Re: generating attributes and content for an element
Michael Ludwig <[email protected]>
| Newsgroups | gmane.comp.lang.perl.xml |
|---|---|
| Message-ID | <[email protected]> |
Tod Harter schrieb: > Yeah, I haven't used the pull parser myself. The memory issue though > is generally an XSLT issue. You can't process a lot of XSLT template > matches or many XPath expressions in a stream oriented manner, so > essentially the whole input DOM has to be in memory. The whole thing has to be in memory, that's a prerequisite for full-blown XSLT. The Saxon processor has extensions for streaming, but only a subset of XSLT can be done in streaming mode. It cannot be otherwise. > If its a large document, then you can be talking really serious > amounts of RAM. Yes. For 1 MB of input, count 10 MB of memory. Documents exceeding the resources available have to be split up resorting to other techniques, like SAX, or in Perl XML::Twig. > Now maybe libxslt is smart enough to understand what it may need to > hold onto and optimize, but even one match='//...' and you're pretty > much sunk there! That should read "select='//*'" or whatever - "match='//*'" doesn't make sense, it's the same as "match='*'". Michael Ludwig _______________________________________________ Perl-XML mailing list [email protected] To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs