GNOME Bugzilla – Bug 333224
Evolution crashed when modified time of exchange appointment
Last modified: 2013-09-10 14:04:48 UTC
Created a recurring appointment. In day view to increase end time dragged the lower end of the appointment slot, evolution crashed. Traces indicates tool tip crash. Evolution stack traces. Backtrace was generated from '/opt/gnome/bin/evolution' Using host libthread_db library "/lib/libthread_db.so.1". [Thread debugging using libthread_db enabled] [New Thread -1233311392 (LWP 9384)] [New Thread -1332479072 (LWP 9401)] 0xffffe410 in __kernel_vsyscall ()
+ Trace 66648
Thread 1 (Thread -1233311392 (LWP 9384))
Reproducible always. Evolution version 2.5.92.
*** Bug 335443 has been marked as a duplicate of this bug. ***
confirming as per duplicate.
*** Bug 338951 has been marked as a duplicate of this bug. ***
This is not specific to appointment in exchange calendar, it is reproducible with GW & personal calendar as well.
srini, i'm pretty sure this is the same crash as in bug 340312, bug 340416 (which provides an explanation how to reproduce) and bug 339030. this bites many people, it seems.
*** Bug 339381 has been marked as a duplicate of this bug. ***
*** Bug 342200 has been marked as a duplicate of this bug. ***
Ubuntu bug about that: https://launchpad.net/products/evolution/+bug/49576
*** Bug 352327 has been marked as a duplicate of this bug. ***
*** Bug 353055 has been marked as a duplicate of this bug. ***
*** Bug 353156 has been marked as a duplicate of this bug. ***
guess that bug 334692 is the same/similar. targetting to 2.8, many duplicates.
*** Bug 353965 has been marked as a duplicate of this bug. ***
*** Bug 355741 has been marked as a duplicate of this bug. ***
Created attachment 73666 [details] [review] Proposed patch
This patch disables the annoyingly popping tooltip when editing an appointment or dragging it, which is the source of the crash. This should solve multiple problems around tooltip in day/work-week view.
Srini - I will let you make a call if you want the tooltips disabled while dragging an event but it is not the source of the crash or the multiple problems around tooltip. From the code : e_calendar_view_get_tooltips (ECalendarViewEventData *data) { <snip> const char *str; <snip> str = icalcomponent_get_summary (pevent->comp_data->icalcomp); /* Here, str is assigned a const char * to the memory belonging to ical */ <snip> e_cal_component_get_location (newcomp, &str); /* This call attempts to overwrite that memory - obviously leading to a crash*/ The variable str must not be reused for the second call or atleast should not be pointing to ical *AND* sufficient memory should be allocated before &str is used.
mea culpa. Guess I need to catch up with my sleep :-/. The above usage of &str is indeed valid and does not need allocation as str only gets assigned to a different pointer. The issue seems to be this part : if (!e_cal_component_set_icalcomponent (newcomp, clone_comp)) g_warning ("couldn't update calendar component with modified data from backend\n"); where even if clone_comp fails to get set into newcomp - the rest of the code attempts to use the values inside newcomp.
if (!e_cal_component_set_icalcomponent (newcomp, clone_comp)) { g_warning ("couldn't update calendar component with modified data from backend\n"); g_object_unref (newcomp); g_free (data); return FALSE; } should solve the problem instead. Srini ?
Harish, if you lookinto the stack trace bug 336555 , bug 340416 and many more you can see the crash occuring at line number 1991 (old line number which is) [Switching to Thread 1099448800 (LWP 22684)] 0x41d40e5e in e_calendar_view_get_tooltips (data=0x8d561a0) at e-calendar-view.c:1991 1991 clone_comp = icalcomponent_new_clone (pevent->comp_data->icalcomp); This particular line is just before what you pointed out. Im sure that what you pointed out is also a issue, but not at all related to the tooltip crashers. In this particular case, the PEVENT structure points to a PTRArray which has the relative memory removed for the event and the tooltip time out triggered for a removed event. that is the reason for all these crashers. Ive fixed a lot of similiar issues in tooltip for month view too. Im sure, this is the best problem. When a event is dragged or resized the event is removed from ptrarray, also tooltip will annoying at that point. Hope Ive clarified ur point.
Created attachment 73922 [details] [review] Updated patch
The updated patch looks good. Please add the ChangeLog before committing it into the gnome-2-16 branch and the trunk.
Committed to HEAD.
to stable also.
*** Bug 372297 has been marked as a duplicate of this bug. ***
*** Bug 374662 has been marked as a duplicate of this bug. ***
*** Bug 382208 has been marked as a duplicate of this bug. ***
*** Bug 404058 has been marked as a duplicate of this bug. ***
*** Bug 421570 has been marked as a duplicate of this bug. ***
*** Bug 425642 has been marked as a duplicate of this bug. ***
*** Bug 429247 has been marked as a duplicate of this bug. ***
reopening as per duplicates.
*** Bug 446306 has been marked as a duplicate of this bug. ***
*** Bug 461715 has been marked as a duplicate of this bug. ***
*** Bug 470887 has been marked as a duplicate of this bug. ***
*** Bug 476236 has been marked as a duplicate of this bug. ***
*** Bug 496673 has been marked as a duplicate of this bug. ***
Anything on 2.12? IIRC I changed a few things in that area. Donno if it is fixed with 2.12.
*** Bug 518646 has been marked as a duplicate of this bug. ***
last dup in debian (gnome 2.20.x)
*** Bug 531539 has been marked as a duplicate of this bug. ***
last dup in debian (gnome 2.22.x)
*** Bug 532238 has been marked as a duplicate of this bug. ***
*** Bug 538892 has been marked as a duplicate of this bug. ***
*** Bug 523593 has been marked as a duplicate of this bug. ***
*** Bug 513428 has been marked as a duplicate of this bug. ***
*** Bug 485074 has been marked as a duplicate of this bug. ***
*** Bug 344960 has been marked as a duplicate of this bug. ***
Created attachment 113987 [details] [review] proposed evo patch for evolution; During some work on other bug I was able to reproduce this and did a patch too. The patch contains also a little optimalization, but I doubt it will be noticeable. Anyway, as I found some other duplicates then I found also that the reason was given by Li in bug #344960 comment 2 some time ago.
*** Bug 542073 has been marked as a duplicate of this bug. ***
*** Bug 543912 has been marked as a duplicate of this bug. ***
*** Bug 381183 has been marked as a duplicate of this bug. ***
*** Bug 526568 has been marked as a duplicate of this bug. ***
*** Bug 552819 has been marked as a duplicate of this bug. ***
*** Bug 556450 has been marked as a duplicate of this bug. ***
*** Bug 556096 has been marked as a duplicate of this bug. ***
Committed to trunk. Committed revision 36858. ... I hate to do this but this takes way too long. Please reopen if you face in 2.25.3 or later. Thanks.
*** Bug 341430 has been marked as a duplicate of this bug. ***
*** Bug 566901 has been marked as a duplicate of this bug. ***
*** Bug 568238 has been marked as a duplicate of this bug. ***
*** Bug 569972 has been marked as a duplicate of this bug. ***
*** Bug 570970 has been marked as a duplicate of this bug. ***
*** Bug 564788 has been marked as a duplicate of this bug. ***