> Generally, the mapping of an incoming request URI to obtain a 
> key into the DB of registered contact is a matter of local
> policy,

Agreed.

> but it needs to be done in a consistent and coherent
> manner, in order to avoid interop problems. This particular
> item, use of user=phone, has been a source of trouble for some
> time.

Well, there needs to be alignment between a Registrar's
usage of the address-of from the To in a REGISTER, and
corresponding lookups based on the Request-URI, or
whatever.

Given this and some other resolutions, it seems to me
that strictly only the user and host parts can be of
significance now.  Which is nice.

> I believe that the correct thing to do in this particular 
> case is that the registrar should not user the user=phone
> when using the address-of-record from the REGISTER. Furthermore,
> it ought to be stripped from the incoming r-uri,

I'm so not against this.
But I fear that I am missing something, since I read it
as conflicting somewhat with your summary of the fix for #281:

> There were two comments on this proposal, from Jo Hornsby
> and from Jon Peterson (enclosed below). Both support the
> proposal. Jon recommended some additional text on reasonable
> choices for the "transformation" as he describes it. I think
> thats OK; discussing the issue in particular with user=phone
> (whether or not that is part of the key, and I believe 
> it generally is).

Or do I have post-New Year Celebrations "flu"? &:)


 - Jo.
_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors

Reply via email to