The bayesian filter in SA 2.60 is broken on Red Hat 9, or, more likely, Red Hat 9 is broken.
I am seeing the same symptoms as here: http://sourceforge.net/mailarchive/message.php?msg_id=6154723 sa-learn --dump Use of uninitialized value in numeric lt (<) at thinking my ~/.spamassassin/bayes* files were corrupted (I have a number of different SA versions running) I deleted them, and found that no matter what I did, sa-learn --spam/ham --mbox/dir will always give "Learned from 0 message" and not create any of the ~/.spamassassin/bayes* files. It also prints the dots faster than it should, from this I would conclude that somehow accessing DB_whatever throws an error. Running the exact same files ($HOME mounted) on a RH7.3 box makes it work fine. The installation process (on RH9) did not give any errors. Nothing else unusual about SA that I have noticed. No trouble with SA 2.60 on SuSE 8.2. Ah, sa-learn -D does wonders: debug: bayes: DB_File module not installed, cannot use Bayes But I don't seem manage to install DB_File. In any case, sa-learn should give an error if essential components are missing. Volker -- Volker Kuhlmann is possibly list0570 with the domain in header http://volker.dnsalias.net/ Please do not CC list postings to me. ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
