[FFmpeg-devel] [PATCH] Make ffmpeg.c use print_error() when it can't read a ffserver stream

Stefano Sabatini stefano.sabatini-lala
Sat Jun 7 23:48:18 CEST 2008


On date Saturday 2008-06-07 23:00:41 +0200, Michael Niedermayer encoded:
> On Sat, Jun 07, 2008 at 08:04:17PM +0200, Stefano Sabatini wrote:
> > Hi, as in subject.
> > 
> > This is more consistent with the other use of av_open_input_file() in
> > ffmpeg.c in opt_input_file, which is:
> > 
> >     /* open the input file with generic libav function */
> >     err = av_open_input_file(&ic, filename, file_iformat, 0, ap);
> >     if (err < 0) {
> >         print_error(filename, err);
> >         av_exit(1);
> >     }
> > 
> > also the error message reported by print_error() should be at least
> > theoretically more accurate than the generic one currently used in
> > read_ffserver_streams():
> > 
> > fprintf(stderr, "Could not read stream parameters from '%s'\n", filename);
> > 
> > This will lead to error messages of the kind:
> > stefano at geppetto ~/s/ffmpeg> ffmpeg -f video4linux -s 640x480 -r 25 -i /dev/video0 http://localhost:8090/feed1.ffm
> > [...]
> > Input #0, video4linux, from '/dev/video0':
> >   Duration: N/A, start: 1212861538.055957, bitrate: 92160 kb/s
> >     Stream #0.0: Video: rawvideo, yuv420p, 640x480, 92160 kb/s, 25.00 tb(r)
> > http://localhost:8090/feed1.ffm: I/O error occurred
> > Usually that means that input file is truncated and/or corrupted.
> > 
> > BTW, I also think that the EIO error code returned in this case when
> > we try to open a closed TCP port is still too vague, I'd like
> > something more meaningful, which leads to the old discussion about the
> > error codes managed by FFmpeg.
> > 
> > Do you have you any hint (maybe just another more specific error code
> > describing the current situation)?
> > 
> > Best regards.
> > -- 
> > FFmpeg = Frenzy & Fast MultiPurpose EnGine
> 
> > Index: ffmpeg.c
> > ===================================================================
> > --- ffmpeg.c	(revision 13687)
> > +++ ffmpeg.c	(working copy)
> > @@ -423,14 +423,16 @@
> >      return ret;
> >  }
> >  
> > -static int read_ffserver_streams(AVFormatContext *s, const char *filename)
> > +static void read_ffserver_streams_or_die(AVFormatContext *s, const char *filename)
> >  {
> >      int i, err;
> >      AVFormatContext *ic;
> >  
> >      err = av_open_input_file(&ic, filename, NULL, FFM_PACKET_SIZE, NULL);
> > -    if (err < 0)
> > -        return err;
> > +    if (err < 0) {
> > +        print_error(filename, err);
> > +        av_exit(1);
> > +    }
> 
> Wwen all else is equal (like here) i prefer 'return some error' over exit()

Mmh.., so what about this one? It still addresses my main concern
which is to use the print_error() rather than a generic error message.

Regards.
-- 
FFmpeg = Fiendish and Frenzy MultiPurpose EnGraver
-------------- next part --------------
A non-text attachment was scrubbed...
Name: use-print-error-when-reading-ffserver-streams-01.patch
Type: text/x-diff
Size: 639 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20080607/76a1b437/attachment.patch>



More information about the ffmpeg-devel mailing list