Juha,

o Juha Heinanen on 09/29/2013 08:39 AM:
> it turned out that it is sems ivr that is broken.  when connectCallee
> call is connected, sems sends out two re-invites with different branch
SEMS detects changed SDP on the other side, so it re-invites the
callee leg after receiving the SDP in 200 from caller side. (This
probably happens then one time more, because the other side may change
SDP as well.). Those part of the call flow looks perfectly OK.

> values to caller and when sems gets bye from caller, it sends out two
> byes to callee.


BYE is handled end-to-end now, so you don't need to call
terminateOtherLeg in onBye(self). remove this one and there's only one
BYE sent out. (The other effect of this is that in ivr you can't any
more prevent the callee leg from getting disconnected if caller leg
disconnects.)

would you care to start a 1.5 to 1.6 migration doc which documents
these changes? Thanks.

Stefan

> 
> the problem can be reproduced by this minimal test app when caller hangs
> up after call is connected to callee.
> 
> ----------------------------------------------------------------------------------
> from log import *
> from ivr import *
> 
> WELLCOME_MSG = "/home/jh/test/gong.wav"
> CALLEE_URI = "sip:[email protected]"
> 
> class IvrDialog(IvrDialogBase):
>     
>     def onSessionStart(self):
> 
>         info("starting b2bua test ...")
> 
>         self.setNoRelayonly()
>         
>         self.welcome_msg = IvrAudioFile()
>         self.welcome_msg.open(WELLCOME_MSG, AUDIO_READ)
>         self.enqueue(self.welcome_msg,None)
> 
>     def onEmptyQueue(self):
>         
>         info("connecting " + CALLEE_URI + " to " + CALLEE_URI)
> 
>         self.disconnectMedia()
>         self.mute()
> 
>         self.connectCallee(CALLEE_URI, CALLEE_URI)
> 
>         return
> 
>     def onBye(self):          
> 
>         self.terminateOtherLeg()
>         self.stopSession()
> -----------------------------------------------------------------------------------
> 
> if needed, i can provide debug and pcap of the test.
> 
> -- juha
> _______________________________________________
> Semsdev mailing list
> [email protected]
> http://lists.iptel.org/mailman/listinfo/semsdev
> 

_______________________________________________
Semsdev mailing list
[email protected]
http://lists.iptel.org/mailman/listinfo/semsdev

Reply via email to