[Mimedefang] SA/MIMEDefang performance tweaking

[Kevin A. McGrail]

I think there is a point to using Bayes in SpamAssassin without
SpamAssassin's built-in autolearn.  In MD, just copy emails that are above
the autolearn threshold to a maildrop and cron the learning process to
prevent lock issues with the Bayes database with an automatic purge of the
learnt emails.

Essentially the same thing with a minor time-shift.

Regards,
KAM


> > Is there any way to continue to use Bayes WITHOUT autolearn
>
> There's no point in using Bayes without autolearn.  Just turn off Bayes.