[Live-devel] Opportunity for unpredictable behaviour in MPEG2TransportStreamFramer

[Tim J Shackleton]

Hi Ross,

I have been working on a DeviceSource that I'm feeding into an RTP sink 
via the MPEG2TransportStreamFramer.

The DeviceSource never runs out of data to provide the framer, so it 
runs for extended periods of time.  What I am finding is that after a 
reasonable period of operation, suddenly the task starts spitting out 
MPEG frames as fast as the network can carry them.  It's as though the 
framer is incorrectly calculating the playout rate required...

I have noticed that such vars as fTSPCRCount, fTSPacketCount, 
lastPacketNum and so forth are unsigned ints giving them a wrapping 
capacity of 4 and a bit billion.  If a wrap happens, I am expecting the 
calculations to become wildly wrong.  I have stood up a separate 
instance of the live555 source tree exchanging the variables involved to 
doubles in order to see if the spurious behavior stops.  A hack rather 
than a fix, I know.

If this wrap is in fact causing the issue, would you consider it a bug?  
If so, is this something you would prefer to address in a later release 
or something I should provide a patch for?

Thanks again for your hard work and patience.

Regards,

Tim