Bug 100391

Summary:	crash when deleting a task
Product:	[Applications] ktimetracker	Reporter:	Martin Honermeyer <maze>
Component:	general	Assignee:	Mark Bucciarelli <mark>
Status:	RESOLVED FIXED
Severity:	crash	CC:	dev+kde, rscrawford
Priority:	NOR
Version:	unspecified
Target Milestone:	---
Platform:	unspecified
OS:	Linux
Latest Commit:		Version Fixed In:

Description Martin Honermeyer 2005-02-27 19:40:03 UTC

Version:           1.5.0 (using KDE 3.3.92 (beta2), Gentoo)
Compiler:          gcc version 3.4.3 (Gentoo Linux 3.4.3, ssp-3.4.3-0, pie-8.7.6.6)
OS:                Linux (x86_64) release 2.6.10-gentoo-r7

When I click on the delete button to delete a task, karm crashes. When I start it again, I can see that the task has been deleted though. Output from strace:

...
close(20)                               = -1 EBADF (Bad file descriptor)
close(21)                               = -1 EBADF (Bad file descriptor)
close(22)                               = -1 EBADF (Bad file descriptor)
close(23)                               = -1 EBADF (Bad file descriptor)
close(24)                               = -1 EBADF (Bad file descriptor)
close(25)                               = -1 EBADF (Bad file descriptor)
close(26)                               = -1 EBADF (Bad file descriptor)
close(27)                               = -1 EBADF (Bad file descriptor)
close(28)                               = -1 EBADF (Bad file descriptor)
...
--- SIGCHLD (Child exited) @ 0 (0) ---
write(8, "\0", 1)                       = -1 EBADF (Bad file descriptor)
rt_sigreturn(0x8)                       = 12651
exit_group(253)                         = ?


=> many strange close() calls!


Martin

Comment 1 Mark Bucciarelli 2005-02-27 20:10:43 UTC

yikes.  thanks for your report!

Comment 2 Mark Bucciarelli 2005-03-20 05:01:05 UTC

I cannot duplicate this with 3_4_BRANCH.

Would it be possible for you to compile with --enable-debug=full and generate a backtrace for me?

If not, please provide any information you think may be relevant so I can duplicate the behavior.

Comment 3 Martin Honermeyer 2005-04-28 21:47:07 UTC

Recompiled! Sorry, how do I generate a backtrace?

Comment 4 Mark Bucciarelli 2005-04-28 21:59:30 UTC

Great, thanks for taking that on.

Just make KArm crash, and KDE will generate  a backtrace.  You can then 
upload it here.

Comment 5 Thorsten Staerk 2005-05-20 17:51:15 UTC

I have this here, too. It occurs on about every 2nd deletion. The following seems to resolve the error:

get the sources.
In taskview.cpp, change the order of the lines
      save();
      deleteItemState( task );
so they look
      deleteItemState( task );
      save();
then, recompile && reinstall.

Does this help at you, too ?

Comment 6 Thorsten Staerk 2005-05-21 08:39:02 UTC

*** Bug 105445 has been marked as a duplicate of this bug. ***

Comment 7 Thorsten Staerk 2005-05-21 08:48:07 UTC

I checked this in and backported it yesterday. The tests of make check still succeed. The fix will be in tomorrow's KDE 3.4.1.

What a luck - this bug blocked my karm-kontact-plugin-testing.

Comment 8 Martin Honermeyer 2005-06-01 19:13:23 UTC

Just installed 3.4.1.

Upon adding a new task, it says: "Error storing new task. Your changes were not saved."

My .xsession-errors says the following
* upon starting:
korgac: ERROR: Received DCOP: resource added for unknown resource Wjp98sv1Ui
kontact: ERROR: Received DCOP: resource added for unknown resource Wjp98sv1Ui

* upon exiting:
korgac: ERROR: Received DCOP: resource deleted for unknown resource Wjp98sv1Ui
kontact: ERROR: Received DCOP: resource deleted for unknown resource Wjp98sv1Ui


I should add that I am using a Kolab mail server (set up with kolabwizard), so I am using the contacts, calendar and notes resources from the IMAP server. Please also note that I am using an AMD64 machine!


Martin

Comment 9 Martin Honermeyer 2005-06-26 16:13:54 UTC

With the help of #106737, it works now! Had to delete the lock file manually. So my problem concerns #106737 only.

Comment 10 Thorsten Staerk 2005-08-27 14:51:02 UTC

*** Bug 110697 has been marked as a duplicate of this bug. ***