- Haskell-Cafe - Haskell.org

ANNOUNCE: mathlink-2.0.0.3
by Tracy Wadleigh 18 May '09

18 May '09

Already posted to the Haskell list but dons suggested I post to the cafe, as well... All: mathlink is a library for writing Mathematica packages in Haskell. One simply writes some functions of type: (MLGet a, MLPut b) => a -> IO b and provides a package specification in a simple DSL that mimics that of Mathematica's mprep utility. The result is a program that exposes functions that can be called from Mathematica. This is a complete rewrite of my original implementation and more closely captures the functionality I originally intended. I've only tested it on my own platform (64-bit Linux), but I've taken pains to make sure the code should run on any platform. (The only real issue in the code is picking the right functions to call when marshaling Ints.) Please report any tweaks required to get it to work on your platform. (That is, of course, only if, in fact, there are any other users out there. Right now, as far as I know, I'm it.) I'd also just like to hear from any Haskellers out there that also use Mathematica on a regular basis. Ping me if you're one. --Tracy

2 1

Re: [Haskell-cafe] Building a better dog house?
by michael rice 17 May '09

17 May '09

Hi Paul, I've been away from software design for more than a decade and have fallen behind on what's currently being used. I wasn't even aware of RUP. Thanks for the info. Michael --- On Sun, 5/17/09, Paul Johnson <paul(a)cogito.org.uk> wrote: From: Paul Johnson <paul(a)cogito.org.uk> Subject: Re: [Haskell-cafe] Building a better dog house? To: "michael rice" <nowgate(a)yahoo.com> Cc: haskell-cafe(a)haskell.org Date: Sunday, May 17, 2009, 3:26 PM michael rice wrote: > I was just looking at my UML (Unified Modeling Language) User Guide and discovered this: > > "The number of object-oriented methods increased from fewer than 10 to more than 50 during the period between 1989 and 1994." pg. xviii, Booch, Rumbaugh, Jacobson, 1999 > > Is there a modeling methodology recommended for functional languages? > > Michael > > UML of course is not a methodology, its a language. "Rational Unified Process" (RUP) is a methodology. There is no recommended methodology for functional programming, but large chunks of RUP and most similar methodologies have little to do with OO programming, and therefore could be used as-is. All the project planning, configuration management, requirements management and so on will work just fine. When it comes to the software design in functional languages I find it best to start by looking for a domain analysis of the problem (something that RUP includes as well, if I recall correctly). Then try to translate that domain analysis into an embedded domain specific language (EDSL). Ideally the EDSL should allow you to describe anything that is physically or logically possible in the domain, but nothing that is impossible. Then you can go ahead and create your software by translating the requirements directly into the EDSL. Paul.

1 0

Building a better dog house?
by michael rice 17 May '09

17 May '09

I was just looking at my UML (Unified Modeling Language) User Guide and discovered this: "The number of object-oriented methods increased from fewer than 10 to more than 50 during the period between 1989 and 1994." pg. xviii, Booch, Rumbaugh, Jacobson, 1999 Is there a modeling methodology recommended for functional languages? Michael

2 1

Can't install happs-tutorial because syb-with-class fails to install
by Colin Paul Adams 17 May '09

17 May '09

Data/Generics/SYB/WithClass/Derive.hs:187:26: Couldn't match expected type `Pred' against inferred type `Type' Expected type: PredQ Inferred type: TypeQ In the first argument of `map', namely `dataCxt' In the first argument of `(++)', namely `map dataCxt dataCxtTypes' Data/Generics/SYB/WithClass/Derive.hs:198:36: Couldn't match expected type `Pred' against inferred type `Type' Expected type: Q Pred Inferred type: TypeQ In the expression: conT 'Data `appT` typ In the first argument of `map', namely `(\ typ -> conT 'Data `appT` typ)' Data/Generics/SYB/WithClass/Derive.hs:227:38: Couldn't match expected type `Name' against inferred type `TyVarBndr' Expected type: [Name] Inferred type: [TyVarBndr] In the expression: ps In the first argument of `return', namely `(n, ps, map conA cs)' cabal: Error: some packages failed to install: syb-with-class-0.5.1 failed during the building phase. The exception was: exit: ExitFailure 1 -- Colin Adams Preston Lancashire

1 1

Haskell Weekly News: Issue 118 - May 16, 2009
by Brent Yorgey 17 May '09

17 May '09

--------------------------------------------------------------------------- Haskell Weekly News http://sequence.complete.org/hwn/20090516 Issue 118 - May 16, 2009 --------------------------------------------------------------------------- Welcome to issue 118 of HWN, a newsletter covering developments in the [1]Haskell community. Welcome to the Google Summer of Code special edition! I asked each of the five students with accepted GSoC projects to describe what they plan to work on. You'll find their descriptions below, with links to their blogs. And keep watching this space: as I did last summer, I plan to provide readers of the HWN with weekly updates on the progress of the GSoC projects. Google Summer of Code Haddock improvements! Isaac Dupree is [2]working on improvements to [3]Haddock. "Besides the various inevitable small fixes/improvements, my specific projects are to make cross-package documentation work, and to refactor the comment-parsing out of GHC and into the Haddock code-base." EclipseFP. Thomas Ten Cate will be [4]working on EclipseFP: "Compared to more mainstream languages, Haskell has surprisingly poor IDE support, even though its static typing system allows for much more help from the IDE than in the case of dynamic languages. For the Java language, a very mature and powerful IDE exists in the form of Eclipse. A plugin for Haskell support in Eclipse, called EclipseFP, is in the works, but its development has been standing still for some time. I will bring EclipseFP to a more usable state. For this, I will use the Scion IDE library, which interfaces with the GHC API, so that more advanced features like type inference become possible. I will also add support for Cabal. Hopefully, this type of IDE support will lead to greater acceptance and use of Haskell, and be useful for development as well as education." Improving the Haskell space profiling experience. Gergely Patai's [5]project will be focused on space profiling: "At the present moment, heap profiling Haskell programs means analysing logs off-line, using conversion tools to visualise data. However, instead of generating graphs with hp2ps, it should be possible to present the data in a graphical application in real time, which is useful while developing interactive applications, and it should also be made easier to export profiler output in different formats. The aim of the project is to create a set of tools that make heap profiling of Haskell programs easier in various ways. In particular, the following components are planned: a library to process profiler output in an efficient way and make it easily accessible for other tools in the future; a real-time visualiser (most likely using OpenGL); some kind of history manager to keep track of profiling data and make it possible to perform a comparative analysis of performance between different versions of your program; a maintainable and extensible replacement for hp2ps; and converters to provide input for other profiling tools." haskell-src-exts -> haskell-src. Niklas Broberg: "My [6]project, dubbed 'haskell-src-exts -> haskell-src' is really two projects in one wrapping. The first milestone is to bring my haskell-src-exts library to the point where it can supersede the old haskell-src library as the de facto package for haskell source manipulation. The main problem that I need to solve is to implement a scheme that lets the user decide what extensions to recognize when parsing a source document. Currently, haskell-src-exts assumes all extensions are always on, which means that some valid H98 programs will be incorrectly parsed due to stolen syntax by e.g. Template Haskell. The second milestone is to extend the focus from source code to full source documents, and implement a scheme for handling comments as well. The ultimate goal here is to have (pretty . parse) == id, to allow haskell-src-exts to be run on source documents without changing them. This would open up for some really interesting applications, in particular refactoring tools that could automatically apply transformations to a source document while still preserving comments." darcs. Last but not least, Petr Rockai will be [7]working on improvements to [8]darcs: "My project revolves around the idea of fast darcs for medium and large repositories. Three are quite a few haskellers who use darcs in their day to day (haskell) work. A fair number of hackage packages is maintained in darcs. Even though many of these repositories are of a relatively modest size, there is a number of relatively large real-world darcs repositories out there. The primary target of the project is to improve scalability of darcs for large working trees. This should help those users with existing large darcs repositories, as well as encourage people to use darcs for larger projects, whenever the development model fits. I intend to make the darcs working tree handling comparably fast to git. And then, git is written in C, hand-tuned for a specific operating system. And unlike mercurial, I do not plan to introduce a C library for low level routines. So let's prove that Haskell is up to the challenge." Announcements 2009.2.1: version freeze for Haskell Platform approaching on Monday. Don Stewart [9]announced that the last chance to propose bug fix version bumps to be included in the first minor release (2009.2.1) of the [10]Haskell Platform is Monday. Please ensure that, as maintainer for one of the 2009.2.x series of packages, any bug fixes are in place by Monday, or they'll be bumped to the next platform release. OpenGL 2.2.3.0. Sven Panne [11]announced the release of a new version of the [12]OpenGL package. This is a feature release, containing a number of changes and additions. Programming in Haskell -- solutions to exercises. Graham Hutton [13]announced that solutions to the exercises from "Programming in Haskell" are now [14]available online. Bookshelf. Emil Axelsson [15]announced the first release of [16]Bookshelf, a simple document organizer with some wiki functionality. Documents in a directory tree are displayed as a set of HTML pages. Documents in Markdown format are converted to HTML automatically using Pandoc. Request for feedback: HaskellDB + HList. Brian Bloniarz [17]requested feedback on a branch of [18]HaskellDB which replaces the home-grown Record code with HList records. RESTng 0.1 + RedHandlers 0.1 (request handlers) + YuiGrids 0.1 (yahoo grids). Sergio Urinovsky [19]announced the release of three new packages developed for a RESTful web framework called RESTng: [20]RESTng, [21]redHandlers, and [22]yuiGrid. #haskell.pt IRC channel. Marco Túlio Gontijo e Silva [23]announced the formation of the #haskell.pt channel on irc.freenode.net for Portuguese-speaking Haskellers. Fun with type functions. Simon Peyton-Jones [24]requested feedback on a [25]draft tutorial paper about type families (aka associated data types, or type functions). Discussion conflicting variable definitions in pattern. Martin Hofmann [26]asked about the possibility of repeated variables in patterns, resulting in an interesting discussion. Removing mtl from the Haskell Platform. Russell O'Connor began a [27]discussion around the possibility of removing the mtl package from the Haskell Platform, and replacing it with something more modern. Jobs PhD position in Nottingham. vxc [28]announced the availability of a new PhD position in the Functional Programming Laboratory at the University of Nottingham. The topic of research for the project is "Programming and Reasoning with Infinite Structures": it consists in the theoretical study and development of software tools for coinductive types and structured corecursion. Blog noise [29]Haskell news from the [30]blogosphere. Blog posts from people new to the Haskell community are marked with >>>, be sure to welcome them! * Leif Frenzel: [31]EclipseFP is going to be reloaded. * Tom Schrijvers: [32]Dictionaries: Eager or Lazy Type Class Witnesses?. Can type class dictionaries be optimized by treating them strictly? * Real-World Haskell: [33]RWH Now In The Kindle Store. * JP Moresmau: [34]Adding a Writer Monad transformer. * Ketil Malde: [35]Using a phantom type to label different kinds of sequences. * Ivan Lazar Miljenovic: [36]Functions All The Way Down. Ivan's talk on lambda calculus. * Thomas M. DuBuisson: [37]Fun with Distributed Hash Tables. Distributed hash tables in Haskell. * Darcs: [38]darcs joins the Software Freedom Conservancy. * Mark Wassell: [39]Just Grapefruit. Mark's first impressions of the Grapefruit library. * Roman Cheplyaka: [40]LambdaCube accepted to JSSP. Jane Street is funding development of the LambdaCube 3D rendering engine. * >>> Joel Neely: [41]BuilderBuilder: The Model in Haskell. * Brandon Simmons: [42]directory-tree module released. * >>> Sadek Drobi: [43]Paul Hudak on Haskell. An interview with Paul Hudak. * >>> dayvan cowboy: [44]Blast from the past: a stochastic monad in Haskell. * Matthew Podwysocki: [45]Type Classes Are The Secret Sauce . * Remco Niemeijer: [46]Programming Praxis - Priority Queues. A priority queue implementation using a leftist heap. * >>> Y. Liang: [47]A Lambda Calculus Interpreter in Haskell. Quotes of the Week * seydar: what's the nick of the drug addict who wrote learn you a haskell? and i mean that in the best possible way. * roconnor: String is kinda a poor data type for strings. * kyevan: I had a haskell-related dream last night. Sorta. I was beaten up by some kids because I tried to go somewhere my type didn't match, apparently. * edwardk: Haskell 98 is the Windows 98 of standards ;) * PhilipWadler: I'm delighted to learn that "a monad is a monoid in the category of endofunctors"---anyone know where I can find a good tutorial? * David Leimbach: Don't play with your monads... eventually you'll go bind. About the Haskell Weekly News New editions are posted to [48]the Haskell mailing list as well as to [49]the Haskell Sequence and [50]Planet Haskell. [51]RSS is also available, and headlines appear on [52]haskell.org. To help create new editions of this newsletter, please see the information on [53]how to contribute. Send stories to byorgey at cis dot upenn dot edu. The darcs repository is available at darcs get [54]http://code.haskell.org/~byorgey/code/hwn/ . References 1. http://haskell.org/ 2. http://haddock2009.wordpress.com/ 3. http://www.haskell.org/haddock/ 4. http://eclipsefp.wordpress.com/ 5. http://just-bottom.blogspot.com/ 6. http://nibrofun.blogspot.com/ 7. http://web.mornfall.net/tag/darcs.html 8. http://darcs.net/ 9. http://article.gmane.org/gmane.comp.lang.haskell.libraries/11120 10. http://trac.haskell.org/haskell-platform/ 11. http://article.gmane.org/gmane.comp.lang.haskell.general/17194 12. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/OpenGL 13. http://article.gmane.org/gmane.comp.lang.haskell.general/17193 14. http://www.cs.nott.ac.uk/~gmh/book.html 15. http://article.gmane.org/gmane.comp.lang.haskell.general/17182 16. http://www.cs.chalmers.se/~emax/bookshelf/Manual.shelf.html 17. http://article.gmane.org/gmane.comp.lang.haskell.cafe/58615 18. http://haskelldb.sourceforge.net/ 19. http://article.gmane.org/gmane.comp.lang.haskell.cafe/58589 20. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/RESTng 21. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/redHandlers 22. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/yuiGrid 23. http://article.gmane.org/gmane.comp.lang.haskell.cafe/58581 24. http://article.gmane.org/gmane.comp.lang.haskell.cafe/58522 25. http://haskell.org/haskellwiki/Simonpj/Talk:FunWithTypeFuns 26. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/58560 27. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/58430 28. http://article.gmane.org/gmane.comp.lang.haskell.cafe/58526 29. http://planet.haskell.org/ 30. http://haskell.org/haskellwiki/Blog_articles 31. http://cohatoe.blogspot.com/2009/05/eclipsefp-is-going-to-be-reloaded.html 32. http://tomschrijvers.blogspot.com/2009/05/dictionaries-eager-or-lazy-type-c… 33. http://www.realworldhaskell.org/blog/2009/05/15/rwh-now-in-the-kindle-store/ 34. http://jpmoresmau.blogspot.com/2009/05/adding-writer-monad-transformer.html 35. http://blog.malde.org/index.php/2009/05/14/using-a-phantom-type-to-label-di… 36. http://ivanmiljenovic.wordpress.com/2009/05/14/functions-all-the-way-down/ 37. http://tommd.wordpress.com/2009/05/13/fun-with-distributed-hash-tables/ 38. http://blog.darcs.net/2009/05/darcs-joins-software-freedom.html 39. http://www.poundstone.org/blog/?p=165 40. http://physics-dph.blogspot.com/2009/05/lambdacube-accepted-to-jssp.html 41. http://joelneely.wordpress.com/2009/05/09/builderbuilder-the-model-in-haske… 42. http://coder.bsimmons.name/blog/2009/05/directory-tree-module-released/ 43. http://www.infoq.com/interviews/paul-hudak-haskell-Qcon-SF-08 44. http://dayvancowboy.org/2009/05/08/blast-from-the-past-a-stochastic-monad-i… 45. http://codebetter.com/blogs/matthew.podwysocki/archive/2009/05/08/type-clas… 46. http://bonsaicode.wordpress.com/2009/05/05/programming-praxis-priority-queu… 47. http://yiopposite.blogspot.com/2009/05/lambda-calculus-interpreter-in-haske… 48. http://www.haskell.org/mailman/listinfo/haskell 49. http://sequence.complete.org/ 50. http://planet.haskell.org/ 51. http://sequence.complete.org/node/feed 52. http://haskell.org/ 53. http://haskell.org/haskellwiki/HWN 54. http://code.haskell.org/~byorgey/code/hwn/

1 0

Re: [Haskell-cafe] Data.FiniteMap deprecrated
by Vasili I. Galchin 17 May '09

17 May '09

Yevgeni, I am specifying "base" as a Cabal dependency ... dumb question ... I have forgotten whether I need to import Data.Map also? Vasili On Sat, May 16, 2009 at 1:05 AM, Eugene Kirpichov <ekirpichov(a)gmail.com>wrote: > It's Data.Map. It's in the base package. > > 2009/5/16 Vasili I. Galchin <vigalchin(a)gmail.com>: > > Hello, > > > > > http://hackage.haskell.org/cgi-bin/hackage-scripts/package/FiniteMap > > ... since this is deprecated what is the orthodox way to implement > "finite > > map"?? > > > > Thanks, > > > > Vasili > > > > _______________________________________________ > > Haskell-Cafe mailing list > > Haskell-Cafe(a)haskell.org > > http://www.haskell.org/mailman/listinfo/haskell-cafe > > > > > > > > -- > Eugene Kirpichov > Web IR developer, market.yandex.ru >

2 3

Request for feedback: HaskellDB + HList
by Brian Bloniarz 16 May '09

16 May '09

Hi, It's come time to share something that I've been playing around with recently: a branch of HaskellDB which replaces the home-grown Record code with HList records. It's definitely not ready for primetime, but I thought it'd be a good time to post the code and solicit some feedback from the community. HaskellDB the concept is very promising, but IMHO the code still falls short of that promise. Hopefully this is a small step in the right direction -- the advantages of using HList: * Shared implementation of extensible records * Additional features from HList * Better error messages for record misuse * "Lacks" predicates * Simpler code As an example of how this can be better, a DB insert looks like so: > insert db table $ constantRecord $ > film .=. "Munchie" .*. > director .=. Just "Jim Wynorski" .*. > emptyRecord The columns need not appear in the same order as in the database. If you forget a column, you'll get "error: No instance for (Fail (FieldNotFound (Proxy Director)))" rather than an opaque error. Using the new "insertOpt" function, Maybe columns will default to Nothing rather than needing to be specified. The details: I haven't updated everything, but there's enough to run test/TestCases.hs under Postgresql. TestCases is probably the best place to look for examples of the new syntax for now. HList had name conflicts with HaskellDB's SQL expression language ((.*.), (.++.), etc.) My temporary band-aid is to move the expression functions to Database.HaskellDB.SqlExpr, and require people to import qualified. The Attr type is gone, columns labels are untyped now. I also replaced a few instances of primitive type-level recursion with HMap/HMapOut. This makes the code simpler, and the type signatures more complex -- type families would help a lot here, I think. Feedback welcome! You can find my darcs tree at: http://mysite.verizon.net/vzewxzuh/sitebuildercontent/sitebuilderfiles/hask… It also requires minor changes to HList, available at: http://mysite.verizon.net/vzewxzuh/sitebuildercontent/sitebuilderfiles/hlis… I'll talk to the HList people about getting those merged. Thanks! Brian Bloniarz _________________________________________________________________ Hotmail® has a new way to see what's up with your friends. http://windowslive.com/Tutorial/Hotmail/WhatsNew?ocid=TXT_TAGLM_WL_HM_Tutor…

1 0

Graphs and graph algorithms?
by Cory Knapp 16 May '09

16 May '09

Hey, Besides fgl, are there any graph libraries in Haskell that are still maintained? Are there other papers (or books) besides Erwig's that I could use to understand how graph algorithms have been implemented in functional languages? Has anything even been published on the topic since Erwig's paper? Thanks, Cory

3 2

typeclasses comprehension problems: situation classes?
by Belka 16 May '09

16 May '09

Hello, cafe visitors! I'm trying to learn Haskell typeclasses, - about how to use them, - but can't handle some conceptiual problems, which confuses me a lot. I took one real problem (ErrorInfo gragual gathering), to tackle it in my studies: I have a class of situations: there is an object, and it gets cumulatively filled (or updated) with content. The code is in the end - it consists of 3 versions: 1. My first try. Fast written, based on intuitive understanding. Failed. 2. The second try - exploring an open world assumption. I hoped this would set me on the right path. Failure. 3. Surrendered to compiler - statisfied all it's requirements. This code looks absurd to me: parameter-never-to-be-used, unwanted-defaults. Compiler accepted this one though. Problems: 1. How to define *fillerRole* correctly, so that it depends on the type-value of "src_t"? 2. How to define *initFillable* correctly, so that it depends only on the type-value "filled_t", which is specified by the context of evaluation? 3. What are my misconcepts in the use of Haskell typeclasses here? 4. Maybe I should distinguish *situation class* (as this one), as something unavailable in Haskell? This assumption is the last one to make... I'd rather belive, that there is something I'm not aware of (for a considerably long time already) in Haskell. A lack of some programming technique Please, Help! Regards, Belka ==TRY=1===DOESN'T=COMPILE================================================== {-# LANGUAGE MultiParamTypeClasses #-} class FillsConsideringRoles src_t filled_t role_t where initFillable :: filled_t fillerRole :: role_t fill :: src_t -> filled_t -> filled_t ------------------------------------------------------------ data Role = Role1 | Role2 deriving (Show) data FillableObject = FillableObject { foData1 :: Maybe (Int, Role) , foData2 :: Maybe (Int, Role) } deriving (Show) newEmptyFillableObject :: FillableObject newEmptyFillableObject = FillableObject Nothing Nothing data Constructor1 = Constructor1 Int data Constructor2 = Constructor2 Int instance FillsConsideringRoles Constructor1 FillableObject Role where initFillable = newEmptyFillableObject fillerRole = Role1 fill c fo = let (Constructor1 i) = c in fo { foData1 = Just (i, fillerRole) } instance FillsConsideringRoles Constructor2 FillableObject Role where initFillable = newEmptyFillableObject fillerRole = Role2 fill c fo = let (Constructor2 i) = c in fo { foData2 = Just (i, fillerRole) } main = putStrLn $ show $ fill c2 $ fill c1 initFillable where c1 = Constructor1 76 c2 = Constructor2 43 ==TRY=1==[END]=================================================== ==TRY=2===DOESN'T=COMPILE======================================== {-# LANGUAGE MultiParamTypeClasses #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE OverlappingInstances #-} class FillsConsideringRoles src_t filled_t role_t where initFillable :: filled_t fillerRole :: role_t fill :: src_t -> filled_t -> filled_t ---------------------------------------------------------------------- data Role = DefaultRole | Role1 | Role2 deriving (Show) data FillableObject = FillableObject { foData1 :: Maybe (Int, Role) , foData2 :: Maybe (Int, Role) } deriving (Show) newEmptyFillableObject :: FillableObject newEmptyFillableObject = FillableObject Nothing Nothing data Constructor1 = Constructor1 Int data Constructor2 = Constructor2 Int instance FillsConsideringRoles filler_t filled_t Role where fillerRole = DefaultRole instance FillsConsideringRoles Constructor2 filled_t Role where fillerRole = Role2 instance FillsConsideringRoles Constructor1 filled_t Role where fillerRole = Role1 instance FillsConsideringRoles filler_t FillableObject role_t where initFillable = newEmptyFillableObject instance FillsConsideringRoles Constructor1 FillableObject Role where fill c fo = let (Constructor1 i) = c in fo { foData1 = Just (i, fillerRole) } instance FillsConsideringRoles Constructor2 FillableObject Role where fill c fo = let (Constructor2 i) = c in fo { foData2 = Just (i, fillerRole) } main = putStrLn $ show $ fill c2 $ fill c1 initFillable where c1 = Constructor1 76 c2 = Constructor2 43 ==TRY=2==[END]=================================================== ==TRY=3===WORKS================================================= {-# LANGUAGE MultiParamTypeClasses #-} class FillsConsideringRoles src_t filled_t role_t where initFillable :: ((),src_t, role_t) -> filled_t fillerRole :: ((),src_t, filled_t) -> role_t fill :: ((),role_t) -> src_t -> filled_t -> filled_t ------------------------------------------------------------ data Role = DefaultRole | Role1 | Role2 deriving (Show) data FillableObject = FillableObject { foData1 :: Maybe (Int, Role) , foData2 :: Maybe (Int, Role) } deriving (Show) newEmptyFillableObject :: FillableObject newEmptyFillableObject = FillableObject Nothing Nothing data Constructor1 = Constructor1 Int data Constructor2 = Constructor2 Int instance FillsConsideringRoles Constructor1 FillableObject Role where initFillable _ = newEmptyFillableObject fillerRole _ = Role1 fill _ c fo = let (Constructor1 i) = c in fo { foData1 = Just (i, fillerRole ((), c, fo)) } instance FillsConsideringRoles Constructor2 FillableObject Role where initFillable _ = newEmptyFillableObject fillerRole _ = Role2 fill _ c fo = let (Constructor2 i) = c in fo { foData2 = Just (i, fillerRole ((), c, fo)) } main = putStrLn $ show $ fill ((), DefaultRole) c2 $ fill ((), DefaultRole) c1 ((initFillable ((), Constructor1 (-1), DefaultRole)) :: FillableObject) where c1 = Constructor1 76 c2 = Constructor2 43 ==TRY=3==[END]=================================================== -- View this message in context: http://www.nabble.com/typeclasses-comprehension-problems%3A-situation-class… Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com.

3 2

Foldable for BNFC generated tree
by Deniz Dogan 16 May '09

16 May '09

Hi I have a bunch of data types which are used to represent a JavaScript program. The data types and a lexer and a parser have all been generated using BNFC. So basically an entire JavaScript program is represented as a tree using these data types. Ideally I'd like to be able to "fold" over this data structure, but I can't instantiate Foldable for my data types, since the data types all have kind *, if I'm not completely lost. Here's an example data type: data Statement = StmtFunDecl JIdent [JIdent] ExprOrBlock | StmtVarDecl [VarDecl] | StmtLetDecl [VarDecl] | StmtWhile Expr Statement | ... So, basically I'd like some sort of folding functionality for these data types, without having to hack the lexer/parser myself (parameterising the data types), because as I said they're being generated by BNFC. I noticed that you can make BNFC generate GADTs instead of normal ADTs, which would allow me to instantiate Foldable, but I'm not entirely sure that this is the best way to do this. Any help is appreciated, Deniz Dogan

3 3