I am not sure I understand your objection to the keep param in TCP. Christer has already said that he will change the wording to indicate that if the response does not have keep=yes, it doesn't mean a UAC must not send double-CRLF. In my mind that's just admitting that local configuration on the UAC can trump negotiation, as it can for pretty much anything.
The draft should NOT recommend the UAC sends double-CRLF in such a case, imo. It should not recommend anything. It's out of the scope of the draft at that point. -hadriel > -----Original Message----- > From: Juha Heinanen [mailto:[EMAIL PROTECTED] > > Christer Holmberg writes: > > > >proxy NEVER needs to "provide" anything for tcp ua's nat problem. > > >as i wrote already, if tcp ua is behind nat it simply sends CRLFs or > > >(if it is not able to send CRLFs) uses short enough registration > > >interval. > > > Where is that standardized? Is that how all UAs work? IF so, we can > > remove it... > > christer, > > UA vendors have read outbound draft and figured it out that sending > CRLFs is a good thing to do for tcp UAs and have ignored the rest if > outbound as crap. > > if you want to formalize sending of CRLFs in your draft, then please do > it. > what we are opposing is the keep protocol that is totally unnecessary > for tcp UAs. > > -- juha _______________________________________________ Sip mailing list https://www.ietf.org/mailman/listinfo/sip This list is for NEW development of the core SIP Protocol Use [EMAIL PROTECTED] for questions on current sip Use [EMAIL PROTECTED] for new developments on the application of sip
