On Tuesday 03 Jan 2012 17:12:54 Martin Kho wrote:
> On Sunday 01 January 2012 18:14:15 Colin J Thomson wrote:
> > On Sunday 01 Jan 2012 11:54:55 Rex Dieter wrote:
> > > Colin J Thomson wrote:
> > > > On Sunday 01 Jan 2012 00:02:57 Colin J Thomson wrote:
> > > >> On Sunday 01 Jan 2012 00:39:12 Kevin Kofler wrote:
> > > >> > Colin J Thomson wrote:
> > > >> > > And Kmail2 still crashes, however in #289693 it is
marked as
> > > >> > > fixed
> > > >> > > in
> > > >> > > 4.8
> > > >> >
> > > >> > I already reopened the bug. All the fixes so far don't
work.
> > > >>
> > > >> If another backtrace is needed after the updates applied on this
box
> > > >> let
> > > >> me know.
> > > >
> > > > I have just found another report + fix possibly related to #289693
> > > >
> > > >
https://bugs.kde.org/show_bug.cgi?id=289831
> > >
> > > That's the patch we applied to kdepimlibs-4.7.95-2 , doesn't seem
to
> > > help
> > > any (for me).
> >
> > I was not sure if it was the same patch, fyi it does not help here either.
> >
> > > It's fairly obvious by now the qList is getting corrupted somehow, so
> > > efforts to change how it's items are free'd essentially don't
help much
> > > at
> > > this point.
> >
> > OK, I'll keep an eye on kde.bugs, maybe a comment and new back trace could
> > be added to #289831 (which is marked as fixed) saying that the crash still
> > happens.
>
> Today for me too, Kmail2 crashed as described above. The crash happened
> after I had removed some email messages. Kmail2 got really lost. Header
> info and message bodies were mixed up, sometimes kmail2 hung up or crashed.
> So I think the problem - as Kevin already said - is not in the destroy
> methode (c.q. qDeleteAll), but is somewhere in the manipulating
> (creation/removing) methodes of the qList.
One thing I have been trying to avoid the crash is closing Kontact is via the
system activity and ending the process, not ideal but seems to work.
Colin
Hi,
May be commit [1] causes the crashes in kdepimlibs-4.7.95-1(2). The next
commit [2] reverts commit 1, but isn't in 4.7.95-1(2). Allen Winter says
about these commits:
"Revert "Add a bunch of validity checks to avoid crashes."
This reverts commit baa76292fbbe4918a630203269e0822905db3503.
I think I might have broken some things with this commit
so I'm reverting it so I can test further."
Martin Kho
[1]