[Live-devel] DeviceSource.cpp impacts on sockets

[Ross Finlayson]

>I'm adapting live555 to stream from a live source and using 
>testMPEG2TransportStreamer.cpp and 
><http://www.live555.com/liveMedia/faq.html#liveInput>http://www.live555.com/liveMedia/faq.html#liveInput 
>as reference.
>
>I got all necessary code to startup a RTP server and start playing 
>and put on my source code.
>
>As explained on the FAQ, i substitue ByteStreamFileSource 
>and MPEG2TransportStreamFramer layers by the DeviceSource one.

That's correct.  However, it's also a good idea to deliver - to the 
"RTPSink" object - a sufficiently large chunk of MPEG Transport 
Stream data to make efficient use of each outgoing RTP packet. 
(Packing just a single 188-byte Transport 'packet' into each outgoing 
RTP packet will work, but is inefficient.)  To overcome this, either 
have your 'DeviceSource' accumulate a reasonable number (usually 7) 
of Transport 'packets' before delivering them, or else insert a 
"MPEG2TransportStreamAccumulator" object (see, for example, the 
source code in "wis-streamer": http://www.live555.com/wis-streamer/ )


>Using DDD to degub a segmentation fault bug, i found thats is caused 
>by a call to socketErr() made by writeSocket() ( more precisely, "if 
>(setsockopt(socket, IPPROTO_IP, IP_MULTICAST_TTL,(const char*)&ttl, 
>sizeof ttl) < 0)" line call).

If this is correct, then it may be caused by a bad 'socket' parameter 
(perhaps a socket that has already been closed.  However, I don't see 
what could be causing this...


>How could DeviceSource changes impact on socket issues??

Unfortunately I don't see how they could.  Sorry.
-- 

Ross Finlayson
Live Networks, Inc.
http://www.live555.com/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.live555.com/pipermail/live-devel/attachments/20110425/54ff3899/attachment-0001.html>