On 22.06.11 15:29, Stefan Sayer wrote:
o Raphael Coeffic on 06/22/2011 03:20 PM:
On 22.06.11 15:15, Stefan Sayer wrote:
Module: sems
Branch: master
Commit: 38a3109950d94bcec96110a70aa8dce820a01c31
URL:
http://git.sip-router.org/cgi-bin/gitweb.cgi/sems/?a=commit;h=38a3109950d94bcec96110a70aa8dce820a01c31
Author: Stefan Sayer<[email protected]>
Committer: Stefan Sayer<[email protected]>
Date: Wed Jun 22 15:13:25 2011 +0200
b/f: don't change OA state to completed in early dialog
if OA state is changed to completed already when early media SDP
received, final SDP received will be seen as another offer
Der punkt ist: der zweite offer (in 200) ist:
s/offer/answer i guess.
- unnötig.
but still it must be there - even in the same branch - right?
not that I know... but we can check. I have seen examples where it was
not there.
- sollte deckungsgleich mit dem in dem 183 sein.
IMO we should (at least optionally) support changing SDP from early
dialog to real dialog, even if its non-standard; in fact I don't know
why the standdard mandates that it must not change.
because it allows you to throw it away ;-)
Ergo, der korrekt fix sollte den SDP im 200 ignorieren. Wenn du nicht
does that work with SDP in early session from a different branch, too?
in that case we would have to go back to offersent, right?
well, we would be supposed to create another dialog, which consists of
INV and 200. How you connect the media is another story.
If we don't create another dialog, we might end-up in a huge mess, while
trying to handle these kind of cases. Though I don't have a straight
forward solution to that problem. The easiest solution is to not support
a 200 with a different to-tag as the 183.
-Raphael.
zu OA_completed wechselst, ist es eher falsch, da offer/answer schon
augetauscht wurde. Es wird z.B. nicht funktionnieren, wenn nach dem
INV/183, einen UPDATE mit einem neuen offer kommt. Dieses scenario ist
normal bei 3GPP.
ok, understood.
_______________________________________________
Semsdev mailing list
[email protected]
http://lists.iptel.org/mailman/listinfo/semsdev
