Re: Debian and spam
On Sun, 2004-11-14 at 17:21, s. keeling wrote:
> Incoming from Chris Lale:
> > On Sat, 2004-11-13 at 23:04, Carl Fink wrote:
> > > On Sat, Nov 13, 2004 at 12:15:00PM -0800, Paul Johnson wrote:
> > >
> > > > I tend to prefer real email management over fake email address hacks.
> > > > Keeps everything simpler, makes the spam easier to report, etc.
> > >
> > > Who are you reporting spam to, anyway? I'd like to contribute but I'm
> > > woefully out-of-touch.
> >
> > OK, this looks like a Good Idea. So I got hold of the adcomplain Perl
> > script from http://www.rdrop.com/users/billmc/adcomplain.html. I pipe a
> > spam message to it
> > cat spam-file | perl adcomplain.pl
>
> Useless use of cat. The adcomplain documentation says:
>
> adcomplain <file
>
> > maildir files with mb2md (from Testing). The files are in a subdirectory
> > called cur with names like these:
> > 1100447087.000000.mbox:2,S
> > 1100447087.000001.mbox:2,S
> > 1100447087.000002.mbox:2,S
>
> # untested!
> #
> for f in cur/1100447087.000000.mbox*; do
> adcomplain.pl < $f
> done
Thanks! Works with slight modification:
for f in cur/*; do
perl adcomplain.pl < $f
done
>
> I haven't used adcomplain, so ymmv. Consider going to Spamcop.net,
> getting a _free_ spam reporting address, and sending your Spam to
> them. They'll analyze it to death and mail you back a URL where you
> can go to see what they came up with, and finish (or cancel) the
> report depending on what they found. If you go that way, they have a
> perl script you can use to auto-report Spam.
OK. It seems that I could use Spamcop.net to provide a blacklist for
Spamassassin. This would basically improve my filtering. Actually, the
Spamassassin wiki indicates that blacklisting is already built in for a
number of DNSBLs including Spamcop
(http://wiki.apache.org/spamassassin/DnsBlocklists). Probably, I just
need to enable it somewhere.
I am already using Vipul's Razor which is a non-commercial collaborative
database which seems to do something similar, although it is not a DNS
blacklist.
The reason that I was interested in Adcomplain is that you identify and
report direct to the originator's ISP. This was the argument put forward
by http://www.interhack.net/pubs/munging-harmful/ earlier in the thread.
This means an investment of time, but I can now automate the process
sufficiently to make this feasible for each spam message that has
slipped through the filtering process.
# script fragment to convert mbox spam file to separate files
# in subdir cur
mb2md -s \
/home/chris/evolution/local/Inbox/subfolders/SPAM-to-report/mbox -d spam
# script fragment to set environment for adcomplain.pl
export ADCOMPLAIN_MDOMAIN=mydomain
export ADCOMPLAIN_FROM=username@mydomain
#script fragment to process maildir format messages in subdir cur.
for f in cur/*; do
perl adcomplain.pl < $f
done
Reply to: