February 2013 - Haskell-Cafe

FunPtr to C function with #arguments determined at runtime?
by Ryan Newton 19 Feb '13

19 Feb '13

Hello cafe, I've been poking around and I haven't seen this addressed anywhere except obliquely in the end of section 8.5.1 of the report, where it says that variable argument C functions aren't supported: http://www.haskell.org/onlinereport/haskell2010/haskellch8.html The scenario is pretty simple. I generate C code at runtime. I compile it to a .so. I know how many arguments it expects (but only at runtime), and I get a FunPtr back from 'dlsym'. How do I call it? I was hoping there would be some "Raw" lower level FFI layer that I could use to invoke a C function without automagic marshaling and all the other goodness provided by the normal "foreign import" mechanism. Failing that, will I just have to generate complex wrappers on the C side which I call N times to load up N arguments into some stateful container before finally launching the function? Thanks, -Ryan

5 6

adding recursion to a DSL
by fritsch＠joerg.cc 19 Feb '13

19 Feb '13

I have a tiny DSL that actually works quite well. When I say import language.CWMWL main = runCWMWL $ do out (matrixMult, A, 1, row, matrix-row) then runCWMWL is a function that is exported by language.CWMWL. This parses the experession and takes some action. Now, A is the name of the matrix and the third tuple element would represent the numbe of the row. For example 1 to 10000. I want to achieve some sort of "elegant" (means readable code, a good representation) recursion that would let me do something like sequence [ out (matrixMult, A, n, row, matrix-row) | n <- [1..1000] ] but in a nicer manner an without expending this to 10000 lines of code at compile time. How can I best introduce recursion into my DSL or borrow this from the host language Haskell "effectively"? --Joerg

3 2

Aeson + MongoDB, how to effortlessly store and retrieve json?
by Alfredo Di Napoli 19 Feb '13

19 Feb '13

Good evening guys, suppose I write a very simple parser using Aeson with these types and ToJSON / FromJSON instances: https://github.com/cakesolutions/the-pragmatic-haskeller/blob/master/01-jso… https://github.com/cakesolutions/the-pragmatic-haskeller/blob/master/01-jso… My question is simple: is it possible to "automagically" store the haskell data structure produced from the aeson encoding using http://hackage.haskell.org/packages/archive/mongoDB/1.3.2/doc/html/Database… Any tip/suggestion is welcome, Thanks, Alfredo

2 2

~ operator ?
by briand＠aracnet.com 18 Feb '13

18 Feb '13

Hi all, I was creating "bigger" uncurries which I am simply extending from an existing uncurry I found some where, e.g. uncurry4 :: (a -> b -> c -> d -> e) -> ((a, b, c, d) -> e) uncurry4 f ~(a,b,c,d) = f a b c d when I realized, what's the "~" for ? I've only been able to find a partial explanation that it involves preserving laziness, or something, maybe ? I was hoping someone could enlighten me. Thanks Brian

3 3

generalized, tail-recursive left fold that can finish tne computation prematurely
by Petr Pudlák 18 Feb '13

18 Feb '13

Dear Haskellers, while playing with folds and trying to implement `!!` by folding, I came to the conclusion that: - `foldr` is unsuitable because it counts the elements from the end, while `!!` needs counting from the start (and it's not tail recursive). - `foldl` is also unsuitable, because it always traverses the whole list. I came up with the following tail-recursive generalization of `foldl` that allows exiting the computation prematurely: foldlE :: (a -> c) -> (a -> b -> Either c a) -> Either c a -> [b] -> c foldlE f g = fld where fld (Left c) _ = c fld (Right a) [] = f a fld (Right a) (x:xs) = fld (g a x) xs `foldl` can be defined from it as foldl'' :: (a -> b -> a) -> a -> [b] -> a foldl'' f z = foldlE id ((Right .) . f) (Right z) and `!!` as: -- Checks for a negative index omitted for brevity. index :: Int -> [a] -> a index i = foldlE (error $ "No such index") f (Right i) where f 0 x = Left x f n _ = Right (n - 1) Is something like that already available somewhere? Best regards, Petr Pudlak

4 5

generalized, tail-recursive left fold that can finish tne computation prematurely
by Petr Pudlák 18 Feb '13

18 Feb '13

Dear Haskellers, while playing with folds and trying to implement `!!` by folding, I came to the conclusion that: - `foldr` is unsuitable because it counts the elements from the end, while `!!` needs counting from the start (and it's not tail recursive). - `foldl` is also unsuitable, because it always traverses the whole list. I came up with the following tail-recursive generalization of `foldl` that allows exiting the computation prematurely: foldlE :: (a -> c) -> (a -> b -> Either c a) -> Either c a -> [b] -> c foldlE f g = fld where fld (Left c) _ = c fld (Right a) [] = f a fld (Right a) (x:xs) = fld (g a x) xs `foldl` can be defined from it as foldl'' :: (a -> b -> a) -> a -> [b] -> a foldl'' f z = foldlE id ((Right .) . f) (Right z) and `!!` as: -- Checks for a negative index omitted for brevity. index :: Int -> [a] -> a index i = foldlE (error $ "No such index") f (Right i) where f 0 x = Left x f n _ = Right (n - 1) Is something like that already available somewhere? Best regards, Petr Pudlak

1 0

JSON querying
by Sergey Mironov 18 Feb '13

18 Feb '13

Hi folks. Hackage contains several JSON packages but as far as I see, they all provide 'static' conversion from JSON format to Haskell data type. Is there a method of converting object containing optional filed 'a' to for example Maybe a. Thanks, Sergey

3 3

Re: [Haskell-cafe] ifdef based on which OS you're on
by Doug McIlroy 17 Feb '13

17 Feb '13

With apologies for prolonging a tangential topic, I'd like to sharpen anti-ifdef comments that have been posted already. 1. An ifdef for portability is an admission of nonportability. What it does is point out a nonportability--a useful crutch for maintainers, but a crutch nonetheless. 2. Ifdefs violate program structure. All ifdefs appear at top level. Yet, save for those that control top-level declarations, they embody conditionals at inner levels of structure. Doug McIlroy

1 0

Terminal Win32/Building Haskeline/8.3 file issue
by Matthew Lamari 17 Feb '13

17 Feb '13

Hit this compiling darcs on a new machine. . . . But it may be more generic to Haskell platform itself. Bottom line, it presents a situation where someone could end up dead-in-the-water with respect to using haskell libraries. I burned a lot of time tracking this down, it may discourage users from the platform. The problem manifests itself when your Win32 home directory has a long name (or one with a space) and you're compiling haskeline, linking on the 15th step. Basically, ld.exe croaks when fed one of these names rather than an 8.3 I can provide more specific details on the point of failure on request; but it was a pain to track down. Maybe a change to the linker, or some specific project, may get rid of the 8.3 dependency. A solution that got me out was to enable 8.3 on the volume, and on the user directory. A fix: Logging in as a *different* user: Open a command prompt with Administrator priveleges fsutil 8dot3name query c: (tells you the current status - you probably see a 2 - volume-select of the feature) fsutil 8dot3name set c: 1 (This sets volume c to create 8.3 names - for future files only) fsutil file setshortname "c:\Users\YouHaveA VeryBigName" YOUHAV~1 (This only works if you're not logged in as 'YouHaveA VeryBigName', otherwise the directory will be in-use). You need this step as even though you've changed the policy on 8.3 naming, this directory was probably made during the prior policy) dir /-n (This shows you the 8.3-sensitive names. You see YOUHAV~1 in the list, you got it right) The user may need to uninstall and reinstall haskell-platform, and wipe all Cabal and GHC directories under AppData/Roaming

1 0

MonadReader laws
by Roman Cheplyaka 17 Feb '13

17 Feb '13

Has anyone formulated a reasonable set of laws for MonadReader? The instances for continuation-based monad transformers contradict any intuition about local and ask. E.g. for LogicT: > flip runReader 1 $ observeAllT $ local (const 2) ask [1] > flip runReader 1 $ observeAllT $ local (const 2) mzero <|> ask [2] I wonder whether there are any real use cases for such instances (and, consequently, whether they should be removed). Roman

1 0