[FFmpeg-devel] rtmp_read_seek unreachable?

[Reimar Döffinger]

On Sat, Jan 28, 2012 at 07:53:54PM +0100, Michael Niedermayer wrote:
> On Sat, Jan 28, 2012 at 07:41:36PM +0100, Reimar Döffinger wrote:
> > That is a bit of a problem.
> > I don't need it since I don't use external FFmpeg.
> > But distributions will complain when MPlayer no longer works
> > with external FFmpeg (some also if it just doesn't work with libav).
> > And they probably won't be doing the work.
> > I'd be happy for someone else to look into it, but otherwise
> > I'll try to keep it on my TODO list.
> 
> iam happy to try to put it back but iam not a user of this feature
> so i wouldnt know if it works nor if it breaks in the future. and
> this isnt good

Hm, I admit I wasn't sure about whether it is better to bring it back
or to rather make the internal API public, which I admit has more risks
right now.
MPlayer uses it, but only really for rtmp and I'm not sure people will
really notice.
There are no automated tests yet for that, though it might work adding
something that tests at least ffmpeg://file://something
Though that still leaves the problem that someone would have to run it
regularly, and experience shows unless you completely automate it it
won't be done.
I don't think I'll have enough time for it any time soon, but does
someone know if it seems realistic to hack FATE to work for other
programs as well?
Ideally with no changes to the server side, only a changed run script.
We'd still need people to host it all after that, but it would be
a first step.

Reimar