On Mon, Jun 8, 2015 at 5:28 AM, John Högberg <john.hogb...@ericsson.com> wrote: > Luca Barbato wrote: >> Please do. I thought about the alternative to support your scenario (beside >> having a callback API that I'll blog about soonish) and another way to do >> that >> is to have an AVSTREAM_EVENT_FLAG_LOST to signal that the stream is gone and >> an AVFMT_EVENT_FLAG_ACTIVE_STREAM_CHANGED to signal the fact the number of >> active stream changed. >> >> The semantics compared to metadata update is changed slightly, since we want >> to keep the AVSTREAM_EVENT_FLAG_LOST as they are until they change while >> wiping the AVFMT_EVENT_FLAG_ACTIVE_STREAM_CHANGED. >> >> This should be enough to avoid to introduce a function to explicitly to check >> and reset even if I'd consider adding it nonetheless. > > A general notification above stream level would be useful, yes. As an aside, > the new semantics are "present" rather than "lost" as per Anton's feedback. It > feels like I've submitted every possible way of implementing this right now. > :| > >> Another option is having a generic AVFMT_EVENT_FLAG_STREAM_EVENT to signal >> that the stream event flags had been set (and spare some iterations over the >> stream list). >> >> How does it sound? > > I'd prefer the generic solution. There's little to be gained by "inheriting" > the > specific flags of the underlying streams since you'll need to check them all > anyway. > >> lu
Is this patch still needed? -- Vittorio _______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel
