MailScanner and SA auto-learning

[Matt Kettler]

Kai Schaetzl wrote:
> Is it possible that MailScanner interferes in any way with Bayes 
> auto-learning? I have "bayes_auto_learn_threshold_spam 8" on my new 
> machine and nothing gets learned, not even spam over 20. It's possible 
> that it's not learned because partial scores for header or body or so 
> didnät reach the required minimum, of course. But I first wanted to make 
> sure that MailScanner doesn't tell SA this value when scanning. There's 
> nothing in the MailScanner.conf that looks like that, so I think 
> MailScanner settings don't matter here at all, correct?

Bayes autolearning works fine in my system, and MailScanner even reports it in
the spamcheck headers:

X-EVI-MailScanner-SpamCheck: spam, SpamAssassin (score=44.262, required 5,
	autolearn=spam, BAYES_99 3.50, DATE_IN_FUTURE_12_24 2.77,
	<snip, massive list of hits>

MailScanner 4.50.15, SpamAssassin 3.1.0.