Hold the Spam
Ugh.
I've had a flood of comment spam that the new MT 3.2 anti-spam features just yawned at and turned the other cheek.
These spam comments were rated with three criteria: IP known as a spam IP? Commenter URL exists on a previously published comment? Commenter email exists on a previously published comment?
Now, because a whole slew of comments got through and published, then each one automatically got a thumbs up on the last two criteria, which completely overrode the first criteria. Good thinking, there.
And it looks like the "trusted" and "banned" commenter feature is only applicable if I used Typekey.
I hope this spam thing doesn't keep me busy this weekend.