January 2016 - Haskell-Cafe

Hackage hoq?
by Vasili I. Galchin 27 Jan '16

27 Jan '16

Hello, I successfully built hoq. When I run th hoq executable, I get ">" prompts but am not sure what to do. I read some of the code and I tried putting .hoq on the command line, That didn't work. Has anyone tried hoq?? Vasily

1 0

[Ann] A new, fun, easy book/tute on Haskell
by Julian Leviston 27 Jan '16

27 Jan '16

[Ann] A new, fun, easy book/tute on Haskell http://www.happylearnhaskelltutorial.com Just released this two days ago, roughly. We've released 7 out of 20 sections on the site so far, and the book is up on leanpub ($8 or $4 - $20 set your own price), at about 95% done. We don't want to put it to 100% until we've had at least 20 or more technical people read it and give us feedback, and until we've done a thorough revision for corrections. We felt like there's a need for an updated education resource that's both fun and for super beginners who have never programmed before. Essentially a slightly less drawing-focussed and slightly more technically advanced version of the 1980's Usborne Computer Guides http://www.usborne.com/catalogue/feature-page/computer-and-coding-books.aspx something that appeals to people who are more visually focussed as we are. Our plan is to keep expanding it into more volumes if there's enough interest and paid support. We're also very interested in hearing critical criticism and or praise! :) Here is a rough breakdown of our take on approaching education: 1. Tackling reading first, then slowly introducing writing when enough examples have been seen to increase confidence, because reading & writing are separate skills 2. Many small, fun examples for each thing to keep interest high: fun examples helps with motivation 3. Gradual, partial introduction of topics, in context at first: graded, less to take in at once, using the writing phase to soldify understanding, will possibly add a revision phase later 4. Not so much theory introduced before the practical has been introduced (examples-first), which gives a concrete context for the theory 5. Pictures. Some of these are visual aides, which are useful as explanations: a picture tells 1000 words, so they say 6. No assumption of previous programming experience. Almost every other guide available assume some programming 7. Smaller sections because completing small things gives a real sense of achievement, which increases motivation 8. Will be tackling how to deconstruct problems using both top down and bottom up approaches: most guides don't tackle this in a simple or basics-first way 9. It's free to read and online, so able to be discussed in public - some guides are, others aren't 10. Not so math focussed including "mathy jargon". We keep away from terms like catamorphism, lambda calculus, monad, until we need them, and these are/will be only introduced when appropriate amounts of concrete practical knowledge are present through repeated exposure to examples, so that it's obvious what is meant Thanks all! Julian - (of GetContented)

1 0

.cabal: API-compatible library versions
by Imants Cekusins 26 Jan '16

26 Jan '16

Stack may already do something like this, I don't know. Anyway, here is an idea. Currently .cabal lists version or version range. What if package info included lists of API-compatible versions of this package? Let's say one library requires v10 of packageA. Another library requires v15 of the same packageA. This is a hypothetical scenario :-P However v10, v15 are API-compatible as far as package maintainer knows. Package info specifies v10 and v15 as API-compatible too. What if it were possible to issue: cabal install packageA .. and then if v10, v15 are compatible, v15 is returned without warnings. If v10 and v15 are not compatible, cabal would warn. Basically instead of library clients needing to test multiple library versions or requiring 1 exact version, deps could be specified as 1 version (with which development took place) and cabal (with hints from package maintainers) would figure this out. It might even be possible to fine-tune it to check modules & symbols actually used by library consumer app. ?

3 6

Local types
by M Farkas-Dyck 26 Jan '16

26 Jan '16

I often wish to be able to define local types and instances, so, for example: import Data.Set as Set ordNubBy :: (a -> a -> Ordering) -> [a] -> [a] ordNubBy cmp = go Set.empty where newtype T = T a instance Ord T where T x `compare` T y = cmp x y go _ [] = [] go ys (x:xs) = bool (x:) id (T x ∈ ys) $ go (Set.insert (T x) ys) xs The notion is that type and instance declarations would become legal in `let` or `where` bonds; they would effectively declare new types and instances in that scope for each use of that term; and such local types and instances could never leave the scope they were defined in, by enforcement, so consistency of instances would be preserved. I know some related work which proposes local instances [0] but none i know proposes also local types. I'm seeking some feedback here before i formalize further and potentially start making implements. Thoughts? Questions? Better ideas? Critical flaws? [0] http://okmij.org/ftp/Haskell/TypeClass.html#local

4 8

[Template Haskell]
by Dominik Bollmann 25 Jan '16

25 Jan '16

Hi all, I'm just getting my feet wet with template haskell, and I tried to write a tmap function which maps a function over the ith component of an n-tuple (which uses a slightly different approach than the given version on the TH wiki): -- | Selects the ith component of an n-tuple tsel :: Int -> Int -> ExpQ -- n-tuple a -> a tsel i n = [| \t -> $(caseE [| t |] [alt]) |] where alt = match (tupP pats) body [] pats = map varP xs xs = [ mkName ("x" ++ show k) | k <- [1..n] ] body = normalB . varE $ xs !! (i-1) -- | Maps a function over the ith component of an n-tuple tmap :: Int -> Int -> ExpQ -- :: (a -> b) -> n-tuple -> n-tuple tmap i n = do f <- newName "f" t <- newName "t" lamE [varP f, varP t] $ [| let prefix = map extract [1..(i-1)] new = $f ($(tsel i n) $t) suffix = map extract [(i+1)..n] extract k = $(tsel k n) t in tupE $ prefix ++ [new] ++ suffix |] However, this code results in the following error: Sandbox.hs:26:29: Stage error: ‘k’ is bound at stage 2 but used at stage 1 … In the splice: $(tsel k n) In the Template Haskell quotation [| let prefix = map extract [1 .. (i - 1)] new = $f ($(tsel i n) ($t)) suffix = map extract [(i + 1) .. n] extract k = $(tsel k n) $t in tupE $ prefix ++ [new] ++ suffix |] Compilation failed. Could anyone explain to me what stage 2 and stage 1 refer to, and further, what the logical flaw in the above snippet is? What exactly is wrong with line `extract k = $(tsel k n) $t' ? Thanks! Dominik.

2 1

Re: [Haskell-cafe] Local types
by Johannes Waldmann 25 Jan '16

25 Jan '16

> I often wish to be able to define local types and instances ... https://mail.haskell.org/pipermail/haskell-cafe/2014-October/116291.html

3 2

ANN: toysolver 0.4.0 released
by Masahiro Sakai 25 Jan '16

25 Jan '16

I'm announcing the release of the toysolver package, version 0.4.0. The toysolver provides solver implementations of various problems including SAT, SMT, Max-SAT, PBS (Pseudo Boolean Satisfaction), PBO (Pseudo Boolean Optimization), MILP (Mixed Integer Linear Programming) and non-linear real arithmetic. The highlight of this release is the introduction of SMT (Satisfiablity Modulo Theories) solver 'toysmt'. At the moment, toysmt is very experimental and only supports the theory of uninterpreted functions and the theory of linear real arithmetic. http://hackage.haskell.org/package/toysolver https://github.com/msakai/toysolver/releases/tag/v0.4.0 Thanks, -- Masahiro Sakai

2 2

trying to build Hackage hoq-0.3
by Vasili I. Galchin 25 Jan '16

25 Jan '16

1) ghc version 7.6.3 2) cabal version 1.16.0.2 3) when i run "cabal install", I receive on stdout : Resolving dependencies... Configuring readline-1.0.3.0... checking for gcc... gcc checking for C compiler default output file name... a.out checking whether the C compiler works... yes checking whether we are cross compiling... no checking for suffix of executables... checking for suffix of object files... o checking whether we are using the GNU C compiler... yes checking whether gcc accepts -g... yes checking for gcc option to accept ISO C89... none needed checking for GNUreadline.framework... checking for readline... no checking for tputs in -lncurses... no checking for tputs in -ltermcap... no checking for tputs in -lcurses... no checking for rl_readline_version... no configure: error: readline not found, so this package cannot be built See `config.log' for more details. Failed to install readline-1.0.3.0 cabal: Error: some packages failed to install: hoq-0.3 depends on readline-1.0.3.0 which failed to install. readline-1.0.3.0 failed during the configure step. The exception was: ExitFailure 1 4) I googled Haskell and readline and I found the same error with other Haskell packages like Twitter. How do I resolve my build readline problem? Vasya

2 3

[ANN] nano-erl
by amindfv＠gmail.com 23 Jan '16

23 Jan '16

I've written a tiny library for Erlang-style actor semantics in Haskell: hackage.haskell.org/package/nano-erl It's fast, and meant to be a lightweight abstraction that's easy to integrate with existing code. Enjoy! Tom

1 0

question on the design of protohaskell (in Write You a Haskell)
by Maurizio Vitale 23 Jan '16

23 Jan '16

I know this part has been left incomplete when Stephen got a real job, but I have a question. In http://dev.stephendiehl.com/fun/007_path.html, compiler steps are piped together using the Kleisli (>=>) operator, presumably the one from Control.Monad. they also return some flavour of AST (typically Syn.Module or Core.Module). Assuming that the >=> is the one from Control.Monad, what is the purpose of this returned value (all the steps are in a compilerMonad and presumably would have to update the state w/ the new AST anyhow)? Or is the intention to have a special >=> that also updates the state (but I wouldn't know how to deal w/ Syn.Core vs Core.Module) I would understand some form of phantom type to ensure that compilation steps are in the right order, but as it is I'm puzzled. Anybody has some insight? Thanks, Maurizio

2 1