October 2002 - Haskell-Cafe

FFI and GHC
by João Ferreira 20 Oct '02

20 Oct '02

Hello all, I am trying to learn how to use FFI with GHC. I've tried the example in GHC's user's guide but it didn't worked. Here it is: 1) The haskell code =================== [bracaman@tucano hib]$ cat foo.hs foreign export ccall foo :: Int -> IO Int foo :: Int -> IO Int foo n = return (length (f n)) f :: Int -> [Int] f 0 = [] f n = n:(f (n-1)) 2) The error ============ [bracaman@tucano hib]$ ghc -c foo.hs foo.hs:3: Type signature given for an expression [bracaman@tucano hib]$ I just can't understand why this is happening... can someone, please, help? Thank you in advance, João Ferreira

3 2

Re: Haskell-Cafe digest, Vol 1 #518 - 4 msgs
by Berron Lara Virginia 19 Oct '02

19 Oct '02

Hi everybody! Is there any tool for draWING the graph of a function in one or two variables in Haskell? Thanks in advance Virgin

1 0

Any way to read in a time?
by David Roundy 18 Oct '02

18 Oct '02

I've been looking at the standard Time module, and it seems pretty easy to convert a date into a string (using either calendarTimeToString or formatCalendarTime), but there doesn't seem to be any way to reverse the process. I understand that there are so many date and time formats that it may be hard to parse, and that for many formats there may be no unambiguous translation back into a CalendarTime (e.g. if the time zone isn't included in the string), but it seems like a pretty fundamental (and commonly needed) ability to be able to output a human-readable date and then read it back in. Does anyone have a suggestion how best to go about this? I guess the easiest way might be to just convert to UTC and then use formatCalendarTime to choose what formatting I want and then write my own parser to reverse the process. But this seems like a process with a relatively high probability of introducing rarely seen bugs, which ought to be dealt with by a standard library. Of course, there are also a couple of other workarounds I could use. CalendarTime derives from Show, so I could just output the result of show, but that's an awfully verbose format. I could also give up on comparing dates (which I don't necesarily need, although it is nice) and just treat dates as strings. Any ideas or commentary would be welcome. -- David Roundy http://civet.berkeley.edu/droundy/

3 2

(no subject)
by Chris.Angus＠Lawson.com 17 Oct '02

17 Oct '02

Does anyone know what happened to the wiki on haskell.org? There was lots of stuff there last year (mostly to do with design patterns) which seems to have been lost. Some of the links are still there but they don't go anywhere. Cheers Chris

3 2

question about lazy io
by David Roundy 12 Oct '02

12 Oct '02

Hello. I have a problem I ran into when I tried writing a function to use lazy IO . Specifically, I would like to be able to compare two directory trees, and I thought the best way to do this would be to compartmentalize things by writing an IO function to read in a whole directory tree (including the contents of its files) in a lazy manner, and then the code to manipulate this tree could be purely functional and non-monadic. The problem is that while reading a file in with readFile (or openFile followed by hGetContents) is lazy regarding the reading of the actual file data, it leaves the file handle open, and (at least on ghc under linux) quickly exhausts the available file handles. I guess this makes sense, since if the file wasn't opened until it was actually used, a file open error couldn't flag an exception at the appropriate time... I think this is a related question to the thread on config files and pure file reading. Certainly if there was a pure functional file read (i.e. readFileOnce), that would solve my problem. I hate to have to write monadic IO code right into the core of my code. The Haskell 98 Library Report describes the state of a file handle after hGetContents as 'semi-closed', meaning it will close as soon as it is finished with the current read. Is there any way to get the file handle into a 'semi-opened' state, in which it will be opened as soon as it starts getting read? That's what I'd really like. fwiw, here's the code I'd like to use to slurp up a directory and its contents: >slurp dirname = do > isdir <- doesDirectoryExist dirname > if isdir > then do > former_dir <- getCurrentDirectory > fnames <- getDirectoryContents dirname > setCurrentDirectory dirname > sl <- slurp_list $ skip_hidden fnames > setCurrentDirectory former_dir > return $ SlurpDir dirname sl > else do > isfile <- doesFileExist dirname > if isfile > then do > contents <- readFile dirname > return $ SlurpFile dirname (lines contents) > else return $ SlurpFile "Oops" [] > >slurp_list :: [FilePath] -> IO [Slurpy] >slurp_list [] = return [] >slurp_list (f:fs) = do > s <- slurp f > ss <- slurp_list fs > return (s:ss) > >skip_hidden :: [FilePath] -> [FilePath] >skip_hidden [] = [] >skip_hidden (('.':_):fps) = skip_hidden fps >skip_hidden (fp:fps) = fp : skip_hidden fps To restate my question (which I may have forgotten to ask...), is there a good way to do this? The only thing that comes to mind is to use unsafeIO, which sounds scary. (Mostly because I don't think I would really understand the consequences.) -- David Roundy http://civet.berkeley.edu/droundy/

2 1

Dealing with configuration data
by Liyang Hu 11 Oct '02

11 Oct '02

Evening, I'm trying to write a utility that reads in some user preferences from a pre-determined file, does some work, and exits. Sounds simple enough. The problem I'm having is with the preferences: How do I make it available throughout the entire program? (FWIW, most of the work is effectively done inside the IO monad.) I could explicitly pass the record around everywhere, but that seems a trifle inelegant. My current solution is to use a global ('scuse my terminology, I'm not sure that's the right word to use here) variable of type IORef Config obtained through unsafePerformIO. It works, but strikes me as a rather barbaric solution to a seemingly tame enough problem... Intuition tells me I should be able to `embed', if you will, the config record somehow within or alongside the IO state, and retrieve it at will. (Is this what MonadState is for?) However it also tells me that this will /probably/ involve lots of needless lifting and rewriting of the existing code, which makes it even less enticing than passing everything around explicitly. Any opinions or suggestions? Cheers, /Liyang -- .--{ Liyang HU }--{ http://nerv.cx/ }--{ Caius@Cam }--{ ICQ: 39391385 }--. | :: zettai unmei mokusiroku :::: absolute destined apocalypse ::::::::: |

8 17

Re: infinite (fractional) precision
by Ashley Yakeley 10 Oct '02

10 Oct '02

At 2002-10-10 01:29, Ketil Z. Malde wrote: >I realize it's probably far from trivial, e.g. comparing two equal >numbers could easily not terminate, and memory exhaustion would >probably arise in many other cases. I considered doing something very like this for real (computable) numbers, but because I couldn't properly make the type an instance of Eq, I left it. Actually it was worse than that. Suppose I'm adding two numbers, both of which are actually 1, but I don't know that: 1.000000000.... + 0.999999999.... The trouble is, as far as I know with a finite number of digits, the answer might be 1.9999999999937425 or it might be 2.0000000000013565 ...so I can't actually generate any digits at all. So I can't even make the type an instance of my Additive class. Not very useful... -- Ashley Yakeley, Seattle WA

6 6

infinite (fractional) precision
by ketil＠ii.uib.no 10 Oct '02

10 Oct '02

Hi I was just browsing around on comp.arch a bit, and there was this discussion about various ways to represent non-integer numeric values. It seems one could easily (I'll get back to that in a moment) calculate the fractional part of numbers lazily, generating the needed precision, and nothing more. Does any such implementation exist in Haskell? I realize it's probably far from trivial, e.g. comparing two equal numbers could easily not terminate, and memory exhaustion would probably arise in many other cases. So, if such implementations exist, how does one deal with these issues? Is it indeed possible to make this useful in practice? (And if it hasn't been implemented, would anybody be interested if I gave it a try?) -kzm -- If I haven't seen further, it is by standing in the footprints of giants

3 2

need help optimizing a function
by David Roundy 09 Oct '02

09 Oct '02

Hello. If you followed my previous post (which was about a bug in my lcs function), this is still the same code (only now bug-free, it seems). It turns out that my lcs diff function is MUCH slower than GNU diff on some test files (much worse than 1000 time slower!).The test case where its speed was close was IO bound on the output to an xterm (silly me). I've profiled my lcs code, and it is spending about 95% of its time in the "hunt_one_char" function. It also claims to spend 0.0% of its time in my_bs, which does a binary search on the array, and is the only other function that gets called by hunt_one_char (not counting itself). I've told the compiler (ghc) to inline hunt_one_char and my_bs to no avail. I guess maybe it can't or won't inline a recursive function, or maybe that just doesn't gain me anything. My only guess is that for some reason it is making an entirely new copy of the array each time it recurses, but that doesn't make any sense, since the array is getting threaded through, so it should be able to just modify the array in place. But then, I'm not entirely clear as to what the compiler can and cannot do as it optimizes. >>data Threshold a = Thresh Int! [a] >> deriving (Show) >> >>hunt_one_char :: String -> [Int] -> Array Int (Threshold String) -> >> Array Int (Threshold String) >>hunt_one_char c [] th = th >>hunt_one_char c (j:js) th = >> case my_bs (Thresh j [c]) th of >> Nothing -> hunt_one_char c js th >> Just k -> >> case th!(k-1) of >> Thresh _ rest -> >> hunt_one_char c js th//[(k,Thresh j (c:rest))] For what it's worth, the calling function is: >>hunt_internal :: [String] -> [[Int]] -> >> Array Int (Threshold String) -> >> Array Int (Threshold String) >>hunt_internal [] _ th = th >>hunt_internal _ [] th = th >>hunt_internal (c:cs) (m:ms) th = >> hunt_internal cs ms $ hunt_one_char c m th Each string in this list of strings is a line from one of the files (and the [[Int]] is a list of line numbers in the other file that that line matches). The array 'th' has size (1,n) where n is the number of lines in the file. My simplest test case consists of a 20k line file, each line of which is unique and a second file which is simply a permutation of the first (moving the first line to the last), so hunt_one_char should be called exactly once for each line, and it is always called with its second argument having exactly one element. In this test case, according to the ghc profiler, 89.7% of the time is spent in hunt_one_char: individual inherited COST CENTRE MODULE entries %time %alloc %time %alloc hunt_internal Main 20001 0.2 0.0 91.4 95.3 hunt_one_char Main 40000 89.7 95.0 91.1 95.3 my_bs Main 20000 0.2 0.0 1.4 0.3 my_helper_bs Main 307231 1.2 0.3 1.2 0.3 This has me at a loss. It seems like such a simple function... of course, it is called 20k times, which could certainly add up, but while each call of my_bs should take O(log 20k) time, each call of hunt_one_char (apart from its one call to my_bs) should only take O(1), so most of the time should be spent in my_bs unless something is terribly wrong (which must be the case here). If it would help, I'd be happy to send a listing of the complete code, but it's 6.5k so I figured it'd be better not to send it, since it seems unlikely that anyone would want to run it anyways. -- David Roundy http://civet.berkeley.edu/droundy/

5 5

Data structure definitions
by Mark T.B. Carroll 09 Oct '02

09 Oct '02

I have a program that basically has, data Expression = Value Value | EVariable Variable | other stuff ... data Value = VNumber Number | other stuff ... data Variable = Variable { variable_name :: String, variable_time :: Expression } data Number = Number { value :: Double, dimension :: Dimension } newtype VariableCount = VariableCount (Variable, Number) The VNumber and EVariable constructors are ugly, though, so I was wondering if I should be using typeclasses - e.g., class Expression a class Expression a => Value a instance Value Number instance Expression Variable ... but I don't see how to define Variable in such a scheme. Maybe I shouldn't be using typeclasses? (Obviously, I actually have lots more type constructors in Expression and Value - dyadic expressions, booleans, etc. - the above with just numbers and variables is somewhat truncated, but should suffice.) -- Mark

3 3