On 21 Mar 2017, at 12:05, Alan Johnston wrote:
If you are proposing using a SIP response code 603 and with a header
field
Decline-Type: spam, the problem with this is that in SIP, failure
responses
(non-2xx) are delivered hop-by-hop and not end-to-end. This means
that
although the first hop (proxy) will get the Decline-Type:spam header
field,
any future hops will not. Instead, they will just get the 603.
A different response code such as 666 will be conveyed end-to-end, so
every
proxy and the calling UA will get the semantics.
Adam walked me through the last few paragraphs of 3261 section 6. It's
not clear in that text whether proxies will or won't preserve the
headers, but if they don't, I expect that's a showstopper for my
proposal. That's a bummer, because I do think the status code is going
to cause future heartburn. However, if you SIP folks conclude that the
header will not get through proxies (and I will take you all at your
word if you so conclude), and this mechanism does need to survive
proxies (which I suspect it does), then I think we're stuck with the
status code.
If so, I will happily return to you all to your discussion of whether
particular numbers do or do not belong in IETF specifications. :-) (For
the record, I thought 451 set a poor precedent, but I'm a curmudgeon.)
pr
--
Pete Resnick <http://www.qualcomm.com/~presnick/>
Qualcomm Technologies, Inc. - +1 (858)651-4478