Lately, I have been using bogofilter to scan my email and detect spam.
However, it is a rather clunky solution to kick off an executable behemoth
for each mail.  I like the daemonized service that SpamAssassin offers.  So,
I have discovered Bayes learning in Spam Assassin.  Autolearning was always
on, but I am not trying to train it when it makes a mistake, using sa-learn.

My question to the list if I scan an email that was marked as spam by Spam
Assassin, what is the best way to fix it.

Currently, I think the best way is likely to be:

cat emailfile | sa-learn --forget
cat emailfile | sa-learn --spam

However, what happens if the original email was munged by SpamAssassin
already (rewrite subject, and envelope the original spam)?  Is sa-learn
smart enough to ignore its own handy work, or do I have to pass it the
enveloped version?

Thanks in advance,

Tom Veldhouse


_______________________________________________
TCLUG Mailing List - Minneapolis/St. Paul, Minnesota
http://www.mn-linux.org tclug-list at mn-linux.org
https://mailman.real-time.com/mailman/listinfo/tclug-list