Bug 442230

Summary: korganizer crashes on window activation
Product: [openSUSE] openSUSE 11.1 Reporter: Kevin Wolf <kwolf>
Component: KDE3Assignee: Will Stephenson <wstephenson>
Status: RESOLVED UPSTREAM QA Contact: E-mail List <qa-bugs>
Severity: Critical    
Priority: P2 - High CC: meissner
Version: FinalFlags: coolo: SHIP_STOPPER-
Target Milestone: ---   
Hardware: x86-64   
OS: Other   
URL: https://bugs.kde.org/show_bug.cgi?id=171787
Whiteboard:
Found By: Development Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: valgrind log file
backtrace from debug version

Description Kevin Wolf 2008-11-06 11:02:39 UTC
I'm using KOrganizer on KDE 3.5 with openSUSE 11.1 Beta 4. When I open korganizer, switch to a different application, wait a while (this seems to be important, it doesn't work immediately) and then switch back to the KOrganizer window, KOrganizer crashes (probably during the repaint).

Unfortunately, I don't seem to be able to get much useful information. As soon as I install kdepim3-debuginfo I don't get a stacktrace at all. So maybe the issue is trivial and this is enough for you, otherwise please explain how I can provide more useful information.


[KCrash handler]
#5  0x0000000000000000 in ?? ()
#6  0x00007ff8a8735889 in KOAgendaItem::paintEvent(QPaintEvent*) ()
   from /opt/kde3/lib64/libkorganizer.so.1
#7  0x00007ff8a66ce425 in QWidget::event(QEvent*) ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#8  0x00007ff8a663a6c5 in QApplication::internalNotify(QObject*, QEvent*) ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#9  0x00007ff8a663b4aa in QApplication::notify(QObject*, QEvent*) ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#10 0x00007ff8a71d3a42 in KApplication::notify(QObject*, QEvent*) ()
   from /opt/kde3/lib64/libkdecore.so.4
#11 0x00007ff8a65ce38d in QETWidget::translatePaintEvent(_XEvent const*) ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#12 0x00007ff8a65d7bf8 in QApplication::x11ProcessEvent(_XEvent*) ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#13 0x00007ff8a65ea89a in QEventLoop::processEvents(unsigned int) ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#14 0x00007ff8a6651259 in QEventLoop::enterLoop() ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#15 0x00007ff8a66510e2 in QEventLoop::exec() ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#16 0x000000000040ab3c in ?? ()
#17 0x00007ff8a5cd6586 in __libc_start_main () from /lib64/libc.so.6
#18 0x000000000040a8e9 in ?? ()
#19 0x00007fffb0d0af38 in ?? ()
#20 0x000000000000001c in ?? ()
#21 0x0000000000000001 in ?? ()
#22 0x00007fffb0d0c104 in ?? ()
#23 0x0000000000000000 in ?? ()
Comment 1 Kevin Wolf 2008-11-06 11:16:18 UTC
Ah, there seems to be a newer package today. The new one gives me a stacktrace with more useful information:


Überprüfung der Systemkonfiguration beim Start deaktiviert.

 [?1034h[Thread debugging using libthread_db enabled]
[KCrash handler]
#5  0x00007fe63d45e886 in KOAgendaItem::paintEvent (this=0x137c240, 
    ev=<value optimized out>)
    at /usr/src/debug/kdepim-3.5.10/korganizer/koagendaitem.cpp:753
#6  0x00007fe63b3f74e5 in QWidget::event (this=0x137c240, e=0x7fff45a34df0)
    at kernel/qwidget.cpp:4808
#7  0x00007fe63b363745 in QApplication::internalNotify (
    this=<value optimized out>, receiver=0x137c240, e=0x7fff45a34df0)
    at kernel/qapplication.cpp:2639
#8  0x00007fe63b36452a in QApplication::notify (this=0x7fff45a35450, 
    receiver=0x137c240, e=0x7fff45a34df0) at kernel/qapplication.cpp:2527
#9  0x00007fe63befca42 in KApplication::notify(QObject*, QEvent*) ()
   from /opt/kde3/lib64/libkdecore.so.4
#10 0x00007fe63b2f738d in QETWidget::translatePaintEvent (this=0x137c240, 
    event=<value optimized out>) at kernel/qapplication.h:526
#11 0x00007fe63b300bf8 in QApplication::x11ProcessEvent (this=0x7fff45a35450, 
    event=0x7fff45a35270) at kernel/qapplication_x11.cpp:3545
#12 0x00007fe63b31389a in QEventLoop::processEvents (this=0x680570, flags=4)
    at kernel/qeventloop_x11.cpp:195
#13 0x00007fe63b37a2d9 in QEventLoop::enterLoop (this=0x680570)
    at kernel/qeventloop.cpp:201
#14 0x00007fe63b37a162 in QEventLoop::exec (this=0x7fff45a34700)
    at kernel/qeventloop.cpp:148
#15 0x000000000040ab3c in main (argc=<value optimized out>, 
    argv=<value optimized out>)
    at /usr/src/debug/kdepim-3.5.10/korganizer/main.cpp:58
Comment 2 Stephan Kulow 2008-11-26 06:40:50 UTC
try valgrind, the backtrace doesn't look helpful
Comment 3 Kevin Wolf 2008-11-26 09:25:39 UTC
Created attachment 255591 [details]
valgrind log file

Attaching the log file of a korganizer run with valgrind.
Comment 4 Will Stephenson 2008-11-27 19:38:28 UTC
Upstream bugs:
https://bugs.kde.org/show_bug.cgi?id=84330  - on removing file resource
https://bugs.kde.org/show_bug.cgi?id=104362 - imap resource 2 simultaneous clients
https://bugs.kde.org/show_bug.cgi?id=115675 - on switching kontact component
https://bugs.kde.org/show_bug.cgi?id=136171 - on nfs with simultaneous other user!
https://bugs.kde.org/show_bug.cgi?id=137303 - nonspecific crashes when using kontact
https://bugs.kde.org/show_bug.cgi?id=170273 - using remote calendar
https://bugs.kde.org/show_bug.cgi?id=171651 - using google calendar proxy
https://bugs.kde.org/show_bug.cgi?id=171787 - on scrolling, "reproducible"
https://bugs.kde.org/show_bug.cgi?id=173225 - random crash on webdav

"That's a typical backtrace of a crash, when the calendar has been reloaded by libkcal in the background, but KOrganizer has still pointers to the old, now deleted events." - https://bugs.kde.org/show_bug.cgi?id=128659

Sideways bugs:
https://bugzilla.redhat.com/show_bug.cgi?id=465211 'reproducible', http calendar
https://www.intevation.de/roundup/kolab/issue3065 - kolab groupware server

Observations:
Manifests with various resources
Several of the above have been closed optimistically
Sometimes it gets blamed on recent point releases but my hunch is that the bug has been there for years

Conclusions:
1) It's an upstream bug but severe enough for us to look at
2) Seems to have a variety of causes
3) It's a design flaw in KOrganizer/KResources
4) It will be fixed by Akonadi since the resource does not pokle the app's internal data structures
5) We still have to deal with it for all our SLEs
Comment 5 Kevin Wolf 2008-11-28 08:30:22 UTC
I guess the stack trace isn't that helpful then...

Meanwhile I observed the same crash (or better: the same stack trace) with having KOrganizer active which could be related to the original crash. This one happens if I select a Groupwise calendar entry so that its details are displayed on the left (it doesn't seem to happen with an entry from a local ical file, and these are the only two types I use). After a few seconds the details disappear (which makes it pretty difficult to dial the phone number in there, btw...) and it shows "No entry selected". The next click into the calendar causes the crash then.
Comment 6 Will Stephenson 2008-12-05 13:31:39 UTC
*** Bug 441543 has been marked as a duplicate of this bug. ***
Comment 8 Rasmus Plewe 2008-12-17 09:07:07 UTC
Created attachment 260511 [details]
backtrace from debug version

FWIW, here's a backtrace from a crash with the debug version kdepim3-debuginfo-3.5.10-8.2
This debuginfo version seems to be more stable, crashing every hour instead of every 10 minutes.
Comment 10 Will Stephenson 2009-09-14 19:30:41 UTC
Upstream's design flaw, can't do anything about it