Re: Does an event generated by strict=state stop signals?

Thank you, Tom and Christian, for your replies.  The actual case that
prompted my question was one in which conclusion 3 applies.  As you
said, "unless provisioned otherwise" is very relevant, as we have had
interoperations in which it is necessary to assume a default of
strict=state.

Regards,
Raphael

-----Original Message-----
From: Christian Groves [mailto:[email protected]] 
Sent: Wednesday, 16 June 2010 2:56 AM
To: Raphael Tryster
Cc: megaco ietf
Subject: Re: [Megaco] Does an event generated by strict=state stop
signals?

Hello Raphael,

I think it is a real event so the behaviour would be to send the notify 
and stop the dial tone. So I think conclusion 2 is the one to take. 
However I think conclusion 3 is also applicable. Megaco v1, v2, or v3 
really is only for the syntax and procedures surrounding that. In the 
case of the Analog Line Supervision Package (al) the package version has

remained at v1 for all Megaco versions. Therefore the inclusion of the 
default should be seen as correcting an omission. So the MG should 
assume strict=exact if strict is not specified (unless provisioned 
otherwise :-) ).

Regards, Christian

On 15/06/2010 5:26 PM, Raphael Tryster wrote:
>
> Megaco stipulates that detection of an event should stop any signal 
> that is playing.  Does an existing hook state, reported immediately 
> with init=true because the MGC requested strict=state, count as an 
> event for the purpose of stopping signals?
>
> If the answer is yes, then we could have the following scenario:
>
> 1. MG sends Notify al/of.
>
> 2. MGC sends Modify, requesting signal cg/dt, and events 
> al/of{strict=state},al/on(strict=state}.
>
> 3. MG immediately sends Notify al/of{init=true}, and stops the dial
tone!
>
> A variation of the above is when the strict parameter is omitted, but 
> the MG decides to assume strict=state because it is working with 
> Megaco version 1 or 2, where no default is specified.
>
> Possible conclusions:
>
> 1. This is not a real event and the signal should not be stopped.
>
> 2. This is a real event and the MGC is stupid to send such a command.
>
> 3. This is a real event, and if strict is not specified, the MG must 
> assume strict=exact.
>
> I await the experts' opinion.
>
> Raphael Tryster
>
>
> _______________________________________________
> Megaco mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/megaco
>