Opened 15 years ago
Closed 15 years ago
Last modified 15 years ago
#9602 closed Bug Report - Crash (fixed)
Occational fe crash on playback startup / MythSystem?
| Reported by: | Owned by: | beirdo | |
|---|---|---|---|
| Priority: | minor | Milestone: | 0.25 |
| Component: | MythTV - General | Version: | Master Head |
| Severity: | medium | Keywords: | |
| Cc: | Ticket locked: | no |
Description
Frontend occationally crashes when playback is started. See attached backtrace which suggest that this might be MythSystem related. Didn't have time to test/report this on ticket #9421.
As shown on version output, local version is "dirty" but those modifications should have nothing to do with this.
Attachments (4)
Change History (12)
by , 15 years ago
| Attachment: | frontend-version.txt added |
|---|
by , 15 years ago
| Attachment: | playback-startup-bt.txt added |
|---|
comment:2 by , 15 years ago
| Owner: | set to |
|---|---|
| Status: | new → assigned |
by , 15 years ago
| Attachment: | fe-log.txt added |
|---|
comment:3 by , 15 years ago
Replying to beirdo:
I also need the log output from the frontend, running -v system,extra
Attached the requested log (not from the same crash as the backtrace). Started frontend, started playback of recording twice and the second one crashed.
by , 15 years ago
| Attachment: | 0001-Try-to-keep-from-another-apparent-race-condition.patch added |
|---|
follow-up: 6 comment:4 by , 15 years ago
OK, give this patch a try if you could, please. You seem to do quite well at finding the little tiny edge cases that I just can't get to fail here. :)
comment:6 by , 15 years ago
Replying to beirdo:
OK, give this patch a try if you could, please. You seem to do quite well at finding the little tiny edge cases that I just can't get to fail here. :)
Patch seems to fix the issue, no crashes with it :)
comment:7 by , 15 years ago
| Milestone: | unknown → 0.25 |
|---|---|
| Resolution: | → Fixed |
| Status: | assigned → closed |
Committed to master in 6fe46da4735. Thanks.
comment:8 by , 15 years ago
| Resolution: | Fixed → fixed |
|---|
Try to keep from another apparent race condition
Not quite sure why the QMap is going crazy, but let's try reading the AutoCleanup flag before unlocking so we can mitigate the race some. Also, in both myth_system() and in the signal handler, use deleteLater() so we don't try to double-delete, as that will just cause other messes.
Fixes #9602
Changeset: 6fe46da473563a3586ca9ead45870734abdc0fbe
I also need the log output from the frontend, running -v system,extra