Hi, On Fri, Apr 6, 2012 at 2:56 PM, Martin Storsjö <[email protected]> wrote: > On Fri, 6 Apr 2012, Ronald S. Bultje wrote: >> On Fri, Apr 6, 2012 at 2:45 PM, Ronald S. Bultje <[email protected]> >> wrote: >>> On Fri, Apr 6, 2012 at 7:57 AM, Martin Storsjö <[email protected]> wrote: >>>> On Fri, 6 Apr 2012, Ronald S. Bultje wrote: >>>>> On Fri, Apr 6, 2012 at 7:03 AM, Martin Storsjö <[email protected]> >>>>> wrote: >>>>>> >>>>>> >>>>>> Some streams don't contain these. >>>>>> --- >>>>>> I also tried changing the code in ff_asf_get_packet to seek back >>>>>> and try to restart with the assumption that there isn't any start >>>>>> marker (the code with a FIXME remark), but it broke in various >>>>>> odd cases, this seems to be a more straightforward solution >>>>>> which shouldn't break the existing error tolerance/resync code. >>>>>> >>>>>> libavformat/asfdec.c | 21 ++++++++++++++++++++- >>>>>> 1 file changed, 20 insertions(+), 1 deletion(-) >>>>> >>>>> >>>>> >>>>> Example of such a stream? >>>> >>>> >>>> >>>> rtsp://208.86.19.32:8000/crtn is such a stream. >>> >>> >>> I remember from when I wrote this stuff that in general, ASF-in-RTSP >>> doesn't need these markers at all anyway, and each RPT packet simply >>> contains an ASF packet. My original implementation skipped that >>> altogether and directly called parse_packet() instead of get_packet(). >>> Maybe we should go back to that approach by default in ASF/RTSP? >> >> >> Behaviour, not approach, i.e. set this option to TRUE by default from >> the RTP/ASF parser. > > > Yeah, that's exactly what I do in patch 2/2 in this series.
OMG I feel to stupid now. Patch (both) OK. Ronald _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
