Re: [sipcore] REMINDER: Re: WGLC: draft-ietf-sipcore-rfc4244bis

[Paul Kyzivat <[e-mail]>]

On 11/2/2010 3:50 PM, Mary Barnes wrote:

>>>> Does this include 100 responses?
>>> [MB] I would think so. Can you think of a reason why it shouldn't?  I>>  would think it could be useful for debug. [/MB]
>> [JRE] The 100 response typically only travels back a single hop, so the benefit of H-I in a 100 is marginal. Also some implementations might send back a 100 response quite early during processing, perhaps even before they decide whether to act as a proxy, a UAS or a redirect, so knowing how to populate H-I in a 100 might be difficult.
> [MB2] If it has no additional hi-entries, then there is nothing
> additional to send back, so I don't think it hurts anything - I would
> agree that in most cases it's of marginal value. We can include it as
> an exception in that sentence if you would prefer, although per the
> next comment, it seems that you are okay with leaving it as is.
> [/MB2]

IMO its unreasonable to expect the server to include *any* H-I in the 
100 response. As Hadriel mentioned, 100 is often sent from a "lower" 
layer of the stack, perhaps before the message has been fully parsed.

IMO H-I MAY be present in 100 responses, and SHOULD be ignored by the 
recipient if present. Since ignored, there should be no requirement on 
what content the H-I is to have in that case.

	Thanks,
	Paul