- Beginners - Haskell.org

ghci and randomRs
by Jeff C. Britton 16 Sep '14

16 Sep '14

I am trying to use randomRs at the ghci prompt like so Prelude System.Random> take 10 $ (randomRs (1, 6) newStdGen) but I get the following error <interactive>:16:12: Could not deduce (RandomGen (IO StdGen)) arising from a use of `randomRs' from the context (Random a, Num a) bound by the inferred type of it :: (Random a, Num a) => [a] at <interactive>:16:1-37 In the second argument of `($)', namely `(randomRs (1, 6) newStdGen)' In the expression: take 10 $ (randomRs (1, 6) newStdGen) In an equation for `it': it = take 10 $ (randomRs (1, 6) newStdGen) I have tried a variety of options, like wrapping it in a "do" or adding type annotations. Nothing seems to work. -----Original Message----- From: Beginners [mailto:beginners-bounces@haskell.org] On Behalf Of martin Sent: Friday, September 12, 2014 10:05 AM To: The Haskell-Beginners Mailing List - Discussion of primarily beginner-level topics related to Haskell Subject: Re: [Haskell-beginners] How to add a "method" to a record Am 09/10/2014 08:50 PM, schrieb Corentin Dupont: > If the field "label" can be deduced from "payload", I recommend not to > include it in your structure, because that would be redundant. > > Here how you could write it: > > data Foo pl = Foo { payload :: pl} > > labelInt :: Foo Int -> String > labelInt (Foo a) = "Int payload:" ++ (show a) > > labelString :: Foo String -> String > labelString (Foo a) = "String payload" ++ a > > You are obliged to define two separate label function, because "Foo Int" and "Foo String" are two completly separate types. This is exactly my problem: Someone will use this type an define the type of pl. How can I know what type she'll use? What I'd like to express is that whoever creates a concrete type should also provide the proper label function. > > On Wed, Sep 10, 2014 at 2:06 PM, martin <martin.drautzburg(a)web.de> wrote: > > Hello all > > if I have a record like > > data Foo pl = Foo { > label :: String, > payload :: pl > } > > how can I create a similar type where I can populate label so it is not a plain string, but a function which > operates on > payload? Something like > > label (Foo pl) = show pl > _______________________________________________ Beginners mailing list Beginners(a)haskell.org http://www.haskell.org/mailman/listinfo/beginners

4 5

What am I reinventing here
by Dániel Arató 16 Sep '14

16 Sep '14

Hey guys, I was working through 99 Haskell Problems a while back and while refactoring my solution to Problem 9 I couldn't help feeling I was reinventing something basic. My thinking was, this problem can be generalized to letting a function recursively process arbitrary chunks of a given data structure in several passes until the whole data is consumed while some function is extracting some information in an accumulator. So I wrote a mini-typeclass like so: class Exhaustible e where isEmpty :: e -> Bool instance Exhaustible [a] where isEmpty = null exhaust :: Exhaustible e => (e -> (b, e)) -> (a -> b -> a) -> a -> e -> a exhaust f g z xs | isEmpty xs = z | otherwise = let (val, rest) = f xs in exhaust f g (g z val) rest Now the solution to Problem 9 is easy to define in terms of "exhaust". Then I thought, "Why am I treating empty as a special value again?", and wrote this instead: stabilize :: Eq a => (a -> (b, a)) -> (c -> b -> c) -> c -> a -> (c, a) stabilize f g z d | d == d' = (z', d') | otherwise = stabilize f g z' d' where (val, d') = f d z' = g z val This looks both really useful and extremely basic to me, so I'm wondering what the standard way to do the same thing might be. I looked at Foldable + foldr but the idea there seems to be to process a data structure _one_ element at a time. So my question is, is there an official implementation for this combinator I called "stabilize"? Also, any tips on improving the definitions and comments on how situations like Problem 9 should be approached are welcome. Daniel

2 1

How to add a "method" to a record
by martin 12 Sep '14

12 Sep '14

Hello all if I have a record like data Foo pl = Foo { label :: String, payload :: pl } how can I create a similar type where I can populate label so it is not a plain string, but a function which operates on payload? Something like label (Foo pl) = show pl

5 5

Re: [Haskell-beginners] [Haskell-cafe] 2014 haskell cabal update hangs on mac
by Yitzchak Gale 09 Sep '14

09 Sep '14

Daniel Gorín wrote: > I had the same problem. It turned out to be that for some unknown reason I ended up with: > > /Library/Haskell/ghc-7.8.3-x86_64/bin/cabal -> cabal.wrap > /Library/Haskell/ghc-7.8.3-x86_64/bin/cabal.real -> cabal.wrap > /Library/Haskell/ghc-7.8.3-x86_64/bin/cabal.wrap > > where cabal.wrap is a shell script that checks if a new cabal.config > should be created and then calls… cabal.real. Replacing cabal.real with a binary version of cabal solves it. > > I don’t know why this happened, I suspect that it might occur if you had ghc 7.8.2 already installed... I believe this is a bug in the scripts activate-hs and uninstall-hs provided by the Haskell Platform installer for Mac OS X. I have reported it here: https://github.com/haskell/haskell-platform/issues/147 -Yitz

1 0

hpc doctest
by ryan 08 Sep '14

08 Sep '14

Hello, Are there any known issues with doctest and hpc? Cabal test runs correctly, but hpc doesn't seem to see anything. I have tried to make some of the test fail on purpose so see it hpc would pick it up, since there is additional output from cabal test when there is a failure, but it doesn't help. -- -- test file -- module Doctests where import Test.DocTest main :: IO () main = doctest ["-isrc", "src/Ch1.hs"] -- -- command line -- $ cabal test Building DiscreteMathematicsUsingAComputer-0.1.0.0... Preprocessing library DiscreteMathematicsUsingAComputer-0.1.0.0... In-place registering DiscreteMathematicsUsingAComputer-0.1.0.0... Preprocessing test suite 'doctests' for DiscreteMathematicsUsingAComputer-0.1.0.0... Warning: output was redirected with -o, but no output will be generated because there is no Main module. Running 1 test suites... Test suite doctests: RUNNING... Test suite doctests: PASS Test suite logged to: dist/test/DiscreteMathematicsUsingAComputer-0.1.0.0-doctests.log Writing: hpc_index.html Writing: hpc_index_fun.html Writing: hpc_index_alt.html Writing: hpc_index_exp.html Test coverage report written to dist/hpc/html/doctests/hpc_index.html 1 of 1 test suites (1 of 1 test cases) passed. Writing: hpc_index.html Writing: hpc_index_fun.html Writing: hpc_index_alt.html Writing: hpc_index_exp.html Package coverage report written to dist/hpc/html/DiscreteMathematicsUsingAComputer-0.1.0.0/hpc_index.html Ryan

1 0

Check constructor's field numeric value at compile time
by Dmitriy Matrosov 08 Sep '14

08 Sep '14

Hi. I have a type data Interval = Seconds Float | MicroSeconds Int The Float field of data constructor Seconds should be >= 1, and Int field of constructor MicroSeconds should be in the range from 0 to 1000000. How can i write this constraints so they're checked at compile time, not at runtime? -- Dmitriy Matrosov

6 7

Continuations vs. Values
by martin 07 Sep '14

07 Sep '14

Hello all, in the continuation monad, I can turn a value into a Cont by means of return. Conversely I can convert a Cont into a value by passing id as the continuation. This dualism of Cont and plain values is well known. The bind operator kind of(!) takes a value a and produces a new value b. This is all wrapped in Cont stuff, so the result of bind is not really a plain a, but a Cont which takes a function from b -> r. But this is just the same dualism as above. Most Cont examples I've seen could much more easily be implemented with just values or calling functions on values. Then again there is "a poor man's concurrency" which I don't get at all. In between there is a huge gap of (my) understanding. Can someone help me cross this gap? There must be certain things I can do with Conts which I cannot do easily with values and functions.

2 1

Haskeline and forkIO
by Jeff C. Britton 03 Sep '14

03 Sep '14

I am trying to modify an example in RealWorldHaskell from Chapter 24. The example is the first code snippet labeled -- file: ch24/Compressor.hs I am trying to replace the use of Readline with Haskeline. In my code the forkIO thread does not run. I guessed that since the result of the worker thread was thrown away that perhaps laziness was the problem. So, I attempted to use `seq`, but that does not work either. I am able to run the RealWorldHaskell example. I am using GHC 7.8.3. I have tried runhaskell with and without the -threaded option and on both Linux and Windows 7. import Control.Concurrent (forkIO) import Control.Exception import qualified Data.ByteString.Lazy as L import System.Console.Haskeline hiding (handle) -- Provided by the 'zlib' package on http://hackage.haskell.org/ import Codec.Compression.GZip (compress) -- Read the file, compress the data, write the compressed data worker :: FilePath -> IO () worker path = L.readFile path >>= L.writeFile (path ++ ".gz") . compress -- Run the worker on a new thread runWorker :: FilePath -> IO() runWorker path = handle (print :: SomeException -> IO ()) $ do forkIO (worker path) return () loop :: InputT IO () loop = do maybeLine <- getInputLine "Enter a file to compress> " case maybeLine of Nothing -> return () -- user entered EOF Just "" -> return () -- treat no name as "want to quit" Just path -> let f = runWorker path in f `seq` do return f loop main = runInputT defaultSettings loop

4 6

Does Haskell work on Free BSD?
by KC 03 Sep '14

03 Sep '14

Does Haskell work on Free BSD? If not which is the better Linux to use? -- -- Sent from an expensive device which will be obsolete in a few months! :D Casey

2 1

Stackage latest snapshot
by Jeremy 03 Sep '14

03 Sep '14

I would like to point my cabal file to the latest Stackage snapshot. I seem to recall that there is a link which will automatically resolve to this, but I can no longer find it. Does this link still exist? Is there a good reason for not doing so? I understand that this will introduce a little instability, but I'm OK dealing with that in return for not getting stuck on the same package versions forever.

3 4