- Haskell-Cafe - Haskell.org

profiling and strictness (was: Re: [Haskell-cafe] An interesting paper from Google)
by Evan Laforge 03 Nov '10

03 Nov '10

> Any time you see something "inexplicable" like lots of time being attributed > to something simple like "get", it means that something isn't strict enough > and "get" is having to force a bunch of lazy evaluations to do its job. > Since you're using State.Strict but lift-ing to get there, I'd first look at > the strictness of the monad you're lift-ing from. (I'm assuming > State.Strict does what the label says, but it's possible that it's not > strict in the way you need; strictness is kinda tricky.) Ahh, thanks, this makes sense to me. The monads involved are an ErrorT and one other strict StateT. When the monads say they are strict, they seem to mean that >>= will force the evaluation of the monadic mechanics of the first argument when the second is demanded. By "monadic mechanics" I mean the work done by >>= itself, so for StateT it binds the result of the first argument with either 'let' or '~(a, s') <-' instead of 'case' or '(a, s') <-'. So in my understanding, a lazy State will build thunks on every >>= and they won't ever be forced until you pull the final value and pass it to a strict function in IO like print or an FFI call, and then they are forced recursively, at which point there is a stack overflow unless it's a short sequence. Talking about strictness seems complicated to me because it's not really about what is actually forced, it's about who is forced by who (i.e., when 'a' is forced then 'b' will be) and whether it's forced in sequence or nested. So a strict State is also not forced until the strict IO function, but it turns into a sequential series of thunks rather than a nested one. If this is accurate, why would anyone want to use the lazy State? Anyway, State doesn't provide the really critical bit of strictness, which is in the actual state being modified. So 'State.modify' will inevitably lead to stack overflow if you don't intersperse 'get's in there. What is needed is both $! on State.put (or define your own strict modify that puts $! on put), and strictness annotation of the fields of the record that is being modified (provided it's a record and not a scalar type). In any case, ErrorT has a strict >>= by necessity, both States have a strict >>=, and all modifys and puts (I think) are strict, as are all the record fields. In any case, a simple 'get' should only have to force the state to whnf, so it shouldn't matter if the fields are huge thunks or not, so really all that should matter is if the state is already in whnf or is a huge nest of thunks. Which a strict >>= should prevent, right? This is right next to a 'lookup' function which is called twice as many times and should be doing more work since it actually looks in a Map, but is credited with less cpu and alloc. A lift . get should turn into -- ErrorT a <- do a <- StateT (\s -> return (s, s)) -- 'get' of the outer StateT return (a, s) -- >>= belongs to inner StateT return (Right a) Anyway, it's sort of a mishmash of semi inlined functions since this is especially nested and hard to understand with transformers, but it looks like a lot of applications of (\a -> (a, s)) type functions, so (,) constructors and a 'Right' constructor, inside of 'case's to make sure this are evaluations and not allocations (and I remember from the STG paper that a constructor inside a case need not even be constructed). So I don't totally understand where the forcing is happening to make credit the 'get' with so much work. BTW, I have a theory about what the '0' might mean... perhaps if a function is inlined away, it still appears in the profile list, but with 0 entries. > Moral of the story: time is accounted to the function that forces > evaluation of lazy thunks, not to the thunks themselves or the function that > created the lazy thunks. (I think the latter is impossible without passing > around a lot of expensive baggage, and in any case doesn't tell you anything > useful; unexpected functions taking a lot of time, on the other hand, tells > you right away that there's excessive laziness in the invocation somewhere > and gives you a starting point to track it down.) Indeed, this is a good point that I hadn't realized in quite that way before, but tracking it down is still the tricky part.

1 1

problems installing latest parsec
by Michael Litchard 03 Nov '10

03 Nov '10

Here's what cabal says I have installed * parsec Synopsis: Monadic parser combinators Latest version available: 3.1.0 Latest version installed: 2.1.0.1 Homepage: http://www.cs.uu.nl/~daan/parsec.html License: BSD3 but here is what happens when I try to upgrade cabal upgrade parsec Resolving dependencies... No packages to be installed. All the requested packages are already installed. If you want to reinstall anyway then use the --reinstall flag. mlitchard@kether:~/projects/perf/autoperf/session_creator$ Any clues as to what I can try next to fix things would be appreciated.

2 2

trying to use Tag Soup - fromAttrib
by Michael Litchard 02 Nov '10

02 Nov '10

I have the following TagOpen [TagOpen "a" [("href","/launchWebForward.do?resourceId=4&policy=0&returnTo=%2FshowWebForwards.do")]] I would like to get the attribute resourceId=4 from that. My understanding is that fromAttrib is the right thing to use. But I'm having difficulty understanding the type signature fromAttrib :: (Show str, Eq str, StringLike str) => str -> Tag str -> str I'm sure if I did, the way to use it would be clear. Could someone help me make sense of this?

5 6

Linker errors when using the Hipmunk library
by Ulrik Rasmussen 02 Nov '10

02 Nov '10

Hello, I have installed the Hipmunk physics library (5.2.0.1) using cabal-install, but whenever I try to compile a program that uses it (e.g. the HipmunkPlayground package), I get the following linker error: > $ cabal install hipmunkplayground > Resolving dependencies... > Configuring HipmunkPlayground-5.2.0... > Preprocessing executables for HipmunkPlayground-5.2.0... > Building HipmunkPlayground-5.2.0... > [1 of 1] Compiling Main ( Playground.hs, dist/build/HipmunkPlayground/HipmunkPlayground-tmp/Main.o ) > Linking dist/build/HipmunkPlayground/HipmunkPlayground ... > /home/ulrik/.cabal/lib/Hipmunk-5.2.0.1/ghc-6.12.3/libHSHipmunk-5.2.0.1.a(Callbacks.o): In function `sr0l_info': > (.text+0x1062): undefined reference to `cpArbiterIsFirstContact' > /home/ulrik/.cabal/lib/Hipmunk-5.2.0.1/ghc-6.12.3/libHSHipmunk-5.2.0.1.a(Callbacks.o): In function `srBn_info': > (.text+0x37f8): undefined reference to `cpArbiterGetShapes' > collect2: ld returned 1 exit status > cabal: Error: some packages failed to install: > HipmunkPlayground-5.2.0 failed during the building phase. The exception was: > ExitFailure 1 The problem does not seem to be with HipmunkPlayground, since loading Hipmunk in ghci and evaluating `initChipmunk' also generates an error: > Prelude> import Physics.Hipmunk > Prelude Physics.Hipmunk> initChipmunk > Loading package StateVar-1.0.0.0 ... linking ... done. > Loading package array-0.3.0.1 ... linking ... done. > Loading package containers-0.3.0.0 ... linking ... done. > Loading package transformers-0.2.2.0 ... linking ... done. > Loading package Hipmunk-5.2.0.1 ... linking ... <interactive>: /home/ulrik/.cabal/lib/Hipmunk-5.2.0.1/ghc-6.12.3/HSHipmunk-5.2.0.1.o: unknown symbol `cpArbiterIsFirstContact' > ghc: unable to load package `Hipmunk-5.2.0.1' Does anyone have a clue as to what might be wrong? I am using Ubuntu 10.10, gcc 4.4.5 and ghc 6.12.3. Thanks :) /Ulrik Rasmussen

2 1

Am I using type families well?
by Yves Parès 02 Nov '10

02 Nov '10

Hello, I'm trying to make a simple monad (built on operational's ProgramT) for resource loading. I have classes featuring type families : {-# LANGUAGE TypeFamilies, FlexibleContexts, GADTs #-} -- | A ResourceId is something that identifies a resource. -- It should be unique for one resource, and should be used to find the location (the path) of the resource, -- possibly by using a configuration datatype class (Ord id) => ResourceId id where type LocOf id type CfgOf id retrieveLoc :: CfgOf id -> id -> LocOf id -- | Class describing a resource of type @rsc@ class (ResourceId (IdOf rsc)) => Resource rsc where type IdOf rsc load :: LocOf (IdOf rsc) -> IO (Maybe rsc) -- ^ Called when a resource needs to be loaded unload :: rsc -> IO () -- ^ Idem for unloading -- | Then, the operations that the loader can perform data EDSL id a where Load :: id -> EDSL id () IsLoaded :: id -> EDSL id Bool Unload :: id -> EDSL id () -- | The loader monad itself type RscLoader rsc m a = ProgramT (EDSL (IdOf rsc)) m a -- | And finally, how to run a loader runLoader :: (Monad m, Resource rsc) => CfgOf (IdOf rsc) -> RscLoader rsc m a -> m a runLoader cfg loader = viewT loader >>= eval M.empty where eval :: (Monad m, Resource rsc) => M.Map (IdOf rsc) rsc -> ProgramViewT (EDSL rsc) m a -> m a eval _ (Return x) = return x eval rscs (instr :>>= k) = case instr of Load id -> do let loc = retrieveLoc cfg id -- open and load from loc will go here viewT (k ()) >>= eval rscs -- -- -- Other cases yet to come... Well, there is no way I can get it type-check. I think I must be misusing the type families (I tried with multi-param typeclasses and functional dependencies, but it ends up to be the same kind of nightmare...).

4 7

Can a fundep force parametricity
by Brandon Moore 02 Nov '10

02 Nov '10

I would like to make a two-parameter type class where some choices of the first type uniquely determine the second type, and other choices of the first type promise to work with any choice of the second type. This code seems a bit suspicious, but is accepted thanks to UndecidableInstances. >{-# LANGUAGE MultiParamTypeClasses, FunctionalDependencies, FlexibleInstances, > UndecidableInstances #-} > >class C a b | a -> b where > update :: b -> a -> a >instance C (Maybe a) a where > update x m = fmap (const x) m > >instance C Int b where > update _ n = n The application is a program transformation problem, where I'm using distinct types for Statements, Expressions, etc, but want a single environment that contains the bindings for variables of each different type for implementing rewriting rules. I'm trying to make a type class for binding things in the environment. Some of the terms, such as Statement, have a type parameter for semantic analysis information, so the Environment type has the type parameter as well. Then the instance a statement variable needs to require that the info type of the statement matches that of the environment, but the operation for binding an expression should be polymorphic in the info type of the environment. I usually have enough type information around that type checking succeeds without the fundep, so it's not currently a big problem. Brandon

2 1

template haskell for typeclass synonyms
by Paolino 02 Nov '10

02 Nov '10

Hello. I'd like to have a template haskell function that take some constraints and a class name and write an empty class from those and relative empty instance to simulate typeclass synonyms. As I've never written TH and couldn't find a easily adaptable code around, I ask here for the code, or some hints on how to arrive there. Thanks paolino

4 7

Re: [Haskell-cafe] [ANNAUNCE] ghcjs-0.1.0 Haskell to Javascript compiler
by Martijn Schrage 02 Nov '10

02 Nov '10

On 01-11-10 22:35, Aaron Gray wrote: > On 1 November 2010 16:04, Martijn Schrage <martijn(a)oblomov.com > <mailto:martijn@oblomov.com>> wrote: > > On 29-10-10 22:20, Aaron Gray wrote: >> ... >> What browser are you using, IE8, IE9, FF and Chrome on Windows >> throw up errors, both locally and to your above code. > That's weird, since http://tryout.oblomov.com/ghcjs/ghcjs.html > works fine for me with Firefox both on Mac and Windows (XP & 7). > There is only a small bug that causes a key event to get lost when > the Haskell JavaScript modules are loaded. > > What error messages do you get from Firefox? > > > Right, FF seems okay now on Vista too. > > Chrome :- > ... It works on Firefox, Chrome and Safari now, both on Windows and Mac. The IE problems seem a bit more subtle, as the script goes into an infinite loop. -- Martijn

1 0

DSL libraries (Was: Map constructor in a DSL)
by Dupont Corentin 02 Nov '10

02 Nov '10

Hello Steffen, can you compile AwesomePrelude? I've got error messages (see below). By the way, I've looked at some DSLs made in Haskell, if I don't mistake there are lots of similarities between them. There similarities could be put in a library to help the implementors of a DSL, more or less like AwesomePrelude. Is there already packages on Hackage dealing with that? Cheers, Corentin AwesomePrelude compilation error: I'm using GHC 6.12.1. > cd tomlokhorst-AwesomePrelude-9819315 > cabal install (...) Warning: Lang.Haskell: Instances of type and data families are not yet supported. Instances of the following families will be filtered out: H Then, when trying with a very simple GATD in GHCI, I've got: <interactive>: HSAwesomePrelude-0.1.0.o: unknown symbol `AwesomePreludezm0zi1zi0_ CompilerziLiftDefinitions_inline_closure' ghc: unable to load package `AwesomePrelude-0.1.0' On Thu, Oct 28, 2010 at 2:02 PM, steffen <steffen.siering(a)googlemail.com>wrote: > I think you would love to have a look at AwesomePrelude[1] or a fork > of AwesomePrelude using associated types[2] > Some more background information by Tom Lokhorst [3][4]. > > [1] http://github.com/tomlokhorst/AwesomePrelude > [2] http://github.com/urso/AwesomePrelude > [3] http://tom.lokhorst.eu/2009/09/deeply-embedded-dsls > [4] http://tom.lokhorst.eu/2010/02/awesomeprelude-presentation-video > > On 28 Okt., 12:09, Dupont Corentin <corentin.dup...@gmail.com> wrote: > > Thank you for your rich responses. > > > > Indeed I think I miss some thinks in my DSL, that would make things > easier > > to deal with lists and first class functions. > > I don't really know what for now. > > Perhaps a List Constructor? Or a constructor on functions like yours > Ryan? > > EAp :: Exp ref (a -> b) -> Exp ref a -> Exp ref b > > It's from which DSL? It is accessible on the net? > > > > Chris suggested me that I can only define the Foldr constructor and > deduce > > Map from it. > > But maybe I have to add a List constructor for that. > > > > But in the suggestions from Ryan and Brandon I don't understand why I > should > > add an extra type parameter and what it is! > > > > Steffen: Wow nice. I'll integrate that ;) > > > > I'm also looking at the Atom's DSL to get inspiration. > > Something I don't understand in it is that it has two languages, on > typed: > > > > data E a where > > VRef :: V a -> E a > > Const :: a -> E a > > Cast :: (NumE a, NumE b) => E a -> E b > > Add :: NumE a => E a -> E a -> E a > > etc. > > > > And, along with it, an untyped counterpart: > > > > -- | An untyped term. > > data UE > > = UVRef UV > > | UConst Const > > | UCast Type UE > > | UAdd UE UE > > etc. > > > > What that for? What's the use of having beautiful GADT if you have to > > maintain an untyped ADT aside?? > > > > Corentin > > > > _______________________________________________ > > Haskell-Cafe mailing list > > Haskell-C...@haskell.orghttp:// > www.haskell.org/mailman/listinfo/haskell-cafe > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe >

1 0

install GHC 7.0.1 RC on OS X?
by Edward Amsden 02 Nov '10

02 Nov '10

I'd like to install the GHC 7.0.1 RC on my mac (Snow Leopard, x86_64), but I don't want it to run over my current GHC 6.12.2 install. If I use the .pkg installer, it doesn't allow me to select the destination, and I worry that it will overwrite my 6.12.2 install. If I try to build from source, that requires me to use my current GHC, which (as I understand) would build for 32 bit. What other options do I have, or am I misunderstanding something? ------ Edward Amsden Undergraduate Computer Science Rochester Institute of Technology

3 2