I'm using Debian's 3.0.3-2sarge1 spam assassin package and I'm attempting to use sa-learn to train the bayesian filter. I've built up a corpus of spam in an IMAP/mbox folder using Thunderbird. The folder has approximately 500 messages.
Something appears to be going wrong though, sa-learn only sees a single message. $ sa-learn --spam Mail/Junk Learned from 1 message(s) (1 message(s) examined). $ Any ideas what it is? Thanks, James -- http://www.freecharity.org.uk/ - Free hosting for charities http://picasaweb.google.com/jrwdavis - Photography
