[Serusers] loose_route()

[JF]

Thank you, Jan.
So it seems like a user agent problem. UA1 is inserting in the
in-dialog request the initial request's Request-URI and not the remote
target for the dialog (UA2's contact) as per RFC3261.

JF

On 1/30/06, Jan Janak <jan at iptel.org> wrote:
> The decision whether after_loose or after_strict will be called is based on the
> Request-URI of the request. If the Request-URI hostname part contains
> the IP or hostname of the proxy which received the request then it
> assumes that the previous hop was strict router. If the previous hop was
> loose router then the Request-URI should contain the Contact header of
> UA2 and that should not be matched by (uri==myself).
>
>   Jan.
>
> On 30-01-2006 19:34, JF wrote:
> > Hi,
> >
> > One of my (Open)SER machines (P1) is acting as registrar and so I have
> > alias=mydomain.tld in the script; it is used in the (uri == myself)
> > check to handle REGISTER requests.
> >
> > I have a setup like this:
> >
> > UA1 (user1 at mydomain.tld) - P2 - P1 - P3 - P1 - P2 - UA2 (user2 at mydomain.tld)
> > (all proxies are record-routing except P3)
> >
> > For initial requests (INVITE from UA1 to UA2) all is fine. But I'm
> > having problems with the loose_route() function, because when an
> > in-dialog (subsequent) request comes into P1 (from UA1, P2) with a
> > Route header containing (<P1>, <P1>, <P2>), the after_strict()
> > function is called, when I would expect the after_loose() function to
> > be called since P1 is a loose router.
> > Within loose_route(), is_preloaded() returns 0, which is correct. But
> > then is_myself returns true because of the alias...
> >
> > Is this a bug in loose_route() which doesn't account for this kind of
> > setup (Routeing in-dialog requests through the registrar)? Maybe this
> > could be avoided, if there was a parameter forcing loose_route to
> > treat previous hop as a loose router... Or am I just doing something
> > wrong?
> >
> > Thanks in advance,
> > JF
> >
> > _______________________________________________
> > Serusers mailing list
> > serusers at lists.iptel.org
> > http://lists.iptel.org/mailman/listinfo/serusers
>