GNOME Bugzilla – Bug 267519
evolution-exchange-storage crashes when trying to view Exchange calendar
Last modified: 2009-08-15 18:40:50 UTC
evolution-exchange-storage crashes when I try to view my Exchange calendar. Only by unselecting (unchecking) my Exchange calendar can I prevent the crash. I still get alarm notifications and can also accept/reject meeting notices, but I cannot view my calendar without it crashing. I saw a bugzilla which seemed to indicate that this is a problem with a large cache.ics so I deleted my file. It was 4.2MB but my new file has grown to 1.7MB. Unfortunately, this did not resolve the problem. I upgraded from Evolution 1.4 to 2.0.1 using the Fedora Core 2 yum repository at: [evolution2] name=Evolution 2.0 for Fedora Core 2 baseurl=http://petrix.se/evolution2 The packages installed are: evolution-2.0.1-0.mozer.2 evolution-connector-debuginfo-1.4.7-5 evolution-data-server-1.0.1-0.mozer.1 evolution-connector-2.0.0-0.mozer.2
Confirmed. I get exactly the same issue Running gdb /usr/libexec/evolution/2.0/evolution-exchange-storage on Gentoo gives me ------ [Thread debugging using libthread_db enabled] [New Thread 1097749168 (LWP 32027)] [New Thread 1106262960 (LWP 32030)] Evolution Exchange Storage up and running /usr/share/themes/Nuvola/gtk-2.0/iconrc:23: Unable to locate image file in pixmap_path: "stock_stock.svg" /usr/share/themes/Nuvola/gtk-2.0/iconrc:54: Unable to locate image file in pixmap_path: "stock_ok.svg" Detaching after fork from child process 32031. [New Thread 1116736432 (LWP 32038)] [New Thread 1125129136 (LWP 32043)] [New Thread 1133521840 (LWP 32057)] [New Thread 1141914544 (LWP 32062)] [New Thread 1150307248 (LWP 32063)] [Thread 1150307248 (LWP 32063) exited] (evolution-exchange-storage:32027): libedata-cal-WARNING **: e-data-cal-view.c:249: Starting new query [Thread 1133521840 (LWP 32057) exited] (evolution-exchange-storage:32027): libedata-cal-WARNING **: e-data-cal-view.c:249: Starting new query [Thread 1116736432 (LWP 32038) exited] (evolution-exchange-storage:32027): libedata-cal-WARNING **: e-data-cal-view.c:249: Starting new query (evolution-exchange-storage:32027): libecal-CRITICAL **: file e-cal-component.c: line 2321 (e_cal_component_get_dtstart): assertion `priv->icalcomp != NULL' failed (evolution-exchange-storage:32027): libecal-CRITICAL **: file e-cal-component.c: line 2205 (e_cal_component_get_dtend): assertion `priv->icalcomp != NULL' failed [Thread 1125129136 (LWP 32043) exited] (evolution-exchange-storage:32027): libecal-CRITICAL **: file e-cal-component.c: line 3396 (e_cal_component_has_rdates): assertion `priv->icalcomp != NULL' failed (evolution-exchange-storage:32027): libecal-CRITICAL **: file e-cal-component.c: line 3489 (e_cal_component_has_rrules): assertion `priv->icalcomp != NULL' failed (evolution-exchange-storage:32027): libecal-CRITICAL **: file e-cal-component.c: line 2677 (e_cal_component_has_exdates): assertion `priv->icalcomp != NULL' failed (evolution-exchange-storage:32027): libecal-CRITICAL **: file e-cal-component.c: line 2838 (e_cal_component_has_exrules): assertion `priv->icalcomp != NULL' failed free(): invalid pointer 0x412621b4! free(): invalid pointer 0x416697c0! Program received signal SIGSEGV, Segmentation fault.
+ Trace 50776
Thread 1097749168 (LWP 32027)
It certainly used to work with Evo 1.4.7 and Evo 1.5.94 - so something happened with ximian-connector 2.0
When I tick the box next to my calendar, evolution crashes. I have to kill evolution and evolution-exchange-storage and restart it three times before evolution will reopen, without the calendar selected. Here is the gdb info: Evolution Exchange Storage up and running (no debugging symbols found)...Detaching after fork from child process 11249. [New Thread -163988560 (LWP 11311)] [Thread -163988560 (LWP 11311) exited] [New Thread -163988560 (LWP 11318)] (evolution-exchange-storage:11245): libedata-cal-WARNING **: e-data-cal-view.c:249: Starting new query (evolution-exchange-storage:11245): libecal-CRITICAL **: file e-cal-component.c: line 2321 (e_cal_component_get_dtstart): assertion `priv->icalcomp != NULL' failed (evolution-exchange-storage:11245): libecal-CRITICAL **: file e-cal-component.c: line 2205 (e_cal_component_get_dtend): assertion `priv->icalcomp != NULL' failed [Thread -163988560 (LWP 11318) exited] (evolution-exchange-storage:11245): libecal-CRITICAL **: file e-cal-component.c: line 3396 (e_cal_component_has_rdates): assertion `priv->icalcomp != NULL' failed (evolution-exchange-storage:11245): libecal-CRITICAL **: file e-cal-component.c: line 3489 (e_cal_component_has_rrules): assertion `priv->icalcomp != NULL' failed (evolution-exchange-storage:11245): libecal-CRITICAL **: file e-cal-component.c: line 2677 (e_cal_component_has_exdates): assertion `priv->icalcomp != NULL' failed (evolution-exchange-storage:11245): libecal-CRITICAL **: file e-cal-component.c: line 2838 (e_cal_component_has_exrules): assertion `priv->icalcomp != NULL' failed *** glibc detected *** double free or corruption: 0x0822aa90 *** Program received signal SIGABRT, Aborted. [Switching to Thread -151147616 (LWP 11245)] 0x4c7197a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 (gdb) thread apply all bt
+ Trace 52167
Thread 2 (Thread -153375824 (LWP 11248))
I am running FC3 BTW. I was running FC2 with the mozer packages, and started having problems upon upgrade.
I installed Fedora Core 3 which includes evolution 2.0.2 and connector 2.0.2 and the application no longer crashes in the same manner as first reported. Instead, it becomes completely inactive/unresponsive and I have to run evolution --force-shutdown. Everytime I try to access my Exchange calendar I have the same results.
Bryan Christ: re the FC3 crash, is this really the same bug? Please can you install the debuginfo packages, and obtain a backtrace; see this page for more instructions: http://fedora.linux.duke.edu/wiki/index.cgi/StackTraces If you think it's actually a different bug, please open a new bug for it; if you think it's still related to this one, then please add the backtrace here. Thanks.
This seems to be having for people who have upgraded from evolution-exchange 1.4 to 2.0. Though, this is not true for all. Which makes me think that this should definately have something to do with the data as well. And another clue is that users facing this problem have had huge calendar/cache.ics files. This is a bit tricky to reproduce. But, my guess is that, its most probably a recurrence meeting [ recurring over a week or month ] which was created using 1.4 client, and then the client was upgraded to 2.0 , and this somehow created a real huge cache.ics file with some really corrupt calendar info in it and this triggers this crash [ and sometimes a memory leak as well ]. I have not been able to reproduce this as such. It would be very helpful if we could get a peek into the corrupted cache.ics file, or any hints to reproduce this, as to what [ the properties such as, its recurrence, its duration, its attendees, etc ] was the last appointment/meeting you created/recieved before upgrading to 2.0
The fix for #70414 should now solve this. Please verify this with the new builds [ 2.0.3 ]
I am using 2.0.3-r2 under gentoo and I get the same results. This referrs to a "huge" cache.ics. What constitues "huge"? 1Meg, 10Meg? I _never_ used 1.4 on my system, it has been 2.0 from the beginning and I get the same problem. All of my calendar info is coming from true Outlook clients, not Evolution clients. I can supply a cache.ics file if it's needed.
I am now using Evolution/Connector version 2.1.5 and the issue appears resloved.
Closing the bug based on reporters comments