August 2016 - Haskell-Cafe

Haskell Job In Cambridge UK Working On Compilers, Machine Learning And Computer Vision!
by Joe Hermaszewski 16 Aug '16

16 Aug '16

We have a Haskell job opening in Cambridge, UK. Please see https://www.myrtlesoftware.com/vacancies/ Here is a video of the computer vision algorithms running in hardware we have produced using Haskell: https://www.myrtlesoftware.com/case-studies/fpga/ Myrtle is currently working as part of the UK government’s autonomous vehicles program and we are looking to hire more Haskell developers to join our expanding team. You will be working directly on our technology that translates high-level signal processing code into efficient hardware designs. Applicants should have a deep understanding of Haskell, modern compiler technology and should be familiar with git and CI based development. We are targeting GPU and FPGA backends so experience with either of these would be advantageous although we’re happy to consider strong developers without direct experience of these two areas. Similarly, a familiarity with an HDL, such as VHDL or Verilog, would be viewed positively but is not essential. Applicants should send a resume and covering email, stating your suitability for this role, to jobs(a)myrtlesoftware.com Myrtle is a software company based in Cambridge UK with a long history of working at the forefront of computer graphics and compiler technology. We are best known for having helped produce computer generated images for over twenty major Hollywood movies including the Transformers franchise. Myrtle’s clients have included NYSE and NASDAQ listed companies in LA, Vancouver and London. Closing date for applications 5pm, 27th August 2016

1 0

Which is the most "hackable" compiler?
by MarLinn 15 Aug '16

15 Aug '16

So I have some crazy ideas for language extensions/reinterpretations and ways I might try to implement them. But ghc is a complex beast and my crazy ideas have too many interrelationships for a noob like me to try to bold them onto it. After all, I have never worked on ghc - or any compiler for that matter. I did look at the code, but I reckon maybe I should gather some experience first before diving into that one. It would be even worse to offload my ideas onto someone else before I at least tried them and got some feedback. No fun for anyone. But of course I don't want to start from scratch. As fun as that could be, there's no need to reinvent all the wheels just for a different axle. And I hear rumors that there are other Haskell compilers out there, even if most of them live in the shadow/slipstream of their brother. Alas, I don't have any idea what their statuses and philosophies are. So maybe you can help me here: Have you experienced any of the alternative compilers as especially easy for a newcomer to pick up and play around with? If it helps, I would be satisfied with plain Haskell2010 or even Haskell 98, although some GADT and/or TypeFamilies code to butcher would be nice, too. The ideas are mostly about larger scale structures like whole functions. One representative example idea is "Could it help the implementation - and does it even make sense - to view a module as just a weirdly written zero parameter type class?" As I said: crazy ideas. I'm thankful for any thoughts and ideas. MarLinn

4 5

Re: [Haskell-cafe] The relationship between F-algebras and the Free Monad
by Olaf Klinke 15 Aug '16

15 Aug '16

Dear Will, I used to think of the extra parameter "a" in Free f a as the type of variables in the terms. So in short, while Fix f gives you the closed terms of your AST, Free f a gives you open terms with variables of type a. Indeed the equations Fix f = f (Fix f) Free f a = Pure a | Free (f (Free f a)) are related: Let Var a f x = Var a | App (f x) Then Free f a = Fix (Var a f) Hence the free monad is the fixed point of the functor f where you first sneak in free variables of type a. Moreover, given a function from variables to closed terms, you can turn an open term to a closed term (see below). Your exercise: Implement the below using catamorphisms and so on. -- Olaf newtype Fix f = Fix (f (Fix f)) data Free f a = Pure a | Free (f (Free f a)) data Var a f x = Var a | App (f x) -- tofree.tofix = id -- tofix.tofree = id tofree :: Functor f => Free f a -> Fix (Var a f) tofree (Pure a) = Fix (Var a) tofree (Free fx) = Fix (App (fmap tofree fx)) tofix :: Functor f => Fix (Var a f) -> Free f a tofix (Fix (Var a)) = Pure a tofix (Fix (App fx)) = Free (fmap tofix fx) close :: Functor f => (a -> Fix f) -> Free f a -> Fix f close evaluate (Pure a) = evaluate a close evaluate (Free fx) = Fix (fmap (close evaluate) fx)

1 0

Typeclass resolution errors quite puzzling
by Tom Ellis 15 Aug '16

15 Aug '16

Here's a program with an odd error message (GHC 8.0.1): data A a = A a deriving Eq data B = B main :: IO () main = print (A B == A B) test/main.hs:5:15: error: • No instance for (Eq B) arising from a use of ‘==’ • In the first argument of ‘print’, namely ‘(A B == A B)’ In the expression: print (A B == A B) In an equation for ‘main’: main = print (A B == A B) I get an error about Eq B even though it's Eq A that is manifestly required at the call site. This error is odder when A and B are defined far away from the use of '=='. This is even odder: data A a = A a data B = B instance Ord a => Eq (A a) where main :: IO () main = print (A B == A B) test/main.hs:7:15: error: • No instance for (Ord B) arising from a use of ‘==’ • In the first argument of ‘print’, namely ‘(A B == A B)’ In the expression: print (A B == A B) In an equation for ‘main’: main = print (A B == A B) Now not only is the type puzzling (B instead of A) but the *class* is puzzling (Ord instead of Eq). This occurred to me in practice because 'Data.Graph.Inductive.PatriciaTree.Gr' has '(Eq a, Ord b) => Eq (Gr a b)'. It would have been a lot more helpful to see * No instance for (Ord B) * arising from (Eq A) * arising from the use of '==' Does anyone agree with me that GHC should produce the full trace when it fails to resolve instances rather than just the proximal failure? Tom

6 11

C call blocking indefinitely/behaving differently when called from Haskell
by Oliver Charles 15 Aug '16

15 Aug '16

Hi all, I'm trying to interface with a Basler USB3 camera <http://www.baslerweb.com/en/products/> from a Haskell application, but I'm experiencing some difficulty. The camera comes with a C++ library that makes it fairly straight forward. The following code can be used to acquire a camera source: PylonAutoInitTerm pylon; CInstantCamera camera( CTlFactory::GetInstance().CreateFirstDevice()); camera.RegisterConfiguration( (CConfigurationEventHandler*) NULL, RegistrationMode_ReplaceAll, Cleanup_None); cout << "Using device " << camera.GetDeviceInfo().GetModelName() << endl; Using the inline-c and inline-c-cpp libraries, I am trying to perform this same call from a Haskell executable. I have the following: {-# LANGUAGE QuasiQuotes, TemplateHaskell #-} import qualified Language.C.Inline as C import qualified Language.C.Inline.Cpp as C C.context C.cppCtx C.include "pylon/PylonIncludes.h" C.include "iostream" C.using "namespace Pylon"; C.using "namespace std"; main :: IO () main = [C.block| void { PylonAutoInitTerm pylon; CInstantCamera camera( CTlFactory::GetInstance().CreateFirstDevice()); camera.RegisterConfiguration( (CConfigurationEventHandler*) NULL, RegistrationMode_ReplaceAll, Cleanup_None); cout << "Using device " << camera.GetDeviceInfo().GetModelName() << endl; }|] However, when the resulting executable is ran, it hangs - failing to print "Using device". inline-c generates a Test.cpp file, which contains #include "pylon/PylonIncludes.h" #include "iostream" using namespace Pylon; using namespace std; extern "C" { void inline_c_Main_0_e23f0c6a0bc10367be76252744bf4e6c74493314() { PylonAutoInitTerm pylon; CInstantCamera camera( CTlFactory::GetInstance().CreateFirstDevice()); camera.RegisterConfiguration( (CConfigurationEventHandler*) NULL, RegistrationMode_ReplaceAll, Cleanup_None); cout << "Using device " << camera.GetDeviceInfo().GetModelName() << endl; } } If I modify this exact file, I can turn it into something that can be run, by simply adding: int main() { inline_c_Main_0_e23f0c6a0bc10367be76252744bf4e6c74493314(); } At this point, I can now compile that very C++ code with G++ and run it: [nix-shell:~/work/circuithub/receiving-station]$ g++ Test.cpp -I/home/ollie/work/circuithub/receiving-station/pylon5/include -L/home/ollie/work/circuithub/receiving-station/pylon5/lib64 -Wl,-E -lpylonbase -lpylonu tility -lGenApi_gcc_v3_0_Basler_pylon_v5_0 -lGCBase_gcc_v3_0_Basler_pylon_v5_0 -Wl,--enable-new-dtags -Wl,-rpath,/home/ollie/work/circuithub/receiving-station/pylon5/lib64 [nix-shell:~/work/circuithub/receiving-station]$ ./a.out GetDeviceDiscoveryInfo: bFunctionClass = 14, bFunctionSubClass = 3, bFunctionProtocol = 0. Device is not an U3V device. GetDeviceDiscoveryInfo: bFunctionClass = 14, bFunctionSubClass = 3, bFunctionProtocol = 0. Device is not an U3V device. Using device acA3800-14um So as we can see, the C++ code is not the problem. Does anyone know why I don't see the same behaviour from my Haskell application? I imagine there might be some thread trickery happening behind the scenes, but I don't know what to even begin looking for. Thanks, let me know if you need any more information. - Ollie

2 1

Re: [Haskell-cafe] library for set of heterogeneous values?
by Anthony Clayden 15 Aug '16

15 Aug '16

> Frames - http://acowley.github.io/Frames/ - is built on > top of vinyl and provides one way to access data either as > an array of structures or a structures of arrays. Thanks Doug, Nice ability to define column types on the fly, and interpret them from incoming CSV. Yes and treat as structure of arrays. AFAICT Frames keeps each set of records distinct; and wants each set to be same type/same columns and types. So there's no heterogeneous repository. > whether it allows the kind of filtering you were asking for. There's standard filtering features you'd get with Set or List. I'm not seeing any ability to match/join across different datasets. That is, you'd have to write your own code. AntC

1 0

Unable to correctly infer function type, via pattern matching?
by David Banas 14 Aug '16

14 Aug '16

Hi all, I wonder if anyone can help me find my error, in this code: It’s note clear to me why x :: a is an error, except that maybe the type of f is not being correctly inferred? Am I, maybe, not allowed to use pattern matching, on the right side of ‘=‘? Thanks, -db data Stream a = Cons a (Stream a) newtype StreamMap a b = SM (Stream a -> Stream b) instance Arrow StreamMap where pure f = SM g where g (Cons x xs) = Cons (f x) (g xs) SM f >>> SM g = SM (g . f) first (SM g) = SM h -- h :: Stream (a, c) -> Stream (b, c) where h :: Stream (a, c) -> Stream (b, c) h (Cons (x, y) zs) = Cons ((f x), y) (h zs) g :: Stream a -> Stream b g (Cons x xs) = Cons (f x) (g xs) Couldn't match expected type ‘b1’ with actual type ‘a’ ‘a’ is a rigid type variable bound by the type signature for g :: interactive:IHaskell152.Stream a -> interactive:IHaskell152.Stream b1 at :10:17 ‘b1’ is a rigid type variable bound by the type signature for g :: interactive:IHaskell152.Stream a -> interactive:IHaskell152.Stream b1 at :10:17 Relevant bindings include xs :: interactive:IHaskell152.Stream a (bound at :11:27) x :: a (bound at :11:20) g :: interactive:IHaskell152.Stream a -> interactive:IHaskell152.Stream b1 (bound at :11:12) In the first argument of ‘f’, namely ‘x’ In the first argument of ‘IHaskell152.Cons’, namely ‘(f x)’

2 1

Re: [Haskell-cafe] Twitter coding kata
by Damian 14 Aug '16

14 Aug '16

I don't know about the feed, but a while ago I was playing with some exercises at exercism.io. Maybe you'd find it interesting. Regards, Damian Op 14 aug. 2016 21:30 schreef "Michael Litchard" <michael(a)schmong.org>: I've been told that there is a twitter feed that gives out little haskell exercises. Can't find it. Anyone know about it? _______________________________________________ Haskell-Cafe mailing list To (un)subscribe, modify options or view archives go to: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe Only members subscribed via the mailman list are allowed to post.

4 3

Twitter coding kata
by Michael Litchard 14 Aug '16

14 Aug '16

I've been told that there is a twitter feed that gives out little haskell exercises. Can't find it. Anyone know about it?

1 0

Function and pattern bindings
by James Brown 14 Aug '16

14 Aug '16

I am reading Haskell report 2010 and I can't understand the differences between function bindings and pattern bindings. Can anyone explain the differences?

2 1