Re: [evolution-patches] patch for #61273 (changes in e-source-group.c)
- From: Sivaiah N <snallagatla novell com>
- To: JP Rosevear <jpr novell com>
- Cc: Evolution Patches <evolution-patches lists ximian com>
- Subject: Re: [evolution-patches] patch for #61273 (changes in e-source-group.c)
- Date: Fri, 16 Jul 2004 20:18:49 +0530
i have committed the fix and closed the bug
Thanks,
Siva
On Fri, 2004-07-16 at 20:03, JP Rosevear wrote:
> On Fri, 2004-07-16 at 12:58 +0530, Sivaiah N wrote:
> > Hi,
> > The attached patch removes the signal handler registered (for CHANGED
> > singal of ESource) on Esource object by ESourceGroup when the ESource
> > group object is removed from ESourceGroup object.
> > There is a problem if we do not do this, when somebody removes a
> > ESource object from group but does not unref the ESource object. This
> > ESource Object gets a weak_ref notification (group_weak_notify) when
> > EsourceGroup object is being disposed. group_weak_notify emits CHANGED
> > signal on source, resulting in call to source_changed_callback (in
> > e-source-group.c) which emits signal on group which is just being
> > disposed thus resulting in crash. I could not find out though where in
> > evolution code ESource object is removed from group but not cleaned up,
> > but this patch solves the crash.
> >
> > Hope my understanding is correct :)
> >
> > Interestingly this crash happens with snaps but not with evolution
> > which i have built from sources. The different versions of GLib is
> > making the difference i guess. Version in Gobject.pc from snaps is 2.2.3
> > and what i have from sources is 2.4.0
>
> This looks correct to me, I think we should probably be more generic
> when disconnecting from the source though (in the other spots, not just
> in your changes), disconnecting all signals by data. However the
> current patch is consistent with the existing source, so "approved".
>
> -JP
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]