- Haskell-Cafe - Haskell.org

Phantoms
by Andrew Coppin 18 Aug '08

18 Aug '08

I just (re)discovered that I can do things like data Foo x = Foo Int Int Now "Foo Int" and "Foo Double" are, as far as the type checker cares, two completely different types, even though in fact they are the same. This is actually Quite Useful, in the particular case I'm working on. Here's the thing though... How do I get it so that "Foo Int" and "Foo Double" produce slightly different strings when printed?

7 11

the cabal category field and "uploads"
by Galchin, Vasili 18 Aug '08

18 Aug '08

Hi Ross, When I upload a code contribution, how do I specify which category in the Hackage database, e.g. System? Kind regards, Vasili

2 2

Houston area FPers....
by Galchin, Vasili 18 Aug '08

18 Aug '08

Hi Austin, I live next to the NASA Johnson Space Center. JSC is minutes away from Uof H Clear Lake. Which campus are you going to? (main campus?). Kind regards, Vasili

1 1

Fwd: Haskell job opportunity: Platform Architect at Peerium, Inc.
by Don Stewart 18 Aug '08

18 Aug '08

Hey all, Below is a job opportunity for a Haskell programmer at Peerium, Inc. I was asked for forward to the community. Enjoy! -- Don ------------------------------------------------------------------------ Subject: Haskell Job Opportunity Platform Architect at Peerium, Inc. Please mail resumes to: hr(a)peerium.com Location: Cambridge, MA (Greater Boston Area) http://www.peerium.com Type: Full-time Skills: Proficiency and a strong interest in Haskell programming :) Bachelor's degree in computer science or equivalent from a four-year institution. Job Description Platform architects work with the founders to design and implement components of the runtime and supporting libraries for our software platform. Company Description At Peerium, we're striving to bring a new level of quality and efficiency to online communication and collaboration within virtual communities, social networks, and business environments. We believe that a new environment that supports the effortless sharing of both information and software will enable a level of online cooperation far beyond current Web-based technologies -- modern programming techniques will enable the creation of more robust and more powerful programs within these environments. ----- To this end, we're building a new software platform for direct, real- time communication and collaboration within graphically rich environments. We're currently building our first publicly available software product using many in-house technologies, including a powerful new programming language. Our first software product brings together a broad range of technologies, including: * Concurrent, distributed, functional programming * Graphically-rich, OpenGL interfaces * Peer-to-peer network protocols * Cryptographically secure communication * Distributed computation * Ubiquitous online agents ----- Peerium is located in the heart of Harvard Square in Cambridge, Massachusetts.

2 3

Re: [Haskell-cafe] permutations and performance
by Henning Thielemann 18 Aug '08

18 Aug '08

John D. Ramsdell schrieb: > Try deleting it and see what happens. Erm, yes, this case is wrong: >>>> | n <= 0 = [] There is _one_ permutation with no elements, namely [], thus it must be | n == 0 = [[]]

2 1

permutations and performance
by John D. Ramsdell 18 Aug '08

18 Aug '08

I tried to replace a permutation generator with one that generates each permutation from the previous one, in a stream-like fashion. I had hoped the stream-based algorithm would be more efficient because I use only one permutation at a time, so only the permutation and the previous one need be in memory at one time. I thought I'd share the results of testing the two algorithms. I first forced the algorithms to produce answers by printing the length of their results. Bad idea. The stream-based algorithm produces a stack overflow on an input that it can handle when the contents of every permutation is forced. In this run, touch = length. $ ghc -O perms.lhs $ echo '(True, 9)' | ./a.out Stack space overflow: current size 8388608 bytes. Use `+RTS -Ksize' to increase it. $ echo '(False, 9)' | ./a.out 362880 $ I forced all parts of the computation by summing all of the numbers in the output. The result show the more obvious algorithm is faster. $ ghc -O perms.lhs $ echo '(True, 12)' | time ./a.out 31614105600 299.56user 0.97system 5:00.75elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+479minor)pagefaults 0swaps $ echo '(False, 12)' | time ./a.out 31614105600 213.86user 0.55system 3:34.90elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+841minor)pagefaults 0swaps $ > module Main(main) where > main = > do (new, n) <- readLn :: IO (Bool, Int) > case new of > True -> print $ touch $ npermutations n > False -> print $ touch $ permutations n Touch all the numbers in the output. Originally, touch = length. > touch :: [[Int]] -> Int > touch xs = > sum (map sum xs) The permutation algorithm used by Serge Mechveliani in The Algebraic Domain Constructor DoCon. The idea of the algorithm was suggested to him by S.M.Abramov. > npermutations :: Int -> [[Int]] > npermutations n = > first : next (spanMonotoneous first) > where > first = take n [0..] > next (_ , []) = [] > next (decr, j:js) = > p : next (spanMonotoneous p) > where > p = concat [reverse smallers, [j], reverse greaters, [i], js] > (greaters, i:smallers) = span (> j) decr > spanMonotoneous (x:y:xs) > | x <= y = ([x], y:xs) > | otherwise = (x:ys, zs) > where > (ys,zs) = spanMonotoneous (y:xs) > spanMonotoneous xs = (xs, []) > p : next (spanMonotoneous p) > where > p = concat [reverse smallers, [j], reverse greaters, [i], js] > (greaters, i:smallers) = span (> j) decr > spanMonotoneous (x:y:xs) > | x <= y = ([x], y:xs) > | otherwise = (x:ys, zs) > where > (ys,zs) = spanMonotoneous (y:xs) > spanMonotoneous xs = (xs, []) Straight forward permation algorithm. > permutations :: Int -> [[Int]] > permutations n > | n <= 0 = [] > | n == 1 = [[0]] > | otherwise = > concatMap (insertAtAllPos (n - 1)) (permutations (n - 1)) > where > insertAtAllPos x [] = [[x]] > insertAtAllPos x (y : l) = > (x : y : l) : map (y :) (insertAtAllPos x l)

3 3

Houston-area FPers?
by Austin Seipp 17 Aug '08

17 Aug '08

Hi, In less than a week I'll be moving to Houston TX in order to start school at university (University of Houston.) I'm wondering if there are any functional programmers (particularly haskellers) in that part of the state? If so, a group meeting and perhaps eventually a user-group would be lovely - if only to chat and talk about what everybody is working on for the time being until something more formal can be set up. I'm posting this a little early to try and get the most responses possible - unfortunately at the moment I will not have my own mode of transportation around town (see: car) and I'm not a houston native as you can see, so it will take time for me to get up to speed with college *and* getting around town, but if anybody is in that area I'd like to hear back from you. Austin

1 0

GHC 6.8.3 binary for older linux
by Agent Zhang 17 Aug '08

17 Aug '08

Halo~ I ran into the following error while trying to run a binary generated by GHC 6.8.x on our production machines with a not-so-recent linux installed (kernel 2.6.9). $ ./restyscript restyscript: timer_create: Invalid argument Evan building a pure static-linking executable did not solve the problem. Then I downloaded the binary GHC 6.8.3 from haskell.org on that machine, but the ghc crashes as well: $ ghc Floating point exception My teammate chaos++ found that the last time precision argument passed to the timer_create function was just too big. Manually editing the binary executable (the restyscript file in the previous example) solved this issue completely, but, yeah, it's terribly hacky. So I decided to build a GHC 6.8.3 from source on that old system using GHC 6.4.2. Fortunately, the binary GHC 6.4.2 from haskell.org does work there. The newly-built GHC solves all the problem. No floating-point exception nor invalid argument for timer_create. I've put my binary distribution on my site here: http://agentzh.org/misc/ghc-6.8.3-i386-old-linux.tar.bz2 Hopefully it'll be useful for someone else ;) Not sure if it's worth putting to the official download page as well :) Ideally GHC should inspect the kernel version and timer_create support at *runtime*, rather compile-time. It'll make our lives much easier ;) Thanks! -agentzh

2 1

Cabal has trouble with projects that have both src and lib directories?
by Nicholas Andrews 17 Aug '08

17 Aug '08

Hi! I'm trying to set up a project with both src and lib directories with cabal. My blah.cabal file looks like: Library Build-Depends: base Exposed-Modules: Foo hs-source-dirs: lib/foo Executable hai Build-depends: base Main-is: Bar.hs ghc-options: -O hs-source-dirs: src/bar Other-Modules: Foo extra-libraries: Foo extra-lib-dirs: lib/bar But I get: $ runhaskell blah.cabal configure $ runhaskell blah.cabal build src/bar/Bar.hs:6:7: Could not find module `Foo': Use -v to see a list of the files... In Bar.hs, line 6 says "import Svm" What's wrong? Thanks!

3 3

duplicate symbols for ODBC-1.1.4.4 when running with GHCi
by Andrew U. Frank 17 Aug '08

17 Aug '08

when i run a program (which works with GHC) under GHCi (in eclipse with FP support) then i get GHCi linker duplicate definition for symbol _sqlSucceeded whilst processing .... HSHDBC-odbc-1.1.4.4.o what is wrong here? any suggestion what i can change? thank you andrew

1 1