[Mimedefang] MD and fetchmail

[Tony Nugent]

On Fri May 16 2003 at 10:44, David Hollis wrote:

> I've been running MD with great success on my system for almost a year
> now, I still have one  thing that is puzzling me.  In addition to my
> normal incoming SMTP traffic, I run fetchmail to pull down mail from some
> stray POP accounts.  That mail then runs through my local spooler into my
> normal inbox.  However, I've noticed that a lot of spam is able to get
> through and it seems it is always the mail I get from fetchmail.  In
> scanning the headers, the X-Scanned-By header is there, but the no
> spamassassin.  If I export the message and hand run it through
> SpamAssassin, all the sirens go off.  My filter is almost stock with minor
> additions on my end to drop spam messages with high scores so I don't even
> have to deal with it at the MUA.  Anyone else see this before?

Not here.  I assume that you have fetchmail configured to give email
to sendmail via 127.0.0.1:25 (port 25 on localhost) and not deliver
directly to the local MDA (eg procmail).

Perhaps the SA tests are being skipped if the email is locally
generated (via localhost)?  Recheck your mimedefang filter to see if
this is what is happening.

If you are using procmail, then it is possible to create a
procmailrc to filter email via spamassassin (spamc/spamd).  I use
this on my own workstation for email that I get from places that
don't have virus or spam filters on their mail systems.

This is an (slightly modified) extract from my ~/.procmailrc....

# SpamAssassin checks - only if mail <10000 bytes
:0 fw
* < 100000
| spamc -p 2222
:0 e
{
  EXITCODE=$?
}
:0H :spammer/$LOCKEXT
* ^X-Spam-Status:.*Yes
spam.IN

The EXITCODE recipe is not strictly necessary, but useful if you
want to delay delivery until other checks are made and then act on
the value of the exit status from spamc (I'm also running spamd on
port 2222).

Cheers
Tony