http://bugzilla.spamassassin.org/show_bug.cgi?id=3705
------- Additional Comments From [EMAIL PROTECTED] 2004-08-24 17:50 ------- I'm not the one who submitted the original bugzilla entry, but re: "This feature is barely used (did it even work in 2.64?), maybe we should just nuke it if we're not going to test it." Since I run with a required-hits of 9 instead of 9, I need to increase a fair number of default scores to keep things in balance. I expect to make fair use of this capability. The exercise of changing score over-rides took a fair amount of manual/personal analysis in the move from 2.5x to 2.6x, and will do so again going from 2.6x to 3.0. I don't adjust EVERY score, of course, or else there'd be no benefit to changing req-hits. I find those that contribute significantly here, decent S/O on the spam my specific system gets, and bump those scores. I then decrease those that contribute to FPs on my specific systems. This function can be useful for me, and while I can't put together an installation test (don't know how), I'll certainly download the latest SVN this weekend and test the functionality itself. I don't suppose that in addition to a numeric score modification score WEB_BUGS (+.5) score CLICK_BELOW (+1) score KNOWN_MAILING_LIST (-.75) it's also possible to get a percentage modification score WEB-BUGS (+180%) score CLICK_BELOW (+0.10%) score KNOWN_MAILING_LIST (-100.0%) My thinking is that a Good Rule, like SUBJ_ILLEGAL_CHARS (good here) should get a full 180% increase here (9/5), while less good rules would get smaller increments. Then, moving from 3.0 to 3.1 to 3.2, each with its own scoring, I'd not have to change these scores, knowing that they'd adjust in the same proportion. If not available here, it'd be nice for a 3.1 enhancement... Bob Menschel ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
