https://issues.apache.org/SpamAssassin/show_bug.cgi?id=7013
--- Comment #10 from Kevin A. McGrail <[email protected]> --- (In reply to Mark Martinec from comment #7) > I'd be happier if this change were reverted and the definition > of the BAYES_99 rule reverted to what we had before. > > The new BAYES_999 (under whatever name) would then just match > a range from 0.999 to 1.0 and just contribute some additional > score points, complementing the BAYES_99. > > It would fix/regain compatibility with existing code base > and non-disruptively add another scoring possibility. > > Inventing a new tag could still be done later if needed. +1 as well. I'm running a make test now and will commit in a few. Can one of you please make an update on dev/users announcing the change that BAYES_99 is being reverted and BAYES_999 will be scored as an overlapping additive rule? BAYES_999 will lower to 0.2 tomorrow so people with high scores on it might want to lower them. John, the scores are static, FYI. The rules are in 23_bayes and scores in 50_scores. The sandbox work I did is what started this mess ;-) Regards, KAM -- You are receiving this mail because: You are the assignee for the bug.
