15 Mar
2005
15 Mar
'05
11:02 p.m.
Ahhh, I found my mistake, but can someone verify my thought process??
I'm using mediaproxy and I only proxy RTP streams if the caller or the
callee are behind a NAT device.
So during a re-INVITE, I cannot just t_relay() the message because the
RTP streams will end up at unreachable IP addresses if either SIP UA
is NATed.
I can use client_nat_test("3") to test the caller's NAT settings, and
for the callee I have to use lookup("location") and test my nat_flag
setting.
But do I need to call revert_uri() after lookup("location") when
processing re-INVITEs???
Regards,
Paul
On Tue, 15 Mar 2005 16:17:38 -0500, Java Rockx <javarockx(a)gmail.com> wrote:
Yep. I agree.
But does this mean for **any** SIP message that loose_route() returns
TRUE, I simply call t_relay() and then stop processing via a break
command?
I can't remember, but I thought I ran in to some cases where even when
loose_route() returned TRUE, I still had to do some other processing.
Regards,
Paul
On Tue, 15 Mar 2005 20:13:40 +0200, Juha Heinanen <jh(a)tutpro.com> wrote:
Java Rockx writes:
I thought I had mine right, but then I realized
that some re-INVITE
messages were not being handled properly and calls were being
disconnected.
you should not use or need to use lcr modules functions for re-invites.
re-invites are in-dialog requests that are forwarded based on route
headers.
-- juha