home *** CD-ROM | disk | FTP | other *** search
- # These are the settings for SpamHeaders
- # Lines starting with a '#' are ignored
- # Each line defines a spam filter
- # A filter scores a header, and gives a score between 0 and 100, with 0
- # meaning "Not Spam", and 100 meaning "Spam"
- #
- # A filter consists of 5 parts:
- # A pattern to match the entire header against
- # Score - this is only used in the case of "kTextTag" - if the text matches use this score.
- # Offset - offset the score from this value.
- # Scale - Scale the value by this value.
- # A description of what the sub-patterns are. This takes one of four types:
- # kTextIntMatch - value is an integer
- # kTextFloatMatch - value is a floating point number
- # kTextTag - value is a string literal
- # kCharacterCount - value's length is the important thing
- #
- # For example, Postini gives you a score that is between 0 .. 100, with low scores being spam.
- # We give them a scale of -1, and an offset of 100; between them, this scales the scores into
- # the range that we want.
-
- # PerlMX - no scaling, since the scores are already 0 .. 100
- "X-Perlmx-Spam:.*Probability=([0-9]+).*" 0 0 1 kTextIntMatch
-
- # Pure Message - no scaling, since the scores are already 0 .. 100
- "X-Puremessage-Spam:.*Probability=([0-9]+).*" 0 0 1 kTextIntMatch
-
- # Spam Assassin (both forms)
- # The actual score is 0 .. 10, so we scale by a factor of 10
- "X-Spam-Score:.*\((\*+)\).*" 0 0 10 kCharacterCount
-
- # If the text matches, we give it a score of 99
- "X-Spam-Flag: YES.*" 99 0 1 kTextTag
-
- # Postini - several differenct scores; we picke the one that gives us the highest score.
- "X-pstn-levels: +\(C:([0-9\.]+) M:([0-9\.]+) P:([0-9\.]+) R:([0-9\.]+) +S:([0-9\.]+) \).*" 0 100 -1 kFloatIntMatch
-