Sorry for replying to my own message, but that should have read: http://sandgnat.com/cmos/cmos.jsp?sourceRules=body+CUM+/\bcum\b/
(http, not https) Chris Thielen said: > Lyle Evans said: >> At 10:52 AM 11/10/03, Jason wrote: >> >>>... >>> >>>I was thinking today wouldn't it be better to just ignore all the >>> periods, >>>commas, and what have you in the text? Inside SA we could just drop >>> those >>>and then search the message from that. >>> >>>I've had one spammer who just puts a random period in the message and it >>>doesn't get tagged. Taking out all the periods in the message and it >>>scored a 10.6 just from the body of the message. >> >> Yes I am being hit hard by he same type of spam. >> >> While I think chasing variants is in general a losing battle >> and that instead more general rules are needed such as ones to eliminate >> internal periods, I just did a quick and dirty rule: >> >> body LE_bp_Naughtydot / CU\.M/i >> describe LE_bp_Naughtydot Body Naughty word with inserted dot >> score LE_bp_Naughtydot 2.85 >> >> The score is more or less arbitrary. The logic is tackle the >> the short words that can't have as many variants. >> A N letter word can have N-1 internal single dot variants. >> Suggestions for improvement strongly encouraged. > > Try this: > > https://sandgnat.com/cmos/cmos.jsp?sourceRules=body+CUM+/\bcum\b/ > > tweak as necessary.. I'm still trying to track down some false positives I > have as the generated rules catch many permutations. ------------------------------------------------------- This SF.Net email sponsored by: ApacheCon 2003, 16-19 November in Las Vegas. Learn firsthand the latest developments in Apache, PHP, Perl, XML, Java, MySQL, WebDAV, and more! http://www.apachecon.com/ _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
