Michal Libra wrote: > Hi there, > have a question about methods of spamassassin bayes learning. > I found some HOWTO where was described that admin created two special > mailboxes- for example [EMAIL PROTECTED] and [EMAIL PROTECTED] > The users was instructed to send bad marked emails to the > corresponding mailbox. Some kind of script regularly gets email from > mailboxes and runs sa-learn to learn new hams and spams. > It seems to me like a good idea how to let users manage the anti spam > control. > > And finally the question. I'm not sure if the "forward process" does > not make any damage to spams. Is it safe? No > Does it work? Sorta, but may have undesirable results. > > Other tips are welcome. Suggestion: have them "forward as attachment" instead, and use a script that rips attachments and feeds those to sa-learn.
