SA-Learn Question

["Mark A. DeMichele" <de...@intellipro.com>]

I've been using SA for about a month now without bayes.   I think I'm
ready to try it.  I have a question though.  Right now my SA is set up
to modify messages it thinks are spams to have the original message in
an attachment.  Is SA-Learn smart enough to recognize this and scan only
the attachment?  I would think it would have to be.

Thanks,
Mark DeMichele

Re: SA-Learn Question

[Matt Kettler <mk...@evi-inc.com>]

At 09:06 AM 2/20/2004, Mark A. DeMichele wrote:
>I've been using SA for about a month now without bayes.   I think I'm 
>ready to try it.  I have a question though.  Right now my SA is set up to 
>modify messages it thinks are spams to have the original message in an 
>attachment.  Is SA-Learn smart enough to recognize this and scan only the 
>attachment?  I would think it would have to be.

 From the BayesFAQ on the spamassassin wiki:

http://wiki.spamassassin.org/w/LearningMarkedUpMessages

In short: as long as the markup, modifications, etc are those done by 
spamassassin itself, and not some other tool, sa-learn can recognize and 
reverse them. 

Re: SA-Learn Question

[Pedro Sam <p2...@yahoo.com>]

On Friday February 20 2004 09:06, Mark A. DeMichele wrote:
> I've been using SA for about a month now without bayes.   I think I'm
> ready to try it.  I have a question though.  Right now my SA is set up
> to modify messages it thinks are spams to have the original message in
> an attachment.  Is SA-Learn smart enough to recognize this and scan only
> the attachment?  I would think it would have to be.

yes... according to "man sa-learn" ...

-- 
You have an unusual magnetic personality.  Don't walk too close to
metal objects which are not fastened down.