On Wed, 2005-09-14, Paul Smith wrote:
(BTW, a potential "problem" with 'formalising' a 'keep-alive' is that
people will then start writing clients which detect it and, for instance,
make their configured timeout act from the final '.' being sent until the
last line of the reply, rather than between lines of the reply)
As a possible fly in this keep alive ointment is RFC 2821 section 22.214.171.124
which *already* implies that the timeout should be until the FINAL reply
code rather than between each line of the reply. In particular, see the
description of the DATA timeout:
DATA Termination: 10 minutes.
This is while awaiting the "250 OK" reply.
This looks like the final reply code to me and I can believe that many SMTP
Client writers would interpret it this way.
Bill McQuillan <McQuilWP(_at_)pobox(_dot_)com>