GNOME Bugzilla – Bug 672518
Evolution crashes when loading Calendar
Last modified: 2012-05-16 13:57:09 UTC
Hello, Starting tonight, Evolution 3.2.2 has started crashing whenever I click the Calendar button. After this crash, Evo will crash on startup every time. I can workaround by starting with "evolution -c mail", but clicking Calendar will crash. The crash is in libical. I've installed Evo debugging symbols and build libical with debugging symbols. Here is the backtrace: Starting program: /usr/bin/evolution [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/i386-linux-gnu/i686/cmov/libthread_db.so.1". [New Thread 0xb4b8cb70 (LWP 3619)] [New Thread 0xb438bb70 (LWP 3620)] [New Thread 0xae6d0b70 (LWP 3621)] RSS Plugin enabled (evolution 3.2, evolution-rss 0.2.90) [New Thread 0xb35b4b70 (LWP 3622)] (evolution:3616): Tracker-DEBUG: Tracker Evolution plugin enabled (evolution:3616): Tracker-DEBUG: Tracker plugin creating new object... [New Thread 0xa9fb9b70 (LWP 3623)] [New Thread 0xa97b8b70 (LWP 3626)] [Thread 0xb35b4b70 (LWP 3622) exited] [New Thread 0xa8faab70 (LWP 3627)] [Thread 0xa8faab70 (LWP 3627) exited] Program received signal SIGSEGV, Segmentation fault. 0xb7ddb423 in icaltimezone_ensure_coverage () from /usr/lib/libical.so.0 (gdb) thread apply all bt
+ Trace 229919
Thread 1 (Thread 0xb5d798b0 (LWP 3616))
Please let me know any other information I can provide. Thanks, Ross
I just tested against libical 0.48, similar crash in the relevant thread:
+ Trace 229920
Thread 1 (Thread 0xb5d7b8b0 (LWP 8427))
Testing against 3.2.3 and installing debug packages for evolution-data-server and libglib2 to provide a better stacktrace is welcome.
Here is the 3.2.2 backtrace with the additional symbols loaded. No Debian packages for 3.2.3 yet, I'll try to build and test this later tonight.
+ Trace 229925
Thread 1 (Thread 0xb5d7b8b0 (LWP 16970))
3.2.3 still suffers from this issue. Here is the backtrace: (gdb) run [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/i386-linux-gnu/i686/cmov/libthread_db.so.1". [New Thread 0xb5f99b70 (LWP 7293)] [New Thread 0xb5798b70 (LWP 7294)] [New Thread 0xafb1bb70 (LWP 7295)] [New Thread 0xaed1cb70 (LWP 7296)] [New Thread 0xae42eb70 (LWP 7298)] [New Thread 0xadc20b70 (LWP 7299)] [Thread 0xadc20b70 (LWP 7299) exited] Program received signal SIGSEGV, Segmentation fault. 0xb78d06c7 in icalarray_free () from /usr/lib/libical.so.0 (gdb) thread apply all bt
+ Trace 229928
Thread 2 (Thread 0xb5f99b70 (LWP 7293))
Thread 1 (Thread 0xb6280870 (LWP 7290))
Similar downstream bug report from e-calendar-factory of eds 3.2.2: https://bugzilla.redhat.com/show_bug.cgi?id=819979
+ Trace 230197
Thread 1 (Thread 0x7f03e0a99700 (LWP 4383))
I figured this out: Evo crashes when trying to display certain calendar items created on my cellphone. My cellphone calls it a "Reminder" - it appears as a meeting with no length, but still appears on the calendar and can have alarms. Evo only crashes if one of these items is in the display. This worked fairly recently, but I don't recall which version. I'll attach an example vcalendar file captured from syncevolution's data dump.
Created attachment 214002 [details] Evo crashes when rendering this calendar item
Thanks for the update. Your event works for me fine with 3.4.2 and libical 0.48. I guess this is introduced by an update of libical. If you downgrade it to 0.46, then it'll start working again.
Hmm, I downgraded to libical 0.46 a while back and was still having the issue. I only upgraded to 0.48 since Debian has debug packages for it already, but ended up building my own against 0.46. I just tried to recreate this against the current Debian packages, and I can no longer recreate the crash. When I discovered yesterday what events needed to be scrolled into view, I removed them with my phone so I could continue using my calendar. The issue must've been particular to those now-deleted events. Closing.
Either that or some libical quirk about timezones (see an investigation at bug #673197).