Thanks for all the references Brett!
I don't know how you manage to dredge them all up.
I haven't bothered to read through them all now - sorry.
But the requirement for an offer in the first reliable response was
debated at length a *long* time ago. (I recall a live discussion at some
meeting, with Rohan taking the active role.)
I do not know the reasoning behind this requirement. Many would find it
convenient if it wasn't there. But the gist of the discussion was that
relaxing this requirement after its been along is more trouble than its
worth. Its really not that hard to come up with some sort of offer.
Thanks,
Paul
On 2/22/12 9:29 AM, Brett Tate wrote:
> Issues with RFC 3262 have been raised over the years including the long 2006
> SIP Tread "The Problem with PRACK".
>
> http://www.ietf.org/mail-archive/web/sip/current/msg13221.html
>
> Some of them have been addressed by RFC 6337, RFC 6141, and RFC 6228.
> However as RFC 6337 indicates, some issues still exist.
>
> I don't recall if the potential deprecation of RFC 3262's requirement of
> offer within 18x was one of the things proposed within the above thread;
> however it was the topic of the following thread.
>
> http://www.ietf.org/mail-archive/web/sipping/current/msg17210.html
>
> If you are motivated to do so, raise the issues with RFC 3262 within SIPCORE
> or DISPATCH to see if there is any current interest in an rfc3262bis.
>
> The following are some other the old RFC 3262 threads which might help
> present some of reasons an rfc3262bis might be useful.
>
> http://www.ietf.org/mail-archive/web/sip/current/msg13240.html
>
> http://www.ietf.org/mail-archive/web/sipping/current/msg15146.html
>
>
>> -----Original Message-----
>> From: [email protected] [mailto:sip-
>> [email protected]] On Behalf Of Iñaki Baz
>> Castillo
>> Sent: Wednesday, February 22, 2012 8:00 AM
>> To: Kevin P. Fleming
>> Cc: [email protected]
>> Subject: Re: [Sip-implementors] Expected behavior for an INVITE sent
>> without an offer
>>
>> 2012/2/22 Kevin P. Fleming<[email protected]>:
>>> Well, this scenario is not very realistic anyway, because as others
>> have
>>> pointed out, "Require: 100rel" in an INVITE is a fairly bad idea to
>>> begin with.
>>
>> "Require: 100rel" is a bad idea for anything, right. And also the
>> whole PRACK specification. A much more easier approach to get "the
>> same" is just to retransmit the 1XX responses every X seconds. That's
>> advised by other SIP related specifications, i.e: ICE usage. PRACK
>> adds lot of complexity (a spec coming from 3GPP world, of course).
>>
>> --
>> Iñaki Baz Castillo
>> <[email protected]>
>
>
> _______________________________________________
> Sip-implementors mailing list
> [email protected]
> https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
_______________________________________________
Sip-implementors mailing list
[email protected]
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
