[FFmpeg-cvslog] r11815 - trunk/libavformat/asf.c

[Reimar Döffinger]

Hello,
On Sun, Feb 03, 2008 at 12:32:25AM +0100, Michael Niedermayer wrote:
> On Sat, Feb 02, 2008 at 11:57:28PM +0100, Reimar D?ffinger wrote:
> > On Sat, Feb 02, 2008 at 11:36:29PM +0100, Michael Niedermayer wrote:
> > > And the #ifdef makes the code much more obfuscated. A unused function never
> > > points to a bug except by coincidence. It might point to some forgotten old
> > > code but not more. Theres attribute(unused) for silencing such warnings.
> > 
> > well, it's no problem to change it to av_unused, though this whole thing
> > IMO raises a few other questions, like what is a CONFIG_MMSH_PROTOCOL
> > doing in a demuxer at all.
> 
> hmm, ive thought it would be needed, but after looking at the mms patch again
> i dont see why it was implemented like that. Feel free to remove the is_mms()
> and the case where its used.

Well, as far as I can tell it is used to let a demuxer-level AV_DISCARD
filter down to the stream level.
But I doubt that this would work too well anyway, because in my
experience quite a few mms severs will not accept any combination of
selected streams but instead just send some that are suitable to it.
Maybe this is due to a bug in MPlayer's code, but if not applications
will have to implement a way to find which streams are _actually_
provided (lack of this is btw. the reason with MPlayer with -demuxer
lavf and mms streams will usually hang, it tries autoselects a stream
that the server does not send).
If you have an idea for a good and simple API to provide such "check if
stream is actually available" functionality I'd be interested in it ;-)

Greetings,
Reimar D?ffinger