Hi Matthias,
* Matthias Fuhrmann <[EMAIL PROTECTED]> [10-05-04 06:11]:
> yes, should have been happen.
> try sa-learn with option -D when learning ham/spam. maybe there is
> something faulty.
I ran the following (part of a script called by cron) as root:
sa-learn --spam --config-file=/etc/mail/spamassassin/local.cf \
--dir /home/roland/Maildir/spam/cur/ -D \
2>&1 | mail -s "Spam Learn" [EMAIL PROTECTED]
-D added for debugging this time round.
It gave the output that can be viewed here:
http://pl.net/~rnrhill/help.htm
I then ran as root:
rrl03:~ # sa-learn --dump magic --dbpath /etc/mail/spamassassin/bayes
0.000 0 2 0 non-token data: bayes db version
0.000 0 81 0 non-token data: nspam
0.000 0 355 0 non-token data: nham
0.000 0 14249 0 non-token data: ntokens
0.000 0 1079235736 0 non-token data: oldest atime
0.000 0 1079235736 0 non-token data: newest atime
<--snip---snip--snip--snip-->
Using --dbpath to ensure the correct db location was being used.
PROBLEM: Despite sa-learn declaring "Learned from 27 message(s)
(30 message(s) examined)" the nspam count has not risen, i.e it was 81
before and after running sa-learn.
Can anyone pick out anything from the log, or suggest why this may be
happening? I wonder if I messed something up when I went to a sitewide
bayes db, although double checking the archives seems to indicate that I
have done it correctly.
Any thoughts are most welcome.
--
Regards,
Roland Hill Registered Linux User #330226
http://pl.net/~rnrhill/rhweb.htm @ http://counter.li.org
