I'm more certain now that this crash was caused by the copy/paste problem - I triggered it a second time, but again did not quite manage to capture the test case... but I am now sure it involved transaction pasting, and transaction journal mode.
Here's a brief test case for a slightly different assertion I just
managed to trigger:
#3 0xb6d4b32f in IA__g_assertion_message (domain=0xb7e9911d
"gnc.register.ledger", file=0xb7e99fd5 "split-register-control.c", line=419,
func=0xb7e9a8a4 "gnc_split_register_move_cursor", message=0xb6d7483a "code
should not be reached")
at
/build/buildd-glib2.0_2.24.2-1-i386-AScyie/glib2.0-2.24.2/glib/gtestutils.c:1318
#4 0xb7e8c6c8 in gnc_split_register_move_cursor (p_new_virt_loc=0xbfffe9e0,
user_data=0x83d2818) at split-register-control.c:419
#5 0xb7e563aa in gnc_table_move_cursor_internal (table=<value optimized out>,
new_virt_loc=..., do_move_gui=1) at table-allgui.c:782
#6 0xb7e56485 in gnc_table_move_cursor_gui (table=0x0, new_virt_loc=...) at
table-allgui.c:900
#7 0xb7e56574 in gnc_table_verify_cursor_position (table=0x847b018,
virt_loc=...) at table-allgui.c:928
#8 0xb7e5661b in gnc_table_wrap_verify_cursor_position (table=0x847b018,
virt_loc=...) at table-allgui.c:991
#9 0xb7e74f8d in gnucash_sheet_cursor_move (sheet=0x84a8060, virt_loc=...) at
gnucash-sheet.c:321
#10 0xb7e754ec in gnucash_button_press_event (widget=0x84a8060,
event=0x84ae558) at gnucash-sheet.c:1479
#11 0xb7353e24 in _gtk_marshal_BOOLEAN__BOXED (closure=0x8098980,
return_value=0xbfffed24, n_param_values=2, param_values=0x84ffe78,
invocation_hint=0xbfffed10, marshal_data=0xb7e75090) at
/build/buildd-gtk+2.0_2.20.1-1+b1-i386-jmql5R/gtk+2.0-2.20.1/gtk/gtkmarshalers.c:84
1. View > Transaction Journal
2. Copy a transaction
3. Edit another transaction (and copy/paste a value from one column to another)
4. Click on first row of this transaction, and select Transaction > Paste
5. Yes, you want to overwrite
6. Click on another transaction
7. Yes, Record changes
Boom.
So with this test case and the one from the upstream bug report, I think
it's enough to backport the patch and check these have gone.
--
Tim Retout <dioc...@debian.org>
signature.asc
Description: This is a digitally signed message part
