GNOME Bugzilla – Bug 518551
Evolution crashed on exit
Last modified: 2010-09-14 09:06:07 UTC
Evolution 2.21.92 I was doing sanity of Evolution 2.21.92 tarballs and had evolution configured with exchange back end. I clicked on 'X' button (right hand top corner) to exit and it crashed. Gdb traces of evolution process (evolution:21041): e-table-CRITICAL **: e_table_model_row_count: assertion `E_IS_TABLE_MODEL (e_table_model)' failed (evolution:21041): e-table-CRITICAL **: e_table_selected_count: assertion `E_IS_TABLE(e_table)' failed Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0xb65a8a60 (LWP 21041)] 0xb6862432 in g_type_check_instance_cast (type_instance=0x49, iface_type=134856872) at gtype.c:3172 3172 if (type_instance->g_class) (gdb) thread apply all bt
+ Trace 190367
Thread 1 (Thread 0xb65a8a60 (LWP 21041))
gdb) thread apply all bt full
+ Trace 190371
Got same crash in Evolution 2.23.1 (evolution:20879): e-table-CRITICAL **: e_table_model_row_count: assertion `E_IS_TABLE_MODEL (e_table_model)' failed (evolution:20879): e-table-CRITICAL **: e_table_selected_count: assertion `E_IS_TABLE(e_table)' failed Program received signal SIGSEGV, Segmentation fault.
+ Trace 195996
Thread 3057678032 (LWP 20879)
Thread 1 (Thread 0xb64076d0 (LWP 20879))
Evolution-data-server also crashed with same traces (evolution:17785): e-utils-WARNING **: Error file '/home/build/opt/gnome2/share/evolution/2.24/errors/org-gnome-shared-folder.errors.xml': missing button details in error 'recurrence' (evolution:17785): Gtk-CRITICAL **: gtk_window_set_transient_for: assertion `parent == NULL || GTK_IS_WINDOW (parent)' failed Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0xb64db6c0 (LWP 17785)] 0xb676c052 in g_type_check_instance_cast (type_instance=0xaaaaaaaa, iface_type=134957488) at gtype.c:3172 3172 if (type_instance->g_class) (gdb) thread apply all bt
+ Trace 195997
Thread 1 (Thread 0xb64db6c0 (LWP 17785))
Created attachment 109807 [details] Gdb traces of evolution-data-server under 'bt full'
Right. I fixed it for mailer. We shouldn't set the info messages while the Component is finalizing. I will fix it.
this bug is in evolution, not e-d-s (In reply to comment #5) > Right. I fixed it for mailer. We shouldn't set the info messages while the > Component is finalizing. I will fix it. srini.. ping ;-)
*** Bug 546387 has been marked as a duplicate of this bug. ***
I think I saw this somewhere, the thing is not so simple, it's problem with backend_died_cb, when emitting lost/broken connection signal from linc. In comment #0 is this signal delivered on idle, in the last during application exit. I do not know how to fix this properly, really.
*** Bug 385189 has been marked as a duplicate of this bug. ***