> That's something that a number of people have requested, but has two
> drawbacks: It requires lots of programming time to create, and lots of CPU
> time.

Hmmm...  Not to be a pest, but I'm wondering if this wouldn't actually
IMPROVE performance?
The problem is that it requires going through the E-mail one character at a time and running a test against each of the filters. Each of those tests is much more involved than a string match (which most of the time just requires comparing 2 bytes).

If all that is being added is a single character that is used to replace a single character, it wouldn't be so bad. But once you go a step beyond that -- a single character representing punctuation but not letters, for example, or "*", or regexp expressions, it can get much more complex quickly.

RULE "free~" finds "free" "free!" "free." "free?" etc. but not "freedom"
or "freeze" -- all in one pass.  It covers STARTSWITH, CONTAINS, IS and
ENDSWITH in one shot.

RULE "~sex" finds "sex" "sexy" "sexiest" "sexaholic" "sex!!!" etc. but
not "Essex" or "unisex" -- all in one pass.  Again, operators
STARTSWITH, CONTAINS, IS and ENDSWITH are all covered.  One rule.
True -- it would likely save CPU time over having multiple filter entries.

Again, this is something that we are looking into, but we just haven't made any final decisions about.
-Scott

---
[This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]

---
This E-mail came from the Declude.JunkMail mailing list. To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail". The archives can be found
at http://www.mail-archive.com.

Reply via email to