- Haskell-Cafe - Haskell.org

StateWriter: a monad-writing exercise
by Anthony LODI 26 Sep '08

26 Sep '08

Hello haskell-cafe, In my application I have a complex state threaded through long computation chains, and I need a way to log all state changes (so that the evolving state can be animated/replayed somewhere else). Initially I tried combining State and Writer monads but this allows for the possibility to change the state, and forget to write a log entry, etc. So I decided to write a separate monad, StateWriter l s, that takes a state-modifying function, l->s->s (l is an ADT for all the allowable state transitions), an initial state s, and only allows s to change by appending 'l' log entries inside the monad. The net result is that I should have read-only access to the current state inside the monad, and all state transitions should be logged (by going through the one function, the log entries serve as witnesses to all state transitions). Anyway, here's my (very rough!) first stab at the problem. This is the first time I've tried writing a monad so any comments/critiques are much appreciated. Also, about the 'StateWriter' idea in general: am I just (poorly?) reimplementing something obvious? Is it unlikely to scale well on real-world problems? Is there some way to combine the existing State and Writer monads to avoid having to do this? If there's nothing seriously wrong here, I was thinking my next step would be to try changing the lists to monoids (like in the Writer monad), and then to try writing a transformer version of the whole thing. Cheers, - Anthony LODI ================================================================================ {-# LANGUAGE MultiParamTypeClasses, FunctionalDependencies, FlexibleInstances #-} newtype StateWriter l s a = StateWriter { _runSWriter :: (l -> s -> s) -> [l] -> s -> (a, [l], s) } instance Monad (StateWriter l s) where return a = StateWriter $ \_ ls s -> (a, ls, s) (StateWriter x) >>= f = StateWriter $ \fn ls s -> let (v, ls', s') = x fn ls s in _runSWriter (f v) fn ls' s' class MonadStateWriter m l s | m -> l s where put :: l -> m () get :: m s instance MonadStateWriter (StateWriter l s) l s where put l = StateWriter $ \fn ls s -> ((), ls ++ [l], fn l s) get = StateWriter $ \fn ls s -> (s, ls, s) runSWriter :: StateWriter l s a -> (l -> s -> s) -> s -> (a, [l], s) runSWriter sw fn = _runSWriter sw fn [] -------------------------------------------------------------------------------- data Ops = Inc | Dec deriving (Show) test :: StateWriter Ops Int String test = do put Inc put Inc put Inc val <- get let op = if val == 3 then Dec else Inc put op return "done" stateFn :: Ops -> Int -> Int stateFn Inc s = s + 1 stateFn Dec s = s - 1 runtest = runSWriter test stateFn 0 -- ("done",[Inc,Inc,Inc,Dec],2)

2 2

Re: [Haskell-cafe] Injecting Haskell into C
by Anatoly Yakovenko 26 Sep '08

26 Sep '08

you'll find this example really helpfull ---------- Forwarded message ---------- From: Claude Heiland-Allen <claudiusmaximus(a)goto10.org> Date: 2008/6/5 Subject: Re: [Haskell-cafe] example of FFI FunPtr To: "Galchin, Vasili" <vigalchin(a)gmail.com> Cc: haskell <haskell-cafe(a)haskell.org> Galchin, Vasili wrote: > > Hello, > > I want to model a Haskell function that is a callback from C. I have > only found one example in the unix package's Semaphore.hsc, which apparently > is not used. I want to be able to marshall a Haskell function that is a > first class citizen residing in a Haskell data type and pass to a C function > via FFI. Are there examples of this? Attached is a simple example. The main thing to note is 'foreign import ccall "wrapper"' which gives you a factory for turning Haskell functions into foreign function pointers. More information: http://www.cse.unsw.edu.au/~chak/haskell/ffi/ Claude -- http://claudiusmaximus.goto10.org CallBacker: CallBacker.hs callerback.c callerback.h ghc -O2 -Wall -fffi -o CallBacker CallBacker.hs callerback.c _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe(a)haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

2 3

GHC 6.10, OS X, Fink and CPPFLAGS
by Svein Ove Aas 26 Sep '08

26 Sep '08

While trying to test 6.10, I ran into the issue that a number of the libraries it wants are not installed by default, and are most conveniently added via Fink. This includes gmp, binutils, readline, etc. /sw/ is not in the default gcc path, of course, so the usual solution is to add -I/sw/include, -L/sw/lib, etc. to it via the CPPFLAGS and LDFLAGS environment variables, which are supported by most[1] autoconf-based packages. The help text for ghc's configure script does in fact suggest that this is the case, and it has some code meant to read it; broken code, as it happens. As it is, the script fails to propagate those variables to whatever is responsible for compiling the RTS, which fails with a missing gmp.h, bfd.h, and so on. All those files exist live quite happily in /sw/include, and would no doubt appreciate visiting makefiles. So my question is two-fold: Is the help file wrong? Is there some other way to do it? Or is this in fact a bug, and if so, will any of you volunteer to fix it for me? ^_^

1 0

Where is OpenGL loadMatrix ?
by minh thu 25 Sep '08

25 Sep '08

Hi, I can't find the loadMatrix function in http://hackage.haskell.org/packages/archive/OpenGL/latest/doc/html/Graphics… Should I use loadIdentity then multMatrix instead ? Thank you, Thu

4 6

Google Android
by Galchin, Vasili 25 Sep '08

25 Sep '08

Hello, Do there currently (or in the works) exist FFI bindings for Google's Android API? Kind regards, Vasili

3 2

Am I doing it right?
by Creighton Hogg 25 Sep '08

25 Sep '08

Hi Haskellers, So as another step in my education, I wanted to ask if the following code at least feels 'morally' correct as a find/replace function, replacing each occurrence of the sub-list before with after in the list. unfoldr felt like the most natural higher order function to capture the recursion, but I imagine there are still slicker ways to have done this because I am creating an intermediate list-of-lists. However, doing it this way seems at least moderately speedy given that I was able to read in, find/replace, and write out a 120 MB file in 30 seconds on a fairly dinky linux box. replace :: (Eq a) => [a] -> [a] -> [a] -> [a] replace before after = concat . unfoldr aux where aux [] = Nothing aux str | take l str == before = Just (after,drop l str) | otherwise = Just (take 1 str,drop 1 str) l = length before Thanks, Creighton

4 5

Haskell board game
by Rafael C. de Almeida 25 Sep '08

25 Sep '08

Hello, I'm interested in doing a simple board game on haskell. For that I want to be able to draw stuff like the possible player movements and I want to be able to display very simple animations. I want to know what graphical interface library you suggest to me. I have almost no prior experience with graphical interfaces of any kind, so I rather start with something easy and straightforward. I have no need for great performance or anything like that. I'd like if it runs on windows with much trouble, that is, it'll be easy to package it for windows without requiring the user to install anything besides my game. My first thought was to use GTK's gtkTable, but I'm unsure how easy it is to make it work on windows. Beside that, I'm not sure it would be the easiest API for me to use. []'s Rafael

4 3

need help making sense of the type ...
by Daryoush Mehrtash 25 Sep '08

25 Sep '08

I am having hard time making sense of the types in the following example from the Applicative Programming paper: http://www.cs.nott.ac.uk/~ctm/IdiomLite.pdf ap :: Monad m ⇒ m (a → b ) → m a → m b ap mf mx = do f ← mf x ← mx return (f x ) Using this function we could rewrite sequence as: sequence :: [ IO a ] → IO [ a ] sequence [ ] = return [ ] sequence (c : cs ) =* return (:) 'ap' c *'ap' sequence cs I am specifically confused over the type of "m" in: return (:) 'ap' c "c" is obviously an instance of IO a monad. "return (:)" on the other hand (at least as I would expect it) is an instance of " ->" monad. a) are the above statements correct? b) if so, does it make sense for the "ap" function to have two different instances of the "m"? thanks for you help daryoush

2 1

Injecting Haskell into C
by Roman Cheplyaka 25 Sep '08

25 Sep '08

I have a C function whose behaviour is customized by user-supplied function (think of libc qsort). Typically these user-supplied functions are written in C, but I'd like to use FFI to write them in Haskell. Precisely, I'd like to write high-order function which will generate these functions (e.g. mkCompare :: ... -> (Ptr a -> Ptr a -> Int) for feeding the result to qsort). As I understand, there are two ways to do that. Either Haskell code is called from C, or C code is called for Haskell. So my questions are: 1. Are they both possible? 2. If yes, which is better performance-wise? (C function is performance-critical). If generated function is called many times, how big an overhead is going to be? -- Roman I. Cheplyaka :: http://ro-che.info/ kzm: My program contains a bug. How ungrateful, after all I've done for it.

3 2

Re: [Haskell] ICFP programming contest results
by David Roundy 25 Sep '08

25 Sep '08

On Wed, Sep 24, 2008 at 1:06 AM, Malcolm Wallace <malcolm.wallace(a)cs.york.ac.uk> wrote: > The ICFP programming contest results presentation: > http://video.google.com/videoplay?docid=-4697764813432201693 > > Feel free to pass on this link to any other appropriate forum. Yikes. Haskell did pretty terribly! Anyone have an idea why, given our success in previous contests? David (speaking as someone who was traveling that weekend and didn't have time to compete)

4 3

the way to fail
by Jason Dusek 24 Sep '08

24 Sep '08

Whenever `fail` comes up, there are usually remarks to the effect that it doesn't really belong in the definition of `Monad`. Where does `fail` belong? Could it go in `Arrow`? -- _jsn

2 1

Haskell Weekly News: Issue 86 - September 20, 2008
by Brent Yorgey 24 Sep '08

24 Sep '08

--------------------------------------------------------------------------- Haskell Weekly News http://sequence.complete.org/hwn/20080920 Issue 86 - September 20, 2008 --------------------------------------------------------------------------- Welcome to issue 86 of HWN, a newsletter covering developments in the [1]Haskell community. Lots of NEW stuff this week! A new generics library, new versions of Pandoc and darcs, a new website for xmonad, a new GADT/type family inference engine for GHC, a Haskell binding for Qt, and some new, astonishingly elegant ideas from Oleg. Also, here's hoping that everyone has a lot of fun at ICFP! Announcements GHC version control. Simon Peyton-Jones [2]sent out a revised proposal for GHC version control. darcs 2.0.3pre1. Eric Kow [3]announced the first pre-release of [4]darcs 2.0.3, featuring a few major bug fixes and a handful of interesting features. EMGM. Sean Leather [5]announced a release of [6]Extensible and Modular Generics for the Masses (EMGM), a library for generic programming in Haskell using type classes. Pandoc 1.0.0.1. John MacFarlane [7]announced the release of [8]pandoc 1.0.0.1, the swiss army knife of text markup formats. Iteratee-based IO. oleg [9]described a [10]safe, declarative approach to input processing which will be the subject of a talk at DEFUN08 on September 27. MetaHDBC paper. Mads Lindstroem [11]announced a [12]draft version of a paper about the [13]MetaHDBC library, which uses Template Haskell to do type-safe database access. Comments are welcomed, especially about the overall quality of the paper, whether it can be called scientific, and anything Mads could do to improve the paper. qtHaskell 1.1.2. David Harley [14]announced a second preview release of [15]qtHaskell, a set of Haskell bindings for Trolltech's Qt. Discussion Library design question. Andre Nathan [16]asked for advice on designing a simple graph library. The resulting discussion included an analysis of using the State monad versus a more functional approach. A round of golf. Creighton Hogg [17]learns about laziness by [18]making grown men cry. XML (HXML) parsing :: GHC 6.8.3 space leak from 2000. Lev Walkin [19]discovers a nice example of an obscure class of space leaks while writing some XML-processing code, prompting an in-depth analysis by Simon Marlow. Proofs and commercial code. Daryoush Mehrtash [20]asked about automated proof tools and techniques, and their uses in the real world. Blog noise [21]Haskell news from the [22]blogosphere. * Creighton Hogg: [23]Haskell Cafe or: How I learned to stop worrying & love laziness. * Douglas M. Auclair (geophf): [24]Animal as RDR, part II. Doug continues his posts on RDR expert systems. * Ivan Lazar Miljenovic: [25]Getting Real World Haskell Down Under. * Douglas M. Auclair (geophf): [26]Animal: an RDR implementation study. Doug describes "ripple-down rules" expert systems, and illustrates the types needed to encode one in Haskell. * Mark Jason Dominus: [27]data Mu f = In (f (Mu f)). Mark writes about fixpoints of type constructors. * John Goerzen (CosmicRay): [28]Switched from KDE to xmonad. John has taken the plunge to xmonad and seems to like it so far! * Eric Kow (kowey): [29]darcs weekly news #4. Pre-release of darcs 2.0.3; darcs hacking sprint next month; code.haskell.org upgrades to darcs 2; and other news. * Mads Lindstroem: [30]MetaHDBC paper (draft). Mads's first paper ever, on using Template Haskell for type-safe database access. Comments welcome! * Braden Shepherdson: [31]xmonad-light 0.8 Released. * Manuel M T Chakravarty: [32]GHC HEAD just got a new inference engine for GADTs and type families.. * Magnus Therning: [33]Haskell and Time. Magnus describes the solution to a problem with Data.Time. * Dan Piponi (sigfpe): [34]Two Papers and a Presentation. * Xmonad: [35]New xmonad website launched. xmonad has a shiny new website! Quotes of the Week * Botje: GHC 11 will have shootout entries as primitives. * wjt: oh, i see what you're doing. ...no, i don't. *splode* * Benjamin Pierce: [on existential types] I have a term, and it has a type. So there. * bos: come on, real programmers use "(((,) <$>) .) . (<*>)" * quicksilver: #haskell : Sometimes we answer your question, sometimes we lay hideous traps which will devour your soul. It's a risk you take. * harrison: [on computing 1000000!] it is the same as factorial 999999 * 1000000, big deal About the Haskell Weekly News New editions are posted to [36]the Haskell mailing list as well as to [37]the Haskell Sequence and [38]Planet Haskell. [39]RSS is also available, and headlines appear on [40]haskell.org. To help create new editions of this newsletter, please see the information on [41]how to contribute. Send stories to byorgey at cis dot upenn dot edu. The darcs repository is available at darcs get [42]http://code.haskell.org/~byorgey/code/hwn/ . References 1. http://haskell.org/ 2. http://www.haskell.org//pipermail/libraries/2008-September/010661.html 3. http://www.haskell.org//pipermail/haskell-cafe/2008-September/047664.html 4. http://www.darcs.net/darcs-2.0.3pre1.tar.gz 5. http://www.haskell.org//pipermail/haskell-cafe/2008-September/047751.html 6. http://www.cs.uu.nl/wiki/GenericProgramming/EMGM 7. http://www.haskell.org//pipermail/haskell-cafe/2008-September/047437.html 8. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/pandoc 9. http://article.gmane.org/gmane.comp.lang.haskell.cafe/44828 10. http://okmij.org/ftp/Haskell/Iteratee/ 11. http://article.gmane.org/gmane.comp.lang.haskell.cafe/44713 12. http://lindstroem.files.wordpress.com/2008/09/metahdbc.pdf 13. http://www.haskell.org/haskellwiki/MetaHDBC 14. http://article.gmane.org/gmane.comp.lang.haskell.general/16425 15. http://qthaskell.berlios.de/ 16. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/44733 17. http://abstractabsurd.blogspot.com/2008/09/haskell-cafe-or-how-i-learned-to… 18. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/44724 19. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/44708 20. http://www.haskell.org//pipermail/haskell-cafe/2008-September/047440.html 21. http://planet.haskell.org/ 22. http://haskell.org/haskellwiki/Blog_articles 23. http://abstractabsurd.blogspot.com/2008/09/haskell-cafe-or-how-i-learned-to… 24. http://logicaltypes.blogspot.com/2008/09/animal-as-rdr-part-ii.html 25. http://ivanmiljenovic.wordpress.com/2008/09/19/getting-real-world-haskell-d… 26. http://logicaltypes.blogspot.com/2008/09/animal-rdr-implementation-study.ht… 27. http://blog.plover.com/prog/springschool95-2.html 28. http://changelog.complete.org/posts/756-Switched-from-KDE-to-xmonad.html 29. http://koweycode.blogspot.com/2008/09/darcs-weekly-news-4.html 30. http://lindstroem.wordpress.com/2008/09/18/metahdbc-paper-draft/ 31. http://braincrater.wordpress.com/2008/09/17/xmonad-light-08-released/ 32. http://justtesting.org/post/50500880 33. http://therning.org/magnus/archives/389 34. http://sigfpe.blogspot.com/2008/09/two-papers-and-presentation.html 35. http://xmonad.wordpress.com/2008/09/14/new-xmonad-website-launched/ 36. http://www.haskell.org/mailman/listinfo/haskell 37. http://sequence.complete.org/ 38. http://planet.haskell.org/ 39. http://sequence.complete.org/node/feed 40. http://haskell.org/ 41. http://haskell.org/haskellwiki/HWN 42. http://code.haskell.org/~byorgey/code/hwn/

3 3

A question about parsec implementation
by Xinyu Jiang 24 Sep '08

24 Sep '08

Hi everybody, I am trying to port (a part of) the parsec library to scheme. I have read the paper and the haskell source of parsec, but I still have a problem. May be the problem is from my lack of parser knowledge, but I have to bother you. Scheme does not support lazy evaluation by default, so I simulated laziness when implementing finite lookahead. The result was not what I expected: it slowed down the parser and ate more memory, not speeding it up and reducing its space leak. If finite lookahead is a real gain, it would beat the cost of simulating lazy evaluation, which is not much. After analysis my code, I drew a conclusion (which may be wrong) that if the grammar you write is LL(1), that is, you don't use the "try" combinator, the time complexity of the predictive parser and the backtracking parser will be the same when they both suceed. Because in an LL(1) grammar, any consumed errors will make the whole parser return errors. Since the behaviors of the predictive parser and the backtracking parser are the same on LL(1) grammars, the backtracking parser will only eat more memory than the predictive parser when it fails. Is my conclusion right? Does parsec perform better than backtracking parsers on a pure LL(1) grammar when it suceeds? I am very grateful for your answer, thanks.

1 0

what is the magic hash?
by Jason Dusek 24 Sep '08

24 Sep '08

It is not much covered in the docs. It has something to do with magic triggered by a postfix octothorpe? -- _jsn |...docs.| http://haskell.org/ghc/docs/latest/html/users_guide/flag-reference.html

3 3

ANN: A Functional Implementation of the Garsia-Wachs Algorithm
by Nicolas Pouillard 24 Sep '08

24 Sep '08

Hi All, Here is an Haskell implementation of an algorithm that builds a binary tree with minimum weighted path length from weighted leaf nodes given in symmetric order. This can be used to build optimum search tables, to balance a 'ropes' data structure in an optimal way. This module a direct translation from OCaml of a functional pearl by Jean-Christophe Filliâtre yesterday on ML Workshop 2008. There was an interesting point to porting it to Haskell, indeed there is a crucial use of first-class references used only locally. A good reason to show the power of the ST monad and it's runST function! Here is the hackage URL: http://hackage.haskell.org/cgi-bin/hackage-scripts/package/garsia-wachs And the darcs 2 URL: http://darcs.feydakins.org/garsia-wachs -- Nicolas Pouillard aka Ertai

3 3

Re: ANNOUNCE: protocol-buffers-0.2.9 for Haskell is ready
by Chris Kuklewicz 24 Sep '08

24 Sep '08

I am cross-posting this message to several lists. I had learned the trick before the documentation was updated. It seems I have used a very unreliable trick. And the "use castToSTUArray" suggested alternative is a really poor one since I am not using arrays at all. Who can suggest a way to cast from Float to Word32 and Double to Word64 using ghc? The actual task is that I need to write out the Float as a little endian sequence of four bytes and also be able to read it back in. The writing and reading are done in Put and Get monads to ByteString (from the "binary" package). The alloca/poke/peek work around I have looks like castWord32ToFloat :: Word32 -> Float castWord32ToFloat x = unsafePerformIO $ alloca $ \p -> poke p x >> peek (castPtr p) castFloatToWord32 :: Float -> Word32 castFloatToWord32 x = unsafePerformIO $ alloca $ \p -> poke p x >> peek (castPtr p) The unsafeCoerce trick that is no longer working looks like: castWord64ToDouble :: Word64 -> Double castWord64ToDouble (W64# w) = D# (unsafeCoerce# w) castDoubleToWord64 :: Double -> Word64 castDoubleToWord64 (D# d) = W64# (unsafeCoerce# d) Any ideas? Or is the alloca trick the only way to do this? Chris Ian Lynagh wrote: > Hi Chris, > > On Sun, Sep 21, 2008 at 05:37:33PM +0100, Chris Kuklewicz wrote: >> Also, I tried two tricks: >> (D# x) <-> (W64# x) which works fine >> (F# x) <-> (W32# x) which produced garbage, so I had to replace it with >> alloca/poke/peek. > > This isn't supported, and I suspect is the cause of the -fasm problems. > > Please see > http://hackage.haskell.org/trac/ghc/ticket/2209 > for more details and suggested alternative. > > > Thanks > Ian >

4 3

hGetContents and lazyness
by Max Vasin 24 Sep '08

24 Sep '08

Hello, haskellers! Suppose we have function (it filters package filenames from apt Packages file): > getPackageList :: FilePath -> IO [FilePath] > getPackageList packageFile = withFile packageFile ReadMode $ > \h -> do c <- hGetContents h > return $ map (drop 10) $ filter (startsWith "Filename:") $ lines c -- (1) > where startsWith [] _ = True > startsWith _ [] = False > startsWith (x:xs) (y:ys) | x == y = startsWith xs ys > | otherwise = False When, I apply it to a Packages file I (surely) get an empty list. This is an expected result due to lazyness of hGetContents. I tried changing line (1) to > return $ map (drop 10) $ filter (startsWith "Filename:") $! lines c or > return $ map (drop 10) $! filter (startsWith "Filename:") $! lines c with no success. Chaning it to > return $! map (drop 10) $ filter (startsWith "Filename:") $ lines c makes getPackageList function return several (but not all) filenames. What I'm missing? And how can I fix my code? -- WBR, Max Vasin.

2 3

"piping" to system call
by Janis Voigtlaender 23 Sep '08

23 Sep '08

Hi, assume I have a program taking input from stdin. How do I call it from Haskell while feeding to it a string as input. That is, I want a function like system' :: String -> String -> IO ExitCode such that system' cmd inp would be equivalent to first writing inp to a file, say "temp", and then calling system (cmd ++ " < temp") Of course, without actually creating that temporary file and having to clean it up afterwards. Does such a system' exist? Thanks, Janis. -- Dr. Janis Voigtlaender http://wwwtcs.inf.tu-dresden.de/~voigt/ mailto:voigt@tcs.inf.tu-dresden.de

5 7

Re[10]: [Haskell-cafe] Climbing up the shootout...
by Bulat Ziganshin 23 Sep '08

23 Sep '08

Hello Donnie, Tuesday, September 23, 2008, 7:16:04 PM, you wrote: > I don't understand why you are willing to criticize GHC, but > unwilling to help improve GHC. Personally, I think it is a waste of > everyone's time for you to just complain about GHC without offering > suggestions on how to improve GHC -- What is the point? i don't criticize nor complain. haskell is much more higher-level language than C and that's the reason why it's slower. i just cold down heroes that lie to novices saying that Haskell is as fast or even faster than C. if i will say that Ford is slower than F1 cars, is it complaints or critics? :) when i worked on high-speed Haskell libs, i helped GHC team a bit by showing concrete cases. now i just don't work in this area so i have a little to offer ps: xposted to cafe -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com

1 0

XML (HXML) parsing :: GHC 6.8.3 space leak from 2000
by Lev Walkin 23 Sep '08

23 Sep '08

Recently I had to process some multi-megabyte XML files. Tried a few Haskell XML parsers (HaXML, HXT, HXML) but all of them were exhibiting very pronounced space leaks, and all but HXML were too strict for my input. Judging by the code and stated objectives, Joe English's HXML (0.2, circa 2003) looked more promising for hacking around so I tried to figure out the space leak problem. It wasn't too long to find out the source of a problem, the buildTree function in TreeBuild.hs. In fact, the very annotation to that function reads as follows: -- %%% There is apparently a space leak here, but I can't find it. -- %%% Update 28 Feb 2000: There is a leak, but it's fixed -- %%% by a well-known GC implementation technique. Hugs 98 happens -- %%% not to implement this technique, but STG Hugs (and most other -- %%% Haskell systems) do implement it. -- %%% Thanks to Simon Peyton-Jones, Malcolm Wallace, Colin Runcinman -- %%% Mark Jones, and others for investigating this. And there's some more in the accompanying documentation: + Under Hugs 98 only, suffers a serious space fault. I wondered why would a contemporary GHC 6.8.3 exhibit such a leak? After all, the technique was known in 2000 (and afir by Wadler in '87) and one would assume Joe English's reference to "most other Haskell systems" ought to mean GHC. But here we are, in 2008 I still can't get HXML to not to leak like a hose while lazily parsing my file. In fact, I can't get my 45-megabyte file parsed on my 1GB RAM system without swapping. So I went ahead and extracted the code and stripped all XML related junk to reproduce the problem with a minimal test case. Attached please find a single tree.hs module which is just sufficient to demonstrate a memory leak. Here's a culprit function: data TreeEvent = Start String -- Branch off a new subtree | Stop -- Stop branching and return 1 level | Leaf String -- A simple leaf without children deriving Show -- Lazily build a tree out of a sequence of tree-building events build :: [TreeEvent] -> ([UnconsumedEvent], [Tree String]) build (Start str : es) = let (es', subnodes) = build es (spill, siblings) = build es' in (spill, (Tree str subnodes : siblings)) build (Leaf str : es) = let (spill, siblings) = build es in (spill, Tree str [] : siblings) build (Stop : es) = (es, []) build [] = ([], []) In fact, the attached module implements almost verbatim the code from an old Joe's request (circa 2000): http://www.cse.unsw.edu.au/~dons/haskell-1990-2000/msg06086.html but my version is a bit more self-sufficient for the new folks who'd like to quickly test it on their system. Am I really ignorant of some important GHC optimization options (tried -O2/-O3), or is this indeed a serious problem to tackle? -- Lev Walkin vlm(a)lionet.info module Main where -- A simple tree to hold hierarchial XML-like data structure data Tree a = Tree a [Tree a] deriving Show -- A member of the event sequence which attempts to build a tree. -- For example, the following sequence -- [Start "top", Leaf "leaf", Start "sub", Stop, Stop] -- should correspond to the following tree: -- Tree "top" [Tree "leaf" [], Tree "sub" []] data TreeEvent = Start String -- Branch off a new subtree | Stop -- Stop branching and return 1 level | Leaf String -- A simple leaf without children deriving Show -- Lazy printing of an infinite tree building process main = print . snd . build $ Start "top" : cycle [Leaf "sub"] -- Convert a stream of tree building events -- into a list of unconsumed events and a constructed tree body. type UnconsumedEvent = TreeEvent -- Alias for program documentation build :: [TreeEvent] -> ([UnconsumedEvent], [Tree String]) build (Start str : es) = let (es', subnodes) = build es (spill, siblings) = build es' in (spill, (Tree str subnodes : siblings)) build (Leaf str : es) = let (spill, siblings) = build es in (spill, Tree str [] : siblings) build (Stop : es) = (es, []) build [] = ([], []) -- Stricter version of build, never terminates on infinite input, -- but exhibits no space leaks whatsoever. build' f (Start str : es) = let (es', subnodes) = build' id es in build' ((Tree str subnodes) :) es' build' f (Leaf str : es) = build' ((Tree str []) :) es build' f (Stop : es) = (es, f []) build' f [] = ([], f [])

4 7

how do i use quickcheck in the IO monad?
by Anatoly Yakovenko 23 Sep '08

23 Sep '08

If i have functions in the IO monad, is there a way to use quickcheck to test them? I have a bunch of C bindings that unfortunately are not "safe". But i would like to be able to use QuickCheck to test them. Thanks, Anatoly

4 3

Re: [Haskell-cafe] Updated formlets sample?
by Martin Huschenbett 23 Sep '08

23 Sep '08

Hi Chris, you're absolutely right. The mistake was in the where-part of withForm. The function handleOk' gets an environment d as argument but uses an extractor that was created without passing d to runFormState. I've put a corrected version on hpaste [1] and also posted it to the wiki on haskell.org [2]. Hope this is ok for you? Regards, Martin. [1] http://hpaste.org/10568#a1 [2] http://haskell.org/haskellwiki/Formlets Chris Eidhof schrieb: > That means that you don't have input0 in your environment, maybe you're > passing in an empty environment? > > -chris > > On 21 sep 2008, at 12:11, Martin Huschenbett wrote: > >> Hi Chris, >> >> thanks for the updated example. Compiling works now. But when I try to >> run it I alway get error messages like >> >> ["input0 is not in the data","input1 is not in the data"] >> >> Regards, >> >> Martin. >> >> Chris Eidhof schrieb: >>> Hey Martin, >>> On 19 sep 2008, at 04:14, Martin Huschenbett wrote: >>>> I found a blog post concerning formlets [1] in the web. Since looks >>>> very interesting I tried to compile the sample code with recent >>>> versions of HAppS and formlets from hackage. But this didn't work as >>>> the API of formlets has changed since this post. >>>> >>>> I tried to adopt the code to the new API but I was unable to finish >>>> this since there is a new monadic context I don't know to handle in >>>> the right way. >>>> >>>> So my question is, is there an updated version of this sample code >>>> in the web or has anybody tried to adopt it and can send me the >>>> results? >>> Yes, I'm sorry for that. The API is still very immature and due to >>> changes, that's also why it hasn't been officially announced yet. >>> I've just put an updated example at http://hpaste.org/10568, hope >>> that'll work for you. I guess we should build a small homepage / >>> wikipage that always has an up-to-date example. >>> -chris

2 1

Re: [Haskell-cafe] Bird-style and blank lines
by Ryan Ingram 23 Sep '08

23 Sep '08

Oops, meant to send this to the whole list. > You can add "-optL -q" to your ghc command line to disable that > behavior; blank lines will no longer be required. -- ryan

2 1

Live twittering from ICFP
by Don Stewart 23 Sep '08

23 Sep '08

ICFP is on now, for those wondering why the list is a bit quiet :) http://www.icfpconference.org/icfp2008/schedule.html I'm doing live updates of events at ICFP/Haskell Symposium here in Canada. http://twitter.com/galoisinc So you can follow events remotely. Cheers, Don

1 0

Fwd: [Haskell-cafe] hGetContents and lazyness
by Rafael Gustavo da Cunha Pereira Pinto 23 Sep '08

23 Sep '08

---------- Forwarded message ---------- From: Rafael Gustavo da Cunha Pereira Pinto <rafaelgcpp(a)gmail.com> Date: Mon, Sep 22, 2008 at 15:46 Subject: Re: [Haskell-cafe] hGetContents and lazyness To: Max Vasin <max.vasin(a)gmail.com> Why don't you use OpenFile? > getPackageList packageFile = do > h <- OpenFile packageFile ReadMode > c <- hGetContents h > return $ map (drop 10) $ filter (startsWith "Filename:") $ lines c I am not at home so I don't have GHC around here, but this usually works for me! Anyway, it also depends on who is consuming the IO [FilePath] data returned. It will not be evaluated unless asked for! On Mon, Sep 22, 2008 at 14:52, Max Vasin <max.vasin(a)gmail.com> wrote: > Hello, haskellers! > > Suppose we have function (it filters package filenames from apt Packages > file): > > > getPackageList :: FilePath -> IO [FilePath] > > getPackageList packageFile = withFile packageFile ReadMode $ > > \h -> do c <- hGetContents h > > return $ map (drop 10) $ filter > (startsWith "Filename:") $ lines c -- (1) > > where startsWith [] _ = True > > startsWith _ [] = False > > startsWith (x:xs) (y:ys) | x == y = startsWith xs ys > > | otherwise = False > > When, I apply it to a Packages file I (surely) get an empty list. This is > an expected result due to > lazyness of hGetContents. I tried changing line (1) to > > > return $ map (drop 10) $ filter (startsWith "Filename:") $! lines c > > or > > > return $ map (drop 10) $! filter (startsWith "Filename:") $! lines c > > with no success. > > Chaning it to > > > return $! map (drop 10) $ filter (startsWith "Filename:") $ lines c > > makes getPackageList function return several (but not all) filenames. > > What I'm missing? And how can I fix my code? > > -- > WBR, > Max Vasin. > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe > -- Rafael Gustavo da Cunha Pereira Pinto Electronic Engineer, MSc. -- Rafael Gustavo da Cunha Pereira Pinto Electronic Engineer, MSc.

1 0

RE: readFile and closing a file
by John Lato 23 Sep '08

23 Sep '08

> On Wed, 17 Sep 2008, Mitchell, Neil wrote: > >> I tend to use openFile, hGetContents, hClose - your initial readFile >> like call should be openFile/hGetContents, which gives you a lazy >> stream, and on a parse error call hClose. > > I could use a function like > withReadFile :: FilePath -> (Handle -> IO a) -> IO a > withReadFile name action = bracket openFile hClose ... > > Then, if 'action' fails, the file can be properly closed. However, there > is still a problem: Say, 'action' is a parser which produces a data > structure lazily. Then further processing of that data structure of type > 'a' may again stop before completing the whole structure, which would also > leave the file open. We have to force users to do all processing within > 'action' and to only return strict values. But how to do this? I used rnf from Control.Parallel.Strategies when dealing with a similar problem. Would it work in your case? To merge discussion from a related thread: IMO, the question is how much should a language/library prevent the user from shooting himself in the foot? The biggest problem with lazy IO, IMO, is that it presents several opportunities to do so. The three biggest causes I've dealt with are handle leaks, insufficiently evaluated data structures, and problems with garbage collection as in the naive 'mean xs = sum xs / length xs' implementation. There are some idioms that can help with the first two cases, namely the 'with*' paradigm and 'rnf', but the third problem requires that the programmer know how stuff works to avoid poor implementations. While that's not bad per se, in some cases I think it's far too easy for the unwitting, or even the slightly distracted, to get caught in traps. I'm facing a design decision ATM related to this. I can use something like lazy bytestrings, in which the chunkiness and laziness is reified into the datastructure, or an Iterator-style fold for consuming data. The advantage of the former approach is that it's well understood by most users and has proven good performance, while on the downside I could see it easily leading to memory exhaustion. I think the problem with lazy bytestrings, in particular, is that the foldChunks is so well hidden from most consumers that it's easy to hold references that prevent consumed chunks from being reclaimed by the GC. When dealing with data in the hundreds of MBs, or GB range, this is a problem. An Enumerator, on the other hand, makes the fold explicit, so users are required to think about the best way to consume data. It's much harder to unintentionally hold references. This is quite appealing. Based on my own tests so far performance is certainly competitive. Assuming a good implementation, handle leaks can also be prevented. On the downside, it's a very poor model if random access is required, and users aren't as familiar with it, in addition to some of the questions Don raises. Back onto the topic at hand - 'action' could be a parser passed into an enumerator. Since it would read strictly, the action could end the read when it has enough data. That's another approach that I think would work with your problem. Well, that's my 2cents. John Lato

2 1

Ropes
by Rafael Gustavo da Cunha Pereira Pinto 22 Sep '08

22 Sep '08

Hi all, Is there any implementation of the rope data structure in Haskell? I couldn't find any on Hackage, and I am intending to implement it. Regards, Rafael Gustavo da Cunha Pereira Pinto Electronic Engineer, MSc.

5 6

Munging wiki articles with tagsoup
by Gwern Branwen 22 Sep '08

22 Sep '08

Hiya Neil. So recently I've been trying to come up with some automated system to turn The Monad Reader articles like those in <http://sneezy.cs.nott.ac.uk/darcs/TMR/Issue11> into wiki-formatted articles for putting on Haskell.org. Thus far, I've had the most success with SVN Pandoc. Pandoc does a good job - you can see an example conversion at <http://haskell.org/haskellwiki/?title=User:Gwern/kenn&oldid=22808>. Modulo the errors which are largely due to haskell.org problems and a few limitations in Pandoc (no comments, no real support for references), it's fine. But Pandoc's author will not support <haskell></haskell> tags inasmuch as they are an extension to MediaWiki and not universal; he prefers <pre> or <pre class="haskell"> tags. He suggested I use TagSoup to convert them into <haskell> tags. Well, alright. They're tags, TagSoup does tags - seems natural. After an hour, I came up with a nice clean little script: ---- import Text.HTML.TagSoup.Render import Text.HTML.TagSoup main :: IO () main = interact convertPre convertPre :: String -> String convertPre = renderTags . map convertToHaskell . canonicalizeTags . parseTags convertToHaskell :: Tag -> Tag convertToHaskell x | isTagOpenName "pre" x = TagOpen "haskell" (extractAttribs x) | isTagCloseName "pre" x = TagClose "haskell" | otherwise = x where extractAttribs :: Tag -> [Attribute] extractAttribs (TagOpen _ y) = y extractAttribs _ = error "The impossible happened." ---- On an aside, may I note that TagSoup doesn't seem to support transformations particularly well? Or if it does, I didn't notice any examples. I spent most of my time just figuring out how to convert the 'x' from a <pre>stuff to <haskell>stuff. Also, it might be nice to define an 'interact' alike, which is (String -> String), and defined, I supposed, as 'interact f = renderTags . f . canonicalizeTags . parseTags'. Extraction functions would be good as well - you'd only need 3 groups, I think; 1 for the 2 items in TagOpen, 1 for TagPosition's 2 positions, and 1 which extracts the String from the rest. Anyway, so my script seems to work. I ran the wiki output through it and this is the diff: <http://haskell.org/haskellwiki/?title=User%3AGwern%2Fkenn&diff=22827&oldid=…>. Ok, good, it replaces all the tags... But wait, what's all this other stuff? It is replacing all my apostrophes with '! No doubt this has something to do with XML/HTML/SGML or whatever, but it's not ideal. Even if it doesn't break the formatting (as I think it does), it's still cluttering up the source. So, how can I fix this? Am I just barking up the wrong tree and should be writing a simple-minded search-and-replace sed script which replaces <pre> with <haskell>, </pre> with </haskell>...? -- gwern USS Enforcers SORO Morwenstow MOD Albright MI5 AOL 701 GCHQ

2 3

a question about concurrent haskell
by Manlio Perillo 22 Sep '08

22 Sep '08

Hi. I have a question about concurrent Haskell in GHC. Suppose I want to write a native pure Haskell PostgreSQL client. I have done this for Python (using Twisted): http://hg.mperillo.ath.cx/twisted/pglib/ and I would like to do this in Haskell, as an exercise. The main problem is with multiple concurrent queries to the same connection, from multiple threads. PostgreSQL supports multiple requests but it's better to execute only one request at a time: http://www.postgresql.org/docs/8.3/interactive/protocol-flow.html#AEN73647 In the Twisted version I queue all the requests, and every time a request completes I call the callback associated with the request and execute the next available request. But how do I solve this problem with concurrent Haskell? Suppose thread A execute a query, and while this query is still active, a thread B execute another query. I should suspend thread B (calling yield), but how do I resume it when the query executed by thread A completes? The GHC concurrent Haskell does not have a function to resume a given thread (as found, as an example, in Lua). P.S.: another question. Why, in ghci, every time I call myThreadId, I get a different value? Prelude Control.Concurrent> myThreadId ThreadId 40 Prelude Control.Concurrent> myThreadId ThreadId 41 Thanks Manlio Perillo

5 5

readFile and closing a file
by Henning Thielemann 22 Sep '08

22 Sep '08

Say I acquire a text by readFile, feed it to a lazy parser and the parser stops reading because of a parse error. After the parser error I won't fetch more characters from the text file, but readFile does not get to know that and cannot close the file. Actually, when encountering a parser error I could read the remaining characters from the file in order to let readFile close the file eventually, but this sounds rather inefficient. So, what is the right way to go, or do I have to stay away from readFile (and maybe unsafeInterleaveIO at all)?

4 6

ANNOUNCE: citeproc-hs, a Haskell implementation of the Citation Style Language designed for Pandoc
by Andrea Rossato 22 Sep '08

22 Sep '08

Hello, I'm happy to announce the first release of citeproc-hs, a Haskell implementation of the Citation Style Language. citeproc-hs adds to Pandoc, the famous Haskell text processing tool, a Bibtex like citation and bibliographic formatting and generation facility. ABOUT The Citation Style Language (CSL) is an XML language for specifying citation and bibliographic formatting, similar in principle to BibTeX .bst files or the binary style files in commercial products like Endnote or Reference Manager. CSL is used by Zotero for bibliographic style formatting, and a huge number of CSL styles have been developed by the Zotero community. There are plans to use CSL in the future release of OpenOffice: http://bibliographic.openoffice.org/ citeproc-hs is a library that exports functions to parse CSL styles and MODS collections, to process lists of citation groups and to format the processed output. The output is a Haskell data type that can be further processed for conversion to any kind of formats (at the present time plain ASCII and the Pandoc native format) citeproc-hs was developed in order to add to Pandoc Bibtex like citations and automatic reference and bibliography generation. More information, with installation instructions, can be found here: http://code.haskell.org/citeproc-hs/ DOWNLOADS citeproc-hs can be downloaded from Hackage: http://hackage.haskell.org/cgi-bin/hackage-scripts/package/citeproc-hs To get the darcs source run: darcs get http://code.haskell.org/citeproc-hs/ KNOWN ISSUES citeproc-hs is in an early stage of development and the CSL is not complete yet. Specifically, citation collapsing is not implemented, and some formatting options are not working neither. The MODS parser needs some refinement too. BUG REPORTS To submit bug reports you can you the Google code bug tracking system available at the following address: http://code.google.com/p/citeproc-hs/issues CREDITS Bruce D'Arcus, the author of CSL, has been very kind and patient with me when I was trying to understand the CSL schema, and provided me with ideas, comments and suggestions that made it possible to come to something usable. John MacFarlane, the author of Pandoc, has been very supportive of the project and provided a lot of useful feed back, comments and suggestions. Hope you'll enjoy, Andrea Rossato

2 2

Re: [Haskell-cafe] Line noise
by Richard A. O'Keefe 22 Sep '08

22 Sep '08

On 22 Sep 2008, at 9:19 am, Albert Y. C. Lai wrote: > Everyone should use exclusively ancient Egyptian iconography. It is > the only universally readable language. I know this was full of (:-) (:-), but it seems like the perfect chance to mention Blissymbols. http://unicode.org/roadmaps/smp/smp-4-14.html (the Unicode SMP road map) has a link to a proposal for adding them to Unicode, and they _are_ in active use... (A Haskell program in Blissymbols would be, well, bigger.)

1 0

Bird-style and blank lines
by Paulo Tanimoto 22 Sep '08

22 Sep '08

Hi, I'm a big fan of literate haskell, especially Bird-style, but there's one behavior that I find rather annoying: the requirement that code be preceded and followed by a blank line. Quoth the Report: "To capture some cases where one omits an ">" by mistake, it is an error for a program line to appear adjacent to a non-blank comment line, where a line is taken as blank if it consists only of whitespace." While that is something I do anyway, since it looks better, there are some cases where it would be nice to relax that. For example: 1. I see a blog post in literate style, but the author did not provide a .lhs file, sometimes copying and pasting the text will result in code being adjacent to the text. Then I have to fix that manually. 2. It would be nice for prototyping if you could quickly comment out code just by removing a ">" character, instead of inserting "--". 3. Pandoc, via markdown, is a very interesting alternative to latex. Right now you can write a block of code like this: ~~~ > fact :: Integer -> Integer > fact 0 = 1 > fact n = n * fact (n-1) ~~~ except that that won't work, precisely because you must have a blank line before and after the code. Unfortunately, when you go and add the blank lines, the output will show them. You may tell me to change what Pandoc does, but my point is that getting any markup to work with Bird-style will have this issue. The only argument I can see for keeping the requirement is what the report says, but feel free to provide others [1]. However, isn't that a bit like saying "we will require two semicolons because forgetting a semicolon is the number one source of mistake for our programmers"? Besides, I suspect most of us uses an editor with syntax highlighting (vi, emacs, yi, etc), and then it's immediately obvious you forgot something. Would it be hard to change this behavior in GHC? Is this open for discussion for Haskell Prime? Thanks, Paulo [1] I don't think it's that common to add a ">" by accident.

3 4

Language.Haskell and strings
by Maurício 22 Sep '08

22 Sep '08

Hi, I'm using Language.Haskell.* and would like to know if it's possible to pretty-print big strings like this: "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa" into something like this: "aaaaaaaa\ \aaaaaaaa\ \aaaaaaaa\ \aaaaaaaa\ \aaaaaaaa\ \aaaaaaaa\ \aaaaaaaa" to respect the limit on number of columns. Can you help me? Is it possible to do that? Thanks, Maurício

3 3

Parsing arguments and reading configuration
by Magnus Therning 22 Sep '08

22 Sep '08

Hi all, I'm looking for some inspiration for an elegant solution to a silly little problem I have. This might have a general well-known solution, or maybe there's something particularly elegant possible in Haskell. I just thought I'd ask. When writing a command line tool I want to use a configuration file and then have the ability to override the configuration using command line arguments. When I've worked with command line arguments before I've used the trick of folding (>>=) over a list of functions that modify the "members" of a type, using the default values as the starting point. I like that, it's cute. First I thought I'd treat the configuration in a similar way, but then I noticed a slight ordering problem. The command line arguments should take priority over the contents of the configuration file, but the location of the configuration can be given as an argument. I could read the arguments twice, first to get the correct location of the config file, then load the config, and then read the arguments again to make sure they take priority. But that feels a little silly. Are there any more elegant solutions people are using? /M -- Magnus Therning (OpenPGP: 0xAB4DFBA4) magnus＠therning．org Jabber: magnus＠therning．org http://therning.org/magnus Haskell is an even 'redder' pill than Lisp or Scheme. -- PaulPotts

4 5

broken link in documentation
by Manlio Perillo 22 Sep '08

22 Sep '08

Hi. Sorry if I use the mailing list for this, but in the documentation of Control.Monad.RWS (and the other Control.Monad.* modules), the link http://www.cse.ogi.edu/~mpj/ is broken. Manlio Perillo

2 2

Updated formlets sample?
by Martin Huschenbett 22 Sep '08

22 Sep '08

Hi all, I found a blog post concerning formlets [1] in the web. Since looks very interesting I tried to compile the sample code with recent versions of HAppS and formlets from hackage. But this didn't work as the API of formlets has changed since this post. I tried to adopt the code to the new API but I was unable to finish this since there is a new monadic context I don't know to handle in the right way. So my question is, is there an updated version of this sample code in the web or has anybody tried to adopt it and can send me the results? Thanks in advance, Martin. [1] http://blog.tupil.com/formlets-in-haskell/

2 2

Drawing an existing image into a cairo surface?
by Rafal Kolanski 22 Sep '08

22 Sep '08

Hi, I'm trying to, in addition to all the other things on my cairo surface, pull some images in from .png files. I've googled, I've looked in the docs, but there isn't any way I can see to load a .png file into a surface. The best I can find is withImageSurfaceFromPNG, but I can't make it work. The C guys can do: image = cairo_image_surface_create_from_png("blah.png"); and then: cairo_set_source_surface(cr, image, 10, 10); cairo_paint(cr) I tried playing around with unsafePerformIO but that just gets me into: /build/buildd/cairo-1.6.0/src/cairo-surface.c:382: cairo_surface_reference: Assertion `((*&(&surface->ref_count)->ref_count) > 0)' failed. Are any gtk2hs people around who can throw me in the right direction? Sincerely, Rafal Kolanski.

4 7

The ultimate insult?
by Andrew Coppin 22 Sep '08

22 Sep '08

"Delphi is usable. If you're a single hobbiest programmer, use whatever gets the job done. But as far as a career goes, I'd say even Haskell has better prospects than Delphi." So "even Haskell" is better? Ouch! Still, in a bank-handed kind of way, I guess that means Haskell isn't as dead as Delphi yet, so it's also kind of a complement... of sorts...

2 1

haskell shootout -- mandelbrot
by Tim Newsham 22 Sep '08

22 Sep '08

Since this one's trivially parallizable, I took a crack at the mandelbrot test case. It was fairly easy to thread it on a per-line basis in Haskell without changing the original too much. It might be more efficient to break the work into larger chunks, but that would require some slightly more complicated changes so I didn't try that yet. Anyway, the code and the results are here. See README for details: http://www.thenewsh.com/~newsham/shootout/ Tim Newsham http://www.thenewsh.com/~newsham/

2 4

OpenSPARC project applicant chosen
by Duncan Coutts 22 Sep '08

22 Sep '08

I am very pleased to announce that we have chosen Ben Lippmeier for the OpenSPARC project. Congratulations Ben! Ben will spend three months hacking on GHC to make it perform well on the latest multi-core OpenSPARC chips. I would also like to thank the other people who applied. The reviewers were very impressed by the number of strong applications. About the project ----------------- http://haskell.org/opensparc/ It is a joint project between Sun Microsystems and the Haskell.org community. Sun has provided the funding for Ben to work on this full time for three months and has donated a powerful SPARC server for him and the rest of us to use. Ben will be working with Roman Leshchinskiy as a mentor and Darryl Gove as an adviser. Roman works on Data Parallel Haskell at UNSW and Darryl is a senior staff engineer in the SPARC compiler team at Sun. If you want to follow the progress we will be using the existing ghc development mailing list: http://www.haskell.org/mailman/listinfo/cvs-ghc and a corner of the ghc development wiki: http://hackage.haskell.org/trac/ghc/wiki/OpenSPARC Duncan (project coordinator)

4 4

Hugs on the iphone
by Alberto R. Galdo 21 Sep '08

21 Sep '08

Hi, is there any chance of having hugs compile for the iPhone? Cross-compiling? Compiling directly on the iPhone? Greets, Alberto

6 12

Iteratee-based IO
by oleg＠okmij.org 21 Sep '08

21 Sep '08

brian wrote: > I want to use Parsec to parse NNTP data coming to me from a handle I > get from connectTo. > One unworkable approach I tried is to get a lazy String from the > handle with hGetContents. It seems there is another approach, which is neither unsafe nor imperative. It relies neither on lazy IO nor on Handles. The input data can come from a file or from an embedded (e.g., chunk-encoded or encrypted) stream; the depth of embedding is arbitrary. The approach is naturally incremental. It permits IO interleaving without any unsafe operations. The approach is algebraic and declarative. The approach is the topic of the DEFUN08 talk in the morning of September 27. The code is already available http://okmij.org/ftp/Haskell/Iteratee/ The file http://okmij.org/ftp/Haskell/Iteratee/README.dr describes the other files in that directory. The running example is reading lines (terminated by CR, LF or CRLF) from a file descriptor and then from the chunk-encoded body. The main example illustrates multiplexing across two file descriptors and the full IO interleaving. The same line parser is used to process data from the file descriptor stream and from the embedded chunk-encoded stream, which is incrementally decoded. The whole code is Haskell98. It is not optimized at all and has no GHC-specific pragmas and options. The code has been used for the Wc program demonstrated yesterday. Perhaps the code answers the questions posed yesterday by Don. Hopefully one can see several composition modes for the iteratees and enumerators; enumerators are just iteratee transformers and compose as such. Incidentally, the operator ==<< is flipped >>==. Just like =<< (which is flipped >>=), it is like a `call-by-value application'. When a call-by-value language evaluates the application (f e), the argument e and all of its effects are executed first. Because of this analogy, I'm tempted to rename <<== into something like $. or .$ (or perhaps <$>, although the latter is taken).

3 2

Configure bug? Re: SOLVED. Re: [Haskell-cafe] Linker Errors For OpenGL / GLUT 'Hello World' Program.
by Donnie Jones 21 Sep '08

21 Sep '08

Hello Brandon, Maybe this is a bug in the configure script of Haskell OpenGL or GLUT packages? Any suggestion from the package maintainers (or someone more familiar with these packages)? Thanks! :) __ Donnie On Sat, Sep 20, 2008 at 10:13 PM, Brandon S. Allbery KF8NH < allbery(a)ece.cmu.edu> wrote: > On 2008 Sep 20, at 22:10, Donnie Jones wrote: > > ghc -package GLUT -lglut Hello1.hs -o Hello1 --- works! :) > > I'm not sure why I must specify "-package GLUT" and "-lglut" but that > prevents the linker errors. Also, shouldn't configure correctly figure out > how to link the GLUT libraries? Can someone explain? > > > You're overriding the information recorded by ghc-pkg (which is used if you > use ghc --make). > > As for why configure doesn't work, I think you'd have to let someone who > knows the HOpenGL build stuff look at your system. > > -- > brandon s. allbery [solaris,freebsd,perl,pugs,haskell] allbery(a)kf8nh.com > system administrator [openafs,heimdal,too many hats] allbery(a)ece.cmu.edu > electrical and computer engineering, carnegie mellon university KF8NH > > >

1 0

SOLVED. Re: [Haskell-cafe] Linker Errors For OpenGL / GLUT 'Hello World' Program.
by Donnie Jones 21 Sep '08

21 Sep '08

Hello, ghc -package GLUT -lglut Hello1.hs -o Hello1 --- works! :) I'm not sure why I must specify "-package GLUT" and "-lglut" but that prevents the linker errors. Also, shouldn't configure correctly figure out how to link the GLUT libraries? Can someone explain? Thank you. __ Donnie On Sat, Sep 20, 2008 at 9:47 PM, Donnie Jones <donnie(a)darthik.com> wrote: > Hello Clifford, > > Thank you for the quick reply. > I was able to get a test C program that draws a triangle with GLUT to work: > gcc -lglut triangle.o -o triangle.exe > > However, when building the Haskell GLUT 'Hello World' it uses, -lGLU to > link GLUT, but that does not work; however, from the config.log output > configure seems to think that -lGL should work. Is there a way I can change > the build to use -lglut? Maybe with configure I can specify that -lglut is > the correct way to link the GLUT libraries? > > Thank you! > __ > Donnie > > On Sat, Sep 20, 2008 at 3:42 PM, Clifford Beshers < > clifford.beshers(a)gmail.com> wrote: > It works for me. I'm running Ubuntu Hardy with some packages built by us > at SeeReason (see debs.seereason.com) > > It looks like ghc can't find your glut libraries (not the Haskell wrappers, > but the real GLUT libs). Try 'ghc -v --make ' and look at the linking lines > and see what GHC is specifying (mine shows a -lglut). Try compiling an > example C program against GLUT and see if that works. > > My package list looks like this: > beshers@blue:~/haskell$ dpkg -l \*glut\* > Desired=Unknown/Install/ > > Remove/Purge/Hold > | > Status=Not/Installed/Config-f/Unpacked/Failed-cfg/Half-inst/t-aWait/T-pend > |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: > uppercase=bad) > ||/ Name Version Description > > +++-==============-==============-============================================ > un freeglut-dev <none> (no description available) > ii freeglut3 2.4.0-6 OpenGL Utility Toolkit > ii freeglut3-dev 2.4.0-6 OpenGL Utility Toolkit development files > un glut-doc <none> (no description available) > un glutg3 <none> (no description available) > un glutg3-dev <none> (no description available) > ii libghc6-glut-d 2.1.1.1-2+3see Haskell GLUT binding for GHC > un libghc6-glut-d <none> (no description available) > un libghc6-glut-p <none> (no description available) > un libglut <none> (no description available) > un libglut-dev <none> (no description available) > pn libglut3 <none> (no description available) > un libglut3-dev <none> (no description available) > > > > On Sat, Sep 20, 2008 at 1:13 PM, Donnie Jones <donnie(a)darthik.com> wrote: > >> Hello Brandon, >> >> On Sat, Sep 20, 2008 at 1:02 PM, Brandon S. Allbery KF8NH < >> allbery(a)ece.cmu.edu> wrote: >> >>> On 2008 Sep 20, at 12:57, Donnie Jones wrote: >>> >>> checking GL/gl.h usability... yes >>> checking GL/gl.h presence... yes >>> checking for GL/gl.h... yes >>> checking OpenGL/gl.h usability... no >>> checking OpenGL/gl.h presence... no >>> checking for OpenGL/gl.h... no >>> checking GL/glu.h usability... yes >>> checking GL/glu.h presence... yes >>> checking for GL/glu.h... yes >>> checking OpenGL/glu.h usability... no >>> checking OpenGL/glu.h presence... no >>> checking for OpenGL/glu.h... no >>> >>> That looks like to me that the gl.h and glu.h header files were found and >>> are usable (in some cases). I am able to build >>> >>> >>> Yes, you have the header. But it says nothing about whether -lGLU was >>> found, and that's where your problems are. >>> Also, that's configure output, not config.log output (which has a >>> transcript of the tests, not just the summary). >>> >>> >> ### Relevant lines that include "-lGL" or "-lGLU" in config.log ### >> configure:4634: checking for GL library >> configure:4670: gcc -o conftest -g -O2 conftest.c -lGL -lm >&5 >> configure:4676: $? = 0 >> configure:4696: result: -lGL -lm >> configure:4773: checking for GLU library >> configure:4809: gcc -o conftest -g -O2 conftest.c -lglu32 -lGL -lm >> >&5 >> /usr/bin/ld: cannot find -lglu32 >> collect2: ld returned 1 exit status >> configure:4815: $? = 1 >> configure: failed program was: >> | /* confdefs.h. */ >> | #define PACKAGE_NAME "Haskell OpenGL package" >> | #define PACKAGE_TARNAME "OpenGL" >> | #define PACKAGE_VERSION "2.2.1" >> | #define PACKAGE_STRING "Haskell OpenGL package 2.2.1" >> | #define PACKAGE_BUGREPORT "sven.panne(a)aedion.de" >> | /* end confdefs.h. */ >> | #include <GL/glu.h> >> | int >> | main () >> | { >> | gluNewQuadric() >> | ; >> | return 0; >> | } >> configure:4809: gcc -o conftest -g -O2 conftest.c -lGLU -lGL -lm >&5 >> configure:4815: $? = 0 >> configure:4835: result: -lGLU -lGL -lm >> ... >> fp_cv_check_GLU_lib='-lGLU -lGL -lm ' >> fp_cv_check_GL_lib='-lGL -lm ' >> ... >> GLU_LIBS=' -lGLU -lGL -lm ' >> ... >> GL_LIBS=' -lGL -lm ' >> ... >> #define GLU_LIBS "-lGLU" ,"-lGL" ,"-lm" >> ######################################## >> >> It seems like the OpenGL and GLUT libraries are found (after -lglu32 >> fails, I am using Debian Linux). I am not sure what to try now. >> >> Thank you for the help. :) >> __ >> Donnie >> >> >

2 1

Re: Ropes
by Rafael Gustavo da Cunha Pereira Pinto 21 Sep '08

21 Sep '08

I have been told that you could pretty much literally implement the > algorithms > from the problem specification with Seq from Data.Sequence and achieve > acceptable speed (IIRC ~ one minute for generating a whole picture). > Yes, it is straightforward to implement the algorithm when using sequences. > > Are you sure that there is no unintentional bug in your implementation that > slows things down? > The test cases on the problem definition all worked, but they touch very little of the code. I added some "trace" calls, but could not see any trouble. I also did some runs "by hand" and it seemed ok. If only Gödel was wrong... :-) -- Rafael Gustavo da Cunha Pereira Pinto Electronic Engineer, MSc. -- Rafael Gustavo da Cunha Pereira Pinto Electronic Engineer, MSc.

1 0

ANNOUNCE: protocol-buffers-0.2.9 for Haskell is ready
by ChrisK 21 Sep '08

21 Sep '08

Hello one and all, Amid much rejoicing, my Haskell version of protocol-buffer is now released (version 0.2.9). What is this for? What does it do? Why? Shorter answer: It generates Haskell data types that can be converted back and forth to lazy ByteStrings that interoperate with Google's generated code in C++/Java/python. It is a pure Haskell re-implementation of the Google code at http://code.Google.com/apis/protocolbuffers/docs/overview.html which is "...a language-neutral, platform-neutral, extensible way of serializing structured data for use in communications protocols, data storage, and more." Google's project produces C++, Java, and Python code. This one produces Haskell code. The release tarball (with 3 Haskell packages inside, see README in source) is at http://hackage.haskell.org/cgi-bin/hackage-scripts/package/protocol-buffers The darcs repository has moved to http://darcs.haskell.org/packages/protocol-buffers2/ You will also need a recent ghc compiler, the "binary" package and the "utf8-string" package from hackage.haskell.org (same site as mentioned above). The source compiles to 3 things: 1) the package "protocol-buffers" with the library API 2) the package "protocol-buffers-descriptor" with the descriptor.proto code 3) The 'hprotoc' executable which is a command line program similar to 'protoc'. The "examples" sub-directory in the code has the Haskell version of the "addressbook.proto" example and is compatible with Google's similar example code. The code generated from unittest.proto (and unittest_import.proto) includes messages TestAllTypes and TestAllExtensions which have been extensively tested by QuickCheck to ensure they can be wire encoded and decoded (see the "tests" sub-directory in the code). The user API, as exported by Text.ProtocolBuffers, allows for converting messages back and forth to the lazy ByteString type. And such messages can be merged, and the defaults accessed via the MessageAPI type class. The messages in Haskell as just regular data types and are thus immutable. Required types are simple record fields, optional types are Maybe, and repeated types are Seq (from Data.Sequence). Extensions are supported via Key data that allows access to the extension fields. Extensible messages contain an opaque ext'field entry of type ExtField that contains the map data structure to contain the extension field values. The User API allows for serializing messages as the usual series of fields. It also provides for a length prefix to be written to create delimited messages. It also provides to write a wire tag with any field number before the length and message data. This last form looks like a field on the wire, and there is a special api call to read back just the one message and its field number. This last API is similar to the one that is part of the C# API. No benchmarks have been run yet. Any suggestions? Unsupported for the moment is loading and storing "unknown" fields. It can be added sooner if someone has a use for this. Unsupported indefinitely is code generation for Services and Methods. I have yet to look into how this is presented in the other languages. The API to read a single message field, as mentioned above, might be extended to read any type instead of just messages. optional clever_quote { <autrijus> Perl: "Easy things are easy, hard things are possible" <autrijus> Haskell: "Hard things are easy, the impossible just happened" } Cheers! Chris Kuklewicz

1 0

[ANN] panda blog engine
by Jinjing Wang 21 Sep '08

21 Sep '08

For some time, I wanted a blog engine that I can * use plain text * never lose my data * configurable at ease * write in any format I want * never care about database schema * them-able without caring about programming logic panda doesn't do all of it, and is worse at these: * no web interface, except for comments * requires a virtual host, or at least be able to start a lighttpd instance * need to : troubleshoot fcgi processes ( kill them ); use SCM to manage resources; don't hate OOP (for hacking, sorry about that) those being said, * demo: http://jinjing.blog.easymic.com * code: http://github.com/nfjinjing/panda/tree/master Cheers, - jinjing

1 0

Library design question
by Andre Nathan 21 Sep '08

21 Sep '08

Hello I'm trying to write a simple graph library for a project of mine (actually, I'm porting it from OCaml) but I've got a design question right in the beginning. My Graph type is the following. data Graph a b = Graph { adjacencies :: Map Int (a, (Map Int b)) , numVertices :: Int , numEdges :: Int } Types "a" and "b" refer to vertex and edge labels (I know it's kind of weird to use a Map of Maps to represent a graph, but it helps me removing vertices later, which is something I'll need to do.) Creating an empty graph is trivial: empty :: Graph a b empty = Graph Map.empty 0 0 The issue I hit was when writing the function to add a vertex to the graph. Since I need to update the vertex counter, I've used the state monad: addVertex :: Int -> a -> State (Graph a b) () addVertex vertex label = do g <- get let adj = Map.insert vertex (label, Map.empty) (adjacencies g) put $ g { adjacencies = adj, numVertices = numVertices g + 1 } That works fine, but from the point of view of a user of the library, the addVertex function should take a "Graph a b" as its first argument, as one would expect to be able to say which graph is going to be modified. So I'm confused right now about how I should proceed. Any hints regarding that? Best regards, Andre

8 16

Ropes
by Rafael Gustavo da Cunha Pereira Pinto 20 Sep '08

20 Sep '08

On Fri, Sep 19, 2008 at 23:18, Don Stewart <dons(a)galois.com> wrote: > RafaelGCPP.Linux: > > Hi all, > > > > Is there any implementation of the rope data structure in Haskell? > > > > I couldn't find any on Hackage, and I am intending to implement it. > > There's no mature rope implementation. Can you write a bytestring-rope > that outperforms lazy bytestrings please :) > I'll give it a try, but cannot promise anything on the "outperform" part! :-) -- Rafael Gustavo da Cunha Pereira Pinto Electronic Engineer, MSc.

1 0