- From: Adrien de Croy <adrien@qbik.com>
- Date: Tue, 08 Jun 2010 00:24:45 +1200
- To: Mark Nottingham <mnot@mnot.net>
- CC: Julian Reschke <julian.reschke@gmx.de>, HTTP Working Group <ietf-http-wg@w3.org>, Jamie Lokier <jamie@shareable.org>
On 3/06/2010 7:06 p.m., Mark Nottingham wrote: > On 02/06/2010, at 6:45 PM, Julian Reschke wrote: > > Furthermore, the BNF doesn't make it clear that request and response > directives shouldn't be intermixed; >>> Cache-Control-v = 1#cache-directive >>> >>> cache-directive = cache-request-directive >>> >>> / cache-response-directive >>> >> That could be addressed by saying: >> >> Cache-Control-v = 1#cache-request-directive >> / 1#cache-response-directive >> >> However, I'm not sure we need to express this as a syntactical constraint (and it also might confuse people with respect to whether this header uses the "list style" or not). >> > > Yes; I'm not sure we can really do anything about this, except make sure it's clear in the prose. > > hang on a minute. The existing BNF is logically incorrect, the proposed replacement is logically correct, but we worry about confusion? I think it's more likely to create implementor confusion by having the ABNF disagree with the prose. It should still be clear to anyone that the choice is between a list of request directives or a list of response directives. Maybe the problem is that in fact the Cache-control request header is different to the Cache-control response header. You shouldn't put the Cache-control request header in a response message and vice-versa. Probably going to far to solve that one though, since it would require something like Cache-control-v = Cache-control-req / Cache-control-resp Cache-control-req = 1#cache-request-directive Cache-control-resp = 1#cache-response-directive Regards Adrien > Cheers, > > > -- > Mark Nottingham http://www.mnot.net/ > > > -- Adrien de Croy - WinGate Proxy Server - http://www.wingate.com
Received on Monday, 7 June 2010 12:26:02 UTC