[FFmpeg-cvslog] r11335 - trunk/libavdevice/v4l.c

Kostya kostya.shishkov
Fri Dec 28 13:52:28 CET 2007


On Fri, Dec 28, 2007 at 11:00:26AM +0100, Michael Niedermayer wrote:
> On Fri, Dec 28, 2007 at 10:14:29AM +0100, Luca Barbato wrote:
> > michael wrote:
> > > Author: michael
> > > Date: Fri Dec 28 04:30:23 2007
> > > New Revision: 11335
> > > 
> > > Log:
> > > Remove AIW support code, it is unfit for ffmpeg svn, doing usleep() and
> > > colorspace convertion and deinterlacing in the demuxer. Whoever wants
> > > AIW support has to implement this cleanly!
> > 
> > I agree that this should be removed, but:
> > 
> > - v4l/v4l2 now have a cleaner way to support AIW?
> 
> There never was any need to do deinterlacing or colorspace convertion in
> the demuxer. The original patch was bad and i made a serious mistake in
> applying it years ago.
> 
> 
> > - Removing this that was causes a regression?
> 
> I dont have a AIW card to test. Its possible that it works just fine. If
> not it has to be implemented cleanly. The messy code was in svn for years
> and people did not clean it up ... now they will if its still needed ...
> If we leave the code there it will never be cleaned up by anyone ...

Well, I have AIW (I have not been using it for quite a while though).
>From my experience km driver (v4l) is extremely incompatible with v4l interface.
And IIRC, AIW v4l2 driver by another developer was working good without.

Not that it matters to me now...
 
> [...]
> 
> -- 
> Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB




More information about the ffmpeg-cvslog mailing list