Date: Mon, 27 Apr 1998 11:58:49 +0100 (BST)
From: Philip Hazel <ph10@cus.cam.ac.uk>
To: drums@cs.utk.edu
Subject: Re: RSET inconsistency
In-Reply-To: <SIMEON.9804171252.Y@p6.mci.net>
Message-ID: <Pine.SOL.3.96.980427114143.8422E-100000@taurus.cus.cam.ac.uk>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Fri, 17 Apr 1998, John C Klensin wrote:

> On Fri, 17 Apr 1998 23:58:06 +1000 Robert Elz 
> <kre@munnari.oz.au> wrote:
> 
> >   | which is a contradiction. Presumably the E: line there should go? 
> 
> Two things here.  First, as a consequence of some changes 
> elsewhere that the WG insisted upon, my recollection is that 
> I've worked over RSET.  You might be sure that the next draft 
> doesn't break something you consider important.
> 
> > These errors are the basic set that can occur in response to any command,
> > and should be listed as possible responses to all of them.   Or that set
> > could be excerpted from each of the commands and be listed just once.
> 
> Hmm.  This strikes me as a good idea, with the only disadvantage 
> being that people will not notice the paragraph that says "these 
> aren't listed below, but apply to every command".  But, unless 
> there are serious objections one way or the other, I'm probably 
> going to change the text to pull the "any command can return..." 
> codes out separately.

Thinking over this while I've been away, it still seems to me that if
you specify "these errors apply to every command" and then you specify
that RSET MUST respond with 250, there is a contradiction.

The comment that a server has to be prepared to receive these other
error codes in case a broken client sends them is pointless. A server 
has to be prepared to receive *any* error code (or none). It also seems 
very odd to be trying to specify the behaviour of a broken server! If a 
server that returns 501 to RSET is "broken", what do you call a server 
that returns 999? "Seriously broken"??

-- 
Philip Hazel                   University Computing Service,
ph10@cus.cam.ac.uk             New Museums Site, Cambridge CB2 3QG,
P.Hazel@ucs.cam.ac.uk          England.  Phone: +44 1223 334714