[FFmpeg-devel] [PATCH/v2] Realmedia RTSP (RDT) support

[Ronald S. Bultje]

Hi,

On 7/19/07, Michael Niedermayer <michaelni at gmx.at> wrote:
>
> On Mon, Jul 16, 2007 at 12:33:04AM -0400, Ronald S. Bultje wrote:
> > Index: ffmpeg/libavformat/rtsp.c
> > ===================================================================
> > --- ffmpeg.orig/libavformat/rtsp.c    2007-07-14 14:24:43.000000000-0400
> > +++ ffmpeg/libavformat/rtsp.c 2007-07-16 00:24:38.000000000 -0400
> > @@ -609,11 +609,12 @@
> >              p++;
> >          get_word_sep(profile, sizeof(profile), "/;,", &p);
> >          lower_transport[0] = '\0';
> > -        if (*p == '/') {
> > +        if (*p == '/') { /* rtp/avp/<protocol> */
> >              p++;
> >              get_word_sep(lower_transport, sizeof(lower_transport),
> >                           ";,", &p);
> > -        }
> > +        } else /* x-pn-tng/<protocol> */
> > +            av_strlcat(lower_transport, profile,
> sizeof(lower_transport));
> >          if (!strcasecmp(lower_transport, "TCP"))
> >              th->protocol = RTSP_PROTOCOL_RTP_TCP;
> >          else
>
> this will totally break normal rtp but iam not rt*p maintainer


I'm probably missing something - why would it break? The first half of the
function should take care of normal RTP (which always has
rtp/avp/<protocol>, at least as far as I've seen). Unless someone uses tcp
as profile, outcome is always the same, no? I can handle this differently to
allow that (tcp as profile...) also, if you'd prefer.

(I'll take care of the rest of the comments, the patches are a bit rough, I
admit - will fix it up.)

Ronald