spam.actions.rules

[Pentland G.]

Julian Field wrote:
> What you have just witnessed is a problem raised by having messages
> with multiple recipients. MailScanner doesn't generate mail messages,
> so if you have 5 recipients with different actions, it has to make
> some decision. In this case, I believe it uses the result from the
> 1st recipient. Take a look in the Advanced Settings section of
> MailScanner.conf where you will find this:

This cannot be right as the ruleset was...

To:     gp397 at soton.ac.uk       delete
To:     g.pentland at soton.ac.uk  delete
To:     jw at soton.ac.uk          delete
To:     J.Watts at soton.ac.uk     delete
To:     eks at soton.ac.uk         delete
To:     E.K.Struzyna at soton.ac.uk        delete
To:     lb3 at soton.ac.uk         delete
To:     L.Williams at soton.ac.uk  delete
FromorTo:       default deliver

and j.watts was not the first recipient in that message...

Feb 23 14:00:57 mta2.sucs.soton.ac.uk sendmail[10428]: i1NE0s7s010428:
from=<acutebabe192kkoy at aol.com>, size=3898, class=0, nrcpts=5,
msgid=<000611d7be47$dab24652$21337435 at efvfxrq.qdi>, proto=SMTP,
daemon=MTA, relay=adsl-065-082-235-059.sip.btr.bellsouth.net
[65.82.235.59]
Feb 23 14:00:57 mta2.sucs.soton.ac.uk sendmail[10428]: i1NE0s7s010428:
to=<j.w.wan at soton.ac.uk>, delay=00:00:02, mailer=esmtp, pri=153898,
stat=queued
Feb 23 14:00:57 mta2.sucs.soton.ac.uk sendmail[10428]: i1NE0s7s010428:
to=<j.watts at soton.ac.uk>, delay=00:00:02, mailer=esmtp, pri=153898,
stat=queued

...Sendmail thought j.watts was second.

Maybe it's the first matching rule for any recipient... here the default
being last so the j.watts rule matched first?

Having re-read that extract I think this is the correct description of
what happened here.


Obviously Mailscanner doesn't create the mail but Mailscanner could
remove a recipient from the mail but leave the rest of it/them in the
queue.  I realise this is not easy to implement.

> The other way of solving it is to use sendmail "Queue Groups" to
> limit the number of recipients per message to a maximum of 1. How to
> do this has been discussed here in the past, should be in the list
> archive.

This is not good for disk usage or speed in general as MS would have to
scan a lot more messages but it would indeed solve the problem.

Gary