Bug 91186 - HotSync causes crash of kpilotDaemon
Summary: HotSync causes crash of kpilotDaemon
Status: RESOLVED WORKSFORME
Alias: None
Product: kpilot
Classification: Applications
Component: General (show other bugs)
Version: 4.4.4
Platform: openSUSE Linux
: NOR crash
Target Milestone: ---
Assignee: groot
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2004-10-12 16:16 UTC by Christoph Ludwig
Modified: 2007-03-20 04:31 UTC (History)
3 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments
backtrace generated by the KDE crash handler (2.38 KB, text/plain)
2004-10-12 16:18 UTC, Christoph Ludwig
Details
kpilot's log up to the crash (1.75 KB, text/plain)
2004-10-12 16:18 UTC, Christoph Ludwig
Details
backtrace of kpilotDaemon (1.35 KB, text/plain)
2005-02-23 23:42 UTC, Nicola Di Nisio
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Christoph Ludwig 2004-10-12 16:16:32 UTC
Version:           4.4.4 (using KDE KDE 3.3.0)
Installed from:    SuSE RPMs
Compiler:          gcc 3.3.1 
OS:                Linux

Today kpilotDaemon started crashing whenever I do a HotSync. I didn't change the kpilot configuration at all and the only data that needs sync'ing is some changes in the calendar. I didn't experience similar problems since I installed KDE 3.3 some weeks ago. My PDA is a Palm Tungsten E.

According to both the kpilot log and the log in the Palm the last conduit started was the todo-conduit.

I am not sure what information I can provide for you to reproduce the problem given that until today I didn't observe the problem myself and I didn't change the software or my configuration... I am going to attach the backtrace reported by the KDE crash handler. Let me know if there is anything else I can provide.

Regards

Christoph
Comment 1 Christoph Ludwig 2004-10-12 16:18:02 UTC
Created attachment 7845 [details]
backtrace generated by the KDE crash handler
Comment 2 Christoph Ludwig 2004-10-12 16:18:50 UTC
Created attachment 7846 [details]
kpilot's log up to the crash
Comment 3 Christoph Ludwig 2004-10-15 14:52:16 UTC
Today I installed SuSE's KDE 3.3.1 rpms with kpilot 4.4.5. 

kpilotDaemon does not crash any more, so the problem seems to be fixed.

Christoph
Comment 4 Christoph Ludwig 2004-11-03 08:51:22 UTC
Today the crash happened again with kpilot 4.4.5. This time the KDE crash handler was unable to generate a backtrace due to a corrupted stack.

As before, I didn't change my configuration at all. AFAICT, the only changes that needed updating were some entries in the calendar on the Palm. The last entry in the log says the todo-conduit was started.

Please tell me if there is anything I can do to help you track down the cause of this crash.

Christoph
Comment 5 groot 2004-11-03 10:32:34 UTC
On Wednesday 03 November 2004 08:51, Christoph Ludwig wrote:
> Please tell me if there is anything I can do to help you track down the
> cause of this crash.

Can you try it with a fresh user? Can you move away your existing calendar and 
todo files and try again from there? Since it appears to be crashing in 
libkcal, it might be a weirdly-corrupt entry in one of those files. Don't 
forget to make backups and to remember how to move the files back.

Comment 6 Christoph Ludwig 2004-11-04 13:43:39 UTC
On Wed, Nov 03, 2004 at 09:32:35AM -0000, groot@kde.org wrote:
> Can you try it with a fresh user? 

I created a new user account on my system and HotSynced my Palm from
there. No crash, no problem...

> Can you move away your existing calendar and 
> todo files and try again from there? Since it appears to be crashing in 
> libkcal, it might be a weirdly-corrupt entry in one of those files. Don't 
> forget to make backups and to remember how to move the files back.

Then I intended to perform this test. But even before I touched
the calendar and todo files, kpilotDaemon did not crash anymore! For
some reason that I don't understand the previous succesful sync solved
the problem (for now). 

If the crash happens again, is it possible to log the communication
between kpilotDeamon and the PDA?

Regards

Christoph

Comment 7 Nicola Di Nisio 2005-02-23 23:42:27 UTC
Created attachment 9799 [details]
backtrace of kpilotDaemon
Comment 8 Nicola Di Nisio 2005-02-23 23:44:32 UTC
I confirm this problem.
I have a Zire 72 and SuSE 9.2 Pro. I use the original KDE shipped by SuSE, the 3.3.0.
I have attached my backtrace of the kpilotDaemon, see comment #7.
Comment 9 Christoph Ludwig 2005-02-24 09:00:35 UTC
Here, too, kpilotDaemon still crashes from time to time. But I am unable to reproduce the problem: When kpilotDaemon crashes then it also crashes after a restart. But if I try it again some days later then usually the sync succeeds.
(Note: I didn't touch the calendar and todo files between the crash and the successful try.)
Comment 10 Nicola Di Nisio 2005-02-24 09:05:14 UTC
In my case the crash is sistematic and happens always when the progress bar is at 98%. By the way, I live in Darmstadt, probably there is a special magnetic here field that disturbs  kpilotDaemon ;)
Comment 11 Mark 2005-06-04 17:43:58 UTC
Can confirm this behaviour here. KPilot 4.5.2

KPilotDaemon log up to crash:

12:39:11 Using encoding ISO 8859-15 on the handheld.
12:39:11 [File Installer]
12:39:11 No Files to install
12:39:11 [Conduit time_conduit]
12:39:11 Setting the clock on the handheld
12:39:11 [Conduit vcal-conduit]
12:39:11 Using non-local time zone: America/Halifax
12:39:11 Syncing with file "/home/vmark/.kde/share/apps/korganizer/std.ics"
12:39:11 Copying records to PC ...
12:39:11 Copying records to Pilot ...


On Palm HotSync Log shows:
[Conduit vcal-conduit] Syncing with file "/home/vmark/.kde/share/apps/korganizer/std/ics"

as the last entry.
Comment 12 Aurelien Bompard 2005-07-31 23:56:59 UTC
Same crash here, with KPilot 4.5.2 :

KPilot log :
---------
Version: KPilot 4.5.2 (kind)
Version: pilot-link 0.12.0-pre3
Version: KDE 3.4.2-1.0.fc4.kde
Version: Qt 3.3.4

Journal de la synchronisation

23:44:38 Démarrage du démon KPilot...
23:44:38 L'état du démon est « stoppé »
23:44:38 Le périphérique « /dev/pilot » n'existe pas. Il s'agit probablement d'un périphérique USB et apparaîtra lors d'une synchronisation.
23:52:00 Liaison vers le périphérique prête.
23:52:00 Vérification de l'ordinateur...
23:52:00 KPilot 4.5.2 (kind) Démarrage de la synchronisation...

23:52:00 Utilisation du jeu de caractères ISO 8859-15 sur le périphérique.
23:52:00 [Canal vcal-conduit]
23:52:00 Utilisation d'un fuseau horaire non local : Europe/Paris
23:52:00 Synchronisation avec le fichier « /home/gauret/systeme/agenda/agenda.ics »
23:52:00 Copie des enregistrements vers le PC...
23:52:00 Copie des enregistrements vers le Palm...
23:52:00 [Canal abbrowser_conduit]
---------

And in the KCrash dialog, the backtrace is :
[KCrash handler]
#4  0x00c42bb5 in memcpy () from /lib/libc.so.6
#5  0x0051ebc9 in PilotAddress::pack_ () from /usr/lib/libkpilot.so.0
#6  0x0051efa2 in PilotAppCategory::pack () from /usr/lib/libkpilot.so.0
#7  0x00667ec8 in AbbrowserConduit::_savePalmAddr ()
   from /usr/lib/kde3/conduit_address.so
#8  0x0067b6c8 in AbbrowserConduit::_copyToHH ()
   from /usr/lib/kde3/conduit_address.so
#9  0x0067f099 in AbbrowserConduit::syncAddressee ()
   from /usr/lib/kde3/conduit_address.so
#10 0x0067f8c2 in AbbrowserConduit::slotPCRecToPalm ()
   from /usr/lib/kde3/conduit_address.so
#11 0x0068158a in AbbrowserConduit::qt_invoke ()
   from /usr/lib/kde3/conduit_address.so
#12 0x0574c574 in QObject::activate_signal ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#13 0x05ac50c6 in QSignal::signal () from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#14 0x05769df3 in QSignal::activate () from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#15 0x057726d5 in QSingleShotTimer::event ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#16 0x056e77dd in QApplication::internalNotify ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#17 0x056e815c in QApplication::notify ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#18 0x05e731c7 in KApplication::notify () from /usr/lib/libkdecore.so.4
#19 0x056dc082 in QEventLoop::activateTimers ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#20 0x0569009a in QEventLoop::processEvents ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#21 0x056fff02 in QEventLoop::enterLoop ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#22 0x056ffe26 in QEventLoop::exec () from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#23 0x056e71a9 in QApplication::exec () from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#24 0x080645ed in ?? ()
#25 0x00bebde6 in __libc_start_main () from /lib/libc.so.6
#26 0x08053c81 in ?? ()
---------

I'm using Fedora Core 4 with KDE rpms from kde-redhat.sf.net
Comment 13 groot 2005-08-01 00:03:21 UTC
On Sunday 31 July 2005 23:57, Aurelien Bompard wrote:
> Version: KPilot 4.5.2 (kind)
> Version: pilot-link 0.12.0-pre3
> Version: KDE 3.4.2-1.0.fc4.kde


Just don't. Pilot-link 0.12-pre3 (which FC4 shipped despite the author's 
explicit instructions not to package it) is just broken with KPilot. Revert 
to an earlier version, or build everything -- pilot link and kpilot -- from 
source.
Comment 14 Jason 'vanRijn' Kasper 2005-08-20 06:57:37 UTC
Can someone please either verify or state that they can't verify this bug with a recent version of kpilot (4.5.3 is the latest, currently), the latest version of KDE (3.4.2), and the latest stable pilot-link release (0.11.8)? 

Thanks!
Comment 15 Christoph Ludwig 2005-08-20 12:06:00 UTC
On Sat, Aug 20, 2005 at 04:57:39AM -0000, Jason 'vanRijn' Kasper wrote:
> Can someone please either verify or state that they can't verify this bug
> with a recent version of kpilot (4.5.3 is the latest, currently), the latest
> version of KDE (3.4.2), and the latest stable pilot-link release (0.11.8)?  


I observed a kpilot crash with KDE 3.3.? three weeks ago. That made me upgrade
to KDE 3.4.2 (SuSE rpms). After the upgrade, I could hotsync again, but in the
beginning kpilot insisted on removing a new entry in the addressbook on the
PDA. I assume that was due to a corrupted address database on the notebook
because after two full syncs kpilot left the new entry alone.

I did not observe a kpilot crash since the upgrade but I don't think that
implies the problem is gone: Even before, the crash happened sporadic, and I
was not able to trigger it on purpose. (*If* kpilot crashed, then it always
involved a corrupted address database and it would work again only if I synced
from another account or upgraded kpilot.)

Regards

Christoph
Comment 16 Alan Hearnshaw 2006-11-01 07:03:56 UTC
KPilot is now crashing for me. The only thing that has changed is that I have added quite a few new contact records on the palm and it will now crash every time I try to hotsync.
Stack Trace:

(no debugging symbols found)
Using host libthread_db library "/lib/libthread_db.so.1".
(no debugging symbols found)
`shared object read from target memory' has disappeared; keeping its symbols.
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
[Thread debugging using libthread_db enabled]
[New Thread -1209153872 (LWP 15856)]
[New Thread -1213494368 (LWP 16388)]
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
[KCrash handler]
#6  0x0054a94a in KABC::AddressBook::clear () from /usr/lib/libkabc.so.1
#7  0x00571eff in KABC::AddressBook::load () from /usr/lib/libkabc.so.1
#8  0x006d9ee7 in AbbrowserConduit::_loadAddressBook ()
   from /usr/lib/kde3/conduit_address.so
#9  0x006da96b in AbbrowserConduit::exec ()
   from /usr/lib/kde3/conduit_address.so
#10 0x004a5ae5 in SyncAction::execConduit () from /usr/lib/libkpilot.so.0
#11 0x004a5c38 in SyncAction::qt_invoke () from /usr/lib/libkpilot.so.0
#12 0x004a2b2b in ConduitAction::qt_invoke () from /usr/lib/libkpilot.so.0
#13 0x006e40e3 in AbbrowserConduit::qt_invoke ()
   from /usr/lib/kde3/conduit_address.so
#14 0x04aed70a in QObject::activate_signal ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#15 0x04e7752e in QSignal::signal () from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#16 0x04b0bdd7 in QSignal::activate () from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#17 0x04b14983 in QSingleShotTimer::event ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#18 0x04a84beb in QApplication::internalNotify ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#19 0x04a86039 in QApplication::notify ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#20 0x008a7ae3 in KApplication::notify () from /usr/lib/libkdecore.so.4
#21 0x04a789f2 in QEventLoop::activateTimers ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#22 0x04a2d05f in QEventLoop::processEvents ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#23 0x04a9e090 in QEventLoop::enterLoop ()
   from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#24 0x04a9df56 in QEventLoop::exec () from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#25 0x04a846ff in QApplication::exec () from /usr/lib/qt-3.3/lib/libqt-mt.so.3
#26 0x080586fa in ?? ()
#27 0x00c467e4 in __libc_start_main () from /lib/libc.so.6
#28 0x08053c01 in ?? ()

Versions:
Qt: 3.3.6
KDE: 3.5.3-0.3.fc5 Fedora-Core
KPilot: 4.6.0 (blivit)
Comment 17 Aurelien Bompard 2006-11-01 07:39:18 UTC
Please run "yum --enablerepo=core-debuginfo install kdepim-debuginfo" and reproduce the crash, you'll get a better stack trace.
Comment 18 subscryer 2006-12-17 03:53:41 UTC
For me it always crashes on the "web" item, the last log entry in kpilot is "backing up blazer".
I can't post a backtrace now, will do if possible.
Comment 19 groot 2006-12-17 22:19:03 UTC
On Sunday 17 December 2006 03:53, subscryer@gmail.com wrote:
> ------- For me it always crashes on the "web" item, the last log entry in
> kpilot is "backing up blazer". I can't post a backtrace now, will do if
> possible.


Easiest thing to do is to add the web item to the skip list (in KPilot 
configuration dialog) of databases that are not backed up. With pilot-link 
0.12 and a recent KPilot, this should no longer be necessary.
Comment 20 Jason 'vanRijn' Kasper 2007-03-20 04:31:51 UTC
Okay... this bug has been out there for a loooooooong time.  It has a variety of problems in it.  The latest one from Alan (Comment #16) looks like it might be caused by a non-local calendar resource(?) which is a well-known and difficult to fix problem documented in a different bug report.

Anyway, I'm going to close this bug, since it has become unruly and doesn't help point to any single cause.  KPilot has changed significantly in the last 2 years.   If anyone CC'd is still seeing issues with instability and you're running the latest KDE (3.5.6) and latest kpilot, please open a new bug report with a more specific summary and very detailed debugging information and we'll look into it.

Thanks!!!