asterisk users - Jul 2004 - Params on SIP URI REGISTER/INVITE

We're doing some SIP development and have a question on "additional
parameters"
supplied to the register (in this case maddr= and the non-standard clport= in
our example below).

What we're experiencing is the INVITE doesn't included these parameters
and they get dropped when the INVITE is sent to the 10.1.1.97 address.

Ideas?  Supported?  SIP Bug?

REGISTER sip:test1.mydomain.com SIP/2.0
Via:
SIP/2.0/UDP10.1.1.97:5060;branch=z9hG4bKd1f1eb5acc28043b83a28ca2ee1e5f15,SIP/2.0/UDP
192.168.0.2:5061;branch=z9hG4bk-8c166b93
From: "JohnDoe"
<sip:2016@test1.mydomain.com>;tag=6f0ecbcb3a5e62c4
To: "JohnDoe" <sip:2016@test1.mydomain.com>
Call-ID: 6c4a901f-fb034a80@192.168.0.2
CSeq: 1 REGISTER
Max-Forwards: 69
Contact:
"JohnDoe"<sip:8994f3ae412133e2056757d2fcccfbed5f9642da@10.1.1.97:5060;maddr=192.168.0.2;clport=5061>;expires=3600
User-Agent: Sipura/SPA2000-1.0.15
Allow: ACK,BYE,CANCEL,INFO,INVITE,NOTIFY,OPTIONS,REFER
Content-Length: 0

Lenny Tropiano / asterisk.org Mailing list wrote:
> We're doing some SIP development and have a question on
"additional parameters"
> supplied to the register (in this case maddr= and the non-standard clport=
in
> our example below).
> 
> What we're experiencing is the INVITE doesn't included these
parameters
> and they get dropped when the INVITE is sent to the 10.1.1.97 address.
> 
> Ideas?  Supported?  SIP Bug?
It's a bug. This is the same problem as the Snom Line= problem and some
middleman box traversal errors. The SIP proxy should really send the same
Contact: header as the client provided.

By this time, this is a well-known hidden feature of chan_sip2. I've added
some code in that test version of the SIP channel to save the contact:
supplied by the client on register and using all of that on subsequent
invites. However, I haven't got feedback enough if that patch worked as
expected or not to port it back into chan_sip in CVS.

Sadly, I've been on holiday and during that period chan_sip and Asterisk
changed a lot (not any surprice) so I have to update chan_sip2 quite a lot
to compile with cvs_head. This will propably happen next week, after we
got the Astricon registration up and running :-)

I would very much like feedback on that part of chan_sip2 in order to
decide whether I should backport it to chan_sip in CVS.

/Olle