Return-Path: Delivered-To: new-httpd-archive@hyperreal.org Received: (qmail 26670 invoked by uid 6000); 13 Jun 1999 06:44:39 -0000 Received: (qmail 26664 invoked from network); 13 Jun 1999 06:44:38 -0000 Received: from paris.ics.uci.edu (mmdf@128.195.1.50) by taz.hyperreal.org with SMTP; 13 Jun 1999 06:44:38 -0000 Received: from kiwi.ics.uci.edu by paris.ics.uci.edu id aa25352; 12 Jun 99 23:40 PDT To: new-httpd@apache.org Subject: Re: Misspelt protocol version treated as HTTP/1.0? In-reply-to: Your message of "Sat, 12 Jun 1999 15:44:04 EDT." <3762B884.F741E97E@Golux.Com> Date: Sat, 12 Jun 1999 23:40:49 -0700 From: "Roy T. Fielding" Message-ID: <9906122340.aa25352@paris.ics.uci.edu> Sender: new-httpd-owner@apache.org Precedence: bulk Reply-To: new-httpd@apache.org >I'm inclined to think this is broken, and that we should pitch >with a 400 anything that doesn't match "HTTP/[0-9]+\.[0-9]+" >for the protocol. The last time I looked, ":" didn't count >as implied LWSP.. There are times when I'd like a switch to turn on strict parsing and extended diagnostics, but in general it is better to just be robust in handling the unexpected. I wouldn't object to sending a 400 response, but there are many more improtant things to fix. ....Roy