Re: [Live-devel] Cleanly close RTSP Server

2019-02-19 Thread Ross Finlayson
> On Feb 20, 2019, at 7:53 AM, Kevin Bailey wrote: > > My particular use case has my program started and stopped rapidly in short > periods of time. But I can't start and stop it like this if I can't quickly > re-bind to the rtsp port. Maybe this doesn't have anything to do with my > clean u

Re: [Live-devel] UDP RTSP Stream

2019-02-19 Thread Ross Finlayson
> The part you’re concerned about is the stream that’s negotiated via RTSP, > which is most probably going to be either RTP — which almost always runs over > UDP — or raw UDP. You have to go out of your way in Live555 to get > RTP-over-HTTP, which is the only TCP-based protocol I believe Live55

Re: [Live-devel] UDP RTSP Stream

2019-02-19 Thread Warren Young
A couple of clarifications: On Feb 19, 2019, at 2:23 PM, Warren Young wrote: > > ...RTP-over-HTTP…is the only TCP-based protocol I believe Live555 supports. I mean the only A/V streaming protocol, of course. I’m excluding RTSP in that characterization. > you’ll need to sniff the RTSP protoco

Re: [Live-devel] UDP RTSP Stream

2019-02-19 Thread Warren Young
On Feb 19, 2019, at 11:57 AM, Kevin Bailey wrote: > > I wanted my RTSP stream to be UDP based not TCP based to take advantage of > the latency advantage UDP has over TCP. However I noticed the following when > I took a look at netstat while my program was running: > > ~# netstat | grep 8554 >

Re: [Live-devel] Cleanly close RTSP Server

2019-02-19 Thread Warren Young
On Feb 19, 2019, at 11:53 AM, Kevin Bailey wrote: > > Maybe this doesn't have anything to do with my clean up code? Indeed it doesn’t. The “problem” is that Live555 doesn’t set SO_REUSEADDR, which it is technically correct to do: https://stackoverflow.com/a/3233022/142454 You can set it

Re: [Live-devel] Cleanly close RTSP Server

2019-02-19 Thread Kevin Bailey
And I am also noticing now that netstat says that the rtsp port I am using is in the following state "FIN_WAIT2". After I kill my client then it moves to "TIME_WAIT". Which leads me to believe that the clean up code listed in testProgs/testRTSPClient.cpp does not sufficiently close the rtsp connect

[Live-devel] UDP RTSP Stream

2019-02-19 Thread Kevin Bailey
I wanted my RTSP stream to be UDP based not TCP based to take advantage of the latency advantage UDP has over TCP. However I noticed the following when I took a look at netstat while my program was running: ~# netstat | grep 8554 tcp0 0 x.x.x.x:8554wsip-x-x-x-x:39224 ESTABLISHED

Re: [Live-devel] Cleanly close RTSP Server

2019-02-19 Thread Kevin Bailey
My particular use case has my program started and stopped rapidly in short periods of time. But I can't start and stop it like this if I can't quickly re-bind to the rtsp port. Maybe this doesn't have anything to do with my clean up code? Is it related to this: https://live-devel.live555.narkive.co

Re: [Live-devel] Cleanly close RTSP Server

2019-02-19 Thread Ross Finlayson
In general, you can clean up by deleting things in the reverse order from which they were created. But why don’t you just call “exit()”? Why do you need to keep using the process after you’ve gotten rid of the RTSP server? Ross Finlayson Live Networks, Inc. http://www.live555.com/ _

[Live-devel] Cleanly close RTSP Server

2019-02-19 Thread Kevin Bailey
Hi all, in the code for testProgs/testOnDemandRTSPServer.cpp there is no code specified that should be used to clean up the RTSP server cleanly. testProgs/testRTSPClient.cpp actually does specify some code to clean up the program RTSP client. I have tried adding "Medium::close" and an "eventLoopWa