[Webkit-unassigned] [Bug 118307] [ATK] Leak: AtkAttributeSet* should be freed

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=118307





--- Comment #8 from Christophe Dumez <dchris at gmail.com>  2013-07-03 02:52:59 PST ---
(In reply to comment #7)
> (In reply to comment #6)
> > (In reply to comment #5)
> > > [...]
> > > > Maybe we should stick with atk_attribute_set_free() then?
> > > > unless someone has a better proposal.
> > > 
> > > I would propose to investigate why this is a problem in EFL and not in GTK,
> > > and try to get it sorted out, because using a GOwnPtr specialization is, in
> > > my opinion, clearer and cleaner than the current approach we have now.
> > 
> > Brian just explained to me that he jsut updated PlatformGtk.cmake but not GNUmakefile.list.am, which I believe is what the EWS is using now.
> > 
> > I guess that explains why it was not failing in the GTK bot, which is probably using the GSList specialization after all...
> 
> Not quite - it doesn't fail when using the autotools make system on Gtk.  Perhaps the EFL cmake build system is trying to link the 2 object files into the same shared object and Gtk is putting them in their own .so files?

I don't think it matters. Since a AtkAttributeSet is the same thing as a GSList, how can you be sure the right GOwnPtr would be used? There is no way the differentiate the 2 types since they are the same type. You cannot do template specialization here. Even if it compiles on GTK, the behavior would be undetermined.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.