At 12:12 PM 1/27/97 -0500, Timothy J Luoma wrote:
I get email from a SPAM fighting group with the subject like this:
Subject: spam: waltercan(_at_)worldnet(_dot_)att(_dot_)net: Hello
Where the first word is "spam: " the second word is the offending
email address, and the third word is the subject of the message that
this offendee sent.
I would like to take the second argument (the email address) and
REMOVE the ":" at the end of the email address and paste that to a
file of known spammers:
I've been thinking about this:
:0
* ^From spam-list-owner(_at_)toby(_dot_)han(_dot_)de
* ^Subject: spam: \/.*$
{
spammer=`echo $MATCH | awk '{print $1}' |tr -d ":"`
:0hi: spamfile.lock
* ! ? grep -s -i -w "$spammer" $PROCDIR/killfile-by-from-line
|/bin/echo "$spammer" >> $PROCDIR/killfile-by-from-line
:0
|appnmail DuplicateSpam
}
The only thing I am worried about is the "tr -d" line. Should I be
worried about that? is there a better way to do this?
If you're sure the "word" ends in a ":" then you can skip the
tr process altogether, using a little more of awk:
spammer=`echo $MATCH | awk '{print substr($1,1,length($1)-1)}'`
This drops the final character from $1, whatever it is. Should perform
better, but I don't know if it's "best" :-)
and about the "DuplicateSpam" thing... I put that there to catch if
I get two reports about the same guy (ie they are already in the
killfile). Is there another way of doing this? I don't really care
to see the message, I suppose I could send it to /dev/null.
If the duplicates are rare, they won't hurt performance much (the
extra checks may cost more); you might consider just running your file
through "sort -u" (while locked, of course) once in a while.
Cheers,
Stan