June 2015 - Beginners

lens-aeson and error messages
by Thomas Koster 14 Jun '15

14 Jun '15

Hi list, tl;dr - I am using aeson and have an unusual JSON data structure to parse, so I have implemented the parseJSON function using the prisms and traversals in the lens-aeson package. Unfortunately, the only error message I ever get from my parser is "mempty". How can my parser give better error messages when using lens-aeson? Consider a need to implement parseJSON for the following type. data Row = Row [Text] Int Text The twist is that the values are in a JSON list, not an object. [["a","b","c"], 4, "a title"] The JSON above should be decoded to the following Haskell value. Row ["a", "b", "c"] 4 "a title" My first FromJSON instance was naive. instance FromJSON Row where parseJSON = withArray "Row" $ \ a -> case Vector.toList a of (x : y : z : []) -> Row <$> parseJSON x <*> parseJSON y <*> parseJSON z _ -> fail "Invalid Row." Then I decided that pattern matching a vector like that via an intermediate list was probably rookie stuff, so I tried lenses (this is my first attempt at lenses, btw). instance FromJSON Row where parseJSON v = Row <$> (v ^. nth 0 . to parseJSON) <*> (v ^. nth 1 . to parseJSON) <*> (v ^. nth 2 . to parseJSON) This looks like it works but now the only error message I get is "mempty". I think this is because the lens operators are folding with the Monoid Parser instance where mempty = fail "mempty". Can I continue to use lenses but produce better error messages? If so, how? The lens package haddocks baffle me. Any suggestions to improve my use of lenses are also welcome (e.g. the expression "to parseJSON" seems so useful that I thought it might be named in the lens-aeson package. But it isn't, so its absence makes me suspect that I'm using it incorrectly). Alternative, non-lens-based suggestions for de-constructing a Vector, pattern matching its elements, without too many intermediate lists or tuples are also welcome. Thanks in advance. -- Thomas Koster

2 2

Packages
by Mike Houghton 13 Jun '15

13 Jun '15

I’m using Cabal to build a package from some source I’m writing. It is not an executable but rather a library. How can I test locally that the package I’m making is complete? ie How do I reference the package I’ve just buiilt in a haskell source file? Thanks Mike

3 3

Escaping special characters in text
by Thomas Koster 11 Jun '15

11 Jun '15

Hi list, My program needs to escape and unescape "special characters" in text (Data.Text.Text), using my own definition of "special character" (isSpecial :: Char -> Bool). I am looking for a library that provides functions that implement or help me implement this functionality. I don't really care exactly how the special characters are escaped, but my preference is to prefix them with backslashes. While "attoparsec" does technically answer my question, it is as unimpressive an answer as "Prelude" unless the answer comes with a particularly clever and concise parser that blows my mind (and then kudos to the author). I am looking for a higher level library where I don't need to re-invent this wheel. That is, I don't want to write an unescaping parser if somebody has already published one on Hackage in a clean, well-tested library. My searches on Hoogle have turned up only network-uri, which offers percent-encoding with the definition of "special character" accepted as an argument [1]. This is the sort of thing I am after, although to use network-uri I would have to round-trip via String, something that I feel I should avoid. Functions of text types that return lazy text builders would be ideal. Also, percent-encoding is not my favourite encoding scheme. Thanks in advance. [1] https://hackage.haskell.org/package/network-uri/docs/Network-URI.html#v:esc… -- Thomas Koster

2 2

Proper use of exit{Success,Failure}
by Vale Cofer-Shabica 09 Jun '15

09 Jun '15

Dear all, I'm writing a command line program which may need to exit early based on some condition. Because of the pipeline the program is a part of, I want to be able to control the exit status when it does terminate. I can successfully use 'error' to indicate failure, but also want to exit early with success (or any other status). Here's what I've managed: >import System.Exit (exitSuccess) >import Control.Monad (when) >main = do ... code to reasonably initialize condition, e.g.: > let condition = True > when condition (putStrLn "Bailing out.")>>exitSuccess>>(return ()) ... program continues If I remove ">>(return ())" my code will not typecheck because "exitSuccess :: IO a". But this looks ugly and smells sufficiently like a kludge that I presume there is a better way to do it. Any suggestions appreciated, vale

3 3

parallel map/filter
by Csaba Marosi 09 Jun '15

09 Jun '15

Hi, I'm trying to use parallelism in Haskell, parMap in particular. My code is essentially this: import Control.Parallel.Strategies slow :: String -> (String, [Int]) fast x = (x, [0..255]) main :: IO () main = do content <- readFile "4.txt" --a 327 lines long file print $ parMap rpar slow $ lines content For the full/compiling code, see: https://github.com/csmarosi/cryptopals/blob/bf1ca794f897858f9008fb8740a1c3e… What do I miss here? The code should be trivial to run on multiple cores, but it seem to use only a single CPU core, i.e. $ ghc -O2 -v0 -threaded --make s01p04.hs slow gives: $ time ./s01p04 +RTS -N2 | wc 1 1 64028 real 0m8.903s user 0m7.888s sys 0m1.792s whereas fast is: $ time ./s01p04 +RTS -N2 | wc 1 1 320787 real 0m0.016s user 0m0.012s sys 0m0.012s Bonus question: is there a parallel filter? I use Debian sid packages, ghc version 7.8.4 Some multi-threading seems to go on, since without -N2, the sys time is below 0.1s. BTW, the slow code use ~300K heap according to valgrind. Thanks, Csabi

2 2

grouping functions together
by Dimitri DeFigueiredo 09 Jun '15

09 Jun '15

Hello! I am trying to tie together a group of functions that turn an unreliable remote network call into a reliable one. For each different network request I make, a specific group of these functions should always work together, but their type signatures are quite different. My first thought was to put them all in a typeclass: import Control.Monad class Reliable1 m req attempt ack failure result where getRequests1 :: Monad m => m [req] mkAttempt1 :: Monad m => req -> m (Maybe attempt) action1 :: Monad m => attempt -> m (Maybe ack) getAcks1 :: Monad m => [attempt] -> m [ack] mkResult1 :: Monad m => req -> ack -> m (Either failure result) run1 :: Monad m => req -> m result That doesn't work because not all functions use all parameters. For example, getAcks1 has no idea of what the final 'result' type parameter is. This lead me to my second attempt. Defining a 'service' type with the sole purpose of tying them all together. Here's my current attempt: {-# LANGUAGE MultiParamTypeClasses #-} import Control.Monad class Reliable m service where getReqs :: Monad m => service -> m [req] mkAttempt :: Monad m => service -> req -> m (Maybe attempt) action :: Monad m => service -> attempt -> m (Maybe ack) getAcks :: Monad m => service -> [attempt] -> m [ack] mkResult :: Monad m => service -> req -> ack -> m (Either failure result) run :: Monad m => service -> req -> m result data RemoteCall = RemoteCall instance Reliable IO RemoteCall where getReqs = undefined mkAttempt = undefined action = undefined getAcks = undefined mkResult = undefined run = undefined This works, but I have to explicitly pass the 'service' argument in every call. Can I avoid passing this parameter every time? Question, is there a better way to do this? I wanted to have a wrapper to make my remote calls reliable. Thanks, Dimitri

2 2

Dynamic programming (memoization) w/ string keys.
by Sourabh 08 Jun '15

08 Jun '15

I haven't tried in another language yet. I wasn't aware of nim-sum, and looked up the game of nim. And yes, this sounds like a tweak on the nim problem, so I will take a look at game theory. Thank you for the pointers. Coursera has a course on combinatorial game theory, and I've put it on my watch list now! I'll try to tweak my program to use bits instead of strings. I don't think I've ever used bit shifting in Haskell (I've used it plenty in C), so it should be interesting. Thanks for the suggestions Stefan! > Do you have a dynamic programming solution which runs reasonably fast > in another language? Because this game of kyles looks like an example > from combinatoric game theory to me and as such it might be much more > efficient to solve it using nim-sums (if possible; I haven't given this > too much thought). > > Also note that, since your strings only consist of X and I, you could > also use Integers and bit-operations instead of lists in your > algorithms. You'd still need a Map for memoization, but other operations > might be much faster. > > Cheers

1 0

(no subject)
by steve barak 08 Jun '15

08 Jun '15

Unsuscribe me from this list Steve Barak

2 1

cabal install glfw failing
by Martin Vlk 08 Jun '15

08 Jun '15

Hi, I am trying to install the Euterpea library, which depends on glfw, but it is failing to install with a compilation failure. Could someone take a look at the error and tell me what might be wrong? Many Thanks Martin The Glorious Glasgow Haskell Compilation System, version 7.10.1 cabal-install version 1.22.4.0 using version 1.22.3.0 of the Cabal library $ cabal update $ cabal install glfw Resolving dependencies... Configuring GLFW-0.5.2.2... Failed to install GLFW-0.5.2.2 Build log ( /home/martin/.cabal/logs/GLFW-0.5.2.2.log ): cabal: Error: some packages failed to install: GLFW-0.5.2.2 failed during the configure step. The exception was: user error ('/home/martin/.ghc/bin/ghc' exited with an error: /tmp/GLFW-0.5.2.2-20349/GLFW-0.5.2.2/dist/setup/setup.hs:106:14: No instance for (Applicative (StateT ConfState IO)) arising from a use of ‘modify’ In the expression: modify In the expression: modify $ \ (ConfState fs ls) -> ConfState fs (lib : ls) In an equation for ‘addLib’: addLib lib = modify $ \ (ConfState fs ls) -> ConfState fs (lib : ls) /tmp/GLFW-0.5.2.2-20349/GLFW-0.5.2.2/dist/setup/setup.hs:242:10: Could not deduce (Applicative (StateT s m)) arising from the superclasses of an instance declaration from the context (Monad m) bound by the instance declaration at /tmp/GLFW-0.5.2.2-20349/GLFW-0.5.2.2/dist/setup/setup.hs:242:10-40 In the instance declaration for ‘Monad (StateT s m)’ /tmp/GLFW-0.5.2.2-20349/GLFW-0.5.2.2/dist/setup/setup.hs:252:10: Could not deduce (Applicative (StateT s m)) arising from the superclasses of an instance declaration from the context (Monad m) bound by the instance declaration at /tmp/GLFW-0.5.2.2-20349/GLFW-0.5.2.2/dist/setup/setup.hs:252:10-47 In the instance declaration for ‘MonadState s (StateT s m)’ )

2 1

Usage of $
by Shishir Srivastava 08 Jun '15

08 Jun '15

Hi, Can someone please highlight how '$' is making the difference in the execution of this function. With $ in place the function runs fine but without $ for large input values the function overflows the stack. -------- s mp 1 = 4 s mp n = ((s mp $ n-1)^2-2) `mod` mp -------- I cannot understand the difference between (s mp $ n-1) and (s mp n-1) Thanks, Shishir Srivastava +44 (0) 750 127 5019

5 11