Hi Matt, It was my typo, it was "--no-rebuild --spam".
I have done "spamassassin -D < a_spam_message", it told me: debug: is spam? score=5.368 required=4.3 tests=ADDR.... I guess that SpamAssassin picked it up from Bayesian database. However, I sent a_spam_message from yahoo.com to my SpamAssassin box, the "hits" is only 1.6. Why? Thank you. ming -----Original Message----- From: Matt Kettler [mailto:[EMAIL PROTECTED] Sent: Thursday, March 18, 2004 4:55 PM To: Hou, Ming; [EMAIL PROTECTED] Subject: Re: sa-learn with spam At 04:45 PM 3/18/2004, Hou, Ming wrote: >I am wondering when I use "sa-learn --rebuild --spam a_spam_message", then >"sa-learn --rebuild". sa-learn printed "Learned from 1 message...", and I >assumed that SpamAssassin 2.63 will learn this spam by saving to Bayesian >database. Then, I sent "a_spam_message" through SpamAssassin again, >"a_spam_messages" was not marked as spam, why? First question.. why did you use --rebuild and --spam at the same time? Second question, what bayes score did it get? Third question, is bayes even enabled or have you only trained a few emails (you must train 200 spam and 200 ham messages before bayes will run). You can check if bayes is in use by looking at the debug output: spamassassin -D < a_spam_message
