W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > September to December 1995

Re: Decision about Host?

From: Ari Luotonen <luotonen@netscape.com>
Date: Fri, 6 Oct 1995 13:23:43 -0700 (PDT)
Message-Id: <199510062023.NAA18077@starfish.netscape.com>
To: David Robinson <drtr1@cus.cam.ac.uk>
Cc: bne@bne.ind.eunet.hu, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com

> False. The server may be behind a firewall, proxy or accelerator; the
> port on which it is listening may not be the port the client used.

None of these are valid arguments for having the port number there:

Behind a proxy: full URL is sent to the proxy, and Host: is
unnecessary anyway; in any case the port number would be that
indicated in the URL, not the port the client happened to connect to a
proxy server.  The proxy then adds the Host: header (although I think
in the transition period, the client will need to send it in case the
proxy doesn't support it yet).

Behind an accelerator: the accelerator port is the one advertized in
the URL -- how would the client have known to connect there anyway?

Or, to put it differently, if your server is started on port X, it
*knows* the connections that come to that port are coming to port X.
Where else would they be going to?  Think about it.

Cheers,
--
Ari Luotonen				ari@netscape.com
Netscape Communications Corp.		http://home.netscape.com/people/ari/
501 East Middlefield Road
Mountain View, CA 94043, USA		Netscape Server Development Team
Received on Friday, 6 October 1995 13:28:59 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:33 EDT