[prev in list] [next in list] [prev in thread] [next in thread] 

List:       vchkpw
Subject:    Re: [vchkpw] SpamAssassin, Vpopmail, conditional handling on Spam
From:       Andrew Lawton <andrew () exit0 ! us>
Date:       2003-05-29 2:48:59
[Download RAW message or body]

Remo Mattei wrote:

> I got some of 40+ just for your info. This will be a good extra thing :)
> Let us know, 
> 
> 
> -- Remo Mattei   --cell 801-209-8554
> http://www.italy1.com
> Freelance Networking-Security/Consultant
> MCSE, MCP, MCP+I, MCT
> Linux Trainer & Firewall Development
> remo@italy1.com
> 
> 
> 
> 
> > From: Dave Richardson - Lists <isp@derdev.com>
> > Reply-To: isp@derdev.com
> > Date: Tue, 27 May 2003 16:42:33 -0500
> > To: vchkpw@inter7.com
> > Subject: [vchkpw] SpamAssassin, Vpopmail, conditional handling on Spam Score?
> > 
> > I'm running the QMAILQUEUE patch with SpamAssassin into Vpopmail with
> > Mysql and Courier.
> > 
> > I'm running SpamAssassin with spam threashold "required_hits" = 8.
> > 
> > I allow direct delivery to the mailbox, but it is subject prefixed
> > "[SPAM]"  (thus it is Spam Tagged)
> > 
> > What I'd like to do is add another "threashold" (probably in SA?) where
> > the spam score of >15 is sent to /dev/null.  Scores from 7-14.99 are
> > spam tagged (and delivered) and scores less than 7 are not spam tagged.
> > (I've been seeing spam scores as high as 28.8 recently BTW)
> > 
> > Is there a "normal" way to do this?
> > D.
> > 
> > 
> > 
> > 
> > 
If you asked this question over at the SpamAssassin list, they would advise against \
sending high scoreing hits to /dev/null. There are two main reasons. One is (believe \
it or not) you actually could get a false positive removed. This usually happens with \
html based newsletters. The other reason is that high scores are useful for training \
the Bayes filtering.


Just something to keep in mind.

> 
> 
> 


[prev in list] [next in list] [prev in thread] [next in thread]