ietf-smtp
[Top] [All Lists]

Re: rfc2821bis-01 Issue 18: Usability of 1yz replies -- proposed text

2007-04-11 15:33:59



--On Wednesday, 11 April, 2007 16:31 -0400 Jeff Macdonald
<jmacdonald(_at_)e-dialog(_dot_)com> wrote:

On Wed, Apr 11, 2007 at 09:25:04AM -0400, John C Klensin wrote:
Proposed text:

             The command has been accepted, but the requested action
             is being held in abeyance, pending confirmation of the
             information in this reply.  The SMTP client should send
             another command specifying whether to continue or abort
             the action.  

Does this allow the following?

s: 150
<time goes by>
s: 250

Not without either an extension or other changes, for two
reasons:

(1) 1yz replies are not now permitted _at all_ except by
extension. As I understand a slightly extended form of Hector's
argument, if one can have different codes on different
continuation lines and only the last one counts, then one can
have _anything_ (or at least any three-digit code) on the
previous lines with no semantic impact as far as the standard is
concerned other than the hope that the client will wait until
the sequence of continuation lines is complete.  The more I
think about that argument, the less I buy it, largely because of
longstanding personal opposition to standardizing tricks and
hacks.  But it is not applicable to this case in any event.

(2) SMTP is designed as strictly a synchronized one-command /
one-response protocol.  Pipelining and some other extension
ideas that are floating around may permit delaying the responses
or otherwise changing the synchronization, but the model
remains.  So, once the server issues a reply (or the last line
of a multiline reply), it needs to get a new command from the
client before it can send out another response code/line.  

     john