[Mimedefang] MD + SA + bayes problem
Paul
freebsd at pksnet.com
Mon Sep 29 20:15:01 EDT 2003
>> I use 'sa-learn -C /etc/mail/spamassassin/sa-mimedefang.cf --spam
>> --rebuild -D 5 --single spam'.
>
>I don't think you can use --rebuild at the same time as training.
>Also, "--single" is not recognized syntax for the version you're
>running -- use "sa-learn --spam --file [filename]".
Hmmm, I noticed the --single problem as I just recently upgraded to 2.60 and tried the --file option.
Your tip above worked in so far that now at least I get a message at the end telling me it learned from 1 message. Something I didn't get before. Guess I did not read the perldoc coirrectly as to me it appeared you could do a --rebuild and have it rebuild the database at the same time...
>What output do you get from "sa-learn --dump magic"?
before '"sa-learn --spam --file spam"
dax# sa-learn --dump magic
0.000 0 2 0 non-token data: bayes db version
0.000 0 0 0 non-token data: nspam
0.000 0 0 0 non-token data: nham
0.000 0 0 0 non-token data: ntokens
0.000 0 1064611203 0 non-token data: oldest atime
0.000 0 1064611203 0 non-token data: newest atime
0.000 0 1064611203 0 non-token data: last journal sync atime
0.000 0 1064611203 0 non-token data: last expiry atime
0.000 0 0 0 non-token data: last expire atime delta
0.000 0 0 0 non-token data: last expire reduction count
after '"sa-learn --spam --file spam" with 1 spam in it.
dax# sa-learn --dump magic
0.000 0 2 0 non-token data: bayes db version
0.000 0 1 0 non-token data: nspam
0.000 0 0 0 non-token data: nham
0.000 0 359 0 non-token data: ntokens
0.000 0 1064611203 0 non-token data: oldest atime
0.000 0 1064718512 0 non-token data: newest atime
0.000 0 1064611203 0 non-token data: last journal sync atime
0.000 0 1064611203 0 non-token data: last expiry atime
0.000 0 0 0 non-token data: last expire atime delta
0.000 0 0 0 non-token data: last expire reduction count
Thanks! Seems that was what I was missing. I'm going to adjust my 'learnspam' scriptaccordingly
Paul
More information about the MIMEDefang
mailing list