[asterisk-dev] Dialog Subscriptions, Bug 6728 and snom dialog; purpose=call-completion

Kevin P. Fleming kpfleming at digium.com
Fri Feb 12 09:45:00 CST 2010


Martin Koenig wrote:

> I believe relevant in this case is RFC3265, 7.2.1.
> 
> <snip>
> "Event" header
> 
>    Event is added to the definition of the element "message-header" in
>    the SIP message grammar.
> 
>    For the purposes of matching responses and NOTIFY messages with
>    SUBSCRIBE messages, the event-type portion of the "Event" header is
>    compared byte-by-byte, and the "id" parameter token (if present) is
>    compared byte-by-byte.  An "Event" header containing an "id"
>    parameter never matches an "Event" header without an "id" parameter.
>    No other parameters are considered when performing a comparison.
> 
>       Note that the forgoing text means that "Event: foo; id=1234" would
>       match "Event: foo; param=abcd; id=1234", but not "Event: foo" (id
>       does not match) or "Event: Foo; id=1234" (event portion does not
>       match).
> 
>    This document does not define values for event-types.  These values
>    will be defined by individual event packages, and MUST be registered
>    with the IANA.
> 
>    There MUST be exactly one event type listed per event header.
>    Multiple events per message are disallowed.
> </snip>
> 
> From my point of view it appears that Bug 6728 actually broke the rfc 3265
> compliant implementation in asterisk.

Note that the text you quoted does not apply to SUBSCRIBE requests, it
only applies to the responses to SUBSCRIBE requests and subsequent
NOTIFY messages generated by the subscription.

Can you post the entire SIP header set of each of these two requests
(suitably anonymized, if you need to)? It would be good to be able to
see the RURI and Call-ID as well to know whether these are supposed to
be independent subscriptions or not.

-- 
Kevin P. Fleming
Digium, Inc. | Director of Software Technologies
445 Jan Davis Drive NW - Huntsville, AL 35806 - USA
skype: kpfleming | jabber: kpfleming at digium.com
Check us out at www.digium.com & www.asterisk.org



More information about the asterisk-dev mailing list