[MPlayer-dev-eng] [PATCH] split filter_n_bytes function a bit

Reimar Döffinger Reimar.Doeffinger at stud.uni-karlsruhe.de
Sun Nov 4 12:17:45 CET 2007


On Sun, Nov 04, 2007 at 01:02:28PM +0200, Uoti Urpala wrote:
> On Sun, 2007-11-04 at 11:39 +0100, Reimar Döffinger wrote:
> > Here I disagree, the current code pretends to check for errors/EOF,
> > while it actually only checks if for some reason the decoder didn't
> > return any data. Most likely that means EOF or error but not
> > necessarily.
> 
> Is there some decoder that returns <= 0 except on error/EOF? If there is
> then it's broken IMO.

Well, e.g. ad_libmad checks if maxlen is too small. On second thought I
must admit that this case could be considered an error.
When I said "error" I was only thinking of things like broken bitstream.
But in the case of broken bitstream it seems not to make sense to me to
behave like on EOF and skip filtering, just ignoring it currently seems
to make more sense. Except that this would cause A-V sync/buffering
issues. I guess I currently just don't know how exactly codecs and this
filtering stuff should behave for which kind of errors :-(

Greetings,
Reimar Döffinger



More information about the MPlayer-dev-eng mailing list