Hi Steffen, on Wed, 2 Apr 2003 15:07:05 +0200 (CEST) you wrote:
[moving from sendmail]
> The problem I see is that we use MIMEDefang
<snip>
> The nice and easy thing with MIMEDefang is that you have full control
> over the conditions and actions as they are perl functions getting
> called at certain circumstances, e.g. when a MAIL FROM or RCPT TO line
> has been encountered, or when all the mail body is acquired (but before
> the the SMTP dialogue returns any success/failure code) etc.
I believe you will be able to do everything you want with:
a) Exim's own (built-in) ACL's, which allow you to make all kinds of
decisions at each point in the SMTP dialogue. If you read this list for a
while, you will see all kinds of weird and wonderful concoctions that
people are using, but the key thing is that these rules aren't set in
stone; Exim is *extremely* flexible and you can almost certainly adapt the
ACL rules to do whatever necessary.
and
b) exiscan, a widely-used, clean and well-supported patch for Exim (
http://duncanthrax.net/exiscan/), which does realtime (post-DATA) mail
scanning (including MIME deconstruction if you want) and incorporates a
whole array of mail scanning features including direct interaction with
many popular virus scanners, SpamAsassin, regular expression rules,
attachment filtering etc.
You will also find some general discussion on the specific point of
virus/spam scanning at SMTP time (but not ACLs) in my mini-HOWTO at
http://www.timj.co.uk/linux/exim.php
Tim
