I use SpamAssassin 3.0.1
I was a bit surprised to notice in my Exim logs that spamassassin --report
emails the spam to SpamCop (among other things, of course). So I looked at
the config options connected with this:
spamcop_from_address [EMAIL PROTECTED] (default: none)
This address is used during manual reports to SpamCop as the From:
address. You can use your normal email address. If this is not set, a guess
will be used as the From: address in SpamCop reports.
spamcop_to_address [EMAIL PROTECTED] (default: generic reporting address)
Your customized SpamCop report submission address. You need to obtain
this address by registering at http://www.spamcop.net/. If this is not set,
SpamCop reports will go to a generic reporting address for SpamAssassin
users and your reports will probably have less weight in the SpamCop system.
I hadn't set either of these (in particular the second one), so the generic
reporting address was being used. I thought of setting spamcop_to_address to
be my own registered SpamCop reporting address (indeed I might at some point
choose this option), but this means that I get an email response from
SpamCop for each report, and then need to log into the SpamCop website to
complete reporting for each of them. If reporting to SpamCop myself (_not_
using SpamAssassin), then I like to attach a load of spams to the same
report, then I only get a single email response.
But what I was wondering is what SpamCop do with all the reports to the
SpamAssassin generic reporting address? Presumably nobody is receiving email
responses for these, or completing reporting on the web page. Did the
SpamAssassin team make a special arrangement with SpamCop so that these
generic reports would at least carry some weight, even though nobody gets to
confirm them? Or are they ignored, which would mean it was rather pointless
SpamAssassin sending them?
Finally: I've set up a 'reporting' email address on my server, so that spam
can be 'bounced' to it, and the server automatically pipes it into spamassas
sin --report (mainly to train my Bayes, but of course it also goes to
Razor/SpamCop/etc.). I can see that I'll have to be very careful here that
the host _from which_ I bounce the spam doesn't become implicated as the
source of the spam (perhaps by only bouncing from localhost). Has anyone any
thoughts on this?
Regards,
Clarke Brunt
