On 11/20/2013 08:59 PM, Ross Finlayson wrote: If the continueAfterOPTIONS command is late (longer than the liveness timeout) I assume the camera would time out and shutdown the session? I don't see anywhere in continueAfterLivenessCommand that it handles this?
Phew... 1/ The interval between two successive "OPTIONS" commands will almost never exceed the "timeout=" parameter that the server reported. (A future release of the software will ensure that it *never* exceeds this - because the interval will be randomly chosen from "[delayMax/2,delayMax-1)" instead of the current "[delayMax/2,delayMax)".) I made a similar change for the delayMax, but instead subtracted 10, because I would see latency when I was logging OPTION commands. I have a extremely busy WIFI network for testing. Thanks for that change as it is along the lines I was think and confirms what I was seeing. 2/ In any case, this shouldn't matter, because - as I've said AD NAUSEAM - the proxy server will send RTCP "RR" packets far more frequently than this, and standards-compliant (i.e., correctly implemented) server will use these as an indication that the proxy server is still alive. 3/ Should the camera "time out the shutdown the session" - for any reason - then the subsequent "OPTIONS" command from the proxy server will fail, and the proxy server will then reopen the connection. That's the PRIMARY PURPOSE of the periodic "OPTIONS" commands: To check whether the server still thinks that the session is alive. Ross Finlayson Live Networks, Inc. http://www.live555.com/
_______________________________________________ live-devel mailing list [email protected] http://lists.live555.com/mailman/listinfo/live-devel
