I'm wondering if there's a way to accept a message if multiple attempts
are made:
Some account on an ISP uses procmail to redirect mail, and
the fall-through address is an email address on my machine running
Exim4.
Spam got delivered to that account and it was forwarded via procmail to
my machine running exiscan-acl. The mail was rejected as spam at data
time.
----- Transcript of session follows -----
... while talking to mail.hank.org.:
>>> DATA
<<< 550 This message scored 32.2 spam points.
The message had this envelope sender:
Return-Path: <csc>
which is the account on the ISP. So, the ISP sent the bounce back to that
account, which started the mail loop resulting in this
demime acl condition: too many MIME parts (max 1024)
as the message grew with each bounce.
I'm assuming the answer is no, but is there any way Exim could notice
that the the same message was trying to be delivered over and over and
accept and drop it? I assume no because, for one thing, each message is
new and has a unique Message-Id.
Anything I can do to protect against this?
FWIW:
The ISP is running Sendmail 8.8.6p (from the connection id string). I
tried to send mail with a "mail from: csc" and to that same account but
they refuse to relay:
Relaying Denied. Check your mail using POP3 first to enable SMTP.
So I'm not sure how the spam got delivered in the first place (unless,
of course, they authenticated with POP3 first). It was not generated
from a local user.
--
Bill Moseley
moseley@???
