November 2006 - Haskell-Cafe

Difficult memory leak in array processing
by Niko Korhonen 03 Dec '06

03 Dec '06

Hi everyone, I have the following code whose purpose is to add dither (noise) to a given array. The code looks very straightforward but apparently it has a memory leak somewhere. Here I try to run the algorithm for an array of 10,000,000 integers. Ten million unboxed strict integers should equal to 40MB which should pose no problems to any modern system. However, the program fails with a stack overflow error. I'm using GHC 6.6 on Windows with 1 GB of RAM. I've tried applying seq and some other strictness tricks (such as x == x) pretty much everywhere on the code with no results. Could you please help me understand what is going on here? Have I misunderstood something critical in how Haskell works? Here is the relevant portion of the code: module Main where import Data.Array.IO import System.Random type Buffer = IOUArray Int Int -- | Triangular Probability Density Function, equivalent to a roll of two dice. -- The number sums have different probabilities of surfacing. tpdf :: (Int, Int) -> IO Int tpdf (low, high) = do first <- getStdRandom (randomR (low, high)) second <- getStdRandom (randomR (low, high)) return ((first + second) `div` 2) -- | Fills an array with dither generated by the specified function. genSeries :: Buffer -> ((Int, Int) -> IO Int) -> (Int, Int) -> IO () genSeries buf denfun lims = let worker low i | i >= low = do r <- denfun lims writeArray buf i r worker low (i - 1) | otherwise = return () in do (lo, hi) <- getBounds buf worker lo hi main = do -- This should allocate a 40 MB array buf <- newArray_ (0, 10000000) :: IO Buffer -- Fill the array with dither genSeries buf tpdf (2, 12) -- niko.korhonen(a)gmail.com

10 18

Re: [Haskell] ANNOUNCE: Visual Haskell prerelease 0.2
by Krasimir Angelov 02 Dec '06

02 Dec '06

Hi Shelarcy, Could you check whether you have this registry key: HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\VisualStudio\8.0\InstallDir and tell me its value? Typically its value should be such that the following script to work. Set shell = CreateObject("WScript.Shell") vstudioPath = shell.RegRead ("HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\VisualStudio\8.0\InstallDir") shell.Run ("""" & vstudioPath & "devenv.exe"" /Setup",0,true) Cheers, Krasimir On 11/30/06, shelarcy <shelarcy(a)gmail.com> wrote: > Hi Krasimir, > > > On 11/30/06, shelarcy <shelarcy(a)gmail.com> wrote: > >> But ... I can't install Visual Haskell prerelease 0.2. Near > >> the end of install process, Microsoft Development Environment > >> cause error. > On Thu, 30 Nov 2006 17:03:22 +0900, Krasimir Angelov <kr.angelov(a)gmail.com> wrote: > > Could you tell me what error message you see during the installation? > > If it is in Japan then translate it in English ;-). > > It's not good error message. Anyway, I translate it. > > Near the end of install process, error dialog opened and says: > > --- > The problem happende, so exit Microsoft Development Environment. > I'm sorry for causing inconvenience to you. > > (under its message, error dialog has form that send error report > for Microsoft or shows error detail. These messages are not > important, so I don't translate that.) > --- > > And click form that shows error detail, another dialog opened. > It shows: > > --- > :Error ditail: > An unhandled exception has been caught by the VSW exception filter. > :Error Signature: > AppName: devenv.exe AppVer: 7.10.6030.0 ModName: unknown > ModVer: 0.0.0.0 Offset: 00bbbacc > :Report Detail: > (Below meassage attetion to user that error report send what. > So these messages are not important, too.) > --- > > > Also it can help > > if you run the installer with logging: > > > > $ msiexec VSHaskell71.msi /l log.txt > > msiexec doesn't run its command. And error dialog noticed that > you forgot /i optio. So I used below command. > > $ msiexec /i VSHaskell71.msi /l log.txt > > I think log.txt is much more useful than previous messages. > log.txt also has Japanese messages. So I translated that part. > > --- > (snip) > > Action 20:38:17: CA_RegisterHelpFile.3643236F_FC70_11D3_A536_0090278A1BB8. > IHxRegisterSession::ContinueTransaction() returned 0. > Helpfile: C:\Program Files\Visual Haskell\doc\alex.HxS was successfully registered to namespace vs_haskell. > ????elpfile: C:\Program Files\Visual Haskell\doc\building.HxS was successfully registered to namespace vs_haskell. > Helpfile: C:\Program Files\Visual Haskell\doc\Cabal.HxS was successfully registered to namespace vs_haskell. > Helpfile: C:\Program Files\Visual Haskell\doc\haddock.HxS was successfully registered to namespace vs_haskell. > Helpfile: C:\Program Files\Visual Haskell\doc\happy.HxS was successfully registered to namespace vs_haskell. > Helpfile: C:\Program Files\Visual Haskell\doc\libraries.HxS was successfully registered to namespace vs_haskell. > Helpfile: C:\Program Files\Visual Haskell\doc\users_guide.HxS was successfully registered to namespace vs_haskell. > Helpfile: C:\Program Files\Visual Haskell\doc\vh.HxS was successfully registered to namespace vs_haskell. > Action 20:38:17: CA_RegisterPlugIn.3643236F_FC70_11D3_A536_0090278A1BB8. > IHxRegisterSession::ContinueTransaction() returned 0. > IHxPlugIn::RegisterHelpPlugIn() returned 0. > Namespace: vs_haskell was successfully plugged into namespace MS.VSCC.2003. > Action 20:38:17: CA_CommitHelpTransaction.3643236F_FC70_11D3_A536_0090278A1BB8. > Action 20:38:17: RegisterProduct. Registering product > RegisterProduct: {FEC3263A-9034-49C5-8C5D-902231009894} > Action 20:38:18: PublishFeatures. Publishing Product Features > PublishFeatures: Feature: Complete > Action 20:38:18: PublishProduct. Publishing product information > 1: {FEC3263A-9034-49C5-8C5D-902231009894} > Action 20:38:18: RollbackCleanup. Removing backup files > IHxRegisterSession::ContinueTransaction() returned 0. > Registration session: {FEC3263A-9034-49C5-8C5D-902231009894} was successfully committed. > RollbackCleanup: File: C:\Config.Msi\fc3d12.rbf > RollbackCleanup: File: C:\Config.Msi\fc3d13.rbf > RollbackCleanup: File: C:\Config.Msi\fc3d14.rbf > RollbackCleanup: File: C:\Config.Msi\fc3d15.rbf > RollbackCleanup: File: C:\Config.Msi\fc3d16.rbf > Action ended 20:38:18: InstallFinalize. Return value 1. > Action 20:38:18: CA_HxMerge_VSCC.3643236F_FC70_11D3_A536_0090278A1BB8. > Action start 20:38:18: CA_HxMerge_VSCC.3643236F_FC70_11D3_A536_0090278A1BB8. > ????tion ended 20:40:09: CA_HxMerge_VSCC.3643236F_FC70_11D3_A536_0090278A1BB8. Return value 1. > Action 20:40:09: CA_RemoveTempHxDs.3643236F_FC70_11D3_A536_0090278A1BB8. > Action start 20:40:09: CA_RemoveTempHxDs.3643236F_FC70_11D3_A536_0090278A1BB8. > Action ended 20:40:09: CA_RemoveTempHxDs.3643236F_FC70_11D3_A536_0090278A1BB8. Return value 1. > Action 20:40:09: VSHaskellInstall. Register Visual Haskell Plugin > Action start 20:40:09: VSHaskellInstall. > Error 1720. There is a problem with this Windows Installer package. A script required for this install to complete could not be run. Contact your support personnel or package vendor. Custom action VSHaskellInstall script error -2146828275, Microsoft VBScript Runtime Error: Couldn't match types.: 'Return' Line 55, Column 5, > MSI (s) (0C:D4) [20:51:43:359]: Product: Visual Haskell 0.2 for Visual Studio 2003 -- Error 1720. There is a problem with this Windows Installer package. A script required for this install to complete could not be run. Contact your support personnel or package vendor. Custom action VSHaskellInstall script error -2146828275, Microsoft VBScript Runtime Error: Couldn't match types.: 'Return' Line 55, Column 5, > > Action ended 20:51:43: VSHaskellInstall. Return value 3. > Action ended 20:51:43: INSTALL. Return value 3. > Action ended 20:51:43: ExecuteAction. Return value 3. > Action 20:51:43: FatalError. > Action start 20:51:43: FatalError. > Action 20:51:43: FatalError. Dialog created > Action ended 20:51:46: FatalError. Return value 2. > Action ended 20:51:46: INSTALL. Return value 3. > === Logging stopped: 2006/11/30 20:51:46 === > MSI (c) (B8:C4) [20:51:46:046]: Product: Visual Haskell 0.2 for Visual Studio 2003 -- Installation failed. > --- > > > > In the Visual Haskell\bin directory you can see one directory called > > 1033 with vs_haskell_ui.dll in it. vs_haskell_ui.dll contains various > > string resources used in Visual Haskell. For each supported language > > you need to have a different directory and 1033 is the code used for > > English. I don't know the code required for Japan but you can look at > > C:\Program Files\Microsoft Visual Studio 8\VC\bin directory and see > > what code VC++ is using. After that try to rename the 1033 directory > > to what is required for Japan. Tell me whether that works. > > Japanese diretory is 1041. > > After rename it, I tried to use Visual Haskell 0.2. But same error dialog > opened and exit Visual Studio, too. > > > -- > shelarcy <shelarcy capella.freemail.ne.jp> > http://page.freett.com/shelarcy/ >

3 12

Haskell source transformer?
by Dimitry Golubovsky 02 Dec '06

02 Dec '06

Hi, In order to automatize the creation of W3C DOM typesafe wrapper (this is needed for my Haskell->Javascript stuff) I am processing the OMG IDL files that contain interface definitions for DOM methods and attributes with HDirect. It works in general (for some reason it didn't like "boolean" type, so I used a preprocessor to redefine it as "Boolean"), and outputs some Haskell code full of foreign calls, yet it contains what I need: type signatures. For example, for the method "appendChild" it outputs: appendChild :: Node a1 -> Node a0 -> Prelude.IO (Node ()) appendChild newChild iptr = do o_appendChild <- Com.invokeIt (\ methPtr iptr -> Foreign.ForeignPtr.withForeignPtr newChild (\ new unmarshallNode o_appendChild (above lines of code may be truncated, they are just for illustration) This is almost it, but I need to replace the return type (using the JS monad instead of IO), and replace the method implementation with something else, based on unsafeJS. In some other cases I may need to modify type declarations, etc. I know there is a Haskell syntax parser around (maybe, more than one). Does anybody know of any utility based on such parser that does things I need, or rather a library on top of the parser? I just would like to avoid reinventing the wheel. Last thing I want to do is to change sources of HDirect. Thanks. -- Dimitry Golubovsky Anywhere on the Web

3 2

why are implicit types different? (cleanup)
by S. Alexander Jacobson 02 Dec '06

02 Dec '06

Why do g and g' have different types? g x y = let ?f = \x-> x in ?f x ++ (show (?f y)) g :: [Char] -> [Char] -> [Char] g' :: (Show t) => [Char] -> t -> [Char] g' x y = let f = \x-> x in f x ++ (show (f y)) Is there a way I can use implicit types and let g be as general as g'? -Alex- ______________________________________________________________ S. Alexander Jacobson tel:917-770-6565 http://alexjacobson.com

4 6

Re: [Haskell] ANNOUNCE: Visual Haskell prerelease 0.2
by Johannes Waldmann 01 Dec '06

01 Dec '06

> > Not only the interfaces [Visual Studio vs. Eclipse] > > are completely different, but an entirely new > > set of interoperability problems would need to be solved. ... I still don't see what would be the fundamental difference. (Except perhaps that the Eclipse interfaces are easily available and well documented so it is at least possible to describe the interface problems...) The main advantage (Visual Haskell over eclipsefp) at the moment is that VH uses incremental (on-the-fly) typechecking/compilation while eclipsefp calls the compiler for whole modules? What source text transformations (refactorings) does VH support? Best regards, -- -- Johannes Waldmann -- Tel/Fax (0341) 3076 6479/80 -- ---- http://www.imn.htwk-leipzig.de/~waldmann/ -------

5 7

Building Binaries with Cabal
by John Goerzen 01 Dec '06

01 Dec '06

Hi folks, I'm in need of some Cabal assistance. I want to build the unit tests for MissingH using Cabal. According to the docs, this should require me to list all of the exposed modules from the library as other modules to the binary. Since there are dozens of these, I thought a simple hook could do the trick. I tried hooking into the confHook (which I have already used successfully to add "unix" as a build-dep on non-Windows platforms), and thought I could just pull the exposedModules list from the package and add those as otherModules to the executable. But it had no effect. So I tried hooking in to customBuildHook to do the same thing. Again, no effect. Here's the code I've tried. Suggestions appreciated. import Distribution.Simple import Distribution.PackageDescription import Distribution.Version import System.Info import Data.Maybe winHooks = defaultUserHooks {confHook = customConfHook} customConfHook descrip flags = let mydescrip = case System.Info.os of "mingw32" -> descrip _ -> descrip {buildDepends = (Dependency "unix" AnyVersion) : buildDepends descrip} in (confHook defaultUserHooks) mydescrip flags customBuildHook descrip lbi uh flags = let myexecutables = map bdfix (executables descrip) bdfix exe = exe {buildInfo = (buildInfo exe) {otherModules = exposedModules . fromJust . library $ descrip}} mydescrip = descrip {executables = myexecutables} in (buildHook defaultUserHooks) mydescrip flags main = defaultMainWithHooks winHooks

2 4

New Name for MissingH?
by John Goerzen 01 Dec '06

01 Dec '06

Quick feedback time... One comment people made in the Future of MissingH thread was that the name isn't very suggestive of what the library does. I'm planning to follow the advice of many people and split the major MissingH components off into smaller bits (ConfigParser, HVFS, etc). MissingH itself will then contain any small utility functions (probably mainly string and list-related) that for whatever reason aren't suitable to go into base. What should it be called? Should it keep the MissingH name? The alternative I've been thinking of is something like Haskell Utility Library (HUL). Ideas?

5 6

Command line utility that shrinks/simplifies functions applications ?
by Nicola Paolucci 01 Dec '06

01 Dec '06

Hi All! Haskell newbie here with a very simple question because google and hoogle are of no help. On the IRC channel #haskell (which I cannot access now from work) I saw somebody using a tool which automatically simplifies expressions,composition of multiple functions to the bare minimum. It was a query to lambdabot I think. Is that tool/library also standalone ? What's its name ? Where can I find it ? That really rocked ... Thanks, Nick

4 4

[Redirect] polymorphism and existential types
by dons＠cse.unsw.edu.au 30 Nov '06

30 Nov '06

Redirected to haskell-cafe(a)haskell.org, the right list. -- Don ----- Forwarded message from Louis-Julien Guillemette <guillelj(a)iro.umontreal.ca> ----- Date: Mon, 27 Nov 2006 16:15:26 -0500 (EST) From: Louis-Julien Guillemette Subject: [Haskell] polymorphism and existential types Supposing a polymorphic value (of type, say, forall a . ExpT a t) is stored inside an existential package (of type, say, forall a . Exp a), I wonder how to recover a polymorphic value when eliminating the existential. The ``natural way'' to write this doesn't work: {-# OPTIONS -fglasgow-exts #-} data ExpT a t data Exp a = forall t . Exp (ExpT a t) f :: (forall a . ExpT a t) -> () f e = () g :: (forall a . ExpT a t) -> () g e = let e1 :: forall a . Exp a e1 = Exp e in case e1 of Exp e' -> f e' {- Test.hs:18:17: Inferred type is less polymorphic than expected Quantified type variable `a' is mentioned in the environment: e' :: ExpT a t (bound at Test.hs:18:11) In the first argument of `f', namely `e'' In the expression: f e' In a case alternative: Exp e' -> f e' Failed, modules loaded: none. -} In particular, the solution of pushing the forall inside doesn't work for me (it breaks other parts of my code...) : data Exp' = forall t . Exp' (forall a . ExpT a t) g' :: (forall a . ExpT a t) -> () g' e = let e1 :: Exp' e1 = Exp' e in case e1 of Exp' e' -> f e' Any help welcome! Louis-Julien _______________________________________________ Haskell mailing list Haskell(a)haskell.org http://www.haskell.org/mailman/listinfo/haskell

2 1

The Future of MissingH
by John Goerzen 30 Nov '06

30 Nov '06

Hi, Josef Svenningsson posted a comment on my blog today that got me to thinking. He suggested that people may be "intimidated by the size of MissingH, confused by the undescriptive name, and don't quite know what's in there." And I think he's right. I've been passively thinking about what MissingH should be for awhile now, and wonder if you all would have some suggestions. First, let me back up and describe a bit of what MissingH is. It gets its name because I think that it has useful functions that are "missing" from the standard Haskell libraries. But I can see how it's not very descriptive. The major features are: * A general-purpose modular logging infrastructure * A virtual filesystem component (similar to VFS in Gnome, but written in Haskell) * A configuration file parser, compatible with Python's and Perl's, which can also modify and generate config files and do variable interpolation * A bunch of tools for setting up pipes to/from processes * Tools to format numbers as quantities, with default configs for the SI system (1000-based) and binary system (1024-based) * Tools to track the progress of long-running actions and display a status bar on the terminal * Pure-Haskell un-gzip code, *DBM infrastructure, etc * Assorted list, string, regexp, bit, pointer, etc. utilities * MIME and email parsers. I wrote most of the code myself, but did pull some large chunks from others where licenses were compatible. The API reference is at http://gopher.quux.org:70/devel/missingh/html/index.html So, some of the questions are: Should this all be one library? I initially wrote it that way to make resolving dependencies easier for end users. Also, it still has utility because modules make use of each other's features. The list and string functions are used almost everywhere, for instance. HVFS support is also fairly pervasive. Should the module naming scheme be changed? Modules are currently named things like MissingH.List, MissingH.IO.HVFS, MissingH.ProgressMeter, etc. I guess these could go to System.ListExt, System.IO.HVFS, System.Console.ProgressMeter, etc. Could, and should, any of this be integrated into the Haskell libraries project? (That set of libraries formerly known as fptools) A few bits of code have already found their way into them. But I bet there could be a lot more. Maybe even the majority of it. How could greater community participation be encouraged, while still encouraging quality control? I have received some very good contributions to MissingH from people, and that's been great. I've also received some that just aren't that great -- they don't have Haddock docs, the code is opaque, they don't come with unit tests, etc. But by and large, I've been maintaining it mostly myself. Whenever I write some code, I think about whether this is generic code that could be useful in other projects. If so, I consider whether it would be appropriate for MissingH (and it usually is). It would be wonderful if others would be interested in contributing code that solves some need as well. There is a public darcs repo already, and I have received some darcs patches from people (thanks!) I'm not trying to rid myself of MissingH, but I think that it could be a nice resource for small tools from the entire community, if we can maintain a structure and QA footprint to it. What else should be done to make this a valuable resource for Haskell programmers? And a showcase for what is possible with Haskell? The floor's open... Thanks, -- John

11 19