> If we end up with a list of ignored events, as I think we should, we
 > should consider making the same list control all of the relevant
 > features.

Not sure whether the attached is what you meant - it's just a first stab
in that direction.

martin