On 1/2/13 9:02 AM, Brett Tate wrote:
>> Contact:<sip:anonymous.invalid;phone-context=+1
> @domain.com:5060;maddr=10.10.10.10;transport=udp;user=phone>
>>
>> Please note I have changed the domain and the IP.
>>
>> We reject the INVITE with SIP/2.0 400 Bad Request
>> - 'Malformed/Missing Contact field'
>>
>> In RFC3261 I see that the INVITE must contain a
>> Valid Contact Header to be used for subsequent
>> requests, but I'm not sure if the above is valid.
>>
>> Can anyone comment?
>
> The contact's validity depends upon the ambiguous meaning of user=phone. The
> user=phone is present; however the user portion is not a valid
> telephone-subscriber per RFC 2806/3966.
>
> Unfortunately since I don't think that sipcore has clarified the meaning of
> user=phone, you should likely accommodate user=phone without requiring the
> user portion to be present and/or decode as telephone-subscriber.
ISTM that the URI is *syntactically* valid, though perhaps semantically
invalid.
But who should detect this semantic invalidity?
The user=phone is all about interpretation of the user part of the URI.
That interpretation is domain-specific. So I will argue that only a
server responsible for the domain should be making such a judgement
about the user part of the URI. Servers in other domains don't need need
to know this to route to the specified domain.
And in this case, the server that is objecting doesn't even need to
resolve the domain, since maddr overrides that. If the maddr were not
present, then the server would be unable to resolve the domain, and
should fail the request on that basis.
Thanks,
Paul
_______________________________________________
Sip-implementors mailing list
[email protected]
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
