May 2008 - Haskell-Cafe

hsc2hs: expanding macros in the .hsc file
by Olivier Boudry 13 May '08

13 May '08

Hi all, Is it possible to expand macros defined in includes into the .hsc file? I'm trying to call functions from a library written in C. The library can be used with or without Unicode chars, depending on #define instructions. The library has macros for all the standard functions used to work on strings (malloc, strlen, strcpy, strtok, ...). So that in code using the library one can always call the macro and don't care about the char type (strlenU instead of wcslen or strlen). Macro expansion will define which function will be used depending on the SAPwithUNICODE #define. I put a very simplified example below. The original file contains much more conditions (OS, Unicode or not, little or big endian, ...). #define strlenU SAP_UNAME(strlen) #ifdef SAPwithUNICODE #define SAP_UNAME(name) name##U16 #else #define SAP_UNAME(name) name #endif #if defined(WCHAR_is_2B) #define strlenU16 wcslen #endif I would like to be able to expand strlenU in my .hsc file and get the correct function. foreign import ccall "static sapuc.h strlenU" f_strlenU :: Ptr (#type SAP_UC) -> IO (#type size_t) I would like to be able to expand the strlenU macro to the real function name (wcslen or strlen). Is there a way to do this? Thanks, Olivier.

1 0

M1 + M2 = M3 where both computations in M1 and M2 can be used?
by sam lee 12 May '08

12 May '08

Hi. I want to compose two monads to build another monad where computations of the two monads can be used inside. I have: - MonadTypeInfer : interface (class) for TypeInfer monad - TypeInfer : a monad that has Map String Type (association of names and types) - TypeInferT : transformer of above monad - MonadEval : interface (class) for Eval monad - Eval : a monad that has Map String Expr (association of names and code/function body) - EvalT : transformer of Eval - tInfer :: Expr -> TypeInfer Type -- given expr, returns type of it in TypeInfer monad - eval :: Expr -> Eval Expr -- given expr, returns normalized expr in Eval monad Problem: in repl, when user defines a function, it should type check and register type of the function to TypeInfer monad's Map String Type. Also, it should store the expression of the function in Eval monad. I build REPL monad using TypeInferT and EvalT. > newtype REPL a = REPL { runREPL :: TypeInferT (EvalT IO) a } > deriving(Monad, Functor, MonadIO, MonadTypeInfer, MonadEval) > repl :: REPL () > repl = do > input <- prompt ">>> " > case parse input of > Left err -> -- handle error > Right expr -> do > t <- tInfer expr -- BAD!! tInfer :: TypeInfer Type > println (show t) > result <- eval expr -- BAD!! eval :: Eval Expr > println (show result) > repl Should I make tInfer :: REPL Type, eval :: REPL Expr? Is there a way to build a monad where you could use sub-monads' (monads used to build current monad) computations? I prefer keeping tInfer :: TypeInfer Type, eval :: Eval Expr because tInfer never uses actions in Eval monad and vice versa. It seems like what I am asking is to break the type system. Maybe I should just make them run in REPL monad. Thank you. Sam.

3 3

Re: [Haskell-cafe] List concat
by Lennart Augustsson 12 May '08

12 May '08

Well, if you want to be picky evaluating (calling) (xs ++ ys) takes whatever time it takes to evaluate xs to WHNF and if xs is null then it in addition takes the time it takes to evaluate ys to WHNF. Saying anything about time complexity with lazy evaluation requires a lot of context to be exact. -- Lennart On Mon, May 12, 2008 at 7:36 AM, <ajb(a)spamcop.net> wrote: > G'day all. > > Quoting Andrew Coppin <andrewcoppin(a)btinternet.com>: > > The function (++) :: [x] -> [x] -> [x] has O(n) complexity. >> > > That's not entirely true. > > When you call (++), it does O(1) work. If you evaluate k cons cells. > it takes O(min(k,n)) work. > > Cheers, > Andrew Bromage > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe >

1 0

On syntactic sugar
by Lauri Oksanen 12 May '08

12 May '08

Hi, I'm writing my first real Haskell program and I came up with the following code snippet. --- let x' = x \+ dist \* dir nx' = normal geometry wi = (-1) \* dir in do (p, wo) <- brdfCosSampling reflector nx' wi let color' = p \** color q = min 1 (scalarContribution p) in do sampler <- biasedCoinSampler q (radianceSampler surfaces x' wo (q \* color')) (terminalRadianceSampler surfaces x' nx' ((1-q) \* color')) sampler --- This works just fine but I don't like the way I had to indent the code because of alternating lets and dos. I would like to indent the code more like an imperative code i.e. like this --- let { x' = x \+ dist \* dir ; nx' = normal geometry ; wi = (-1) \* dir ; } in do { (p, wo) <- brdfCosSampling reflector nx' wi ; let { color' = p \** color ; q = min 1 (scalarContribution p) ; } in do { sampler <- biasedCoinSampler q (radianceSampler surfaces x' wo (q \* color')) (terminalRadianceSampler surfaces x' nx' ((1-q) \* color')) ; sampler ; }} --- but without braces and semicolons. Following works also, but is ugly (and probably less efficient?). --- do x' <- return $ x \+ dist \* dir nx' <- return $ normal geometry wi <- return $ (-1) \* dir (p, wo) <- brdfCosSampling reflector nx' wi color' <- return $ p \** color q <- return $ min 1 (scalarContribution p) sampler <- biasedCoinSampler q (radianceSampler surfaces x' wo (q \* color')) (terminalRadianceSampler surfaces x' nx' ((1-q) \* color')) sampler --- Is there some nice trick to do this?

4 3

Re: [Haskell-cafe] List concat
by Derek Elkins 12 May '08

12 May '08

On Mon, 2008-05-12 at 02:36 -0400, ajb(a)spamcop.net wrote: > G'day all. > > Quoting Andrew Coppin <andrewcoppin(a)btinternet.com>: > > > The function (++) :: [x] -> [x] -> [x] has O(n) complexity. > > That's not entirely true. > > When you call (++), it does O(1) work. If you evaluate k cons cells. > it takes O(min(k,n)) work. I think we can safely simplify that to O(k) (at least for sequential programs).

1 0

help(!!)
by Galchin, Vasili 12 May '08

12 May '08

Hello, As I said in a Friday posting, I changed the version # of the cabal file that I am using, unix.cabal. Now my build environment is broken and I am dead in the water: vigalchin@ubuntu:~/FTP/Haskell/unix-2.2.0.0$ runhaskell Setup.hs configure Configuring unix-2.3.0.0... vigalchin@ubuntu:~/FTP/Haskell/unix-2.2.0.0$ runhaskell Setup.hs build Setup.hs: unix.cabal has been changed, please re-configure. vigalchin@ubuntu:~/FTP/Haskell/unix-2.2.0.0$ It seems like "runhaskell Setup.hs configure" isn't doing much after the unix.cabal change. How can I force the configure to really run? Do something with some of the config files?? Thanks, Vasili

2 1

Re: [Haskell-cafe] saner shootout programs
by PR Stanley 12 May '08

12 May '08

>I don't know Haskell very well, but Paul: "I'm not racist but . . ." :-) >even I can tell, looking at, for >example, the N-body benchmark, that the Haskell code is probably not >type-safe, and the tricks used in it would not be usable in a larger >program (see below). > >The task is essentially a pure computation: take a list of bodies >having mass, position and velocity; apply Newton laws at discrete >intervals for a large number of times; return new positions and >velocities. > >I could write a C++ procedure that performs this task and have some >piece of mind regarding its type correctness, exception safety and >functional purity: side effects would be local to the procedure, >arguments passed as const or by value, the result returned by value, >no type casts or new/delete operators used. > >On the other hand, the Haskell code makes assumptions about the size >of double-precision floats (obviously not type-safe). Further, the >simulation is not a pure function. > >It is often argued that one only needs these dirty tricks in the most >time-consuming functions. However, if using imperative programming in >these "inner loop" procedures places them in IO monad, the "outer >loops" (the rest of the program - procedures that call it) will have >to go there as well. This makes me doubt the Haskell approach to >functional programming. > >If anyone has a version of the N-body benchmark, where the simulation >is a type-safe pure function, I would very much like to see and time >it. >_______________________________________________ >Haskell-Cafe mailing list >Haskell-Cafe(a)haskell.org >http://www.haskell.org/mailman/listinfo/haskell-cafe

2 1

List concat
by Andrew Coppin 12 May '08

12 May '08

The function (++) :: [x] -> [x] -> [x] has O(n) complexity. If somebody were to invent some type that looks like [x] but actually uses some sort of tree rather than a linked list, you should be able to get O(1) concatenation. Has anybody ever implemented such a thing?

7 7

Interesting critique of OCaml
by Don Stewart 12 May '08

12 May '08

An interesting critique of OCaml. http://enfranchisedmind.com/blog/2008/05/07/why-ocaml-sucks/ One phrase that stood out, regarding GHC's support for deforestation transformations like build/foldr and stream fusion: "Haskell is doing data structure level optimizations with the ease that most other compiler do peephole instruction optimization. This is a non-trivial result." Which I think really captures the joy of being able to write algebraic and data structure transformations, via rewrite rules, without having to extend the compiler -- all thanks to purity, laziness, and static typing. -- Don

12 16

dropping hyphens and \n in words
by Ivan Amarquaye 11 May '08

11 May '08

Generally a hyphen is written at the end of the sentance when moving on to the next line and i managed to achieve this in haskell by using the "\n"- newline which places an index word in the next line i.e. if the words appear indexed like this...([1]),[mangoes] and a hyphen is applied, it becomes ([1],[mang-oes]) and it is valid in my function as i made it accept hyphens as part of a single word. Now my problem is this...I'm assuming that the hyphen normally comes at the end of a sentence like this: "there are so many guys ravis-hing our women" and this can be demonstrated in haskell by "\n" which places the words or characters following it in a new line like this: input: makeIndex"there are so many guys ravis\nhing our women" and output is: (([1],[there]),([1],[ravis]),([2],[hing])) where 1 means the first line and 2 the next. Now i want to write a function that would take away the hyphen and \n from all the words supposed to end on the first line and continue on the next and make all appear on the first line like this: all words in this form: "chip-\nheater" should become "chipheater". hope i can get some guidance on doing this. _________________________________________________________________ Connect to the next generation of MSN Messenger http://imagine-msn.com/messenger/launch80/default.aspx?locale=en-us&source=…

4 4