- Haskell-Cafe - Haskell.org

Job offer at Create-Net, Italy
by Corentin Dupont 22 Apr '14

22 Apr '14

Hi guys, I'm surprised I didn't got any reply from the community to this job offer, that I posted some time ago: http://www.create-net.org/work-with-us/postdoctoral-researcher-cloud-comput… While it's not 100% guarantee to work in Haskell, the position is still very interesting with technologies such as Constraint Programming (I'm currently exploring SMT with the SBV library), Cloud computing and energy efficiency. Ciao, Corentin

1 0

Correctly setting up emacs + ghc-mod
by Arnaud Bailly 22 Apr '14

22 Apr '14

Hello, I am struggling to get my emacs settings right for doing significant Haskell development, eg. something more involved than a bunch of source files in a single directory. The issue I am running in currently is that when trying to load files in the interpreter, most source paths are missing which means loading inevitably fails. This is especially annoying with tests: To get things working correctly I need to explicitly add in the console > :set -iwhere/my/src/are Is there any configuration I should be aware of in my cabal files or .emacs file that would correctly set the source roots? Am I missing something? Here is my .emacs configuration. Thanks for any help. ;; haskell coding (require 'auto-complete) (require 'haskell-mode) (require 'haskell-cabal) (autoload 'ghc-init "ghc" nil t) (add-hook 'haskell-mode-hook (lambda () (ghc-init))) (eval-after-load "haskell-mode" '(progn (setq haskell-stylish-on-save t) (setq haskell-process-args-cabal-repl '("--ghc-option=-ferror-spans" "--with-ghc=ghci-ng")) (define-key haskell-mode-map (kbd "C-,") 'haskell-move-nested-left) (define-key haskell-mode-map (kbd "C-.") 'haskell-move-nested-right) (define-key haskell-mode-map (kbd "C-c v c") 'haskell-cabal-visit-file) (define-key haskell-mode-map (kbd "C-c C-c") 'haskell-compile) (define-key haskell-mode-map (kbd "C-x C-d") nil) (setq haskell-font-lock-symbols t) ;; Do this to get a variable in scope (auto-complete-mode) ;; from http://pastebin.com/tJyyEBAS (ac-define-source ghc-mod '((depends ghc) (candidates . (ghc-select-completion-symbol)) (symbol . "s") (cache))) (defun my-ac-haskell-mode () (setq ac-sources '(ac-source-words-in-same-mode-buffers ac-source-dictionary ac-source-ghc-mod))) (add-hook 'haskell-mode-hook 'my-ac-haskell-mode) (defun my-haskell-ac-init () (when (member (file-name-extension buffer-file-name) '("hs" "lhs")) (auto-complete-mode t) (setq ac-sources '(ac-source-words-in-same-mode-buffers ac-source-dictionary ac-source-ghc-mod)))) (add-hook 'find-file-hook 'my-haskell-ac-init))) (add-hook 'haskell-mode-hook 'turn-on-haskell-decl-scan) (add-hook 'haskell-mode-hook 'turn-on-haskell-indentation) (eval-after-load "which-func" '(add-to-list 'which-func-modes 'haskell-mode)) (eval-after-load "haskell-cabal" '(define-key haskell-cabal-mode-map (kbd "C-c C-c") 'haskell-compile))

3 7

Re: [Haskell-cafe] "add ifM and whenM to Control.Monad" and other discussions like that
by Yuri de Wit 22 Apr '14

22 Apr '14

I haven't been around for long in the Haskell community, but I have seen a few of these discussions taking long email rounds in different mailing lists. I don't have a strong opinion at this point to add my vote here or there one way or the other, and I do appreciate the effort made by everyone to find a sort of consensus, as hard as that can be sometimes. However, the question I keep asking myself is why can't the compiler disambiguate between two definitions with the same name based on the actual argument types? This would allow packages to choose their own names for functions without such global-naming discussions and without much pain when using them. e.g.: > import PackageA (when) > import PackageB (when) > when (Just True) doSomething versus: > when True 1 Is there a theoretical or practical issue that makes this an undesirable feature in Haskell?

2 1

Does this typeclass have a name?
by Bardur Arantsson 21 Apr '14

21 Apr '14

Hi all, I was wondering if anyone out there knows if this type class has a name? class Foo a e where foo :: e -> a -> a bar :: a (I also have a fundep a -> e, but that's not essential.) Essentially the usage is: You have a sequence of "events" e_i and a starting value "bar" and can use the "foo" function to apply all events to that starting value foo e_n $ foo e_{n-1} $ ... $ foo e_0 $ bar and thus get the final value of the a induced by the events e_i. (I've included the "bar" starting value in the type class, but I suppose it could be supplied by a Default instance.) Regards,

6 10

Re: [Haskell-cafe] conduit zips
by Ben Foppa 21 Apr '14

21 Apr '14

Don't know how I missed those. Thanks! On Mon, Apr 21, 2014 at 3:04 PM, Michael Snoyman <michael(a)snoyman.com>wrote: > Instead of functions, they're now newtype wrappers: ZipSource, ZipSink, > and ZipConduit (that last one never existed as a function). Some > information on each of these: > > * http://www.yesodweb.com/blog/2013/09/zipsinks-conduit-extra > * http://www.yesodweb.com/blog/2014/02/conduit-zip-source > * > http://stackoverflow.com/questions/21671688/single-stepping-a-conduit/21684… > > > On Mon, Apr 21, 2014 at 3:52 AM, Ben Foppa <benjamin.foppa(a)gmail.com>wrote: > >> Since conduit-1.1.*, the package hasn't contained conduit zipping >> functions anymore. They don't appear to have been moved to conduit-extra, >> either. Did they get lost in the ether, or are they in a separate package? >> >> _______________________________________________ >> Haskell-Cafe mailing list >> Haskell-Cafe(a)haskell.org >> http://www.haskell.org/mailman/listinfo/haskell-cafe >> >> >

1 0

ANN: v0.0.8 of treeviz package released.
by David Banas 21 Apr '14

21 Apr '14

- Split code into multiple files. - Completed Haddoc documentation. - Changed decimation style flag from Boolean to custom type, DecimationStyle, w/ constructors: DIT & DIF, as per Mark L.’s suggestion. - Introduced type alias: type Radix = Int, for readability. cabal install treeviz http://www.haskell.org/haskellwiki/Treeviz Cheers, -db

1 0

GHC-7.8.2 & Win
by Cédric Chantepie 21 Apr '14

21 Apr '14

Hi, Try to setup GHC 7.8.2 & cabal 1.8 on Windows (7/64bits), using binary distributions from haskell.org I get error when I try to install glib package using cabal : cabal: Missing dependency on a foreign library: * Missing (or bad) header file: primitive-memops.h This problem can usually be solved by installing the system package that provides this library (you may need the "-dev" version). If the library is already installed but in a non-standard location then you can use the flags --extra-include-dirs= and --extra-lib-dirs= to specify where it is. If the header file does exist, it may contain errors that are caught by the C compiler at the preprocessing stage. In this case you can re-run configure with the verbosity flag -v3 to see the error messages. Failed to install primitive-0.5.2.1 cabal: Error: some packages failed to install: Command cabal install primitive raise the same error. I have tried either within cygwin or from plain cmd. Thanks for any hint, best.

2 1

conduit zips
by Ben Foppa 21 Apr '14

21 Apr '14

Since conduit-1.1.*, the package hasn't contained conduit zipping functions anymore. They don't appear to have been moved to conduit-extra, either. Did they get lost in the ether, or are they in a separate package?

1 0

A list library without fixpoints and recursive types.
by flicky frans 21 Apr '14

21 Apr '14

Hello. I read some Oleg Kiselyov's papers and tried to write more or less efficient list library with lists represented as their right fold, as mentioned here http://okmij.org/ftp/Haskell/zip-folds.lhs : It would be interesting to see, constructively, how to build the full-fledged (FR)-list library and avoid general recursion. Let the FR-list be the only driver of iterations. Here is the code: http://ideone.com/NKNeWy I don't post it to hackage, since I'm not sure, these lists are really efficient. There are two main ideas: 1) List is represented as an accumulating right fold, and an accumulator passes from left to right. 2) For efficiency purposes elements skipping is incorporated into the cons (($:) in the code) function. So there is only one isSkipped test (d == 0 in the code) for every element, no matter, how many times drop (or tail or dropWhile) was applied. So there are no redundant (<= 0) checks, as described in the zip-folds paper: The function sdrop shows the major deficiency: we're stuck with the (<=0) test for the rest of the stream. In this case, some delimited continuation operators like `control' can help, in limited circumstances. Here are the pluses: 1) No fixpoints and recursive types. 2) These lists are totally isomorphic to ordinary haskell's, even in a strict monad. 3) No redundant (<= 0) checks. There are some not very big problems with this encoding: 1) Code becomes more complicated than with the ordinary right fold. 2) In fscanl and fsndMapAccumL functions skipped elements are traversed twice, because you need to traverse an accumulator through skipped elements too. And the same holds for the fdropWhile function with additional skipping overhead. 3) Since an accumulator passes from left to right, it's not possible to efficiently fuse fscanr and fmapAccumR functions, so they are defined in the terms of ($:) and ffoldr. And one big problem: zipWith and all other functions, that require parallel lists processing, are quadratic. And the same holds for the ftails function and all other functions, that deconstruct a list by repeatedly applying (fhead . ftail) or collect lists, obtained by repeteadly applied ftail function. However I have never seen any other representation, expressible in pure System F with higher-rank types, which allows such deconstruction in a linear time. If someone knows, give me a link, please. Any suggestions are welcome.

1 0

Re: [Haskell-cafe] Syntax proposal for "reverse apply"/"pipeline apply" (flip ($))
by Doug McIlroy 20 Apr '14

20 Apr '14

> '$$' should be strongly considered That breaks time-honored practice in hugs and ghci. Doug McIlroy

3 2