In my continuing effort to fight spam that gets through my mx gateways (MIMEDefang 2.31, Spamassassin 2.60, Sendmail 8.12), I had set up an IMAP folder on the one of the gateways for several of our users to copy SPAM that made it through.
I had set up a job that weekly runs sa-learn against the inbox, then clears the box out. I had been doing this for several weeks and it gave end users a warm fuzzy feeling that they were helping fight the spam problem :<).. I didn't feel like it (bayesian filtering) was making any difference, so upon taking a second look at my sa-mimedefang.cf , I realized there was nothing in there to enable the bayes checks. I added the following to my sa-mimedefang.cf and now I believe it is using the bayes check: # Trying to enable bayesian filtering 02-03-2004 auto_learn 1 use_bayes 1 bayes_path /var/spool/MIMEDefang/bayes basyes_file_mode 0666 #Enable AWL autowhitelisting auto_whitelist_path /var/spool/MIMEDefang/auto_whitelist auto_whitelist_file_mode 0666 My question is how do make sure my manually learned samples get included into the database that MIMEDefang/Spamassassin use? MIMEDefang runs as user defang so I'm thinking if I run sa-learn as that user then it would get added to the same database. Any thoughts or suggestions would be greatly appreciated. -- _______________________________________________ James Whittington Systems Administrator Excelorant, LLC A TM Floyd Affiliate Company 1800 St. Julian Place P.O. Box 8263 Columbia, SC 29202 803.765.1310 | Phone 800.780.1170 | Toll Free 803.765.1431 | Fax [EMAIL PROTECTED] _______________________________________________ _______________________________________________ Visit http://www.mimedefang.org and http://www.canit.ca MIMEDefang mailing list [EMAIL PROTECTED] http://lists.roaringpenguin.com/mailman/listinfo/mimedefang
