"Timothy J. Luoma" <luomat(_at_)peak(_dot_)org> wrote:
I now have 110 domains that I refuse to accept mail from.
Really?
I have been using this:
:0
* ^From.*(telysis\.com|\
[snip for brevity]
usa\.net|\
And what do you have against my provider netusa.net? This is
not a spam tollerant domain.
<rant>
I find it VERY rude when people who do not understand regular
expressions (ab)use them in spam filters. If you don't understand
how pattern matching works, you should not be using it.
Just because you have seen some spam that comes from a site called
"usa.net" does not mean that all mail from sites with "usa.net"
in the name are spam sites.
(Cf. n.a.n-a.usenet discussion of alt.sex.stories.) Just because
you have seen some spam with the word "free" in it does not mean
you should issue spam cancels for all posts with the word "free"
in them.
</rant>
At the very least you should have the procmail fake word boundry
marker before the open paren and "(\>|$)" at the end of this:
softcell\.net)
I would like to dump the messages if any of these hosts are found
in any of these headers:
From
Received
From:
Message-ID:
Return-Path:
X-Sender:
Return path and From contain information dervied from the same
source. You only have to check one of the two.
Is there a way to put all the domains in a separate file (which
would be easier to maintain) and have them match any of those
headers?
From my procmailrc:
# grep options: -i case insensitive, -q quiet, -F fgrep, -w check for
# word boundries, -f read patterns from file
:0 Wi
* ? (formail -x from: -x sender: -x received: | grep -iqFwf $SPAMREJECTS)
$MAILDIR/spam-output
I am using gnu grep. Pretty much the only stuff in the $SPAMREJECTS file
are the Cyberpromo domains, for me. I often get real mail from usa.net,
for instance. This is the last recipe in my procmailrc and it catches
very little, my earlier tests are fairly effective.
Elijah
------
please do not CC me when replying to the list