March 2004 - Haskell-Cafe

Extreme Markup Languages Conference 2004
by srn＠coolheads.com 26 Mar '04

26 Mar '04

Two kinds of information are necessary for human life on this planet: (1) Genetic information. No batteries required. (2) Everything else, i.e., civilization. Batteries required. Please be welcome to participate in the 11th annual Extreme Markup Languages Conference --> August 2-6, 2004 <-- Hotel Europa Montreal, Canada www.extrememarkup.com As always, it's a family gathering for rubber-meets-the-road technical luminaries interested in doing a better job of supporting civilization. Come share your light, and be brightened, too. You'll be glad you did, and not only because Montreal in August is great fun. Peer-reviewed paper submissions are due --> April 16, 2004 <-- Submission guidelines: http://www.mulberrytech.com/Extreme (e-mail questions to extreme(a)mulberrytech.com) Peer Reviewer Applications are due TODAY, Friday March 26, 2004. Tutorial Proposals are due TODAY, Friday March 26, 2004. -- Steve Steven R. Newcomb, Co-chair Extreme Markup Languages 2004 -- An IDEAlliance Event www.extrememarkup.com srn(a)coolheads.com ------------------------------------------------------- Coolheads Consulting http://www.coolheads.com direct: +1 540 951 9773 main: +1 540 951 9774 fax: +1 540 951 9775 208 Highview Drive Blacksburg, Virginia 24060 USA

1 0

RE: [Haskell-cafe] looking for optimization advice
by Simon Marlow 26 Mar '04

26 Mar '04

> > data PackedString = PS !(ForeignPtr Word8) !Int !Int A ForeignPtr is a sum type, so it can't be unboxed unfortunately. This is undoubtedly one cause of inefficiency in the code. Depending on how you're using ForeignPtrs, you could possibly define a specialiased kind of ForeignPtr that would be unboxable. This would require using low-level non-portable stuff though. Are you using mallocForeignPtr? Or adding extra finalizers to ForeignPtrs? Cheers, Simon

2 1

Implementing and indexing 2 dimen arrays
by Alex Gontcharov 25 Mar '04

25 Mar '04

I am writing a function, It's my first Haskell program in output function I will be working on [[3,4],[1,2,3,2],[3,9,9,4],[8,3,3,4] type structures I'd like to know how to index a particular element at run-time as I don't know on which element of which sublist I am at the moment in pullHelper function which starts with values !! ind p xs !! in C it would look something like values[][10]={{....}.{....}..}; int index[4]={3,4,6,7}; int i; i for(i = 0; i<4; i++) { int c = index[i]; if(values[i][c-1] == 0) /* do something */ else /*other things*/ } Any help would be appreciated --Stack for the digits of numbers, a modulo b, stack :: Int->Int->[[Int]] stack x y = reverse (split (subStack x y)) where --Split the digits further and constructs lists of lists split :: [Int]->[[Int]] split xs | length xs == 0 = [] | otherwise = (reverse (subStack (head xs) 10)): (split (tail xs)) subStack :: Int->Int->[Int] subStack x y | x == 0 = [] | otherwise = (mod x y):(subStack (div x y) y) -- It should work for 999999999 range but since I am using -- Int maximum is 2^31-1 output :: Int->[Char] output n = loop (stack n 10000) where -- goes through sets of at most 4 digits in the list, loop loop :: [[Int]]->[Char] loop xs | length xs == 0 = [] | otherwise = (strCon (head xs)) : (loop (tail xs)) strCon :: [Int]->[Char] strCon xs | length xs == 0 = [] | otherwise = (pullStr (head xs)) : (strCon (tail xs)) pullStr :: Int->[Char] pullStr s | s == 0 = [] | otherwise = pullHelper s --Helping to pull elements from the set, nested if pullHelper :: Int->[Char] pullHelper p | values !! ind p xs !! p - 1 == 0 = ones !! p - 1 ++ values !! ind p xs !! 0 | otherwise = values !! ind p xs !! p - 1 --Nasty !

3 2

zip comprehensions and local decls
by Christian Sievers 25 Mar '04

25 Mar '04

Hello! This is about a misconception I had about zip comprehensions and a possible extension to get the desired result. As zip compr. are implemented in hugs and in ghc (though I only tried hugs), I thought that haskell-cafe might be a good place for it. In hugs mode, I expected something like the following to work: [ (a,a*x) | x <- [1..5] | y <- [10..], let a=x+y ] but hugs says: ERROR - Undefined variable "x" I think I can see why: The expression is rewritten to [ (a,a*x) | (x,y) <- zip [ x | x<-[1..5] ] [ y | y <- [10..], let a=x+y ] ] This scheme is useful if one has ...|y<-list, let a=..., sometest a which drops some elements from the list before zipping. But it would be nice to have a way to drop already zipped tuples and to allow such usages of local declarations. Maybe a final '|' branch without generator, so to get what was intended with the expression above one might write [ (a,a*x) | x <- [1..5] | y <- [10..] | let a=x+y ] or, as another example, one could have [ (x,y) | x <- [1..5] | y <- [10..] | even x ] giving [(2,11),(4,13)]. So the idea is to rewrite [ e | oldstyle parallel list comprehension | declsandpreds ] to [ e | zipN <as before...> , declsandpreds ] Is this clear enough? Does it sound useful? Can you think of a nicer syntax? All the best Christian Sievers

1 0

looking for optimization advice
by David Roundy 25 Mar '04

25 Mar '04

Hello everyone, I've got a function that I can't think how it can be optimized any more, and perhaps it can't, but just in case, I figured I'd run it by you all, since it's currently taking 70% of my CPU time. The main problem is an unrelated scaling issue (i.e. it's getting called too many times... this is a separate issue I'm working on), but even in "reasonable" cases, it takes 20% of the processor time. The function is a simple packed string compare, and I basically just call the C standard library function memcmp for this. Without further ado: data PackedString = PS !(ForeignPtr Word8) !Int !Int instance Eq PackedString where (==) = psniceq foreign import ccall unsafe "static string.h memcmp" c_memcmp :: Ptr Word8 -> Ptr Word8 -> Int -> IO Int {-# INLINE psniceq #-} psniceq :: PackedString -> PackedString -> Bool psniceq a b | nullPS a && nullPS b = True psniceq (PS x1 s1 l1) (PS x2 s2 l2) = (l1 == l2 &&) $ unsafePerformIO $ withForeignPtr x1 $ \p1-> withForeignPtr x2 $ \p2 -> liftM (==0) $ c_memcmp (p1 `plusPtr` s1) (p2 `plusPtr` s2) l1 Is there anything obvious that I might have missed? Or is there anything incredibly subtle that I missed? :) BTW, I compile this with -O -funbox-strict-fields, so the packed string contents are unboxed, if that makes any difference. -- David Roundy http://www.abridgegame.org

3 3

Indentation-sensitive Parsec components?
by Graham Klyne 25 Mar '04

25 Mar '04

I'm wondering if anyone has implemented any indentation-sensitive parser components that can be used with Parsec. I'm thinking of components along these lines: [[ blockStart :: Parser () -> Parser Block blockSep :: Block -> Parser () -> Parser () blockEnd :: Block -> Parser () -> Parser () ]] which might be used (somewhat self-referentially) like this: [[ doSequence :: Parser () doSequence = do { symbol "do" ; b <- blockStart (symbol "{") ; e1 <- expression ; es <- many ( do { blockSep b (symbol ";") ; expression } ; blockEnd b (symbol "}") ; dostuff (s1:ss) } ]] The intent is that blockStart would look for the separator component, and if not present would determine the appropriate indentation level from the input character stream. The other components would match or fail according to the way that the corresponding blockStart was matched. #g ------------ Graham Klyne For email: http://www.ninebynine.org/#Contact

1 0

HToolkit HSQL on Windows/GHC
by Clive Brettingham-Moore 25 Mar '04

25 Mar '04

I realise that this is a little technical for this list, but I thought that I'd check here before starting on the hard-core lists. Is there an easy way to build the HToolkit HSQL library for GHC (6.0.1 when I tried, 6.2.1 now) on windows (particularly MySQL/Windows XP, but I wont say no to postgres, ODBC & sqlite)? I assume people have tired and succeded, but I got lost [I think] in the nasty place between the unix/cygwin libraries in GHC and the Windows focus of the compiled MySQL client library. Since haskell is mainly a hobby for me, and I don't have a lot of free time I haven't dug very deep, but not being able to interface my two of my favourite constructs (functional programming and relational databases) is getting to me. So any takers? C

1 0

Implementing C program in Haskell
by Alex Gontcharov 24 Mar '04

24 Mar '04

Hi, Is anyone bored and wants to rewrite a small C program in Haskell?:) http://users.tpg.com.au/galutva/jap.c I am new to Haskell and it'd take me sometime to do it all myself. Thanks

2 1

Where's the error in this snippet of code?
by Alex Gontcharov 24 Mar '04

24 Mar '04

Ignore the layout I can't find the error, running it gives parse error during compile on pStack, it is not very descriptive and I don't what is wrong. --Stack for the digits for numbers, a modulo b digStack :: Integer->Integer->[Integer] digStack a b | a == 0 = [] | otherwise = (mod a b):(digStack (div a b) b) --Takes a list of remainders and pStack :: [Integer] -> Integer -> [Integer] pStack [] _ = [] pStack (x:xs) a = digStack x a ++ pStack xs I think I will never like Haskell :-)

4 3

Re: [Haskell-cafe] Where's the error in this snippet of code?
by Duncan Coutts 24 Mar '04

24 Mar '04

On Wed, 2004-03-24 at 18:00, Alex Gontcharov wrote: > I don't want to hide it within the function, I need the functions > to have file scope. Oh, ok. I assumed from the indentation (you see layout matters!) that it was a nested function. Haskell (like Python) is layout sensitive, the indentation makes a difference. All function definitions in the same scope have to line up in the same column. So for global/file scope functions they should not be indented at all. So the problem with your example was that your second function was indented. Write: globalFunc1 = ... globalFunc2 = ... and not: globalFunc1 = ... globalFunc2 = ... > So, If you want to call a function from another function b I have to > nest it using where? No, only if you want it nested rather than whole file scope. Duncan

1 0