- Haskell-Cafe - Haskell.org

Can't compile hmake
by Thomas Davie 29 Jul '03

29 Jul '03

Hi, I'm just trying to get started with haskell and have been having some problems getting hmake to compile. For some reason it doesn't detect libreadline. Lib readline is installed as evidenced by this: [computer:~/Desktop/hmake-3.08] username% fink list | grep readline i readline 4.3-5 Comfortable terminal input library i readline-shlibs 4.3-5 Comfortable terminal input library I've been attempting to force it to link against the library by using ./configure --buildopts="-L/sw/lib/ -lreadline -I/sw/include/ -include readline/readline.h" but with no success, I'm getting the error "gcc: readline/readline.h: No such file or directory", despite the file /sw/includes/readline/readline.h very definitely existing. Has anyone got any ideas for where I could go from here? FYI, I'm running Mac OS X.2.6 Thanks Thomas Davie

2 2

lazy Printing question
by AJ 26 Jul '03

26 Jul '03

Hi everyone, I have written the following program to find "magic numbers" i.e. integers 'n' such that both (n+1) and (n/2+1) are perfect squares. -- Program to find magic numbers Import IO main :: IO () main = do print (filter magicP sqs) sqs :: [Int] sqs = [x*x | x <- [1,3..mAX]] magicP :: Int -> Bool magicP x | ((x-1) `mod` /= 0 = False | otherwise = (((x-1) `div` 2) + 1) `elem` sqs) mAX :: Int mAX = 20000 -- End of listing If I try to run the program (compiled using GHC 6), it calculates all members of the list and then prints the whole list in the end. Since Haskell is 'lazy' I was expecting behaviour similar to HUGS where it prints the numbers as it finds them. Does this behaviour have something to do with the monadic IO in Haskell? I am a Haskell newbie and can't even claim having understood the "gentle introduction" properly. Now my question is - How do I reproduce such "lazy printing" behaviour in GHC? Wait... I thought of another one :-), how do I speed up this program? - AJ

5 6

Type design question
by Konrad Hinsen 25 Jul '03

25 Jul '03

I am a Haskell newbie working on my first serious test case, and I would like some feedback from the experts to make sure I am not doing anything stupid ;-) My applications are numerical (one goal of my current tests being to check how much of a performance penalty I will pay for using Haskell as opposed to C/C++) and involve a lot of geometry. Therefore I started by defining some appropriate types, in particular vectors. So here we go: -- Most general: element of a vector space. -- Operations: addition, subtraction, and multiplication by a scalar class Vect v where (<+>) :: Floating a => v a -> v a -> v a (<->) :: Floating a => v a -> v a -> v a (<*>) :: Floating a => a -> v a -> v a -- Vector space with a scalar product -- Adds dot product and norm class Vect v => VectSProd v where dot :: Floating a => v a -> v a -> a norm :: Floating a => v a -> a norm x = sqrt (dot x x) data Floating a => Vector a = Vector !a !a !a deriving (Eq, Show) -- Standard vector in 3D space instance Vect Vector where (<+>) (Vector x1 y1 z1) (Vector x2 y2 z2) = Vector (x1+x2) (y1+y2) (z1+z2) (<->) (Vector x1 y1 z1) (Vector x2 y2 z2) = Vector (x1-x2) (y1-y2) (z1-z2) (<*>) s (Vector x y z) = Vector (s*x) (s*y) (s*z) instance VectSProd Vector where dot (Vector x1 y1 z1) (Vector x2 y2 z2) = x1*x2+y1*y2+z1*z2 cross :: Floating a => Vector a -> Vector a -> Vector a cross (Vector x1 y1 z1) (Vector x2 y2 z2) = Vector (y1*z2-z1*y2) (z1*x2-x1*z2) (x1*y2-y1*x2) I would like to keep the numerical representation as general as possible, in particular to postpone decisions about precision (Float, Double, eventually arbitrary precision arithmetic) as much as possible. Therefore the "Floating a" constraint. Which leads to my first question: doing it the way I did, I find myself having to add a "Floating a" constraint to almost every function specification with a vector argument. I would prefer to specify once and for all that vector elements need to be "Floating", everywhere. Is there a way of doing that? Next, I would like to define some stuff for simulations of atomic systems. There are two kinds of environments for such simulations, infinite (no boundaries) and periodic (topologically a torus surface). The essential difference is the way that distances are calculated. -- General simulation universe properties class Universe u where distanceVector :: Floating a => u -> (Vector a) -> (Vector a) -> (Vector a) distance :: Floating a => u -> (Vector a) -> (Vector a) -> a distance u v1 v2 = norm (distanceVector u v1 v2) -- Infinite universe data InfiniteUniverse = InfiniteUniverse instance Universe InfiniteUniverse where distanceVector u v1 v2 = v2 <-> v1 -- Periodic universe data Floating a => OrthorhombicUniverse a = OrthorhombicUniverse a a a instance Floating a => Universe (OrthorhombicUniverse a) where distanceVector (OrthorhombicUniverse a b c) (Vector x1 y1 z1) (Vector x2 y2 z2) = Vector (fmod (x2-x1) a) (fmod (y2-y1) b) (fmod (z2-z1) c) where fmod x y = x - y*truncate (x/y) And here I run into a problem. The "Floating a" constraint on OrthorhombicUniverse has the same purpose as for vectors, the three arguments are the dimensions of the box that is periodically replicated. The compiler complains about the definition of distanceVector: ERROR "/home/hinsen/haskell/md/Universe.hs":24 - Inferred type is not general enough *** Expression : distanceVector *** Expected type : (Universe (OrthorhombicUniverse a), Floating b) => OrthorhombicUniverse a -> Vector b -> Vector b -> Vector b *** Inferred type : (Universe (OrthorhombicUniverse a), Floating a) => OrthorhombicUniverse a -> Vector a -> Vector a -> Vector a I do, in fact, want "Floating a" and "Floating b" to enforce the same type, but how can I do that? BTW, in case anyone wonders why I used a class for "Universe" instead of simply defining data Universe = InfiniteUniverse | OrthorhombicUniverse a a a the reason is that I want to leave the option of adding differently shaped universes later on. Konrad.

1 0

RE: segfault in haskell!
by Simon Marlow 25 Jul '03

25 Jul '03

> I've managed to get a segfault in haskell! And without even using the > FFI... actually my code uses the FFI, but the changes that > triggered the > segfault don't involve that, they just use Text.Regex. > > The code that triggers the segfault is the function produced by: > > filetype_function :: IO (FilePath -> FileType) > filetype_function = do > binsfile <- def_prefval "binariesfile" "_darcs/prefs/binaries" > bins <- (liftM lines $ readFile binsfile) `catch` > (\e-> if isDoesNotExistError e then return [] > else ioError e) > let rs = map (mkRegex.fixregex) bins > isbin f = or $ map (\r-> matchRegex r f /= Nothing) rs > ftf f = if isbin f then BinaryFile else TextFile > in > return ftf > > and the valgrind report on the segfault is: > > ==8705== Invalid free() / delete / delete[] > ==8705== at 0x4015D6A4: free (vg_clientfuncs.c:185) > ==8705== by 0x40399A34: (within /lib/libc-2.3.1.so) > ==8705== by 0x40399C13: regfree (in /lib/libc-2.3.1.so) > ==8705== by 0x8117683: s67U_entry (in /home/droundy/darcs/darcs) > ==8705== Address 0x1 is not stack'd, malloc'd or free'd > ==8705== > ==8705== Invalid read of size 4 > ==8705== at 0x40399A58: (within /lib/libc-2.3.1.so) > ==8705== by 0x40399C13: regfree (in /lib/libc-2.3.1.so) > ==8705== by 0x8117683: s67U_entry (in /home/droundy/darcs/darcs) > ==8705== Address 0x22 is not stack'd, malloc'd or free'd I've found one possible cause of this: we were calling regfree on the regular expression structure even if regcomp failed. This will be fixed in GHC 6.0.1. Workaround: don't pass any bogus regular expressions to regcomp. Cheers, Simon

1 0

ENQUIRY
by MR. DAVID LARSON 25 Jul '03

25 Jul '03

FROM: MR. DAVID LARSON ANTHONY EZE & CO. SIR, SUBJECT: TRUST NEED YOUR KIND RESPONSE TO THE FOLLOWING REQUEST: I HAVE A MANDATE TO SOURCE FOR A FINANCIAL MANAGEMENT PERSONNEL ASIDE AFRICA, BE IT YOUR PROFFESSION OR NOT, ITS NATURE DOES NOT COUNT, BUT ONLY HONESTY IS NEEDED AT ITS PEAK AND WANTED AS MOST PRIORTY TO ACTUALIZE THIS EXPECTED GOAL. HAS CLOSE LINK WITH THE PERSONNELS IN THE CORRIDORS OF POWER IN MY COUNTRY HAS BEEN DELEGATED TO SOURCE FOR A COMPETENT ASSOCIATE OUTSIDE AFRICA IF YOU ARE WORTHY OF BEEN TRUSTED WITH HUGE CAPITAL FOR ESTABLISHMENT IN YOUR COUNTRY CONTACT ME FOR FULL PERSONAL SECRETIVE INFORMATION OR DETIALS FOR COMMENCEMENT AND TERMS OF REWARD FOR YOUR EXPERTED INTEREST AND CONTRIBUTION. THE VOLUME OF CAPITAL WILL BE DISCLOSED TO YOU ON MY RECEIPT OF YOUR RETURN CORRESPONDENCE INDICATING YOUR WILLINGNESS TO PER-TAKE. KINDLY RESPOND TO THIS MAIL INCLUDING YOUR PERSONAL CONTACT DATA'S FOR FULL DISCLOSURE OF PROCEEDURES. JUST WAITING FOR YOUR REVERT THANKS MR. DAVID LARSON

1 0

RE: How do I create dynamic exceptions
by Bayley, Alistair 23 Jul '03

23 Jul '03

> From: Ross Paterson [mailto:ross@soi.city.ac.uk] > > On Tue, Jul 22, 2003 at 02:59:12PM +0100, Bayley, Alistair wrote: > > > handler :: Dynamic -> IO () > > > handler e = do > > > case (fromDynamic e) of > > > Nothing -> putStrLn ("dynamic cast failed") > > > Just (MkExc n s) -> putStrLn ("exception: " ++ (show > n) ++ " " ++ s) > > > > > main :: IO () > > > main = catchDyn temp handler > > You're catching a Dynamic wrapped as a Dynamic (catchDyn already does > fromDynamic, as indicated by its type). Change the handler to > > > handler :: MyException -> IO () > > handler (MkExc n s) = putStrLn ("exception: " ++ (show n) > ++ " " ++ s) > > any other dynamic exception fails through to the top level. Thanks. I was applying some advice from Hal Daume: > so, what you want is probably something like: > > > if rc < 0 > > then throwIO (DynException (toDyn ("Err", rc))) > > else ... > > and then when you catch, use fromDyn(amic) to get the value out if you can > predict its type... ... but in this case what I wanted was (just one more layer of indirection): > handler :: Exception -> IO () > handler e = do > case e of > DynException err -> case (fromDynamic err) of > Nothing -> putStrLn "dynamic cast failure" > Just (MkExc n s) -> putStrLn ("exception: " ++ (show n) ++ " " ++ s) > _ -> putStrLn "unhandled exception" > main :: IO () > main = catch temp handler -- use catch rather than catchDyn ***************************************************************** The information in this email and in any attachments is confidential and intended solely for the attention and use of the named addressee(s). This information may be subject to legal professional or other privilege or may otherwise be protected by work product immunity or other legal rules. It must not be disclosed to any person without our authority. If you are not the intended recipient, or a person responsible for delivering it to the intended recipient, you are not authorised to and must not disclose, copy, distribute, or retain this message or any part of it. *****************************************************************

1 0

RE: instance Typeable (Re: How do I create dynamic exceptions)
by Simon Peyton-Jones 22 Jul '03

22 Jul '03

| The above should work, but I wonder when it is possible to simply derive | Typeable? GHC 6.0 supports deriving( Typeable ), but for some strange reason it's not in the manual. It will be. Simon

2 1

RE: How do I create dynamic exceptions
by Bayley, Alistair 22 Jul '03

22 Jul '03

> The documentation for Data.Dynamic is pretty clear for this and I > believe with GHC6 you can derive Typeable. > > {-# NOINLINE myExceptionTyCon #-} > -- the documentation suggests a fully qualified name > myExceptionTyCon = mkTyCon "MyException" > > instance Typeable MyException where > typeOf _ = mkAppTy myExceptionTyCon [] > I disagree that the docs are clear. I could see that I needed a value of type TypeRep (a datatype with no constructors - so I can't create one directly), and also the mkApplyTy, mkFunTy, and applyTy functions, but it wasn't clear how to use them to get a TypeRep (which one do I use? and how?). An example like this is more helpful. Is there somewhere else I could have read about this? (I took a quick look at the wiki first.) I should have realised I could derive Typeable with ghc6; there was an itch in the back of my brain (now I remember reading about deriving Typeable in ghc6) that was trying to tell me there was an easy way... This is what I have at present, but it still doesn't catch the exception: > module Main where > import Prelude hiding (catch) > import Control.Exception > import Data.Dynamic > data MyException = MkExc Int String > deriving Typeable > mkErr :: Int -> String -> MyException > mkErr n s = MkExc n s > temp :: IO () > temp = do > putStrLn "line 1" > throwIO (DynException (toDyn (mkErr 2 "my temp"))) > handler :: Dynamic -> IO () > handler e = do > case (fromDynamic e) of > Nothing -> putStrLn ("dynamic cast failed") > Just (MkExc n s) -> putStrLn ("exception: " ++ (show n) ++ " " ++ s) > main :: IO () > main = catchDyn temp handler ***************************************************************** The information in this email and in any attachments is confidential and intended solely for the attention and use of the named addressee(s). This information may be subject to legal professional or other privilege or may otherwise be protected by work product immunity or other legal rules. It must not be disclosed to any person without our authority. If you are not the intended recipient, or a person responsible for delivering it to the intended recipient, you are not authorised to and must not disclose, copy, distribute, or retain this message or any part of it. *****************************************************************

2 1

How do I create dynamic exceptions
by Bayley, Alistair 22 Jul '03

22 Jul '03

Exceptions, again... The docs for Control.Exception say (for Dynamic exceptions): "When using dynamic exceptions it is advisable to define a new datatype to use for your exception type, to avoid possible clashes with dynamic exceptions used in other libraries." So I went with that and created a datatype for my exceptions: > data MyException = MkExc Int String ... but now it seems I have to install this as an instance of Typeable, and I don't see how to do that. > instance Typeable MyException where > typeOf ? = ? ***************************************************************** The information in this email and in any attachments is confidential and intended solely for the attention and use of the named addressee(s). This information may be subject to legal professional or other privilege or may otherwise be protected by work product immunity or other legal rules. It must not be disclosed to any person without our authority. If you are not the intended recipient, or a person responsible for delivering it to the intended recipient, you are not authorised to and must not disclose, copy, distribute, or retain this message or any part of it. *****************************************************************

3 2

Ignoring ExitExceptions
by Tim Jolly 21 Jul '03

21 Jul '03

I'm looking for a way to return an arbitrary exit code to the OS. The standard `exitWith' function does the job, but unfortunately it also has the side-effect of reporting an exception (I'm using GHC 6.0, with Control.Exception imported for other purposes). When a fatal error occurs, my program will already have issued a meaningful error message, and the added "*** Exception: exit: ExitFailure 2" will serve only to confuse users. Unfortunately it appears that the only way to return an exit code is to let an ExitException pass unhandled. All I'm really asking for is a way to use the exitWith function as defined in Haskell 98, along with the exception extensions provided in Control.Exception. Is there perhaps some way to prevent ExitExceptions from being reported, or a way to exclude certain types of exception from the handling in Control.Exception? Or can anyone suggest an alternative approach? -- TJ

1 0