- Beginners - Haskell.org

monad nomad gonad gomad
by prad 18 Aug '10

18 Aug '10

ok so taking MAN's suggestion: "I think It's time for you to get serious with the monads" that's just what i'm going to do! i found the following references: You Could Have Invented Monads! http://blog.sigfpe.com/2006/08/you-could-have-invented-monads-and.html an interesting way to become familiar with the idea Monads for the Working Haskell Programmer http://www.engr.mun.ca/~theo/Misc/haskell_and_monads.htm seems to have several practical ideas there (though i don't understand them yet) A tour of the Haskell Monad functions http://members.chello.nl/hjgtuyl/tourdemonad.html a nice reference with brief explanations Explaining Haskell IO without Monads http://neilmitchell.blogspot.com/2010/01/haskell-io-without-monads.html possibly a good conceptual aid for background All About Monads http://www.haskell.org/all_about_monads/html/index.html this seems to have everything though it'll take some work to dig in in any case, monads seem to be a rather important concept to getting anything done in haskell. for instance, i have a program which is generating the output it is supposed to (i can print it), but i can't seem to get it into another function and keep getting the error i've seen so many times: Couldn't match expected type `[String]' against inferred type `IO [String]' so it's time to understand them. besides, the stuff looks rather intriguing and certainly appears to take one into computer language design theory. any sources others have found useful would be appreciated as well as suggestions on how to proceed through the above. in friendship, prad -- In friendship, prad ... with you on your journey Towards Freedom http://www.towardsfreedom.com (website) Information, Inspiration, Imagination - truly a site for soaring I's

6 15

hClose: illegal operation (handle is finalized) and/or <<loop>>
by Travis Erdman 18 Aug '10

18 Aug '10

> Perhaps in your {BOOLEANS} you are referring to > 'recurseChildren', which in turn needs the {BOOLEANS}? Uggh! Indirectly, yes, I am. OK, back to the drawing board for me on this one. Thanks!

1 0

hClose: illegal operation (handle is finalized) and/or <<loop>>
by Travis Erdman 18 Aug '10

18 Aug '10

> > putStrLn $ show result > > > > <<loop>> > > That is the *real* error message. > > It means that where you define result, you refer to result > also on the right-hand side in a way that is an endless > loop instead of well-behaved recursion. > > We would have to see more of your code to be able to > say much more. Thank you very much ... now I have isolated the offending "loop" code: I am folding with a function that takes a tree and the input data, and delivers a "processed"/"updated" tree. The function that is used in the fold is "calcTree". calcTree is also recursive ... it updates the root node, then recursively calls itself to update each of its children, which is all working fine. Here are simplified versions of my structure: data Tree a = Node a [Tree a] calcTree (Node oldData oldsubtree) input = Node newData recurseChildren where {-calculate newData from oldData and input-} recurseChildren = if {BOOLEANS} then oldsubtree else map (\old -> calcTree old input) oldsubtree And, in words: if certain conditions become true, there will be no change in any decedent node from the current node (ie all newData = oldData, and tree structure never changes), so rather than waste exponential time traversing all the decedent nodes, I'd like to just re-use the ones I already have, as they are identical. This is "pruning" the tree. If I remove this pruning, the "<<LOOP>>" objection ceases and the code executes as expected ... except it takes lots longer than it should, because there are usually lots of oppt'ys for pruning, and I'd rather not explicitly traverse all of them. How can I make this work? Thanks again for all the A's to my Q's !

3 2

Man/dir/help pages for haskell?
by Kevin Burke 17 Aug '10

17 Aug '10

Hello, I've been learning Haskell by working through the O'Reilly book, but a lot of the coding I do is offline. One thing that's been really helpful in other languages are the terminal instructions and detail for functions. Is there something similar for Haskell? I know I can type :help at the prompt but that only gives you info on the broad commands. Thanks for your help, Kevin

4 4

hClose: illegal operation (handle is finalized) and/or <<loop>>
by Travis Erdman 17 Aug '10

17 Aug '10

I read in some data from a file (via decodeFile), process it, and write the result back to a file. My last line of code is .... encodeFile (filename++".sol") (gametype,gstree,newiterlist,newstamplist,result) and it is all of the sudden now causing this run-time error: mincfr: kclv5.sol: hClose: illegal operation (handle is finalized) It has been working fine for quite a while, but now quite mysteriously gives this error. (But obviously I have been heavily tinkering with the processing part of the program). If I remove the encodeFile line and instead ... putStrLn $ show result I get this run-time error: <<loop>> Help? thanks, Travis

2 1

very detailed monad tutorials
by Michael Vanier 17 Aug '10

17 Aug '10

Hi everyone, I've been writing a bunch of in-depth monad tutorials that I thought some of you might find useful. So far there are six, with at least three more planned. Here are the links: Part 1 (Basics): http://mvanier.livejournal.com/3917.html Part 2 (>>= and return): http://mvanier.livejournal.com/4305.html Part 3 (The monad laws): http://mvanier.livejournal.com/4586.html Part 4 (The Maybe and list monads): http://mvanier.livejournal.com/4647.html Part 5 (Error-handling monads, part 1): http://mvanier.livejournal.com/5103.html Part 6 (Error-handling monads, part 2): http://mvanier.livejournal.com/5343.html The next ones will deal with state monads. I would love to get some feedback on these. Cheers, Mike

1 0

global variables
by prad 16 Aug '10

16 Aug '10

i'm curious about the best way to deal with conn to make db connections. i define it in main like this, then when i want to use it in a function i have to pass it as a parameter: main = do conn <- connectPostgreSQL ... fn conn something fn conn x = ... now conn is a global variable i guess because it is defined in main, but perhaps main isn't considered a global namespace? i recall in python a global variable could be accessed by functions, but in php they can't unless you use "global" within the function. so is this the correct way to handle such things? right now it seems to be the only way in haskell. -- In friendship, prad ... with you on your journey Towards Freedom http://www.towardsfreedom.com (website) Information, Inspiration, Imagination - truly a site for soaring I's

6 7

= vs <-
by prad 15 Aug '10

15 Aug '10

i'm using them right i think because it works, but my understanding is fuzzy. it seems <- is used when you do things like load a file or get arguments from outside or if you are return something from a function these seem to take the form of something other than an internal type like Int or String and have to be 'massaged' into use with things like show or fromSql etc the = is used with let and seems to work for any assignments that are internal. are these assessments correct? -- In friendship, prad ... with you on your journey Towards Freedom http://www.towardsfreedom.com (website) Information, Inspiration, Imagination - truly a site for soaring I's

4 6

leaky folding
by Travis Erdman 15 Aug '10

15 Aug '10

in the code below, finalmap seems fast enough ... but it has a space leak. otoh, finalmap'rnf runs in constant space, but its performance is terrible, at least 4x slower than finalmap. this is a common problem i'm having ... foldl' isn't strict enough, but foldl'rnf kills performance. and not only with IntMap as the cumulating data structure, but others as well. any ideas on this one? how can i get a fast fold in constant space? thanks again, travis {-# LANGUAGE BangPatterns #-} import System.Environment import Foreign (unsafePerformIO) import System.Random.Mersenne import Data.List import Control.DeepSeq import Control.Parallel.Strategies import qualified Data.IntMap as IntMap mersennegen = unsafePerformIO $ newMTGen Nothing infrandoms = unfoldr ( Just . splitAt 3) $ map (\x -> abs (x `mod` n)) (unsafePerformIO $ (randoms mersennegen)::[Int]) n = 200 foldl'rnf :: NFData a => (a -> b -> a) -> a -> [b] -> a foldl'rnf f z xs = lgo z xs where lgo z [] = z lgo !z (x:xs) = lgo (runEval (rdeepseq (f z x))) xs startmap = IntMap.fromDistinctAscList $ zip [0..] [1..n] finalmap x = foldl' g startmap (take x infrandoms) finalmap'rnf x = foldl'rnf g startmap (take x infrandoms) g:: IntMap.IntMap Int -> [Int] -> IntMap.IntMap Int g !a [x,y,z] = IntMap.adjust (const $ y + (a IntMap.! z) `mod` n) x a main = do args <- getArgs print $ finalmap (read $ head args)

2 1

how to access command line arguments
by prad 15 Aug '10

15 Aug '10

here's one way: import System (getArgs) args <- getArgs first = head args last = last args and so on. i can also do this (first:last:z) <- getArgs and avoid using head and tail. in fact, this seems nicer because i can pattern match for a specific series of inputs. however, if the inputs aren't there (eg say program arg1 only instead of program arg1 arg2), then the runtime pattern match failure error: user error (Pattern match failure in do expression at Tests.hs:14:4-14) is understandably generated. this of course doesn't happen if i just use args because there is no pattern matching. is the former the better way to do it then? or are there other alternatives? -- In friendship, prad ... with you on your journey Towards Freedom http://www.towardsfreedom.com (website) Information, Inspiration, Imagination - truly a site for soaring I's

3 2