December 2008 - Haskell-Cafe

Haskell Weekly News: Issue 98 - December 25, 2008
by Brent Yorgey 26 Dec '08

26 Dec '08

--------------------------------------------------------------------------- Haskell Weekly News http://sequence.complete.org/hwn/20081225 Issue 98 - December 25, 2008 --------------------------------------------------------------------------- Welcome to issue 98 of HWN, a newsletter covering developments in the [1]Haskell community. Happy holidays! An exciting HWN for you this week, including a number of cool new libraries, the [2]public release of [3]Cryptol, a [4]Haskell logo contest, and the [5]second most awesome GHC bug ever (see augustss's quote at the end of the Quotes section for the most awesome GHC bug ever). Announcements Hieroglyph 0.85. Jeff Heard [6]announced that the Thingie library has been renamed [7]Hieroglyph, and now has support for displaying images on the Cairo canvas. Cryptol now freely available. Don Stewart [8]announced that [9]Cryptol, the language of cryptography, is [10]now available to the public! Cryptol is a domain specific language for the design, implementation and verification of cryptographic algorithms, developed over the past decade by Galois for the United States National Security Agency. It has been used successfully in a number of projects, and is also in use at Rockwell Collins, Inc. Cryptol is implemented in Haskell. Control.Monad.IfElse. Jeff Heard [11]announced the [12]Control.Monad.IfElse module, which provides useful anaphoric and monadic versions of if-else and when. llvm-0.4.0.1. Lennart Augustsson [13]announced version 0.4.0.1 of the [14]release that is quite incompatible with the old 0.0.2 release.) Haskell LLVM bindings. LLVM is a virtual machine and the bindings allow you to generate code for this virtual machine. This code can then be executed by a JIT or written to a file for further processing by the LLVM tools. bytestring-trie 0.1.0. wren ng thornton [15]announced the release of [16]bytestring-trie 0.1.0, an efficient finite map from (byte)strings to values. The implementation is based on big-endian patricia trees, like Data.IntMap. RWH book club. Don Stewart [17]announced that Matt Podwysocki has set up a [18]Real World Haskell book club, a mailing list on google groups with already some 200 members discussing typical new user Haskell questions. Feel free to join if you like talking about Haskell, or teaching new users. Thingie-0.80. Jeff Heard [19]announced the release of [20]Thingie, a library for creating 2D visualizations in a purely functional manner. It supports static visualizations and animation, and like most vis libraries, can probably do games as well as simple viz graphics. typehash version 1.3. Lennart Augustsson [21]announced the release of the [22]typehash library, which allows you to produce a unique identifier (a cryptographic hash) for a type. This is useful if you save values of some type to a file (text, binary, whatever format you wish) and then when you read it back in again you want to verify that the type you want to read is the one you actually wrote. The library also supports type codes, which encode the complete structure of a type and can be used for finer comparison than just equality. uvector-algorithms 0.1. Dan Doel [23]announced the release of [24]uvector-algorithms, a library of algorithms (mostly sorting) for the mutable arrays defined in uvector. It has several varieties of sorting, including introsort (quicksort which falls back on heapsort in bad cases), heapsort, a simple top- down merge sort and a radix sort. Also exposed are the operations that allow you to use the arrays as heaps and a combinator for safely using these mutable array algorithms to sort immutable arrays. All algorithms have been painstakingly profiled and optimized. Data.List.Split. Brent Yorgey [25]announced the release of [26]Data.List.Split, which provides a wide range of strategies and a unified combinator framework for splitting lists with respect to some sort of delimiter. Hoogle with more libraries. Neil Mitchell [27]announced that [28]Hoogle will now search lots of the libraries present on hackage! HLint 1.0. Neil Mitchell [29]announced the re-release of [30]HLint, a tool for making suggestions to improve your Haskell code. Previously this tool was called Dr Haskell and depended on a working installation of Yhc; now it depends on GHC 6.10.1. rangemin-1.0. Louis Wasserman [31]announced the release of [32]rangemin, a library for efficiently preprocessing an array to find minimum elements of subranges of the array in constant time. Discussion length of module name affecting performance??. Daniel Gorin [33]reported a GHC bug where in certain cases, changing the name of a module to something longer results in a 2x-3x performance hit! Strange but true. Time for a new logo?. Don Stewart [34]proposed a competition to produce a new Haskell logo! Submissions should go on the [35]wiki page; the deadline for logo submissions is December 31. Pattern combinators. Andrew Wagner [36]started a thread turning a [37]paper on pattern-matching in Haskell into actual code for hackage. Coroutines. Ryan Ingram [38]posted some interesting code showing how to implement coroutines with session types. Type wildcards. Eyal Lotem [39]proposed a 'type wildcards' extension to the language. Blog noise [40]Haskell news from the [41]blogosphere. * Galois, Inc: [42]Cryptol, the language of cryptography, now available. * Dan Piponi (sigfpe): [43]The Mother of all Monads. * Real-World Haskell: [44]Pat Eyler Interviews the Real World Haskell Team. * Philip Wadler: [45]Unsafe. * Conrad Parker: [46]Release: HOgg 0.4.1. * ezekiel smithburg: [47]fast string appending/concatenation in haskell. * Brent Yorgey: [48]Data.List.Split. * Alson Kemp: [49]Turbinado: Implementing a poor-man's wiki. * Luke Palmer: [50]Reactive spaces. * Conal Elliott: [51]Smarter termination for thread racing. * Real-World Haskell: [52]RWH on Twitter. * >>> Muharem Hrnjadovic: [53]My new favourite book. * Alson Kemp: [54]Turbinado update. * Philip Wadler: [55]Type Safe Pattern Combinators, by Morten Rhiger. * Osfameron: [56]Crossword puzzles in Haskell. * Brent Yorgey: [57]QuickCheck rocks my socks. * >>> Matt Youell: [58]If programming languages were martial arts. * >>> Sebastian Fischer: [59]Haskell idioms I did not understand before hacking them on my own. * Alson Kemp: [60]Thinking about Haskell: You Know Lazy Evaluation; You Just Don't Know It. * >>> Brian McCallister: [61]Real World Haskell, for Jon. * Ivan Lazar Miljenovic: [62]RWH Arrives Down Under!. * >>> Thomas ten Cate: [63]XMonad with Ubuntu, dvorak, Pidgin and Skype. * >>> Jeremy Frens: [64]PE Problem #3 in All Languages. * Jeremy Shaw: [65]Data Migration with HApps-Data. * Chris Done: [66]Haskell Formlets: Composable web form construction and validation. * Real World Haskell: [67]Haskell around the world. * Dougal Stanton: [68]A brief look at fingertrees. Quotes of the Week * luqui: no!!! I was building a joke, but then I sent it before I thought of one. * roconnor: We put up a clothes line and made a turing machine by hand at a party once ... that is the sort of parties I go to. * cjs: In what other language could I have learned so much about Win32 programming (summary: basically, the Windows 3.0 API but with all sorts of hacks to deal with having more than one thread in the system), and come out having *enjoyed* myself? Praise to the Lord! * PaulJohnson: A paradox of the Haskell world is that, while the language is Vulcan, the community around it is dominated by Warm Fuzziness. Clearly the two are not mutually exclusive. * Botje: Caleskell even has unsafeSolveHaltingProblem? * Taejo: * Taejo needs to write Sitar Hero in Yampa * dons: it is safer for incompetent people to be working in Haskell than C++. * mpeter: the quality of my code increased drastically when i realized i should stop telling the computer to do things which were stupid. * byorgey: <Cale> RandomT/Random are effectively state monads. (in fact, they're thin candy shells around StateT/State.) <byorgey> "newtype: melts in the compiler, not in your hands" * quicksilver: [on classes having the same name as constructs in other languages] it's like having a laxative called "after dinner mint", and people being upset when they were looking for something nice to eat after dinner. * quicksilver: #haskell is a loquacracy! * quicksilver: It's also the same thing as the Yoneda lemma. That's the thing about maths. Everything is actually the same. * hugo: i feel like i was drugged with imperative programming, and now im in rehab. * chrisdone: yo dawg we heard you like haskell so we installed a lambdabot in your ghci so you can monad while you monad * augustss: ghc had a bug once where it deleted the source file if it had a type error. Quite sensible, I think. About the Haskell Weekly News New editions are posted to [69]the Haskell mailing list as well as to [70]the Haskell Sequence and [71]Planet Haskell. [72]RSS is also available, and headlines appear on [73]haskell.org. To help create new editions of this newsletter, please see the information on [74]how to contribute. Send stories to byorgey at cis dot upenn dot edu. The darcs repository is available at darcs get [75]http://code.haskell.org/~byorgey/code/hwn/ . References 1. http://haskell.org/ 2. http://www.galois.com/blog/2008/12/24/cryptol-the-language-of-cryptography-… 3. http://www.cryptol.net/ 4. http://haskell.org/haskellwiki/Haskell_logos/New_logo_ideas 5. http://thread.gmane.org/gmane.comp.lang.haskell.glasgow.user/16037 6. http://article.gmane.org/gmane.comp.lang.haskell.cafe/49528 7. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/Hieroglyph 8. http://article.gmane.org/gmane.comp.lang.haskell.cafe/49516 9. http://www.cryptol.net/ 10. http://www.galois.com/blog/2008/12/24/cryptol-the-language-of-cryptography-… 11. http://article.gmane.org/gmane.comp.lang.haskell.cafe/49365 12. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/IfElse 13. http://article.gmane.org/gmane.comp.lang.haskell.cafe/49340 14. file://localhost/home/brent/hacking/hwn/(This 15. http://article.gmane.org/gmane.comp.lang.haskell.cafe/49234 16. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/bytestring-trie 17. http://article.gmane.org/gmane.comp.lang.haskell.cafe/49168 18. http://groups.google.com/group/real-world-haskell-book-club 19. http://article.gmane.org/gmane.comp.lang.haskell.cafe/49147 20. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/Thingie 21. http://article.gmane.org/gmane.comp.lang.haskell.cafe/48902 22. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/typehash 23. http://article.gmane.org/gmane.comp.lang.haskell.cafe/48889 24. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/uvector%2Dalgori… 25. http://article.gmane.org/gmane.comp.lang.haskell.general/16700 26. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/split 27. http://article.gmane.org/gmane.comp.lang.haskell.general/16698 28. http://haskell.org/hoogle 29. http://article.gmane.org/gmane.comp.lang.haskell.general/16692 30. http://www-users.cs.york.ac.uk/~ndm/hlint/ 31. http://article.gmane.org/gmane.comp.lang.haskell.general/16686 32. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/rangemin 33. http://thread.gmane.org/gmane.comp.lang.haskell.glasgow.user/16037 34. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/48892 35. http://haskell.org/haskellwiki/Haskell_logos/New_logo_ideas 36. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/49267 37. http://www.itu.dk/people/mir/typesafepatterns.pdf 38. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/49108 39. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/48987 40. http://planet.haskell.org/ 41. http://haskell.org/haskellwiki/Blog_articles 42. http://www.galois.com/blog/2008/12/24/cryptol-the-language-of-cryptography-… 43. http://sigfpe.blogspot.com/2008/12/mother-of-all-monads.html 44. http://www.realworldhaskell.org/blog/2008/12/24/pat-eyler-interviews-the-re… 45. http://wadler.blogspot.com/2008/12/unsafe.html 46. http://blog.kfish.org/2008/12/release-hogg-041.html 47. http://tehgeekmeister.wordpress.com/2008/12/22/fast-string-appendingconcate… 48. http://byorgey.wordpress.com/2008/12/21/datalistsplit/ 49. http://www.alsonkemp.com/haskell/turbinado-implementing-a-poor-mans-wiki/ 50. file://localhost/home/brent/hacking/hwn/20081225.html 51. http://conal.net/blog/posts/smarter-termination-for-thread-racing/ 52. http://www.realworldhaskell.org/blog/2008/12/18/rwh-on-twitter/ 53. http://muharem.wordpress.com/2008/12/18/my-favourite-book/ 54. http://www.alsonkemp.com/haskell/turbinado-update/ 55. http://wadler.blogspot.com/2008/12/type-safe-pattern-combinators.html 56. http://greenokapi.net/blog/2008/12/18/crossword-puzzles-in-haskell/ 57. http://byorgey.wordpress.com/2008/12/16/quickcheck-rocks-my-socks/ 58. http://www.youell.com/matt/writing/?p=46 59. http://www-ps.informatik.uni-kiel.de/~sebf/projects/haskell-idioms-i-did-no… 60. http://www.alsonkemp.com/?p=110 61. http://kasparov.skife.org/blog/src/haskell/real-world-haskell-for-jon.html 62. http://ivanmiljenovic.wordpress.com/?p=30 63. http://typethinker.blogspot.com/2008/12/xmonad-with-ubuntu-dvorak-pidgin-an… 64. http://jdfrens.blogspot.com/2008/12/pe-problem-3-in-all-languages.html 65. http://nhlab.blogspot.com/2008/12/data-migration-with-happs-data.html 66. http://chrisdone.com/blog/2008/12/14/Haskell-Formlets:-Composable-web-form-… 67. http://www.realworldhaskell.org/blog/?p=154 68. http://www.dougalstanton.net/blog/index.php/2008/12/12/a-brief-look-at-fing… 69. http://www.haskell.org/mailman/listinfo/haskell 70. http://sequence.complete.org/ 71. http://planet.haskell.org/ 72. http://sequence.complete.org/node/feed 73. http://haskell.org/ 74. http://haskell.org/haskellwiki/HWN 75. http://code.haskell.org/~byorgey/code/hwn/

2 1

How does one use Text.Regex.Base.RegexLike?
by Lyle Kopnicky 26 Dec '08

26 Dec '08

I'm trying to migrate code from using the old Text.Regex to the new Text.Regex.Base. But, I'm getting type errors. I can't even create a regex. Looking at the docs, it seems like this should print "bcd": import Data.Array import Text.Regex.Base import Text.Regex.Posix rx = makeRegex "a(.*)A" Just (_, mt, _) = matchOnceText rx "abcdA" main = putStrLn (fst (mt ! 0)) But I get an error: src\regex.hs:5:5: No instance for (RegexMaker regex compOpt execOpt [Char]) arising from a use of `makeRegex' at src\regex.hs:5:5-22 Possible fix: add an instance declaration for (RegexMaker regex compOpt execOpt [Char]) In the expression: makeRegex "a(.*)A" In the definition of `rx': rx = makeRegex "a(.*)A" src\regex.hs:7:18: No instance for (RegexLike regex [Char]) arising from a use of `matchOnceText' at src\regex.hs:7:18-41 Possible fix: add an instance declaration for (RegexLike regex [Char]) In the expression: matchOnceText rx "abcdA" In a pattern binding: Just (_, mt, _) = matchOnceText rx "abcdA" Why does it say there is no instance? Isn't the instance imported by Text.Regex.Posix? Why in the world is it so complicated just to get a matched substring out of the text? Is there an easier way? Thanks, Lyle

4 5

forkIO on multicore
by Paul Keir 25 Dec '08

25 Dec '08

Hi all, I'm seeing no performance increase with a simple coarse-grained 2-thread code using Control.Concurrent. I compile with: > hc conc.hs -o conc --make -threaded and I run with > time ./conc +RTS -N2 But using either "-N1" or "-N2", the program runs in about 1.8secs. (I'd prefer a longer running thread task, but my fib function currently runs out of memory). Anyway, my program is below, and I'm using GHC version 6.8.2 on a 2-core Pentium D. Can anyone help? module Main where import Control.Concurrent fibs = 0 : 1 : zipWith (+) fibs (tail fibs) heavytask m = putMVar m (fibs !! 100000) main = do ms <- sequence $ replicate 2 newEmptyMVar mapM_ (forkIO . heavytask) $ tail ms heavytask $ head ms ms' <- mapM takeMVar ms mapM_ print ms' Regards, Paul

7 14

Hackage packages for indexing by Google
by Henning Thielemann 25 Dec '08

25 Dec '08

When testing the ShowMeta parallel web: http://hackage.haskell.org.showmeta.parallelnetz.de/packages/archive/pkg-li… I found, that all HackageDB pages forbid crawling and indexing. Actually, the content of this page is dynamic - but it changes only incrementally. It would be very worth when I search for "markov chain haskell" with Google, that I will be directed to the according package page of HackageDB.

1 0

ANN: Hieroglyph 0.85
by Jeff Heard 25 Dec '08

25 Dec '08

I have changed the name of Thingie to Hieroglyph and added support for displaying images on the Cairo canvas. I'm still working on using template haskell to derive the basic UIState class, but I don't understand template haskell all that well, so it's taking some time. Hopefully I can get some examples of the library up over the christmas holiday. -- Jeff

1 0

Re: [Haskell-cafe] Coroutines
by Ryan Ingram 25 Dec '08

25 Dec '08

On Thu, Dec 18, 2008 at 2:00 AM, Nicolas Pouillard <nicolas.pouillard(a)gmail.com> wrote: > I don't see why one would need session types, channels... to express that. > I maybe need a more complicated coroutines (ruby) example that would require > using this system. OK, how would you type these routines in Haskell? def simple yield "hello" yield 1 yield (lambda { |x| x + 1 }) end def useSimple state = 0 result = nil simple { |x| if (state == 0) then result = x else if (state == 1) then result += (x * 4).toString else if (state == 2) then result += x.call(10).toString state = state + 1 } result end I know it's a bit contrived, but you get the idea. In Haskell using Control.Coroutine: simple :: forall rest. Session (String :!: Int :!: (Int -> Int) :!: rest) rest () simple = do put "hello" put 1 put (\x -> x + 1) useSimple :: forall rest. Session (String :?: Int :?: (Int -> Int) :?: rest) rest String useSimple = do string <- get int <- get func <- get return (string ++ show (int * 4) ++ show (func 10)) result :: String result = snd $ connects simple useSimple -- result = "hello411"

7 8

Stupid question #374: why is MaybeT not in the standard library?
by Brian Hurt 25 Dec '08

25 Dec '08

I wrote my own implementation of MaybeT (which was a usefull exercise), but a quick google showed: http://www.haskell.org/haskellwiki/New_monads/MaybeT But I'm wondering why it's not in the standard library. The standards committee just hasn't gotten around to it yet? Or was there some discussion of this in the past on some (public) maillist, that my admittedly shallow googling failed to uncover, that someone could point me at? Thanks. Brian

6 9

ANN: Control.Monad.IfElse
by Jeff Heard 25 Dec '08

25 Dec '08

Provides useful anaphoric and monadic versions of if-else and when, as well as infix operators for the actions to allow things like this: -- Allow the clause to be in m a whenM (Gtk.widgetIsFocus win) $ do this that -- anaphoric if. If the condition is nonempty, pass the conditional value to the "then" clause, otherwise do an action that requires no parameter. aif (lookup name list) (\file -> Right `liftM` hGetLine file) (do f <- openFile name l <- hGetLine file return $ Right (f,l)) -- infix version of Control.Monad.When (x>5) >>? putStrLn x -- infix version of anaphoric when lookup name list >>=? \value -> putStrLn value -- infix version of anaphoric whenM Gtk.widgetIntersect win rect >>=>>? \intersectingRectangle' -> Gtk.invalidate intersectingRectangle'

2 1

ANNOUNCE: Cryptol now freely available
by Don Stewart 25 Dec '08

25 Dec '08

Galois is pleased to announce that Cryptol, the language of cryptography, is now available to the public! Cryptol is a domain specific language for the design, implementation and verification of cryptographic algorithms, developed over the past decade by Galois for the United States National Security Agency. It has been used successfully in a number of projects, and is also in use at Rockwell Collins, Inc. Cryptol is implemented in Haskell. Domain-specific languages (DSLs) allow subject-matter experts to design solutions in using familiar concepts and constructs. Cryptol, as a DSL, allows domain experts in cryptography to design and implement cryptographic algorithms with a high degree of assurance in the correctness of their design, and at the same time, producing a high performance implementation of their algorithms. In this release, Galois has made a implementation of the Cryptol language available free of charge for non-commercial uses. More information on the blog: http://www.galois.com/blog/2008/12/24/cryptol-the-language-of-cryptography-… And at http://www.cryptol.net Enjoy!

1 0

When using functional dependencies to combine...
by Martijn van Steenbergen 24 Dec '08

24 Dec '08

Good evening everyone, My program reads: > module Boom where > > import Control.Monad.State > > type SucParser s = StateT [s] [] > > newtype WithUnit s a = WithUnit (SucParser s (a, ())) > > foo :: SucParser s [s] > foo = get > > bar :: WithUnit s [s] > bar = WithUnit get The compiler complains: Boom.hs:13:0: Couldn't match expected type `([s], ())' against inferred type `[s]' When using functional dependencies to combine MonadState s (StateT s m), arising from the instance declaration at <no location info> MonadState ([s], ()) (StateT [s] []), arising from a use of `get' at Boom.hs:13:15-17 When generalising the type(s) for `bar' I'm wondering if I'm making a silly mistake or if there's something less trivial going on here. Could someone please explain the error and give a hint on how to fix it? Thanks much. :-) Martijn.

2 3