Spam autolearning and spam blacklists

[Peter Bonivart]

Magnusson, Andrew wrote:
> We're seeing a lot of these lately:
>
> Apr 25 10:33:32 external-smtp MailScanner[23350]: Message 75AE584846
> from 4.7.254.117 (lnwbompvnxmso at attorney.ac) to localdomain.com is spam,
> spamcop.net, CBL, DSBL, SBL+XBL, SpamAssassin (score=0, required 7,
> autolearn=not spam)
>
> This mail is certainly spam, but SpamAssassin doesn't have rules for it
> and hence not only gives it a 0 score but autolearns it as not spam.
> This is, of course, badly messing with our Bayes database. Is there any
> way to tell it (via MailScanner?) that if it hits 1 or more spam
> blacklists (4 in this case) it sohuld be autolearned as spam, regardless
> of SpamAssassin score?

The autolearn mechanism in SA is based on the score, it's not MS job to
manipulate that. Do your RBL checks in SA and they will add to the score.

By the way, if you're checking SBL+XBL you shouldn't check CBL since
that's included.

--
/Peter Bonivart

--Unix lovers do it in the Sun

Sun Fire V210, Solaris 9, Sendmail 8.12.10, MailScanner 4.29.7,
SpamAssassin 2.63 + DCC 1.2.39, ClamAV 0.70 + GMP 4.1.2, Vispan 1.3

-------------------------- MailScanner list ----------------------
To leave, send    leave mailscanner    to jiscmail at jiscmail.ac.uk
For further info about MailScanner, please see the Most Asked
Questions at    http://www.mailscanner.biz/maq/     and the archives
at    http://www.jiscmail.ac.uk/lists/mailscanner.html