I've upgraded a personal system from 2.43 to 2.50 in order to play with
the Bayesian stuff. The configuration is Redhat 7.2, Sendmail 8.12.6,
spamassassin_milter with a lot of my own hacks in. 2.43 and
predecessors have been running correctly for the past year.
I've encountered two issues:
1. Even though I have terse reporting turned on, which under 2.43 and
predecessors yielded headers like ``No, hits=-1.6 required=5.0
tests=AWL,CARRIAGE_RETURNS,SPAM_PHRASE_00_01,SUBJECT_MONTH,
SUBJECT_MONTH_2'', I am now getting ``7.40 hits, 5 required; * 2.0
-- BODY: Nigerian scam key phrase (million dollars) * 1.7 -- BODY:
Contains 'Dear (something)' * 0.6 -- BODY: illegal'' inserted as
well. /etc/mail/spamassassin/local.cf is as follows.
rewrite_subject 0
use_terse_report 1
defang_mime 0
auto_whitelist_path /var/spamd/auto-whitelist
auto_whitelist_file_mode 0666
bayes_path /var/spamd/bayes
bayes_file_mode 0777
clear-terse-report-template
terse-report _HITS_ hits, _REQD_ required;
terse-report _SUMMARY_
2. Although the Bayesian filtering, which I have trained on some large
corpora, works from the command line with ``spamassassin < xxx >
yyy'', and inserts BAYES_70 or whatever to prove its down its thing,
it doesn't appear to work from spamd.
Please cc: me on any replies.
ian
-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility
Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
