From: "David F. Skoll" <dfs(_at_)roaringpenguin(_dot_)com>
There are existing solutions to that problem. One is to reject with
a 4yz temporary failure codes any Rcpt_To address whose filtering
preferences might yield a different answer than the preferences of
the first recipient.
That was an option I considered, but rejected. If you're the tenth
recipient, you might wait a long time to get your e-mail.
Only 10 times the delay between runs of the sender's queue.
Still, as anyone with an address near the end of the alphabet will
probably tell you, that is a good point I'd not thought of. I usuall
get my copy of mail from the main IETF list hours to to days after
people at the front of the alphabet.
A second solution is to resolve conflicts in the correct answer to
the DATA command with a 2yz OK answer but to deliver only to mailboxes
that want the message and silently discard instead of rejecting the
message for other mailboxes.
That is the method I chose also. Actually, I group recipients by
filter preferences, and then re-mail the message to each group with
identical preferences, discarding the original. The re-mailed
messages get filtered, but by then everyone is known to have the same
set of preferences, so the problem is solved.
Remailing forces DSNs on failure, which has various costs including
double bounces into your postmaster mailbox and bounces sent to
innocent people when the return address really was forged.
Or do you discard instead of bounce during the re-mailing?
Vernon Schryver vjs(_at_)rhyolite(_dot_)com
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg