FH wrote: > It was less than 1/2 hour because I was experimenting w/ the commands > and the new email came in so I decided to use that one ;) > > Initial email: > X-Spam-Status: No, score=1.9 required=4.0 tests=BAYES_99 autolearn=no > version=3.0.2 > > After running it through the spamassassin -D command: > X-Spam-Status: Yes, score=12.5 required=4.0 tests=BAYES_99, > RCVD_IN_BL_SPAMCOP_NET,URIBL_AB_SURBL,URIBL_OB_SURBL, > URIBL_SC_SURBL,URIBL_WS_SURBL autolearn=no version=3.0.2
Half an hour is plenty of time to show this kind of difference in scores. Note that both times, the message hit BAYES_99, but the second time around it hit the "normal" SpamCop RBL, and a set of URIRBL tests that trigger on URLs in the message body. You may want to bump up the score for BAYES_99; certainly in the email flow I see there are often messages like this that only hit Bayes or Bayes plus a few very low-scoring "standard" rules. Checking later reveals hits on quite a few RBLs - in the case of SpamCop and the various URIRBLs, even 15 minutes later often shows a noticeable difference. > BTW isn't the default autolearn spam threshold supposed to be 12? Yep, IIRC. But that's using a non-Bayes score set (don't recall the exact details - see the man page) and even without different scores, this would have scored less than 12 without the BAYES_99 hit. -kgd -- Get your mouse off of there! You don't know where that email has been!
