Re: [exim] Exim & Spamassassin at SMTP time

Hi,

Just a quick zoom through your questions:

On Tue, Mar 09, 2010 at 10:40:25AM +0000, exim.ml@??? wrote:
> I'm told that Exim has a steep learning curve and 'ordinary' admins
> struggle with it when compared to Postfix - does anyone here hold a view
> on that?

I've been using it for over 10 years, and have never seriously
used anything else, but found it no problem to learn. But I guess
if you're moving from something else it may be harder. Exim has
definitely got more flexible over the years, which probably also
makes it harder - in some ways it's a programming system for
e-mail. However, the documentation is some of the best of any of
the software I use, it comes with a basic working config, and
examples are to be found all over the place.


> My cursory searches suggest that Spamassassin can be implemented with
> Exim but I'm keen to know if it's possible to get Exim to reject
> messages that Spamassassin scores at 'x' during the SMTP session {rather
> than after accepting the message} Currently I can get Postfix to do this
> but have to plug a slightly buggy and exploitable milter into it to do
> this. I'd be very keen on an MTA that could do this easily - does Exim
> tick this box?

Yes, easily. You could set it to, for example, reject messages
with a score over 10, and tag messages with a score over 5.

> I would like to make use of Bounce Address Tag Validation. I've seen it
> in my logs but it appears that Postfix has only an experimental
> milter/PD for this - does Exim have a solution for this and is it easy
> to implement?

Yes. Here at work I run BATV selectively - it only applies (and
checks) if the user is listed in a config file; you don't have to
run it for everyone.

> Postfix allows some basic header and body checking (but does not really
> have an effective way to white list such rules) - does Exim have similar
> or better features in this area?

You can check headers easily while mail is being handled, and
could accept/reject dependent on them (or even route mail
differently depending on the header contents, for example).

> As far as SASL goes with Exim, is it limited to Cyrus or can it use
> Dovecot?

Not sure about this - I do auth at home by having exim and dovecot
both look at the same passwd file, and at work by exim checking
the password with LDAP.

> Can Exim support 'virtual' domains with this kind of scenario:
> Accept mail for catchall@domain for virtual mailbox on the local machine
> Accept mail for catchall@domain and relay to a final destination server
> Accept mail for x.recipient@domain1 and deliver to local virtual mailbox
> Accept mail for y.recipient@domain1 and relay to remote final
> destination.

Yes.

> Can I get Exim to archive a copy of all mail that passes through it by
> piping it to another indexing program?

Yes.

> How does Exim fair as far as DKIM signing / verification goes?

As of the latest version, it's built in and works great. I've got
it running on my home system, and hope to add it to my work system
this year sometime.

> Finally, I'm guessing and taking as read Exim supports rejecting on
> DNSBL's, missing/errors in PTR and invalid recipients (either by looking
> to LDAP, a PostgreSQL or MySQL database)

Yes.

> I appreciate that this is a long post with lots of questions. I want to
> be 'patriotic' and try an English MTA but need to be sure I'm not going
> to get into several months of 'learning' to find that I've gone the
> wrong way!

I'm sure you won't find you've gone the 'wrong way' with exim (but
then, I can't really compare to anything else).

Hope that helps,

Matthew


--
Matthew Newton, Ph.D. <mcn4@???>

Systems Architect (UNIX and Networks), Network Services,
I.T. Services, University of Leicester, Leicester LE1 7RH, United Kingdom

For IT help contact helpdesk extn. 2253, <ithelp@???>