December 2009 - Haskell-Cafe

Instances of `IsString`.
by Jason Dusek 24 Dec '09

24 Dec '09

A quick check on Hayoo! and in my interpreter shows that there are basically no instances of `IsString`. Is it really so little used? -- Jason Dusek

5 5

ANN: Hemkay, the 100% Haskell MOD player
by Patai Gergely 24 Dec '09

24 Dec '09

Hello all, I just uploaded the fruit of a little side project. Hemkay [1] is an oldschool module music [2] player that performs all the hard work in Haskell. If there was any goal, it was to express the transformation from the song structure to the output of the mixer as a series of function compositions, maintaining a style that one might call idiomatic Haskell. Considering the dirtiness of the format in question, I'm quite pleased with the initial version. Still, I'd be curious to see how the overall quality of the code could be improved. In particular, retrieving and updating record fields is somewhat inconvenient. Also, the actual mixing (limited to the mixChunk function) is embarrassingly slow, and I wonder how much it could be improved without leaving the pure world. The program uses Portaudio for playback, but that might easily change in the future. The problem is that I couldn't get sound to work smoothly when producing samples in batches, so I'm sending them off one by one (!) at the moment, which doesn't help with performance either. I'm open to suggestions as to what library to use to push data to the sound card. Gergely [1] http://hackage.haskell.org/package/hemkay [2] http://en.wikipedia.org/wiki/MOD_(file_format) -- http://www.fastmail.fm - Or how I learned to stop worrying and love email again

4 13

Space Leak with semi-implicit parallelization and the nasty Garbage collector
by Michael Lesniak 24 Dec '09

24 Dec '09

Hello haskell-cafe (and merry christmas!), I have a strange problem with the garbage collector / memory which I'm unable to find a solution for. I think the source of my problems has to do with lazy evaluation, but currently I'm unable to find it. Using the attached program and threadscope I see that the GC is using a lot of time and the program comes (more or less) to a halt (see exa-1.png). When I increase the heap the program takes much longer and the GC is running more or less all the time (see exa-2.png). Some more detailled information: * I can see the described behaviour under both GHC 10.4 and GHC 12.1 * Linux kernel 2.6.31-16 on a dualcore * Program compiled with ghc --make -O2 -threaded -eventlog Example.hs -o exa * Started with exa +RTS -ls and one of { -N, -N1, -N2 } Any help (pointing into the right direction, mention possibly helpful blog articles or paper, constructive critic in general) is appreciated! Best wishes, Michael PS: 1. I hope the graphs are good enough to see the problem 2. Is it a known bug that threadscope eats 100%+ CPU when I just view an eventlog? ------------------------------------------------------------------------------- -- -- Minimal test example for problems with garbage collection (both under -- GHC6.10.4 and GHC6.12.1). We calculate Pi to an arbitrary length (here -- 50.000 + n) for n times using Machin's formula[1]. -- -- [1] http://en.literateprograms.org/Pi_with_Machin's_formula_(Haskell) ------------------------------------------------------------------------------- module Main where import Control.Parallel import Control.Parallel.Strategies main = do -- 20 tasks, length 50000 + [1..20] let values = implicit (initTasks 20 50000) -- Dirty trick to force evaluation: mapM_ (print . (`mod` 10)) values ------------------------------------------------------------------------------- -- Using semi-implicit parallelization here. implicit :: [PiTask] -> [Integer] implicit tasks = parMap rdeepseq (\(PiTask k) -> calcPiPure k) tasks ------------------------------------------------------------------------------- -- Definition of a task to calculate \pi to an arbitrary length and its -- implementation. data PiTask = PiTask Integer instance Show PiTask where show (PiTask i) = "PiTask <" ++ show i ++ ">" -- Returns @number@ tasks of length @len@ up to @len+number@. We add one for -- each successive task to prevent additional compiler or runtime optimizations -- through referential transparency. In dimensions of 10^5 this should not make a -- measureable difference. initTasks :: Int -> Int -> [PiTask] initTasks number len = let len' = toEnum len num' = toEnum number in map PiTask [len'..len'+num'] -- Not used here calcPi :: Integer -> IO () calcPi digits = calcPiPure digits `pseq` return () calcPiPure :: Integer -> Integer calcPiPure digits = pi' `div` (10 ^ (10 :: Integer)) where unity = 10 ^ (toInteger digits + 10) pi' = 4 * (4 * arccot 5 unity - arccot 239 unity :: Integer) arccot :: Integral t => t -> t -> t arccot x unity = arccot' x unity 0 start 1 1 where start = unity `div` x arccot' x' u sm xpower n sign | xpower `div` n == 0 = sm | otherwise = arccot' x' u (sm + sign*term) (xpower `div` (x'*x')) (n+2) (-sign) where term = xpower `div` n

2 2

Generating Haskell From a XSD
by jonathanGfischoff＠gmail.com 23 Dec '09

23 Dec '09

I would like to generate Haskell data types and xml serialization code from xsd. I know of DtdToHaskell but unfortunately I yet to be able to generate a valid dtd from my xsd. Is there a way to generate Haskell from a xsd? -Jonathan

5 7

ANNOUNCE: regions-0.1.0.1
by Bas van Dijk 23 Dec '09

23 Dec '09

Hello, As explained in my previous email, I generalized the DeviceRegion monad transformer from my usb-safe package and put it in its own package: http://hackage.haskell.org/package/regions-0.1.0.1 Hackage does not seem to build documentation at the moment so if you're interested please look at the source code which you can get using: darcs get http://code.haskell.org/~basvandijk/code/regions The main goal of regions is to allow opening of scarce resoures and operating on them but to prevent operating on closed resources. The primary technique used in this package is called "Lightweight monadic regions" which was invented by Oleg Kiselyov and Chung-chieh Shan. See: http://okmij.org/ftp/Haskell/regions.html#light-weight The package provides the monad transformer: newtype RegionT resource s (pr ∷ * → *) α And the following operation for opening scarce resources (like files, memory pointers, database connections, USB devices, etc.): open ∷ (Resource resource, MonadCatchIO pr) ⇒ resource → RegionT resource s pr (RegionalHandle resource (RegionT resource s pr)) So you give it the resource you wish to open and it yields a computation in the region monad that returns a handle to the opened resource. Note that this handle is parametrized by the region itself. Finally there's the ability to run the region: runRegionT ∷ (Resource resource, MonadCatchIO pr) ⇒ (∀ s. RegionT resource s pr α) → pr α 'runRegionT' runs the given region and finally closes all opened resources automatically. Note that the 's' is quantified over the region but not over the return value. So all values that have this 's' in their type can not be returned from the given region. Remember that the handle to a resource, which we got from 'open', was parametrized by the region and so has this 's' in it. This prevents us from returning a handle to a closed resource from this function. So it's never possible to accidentally perform I/O with a closed resource. There's also the ability to concurrently run a region inside another region using: forkTopRegion ∷ (Resource resource, MonadIO pr) ⇒ TopRegion resource s () → RegionT resource s pr ThreadId Alo note the packages: http://hackage.haskell.org/package/regions-monadsfd-0.1.0.1 http://hackage.haskell.org/package/regions-monadstf-0.1.0.1 Which defines instances for the monads classes in monads-fd and monads-tf respectively. Finally, if you want to open your own scarce resource in a region, the only thing you have to do is define an instance for: class Resource resource where data Handle resource ∷ * openResource ∷ resource → IO (Handle resource) closeResource ∷ Handle resource → IO () That's it! I'm looking forward to see instances for file Handles, memory Ptrs, database connections, etc. regards, Bas

1 0

ANNOUNCE: usb-safe-0.4.1
by Bas van Dijk 23 Dec '09

23 Dec '09

Hello, I just released usb-safe-0.4.1: http://hackage.haskell.org/package/usb-safe-0.4.1 Changes compared to 0.3 (I didn't make an announcement for 0.4): * Tested with ghc-6.12.1 and base-4.2.0.0. * It turned out that an Interface can also be regarded as a scarce resource: it needs to be claimed before use and it needs to be released when you're done with it. So just as with USB devices I had to create a new region monad that supported claiming of interfaces. So now I had two region monads: a DeviceRegion and an InterfaceRegion which both were implemented exactly the same and supported the exact same operations. So the next thing I did was to generalize the region code and put it in its own package: 'regions'. This package will be announced separately. * Removed the FilteredEndpoint type. Specific endpoints can now be retrieved from the alternate directly using: 'getEndpoints'. * Added missing language extension: OverlappingInstances (The package was unusable without it) * Added convenience functions 'setConfigWhich', 'withInterfaceWhich' and 'setAlternateWhich' for quickly setting the first setting that satisfies the given predicate on its descriptor. * Hidden the ReadEndpoint and WriteEndpoint type classes so users can't write their own instances for it and thereby potentially messing up my safety guarantees. * Lots of little refactorings (is that a real word?): documentation updates, explicit imports of all used symbols, Unicode syntax and symbols (using the base-unicode-symbols package) * Finally I packaged up some example code in: darcs get http://code.haskell.org/~basvandijk/code/usb-safe-examples/ regards, Bas

1 0

ANNOUNCE: explicit-iomodes-0.1.2
by Bas van Dijk 23 Dec '09

23 Dec '09

Hello, I just released a minor update to my explict-iomodes package: http://hackage.haskell.org/package/explicit-iomodes-0.1.2 Changes: * Tested with ghc-6.12.1 and base-4.2.0.0 * Re-exported 'isEOF' from System.IO which I forgot in the previous version. * Internal changes: Explicit imports of all used symbols and Unicode syntax and symbols everywhere. regards, Bas

1 0

Is it just me... or is cabal/hackage a little broken?
by Bardur Arantsson 23 Dec '09

23 Dec '09

Hi all, Sorry about the inflammatory title, but I just got this message from an uploaded package ("hums"): Warning: This package indirectly depends on multiple versions of the same package. This is highly likely to cause a compile failure. The thing is, I got the same message while trying to compile locally and it turned out that all I had to do was to $ cabal install <PKG-X> on all the packages that cabal complained about. So why doesn't hackage do this automagically when I upload a package? How am I supposed to know which versions of my package's dependencies (or their dependencies) are the most recently compiled by hackage? For the record: I did do a "Check package" upload first. It didn't complain. Is this an intractable problem? Am I being overly demanding (probably)? Cheers, Bárður

3 2

simple type?
by Jamie Morgenstern 23 Dec '09

23 Dec '09

Hello; I am writing some vectorized code, and I wondered exactly which types can be used in DPH? Is it true one cannot use data constructors and recursive types? Thanks, -Jamie

1 0

Haskell in image processing
by Ralf Laemmel 23 Dec '09

23 Dec '09

Disclaimer: this is an Xmas gift as opposed to serious stuff. http://professor-fish.blogspot.com/2009/12/major-breakthrough-in-image-proc… Greetings, Ralf

1 0