[Mimedefang] SA 2.50

[James Olsen]

Hello all,

>> I haven't done a terrible amount of testing/monitoring of it.  But from the
>> get-go, it appears to be a lot more efficient in catching spam.  I've
>> already noticed a decline in false positives.  So it's been good for me.

DFS> I'll second that; even with the Bayesian rules disabled, SA 2.50 seems
DFS> a lot better at catching spam than 2.44.

DFS> The Perl API used by MIMEDefang hasn't changed, so you can simply install
DFS> it over 2.44 without changing any MD settings.

I've installed over the old version, so far everything seems fine.

However, I installed 2.50 to specifically get Bayesian filtering up
and going. I haven't had much time to dink with it, but so far I don't
seem to be able to get it to work. I've fed a bunch of spam/ham
messages to "teach" SpamAssassin via sa-learn, but when a spam message
goes through I don't see any BAYES* flags raised in the "X-Spam-Tests"
header.

I don't know if anything needs to be changed in the mimedefang filter
to invoke the bayes tests or if I've just not configured things
correctly. I tried setting a specific path in the sa-mimedefang.cf and
it would appear to be using the files in that path (i was getting
permission errors until I fixed the permissions, so it appears to be
trying to use them). But I'm not sure what I should try next.

Has anyones got the bayes tests to work, and would you share how you
did it? :)

Thanks..

-- 
 James