I am new to SpamAssassin and have collected quite a bit of mail to send to sa-learn. The problem is that all my mail is in Maildirs not in mboxes. How might I go about teaching SpamAssassin?
sa-learn supports maildirs just fine.
By default sa-learn assumes that if a name passed is a directory, it must be a maildir and will learn all the mail in it.
However, be sure to pass names to sa-learn, don't use redirection.
ie: sa-learn --ham /home/joe/maildirham not: sa-learn --ham < /home/joe/maildirham
Unlike spamassassin, sa-learn understands the idea of being passed filenames so redirection, while supported, merely limits sa-learns abilities.
The only issue is you're going to have to generate separate ham and spam directories, but that's not too hard..
While I am asking, the mail sits on one computer and the SpamAssassin gateway is sitting on another. I was going to simply tar up the mail and copy it to the gateway machine, is there a better way?
That's the way I would do it. In theory you could do some kind of SMB or NFS share, but that seems like way too much work.