We are already very successfully using SA, but there a few things which I am still in the dark about.
I have had a look at the FAQ, but have not been able to find an answer to the following:
I read references to scoresets, for example:
"Also note that auto-training occurs using scores from either scoreset 0 or 1, depending on what scoreset is used during message check."
What are these scoreset and how do you know which one you are using?
There are 4 scoresets, used in the following situations:
0 - used when both bayes and network checks are disabled
1 - used when bayes is disabled, but network checks are enabled
2 - used when there bayes is enabled, but network checks are disabled
3 - used when both bayes and network checks are enabled.Bayes autolearning always uses 0 or 1 to calculate a score, because autolearning is done based on the score the message would have if bayes was disabled. (to prevent bayes self-feedback)
If you need to know which one you are using for sure, you can always run a message through spamassassin -D and look at the debug output. It will mention which scoreset it is using.
The second thing is, I read on the mailing list people referring to, for example, bigevil.cf - what is this - I see a bunch of .cf files in /usr/share/spamassassin, but nothing like the total number of rules and no bigevil. What am I missing here?
bigevil, antidrug, etc are add-on rulesets written by users, and aren't part of the SA distro. Some are destined to become a part of the standard distro, but aren't included yet, others are too specialized or require frequent updates and can't be made a part of the standard distro.
In general these rulesets are handy, but aren't as rigorously tested as the rules in SA itself.
A list of the add-on rulesets is maintained on the wiki: http://wiki.spamassassin.org/w/CustomRulesets
There's also some sample test data from mass-check, so you can get some feel for how prone to false positives the ruleset is.
