At 8/27/03 09:37 AM , Robin Witkop-Staub wrote:
I was trying the following filter found on the suggested script sharing
page.  For some reason it was tagging emails that had the word "pizza" in
them.  Why?

body     RAVEN_MaskedWordsF
/\b(?:excIusive|GiangBiang|sIut|ganigbainged|duides|hairdciore|ExcIude|pIz)/
i

Looking at the other uses of capital I in the regex, I think the regex writer was trying to catch obfuscated versions of "plz". But I think the logic behind this whole regex is kind of shaky -- I often gets substituted for L, but the other way around seems far less common.


describe RAVEN_MaskedWordsF    masked spam word(s)
score    RAVEN_MaskedWordsF    10.0

And giving something 10 points for this seems a little strange. Since this is a case-insensitive match, the "sIut" part could easily be matched by "siut" -- an obvious transposition typo for "suit". Or by "Siuth" (typo for "South"), since there's no \b anchor on the tail end.


This rule seems ready for some very interesting FPs. (But then again, you've already noticed that.)

                                                --Kai MacTane
----------------------------------------------------------------------
"These eyes see more than window dressings
 and stoic virtue of mannequins:
 Somewhere beyond the world of single dimension,
 Beyond the façade of the sensual,
 Where nothing cuts you deeper than your own blade."
                                                --The Last Dance,
                                                 "Fairytale (the Storm)"



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to