Parse bits easily. Parsing can be done either in a monadic style, or more efficiently, using the Applicative style.

For the monadic style, write your parser as a BitGet monad using the

• getBool
• getWord8
• getWord16be
• getWord32be
• getWord64be
• getByteString

functions and run it with runBitGet.

For the applicative style, compose the fuctions

• bool
• word8
• word16be
• word32be
• word64be
• byteString

to make a Block. Use block to turn it into the BitGet monad to be able to run it with runBitGet.

Parse bits using a monad.

myBitParser :: Get (Word8, Word8)
myBitParser = runGetBit parse4by4

parse4by4 :: BitGet (Word8, Word8)
parse4by4 = do
   bits <- getWord8 4
   more <- getWord8 4
   return (bits,more)
 

Parse more efficiently in blocks. Each block is read with only one boundry check (checking that there is enough input) as the size of the block can be calculated statically. This is somewhat limiting as you cannot make the parsing depend on the input being parsed.

data IPV6Header = IPV6Header {
     ipv6Version :: Word8
   , ipv6TrafficClass :: Word8
   , ipv6FlowLabel :: 'Word32
   , ipv6PayloadLength :: Word16
   , ipv6NextHeader :: Word8
   , ipv6HopLimit :: Word8
   , ipv6SourceAddress :: ByteString
   , ipv6DestinationAddress :: ByteString
 }

 ipv6headerblock =
         IPV6Header <$> word8 4
                    <*> word8 8
                    <*> word32be 24
                    <*> word16be 16
                    <*> word8 8
                    <*> word8 8
                    <*> byteString 16
                    <*> byteString 16

ipv6Header :: Get IPV6Header
ipv6Header = runBitGet (block ipv6headerblock)