- Haskell-Cafe - Haskell.org

Partial statical linking
by Jason Dusek 26 Sep '12

26 Sep '12

Some time ago, I wrote to this list about making shared libraries with GHC, in such a way that the RTS was linked and ready to go. Recently, I've been looking a similar but, in a sense, opposite problem: linking Haskell executables with some of their non-Haskell dependencies, for distribution. I tried passing a few different sets of options to the linker through GHC, with -optl: -optl'-Wl,-r' -optl'-Wl,-r,-dy' -optl'-Wl,-static,-lffi,-lgmp,-dy' None of these had the desired effect. In the end, running GHC with -v and carefully editing the linker line produced the desired change (I have linked to and provided the diff below). The effect -optl seems to be to introduce options in the linker line just before -lHSrtsmain, which would seem to prevent one from linking libffi and libgmp differently. Is editing and storing away the linker script the best option at present for partially static linking? -- Jason Dusek () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments https://github.com/solidsnack/arx/commit/90ec5efdb0e991344aa9a4ad29456d466e… #@@ -122,10 +122,8 @@ # -lHSarray-0.3.0.2 \ # -lHSbase-4.3.1.0 \ # -lHSinteger-gmp-0.2.0.3 \ #- -lgmp \ # -lHSghc-prim-0.2.0.0 \ # -lHSrts \ #- -lffi \ # -lm \ # -lrt \ # -ldl \ #@@ -136,4 +134,7 @@ # -lgcc_s --no-as-needed \ # /usr/lib/gcc/x86_64-linux-gnu/4.6.1/crtend.o \ # /usr/lib/gcc/x86_64-linux-gnu/4.6.1/../../../x86_64-linux-gnu/crtn.o \ #+ -static \ #+ -lgmp \ #+ -lffi \

5 10

Re: [Haskell-cafe] Church vs Boehm-Berarducci encoding of Lists
by oleg＠okmij.org 26 Sep '12

26 Sep '12

Dan Doel wrote: > >> P.S. It is actually possible to write zip function using Boehm-Berarducci > >> encoding: > >> http://okmij.org/ftp/Algorithms.html#zip-folds > > If you do, you might want to consider not using the above method, as I > seem to recall it doing an undesirable amount of extra work (repeated > O(n) tail). It is correct. The Boehm-Berarducci web page discusses at some extent the general inefficiency of the encoding, the need for repeated reflections and reifications for some (but not all) operations. That is why arithmetic on Church numerals is generally a bad idea. A much better encoding of numerals is what I call P-numerals http://okmij.org/ftp/Computation/lambda-calc.html#p-numerals It turns out, I have re-discovered them after Michel Parigot (so my name P-numerals is actually meaningful). Not only they are faster; one can _syntactically_ prove that PRED . SUCC is the identity. The general idea of course is Goedel's recursor R. R b a 0 = a R b a (Succ n) = b n (R b a n) which easily generalizes to lists. The enclosed code shows the list encoding that has constant-time cons, head, tail and trivially expressible fold and zip. Kim-Ee Yeoh wrote: > So properly speaking, tail and pred for Church-encoded lists and nats > are trial-and-error affairs. But the point is they need not be if we > use B-B encoding, which looks _exactly_ the same, except one gets a > citation link to a systematic procedure. > > So it looks like you're trying to set the record straight on who actually > did what. Exactly. Incidentally, there is more than one way to build a predecessor of Church numerals. Kleene's solution is not the only one. Many years ago I was thinking on this problem and designed a different predecessor: excerpted from http://okmij.org/ftp/Haskell/LC_neg.lhs One ad hoc way of defining a predecessor of a positive numeral predp cn+1 ==> cn is to represent "predp cn" as "cn f v" where f and v are so chosen that (f z) acts as if z == v then c0 else (succ z) We know that z can be either a numeral cn or a special value v. All Church numerals have a property that (cn combI) is combI: the identity combinator is a fixpoint of every numeral. Therefore, ((cn combI) (succ cn)) reduces to (succ cn). We only need to choose the value v in such a way that ((v I) (succ v)) yields c0. > predp = eval $ > c ^ c > # (z ^ (z # combI # (succ # z))) -- function f(z) > # (a ^ x ^ c0) -- value v {-# LANGUAGE Rank2Types #-} -- List represented with R newtype R x = R{unR :: forall w. -- b (x -> R x -> w -> w) -- a -> w -- result -> w} nil :: R x nil = R (\b a -> a) -- constant type cons :: x -> R x -> R x cons x r = R(\b a -> b x r (unR r b a)) -- constant time rhead :: R x -> x rhead (R fr) = fr (\x _ _ -> x) (error "head of the empty list") -- constant time rtail :: R x -> R x rtail (R fr) = fr (\_ r _ -> r) (error "tail of the empty list") -- fold rfold :: (x -> w -> w) -> w -> R x -> w rfold f z (R fr) = fr (\x _ w -> f x w) z -- zip is expressed via fold rzipWith :: (x -> y -> z) -> R x -> R y -> R z rzipWith f r1 r2 = rfold f' z r1 r2 where f' x tD = \r2 -> cons (f x (rhead r2)) (tD (rtail r2)) z = \_ -> nil -- tests toR :: [a] -> R a toR = foldr cons nil toL :: R a -> [a] toL = rfold (:) [] l1 = toR [1..10] l2 = toR "abcde" t1 = toL $ rtail l2 -- "bcde" t2 = toL $ rzipWith (,) l2 l1 -- [('a',1),('b',2),('c',3),('d',4),('e',5)]

4 6

Help me with space leaks
by Alexey Uimanov 25 Sep '12

25 Sep '12

Hello. I am trying to write some thing in haskell and i need fast storage to store and select this things from storage. https://github.com/s9gf4ult/projs/tree/master/haskell/teststorage I am writing simple testing package to determine my needs and select the fastest storage and i have encountered problems. Firstly i decided to check out postgresql-simple, but when i am doing "executeMany" i see space leaks, here is the picture http://bayimg.com/NAdjHAaeA http://bayimg.com/NADjKAAea I dont fully understand what is happening here, but i belive this is because of lazy consuming of "executeMany" or something. So things im trying to insert do not calculate one by one, but this is creating many thunks for calculate them. I would understand how to narrow the cause of this problem and create more strict function "executeMany" which would work in constant space (or space of data). Here the insertMany from postgresql-simple executeMany :: (ToRow q) => Connection -> Query -> [q] -> IO Int64 executeMany _ _ [] = return 0 executeMany conn q qs = do result <- exec conn =<< formatMany conn q qs finishExecute conn q result And this is not looks like a problem case go to formatMany formatMany :: (ToRow q) => Connection -> Query -> [q] -> IO ByteString formatMany _ q [] = fmtError "no rows supplied" q [] formatMany conn q@(Query template) qs = do case parseTemplate template of Just (before, qbits, after) -> do bs <- mapM (buildQuery conn q qbits . toRow) qs return . toByteString . mconcat $ fromByteString before : intersperse (fromChar ',') bs ++ [fromByteString after] Nothing -> fmtError "syntax error in multi-row template" q [] Here bs is the map of bs and must stay lazy evaluated and here we see mconcat which must be strict i think, and if i am right so problem must disapear whan i replace it with strict mconcat mconcat' :: (Monod a) => [a] -> a mconcat' [] = mempty mconcat' (x:xs) = x `seq` (mappend x $ mconcat' xs) but it doesnt. I assume i must write the same for "deepseq" to realy calculate each parameter up to value. But if so, this must touch several projects such as blaze and maybe bytestring because of the need to make NFData instances. How to solve this problem right - way ? Maybe haskell have elegant solution for this ?

1 0

How to take a minimum sub list that only contain certain number of elements of certain type?
by Magicloud Magiclouds 25 Sep '12

25 Sep '12

Hi, For example, I have an array [0..]. Now I want to take a sub list that starts from index 0, and only contain 4 odds, and is minimum result. The answer should be [0, 1, 2, 3, 4, 5, 6, 7]. How to do that? Combining lazy computing, I cannot figure out an efficient algorithm. -- 竹密岂妨流水过山高哪阻野云飞 And for G+, please use magiclouds#gmail.com.

7 11

Re: [Haskell-cafe] [Haskell] ANN: HERMIT for GHC 7.6
by Jay Sulzberger 25 Sep '12

25 Sep '12

[Note that original post was on the Haskell list, and not the Haskell-cafe list.] On Tue, 25 Sep 2012, Andrew Farmer <afarmer(a)ittc.ku.edu> wrote: > A new version of HERMIT (0.1.2.0) that is compatible with GHC 7.6 is now > available on hackage. This release has essentially the same functionality > as the 7.4-only pre-alpha release made before ICFP 2012. > > HERMIT (Haskell Equational Reasoning Model-to-Implementation Tunnel) is a > plugin for GHC that provides an interactive interface for applying > transformations directly to GHC's internal intermediate language, Core. > This plugin is part of a larger HERMIT toolkit, which is being developed at > the University of Kansas with the aims of supporting equational reasoning > and allowing custom optimizations to be applied without modifying either > GHC or the Haskell source code. > > Introduction to HERMIT via Neil Sculthorpe's Haskell Symposium 2012 talk: > > http://www.youtube.com/watch?v=x2QH3jJCJso > > Example transformations can be found in the examples folder, contained in > the cabal source package. > > Hackage: http://hackage.haskell.org/package/hermit > Github (for bug reports): https://github.com/ku-fpg/hermit > > Andrew Farmer Thank you and all the HERMIT Team for this! Does the HERMIT system provide an executable Haskell Core, that is, we can get an ordinary text file which can be executed by some version of GHC+HERMIT? oo--JS.

1 0

Re: [Haskell-cafe] [Haskell] Well-Typed and Skills Matter offer Haskell courses in London in October
by Claude Heiland-Allen 25 Sep '12

25 Sep '12

On 19/09/12 08:52, Andres Löh wrote: > Registration for all these events is open. I hope to see many > of you there. Is there an informal hangout without the £225 price-tag > > Cheers, > Andres >

3 5

in the vector library, why define unbox tuples > 2?
by Jeff Shaw 25 Sep '12

25 Sep '12

I'm looking at storing a data type with 7 fields in an unboxed vector, which means that I'll have to use GenUnboxTuple to create an instance for Unbox (a,b,c,d,e,f,g), but I was thinking that another solution is to use "instance (Unbox a, Unbox b) => Unbox (a,b)" recursively to create "instance Unbox (a,(b,(c,(d,(e,(f,g))))))". Is there any reason I shouldn't say something like, import qualified Data.Vector.Unboxed as Vector data X = X Int Int Int Int Int Int Int toTuple (X a b c d e f g) = (a,(b,(c,(d,(e,(f,g)))))) toVector = Vector.fromList . map toTuple I guess I'm wondering if there is a technical reason why I should use an explicit Unbox instance for 7-tuples, or if it's just a convenience to have Unbox defined for tuples > 2? Thanks, Jeff

1 0

Fwd: [Haskell] ANNOUNCE: MFlow-0.1.5
by Alberto G. Corona 25 Sep '12

25 Sep '12

Oh, I´m stupid. You mean web pages with multiple tabs.... I have not tested it. but each tab can be handled easily by a different server process.. or it can be handled in a single server process, like in a menu. For example, this code present different options, and the process renders different things depending on the response. The last option is a link to a different process, while the others( wlinks) are links that return back to the same process. The operator <|> is the applicative operator. a breakline is prepended to each link: data Ops= Ints | Strings | Actions | Ajax | Opt deriving(Typeable,Read, Show) mainf= do r <- ask $ wlink Ints (bold << "increase an Int") <|> br ++> wlink Strings (bold << "increase a String") <|> br ++> wlink Actions (bold << "Example of a string widget with an action") <|> br ++> wlink Ajax (bold << "Simple AJAX example") <|> br ++> wlink Opt (bold << "select options") <++ (br +++ linkShop) -- this is an ordinary XHtml link case r of Ints -> clickn 0 Strings -> clicks "1" Actions -> actions 1 Ajax -> ajaxsample Opt -> options mainf where linkShop= toHtml $ hotlink "shop" << "shopping" . Alberto 2012/9/18 Alberto G. Corona <agocorona(a)gmail.com>: > Hi Jake > > I don´t know what you mean with multiple tabs. The user management is > simple, anonymous clients are identified with a cookie. if the user > is logged (MFlow has widgets for logging-validation) the user is the > identifier. > > The state of a process is associated to the client identifier and to > the path invoked in the url requested. > > I don´t know if this answer your question.... > > Alberto > > 2012/9/18 Jake McArthur <jake.mcarthur(a)gmail.com>: >> This sounds really cool. >> >> How do you handle users having multiple tabs? >> >> On Tue, Sep 18, 2012 at 11:26 AM, Alberto G. Corona <agocorona(a)gmail.com> wrote: >>> Hi haskellers and specially the web developers. >>> >>> http://hackage.haskell.org/package/MFlow-0.1.5.3 >>> >>> MFlow is a is a Web framework with some unique, and I mean unique, >>> characteristics that I find exciting: >>> >>> - It is a Web application server that start and restart on-demand >>> stateful web server processes (not request.-response) >>> This means that all the page navigation can be coded in a single >>> procedure. This increases readability of the programmer code. I woul >>> call it >>> a anti-node.js. Buit usual request-response (stateless) server >>> processes are also allowed >>> >>> - When the process is invoqued as result of an URL request, the Web >>> app server not only restart the process but also recover its execution >>> state. The enclosing Workflow monad provides the thread state >>> persistence. There are state timeouts and process timeouts defined by >>> the programmer. Processes with no persistent state (transient) are >>> possible. >>> >>> -The user interface is made of widgets. They are formlets with added >>> formatting, attributes, validations, modifiers and callbacks, that >>> are composable, so the pieces are reusable and return type safe >>> responses to the calling process. Even the links are part of widgets >>> and return back type safe inputs at compile time to the calling server >>> process. Tho glue these components, ordinary applicative combinators >>> and other extra combinators are used. >>> >>> - The widgets and the communication don´t make assumptions about the >>> architecture, so it can be adapted to non-web environments. This >>> versions has interface for WAI-warp, Hack, Text.XHtml (xhtml) , and >>> Haskell Server Pages. >>> >>> -The widget rendering can be converted to ByteStrings automatically >>> with special combinators. A mix of widgets with different formats can >>> be combined in the same source file. For example Text.Html and HSP >>> (Haskell server pages) >>> >>> -These widgets can be cached, to avoid widget rendering on every interaction. >>> >>> -To handle the back button, and because the processes are stateful, >>> they can run backwards until the response match. This is transparent >>> for the programmer, thanks to the embedded FlowM monad. >>> >>> -All the code is in pure Haskell. No deployment, special scripts, >>> formats etc are necessary. >>> >>> -Besides automatic state persistence, TCache provides transactions and >>> user data persistence, that can be configured for SQL databases. >>> Default persistence in files permit very rapid prototyping. Just code >>> and run it with runghc. >>> >>> -Has AJAX support >>> >>> All of this sounds very complicated, but really it is simple!. Most of >>> these things are transparent. The resulting code is quite readable and >>> has very little plumbing! >>> >>> There is a non trivial example that some of these functionalities >>> embedded here that you can run: >>> >>> http://hackage.haskell.org/packages/archive/MFlow/0.1.5.3/doc/html/MFlow-Fo… >>> >>> Take a look and tell me your opinion. I hope that you find it as >>> exciting as me. >>> >>> I´m looking for people to collaborate in the development of MFlow. >>> >>> Although still it is experimental, it is being used in at least one >>> future commercial project. So I have te commitment to continue its >>> development. There are many examples in the documentation and in the >>> package. >>> >>> Alberto >>> >>> _______________________________________________ >>> Haskell mailing list >>> Haskell(a)haskell.org >>> http://www.haskell.org/mailman/listinfo/haskell

2 4

Getting started with Shake
by Richard Wallace 24 Sep '12

24 Sep '12

Hey all, I'm going to be working on a webapp in Haskell in the upcoming months and am thinking Shake would be a good fit - I'll need to do js, css, and probably some graphics processing as part of the build and would like to use Shake to automate deployment. I'm not entirely sure how to get started with Shake. I've used to make, jam, maven, sbt and other build tools in the past. All of these keep a build configuration in the project and use that to build the project. In sbt, which is most like Shake, the build configuration is written in Scala and compiled by sbt when you run the build. The result of the compilation is saved and reused unless it detects that the build source files changed. With shake I'm not sure exactly how to get started. Should I have a separate project where I create the build system for the webapp? Or can I setup something similar to sbt? Also, how do I handle dependencies with shake? cabal will pull in packages from hackage and do the needful, is there anything in shake to do the same? If not, how is it normally done? Thanks, Rich

2 1

ANN: New wiki page and mailing list for combinatorics and Haskell
by Brent Yorgey 24 Sep '12

24 Sep '12

Quite a while ago there was a mailing list thread where the idea was brought up of creating a mailing list for those interested in combinatorics in the setting of Haskell. Well, I've finally gotten around to it. If you're interested in (probably extremely low-traffic) discussion and announcements relating to doing combinatorics in Haskell (computing combinatorial/ number-theoretic functions; exhaustive or random generation of combinatorial objects; type algebra; etc.), please join! http://projects.haskell.org/cgi-bin/mailman/listinfo/combinatorics There is also a wiki page for collecting information here: http://www.haskell.org/haskellwiki/Combinatorics -Brent

1 0