On Wed, 2005-03-02 at 20:21 -0800, Jonathan Nichols wrote: > Ow Mun Heng wrote: > > On Wed, 2005-03-02 at 16:13 -0500, Thomas Kirchner wrote: > > > >>* On Mar 2 21:50, Matthias F. Brandstetter ([email protected]) wrote: > >> > >>>after upgrading SA to version 3 I now do get lots of false negative. I > >>>trained my filter via "sa-learn", and many of these FP do have "BAYES_99", > >>>but they do not reach 5.0 points. > >> > >>In your ~/.spamassassin/user_prefs file, add some lines like the > >>following, for whatever tests you want to modify: > >> > >>score BAYES_50 1.0 > >>score BAYES_60 1.5 > >>score BAYES_80 2.5 > >>score BAYES_95 4.0 > >>score BAYES_99 4.5 > > > > > > How can this be implemented to be done system wide?? > > > > > > /etc/mail/spamassassin/local.cf > > Also make sure that you're using the new surbl features of SpamAssassin > 3.x - they catch a LOT of spam. > It's a network check, though. Keep that in mind.
Cool.. I've added is as per below. Don't have I-Net access so surbl (or any sort of RDL/network checks) is of no use for me. required_hits 5 rewrite_header Subject [N-SPAM] report_safe 0 use_bayes 1 bayes_path /var/cache/spamassassin/bayes auto_learn 1 lock_method flock score BAYES_50 1.0 score BAYES_60 1.5 score BAYES_80 2.5 score BAYES_95 4.0 score BAYES_99 4.5 -- Ow Mun Heng Gentoo/Linux on DELL D600 1.4Ghz 98% Microsoft(tm) Free!! Neuromancer 13:35:43 up 11:34, 6 users, load average: 0.22, 0.28, 0.18 -- [email protected] mailing list
