- Haskell-Cafe - Haskell.org

Organizaing tests in Haskell
by Jan Stolarek 24 Sep '12

24 Sep '12

Dear list, during many years of Java programming I've been faithful to TDD methology. Recently I've been trying to figure out how to do tests in Haskell. Thanks to RWH and help from great folks at #haskell I've managed to get on my feet. There is however one issue I wasn't able to solve. In Java it is very easy to separate test code from the actual source. A Java project simply contains two folders: src and tests. These two are treated as the source directories. Both these directories have the same subdirectory structure. For example I have a source file src/myPackage/mySubpackage/MyClass.java and test for it are kept in file tests/myPackage/mySubpackage/MyClassTest.java. These two files are considered by Java to be in the same package. Here's the main trick: fileds and methods that are marked as protected in Java are accessible to other classes in the same package. This allows to test internal methods of a class by marking them as protected instead of private. This breaks encapsulation but only within a package, which is acceptable. Now I'd like to achieve something similar in Haskell. I'm using cabal's support for testing. I created separate src and tests directories, both with the same subdirectory structure. I keep tests for each module in a separate file (e.g. I have src/Math/MyModule.hs and tests/Math/MyModuleTest.hs) and I have one file that assembles all the tests into a single test suite (I use test-framework for that). The only problem is that in order to test some function from a module I have to expose that function, which pretty much forces me to give up on encapsulation. Is there any better solution to organize tests in Haskell? Should I just give up on module encapsulation, or should I only test functions exposed by the module and don't worry about internal functions? Perhaps I should use some different approach? Jan

4 8

Darcs on Windows 7
by Vasili I. Galchin 24 Sep '12

24 Sep '12

Hello Haskellers, I installed darcs on a Windows 7 machine. A darcs folder was created under "Program Files(x86)" folder. However, when I pull up "Program..." on the left side, darcs not there for me to run it. Why? Is darcs run only from the CLI? Regard, Vasili

4 3

Platform Versioning Policy: upper bounds are not our friends
by Bryan O'Sullivan 23 Sep '12

23 Sep '12

Hi, folks - I'm sure we are all familiar with the phrase "cabal dependency hell" at this point, as the number of projects on Hackage that are intended to hack around the problem slowly grows. I am currently undergoing a fresh visit to that unhappy realm, as I try to rebuild some of my packages to see if they work with the GHC 7.6 release candidate. A substantial number of the difficulties I am encountering are related to packages specifying upper bounds on their dependencies. This is a recurrent problem, and its source lies in the recommendations of the PVP itself (problematic phrase highlighted in bold): When publishing a Cabal package, you should ensure that your dependencies > in the build-depends field are accurate. This means specifying not only > lower bounds, *but also upper bounds* on every dependency. I understand that the intention behind requiring tight upper bounds was good, but in practice this has worked out terribly, leading to depsolver failures that prevent a package from being installed, when everything goes smoothly with the upper bounds relaxed. The default response has been for a flurry of small updates to packages in which the upper bounds are loosened, thus guaranteeing that the problem will recur in a year or less. This is neither sensible, fun, nor sustainable. In practice, when an author bumps a version of a depended-upon package, the changes are almost always either benign, or will lead to compilation failure in the depending-upon package. A benign change will obviously have no visible effect, while a compilation failure is actually *better* than a depsolver failure, because it's more informative. This leaves the nasty-but-in-my-experience-rare case of runtime failures caused by semantic changes. In these instances, a downstream package should *reactively* add an upper bound once a problem is discovered. I propose that the sense of the recommendation around upper bounds in the PVP be reversed: upper bounds should be specified *only when there is a known problem with a new version* of a depended-upon package.

37 59

[Snap] Argument Substitution in Heist Templates with Splices
by Sebastian Fischer 23 Sep '12

23 Sep '12

Hello, the following program demonstrates that arguments in Heist templates are sometimes not substituted in presence of splices: {-# LANGUAGE OverloadedStrings #-} import Blaze.ByteString.Builder (toByteString) import qualified Data.ByteString.Char8 as BS import Data.Functor ((<$>)) import Data.Maybe (fromJust) import Text.Templating.Heist -- just return input node unchanged testSplice :: Splice IO testSplice = (:[]) <$> getParamNode main = do writeFile "test.tpl" "<arg /><test attr='${arg}'><arg /></test>" state <- either error id <$> loadTemplates "." defaultHeistState (builder,_) <- fromJust <$> renderWithArgs [("arg","42")] state "test" BS.putStrLn $ toByteString builder -- 42<test attr='42'>42</test> let state' = bindSplices [("test",testSplice)] state (builder',_) <- fromJust <$> renderWithArgs [("arg","42")] state' "test" BS.putStrLn $ toByteString builder' -- 42<test attr='42'><arg></arg></test> Without using splices, all occurrences of 'arg' in the template are substituted. When using a splice, 'arg' is not substituted underneath the input node of the splice. It is substituted in an attribute of the input node. Is this intentional? How can I ensure substitution also underneath the input node? Best, Sebastian

2 2

Fwd: ANNOUNCE: hsqml-0.1.1 (and hsqml-morris-0.1.0)
by Matthew West 22 Sep '12

22 Sep '12

Hello, In case anyone is wondering this does not compile on vanilla Mac OS 10.6 because XCode 3.2 doesn't support C++11. I believe that XCode 4.2 does support it however I've not yet paid Apple the $100 they want before I can download it for 10.6. Macports has more recent versions of GCC however this option turned out to be fraught with difficulties[1][2]. Has anybody got HsQML working with XCode 4.2? Matt [1] https://trac.macports.org/wiki/UsingTheRightCompiler [2] http://www.haskell.org/haskellwiki/Mac_OS_X_Strike_Force On 11 Sep 2012, at 08:47, Robin KAY wrote: > 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

Haskell Wiki News
by Colin Adams 22 Sep '12

22 Sep '12

This doesn't seem to be up-to-date. It announces GHC 7.4 released (but 7.6.1 was released a couple of weeks ago).

2 1

Munich Haskell Meeting
by Christian Neukirchen 22 Sep '12

22 Sep '12

Dear all, Heinrich is on holiday next week, thus I'll manage it this time: Once again, it's time for our monthly get-together. Functional programmers from Munich and other cities will meet on Thu, 27th of September, at 19h30 at Cafe Puck in Munich. If you plan to join, please go to http://www.haskell-munich.de/dates and click the button! With hope to meet many of you there, -- Christian Neukirchen <chneukirchen(a)gmail.com> http://chneukirchen.org

1 0

Question about type inference of a GADT term
by Florian Lorenzen 22 Sep '12

22 Sep '12

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello cafe, I have the following GADT definitions capturing the simply typed lambda calculus with de Bruijn indices for variables and explicitly annotated types for variables: {-# LANGUAGE GADTs #-} {-# LANGUAGE KindSignatures #-} {-# LANGUAGE DataKinds #-} - -- Typing contexts data Ctx = CtxNil | CtxCons Ctx Type - -- Types data Type = TyInt | TyArrow Type Type - -- Variable indices data Ix (ctx :: Ctx) (ty :: Type) where IxZero :: Ix ctx ty IxSucc :: Ix ctx ty1 -> Ix (CtxCons ctx ty2) ty1 - -- Type representations data TyRep (ty :: Type) where TyRepInt :: TyRep TyInt TyRepArrow :: TyRep ty1 -> TyRep ty2 -> TyRep (TyArrow ty1 ty2) - -- Terms data Term (ctx :: Ctx) (ty :: Type) where TmInt :: Integer -> Term ctx TyInt TmVar :: Ix ctx ty -> Term ctx ty TmAdd :: Term ctx TyInt -> Term ctx TyInt -> Term ctx TyInt TmApp :: Term ctx (TyArrow ty1 ty2) -> Term ctx ty1 -> Term ctx ty2 TmAbs :: TyRep ty1 -> Term (CtxCons ctx ty1) ty2 -> Term ctx (TyArrow ty1 ty2) For the following definition test1 = TmAbs TyRepInt (TmVar IxZero) GHCi infers the type test1 :: Term ctx (TyArrow 'TyInt ty2) I was a bit puzzled because I expected Term ctx (TyArrow TyInt TyInt) Of course, this more special type is an instance of the inferred one, so I can assign it by a type signature. Can someone explain why the inferred type is more general? Terms like test2 = TmAbs TyRepInt (TmAdd (TmVar IxZero) (TmInt 5)) have the type I expected: test2 :: Term ctx (TyArrow 'TyInt 'TyInt) Thank you and best regards, Florian - -- Florian Lorenzen Technische Universität Berlin Fakultät IV - Elektrotechnik und Informatik Übersetzerbau und Programmiersprachen Sekr. TEL12-2, Ernst-Reuter-Platz 7, D-10587 Berlin Tel.: +49 (30) 314-24618 E-Mail: florian.lorenzen(a)tu-berlin.de WWW: http://www.user.tu-berlin.de/florenz/ -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://www.enigmail.net/ iEYEARECAAYFAlBcXs4ACgkQvjzICpVvX7b1WQCePiL+SFNj9X+U0V2fnykuatLX pIcAn1VDNRiSR18s7UgctdPeNzFgStbi =LBGb -----END PGP SIGNATURE-----

2 2

Church vs Boehm-Berarducci encoding of Lists
by oleg＠okmij.org 22 Sep '12

22 Sep '12

There has been a recent discussion of ``Church encoding'' of lists and the comparison with Scott encoding. I'd like to point out that what is often called Church encoding is actually Boehm-Berarducci encoding. That is, often seen > newtype ChurchList a = > CL { cataCL :: forall r. (a -> r -> r) -> r -> r } (in http://community.haskell.org/%7Ewren/list-extras/Data/List/Church.hs ) is _not_ Church encoding. First of all, Church encoding is not typed and it is not tight. The following article explains the other difference between the encodings http://okmij.org/ftp/tagless-final/course/Boehm-Berarducci.html Boehm-Berarducci encoding is very insightful and influential. The authors truly deserve credit. P.S. It is actually possible to write zip function using Boehm-Berarducci encoding: http://okmij.org/ftp/ftp/Algorithms.html#zip-folds

6 10

Passing CString array to Haskell shared library
by Alexander Mumme 21 Sep '12

21 Sep '12

Hello everyone! Long time reader, first time poster. Was wondering if someone could give me some direction or hints on how I might go about passing a CString array into an exported Haskell function. What I'm trying to do is augment the RecordLinkage package from R using Haskell. Seems pretty straight forward. With some help from Neil Mitchell's excellent blog post on the subject ( http://neilmitchell.blogspot.com/2011/10/calling-haskell-from-r.html) I've managed to make calls into Haskell and utilize the Text.EditDistance package. This is working rather well, except that I'm trying to calculate the edit distance from each of 3000 strings to each of 780,000 strings. Since I'm calling Haskell from R once for every comparison (and allocating a result record in R for each return) I find myself running out of memory rather quickly. What I'd like to do is send both complete string lists into Haskell to process, and have it pass me back a result vector. Below is how I'm currently making the call using individual strings. I've tried looking through the GHC User's Guide, through the Wiki on usingthe FFI, and a number of other resources, but I seem to have come to an impasse. Could anyone lend some assistance? ------------- Code start ----- {-# LANGUAGE ForeignFunctionInterface #-} module Levenshtein where import Foreign.C.Types import Foreign.C.String import Foreign.Ptr import Foreign.Storable import Text.EditDistance levenshteinWeight :: Ptr Int -> Ptr Int -> Ptr Int -> Ptr CString -> Ptr CString -> Ptr Int -> IO () levenshteinWeight del ins subs str1 str2 result = do del <- peek del ins <- peek ins subs <- peek subs str1' <- peekCString =<< peek str1 str2' <- peekCString =<< peek str2 poke result $ levenshteinDistance EditCosts { deletionCosts = ConstantCost del, insertionCosts = ConstantCost ins, substitutionCosts = ConstantCost subs, transpositionCosts = ConstantCost 1} str1' str2' foreign export ccall levenshteinWeight :: Ptr Int -> Ptr Int -> Ptr Int -> Ptr CString -> Ptr CString -> Ptr Int -> IO () ------------- Code end ----- Thanks in Advance! Alex

1 0