> One last question on this, as I'm in the middle of reorganizing the code.
> Shouldn't it be safe to create the rtspClient object in another thread?
Once again, NO! "RTSPClient" objects - like all subclasses of class "Medium" -
update shared data structures (stored within the "UsageEnvironment" object)
when they are constructed. A single "UsageEnvironment" object MUST NOT be
accessed from multiple threads (except that the "triggerEvent()" member
function may be called from a non-LIVE555-event-loop thread).
> Having to create the rtspClient objects using triggers would introduce
> additional callbacks and making passing arguments a bit more complex.
Nonetheless, if you are creating these "RTSPClient" objects dynamically, then
the creation of each of these objects has to be done as an 'event' - i.e., done
from a handler function that's called from the LIVE555 thread. And the only
way to signal this from an external thread is to call "triggerEvent()", or to
use the 'watch variable' mechanism.
> This is now the "main startup code":
>
> void InitStreamEngine()
> {
> TaskScheduler* scheduler = BasicTaskScheduler::createNew();
> usageEnvironment = BasicUsageEnvironment::createNew(*scheduler);
>
> myStartStreamEvent =
> usageEnvironment->taskScheduler().createEventTrigger((TaskFunc*)StartStreamEvent);
> myStopStreamEvent =
> usageEnvironment->taskScheduler().createEventTrigger((TaskFunc*)StopStreamEvent);
> mySeekAbsoluteEvent =
> usageEnvironment->taskScheduler().createEventTrigger((TaskFunc*)SeekAbsoluteEvent);
>
> usageEnvironment->taskScheduler().doEventLoop(&eventLoopWatchVariable);
> }
This looks fine.
Ross Finlayson
Live Networks, Inc.
http://www.live555.com/
_______________________________________________
live-devel mailing list
[email protected]
http://lists.live555.com/mailman/listinfo/live-devel
