- Haskell-Cafe - Haskell.org

Reinventing the wheel with HList.
by Nicholls, Mark 24 Jun '15

24 Jun '15

My latest confusion.... I've got a typeclass (with associated types) monoid sort of thing.... I can get a Singleton family of Nat to be an instance of the typeclass. Lists are sort of Natural numbers...with a bit of extra functionality... So I should be able to take HList and make that a member of the same typeclass... I fail for 2 reasons...1...I have to wrestle with the kinds...in a way I find ununderstandbly irritating. And 2...because BOOM. "is a rigid type variable bound by" blab la. My Haskell in many ways is primitive...I seem to be messing with things beyond my ability....but that's the way to learn I suppose. > {-# LANGUAGE DataKinds #-} > {-# LANGUAGE ExplicitForAll #-} > {-# LANGUAGE FlexibleContexts #-} > {-# LANGUAGE FlexibleInstances #-} > {-# LANGUAGE GADTs #-} > {-# LANGUAGE MultiParamTypeClasses #-} > {-# LANGUAGE PolyKinds #-} > {-# LANGUAGE StandaloneDeriving #-} > {-# LANGUAGE TypeFamilies #-} > {-# LANGUAGE TypeOperators #-} > {-# LANGUAGE UndecidableInstances #-} > {-# LANGUAGE ScopedTypeVariables #-} > import Control.Effect > import Control.Effect.Helpers.List > import Control.Effect.WriteOnceWriter > import Prelude hiding ((>>=),(>>),return,fmap,fail) > --import GHC.TypeLits > import Data.Type.Equality > import Data.Type.Bool > --import Data.HList.HList > data Nat where > Z :: Nat > S :: Nat -> Nat > data family Sing (a :: k) > data SNat (a :: Nat) where > SZ :: SNat 'Z > SS :: SNat a -> SNat ('S a) > type family (n :: Nat) :+ (m :: Nat) :: Nat > type instance 'Z :+ m = m > type instance ('S n) :+ m = 'S (n :+ m) Define something like a monoid over the types... (note I have to wrestle with the kind signatures to get the instance of SNat to work, which is fine...but now closes doors to other instances) > class TMonoid (m :: k -> *) where > type O m :: k > zero :: m (O m) > type TPlus m (d :: k) (e :: k) :: k > add :: m a -> m b -> m (TPlus m a b) And this works quite nicely > instance TMonoid SNat where > type O SNat = 'Z > zero = SZ > type TPlus SNat a b = a :+ b > add SZ b = b > add (SS a) b = SS (add a b) Now.... (I have to declare different kind?....I can't get a single simple declaration of TMonoid?) > class TMonoid' (m :: [*] -> *) where > type O' m :: [*] > zero' :: m (O' m) > type TPlus' m (d :: [*]) (e :: [*]) :: [*] > add' :: m a -> m b -> m (TPlus' m a b) Steal the definition of HList > data family HList (l::[*]) :: * > data instance HList '[] = HNil > data instance HList (x ': xs) = x `HCons` HList xs And try to construct an analogous TMonoid of it... > instance TMonoid' HList where > type O' HList = '[] > zero' = HNil > type TPlus' HList a b = a :++ b > add' HNil b = b > add' (HCons x xs) ys = HCons x (add' xs ys) BOOM Couldn't match type 'a' with ''[]' 'a' is a rigid type variable bound by the type signature for add' :: HList a -> HList b -> HList (TPlus' HList a b) at broadcast2.lhs:90:5 Expected type: HList a Actual type: HList '[] Relevant bindings include add' :: HList a -> HList b -> HList (TPlus' HList a b) (bound at broadcast2.lhs:90:5) In the pattern: HNil In an equation for 'add'': add' HNil b = b In the instance declaration for 'TMonoid' HList' Hmmm....confused...why does add SZ b = b work, but the analogous add' HNil b = b fail CONFIDENTIALITY NOTICE This e-mail (and any attached files) is confidential and protected by copyright (and other intellectual property rights). If you are not the intended recipient please e-mail the sender and then delete the email and any attached files immediately. Any further use or dissemination is prohibited. While MTV Networks Europe has taken steps to ensure that this email and any attachments are virus free, it is your responsibility to ensure that this message and any attachments are virus free and do not affect your systems / data. Communicating by email is not 100% secure and carries risks such as delay, data corruption, non-delivery, wrongful interception and unauthorised amendment. If you communicate with us by e-mail, you acknowledge and assume these risks, and you agree to take appropriate measures to minimise these risks when e-mailing us. MTV Networks International, MTV Networks UK & Ireland, Greenhouse, Nickelodeon Viacom Consumer Products, VBSi, Viacom Brand Solutions International, Be Viacom, Viacom International Media Networks and VIMN and Comedy Central are all trading names of MTV Networks Europe. MTV Networks Europe is a partnership between MTV Networks Europe Inc. and Viacom Networks Europe Inc. Address for service in Great Britain is 17-29 Hawley Crescent, London, NW1 8TT.

2 2

ANNOUNCE: stack 0.1.0.0
by Emanuel Borsboom 24 Jun '15

24 Jun '15

Announcing stack, a new build tool, first stable release. More information at: https://www.fpcomplete.com/blog/2015/06/stack-0-1-release

1 0

RebindableSyntax on individual do blocks
by Mark Roberts 23 Jun '15

23 Jun '15

I have a program that uses both monads and indexed monads, and I'd like to use do-notation for each in the same source file. Is there a way to rebind syntax for only the do blocks that make use of an indexed monad? Thanks, Mark

4 4

Turning Monoids into Categories
by David Feuer 23 Jun '15

23 Jun '15

Is there a newtype defined in some normal sort of place for turning Monoids into Categories? I'm looking for something like this: newtype ConstConst m a b = ConstConst m instance Monoid m => Category (ConstConst m) where id = ConstConst mempty ConstConst x . ConstConst y = ConstConst (x <> y) Thanks, David Feuer

2 2

Mild confusion around type family
by Nicholls, Mark 23 Jun '15

23 Jun '15

Below is a very simple almost canonical natural number model. I'm slightly confused by how type families work Wikipedia says.... "Type families are a feature of some type systems<https://en.wikipedia.org/wiki/Type_system> that allow PARTIAL functions between types to be defined by pattern matching<https://en.wikipedia.org/wiki/Pattern_matching>" So... > {-# LANGUAGE DataKinds #-} > {-# LANGUAGE ExplicitForAll #-} > {-# LANGUAGE FlexibleContexts #-} > {-# LANGUAGE FlexibleInstances #-} > {-# LANGUAGE GADTs #-} > {-# LANGUAGE MultiParamTypeClasses #-} > {-# LANGUAGE PolyKinds #-} > {-# LANGUAGE StandaloneDeriving #-} > {-# LANGUAGE TypeFamilies #-} > {-# LANGUAGE TypeOperators #-} > {-# LANGUAGE UndecidableInstances #-} > {-# LANGUAGE ScopedTypeVariables #-} lets do something simple > data Z > data S n > type family Sub a b > -- a - 0 == a > type instance Sub a Z = a > -- (a + 1) - (b + 1) == a - b > type instance Sub (S a) (S b) = Sub a b > oneMinusOne :: Z > oneMinusOne = undefined :: Sub (S Z) (S Z) I want this to be a type error!...but the above type family appears to not be partial...this IS defined...."Sub Z (S Z)" is a type! (I thought it wasn't) > oneMinusTwo :: Sub Z (S Z) > oneMinusTwo = undefined :: Sub (S Z) (S (S Z)) CONFIDENTIALITY NOTICE This e-mail (and any attached files) is confidential and protected by copyright (and other intellectual property rights). If you are not the intended recipient please e-mail the sender and then delete the email and any attached files immediately. Any further use or dissemination is prohibited. While MTV Networks Europe has taken steps to ensure that this email and any attachments are virus free, it is your responsibility to ensure that this message and any attachments are virus free and do not affect your systems / data. Communicating by email is not 100% secure and carries risks such as delay, data corruption, non-delivery, wrongful interception and unauthorised amendment. If you communicate with us by e-mail, you acknowledge and assume these risks, and you agree to take appropriate measures to minimise these risks when e-mailing us. MTV Networks International, MTV Networks UK & Ireland, Greenhouse, Nickelodeon Viacom Consumer Products, VBSi, Viacom Brand Solutions International, Be Viacom, Viacom International Media Networks and VIMN and Comedy Central are all trading names of MTV Networks Europe. MTV Networks Europe is a partnership between MTV Networks Europe Inc. and Viacom Networks Europe Inc. Address for service in Great Britain is 17-29 Hawley Crescent, London, NW1 8TT.

4 9

ANNOUNCE: fgl-arbitrary
by Ivan Lazar Miljenovic 22 Jun '15

22 Jun '15

I've just released the first version of fgl-arbitrary [1], which provides Arbitrary instances for fgl graphs for use with QuickCheck. Also provided are some wrapper newtypes to produce specific types of graphs, and also the ability to generate just valid node and edge lists for use with other graph-like data structures. [1]: http://hackage.haskell.org/package/fgl-arbitrary -- Ivan Lazar Miljenovic Ivan.Miljenovic(a)gmail.com http://IvanMiljenovic.wordpress.com

1 0

"SameConstraints?" type constraints
by Leza Morais Lutonda 22 Jun '15

22 Jun '15

Hi All, I have the following data type: data S e where SC :: d -> e -> S e And I want to declare a Show instance for it in the following way: instance Show e => Show (S e) where show (SC x y) = show x ++ show y But, of course it don't typechecks because: could not deduce `Show d` arising from a use of `show`. Is there a way to constraint the `d` type in the `SC` constructor definition to have the same constraints of the `e` type? Something like: SC :: SameConstraints d e => d -> e -> S e ??? Thanks! -- Leza Morais Lutonda, Lemol-C http://lemol.github.io 50 Aniversario de la Cujae. Inaugurada por Fidel el 2 de diciembre de 1964 http://cujae.edu.cu

7 13

stack mailing list
by Michael Snoyman 22 Jun '15

22 Jun '15

We recently started a mailing list for discussions of stack[1], a new Haskell build tool (see initial release announcement[2]). This list is open for questions, requests, discussions, and internal development (though if that last becomes too chatty, we can always move it to a separate list). You can sign up for the mailing list at: https://groups.google.com/d/forum/haskell-stack [1] https://github.com/commercialhaskell/stack#readme [2] https://www.fpcomplete.com/blog/2015/06/announcing-first-public-beta-stack

1 0

Do people use visitor pattern in Haskell...
by Nicholls, Mark 22 Jun '15

22 Jun '15

My initial guess would be no...its built into the language But.... I find myself tempted to use it (in order to match on a type that's the instance of a typeclass)...which makes me think I've missed something. So... > data Foo = Foo > data Bar = Bar Ooo...this looks like an OO design pattern...surely wrong? > class Wibble a where > visit :: (Foo -> b) -> (Bar -> b) -> a -> b > instance Wibble Foo where > visit f _ x = f x > instance Wibble Bar where > visit _ f x = f x I want a list of Wibbles.... hmmm... (Wibble a) => [a] is clearly wrong... I want [(Wibble a) =>a] so I package it up? > data WibblePackage where > WibblePackage :: (Wibble a) => a -> WibblePackage lets try this now...so pointless function across Wibbles in a list. > fizzBuzz :: [WibblePackage] -> Integer > fizzBuzz [] = 0 > fizzBuzz ((WibblePackage x) : xs) = (visit (\_ -> 1) (\_ -> 2) x) + (fizzBuzz xs) > help :: Integer > help = fizzBuzz [WibblePackage Foo,WibblePackage Bar] That works! OO nerds will get uptight about "closing" the Wibble typeclass....but I've managed to create a list of different types and have a mechanism for recovering the type. How would a Haskell nerd do this? (I have looked, but they look more complicated....maybe to my OO eye). CONFIDENTIALITY NOTICE This e-mail (and any attached files) is confidential and protected by copyright (and other intellectual property rights). If you are not the intended recipient please e-mail the sender and then delete the email and any attached files immediately. Any further use or dissemination is prohibited. While MTV Networks Europe has taken steps to ensure that this email and any attachments are virus free, it is your responsibility to ensure that this message and any attachments are virus free and do not affect your systems / data. Communicating by email is not 100% secure and carries risks such as delay, data corruption, non-delivery, wrongful interception and unauthorised amendment. If you communicate with us by e-mail, you acknowledge and assume these risks, and you agree to take appropriate measures to minimise these risks when e-mailing us. MTV Networks International, MTV Networks UK & Ireland, Greenhouse, Nickelodeon Viacom Consumer Products, VBSi, Viacom Brand Solutions International, Be Viacom, Viacom International Media Networks and VIMN and Comedy Central are all trading names of MTV Networks Europe. MTV Networks Europe is a partnership between MTV Networks Europe Inc. and Viacom Networks Europe Inc. Address for service in Great Britain is 17-29 Hawley Crescent, London, NW1 8TT.

7 11

"Universal" constraints?
by Michael Walker 22 Jun '15

22 Jun '15

Hi all, I was wondering about constraints in class declarations. Suppose we have something like this: class Monad m => Store m where type Var m :: * -> * new :: a -> m (Var m a) read :: Var m a -> m a write :: Var m a -> a -> m () Then some reasonable implementations would be IO and ST, using IORefs and STRefs, instance Store IO where type Var IO = IORef new = newIORef read = readIORef write = writeIORef instance Store (ST s) where type Var (ST s) = STRef s new = newSTRef read = readSTRef write = writeSTRef Now, both IORefs and STRefs have equality, which doesn't require equality of the contained type, instance Eq (IORef a) where ... instance Eq (STRef s a) where ... This corresponds to pointer equality and, because reading the values is a monadic action, is as good an Eq instance as we can get. Given this, it would be nice to be able to specify that all instances of our Store class have this property. Something like, class (Monad m, Eq (Var m a)) => Store m where ... But we can't write this! The `a` isn't in scope! What I really want to do is to be able to write class constraints like this, class (forall a. C (T a)) => D x where ... Or why not get really crazy: class (forall a. C a => D (T a)) => E x where ... Is there some extension or combination of extensions that would make this work without completely sacrificing type safety? FlexibleContexts allows it (the simpler form, at least) for function constraints, but not for class constraints. If not, has anyone thought about this sort of thing before? -- Michael Walker (http://www.barrucadu.co.uk)

4 3