February 2009 - Haskell-Cafe

ANN: HLint 1.2
by Neil Mitchell 07 Feb '09

07 Feb '09

Hi, I am pleased to announce HLint version 1.2. HLint is a lint-like tool for Haskell that detects and suggests improvements for your code. HLint is compatible with most GHC extensions, and supports a wide variety of suggestions, and can be extended with additional user suggestions. To install: cabal update && cabal install hlint Web page: http://www-users.cs.york.ac.uk/~ndm/hlint/ Biggest new feature: List recursion suggestions. For example, running HLint over the GHC compiler now suggests things such as: Example.hs:3:1: Warning: Use foldr Found: seqList [] b = b seqList (x : xs) b = x `seq` seqList xs b Why not: seqList xs b = foldr seq b xs Example.hs:6:1: Warning: Use foldr Found: seqIds [] = () seqIds (id : ids) = seqId id `seq` seqIds ids Why not: seqIds ids = foldr (seq . seqId) () ids Example.hs:13:1: Warning: Use foldl Found: rev_app [] xs = xs rev_app (y : ys) xs = rev_app ys (y : xs) Why not: rev_app ys xs = foldl (flip (:)) xs ys HLint will automatically detect if you should have used a map, a foldr or a foldl and suggest how to change your code. In the GHC, darcs and Hoogle code bases there are no obvious map-like functions, which is a good sign :-) Changes from last time: * More GHC extensions supported - including record wild cards and fixes to a number of existing extensions. These changes are thanks to Niklas and the haskell-src-exts package. * Many infix operators are now correctly associated with the right priority. * All hints are now ignored, warnings or errors. Hopefully this should make it clearer what hints are most important. As part of this, the format for defining ignore files has changed completely - see the manual for the new format. * Lots of new hints have been added - many from users of HLint. * Many bugs have been fixed. Can any follow-up discussions please be directed to haskell-cafe@. All comments, questions, bug reports are welcome! Neil

13 33

Function const (Binding)
by TKM 07 Feb '09

07 Feb '09

Hello, I've a small question about the function const. I'm a bit of confused about how it binds. Let me take the following expression as example: const id 1 2 If I execute this expression, I will get as answer 2 with Helium. Now is my question, why doesn't it give me 1 as the answer? Because the type of id would be: a -> a. So first it would execute id 1 in my opinion. That gives us 1. And after executing const 1 2 it should give us 1. Can somebody explain to me why it does not bind as I expect? (I know I can do: const (id 1) 2 to get what I want) Thank you for your answers. Greetz TKM

3 2

HSGI: Haskell Server Gateway Interface
by Manlio Perillo 07 Feb '09

07 Feb '09

Lately I have started to think about how I would like to implement a web application framework in Haskell. To keep my thoughts well defined, I have decided to write some documents. My ideas are heavily based on my experience with Python WSGI, and a Python WSGI framework I'm implementing (wsgix). Note that there is already another attempt to define an Haskell Gateway interface, and it may also have the same name! However I'm not trying to write a formal specification. The core of my idea is in the Haskell Web Application Framework, I just need a gateway interface as a building block. I apologize for my not very good experience at writing in english and Haskell. Also note that the document is in a very rough state. I'm announcing it here, to receive feedback, just to make sure that what I would like to do is possible/convenient to do in Haskell. The documents (only the first one - HSGI - available at the moment) are available here: http://mperillo.ath.cx/haskell/ Thanks and regards Manlio Perillo

2 4

morphisms in IO
by Gregg Reynolds 07 Feb '09

07 Feb '09

I'm working on a radically different way of looking at IO. Before I post it and make a fool of myself, I'd appreciate a reality check on the following points: a) Can IO be thought of as a category? I think the answer is yes. b) If it is a category, what are its morphisms? I think the answer is: it has no morphisms. The morphisms available are natural transformations or functors, and thus not /in/ the category. Alternatively: we have no means of directly naming its values, so the only way we can operate on its values is to use morphisms from the outside (operating on construction expressions qua morphisms.) c) All categories with no morphisms ("bereft categories"?) are isomorphic (to each other). I think yes. Thanks, gregg

4 4

Haskell Weekly News: Issue 104 - February 7, 2009
by Brent Yorgey 07 Feb '09

07 Feb '09

--------------------------------------------------------------------------- Haskell Weekly News http://sequence.complete.org/hwn/20090207 Issue 104 - February 07, 2009 --------------------------------------------------------------------------- Welcome to issue 104 of HWN, a newsletter covering developments in the [1]Haskell community. Community News Andre Pang (ozone) will be soon [2]moving to San Fransisco to begin work with [3]Pixar! Announcements Mutually recursive modules. Henning Thielemann [4]announced a [5]small writeup explaining how mutually recursive modules are currently supported, and how they can be avoided. Please add information about other compilers and more ideas on breaking cycles. UrlDisp, a friendly URL dispatching library. Artyom Shalkhakov [6]announced the first release of [7]UrlDisp, a small library for URL dispatching (aka routing). Right now it works with CGI, and should be compatible with FastCGI as well (not tested); Happstack compatibility is planned. [8]Documentation and usage examples are available. Purely functional LU decomposition. Rafael Gustavo da Cunha Pereira Pinto [9]released some code to perform purely functional LU decomposition. Ready for testing: Unicode support for Handle I/O. Simon Marlow [10]announced that proper Unicode support in Handle I/O is ready for testing in GHC. Just download the [11]set of patches, compile GHC with them, and test away! Comments and discussion welcome. HaskellWiki Accounts. Ashley Yakeley can [12]create a HaskellWiki account for anyone who wants one (account creation has been disabled as a spam-fighting measure). multiplicity 0.1.0 released. Dino Morelli [13]announced the release of [14]multiplicity 0.1.0, a configuration file driven wrapper around [15]duplicity. It allows you to easily define backup sets as config files and avoid long, repetitive command lines. Happstack 0.1 Released!. Matthew Elder [16]announced the [17]0.1 release of [18]Happstack, the successor for the HAppS project. #haskell-in-depth IRC channel. Philippa Cowderoy [19]announced the creation of a new IRC channel, #haskell-in-depth. The new channel is open to everyone, just like #haskell, but is intended for more in-depth conversations, to allow the #haskell channel to be a more newbie-friendly place. regex-posix-unittest-1.0 AND regex-posix-0.94.1 AND regex-tdfa-0.97.1. ChrisK [20]announced an update to the [21]regex-posix package which provides better semantics for multiple matches; an update to the [22]regex-tdfa package, which provides the same new multiple match semantics and fixes a bug; and finally, a new package, [23]regex-posix-unittest, along with an [24]accompanying wiki page; it runs a suite of unit tests which regex-tdfa passes, but reveals bugs in the standard glibc, OS X, FreeBSD, and NetBSD implementations! Jane Street Summer Project 2009. Yaron Minsky [25]announced the [26]Jane Street Summer Project for 2009, the goal of which is to make functional programming languages into better practical tools for programming in the real world. Students will be funded over the summer to work on open-source projects which aim at improving the practical utility of their favorite functional language. gitit 0.5.1. John MacFarlane [27]announced the release of [28]gitit 0.5.1, a wiki program that uses git or darcs as a filestore and HAppS as a server. Changes include major code reorganization, bug fixes, new debugging features, and more. regex-xmlschema. Uwe Schmidt [29]announced the [30]release of [31]regex-xmlschema, (yet another) package for processing text with regular expressions, containing a complete implementation of the W3C XML Schema specification language for regular expressions. diagrams 0.2. Brent Yorgey [32]announced version 0.2 of the [33]diagrams package, an embedded domain-specific language for creating simple graphics in a compositional style. New features include support for arbitrary paths, text, multiple output formats, and support for the [34]colour library. Discussion Haddock Markup. David Waern began a [35]discussion on Haddock markup syntax: should it support (La)TeX for embedded mathematics? Should it support other stuff? Elegant & powerful replacement for CSS. Conal Elliott began a [36]discussion on an elegant replacement for CSS that is consistent, composable, orthogonal, functional, and based on an elegantly compelling semantic model---what might such a thing look like? type metaphysics. Gregg Reynolds began a long and interesting [37]discussion on the type system, denotational semantics, and related matters. Jobs Postdoc Positions at the CLIP group, Spain. CFP [38]announced the availability of [39]postdoctoral research positions within the [40]CLIP (Computational Logic, Implementation and Parallelism) group in Madrid, Spain. The application deadlines are February 13th and 18th; see the original email for more details. Multiple funded Ph.D. positions available. Martin Erwig [41]announced the availability of multiple funded Ph.D. positions in the [42]school of EECS at Oregon State University, in the areas of programming languages (focusing on DSLs and language design), software engineering, and HCI. If you are interested, email Martin with a resume and contact information by February 15. Blog noise [43]Haskell news from the [44]blogosphere. * Erik de Castro Lopo: [45]libsndfile 1.0.18.. * >>> Gregg Reynolds: [46]Moggi :: CT -> Hask. An interesting perspective on Haskell and category theory. * Osfameron: [47]Functional Pe(a)rls v2 (now with Monads!) at the London Perl Workshop 2008. * ezekiel smithburg: [48]google code project, google group, and darcs repository for WikimediaParser. * ezekiel smithburg: [49]releasing WikimediaParser 0.1. * Andre Pang (ozone): [50]Change. * Lennart Augustsson: [51]Is Haskell fast?. * Lennart Augustsson: [52]Regression. A library for writing embedded BASIC programs in Haskell? What is the world coming to? * Galois, Inc: [53]Equivalence and Safety Checking in Cryptol. * GHC / OpenSPARC Project: [54]Just Testing. * Jeff Heard: [55]Hieroglyph: Interactively graph a spreadsheet in 99 lines of code.. * Jeff Heard: [56]Snippet: sorting the elements of a map. * LHC Team: [57]Grin a little.. * Alex Mason: [58]Other shootout news. * Xmonad: [59]contribs review: SpawnOn. * Jeff Heard: [60]Hieroglyph Haddock docs. * Alex Mason: [61]More n-bodies speedups. * Alex Mason: [62]N-bodies evolution. * Eric Kow (kowey): [63]practical quickcheck (wanted). Eric is seeking practical advice on how to effectively use QuickCheck. * Alex Mason: [64]TextMate haskell bundle improvements. * Haskell application server stack: [65]Happstack 0.1 Released (one day early!). * Haskell application server stack: [66]SOCALFP presentation: patch-tag as a startup using Happstack. * Neil Mitchell: [67]Monomorphism and Defaulting. * Jeff Heard: [68]Hieroglyph Part II, basic mouse/keyboard interactivity. * Jeff Heard: [69]Introducing Hieroglyph. Purely functional 2D visualization on top of Cairo.. The Hieroglyph library looks pretty cool! * Jeff Heard: [70]Hieroglyph HOWTO, part I: Sparklines. * Alex Mason: [71]N-bodies speedup (50%!). * >>> John Lato: [72]Build a better WAVE reader, pt 2. * GHC / OpenSPARC Project: [73]Join Points. * Luke Palmer: [74]Parallel Rewrite System. * >>> John Lato: [75]Build a better WAVE reader, pt 1. Notes on building an 'optimal audio I/O library in Haskell'. * Luke Palmer: [76]Dana update: System U-Box compiler is go. * Luke Palmer: [77]Dana needs a self-embedding of U-Box. * Beelsebob: [78]Simulating n-bodies and functional programming. A beautiful implementation of an n-bodies simulation using reactive. * Xmonad: [79]New layout: GridVariants.SplitGrid. * Lane: [80]Christopher Lane Hinson: MaybeArrow?. An arrow with possible failure, which still allows later effects to happen after failure. * Don Stewart (dons): [81]#haskell is a busy place. Some statistics on the #haskell IRC channel. * Alson Kemp: [82]Cabal-install is awesome. Enough said! * FP-Syd: [83]Sydney FP Group: FP-Syd #11.. * Brent Yorgey: [84]Diagrams 0.2 release. * Dan Piponi (sigfpe): [85]Beyond Regular Expressions: More Incremental String Matching. More incremental brain expansion courtesy of Dan. * Don Stewart (dons): [86]Reviving the Gofer Standard Prelude (circa. 1994). Haskell, retro-style! Quotes of the Week * sigfpe: If I took some Ritalin, maybe I could write an entire book on Haskell and algebra. * cjb: It's all fun and games until somebody loses an IOVar. * Gracenotes: You are likely to be eaten by a poset * Gracenotes: A public service announcement: if you find yourself overusing the Writer monad, tell (Sum 1) * Anonymous: Haskell, the world's leading purely fictional programming language * ehird: 2009: The Year of the Combinatorial Explosion of Haskell Web Frameworks. Also, the Linux Desktop. * ehird: [on the previous quote] Someone re-remember that quote when lambdabot's back so I don't have to and thereby look egotistical, thanks * luqui: Hmph! My program which uses unsafeCoerce everywhere is not working properly! Who'd have thought... About the Haskell Weekly News New editions are posted to [87]the Haskell mailing list as well as to [88]the Haskell Sequence and [89]Planet Haskell. [90]RSS is also available, and headlines appear on [91]haskell.org. To help create new editions of this newsletter, please see the information on [92]how to contribute. Send stories to byorgey at cis dot upenn dot edu. The darcs repository is available at darcs get [93]http://code.haskell.org/~byorgey/code/hwn/ . References 1. http://haskell.org/ 2. http://www.algorithm.com.au/blog/files/change.php#unique-entry-id-605 3. http://www.pixar.com/ 4. http://article.gmane.org/gmane.comp.lang.haskell.cafe/52272 5. http://haskell.org/haskellwiki/Mutually_recursive_modules 6. http://article.gmane.org/gmane.comp.lang.haskell.cafe/52045 7. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/UrlDisp 8. http://www.haskell.org/haskellwiki/UrlDisp 9. http://article.gmane.org/gmane.comp.lang.haskell.cafe/52012 10. http://article.gmane.org/gmane.comp.lang.haskell.cafe/51988 11. http://www.haskell.org/~simonmar/base-unicode.tar.gz 12. http://article.gmane.org/gmane.comp.lang.haskell.general/16846 13. http://article.gmane.org/gmane.comp.lang.haskell.general/16840 14. http://ui3.info/d/proj/multiplicity.html 15. http://duplicity.nongnu.org/index.html 16. http://article.gmane.org/gmane.comp.lang.haskell.general/16838 17. http://blog.happstack.com/2009/02/03/happstack-01-released 18. http://happstack.com/ 19. http://article.gmane.org/gmane.comp.lang.haskell.general/16833 20. http://article.gmane.org/gmane.comp.lang.haskell.general/16829 21. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/regex%2Dposix 22. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/regex%2Dtdfa 23. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/regex%2Dposix%2D… 24. http://www.haskell.org/haskellwiki/Regex_Posix 25. http://article.gmane.org/gmane.comp.lang.haskell.general/16822 26. http://janestreetsummer.com/ 27. http://article.gmane.org/gmane.comp.lang.haskell.cafe/51783 28. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/gitit 29. http://article.gmane.org/gmane.comp.lang.haskell.cafe/51764 30. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/regex-xmlschema 31. http://www.haskell.org/haskellwiki/Regular_expressions_for_XML_Schema 32. http://article.gmane.org/gmane.comp.lang.haskell.cafe/51726 33. http://code.haskell.org/diagrams 34. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/colour 35. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/52219 36. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/51998 37. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/51858 38. http://article.gmane.org/gmane.comp.lang.haskell.general/16837 39. http://clip.dia.fi.upm.es/Job_Openings/RyC2009.html 40. http://clip.dia.fi.upm.es/ 41. http://article.gmane.org/gmane.comp.lang.haskell.general/16841 42. http://eecs.oregonstate.edu/ 43. http://planet.haskell.org/ 44. http://haskell.org/haskellwiki/Blog_articles 45. http://www.mega-nerd.com/erikd/Blog/CodeHacking/libsndfile/rel_18.html 46. http://syntax.wikidot.com/blog:1 47. http://greenokapi.net/blog/2008/12/01/functional-pearls-v2-now-with-monads-… 48. http://tehgeekmeister.wordpress.com/2009/02/06/google-code-project-google-g… 49. http://tehgeekmeister.wordpress.com/2009/02/05/releasing-wikimediaparser-01/ 50. http://www.algorithm.com.au/blog/files/change.php#unique-entry-id-605 51. http://augustss.blogspot.com/2009/02/is-haskell-fast-lets-do-simple.html 52. http://augustss.blogspot.com/2009/02/regression-they-say-that-as-you-get.ht… 53. http://www.galois.com/blog/2009/02/05/equivalence-and-safety-checking-in-cr… 54. http://ghcsparc.blogspot.com/2009/02/just-testing.html 55. http://vis.renci.org/jeff/2009/02/05/heiroglyph-interactively-graph-a-sprea… 56. http://vis.renci.org/jeff/2009/02/05/snippet-sorting-the-elements-of-a-map/ 57. http://lhc-compiler.blogspot.com/2009/02/grin-little.html 58. http://axman6.homeip.net/blog/2009/02/other-shootout-news.html 59. http://xmonad.wordpress.com/2009/02/04/contribs-review-spawnon/ 60. http://vis.renci.org/jeff/2009/02/04/hieroglyph-haddock-docs/ 61. http://axman6.homeip.net/blog/2009/02/more-n-bodies-speedups.html 62. http://axman6.homeip.net/blog/2009/02/n-bodies-evolution.html 63. http://koweycode.blogspot.com/2009/02/practical-quickcheck-wanted.html 64. http://axman6.homeip.net/blog/2009/02/textmate-haskell-bundle-improvments.h… 65. http://blog.happstack.com/2009/02/03/happstack-01-released/ 66. http://blog.happstack.com/2009/02/03/socalfp-presentation-patch-tag-as-a-st… 67. http://neilmitchell.blogspot.com/2009/02/monomorphism-and-defaulting.html 68. http://vis.renci.org/jeff/2009/02/03/hieroglyph-part-ii-basic-mousekeyboard… 69. http://vis.renci.org/jeff/2009/02/03/introducing-hieroglyph-purely-function… 70. http://vis.renci.org/jeff/2009/02/03/hieroglyph-howto-part-i-sparklines/ 71. http://axman6.homeip.net/blog/2009/02/n-bodies-speedup-50.html 72. http://johnlato.blogspot.com/2009/02/build-better-wave-reader-pt-2.html 73. http://ghcsparc.blogspot.com/2009/02/join-points.html 74. http://lukepalmer.wordpress.com/2009/02/02/parallel-rewrite-system/ 75. http://johnlato.blogspot.com/2009/02/build-better-wave-reader.html 76. http://lukepalmer.wordpress.com/2009/02/02/dana-update-system-u-box-compile… 77. http://lukepalmer.wordpress.com/2009/02/02/dana-needs-a-self-embedding-of-u… 78. http://noordering.wordpress.com/2009/02/01/simulating-n-bodies-and-function… 79. http://xmonad.wordpress.com/2009/02/01/new-layout-gridvariantssplitgrid/ 80. http://blog.downstairspeople.org/2009/01/04/maybearrow/ 81. http://donsbot.wordpress.com/2009/02/01/haskell-is-a-busy-place/ 82. http://www.alsonkemp.com/haskell/cabal-install-is-awesome/ 83. http://www.mega-nerd.com/erikd/Blog/FP-Syd/fp-syd-11.html 84. http://byorgey.wordpress.com/2009/01/31/diagrams-02-release/ 85. http://sigfpe.blogspot.com/2009/01/beyond-regular-expressions-more.html 86. http://donsbot.wordpress.com/2009/01/31/reviving-the-gofer-standard-prelude… 87. http://www.haskell.org/mailman/listinfo/haskell 88. http://sequence.complete.org/ 89. http://planet.haskell.org/ 90. http://sequence.complete.org/node/feed 91. http://haskell.org/ 92. http://haskell.org/haskellwiki/HWN 93. http://code.haskell.org/~byorgey/code/hwn/

1 0

Re: [Haskell-cafe] Re: [Haskell-beginners] Just how unsafe is unsafe
by Kalman Noel 07 Feb '09

07 Feb '09

As I didn't catch the whole thread, I hope I'm not just repeating everyone else: Roel van Dijk wrote: > I guess what unsafe should mean is a matter of taste. Personally I > find correctness more important that pureness. An unsafe function will > crash your program if evaluated when its preconditions do not hold. > Whether that is because of impurity (segmentation fault?), a partial > pattern match or a direct error "bla" is not that important. It might > be important when determining why your program crashed, but the result > is still the same. Maybe for the purposes of naming functions it's sufficient to argument that a crash is a crash whatsoever, but as for terminology in general, I think it's good to distinguish (1) partial functions from (2) functions that break purity, may lead to segmentation faults etc. I suppose (2) is the »traditional« meaning of unsafe. The motivation for this distinction is that you can still reason to some extent about (1) if you consider _|_, and that there are many functions that, although useful and correct, aren't guaranteed to terminate on infinite input, thus are partial. On the other hand, functions from (2) should always be called unsafeFoo, and/or wrapped by safe functions. Kalman

1 0

1,000 packages, so let's build a few!
by Andrew Coppin 07 Feb '09

07 Feb '09

In celebration of Hackage reachin over 1,000 unique packages, I decided that I would re-visit the problem of attempting to build them on Windows. I began by removing all existing Haskellness from my PC. I now have a vanilla Windows XP (32-bit) system with Service Pack 3. So, let's see what we can do here... - Install GHC 6.10.1. As you'd expect, no issues here. - Now, let's install stream-fusion. First let me ch... woah! What the hell? OK, it seems that Google is temporarily "broken". (All websites show up as "this site may damange your computer".) Obviously this has nothing to do with Haskell, but it makes navigating Hackage moderately more tedious. Oh well, anyway, where was I? Ah yes, I already have the tarball for stream-fusion-0.1.1, but I see that the latest release is 0.1.2.1. (Unfortunately, there doesn't appear to be any way to determine what the difference is between the two versions...) - Right, I've got the tarball for stream-fusion. Now let's see if it will install... J:\Haskell\unpack> runhaskell Setup configure J:\Haskell\unpack> runhaskell Setup build J:\Haskell\unpack> runhaskell Setup install J:\Haskell\unpack> ghc-pkg list ...stream-fusion-0.1.2.1... OK, nothing much wrong with that then. :-D - Hmm, doesn't GHC 6.10.1 now come with Haddock included? I wonder if there's a way to ask Cabal to build the docs for me... [Grr! Google is still busted.] Ah, yes there is. OK, let's try this: J:\Haskell\unpack> runhaskell Setup haddock ... Data\Stream.hs:292:33: parse error on input `!' haddock: Failed to check module: Data.Stream Uh... OK. So I guess I *won't* build the docs for that one then. :-/ Oh well, maybe it's a one-off failure? Let's try some more packages... - Next up, let's try the excellent ansi-terminal package. Checking Hackage [again without Google] and I see that the tarball I've got is still the latest version (0.5.0). Let's see what happens now... J:\Haskell\unpack> runhaskell Setup configure J:\Haskell\unpack> runhaskell Setup build J:\Haskell\unpack> runhaskell Setup haddock J:\Haskell\unpack> runhaskell Setup install J:\Haskell\unpack> ghc-pkg list ...ansi-terminal-0.5.0... Well, that installed OK. And look... it's given me a little folder full of Haddock documentation. How nice! Pitty it's burried 5 subfolders deep on the system drive with no hyperrefs to it... but I guess I could just bookmark the root folder or something? That'll work. (I.e., when I add more packages, it'll still be easy to navigate to their documentation this way.) - Now on to the ambiguously-named "binary" package. Again, I have a tarball for 0.4.2 but the latest is 0.4.4. (And, again, no indication of how this is different. But let's hope the difference is that it's fixed to work with GHC 6.10...) Configure, build, Haddock, install... no issues. That's refreshing. - OK, so what have we done? We've done stream-fusion, which is 100% Haskell. We've done ansi-terminal, which only links to standard Win32 headers [which, oddly, seem to come bundled with GHC]. And we've done binary, which (I think!) comes bundled with a few small snippets of C. Now let's try something harder... - Gtk2hs doesn't seem to want to work with GHC 6.10 yet; that's a pitty. It's a nice library. OTOH, how many Windows users have GTK+ installed? It's fairly uncommon. Fortunately you can just bundle the DLLs along with your app; unfortunately there's *a lot* of them, and they're quite large. And, obviously, you get a very non-native look & feel. Still, it *is* a nice library, and I'm sure at some point people will hassle Duncan to the point where an update appears. ;-) - A better choice, theoretically, should be wxHaskell. As I understand it, wxWidgets is supposed to look native. (Do you still have to bundle a DLL? Or is it statically linked in? Or...?) Oh well, I'd like to at least try it out and see how it works... Hmm, looks like you can obtain it via Cabal, or there's a special Windows build where you just run a BAT file and it registers everything for you. Let's try that first... "Click on the newly created wxhaskell-0.11.0\bin\wxhaskell-register.bat program to register the wxHaskell package. To uninstall, you can click on wxhaskell-0.11.0\bin\wxhaskell-unregister.bat and than remove the wxhaskell-0.11.0 directory." That sounds easy. Let's try it... 1 file(s) copied. 1 file(s) copied. 1 file(s) copied. - register packages 'ghc-pkg' is not recognized as an internal or external command, operable program or batch file. 'ghc-pkg' is not recognized as an internal or external command, operable program or batch file. error: Unable to register the package using "ghc-pkg". Maybe you have an incompatible version of ghc installed? Press any key to continue . . . Uh... what? OK, let's see if I can figure out what this batch file is trying to do... Er... wow. o_O Alright, there's no way I'm going to figure this out. I can't even tell where it's trying to install to or anything. Hmm. Well, it looks like it's all compiled and it's just trying to copy a few files and register them with GHC. That's probably a lot simpler than trying to actually compile from source using Cabal. I don't know what's confusing it, but maybe somebody in HCafe will know? I'm going to leave this one for now... - Another exciting package is the Haskell binding for libSDL. I am just *dying* to start doing some sound coding with Haskell, so I'm very interested in getting this one to work. Last time I tried it failed miserably, but let's see what we get this time around... Hmm, OK. Well let me see now... There's a project on Sourceforge called "hSDL" which seems to be dated Apr 2005. Then there's a thing called "hsSDL" on somebody's website dated around Feb 2008. And finally on Hackage itself there's a "SDL-0.5.4" dated Apr 2008. Oh... goodie. Well, let's go with the latest date, shall we? OK, download SDL-0.5.4. Unpack. Let's see if there's any useful documentation in here... Ah, "Win32". What does that contain? Mmm, looks useful. OK, let's follow these instructions and see what I get... Need to grab libSDL... Downloaded SDL-devel-1.2.13-mingw32.tar.gz. Unpacked it to C:\SDL. OK, now what? Ooo, manually edit the Cabal file... goodie! OK, well this shouldn't be too hard... Right, now... configure... That looked OK... build... Preprocessing library SDL-0.5.4... Graphics\UI\SDL\General.hsc:1:17: SDL.h: No such file or directory Graphics\UI\SDL\General.hsc: In function `main': Graphics\UI\SDL\General.hsc:57: error: `SDL_INIT_TIMER' undeclared (first use in this function) Graphics\UI\SDL\General.hsc:57: error: (Each undeclared identifier is reported o nly once Graphics\UI\SDL\General.hsc:57: error: for each function it appears in.) Graphics\UI\SDL\General.hsc:58: error: `SDL_INIT_AUDIO' undeclared (first use in this function) Graphics\UI\SDL\General.hsc:59: error: `SDL_INIT_VIDEO' undeclared (first use in this function) Graphics\UI\SDL\General.hsc:60: error: `SDL_INIT_CDROM' undeclared (first use in ...etc... OK, that's fantastic. Now what? So where *is* SDL.H then? Ah, it's in C:\SDL\include\SDL. Maybe if I adjust the Cabal path... configure... build... OMG! It looks like that might even have worked... Haddock... install... Well, well! That seems to have worked! Right, are there any example programs with this puppy?... No, there aren't. Apparently. Well that keeps things interesting... OK, well it looks like Graphcs.UI.SDL.CPUInfo is trivial, so let's see if I can compile a short program that calls it. Right, well it compiles. Does it run? ...no. It can't find SDL.dll. Presumably if I insert that into my search path somewhere... woo! It works! :-D Right, well. I guess if I can find some nontrivial examples from somewhere, I can try this thing out properly. (Looking at the API documentation, it's not very obvious how to do anything with it...) - OK, well I guess this message is long enough already. I notice there's a brand new version of HDBC out, so tomorrow I'm going to attempt to finally make a Haskell program actually talk to a database. ;-) TO BE CONTINUED...

11 25

Fastest regex package?
by Eugene Kirpichov 07 Feb '09

07 Feb '09

Hello. What is the fastest regex package of the multitude of packages present on Hackage (namely, PCRE, PCRE-light, DFA, TDFA and many more)? My benchmark (parsing a huge logfile with a regex like "GET /foo.xml.*fooid=([0-9]++).*barid=([0-9]++)") shows that plain PCRE is the fastest one (I tried PCRE, PCRE-light and TDFA; DFA can't do capturing groups at all, TDFA was abysmally slow (about 20x slower than PCRE), and it doesn't support ++), but maybe have I missed any blazing-fast package? Even PCRE seems somewhat too slow to me - matching this regex eats 50% of the program time; considering that the only things it does is ungzipping the log and searching for the regex, this seems like far too much, the regex is not that complex to my mind. I tried to replace this regex by Data.ByteString.Search.KnuthMorrisPratt and BoyerMoore from stringsearch, but they work even twice slower. I also noticed that PCRE does about 2x fewer allocations than PCRE-Light, but is not much faster. However, PCRE, strangely, has a non-pure interface, I had to use it with unsafePerformIO. It's a pity I can't put these findings on haskellwiki to http://www.haskell.org/haskellwiki/Regular_expressions since I don't have an account and new accounts can't be created right now. Could anyone put them there? All in all, my question remains: what is the fastest way to do this kind of parsing on a lazy bytestring?

4 5

Haskell tutorial for pseudo users?
by Emil Axelsson 06 Feb '09

06 Feb '09

Hello, Are there any Haskell tutorials suitable for people who don't (and possibly don't want to) know Haskell, but just want to use an embedded language that happens to be in Haskell? Such a tutorial would focus on using libraries rather than defining them. For example, it might explain how to interpret a type signature involving type classes, but not how to write one's own type class. Thanks, / Emil

8 13

about integer and float operations
by Manlio Perillo 06 Feb '09

06 Feb '09

Hi. During some experiments with Python and Haskell I found some important differences about how some integer and float operations are implemented. The first difference is about a `mod` b, when a and b are Float types. Python use the fmod function, and it also implement divmod; Haskell seems to lack support for this operation. The second difference is about the division of two integers. Consider this Python code: >>> from __future__ import division >>> def fac(n): ... return reduce(lambda a, b: a * (b + 1), xrange(n), 1) ... >>> float(fac(777)) Traceback (most recent call last): File "<stdin>", line 1, in <module> OverflowError: long int too large to convert to float >>> fac(777) / fac(777) 1.0 Here CPython does not convert the two integers to float before to divide them, but make use of a special algorithm. GHC, instead, returns NaN I don't know if the implementations of divMod and "true" integer division in CPython are "robust", but there is some important reason why these operations are not supported in Haskell? Thanks Manlio Perillo

9 32