changes:

- (change)    The cache now uses only one query to ask for HAM|SPAM.
              Thanks to H. Krohns.

- (change)    the $dnsbl_hits score get only increases if the RBL is a
              blacklist. This is required to use whitelists in the $dnsbl_score
              list. Thanks to 'Steve'.

- (change)1   The cache can now return hard-whitelists.
              Enforce N days/hours RBL checks for HAM clients.
              After N days do each P hours or each R requests a RBL check for
              HAM cached clients. This should save a good amount of RBL DNS
              traffic.

- (change)    The delay in seconds for each policy request is now logged via
              "decided action"

- (change)    The time required for a cache cleanup is now logged, too.

- (change)2   $ENFORCE_DYNDNS config setting added. Default 1 (on).
              This tells policyd-weight to enforce clients who use dynamic
              hostnames in its HELO to use a DynDNS hostname.
              This is skipped if the client is a MX for HELO or SENDER.
              In earlier versions this check was only performed if the client
              was NOT on a DUL AND the resolved client hostname appeared to
              be NOT dynamic.
              WARNING: such clients may become very high scorers

- (scores)    @helo_seems_dialup = (1.5,0)
              'dsn.rfc-ignorant.org', 3.5, 0, 'DSN_RFCI',
              ordb.org removed



Additional info to 1:   The usage of a hard cache is logged via "rate:hard:"
                        The times when to use a hard cache, and when to
                        temporarily expire a hard cache are documented in the
                        'defaults' output ($PTIME, $TEMP_PTIME, $PTTL)

Additional info to 2:   This was necessary as some spammers started to change
                        their software (www.policyd-weight.org/new-stuff.txt)
                        Given that we are low volume the relation of new stuff 
                        which passed vs normal spam was rather high.


p0f has not yet been built it. I expect this to happen with 0.1.15

-- 
    Robert Felber (PGP: 896CF30B)
    Munich, Germany

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

Reply via email to