W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > May to August 1999

RE: Fragments in "Location" field

From: Henrik Frystyk Nielsen <henrikn@microsoft.com>
Date: Fri, 13 Aug 1999 11:43:13 -0700
Message-ID: <FD8F41F04D35D111899600805F19A16C1C3F4CA9@RED-MSG-48>
To: "'Rodent of Unusual Size'" <Ken.Coar@Golux.Com>, "Roy T. Fielding" <fielding@kiwi.ics.uci.edu>
Cc: http-wg@hplb.hpl.hp.com

> Does this mean that we (in the CGI work) can assume that fragments
> are legal in our overload definition of Location?

If it is that a CGI script can set the location header field with a value
containing a fragment then yes.

Btw, I have seen several servers sending relative Location header field
values. This does make sense but does it break any existing applications we
know of?

Regarding allowing fragments, I believe the same is the case for
Content-Location:

       Content-Location = "Content-Location" ":"
                         ( absoluteURI | relativeURI )

should be

       Content-Location = "Content-Location" ":"
                         ( absoluteURI | relativeURI ) [ "#" fragment ]

I don't like introducing a lot of "sanity rules" for when they are allowed
and when not. I think that proxy redirection (305 Use Proxy) is the only one
that shouldn't be allowed.

Henrik Frystyk Nielsen
frystyk@microsoft.com
Received on Friday, 13 August 1999 19:52:53 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:33:32 EDT