Re: [cobalt-users] SpamAssassin forwarding mailbox [SCANNED]

["VFastlink Support Team" <support@xxxxxxxxxxxxx>]

----- Original Message ----- 
From: "Dave's List Addy" <listonly@xxxxxxxxxxxxxxxxxxxx>
To: "Users" <cobalt-users@xxxxxxxxxxxxxxx>
Sent: Friday, April 09, 2004 11:14 AM
Subject: Re: [cobalt-users] SpamAssassin forwarding mailbox [SCANNED]


> On 4/9/04 11:05 AM, "Jeff Lasman" wrote:
>
> >
> >> SpamAssassin uses the header information
> >> as part of it's criteria.  We were hoping someone had made a script
> >> that would strip out that extra header information from the mal and
> >> then run the SpamAssassin command.  There is a sample one on the
> >> SpamAssassin site, but it doesn't work very well.
> >
> > If you use procmail instead of forwarding there won't be any extra stuff
> > added to the headers to strip out.
> >
> > As long as you're staying on the same server this will work fine.  If
> > you're not staying on the same server then once a day zip up the
> > mailbox and move it the box you need it on, and unzip it there.
> >
> > That script would be so simple even we could do it <smile>.
>
> I think the problem he is running into is that spamassassin will allow
users
> to submit what they consider spam or non-spam, there by creating 2
> mailboxes, and letting Bayes auto-learn from that, If I am not mistaken.
>
> The problem is as we all know, headers get munged when they are forwarded
to
> a new box. The script they show at SA is a little rough, and while your
idea
> is good Jeff, it still wouldn't address the users trying to submit there
> spam/non-spam emails and not alter the header which is a large part of
what
> SA will learn from.
>
> But then again I could be or usually am wrong...
> -- 
> Thanks!!
> David Thurman
> List Only at Web Presence Group Net
>
>
> _____________________________________
> cobalt-users mailing list
> cobalt-users@xxxxxxxxxxxxxxx
> To subscribe/unsubscribe, or to SEARCH THE ARCHIVES, go to:
> http://list.cobalt.com/mailman/listinfo/cobalt-users
>

Right you are David!  Thank you for explaining better than I can. ;-)