- Haskell-Cafe - Haskell.org

Need ideas how to model the lack of something
by martin 17 Dec '15

17 Dec '15

Hello all, here is a problem where I did not manage to find a suitable abstraction. The main idea goes like this: a List (and many other containers) can be seen as something containing "stuff". There is a function (:) that unconditionally adds an element to the container and returns a new container Now suppose the container has the possiblility to refuse having another element added to it, e.g. because it has only limited "space". In that case the corresponding function would have a signature of insert :: a -> C a -> Maybe (C a). If an item can successfully be added, then the returned container will be less space avaiable. I'd like stuff and space to be symmetrical (maybe there lies the first flaw, because I can enumerate the elements, but I cannot enumerate the space). A symmetry like electrones and holes. I started like this data C a = C { insert :: a -> Maybe (C a), remove :: Maybe (a, C a) } but I could not implement anything sensible on top of this. I'd be happy to hear any comments on this, including loud thinking and random ramblings.

10 14

Haskell safe ccalls
by Valentino Picotti 17 Dec '15

17 Dec '15

Hi everybody, I’m new on this list and have a newbie question about Haskell FFI and safe calls. I have written the following lines of C code that rely on the RTS: Condition* init_cond(){ Condition* c = malloc(sizeof(Condition)); initCondition(c); return c; } Mutex* init_mutex(){ Mutex* m = malloc(sizeof(Mutex)); initMutex(m); return m; } void wait_condition(Condition* condition, Mutex* mutex) { ACQUIRE_LOCK(mutex); if(!waitCondition(condition, mutex)){ fprintf(stderr, "A: Wait failed!"); }; RELEASE_LOCK(mutex); } And the Haskell program is written as follow: data Condition = Condition data Mutex = Mutex foreign import ccall "init_cond" init_cond :: IO (Ptr Condition) foreign import ccall "init_mutex" init_mutex :: IO (Ptr Mutex) foreign import ccall "wait_condition" wait_condition:: Ptr Condition -> Ptr Mutex -> IO () main :: IO () main = do c <- init_cond m <- init_mutex wait_condition c m putStrLn “Hello!” The behaviour that i want from my application is the same as the following C program: int main(int argc, char** argv){ pthread_cond_t c = PTHREAD_COND_INITIALIZER; pthread_mutex_t m; pthread_mutex_init(&m, 0); pthread_mutex_lock(&m); pthread_cond_wait(&c, &m); printf(“I didn’t block!\n"); pthread_mutex_unlock(&m); return 0; } What I expect is that the program doesn’t reach the putStrLn line, but it prints “Hello!” and terminates with exit code 0. I can’t figure out what’s going on under the hood. Can someone explain me what i’m missing? (I’m familiar with Capabilities, Tasks, Haskell threads) Thanks, Valentino P.

1 0

Hyperlinked source docs
by Johannes Waldmann 16 Dec '15

16 Dec '15

As I said, I admire hyperlinked sources in the new "base" documentation https://hackage.haskell.org/package/base-4.8.1.0/docs/ and I want to know how I can produce such documents. What is the magic haddock/hscolour invocation? (I actually use standalone-haddock.) On hackage, not every package gets the special treatment? (In fact, only "base" does?) - J.W.

5 4

IHaskell trouble, after upgrading to Haskell Platform 7.10.3?
by David Banas 16 Dec '15

16 Dec '15

Hi all, Are there any other IHaskell users out there, who are having trouble re-installing IHaskell, after upgrading to Haskell Platform 7.10.3? I’m getting this, at the tail end of the actual ihaskell build: [24 of 25] Compiling IHaskell.Eval.Info ( src/IHaskell/Eval/Info.hs, dist/build/IHaskell/Eval/Info.o ) [25 of 25] Compiling IHaskell.Publish ( src/IHaskell/Publish.hs, dist/build/IHaskell/Publish.o ) In-place registering ihaskell-0.8.3.0... Preprocessing executable 'ihaskell' for ihaskell-0.8.3.0... main/Main.hs:29:18: Could not find module ‘Data.String.Here’ It is a member of the hidden package ‘here-1.2.7@here_63mTmLx8Sz20TyEJBVCnnC’. Perhaps you need to add ‘here’ to the build-depends in your .cabal file. Use -v to see a list of the files searched for. Updating documentation index /Users/dbanas/Library/Haskell/share/doc/x86_64-osx-ghc-7.10.3/index.html cabal: Error: some packages failed to install: ihaskell-0.8.3.0 failed during the building phase. The exception was: ExitFailure 1 Thanks, and Happy Holidays! -db

2 2

Parallel Haskell on RPi2
by spacestation＠venussociety.org 16 Dec '15

16 Dec '15

Hello to everone, Did anybody manage to install parconc-examples on a Raspberry Pi2 without the use of --force-reinstalls. As it would break to many packages on my end.I tried it the way describe from Simon Marlow s book "Parallel and concurrent programming in haskell". Or is there any option within cabal to seperate between new,new version and reinstall packages listed with pi@opencvrpi2:~/parconc-examples-master$ cabal install --only-dependencies I m using ghc 7.4.1 Thanks for any feedback Gottfried F. Zojer

1 0

sum $ map f xs ... ghc-7.10 performance regression?
by Johannes Waldmann 15 Dec '15

15 Dec '15

Is there a difference between "sum $ map f xs" and "sum $! map f xs"? I would think no, since "$!" just evaluates the argument to WHNF, so it would evaluate just the first cons of the list. This little program main = print $ sum $ map bitcount [0, 4 .. 2^24-1 ] bitcount :: Int -> Int bitcount x = if x > 0 then let (d,m) = divMod x 2 in bitcount d + m else 0 runs in 1.6 seconds when compiled with -O2 on ghc-7.6.3, ghc-7.8.4, but takes 3.2 seconds on ghc-7.10.[1-3]. Why? when I change the main function (note: $ => $!) to main = print $ sum $! map bitcount [0, 4 .. 2^24-1 ] and compile with 7.10.[1-3], then it also runs in 1.6 seconds. (I can write a bug report but I want to check whether I am missing something obvious) - J.W.

5 8

Does Backwards admit a Monad instance?
by David Feuer 15 Dec '15

15 Dec '15

In particular, I'm thinking about instance MonadFix m => Monad (Backwards m) where m >>= f = Backwards $ do fin <- forwards (f int) int <- forwards m return fin It looks to me like this should be valid, and compatible with the Applicative instance, but maybe I'm missing something.

1 1

Re: [Haskell-cafe] Trouble installing ghc on CentOS 6
by Johannes Waldmann 15 Dec '15

15 Dec '15

I think that "--with-gmp-libraries" states the location of libgmp that the installed ghc will use when linking programs. Your error indicates that a .so file referenced from the executable (contained in the binary package) "utils/ghc-pwd/dist-install/build/tmp/ghc-pwd" could not be resolved. $ ldd utils/ghc-pwd/dist-install/build/tmp/ghc-pwd if it says ... libgmp.so.3 => not found then set LD_LIBRARY_PATH or check values in /etc/ld.so.conf and run /sbin/ldconfig - J.W.

2 1

Trouble installing ghc on CentOS 6
by Lyle Kopnicky 14 Dec '15

14 Dec '15

Hi folks, I'm trying to install the 64-bit generic Linux distribution of ghc-7.10.3 on CentOS 6, as a non-privileged user in my home dir. I've unpacked the files, and I ran ./configure, but it told me it couldn't find libgmp.so.3. I searched for and found that library on the system. Looking at the help for ./configure, I saw the --with-gmp-libraries option, so I tried again with that, but it still can't seem to find the file: $ ./configure --prefix=/home/kopnicky/ghc --with-gmp-libraries=/usr/lib64 checking for path to top of build tree... utils/ghc-pwd/dist-install/build/tmp/ghc-pwd: error while loading shared libraries: libgmp.so.3: cannot open shared object file: No such file or directory configure: error: cannot determine current directory Strange, because that library is definitely there in /usr/lib64. Any thoughts on what I can try next? Thanks, Lyle

1 0

a problem I can't solve simply
by Dennis Raddle 14 Dec '15

14 Dec '15

I have a problem that doesn't seem hard to state but I can't seem to solve without a bunch of complex code. This relates to my musical score playback. In using it to play music, I don't always want to play back the entire source musical document, but rather play a range of measures. So I might give a command to my app like "play 1-3" which means play measures 1 through 3. There is a time saving feature, which is that I can type "play 10" which means start the playback at measure 10 and continue until the first occurrence of two empty measures. This is a common use case. So I have to write a function that takes a start measure and computes the end measure by scanning for two empty measures. Let's say for simplicity's sake that we'll forget about "measures" and just say that notes have a start time and end time, which will be integers. type Note = (Int,Int) A musical score can have several individual staves (notes for individual instruments), so it will look like this: type Staff = [Note] type Score = [Staff] I need to write a function as follows computeEndMsr :: Int -> Score -> Int computeEndMsr beginMsr score = ... Some examples: Here's a score with just one staff, to give you an idea. score1 = [ [(1,3), (2,4), (7,10)] ] -- In the following case a two-unit gap is found at units 5 and 6. computeEndMsr 1 score1 = 4 computeEndMsr 5 score1 = should throw an error indicating that a gap was found immediately and no actual notes were included -- In the following case, the maximum unit of any note is 10, so that is what is computed computeEndMsr 6 score1 = 10 -- This case illustrates how it's okay if the computed end measure is equal to the begin msr computeEndMsr 10 score1 = 10 computeEndMsr 11 score1 = should throw an error indicating that the given begin msr is past the end of any note in the score This example has only one staff, but a score can have multiple staves. Also the timing and duration of notes can overlap, either on one staff or across multiple staves.

2 2