I allow our users to opt-in and out of spam rejection at SMTP time. I
use Exim and SpamAssassin to score messages.
I use the scheme where I record the opt-in status of the first RCPT user
and do a retry-later rejection for subsequent users who do not have the
same status. If the first user has opted in then I can reject at the
end of the DATA phase if the message exceeds the rejection threshold.
This scheme has been working fine for several months.
I now have one site which re-queues for all recipients if any one of
them gets a retry-later, so they never get a message to us if it has
multiple recipients with mixed opt-in statuses.
I would prefer not to add a whitelisting scheme to let them by-pass spam
rejection. They claim that nowhere else can be doing what we are doing
because we are the only site they have a problem with. I claim that
they are acting differently from everywhere else because we have not had
the problem reported from anywhere else.
I bit of an imapasse!
It appears to me from statements in their e-mail that their MTA software
is home-grown.
I am looking for concrete evidence that what I am doing is not unusual,
so can you please let me know if your system uses the same scheme to
allow opt-in for SMTP-time spam rejection?
Thanks in advance,
Phil.
--------------------
Phil Chambers
Postmaster
University of Exeter
