August 2015 - Haskell-Cafe

Reactive config file generation
by David Turner 20 Aug '15

20 Aug '15

Hi all, I am using Apache Zookeeper as a service directory: I have a bunch of services which announce their presence by making nodes in Zookeeper, so that dependent services can update their configuration to make use of the available services, and stop trying to use services that have died. Zookeeper is a pretty nice fit for this because it supports watching a node for changes, so in theory there is no need to poll Zookeeper periodically. The services that I control work with this just fine, but there are some (e.g. an Nginx reverse-proxy) that are reconfigured using the rather more common approach of updating a file (or files) and then sending the process a signal. I am currently pondering how to make this work without polling, or manually triggering a refresh script, which is how it is currently done. I've never used FRP, but am at least vaguely aware of it and from my high-level understanding it seems like this could be a very good fit. The Zookeeper state is a time-varying value which I want to convert into a time-varying set of files, ideally as declaratively as possible. So, two questions to the FRP congnoscenti out there. Firstly, is it worthwhile to attempt this using FRP at all? I'm sure I could do it by hand too. Secondly, which of the many FRP libraries would you recommend for it? There are loads on the Wiki, seemingly in various states of repair and documentation. Any pointers on how to choose one that's featureful enough, performant enough, and being actively maintained? Thanks in advance, David

3 3

ANN: cabal-bounds 1.0.0
by Daniel Trstenjak 20 Aug '15

20 Aug '15

cabal-bounds[1] is a command line programm for managing the bounds/versions of the dependencies in a cabal file. Changes for 1.0.0 ================= * automatically find the cabal and setup-config files * ignore the base library by default Perhaps the two most relevant use cases: Initialize Bounds ================= If you have started a new project, created a cabal file, added dependencies to it, build it, and now want to set the lower and upper bounds of the dependencies according to the currently used versions of the build, then you can just call: $> cabal-bounds update This call will update the bounds of the dependencies of the cabal file in the working directory. Raise the Upper Bounds ====================== If you have several cabalized projects, then it can be quite time consuming to keep the bounds of your dependencies up to date. Especially if you're following the [package versioning policy](<http://www.haskell.org/haskellwiki/Package_versioning_policy>), then you want to raise your upper bounds from time to time, to allow the building with newer versions of the dependencies. `cabal-bounds` tries to automate this update process to some degree. So a typical update process might look like: # update the version infos of all libraries $> cabal update # drops the upper bound of all dependencies of the cabal file in the working directory $> cabal-bounds drop --upper # create a cabal sandbox for building your project, this ensures that you're really using # the newest available versions of the dependencies, otherwise you would be constraint # to the already installed versions $> cabal sandbox init # build your project $> cabal install # update the upper bound of all dependencies of the cabal file in the working directory $> cabal-bounds update --upper Please consult the README for further informations. [1] https://github.com/dan-t/cabal-bounds

1 0

octree implementation - followup to "How to model outer space for a MUD"
by Michael Litchard 20 Aug '15

20 Aug '15

It looks like Octree is what I want if I can solve a particular problem. First let me articulate what I am looking for. (1) Objects in 3-d space with no spatial extent. Collision is determined by two objects occupying the same point.' (2) Efficient insert-delete-update #2 seems to be the problem. I found this library, thankfully. I believe it's exactly what I am looking for, but for quadtrees. https://github.com/AshleyMoni/QuadTree I'm a little overwhelmed as this is a new data structure for me. I believe if I can grok what is going on in this library, I can take these ideas and extend them to the octree. I'm referring specifically to this module https://github.com/AshleyMoni/QuadTree/blob/master/Data/QuadTree/Internal.hs And these functions setLocation :: Eq a => Location -> a -> QuadTree a -> QuadTree a setLocation = set . atLocation atLocation (Having difficulty cutting and pasting this to mail.) I'd like to be able to visualize how the tree is being compressed, it may help to be able to see what happens to the tree when it's not compressed. Does anyone have a pointer or advice in how to break this library down in parts to help me grok, in a way that will help me extend to octree?

3 3

List of nitpicks
by William Yager 20 Aug '15

20 Aug '15

Hello all, Inspired by the recent thread re. modifying data/type/newtype syntax, I've made a wiki article dedicated to Haskell-related nitpicks. I've added a few entries; one from the thread I just mentioned, and a classic from way back in the 90s (map vs fmap). https://wiki.haskell.org/Nitpicks Hopefully, this page will help us keep track of what people are annoyed about (especially people who aren't already inured to the quirks of Haskell) as well as give us inspiration for future proposals. Cheers, Will

7 7

Problem parallelising simple algorithm
by Jens Blanck 20 Aug '15

20 Aug '15

The following code simply computes the Euler constant by binary splitting. I'm struggling to get any speed-up on multi-core. Why do I get a fair number of fizzled? Why don't I get any real speed-up even if it claims to run tasks in parallel? Could I have exhausted the integer arithmetic unit(s) on my chip (i7-4790)? How would I verify that? The following is the code and typical runs on 1 and 4 cores respectively. --------- import Control.Parallel.Strategies import Control.DeepSeq import Data.Ratio divConq :: (NFData b) => (a -> b) -> a -> (a -> Bool) -> (b -> b -> b) -> (a -> Maybe (a,a)) -> b divConq f arg threshold conquer divide = go arg where go arg = case divide arg of Nothing -> f arg Just (l0,r0) -> conquer l1 r1 `using` strat where l1 = go l0 r1 = go r0 strat x = do r l1; r r1; return x where r | threshold arg = rdeepseq | otherwise = rpar pqCombine :: (Integer, Integer) -> (Integer, Integer) -> (Integer, Integer) pqCombine (pl, ql) (pr, qr) = (pl*qr+pr, ql*qr) pq :: (Integer, Integer) -> (Integer, Integer) pq (a, b) = (\t -> (sum t, last t)) $ scanl1 (*) [b,b-1..a+1] euler :: Integer -> Rational euler n = let (p,q) = divConq pq (0,n) (\(a,b) -> b-a < 10000) pqCombine (\(a,b) -> if b-a > 5 then let m = (a+b+1) `div` 2 in Just ((a,m), (m, b)) else Nothing) in p%q main = print $ euler 320000 `seq` () ---------- > ./BinSplit +RTS -s -N1 () 178,375,880 bytes allocated in the heap 2,452,040 bytes copied during GC 3,222,696 bytes maximum residency (7 sample(s)) 883,040 bytes maximum slop 11 MB total memory in use (2 MB lost due to fragmentation) Tot time (elapsed) Avg pause Max pause Gen 0 333 colls, 0 par 0.004s 0.002s 0.0000s 0.0000s Gen 1 7 colls, 0 par 0.001s 0.001s 0.0001s 0.0003s TASKS: 4 (1 bound, 3 peak workers (3 total), using -N1) SPARKS: 126 (0 converted, 0 overflowed, 0 dud, 0 GC'd, 126 fizzled) INIT time 0.001s ( 0.000s elapsed) MUT time 0.928s ( 0.936s elapsed) GC time 0.005s ( 0.003s elapsed) EXIT time 0.001s ( 0.000s elapsed) Total time 0.935s ( 0.939s elapsed) Alloc rate 192,215,387 bytes per MUT second Productivity 99.4% of total user, 98.9% of total elapsed gc_alloc_block_sync: 0 whitehole_spin: 0 gen[0].sync: 0 gen[1].sync: 0 > ./BinSplit +RTS -s -N4 () 178,727,480 bytes allocated in the heap 3,506,488 bytes copied during GC 3,650,032 bytes maximum residency (7 sample(s)) 934,976 bytes maximum slop 12 MB total memory in use (1 MB lost due to fragmentation) Tot time (elapsed) Avg pause Max pause Gen 0 141 colls, 141 par 0.009s 0.002s 0.0000s 0.0001s Gen 1 7 colls, 6 par 0.003s 0.001s 0.0001s 0.0001s Parallel GC work balance: 38.80% (serial 0%, perfect 100%) TASKS: 10 (1 bound, 9 peak workers (9 total), using -N4) SPARKS: 126 (12 converted, 0 overflowed, 0 dud, 0 GC'd, 114 fizzled) INIT time 0.002s ( 0.002s elapsed) MUT time 2.104s ( 0.946s elapsed) GC time 0.012s ( 0.003s elapsed) EXIT time 0.001s ( 0.000s elapsed) Total time 2.119s ( 0.951s elapsed) Alloc rate 84,946,520 bytes per MUT second Productivity 99.3% of total user, 221.3% of total elapsed gc_alloc_block_sync: 600 whitehole_spin: 0 gen[0].sync: 9 gen[1].sync: 1073

1 1

Haskell developer position in Jerusalem
by Yitzchak Gale 20 Aug '15

20 Aug '15

We have a position open for a Haskell developer. This is an on-site position at our development center in Jerusalem. You will be working on our enterprise customer-tailored content delivery products - see the "Products" tab on our home page - which are implemented (mostly) in Haskell. Search for Haskell on our "Careers" page for a fuller description of the position. Submit your CV to [jobs@suite-sol.com](mailto:jobs@suite-sol.com), but also feel free to contact me if you have any questions. Thanks, Yitz

2 2

haddock complains about unknown symbol in .a file
by Johannes Waldmann 20 Aug '15

20 Aug '15

Dear all, what is happening here? http://hackage.haskell.org/package/satchmo-2.9.9.1/reports/1/log The library is built, but haddock fails with haddock: /home/builder/hackage-server/build-cache/tmp-install/lib/x86_64-linux-ghc-7.8.3/minisat-0.1.2/libHSminisat-0.1.2.a: unknown symbol `_ZGVZN7Minisat6Option13getOptionListEvE7options' Why is haddock complaining, but ghc is not? Any pointers appreciated (to explain and solve the problem - I think there's a work-around of uploading documentation to hackage but that does not feel right) - J.W.

1 0

A type level puzzle
by Tom Ellis 19 Aug '15

19 Aug '15

For the type level gurus: I would like to define types Sum :: [Field] -> * (:::) :: String -> * -> Field used like this Sum '[ "foo" ::: a, "bar" ::: b, "baz" ::: c ] (I hope this makes sense. I'm not quite familiar with the syntax.) such that Sum '[ s1 ::: a1, ... ] unifies with Sum '[ s1 ::: b1, ... ] to give Sum '[ s1 ::: c1, ... ] where c1 is the unification of a1 and b1. If sn is absent from exactly one of the unificands then its type is copied over unchanged. If sn is absent from both then it is absent from the unification. The types should also be invariant under permutation of the list. This is perhaps a bit obscure so I'll explain the application. This is intended for the sumtype equivalent of a record type, so if we have Sum '[ "foo" ::: a, "bar" ::: b ] and Sum '[ "foo" ::: a, "baz" ::: c ] then I take take the sum of these sums and get something of type Sum '[ "foo" ::: a, "bar" ::: b, "baz" ::: c ] If it makes it easier than I am happy for Sum '[ s1 ::: a1, ... ] to unify with forall ak. Sum '[ s1 ::: a1, ..., sk ::: ak ] This is justified by forall b. Either a b being isomorphic to a. Is such a type possible? If so my next question will be about how to type deconstructors of such things, but let's start with a small first step! Thanks, Tom

6 10

Kill forked threads in ghci
by amindfv＠gmail.com 17 Aug '15

17 Aug '15

Is there a way in ghci to kill all running background processes without quitting ghci itself? E.g. if I do > forkIO . forever $ print 5 >> threadDelay 1000000 If I don't have the ThreadId, is there any way for me to stop printing "5"s without killing ghci? tom

2 1

oddsFrom3 function
by Debdut Karmakar 17 Aug '15

17 Aug '15

I am a haskell beginner and wondering how the following function works (in procedure) : oddsFrom3 :: [Integer] oddsFrom3 = 3 : map (+2) oddsFrom3 Thanks for your help. -- A GNU [1] LINUX [2] Patron Links: ------ [1] http://gnu.org [2] http://www.linuxfoundation.org/

3 2