September 2006 - Haskell

BitSyntax for Haskell
by Adam Langley 23 Sep '06

23 Sep '06

Erlang's bit syntax[1] is a great for building and breaking up binary structures. I've knocked up something similar (although a little clumsy) for Haskell: http://www.imperialviolet.org/binary/bitsyntax/ http://www.imperialviolet.org/binary/bitsyntax/BitSyntax.hs I'm sure that this isn't the best possible way to do this, but it suffices at this stage for many problems. [1] http://www.erlang.org/doc/doc-5.4.12/doc/programming_examples/bit_syntax.ht… -- Adam Langley agl(a)imperialviolet.org

1 0

ANNOUNCE: Dr Haskell 0.1
by Neil Mitchell 23 Sep '06

23 Sep '06

Hi Haskell, This is the first release of Dr Haskell, a tool to help suggest improvements to your code. GETTING DR HASKELL See http://www-users.cs.york.ac.uk/~ndm/projects/drhaskell.php for details on how to download and install the program. There are prebuilt Windows binaries. This project requires Yhc to be installed and working, of which there are also prebuilt binaries (see the above web page for pointers). This program should work fine on Linux/Mac etc, provided you first have Yhc installed. AN EXAMPLE SESSION Imagine you have a program, and somewhere deep inside the program it has the expression: take 5 (repeat '!') An experienced Haskell user might realise that this is in fact replicate, but a beginner has no way of discovering this function for themselves, besides reading the standard libraries! Put that definition in a Haskell source file, say Test.hs, then run: > drhaskell Test.hs I can apply use_replicate in check_elem (5:10-6:80) Dr Haskell gives a suggestion (use_replicate) to a specific function (check_elem) and a given source location (line 5, column 10) Of course, they might have written take 5 $ repeat '!', or take n . repeat, or quite a few other variations. Dr Haskell can see through these things nicely. Of course, Dr Haskell can make more powerful suggestions, for example: if f x then 4 else if g x then 4 else primes !! 10 And Dr Haskell will suggest you use f x || g x instead The hints are isolated from the program, and can be easily changed by the user. More hint patches are most welcome. See http://www.cs.york.ac.uk/fp/darcs/drhaskell/Hints.hs for a list of the hints currently available in the code. LIMITATIONS Requires a working copy of Yhc. You must first have your program/file successfully compiling with Yhc. Once you have got over the admin side (compiling etc, invoking correctly), the actual hint algorithm is pretty robust. Bug reports/suggestions/patches most welcome. YHC INFORMATION Home page: http://haskell.org/haskellwiki/Yhc Build details: http://haskell.org/haskellwiki/Yhc/Building Windows snapshot: http://haskell.org/haskellwiki/Yhc/Snapshots If Yhc won't build for you, or you have other difficulties, please bug the Yhc Team on their mailing list, yhc -at- haskell -dot- org. Thanks Neil

1 0

ANNOUNCE: new release of Hugs
by Ross Paterson 23 Sep '06

23 Sep '06

We are pleased to announce a new minor release of Hugs, fixing a few bugs with the May 2006 release, and with libraries roughly matching the forthcoming GHC 6.6 release. It is available from the Hugs page: http://www.haskell.org/hugs/ As with the May 2006 release, two sizes of the Windows distribution and two sizes of source distribution are available for download. Distributions for other platforms will be linked to when they are provided by third parties. Packagers can choose between an omnibus package and a more modular arrangement.

3 2

Implicit type of numeric constants
by Robert Stroud 22 Sep '06

22 Sep '06

Hello, I'm a newcomer to Haskell and I've been typing some simple expressions into Hugs in an attempt to understand how the built-in numeric types work . However, I'm puzzled by the following example: k = 2 f :: Int -> Int -> Int f x y = x * y After loading a file containing this code into Hugs, if I type "f 2 2" I get the value 4 as expected, but if I type "f k k" I get a type error because the type of k is inferred to be Integer. This seems like a violation of referential transparency to me - I would expect the inferred type of k to be the same as the type of 2. However, if I type an apparently equivalent let expression into Hugs directly, then I get the value 4 as expected let k = 2 ; f :: Int -> Int -> Int ; f x y = x * y in f k k Why is there a difference in behaviour? For that matter, if the type of 2 is "Num a => a", which I understand to mean "an arbitrary numeric type", why is it OK to pass 2 to a function that expects an Int? Thank you. Robert Stroud

4 12

excel crashes when using dll
by nikunj 22 Sep '06

22 Sep '06

I have created a dll that works perfectly in the debug mode. But when i try to build it for release or for that matter even use the debug version of the library on any other computer without the debug excel crashes and gives me the "excel has encountered a problem window". do any of you guys have any ideas what might be the reason? regards nikunj -- View this message in context: http://www.nabble.com/excel-crashes-when-using-dll-tf2319001.html#a6451273 Sent from the Haskell - Haskell mailing list archive at Nabble.com.

1 0

Re[2]: [Haskell-cafe] Re: Haskell web forum
by Bulat Ziganshin 21 Sep '06

21 Sep '06

This message (from haskell-cafe) by Kurt Hutchinson exactly explains that i want and why i propose do to it. sorry for inconvenience, i don't propose to stop mail lists and plan to present in both medias (while others will continue to use irc, blogs, LJs, rss feeds and all other ways to let Haskell rule the world :) > I was just about to ask the same thing. Clearly, some prefer email > and some prefer web forums. This is undoubtably true of new Haskell > users with questions, as well. Let those interested in a web forum > set one up and run it. Those interested in email can ignore the web > forum. The Perl community has a similar setup with the popular > Perlmonks web forum, and their hundreds of mailing lists. There are > many users who read and post on both, and many who do not. > I believe Bulat was just trying to find out if anyone else would like > to help set one up, not suggesting that it replace the mailing lists. Thank you, Kurt, for detailed explanation -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com

1 0

Injecting stateful computations into pure Haskell code.
by szefirov＠ot.ru 21 Sep '06

21 Sep '06

It is widely spread convention to simulate electronic circuits using infinite lazy lists. This works well for high level modeling, but when engineers ask to cosimulate it with Verilog or VHDL modules it goes somewhat off of pure functional programming. The main obstacle is that those modules have their own internal state. Is it possible to represent a result of several impure calculation as a pure list? The only way I can think of is the use of unsafePerformIO. Maybe State monad will help? And how?

2 1

ANNOUNCE: Codec.Compression.GZip & .BZip
by Duncan Coutts 21 Sep '06

21 Sep '06

I'm pleased to announce two new packages: zlib and bzlib which provide functions for compression and decompression in the gzip and bzip2 formats: Both provide pure functions on streams of data represented by lazy ByteStrings: compress, decompress :: ByteString -> ByteString This makes it easy to use either in memory or with disk or network IO. For example a simple gzip compression program is just: > import qualified Data.ByteString.Lazy as ByteString > import qualified Codec.Compression.GZip as GZip > > main = ByteString.interact GZip.compress Or you could lazily read in and decompress @.gz@ file using: > content <- fmap GZip.decompress (readFile file) The code is available via darcs: darcs get http://haskell.org/~duncan/zlib darcs get http://haskell.org/~duncan/bzlib (Note that if you are using GHC-6.5 then you'll need to edit the .cabal file to remove the dependency on fps, since ByteStrings are provided in the base package in GHC-6.5.) There is API documentation too: http://haskell.org/~duncan/zlib/docs http://haskell.org/~duncan/bzlib/docs Both packages are bindings to the corresponding C libs, so they depend on those C libraries. Fortunately both zlib and bzlib2 are available on every OS. It also means that the compression speed is as you would expect since it's the C lib that is doing all the work. I'm very happy to get feedback on the API, the documentation or of course any bug reports. Duncan

2 1

Deferred instance declarations (serialization of existential boxes)
by Misha Aizatulin 21 Sep '06

21 Sep '06

hello all, I asked this question on haskell-cafe once, but my post didn't get any responses, so I try once again here explaining a bit more where my problem comes from. Suppose I have a monomorphic datatype, say T and want to do something with it, which depends on an instance of class C being defined for T. However I can only provide this instance later, in a different module. So I would like to write something like > f :: C T => T -> T or > instance C T => C T' where > -- a body that uses functions from C T By this I mean that f can only be called or C T' can only be derived in a module that defines C T. However, I don't see any way to make this work with ghc. Note that if I had a polymorphic type like (T a), everything would be fine. What I need it for is serialization of datatypes with existential constructors (see http://haskell.org/pipermail/haskell-cafe/2006-September/018041.html) I am having a box like > data Box = forall a. Cxt a => Box a and want to write a Read instances for it. The problem is that currently I can only do that when I am aware of all types that will ever go into the box, which is not in the same module where the Box is defined. But there is code in the same module that needs to refer to this instance. I hope you see now where my problem comes from. I would be very thankful for any help or suggestions. If someone has implemented serialization with existentials before, it would be really interesting, if they had a similar problem. Another question: would it be possible to allow definitions of f and C T' above in GHC? Cheers, Misha P.S. Currently I am solving this problem by introducing recursively dependent modules.

3 2

Haskell Weekly News: September 18, 2006
by dons＠cse.unsw.edu.au 20 Sep '06

20 Sep '06

--------------------------------------------------------------------------- Haskell Weekly News http://haskell.org/haskellwiki/HWN Issue 41 - September 18, 2006 --------------------------------------------------------------------------- Welcome to issue 41 of HWN, a weekly newsletter covering developments in the Haskell community. Each week, new editions are posted to [1]the Haskell mailing list as well as to [2]the Haskell Sequence and [3]Planet Haskell. [4]RSS is also available, and headlines appear on [5]haskell.org. The 2006 Haskell Workshop was held today in Portland, Oregon. Thanks to Edward Kmett for a report on the event. 1. http://www.haskell.org/mailman/listinfo/haskell 2. http://sequence.complete.org/ 3. http://planet.haskell.org/ 4. http://sequence.complete.org/node/feed 5. http://haskell.org/ Announcements * Haskell98 Termination Analyser . Stephan Swidersk [6]announced the integration of an automatic Haskell98 termination analyzer in the termination tool AProVE. The tool accepts full Haskell as specified in the Haskell 98 Report and is available through our web interface. [7]More 6. http://article.gmane.org/gmane.comp.lang.haskell.general/14193 7. http://aprove.informatik.rwth-aachen.de/ * Free theorems . Janis Voigtlaender [8]announced that Sascha Boehme has done a project to implement the Reynolds/Wadler algorithm generating theorems from polymorphic types, plus simplifications and postprocessings for such free theorems. [9]More info 8. http://article.gmane.org/gmane.comp.lang.haskell.general/14160 9. http://haskell.as9x.info/ * Haddock/GHC SoC . David Waern [10]announced a short status report of the "Port Haddock to use GHC" Summer of Code project. The GHC modifications, are finished and will be included in the GHC head repository soon. 10. http://article.gmane.org/gmane.comp.lang.haskell.general/14149 * AutoForms release 0.2 . Mads Lindstr�m [11]released AutoForms 0.2, a library to ease the creation of GUIs. It does this by using generic programming (SYB) to construct GUI components. [12]More info 11. http://article.gmane.org/gmane.comp.lang.haskell.general/14148 12. http://autoforms.sourceforge.net/ * HSPClientside 0.2 . Joel Bj�rnson [13]announced a new version of HSPClientside (0.2) ,developed as a GSoC project during this summer. HSPClientside is a Haskell Server Pages library for generating JavaScript code. [14]More info 13. http://article.gmane.org/gmane.comp.lang.haskell.general/14133 14. http://darcs.haskell.org/SoC/hsp.clientside/ * SOE implementation based on Gtk2Hs . Duncan Coutts [15]Due to popular demand the new SOE implementation based on Gtk2Hs is [16]available. The rendering quality is better than the original HGL version. [17]Here's a side-by-side comparison 15. http://article.gmane.org/gmane.comp.lang.haskell.general/14132 16. file://localhost/home/dons/dons/src/hwn/now 17. http://haskell.org/~duncan/gtk2hs/SOE-cairo.png * The experimental GHCi debugger . Pepe [18]announced the results of his SoC project, the experimental Haskell debugger. [19]More details 18. http://article.gmane.org/gmane.comp.lang.haskell.general/14131 19. http://haskell.org/haskellwiki/GHC/GHCiDebugger * SmallCheck . Colin Runciman [20]released a prototype tool that is similar in spirit, and in some of its workings, to QuickCheck. SmallCheck is, though, based on exhaustive testing in a bounded space of test values. [21]More info 20. http://article.gmane.org/gmane.comp.lang.haskell.general/14129 21. http://www.cs.york.ac.uk/fp/smallcheck0.0.tar * Frisby: composable, linear time parser for arbitrary PEG grammers . John Meacham [22]released Frisby, an implementation of the 'packrat' parsing algorithm, which parse PEG grammars and have a number of very useful qualities, they are a generalization of regexes in a sense that can parse everything in LL(k), LR(k), and more, including things that require unlimited lookahead, all in guaranteed linear time. [23]More information 22. http://article.gmane.org/gmane.comp.lang.haskell.general/14128 23. http://repetae.net/computer/frisby/ * HaskellNet . Jun Mukai [24]published a status report on the state of his SoC project, HaskellNet 24. http://article.gmane.org/gmane.comp.lang.haskell.general/14126 * GHC's new support engineer . Simon Marlow [25]announced that GHC now has a full-time support engineer, Ian Lynagh (aka Igloo on IRC). He'll be helping with all aspects of GHC, especially release management, bug diagnosis and tracking, documentation, packaging, and supporting other GHC hackers. Welcome Ian! 25. http://article.gmane.org/gmane.comp.lang.haskell.general/14125 Haskell' This section covers the [26]Haskell' standardisation process. * [27]Writing / status teams - call for volunteers * [28]do and if then else * [29]map and fmap * [30]All monads are functors 26. http://hackage.haskell.org/trac/haskell-prime 27. http://article.gmane.org/gmane.comp.lang.haskell.prime/1657 28. http://article.gmane.org/gmane.comp.lang.haskell.prime/1599 29. http://thread.gmane.org/gmane.comp.lang.haskell.prime/1625/focus=1625 30. http://thread.gmane.org/gmane.comp.lang.haskell.prime/1590/focus=1590 Discussion The Haskell Workshop . Here is a summary of the talks presented today at the Haskell Workshop * RepLib: A Library for Derivable Type Classes . Stephanie Weirich [31]presented a neat library for generic traversals. It covered just about everything you'd need and seemed pretty painless to use. From the abstract: Some type class instances can be automatically derived from the structure of types. As a result, the Haskell language includes the "deriving" mechanism to automatic generates such instances for a small number of built-in type classes. In this paper, we present RepLib, a GHC library that enables a similar mechanism for arbitrary type classes. 31. http://www.cis.upenn.edu/~sweirich/RepLib/ * A Generic Recursion Toolbox for Haskell (Or: Scrap Your Boilerplate Systematically) . Deling Ren and Martin Erwig [32]talked about another generics library, with a neat core "mother traversal" that abstracted over different data structure traversals. The Recursion Library for Haskell provides a rich set of generic traversal strategies to facilitate the flexible specification of generic term traversals. The underlying mechanism is the Scrap Your Boilerplate (SYB) approach. Most of the strategies that are used to implement recursion operators are taken from Stratego. 32. http://web.engr.oregonstate.edu/~erwig/reclib/ * Strong Types for Relational Databases . Alexandra Silva and Joost Visser [33]presented a relational database Using the type system to encode not only the types of fields it was joining, but a bit more about their meaning and role in the key. Type-level programming is used to maintain the consistency between header and rows, and between functional dependencies and tables. Type-level predicates are provided to check whether tables and functional dependencies are in particular normal-forms, or not. 33. http://lmf.di.uminho.pt/wiki/pub/PURe/CoddFish/StrongTypesForRDBs-draft-04-… * Polymorphic Variants in Haskell . Koji Kagawa [34]presented a paper on polymorphic variants in Haskell, describing some interesting type system extensions. In languages that support polymorphic variants, a single variant value can be passed to many contexts that accept different sets of constructors. Polymorphic variants are potentially useful for application domains such as interpreters, graphical user interface (GUI) libraries and database interfaces, where the number of necessary constructors cannot be determined in advance. 34. http://guppy.eng.kagawa-u.ac.jp/~kagawa/PVH * Extended Static Checking for Haskell . Dana Xu [35]gave a great talk on ESC/Haskell, proving the partial correctness of programs with compile time checks for pre/postcondition violations, and using an external theorem prover to prove bits about arithmetic expressions. Very practical stuff. The extended static checking is beautiful in that it opens up a whole new arena in Haskell for machine checkable documentation. 35. http://www.cl.cam.ac.uk/~nx200/ * Running the Manual: An Approach to High-Assurance Microkernel Development . Philip Derrin, Kevin Elphinstone, Gerwin Klein, David Cock and Manuel M.T. Chakravarty [36]presented a microkernel paper. This was really good. Before I saw the talk I really questioned the judgement of prototyping a microkernel in Haskell. Now I want to go prototype one myself. We propose a development methodology for designing and prototyping high assurance microkernels, and describe our application of it. The methodology is based on rapid prototyping and iterative refinement of the microkernel in a functional programming language. 36. http://www.cse.unsw.edu.au/~chak/papers/DEKC+06.html * Strongly Typed Memory Areas -- Programming Systems-Level Data Structures in a Functional Language . Iavor S. Diatchki and Mark P. Jones [37]gave a presentation on how you could extend the type system to include an extra kind for physical memory regions and data representations in memory, with endianness, and the like. He implemented it in a kind of strict pseudo-Haskell, but the general notation was pretty clean and would be usable for system level programming in Haskell. 37. http://www.cse.ogi.edu/~diatchki/papers/haskell007-diatchki.pdf * User-Level Transactional Programming in Haskell . Peter Thiemann wrote on generalizing the interface for STM to other kinds of database and user-interface transactions, where he gave up the ability to automatically retry blocking on a changeset of TVars, producing a kind of transaction that could actually fail and be manually retried. * An Extensible Dynamically-Typed Hierarchy of Exceptions . Simon Marlow [38]presented an extensible exception hierarchy using type classes. In this paper we address the lack of extensibility of the exception type in Haskell. We propose a lightweight solution involving the use of existential types and the Typeable class only, and show how our solution allows a fully extensible hierarchy of exception types to be declared, in which a single overloaded catch operator can be used to catch either specific exception types, or exceptions belonging to any subclass in the hierarchy. 38. http://www.haskell.org/~simonmar/papers/ext-exceptions.pdf * Demo: Introducing the Haskell Equational Reasoning Assistant . Andy Gill came in and did a really neat, high energy, demo of an interactive equational reasoning/code refactoring tool. He actually loaded up the application and showed it refactoring live. * GenI: Natural Language Generation in Haskell . Eric Kow [39]covered genI and gave a user's level overview of the features in Haskell that are useful in real code. The natural language generation stuff there was pretty nifty. 39. http://hal.inria.fr/docs/00/08/87/87/PDF/haskell01-kow.pdf * Demo: Typed Linear Algebra . Frederik Eaton covered a linear algebra library with ATLAS and LAPACK bindings in Haskell that uses type level enumerable values to provide its bounds. Neat stuff. * Demo: Interactive Debugging . David Himmelstrup unforutnately was unable to present [40]his talk, due to illness. From the abstract: Haskell's lazy nature is a large part of its beauty. Unfortunately, laziness complicates debugging in the "stop, examine, continue" sense, and hence tools such as Freja, Hat and Hood have risen. This paper describes the implementation of a "stop, examine, continue" debugger for Glasgow Haskell 40. http://darcs.haskell.org/~lemmih/hw2006.ps * Demo: The Galois Haskell Program Coverage tool, HPC . A demo was presented on how you can use HPC to interactively control QuickCheck, get it to run some cases, and check to see how much coverage you have had, then have it run more cases if thats not enough. The notion of looking at the difference in the coverage sets between a good run and a bad run to try to isolate bugs seems really useful. * Demo: On computable types . Oleg Kiselyov [41]presented a "proof by example" that the Haskell type system is Turing complete, with a 5 line type system lambda calculus and a 6 line type level fibonacci. 41. http://www.haskell.org/pipermail/haskell/2006-September/018486.html * Haskell' Status Report -- An Update on the Next Haskell Standard . Isaac Jones Isaac gave a talk on the state of [42]Haskell' 42. http://hackage.haskell.org/trac/haskell-prime Quotes of the Week * Baughn : "I think I'm beginning to understand this language. I just defined: let 1 + 1 = 3" * SamB : "Haskell is the only language I know with first-class support for imperative programming" * David Amos : "The real reason for using Haskell is that the code comes out shorter, and is quicker to write, than in imperative languages .. What that means is, I can get much more done when I use Haskell" * SyntaxNinja : "I recommend seeing if people have a major problem, then pouring concrete on them, and implementing what you want anyway" * SamB : "I was just saying that encasing poeple in concrete is not the best solution to that problem" * Sean Russell : "[On the wmii window manager mailing list] Hah! That's a laugh. Since when does 'it compiles' equate to 'it will run (correctly)'? We're talking about C, after all... not Haskell." * Adaptec : "[In documentation for one of their chipsets] We have a disclaimer because there may be corner cases" * audreyt : "Yes, I'm aware that I'm abusing GHC beyond its designed purpose" * cjeris : "I have a little perl script that aliases gcc -freduced-suckage to ghc" * dibblego : "Using logic in imperative programming language channels is generally considered trolling" * edwardk : "lambdabot is the first irc bot I ever added to my buddy list" * gmh33 : "Haskell makes so much more sense now and all I did was figure out the typing thing" * musasabi : "reallyUnsafePointerEq# seems unsafe" * sieni : "The advantage of Haskell is that it doesn't suck" * therp : "Good morning. I think I have been dreaming of shift/reset continuations" Contributing to HWN To help create new editions of this newsletter, please see the [43]contributing information. Send stories to dons at cse.unsw.edu.au. The darcs repository is available at darcs get http://www.cse.unsw.edu.au/~dons/code/hwn 43. http://haskell.org/haskellwiki/HWN

3 2