- Haskell-Cafe - Haskell.org

Haskell Weekly News: Issue 77 - July 16, 2008
by Brent Yorgey 16 Jul '08

16 Jul '08

--------------------------------------------------------------------------- Haskell Weekly News http://sequence.complete.org/hwn/20080716 Issue 77 - July 16, 2008 --------------------------------------------------------------------------- Welcome to issue 77 of HWN, a newsletter covering developments in the [1]Haskell community. Announcements Takusen 0.8.3. Alistair Bayley [2]announced the release of [3]Takusen 0.8.3, with ODBC support, more Cabal improvements, bug fixes, and some basic result-set validation. Launching Haskell Group in Vancouver, Canada. Jon Strait [4]announced that a [5]Haskell Programmers Group has been created in Vancouver; the first meeting is scheduled for next Monday, July 21st. Feel free to join the Google Groups list to be notified of future events, or just show up to a meeting to bounce ideas and questions off of other Haskell programmers. Sphinx full-text searching client on Hackage. Chris Eidhof [6]announced work on a [7]client for the [8]sphinx full-text search engine. Help hacking on it, testing it or improving documentation is welcome. haskell-src-exts 0.3.5. Niklas Broberg [9]announced that the [10]haskell-src-exts package is now updated to understand the current version of Template Haskell syntax. Bug reports welcome. Prime time for Haskell. Janis Voigtlaender [11]announced that Haskell STM is featured in an [12]article in this month's Communications of the ACM. vector 0.1 (efficient arrays with lots of fusion). Roman Leshchinskiy [13]announced an initial release of the [14]vector library, which will eventually provide fast, Int-indexed arrays with a powerful fusion framework. Galois Tech Talks: Stream Fusion for Haskell Arrays. Don Stewart [15]announced that he was giving this week's Galois Tech Talk, on stream fusion for Haskell arrays. The talk was yesterday, July 15, but hopefully some sort of recording or slides will be made available. protocol-buffers. Chris Kuklewicz [16]announced a very early version of the [17]protocol-buffers package, a Haskell interface to Google's newly released [18]data interchange format. GHC IRC meeting. Simon Marlow [19]announced the (first weekly?) IRC meeting to discuss GHC, a scheduled time when the developers turn up on #ghc, discuss current topics around GHC, and users can chime in with questions, points for discussion, complaints and so on. The first meeting took place on July 16 at 1600 BST (UTC +1)/0800 PDT (UTC-7)/1100 EDT (UTC-4), in the #ghc channel on chat.freenode.net; hopefully this will become a weekly event. Google Summer of Code Progress updates from participants in the 2008 [20]Google Summer of Code. GHC API. Thomas Schilling (nominolo) is working on [21]improvements to the GHC API. He recently [22]asked for comments on a proposed refactoring to the GHC API, creating a new Ghc monad to capture error handling and single-threaded use of Sessions. GHC plugins. Max Bolingbroke is working on dynamically loaded plugins for GHC. [23]This week, he wrote a ton of Haddock documentation for GHC internals. He also added the ability for compiler plugins to generate their own source annotations, to allow plugins to use intermediate results from previous plugins. He's currently working on an exciting, secret feature: tune in next week to find out what it is! Generic tries. Jamie Brandon is working on a library for efficient maps using generalized tries. Jamie is [24]currently working on, tweaking the [25]api, writing tests and writing reference implementations on sorted and unsorted association lists. Hoogle 4. Neil Mitchell (ndm) is working on [26]Hoogle 4. [27]This week, he worked on type search and database generation. Next week he plans to finish up type search and release and command-line version. Language.C. Benedikt Huber (visq) is [28]working on Language.C, a standalone parser/pretty printer library for C99. DPH physics engine. Roman Cheplyaka (Feuerbach) is working on a [29]physics engine using [30]Data Parallel Haskell. Cabal dependency framework. Andrea Vezzosi (Saizan) is working on a [31]make-like dependency analysis framework for Cabal. Libraries Proposals and extensions to the [32]standard libraries. adding split to Data.List. Gwern Branwen [33]proposed adding some split-like functions to Data.List. Will they actually get added this time? Will people be able to agree on one of the seventeen possible sets of semantics? Tune in next time... Discussion GHC API: monad and error handling. Thomas Schilling [34]asked for comments on a proposed refactoring to the GHC API, creating a new Ghc monad to capture error handling and single-threaded use of Sessions. Jobs Research positions on Modeling and Analyzing Software Adaptation, University of Koblenz. Ralf Lammel [35]announced two research positions for a postdoc and a PhD student, available initially for 2 years. The successful applicants will work on the research theme of "ADAPT: Modeling and Analyzing Software Adaptation". The objective of [36]ADAPT is to relate, advance, combine, and challenge adaptation methods and associated methods of modeling and analyzing that are used by the communities of software engineering, programming languages, logic-based modeling, multi-agent systems, formal methods, SOA, web systems, and mobile, autonomous systems. Blog noise [37]Haskell news from the [38]blogosphere. * Real-World Haskell: [39]Real World Haskellers at OSCON next week. John Goerzen and Bryan O'Sullivan will be in Portland, Oregon next week for OSCON (along with Don Stewart, who lives in Portland). * Russell O'Connor: [40]ICFP 2008 Post-Mortem. * Brent Yorgey: [41]Call for an ICFP Mars Server. Will someone make a Mars Server where we can all submit our rovers and new maps, watch them compete, improve them, and generally have fun getting some feedback without bugging the organizing committee? Pretty please? * Jamie Brandon: [42]Week 5 progress. An update on Jamie's Google Summer of Code project. * Brent Yorgey: [43]ICFP programming contest reflections. * >>> Chris Bogart: [44]Help with study of functional programmers. Chris is doing a study as part of a research internship at Microsoft, and is seeking people currently working on a real project in a functional language whom he can observe as part of his study. * >>> eigenclass: [45]Quicksort erratum. A time and space analysis of a classic quicksort implementation in Haskell. * Ralf Lammel: [46]Research positions on Modeling and Analyzing Software Adaptation, University of Koblenz. * Max Bolingbroke: [47]Compiler Plugins For GHC: Week Five. An update on Max's Google Summer of Code project. * Luke Plant: [48]Haskell Blog Rewrite - Session 7. Luke documents setting up CentOS in VirtualBox VM in order to compile Haskell code for a server to which he does not have ssh access. Not for the faint of heart. * Chris Eidhof: [49]Stemming with Haskell. A [50]stemming library and [51]Sphinx client for Haskell. * >>> Duane Johnson: [52]Haskell is Popular on IRC. The large amount of community participation and academic brainshare gives Duane a lot of confidence in Haskell. * >>> Mike Harris: [53]ICFP Programming Contest '08. Mike participated in the ICFP programming contest, and might try porting his solution to Haskell. * Luke Palmer: [54]The Curry-Howard isomorphism and the duality of x and ->. * Magnus Therning: [55]Playing with prefixes. Magnus explores different ways to encode units of storage (bytes, kilobytes, kibibytes...) in Haskell. * Dan Piponi (sigfpe): [56]MSFP 2008. * Shin-Cheng Mu: [57]Tail-Recursive, Linear-Time Fibonacci. * Neil Mitchell: [58]GSoC Hoogle: Week 7. * FP Lunch: [59]CCC-ness of the category of containers. * JP Moresmau: [60]instance Data Map where -- half done!. JP writes an Data.Generics.Data instance for Data.Map. * >>> Rick Carback: [61]AutoKey in Haskell. Rick is learning Haskell and has implemented a simple AutoKey cipher. * >>> Jeremy Frens: [62]PE Problem #2 in All Languages (Part II). * Luke Palmer: [63]Required Optimization. Annotations for specifying expected compiler optimizations? * London Haskell Users Group: [64]AngloHaskell 2008. * >>> Justus: [65]random programming. Solving a simple programming challenge in Haskell. Quotes of the Week * Japsu: segfault cat is watching you unsafeCoerce * z0MB13: who can say hello to me as a md5 or decipher what kind of technique can be used to remove the password * denq: [upon experiencing a moment of enlightenment] oh! something bing in my brain :) * Pseudonym: Pseudonym needs codependent types - useful when you're doing dysfunctional programming. About the Haskell Weekly News New editions are posted to [66]the Haskell mailing list as well as to [67]the Haskell Sequence and [68]Planet Haskell. [69]RSS is also available, and headlines appear on [70]haskell.org. Headlines are available as [71]PDF. To help create new editions of this newsletter, please see the information on [72]how to contribute. Send stories to byorgey at seas dot upenn dot edu. The darcs repository is available at darcs get [73]http://code.haskell.org/~byorgey/code/hwn/ . References 1. http://haskell.org/ 2. http://article.gmane.org/gmane.comp.lang.haskell.libraries/9570 3. http://hackage.haskell.org/packages/archive/Takusen/0.8.3/Takusen-0.8.3.tar… 4. http://article.gmane.org/gmane.comp.lang.haskell.cafe/42287 5. http://groups.google.com/group/hugvan 6. http://article.gmane.org/gmane.comp.lang.haskell.cafe/42263 7. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/sphinx 8. http://www.sphinxsearch.com/ 9. http://article.gmane.org/gmane.comp.lang.haskell.cafe/42233 10. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/haskell-src-exts 11. http://article.gmane.org/gmane.comp.lang.haskell.cafe/42227 12. http://research.microsoft.com/~larus/Papers/p80-larus.pdf 13. http://article.gmane.org/gmane.comp.lang.haskell.cafe/42207 14. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/vector 15. http://article.gmane.org/gmane.comp.lang.haskell.cafe/42196 16. http://article.gmane.org/gmane.comp.lang.haskell.libraries/9575 17. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/protocol-buffers 18. http://code.google.com/apis/protocolbuffers/docs/overview.html 19. http://article.gmane.org/gmane.comp.lang.haskell.glasgow.user/14741 20. http://hackage.haskell.org/trac/summer-of-code/wiki/SoC2008 21. http://hackage.haskell.org/trac/ghc/wiki/GhcApiStatus 22. http://thread.gmane.org/gmane.comp.lang.haskell.libraries/9589 23. http://blog.omega-prime.co.uk/?p=69 24. http://jamiiecb.blogspot.com/2008/07/week-5-progress.html 25. http://code.haskell.org/gmap/api 26. http://code.haskell.org/hoogle/ 27. http://neilmitchell.blogspot.com/2008/07/gsoc-hoogle-week-7.html 28. http://www.sivity.net/projects/language.c/wiki/ 29. http://haskell.org/haskellwiki/Hpysics 30. http://www.haskell.org/haskellwiki/GHC/Data_Parallel_Haskell 31. http://code.haskell.org/~Saizan/cabal 32. http://haskell.org/haskellwiki/Library_submissions 33. http://thread.gmane.org/gmane.comp.lang.haskell.libraries/9558 34. http://thread.gmane.org/gmane.comp.lang.haskell.libraries/9589 35. http://blogs.msdn.com/ralflammel/archive/2008/07/15/research-positions-on-m… 36. http://adapt.uni-koblenz.de/ 37. http://planet.haskell.org/ 38. http://haskell.org/haskellwiki/Blog_articles 39. http://www.realworldhaskell.org/blog/2008/07/16/real-world-haskellers-at-os… 40. http://r6.ca/blog/20080716T131235Z.html 41. http://byorgey.wordpress.com/2008/07/16/call-for-an-icfp-mars-server/ 42. http://jamiiecb.blogspot.com/2008/07/week-5-progress.html 43. http://byorgey.wordpress.com/2008/07/15/icfp-programming-contest-reflection… 44. http://sambangu.blogspot.com/2008/07/help-with-study-of-functional 45. http://eigenclass.org/hiki/quicksort-erratum 46. http://blogs.msdn.com/ralflammel/archive/2008/07/15/research-positions-on-m… 47. http://blog.omega-prime.co.uk/2008/07/14/compiler-plugins-for-ghc-week-five/ 48. http://lukeplant.me.uk/blog.php?id=1107301689 49. http://blog.tupil.com/stemming-with-haskell/ 50. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/stemmer 51. http://www.sphinxsearch.com/ 52. http://blog.inquirylabs.com/2008/07/14/haskell-is-popular-on-irc/ 53. http://thingsthatmikethinks.blogspot.com/2008/07/icfp-programming-contest-0… 54. http://luqui.org/blog/archives/2008/07/13/ch-isomorphism-duality/ 55. http://therning.org/magnus/archives/354 56. http://sigfpe.blogspot.com/2008/07/msfp-2008.html 57. http://www.iis.sinica.edu.tw/~scm/2008/tail-recursive-linear-time-fibonacci/ 58. http://neilmitchell.blogspot.com/2008/07/gsoc-hoogle-week-7.html 59. http://sneezy.cs.nott.ac.uk/fplunch/weblog/?p=108 60. http://jpmoresmau.blogspot.com/2008/07/instance-data-map-where-half-done.ht… 61. http://carback.us/rick/blog/?p=24 62. http://jdfrens.blogspot.com/2008/07/pe-problem-2-in-all-languages-part-ii.h… 63. http://luqui.org/blog/archives/2008/07/10/required-optimization/ 64. http://www.londonhug.net/2008/07/10/anglohaskell-2008/ 65. http://hoostus.blogspot.com/2008/07/random-programming.html 66. http://www.haskell.org/mailman/listinfo/haskell 67. http://sequence.complete.org/ 68. http://planet.haskell.org/ 69. http://sequence.complete.org/node/feed 70. http://haskell.org/ 71. http://code.haskell.org/~byorgey/code/hwn/archives/20080716.pdf 72. http://haskell.org/haskellwiki/HWN 73. http://code.haskell.org/~byorgey/code/hwn/

1 0

A type signature inferred by GHCi that is rejected when written explicitly
by Pablo Nogueira 16 Jul '08

16 Jul '08

I find this interesting, GHCi accepts a function |dmap| which I show below and infers its type, but if I annotate the function with the inferred type, GHCi's type-checker rejects it. I'm trying to generalise the datatype-generic dmap: < dmap :: Bifunctor s => (a -> b) -> Fix s a -> Fix s b < dmap f = In . bimap (dmap f) f . out where > data Fix s a = In { out :: s (Fix s a) a } > > class Bifunctor s where > bimap :: (a -> c) -> (b -> d) -> s a b -> s c d The idea is that recursive types are represented by their (lamba-lifted) functors, eg: > data ListF b a = NilF | ConsF a b > instance Bifunctor ListF where > bimap f g NilF = NilF > bimap f g (ConsF a b) = ConsF (g a) (f b) I now define two classes: > class From a c x where > from :: a x -> c x > > class To a c y where > to :: c y -> a y And a generic |gmap| which given the map for |c| and a mapping for |x| delivers the map for |a|: > type GMap t x y = (x -> y) -> t x -> t y > > gmap :: (From a c x, To a c y) => GMap c x y -> GMap a x y > gmap gmc gmx = to . gmc gmx . from I want to write |dmap| as a special case of |gmap|, but I can't even get there. If I write > dmap f = to . bimap (dmap f) f . from GHCi infers it has type (up to renaming): (From a1 (s (a1 x)) x, Bifunctor s, To a2 (s (a2 y)) y) => (x -> y) -> a1 x -> a2 y But if I cut and paste the type into the code I get type errors: Could not deduce (From a1 (s1 (a11 x)) x) ... Could not deduce (From a11 (s1 (a11 x)) x, To a21 (s1 (a21 y)) y) ... Could not deduce (From a1 (s1 (a11 x)) x) ... Interesting,

3 3

Euler 201 performance mystery
by nadine.and.henry＠pobox.com 16 Jul '08

16 Jul '08

Dear Group, I've spend the last few days figuring out the solution to Euler Problem 201 in haskell. I first tried a relatively elegant approach based on Data.Map but the performance was horrible. I never actually arrived at the answer. I then rewrote the same algorithm using STUArrays and it was lightning. I have posted both versions of the code at: http://www.maztravel.com/haskell/euler_problem_201.html and would appreciate any insights that you master haskellers can provide on why the speed difference is so huge. Thanks in advance. Henry Laxen

2 2

Strange space leak
by Grzegorz Chrupala 16 Jul '08

16 Jul '08

Hi all, I just noticed that a tiny change to the program which I posted recently in the "More idiomatic use of strictness" thread causes a space leak. The code is: {-# LANGUAGE BangPatterns, PatternGuards #-} import Data.List (foldl') import Data.Char split delim s | [] <- rest = [token] | otherwise = token : split delim (tail rest) where (token,rest) = span (/=delim) s main = do putStrLn =<< fmap (show . stats ["the","a","and"] . split "<DOC>" . words) getContents stats ws docs = foldl' f ((map (const 0) ws),0) docs where f (dfs,n) d = let dfs' = zipWith (\w df -> (df + fromEnum (w `elem` d))) ws dfs in sum dfs' `seq` (dfs',n+1) If I change this line: putStrLn =<< fmap (show . stats ["the","a","and"] . split "<DOC>" . words) getContents to this: putStrLn =<< fmap (show . stats ["the","a","and"] . split "<DOC>" . words . map toLower) getContents suddenly the programs starts using tons of memory instead of running in small constant space. What's going on? Best, -- Grzegorz -- View this message in context: http://www.nabble.com/Strange-space-leak-tp18439685p18439685.html Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com.

3 4

win32: haddock: internal Haddock or GHC error?
by Dmitri O.Kondratiev 16 Jul '08

16 Jul '08

Using GHC 6.8.2 and haddock 2.0.0.0. in WinXP (SP3) I am trying to follow instructions from "How to write a Haskell program<http://www.haskell.org/haskellwiki/How_to_write_a_Haskell_program#Build_som…> ". Everything goes well, until I try to generate html documentation for the project. Then I get the following errors: C:\wks\haq>runhaskell Setup.lhs haddock --executables Preprocessing executables for haq-0.0... Running Haddock for haq-0.0... Preprocessing executables for haq-0.0... Warning: Cannot read C:/usr/ghc-6.8.2/doc/libraries/base\base.haddock: "Magic number mismatch: couldn't load interface file: C:/usr/ghc-6.8.2/doc/li braries/base\\base.haddock" Skipping this interface. Warning: main:Main: could not find link destinations for: GHC.IOBase.IO haddock: internal Haddock or GHC error: c:/builds/haddock-2/install\haddock-2.0. 0.0/html/haddock.css: openFile: does not exist (No such file or directory) Any ideas how to solve this? Thanks!

2 2

Mutually recursive modules and google protocol-buffers
by Chris Kuklewicz 16 Jul '08

16 Jul '08

I have reached an impasse in designing a Haskell API for the google's protocol-buffers data language / format. ( http://code.google.com/apis/protocolbuffers/docs/overview.html ) The messages in protobuf are defined in a namespace that nests in the usual hierarchical OO style that Java encourages. To avoid namespace conflicts, I made a hierarchy of modules. But...this is a legal pair protobuf message definitions: > // Test that mutual recursion works. > message TestMutualRecursionA { > optional TestMutualRecursionB b = 1; > optional int32 content = 2; > } > > message TestMutualRecursionB { > optional TestMutualRecursionA a = 1; > optional int32 content = 2; > } And there is no way ghc can compile these in separate modules. But the overlap of record accessors names "content" makes defining these messages in a single module with a common namespace quite verbose. Any opinions on the least worst way to design this? -- Chris

9 10

CAL (OpenQuark) and enterprise
by fero 16 Jul '08

16 Jul '08

Hi Haskellers and CALers, I have the feeling that a lot of code in my jee application can be done better by using functional programming. There is a lot of searching in object trees, transforming objects to another objects, aggregation functions... All written in java. Sequential logic can by done declarative with statemachines and workflows (not handcoded, dut drawed) all other with functions. Has somebody experience with this? Or is that a bad idea? Should I use pure functions, or imperative-functional language such as Scala? Is somebody here using CAL (OpenQuark) in jee application? What is your experience -- View this message in context: http://www.nabble.com/CAL-%28OpenQuark%29-and-enterprise-tp18366397p1836639… Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com.

4 5

Getting latest GHC-head/libraries takes forever
by Austin Seipp 16 Jul '08

16 Jul '08

For the purpose of experimenting with NDP I went through the process of getting the GHC head from darcs.haskell.org. As specified in the developer wiki[1], using darcs get is basically not possible because there're so many patches. So I downloaded http://darcs.haskell.org/ghc-HEAD-2008-06-06-ghc-corelibs-testsuite.tar.bz2 This alone took over *7 hours* using wget because it could never get a connection faster than 5kb/s. After downloading I changed to the directory and ran 'darcs pull -a'. Nothing bad happened here, darcs pulled most of the changes pretty quick so I had an up to date GHC. However, when trying to use the ./darcs-all script to pull in the latest library/testsuite changes, things go horribly, horribly wrong. I let darcs sit for over another hour and a half trying to pull patches from darcs.haskell.org for the libraries and the testsuite. It never got past packages/base. In fact, it never even got to *getting* the patches from the base repository, it sat there, stuck at the 'identifying repository' stage (darcs 2.0.2) for over an hour, making my processor go round and round at 99% CPU usage the entire way, never accomplishing anything. What makes it even stranger is that it got the latest patches for the testsuite and the array package pretty quick. It just got permanently stuck on base for some reason and never went further. Also, I tried pulling package-ndp a few days ago while using an older development snapshot of GHC (not the darcs repository, a nightly HEAD tarball of ghc.) Just pulling 500 patches from http://darcs.haskell.org/packages/ndp took over an hour, but at least it got done. It didn't build properly so I thought I'd get the latest darcs version of GHC instead, and here I am. Has anybody experienced anything like this recently? Off the top of my head the only things that possibly come to mind are: 1) darcs.haskell.org is just really slow or under -constant- heavy load. 2) the darcs repositories are in darcs-1 format, so darcs-2 is having issues when trying to pull the patches. 3) Diabolical connection-killing ninjas. None of these 3 possibilities are really 'great,' and none of them help me get the latest version of GHC, the libs, and NDP any faster then a narcoleptic snail. If instead this is a techincal issue with, say, darcs that someone is aware of, I'm using OS X Leopard, 10.5.4 with a statically linked darcs-2.0.2 binary from [2]. [1] http://hackage.haskell.org/trac/ghc/wiki/Building/GettingTheSources [2] http://wiki.darcs.net/DarcsWiki/CategoryBinaries#head-25c8108e9d719be30a8cc…

3 2

Can't get Haddock 2.0 anchor reference to work
by Mike Gunter 15 Jul '08

15 Jul '08

How are anchor references in Haddock supposed to work? When I use "Dir.Mod#foo", the resulting HTML contains: <A HREF="Dir-Mod#foo.html"> instead of the more desirable: <A HREF="Dir-Mod.html#foo"> User error or bug? thanks, -m

2 1

Remote control of firefox through Haskell?
by Jefferson Heard 15 Jul '08

15 Jul '08

Is there a library out there that will allow me to remote-control the firefox or mozilla browsers, e.g. change the current page, open a new tab? -- I try to take things like a crow; war and chaos don't always ruin a picnic, they just mean you have to be careful what you swallow. -- Jessica Edwards

3 3