[FFmpeg-devel] [PATCH] Redefine AVERROR_NOFMT as a specific FFmpeg error code at the next lavu major bump

[Michael Niedermayer]

On Mon, Mar 15, 2010 at 02:14:21PM +0100, Aurelien Jacobs wrote:
> On Mon, Mar 15, 2010 at 01:20:29PM +0100, Michael Niedermayer wrote:
> > On Mon, Mar 15, 2010 at 12:08:54AM +0100, Stefano Sabatini wrote:
> > > On date Sunday 2010-03-14 20:06:34 +0100, Michael Niedermayer encoded:
> > > > On Sat, Mar 13, 2010 at 09:50:34PM +0100, Stefano Sabatini wrote:
> > > > > $subj.
> > > > 
> > > > >  error.h |    3 ++-
> > > > >  1 file changed, 2 insertions(+), 1 deletion(-)
> > > > > 2941ec2a41518b3a058586be322100cfa2be746b  0007-Change-the-definition-of-AVERROR_NOFMT-at-the-next-l.patch
> > > > > >From dc74163b0948fe0d84f7aa7a96b589c668cc60a8 Mon Sep 17 00:00:00 2001
> > > > > From: Stefano Sabatini <stefano.sabatini-lala at poste.it>
> > > > > Date: Sat, 13 Mar 2010 21:36:14 +0100
> > > > > Subject: [PATCH 7/8] Change the definition of AVERROR_NOFMT at the next libavutil major
> > > > >  bump, using an FFmpeg specific error code rather than EIILSEQ, which
> > > > >  has a quite different semantics.
> > > > 
> > > > do we really need AVERROR_NOFMT?
> > > > does this have any use outside of probe() ?
> > > 
> > > Currently grep reveals these uses:
> > > ./libavformat/utils.c:562:        err = AVERROR_NOFMT;
> > > ./libavformat/matroskadec.c:1147:        return AVERROR_NOFMT;
> > > ./libavcodec/g729dec.c:174:        return AVERROR_NOFMT;
> > > ./libavcodec/g729dec.c:227:        return (AVERROR_NOFMT);
> > > 
> > > Both the uses in matroskadec.c and g729dec.c look like they should be
> > > replaced by one of AVERROR_INVALIDDATA / AVERROR_PATCHWELCOME /
> > > AVERROR_NOTSUPP.
> > 
> > yes
> 
> I replaced the one in matroskadec.
> 
> > > As for the use in av_open_input_file(), what should it return in case
> > > of not found format error? To me it makes sense to keep this error
> > > code, and I can't find valid POSIX candidates for this type of error
> > > so I proposed to use an FFmpeg specific code (while using EILSEQ for
> > > it is plain wrong).
> > 
> > invalid data could be used
> > it feels annoying to have an error code for one single use
> 
> As we are talking about this, AVERROR_NOTSUPP also has only one single
> use, in libavformat/file.c (after its incorrect usage in wmaprodec is
> replaced, as proposed by Stefano).

it could be usefull to indicate features that are disabled at compiletime
i think

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Asymptotically faster algorithms should always be preferred if you have
asymptotical amounts of data
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20100315/e393e7fd/attachment.pgp>