W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > January to April 1998

Re: This may be a stupid question ...

From: Scott Lawrence <lawrence@agranat.com>
Date: Fri, 13 Feb 1998 09:06:24 -0500 (EST)
To: Robert Olsson <robban@bigfoot.com>
Cc: Nick Ambrose <nicka@interdyn.com>, http-wg@cuckoo.hpl.hp.com
Message-Id: <Pine.LNX.3.96.980213085741.17500E-100000@alice.agranat.com>


On Fri, 13 Feb 1998, Robert Olsson wrote:

> >You can't write a transparent HTTP/1.1 proxy.  HTTP/1.1 makes explicit
> >provisions for proxies which requires the client to be aware whether or
> >not a proxy is in between.

> On the other hand, I can't see why a transparent HTTP/1.1-proxy would be
> impossible.
> I just wrote one myself (transparently proxies socks port 80) and it seems
> to work.
> It will rather upgrade on both sides to 1.1 than downgrade to 1.0, and IMHO
> it works like a charm.
> So, I nervously wonder, what paragraph of RFC2068 did I just break, who's
> the victim, and what's the verdict? :)

  You missed at least section 14.44 Via:

   The Via general-header field MUST be used by gateways and proxies to
   indicate the intermediate protocols and recipients between the user
   agent and the server on requests, and between the origin server and
   the client on responses.

  I invite you to join the ongoing interoperability testing by putting
your proxy on an Internet-reachable system so that the rest of us can test
through it.

  This requirement is not just a matter of known problems in the present
protocol - it comes from long experience with protocol development that
has shown that having 'invisible' participants in the protocol makes
debugging the protocol and detecting what features may be expected to work
for a given set of participants difficult or impossible.

  
Received on Friday, 13 February 1998 06:08:04 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:33:12 EDT