Well, here's an attempt at a start on a similar mechanism for Haskell: ---------- (start Packet.hs) module Packet where

import Data.Bits import Data.Word concatBits :: (Integral a, Bits a, Bits b) => [a] -> b concatBits [] = 0 concatBits (x:xs) = shift (fromIntegral x) (sum (map bitSize xs)) + concatBits xs class Packet a where readPacket :: [Word8] -> (a, [Word8]) instance Packet Word8 where readPacket (x:xs) = (x,xs) instance Packet Word16 where readPacket xs = let (ys, zs) = splitAt 2 xs in (concatBits ys, zs) instance Packet Word32 where readPacket xs = let (ys, zs) = splitAt 4 xs in (concatBits ys, zs) instance Packet Word64 where readPacket xs = let (ys, zs) = splitAt 8 xs in (concatBits ys, zs) instance (Packet a, Packet b) => Packet (a,b) where readPacket xs = let (u, xs') = readPacket xs (v, xs'') = readPacket xs' in ((u,v), xs'') instance (Packet a, Packet b, Packet c) => Packet (a,b,c) where readPacket xs = let (u, xs') = readPacket xs (v, xs'') = readPacket xs' (w, xs''') = readPacket xs'' in ((u,v,w), xs''') instance (Packet a) => Packet [a] where readPacket [] = ([],[]) readPacket xs = let (u, xs') = readPacket xs in (u : fst (readPacket xs'), []) -------- (end Packet.hs) With this you can convert lists of Word8's into particular structured forms as you see fit. Additional instances of Packet can be added for other types as needed. (As an easy example, if you have a GID newtype based on Word32, you could just add Packet to the deriving clause, assuming GHC extensions.) For example: readPacket [24,182,64,43,53,10,1] :: ((Word8,Word32,Word16), [Word8]) == ((24,3057658677,2561),[]) readPacket [24,182,64,43,53,10,1,24,197,17,34,200,10,2] :: ((Word8,Word32,Word16), [Word8]) == ((24,3057658677,2561),[24,197,17,34,200,10,2]) readPacket [24,182,64,43,53,10,1,24,197,17,34,200,10,2] :: ([(Word8,Word32,Word16)], [Word8]) -- note the list type == ([(24,3057658677,2561),(24,3306234568,2562)],[]) Anyway, I hope this is useful :) - Cale On 28/08/05, Joel Reymont <[EMAIL PROTECTED]> wrote: > Alistair, > > Thanks alot for your examples. I still have one unanswered question... > > How would you read a tuple of values (24, GID, Seq) like in my Erlang > example, where 24 is one byte, GID is a 4-byte integer and Seq is a 2- > byte word? Is there an elegant way of specifying packet format and > reading/writing Haskell data according to it? > > Thanks, Joel > > On Aug 28, 2005, at 11:58 PM, Alistair Bayley wrote: > > > Below is a contrived, non-optimal first attempt. The server just reads > > seven bytes from the socket, prints them, and quits. And the client > > just sends seven bytes and quits. > > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe@haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe > _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe