- Haskell-Cafe - Haskell.org

Typeclass with an ‘or’ restriction.
by Mateusz Kowalczyk 10 May '13

10 May '13

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Greetings, We can currently do something like > class (Num a, Eq a) => Foo a where bar :: a -> a -> Bool bar = > (==) This means that our `a' has to be an instance of Num and Eq. Apologies for a bit of an artificial example. Is there a way however to do something along the lines of: > class Eq a => Foo a where bar :: a -> a -> Bool bar = (==) > > class Num a => Foo a where bar :: a -> a -> Bool bar _ _ = False This would allow us to make an instance of Num be an instance of Foo or an instance of Eq to be an instance of Foo. The compiler currently complains about multiple declarations. Is there currently a way to achieve this? The main issue I can see with this is that given an instance of both, Num and Eq, it wouldn't be possible to pick the correct default implementation. Purely a theoretical question. - -- Mateusz K. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.19 (GNU/Linux) iQIcBAEBAgAGBQJRjP0hAAoJEM1mucMq2pqX30wP/0d0TQHs4S3G5TBw+T1baI6n g/5k/YlPmSgS3FaO8JMQsb2uqL8dGPZGUN7d2ohwcigtXS88KWH4u4rTjrs1+p8o ktS+vIE4kEEedTAX6wmP2Zn+rvK2zFGboCafaX/a/IxT5CbwYZ97RrWCjzz1jlPs S/VlhNHcTQ7Cf/0pa0xJ1kbao+vBHiWtWjxcdzCT/6zS86+jm9vz8qrYT3TWUd3y AJXPBjRuXeyz+RDv18yrth7hMlAvaeoWzmC4gbGFHC68/Oq2l8kz+4dt6ApN+mIy l73wNjrU185YoZ2dkuKTIph8/BadgkD+9ktkgZZ/NlqxElc596BdbOcMfVk4rz4A 0nWqaLa4QctIthghJ1UNfKS8lQzkVVRT6e03LYdgPkqJm1HQxkjHL/WieV2NEoRf 1K9S4SVW8Aq/ML/Gmx782Z3jMECfnYWntf9gSOwFASB64tVej1iUxb7UsfxJAL1t ysf9MjcbZsHe3M/JAq4f8HtHoZoiIG/TTjD0yo74owssJDfOTDqmYMriyelcnUf6 hDCPZyUqLPMTNVx07T+gwfXJoE1HK20hzVe2o1dPBZ8Kb2KJbNg0+sUSB3/v6O8e EJ6w7aQ3OnZUACd1i2uLZiphMF8d8va4T8eTFUyROODHvaPv6netzr8gPcIXhpVO 5OwyMfO54cinZYen7+HR =iAYY -----END PGP SIGNATURE-----

2 1

Reinventing a solution to configuration problem
by Krzysztof Skrzętnicki 10 May '13

10 May '13

Hello Cafe, I came up with (actually reinvented [1]) a solution to configuration problem. For example configuration: data Config = Config { port :: Int, verbosity :: Int, logfile :: FilePath } deriving Show the basic idea is to use implicit parameters: runServer :: (?config :: Config) => IO () The problem is that it is tiresome to decorate all functions in this way. It is also non composable. For example if we switch to using two separate configuration entities like this: data LogConfig = LogConfig { verbosity :: Int, logfile :: FilePath } data NetConfig = NetConfig { port :: Int } we suddenly have to fix types all over the place. Yuck. The extended idea here (compared to [2]) is to use type synonym with all required parameters. This is simple (but requires at least Rank2Types): type ConfIO a = (?config :: Config) => IO a All the functions requiring config will now be simply in ConfIO monad which is really IO with a bonus. A great feature of this solution is that we don't need to use liftIO anywhere. The core *problem* with this approach is that at least current version of GHC fails to propagate implicit parameters the way it propagate typeclass requirements. For example I would like the following function to have (inferred) type -- printConfig :: (?config :: Config) => IO () printConfig = print (?config :: Config) Instead GHC will complain about unbound implicit parameter. Similar problem from GHCi session: Prelude> let f = print (?conf :: Int) <interactive>:11:16: Unbound implicit parameter (?conf::Int) arising from a use of implicit parameter `?conf' In the first argument of `print', namely `(?conf :: Int)' In the expression: print (?conf :: Int) In an equation for `f': f = print (?conf :: Int) For reason I don't fully understand this is fixed by disabling monomorphic restriction. The documentation mentions the problem but in a different context: http://www.haskell.org/ghc/docs/latest/html/users_guide/other-type-extensio… *TL;DR*: A nice solution to configuration problem is to use implicit parameters and type synonym for IO monad extended with implicit configuration parameters. It requires Rank2Types for defining the synonym and NoMonomorphismRestriction for lightweight use. Best regards, Krzysztof Skrzętnicki [1] I only realised this after starting the writing of this email. The implicit parameters solution is given at least in [2]. Somehow I didn't came across it before. [2] "Global Variables in Haskell (2004)"<http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.23.145&rep=rep1&ty…> by John Hughes

1 0

Flip around type parameters?
by Christopher Howard 10 May '13

10 May '13

Hi. Does Haskell allow you to flip around type parameters somehow? I was playing around with toy code (still learning the fundamentals) and I came up with a class like this: code: -------- class Rotatable2D a where rotate :: (Num b) => (a b) -> b -> (a b) -------- It was easy to make an instance of a generic single-parameter type: code: -------- data Angle a = Angle a deriving (Show) instance Rotatable2D Angle where rotate (Angle a) b = Angle (a + b) -------- But let's say I have something a little more complicated: code: -------- data CircAppr a b = CircAppr a a b -- radius, rotation angle, number of points -------- I think I need something like so: -------- instance Rotatable2D (\x -> CircAppr x a) where rotate (CircAppr a b c) d = CircAppr a (b + d) c -------- But I tried that and it isn't valid syntax. -- frigidcode.com

2 1

Fwd: Error compiling transformers-base-0.4.1 (debian 64 bits, ghc-7.4.2)
by jean-christophe mincke 09 May '13

09 May '13

Thank you Sjoerd, Unfortunately this does not work. I think the problem is linked to the evaluation of the condition (in Base.hs) #if !MIN_VERSION_base(4,4,0) && HS_TRANSFORMERS_BASE__ORPHANS Just as a test, I modified the file and added the following lines A. negation of min_version #if *!*MIN_VERSION_base(4,4,0) rgfyfgruyfgreuy -- making sure the compilation fails #endif B #if MIN_VERSION_base(4,4,0) rgfyfgruyfgreuy -- making sure the compilation fails #endif In both case the compilations failed (cabal clean, configure, build). That is, it seems that the #if always evaluate to true... I am surely missing something here... Regards J-C On Thu, May 9, 2013 at 4:13 PM, Sjoerd Visscher <sjoerd(a)w3future.com> wrote: > There's a flag to turn it off, does that work? Try cabal install > transformers-base -f-OrphanInstances > > Sjoerd > > On May 9, 2013, at 2:47 PM, jean-christophe mincke < > jeanchristophe.mincke(a)gmail.com> wrote: > > Hello Café > > I am running into problems when installing transformers-base-0.4.1. > > Has anyone an idea about what is going on? > > REM : this is a "brand new" debian install, I have just install ghc, cabal > install cabal-install etc. > > Thank you > > very much > > J-C > > cabal install transformers-base > Resolving dependencies... > Downloading transformers-base-0.4.1... > Configuring transformers-base-0.4.1... > Preprocessing library transformers-base-0.4.1... > Building transformers-base-0.4.1... > [1 of 1] Compiling Control.Monad.Base ( src/Control/Monad/Base.hs, > dist/build/Control/Monad/Base.o ) > > src/Control/Monad/Base.hs:63:10: > Duplicate instance declarations: > instance Applicative (L.ST <http://l.st/> s) > -- Defined at src/Control/Monad/Base.hs:63:10 > instance Applicative (L.ST <http://l.st/> s) -- Defined in > `Control.Applicative' > > src/Control/Monad/Base.hs:67:10: > Duplicate instance declarations: > instance Applicative (S.ST <http://s.st/> s) > -- Defined at src/Control/Monad/Base.hs:67:10 > instance Applicative (S.ST <http://s.st/> s) -- Defined in > `Control.Applicative' > cabal: Error: some packages failed to install: > transformers-base-0.4.1 failed during the building phase. The exception > was: > ExitFailure 1 > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe > > >

1 1

Accessing resource files in tests
by Graham Berks 09 May '13

09 May '13

Hi, whats the recommended way to access a resource file for a test within cabal. I have a text file containing data that I use in a test and want to be able to access it in a root path independent manner. Thanks

2 1

Re: [Haskell-cafe] Interfacing Java/Haskell
by Hans Georg Schaathun 09 May '13

09 May '13

On Thu, May 09, 2013 at 11:29:19AM -0400, Dan Mead wrote: > if you're on the jvm already why not consider clojure? This is why (from the clojure site you referenced): «Clojure is impure, in that it doesn't force your program to be referentially transparent, and doesn't strive for 'provable' programs.» I won't say anything about the general suitability of clojure, but I will not learn a new language at this stage unless it gives me purity, strictness, and transparency. But thanks anyway. It is useful to be aware of. > otherwise you'll have to do some fancy JNI calls and worry about calling > haskell from C, like this > http://www.haskell.org/haskellwiki/Calling_Haskell_from_C I was afraid so, and JNI appears to have a rumour for awkwardness, although there might be some simplifying overlays. -- :-- Hans Georg

1 0

What symbol / atom/ interned-string package is currently preferred?
by Ryan Newton 09 May '13

09 May '13

Hi cafe, I use symbols all the time when manipulating code. I vacillate between using stringtable-atom, symbol, and simple-atom. Unfortunately: - Stringtable-atom is has a broken build right now (GHC 7.6). (Also I've had some crashes with it in the past.) - simple-atom has dependency upper bounds that restrict it to GHC <= 7.0 - "symbol" has a design flaw, which is that it exposes the constructor for Symbol, which reveals the (nondeterministic) unique Int IDs. I.e. it is not "SafeHaskell". I submitted a pull request to relax simple-atom, but perhaps there are other packages people use with which I am not aware? Thanks, -Ryan

5 5

ghci: Difference in garbage collection etc. between REPL and function
by Niklas Hambüchen 09 May '13

09 May '13

I have an annoying bug in a C binding that can be triggered like this: handle <- open ... prep <- makePreparedStatement handle "INSERT ..." performGC runStatement prep close handle If I run these steps one by one in ghci, garbage ends up in my handle as expected. However, if I "let main = do ..." this whole block in order to pack it in a test case, it does not happen, neither in ghci nor ghc. What might be the special magic in ghci's REPL that allows me to trigger my bug so easily there? Thanks Niklas

3 5

Error compiling transformers-base-0.4.1 (debian 64 bits, ghc-7.4.2)
by jean-christophe mincke 09 May '13

09 May '13

Hello Café I am running into problems when installing transformers-base-0.4.1. Has anyone an idea about what is going on? REM : this is a "brand new" debian install, I have just install ghc, cabal install cabal-install etc. Thank you very much J-C cabal install transformers-base Resolving dependencies... Downloading transformers-base-0.4.1... Configuring transformers-base-0.4.1... Preprocessing library transformers-base-0.4.1... Building transformers-base-0.4.1... [1 of 1] Compiling Control.Monad.Base ( src/Control/Monad/Base.hs, dist/build/Control/Monad/Base.o ) src/Control/Monad/Base.hs:63:10: Duplicate instance declarations: instance Applicative (L.ST s) -- Defined at src/Control/Monad/Base.hs:63:10 instance Applicative (L.ST s) -- Defined in `Control.Applicative' src/Control/Monad/Base.hs:67:10: Duplicate instance declarations: instance Applicative (S.ST s) -- Defined at src/Control/Monad/Base.hs:67:10 instance Applicative (S.ST s) -- Defined in `Control.Applicative' cabal: Error: some packages failed to install: transformers-base-0.4.1 failed during the building phase. The exception was: ExitFailure 1

2 1

Test access resources files in cabal-dev
by Graham Berks 09 May '13

09 May '13

Hi, whats the recommended way to access a resource file for a test within cabal. I have a text file containing data that I use in a test and want to be able to access it in a root path independent manner. Thanks

1 0