Stefan Sayer writes:
> > when sems sends update, it is recommended that also there sems writes
> > ;refresher=uas.
> if you are the refresher, and you are doing the refresh (as uac) a you
> can't require the other side to be the refresher (put refresher=uas),
> therefore sems doesn't put refresher, and hopes the other side takes
> over the refresher role.
stefan,
perhaps i was not clear enough. if sems sbc has been the refresher and
sends a new refresh request, rfc recommends that the roles remain the same
as before and therefore sems should put refresher=uac in the new
refresh.
how could this break something?
here is what i have tried in bool SessionTimer::onSendRequest:
req.hdrs += SIP_HDR_COLSP(SIP_HDR_SESSION_EXPIRES) +
int2str(session_interval);
if (req.rseq != 0) {
INFO("request is in-dialog\n");
req.hdrs += ";refresher=" +
(session_refresher==refresh_local ? "uac":"uas");
} else {
INFO("request is out-of-dialog\n");
}
req.hdrs += CRLF + SIP_HDR_COLSP(SIP_HDR_MIN_SE) + int2str(min_se) + CRLF;
but my in-dialog test does not work i have not yet figured how to do it.
-- juha
_______________________________________________
Semsdev mailing list
[email protected]
http://lists.iptel.org/mailman/listinfo/semsdev
