GNOME Bugzilla – Bug 734471
asfdemux: Should pass seek events upstream first
Last modified: 2014-09-16 15:24:01 UTC
asfdemux currently discards any seek event it doesn't support. It should pass them upstream instead giving a chance for another element (like dlnasrc) to handle it.
Created attachment 282892 [details] [review] asfdemux: send unsupported seek events upstream Upstream elements like dlnasrc may be able to handle the seek so we should give them a chance rather than discarding the seek event.
Please note that we have been using 'element: xyz' in bug titles for the last couple of years instead of '[element] xyz'. I think what you want to do here is right in principle, but I'm not sure if the patch is right. asfdemux (and any demuxer, at least when operating in push mode) should already try sending a seek upstream first before attempting to process it itself. In your patch you do that as a fallback, but it should be tried first, and I'm sure it does that already for mms and rtsp, so question is why doesn't it do it for your case? Is your stream marked as non-seekable in the headers? Perhaps that's something that should be fixed too?
As Tim mentioned, the upstream seek should be attempted first, regardless of any other settings. If upstream is not seekable (for whatever reason), the upstream element should properly refuse it. Can you flip the logic and try again ?
Created attachment 284171 [details] [review] asfdemux: first try to delegate TIME seek events upstream We should give a chance to upstream to handle TIME seek events before trying to handle it in asfdemux. Fix a bug where fast backward seeks where discarded right away because they were not supported by asfdemux.
asfdemux: first try to delegate TIME seek events upstream We should give a chance to upstream to handle TIME seek events before trying to handle it in asfdemux. Fix a bug where fast backward seeks where discarded right away because they were not supported by asfdemux. https://bugzilla.gnome.org/show_bug.cgi?id=734471