- Beginners - Haskell.org

(no subject)
by Dennis Raddle 03 Sep '12

03 Sep '12

I wanted to run the map function from Data.Map, let's call it M.map, but inside a monad transformer stack including the Error monad. M.map has this type: M.map :: (Ord k) => (a -> b) -> Map k a -> Map k b However, I want to use a mapping function that has type (Monad m) => a -> m b (i.e. errors could be thrown during the computation, a log could be written, etc) I wrote the following. Any comments on this way of doing things? mapMapM :: (Monad m, Ord k) => (a -> m b) -> Map k a -> m (Map k b) mapMapM g mapIn = do let h (k,a) = do b <- g a return (k,b) y <- mapM h (M.toAscList mapIn) return $ M.fromAscList y

2 1

literate programming
by Dennis Raddle 02 Sep '12

02 Sep '12

Does literate programming in haskell (.lhs) support the idea in Knuth's WEB of using aliases for blocks of code? In other words doing something like main = do <part 1> <part 2> <part 1> === putStrLn "foo" <part 2> === putStrLn "bar" (except I don't know what the actual syntax is) -Dennis

2 1

<-
by Patrick Redmond 02 Sep '12

02 Sep '12

I'm reading "Learn You a Haskell for Great Good!", chapter 9, "Input and Output" <http://learnyouahaskell.com/input-and-output>. IO actions are given liberal coverage throughout the chapter, however it is never mentioned whether the value-extractor syntax (<-) has a type or not. main = do x <- getLine putStrLn $ reverse x In this little program, getLine has type "IO String" and x has type "String". This implies to me that (<-) has type "IO a -> a". However, GHCI chokes on ":t (<-)" and Hoogle says it's just a syntactic element <http://www.haskell.org/haskellwiki/Keywords#.3C->. I guess I don't have a specific question, but I was kind of expecting it to be a function with a type because everything seems to be a function with a type in Haskell... Thanks for listening!

5 5

Warp and Yesod benchmark puzzle
by Lorenzo Bolla 01 Sep '12

01 Sep '12

Hi all, This is a question specific to the Yesod framework, but simple enough (I hope) to be considered a beginner question... I am puzzled by the performance of these two very simple web-servers, one written in Warp and another written in Yesod: === YESOD === {-# LANGUAGE TypeFamilies, QuasiQuotes, MultiParamTypeClasses, TemplateHaskell #-} import Yesod data HelloWorld = HelloWorld mkYesod "HelloWorld" [parseRoutes| / HomeR GET |] instance Yesod HelloWorld getHomeR :: Handler RepHtml getHomeR = defaultLayout [whamlet|$newline always Hello World! |] main :: IO () -- main = warpDebug 3000 HelloWorld main = warp 3000 HelloWorld === WARP === {-# LANGUAGE OverloadedStrings #-} import Network.Wai import Network.HTTP.Types import Network.Wai.Handler.Warp (run) import Data.ByteString.Lazy.Char8 () app :: Application app _ = return $ responseLBS status200 [("Content-Type", "text/html")] "Hello, Warp!" main :: IO () main = do putStrLn "http://localhost:8080/" run 8080 app === I've tested both using httperf: $> httperf --hog --client=0/1 --server=localhost --port=3000 --uri=/ --rate=1000 --send-buffer=4096 --recv-buffer=16384 --num-conns=100 --num-calls=100 --burst-length=20 and I got very different results: YESOD: Request rate: 4048.0 req/s (0.2 ms/req) WARP: Request rate: 33656.2 req/s (0.0 ms/req) Now, I understand that Yesod is expected to be slower than the "raw" Warp, but I wasn't expecting a 10x slowdown, especially for such a trivial Yesod app (no db, no auth, etc.). [ Compilation command was: ghc -Wall -O2 --make yesod.hs $ yesod version yesod-core version:1.1.0 ] What is going on? Thanks, L.

4 8

Haskell POP3S
by Franco 01 Sep '12

01 Sep '12

Hello everyone, I am trying to write a simple script in Haskell. Such script is meant to: - contact my email (pop3) - ask the server wheter there are new messages - report the result back to me This was trivial to write, using the "pop3-client" library, but I soon realized the whole thing is pretty unsecure. User and password are sent in plain text and I would like to encrypt those. So I asked Hayoo: it seems there is only one package which supports pop3s, curlhs; still, I didn't manage to get it. So I am asking you: - do you know wheter curlhs is fit for my goal? pop3-client had a nice 'getNumberOfMessages' function. I don't see anything similar in there - do you have examples of querying mail via curl? I found very few of those Thanks in advance. I should mention this thing was meant to be fed to xmobar, to check at intervals of, say, 15 minutes, wheter I got new mail without having a mail client always open -Franco -- Franco <franco00(a)gmx.com>

2 1

Re: [Haskell-beginners] joining lists sharing multiple type classes
by Christopher Howard 01 Sep '12

01 Sep '12

On 08/31/2012 12:16 PM, Sergey Mironov wrote: > Yes, you can't concat [Star] and [Asteroid] because they are of different type. > Lets assume that Animation is defined as follows > > class Animation a where > feed :: GraphicSystem -> a -> IO () -- feeds a to graphic system > > and we have > > instance Animation Star where ... > instance Animation Asteroid where ... > > than we can do > > game_cycle :: ([Star],[Asteroid]) -> GraphicSystem -> IO () > game_cycle world@(stars, asteroids) gs = do > mapM (feed gs) stars > mapM (feed gs) asteroids > return () > This would probably work, though it evades my principle inquiry, i.e., how to purposely downgrade multiple types which belong to the same type classes into a single type. > but not > > game_cycle :: ([Star],[Asteroid]) -> GraphicSystem -> IO () > game_cycle world@(stars, asteroids) gs = do > mapM (feed gs) (stars ++ asteroids) -- type mismatch > return () > > If you absolutly sure, that you really need a single list of all objects, > consider using single type for them! > > data WorldObject = Star ... | Asteroid ... > > Sergey > This approach is not modular... some of my types will be quite complex and I would rather have them as their own separate data types in their own module, rather than one monstrous type. Looking into this some more... I think what was actually looking for was existential quantification. That is, I could define a third type: code: -------- data Displayable = forall a. (Locatable a, Animation a) => Displayable a -------- Then I could map this constructor over the other two lists and concatenate them. This would allow me to use functions from both type classes in operations on the list members, provided that I extract the polymorphic component first. I haven't applied it to my actual code yet, but here is a sort of test case that compiles: code: -------- {-# LANGUAGE ExistentialQuantification #-} class CA a where f :: a -> Double class CB a where g :: a -> Integer data D1 = D1 Integer data D2 = D2 Integer instance CA D1 where f (D1 x) = fromInteger x + 2.0 instance CB D1 where g (D1 x) = x + 3 instance CA D2 where f (D2 x) = fromInteger x + 3.0 instance CB D2 where g (D2 x) = x + 1 data E = forall a. (CA a, CB a) => E a d1 = map E [D1 23] d2 = map E [D2 4] l = d1 ++ d2 r :: E -> Double r (E x) = fromInteger (g x) + f x result = map r l *Main> :load "/scratch/cmhoward/test0/plist.hs" [1 of 1] Compiling Main ( /scratch/cmhoward/test0/plist.hs, interpreted ) Ok, modules loaded: Main. *Main> result [51.0,12.0] -------- Though, I'm not sure if that is simpler than your first suggestion in this particular case. But it is certainly more interesting! -- frigidcode.com indicium.us

1 0

joining lists sharing multiple type classes
by Christopher Howard 01 Sep '12

01 Sep '12

Hi. I've got two data structures, Star and Asteroid (eventually I'll make more) that both belong to the following type classes: Locatable, and Animation. I wanted to do something like so in part of the code: code: -------- let stars = ... in -- of type [Star] let asteroids = ... in -- of type [Asteroid] let visibleObjects = do visibleObject <- (stars ++ asteroids) ... -- prep each object for graphics system -- using funcs from both type classes ... -- feed visibleObjects to graphics system -------- However, this does not work because the two lists are not automatically downgraded when joined together by (++). The compiler complains about "asteroids" not being of type [Star]. What is the simplest way to do what I am trying to do? I tried this, but I think my syntax is off: code: -------- let visibleObjects = do visibleObject <- ((stars :: [(Locatable a, Animation a) => a) ++ (asteroids :: [(Locatable a, Animation a) => a) ) -------- Compiler complains about "Illegal polymorphic or qualified type". -- frigidcode.com indicium.us

2 1

A simple function V2
by Ezequiel Hernan Di Giorgi 01 Sep '12

01 Sep '12

First i want to thank all the persons who responded me yesterday to help me. Thanks! I am so happy with with your friendliness. So i have other beginners question: Now i want a improved version of my* intercalate*. Now i want to call a function with two [t][t] and obtain another one which have only even elements, even because: - [1,2,3,3,4][6] and the ouput [1,6] - [1,2,3,4][5,6,7] output [1,5,2,6,3,7] I tried it: *intercalate :: (Eq t) => [t] -> [t] -> [t]* *intercalate (x:xs) (y:ys)* * | xt == [] = []* * | yt == [] = []* * | otherwise = x : y : intercalate xs ys* * where xt=(x:xs)* * yt=(y:ys)* but i get nice error **Main> intercalate [1][6]* *[1,6*** Exception: baby.hs:(2,1)-(5,51): Non-exhaustive patterns in function intercalate* * * **Main> * (yes...the file's name is baby.hs) Thanks in advance! (: (: (: (: (:

5 8

factorial question
by KMandPJLynch 01 Sep '12

01 Sep '12

Good afternoon, I'm going thru Graham Hutton's book "Programming in Haskell" [and am viewing the associated online lectures by Erik Meijer and Graham]. I find both to be excellent. My problem is with the following statements: factorial' :: Int -> Int factorial' 0 = 1 factorial' (n+1) = (n+1)*factorial' n When I load this into GHC I get the following error: pihch01.hs:128:13: Parse error in pattern: n + 1 I'd appreciate any advice. Good day

3 2

First version of GHC
by Dániel Arató 30 Aug '12

30 Aug '12

Hi all, I'm curious if the first implementation of GHC was in a different language before its current instance could be developed. If yes, what language was used? If not, what compiler was first used to compile GHC? Thank you, Daniel

2 2