>> In my environment, postfix passes the message onto the exchange server >> so once it releases the message, I don't have anything to train bayes >> with since it's deleted. >> >Add an 'always_bcc' directive to your Postfix configuration to grab a copy of >all mail passing through it and send it to a capture mailbox. >Use a procmail recipe to classify mail arriving in the capture mailbox as ham, >spam or indeterminate and file it appropriately for input to sa_learn. > >Martin
That is perfect! I've done that and it saves the mail locallly. The only problem is that when I open the file for the users mailbox, it makes all of the email as one large text file with one email after the next. Is that normal? I wouldn't have to go through it and separate each mail would I? Kaleb
