Re: [Haskell-cafe] I/O system brokenness with named pipes