February 2012 - Haskell-Cafe

Compile one static Haskell binary for all x86_64 Linux distributions, is it possible?
by Morten Olsen lysgaard 10 Feb '12

10 Feb '12

Hi, I've written a program that converts an XML file from a database into another XML compatible with another database. Quite standard. I'm using the mtl library for State and HaXml for XML parsing. Is it possible to compile this program into a single static binary that works on all Linux computers that run x86_64. My development box is an Arch Linux x86_64 running GHC 7.0.3. I tried with: ghc --make -static -rtsopts=all -optl-pthread -optl-static Program.hs When I copy the binary created over to my x86_64 Ubuntu 11.10 desktop and run it I get: Program: mkTextEncoding: invalid argument (Invalid argument) I've googled around and found out this has to do with iconv, but I've never compiled anything to run on another pc than the development machine, so I've got no experience in fixing these kinds of problems. Any help appreciated. -- Morten

3 2

Error in installing dph-examples on Mac OS X 10.7.3
by mukesh tiwari 10 Feb '12

10 Feb '12

Hello all I am trying to install dph-examples on Mac OS X version 10.7.3 but getting this error <http://hpaste.org/57699>. I am using ghc-7.4.1. Error running clang! you need clang installed to use the LLVM backend cabal: Error: some packages failed to install: dph-examples-0.6.1.3 failed during the building phase. The exception was: ExitFailure 1 I have already installed Xcode 4.2 and it shows Macintosh-0026bb610428:tmp mukesh$ clang clang clang++ Macintosh-0026bb610428:tmp mukesh$ clang -v Apple clang version 3.0 (tags/Apple/clang-211.10.1) (based on LLVM 3.0svn) Target: x86_64-apple-darwin11.3.0 Thread model: posix Macintosh-0026bb610428:tmp mukesh$ clang++ -v Apple clang version 3.0 (tags/Apple/clang-211.10.1) (based on LLVM 3.0svn) Target: x86_64-apple-darwin11.3.0 Thread model: posix Macintosh-0026bb610428:tmp mukesh$ l Display all 113 possibilities? (y or n) Macintosh-0026bb610428:tmp mukesh$ llvm- llvm-ar llvm-config llvm-diff llvm-extract llvm-gcc llvm-link llvm-objdump llvm-rtdyld llvm-tblgen llvm-as llvm-cov llvm-dis llvm-g++ llvm-gcc-4.2 llvm-mc llvm-prof llvm-size llvm-bcanalyzer llvm-cpp-4.2 llvm-dwarfdump llvm-g++-4.2 llvm-ld llvm-nm llvm-ranlib llvm-stub Macintosh-0026bb610428:tmp mukesh$ llvm-g++ -v Using built-in specs. Target: i686-apple-darwin11 Configured with: /private/var/tmp/llvmgcc42/llvmgcc42-2336.1~1/src/configure --disable-checking --enable-werror --prefix=/Developer/usr/llvm-gcc-4.2 --mandir=/share/man --enable-languages=c,objc,c++,obj-c++ --program-prefix=llvm- --program-transform-name=/^[cg][^.-]*$/s/$/-4.2/ --with-slibdir=/usr/lib --build=i686-apple-darwin11 --enable-llvm=/private/var/tmp/llvmgcc42/llvmgcc42-2336.1~1/dst-llvmCore/Developer/usr/local --program-prefix=i686-apple-darwin11- --host=x86_64-apple-darwin11 --target=i686-apple-darwin11 --with-gxx-include-dir=/usr/include/c++/4.2.1 Thread model: posix gcc version 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.1.00) Regards Mukesh Tiwari

4 8

Mersenne-random and standard random API
by Yves Parès 10 Feb '12

10 Feb '12

Hi, I've been in the past told that mersenne-random was much better than the standard random package. However, System.Random.Mersenne doesn't follow the general API described in System.Random, MTGen is not an instance of RandomGen. But a sample on System.Random.Mersenne.getStdRandom documentation ( http://hackage.haskell.org/packages/archive/mersenne-random/1.0.0.1/doc/htm…) indicates this: rollDice :: IO Int rollDice = getMTRandom (randomR (1,6)) It looks like wrong documentation (checking of older versions show that getMTRandom have never lived or been exposed by this module), but suggests MTRandom should be an instance of RandomGen (if I assume well that randomR is that of System.Random). So is it possible to use the fast and efficient mersenne generator with the convenient and general random API?

5 11

Re: [Haskell-cafe] [Haskell] Support tools for development with Cabal and Darcs
by Duncan Coutts 10 Feb '12

10 Feb '12

On 10 February 2012 12:06, Henning Thielemann <lemming(a)henning-thielemann.de> wrote: > > I have uploaded three packages to Hackage that shall simplify maintaining > cabal packages under darcs revision control. They simplify tasks like > uploading packages after some tests, cabal version handling, compiling > multiple local packages in the right order, replacing identifiers in > multiple files, renaming modules. > > * http://hackage.haskell.org/package/darcs-scripts > * http://hackage.haskell.org/package/cabal-scripts > * http://hackage.haskell.org/package/cabal-sort Hi Henning, Would you mind giving me a breif explanation of the "compiling multiple local packages in the right order" feature? I'd like to understand what is missing in cabal-install in this respect. Currently with cabal-install you can say: cabal install ./a ./b ./c and it will install the packages in those directories in the right order. (It is also possible to list other install targets including local and remote tarballs, and of course packages from hackage). Duncan

2 4

Fancy REPL
by Heinrich Apfelmus 10 Feb '12

10 Feb '12

Dear list, the Show class is extremely useful for exploring Haskell in a terminal, but sometimes, I just want something fancier. For instance, I'm currently dabbling with sound generation and it is only natural that I want to hear the sound instead of seeing a textual representation. Another example would be graphics, that are simply drawn on screen whenever you evaluate their value. Of course, this is simple to implement with a class class Demonstrable a where demo :: a -> IO () instance Demonstrable Sound where demo = play instance Demonstrable Sound where demo = draw instance Demonstrable GUI where demo = run etc. However, I don't want to reinvent the wheel, small as it may be, hence my question: is there a package on hackage that already defines a class similar to Demonstrable ? Or any other projects in this direction, like, say, a fancy REPL built on wxHaskell? Best regards, Heinrich Apfelmus -- http://apfelmus.nfshost.com

3 4

The State of Testing?
by Michael Craig 10 Feb '12

10 Feb '12

I've been picking up Haskell as a side project for the last few months, and I'm now considering publishing some useful code I've written and reused in several small projects. So far I've done relatively little with testing (these have been non-mission-critical applications) but I feel I should get my act together before pushing my work into the common ecosystem. I'm comfortable writing tests in QuickCheck and HUnit and bundling them as optional executables with cabal, but I understand there's a better way. Specifically, I'm looking at the test-framework package and cabal's (newish) test-suite sections. Are these two used together or each to the exclusion of the other? Is there something else I should be using? Cheers, Mike Craig

9 15

Re: [Haskell-cafe] [Haskell] ANNOUNCE: hs-json-rpc 0.0.0.1
by Felipe Almeida Lessa 09 Feb '12

09 Feb '12

[Redirecting to haskell-cafe] Congrats on your first release! While I haven't looked into your package in more depth, I'd suggest taking a look at http-conduit [1]. While I don't know of any benchmarks, it should be faster or at least as fast the HTTP package. It's also used by many people everyday (myself included), which means that it has no major bugs and that bugs are promptly solved. It also has some features that HTTP doesn't. Cheers! [1] http://hackage.haskell.org/package/http-conduit -- Felipe.

2 1

Yesod and dependencies hell
by Никита Тимофеев 09 Feb '12

09 Feb '12

For three days I can't compile dependencies for my project using yesod, yesod-auth, yesod-persistent, persistent-template, persistent-sqlite, persistent. When I varied version I received a variety of broken dependencies: persistent (0.6.* vs 0.7.*), conduit (0.1.* vs 0.2.*), conduit-pool, path-pieces, attoparsec and something else. In this regard, I have a few questions. How to build these packages together? And, Michael, why so bad? -- Тимофеев Н.Д.

4 4

Re: [Haskell-cafe] Is ListT a valid MonadPlus?
by oleg＠okmij.org 09 Feb '12

09 Feb '12

First of all, ListT is not a monad transformer, since it breaks the law of associativity of bind: *Control.Monad.List> let one = (lift $ putStrLn "1") :: ListT IO () *Control.Monad.List> let two = (lift $ putStrLn "2") :: ListT IO () *Control.Monad.List> let choice = return 1 `mplus` return 2 :: ListT IO Int *Extensible Control.Monad.List> runListT $ choice >> (one >> two) 1 2 1 2 [(),()] *Extensible Control.Monad.List> runListT $ (choice >> one) >> two 1 1 2 2 [(),()] > It appears to me that the MonadPlus instance for ListT breaks the > following MonadPlus law > m >> mzero = mzero The precise set of MonadPlus laws isn't actually agreed upon. The above law is the least agreed upon. Incidentally, this law is certainly violated for _any_ back-tracking monad transformer. Consider (lift <any action in base monad>) >> mzero for example (lift (putStrLn "printed")) >> mzero Is it reasonable to expect that printing should be undone? That the paper should be fed back into the printer and the toner lifted? There are back-tracking monad transformers; for example, LogicT.

1 0

Is ListT a valid MonadPlus?
by Richard Wallace 09 Feb '12

09 Feb '12

It appears to me that the MonadPlus instance for ListT breaks the following MonadPlus law m >> mzero = mzero I've tried with every implementation of ListT I could find - the one in mtl, transformers, List, even "ListT done right" and it's alternative. They all seem to violate the above law. Following is a ghci session that shows this with the mtl impl. Prelude> import Control.Monad.List Prelude Control.Monad.List> let nt = ListT Nothing :: ListT Maybe Int Prelude Control.Monad.List> let e = mzero :: ListT Maybe Int Prelude Control.Monad.List> runListT nt Nothing Prelude Control.Monad.List> runListT e Just [] Prelude Control.Monad.List> runListT $ nt >> e Nothing According to the MonadPlus law above, that last line should evaluate to `Just []`. So is ListT a valid MonadPlus? Or am I missing something? If not, can a valid MonadPlus instance for ListT be written? I've tried a few things and come up short. Thanks, Rich

1 0