[Zope-dev] memory leak

[Marco Nova]

> In either case, it seems like you shouldn't be using ParsedXML's
> ExpatBuilder, which is an interface to PyExpat which is made to create
> ParsedXML's DOM objects.  You should be using your own PyExpat
> interface to parse this information from your XML input, without the
> overhead of generating a DOM tree that you're just going to throw
> away.  The interface is documented in the XML-SIG topic area on
> python.org, and our ExpatBuilder should be a good example on how to
> interface with our PyExpat installation - just don't build the DOM if
> you don't use it.
> 
> Regardless of what you *should* be doing :), or if I'm misreading you,
> if you aren't keeping references to the DOM objects being created,
> then they should be reclaimed by the GC, and you do seem to have DOM
> objects lying around.  I'd like to see your code, is it short enough
> to post?  Or you can chuck it into the collector or ParsedXML tracker.
> 
> For debugging purporses, I'd suggest simplifying by supplying an XML
> string directly, rather than from an external server.

I've replaced ParsedXML's ExpatBuilder with the pyXML package and I used the
sax parser without modifing the code (except for the import), this is the
refcounts results.

Class  April 12, 2001 11:55 am  April 12, 2001 12:00

xml.dom.NodeList.NodeList           2103  3678  +1575  
xml.dom.Text.Text                   1263  2208  +945  
xml.dom.NamedNodeMap.NamedNodeMap    507   885  +378  
xml.dom.Element.Element              499   871  +372  
xml.dom.Attr.Attr                    339   591  +252  
xml.dom.DocumentType.DocumentType      7    10  +3  
xml.dom.Document.Document              7    10  +3

xml.dom.Document.Document increments by 1 each time an xml is procesed (I
tried to add a del doc at the end but it's ineffective).

So the problem is not ParsedXML but Zope itself or my bad methods.

Regards,

- mn