Re[2]: [Haskell-cafe] Re: Optimizing locking with MVars