>>1) Are we sure that no later running tests would pull the score back >>below the threshold? (whitelist, Bayes, local negative scoring >>rules)
IMO a critical point. >>3) How much time and processor savings is there really? Good question; I'm afraid I have no idea... >What about only running the RBLs if the email is below the >spam threshold? How about 1) run all ham tests first, then 2) run the spam tests until required_hits is exceeded. The 2nd part could further be optimized by running the largest scoring rules first, and the network tests last. Interesting thought, anyway. But unless this is pretty easy to do, and the developers feel it would be a pretty big performance gain for folks who elect to use it, their time is probably better spent on other tasks. I vote for finding a way to release more frequent "official", mass-checked updates to the rules :) Of course the "stop processing" feature would need to be made optional, as plenty of other software makes further decisions based upon SpamAssassin's final score. http://marc.merlins.org/linux/exim/sa.html comes to mind. But anything they feel like doing is fine with me; thanks guys for a great tool! Yesterday I had a phone call trying to sell me an anti-spam package; I have fun with them :) -------------------------- John C. Ring, Jr. [EMAIL PROTECTED] Network Engineer Union Switch & Signal Inc.
