April 2006 - Haskell-Cafe

PPM image manipulation
by newb 05 Apr '06

05 Apr '06

having some trouble trying to be able to manipulate PPM images. I want to be able to desaturate and also double the scale of them. im a first time user of haskell so am not very familiar with it. any info would be great! thanks! -- View this message in context: http://www.nabble.com/PPM-image-manipulation-t1382306.html#a3711988 Sent from the Haskell - Haskell-Cafe forum at Nabble.com.

3 3

RE: generics question 2
by Ralf Lammel 04 Apr '06

04 Apr '06

Hi Frederik, [resending; as it bounced because of size.] That’s a tricky one. Let’s first recall that this one is still fine: *Main> :t cast True :: (Typeable a, Typeable b) => Maybe (a,b) cast True :: (Typeable a, Typeable b) => Maybe (a,b) :: (Typeable b, Typeable a) => Maybe (a, b) The type error for your attempt was caused by the attempt to print this polymorphic maybe. The overloading resolution for the show instance cannot possibly deal with this polymorphism. More generally, any dependent SYB functionality or class-based code would run into this problem. However, the above polymorphic cast, as such does not work as you would expect. Here is a clear example: sigh = maybe 0 (\(a,b) -> gsize a + gsize b) (cast True) The component types of this product are *ambiguous*. Cast is monomorphic. No way to use it for polymorphic problems. So we need polymorphic type extension, here: dataCast2 provided by the Data class. So here is how it works with polymorphic type extension. See below and check out the 2nd paper for some deeper discussion. Or if you find this all too complicated, use class-based SYB style (3rd SYB paper). Thanks for keeping me off the street. ☺ Regards, Ralf import Data.Generics -- -- A weird gsize function -- g :: Data a => a -> Int g = gsize `extQ` (\(x::Int) -> x) `ext2Q` f where f :: (Data a, Data b) => (a,b) -> Int f (a,b) = g a + g b + 1 -- -- | Flexible type extension -- Should be in Data.Generics.Aliases -- ext2 :: (Data a, Typeable2 t) => c a -> (forall a b. (Data a, Data b) => c (t a b)) -> c a ext2 def ext = maybe def id (dataCast2 ext) -- -- | Type extension of transformations for unary type constructors -- Should be in Data.Generics.Aliases -- ext2T :: (Data d, Typeable2 t) => (forall d. Data d => d -> d) -> (forall d d'. (Data d, Data d') => t d d' -> t d d') -> d -> d ext2T def ext = unT ((T def) `ext2` (T ext)) -- -- | Type extension of queries for type constructors -- Should be in Data.Generics.Aliases -- ext2Q :: (Data d, Typeable2 t) => (d -> q) -> (forall d d'. (Data d, Data d') => t d d' -> q) -> d -> q ext2Q def ext = unQ ((Q def) `ext2` (Q ext)) -- -- | The type constructor for transformations -- Not exported from Data.Generics.Aliases -- newtype T x = T { unT :: x -> x } -- -- | The type constructor for queries -- Not exported from Data.Generics.Aliases -- newtype Q q x = Q { unQ :: x -> q } -- -- Time to test -- data Foo = Foo1 Int | Foo2 (Int,Int) deriving (Typeable, Data, Show) main = do print $ gmapQ g (Foo1 88) print $ gmapQ g (Foo2 (21,20)) > -----Original Message----- > From: Frederik Eaton [mailto:frederik@a5.repetae.net] > > What I want to do is specialise not to a specific type like Bool but > to the class of all pairs (a,b). But this causes the compiler to > complain, even for simpler examples: > > cast True :: (Typeable a, Typeable b) => Maybe (a,b) > > <interactive>:1:0: > Ambiguous type variable `a' in the constraints: > `Typeable a' arising from instantiating a type signature at > <interactive>:1:0-51 > `Show a' arising from use of `print' at Top level > Probable fix: add a type signature that fixes these type variable(s) > > <interactive>:1:0: > Ambiguous type variable `b' in the constraints: > `Typeable b' arising from instantiating a type signature at > <interactive>:1:0-51 > `Show b' arising from use of `print' at Top level > Probable fix: add a type signature that fixes these type variable(s) > > Is there a way to solve this, or do I have to avoid polymorphism? I > can use 'toConstr' to find out dynamically if a particular type is a > pair, and then use unsafeCoerce, but I hear that unsafeCoerce is > unsafe.

1 0

RE: generics question 2
by Ralf Lammel 04 Apr '06

04 Apr '06

> Hi Ralf, > > I'm looking for a function like extT but with more general type: > > (t a -> s a) -> (t b -> s b) -> (t a -> s a) > > Is there such a thing in the generics library? Hi Frederik, Not sure how you are exactly going to use such an operation ... But here is its implementation anyhow. Thanks for the riddle. Ralf import Data.Generics -- Frederik's weird ext operation :-) ext' :: (Data (t a), Data (s a), Data (t b), Data (s b)) => (t a -> s a) -> (t b -> s b) -> (t a -> s a) ext' f g ta = case cast g of Just g' -> g' ta Nothing -> f ta -- A generic default f (Just x) = [x] f Nothing = [] -- A type-specific case g (Just True) = [True] g (Just False) = [] g Nothing = [] -- A composition using our new type-extension operator test :: Data a => Maybe a -> [a] test = ext' f g -- Let's see whether it works ... main = do print $ test (Just (1::Int)) print $ test (Just False)

2 1

Creating a profiling version of FPS
by Jeremy O'Donoghue 04 Apr '06

04 Apr '06

Hi list, I'm currently working on a pretty simple Haskell program to manipulate memory dumps taken from an embedded environment. The memory dumps are 64MB in length, so I manipulate them using Don Stuart's Fast Packed String library, as it supports mmapped files. I'm a relative Haskell newbie, so unsurprisingly my first attempt has dreadful performance (30% of time spent in GC), and I want to do some profiling to find out why. The difficulty is that I seem to be unable to produce a profiling-enabled version of the FPS library, which means that my application cannot be compiled with profiling enabled. I notice that most of the GHC libraries seem to come with both profiling and non-profiling versions enabled. Does anyone know how I can create both profiling and non-profiling version of FPS? As a matter of interest (feature request), why doesn't Cabal build profiling and non-profiling libraries as a matter of course, given that you can't link non-profiling code with profiling-enabled code? Thanks in advance Jeremy

2 1

Re: Indent for haskell
by dons＠cse.unsw.edu.au 04 Apr '06

04 Apr '06

patc: > > Is there an equivalent of an "indent" program for haskell? I have a > bit of code I want to clean up ... You could run your code through ghc, with -ddump-parsed turned on. Then with a little bit of sed magic, you could recover the original code Before: $ cat B.hs main = do { ;print (f 10) ; print (f 20) ;print (f 0xdeadbeef) } f x = sqrt ( fromIntegral (x * 10133123 `mod` 1231111121 :: Int )) After: $ ghc -fno-code -ddump-parsed B.hs ==================== Parser ==================== main = do print (f 10) print (f 20) print (f 3735928559) f x = sqrt (fromIntegral ((x * 10133123) `mod` 1231111121 :: Int)) Postprocess: $ ghc -fno-code -ddump-parsed B.hs | sed '/^===/d' main = do print (f 10) print (f 20) print (f 3735928559) f x = sqrt (fromIntegral ((x * 10133123) `mod` 1231111121 :: Int)) Yay! Now, I suspect there may be actually be (not well publicised) tools for this. If there are, they should go on the haskell.org "Libraries and tools" page. -- Don

1 0

Strafunski
by Christopher Brown 03 Apr '06

03 Apr '06

Hi, I am trying to use Strafunski with GHC 6.5 and was wondering if someone could help me. I have all the instances for Term and Typeable defined for my data types, but when I try to compile with GHC 6.5 I get lots of "overlapping instance" errors. In particular, it seems the instances I am using (generated by DrIFT) are clashing with the ones in Data.Typeable. Is there a way I can fix this? Also I have heard that it is possible to add "deriving Typeable" to each data type and I don't need to use the instances I have created. However, now it complains that it can't find instances for Term - but I can't derive from Term. Does anyone have any ideas how I can get Strafunski working with GHC 6.5? Thanks. Chris. Christopher Brown PhD Student, University of Kent. http://www.cs.kent.ac.uk/people/rpg/cmb21/ cmb21(a)kent.ac.uk

4 6

Efficient Sets for Small Enumerations
by David F. Place 03 Apr '06

03 Apr '06

Often when writing algorithms which involve set operations on small enumerations, I start off using Data.Set. I soon find performance requires rewriting that code to use bitwise operations. I miss the nice interface of Data.Set and the type checking of using a proper data type. So, as an exercise in writing a library, I wrote out an implementation of bitwise set operations using the interface of Data.Set with a couple of extensions. It provides an abstract interface and type checking. Using GHC -O3, code compiles right down to the primitive bit-twiddling operators. My example program (a sudoku solver) runs several times faster. I'll be grateful for any feedback on this. Perhaps something like it would be useful included in the standard libraries. Cheers, David -------------------------------- David F. Place mailto:d@vidplace.com

3 6

Distributing monadic(?) functions across dyadic functions
by tpledger＠ihug.co.nz 03 Apr '06

03 Apr '06

Nils Anders Danielsson wrote: > A function like this has been suggested for the > standard libraries a couple of times before. > Someone suggested the name "on", which I quite > like: > > (*) `on` f = \x y -> f x * f y Thanks! I always wanted to be someone. :-) Here's the link. http://www.haskell.org//pipermail/haskell-cafe/2004-December/007917.html - Tom

1 0

RE: [Haskell-cafe] generics question, logical variables
by Ralf Lammel 03 Apr '06

03 Apr '06

Frederik, > As for your code example, it looks very interesting, but are you > saying that this could turn into an extension of the Data.Generics > library, or that this is something I could be implementing in terms of > what's already there? The posted code works with GHC 6.4 (SYB2) intentionally and actually. I have attached another attempt (again GHC 6.4, based on SYB2) which might be more useful for your purposes, and it may be useful in general, in fact. What I defined this time is a "certainty-improving" function: idify :: (Typeable1 f, Monad m, Data (a f), Data (a Id)) => (forall a. f a -> m a) -> a f -> m (a Id) That is, the function "idify get" takes a value whose type is parameterized in a type constructor f (such as Maybe or IORef), and the function attempts to establish Id instead of f on the basis of the function argument "get". > What is the 'a' parameter for in "force"? > > force :: ( Data (t Maybe a) > , Data (t Id a) > , Term t Maybe a > , Term t Id a > ) => t Maybe a -> t Id a The previous attempt was a more parameterized blow than required in your case. (I was guessing about what "typed logical variables" could mean. I was assuming that you would need some extra layer of embedding types on top of the Haskell term types. Looking at your code, this was not the case.) > For the part which I asked for help with, to get around my trouble > with generics, I defined a class GenFunctor and an example instance. > The intent is that generics should be able to provide this > functionality automatically later on, but you can see what the > functionality is. Let's look at the type of your GenFunctor: class GenFunctor f where gfmapM :: (Monad m, FunctorM b) => (forall x . a x -> m (b x)) -> f a -> m (f b) This type can be seen as a more relaxed version of the idify operation above. That is, idify fixes GenFunctor's b to Id. The particular encoding of idify (attached) takes advantage of this restriction. I wonder whether I should bother. (Exercise for reader :-)) > However, I am stuck on something else, the program doesn't typecheck > because of use of another function I defined, 'cast1'. Maybe you can > take a look. I had thought that I would be able to write a generic > 'unify' but I get the error: > > GenLogVar.hs:82:19: > Ambiguous type variable `a' in the constraint: > `Data a' arising from use of `cast1' at GenLogVar.hs:82:19-23 > Probable fix: add a type signature that fixes these type variable(s) > > This is because I need to do something special when I encounter a > "Var" variable in unification, but the compiler seems to not like the > fact that the type argument of the Var data type is not known. Please try to avoid new cast operations at all costs. :-) Your code can be arranged as follows: (i) Use extQ1 to dispatch to a special case for "Var x" for the first argument. (ii) In this special case, use again ext1Q to dispatch to a special case for "Var y" for the second argument. (iii) At this point, *cast* the variable value of *one* variable to the type of the other. So the problem with your code, as it stands, is that the target type of cast is ambiguous because you cast *both* arguments. The insight is to make the cast asymmetric. Then, not even polymorphism is in our way. Interesting stuff! Ralf

2 2

Haskell on Pocket PC?
by Dmitri O.Kondratiev 03 Apr '06

03 Apr '06

Any ideas on how much work needs to be done for using Haskell on PPC Windows Mobile platform? It would be interesting to use PPC as: 1) Haskell learning tool, so small code snipets could be entered and run directly on hand-held (REPL). How hard is it to port Hugs to PPC for this? Do any other (then Hugs) implementations exist that are easier to port? 2) Running on PPC Haskell applications cross-compiled on host PC. How much work must be done for GHC to support this? 3) There are several ways both for porting Haskell to PPC as well as cross-compiling Haskell code. Which one makes more sense to implement and will take less effort: a) Porting/compiling to VM byte code (Java, Parrot)? b) Porting/compiling to .NET CLR? c) Porting/compiling source code 4) And finally, do any projects already exist in this area? It would be great to get ideas how such port can be done in reasonable amount of time, in other words to find the most easy way to start with some Haskell implementation on PPC. Thanks for any info!

3 5