GNOME Bugzilla – Bug 687640
[abrt] Crash in caldav's initialize_backend()
Last modified: 2014-01-15 13:51:15 UTC
Moving this from a downstream bug report:
Version-Release number of selected component:
libreport version: 2.0.18
Core was generated by `/usr/libexec/evolution-calendar-factory'.
Program terminated with signal 6, Aborted.
(Thread 0x7f7743b65800 (LWP 1642))
I'm still seeing this with 3.8.2 (on Ubuntu). I don't have a good way to reproduce; setting a breakpoint shows that the code passes through this block successfully most of the time and it *seems* to be coded appropriately defensively...
Thanks for the update, Iain, I'm updating the version information. I'm not able to reproduce this reliably too, it just happens sometimes. I guess it's some use-after-free, but as you said, the code has there some locks, which are used to prevent such situation.
Still happens with 3.8.3
The latest release is 3.8.4, and the line numbers in these back traces don't match what's currently in the gnome-3-8 branch, so I can't tell if the issue is fixed already or not.
It would be helpful if someone could reproduce this with 3.8.4 or later.
It crashes from time to time to me too, even on git master, but I cannot debug it properly, because once I try it stops crashing :-/
Happens to me (and others) on 3.8.5 per this downstream bug report:
Bug 862830 - [abrt] evolution-data-server-3.5.92-1.fc18: initialize_backend: Process /usr/libexec/evolution-calendar-factory was killed by signal 6 (SIGABRT)
Created attachment 262983 [details]
backtrace from evolution-data-server-3.10.2-2 on Fedora 19
Attaching a backtrace from evolution-data-server-3.10.2-2 on Fedora 19
I think I finally found the cause, it was about two threads accessing internal structure members simultaneously, overwriting data one to the other.
Created commit 12a279c in eds master (3.11.5+) 
Created commit 16e77f3 in eds gnome-3-10 (3.10.4+)