTagCompression, $skipDtd and $skipXMLDeclaration)
• more documentation

1.40

• new parsefunction: $parser->parse_xml_chunk($string);
• appendChild( $doc_fragment ) bug fixed
• removed obsolete files (parser.?)
• fixed getElementsByTagName and getElementsByTagNameNS to fit the spec
• new functions in XML::LibXML::Element: getChildrenByTagName getChildrenByTagNameNS getElementsByLocalName
• minor fixes and extensions of the tests
• more docu ;)
• SAX added comment and PI support
• SAX added start_prefix_mapping/end_prefix_mapping
• Fixed find() bug with no results
• Added use IO::Handle so FH reads work
• A number of segfault fixes
• constants added without XML_ prefix

1.31

• Removed C-layer parser implementation.
• Added support for prefixes in find* functions
• More memory leak fixes (in custom DOMs)
• Allow global callbacks

1.30

• Full PI access
• New parser implementation (safer)
• Callbacks API changed to be on the object, not the class
• SAX uses XML::SAX now (required)
• Memory leak fixes
• applied a bunch of patches provided by T.J. Mather

1.00

• Added SAX serialisation
• Added a SAX builder module
• Fixed findnodes in scalar context to return a NodeList object
• Added findvalue($xpath)
• Added find(), which returns different things depending on the XPath
• Added Boolean, Number and Literal data types

0.99

• Added support for $doc->URI getter/setter

0.98

• New have_library implementation

0.97

• Addition of Dtd string parser
• Added support for namespace nodes (e.g. $element->getNamespaces())
• Some memory leak and segfault fixes
• Added $doc->validate([$dtd]) which throws exceptions (augments $doc->is_valid([$dtd]))
• Added doc files and test files to CPAN distro

0.96

• Addition of HTML parser
• getOwner method added
• Element->getAttributes() added
• Element->getAttributesNS(URI) added
• Documentation updates
• Memory leak fixes
• Bug Fixes

0.94

• Some DOM Level 2 cleanups
• getParentNode returns XML::LibXML::Document if we get the document node

0.93

• Addition of DOM Level 2 APIs
• some more segfault fixes
• Document is now a Node (which makes lots of things easier)

0.92

• Many segfault and other bug fixes
• More DOM API methods added

0.91

• Removed from XML::LibXSLT distribution
• Added DOM API (phish)

0.01 Sat Mar 3 17:08:00 2001

• original version; created by h2xs 1.19