[SCM] libav/experimental: rtpdec: Don't call the depacketizer to return more data unless it actually said it has more data

[siretart at users.alioth.debian.org]

The following commit has been merged in the experimental branch:
commit f6e138b4f4fbfb4caf50bcc628d48b6e4cdca269
Author: Martin Storsjö <martin at martin.st>
Date:   Wed Oct 13 08:13:07 2010 +0000

    rtpdec: Don't call the depacketizer to return more data unless it actually said it has more data
    
    It may have returned a negative number for an error (e.g. AVERROR(EAGAIN),
    if more data is required for it to be able to return a complete packet).
    
    Originally committed as revision 25458 to svn://svn.ffmpeg.org/ffmpeg/trunk

diff --git a/libavformat/rtpdec.c b/libavformat/rtpdec.c
index 51583c0..240229a 100644
--- a/libavformat/rtpdec.c
+++ b/libavformat/rtpdec.c
@@ -613,10 +613,10 @@ static int rtp_parse_one_packet(RTPDemuxContext *s, AVPacket *pkt,
     int rv= 0;
 
     if (!buf) {
-        /* If parsing of the previous packet actually returned 0, there's
-         * nothing more to be parsed from that packet, but we may have
+        /* If parsing of the previous packet actually returned 0 or an error,
+         * there's nothing more to be parsed from that packet, but we may have
          * indicated that we can return the next enqueued packet. */
-        if (!s->prev_ret)
+        if (s->prev_ret <= 0)
             return rtp_parse_queued_packet(s, pkt);
         /* return the next packets, if any */
         if(s->st && s->parse_packet) {

-- 
Libav/FFmpeg packaging