Tony Meyer wrote:
There are many ways to train SpamBayes - it's really a case of picking the
one that suits you best.  For example:

  * You can use sb_mboxtrain to train from a mbox or MH mailbox.

* You can use sb_filter (-g or -s) to train a single message.

I've been experimenting with this method but I was puzzled why the .hammiedb was not being updated. Then I realized that since I was having it run through an E-mail alias the sb_filter program was being run as user "mail" and sure enough the ~mail/.hammiedb file was being updated each time, not exactly the effect I was going for.


I tried to run it with -d ~user/.hammiedb but that didn't seem to work either.

  * You can use sb_server as a SMTP proxy (without the POP3 proxy, if you
like) and train via email forwarding.

Right, I'd like to be able to forward messages but for a variety of reasons (mostly afraid of breaking a variety of other goofy special E-mail handling I have) I don't want to put a proxy in front of my real E-mail server.


The latter is similar to what you describe, I believe.  However, it works
best when used in conjunction with the POP3 proxy so that it actually trains
from the stored copy of the message in the cache.  If you use it without
that, then you need to be sure that you are forwarding the message exactly
as it arrived, and not in a forwarded format that most mailers will produce
(ideally this also means that the To: header should be as when arrived, and
only the SMTP FROM should be [EMAIL PROTECTED] or whatever you set
the address to).

Those are the most common ways, but I've probably missed something.  If none
of these suit, please post a message describing what would, and we can see
if we can come up with a way to do what you're after.

Ah, I see. The problem is that my users will be forwarding the messages. Ugh. We also use IMAP here exclusively so the POP3 proxy won't help me anyway.



-- Greg Gulik http://www.gulik.org/greg/ greg @ gulik.org

_______________________________________________
[email protected]
http://mail.python.org/mailman/listinfo/spambayes
Check the FAQ before asking: http://spambayes.sf.net/faq.html

Reply via email to