On Monday 18 June 2007 12:36, ian douglas wrote: > Rob Wright wrote: > > So far I've managed to run ~2500 messages through sa-learn over the > > course of the last week or so, and I've yet to see a single log entry > > with a BAYES rule match of any kind. > > From your own logs: > > [24761] dbg: bayes: not available for scanning, only 0 ham(s) in bayes > DB < 100 > > ... which tells me it hasn't learned the minimum of 100 ham messages > that you've told it to need before Bayes will kick in. > > Maybe post a dump of "sa-learn --dump magic" too?
Thanks Ian. I had run some ham through sa-learn so that's odd. After receiving your messages I ran the "sa-learn --dump magic" and I get this: sa-learn --dump magic 0.000 0 3 0 non-token data: bayes db version 0.000 0 2099 0 non-token data: nspam 0.000 0 0 0 non-token data: nham 0.000 0 188955 0 non-token data: ntokens 0.000 0 1181845178 0 non-token data: oldest atime 0.000 0 1182181807 0 non-token data: newest atime 0.000 0 1182182134 0 non-token data: last journal sync atime 0.000 0 1182182158 0 non-token data: last expiry atime 0.000 0 0 0 non-token data: last expire atime delta 0.000 0 0 0 non-token data: last expire reduction count So, I went back to where my ham is and re-ran sa-learn on that with this result (after first using --forget): sa-learn --ham -C /etc/mail/spamassassin --showdots --spam --no-sync Maildir/new ............................................................................................................ Learned tokens from 108 message(s) (108 message(s) examined) I ran sa-learn --sync, then restarted spamassassin and get this: sa-learn --dump magic 0.000 0 3 0 non-token data: bayes db version 0.000 0 2100 0 non-token data: nspam 0.000 0 0 0 non-token data: nham 0.000 0 188955 0 non-token data: ntokens 0.000 0 1181845178 0 non-token data: oldest atime 0.000 0 1182181807 0 non-token data: newest atime 0.000 0 1182189220 0 non-token data: last journal sync atime 0.000 0 1182182158 0 non-token data: last expiry atime 0.000 0 0 0 non-token data: last expire atime delta 0.000 0 0 0 non-token data: last expire reduction count spamassassin -D --lint still shows: [14187] dbg: bayes: not available for scanning, only 0 ham(s) in bayes DB < 100 So, then, spamassassin isn't seeing the ham that I'm feeding it? Why would it see the spam but not the ham? Thanks Rob