June 2010 - Beginners

Testing a data structure, etc
by matthew coolbeth 19 Jun '10

19 Jun '10

As an excercise, I have created a small module defining a data structure which is intended to be an efficient map from sequences (possibly very long ones) to objects of some other type. The source code is below for anyone interested. Three questions. Hopefully they aren't too naive: 1) Can anyone point me toward whatever standard modules would be most appropriate for timing this (or measuring other kinds of performance) against the naive Data.Map? 2) Any suggestions on how to improve the module itself are more than welcome. 3) In a language like this with meaningful white-space, how do people keep their lines of source code from getting too wide? There is one line below that stretched beyond the desired 80 characters and, in general, I seem to have a hard time avoiding this. Thanks for your help. -- mac -- A map from [a] to p module SequenceMap (SequenceMap, empty, insert, lookup, test) where import Prelude hiding (lookup) import qualified Data.Map import Data.Maybe data Ord a => SequenceMap a p = SequenceMap (Maybe p) (Data.Map.Map a (SequenceMap a p)) empty :: Ord a => SequenceMap a p empty = SequenceMap Nothing Data.Map.empty lookup :: Ord a => [a] -> SequenceMap a p -> Maybe p lookup [] (SequenceMap p m) = p lookup (a:as) (SequenceMap p m) = let msm = Data.Map.lookup a m in case msm of Nothing -> Nothing Just sm -> lookup as sm insert :: Ord a => [a] -> p -> SequenceMap a p -> SequenceMap a p insert [] p (SequenceMap oldP othrs) = SequenceMap (Just p) othrs insert (a:as) p (SequenceMap notP othrs) = let msm = Data.Map.lookup a othrs in case msm of Nothing -> SequenceMap notP (Data.Map.insert a (insert as p empty) Data.Map.empty) Just sm -> SequenceMap notP (Data.Map.insert a (insert as p sm) othrs) test :: Bool test = fromMaybe False (lookup "qwerty" (insert "qwerty" True empty))

4 3

Haskore lib
by Luca Ciciriello 19 Jun '10

19 Jun '10

Hi All. Is there someone using the library Haskore to handle music in Haskell? I'm reading some haskore manuals and now I'm a bit confused about the compositional operators (:+:) and (:=:). These operators are mentioned everywhere in the manuals, but browsing the library code and trying some examples I just found the operators (+:+) and (=:=). Which is the difference between (:+:) and (+:+) and between (:=:) and (=:=)? Thanks in advance. Luca.

3 5

course of action
by prad 19 Jun '10

19 Jun '10

greetings! i am very intrigued with haskell having acquired some experience in pascal, elisp, c, perl, mostly python and some sql. i've been going through various tutorials and the online real world haskell. i've also been reading through the haskell98 report at appropriate points. are there any recommended approaches for learning and studying 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

3 3

Accounting Engine in Haskell
by Amiruddin Nagri 17 Jun '10

17 Jun '10

My current project is about making an accounting engine that handles all the journal entries, transactions, portfolios etc. The communication with the engine is based on simple protocol, the things to be taken care of in the order are consistency, handling large data(performance) and availability. I came across a video lecture by Simon Peyton Jones where he gives an example from Financial domain (derivatives etc) to explain how haskell is being used and the advantages provided. I am interested in knowing if Haskell will be the right fit for my project. My requirements are transactional nature, which I believe is one of the strengths of functional programming, also handling large data set and being available. there is no such requirement for partitioning of data and the application is going to be centrally hosted on a single server. AFAIK OCaml and other functional languages are heavily used in financial domain, some of the reason are same as features I am looking for. I wanted some insight as to how Haskell is going to help me with my project. Also there has been some concerns because of lazy evaluation in Haskell and memory leaks associated with it. http://jlouisramblings.blogspot.com/2010/04/haskell-vs-erlang-for-bittorent… Also, if you have any suggestions of the choice of programming language, we have been looking into other functional languages like Scala and Clojure. But we have not dig deep on the performance aspects of these languages, if someone can shed a light on the pros- cons of these languages, it will help us very much to come to a decision. -Regards, Amir

7 10

Haskell and Niches
by Jordan Cooper 16 Jun '10

16 Jun '10

The recent mailings about using Haskell for an accounting system got me thinking, since people have pointed out the things that would make the language suited for the task. Does this means there are some programming tasks that Haskell is great for, and others that you would pick a different language for? I was under the impression that it was more a "general purpose" language than a "niche" language--I'm using it for programming a decently ambitious (though small) game now, and it's bending my brain but I suspect it is fully possible even though games are traditionally imperatively programmed. Do you try to use Haskell for every project? Is there some sort of "right tool for the right job" adage you follow?

3 3

strange behavior
by Luca Ciciriello 16 Jun '10

16 Jun '10

Hi All. I'm experimenting a strange behavior. I've wrote in Haskell a tool to automatic generate C++ code. All works fine if I use GHCi or compile with GHC on WIndows, but I see the strange behavior if I use my program on MacOS X 10.6.4 In example, the portion of code here below createConstructor :: [String] -> String -> [String] createConstructor xs classname = ["\t\tpublic: " ++ classname ++ "()\n\t\t{"] ++ ["\t\t\t" ++ (snd (makePair x)) ++ " = nullptr;" | x <- xs] ++ ["\t\t}\n"] on Windows generate: public: DAT_CDM_Enum() { fileName = nullptr; filePath = nullptr; fileType = nullptr; } but on MacOS X 10.6.4 I have: public: DAT_CDM_Enum() { = nullptr; fileName = nullptr; filePath fileType = nullptr; } This is a very weird behavior to me. I'm using Haskell platform (ghc 6.12.1) on MacOS X and on Windows. Any idea? Luca

2 2

Extract and integer from a ByteString
by Tom Hobbs 16 Jun '10

16 Jun '10

Hi all, Can anyone help me with this problem I'm been banging my head against for a while? I have a stream of bytes that I want to read, the stream is formatted thus; null, null, null, 4, 'd', 'a', 't', 'a'<end of stream> Where the first four bytes tell me the number of bytes I should read next in order to get some string value. What I have is code similar to the following; readFromStream address port = do h <- connectTo address (PortNumber port) hSetBuffering h NoBuffering L.hPut h (encode (0xFAB10000 :: Word32)) p <- L.hGet h 4 readData h (extractInt((L.unpack p))) extractInt = foldl addDigit 0 where addDigit num d = 10*num + d readData h c = do print c s <- L.hGet h c print s (Note: extractInt is largely copied from a Stack Overflow answer to something else.) The problem with this code in GHCI is; Couldn't match expected type `Int' against inferred type `Word8' In the second argument of `readData', namely `(extractInt ((L.unpack p)))' In a stmt of a 'do' expression: readData h (extractInt ((L.unpack p))) In the expression: do { h <- connectTo a (PortNumber p); hSetBuffering h NoBuffering; L.hPut h (encode (4205903872 :: Word32)); p <- L.hGet h 8; .... } I read this as saying that readData was expecting an Int, but instead extractInt returned it a Word8. How do I fix extractInt to return an Int rather than a Word8? Or is my problem something else? Some other things that bother me is my repeating use of 'do'. How can I go about breaking these functions out of being monadic and instead being pure? Is that something I should be striving for, because otherwise I can see every function I write here being in a 'do' block. Also, the final line of readFromStream bothers me. I feel like I'm starting to violate some similar form of the Lay of Demeter. (spelling?) I don't like writing code that looks like "f1 (f2 (f3 (f4 (f5 e))))" even if that is what is officially going on, as a programmer I'd rather see named variables. How can I fix that inside of a 'do' block? Sorry, this is a bit of a brain dump, but I would really appreciate someone else's opinion. Many thanks, Tom -- Using Opera's revolutionary e-mail client: http://www.opera.com/mail/

5 7

solutions
by Andy Larocque 15 Jun '10

15 Jun '10

I'm working my way thru Simon Thompson's book "The Craft of Functional Programming" and would really like to find solutions to the exercises in chapters 14,15,16 . I have no idea if any of my own solutions are even close. Does anyone have them? AL

2 1

Cabal issues
by Anand Patil 15 Jun '10

15 Jun '10

Hi all, I'm trying to install hmatrix-static using Cabal. I installed the Haskell Platform for Mac yesterday from hackage.haskell.org/platform. I've been having build problems, and one of them is with haskell-src-exts. The output of cabal install haskell-src-exts -v ends in: Building haskell-src-exts-1.9.0... Building library... Creating dist/build (and its parents) /usr/bin/ghc --make -package-name haskell-src-exts-1.9.0 -hide-all-packages -fbuilding-cabal-package -i -idist/build -isrc -idist/build/autogen -Idist/build/autogen -Idist/build -optP-DBASE4 -optP-include -optPdist/build/autogen/cabal_macros.h -odir dist/build -hidir dist/build -stubdir dist/build -package-id array-0.3.0.0-307c95c424058c7b6bd041b809e52ef3 -package-id base-4.2.0.0-0120a3650a6376da8d1b6a4e48a2973a -package-id cpphs-1.11-20b9349b60ba9e187ce20bacef8124a5 -package-id pretty-1.0.1.1-2bfa34c699efff7eed4d6297a2dac12d -O Language.Haskell.Exts Language.Haskell.Exts.Parser Language.Haskell.Exts.Pretty Language.Haskell.Exts.Syntax Language.Haskell.Exts.Extension Language.Haskell.Exts.Build Language.Haskell.Exts.Fixity Language.Haskell.Exts.Comments Language.Haskell.Exts.SrcLoc Language.Haskell.Exts.Annotated Language.Haskell.Exts.Annotated.Syntax Language.Haskell.Exts.Annotated.Fixity Language.Haskell.Exts.Annotated.Build Language.Haskell.Exts.Annotated.ExactPrint Language.Haskell.Exts.Annotated.Simplify Language.Haskell.Exts.ExtScheme Language.Haskell.Exts.ParseMonad Language.Haskell.Exts.ParseSyntax Language.Haskell.Exts.Lexer Language.Haskell.Exts.ParseUtils Language.Haskell.Exts.InternalParser <command line>: cannot satisfy -package-id cpphs-1.11-20b9349b60ba9e187ce20bacef8124a5: cpphs-1.11-20b9349b60ba9e187ce20bacef8124a5 is unusable due to missing or recursive dependencies: haskell98-1.0.1.1-0fdaf3b26bc38c43ce8371edf538dbf6 (use -v for more information) cabal: Error: some packages failed to install: haskell-src-exts-1.9.0 failed during the building phase. The exception was: ExitFailure 1 However, cabal install cpphs --reinstall works fine. Any help getting past this would be much appreciated. Thanks, Anand

2 1

type ambiguity confusion
by Thomas 15 Jun '10

15 Jun '10

Hello all! This is a simplified example: import Control.Exception testHandle :: Integer -> IO (Maybe Integer) testHandle i = handle (\_ -> return Nothing) (return (Just i)) It gives me an "ambiguous type variable" arising from the use of "handle". Ok, I understand that the type variable is ambiguous. And in fact I can resolve this with either wrapping it in a specialised handler: import Control.Exception handleIO :: (IOException -> IO a) -> IO a -> IO a handleIO ec nc = handle ec nc testHandle :: Integer -> IO (Maybe Integer) testHandle i = handleIO (\_ -> return Nothing) (return (Just i)) or resolving it directly: {-# LANGUAGE ScopedTypeVariables #-} import Control.Exception testHandle :: Integer -> IO (Maybe Integer) testHandle i = handle (\ (_ :: IOException) -> return Nothing) (return (Just i)) However: Why does the ambiguity matter in the first place? The corresponding value is never used (\_ -> return Nothing). And there's another doubt: I got this example from a book (Real World Haskell, chapter 9) where actually the first version is used, without resolving the ambiguity in any way. And it does not show up in the errata list. So there's likely something I'm still missing about this issue?!?! Any hints would be really appreciated! Thanks in advance. Thomas

3 2