> Would it be more consistent if the type error message for failing to > admit equality used the "[...]" convention of other type error > messages. Sure. I committed a fix to do that.