- Beginners - Haskell.org

Re: [Haskell-beginners] This year's Google Summer of Code
by Shishir Srivastava 28 Apr '15

28 Apr '15

Brilliant work ! Congrats to ever so helpful Sumit Sahrawat ! Cheers, Shishir

2 1

This year's Google Summer of Code
by Kim-Ee Yeoh 28 Apr '15

28 Apr '15

will see one of the actively helpful members of this list, Sumit Sahrawat, help contribute to making a full-featured Computer Algebra System an open-source reality. He will take the IHaskell package (inspired by IPython) one step closer to feature parity with Mathematica: https://www.google-melange.com/gsoc/project/details/google/gsoc2015/sumit_s… Sumit: Congrats for bagging GSOC ! -- Kim-Ee

1 0

Partition a number or optimize
by Animesh Saxena 27 Apr '15

27 Apr '15

I have the following function to optimize myfunction = max(p `div` c + p `mod` c) Here p is an array of numbers and c is also an array of numbers I need to find c’s which minimize the value of my function. for example if p = [1,10000]. c = 2,8 or c = 1,9 or c=5,5 number of c is length of p which is 2 (in this case). Sum of c is given as a constraint, so for example if sum is 10, then c’s can be 2,8 or 1,9 or 5,5 or 4,6 or…..etc etc. Length of c will be equal to length of p array. Hope the problem statement is clear In haskell terms it will be, let my_min p c = foldl1 max $ zipWith (\p c -> (p `div` c) + (p `mod` c)) p c One way is to use various combinations of c using integer partitioning. Now i wrote a crude parts function to do this nparts 0 = [] nparts n = [n] : [x:xs | x <- [1..n`div`2], xs <- nparts(n - x)] which is a bad bad example of recursion coz it’s slow! *Main> nparts 5 [[5],[1,4],[1,1,3],[1,1,1,2],[1,1,1,1,1],[1,2,2],[1,2,1,1],[2,3],[2,1,2],[2,1,1,1]] I can then filter this with length of p and get the required options of c I can then call the function “ my_min” with this list (simple map application) Problem is if i do *Main> filter (\x->length x == 2) $ nparts 100 [[1,99] and it hangs…for a lot of time….so time constraint? Any other approaches to this problem to make the partitions speedier. I know it’s an optimization problem…but all non linear optizers will use floats…and fail to converge…this is an integer optimizing problem. Any ideas??

3 3

How to "add column" to a Table?
by martin 27 Apr '15

27 Apr '15

Hello all, I suppose the natural representation for a Table is a List of Tuples. Or more generally I'll probably end up with a type "Table a" where a stands for the Columns. I cannot figure out how to add a new Column to a Table. What would be the type of such an operation? If I just want to add an empty Column (which what an RDBMS would do), would I do something like this? addColumn :: ColumnDef -> Table a -> Table b But what is this ColumnDef? It would need to contain the name and type of the new column. To implement this operation I would have to somewehre add a column to something. But what is the type of this something. I don't see how I can add a component to a Tuple in a generic way, or can I? I browsed some code on Hackage and found code which uses Data.Proxy. It this the right direction? Or do I need to look at GADTs, or template haskell? Or am I overcomplicating things, and there is a true haskellish solution?

3 4

fmap Maybe
by Shishir Srivastava 27 Apr '15

27 Apr '15

Hi, Please can someone explain why these two expression behave differently - ---- fmap (\x -> x) Just 2 Just 2 ----- fmap (\x -> x+2) Just 2 *No instance for (Num (Maybe a0)) arising from a use of `it'* *In a stmt of an interactive GHCi command: print it* ---- The first expression evaluates fine whereas the second one fails. However if I do - ---- fmap (\x -> x+2) $ Just 2 Just 4 ---- Then the second expression also returns the Maybe value. Why is $ needed in second expression but not in the first one ? Thanks, Shishir

3 2

IO and purity
by Shishir Srivastava 26 Apr '15

26 Apr '15

Hi, Can someone please explain how IO operations do not fit in the pure category of mathematical function in that they have to be implemented via Monads. For e.g. the getLine function has the type IOString and it reads the input from the user. Now as I see it the output of getLine will always be same if the input remain same (i.e. for input "X" getLine will always return "X" ) which is the constraint on mathematical functions. Therefore I don't see why monads are necessary for implementing IO in pure languages. I can understand why Date and Random functions have be implemented via monads because their output will always change. Thanks, Shishir

7 8

missing parallelism
by Maurizio Vitale 25 Apr '15

25 Apr '15

G'day, I have a test code that I compile with ghc -threaded -eventlog -rtsopts --make parallel.hs and run with ./parallel 2 +RTS -ls -N4 on a laptop with 4 physical cores. I would expect activities in two threads, but threadscope shows only one active thread. Can somebody explain me the program's behaviour? Thanks a lot Maurizio {-# LANGUAGE UnicodeSyntax #-} {-# LANGUAGE TupleSections #-} import Control.Applicative import Control.Concurrent.Async (async, waitAny, Async) import Data.List (delete, sortBy) import Data.Ord (comparing) import System.CPUTime import System.Environment import GHC.Conc (numCapabilities) concurrentlyLimited :: Int -> [IO a] -> IO [a] concurrentlyLimited n tasks = concurrentlyLimited' n (zip [0..] tasks) [] [] concurrentlyLimited' ∷ Int -- ^ number of concurrent evaluations → [(Int, IO b)] -- ^ still to run (ordered by first element) → [Async (Int,b)] -- ^ currently running → [(Int,b)] -- ^ partial results (ordered by first element) → IO [b] concurrentlyLimited' _ [] [] results = return . map snd $ sortBy (comparing fst) results concurrentlyLimited' 0 todo ongoing results = do (task, newResult) <- waitAny ongoing concurrentlyLimited' 1 todo (delete task ongoing) (newResult:results) concurrentlyLimited' _ [] ongoing results = concurrentlyLimited' 0 [] ongoing results concurrentlyLimited' n ((i, task):otherTasks) ongoing results = do t <- async $ (i,) <$> task concurrentlyLimited' (n-1) otherTasks (t:ongoing) results euler :: Int → Int euler n = length (filter (relprime n) [1..n-1]) where relprime x y = gcd x y == 1 sumEuler :: Int → Int sumEuler = sum . (map euler) . mkList where mkList n = [1..n-1] p ∷ IO Int p = return $ sumEuler 3000 numThreads ∷ [String] → IO Int numThreads [] = return numCapabilities numThreads [cores] = return $ read cores main ∷ IO() main = do threads ← getArgs >>= numThreads putStrLn $ "Running up to " ++ show threads ++ " threads in parallel (on " ++ show numCapabilities ++ " cores)" startTime ← getCPUTime f ← concurrentlyLimited threads $ replicate 10 p endTime ← getCPUTime putStrLn $ foldr ((++) . show ) "" f putStrLn $ "Evaluation took " ++ show (fromIntegral (endTime - startTime) / 1000000000000∷Double)

2 5

Re: [Haskell-beginners] missing parallelism
by Grzegorz Milka 25 Apr '15

25 Apr '15

As far as I can tell there's nothing wrong with your code. My hypothesis is that Haskell optimizes call to sumEuler 5000 by calling it only once in one thread. Here's why I think so: The program I used for debugging this is: import Control.Concurrent.Async (async, wait) import System.IO.Unsafe (unsafePerformIO) sumEuler :: Int -> Int sumEuler = sum . map euler . mkList where mkList n = seq (unsafePerformIO (putStrLn "Calculating")) [1..n-1] euler n = length (filter (relprime n) [1..n-1]) where relprime x y = gcd x y == 1 p :: Int -> IO Int p b = return $! sumEuler b p5000 :: IO Int p5000 = return $ sumEuler 5000 main :: IO () main = do a <- async $ p5000 b <- async $ p5000 av <- wait a bv <- wait b print (av,bv) The two main modification are adding a debugging message "Calculating" when a list [1..(n-1)] is evaluated. The second one is making p into a function. Notice that it uses strict application ($! - check how it works with simple $). I will use that function in further examples. Running this program with "time ./Test 2 +RTS -ls -N2" gives me: Calculating (7598457,7598457) real 0m3.752s user 0m3.833s sys 0m0.211s Just to be sure I have almost the same time when doing only one computation with: main :: IO () main = do a <- async $ p5000 av <- wait a print av So it seems like the value returned by p5000 is computed only once. GHC might be noticing that p5000 will always return the same value and might try cache it or memoize it. If this hypothesis is right then calling sumEuler with two different values should run in two different threads. And indeed it is so: main :: IO () main = do a <- async $ p 5000 b <- async $ p 4999 av <- wait a bv <- wait b print (av,bv) Gives: Calculating Calculating (7598457,7593459) real 0m3.758s user 0m7.414s sys 0m0.064s So it runs in two threads just as expected. The strict application ($!) here is important. Otherwise it seems that the async thread returns a thunk and the evaluation happens in print (av, bv) which is evaluated in a single thread. Also the fact that p5000 is a top level binding is important. When I do: main :: IO () main = do a <- async $ p 5000 b <- async $ p 5000 av <- wait a bv <- wait b print (av,bv) I get no optimization (GHC 7.6.3). Best, Greg

1 0

Strange "Not in scope" error message
by Matthew Moppett 24 Apr '15

24 Apr '15

I have a file with the following relevant lines: import qualified Control.Concurrent as C ... isCancelled = C.isEmptyMvar . stopConcurrentProcess Trying to compile this or run it in ghci yields a puzzling error message: Not in scope: `C.isEmptyMvar' Perhaps you meant one of these: `C.isEmptyMVar' (imported from Control.Concurrent), `C.isEmptyChan' (imported from Control.Concurrent) Any idea what could be going wrong?

3 3

Re: [Haskell-beginners] Error when running `cabal install cabal-install`
by Peng Yu 23 Apr '15

23 Apr '15

On Thu, Apr 23, 2015 at 7:39 PM, Ryan Schmidt <ryandesign(a)macports.org> wrote: > >> On Apr 23, 2015, at 7:23 PM, Peng Yu <pengyu.ut(a)gmail.com> wrote: >> >>> Long story short: Try >>> $> cabal install --extra-include-dirs=/opt/local/include --extra-lib-dirs=/opt/local/lib >> >> I tried it but it still shows the error messages. >> >> ~$ cabal install --extra-include-dirs=/opt/local/include >> --extra-lib-dirs=/opt/local/lib cabal-install >> >> Resolving dependencies... >> Configuring cabal-install-1.22.3.0... >> Failed to install cabal-install-1.22.3.0 >> Build log ( /Users/py/.cabal/logs/cabal-install-1.22.3.0.log ): >> cabal: Error: some packages failed to install: >> cabal-install-1.22.3.0 failed during the configure step. The exception was: >> user error ('/usr/bin/ghc' exited with an error: > > You've installed ghc into /usr/bin? That's probably not a good idea. Prefix /usr is only for the OS vendor to install software into. This is what installed from the haskell installer. I agree this might not be a good idea. But I have no control over it. -- Regards, Peng

1 0