November 2002 - Haskell-Cafe

ghc binary, fptools CVS jamming :-(
by Ahn Ki-yung 13 Nov '02

13 Nov '02

Do I have to compile every packages again if I upgrade ghc ? It is a pleasure to make use of the utilities (greencard, HOpenGL, etc). But what When I try to use CVS repository and stand-alone binary distribution together (because I wanted to see hdirect is like and v 0.16 does not work with ghc-5.xx) I felt very much discouraged. What should I do ? Do I always have to compile everything from the fptools source again ? I hope FFI related Haskell module gets stable and standardized soon. -- Ahn Ki-yung

1 0

Re: Record of STRefs better than STRef to a Record?
by Jorge Adriano 13 Nov '02

13 Nov '02

> > I'm passing that 'record' around as an implicit value. The record as > > STRefs that I use to collect info, but it also has some pure fields with > > 'read-only' info. Something like, <opss let me change 'max' to 'm'> > > data E s = E{ > > refi :: STRef s Int, > > refc :: STRef s Char, > > m :: Int > > } > > > > In some functions I might need only some pure fields, and none of the > > STRefs, but since I pass something of type 'E s' around, everything ends > > up beeing monadic. > > there is no reason why that should be the case. you only need to be > within the ST monad when you read the references, so if in some function > you only look at the pure values, it can have a pure type. > hope this helped > iavor That's what I thought, untill I tried it. Well, actually you must be right since the pure field defines a pure (projection) function... Hmmm, ok, can someone explain this to me, data E s = E{ refi :: STRef s Int, refc :: STRef s Char, m :: Int } -- this is fine, obviously... pure :: E s -> Int pure e = m e -- but this is not... pure2 :: (?e :: E s) => Int pure2 = m (?e) Why exactly isn't this allowed? What is the workaround? Error msg: ------------------------------------------------------------------------ Ambiguous constraint `?e :: E s' At least one of the forall'd type variables mentioned by the constraint must be reachable from the type after the '=>' In the type: forall s. (?e :: E s) => Int While checking the type signature for `pure2' Failed, modules loaded: none. -------------------------------------------------------------------------- Thanks, J.A.

1 0

Record of STRefs better than STRef to a Record?
by Jorge Adriano 13 Nov '02

13 Nov '02

Hi all, If I use an STRef to a record, will a new record be created each time I want to update a single field? Or can I expect GHC to optimize it and have the field of the record updated in place? Right now I'm using a record of STRefs, like: data E s = E{ refi :: STRef s Int, refc :: STRef s Char } but it can get a little messy, since thread s propagates to every datatype that uses the record type in it's definition. Thanks, J.A.

2 2

monadic stack to register machine translator
by William Lee Irwin III 13 Nov '02

13 Nov '02

module GT where import Monad import Monoid import MonadState import MonadWriter import MonadRWS -- Just a quick exercise in using monads. -- Thought it'd be nice to share with the class. data GOp = PushVal Integer | Push Integer | Pop Integer | Slide Integer | Update Integer | GAdd | GSub | GMul | GDiv | GMod | GPow | GNeg | GAbs deriving (Eq, Ord, Read, Show) type Tmp = Integer data ROp = LoadImm Tmp Integer | RAdd Tmp Tmp Tmp | RSub Tmp Tmp Tmp | RMul Tmp Tmp Tmp | RDiv Tmp Tmp Tmp | RMod Tmp Tmp Tmp | RPow Tmp Tmp Tmp | RNeg Tmp Tmp | RAbs Tmp Tmp deriving (Eq, Ord, Read, Show) type CounterT m t = StateT Integer m t type StackT t = State [Integer] t type GST t = RWS () [ROp] (Integer, [Integer]) t class Stack f where pushVal, push, pop, update, slide :: Integral t => t -> f () popVal :: Integral t => f t instance Integral t => Stack (RWS () [ROp] (t, [t])) where pushVal n = do (ctr, stk) <- get put (ctr, fromIntegral n : stk) popVal = do (ctr, top:stk) <- get put (ctr, stk) return (fromIntegral top) push n = do (ctr, stk) <- get put (ctr, stk!!fromIntegral n : stk) pop n = do (ctr, stk) <- get put (ctr, drop (fromIntegral n) stk) slide n = do (ctr, top:stk) <- get put (ctr, top : drop (fromIntegral n) stk) update n = do (ctr, top:stk) <- get let (front, _:back) = splitAt (fromIntegral n) stk put (ctr, front ++ [top] ++ back) class Counter f where gen :: Enum t => f t instance Integral t => Counter (RWS () [ROp] (t, [t])) where gen = do (ctr, stk) <- get put (ctr + 1, stk) return . toEnum . fromIntegral $ ctr + 1 instance (Enum t, Monad m) => Counter (StateT t m) where gen = do ctr <- get put $ succ ctr ctr <- get return . toEnum $ fromEnum ctr translate gOps = snd $ evalRWS (mapM trans gOps) () (0,[]) trans :: GOp -> GST () trans i = case i of PushVal n -> do reg <- gen tell [LoadImm reg n] pushVal reg Push n -> push n Pop n -> pop n Slide n -> slide n Update n -> update n GAdd -> doBinOp RAdd GSub -> doBinOp RSub GMul -> doBinOp RMul GDiv -> doBinOp RDiv GMod -> doBinOp RMod GPow -> doBinOp RPow GNeg -> doUnOp RNeg GAbs -> doUnOp RAbs where doUnOp op = do x <- popVal y <- gen tell [op y x] pushVal y doBinOp op = do x <- popVal y <- popVal z <- gen tell [op z x y] pushVal z

1 0

RE: Calling Haskell from Python / C++
by Simon Peyton-Jones 13 Nov '02

13 Nov '02

| So here is what I envision: I write the main | application in Python. I write a (hopefully) small | Haskell module that: | a) Calls back to the main Python app for reading the | text to be parsed, preferably using laziness. | b) Parses the text, and maybe processes a bit. | c) Returns the parsed data-structure. (This may be | tricky, but I think I know how to do it). If Python uses C's calling convention, it might be easy; just use 'foreign import' and 'foreign export' (see the FFI spec at haskell.org) If it doesn't, someone would have to add the Python calling convention to the various implementations. Simon

3 2

Re: 1 line simple cat in Haskell
by C.Reinke 13 Nov '02

13 Nov '02

> main = mapM (>>=putChar) getCharS where getCharS = getChar:getCharS > > How would you suggest to neatly insert the error handling code into ? \begin{code} -- some suggestions for a little zoo of cats module Main where import IO import Monad main0 = interact id main1 = getContents >>= putStr main2 = untilEOF (getChar>>=putChar) catchEOF io = catch io (\e->unless (IO.isEOFError e) (ioError e)) untilEOF io = catchEOF (sequence_ $ repeat io) main = main2 \end{code} Claus PS. I haven't kept up to date with buffering issues, and hugs/ghci may not like this kind of code..

1 0

Haskell98 Report copyright
by Christopher Milton 12 Nov '02

12 Nov '02

I hope we don't have a repeat of the MathWorld website shutdown.* I also can't find a webpage with the definition of Standard ML... only avaible in print from MIT Press? Chris * http://mathworld.wolfram.com/erics_commentary.html ===== Christopher Milton cmiltonperl(a)yahoo.com __________________________________________________ Do you Yahoo!? U2 on LAUNCH - Exclusive greatest hits videos http://launch.yahoo.com/u2

1 0

Re: showsPrec: cui bono?
by C.Reinke 12 Nov '02

12 Nov '02

> has anybody here used in a non-trivial way the showsPrec anti-parser? Isn't the idea to make things trivial while avoiding performance penalties? Perhaps: simple pretty-printing of abstract syntax trees? I often use it to get simple debugging output for complex internal data structures (first, use deriving; then, define showsPrec; if that's still not good enough, do some real thinking..). Anyway, this reminded me of a litte old hack of mine. Only trivial use of showsPrec, but perhaps you'll like it anyway?-) http://www.cs.ukc.ac.uk/people/staff/cr3/toolbox/haskell/R.hs As with anything else in my toolbox, no warranty for nothing.. Cheers, Claus ------------------- cut here {- Representative thingies.. A little hack to pair values with string representations of their expressions. Useful if you want to explain what map (+1) [1..4] or foldr1 (*) [1..5] do, or if you want to demonstrate the difference between foldr (+) 0 [1..4] and foldl (+) 0 [1..4] Load this module into Hugs (Hugs mode) and type in some of these examples to get an idea of what I mean. Also try map (+) [1..4] This could be extended in various directions, but I wanted to keep things simple. I'm not convinced that extra complications would be worth the effort. Claus Reinke -} default (R Integer) data R a = R {rep:: String ,val:: a } instance Show (R a) where showsPrec _ a = showString (rep a) instance Show (R a -> R b) where showsPrec _ f = showString ("\\x->"++(rep (f x))) where x = R{rep="x",val=error "variable"} instance Show (R a -> R b -> R c) where showsPrec _ f = showString ("\\x y->"++(rep (f x y))) where x = R{rep="x",val=error "variable"} y = R{rep="y",val=error "variable"} lift1 op a = R {rep="("++(rep op)++" "++(rep a)++")" ,val= ( (val op) (val a) ) } lift2 op a b = R {rep="("++(rep op)++" "++(rep a)++" "++(rep b)++")" ,val= ( (val op) (val a) (val b) ) } lift2infix op a b = R {rep="("++(rep a)++" "++(rep op)++" "++(rep b)++")" ,val= ( (val a) `iop` (val b) ) } where iop = val op instance (Num a,Show a) => Num (R a) where (+) = lift2infix R{rep="+",val=(+)} (-) = lift2infix R{rep="-",val=(-)} (*) = lift2infix R{rep="*",val=(*)} negate = lift1 R{rep="-",val=negate} fromInteger a = (\fIa->R{rep=show fIa,val=fIa}) (fromInteger a) instance (Eq a,Num a) => Eq (R a) where a == b = (val a) == (val b) instance (Ord a,Num a) => Ord (R a) where a <= b = (val a) <= (val b) instance (Enum a,Num a,Show a) => Enum (R a) where fromEnum = fromEnum.val toEnum a = R{rep=show a,val=toEnum a} enumFrom x = map toEnum [fromEnum x..] -- missing in Hugs Prelude..

1 0

Rational sequence
by Ferenc Wagner 12 Nov '02

12 Nov '02

With GHC-5.02.2, I do $ ghci Prelude> :m Ratio Ratio> [1%2..10%2] [1 % 2,3 % 2,5 % 2,7 % 2,9 % 2,11 % 2] The question is, why is there 11%2 at the end of the list? It's inconsistent with the (good) rules for Integer, since Ratio> [1,3..10] [1,3,5,7,9] Is this intentional? Feri.

5 8

H98 copyright notice
by Ross Paterson 08 Nov '02

08 Nov '02

The new copyright notice is inconsistent. If it is not otherwise changed, I suggest that the first clause, "The publisher intends this Report to belong to the entire Haskell community, ..." be deleted. Maybe the word "belong" could be replaced with "be available", but that doesn't seem completely accurate.

2 2