Hello,
It seems bayes is not working properly for me.
I have sendmail/MailScanner/Spamassassin setup for my users.
MailScanner version 4.56.8
Spamassassin version 3.1.3
autolearn is on
------------------------------
sa-learn --dump magic
0.000 0 3 0 non-token data: bayes db version
0.000 0 487135 0 non-token data: nspam
0.000 0 560705 0 non-token data: nham
0.000 0 139626 0 non-token data: ntokens
0.000 0 1163723330 0 non-token data: oldest atime
0.000 0 1163767343 0 non-token data: newest atime
0.000 0 0 0 non-token data: last journal sync
atime
0.000 0 1163766599 0 non-token data: last expiry atime
0.000 0 43200 0 non-token data: last expire atime
delta
0.000 0 49483 0 non-token data: last expire
reduction count
------------------------------
The last journal sync being 0 doesn't seem correct...shouldn't this be
some positive number?
I even ran sa-learn --force-expire.
When I use 'spamassassin -D spamtest' to feed through a known spam, this
is one section of the debug with respect to bayes that I am confused
about.
The score is 0.996 before bayes.
Bayes score is 1.11022302462516e-16
What causes this type of score?
Is this why this message ends up getting -1.603 after the bayes tests?
----------------------------------------------------------------------
[26760] dbg: rules: running body-text per-line regexp tests; score so
far=0.996
[26760] dbg: rules: ran body rule __NONEMPTY_BODY ======> got hit: "D"
[26760] dbg: uri: running uri tests; score so far=0.996
[26760] dbg: rules: ran uri rule __SARE_URI_ANY ======> got hit: "m"
[26760] dbg: rules: ran uri rule __URI_ABOUT_COM ======> got hit:
"about.com"
[26760] dbg: bayes: DB journal sync: last sync: 0
[26760] dbg: bayes: corpus size: nspam = 487131, nham = 560694
[26760] dbg: bayes: score = 1.11022302462516e-16
[26760] dbg: bayes: DB journal sync: last sync: 0
[26760] dbg: bayes: untie-ing
[26760] dbg: bayes: untie-ing db_toks
[26760] dbg: bayes: untie-ing db_seen
[26760] dbg: plugin: registering glue method for check_uridnsbl
(Mail::SpamAssassin::Plugin::URIDNSBL=HASH(0x1c5a020))
[26760] dbg: rules: ran eval rule __SARE_BODY_BLANKS_5_100 ======> got hit
[26760] dbg: rules: ran eval rule BAYES_00 ======> got hit
[26760] dbg: rules: ran eval rule __SARE_BODY_BLNK_5_100 ======> got hit
[26760] dbg: rules: running raw-body-text per-line regexp tests; score so
far=-1.603
[26760] dbg: rules: ran rawbody rule __VIRUS_WARNING268F ======> got hit:
"From: "Daryl Milton" <[EMAIL PROTECTED]>"
[26760] dbg: rules: running full-text regexp tests; score so far=-1.603
-----------------------------------------------------------------------
Thanks, Curt
