On Thu, 2004-05-20 at 13:57, Lucky Leavell wrote:
autolearn=ham version=2.55
The ONLY place mail is filed in spamSA.$LOGNAME is from this /etc/procmailrc entry. Problem began today when bayseian filtering started working for the first time.? (Why it started working is a mystery but it did. I posted a separate thread on this problem last week and even followed thru with the SA users list to no avail.)
I can understand procmail getting confused by negative SA score BUT SA still correctly says this was No[t] spam which my procmail rule should not file the emails in spamSA.$LOGNAME but fall thru to the other rules and eventually file the mail in the user's INBOX. (This particular misfiled message was from SuSE mailing list SLE.) Please note I have been using this procmail rule for about 10 days with no problems until today.)
Any ideas of why this is happening and what I can do about it?
Thank you, Lucky Leavell
My understanding is that the bayes filtering will not start until you
have processed several hundred messages. Until then the other rules are
the only things scoring the email.
As to why your messages are getting filed as if they are spam, best
guess is the * ^X-Spam-Status.*Yes rule is not doing what you expect.
I use the following rule in my procmailrc file:
* ^X-Spam-Flag: YES
This works on my system. (trapped over 10000 spam in the last two
months).
I am using the 2.63 version of spamassissin. I assume you are using
clamav based on the clamassassin filter.
I don't know if that could be part of the issue or not since I have not
implemented clamav yet.
--
Scot L. Harris