- Haskell-Cafe - Haskell.org

GHC doesn't work
by Andrew Coppin 05 Oct '07

05 Oct '07

Greetings. I have a PC that had GHC 6.6 running on it. Worked fine. Then I uninstalled 6.6 and installed 6.6.1, and now it doesn't appear to work at all. Any attempt to run GHC results in a message that says The entry point OpenThread could not be found in KERNEL32.dll. or something very similar to that. Is this likely to be because GHC 6.6.1 doesn't support Windows NT, or because my PC is messed up?

2 2

Function composition
by Tiago Miguel Laureano Alves 05 Oct '07

05 Oct '07

Hi, I'm playing a little bit with pointfree and function composition and I would like to ask you if the following is theoretical correct and how can I express it in haskell. Imagine that I have the following functions f :: a -> b -> c -> d g :: d -> e I want to compose these two functions such that: (g . f) :: a -> b -> c -> e in which a, b and c are arguments of the composition. Is this theoretically correct? I tried this in haskell and ghc gives me an error saying that the function g should be of type b -> c -> d. With this I assume that when I specify f :: a -> b -> c -> d I'm really specifying f :: a -> (b -> c -> d) and such the composition doesn't work, but is it possible to overcome this? Kind regards, Tiago

5 5

bizarre memory usage with data.binary
by Anatoly Yakovenko 04 Oct '07

04 Oct '07

i am getting some weird memory usage out of this program: module Main where import Data.Binary import Data.List(foldl') main = do let sum' = foldl' (+) 0 let list::[Int] = decode $ encode $ ([1..] :: [Int]) print $ sum' list print "done" it goes up to 500M and down to 17M on windows. Its build with ghc 6.6.1 with the latest data.binary Any ideas what could be causing the memory usage to jump around so much? Thanks, Anatoly

10 17

Why not assign a type to unsafePerformIO?
by Justin Bailey 04 Oct '07

04 Oct '07

One of the "holes" in real-world Haskell is you never know if a library/function is calling unsafePerformIO and you have to trust the library author. I recognize the necessity of the function, but should it announce itself? unsafePerformIO has this type: unsafePerformIO :: IO a -> a Would there be any value to making it have a type that can be stripped off, like some other monads? For example, providing a "runUnsafe" or similar: data UnsafePerformIO a = Unsafe a runUnsafe :: UnsafePerformIO a -> a runUnsafe (Unsafe o) = o and changing unsafePerformIO to have the type: unsafePerformIO :: IO a -> UnsafePerformIO a It seems it would be valuable to have functions announce when they use unsafePerformIO, but additionally allow it to be stripped off. So the classic launchMissiles :: a -- Uses unsafePerfomIO! Would become launchMissiles :: UnsafePerformIO a Which could be stripped off it you wanted: evilDictatator :: a evilDictator = runUnsafe $ launchMissiles But doesn't have to be: incompetentDictator :: a incompetentDictator = launchMissiles -- Doesn't type check! I doubt this is original - does it buy anything? Justin

3 2

Re: [Haskell-cafe] Why not assign a type to unsafePerformIO?
by Justin Bailey 04 Oct '07

04 Oct '07

On 10/3/07, Victor Nazarov <asviraspossible(a)gmail.com> wrote: > > But how would you know that evil dictator uses unsafePerformIO??? You don't. unsafePerformIO can't be taken it away (there are legitimate reasons to strip IO), which is why I wonder if it's useful at all. p.s. CC'ed to haskell-cafe

2 1

Curry and uncurry
by PR Stanley 04 Oct '07

04 Oct '07

Hi The following is from the Hutton book: Without looking at the standard prelude, define the higher-order library function curry that converts a function on pairs into a curried function, and conversely, the function uncurry that converts a curried function with two arguments into a function on pairs. Hint: first write down the types of the two functions. I didn't even know about the curry and uncurry functions. I'm not looking for the answer but some guidance would be much appreciated. Thanks, Paul

4 3

PROPOSAL: New efficient Unicode string library.
by Johan Tibell 03 Oct '07

03 Oct '07

Dear haskell-cafe, I would like to propose a new, ByteString like, Unicode string library which can be used where both efficiency (currently offered by ByteString) and i18n support (currently offered by vanilla Strings) are needed. I wrote a skeleton draft today but I'm a bit tired so I didn't get all the details. Nevertheless I think it fleshed out enough for some initial feedback. If I can get the important parts nailed down before Hackathon I could hack on it there. Apologies for not getting everything we discussed on #haskell down in the first draft. It'll get in there eventually. Bring out your Unicode kung-fu! http://haskell.org/haskellwiki/UnicodeByteString Cheers, Johan Tibell

27 60

"with" and "preserving" for local state
by Jules Bean 03 Oct '07

03 Oct '07

Lots of external libraries contain state, but one that really contains a *lot* of state is the OpenGL libraries, since OpenGL is specified as a statemachine. This means that when you're writing structured code you quite often want to save and restore chunks of state 'automatically'. For the very most common case (coordinate transformations) Sven gives us 'preservingMatrix' which is extremely handy. Unless I've missed something there's no similar API for saving/restoring arbitrary state variables. It's not hard to write: > {-# OPTIONS -fglasgow-exts #-} > import Graphics.Rendering.OpenGL > import Graphics.UI.GLUT > > preserving :: (HasSetter g, HasGetter g) => g a -> IO t -> IO t > preserving var act = do old <- get var > ret <- act > var $= old > return ret This enables us to write preserving lighting $ do ..... Note that, since IORef is an instance of HasGetter and HasSetter, you can do 'preserving' on any old IORef, not just an openGL StateVar. Also note that the 'makeStateVar' interface that Graphics.Rendering.OpenGL.GL.StateVar exports allows you to make a statevar out of any appropriate action pair (not entirely unrelated to http://twan.home.fmf.nl/blog/haskell/overloading-functional-references.deta…) Sometimes you don't only want to preserve a value, but set a specific temporary value, so: > with :: (HasSetter g, HasGetter g) => g a -> a -> IO t -> IO t > with var val act = do old <- get var > var $= val > ret <- act > var $= old > return ret with lighting Enabled $ do .... (of course, with could be written as with var val act = preserving var $ var $= val >> act ) But this gets really clumsy if you have multiple variables to save/restore, which is really what lead me to write this message in the first place. A cute syntax for doing multiple save/restores at once is given by an existential: > data TemporaryValue = forall a g. > (HasGetter g,HasSetter g) => > g a := a > > with' :: [TemporaryValue] -> IO t -> IO t > with' tvs act = do olds <- mapM (\(a := b) -> do old <- get a > return (a := old)) > tvs > ret <- act > mapM_ (\(a := b) -> a $= b) tvs > return ret so we can then write: with' [lighting := Enabled, currentColor := Color4 1 0 1 0] $ do ... and have a type safe list of temporary assignments passed as an argument. And, amazingly, you get decent error messages too: *Main> :t with' [lighting := Enabled, currentColor := Color4 1 0 1 0] with' [lighting := Enabled, currentColor := Color4 1 0 1 0] :: IO t -> IO t *Main> :t with' [lighting := Enabled, currentColor := "Foo"] <interactive>:1:44: Couldn't match expected type `Color4 GLfloat' against inferred type `[Char]' In the second argument of `(:=)', namely `"Foo"' In the expression: currentColor := "Foo" In the first argument of `with'', namely `[lighting := Enabled, currentColor := "Foo"]' Hope someone else finds that useful, Jules

1 0

Parsing R5RS Scheme with Parsec
by Pasqualino 'Titto' Assini 03 Oct '07

03 Oct '07

Hi Alex, I hope not to spoil your fun but have you had a look at this: Write Yourself a Scheme in 48 Hours http://halogen.note.amherst.edu/~jdtang/scheme_in_48/tutorial/overview.html Regards, titto

2 1

Re: [Haskell-cafe] Assignment, Substitution or what?
by PR Stanley 03 Oct '07

03 Oct '07

> > > f x = x + x > > > Is the "x" use to create a pattern in the definition and when f is > > > called it's replaced by a value? > > > >Those equation-like definitions are syntactic sugar for lambda > >abstractions. f could as well be defined as f = \x -> x + x. > >Please elaborate > > >First, the > >f x = > >part says that f is a function which takes a single parameter, >called x. The other side of the = sign gives the function body: in >this case, x + x. This is exactly the same thing that is expressed >by the lambda expression > >\x -> x + x > >This expression defines a function that takes a single parameter >called x, and returns the value of x + x. The only difference is >that with the lambda expression, this function is not given a >name. But you can easily give the function a name (just as you can >give any Haskell expression a name) by writing > >f = \x -> x + x > >In general, writing > >g x y z = blah blah > >is just a shorthand for > >g = \x -> \y -> \z -> blah blah. > >That is, it simultaneously creates a function expression, and >assigns it a name. > >Does that help? Yes and thanks for the reply. When a function is declared in C the argument variable has an address somewhere in the memory: int f ( int x ) { return x * x; } any value passed to f() is assigned to x. x is the identifier for a real slot in the memory (the stack most likely) made available for f(). Is this also what happens in Haskell? Thanks, Paul

4 3