procmail
[Top] [All Lists]

Re: Why did this receipe fail?

1998-05-29 00:57:47
On Thu, 28 May 98 12:39:47 -0400, Timothy J Luoma
<luomat+Lists/procmail(_at_)luomat(_dot_)peak(_dot_)org> wrote:
     Author:        Steve Woodard <woodard(_at_)kodak(_dot_)com>
* ^Subject:.*(sex|xxx)
* ^Subject:.*adults
Why did mail with this subject line get through my filter?-
Subject: Wholesale XXX Videos
It got through because your recipe asks for 4 matches, including 2
Subject lines.
That recipe will never ever match anything.

Except of course something like "xxx adults" which does match both
conditions. Or even "adultsex", which brings up an interesting point:

This can be a useful construct for cases where you look for two
phrases which might e.g. overlap partially. The classical example of
this is word-token scoring, where

    * 1^1 ()\<(big|fat|hairy)\>

would only score two on "big fat hairy" (it gets \<big\> and \<hairy\>
but there is not enough whitespace left to match \<fat\>) whereas this
would get a score of three (as you would perhaps expect, or hope of
the other recipe as well):

    * 1^1 ()\<big\>
    * 1^1 ()\<fat\>
    * 1^1 ()\<hairy\>

It would of course be extremely nice if \< and \> would work like in
other regex implementations, but I digress ...

/* era */

-- 
 Paparazzi of the Net: No matter what you do to protect your privacy,
  they'll hunt you down and spam you. <http://www.iki.fi/~era/spam/>

<Prev in Thread] Current Thread [Next in Thread>