April 2005 - Haskell-Cafe

RE: [Haskell-cafe] FFI and pointers to pointers
by Simon Marlow 26 Apr '05

26 Apr '05

On 25 April 2005 21:40, John Meacham wrote: > On Mon, Apr 25, 2005 at 12:51:04PM +0100, Simon Marlow wrote: >> The syntax you want is >> >> foreign import ccall "&bar" :: Ptr CInt >> >> But note that GHC 6.4 has a bug whereby this doesn't work as expected >> when compiling via C. Either use the native code generator (-fasm) >> or wait for 6.4.1. > > Any ETA on 6.4.1? I am debating whether to rewrite code to work around > this bug or just wait for the point-release? We don't have any firm plans yet. It makes sense to wait until the flow of bug reports has slowed, and that hasn't happened yet. But I'll take your message as a vote for sooner rather than later. Cheers, Simon

1 0

How to debug GHC
by Monique Louise 25 Apr '05

25 Apr '05

Hi, all, I'm developing a back end for GHC and I have the following problem: my program is throwing an "empty list exception" due to head [] and I need to compile GHC with -prof -auto-all in order to see the stack trace when running it with +RTS -xc -RTS. I changed the makefile but the option +RTS -xc -RTS was not recognized as an available RTS option Does anyone have any idea about how I can do that ? Thanks in advance, -- ________________________________ Monique Louise B.Monteiro Msc Student in Computer Science Center of Informatics Federal University of Pernambuco

1 0

RE: [Haskell-cafe] FFI and pointers to pointers
by Simon Marlow 25 Apr '05

25 Apr '05

On 22 April 2005 16:18, Dimitry Golubovsky wrote: > I am trying to generalize my knowledge about FFI declarations when > dealing with pointers to pointers (import from C to Haskell). Maybe > these are silly questions, but It seems to me, I am missing some > understanding. > > Per the FFI Addendum: > > For a variable, we use "&" import: > > int bar; > > foreign import ccall "&" bar :: Ptr CInt The syntax you want is foreign import ccall "&bar" :: Ptr CInt But note that GHC 6.4 has a bug whereby this doesn't work as expected when compiling via C. Either use the native code generator (-fasm) or wait for 6.4.1. Cheers, Simon

2 1

RE: [Haskell-cafe] FFI and pointers to pointers
by Simon Marlow 25 Apr '05

25 Apr '05

On 25 April 2005 12:51, Simon Marlow wrote: > On 22 April 2005 16:18, Dimitry Golubovsky wrote: > >> I am trying to generalize my knowledge about FFI declarations when >> dealing with pointers to pointers (import from C to Haskell). Maybe >> these are silly questions, but It seems to me, I am missing some >> understanding. >> >> Per the FFI Addendum: >> >> For a variable, we use "&" import: >> >> int bar; >> >> foreign import ccall "&" bar :: Ptr CInt > > The syntax you want is > > foreign import ccall "&bar" :: Ptr CInt oops, never mind. Your version was fine. Cheers, Simon

1 0

RE: [Haskell-cafe] String search algorithms
by Bayley, Alistair 25 Apr '05

25 Apr '05

> From: Henning Thielemann [mailto:lemming@henning-thielemann.de] > > But I'm curious if you really need the index. Working with indexes on > lists is quite inefficient. E.g. if you want to replace > substrings you > may want to use this implementation: I'm not mutating the list, but I am extracting sublists. I'm toying with porting FIT and I'm starting with the Parse class, as recommended ( http://fit.c2.com/wiki.cgi?TipsForCoreImplementors ). They warn against using a generic parser: "Parse is the internal representation of the test data. Don't be seduced by generalized html/xml parsers. This one has unique features. Make it work. Consider replacing it only after you've used this one a while." Their parser is in Java, and makes heavy use of String.indexOf(), and I'm just trying to port it more-or-less as-is. I appreciate that'll be somewhat inefficient, but until I understand all of the wrinkles in their implementation, I'll take the simplest approach. Alistair. ----------------------------------------- ***************************************************************** Confidentiality Note: The information contained in this message, and any attachments, may contain confidential and/or privileged material. It is intended solely for the person(s) or entity to which it is addressed. Any review, retransmission, dissemination, or taking of any action in reliance upon this information by persons or entities other than the intended recipient(s) is prohibited. If you received this in error, please contact the sender and delete the material from any computer. *****************************************************************

1 0

String search algorithms
by Bayley, Alistair 25 Apr '05

25 Apr '05

I'm a bit puzzled to find no sub-string search in the Haskell libraries (unless there's some neat composition of the existing Data.List functions that I've missed). Google doesn't help much either. I've found a KMP implementation: http://haskell.org/hawiki/RunTimeCompilation I'm after something that'll report the position of the first occurrence, like Java's String.indexOf(). Alistair. ----------------------------------------- ***************************************************************** Confidentiality Note: The information contained in this message, and any attachments, may contain confidential and/or privileged material. It is intended solely for the person(s) or entity to which it is addressed. Any review, retransmission, dissemination, or taking of any action in reliance upon this information by persons or entities other than the intended recipient(s) is prohibited. If you received this in error, please contact the sender and delete the material from any computer. *****************************************************************

2 2

RE: [Haskell-cafe] String search algorithms
by Bayley, Alistair 25 Apr '05

25 Apr '05

> From: Henning Thielemann [mailto:lemming@henning-thielemann.de] > > > (unless there's some neat composition of the existing > Data.List functions > > List.findIndex (List.isPrefixOf "bla") (List.tails "dfvbdbblaesre") Oh, so there is. Thanks. ----------------------------------------- ***************************************************************** Confidentiality Note: The information contained in this message, and any attachments, may contain confidential and/or privileged material. It is intended solely for the person(s) or entity to which it is addressed. Any review, retransmission, dissemination, or taking of any action in reliance upon this information by persons or entities other than the intended recipient(s) is prohibited. If you received this in error, please contact the sender and delete the material from any computer. *****************************************************************

1 0

Why doesn't this work?
by Michael Vanier 25 Apr '05

25 Apr '05

I've been trying to generate an infinite list of random coin flips in GHC 6.4, and I've come across some strange behavior: ---------------------------------------------------------------------- import System.Random data Coin = H | T deriving (Eq, Show) -- Generate a random coin flip. coinFlip :: IO Coin coinFlip = do b <- getStdRandom random return (bool2coin b) where bool2coin True = H bool2coin False = T -- Generate an infinite list of coin flips. coinFlips :: IO [Coin] coinFlips = sequence cfs where cfs = (coinFlip : cfs) -- Print n of them. test :: Int -> IO () test n = do f <- coinFlips print (take n f) ---------------------------------------------------------------------- Now when I do "test 1" (for instance), it hangs forever. It seems as if there is some kind of strictness constraint going on that I don't understand. My understanding is that cfs is an infinite list of (IO Coin), sequence lifts this to be IO [Coin] where [Coin] is an infinite list, and then test should extract the infinite list of coin flips into f, take some number of them, and print them. But instead, the system appears to be trying to compute all the coin flips before taking any of them. Why is this, and how do I fix it? Thanks, Mike

5 4

Haskell example archive?
by Echo Nolan 25 Apr '05

25 Apr '05

Hello all, I recently read the post about a problem using the list monad, and I was wondering if there was an archive of monad usecases. If there is one, I'd like to see it, and if not it'd be a helpful part of the haskell community. Something like the example of the Maybe monad from "All About Monads".

3 2

a simple algebra for creating & editing values
by Conal Elliott 23 Apr '05

23 Apr '05

I've been playing a simple idea for creating and editing values, and I'd appreciate some feedback on whether or not this idea is well-explored territory. The idea is to construct functions that direct transformations (functions or other arrow-typed values) to the insides of values. This note is a literate Haskell program briefly demonstrating the idea. It runs under hugs and ghc/ghci. > module Path where > import Control.Arrow Consider, for example, a value of type (a->(f,b->(c,g)),e), and Suppose we'd like to apply a function to the c part. We'd want a transformation path > p1 :: (c->c') -> (a->(f,b->(c ,g)),e) > -> (a->(f,b->(c',g)),e) To define p1, simply list the steps taken to get to the type c in the type (a->(f,b->(c,g)),e), as follows: > p1 = first.result.second.result.first Function arguments may be edited also. In general, the part-changer (of type c->c' in p1) will need to have its domain and range types swapped if it appears in a negative position. > p2 :: (b'->b) -> (d->(a,b)->c) -> (d->(a,b')->c) > p2 = result.argument.second > p3a :: (a->a') -> ((a->b)->c) -> ((a'->b)->c) > p3a = argument.argument > p3b :: (a->a') -> ((((e,a)->b),d)->c) -> ((((e,a')->b),d)->c) > p3b = argument.first.argument.second The "first" and "second" path components are simply the Arrow methods of those names. As for "result" and "argument", for functions, we can define > result :: (b->b') -> (a->b)->(a->b') > argument :: (a'->a) -> (a->b)->(a'->b) >From the types, you can guess the definitions: > result = (.) > argument = flip (.) (BTW, Fritz Ruehr pointed out in his Feb 20 haskell-cafe note that compositions of (.) can be used to "compose a 1-argument function with an n-argument function".) Generalizing to arrows, look at the types of the "first" and "second" Arrow methods: > class Arrow arrow where > first :: arrow a a' -> arrow (a, b) (a', b ) > second :: arrow b b' -> arrow (a, b) (a', b ) > ... Similarly, define "result" and "argument" as methods of a new class: > class Arrow arrow => ArrowRA arrow where > result :: arrow b b' -> arrow (a->b) (a ->b') > argument :: arrow a' a -> arrow (a->b) (a'->b ) > instance ArrowRA (->) where > result = (.) > argument = flip (.) (I've also defined ArrowRA instances for GUI construction and for Haskell code generation.) Here's a suitable second argument for p1: > has1 :: (Bool -> (Bool, String -> (String, Int)), Bool) > has1 = (\ a->(not a, \ b->("hello"++b,3)), True) To see the result, fill in function arguments > x1 :: (Bool -> (f, String -> cg), e) -> ((f, cg), e) > x1 = first (second ($" world") . ($True)) and test: > t1a = x1 $ has1 -- ((False,("hello world",3)),True) > t1b = x1 $ p1 reverse has1 -- ((False,("dlrow olleh",3)),True) > t1c = x1 $ p1 length has1 -- ((False,(11,3)),True) Here's a suitable second argument for p2: > has2 :: Bool->(String,Int)->Bool > has2 b (str,n) = b && length str == n To see the result, fill in function arguments > x2 :: (Bool->(String,a)->Bool) -> a ->Bool > x2 f a = f True ("string",a) For testing: > t2a = x2 (p2 length has2) "bow" -- False > t2b = x2 (p2 ($ "fiddle") has2) length -- True These transformation paths allow us to apply a function inside of a value. What about applying a function that itself is inside some value. Transformation paths can do this job also, if we combine them with reverse application. In general, both the function and its argument may be buried inside of values. In that case, we might want the function's context to end up on the outside and the argument's context on the inside, or vice versa. > applyF pathF pathX hasF hasX = > pathF (\ f -> pathX (\ x -> f x) hasX) hasF > applyX pathF pathX hasF hasX = > pathX (\ x -> pathF (\ f -> f x) hasF) hasX As an example, > hasF1 = ("square", (^2)) > hasX1 = (3,"three") > t3a = applyF second first hasF1 hasX1 -- ("square",(9,"three")) > t3b = applyX second first hasF1 hasX1 -- (("square",9),"three") How about some other type constructors, besides (,) and (->)? Try lists. > class Arrow arrow => ArrowL arrow where > elementL :: arrow a a' -> arrow [a] [a'] > instance ArrowL (->) where elementL = map Generalizing from [] to any functor: > class Arrow arrow => ArrowF arrow where > elementF :: Functor f => arrow a a' -> arrow (f a) (f a') > instance ArrowF (->) where elementF = fmap The following example combines all four transformation combinators so far: > p4 :: (b'->b) -> (a->Maybe(b,d)->c) -> (a->Maybe(b',d)->c) > p4 = result.argument.elementF.first Comments? - Conal

2 1