July 2006 - Haskell-Cafe

Separate a string into a list of strings
by Sara Kenedy 18 Apr '08

18 Apr '08

Hi all, I want to write a function to separate a string into a list of strings separated by commas. Example: separate :: String -> [String] separate "Haskell, Haskell, and Haskell" = ["Haskell", "Haskell", "and Haskell"] If anyone has some ideas, please share with me. Thanks. S.

14 35

Creating static libraries using GHC
by Tom Hawkins 16 Mar '07

16 Mar '07

I have a chunk of Haskell code I would like wrap up and distribute as a library. Is there a way to build a static library (*.a) that includes my code plus the Haskell runtime, which C programs can easily link against? Here is what I have tried so far... ghc --make -fffi MyLib # Builds MyLib.o and MyLib_stub.o. gcc -c -I/usr/local/lib/ghc-6.4.1/include MyLibWrapper.c ar -r libMyLib.a MyLib.o MyLib_stub.o MyLibWrapper.o This works fine when I use ghc to compile and link a C program... ghc main.c libMyLib.a But if I use gcc, it throws a lot of unresolved references. I have no problem compiling with ghc, but the folks using my library probably won't have it installed. I added -v to ghc to see how it calls gcc; it seems to link in different libraries based on what Haskell libraries are being used, and it undefines a bunch is symbol references. Are there any switches to have ghc return a single archive with everything included? Please excuse my ignorance of the linking process. Thanks for any help! -Tom

3 2

forall and a parse error
by Neil Mitchell 15 Nov '06

15 Nov '06

Hi, I was experimenting with forall and higher rank types briefly, in particular: x :: [forall a . a] This is illegal because of: http://www.haskell.org/ghc/docs/latest/html/users_guide/type-extensions.htm… Which is fine, however its surprising to compare the error messages: [forall a . a] parse error on input `forall' [(forall a . a)] Illegal polymorphic or qualified type: forall a. a In the type signature: lst :: [(forall a. a)] In normal Haskell, I tend to view [x] as equivalent to [(x)] (provided that x is not a tuple) but in this case it has a different meaning - albeit both are erronous meanings. When running the example with Hugs, they both come out as syntax errors - the first on the forall, the second on the closing square bracket. Thanks Neil

5 6

Computing lazy and strict list operations at the same time
by C Rodrigues 27 Sep '06

27 Sep '06

Here's a puzzle I haven't been able to solve. Is it possible to write the initlast function? There are functions "init" and "last" that take constant stack space and traverse the list at most once. You can think of traversing the list as deconstructing all the (:) [] constructors in list. init (x:xs) = init' x xs where init' x (y:ys) = x:init' y ys init' _ [] = [] last (x:xs) = last' x xs where last' _ (y:ys) = last' y ys last' x [] = x Now, is there a way to write initlast :: [a] -> ([a], a) that returns the result of init and the result of last, takes constant stack space, and traverses the list only once? Calling reverse traverses the list again. I couldn't think of a way to do it, but I couldn't figure out why it would be impossible.

10 14

Why does Haskell have the if-then-else syntax?
by Mike Gunter 16 Aug '06

16 Aug '06

I had hoped the "History of Haskell" paper would answer a question I've pondered for some time: why does Haskell have the if-then-else syntax? The paper doesn't address this. What's the story? thanks, -m

18 21

REALLY simple STRef examples
by Chad Scherrer 05 Aug '06

05 Aug '06

I've looked around at the various STRef examples out there, but still nothing I write myself using this will work. I'm trying to figure out how the s is escaping in really simple examples like x = runST $ return 1 y = runST $ do {r <- newSTRef 1; readSTRef r} Neither of these works in ghci - they both say <interactive>:1:0: Inferred type is less polymorphic than expected Quantified type variable `s' escapes Expected type: ST s a -> b Inferred type: (forall s1. ST s1 a) -> a In the first argument of `($)', namely `runST' In the definition of `it': ... I thought maybe I needed to replace 1 with (1 :: Int) so the state representation didn't force the type, but it still gives the same result. Can someone point me to the simplest possible runST example that actually works? Thanks! -- Chad Scherrer "Time flies like an arrow; fruit flies like a banana" -- Groucho Marx

2 2

RE: ANN: System.FilePath 0.9
by Brian Smith 04 Aug '06

04 Aug '06

[ I tried to send this through the fa.haskell USENET group but apparently it did not go through] I kind of expect that a Haskell library for file paths will use the type system to ensure some useful properties about the paths. For example, when writing security-conscious software I often want to be able to distinguish between absolute, ascending (paths with leading "../" and similar items), and decending paths (paths that contain no "../"). I want to make sure a filename is valid. For example, "prn" and "con" are not valid path elements for normal files on Windows, certain characters are not allowed in filenames (depending on platform), some platforms may require paths to be escaped in different ways. I see there is a "isValid" function and even a (magical) "makeValid" function, but they do not report what was wrong with the filename in the first place. Furthermore, it isn't clear from the documentation how these functions determine whether a filename is valid. Many people might be familiar with Joel Spolsky's article "Making Wrong Code Look Wrong" at http://www.joelonsoftware.com/articles/Wrong.html . If you read that article, you will see that advocates a variable naming convenctions to distinguish between HTML-escaped "safe" and non-escaped "unsafe" strings. In Haskell, we can do much better than that by using convenient strong static typing to enforce such constraints. We should be able to use type signatures to ensure that we are operating on syntactically valid paths of a certain form. For example, a function's type should indicate that it expects an absolute path, or a relative path that doesn't use "../../.." to decend into parent/sibling directories unexpectedly, which is a common exploit technique. IMO, safety is the most important issue regarding file paths and it is not addressed in this library as far as I can see. Writing code to handle these issues is tedious, error-prone, and boring to write despite being critical. It isn't the kind of code that you want to just download off of some guy's webpage. Basically, it is exactly the type of thing that belongs in a standard library. In this library proposal, there are a bunch of "xxxDrive" functions that many Unix-oriented programmers are sure to ignore because they are no-ops on Unixy systems. Even on Windows, they are not very useful: Windows software almost never cares what drive a directory is on, especially because the directory might not be on a drive at all (for example, a UNC network path "\\foo\bar" has no "drive" component but it is an absolute path). Really, Windows software usually works similarly to Unixy software in this regard--the big difference is that Unixy systems have one root path, whereas Windows systems can have many, many root paths (drives and UNC's). - Brian

20 68

[Parsec] Backtracking with try does not work for me?
by Stephane Bortzmeyer 02 Aug '06

02 Aug '06

I'm trying to write a Parsec parser for a language which authorizes (this is a simplified example) "a" or "a,b,c" or "a,c" or "a,b". (I can change the grammar but not the language.) The first attempt was: ***** CUT HERE **** import Text.ParserCombinators.Parsec import System (getArgs) comma = char ',' minilang = do char 'a' optional (do {comma ; char 'b'}) optional (do {comma ; char 'c'}) eof return "OK" run parser input = case (parse parser "" input) of Left err -> putStr ("parse error at " ++ (show err) ++ "\n") Right x -> putStr (x ++ "\n") main = do args <- getArgs run minilang (head args) ***** CUT HERE **** Of course, it fails for "a,c": parse error at (line 1, column 3): unexpected "c" expecting "b" for a reason explained in Parsec's documentation (the parser "optional (do {comma ; char 'b'})" already consumed the input, do note the column number). What puzzles me is that the solution suggested in Parsec's documentation does not work either: ********* CUT HERE ******************* minilang = do char 'a' try (optional (do {comma ; char 'b'})) optional (do {comma ; char 'c'}) eof return "OK" ********* CUT HERE ******************* parse error at (line 1, column 2): unexpected "c" expecting "b" Apparently, "try" was used (do note that the column number indicates that there was backtracking) but the parser still fails for "a,c". Why?

5 10

The difficulty of designing a sequence class
by Brian Hulley 01 Aug '06

01 Aug '06

Hi - Part 1 of 2 - Monoid versus MonadPlus =========================== I've just run into a troublesome question when trying to design a sequence class: class ISeq c a | c -> a where empty :: c single :: a -> c append :: c -> c -> c However I've noticed that people sometimes separate the empty and append operations out as sequences with these ops form a Monoid therefore: class Monoid c => ISeq c a | c -> a where single :: a -> c -- now outside the class append :: ISeq c a => c -> c -> c append = mappend empty :: ISeq c a => c empty = mempty Another option, is the Edison library which uses: class (Functor s, MonadPlus s) => Sequence s where so here MonadPlus is used instead of Monoid to provide empty and append. So I've got three main questions: 1) Did Edison choose MonadPlus just because this fitted in with the lack of multi-parameter typeclasses in H98? 2) Are there any reasons to prefer the Edison design over the MPTC design (apart from H98 compatibility) or vice versa? 3) Is it worth bothering to derive ISeq from Monoid (with the possible extra inefficiency of the indirection through the definitions for append = mappend etc or does the compiler completely optimize this out)? and a fourth more long term question: 4) Would it be worth reconsidering the rules for top level names so that class methods could always be local to their class (ditto for value constructors and field names being local to their type constructor). For example it would be nice imho to be able to write: class Monoid c => ISeq c a | c -> a where length :: c -> Int f x y = Monoid/append x y -- or ISeq/append x y g x = ISeq/length x instead of having all names collide in the top level of a module, since at the moment it is difficult to think of method names that don't collide with the Prelude, and it's not nice to have to write "mempty" in place of "empty". Part 2 of 2 - Monad versus Ancillary result type ================================ Another issue relates to left and right views of a sequence. If a sequence is non-empty, the left view is just the leftmost element and the rest of the sequence. The problem arises when the sequence is empty. In the Edison library, this is solved by returning a monadic value ie: lview :: Monad m => s a -> m (a, s a) whereas from the paper "Finger trees: a simple general purpose data structure" by Ralf Hinze and Ross Paterson they use an ancillary data type to store the result of a view: data ViewL s a = NilL | ConsL a (s a) viewL :: FingerTree a -> ViewL FingerTree a So my question here is: what's the best choice? I can see that the monadic version has the advantage that it could be used in do notation in cases where you expect the sequence to be non-empty, but has the disadvantage that it treats the empty sequence as something special (resulting in Monad/fail) and an extra indirection to find the components when the sequence is non-empty. Anyway these are my main questions for now - any feedback appreciated ;-) Thanks, Brian. -- Logic empowers us and Love gives us purpose. Yet still phantoms restless for eras long past, congealed in the present in unthought forms, strive mightily unseen to destroy us. http://www.metamilk.com

8 19

Trouble compiling wxhaskell
by David F. Place 31 Jul '06

31 Jul '06

Hi: I'm having trouble compiling wxhaskell 0.9.4 under both ghc 6.4.2 and ghc 6.5. Does anyone know where I should direct my queries? Thanks. -------------------------------- David F. Place mailto:d@vidplace.com

6 5