I've isolated the problem a bit more. The script will not leak if line 25: transformer.transform2String(transformer.parseSource(xml), xslt)
is changed to: transformer.transform2String(xml, xslt)
Looks like the DOM tree or at least parts of it are not freed. But there are no need to call parseSource when the XML is different for each iteration, that solves my problem, but the parseSoruce function should free all memory anyways.
Yes it should. It's not obvious to me why it doesn't. I'll investage it in detail a bit later.
Cheers, Brian -- http://mail.python.org/mailman/listinfo/python-list
