Description of problem: Going to calendar. I'm guessing some appointment it's fetching is causing this to crash. Version-Release number of selected component: evolution-3.8.5-2.fc19 Additional info: reporter: libreport-2.2.2 backtrace_rating: 4 cmdline: /usr/bin/evolution crash_function: icaltimezone_load_builtin_timezone executable: /usr/bin/evolution kernel: 3.14.23-100.fc19.x86_64 runlevel: N 5 type: CCpp uid: 1000 Truncated backtrace: Thread no. 1 (10 frames) #0 icaltimezone_load_builtin_timezone at icaltimezone.c:1766 #1 icaltimezone_ensure_coverage at icaltimezone.c:462 #2 icaltimezone_get_utc_offset at icaltimezone.c:818 #3 icaltimezone_convert_time at icaltimezone.c:773 #4 icaltime_as_timet_with_zone at icaltime.c:415 #5 e_cal_model_set_instance_times at e-cal-model.c:4051 #6 process_added at e-cal-model.c:2762 #7 process_event at e-cal-model.c:2666 #8 client_view_objects_added_cb at e-cal-model.c:2938 #9 g_cclosure_marshal_VOID__POINTERv at gmarshal.c:1236 Potential duplicate: bug 875256
Created attachment 971245 [details] File: backtrace
Created attachment 971246 [details] File: cgroup
Created attachment 971247 [details] File: core_backtrace
Created attachment 971248 [details] File: dso_list
Created attachment 971249 [details] File: environ
Created attachment 971250 [details] File: exploitable
Created attachment 971251 [details] File: limits
Created attachment 971252 [details] File: maps
Created attachment 971253 [details] File: open_fds
Created attachment 971254 [details] File: proc_pid_status
Created attachment 971255 [details] File: var_log_messages
I can't tell if this is a duplicate of bug 875256. In that case, the remote calendar system was LDAP based. For this bug, the remote calendars are the Google calendars. Guess it depends on where the bug exactly is.
Thanks for a bug report. As noted in bug #1136778 comment #15, this can be caused by two parts, where one of them is fixed in the most latest evolution-data-server 3.12.9. The other part involves libical itself, to not use putenv() when working with timezones. I'm moving this to libical and I'll use a proposed patch from [1] to fix this issue. The fix itself may change during the time, especially after the libical's upstream review. In any case, the change will be done only for Fedora 21 and rawhide. Fedora 19 may suffer of other issues than the above mentioned. [1] http://lists.infradead.org/pipermail/libical-devel/2014-November/000630.html
Do you forsee any problems with compiling the patched libical manually on Fedora 19 as a way to get around the problem (since I won't be going to Fedora 21 anytime soon)?
libical-1.0-8.fc21 has been submitted as an update for Fedora 21. https://admin.fedoraproject.org/updates/libical-1.0-8.fc21
There is a build for rawhide too, the libical-1.0-8 version as well.
(In reply to Adam Stein from comment #14) > Do you forsee any problems with compiling the patched libical manually on > Fedora 19 as a way to get around the problem (since I won't be going to > Fedora 21 anytime soon)? I think there might not be an issue, it builds fine at least. Here are scratch (test) builds of evolution-data-server and libical with backported changes. Please report any issue you might find with them. Also note that the scratch builds will be autometically erased from the server after few days. If anything will go wrong then downgrade to previous version (`yum downgrade libical evolution-data-server`). You might restart the computer to be sure that all the background processes will be those from the updated packages. libical: http://koji.fedoraproject.org/koji/taskinfo?taskID=8458685 eds: http://koji.fedoraproject.org/koji/taskinfo?taskID=8458697
1) Quit evolution 2) Ran 'evolution --force-shutdown' to ensure no components were running 3) Installed scratch packages 4) Restarted evolution No difference. Calendar still crashes, everything else looks like it works. Thanks for the F19 scratch packages anyway. > sudo yum list libical evolution-data-server Loaded plugins: auto-update-debuginfo, langpacks, refresh-packagekit Installed Packages evolution-data-server.x86_64 3.8.5-7.1.fc19 @/evolution-data-server-3.8.5-7.1.fc19.x86_64 libical.x86_64 0.48-4.1.fc19 @/libical-0.48-4.1.fc19.x86_64 Available Packages evolution-data-server.i686 3.8.5-7.fc19 updates libical.i686 0.48-4.fc19 fedora
Package libical-1.0-8.fc21: * should fix your issue, * was pushed to the Fedora 21 testing repository, * should be available at your local mirror within two days. Update it with: # su -c 'yum update --enablerepo=updates-testing libical-1.0-8.fc21' as soon as you are able to. Please go to the following url: https://admin.fedoraproject.org/updates/FEDORA-2014-17526/libical-1.0-8.fc21 then log in and leave karma (feedback).
I'm running fc19. Would yum show me fc21 packages? I don't see any at updates-testing, including the libical mentioned above so I'm assuming I can't try it out.
(In reply to Adam Stein from comment #18) > 1) Quit evolution > 2) Ran 'evolution --force-shutdown' to ensure no components were running > 3) Installed scratch packages > 4) Restarted evolution > > No difference. Calendar still crashes, everything else looks like it works. > Thanks for the F19 scratch packages anyway. The --force-shutdown didn't kill background processes for couple releases, but some later version will do it again. To ensure no evolution related code is running use `ps ax | grep evolution`, but a simpler way is to re-login or restart the machine completely. (In reply to Adam Stein from comment #20) > I'm running fc19. Would yum show me fc21 packages? I don't see any at > updates-testing, including the libical mentioned above so I'm assuming I > can't try it out. Nope, you cannot use Fedora 21 packages in Fedora 19, due to the dependencies.
On FC19, the new scratch libical you made doesn't do anything (or more accurate to say it doesn't appear to solve my problem). Made sure no evolution component was running (via ps). Evolution still crashes when switching to the calendar. Since I won't be upgrading to FC21 anytime soon, I can't test that.
(In reply to Adam Stein from comment #22) > Evolution still crashes when switching to the calendar. I guess it's with the same (or a similar) backtrace as above. It means that the crash is unrelated to the change in libical, which I hoped it is. There landed too many changes between 3.8.5 and 3.12.9 (current stable in Fedora 21), to be able to identify the right bug, if any, which would influence your crasher. As you can reproduce this reliably, then maybe a run under valgrind would help to identify the cause of the crash, but again, the 3.8.5 is too old and dead for upstream. If you'd give it a try anyway, then please install debuginfo packages for evolution-data-server and evolution (make sure the version will match the binary package version) and then run evolution under valgrind with this command: $ G_SLICE=always-malloc valgring --num-callers=20 evolution -c mail &>log.txt It will run evolution and will log about any spot memory issues. The evolution will be very slow and CPU expensive, due to all the memory checking, but once it'll settle, please switch from the Mail view to the Calendar view. Even if it'll not crash, then the valgrind may log about the issue. It's due to valgrind being able to recognize certain memory issues and prevent the crash.
Thanks for the information. As you mentioned, the FC19 evolution is too old, so there's no point in continuing. I'll upgrade to FC21 sometime this year when I have some time.
libical-1.0-8.fc21 has been pushed to the Fedora 21 stable repository. If problems still persist, please make note of it in this bug report.
*** Bug 973755 has been marked as a duplicate of this bug. ***