[FFmpeg-devel] MPEG-PS demuxer index memory usage[PATCH]

[Paul Kelly]

On Mon, 7 Jan 2008, Michel Bardiaux wrote:

[...]
> And this issue led me to read more carefully the code of 
> mpegps_read_pes_header, and I must say it looks quite strange to me, how can 
> it work for streamed (http) MPEG-PS when it relies so much on url_fseek, 
> ftell, fskip, which should fail for a streamed URL, but the return value is 
> never checked?

Yes I've noticed it attempts to do a lot of seeking in the stream, which 
obviously won't work, but all I can say is it seems not to affect the 
decoded video/audio and I haven't noticed any strange gaps or 
interruptions in the output files I am transcoding too. We have a fixed 
test sequence, which repeats every 70 seconds or so, so it is vaguely 
possible I haven't noticed something. There are no MPEG decode errors 
though which might be expected if it ended up in a position it wasn't 
expecting when reading the stream?

Also, I don't think it's relevant but our application is not http 
streaming, but reading from stdin piped from a hardware mpeg encoder 
device. The MPEG-TS output of the encoder is broken (Hauppauge PVR-150), 
otherwise we would be using that...

Paul