Re: some debugging help needed

[Fred Kiefer]

-------- Original-Nachricht --------
> Datum: Thu, 11 Feb 2010 09:46:50 +0100
> Von: Sebastian Reitenbach <sebastia@l00-bugdead-prods.de>
> An: Fred Kiefer <fredkiefer@gmx.de>
> CC: "Richard Frith-Macdonald" <richard@tiptree.demon.co.uk>, GNUstep general 
> <discuss-gnustep@gnu.org>
> Betreff: Re: some debugging help needed

> Hi,
> On Tuesday 09 February 2010 10:46:15 pm Fred Kiefer wrote:
> > You session object gets release by an autorelease pool. This information
> > wont help you track down your problem. From the code sniplets I would
> > guess that the session should be kept alive by the sessionStore. You
> > could inspect that code, whether it retains the session correctly. If it
> > does then another trick will be needed. You can define retain and
> > release method on session (just let them call the super implementation)
> > and set breakpoints in both methods. This may result in too much
> > information, but could lead to the clue on who is releasing this object.
> >
> > Hope this helps
> That helped. Thanks a lot. After some hours of debugging and wading
> through 
> the code, I found the place where the session gets stored in the 
> WOServerSessionStore. There was an if else statement, where the session
> was 
> only in the if clause retained. Doing the same in the else clause fixed my
> problem.
> 
> However, I did not really understood the things you proposed in the second
> half of your answer. I should add a retain and release method in the 
> OGoSession class, and just call [super retain|release] in them and then
> set a 
> breakpoint on those? 

Yes, this summons up what I proposed, but it wouldn't have helped you much in 
the case of a missing retain. But then, what apart from code inspection would 
have?
Great that you were able to resolve this issue yourself.
Fred
-- 
Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5 -
sicherer, schneller und einfacher! http://portal.gmx.net/de/go/chbrowser