September 2009 - Haskell-Cafe

An encoding of parametricity in Agda
by Eugene Kirpichov 24 Sep '09

24 Sep '09

Hi, I've discovered the following post on the internets: http://gelisam.blogspot.com/2009/09/samuels-really-straightforward-proof-of… It contains an ingenious (to my mind) encoding of parametricity in Agda that makes the Free Theorems a trivial consequence. I've tried to formalize it in Coq, but, as of now, with no success. -- Eugene Kirpichov Web IR developer, market.yandex.ru

4 4

Bug in writeArray?
by Grzegorz Chrupała 24 Sep '09

24 Sep '09

Hi all, This seems like a bug in the implementation of writeArray: when passed an out-of-range index it silently writes to an incorrect index in the array. -- import Data.Array.IO import Data.Array.Unboxed main = do let (l,u) = ((0,10),(20,20)) marr <- newArray (l,u) 0 :: IO (IOUArray (Int,Int) Int) let badi = (10,9) print (inRange (l,u) badi) writeArray marr badi 1 arr <- freeze marr :: IO (UArray (Int,Int) Int) print . filter ((/=0) . snd) . assocs $ arr -- Grzegorz

5 5

Status of GHC as a Cross Compiler
by John Van Enk 24 Sep '09

24 Sep '09

Hi, This may be more appropriate for a different list, but I'm having a hard time figuring out whether or not we're getting a cross compiler in 6.12 or not. Can some one point me to the correct place in Traq to find this information? /jve

4 4

gtk2hs and runghc
by Günther Schmidt 24 Sep '09

24 Sep '09

Hi, I'm trying to test some gtk2hs gui code without compiling it first, just by using runghc. Gtk2hs then complains about running in a multithreaded ghc, ie. one with several "real" OS threads. Is it possible to start runghc single-threaded? Günther

4 8

Ctrl-C handling in Haskell with curl on Linux
by Vasyl Pasternak 24 Sep '09

24 Sep '09

Hi, Yesterday I tried to implement simple tool to download pages, and wanted catch Ctrl-C (and other 'killing' messages) from haskell to handle state saving. Without curl (when I perform some long operation) haskell throws UserInterrupt exception immediately, but if I put long operation, which downloads page from the WEB (from the far-far-away server :) ) than I noticed following issues: - to break my program I have to press Ctrl-C twice - haskell doesn't throw an exception - when I rewrite this code to use signals, haskell, after I press Ctrl-C several times exits with error "too many pending signals" I put the test code in the end of the letter. Shortly the longTask doesn't handle Ctrl-C and longTask' handles it. I couldn't find any solutions to this problem, I am afraid that this problem could occur in other non-native haskell modules (bindings to C libraries) Many thanks in advance, Vasyl pasternak ------------------------------------------ Test code: module Main where import Prelude hiding (catch) import Network.Curl import Control.Exception import Control.Monad import System.IO errorHandler defVal e = do putStrLn $ "Error: " ++ (show (e :: ErrorCall)) return defVal link = "far-far-away-site.com.net" getSite curl l = do r <- do_curl_ curl l method_GET :: IO (CurlResponse) if respCurlCode r /= CurlOK then error "get page failed" else return $ respBody r -- this long task doesn't throw user interrupts longTask = do putStrLn "Long task started" curl <- initialize setopts curl [CurlCookieJar "cookies"] handle (errorHandler ()) $ mapM_ (\_ -> getSite curl link >> return ()) [0..100] return () -- this trows longTask' = do putStrLn "long task started" let fib n = foldr (*) 1 [1..n] h <- openFile "/dev/null" WriteMode -- never ends mapM_ (hPutStr h . show . fib) [1..] return () onAbort e = do let x = show (e :: AsyncException) putStrLn $ "Aborted: " ++ x return () main :: IO () main = do handle onAbort longTask putStrLn "Exiting"

3 4

ANNOUNCE: graphviz-2999.5.1.1
by Ivan Lazar Miljenovic 24 Sep '09

24 Sep '09

I'm pleased to announce version 2999.5.1.1 [1] of the graphviz library, which provides bindings to the GraphViz [2] suite of tools for drawing graphs. [1] http://hackage.haskell.org/package/graphviz-2999.5.1.1 [2] http://www.graphviz.org/ This is yet another bugfix release, fixing the problem spotted by Kathleen Fisher where Dot keywords need to be explicitly quoted if used as labels, etc. Once again, this is done automagically with no change to the API. -- Ivan Lazar Miljenovic Ivan.Miljenovic(a)gmail.com IvanMiljenovic.wordpress.com

1 0

[uvector] derive UA instance for newtype
by Khudyakov Alexey 24 Sep '09

24 Sep '09

Hello I want to derive UA instance for newtypes. Say I have following newtype: > newtype Foo a = Foo a Attempts to derive UA automatically fails because it have associated types. It seems that it's not possible to define from outside. However it's possible to derive instance in the uvector's internals. > deriving instance UAE a => UAE (Foo a) > deriving instance UPrim a => UPrim (Foo a) > > instance UPrim a => UA (Foo a) where > newtype UArr (Foo a) = UAFoo (BUArr (Foo a)) > newtype MUArr (Foo a) s = MUFoo (MBUArr s (Foo a)) > > lengthU = primLengthU > indexU = primIndexU > sliceU = primSliceU > lengthMU = primLengthMU > newMU = primNewMU > readMU = primReadMU > writeMU = primWriteMU > copyMU = primCopyMU > unsafeFreezeMU = primUnsafeFreezeMU > > memcpyMU = primMemcpyMU > memcpyOffMU = primMemcpyOffMU > memmoveOffMU = primMemmoveOffMU Am I moving in right direction and are there other/better solutions to problem? -- Khudyakov Alexey

3 2

best practices for monad transformers and exceptions?
by Evan Laforge 24 Sep '09

24 Sep '09

I don't have an especially deep understanding of monad transformers... basically, I asked around (thanks #haskell!) and struggled and with helpful advice came up with something that seemed to work for me, only now I'm having trouble. What I'm trying to do is hardly revolutionary or even advanced, so I'd like to come up with some kind of standard best practices way to do this. Basically, the question is, "how do you combine multiple sets of monad transformers, each of which may contain their own exception mechanism, and throw and catch exceptions for each one?" This seems like a perfectly normal thing to want to do, so someone must have figured this out. For more details: I have several subsystems, each of which runs in its own monad. For instance, the Ui level has a monad that provides state, logging of updates, and an exception (LoggerT is simply a type alias for a WriterT with DList): module Ui.State where type StateStack m = State.StateT State (Logger.LoggerT Update.Update (Error.ErrorT StateError m)) newtype StateT m a = StateT (StateStack m a) deriving (Functor, Monad, Trans.MonadIO, Error.MonadError StateError) I wished to wrap up this Ui.StateT as its own transformer that can be reused inside other transformers.... because another subsystem has its own set of needs, including its own logging, state, and exception, in addition to Ui.StateT: module Cmd.Cmd where type CmdStack m = Ui.State.StateT (Monad.State.StateT State (Error.ErrorT Abort (Log.LogT m))) newtype CmdT m a = CmdT (CmdStack m a) deriving (Functor, Monad, Trans.MonadIO) I want to provide unlifted access to Ui.State operations, so I have: class (Monad m, Functor m) => UiStateMonad m where get :: m State put :: State -> m () modify :: (State -> State) -> m () update :: Update.Update -> m () throw :: String -> m a Along with instances for Ui.StateT and Cmd.CmdT where Ui.StateT does the appropriate lifting and CmdT just "delegates": 'put st = CmdT (State.put st)' etc. So now I can throw a StateError within Ui.State by calling State.throw. I can throw the error within Cmd with: abort = (CmdT . lift . lift . lift) (Error.throwError Abort) So this all works fine and each "level" of transformers can provide access to its throw by exporting a 'throw' with the appropriate lifts already applied. But then I want to catch an exception, and there's a problem because catchException takes a monadic argument in addition to having a monadic result. I can easily lift the result into the proper layer, but I can't so easily "lower" the result of the monadic argument back down to the StateT or CmdT or whatever. I can use fancy newtype deriving to put Ui.StateT into MonadError and get catchError for free, but if I try to do the same with CmdT then I get: Couldn't match expected type `Abort' against inferred type `State.StateError' When using functional dependencies to combine Error.MonadError State.StateError (State.StateT m), arising from the dependency `m -> e' So evidentally they are not co-existing. In any case, I would need different names for the throw and catch in the different monads. So all I want to do is to be able to catch an Abort. The closest I can think of is something vaguely like: catch_abort :: (Monad m) => CmdT m a -> CmdT m (Maybe a) catch_abort cmd = do ustate <- State.get cstate <- get_state (cmd_result, logs) <- run_cmd ustate cstate mapM_ Log.record logs case cmd_result of Left Abort -> return Nothing Right ui_result -> case ui_result of Left ui_err -> State.throw ui_err Right (val, ustate, updates) -> do State.put ustate mapM_ State.update updates return (Just val) All of this manual "unlifting" is quite ugly and clearly doesn't scale... and it's quite easy to introduce a subtle bug by forgetting to merge one of the many results that come out. So... what should I have done? What do other people who have a project with >1 kind of exception do? Is there an easier way to write 'catch_abort'? Should I break modularity by making a global exception type and putting them all in StateError? Then (I think) I could use catchError directly since I'd only have one occurrance of ErrorT in any transformer stack.

1 0

Re: [Haskell-cafe] accessible layout proposal?
by Jimmy Hartzell 24 Sep '09

24 Sep '09

> I am a bit puzzled here. > This seems to mean something like > "If you take readable code using an operator you can > make it less readable, and when you do that you create > another problem as well, and an even less readable hack > can fix that." > > "I know an old lady who swallowed a fly..." This proposal notwithstanding, I find 'concat [a, b, c]' much more readable than (a) ++ (b) ++ (c), especially since the parentheses in the latter might be necessary (or become necessary unexpectedly upon editing), whereas in the former order of operations is never a concern. Furthermore, why repeat the (++) over and over again, when really you are making a *list* of things to add, and should be using a list?

2 1

help with cabal; trying to escape from configuration hell
by S. Doaitse Swierstra 23 Sep '09

23 Sep '09

I am trying to run happstack on my Mac, but unfortunately I am getting error messages as described in: http://code.google.com/p/happstack/issues/detail?id=88 The cure seems to be to downgrade to network-2.2.0.1, but unfortunately my installed cabal depends on network-2.2.1.4. I tried to re-install happstack using: cabal install happstack --reinstall --constraint="network==2.2.0.2" but unfortunately the ghc happily reports to link against network-2.2.1.4: ... Loading package parsec-2.1.0.1 ... linking ... done. Loading package hsemail-1.3 ... linking ... done. Loading package network-2.2.1.4 ... linking ... done. Loading package SMTPClient-1.0.1 ... linking ... done. Loading package time-1.1.4 ... linking ... done. ... Can someone rescue me? Doaitse

3 2