- Haskell-Cafe - Haskell.org

Re: number-parameterized types and heterogeneous lists
by oleg＠okmij.org 24 Jun '08

24 Jun '08

Luke Palmer wrote in response to Harald ROTTER > > I also wonder if there is some kind of "generalized" foldr such that, e.g. > > D1 $ D0 $ D0 $ Sz = specialFoldr ($) Sz [D1,D0,D0] > > I think that this foldr must be some "special" foldr that augments the data > > type of the result in each foldr step. > > Would this be possible or am I just chasing phantoms ? > Mostly I believe you are. What you are describing is firmly in the > realm of dependent types, far beyond Haskell's type system. See > Epigram or Agda for languages which have attempted to tackle this > problem. First of all, Haskell (as implemented by released versions of GHC for at least two years) already permits dependent types, in the strict sense of a type of an expression being determined by the value of one argument of that expression: http://okmij.org/ftp/Computation/tagless-typed.html#tc-GADT-tc Second, (values of) types like D0 and D1 can be collected in a heterogenous list (HList), which does have the corresponding fold operator, called HFoldr in http://darcs.haskell.org/HList/Data/HList/HListPrelude.hs

1 0

"C" buffer suggestions??
by Galchin, Vasili 24 Jun '08

24 Jun '08

Hello, I want to model an I/O read/write buffer in Haskell and marshall down into ANSI C. 1) ByteString? Or Ptr Word8? 2) What are example values of the types in 1)? Kind regards, Vasili

4 9

Re: [Haskell-cafe] number-parameterized types and heterogeneous lists
by Harald ROTTER 23 Jun '08

23 Jun '08

Hello, sorry for the late answer, I was off for the weekend :-) The paper "Number-parameterized types" by Oleg Kielyov is located at http://okmij.org/ftp/papers/number-parameterized-types.pdf It impressively shows what one can do with Haskell's type system. What I am after is the replacement of a "chain" of digits, like e.g. D1 $ D0 $ D0 $ Sz by a list [D1,D0,D0] so I can effectively use list operations on those "typed numbers". I also wonder if there is some kind of "generalized" foldr such that, e.g. D1 $ D0 $ D0 $ Sz = specialFoldr ($) Sz [D1,D0,D0] I think that this foldr must be some "special" foldr that augments the data type of the result in each foldr step. Would this be possible or am I just chasing phantoms ? Thanks Harald. " Ce courriel et les documents qui y sont attaches peuvent contenir des informations confidentielles. Si vous n'etes pas le destinataire escompte, merci d'en informer l'expediteur immediatement et de detruire ce courriel ainsi que tous les documents attaches de votre systeme informatique. Toute divulgation, distribution ou copie du present courriel et des documents attaches sans autorisation prealable de son emetteur est interdite." " This e-mail and any attached documents may contain confidential or proprietary information. If you are not the intended recipient, please advise the sender immediately and delete this e-mail and all attached documents from your computer system. Any unauthorised disclosure, distribution or copying hereof is prohibited."

3 2

IMAP and NNTP libraries
by Maurício 23 Jun '08

23 Jun '08

Hi, Are there mature libraries for IMAP and NNTP available to Haskell? Thanks, Maurício

6 5

number-parameterized types and heterogeneous lists
by Harald ROTTER 23 Jun '08

23 Jun '08

Dear Haskellers, after reading Oleg Kiselyov's paper on number-parameterized types I started to play around with the class Digits that encodes decimal numbers in types. The "typed number" 10 would e.g. be defined as D1 $ D0 $ Sz I wondered if it would be possible replace the expression above by a heterogeneous list like [D1,D0] so I tried to define data Digit = forall a b.(Digits a, Digits (b a)) => Digit (a -> b a) Loading this into ghci yields: :t Digit D0 <interactive>:1:0: Ambiguous type variable `a' in the constraint: `Digits a' arising from a use of `Digit' at <interactive>:1:0-7 Probable fix: add a type signature that fixes these type variable(s) Removing the type constraints in the definition of "Digit": data Digit = forall a b.Digit (a -> b a) makes it work like this: :t Digit D0 Digit D0 :: Digit :t [Digit D0, Digit D1] [Digit D0, Digit D1] :: [Digit] "Digit", however, is far too general (it also includes e.g. \x -> [x]), but I would like it to be restricted to the Digit class. Any help is appreciated. Thanks Harald. CODE: module Test where data D0 a = D0 a data D1 a = D1 a data D2 a = D2 a data D3 a = D3 a data D4 a = D4 a data D5 a = D5 a data D6 a = D6 a data D7 a = D7 a data D8 a = D8 a data D9 a = D9 a class Digits ds where d2num :: Num a => ds -> a -> a data Sz = Sz -- zero size instance Digits Sz where d2num _ acc = acc instance Digits ds => Digits (D0 ds) where d2num dds acc = d2num (t22 dds) (10*acc) instance Digits ds => Digits (D1 ds) where d2num dds acc = d2num (t22 dds) (10*acc+1) instance Digits ds => Digits (D2 ds) where d2num dds acc = d2num (t22 dds) (10*acc+2) instance Digits ds => Digits (D3 ds) where d2num dds acc = d2num (t22 dds) (10*acc+3) instance Digits ds => Digits (D4 ds) where d2num dds acc = d2num (t22 dds) (10*acc+4) instance Digits ds => Digits (D5 ds) where d2num dds acc = d2num (t22 dds) (10*acc+5) instance Digits ds => Digits (D6 ds) where d2num dds acc = d2num (t22 dds) (10*acc+6) instance Digits ds => Digits (D7 ds) where d2num dds acc = d2num (t22 dds) (10*acc+7) instance Digits ds => Digits (D8 ds) where d2num dds acc = d2num (t22 dds) (10*acc+8) instance Digits ds => Digits (D9 ds) where d2num dds acc = d2num (t22 dds) (10*acc+9) t22 :: f x -> x t22 = undefined --data Digit = forall a b.(Digits a, Digits (b a)) => Digit (a -> b a) data Digit = forall a b.Digit (a -> b a) ------------------------------------------------------------------------------------------------- " Ce courriel et les documents qui y sont attaches peuvent contenir des informations confidentielles. Si vous n'etes pas le destinataire escompte, merci d'en informer l'expediteur immediatement et de detruire ce courriel ainsi que tous les documents attaches de votre systeme informatique. Toute divulgation, distribution ou copie du present courriel et des documents attaches sans autorisation prealable de son emetteur est interdite." " This e-mail and any attached documents may contain confidential or proprietary information. If you are not the intended recipient, please advise the sender immediately and delete this e-mail and all attached documents from your computer system. Any unauthorised disclosure, distribution or copying hereof is prohibited."

3 3

Literal programming in Haskell with rst-literals
by Martin Blais 23 Jun '08

23 Jun '08

Hello Haskell community! I just did a marginally cool thing and I wanted to share it with you. "rst-literals" is a small program I wrote a while ago in order to write documents in reStructuredText format that would embed SQL code for data models in them, a form of literal programming for SQL if you will; I would describe my needs for the schema in prose, and reST literal-blocks were used to embed SQL code, blocks that look like this:: CLASS Employee ( firstname VARCHAR, lastname VARCHAR ) I wrote the script to be entirely generic: it parses the reST documents using the docutils code and outputs only the literal-blocks, with indentation removed; you can then run your favourite interpreter/compiler on the result (in that case, psql to initialize a database). Recently, while experimenting with Haskell, I started using both the literal (.lhs) and non-literal (.hs) styles of Haskell input, and I found the literal style a bit unpleasant to use, in particular, I don't like to have to prefix every line of code I write, despite the help that Emacs' haskell-mode provides. So I tried pulling a similar trick and embedding Haskell code in literal-blocks within reST documents, extracting that code using rst-literals, and it turns out that it works like a charm. Here is an example makefile for doing this:: .SUFFIXES: .rst .hs all: chap6 .rst.hs: rst-literals $< > $@ chap6: chap6.hs ghc --make chap6.hs An example reST document with some embedded Haskell code follows this email. Note that since rst-literals is using the docutils parser, you can make use of all of the recursive reST syntax, sections, bulleted items and much more. Only the literal-blocks are extracted, anywhere they appear. You can also easily process the reST source into HTML pages or LaTeX documents using the tools that come with docutils. You can find rst-literals here: http://furius.ca/pubcode/ Enjoy, -- Martin P.S. If there is a way to output cpp-like directives for GHC, like "#line <filename> <lineno>", it would be easy to modify rst-literals to generate those, so that compilation errors could refer to the source reST document instead of the extracted source. chap6.hs: ---------------------------------------------------------------------- =========================================== Exercises from Hutton book, Chapter 6 =========================================== .. contents:: .. 1 Introduction 2 Exercise 1 3 Exercise 2 4 Exercise 3 5 Exercise 4 6 Exercise 5 7 Exercise 6 Introduction ============ Bla bla bla blablablablablabla bla bla blabla. Bla bla bla blablablablablabla bla bla blabla. Bla bla bla blablablablablabla bla bla blabla. Bla bla bla blablablablablabla bla bla blabla. Bla bla bla blablablablablabla bla bla blabla. Bla bla bla blablablablablabla bla bla blabla. Exercise 1 ========== :: myexp :: Int -> Int -> Int myexp b 0 = 1 myexp b (n+1) = b * (myexp b n) Exercise 2 ========== (Exercise 2 consisted in derivations, so we mark the literal blocks as another type of block with "#!example", so that they don't get included in the output when only the "default" literal blocks get extracted. See rst-literals docstring for details.) Length:: #!example 1 + (length [2, 3]) 1 + 1 + (length [3]) 1 + 1 + (1) 3 Drop:: #!example drop 3 [1, 2, 3, 4, 5] [] ++ drop 3 [2, 3, 4, 5] [] ++ [] ++ drop 3 [3, 4, 5] [] ++ [] ++ [] ++ [4, 5] [4, 5] Init:: #!example init [1, 2, 3] [1] ++ init [2, 3] [1] ++ [2] ++ init [3] [1] ++ [2] ++ [] [1, 2] Exercise 3 ========== These are alternate versions of the example functions defined in the text:: and' :: [Bool] -> Bool and' [x] = x and' (x:xs) = x && and' xs concat' :: [[a]] -> [a] concat' [] = [] concat' (x:xs) = x ++ concat' xs replicate' :: Int -> a -> [a] replicate' 0 x = [] replicate' (n+1) x = (x : replicate' n x) select' :: [a] -> Int -> a select' (x:xs) 0 = x select' (x:xs) (n+1) = select' xs n elem' :: Eq a => a -> [a] -> Bool elem' _ [] = False elem' y (x:xs) | x == y = True | otherwise = elem' y xs Exercise 4 ========== The exercise asked to implement a function to merge two lists:: merge :: Ord a => [a] -> [a] -> [a] merge xs [] = xs merge [] xs = xs merge (x:xs) (y:ys) | x < y = (x : merge xs (y:ys)) | otherwise = (y : merge (x:xs) ys) Exercise 5 ========== :: msort :: Ord a => [a] -> [a] msort [] = [] msort [x] = [x] -- This is necessary to end the recursion. msort xs = merge (msort (fst hh)) (msort (snd hh)) where hh = halve xs halve :: [a] -> ([a], [a]) halve xs = (take n xs, drop n xs) where n = (length xs) `div` 2 Some main program:: main = (putStrLn . show) (halve [1..17]) Exercise 6 ========== (Too basic, I didn't bother.) ----------------------------------------------------------------------

3 2

expected vs inferred type confusion
by Stephen Howard 23 Jun '08

23 Jun '08

I am having issues with type signatures for the resourceHandler action below. I haven't figured out what I'm doing wrong to cause the wrong type to be returned. It looks like Haskell is inferring that I'm returning a function from this action, when from the type signature I've specified, that is not what I'm wanting. The gist of the code is based on the request_method of the passed in HttpRequest, choose an action to run and pass it the needed parameters. the error: renao.hs:6:4: Couldn't match `HttpResponse' against `IO HttpResponse' Expected type: IO HttpResponse Inferred type: IO (IO HttpResponse) In the expression: return $ (handler session request) In the result of a 'do' expression: return $ (handler session request) the code: import HttpMessage import HttpSession resourceHandler :: HttpSession -> HttpRequest -> IO HttpResponse resourceHandler session request = do return $ handler session request where handler = which_handler request which_handler r = case request_method r of Get -> get_resources Head -> last_modified Post -> update_resources Put -> create_resources Delete -> delete_resources get_resources :: HttpSession -> HttpRequest -> IO HttpResponse get_resources session request = undefined -- ... the same stubbed definition for the other handlers (last_modified, etc)

2 1

Planet haskell
by Jamie Brandon 23 Jun '08

23 Jun '08

I was hoping to have my summer of code blog added to planet haskell but planet(a)community.haskell.org no longer seems to exist. Hopefully the owner is subscribed to this list? Jamie

4 4

ANNOUNCE: Pipe 1.0
by Matti Niemenmaa 23 Jun '08

23 Jun '08

Greetings, I hereby announce the release of the Pipe library, a library for piping data through a pipeline of processes. A web page with (hopefully) all the necessary info and a simple example can be found at http://iki.fi/matti.niemenmaa/pipe/ The package is at Hackage: http://hackage.haskell.org/cgi-bin/hackage-scripts/package/Pipe Feel free to comment on any part of it, either here or straight to my e-mail address. I've tested it on Windows XP and Linux (Gentoo) and GHC 6.8.2 and 6.8.3. Since the code is rather low-level it'd be interesting to hear whether it works on other systems, say the BSDs or OS X. -- Matti Niemenmaa

4 4

Safe way to parse arguments?
by Xiao-Yong Jin 23 Jun '08

23 Jun '08

Hi, I'm wondering how usually you parse command line arguments list safely. If the given argument is wrong, the program can still print out some error information instead of giving something like Prelude.read: no parse Let's make the discussion concrete. Suppose we have the following code. -->8----------8<-- import System.Environment data Conf = Conf { number :: Int } parseArgs [] = error "need at least one argument" parseArgs (x:_) = Conf { number = read x } work (Conf { number = n }) = n * (n-1) `div` 2 main = print . work . parseArgs =<< getArgs -->8----------8<-- If the "read x" fails in line 8, the program dies with Prelude.read: no parse What is the standard way to make the error message useful? Xiao-Yong -- c/* __o/* <\ * (__ */\ <

6 6