http://issues.apache.org/SpamAssassin/show_bug.cgi?id=3109
------- Additional Comments From [EMAIL PROTECTED] 2006-04-16 17:17 ------- >> mass-check will not attempt to reuse hits from shortcircuited messages. > >Understood, however if this gets put on one of the trap servers it will >seriously hamper our ability to use realtime results for network tests. hey, I've been thinking about this. I agree that it shouldn't be used on a trap server. However I'm not mass-checking any of the traps currently; I mass-check my own spam/ham. I already reject at SMTP connect time using SBL/XBL, which means that, what is it, 40% of incoming spam is not represented in my spam corpus? I'll bet quite a few of the other mass-checkers are doing the same, or even using other SMTP-time tactics, like greylisting or other DNSBLs. As a result, if we're worried about our filtering tactics affecting our corpora, and affecting mass-check results on network tests, IMO that horse has already bolted, to a substantial degree.... ok, things still to do on this bug: - a test for shortcircuit; - I need to dogfood it on my server as well; - figure out what to do with the branch. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
