On 16 September 2013 00:21, Josh Allmann <[email protected]> wrote: > On 15 September 2013 07:54, Luca Barbato <[email protected]> wrote: >> --- >> libavformat/rtmpproto.c | 8 +++++++- >> 1 file changed, 7 insertions(+), 1 deletion(-) >> >> diff --git a/libavformat/rtmpproto.c b/libavformat/rtmpproto.c >> index 03f5761..e1a7782 100644 >> --- a/libavformat/rtmpproto.c >> +++ b/libavformat/rtmpproto.c >> @@ -1903,6 +1903,13 @@ static int send_invoke_response(URLContext *s, >> RTMPPacket *pkt) >> // Send onStatus(NetStream.Publish.Start) >> return write_status(s, pkt, "NetStream.Publish.Start", >> filename); >> + } else if (!strcmp(command, "play")) { >> + ret = write_begin(s); >> + if (ret < 0) >> + return ret; >> + rt->state = STATE_PUBLISHING; >> + return write_status(s, pkt, "NetStream.Play.Start", >> + filename); >> } else { >> if ((ret = ff_rtmp_packet_create(&spkt, RTMP_SYSTEM_CHANNEL, >> RTMP_PT_INVOKE, 0, >> @@ -2523,7 +2530,6 @@ reconnect: >> } else { >> if (read_connect(s, s->priv_data) < 0) >> goto fail; >> - rt->is_input = 1; >> } >> >> do { >> -- >> 1.8.3.2 >> > > IIRC for compatibility with the Flash player you'll also need to send > a (following the librtmp terminology) type-0 control message with the > stream ID. This would correspond here to (the somewhat incorrectly > named) RTMP_PT_PING, followed by a body of be16(zero), > amf32(streamid). Can supply a proper patch for this in the morning.
Never mind; this is already addressed in write_begin from the previous patch. Sorry for the noise. _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
