Skip Menu |

This queue is for tickets about the POE-Filter-XML CPAN distribution.

Report information
The Basics
Id: 32401
Status: resolved
Priority: 0/
Queue: POE-Filter-XML
People
Owner: nperez [...] cpan.org
Requestors: ewaters [...] uarc.com
Cc:
AdminCc:
Bug Information
Severity: Important
Broken in: 0.33
Fixed in: (no value)

History Show all quoted text
  Thu Jan 17 17:12:31 2008 ewaters [...] uarc.com - Ticket created
Subject: Memory leak still present
While the code changes done recently to avoid the XML::SAX::Expat::Incremental memory leak upon DESTROY() appear to deal with that issue, the parser object is destroyed and recreated upon reset(), incurring a memory leak. In code that calls reset() frequently, this bloats the program. Simply applying the weaken code in DESTROY() to reset() appears to fix the issue in my tests.
  Fri Jan 18 08:26:14 2008 nperez [...] cpan.org - Correspondence added
From: nperez [...] cpan.org
Alrighty, that is a good catch. Definitely need to do the release() dance on the parser object. What a shitty piece of code. I'm this |> <| tempted to rip it out and replace it with something else. I'll see about getting a release out in a bit. Might take awhile though. Thanks :) On Thu Jan 17 17:12:31 2008, ewaters@uarc.com wrote: Show quoted text
> While the code changes done recently to avoid the > XML::SAX::Expat::Incremental memory leak upon DESTROY() appear to deal > with that issue, the parser object is destroyed and recreated upon > reset(), incurring a memory leak. In code that calls reset() > frequently, this bloats the program. Simply applying the weaken code in > DESTROY() to reset() appears to fix the issue in my tests.
  Fri Jan 18 08:26:16 2008 The RT System itself - Status changed from 'new' to 'open'
  Fri Jan 18 08:26:34 2008 nperez [...] cpan.org - Taken
  Fri Jan 18 08:42:47 2008 nperez [...] cpan.org - Correspondence added
Okay, I have this fixed and in svn on googlecode: http://code.google.com/p/pfx/ I went ahead and broke out the parser frobbing into its own method and call it from destroy and from reset On Fri Jan 18 08:26:14 2008, NPEREZ wrote: Show quoted text
> Alrighty, that is a good catch. Definitely need to do the release() > dance on the parser object. What a shitty piece of code. I'm this |> <| > tempted to rip it out and replace it with something else. > > I'll see about getting a release out in a bit. Might take awhile though. > > Thanks :) > > On Thu Jan 17 17:12:31 2008, ewaters@uarc.com wrote:
> > While the code changes done recently to avoid the > > XML::SAX::Expat::Incremental memory leak upon DESTROY() appear to deal > > with that issue, the parser object is destroyed and recreated upon > > reset(), incurring a memory leak. In code that calls reset() > > frequently, this bloats the program. Simply applying the weaken code in > > DESTROY() to reset() appears to fix the issue in my tests.
>
  Fri Jan 18 08:42:51 2008 nperez [...] cpan.org - Status changed from 'open' to 'resolved'