tokenizing of forwarded messages
Mike Lykov
combr at vesna.ru
Wed Nov 5 10:24:47 CET 2003
Hi all.
Continuing questions ;) Thanks for payed attention.
Among some of my users I distribute information about fighting spam by this
manner:
I create special mailbox spamblock@ and users forward to it all spam they
don't need. I receive it on my computer and see, put on server and register
in base by hand (can make it automatically, but it is not guarantee from
user's error).
But in that emails exists following part:
Received: by mail.vesna.ru (Postfix, from userid 501)
id 8BAD62116; Wed, 5 Nov 2003 11:14:30 +0400 (SAMT)
Received: from tdv_mrk14.vesna.local (tdv_mrk14.vesna.local [192.168.4.108])
by mail.vesna.ru (Postfix) with ESMTP id 3A82120FE
for <spamblock at vesna.ru>; Wed, 5 Nov 2003 11:14:28 +0400 (SAMT)
X-Mailer: The Bat! (v1.62r)
Organization: CF Vesna
X-Priority: 3 (Normal)
Message-ID: <4510704438.20031105111426 at vesna.ru>
To: spamblock at vesna.ru
And other headers - far from original spam.
"This is a forwarded message
===8<==============Original message text==============="
in body also not exist in original spam.
How it is influence to classifying other (non-spam) emails with the same
headers/words?
Can this headers to mess up the spambase ? I do not want to put in spam
wordlist headers from normal letters like it, and words like "forwarded
message"
Are there method correct or you not to recommend it ?
--
Mike
registered linux user #315334
jabber id: combr at jabber.ru
More information about the Bogofilter
mailing list