Hola Maximiliano
Greetings
Volker Groll
Am Donnerstag, 21. Juli 2016, 13:59:42 CEST schrieb Maximiliano Curia:
> ¡Hola Volker!
>
> El 2016-07-21 a las 11:03 +0200, Volker Groll escribió:
> > Holla Maximiliano,
> >
> >> It seems to me that the bug is the same as:
> >> https://bugs.kde.org/show_bug.cgi?id=358696
> >
> > Hm, not sure, but there is mentioned the crash occurs, when the
> > birthday calendar is enabled, mine is disabled.
>
> Well, it seems to be the same underlying issue.
>
> >> Could you please provide additional information in the bug upstream?
> >> Currently it seems to require confirmation that this bug is only
> >> reproduceable in 32 bits systems.
> >
> > I have local address books and two on an owncloud server.
> > Changing the owncloud resource to offline with akonadiconsole
> > changes nothing (crash persists). I get also a crash from akonadiconsole
> > when I try to switch to Browser tab.
> >
> >> Also, a full backtrace would be useful, for that you would need to
> >> install
> >> the kaddressbook-dbgsym package (and probably the dbg/dbgsym packages
> >> from
> >> a couple of kaddressbook dependencies) from the debug archive (that needs
> >> to be added in the sources.list as: deb
> >> http://deb.debian.org/debian-debug
> >> unstable-debug main)
> >
> > Most threads are in g_mutex_unlock (), __kernel_vsyscall ()
> > Installing libglib2.0-0-dbg will give only 0xffffffff in output,
> > so I purged this debug package again.
>
> If possible please keep the full backtrace.
>
> > The output of the last thread is:
> > Thread 1 (Thread 0xed0a6ac0 (LWP 5109)):
> > [KCrash Handler]
>
> Shouldn't there be information of 6 more frames?
>
> > #7 0xf4ff1cb6 in Akonadi::Tag::isValid (this=0xff8b137c) at
> > /build/akonadi- uzW5jU/akonadi-16.04.3/src/core/tag.cpp:239
> > #8 0xf5097577 in Akonadi::TagModel::data (this=0x9b1c098, index=...,
> > role=258) at /build/akonadi-uzW5jU/akonadi-16.04.3/src/core/models/
> > tagmodel.cpp:95
>
> Looking at the code, my bet is that changing akonadi's
> src/core/models/tagmodel_p.cpp the function Tag
> TagModelPrivate::tagForIndex line 106:
> return children.at(index.row());
> for:
> return children.value(index.row());
>
> would make the issue to go away (by creating a new instance whenever an
> invalid index is accessed), but the real issue is to be hitting this
> function with invalid index values (the previous fix only takes care of
> negative values, btw). That would need someone that actually understands the
> code.
>
> Please, if possible, add the information to the upstream bug with the full
> backtrace.
>
> Happy hacking,
I filed a bugreport to kde: https://bugs.kde.org/show_bug.cgi?id=365944
It looks like it is a duplicate to https://bugs.kde.org/show_bug.cgi?id=357228
but drkonqi won't accept my duplicate hint.
Thank you
volker
