-
Notifications
You must be signed in to change notification settings - Fork 0
Crash when selecting a turret in the unit design screen #2209
Comments
info@... uploaded file |
info@... uploaded file |
info@... uploaded file |
anonymous uploaded file |
anonymous commented Same here. Crashes with access violation at 77c?e1fe in ntdll.dll |
Buginator commented Can you give any more information on how to replicate this ? It looks like you both are not using English, but something else. It also looks like you both are crashing after SDL_HasAltiVec() so it could also be a SDL bug of some type... |
anonymous commented Indeed, I'm playing the german version, since that's the one I'm most used to ^_^ I'll try replicating the crash on english when I have some spare time. |
cybersphinx changed operating_system which not transferred by tractive |
cybersphinx edited the issue description |
cybersphinx changed title from |
cybersphinx changed priority from |
cybersphinx changed milestone from |
cybersphinx changed keywords from |
cybersphinx uploaded file glibc backtrace, not caught by the crash handler |
cybersphinx uploaded file gdb backtrace from 3.0, German |
cybersphinx uploaded file gdb backtrace from 3.0, Chinese |
cybersphinx uploaded file gdb backtrace from qt-trunk, German |
cybersphinx edited the issue description |
R3act!ve commented Campaign is definitive unplayable (untestable) whit this bug. Main reason is the flamer turret. Replicable. Sometimes there are also random crashes if you click on any button in the design menu. |
cybersphinx uploaded file Valgrind run of 6f5392f14947f0c44cedd40165031f4eb4a819a7. Of course the problem itself didn't occur, but maybe the invalid read/writes are the cause. |
cybersphinx commented Some bisecting was more successful, and indicates 81f07188a3693007621496ac9fa0c0e2e8b24b87. Now we get crashes instead of "error |05:10:12: [iV_GetCharWidth] Couldn't retrieve a bounding box for the character code 252". Well, at least we have some whiny guy on IRC to blame. |
cybersphinx commented Hm, and of course it was one of those "what could possibly happen?" commits that didn't even need another rc... |
buginator <buginator@...> changed status from |
buginator <buginator@...> changed resolution from `` to |
buginator <buginator@...> commented (In Warzone2100/warzone2100@3012fde) CHANGELOG: Correctly handle unicode strings for edit boxes. fixes #2209 |
buginator <buginator@...> commented In Warzone2100/warzone2100@3012fde:
|
Per Inge Mathisen <per@...> commented (In Warzone2100/warzone2100@209e12b) CHANGELOG: Correctly handle unicode strings for edit boxes. fixes #2209 |
Per Inge Mathisen <per@...> commented In Warzone2100/warzone2100@209e12b:
|
buginator <buginator@...> commented (In Warzone2100/warzone2100@5a35eb9) CHANGELOG: Correctly handle unicode strings for edit boxes. fixes #2209 Conflicts:
|
buginator <buginator@...> commented In Warzone2100/warzone2100@5a35eb9:
|
buginator commented Don't try to render a string that is NULL. Unknown why this happens, need someone that has access to non-English systems to debug this matter more. refs #2209 Conflicts:
Changeset: 50aafb279edaed68251079ce463c83c46628d14b |
buginator commented CHANGELOG: Correctly handle unicode strings for edit boxes. fixes #2209 Conflicts:
Changeset: 5a35eb9f89b1d24bcc7abd58094e67496de5708e |
buginator commented Don't try to render a string that is NULL. Unknown why this happens, need someone that has access to non-English systems to debug this matter more. refs #2209 Conflicts:
Changeset: 50aafb279edaed68251079ce463c83c46628d14b |
cybersphinx changed milestone from |
cybersphinx commented Milestone 3.0 deleted |
resolution_fixed
type_bug
| by info@...Steps to reproduce: Go to unit design, select a body and propulsion, at least one of which needs a non-ascii character in its name, select a turret (possibly repeat the last step until the crash).
This is a strange problem. It doesn't happen (in an obvious way) on Linux, and it doesn't happen (so far) with the English language setting. It does happen on qt-trunk though (as well as 2.3.5, 3.0 and master). Happens in both skirmish and campaign. On Linux, it doesn't happen (so far) when the first turret is selected, it can take a dozen clicks (speed doesn't matter). On Windows, it seems to crash instantly in the campaign, but not skirmish. Possibly some memory corruption again...
Issue migrated from trac:2209 at 2022-04-16 06:39:52 -0700
The text was updated successfully, but these errors were encountered: