On Fri, 1 Aug 2003, Simon Marlow wrote:
class InputStream s where streamGet :: s -> IO Word8 streamReadBuffer :: s -> Integer -> Buffer -> IO () streamGetBuffer :: s -> Integer -> IO ImmutableBuffer streamGetContents :: s -> IO [Word8]
class OutputStream s where streamPut :: s -> Word8 -> IO () streamPuts :: s -> [Word8] -> IO () streamWriteBuffer :: s -> Integer -> Buffer -> IO ()
Now that I think about it, we should parameterize these classes over the input/output data type also. Then we could have filterInputStream :: (Storable a, Storable b, InputStream s a, InputStream t b) => BlockRecoder a b -> s -> t filterOutputStream :: (Storable a, Storable b, OutputStream s a, OutputStream t b) => BlockRecoder b a -> s -> t Text streams would fit into this model. Plus, there are many useful stream filters which don't go octet -> char or char -> octet, such as: packBitsLE, packBitsBE :: BlockRecoder Bool Word8 unpackBitsLE, unpackBitsBE :: BlockRecoder Word8 Bool zlibPack, zlibUnpack :: BlockRecoder Word8 Word8 recodeNewlines :: NewlineFormat -> BlockRecoder Char Char serialize :: (Serializable a) => BlockRecoder a Word8 unserialize :: (Unserializable a) => BlockRecoder Word8 a -- Ben