Tim Litwiller wrote:
this is the way I've been doing it in procmail - then I don't have to
count *'s
# -----------------------------------------------
# Spamassassin - certainly spam
# -----------------------------------------------
:0 H:
* ^X-Spam-Status: +(yes|no), +score=\/[^. ]*
* ? (( ${MATCH} > 14 ))
/dev/null
# -----------------------------------------------
# Spamassassin - probably spam
# -----------------------------------------------
:0 H:
* ^X-Spam-Status: Yes.*
\;Junk/new
I've never understood what procmail users have against human readable
code :-) I just use maildrop:
if (/^X-Spam-Level: \*\*\*\*\*\*\*\*/:h)
{
#do something with obvious spam
}
else
{
if (/^X-Spam-Level: \*\*\*\*\*/:h)
{
#do something else with likely spam
}
}
Very simple and easy to understand. My live server actually runs
somewhat more complex where the users "obvious spam" score is stored in
a database and retreived for comparison at local delivery time rather
than being hard coded, but anyway, I get the ability to have multiple
spam categories without source code modification to SpamAssassin.
Jay
hmm after pasting that in I wonder if there is any chance that the
catches large No scores also?
It's just easier for most filtering languages to look at the stars,
that's why there there. The yes/no only gives you a black/white world,
the score number is easy for humans to read but hard for programming
languages.
Jay
--
Jay Lee
Network / Systems Administrator
Information Technology Dept.
Philadelphia Biblical University
--
