- From: Roy T. Fielding <fielding@liege.ICS.UCI.EDU>
- Date: Sat, 01 Jun 1996 01:59:44 -0700
- To: jg@w3.org
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
> Actually, there is a more serious problem with Roy's rewrite of 5.1 and 5.2 > section in the current document (04a) that I missed when vetting his changes > (and mostly improvements). > > The requirement that all servers check that the host part is supplied > one way or the other only is required to servers supporting multiple > hosts in the current text. This requirement should be stronger; that > all servers check and generate errors if host information is not > present one way or the other. Otherwise we won't necessarily get the > desired effect of detecting buggy 1.1 clients until 1.1 servers are > deployed, likely much after when clients are shipped. This would > result in the scenario of none of the host requirements being > effective to solve the problem they were intended to solve. > > So Bullet #3 in 5.2 should be pulled out and put at the end of the first > paragraph of section 5.2. No. We have already discussed this several times. This section has to do with the requirements upon receipt of any message for how the origin server determines which resource has been identified. For that purpose, the origin server must not be required to send an error message just because it hasn't received the Host header field. If you make that change, we will NEVER be able to remove the Host header field. It is sufficient to just require that Host be sent with HTTP/1.1 messages and that servers respond with an error when Host is not found in an HTTP/1.1 message, neither of which are defined by this section -- they are defined elsewhere. ...Roy T. Fielding Department of Information & Computer Science (fielding@ics.uci.edu) University of California, Irvine, CA 92717-3425 fax:+1(714)824-4056 http://www.ics.uci.edu/~fielding/
Received on Saturday, 1 June 1996 02:07:25 UTC