[FFmpeg-devel] Realmedia patch
Wed Sep 10 14:50:04 CEST 2008
Ronald S. Bultje wrote:
> Yes. I'm considering adding a if(rdt&&!tcp) return error; as you
> suggested earlier, I think we're thinking about a case too much that
> may not exist at all, all servers give me a 461 (unsupported protocol)
> for UDP. See #4 attached.
It turned out that this happens because of the ";unicast" parameter we
add to the transport... It is standard, and mandated by rfc2326, but
again real servers live in their own world, and do not like it.
If the transport line is changed in
then the server is happy with it and replies with a "200 OK".
BTW, any reason for the funny way libavformat is using to check for
a supported transport? Maybe this is due to some other known bug in
some crappy server? Or to some other reason?
More information about the ffmpeg-devel