Package: spamassassin Version: 3.1.0a-2 Hello Duncan, I retried with 3.1.0a-2, and this time I deleted the bayesian database and rebuilt it from my spam/ham maiboxes. The results are better but there still the problem of sa-learn --spam having no effect on the rating of some spam, whereas the sarge sa-learn works fine. Sometimes this is even stranger: if you feed the same mail twice in a row to sa-learn, it is not reckognized as seen and the spam level go up, but if you feed it to spamassin -t twice, the spam level go down.
So I suspect a bug in sa-learn. Cheers, -- Bill. <[EMAIL PROTECTED]> Imagine a large red swirl here. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]