April 2011 - Haskell-Cafe

Syntastic support for cabal
by Luke Randall 13 Apr '11

13 Apr '11

Hi all I'm using Haskell mode for Vim, as well as syntastic. I've found syntastic particularly useful as I'm quite new to Haskell and sometimes get the types of my functions wrong. When working on Cabal packages however, syntastic marks import statements which import modules local to the cabal package as being incorrect because the module cannot be found. Has anyone solved this, or do you have pointers on how I can? Thanks -- Luke Randall Sent with Sparrow

1 1

Debugging with gdb?
by Svante Signell 13 Apr '11

13 Apr '11

Hi, As I don't know anything about Haskell, can I make a stupid question: Is there any method to create debug symbols for a Haskell program, and is it possible to debug with gdb? Thanks!

4 4

RFC: Extendable and Easy-To-Use Logger (HLogger)
by Jon Kristensen 13 Apr '11

13 Apr '11

Hello Haskellers! I have developed a very simple logging library to be extendable and easy-to-use. You can see the current release at http://hackage.haskell.org/package/hlogger/. The SimpleHLogger module is an example of what a HLogger implementation could look like. It could be used like this: import System.Log.SimpleHLogger main = do logger<- simpleLogger "test" loggerLog logger (Just ("LoggerTest", "main")) Debug "Test" loggerStop logger As I have now implemented everything that I could think of, I would like to ask the Haskell community to give some feedback on HLogger so that I can continuing improving it. Some questions: 1. Do you have any general/architectural thoughts on things I might be doing wrong? 2. Is the ease-of-use goal contradicting making this library useful in more complex applications? 3. Do I want to redesign HLogger to allow loggers to have a mutable state, and if so, how can I do that? 4. Is there a nice way to provide the module and the function than using (Just ("LoggerTest, "main)) (shortening it somehow, I guess)? Thanks! Warm regards, Jon Kristensen

5 5

Using DPH
by Wilfried Kirschenmann 13 Apr '11

13 Apr '11

Hi, In order to do a performance comparison beetween different approaches for our application, I make different implementation of a simple example (computing the norm of a vector expression. I rely on Repa to do this. However, when I tried to build the parallel version (-threaded -fvectorise -rtsopts), I got an error specifying that dph-par was not available. Indeed, It wasn't. As explained on the dph webpage, I installed a developpment version of ghc (ghc-7.1.20110331). However, when I try to build my application, I get the folowing error: user error (Pattern match failure in do expression at compiler/vectorise/Vectorise/Monad.hs:57:10-20) My source code is attached and compiled with the following line: ghc haskell.hs -O3 -XBangPatterns -msse2 -fforce-recomp -rtsopts -threaded -fvectorise So my question is : is the problem concerning the pattern matching known ? If yes, is there a workaround ? is there an other version of dph that would work with Repa ? Thank you ! ----- Wilfried Kirschenmann "An expert is a person who has made all the mistakes that can be made in a very narrow field."* *Niels Bohr - *Danish physicist (1885 - 1962)*

2 4

ANNOUNCE: TCache 0.8
by Alberto G. Corona 13 Apr '11

13 Apr '11

Tired of your persistence layer? Take persistence the haskelll way. Besides the backward compatible stuff, TCache now defines persistent STM variables (DBRef's) that leverages the traditional haskell reference syntax to perform fast in-memory database transactions and inter-object references. It includes indexing, triggers and an straightforward pure-haskell query language based on record field relations. http://hackage.haskell.org/package/TCache-0.8.0.2 This version includes a lot of changes and must be considered beta. Please submit any bugs, comments and so on. Cheers!

1 0

Request for library recommendations
by Erik de Castro Lopo 13 Apr '11

13 Apr '11

Hi all, I'm currently scoping out a project for work that I'd like to do in Haskell. Basically its a HTTP proxy (ie like the Squid proxy) but with a few wrinkles: - Since the connecting clients are untrusted, the incoming connection is HTTP, the header is modified by the proxy to add BasicAuth (credentials known only to the proxy, not the client) and then forwared upstream with responses pass back to the client. - Needs to work via upstream Squid proxy (so this proxy needs to be able to requirest a CONNECT from the upstream Squid proxy). - In some cases, the client will POST via the proxy and immediately exit and the proxy will be required to re-try the POST if it fails. - Other manipulation of headers and data. I am aware that Squid can satisfy the first two requirements above but not all the others (without actually hacking the Squid code base). I have done stuff like this before in C++ and Ocaml, so the mechanics are not a mystery. However I would like as much as possible to make use of existing Haskell libraries to cut down the development time. Performance is also not critical, so as long as the Haskell version can beat a hypothetical Python implementation I'd be happy. Thanks in advance. Cheers, Erik -- ---------------------------------------------------------------------- Erik de Castro Lopo http://www.mega-nerd.com/

3 2

syntactic sugar for heterogeneous lists
by Alberto G. Corona 12 Apr '11

12 Apr '11

Hi Cafe: http://hackage.haskell.org/trac/ghc/ticket/1245 I also want some kind of syntactic sugar for H. Lists. The absence of syntactic sugar makes heterogeneous list to look scary. People in haskell is accostumed to syntactic sugaring, so people think of not sugared expressions as second class. I´ don't know the details, but it seems that tuples are in the language to cover the heterogeneous flexibility thing that homogeneous lists may not provide, but they introduce its own inflexibilities; These repeating instances for two, three, four and so on tuples makes them artificial and repetitive. In its comparison, a complex but unique H. List instance look elegant. specially if it is sugarized. An alternative to sugarize H.Lists preserving tuples could be to use {} to sugarize H. lists as {x,y,z} and desugarize it into: x :*: y :*: z :*: {} Just like [] means empty list, {} would mean HNil, the empty heterogeneous list. But this alternative , if implemented, would soon render tuples obsolete. These (,,) (,,,) constructors may create marginally faster and compact structures, but they are much less manageable.

2 2

redefining a type class for one instance?
by Edward Amsden 12 Apr '11

12 Apr '11

I am trying as an experiment to write something as follows: data Wrapper a b where Pure :: (a -> b) -> Wrapper a b Nullable :: (Maybe a -> b) -> Wrapper a b class Liftable a b where liftTo :: (b -> c) -> a b c instance Liftable Wrapper a where liftTo = Pure instance Liftable Wrapper (Maybe a) where liftTo = Nullable Obviously this code (with suitable extensions enabled) complains of overlapping type instances, but is there a way to accomplish redefining a class instance for one particular group of types while having it defined for all others? (Maybe a vs. a in this instance) -- Edward Amsden Student Computer Science Rochester Institute of Technology www.edwardamsden.com

2 2

Current heap size and other runtime statistics -- API for accessing in GHC?
by Ryan Newton 12 Apr '11

12 Apr '11

Hi cafe, The rtsopts (-s etc) can provide some nice debugging information regarding memory management. And System.Mem.performGC can initiate garbage collection. But are there APIs for querying the current state of the heap? I've googled and come up dry. In this case I'm running benchmarks and for the sake of fair comparison I want to make sure that everything from a previous run is cleaned up before the next run. Thanks, -Ryan

4 3

ghc6: Re: Update: ... compiling ghc6
by Svante Signell 12 Apr '11

12 Apr '11

Addressing both Haskell and Hurd people here. Any hints by anyone? On Wed, 2011-04-06 at 09:30 +0200, Svante Signell wrote: ... > #5 0x011d3ce0 in __libc_read (fd=DWARF-2 expression error: DW_OP_reg > operations must be used either alone or in conjuction with DW_OP_piece > or DW_OP_bit_piece. > ) at ../sysdeps/mach/hurd/read.c:27 > #6 0x084919c8 in s9qJ_ret () > #7 0x0861f842 in StgRun () > #8 0x087c44e0 in ?? () Looking into this further, this looks like an error. What does it mean? from eglibc-2.11.2/sysdeps/mach/hurd/read.c /* Read NBYTES into BUF from FD. Return the number read or -1. */ ssize_t __libc_read (int fd, void *buf, size_t nbytes) { error_t err = HURD_FD_USE (fd, _hurd_fd_read (descriptor, buf, &nbytes, -1)); return err ? __hurd_dfail (fd, err) : nbytes; } The call is from the haskell binary cabal-bin but I don't have any debug info from that executable. Anybody fluent in Haskell? (How to get debug info in Haskell with gdb?) Below is the (stripped-down) call and the source code cabal-bin.hs: cd libraries/random: ../cabal-bin /usr/bin/ghc6 ../bootstrapping.conf configure --verbose=3 --with-compiler=../../ghc/stage1-inplace/ghc --with-hc-pkg=../../utils/ghc-pkg/install-inplace/bin/ghc-pkg module Main (main) where import Control.Monad import Data.Maybe import Distribution.PackageDescription import Distribution.PackageDescription.Parse import Distribution.Simple import Distribution.Simple.Utils import Distribution.Verbosity import System.Directory import System.Environment import System.FilePath import qualified Distribution.Make as Make import qualified Distribution.Simple as Simple setupProg :: FilePath setupProg = "./Setup" main :: IO () main = do unprocessedArgs <- getArgs let verbosity = verbose case unprocessedArgs of ghc : packageConf : args -> doit verbosity ghc packageConf args _ -> die "Bad args" doit :: Verbosity -> FilePath -> FilePath -> [String] -> IO () doit verbosity ghc packageConf args = do exists <- doesFileExist setupProg if exists then rawSystemExit verbosity setupProg args else do gpdFile <- defaultPackageDesc verbosity gpd <- readPackageDescription verbosity gpdFile let pd = packageDescription gpd case buildType pd of Just Simple -> Simple.defaultMainArgs args Just Make -> Make.defaultMainArgs args Just Configure -> defaultMainWithHooksArgs autoconfUserHooks args _ | packageName pd == PackageName "Cabal" -> -- Cabal is special...*sigh* Simple.defaultMainArgs args | otherwise -> runSetup verbosity ghc packageConf args runSetup :: Verbosity -> FilePath -> FilePath -> [String] -> IO () runSetup verbosity ghc packageConf args = do -- Don't bother building Setup if we are cleaning. If we need to -- build Setup in order to build, and Setup isn't built already, -- then there shouldn't be anything to clean anyway. unless cleaning $ rawSystemExit verbosity ghc ["-package-conf", packageConf, "--make", "Setup", "-o", "Setup"] rawSystemExit verbosity "./Setup" args where cleaning = case args of "clean" : _ -> True _ -> False

1 1