Hector Santos wrote:
Receiver (preliminary: final destination decision made)
C: RCPT TO:<some address>
250 - USER FOUND
45x - USER FOUND (For some policy reason, temporary reject)
55x - USER NOT FOUND (domain is locally hosted)
55x - RELAY AUTHENTICATION REQUIRED (domain is remote)
The point is that in the newer models the Return-Path value increases
and is more reliable at this point. Regardless if NULL or not NULL, in
this model, it is meaningless. It has no value. In fact, adding the
Return-Path for some post smtp mailbot shouldn't be a consideration
because it no longer applies.
I said two different things. :)
Let me be more specific:
I meant it has no value as a "notification" address because it is not
required with a dynamic SMTP receiver making final destination
decisions. For a negative response, the message is not going to be
accepted so there is no issue with post smtp mail bots. This is ideal
to prevent bounce attacks.
If the message is accepted, the value of the return path means
"something." It has value. If there are post smtp mail bots existing
outside of the control of the main SMTP implementation that could create
problems, it is quite possible that the return-path not be added or
change to NULL. I'm not recommending this or suggesting it be part of
some automation logic, but that it is possible in order to minimize post
smtp mailbot mishaps.
--
Sincerely
Hector Santos, CTO
http://www.santronics.com
http://santronics.blogspot.com