changes:

    core:

        -   BOGUS_MX check added
            If the MX of the sender domain is either empty or resolves to
            127. or 192.168. or 10. then we add $bogus_mx_score (1.4)
            and the mail is DEFERed (_not_ REJECTed) unless the overall score
            is higher than $DEFER_LEVEL

            config: @bogus_mx_score = (1.4,        0    );
                    $DEFER_STRING   = "IN_SPAMCOP= BOGUS_MX=";
                    $DEFER_LEVEL    = 5; # defer up to this level


        -   RHSBL check changed
            All RHSBL entries are now queried
            Also some other checks (like RANDOM_SENDER) modify a $rhsbl_penalty
            which will be added to the rhsbl_score. You can change the factor
            of rhsbl_penalties via @rhsbl_penalty_score

            config: @rhsbl_penalty_score = (3.3,        0    );


        -   RANDOM_SENDER check added
            The local part of an e-mail address is now checked against:
            /[bcdfgjklmnpqrtvwxz]{5,}.*\@/i
            and
            /[aeiou]{4,}.*\@/i
            to detect randomized addresses. The default score is kept rather
            low because I know of the existance of such legitime addresses
            (just do a grep on the postfix-user ML).
            This score only high when RBL listed, also it modifies the
            rhsbl_penalty (which will score higher then).

            config: @random_sender_score = (0.1,        0    );


IMPORTANT NOTE:

            All changes imply that usual results may be different now.
            So please watch the logs and report false positives (false
            positives are such ones which appear to be semi-perfect MTAs).
            100% misconfigured MTAs/DNS are no false positives.
            For instance, rejected yahoo groups would mean a false positive.

            Also, everyone is invited to join #policyd-weight on IRCNet
            
            
-- 
    Robert Felber (PGP: 896CF30B)
    Munich, Germany

____________________________________________________________
Policyd-weight Mailinglist - http://www.policyd-weight.org/

Reply via email to