December 2009 - Haskell-Cafe

HP + Gtk2hs?
by Andrew Coppin 06 Dec '09

06 Dec '09

So I decided to try out the Haskell Platform instead of just installing GHC directly. To be honest, I didn't really notice much difference, except for the files being installed in a different place. But then, initially HP ~= GHC; presumably the plan is for it to grow as time goes on. Anyway, I'm sure you can all see where I'm going with this: Latest Gtk2hs binary for Windows doesn't seem to like the latest HP installation. A few things: - It is unclear to be which Gtk2hs binaries require which GHC version. Can we tabulate this information somewhere? - It is also unclear to me which HP versions contain what stuff. The HP page tells you all about the *current* HP version, and contains a link to a raw directory listing for the older versions. (I especially love the way the server truncates the filenames to just before where the version number starts. Very helpful. :-D ) Can we get propper descriptions for _all_ releases of HP? A table summarising the GHC version and major packages/versions in each release would be nice. - Any idea why the HP says "beta" on it? - Apart from HP providing GHC 6.10.4 while Gtk2hs currently requires 6.10.3, it appears that the Gtk2hs installer package doesn't like GHC being installed in a path with spaces. Apparently Gtk2hs has a bug tracker. (I only just discovered this, so perhaps it needs to be more prominent?) Somebody pointed out this bug 6 months ago. Somebody else posted a potential fix a month ago. There is no visible activity from the developers. (I emphasize "visible". Maybe they've seen this and they're working on it, but there's no visible indication of activity.) Gtk2hs is currently the *only* GUI binding that actually works on Windows. It's always been slightly frustrating that every time a new version of GHC comes out, us Windows users have to beg for somebody to build a new binary. (Linux users, on the other hand, can just download the sources and compile them with any compiler they like.) Presumably going forward Gtk2hs will synchronise to the (less frequent) HP releases rather than the individual GHC releases, so many this particular frustration will be reduced. (It would be even less frustrating if Gtk2hs was *in* HP, but I gather this won't ever happen.) Of course, being able to actually build C bindings on Windows would be even better, but this is apparently infeasible... :-( That being the case, I'll settle for just a little bit more documentation.

6 10

universal binary version of Haskell Platform?
by Benjamin L.Russell 06 Dec '09

06 Dec '09

Recently, in changing my work schedule to work mainly from home, I switched from mainly using a work Wintel machine running Windows XP Professional, Service Pack 3, to mainly using my home PowerPC G4 PowerBook Macintosh, currently upgraded to Mac OS X 10.5.8 Leopard. However, to my surprise, there does not seem to be a version of the Haskell Platform that runs natively on my current OS. Does anybody know where to find a universal binary version of the Haskell Platform, or at least of GHC 6.10.4? Otherwise, I'm stuck without a native version. -- Benjamin L. Russell -- Benjamin L. Russell / DekuDekuplex at Yahoo dot com http://dekudekuplex.wordpress.com/ Translator/Interpreter / Mobile: +011 81 80-3603-6725 "Furuike ya, kawazu tobikomu mizu no oto." -- Matsuo Basho^

4 3

Re: Optimizing Parsec 3 -- was: Wiki software?
by Howard B. Golden 05 Dec '09

05 Dec '09

On Friday December 4, 2009, John MacFarlane wrote: > I used criterion to compare pandoc compiled with parsec2 to > pandoc compiled with your version of parsec3. (The benchmark > is converting testsuite.txt from markdown to HTML.) The difference > was minor: > > parsec2: > mean: 67.66576 ms, lb 67.56722 ms, ub 67.88983 ms, ci 0.950 > std dev: 722.3878 us, lb 323.0343 us, ub 1.356013 ms, ci 0.950 > > parsec3: > mean: 68.20847 ms, lb 68.16387 ms, ub 68.26284 ms, ci 0.950 > std dev: 252.7773 us, lb 204.5512 us, ub 325.2424 us, ci 0.950 > > So, once you release the new parsec3, I am prepared to remove the > "parsec < 3" restriction from the libraries I maintain: pandoc, > highlighting-kate, filestore, gitit, and yst. I don't know what the performance of the current parsec3 is compared to parsec2. It would be helpful if you could run your benchmark for that also and include it. If the only issue is performance, I respectfully request that you remove the parsec < 3 requirement even before the new version of parsec3 is released. Thank you. Howard

2 1

Avoiding undecidables
by Michael Snoyman 05 Dec '09

05 Dec '09

Hi all, Well, I've got two problems which both want to be solved with undecidable and overlapping instances. Obviously, I'd like to try and avoid them. For the record, the problems have to do with the control-monad-failure and convertible packages. The code below *should* make clear what I'm trying to accomplish: -- failure should not be limited to just monads, so... class Failure e f where failure :: e -> f v class (Functor f, Failure e f) => FunctorFailure e f instance (Functor f, Failure e f) => FunctorFailure e f -- undecidable, overlaps class (Applicative f, Failure e f) => ApplicativeFailure e f instance (Applicative f, Failure e f) => ApplicativeFailure e f -- undecidable, overlaps class (Monad f, Failure e f) => MonadFailure e f instance (Monad f, Failure e f) => MonadFailure e f -- undecidable, overlaps And now the convertible issue. I want two type classes: Convert for anything which *might* be convertible, but might not. For example, sometimes a String can be converted to an Int (like the string "5"), but sometimes it will fail (like "five"). TotalConvert is when something *is* convertible, such as Int to String (simply the show function). Thus the following: class Convert x y where convert :: x -> Maybe y class Convert x y => TotalConvert x y where totalConvert :: x -> y instance TotalConvert x y => Convert x y where -- Boom! convert = Just . totalConvert Any ideas are most welcome. Both of these seem like cases where the compiler could do some DWIMery, but isn't. Michael

3 3

Re: [Haskell-cafe] HP + Gtk2hs?
by Andrew Coppin 05 Dec '09

05 Dec '09

M Xyz wrote: > > > > - Apart from HP providing GHC 6.10.4 while Gtk2hs currently > requires 6.10.3, it appears that the Gtk2hs installer package > doesn't like GHC being installed in a path with spaces. Apparently > Gtk2hs has a bug tracker. (I only just discovered this, so perhaps > it needs to be more prominent?) Somebody pointed out this bug 6 > months ago. Somebody else posted a potential fix a month ago. > There is no visible activity from the developers. (I emphasize > "visible". Maybe they've seen this and they're working on it, but > there's no visible indication of activity.) > > Thank you so much for this post. I tried to install Gtk2hs yesterday > to play with > its Cairo bindings and I hit the "Setup found what appears to be a > non-working installation of GHC" brickwall. I found a closed ticket > about this problem: http://hackage.haskell.org/trac/gtk2hs/ticket/1117 > I posted this issue on the Gtk2hs-devel mailing list and got a > response today: > http://sourceforge.net/mailarchive/forum.php?thread_name=EF0EE62B-259C-4F62… > > I don't think I have the will to uninstall the Haskell Platform and > everything I've set up just to install in a path with no spaces. :) > > My information is from here: http://hackage.haskell.org/trac/gtk2hs/ticket/1165 In case it makes any difference to you. :-)

1 0

Re: [Haskell-cafe] HP + Gtk2hs?
by M Xyz 05 Dec '09

05 Dec '09

- Apart from HP providing GHC 6.10.4 while Gtk2hs currently requires 6.10.3, it appears that the Gtk2hs installer package doesn't like GHC being installed in a path with spaces. Apparently Gtk2hs has a bug tracker. (I only just discovered this, so perhaps it needs to be more prominent?) Somebody pointed out this bug 6 months ago. Somebody else posted a potential fix a month ago. There is no visible activity from the developers. (I emphasize "visible". Maybe they've seen this and they're working on it, but there's no visible indication of activity.) Thank you so much for this post. I tried to install Gtk2hs yesterday to play with its Cairo bindings and I hit the "Setup found what appears to be a non-working installation of GHC" brickwall. I found a closed ticket about this problem: http://hackage.haskell.org/trac/gtk2hs/ticket/1117 I posted this issue on the Gtk2hs-devel mailing list and got a response today: http://sourceforge.net/mailarchive/forum.php?thread_name=EF0EE62B-259C-4F62… I don't think I have the will to uninstall the Haskell Platform and everything I've set up just to install in a path with no spaces. :)

1 0

Finding HP
by Andrew Coppin 05 Dec '09

05 Dec '09

Today I was setting up a my new, and I wanted to put Haskell on it. Rather than download GHC itself, I decided to install the Haskell Platform instead, just to see what it's like. Much to my surprise, I couldn't actually find any reference to its existence anywhere from the haskell.org home page, and I eventually had to run a search to find it. Subsequently, I realise [as somebody will no doubt point out] that the link is actually there, on the front page, right next to GHC, Hugs, et al. My suggestion is that if we really want people to grab the HP rather than download GHC directly, maybe we could make the link slightly more prominent? It also wouldn't hurt to mention it from the "Implementations" page, and maybe the GHC homepage? Just a suggestion...

11 15

Wikipedia article
by Simon Marlow 05 Dec '09

05 Dec '09

As noted before, the Wikipedia article for Haskell is a disorganised mess. http://en.wikipedia.org/wiki/Haskell_%28programming_language%29 earlier this year, dons suggested reorganising it and posted a template on the Haskell wiki: http://haskell.org/haskellwiki/WikipediaArticleDesign I've made a start on a new version of the page: http://en.wikipedia.org/wiki/User:Simonmar/Haskell_%28programming_language%… I've kept most of the existing information, but reorganised it more or less according to Don's template, and I filled out the overview section. Also I fixed numerous things, but the page still has a long way to go Does anyone mind if I spam the existing Haskell article with this new one, or do people think we should continue editing the sandbox version until it's in better shape? Cheers, Simon

6 9

From function over expression (+, *) derive function over expression (+)
by Radek Micek 05 Dec '09

05 Dec '09

Hello. I have two types for expression: data Expr = Add Expr Expr | Mul Expr Expr | Const Int data AExpr = AAdd AExpr AExpr | AConst Int The first one supports addition and multiplication and the second only addition. I can write a function to simplify the first expression: simplify :: Expr -> Expr simplify = {- replaces: "a*1" and "1*a" by "a", "a+0" and "0+a" by a -} And I would like to use the function simplify for the second type AExpr. What can I do is to convert AExpr to Expr, simplify it and convert it back. But I don't like this solution because conversions take some time. I would prefer following: I say to the compiler that AAdd is like Add and AConst is like Const and the compiler derives function asimplify for AExpr. Is it possible to do this? In fact I want to have two distinct types where one is "extension" of the second (Expr is extension of AExpr) and I want to define a function for the extended type (Expr) and use it for the original type (AExpr). I assume that the function won't introduce Mul to the expression which had no Mul. Thanks in advance Radek Micek

6 9

Haskell Weekly News: Issue 142 - December 05, 2009
by jfredett＠gmail.com 05 Dec '09

05 Dec '09

--------------------------------------------------------------------------- Haskell Weekly News http://sequence.complete.org/hwn/20091205 Issue 142 - December 05, 2009 --------------------------------------------------------------------------- Welcome to issue 142 of HWN, a newsletter covering developments in the [1]Haskell community. Occasionally, I have something silly or clever to say here. Usually I give a summary of the weeks news and end it with a bit of a cliche tagline, however, this is finals week, and as such, there is absolutely no brain cells left working to come up with any cleverness for here. I'll see you next week with (hopefully) more ability for humor. Until next week, the Haskell Weekly News! Announcements error-message. Gregory Crosswhite [2]announced a package for handling errors in haskell. To quote his message directly, 'If there is one thing that we really don't have enough of in Haskell, it is *ways to handle errors*!' atom-0.1.3. Tom Hawkins [3]announced a new version of the Atom DSL for designing hard realtime embedded software. Blueprint 0.1 -- PREVIEW. Gregory Crosswhite [4]announced a preview version of his 'Blueprint' configuration package. haskell-mode 2.7.0. Svein Ove Aas [5]announced another release of Haskell-mode for Emacs graphviz-2999.7.0.0. Ivan Lazar Miljenovic [6]announced a new version of his set of bindings to the graphviz tools for visualizing graphs. Haskell Job Openings (Pune, IN). Tom Hawkins [7]announced some job openings in Eaton's engineering center in Pune, India. SaC Tutorial at PPoPP 2010. Clemens Grelck [8]announced a tutorial to be held at PPoPP on SAC and it's auto-parallelizing compiler SAC2C. Call for papers: PAPP 2010, 7th International Workshop on Practical Aspects of High-level Parallel Programming. Clemens Grelck [9]announced a call for papers for the Seventh International Workshop on Practical Aspects of High-level Parallel Programming. PhD Studentships in Nottingham. Graham Hutton [10]announced some PhD openings at the University of Nottingham. Second Call for Copy: Monad.Reader Issue 15. Brent Yorgey [11]announced a second call for copy for the Monad.Reader. 3 full professor positions at DIKU. Fritz Henglein [12]announced three open professor positions at the the Department of Computer Science at the University of Copenhagen ([13]DIKU). Hubris 0.0.2, the Ruby-Haskell bridge. Mark Wotton [14]announced a new version of the Hubris package is now available on hackage. Discussion You are in a twisty maze of concurrency libraries, all different ... Patrick Caldon [15]asked about the 'right' concurrency package to use for his project. Greetings! 2D Graphics? M Xyz [16]asked (enthusiatically) about different ways to do 2-dimensional graphics in Haskell. Wikipedia article. Simon Marlow [17]rallied us all into fixing the (or at least starting to fix) wikipedia page for Haskell. Great Programs to Read? Michael Lesniak [18]asked about what programs were worth reading. Blog noise [19]Haskell news from the [20]blogosphere. Blog posts from people new to the Haskell community are marked with >>>, be sure to welcome them! * Michael Snoyman: [21]String-like. * Claude Heiland-Allen: [22]Heist: dataflow algebra. * Adam Jones: [23]Lambda Calculus Compiler: Part III: First-Order Functions. * Holumbus: [24]Hayoo! Update âŠ Again. * Mikael Vejdemo Johansson (Syzygy-): [25][MATH198] Lecture 10 (last lecture) posted. * Neil Brown: [26]Automated Wiring of Message-Passing Processes. * Computer Systems Design Laboratory at the University of Kansas: [27]ChalkBoard Release. * Bryan O'Sullivan: [28]Dense? Dense, you say?. * Ivan Lazar Miljenovic: [29]Waddaya know, testing WORKS!. * Ivan Lazar Miljenovic: [30]If wishes were tests, code would be perfect. * Ivan Lazar Miljenovic: [31]The Problems with Graphviz. Quotes of the Week * Badger: one does not simply >>= into mordor * Apocalisp: data CanHaz a = Haz a | ButIEatedIt * Apocalisp: You can't have your baby and eat it too * Dave_Benjamin: please talk to your son or daughter about parametric polymorphism * Cale: Beginners are confusing to beginners. I move that we remove them from the language altogether. * monochrom: I am 17-ary, going on 18-ary, I can take curry of you * jmillikin: [the real world is] an implementation detail of IO, pay it no mind * monochrom,ezyang: The Principle Of Idiot Savant: any sufficiently misguided opinion is indistinguishable from deep insight * Cale: I swear that most of higher-dimensional category theory must have been arrived at by some guys sitting around in a room with a blackboard and saying "What if a drew a diagram like *THIS*!?" and drawing some insane scribble up on the blackboard, and then everyone tries to figure out how to turn it into meaningful mathematics. * jbe: Here I am, happy my code compiles, and the runtime is thumbing its nose at Turing. * Ferdirand: I was TA for a C++ programming course aimed at 1st year physics once. Some girl asked for help "i wrote pseudo-code but I cannot translate it to C++". Her pseudo-code was valid haskell. I cried. * mmorrow: [regarding excessive use of categorical recursion schemes] a morphasm? * copumpkin: I'm on a rollomorphism * dons: but rumours are remarkably common when it comes to haskell * monochrom: If you read a haskell book or an FP book, by chapter 5 it's already doing data structures. It's chapter 10 in imperative books. About the Haskell Weekly News New editions are posted to [32]the Haskell mailing list as well as to [33]the Haskell Sequence and [34]Planet Haskell. [35]RSS is also available, and headlines appear on [36]haskell.org. To help create new editions of this newsletter, please see the information on [37]how to contribute. Send stories to jfredett . at . gmail . dot . com. The darcs repository is available at darcs get [38]http://patch-tag.com/r/jfredett/HWN2/pullrepo HWN2 . References 1. http://haskell.org/ 2. http://article.gmane.org/gmane.comp.lang.haskell.cafe/67149 3. http://article.gmane.org/gmane.comp.lang.haskell.cafe/67090 4. http://article.gmane.org/gmane.comp.lang.haskell.cafe/66924 5. http://article.gmane.org/gmane.comp.lang.haskell.cafe/66923 6. http://article.gmane.org/gmane.comp.lang.haskell.cafe/66920 7. http://article.gmane.org/gmane.comp.lang.haskell.cafe/66904 8. http://article.gmane.org/gmane.comp.lang.haskell.general/17661 9. http://article.gmane.org/gmane.comp.lang.haskell.general/17660 10. http://article.gmane.org/gmane.comp.lang.haskell.general/17659 11. http://article.gmane.org/gmane.comp.lang.haskell.general/17657 12. http://article.gmane.org/gmane.comp.lang.haskell.general/17656 13. http://www.diku.dk/ 14. http://article.gmane.org/gmane.comp.lang.haskell.general/17653 15. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/67211 16. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/67184 17. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/67158 18. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/66917 19. http://planet.haskell.org/ 20. http://haskell.org/haskellwiki/Blog_articles 21. http://snoyberg.wordpress.com/2009/12/04/string-like/ 22. http://claudiusmaximus.goto10.org/cm/2009-12-04_heist_dataflow_algebra.html 23. http://blog.finiteimprobability.com/2009/12/03/lambda-calculus-compiler-par… 24. http://holumbus.fh-wedel.de/blog/?p=28 25. http://blog.mikael.johanssons.org/archive/2009/12/math198-lecture-10-last-l… 26. http://chplib.wordpress.com/2009/12/02/automated-wiring-of-message-passing-… 27. http://www.ittc.ku.edu/csdlblog/?p=25 28. http://www.serpentine.com/blog/2009/12/01/dense-dense-you-say/ 29. http://ivanmiljenovic.wordpress.com/2009/11/17/waddaya-know-testing-works/ 30. http://ivanmiljenovic.wordpress.com/2009/11/26/if-wishes-were-tests-code-wo… 31. http://ivanmiljenovic.wordpress.com/2009/11/30/the-problems-with-graphviz/ 32. http://www.haskell.org/mailman/listinfo/haskell 33. http://sequence.complete.org/ 34. http://planet.haskell.org/ 35. http://sequence.complete.org/node/feed 36. http://haskell.org/ 37. http://haskell.org/haskellwiki/HWN 38. http://patch-tag.com/r/jfredett/HWN2/pullrepo%20HWN2

1 0