[FFmpeg-devel] [PATCH]Only test the first frame for malformed bitstreams

[Carl Eugen Hoyos]

Reimar Döffinger <Reimar.Doeffinger <at> gmx.de> writes:

> The point I am trying to make is that I believe testing only 
> the first frame might still give the same false positives 
> the patch is trying to fix, just more rarely.

Sorry, but I find that argumentation not completely 
understandable since I move the problem from "all frames" to 
"first frame" (and did you test that streams with a broken 
first frame are accepted by the relevant players? I wanted 
to do that once this patch is committed and add a -strict -2 
setting.)

> Wouldn't it make sense to change the test to check for the 
> "bad" format, e.g. if the first few bytes seem to match 
> the packet size for H.264?

How would that avoid the false positives?

And how would that work for aac?
(Sorry, but I was hoping that fixing the bugs for the 
reported use-cases would not involve learning the 
bitstream syntax.)

> It will be less reliable in detecting a missing filter

I would consider that a problem.

Carl Eugen