September 2012 - Haskell-Cafe

Re: [Haskell-cafe] type variable in class instance
by oleg＠okmij.org 12 Sep '12

12 Sep '12

Corentin Dupon wrote about essentially the read-show problem: > class (Typeable e) => Event e > > data Player = Player Int deriving (Typeable) > data Message m = Message String deriving (Typeable) > > instance Event Player > > instance (Typeable m) => Event (Message m) > > viewEvent :: (Event e) => e -> IO () > viewEvent event = do > case cast event of > Just (Player a) -> putStrLn $ show a > Nothing -> return () > case cast event of > Just (Message s) -> putStrLn $ show s > Nothing -> return () Indeed the overloaded function cast needs to know the target type -- the type to cast to. In case of Player, the pattern (Player a) uniquely determines the type of the desired value: Player. This is not so for Message: the pattern (Message s) may correspond to the type Message (), Message Int, etc. To avoid the problem, just specify the desired type explicitly > case cast event of > Just (Message s::Message ()) -> putStrLn $ show s > Nothing -> return () (ScopedTypeVariables extension is needed). The exact type of the message doesn't matter, so I chose Message (). BTW, if the set of events is closed, GADTs seem a better fit > data Player > data Message s > > data Event e where > Player :: Int -> Event Player > Message :: String -> Event (Message s) > > viewEvent :: Event e -> IO () > viewEvent (Player a) = putStrLn $ show a > viewEvent (Message s) = putStrLn $ show s

5 12

IO vs MonadIO
by Sergey Mironov 12 Sep '12

12 Sep '12

Hi. Just a brief question. System.IO functions are defined in IO monad and have signatures like Foo -> IO Bar. Would it be better to have all of them defined as (MonadIO m) => Foo -> m Bar? What are the problems that would arise? Sergey

2 2

Pattern matching: multi constructors to same branch
by Thiago Negri 11 Sep '12

11 Sep '12

Is it possible to capture more than one constructor in a single pattern matching? I mean, is it possible to generalize the following pattern matching of A and B to a single branch? g f C = [f C] g f v@(A _ n) = f v : g n g f v@(B _ n) = f v : g n For example: g f C = [f C] g f v@(A|B _ n) = f v : g n Or: g f v = case v of C -> [f C] A|B -> f v : g (n v) Thanks.

2 1

Either Monad and Laziness
by Eric Velten de Melo 11 Sep '12

11 Sep '12

Hello, I am currently trying to rewrite the Graphics.Pgm library from hackage to parse the PGM to a lazy array. The current implementation parses it straight to UArray, which is strict. The behaviour I want to achieve is like this: I want the program when compiled to read from a file, parsing the PGM and at the same time apply transformations to the entries as they are read and write them back to another PGM file. The problem is that even using a lazy array structure, because the parser returns an Either structure it is only possible to know if the parser was successful or not after the whole file is read, therefore requiring it to read the entire file before applying the transformations, ruining the property of laziness. Is there some way to keep this from happening? Should I even want to make it like this? Not really a real life situation, but imagine I want to read a reaaaaally large PGM file which does not fit into RAM memory and I don't want to be forced to have the whole array in the memory at once. One alternative I thought was parsing only the PGM header and then read the rest of the input without using Parsec and the Either Monad. In the event the data is corrupted, though, I would not know how to recover from it. Any thoughts? Hopefully I'm not saying anything really stupid. Eric

3 2

Destructive updates to plain ADTs
by Milan Straka 11 Sep '12

11 Sep '12

Hi all, is there any way to perform a destructive update on a plain ADT? Imagine I have a simple data Tree a = Nil | Node a (Tree a) (Tree a) I would like to be able to modify right subtree of an existing tree. I can do that for example when using IORefs by changing the datatype to data Tree a = Nil | Node a (IORef (Tree a)) (IORef (Tree a)) and use unsafePerformIO + writeIORef. But the IORefs cause additional complexity when working with the data type. At the moment I am interested in any GHC solution, be it non-portable or version specific. I would like just to run some benchmarks and see the results. Cheers, Milan

5 7

ANNOUNCE: hsqml-0.1.1 (and hsqml-morris-0.1.0)
by Robin KAY 11 Sep '12

11 Sep '12

Dear All, I would like to announce a new package called 'hsqml' [1]. HsQML provides a Haskell binding to the Qt Quick framework. It allows you to write graphical applications where the front-end is written in Qt Quick's QML language (incorporating JavaScript) and the back-end is written in Haskell. The two layers are coupled together via a facility to define custom JavaScript objects through which QML code can call into Haskell. HsQML requires an installation of Qt 4.7 (including QtDeclarative) or later on your path. It has been tested with GHC 7.4 on both Linux and Windows. There is also a simple example program available using HsQML on Hackage called 'hsqml-morris'. It implements a human v computer game of Nine Men's Morris with game logic and AI written in Haskell and the presentation layer in QML. The library is currently limited by a lack of support for Qt signals, so calls can only be made from JavaScript to Haskell and not the other way. I plan to add support for these in a future release. Migration over to Qt 5.x is also likely once it's released. [1] http://hackage.haskell.org/package/hsqml-0.1.1 [2] http://hackage.haskell.org/package/hsqml-morris-0.1.0 Regards, -- Robin KAY

1 0

The Architecture of the Mighttpd High-Speed Web Server
by Kazu Yamamoto 10 Sep '12

10 Sep '12

Hello from Copenhagen, In last may, I wrote an article about Mighttpd in Japanese because a PR staff of IIJ asked me to do so. Since this article got popular in Japan, the PR staff decided to translate it into English. The English version is now open to the public: http://www.iij.ad.jp/en/company/development/tech/mighttpd/ # Mighttpd is based on Yesod's Warp but I did touch it to make the # article simple. I guess that we can use this article to explain why Haskell is suitable for network programming to other programming language communities. FYI, the original Japanese article can be found: http://www.iij.ad.jp/company/development/tech/activities/mighttpd/ Regards, --Kazu

1 0

type variable in class instance
by Corentin Dupont 10 Sep '12

10 Sep '12

Hello everybody! I'm soliciting once again your help! It's been several days I'm blocked by this problem: *{-# LANGUAGE DeriveDataTypeable #-} import Data.Typeable class (Typeable e) => Event e data Player = Player Int deriving (Typeable) data Message m = Message String deriving (Typeable) instance Event Player instance (Typeable m) => Event (Message m) viewEvent :: (Event e) => e -> IO () viewEvent event = do case cast event of Just (Player a) -> putStrLn $ show a Nothing -> return () case cast event of Just (Message s) -> putStrLn $ show s Nothing -> return () * gives me a: * Ambiguous type variable `t0' in the constraint: (Typeable t0) arising from a use of `cast' Probable fix: add a type signature that fixes these type variable(s) In the expression: cast event In the expression: case cast event of { Just (Message s) -> putStrLn $ show s Nothing -> return () }* This is because *Message* has a type variable, while *Player* has not... How to get this to work? I tried everything, existential types, scoped type variables etc. without success... Thanks!! Corentin

3 4

extensible-exceptions no longer a part of GHC 7.6.1?
by Peter Simons 10 Sep '12

10 Sep '12

Hi, 'extensible-exceptions' used to be a part of GHC, but it appears that the package has been dropped from 7.6.1. Yet, the release notes on haskell.org don't say anything about this subject (other than TODO). Was that change intentional? Take care, Peter

2 1

ANNOUNCE: grid-1.1
by Amy de Buitléir 10 Sep '12

10 Sep '12

I'm happy to announce a new package called grid: http://hackage.haskell.org/package/grid https://github.com/mhwombat/grid/wiki (wiki) Grid provides tools for working with regular arrangements of tiles, such as might be used in a board game or self-organising map (SOM). Grid currently supports triangular, square, and hexagonal tiles, with various 2D and toroidal layouts. If you need a tile shape or layout that isn't currently provided, please let me know. See Math.Geometry.Grid for an example of how to use the package. Suggestions for improvement are welcome.

6 8