October 2011 - Haskell-Cafe

Hackage feature request: E-mail author when a package breaks
by Gregory Crosswhite 02 Nov '11

02 Nov '11

Hey everyone, I have uploaded a number of small packages to Hackage that I no longer actively use so that I don't find out immediately when a new version of GHC has broken them. Since Hackage is going to the trouble of finding out when a package no longer builds anyway, could it have a feature where when a working package breaks with a new version of GHC the author is automatically e-mailed? This would make me (and probably others) a lot more likely to notice and proactively fix broken packages. (Heck, I wouldn't even necessarily mind being nagged about it from time to time. :-) ) Cheers, Greg

16 24

Re: [Haskell-cafe] How to implement digital filters using Arrows
by Captain Freako 01 Nov '11

01 Nov '11

Hi John, Thanks for all your help. I've been studying your suggested code: > type FilterAu b c = Automaton (->) b c > liftAu :: ((x,FilterState s)->(y,FilterState s)) -> FilterState s -> FilterAu x y > liftAu f s0 = proc x -> do > rec (y,s') <- arr f -< (x,s) > s <- delay s0 -< s' > returnA -< y runAutomaton is a bit cumbersome, so define a custom run function that takes a list > runAuto a [] = [] > runAuto (Automaton f) (x:xs) = let > (y,a) = f xt > in y:runAuto a xs as well as the various instance definitions for Automaton. I think I understand how the `returnA' in the last line of your `liftAu' function is getting translated by those instance definitions into: c where c = Automaton ( arr id &&& arr (const c) ) and, furthermore, how that is passing the supplied `y' into the first element of the resulting couple. However, I don't understand how the recursively defined `c' is capturing the modified filter state and preserving it for the next call. It seems like the Automaton being referred to by `c' is a newly constructed entity, which knows nothing about the current state of the running Automaton. Any help in understanding this would be greatly appreciated. Thanks! -db

3 7

Organizing big repository
by Konstantin Litvinenko 01 Nov '11

01 Nov '11

I am trying to understand how to organize my code and edit-compile-run cycles. I can't figure out how to setup environment in such why that when I build some program using cabal, cabal will rebuild program dependencies if some was changed. I don't want to configure/build/install manually. Having program 'foo' depends on lib 'bar' I want to edit some files in 'bar' than build 'foo' and get 'bar' rebuilt and 'foo' rebuilt/relink. How can I do this?

5 5

lazy A-star search
by Anton Kholomiov 31 Oct '11

31 Oct '11

Recently I was looking for an A-star search algorithm. I've found a package but I couldn't understand the code. Then I saw some blogposts but they were difficult to understand too. I thought about some easier solution that relies on laziness. And I've come to this: Heuristic search is like depth-first search but solutions in sub-trees are concatenated with mergeBy function, that concatenates two list by specific order: module Search where import Control.Applicative import Data.Function(on) import Control.Arrow(second) import Data.Tree -- | Heuristic search. Nodes are visited from smaller to greater. searchBy :: (a -> a -> Ordering) -> Tree a -> [a] searchBy heur (Node v ts) = v : foldr (mergeBy heur) [] (searchBy heur <$> ts) -- | Merge two lists. Elements concatenated in specified order. mergeBy :: (a -> a -> Ordering) -> [a] -> [a] -> [a] mergeBy _ a [] = a mergeBy _ [] b = b mergeBy p (a:as) (b:bs) | a `p` b == LT = a : mergeBy p as (b:bs) | otherwise = b : mergeBy p bs (a:as) Now we can define specific heuristic search in terms of searchBy: -- | Heuristic is distance to goal. bestFirst :: Ord h => (a -> h) -> (a -> [a]) -> a -> [a] bestFirst dist alts = searchBy (compare `on` dist) . unfoldTree (\a -> (a, alts a)) -- | A-star search. -- Heuristic is estimated length of whole path. astar :: (Ord h, Num h) => (a -> h) -> (a -> [(a, h)]) -> a -> [a] astar dist alts s0 = fmap fst $ searchBy (compare `on` astarDist) $ unfoldTree gen (s0, 0) where astarDist (a, d) = dist a + d gen (a, d) = d `seq` ((a, d), second (+d) <$> alts a) I'm wondering is it effective enough? Anton

5 9

ANN: Monad.Reader Issue 19
by Brent Yorgey 31 Oct '11

31 Oct '11

I am pleased to announce that Issue 19 of The Monad.Reader, a special issue on parallelism and concurrency, is now available: http://themonadreader.files.wordpress.com/2011/10/issue19.pdf Issue 19 consists of the following three articles: * Mighttpd – a High Performance Web Server in Haskell by Kazu Yamamoto * High Performance Haskell with MPI by Bernie Pope and Dmitry Astapov * Coroutine Pipelines by Mario Blažević Feel free to browse the source files. You can check out the entire repository using darcs: darcs get http://code.haskell.org/~byorgey/TMR/Issue19 If you’d like to write something for Issue 20, please get in touch. The deadline will likely be in December; more details will be forthcoming.

4 3

Re: [Haskell-cafe] Amazon AWS storage best to use with Haskell?
by dokondr 31 Oct '11

31 Oct '11

On Mon, Oct 31, 2011 at 6:50 PM, John Lenz <lenz(a)math.uic.edu> wrote: > 4) My code processes hundreds of messages. Every message is processed in > >> exactly the same way as the others. So the code can be easily >> parallelized. Any Haskell frameworks that will allow me to run this code >> in a simple concurrency model? >> >> > Yes, there are many options. > > http://www.haskell.org/**haskellwiki/GHC/Concurrency<http://www.haskell.org/haskellwiki/GHC/Concurrency> > > > John, thanks for detailed reply! I am looking at Haskell Concurrency wiki, but can not figure out which framework - STM, sparks, threads, etc. Amazon AWS will be able to scale? As far as I know, to scale CPU and program memory in Amazon, all you can ask from AWS is to start some number of additional extra VMs. Every VM contains a complete image of your OS and executables, all images are exactly the same. That's fine with me as currently all my workflow tasks are performed by separately compiled Haskell executables communicating via regular files. So to reformulate my question: - Does any Haskell framework exist that allow to orchestrate separate processes (NOT threads that share the same process memory)? On the other hand, in case there is a way to make Amazon AWS to scale Haskell STM, sparks, threads, etc. - I would happily rewrite my Haskell code to use these frameworks. So my second question: - Is there a way to make Amazon AWS to scale Haskell STM, sparks, threads or any other Haskell concurrency frameworks?

1 0

Figures 6 and 7 from `Programming with Arrows'?
by Captain Freako 31 Oct '11

31 Oct '11

Does anyone have figures 6 and 7, as well as the intervening unnumbered figure, from `Programming with Arrows'? Thanks, -db

1 0

is Haskell missing a non-instantiating polymorphic case?
by Adam Megacz 31 Oct '11

31 Oct '11

The title might be a bit more provocative than necessary. I'm starting to suspect that there are very useful aspects of the parametricity of System F(C) which can't be taken advantage of by Haskell in its current state. To put it briefly, case-matching on a value of type (forall n . T n) forces one to instantiate the "n", even though the branch taken within the case cannot depend on "n" (parametricity). I came up with the simplest example I could and posted it to StackOverflow, but there haven't been any successes (despite some excellent attempts!): http://stackoverflow.com/questions/7720108/ This might sound like an obscure problem, but it actually starts to get in the way of the very useful Washburn+Weirich "Boxes Go Bananas" and related PHOAS representation. I've written up a short example of the problems that happen here: The link above also includes a very rough sketch of a proposal for a "pcase" construct that doesn't instantiate its argument. Is it possible to do what I'm attempting without adding "pcase" to the language? Would "pcase" break the soundness of the type system (I don't think so) or complicate type inference (probably)? I'd be interested in any comments/thoughts/help people can offer. Thanks! - a PS, I suspect that this is somehow related to the issue that Guillemette and Monnier mention in section 5.1 of their paper on type-preserving closure conversion, although I wasn't able to find the code online in order to be sure.

4 6

Amazon AWS storage best to use with Haskell?
by dokondr 31 Oct '11

31 Oct '11

Hi, Please share your experience / ideas on AWS storage most friendly to Haskell. So far I store my data mostly in Data.Map structures serialized to text files with write / read functions. Now I was requested to move my app and data to Amazon cloud. As far as I know there are two main storage types that Amazon provides: S3 - basic block storage and SimpleDB ( http://docs.amazonwebservices.com/AmazonSimpleDB/latest/GettingStartedGuide/ ) Questions: 1) I would like to continue working with my data using abstractions similar to the ones that Data.Map provides. Any ideas how to iterate and modify SimpleDB records in a similar powerful way as provided by Data.Map? Or maybe S3? 2) It would be great to do development and testing offline without actually connecting to AWS S3 / SimpleDB. Are there any AWS simulators + Haskell libraries that will allow to do such an offline development? 3) Any experience / ideas with Haskell libs for NoSQL, not AWS-native, storages, that will run well both offline and in AWS? 4) My code processes hundreds of messages. Every message is processed in exactly the same way as the others. So the code can be easily parallelized. Any Haskell frameworks that will allow me to run this code in a simple concurrency model? Thanks! Dmitri

1 0

Efficient mutable arrays in STM
by Ben Franksen 30 Oct '11

30 Oct '11

I have an application in mind where concurrent access to large arrays (up to millions of elements) of mostly small elements (Int or Double) is common. Typical access patterns would be chunk-wise, i.e. reads or writes from index n up to index m. Imagine stuff like images, scientific data, etc. All this suggests that Control.Concurrent.STM.TArray, in its current implementation, is not appropriate. Quoting the docs: "It is currently implemented as Array ix (TVar e), but it may be replaced by a more efficient implementation in the future (the interface will remain the same, however)." An array of TVars is certainly *much* too inefficient for what I have in mind w.r.t. both memory and cpu time. In fact I had already decided to use Data.Vector.Unboxed from the vector package. I see that Data.Vector.Unboxed.Mutable provides slice :: Unbox a => Int -> Int -> MVector s a -> MVector s a Yield a part of the mutable vector without copying it. which is almost what I need... Can I use this together with unsafeIOToSTM internally inside a library to provide shared transactional access to an IOArray? The docs warn that using unsafeIOToSTM is (obviously) "highly dangerous", but for what I have in mind the listed problems are not an issue: * running the IO code multiple times is ok * aborting is ok, too * inconsistent views are ok, too The main question is: does the STM transaction actually "see" that I changed part of the underlying array, so that the transaction gets re-tried? Or do I have to implement this manually, and if yes: how? Has anyone done something like that before? (If you tried this and found it doesn't work, please tell me, it would save me a lot of work repeating the effort). Is someone working on providing a more efficient version of TArray? Would it help if I said I'd be a happy user of a better TArray? ;-) If what I sketched above is infeasible (I would not be surprised if it was, I haven't yet spent much effort trying it out), what other options do I have? Is there an internal API for the STM stuff, i.e. a C header file or something which would make it possible to add efficient TArrays? Or should I use a high-level approach, something like a Data.Sequence.Seq of medium sized chunks (TVar (IOVector e))? Any comments are highly appreciated! Cheers Ben

8 11