Opened 14 years ago
Closed 13 years ago
#9581 closed Bug Report - Crash (Won't Fix)
[CRASH] mythcommflag SIGSEGV in ClassicLogoDetector::DetectEdges()
Reported by: | Owned by: | cpinkham | |
---|---|---|---|
Priority: | major | Milestone: | 0.25 |
Component: | MythTV - Mythcommflag | Version: | 0.24-fixes |
Severity: | medium | Keywords: | |
Cc: | Ticket locked: | no |
Description
Crash while trying to commflag a (very) corrupt capture. I can upload the stream somewhere if it is desired - otherwise feel free to close this ticket if it is not worth pursuing.
Attachments (2)
Change History (12)
Changed 14 years ago by
Attachment: | ThreadStacktrace.txt added |
---|
Changed 14 years ago by
Attachment: | mythbackend.log.gz added |
---|
comment:1 Changed 14 years ago by
Summary: | mythcommflag SIGSEGV in ClassicLogoDetector::DetectEdges() → [CRASH] mythcommflag SIGSEGV in ClassicLogoDetector::DetectEdges() |
---|
comment:2 Changed 13 years ago by
Status: | new → infoneeded_new |
---|
comment:3 Changed 13 years ago by
It's been four months, but you are lucky - I did manage to find the recording. mythcommflag -v filename still crashes 0.24.1, if that is your question. I can upload the recording somewhere if that is desired.
comment:4 Changed 13 years ago by
Sure, if you could. This would help test the extra protection that I'm thinking of putting in. Few of my files get bad enough to still crash commflagging.
comment:5 Changed 13 years ago by
Here is a small sample: http://www.2shared.com/video/4varxuGf/cut.html I'll try to remember to upload a larger one overnight
comment:7 Changed 13 years ago by
If this is still happening, is there any chance you can upload the sample file again? Fixing checks for the null buf pointer won't help solve the underlying problem that we are getting a null pointer from the decoder. We should fix this in the decoder rather than at every point where we touch the buf pointer.
comment:8 Changed 13 years ago by
Priority: | minor → major |
---|---|
Type: | Bug Report → Bug Report - Crash |
comment:9 Changed 13 years ago by
Milestone: | unknown → 0.25 |
---|
comment:10 Changed 13 years ago by
Resolution: | → Won't Fix |
---|---|
Status: | infoneeded_new → closed |
The file in question was not even watchable. In cases like these, while it would be nice to not crash, it's something inside ffmpeg code that is actually crashing, so without extensive debugging inside their code, it's best to assume this will not get fixed. If your recording is that screwed, delete it and re-record, and live with commflagging crashing on it.
Is this repeatable? I am thinking of putting in some more null protection, but without a way to test it, I'm not sure if it's worth the effort.