- From: Foteos Macrides <MACRIDES@sci.wfbr.edu>
- Date: Tue, 18 Nov 1997 13:06:24 -0500 (EST)
- To: frystyk@w3.org
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Henrik Frystyk Nielsen <frystyk@w3.org> wrote: >At 15:03 11/17/97 -0800, Jim Gettys wrote: > >>New text: >> >>Due to interoperability problems with HTTP/1.0 proxies discovered since >>the publication of RFC 2068, caching proxies MUST, gateways MAY, and tunnels >>MUST NOT upgrade the request to the highest version they support. The proxy >>or gateway's response to a request MUST be in the same major version as >>the request. > >We can't say that if the server that the proxy speaks to is buggy and >doesn't understand HTTP/1.1. It may have to speak HTTP/1.0 in some >situations. It also doesn't make sense to say that gateways MAY upgrade as >they are gateways into other protocols. > >I can't see what we get out of this that is not already captured in RFC >2145, section 2.3: >[...] I also don't know if something new was raised in Munich, but as far as discussions on this list are concerned, I similarly feel that Section 2.3 covers the issues quite well and would be difficult to improve. One thing that's missing is an explicit reference to the Via header for the case in which proxies are interposed between the browser and origin server, so that scripts can't rely on the SERVER_PROTOCOL variable, but should also examine the HTTP_VIA variable (or their server-side scripting equivalents) to assess the browser's version. I'm not the appropriate one to attempt a chunk of text about that, but if someone who is, did, I suspect many implementors and providers would be appreciative. Fote ========================================================================= Foteos Macrides Worcester Foundation for Biomedical Research MACRIDES@SCI.WFBR.EDU 222 Maple Avenue, Shrewsbury, MA 01545 =========================================================================
Received on Tuesday, 18 November 1997 10:10:55 UTC