[Mailman-Users] Spamassassin/Mailman - filter messages tolist-owners

[Mark Sapiro]

Michael Berkowski wrote:

>I have recently implemented SpamAssassin with Mailman on
>our server, using the spamd.py and spamassassin.py handlers described here:
>http://www.jamesh.id.au/articles/mailman-spamassassin/
>The spamd/spamassassin handlers are called with GLOBAL_PIPELINE.insert(1,
>'SpamAssassin') in mm_cfg.py.
>
>This method has far exceeded our expectations for reducing the amount of
>spam held for moderation, but it does not seem to have any effect on spam
>messages sent to the *-owner@ or *-request@ addresses.  It would appear that
>all messages to those addresses pass straight through without being scanned
>by the handlers, leaving phishing scams in our moderators' inboxes by the
>hundreds.  I never see anything related to these in the mailman logs.

There is a separate pipeline for messages to the owner, so if you want
to filter these, you need to add

OWNER_PIPELINE.insert(1, 'SpamAssassin')

in mm_cfg.py. Also note that this and the

GLOBAL_PIPELINE.insert(1, 'SpamAssassin')

insert SpamAssassin after Mailman's own SpamDetect module. It doesn't
really matter, but if you want to put it first, use

xxx_PIPELINE.insert(0, 'SpamAssassin').

OWNER_PIPELINE will directly take care of messages to -owner addresses.
Messages to the -admin and -bounces (admin is a synonym for bounces)
address will not be filtered, but if they are unrecognized and
forwarded to the owner, they will be processed through the
OWNER_PIPELINE at that point.

Messages to the -request address will not be filtered, but spam to this
address is normally just returned to the sender with an error message,
not forwarded to an owner/moderator.

-- 
Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan