Kyle Quillen wrote: > Hello all, > > I have a bit of a problem. I am trying to come up with a solution > that I think will help a lot of people who use the qmail toaster mail > server. I have my toaster setup to drop mail tagged as spam into spam > directories in each users mailbox. I am planning on putting a Ham box > in their web mail as well. I would like to come up with a script for > the toaster community that will tell spamassassin to learn from the > mail in each of these directories on a regular basis. I would like > the script to delete mail in the spam and ham dirs on a regular basis.
That is fairly easy to do and I think there has already been one script posted. I just wanted to comment on what it sounds like you are doing. Please make sure that you are only learning messages that have been sorted by a person. Otherwise, false positives and false negatives will very quickly make your Bayes DB useless. Leave the autolearning on if you like, but manual learning should be limited to manually sorted mail, otherwise it is not helpful. -- Bowie
