[Asterisk-Dev] Changes to SIP

[Michael Baird]

Agreed.

Regards
MIKE


On Sat, 2003-04-05 at 17:08, Gregg Lebovitz wrote:
> Mark,
> 
> I would like to see the results for your discussion with james. Can you
> post, once you come to a conclusion?
> 
> Gregg
> 
> On Sat, 2003-04-05 at 16:45, Mark Spencer wrote:
> > > (o)  The From header needs to be set to the account that performs the
> > > REGISTER. I currently have a small hack that allows me to specify what the
> > > >From header should be in the sip.conf file for that peer.
> > 
> > Can you provide an example?  I believe you can use "callerid=" in the
> > "general" section to set the default callerid.  I am not sure this is the
> > problem you are trying to solve though.
> > 
> > > (o)  When a BYE is sent the authorization digest is not sent, cisco replies
> > > with Proxy auth required but by then Asterisk has already destroyed the
> > > channel. From a glance it looks like the cisco phones send the same auth
> > > digest as what was sent for the INVITE. The phone does not go through the
> > > whole transaction of trying without auth, then receiving a proxy-auth
> > > required error, then finally sending the auth digest.
> > 
> > Are we permitted to send the same authorization digest as we sent on the
> > first call?  It would still be possible for us to hang around and handle
> > authentication on the BYE, but this will take a little work to do.  As of
> > CVS this morning, the SIP channel has been modified somewhat to be sure we
> > hang around when transmitting "final" messages until they've been ACK'd.
> > 
> > > (o)  While an INVITE is "Trying" a CANCEL request sent by asterisk results
> > > in a "487 Request Cancelled". Cisco equipment ACKs this response. It seems
> > > that if the proxy does not receive this ACK it does not hang up the other
> > > end until a time out period is reached. Should Asterisk be modified to not
> > > clean up after a CANCEL and only after sending an ACK to a 487? Does
> > > everything always send a 487 or should perhaps a channel timeout if non is
> > > received?
> > 
> > I believe this should have been somewhat incidently fixed by CVS this
> > morning and is related to the above behavior.  The basic problem is that
> > we shouldn't destroy the SIP channel until all outstanding messages have
> > been handled, including waiting for the 487 response to our "CANCEL" which
> > should be dumped just before it's destroyed.
> > 
> > > (o)  The ACK sent on the INVITE's 200 OK are not recognized for some reason.
> > > This may be due to my from header hack I'm not sure. After each ACK sent the
> > > proxy sends another 200 OK. Eventually the proxy drops the connection.
> > 
> > I believe this is a result of improper handling of the record-route
> > header.  Someone sent a patch that I plan to review before applying.
> > Please contact me off list or on IRC if the first items are not resolved
> > and I'll be happy to work with you on it.
> > 
> > Mark
> > 
> > _______________________________________________
> > Asterisk-Dev mailing list
> > Asterisk-Dev at lists.digium.com
> > http://lists.digium.com/mailman/listinfo/asterisk-dev
> _______________________________________________
> Asterisk-Dev mailing list
> Asterisk-Dev at lists.digium.com
> http://lists.digium.com/mailman/listinfo/asterisk-dev
-- 
Michael Baird <mike at tc3net.com>