October 2008 - Haskell-Cafe

I want my free (and good looking) parser!
by Slavomir Kaslev 06 Oct '08

06 Oct '08

I am writing a Parsec parser for a C-like language and I have several datas that look more or less like this one: > import Control.Monad( liftM ) > import Text.ParserCombinators.Parsec > data FooBar = Foo | Bar > deriving (Show,Read,Bounded,Enum) Looking at these, it seems that there should be no need to write a parser for this guy by hand. We already know that it is bounded and enumerable, so we can get a list of all possible FooBars: > enumAll :: (Bounded a, Enum a) => [a] > enumAll = enumFromTo minBound maxBound Also, we know how to show and read each FooBar. Therefore, I should get a free parser! > freeParser :: (Enum a, Bounded a, Show a, Read a) => Parser a Here is one use of freeParser: > paramMod = option Foo freeParser > test = parseTest $ do { x <- paramMod; eof; return x } Not suprisingly: test "Foo" => Foo test "Bar" => Bar test "" => Foo I had a little hard time figuring out how this parser should look. The best I came up with was: > freeParser = freeParser' minBound > where enumAll' :: (Bounded a, Enum a) => a -> [a] > enumAll' _ = enumAll > freeParser' :: (Enum a, Bounded a, Show a, Read a) => a -> Parser a > freeParser' x = liftM read $ choice (map (string . show) (enumAll' x)) [Actually, in my code I use reserved' (reserved' x = reserved x >> return x) instead of string, where reserved is from Parsec's builtin tokenizer (which does some neat things behind the curtains). Here string is used just to illustrate the expamle.] The problem is that freeParser, although useful, is far from elegant. It's something that I came up with by trial and error. In short: it's a hack. I would like to hear your suggestions about how it can be beautified. Thank you in advance. Cheers! -- Slavomir Kaslev

3 3

editing a Alex and Happy file
by Manlio Perillo 06 Oct '08

06 Oct '08

Hi. There is support for editing an Alex or Happy file? I have seen some files that make use of tabs for layout (even Alex template files, and examples from both Alex and Happy), however tabs seems to be not recommended in Haskell. Thanks Manlio Perillo

1 0

Help me prove macid can scale! (Macid stress tests results for happs-tutorial toy job board disappointing so far.)
by Thomas Hartman 06 Oct '08

06 Oct '08

From: tphyahoo <thomashartm...(a)googlemail.com> Subject: Help me prove macid can scale! (Macid stress tests results for happs-tutorial toy job board disappointing so far.) To: HAppS HAppS is a new, relatively unproven technology. So I am asking myself this question. Will HAppS allow me to scale the toy job board I created for happs-tutorial into a high-volume, high concurrency, large-user count type job board? You know, the kind that might make money? I have done some preliminary testing to answer this question, and so far the results have been disappointing. I am hoping that I am doing something wrong, and that there is a way of using macid effectively for more than just toy applications. If there are solution for the problems I'm experiencing I will definitely be integrating this knowledge into the tutorial, so stay tuned. I am seeking feedback from HAppS experts and educated users on the following questions: * Is building a heavy duty website like monster.com in HAppS a realistic goal -- say, in the next twelve months? * Are there certain types of web apps that are unlikely to work well with the HAppS web architecture? * Are there changes I can make to my toy app's architecture -- be it data structures, buying new hardware, whatever -- that will enable me to get good performance against the stress test described below and in the demo? * Are there other HAppS stress tests in the public domain, and what are the results so far? A complete description of the stress test I did (basically, insert 200 users with 200 jobs apiece) along with a runnable demo, is in the tutorial preview at http://happstutorial.com:5002/tutorial/macid-stress-test. (The actual stress test is disabled in the online demo because I don't want to zap my server.) But the summary is, it's slow, slow to start, and creates a huge events file -- 356M. For around 40000 records. That's like, 8 megabytes per job, which is a few lines of text. OK, happs isn't a database -- but still. This is all is checked into darcs and tagged: darcs get --tag='stress test'http://code.haskell.org/happs-tutorial Thanks in advance for helping make happs usable in the real world! thomas.

3 3

Hmm, what license to use?
by Magnus Therning 06 Oct '08

06 Oct '08

Recently I received an email with a question regarding the licensing of a module I've written and uploaded to Hackage. I released it under LGPL. The sender wondered if I would consider re-licensing the code under BSD (or something similar) that would remove the need for users to provide linkable object files so that users can re-link programs against newer/modified versions of my library. Now I have fairly strong feelings about freedom of code and I everything I release is either under GPL or LGPL. What I like about those licenses is it protects freedom in a way that I think it should and it forces a sort of reciprocity which resonates very well with my selfishness. Re-licensing code under BSD is not something I'm willing to do without something that compensates for that reciprocity, and I can think of several kinds of compensation here but they all pretty much boil down to either fame or fortune. ;-) Once GHC supports dynamic binding on all platforms (or at least the major ones) this issue will (largely) go away (thanks Andrew for reporting on the state of this), but until then LGPL does create a large burden for users of my module. Until that happens I wouldn't mind re-licensing the code under a license that has the reciprocity attribute of LGPL on the source level, but does allow for static linking without requiring the availability of linkable object files. Is there such a license? I've heard that the OCaml crowd uses a modified LGPL with a static linking exception. Unfortunately I've also heard that their addition to LGPL hasn't gotten much review by lawyers, I'd much rather use something that feels less ad hoc, if you get what I mean. Any suggestions? /M -- Magnus Therning (OpenPGP: 0xAB4DFBA4) magnus＠therning．org Jabber: magnus＠therning．org http://therning.org/magnus identi.ca|twitter: magthe

37 96

Re: MPFR / FFI - Nefarious (Simple?) bug
by Michal Konecny 06 Oct '08

06 Oct '08

Jared Updike wrote on 6 Oct 02:25: > In order to create an arbitrary precision floating point / drop in > replacement for Double, I'm trying to wrap MPFR > (http://www.mpfr.org/) Have you looked at: http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hmpfr Michal -- |o| Michal Konecny |o| http://www-users.aston.ac.uk/~konecnym/ |o| office: (+42) (0)121 204 3462 |o| PGP key http://www-users.aston.ac.uk/~konecnym/ki.aow

1 0

Health effects
by Andrew Coppin 06 Oct '08

06 Oct '08

The other day, I sat down to eat a 2 Kg block of chocolate - one of those ones that's divided into lots of little squares. I proceeded to recursively subdivide it into smaller and smaller blocks, and then eat the individual squares in depth-first order. It was only after getting through 16 of the things that I stopped to notice that the whole bar just happens to have an exact power of two squares on it. And it was some time after *that* when I thought to myself "...woah, maybe do too much Haskell?" o_O Seriously, who recursively subdivides their food? I think I have something wrong with me...

23 40

Type family oddity
by Florian Weimer 05 Oct '08

05 Oct '08

I can't figure out why the following code doesn't compile with the October 2n GHC 6.10 beta (-XTypeFamilies -XFlexibleContexts) when the type declaration is not commented out. module T where type family RangeTrait c class InputRange r where remaining :: r -> Bool advance :: r -> r class (InputRange (RangeTrait s)) => Sequence s where erase :: RangeTrait s -> IO (RangeTrait s) -- erase_range :: (Sequence s) => RangeTrait s -> IO (RangeTrait s) erase_range r = if remaining r then do r' <- erase r erase_range r' else return r GHCi says the type is precisely as specified in the comment. However, when I activate the type declaration, GHC complains: T.hs:16:22: Couldn't match expected type `RangeTrait s' against inferred type `RangeTrait s2' In the first argument of `erase', namely `r' In a stmt of a 'do' expression: r' <- erase r In the expression: do r' <- erase r erase_range r' T.hs:17:22: Couldn't match expected type `RangeTrait s1' against inferred type `RangeTrait s2' In the first argument of `erase_range', namely `r'' In the expression: erase_range r' In the expression: do r' <- erase r erase_range r' Any suggestions? Is this a bug in GHC?

3 4

Total Functional Programming in Haskell
by Jason Dagit 04 Oct '08

04 Oct '08

Hello, I recently had someone point me to this thread on LtU: http://lambda-the-ultimate.org/node/2003 The main paper in the article is this one: http://www.jucs.org/jucs_10_7/total_functional_programming/jucs_10_07_0751_… It leaves me with several questions: 1) Are there are existing Haskell-ish implementations of the total functional paradigm? 2) Could we restructure Haskell so that it comes in 3 layers, total functional core, lazy pure partial functional middle, and IO outer layer? The idea with #2 is that similar to how we use the type system to separate the purely functional from monadic, and in particular IO actions, can we make the inner most part of Haskell total? Furthermore, could we do this in a way that makes it easy to intermingle the three layers the way we can mingle pure Haskell with IO actions? Maybe instead of using (->) as the function constructor for total functions we use a different symbol, say (|->), and the compiler knows to use the more specialized semantics on those definitions. I'm not sure how make this work for values that are total functional versus values that are just pure (partial) functional. I'm also interested in hearing about the optimizations that would be possible when all your expressions are strongly normalizing. Jason

9 14

Job Opportunity
by Michael Bott 04 Oct '08

04 Oct '08

Afternoon, I have been recommended to use this site from a couple of people I have spoken to. Sorry if this offends anyone but I have an amazing opportunity for a 2 Functional Programmers based in London. Functional Programming - (HASKELL / ERLANG / OCAML) - Financial This is an incredible opportunity for a FUNCTIONAL PROGRAMMER (Haskell / Erlang or Ocaml) to join an expanding London software house who specialise in Asset Management for top financial clients - they specialise in providing algorithmic execution solutions for hedge funds. My client are a leading provider in financial software, they are looking for a top Graduates or 1-2 years experienced Java / Functional Programmers to join the team and really make a difference to the company and work with the biggest clients in the UK. Desirable Skills - * Functional Programming - ( Haskell or Erlang or Ocaml to name 3). * Java * Degree Education - top University - 2:1 and above only. * Experience within the finance and/or banking sector would be highly advantageous. This is an URGENT requirement and my client are keen to get at least 2 exceptional developers into the team ASAP. You will join a relaxed team who work on exciting projects, and use the most up to date technologies........ To learn more about the role and the client please contact Mike at Lawrence Harvey. Please contact me if you are interested. Kind Regards, Mike Michael Bott Java Recruitment Specialist Lawrence Harvey Search & Selection Tel: +44 (0)208 875 4353 Fax: +44 (0)208 870 8116 Mobile: +44 (0)7966 554892 Email: m.bott(a)Lawrenceharvey.com <mailto:m.bott@Lawrenceharvey.com> Web: www.Lawrenceharvey.com <http://www.lawrenceharvey.com/> LinkedIn - http://www.linkedin.com/in/mikebott <http://www.linkedin.com/in/mikebott> The information transmitted contains confidential material and is intended only for the person or organisation to which it is addressed. Any review, retransmission, dissemination or other use of this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. No liability whatsoever is accepted by Lawrence Harvey Search & Selection in respect of the contents of this email.

2 1

Hackage Build Failures
by Cetin Sert 04 Oct '08

04 Oct '08

Hi, what is the best action to take if a package from hackage fails to build? Is there a recommended/established common way to deal with build failures/runtime bugs etc.? For example: cetin@linux-d312:~/Links/Elite/ac> cabal install wired Resolving dependencies... Downloading Wired-0.1.1... Configuring Wired-0.1.1... Preprocessing library Wired-0.1.1... Building Wired-0.1.1... [ 1 of 21] Compiling Data.Hardware.Internal ( Data/Hardware/Internal.hs, dist/build/Data/Hardware/Internal.o ) Data/Hardware/Internal.hs:198:15: Ambiguous occurrence `><' It could refer to either `Data.Hardware.Internal.><', defined at Data/Hardware/Internal.hs:129:4 or `Test.QuickCheck.><', imported from Test.QuickCheck at Data/Hardware/Internal.hs:11:0-21 cabal: Error: some packages failed to install: Wired-0.1.1 failed during the building phase. The exception was: exit: ExitFailure 1 Regards, CS

7 7