>>I agree with Hadriel. It is very difficult to, in a spec, say what an
> entity not supporting the spec should do - because implementors may  >
not even have read the spec. 
>
>i don't remember i have proposed saying something about an entity (UA)
not supporting the spec.  what i have proposed is that an UA supporting
the keep spec should still send CRLFs even it does 
>not get back keep=yes.
>
>>But, again, we can choose to not explicitly forbid sending CRLF, which
means that people not  > supporting "keep" can continue sending CRLF
without anyone being able
>>to tell them that what they are doing goes against some spec etc. 
>
>i still don't understand why we need an addition to sip protocol for
CRLFs keepalives.
>
>why can't you instead say in your draft that before sending a register
or any other sip request, UA sends double-CRLF to find out if ob proxy
supports CRLF keepalives.  it does, if UA gets back 
>a single-CRLF and otherwise not.  
>
>in my opinion, it would be much simpler than inventing yet another
addition to sip protocol on top to the countless additions there already
exists and that no-one has implemented.

That still doesn't indicate to the edge proxy that the UA supports the
keep-alives. I don't think we can expect the edge proxy to remember
every IP address/port from where it has received a double-CRLF, before
it has received a SIP request from that same IP address/port. 

Regards,

Christer
_______________________________________________
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

Reply via email to