Hi, I've pushed a topic branch to the Git repository for MARC/Perl called use-dom-instead-of-sax [1]. This implements some work I've done to use a DOM parser (specifically XML::LibXML) instead of a SAX parser. Testing indicates that using XML::LibXML is faster than the current SAX implementation even when using XML::ExpatNS as the SAX driver (details of the testing can be found in two emails to koha-devel [2], [3]). Since MARC::File::XML is presently geared towards parsing a single record at a time, memory consumption shouldn't be much of an issue.
Any testing would be gratefully appreciated. I would also be curious to hear people's thoughts as to whether the LibXML mode ought to replace the SAX mode entirely or if people feel that both should be supported by MARC::File::XML. [1] http://marcpm.git.sourceforge.net/git/gitweb.cgi?p=marcpm/marcpm;a=shortlog;h=refs/heads/use-dom-instead-of-sax [2] http://lists.koha-community.org/pipermail/koha-devel/2010-November/034718.html [3] http://lists.koha-community.org/pipermail/koha-devel/2010-November/034719.html Regards, Galen -- Galen Charlton gmcha...@gmail.com
