On Tue, Dec 11, 2001 at 12:45:43PM +0100, Thomas Roessler wrote: > On 2001-12-10 14:31:54 -0800, Vineet Kumar wrote: > > >I think your best bet is to check them at delivery time via your > >MDA (maildrop, procmail, etc.) Have it add a header indicating its > >results, and have mutt perform checks based on the presence or > >contents of this header. > > I'd also recommend some kind of reversible spam filtering - after > all, razor has had a couple of bad false positives in the past week > or so, including bugtraq posts and messages to the razor-users > mailing list. > > Possible approaches include using mutt's scoring, or creating a > special folder to which all messages identified as spam by razor are > written.
I installed the spamassassin a week ago and I *love* it. http://spamassassin.taint.org or from CPAN It uses Vipul's Razor as one of its spam scoring methods. By default the weight 3 of a possible 5 so that the Razor alone is insufficient for spamassassin to mark a message as spam, but gets it really close. Some other feature like a forged From addr is enought, so it is damn impressive. The .procmailrc below makes a reversible spam filter for me (I skim with 'mutt -f =caughtmail' periodically to see if it made any mistakes) and added a mutt macro in my .muttrc to report spam it didn't catch to the Razor. Here are the snippets from my dotfiles. I got this from spamassasin's own README file... the URL follows the dotfile snippets. -- my .muttrc -- ... # Give the message to Vipul's Razor macro index X "| spamassassin -r" ... -- my .procmailrc -- :0f | /usr/local/bin/spamassassin -P :0e { EXITCODE=$? } :0: * ^Subject:.*\*\*\*\*SPAM\*\*\*\* mail/caughtspam -- The recipe -- http://www.cpan.org/modules/by-authors/id/J/JM/JMASON/Mail-SpamAssassin-1.2.readme
