Re: NEW ISSUE: message-body in CONNECT response

On Thu, Feb 28, 2008 at 11:10:37AM +1100, Mark Nottingham wrote:

> Is it worth strengthening this language?

Strengthening? Section 4.3. says that the existence of a message body is 
request method dependent. That is mathematically correct, but not 
logically: apart from the historic HEAD exception, the existens of a 
body is *not* method dependant. (Future methods can not change that.)

It is generally a good idea to include "why" some things are the way 
they are. A note that HEAD is different for historic reasons makes 
pretty clear to the reader: don't draw any generic conclusions from 
HEAD.

> 
> 
> On 29/11/2007, at 10:58 AM, Roy T. Fielding wrote:
> 
> >
> >On Nov 28, 2007, at 3:20 PM, Robert Siemer wrote:
> >>I read the last sentence as "all other respones - defined in this  
> >>spec - do
> >>include a message-body..."
> >
> >You read it wrong.  The only possible way that a proxy can forward
> >the response to a method it does not understand is if no method
> >other than HEAD (for legacy reasons ONLY) is allowed to change the
> >message delimiting rules.  This applies to anything that uses or
> >extends HTTP and will not be changed.
> >
> >As it happens, CONNECT can't be forwarded by proxies without having
> >understood the method (because it uses a unique request format).
> >That allows non-compliant behavior to be ignored, but it is still
> >non-compliant.
> >
> >....Roy
> >
> 
> 
> --
> Mark Nottingham     http://www.mnot.net/
> 

Received on Thursday, 28 February 2008 01:21:08 UTC