- Haskell-Cafe - Haskell.org

The "Exp -> Term a" problem (again), how to dynamically create (polymorphic) typed terms in Haskell ??
by Pasqualino 'Titto' Assini 05 Oct '07

05 Oct '07

Hi, I am trying to write an interpreter for a little functional language but I am finding very problematic to dynamically create a typed representations of the language terms. I have googled around and found a few solutions but none seem to solve the problem. This is the example code: > {-# OPTIONS -fglasgow-exts #-} > module Eval where These are my untyped terms (what I get from my parser): > data Exp = EDouble Double | EString String | EPrim String | EApp Exp Exp deriving (Show) And these are the typed terms: > data Term a where > Num :: Double -> Term Double > Str :: String -> Term String > App :: Term (a->b) -> Term a -> Term b > Fun :: (a->b) -> Term (a->b) The problem is to write a function that converts between Exp and Term t as in: > test :: Term Double > test = typecheck $ EApp (EPrim "inc") (EDouble 10.0) So this is the conversion function: > class TypeCheck t where > typecheck :: Exp -> Term t A few primitives: > instance TypeCheck (String->String) where > typecheck (EPrim "rev") = Fun reverse > typecheck (EPrim "show") = Fun show > instance TypeCheck (Double->Double) where > typecheck (EPrim "inc") = Fun ((+1) :: Double -> Double) > instance TypeCheck (Double->String) where > typecheck (EPrim "show") = Fun show > instance TypeCheck Double where > typecheck (EDouble x) = Num x > instance TypeCheck String where > typecheck (EString x) = Str x The problem arises in the conversion of the function application (EApp). It does not seem to be possible to define "typecheck" on EApp in a generic way and is also not possible to distinguish between the different cases: > typecheck (EApp f a) = App (typecheck f :: Term (String->String)) (typecheck a:: Term String) The following pattern overlaps the previous one: > typecheck (EApp f a) = App (typecheck f :: Term (Double->String)) (typecheck a:: Term Double) To avoid this problem I could split my untyped terms in different data types as in: data EDouble = EDouble Double data App a b c = App a b c ... and define TypeCheck separetely on every data type. However, in that case what would be the type of my parser?? parser :: String -> ?? Any suggestion woud be very welcome indeed, titto

3 4

Re: [Haskell] reading from stdin
by Stefan O'Rear 05 Oct '07

05 Oct '07

On Thu, Oct 04, 2007 at 05:46:09PM +0100, Axel Simon wrote: > Hi, > > I'm trying to continuously output data to a file handle while reading > single characters from the user to adjust the speed at which things are > output. I'm interested to get this to work in Hugs on Windows. > I successfully used the following function ghci under Mac OS: > > {{{ > getUserInput :: IO (Maybe Char) > getUserInput = do > hasInput <- hReady stdin > if hasInput then liftM Just (hGetChar stdin) else return Nothing > }}} > > This function returns a character to me if there's one available. > If anybody could give me a hint how to get this working in Hugs under > Windows, please tell me. Something in the 'win32' package, or maybe the FFI. MSDN will be very handy. > P.S.: Sorry to post to haskell@..., but nothing else seemed to match. This is a thread for haskell-cafe@. Stefan

1 0

gtk2hs in Ubuntu Gutsy
by Chad Scherrer 05 Oct '07

05 Oct '07

I just installed the beta release for Ubuntu Gutsy, and I noticed that gtk2hs (provided by libghc6-gtk-dev) is still at version 0.9.10.5-1ubuntu1. Worse, it's apparently not installable; when I try I get this message: libghc6-gtk-dev: Depends: ghc6 (<6.6+) but 6.6.1-2ubuntu2 is to be installed Depends: libghc6-cairo-dev but it is not going to be installed Do any of you guys know why this wouldn't include a newer version? Probably just an oversight, I'm guessing. -- Chad Scherrer "Time flies like an arrow; fruit flies like a banana" -- Groucho Marx

2 1

GHC doesn't work
by Andrew Coppin 04 Oct '07

04 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 04 Oct '07

04 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 03 Oct '07

03 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 03 Oct '07

03 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