>> >> On Dec 1, 2005, at 9:28 PM, Nick Hemmesch wrote: >> >>> If spamassassin is having problems reading bayes_seen, bayes_toks or >>> auto-whitelist, run spamassassin -D --lint and it will unlock >>> everything >>> for you. >> >> Since that didn't work, and some of us are still having some odd >> error messages, any further ideas on what we can look at, or >> investigate, or troubleshoot this? >> >> My 'bayes' file is still zero length, and while there's now a >> user_prefs file, it doesn't seem like anything is happening on the >> auto-learn front. >> >> Harry > > Hi Harry, > > It's likely that you won't see anything until you actually receive a mail > that has a spam score over 12. Take a look at the variables in the > mailfilter file, /etc/mail/mailfilter. > > auto-learn (sa-learn) is controlled by the mailfilter script not by > spamassassin, so your mail header will show autolearn=no. As long you have > an entry that shows "Content-Filter: maildrop-toaster" everything is OK. > > Regards, > > Nick >
Hi again Harry, There is a problem with autolearn running from the mailfilter script. It must have started with spamassassin ver. 3.1.0 since my ver 3.0.4. works fine. Let me take a look at it on Sunday and I'll get back to you with a solution. Regards, Nick --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
