- Haskell-Cafe - Haskell.org

Deriving with generics without values
by JP Moresmau 12 Jul '13

12 Jul '13

Hello all, My problem is the following: I have my own data types, and I'd like to derive automatically instances of some type class from them. I've started looking at GHC.Generics, which offer tools to do exactly that. However, some functions of my typeclass do not take my data type as a parameter, but as a result. Basically: class MyClass where fromString :: String -> a data MyData=MkMyData { myField ::Int } deriving (Generic) and I want to automatically generate the instance instance MyClass MyData, using default methods, etc. The GHC Generic class does say that it uses a from function that convert from the datatype to its representation: from :: a -> Rep<http://www.haskell.org/ghc/docs/7.4.1/html/libraries/ghc-prim-0.2.0.0/GHC-G…> a xfrom :: a -> Rep<http://www.haskell.org/ghc/docs/7.4.1/html/libraries/ghc-prim-0.2.0.0/GHC-G…> a x But I don't have a "a" to start from! I see from the related papers that the automatically generated code from from actually does pattern matches on constructors, so I need a value, undefined won't work. However I see the GHC.Generics also provide :+: (Sums: encode choice between constructors). If I have to provide an value, then the choice between constructor has been done! The examples about generics on http://www.haskell.org/haskellwiki/GHC.Generics do provide an example of defining the instance for :+: but I don't understand how we can get there. If I have a class method that takes a value as a parameter, and I pass undefined to it, the code will crash, since it can't pattern match on undefined. Can somebody shed some light on this? Am I using the wrong tool for the job? How can I achieve what I want? I want the full type representation with sums, but without a value to start from. Thanks a million! JP -- JP Moresmau http://jpmoresmau.blogspot.com/

2 2

Comparing functions
by Vlatko Basic 12 Jul '13

12 Jul '13

Hello Cafe, I have data CmpFunction a = CF (a -> a -> Bool) that contains comparing functions, like ==, <, > ..., and I'm trying to declare the Show instance for it like this instance Show (CmpFunction a) where show (CF (==)) = "== " -- no good show f = case f of -- no good also CBF (==) -> "==" _ -> "Other" but compiler complains for both with This binding for `==' shadows the existing binding imported from `Prelude' at src/Main.hs:6:8-11 (and originally defined in `ghc-prim:GHC.Classes') Is it possible at all to compare two functions or how to solve this problem, to show some string for a specific function? br, vlatko

5 10

How to use cabal-dev ghci and multiple targets/build-depends from cabal file?
by Kirill Zaborsky 12 Jul '13

12 Jul '13

Currently I'm creating a small library and I wanted to create tests for it. So I have a library section in cabal file and also a test-suite section. Everything goes well but when I tried to load file with tests into ghci (actually it was a REPL in emacs) I received errors stating that GHCi can not find e.g. Test.Framework module. Is there any way to use multiple build-depends in GHCi? KInd regards, Kirill Zaborsky

1 2

Netwire bouncing ball
by Just 12 Jul '13

12 Jul '13

Hello, I'm trying to get a grasp of netwire by implementing a bouncing ball simulation and I'm failing. The ball starts from the ground with a given velocity and when hitting the ground the wire inhibits successfully. Now I'm kinda stuck. How can I make the ball bounce? Here is the code: {-# LANGUAGE Arrows #-} module Main where import Control.Wire import Prelude hiding ((.), id) import Control.Concurrent type Pos = Double type Vel = Double type ObjState = (Pos, Vel) testApp :: Pos -> Vel -> WireP () ObjState testApp p0 v0 = proc _ -> do v <- integral_ v0 -< -9.81 p <- integral1_ p0 -< v when (>= 0) -< p returnA -< (p, v) main :: IO () main = loop' (testApp 0 30) clockSession where loop' w' session' = do threadDelay 1000000 (mx, w, session) <- stepSessionP w' session' () case mx of Left ex -> putStrLn ("Inhibited: " ++ show ex) Right x -> putStrLn ("Produced: " ++ show x) loop' w session Thanks in advance!

4 5

There was someone here going to use REPA and a functional way to find eigenvalues :)
by KC 11 Jul '13

11 Jul '13

Have you succeeded? In general, is there a functional way to do matrix manipulations? Linear algebra? Linear programming? Integer Programming? Numerical Analysis? Casey -- -- Regards, KC

1 0

Re: [Haskell-cafe] Non-recursive let
by Doug McIlroy 11 Jul '13

11 Jul '13

By analogy to ML, which has "let" and "let rec", where the latter corresponds to Haskell's "let", one is led to "let nonrec". I would definitely not like "shadow", for it means that new variable does NOT cast a shadow on its definining expression. I fear also that "let nonrec" by any name would introduce another attractive nuisance, just as insidious as the one it is intended to correct. For example x = ... let nonrec { x = someFunction x xsq = x^2} in ... won't do what was probably intended. In my own code, this idiom is more likely than the one that sparked the discussion. Doug

2 1

Well-Typed are hiring: Haskell developer
by Duncan Coutts 11 Jul '13

11 Jul '13

Fellow Haskellers, We are looking to hire a Haskell expert to work with us at Well-Typed as a Haskell developer. This is an exciting opportunity for someone who is passionate about Haskell and who is keen to improve and promote Haskell in a professional context. The role is quite general and could cover any of the projects and activities that we are involved in as a company. The tasks may involve: * working on the Haskell compiler, libraries and tools; * Haskell application development; * working directly with clients to solve their problems; * teaching Haskell, and developing training materials. At the moment, we are particularly hoping to find someone with an interest in supporting the development and maintenance of GHC. Therefore, some knowledge or interest in compiler internals, operating systems, the foreign function interface (FFI), and/or deployment issues would be welcome. Well-Typed has a variety of clients. For some we do proprietary Haskell development and consulting. For others, much of the work involves open-source development and cooperating with the rest of the Haskell community: the commercial, open-source and academic users. Our ideal candidate has excellent knowledge of Haskell, whether from industry, academia, or personal interest. Familiarity with other languages, low-level programming, and good software engineering practices are also useful. Good organisation and ability to manage your own time, and reliably meet deadlines, is important. You should also have good communication skills. Being interested or having experience in teaching Haskell (or other technical topics) is a bonus. Experience of consulting, or running a business, is also a bonus. You are likely to have a bachelor's degree or higher in computer science or a related field, although this isn't a requirement. The offer is initially for a one-year full time contract. We are also happy to receive applications for part-time work. The annual salary is from GBP 34,800 or pro rata for part-time or flexible work. We also operate a bonus scheme. We offer flexible hours and work from home. Living in England is not required. We may be able to offer either employment or sub-contracting, depending on the jurisdiction in which you live. If you are interested, please apply via info(a)well-typed.com. Tell us why you are interested and why you would be a good fit for the job, and attach your CV. Please also indicate how soon you might be able to start. We are more than happy to answer informal enquiries. Contact Duncan Coutts, Ian Lynagh or Andres Löh for further information, either by email or IRC. To ensure we can properly consider your application, please get it to us by July 25th, 2013, though we may be able to consider applications received later. -- Duncan Coutts, Haskell Consultant Well-Typed LLP, http://www.well-typed.com/

1 0

Music / MIDI help
by Mark Lentczner 11 Jul '13

11 Jul '13

I'm a little lost in the bewildering array of music packages for Haskell, and need some help. I'm looking to recreate one of my algorithmic music compositions from the 1980s. I can easily code the logic in Haskell. I'm looking for a the right set of packages and SW so that I can: a) generate short sequences and play them immediately, preferrably in ghci, -- but 'runHaskell Foo.hs | barPlayer' would be acceptable 2) generate MIDI files I'm on OS X. So far what I've found is: Haskore, the midi package, and the jack package - and then I'd need some MIDI software synth for the Mac, and Jack based patcher.... Or perhaps I want SuperCollider, and the Haskell bindings - but that seems rather low level for my needs here (I don't really need to patch together my instruments, and I don't want to have re-write the whole timing framework from scratch.) So - What's a quick easy path here? - Mark

4 3

Problem with pipes interpretation of Lazy MapM program.
by Lyndon Maydwell 11 Jul '13

11 Jul '13

Hi café. I've come up with a little version of 'uniq' that should take into account md5 sums of the file changes... In essence, this: main :: IO () main = getContents >>= mapM check . lines -- PROBLEM!!!! >>= mapM_ (putStrLn . (" --> " ++ )) . strip check :: String -> IO (String, ABCD) check s = (s,) . md5 . Str <$> readFile s strip :: (Ord a, Eq b) => [(a,b)] -> [a] strip = concat . uncurry (zipWith look) . (id &&& maps) look :: (Ord a, Eq b) => (a,b) -> M.Map a b -> [a] look (k,v) m | M.lookup k m == Just v = [] | otherwise = [k] maps :: Ord a => [(a,b)] -> [M.Map a b] maps = scanl (flip (uncurry M.insert)) M.empty Unfortunately mapM isn't lazy, so this doesn't work. I thought this would be a good opportunity to try out the Pipes library for a simple real-world task, but I've come up against some issues with using 'zip' and 'scan' like functions when translating the code. This is what I've got so far, but I'm not sure how to proceed: main :: IO () main = runProxy $ stdinS >-> pipe >-> stdoutD pipe :: () -> ProxyFast () String () String IO () pipe = mapMD check >-> mapScan -- zip, check, output go here >-> mapD ((" --> " ++) . show) mapScan :: () -> ProxyFast () (String, ABCD) () (M.Map String ABCD) IO b mapScan = scanlp (uncurry M.insert) (M.empty) check :: String -> IO (String, ABCD) check s = (s,) . md5 . Str <$> readFile s -- Utils scanlp :: (Monad (p () t a b1 m), Monad m, Functor (p () t a b1 m), Proxy p) => (t -> b1 -> b1) -> b1 -> () -> p () t a b1 m b scanlp f a b = do void $ respond a v <- request () scanlp f (f v a) b There doesn't seem to be any easy zipLike functions, and having to write my own scan function seems odd. Can someone point me in the right direction for this? Thanks! - Lyndon

1 0

Re: [Haskell-cafe] Non-recursive let [Was: GHC bug? Let with guards loops]
by oleg＠okmij.org 11 Jul '13

11 Jul '13

Alberto G. Corona wrote: > I think that a non-non recursive let could be not compatible with the pure > nature of Haskell. I have seen this sentiment before. It is quite a mis-understanding. In fact, the opposite is true. One may say that Haskell is not quite pure _because_ it has recursive let. Let's take pure the simply-typed lambda-calculus, or System F, or System Fomega. Or the Calculus of Inductive Constructions. These calculi are pure in the sense that the result of evaluation of each expression does not depend on the evaluation strategy. One can use call-by-name, call-by-need, call-by-value, pick the next redex at random or some other evaluation strategy -- and the result will be just the same. Although the simply-typed lambda-calculus is quite limited in its expressiveness, already System F is quite powerful (e.g., allowing for the list library), to say nothing of CIC. In all these systems, the non-recursive let let x = e1 in e2 is merely the syntactic sugar for (\x. e2) e1 OTH, the recursive let is not expressible. (Incidentally, although System F and above express self-application (\x.x x), a fix-point combinator is not typeable.) Adding the recursive let introduces general recursion and hence the dependence on the evaluation strategy. There are a few people who say non-termination is an effect. The language with non-termination is no longer pure.

1 0