Bill Randle wrote:
> On Sun, 2008-07-13 at 16:52 -0500, Chris wrote:
> > Bowie, you wrote this script back in 2006, I've been running it
> > since back in Aug 0f 2006 and today just noticed something. The
> > addition doesn't seem to be quite right. For instance:
> >
> > FreeMail.cf:
> > Rule Name Score Ham Spam %of Ham
> > %of Spam
> >
> >
> >
> >
> > -----------------------------------------------------------------------
> > FREEMAIL_REPLYTO 2.00 2 20 1.87%
> > 14.60% FREEMAIL_FROM 1.00 8 56 7.48%
> > 40.88%
> > -----------------------------------------------------------------------
> > OVERALL 8 56 7.48%
> > 40.88%
> >
> > iXhash.cf:
> > Rule Name Score Ham Spam %of Ham
> > %of Spam
> >
> >
> >
> >
> >
> > -----------------------------------------------------------------------
> > LOGINHASH 4.50 0 3 0.00%
> > 2.19% LOGINHASH2 2.50 0 3 0.00%
> > 2.19% IXHASH 2.50 0 8 0.00%
> > 5.84%
> > -----------------------------------------------------------------------
> > OVERALL 0 9 0.00%
> > 6.57%
> >
> >
> > I'm not good at scripts so possibly you could find a way to fix
> > this? I would have emailed you privately however, I don't have your
> > email address so had to post to the list.

>
> Chris,
>
> This is actually working as designed. According to the comments in the
> code, the OVERALL numbers are "number of spams hit by the file, not
> just number of rule hits". In other words, since some spam hit
> multiple rules in the same .cf file, this is a count of the number of
> unique hits to that whole set of rules.
>
> In your first example, of the 8 messages that hit FREEMAIL_FROM, 2 of
> them also hit the FREEMAIL_REPLYTO from the same .cf file, so the
> total is 8 not 10.


Exactly.

I'm glad you find the script useful. I haven't looked at that script in
a while. I notice that it's not particularly useful if you are using
sa-update since the rule files are scattered across the update
directories. Maybe I'll find some time to add a fix for that.

--
Bowie