October 2009 - Haskell-Cafe

difference "cabal configure" and "runghc Setup.lhs configure"
by Andrew U. Frank 08 Oct '09

08 Oct '09

i have a strange error, which does not occur when i run runghc Setup.lhs configure but when i use cabal configure and then build, it occurs. the error is Type constructor Control.Exception.Exception used as a class in the instance declaration. (i have imported qualified Control.Exception) two questions: 1. Exception is a class (in Control.Exception) - why the error? 2. why is there a difference between runghc and cabal (using the same cabal file). i attach the cabal file and the two configuratin files produced.

2 1

Generalizing IO
by Floptical Logic 08 Oct '09

08 Oct '09

The code below is a little interactive program that uses some state. It uses StateT with IO to keep state. My question is: what is the best way to generalize this program to work with any IO-like monad/medium? For example, I would like the program to function as it does now using stdin but I would also like it to function over IRC using the Net monad from <http://haskell.org/haskellwiki/Roll_your_own_IRC_bot>. Thanks for any suggestions. -- begin code -- import Control.Monad import Control.Monad.State import Data.List data PD = PD { pdCount :: Int , pdList :: [String] } deriving (Show) type PDState = StateT PD IO main = runStateT loop (PD { pdCount = 0, pdList = [] }) loop :: PDState a loop = forever $ do cmd <- liftIO getLine runCmd cmd runCmd :: String -> PDState () runCmd "Inc" = increment runCmd "PrintCount" = liftIO . print =<< getCount runCmd "PrintList" = liftIO . print =<< getList runCmd str | "Add " `isPrefixOf` str = addToList $ drop 4 str runCmd _ = return () getCount :: PDState Int getCount = pdCount `liftM` get getList :: PDState [String] getList = pdList `liftM` get increment :: PDState () increment = modify $ \st -> st { pdCount = pdCount st + 1 } addToList :: String -> PDState () addToList str = modify $ \st -> st { pdList = pdList st ++ [str]} -- end code --

7 12

Let it be
by michael rice 08 Oct '09

08 Oct '09

>From Learn You a Haskell ("Let it be" section): 1. cylinder :: (RealFloat a) => a -> a -> a 2. cylinder r h = 3. let sideArea = 2 * pi * r * h 4. topArea = pi * r ^2 5. in sideArea + 2 * topArea =================== What's the proper indentation for LET so these problems (below) don't arise? I thought LET and IN should be aligned in the same column. Also, isn't a LET expression an "expression." Michael ============== This works: import System.Random main = do gen <- getStdGen let (randNumber, newGen) = randomR (1,6) gen :: (Int, StdGen) in putStrLn $ "Number is " ++ show randNumber ============== This works: import System.Random main = do gen <- getStdGen let (randNumber, newGen) = randomR (1,6) gen :: (Int, StdGen) putStrLn $ "Number is " ++ show randNumber ============== This doesn't: import System.Random main = do gen <- getStdGen let (randNumber, newGen) = randomR (1,6) gen :: (Int, StdGen) in putStrLn $ "Number is " ++ show randNumber [michael@localhost ~]$ runhaskell zz.hs zz.hs:4:2: The last statement in a 'do' construct must be an expression

4 6

Test.QuickCheck: generate
by Michael Mossey 08 Oct '09

08 Oct '09

In Test.QuickCheck, the type of 'generate' is generate :: Int -> StdGen -> Gen a -> a I can't find docs that explain what the Int does. Some docs are here: <http://www.haskell.org/ghc/docs/latest/html/libraries/QuickCheck/Test-Quick…>

3 2

On DSLs - one last time
by Günther Schmidt 08 Oct '09

08 Oct '09

Hi all, I'd like to summarize my impressions on the DSL issue, gathered from the responses to my posts. It is absolutely amazing, various people have shown here that it is possible to design languages within haskell that are almost or just as powerful and safe as haskell itself. It is probably even possible to create an EDSL within an EDSL within haskell. Amazing! Uhm... ... but in that case I really wonder why I should not use Haskell directly then. Because out of the EDSL I can clearly see the "E" and I can see the "L" but having slight problems figuring out where the "D" and the "S" come in. In short, yes it's great that you can create a language within Haskell so powerful that it's gone full circle and is general purpose again, but what good is that? The efforts I've seen here where efforts to create a rich and powerful syntax, but aren't we supposed to focus on expressing semantics specific to a problem in a DSL rather? I'm not trying to put the efforts and achievements of the people who have done such amazing work down, but what exactly is it that I missed? Günther

2 1

random question
by Michael Mossey 08 Oct '09

08 Oct '09

My thread about randomness got hijacked so I need to restate my remaining question here. Is it acceptable to write pure routines that use but do not return generators, and then call several of them from an IO monad with a generator obtained by several calls to newStdGen? shuffle :: RandomGen g => g -> [a] -> [a] shuffle = ... foo :: [a] -> [a] -> IO () foo xs ys = do g1 <- newStdGen print $ shuffle g1 xs g2 <- newStdGen print $ shuffle g2 ys Does this kind of thing exhibit good pseudorandomness? Thanks, Mike

6 6

#include path
by Sean McLaughlin 08 Oct '09

08 Oct '09

Hi, I'm trying to compile some code using Cabal. One of the files has a CPP directive #include "undefined.h" The file undefined.h is in the same directory as the file with the directive. If I use the full path name, cabal can compile it. However, if I use the relative path, it complains about not being able to find the file. I tried adding it to extra-source-files, but that didn't help. I see that Hackage packages like Agda use this trick, so I know it's possible. How can I get these paths worked out correctly? There isn't much comment in the Cabal docs. Thanks, Sean

2 1

Cabal, sdist and flags
by Magnus Therning 08 Oct '09

08 Oct '09

A while ago I moved a project away from using a mix of Cabal and make to build (Cabal to build the library and make to build the tests) to using only Cabal. I added a flag and then made a construct like this: flag BuildTests Description: Build unit and quickcheck tests. Default: False ... executable tests if flag(BuildTests) main-is: Test.hs build-depends: test-framework, test-framework-hunit, HUnit, test-framework-quickcheck2, QuickCheck >= 2.1.0.0 hs-source-dirs: test-src src else main-is: Test.hs buildable: False Interestingly that does work for things like 'configure' and 'build', but it doesn't work for 'sdist': % ./Setup.hs sdist Distribution quality errors: Building source dist for dataenc-0.13.0.1... Setup.hs: Test.hsTest.hs doesn't exist To make this work I had to move the 'main-is' outside of the if-else: executable tests main-is: Test.hs hs-source-dirs: test-src src if flag(BuildTests) build-depends: test-framework, test-framework-hunit, HUnit, test-framework-quickcheck2, QuickCheck >= 2.1.0.0 else buildable: False I found this surprising, so I'm tempted to call it a bug. A quick search didn't tell me whether this is well-known already, hence the email. Would you consider it a bug? Has it been raised already? /M -- Magnus Therning (OpenPGP: 0xAB4DFBA4) magnus＠therning．org Jabber: magnus＠therning．org http://therning.org/magnus identi.ca|twitter: magthe

3 2

What *is* a DSL?
by oleg＠okmij.org 08 Oct '09

08 Oct '09

Perhaps it would be appropriate to point out the IFIP conference on exactly that topic, DSL. The conference took place in July, here is the permanent record: http://dsl09.blogspot.com/ with pointers to the slides and the discussions. The panel discussion has debated that very question, what exactly is DSL. http://dsl09.blogspot.com/2009/07/panel.html As you can see, the agreement could not be reached on what is Domain. There will be another DSL conference, in Australia. Perhaps all the Haskell-Cafe debaters may want to participate.

1 0

ANNOUNCE: vty-ui 0.2
by Jonathan Daugherty 07 Oct '09

07 Oct '09

I'm happy to announce the release of vty-ui 0.2. Get it from Hackage: http://hackage.haskell.org/package/vty-ui Or get the source with darcs: http://repos.codevine.org/vty-ui This version of vty-ui is everything you hoped it would be: fewer bugs, more widget types, cleaner code, better abstractions, and more generality. If you've written any applications to use vty-ui, only a few of your type signatures will need to change. Here's a summary of changes: General * Unify HBorder and VBorder types (now called Border) * Unify HFill and VFill types (now called Fill) * Unify HBox and VBox types (now called Box) * Box layout: layout documentation, edge case bug fixes * Widget class: refactor growthPolicy into growHorizontal and growVertical * Add Graphics.Vty.Widgets.Composed for high-level widget compositions * Add Graphics.Vty.Widgets.All convenience module to export everything in the library Lists * List items may now be of any widget type! * Add resize, pageUp, pageDown, scrollBy functions and expose other list functionality * Generalize scrolling API * Add convenience API for using lists of Text widgets * Support internal identifiers for list items * Always consume scrollWindowSize lines of space even if the list is smaller New widgets * WrappedText: automatically wraps text in available space; try resizing the demo. * Borders module: one-row/one-column ASCII borders between boxed widgets; ASCII art box container * Demo: use crazy boxed list items to show off border boxes and generic widget support in Lists! * Demo: support PageUp/PageDown to demonstrate List paging API Enjoy! -- Jonathan Daugherty

1 0