Stig, Would you mind sharing your set of procmail filters?
Regards, Alex. * Stig Brautaset ([EMAIL PROTECTED]) wrote: > > Err... No? Whatever gave you that idea? Did you not read my post? > > My spam filter caught the original spam, based on such things as sender > address, character set used, adjacent crap characters in subject > ([EMAIL PROTECTED]&* etc.) and body, whether the subject is empty or lacks > lower > case characters (all caps subjects are seldom legit mail), certain > words in the body and more. None of these are enough to trigger the > filter on their own--that would create far too many false positives. > > However, when a follow-up to a previously caught spam is detected > (using the references/in-reply-to header) it is instantly filed as spam. > > When the OP reposted the spam he did so with a "legal" subject line, > announcing a (to me) understandable charset, a non-depreciated sender > address and without the references/in-reply-to header that would link it > to the previous spam so I could instantly blackhole it. > > Please explain to me, in the light of this information, how I can > improve my spam-detection routines so that I would be able to catch the > original post of this tread. I am genuinely interested to find out how > you think that is possible. > > Stig > > -- > brautaset.org > Registered Linux User 107343 > > ``Oh, how I wish `undo' was ported to everyday life.'' > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > Oleksandr Moskalenko <[EMAIL PROTECTED]> pub 1024D/6C5F196B 2001-08-17 /* http://www.tagancha.org/pgp */ Oleksandr V. Moskalenko (Alex) <[EMAIL PROTECTED]> Fingerprint = EE63 C471 ADBA 5D80 ADFB 1054 DA28 6F32 6C5F 196B
