Differing Spam Actions for different thresholds?

Julian Field mailscanner at ecs.soton.ac.uk
Wed Jun 26 15:42:46 IST 2002

I am intending to implement a full system where you can specify any minimum
threshold for any rule in spam.actions.conf, to give you full flexibility.
I admit that's not what quite a few of you want, but there are people who do.

However, you have the source, so feel free to DIY :-)

At 15:35 26/06/2002, you wrote:
>I had suggested something along these lines awhile back but it didn't
>happen so I just added it myself. It's a global threshold since it just
>doesn't make sense to do it for each entry in the spam.action file
>because of performance concerns if that list is big.  I added an option
>to mailscanner.conf where I can change the value, then changed config.pl
>to read it in, then added the extra condition to the HandleSpam routine
>in explode.pl. So if a score is 10 or more and there is a spam.action
>for that message such as delete or quaranteen then the action occurs.
>This was mostly just for a delete threshold.
>-----Original Message-----
>From: John Goggan [mailto:jgoggan at DCG.COM]
>Sent: Wednesday, June 26, 2002 8:32 AM
>Subject: Differing Spam Actions for different thresholds?
>I've just recently started using MailScanner.  Almost immediately, I see
>a feature that seems it would be very useful, but that I do not see a
>way to use at this point.  Hopefully someone can tell me if this is
>Basically, I would like different Spam Actions based on the
>required_hits threshold.  I run MailScanner (with SpamAssassin) on a
>site-wide basis.  I've found that I'm getting too many "borderline"
>messages that get marked as spam that are not -- and "whitelisting" all
>of them I don't see as a good solution here.
>What I'd really like to do is set a required_hits threshold at some
>number (say 10) and have all messages over that have "Spam Action =
>delete".  Then I'd like to have a lower threshold (say required_hits =
>6) and have all of THOSE messages have "Spam Action = deliver" (along
>with the subject line modification, of course).
>That way, most of the obvious spam would get deleted.  And the "not
>absolutely sure" stuff would go to the normal user but be marked as
>possible spam.  If it happens to work accurately for them, then they can
>do the final filtering themselves on those messages -- but still never
>have to see the junk that is obviously spam.
>Can that be done already?  If not, any chance of it being implemented
>Worst case, maybe I could somehow configure two instances of
>MailScanner?  The first one would have a higher required_hits and delete
>the messages -- then it runs through a second time with a different
>config (with the lower
>required_hits) and just marks those ones?  Just a thought I had in case
>there isn't a way yet to do it with one config.
>  - John...

Julian Field                Teaching Systems Manager
jkf at ecs.soton.ac.uk         Dept. of Electronics & Computer Science
Tel. 023 8059 2817          University of Southampton
                             Southampton SO17 1BJ

More information about the MailScanner mailing list