- From: Willy Tarreau <w@1wt.eu>
- Date: Thu, 21 Oct 2010 22:52:20 +0200
- To: =JeffH <Jeff.Hodges@KingsMountain.com>
- Cc: IETF HTTP WG <ietf-http-wg@w3.org>, Adam Barth <abarth@eecs.berkeley.edu>
Hello, On Thu, Oct 21, 2010 at 10:42:32AM -0700, =JeffH wrote: > FYI, we're getting very close to requesting RFC publication of > draft-ietf-httpstate-cookie. Please send any feedback to > <http-state@ietf.org> A few months ago, I reported one point to Adam which I think has since been missed. The draft does not indicate that it obsoletes the parts related to the Cookie header syntax of the RFC2965. But this RFC (re-)defined it, explicitly allowing empty lone attributes, as well as quoted strings in the value, both of which don't seem to be allowed anymore in latest draft, which makes the equal sign mandatory and expects a token as the value. So with 2965 not being obsoleted, both RFCs could be used to build possibly incompatible implementations. I don't know how this is normally dealt with. Is it possible to obsolete just one part of an RFC ? (the cookie2 part of 2965 looks fine). Regards, Willy
Received on Thursday, 21 October 2010 20:53:10 UTC