Mat Harris <[EMAIL PROTECTED]> wrote on 22.03.2004 17:44:22:

> On Mon, Mar 22, 2004 at 05:14:23 +0100, [EMAIL PROTECTED]
wrote:
> > Hi
>
> Hello :)
> >
> > I have sent this already a wile ago, maybe my question is stupid? Or
did I
> > miss some documantation? Please tell me...
> >
> > I have setup an email proxy using postfix and spamassassin. All emails
are
> > received by this host and then relayd to our incoming Lotus Domino
server.
> >
> > Now, I'd like to give my users the possibility to submit ham and spam
to my
> > SA host via email, I would create a local address for ham and spam on
the
> > SA host.
> >
> > My question is: How does sa-learn interpret mails. What headers are
> > included in the "calculation". Do I need to do any rewriting of the
emails
> > received on the ham and spam addresses on the SA host, because they
were
> > relayd to a different server? What happens with additional headers
added by
> > our Domino server?
> >
>
> as far as i can see it, you would need something like this:
>
> 1) mail recieved by relay for learning
> 2) pipes mail thru prog listed in /etc/aliases or procmail
>    (or whatever postfix uses)
> 3) prog would parse mime headers and output innermost (oldest)
>    part (the spam/ham)
> 4) Output is piped to sa-learn with the appropriate switches.
>
> I would recommend perl for the filter but I've never got mime
> parsing to work as much as I'd like so I'll leave that part to
> someone else :)
>
> The decision between learning as spam or ham could be made by
> procmail based upon the To: or Subject: lines.

Is bayes learning all header values or ist it limited to certain headers

>
> > Thanks for any answers.
> >
> > Herbert
>
> HTH
>
> --
> Cats land on their feet.
> Toast lands jellyside down.
> A cat glued to some jelly toast will hover in quantum indecision
>
> perl -e'$_=q#: 13_2: 12/o{>: 8_4) (_4: 6/2^-2; 3;-2^\2: 5/7\_/\7: \
>    12m m::#;y#:#\n#;s#(\D)(\d+)#$1x$2#ge;print'
>
> Yes, of course it's the right cabl [le0: NO CARRIER]
> [attachment "attj208m.dat" deleted by Herbert Augustiny/ROTH/SPT]

Reply via email to