Re: [Haskell-cafe] Another question about unsafePerformIO