- From: David Morris <dwm@xpasc.com>
- Date: Mon, 14 Apr 2008 15:15:46 -0700 (PDT)
- cc: HTTP Working Group <ietf-http-wg@w3.org>
Probably way too late in HTTP's life cycle, but the what I'd like is to separate quoting requirements from specific situations and generically require quoting in any case where the value isn't safe w/o quoting. In all cases, quoting should be removed before comparison or other processing. Oh well ... I can dream ... Dave Morris On Fri, 11 Apr 2008, Mark Nottingham wrote: > > Hmm. That's the opposite of the direction I *thought* we were going in. > > If a parameter specifies a token, e.g., charset, I read the discussion > as saying that we should also accept a quoted-string. This allows the > approach that Frank was talking about (quote everything, just for > safety and simplicity's sake). > > If a parameter specifies a quoted-string, there isn't any > corresponding rationale to support also allowing a token. > > That's why I wrote the proposal as I did. > > > On 11/04/2008, at 10:33 AM, Henrik Nordstrom wrote: > > > > > tor 2008-04-10 klockan 17:17 +0200 skrev Julian Reschke: > > > >> I have my doubts. Why add special cases in producers when it's not > >> needed? Or *are* special cases needed? > > > > I wouldn't expect all parsers to deal with quoted parameters in all > > cases.. the current specs is a bit fuzzy on that. > > > > But I do expect the ones who deal with quoted parameters to also > > handle > > the token form.. > > > > Regards > > Henrik > > > > > > > > > > > -- > Mark Nottingham http://www.mnot.net/ >
Received on Monday, 14 April 2008 22:16:29 UTC