[Devel] [ openser-Bugs-1504276 ] Correct handling of branch attributes

[SourceForge.net]

Bugs item #1504276, was opened at 2006-06-11 11:15
Message generated for change (Comment added) made by bogdan_iancu
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=743020&aid=1504276&group_id=139143

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
>Category: modules
Group: ver devel
>Status: Closed
>Resolution: Fixed
Priority: 5
Submitted By: Federico Giannici (giannici)
Assigned to: Bogdan (bogdan_iancu)
Summary: Correct handling of branch attributes

Initial Comment:
The RFC says that:

1) The 2xx ACKs are a new transaction.

2) Each new transaction must have a different branch in
the Via header.

So the "branch" attribute of 2xx ACKs should be
different of the one used for the INVITE/reply.

We should verify that:

1) SER generated ACKs are correct.

2) Transaction maching algoritm is correct (the
"branch" of INVITEs and corresponding ACKs don't have
to be the same).


----------------------------------------------------------------------

>Comment By: Bogdan (bogdan_iancu)
Date: 2006-06-21 20:13

Message:
Logged In: YES 
user_id=1275325

in stateful mode, the building of ACK via branch is fixed -
it guaranteed to be different that the one fro INVITE.

regarding the stateless processing, is still to implement
the behaviour described on the sip-implemetors mailing list
(copying the latest stateful branch). This algorithm will
apply for all requests.
For this I will open a new bug report.

----------------------------------------------------------------------

Comment By: Bogdan (bogdan_iancu)
Date: 2006-06-13 14:01

Message:
Logged In: YES 
user_id=1275325

The 2xx ACK are forwarded by OpenSER using a stateless
function from core and the used branch will be "0" string.
So, it is different from the INVITE branch, but is not
correct since it is not unique.

The same problem is when statelessly forwarding requests -
the banch will be set to "0". There is a solution for this. See:
https://lists.cs.columbia.edu/pipermail/sip-implementors/2006-January/011864.html

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=743020&aid=1504276&group_id=139143