Hi Scott, Thanks a lot for your contribution. That is indeed a bug in rtpbin. It should be able to accept post-PLAYING request pads. I will review your patch and make sure there isn't anything else that needs to be taken into consideration. Then I'll apply it upstream.
Best regards, Philippe On Fri, 12 Jan 2007 23:41:02 -0700 Scott Zuk <[EMAIL PROTECTED]> wrote: > Hi Philippe, > > Ok, I've done some more debugging and what I found was that when I attached a > source to the rtpbin 'sink' pad after the rtpbin had been set to PLAYING, the > gst_rtp_bin_setup_send_elements function would never get called and no data > would hit the network. I also discovered that the reverse issue also exists; > if a sink is attached to the rtpbin 'src' pad after setting rtpbin to > PLAYING, it will not be able to receive any data from the network. I have > attached another test program that demonstrates this second case. > > The reason for these two problems is that when rtpbin changes state from NULL > to READY, it checks to see if at least one of the source or sink request pads > were created before calling gst_rtp_bin_setup_elements. If only one of the > pads was created, than only one of send or recv elements will be created and > initialized. If at a later point in time, an element is linked to the > previously unlinked pad on rtpbin, a request pad will be created in > gst_rtp_bin_request_new_pad however the underlying elements that should read > or write to these pads do not exist! > > I have attached a patch for rtpbin that will create these elements when the > request pads are created if necessary. I don't know if this is the best way > to solve the problem but maybe it will give you a hint for better way if it > exists. So far with my own testing the patch seems to be ok as is. > > If you compile and run my two test programs without my patch, you should see > that the payload type callbacks are only reached when using a fake source or > sink, eg. > ./rtpbinfakesinktest --use-fakesink > ./rtpbinfakesinktest > ./rtpbinfakesrctest --use-fakesrc > ./rtpbinfakesrctest > > With my patch, the callbacks are reached in all cases. > > Regards, > ~Scott > > > On Thursday 11 January 2007 10:24, Philippe Khalaf wrote: > > Hi, > > > > I don't really have time to look at your stuff but you should do some > > basic gstreamer debugging. Run your program with : > > GST_DEBUG="rtpbin:5, rtpsend:5, rtprecv:5" and look for any suspicious > > messages. > > > > Regards, > > Philippe > > > > On Wed, 10 Jan 2007 00:22:03 -0700 > > > > Scott Zuk <[EMAIL PROTECTED]> wrote: > > > Hi, > > > > > > I'm trying to use the rtpbin element in a pipeline that I want to set up > > > initially to receive rtp data but change to send and receive when a > > > destination and send codec is known. I have attached a small test > > > program that shows what I am doing. > > > > > > Basically, I create a pipeline and add rtpbin and rtpdemux elements and > > > set the pipeline to playing to start receiving data. After a few seconds > > > I add an audio source, encoder and payloader to the pipeline, link them > > > up and set a destination ip and port. The problem though is that no rtp > > > data ever gets to the network and hence the rtp payload callbacks are > > > never called (I'm sending to the listen port). Not sure why this is > > > happening. > > > > > > However, if I add a fakesrc element to the pipeline and link it to the > > > sink pad of the rtpbin and remove it immediately after setting the > > > pipeline to playing, sending rtp audio will then be successful. It seems > > > like a hack to me though. You can activate this behavior in my test > > > program by passing the --use-fakesrc argument. It seems that linking the > > > fakesrc to the rtpbin sink causes something to be initialized differently > > > but I don't know what. Any ideas? > > > > > > BTW, I'm using a copy of the latest rtpbin from darcs compiled without > > > jrtp and jingle and gstreamer 0.10.10 > > > > > > Thanks, > > > ~Scott > > > > ------------------------------------------------------------------------- > > Take Surveys. Earn Cash. Influence the Future of IT > > Join SourceForge.net's Techsay panel and you'll get the chance to share > > your opinions on IT & business topics through brief surveys - and earn cash > > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > > _______________________________________________ > > Farsight-devel mailing list > > Farsight-devel@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/farsight-devel ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Farsight-devel mailing list Farsight-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/farsight-devel
