GNOME Bugzilla – Bug 668006
Crash when dragging column headers of Tasks list
Last modified: 2014-06-06 12:17:59 UTC
evolution-3.2.2-1.fc16.i686 STEPS: 1. Go to tasks view 2. As used from the mailer view, try two or three times to drag any of the column headers to the left or right. 3. Crash REPRO: 2/2 Program received signal SIGSEGV, Segmentation fault. 0x49b146b9 in eti_header_structure_changed (eth=0x85a6a40 [ETableHeader], a11y=0x8702d10) at gal-a11y-e-table-item.c:857 857 for (j = 0; j < prev_n_cols && prev_cols[j]; j++) { (gdb) thread apply all bt
+ Trace 229457
Thread 2 (Thread 0xb7dadb40 (LWP 4281))
Thread 1 (Thread 0xb7fc98c0 (LWP 4277))
(gdb) info registers eax 0x0 0 ecx 0xb0b0eef0 -1330581776 edx 0x5 5 ebx 0x49b31ff4 1236475892 esp 0xbfffdf30 0xbfffdf30 ebp 0x0 0x0 esi 0x1 1 edi 0x1 1 eip 0x49b146b9 0x49b146b9 <eti_header_structure_changed+265> eflags 0x10202 [ IF RF ] cs 0x73 115 ss 0x7b 123 ds 0x7b 123 es 0x7b 123 fs 0x0 0 gs 0x33 51 (gdb) list 852 prev_state = g_malloc0 (sizeof (gint) * prev_n_cols); 853 reorder = g_malloc0 (sizeof (gint) * n_cols); 854 855 /* Compare with previously saved column headers. */ 856 for (i = 0; i < n_cols && cols[i]; i++) { 857 for (j = 0; j < prev_n_cols && prev_cols[j]; j++) { 858 if (prev_cols[j] == cols[i] && i != j) { 859 860 reorder_found = TRUE; 861 state[i] = ETI_HEADER_REORDERED;
Interesting, this seems to work in 3.3.x webkit branch for me.
Crash seems to be from ETable a11y code. I'm this -><- close to just ripping out the a11y code once and for all.
Downstream bug report about the same from 3.2.2: https://bugzilla.redhat.com/show_bug.cgi?id=782372
Similar trace for the same issue:
+ Trace 229552
Thread 1 (Thread 0xb7fc7b00 (LWP 23568))
Downstream bug report about the same from 3.10.4: https://bugzilla.redhat.com/show_bug.cgi?id=1104776 Description of problem: I added new column in the task view. Version-Release number of selected component: evolution-3.10.4-2.fc20 Additional info: reporter: libreport-2.2.2 backtrace_rating: 4 cmdline: evolution crash_function: g_malloc0 executable: /usr/bin/evolution kernel: 3.14.4-200.fc20.x86_64
+ Trace 233670
Thread 1 (Thread 0x7f787d8faa40 (LWP 7011))
Use-after-free, caused by the gal-a11y object not disconnecting from a signal handler when it is freed. Created commit a7c87d1 in evo master (3.13.3+) [1] Created commit 69914f7 in evo evolution-3-12 (3.12.3+) [1] https://git.gnome.org/browse/evolution/commit?id=a7c87d1