W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 1999

RE: Host header issue

From: Larry Masinter <masinter@parc.xerox.com>
Date: Tue, 7 Sep 1999 16:41:02 PDT
To: "Josh Cohen (Exchange)" <joshco@exchange.microsoft.com>
Cc: HTTP Working Group <http-wg@hplb.hpl.hp.com>
Message-ID: <001201bef98a$721da0e0$c5d1000d@copper.parc.xerox.com>
X-Mailing-List: <http-wg@cuckoo.hpl.hp.com> archive/latest/562
You know, I'm still not sure under which situation you
would 

a) not require a Host: header to be sent by a client;
   (since there are HTTP/1.1 servers that would error
   if the Host header were omitted)

b) not require that a server respond with an error if the Host
  header were missing (since HTTP/1.1 clients should only
  send absolute URIs to proxies.)

Procedurally, since HTTP/1.1 is Draft Standard, the only way
I can think to institute a change such as this would be to write
a separate document, as Proposed Standard, which updates 2616;
I'd suggest:

   Omitting Host Header from HTTP/1.1 Requests with Full URI

as a title.

Note that:

   1. If Request-URI is an absoluteURI, the host is part of the
     Request-URI. Any Host header field value in the request MUST be
     ignored.

so a valid request would be

GET http://host.dom/path HTTP/1.1
Host: 

and you're saving 7 bytes for the "CRLFHost:" But then, if you're so
concerned about saving 6 bytes, why not also work on encoding the 20 bytes
that comprise "GET http://" and "HTTP/1.1" into something more efficient?

Larry
-- 
http://www.parc.xerox.com/masinter
Received on Tuesday, 7 September 1999 16:44:40 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 2 February 2023 18:43:06 UTC