February 2008 - Haskell-Cafe

Re: static constants -- ideas?
by Don Stewart 01 Mar '08

01 Mar '08

jay: > Don Stewart dons(a)galois.com: > >jay: > >> Don Stewart dons(a)galois.com: > >> >jay: > >> >> I also have constants that are too large to compile. I am resigned to > >> >> loading them from data files--other solutions seem even worse. > >> ... > >> >> Data.Binary eases the irritation somewhat. > >> > > >> >Did you try bytestring literals (and maybe parsing them in-memory with > >> >Data.Binary)? > > I finally squeezed enough time to try it, and it didn't work for me. > > -- > ghc Overflow.hs > [1 of 1] Compiling Overflow ( Overflow.hs, Overflow.o ) Enable optimisations! Compile with ghc -O2. You need this to avoid having a very slow pack call at runtime. > Overflow.hs:8:10:stack overflow: use +RTS -K<size> to increase it > -- > > where Overflow.hs is in the vicinity of 40M and looks like > > -- > {-# LANGUAGE OverloadedStrings #-} > > module Overflow where > > import qualified Data.ByteString.Lazy as S > > bigData :: S.ByteString > bigData = "\0\0\0\0\0\5\67\195\0\0\0\0... > -- > > I didn't compress it, because Codec.Compression.GZip didn't compile for > me. It looked like a library change since 6.6 broke it. Probably you don't have the zlib.h header? Or make sure you have the latest version of zlib from hackage -- it does work. > Is there a handy string escaping function in the libraries somewhere? It > only took a minute to write one, and I spent longer than that looking, > so maybe it's the wrong question.... Surely it's in there somewhere, and > I'm just 2 dum 2 c. The show function?

1 0

to stimulate discussion concerning where Haskell is going .. ...
by Galchin Vasili 01 Mar '08

01 Mar '08

1) Berkeley consternation about the inability of current concurrency to keep up with multicor developments 1) http://www.eetimes.com/news/latest/showArticle.jhtml;jsessionid=CWDB5YTIKPC… http://www.eetimes.com/news/latest/showArticle.jhtml;?articleID=206801229 IMO a gauntlet has been thrown down to the Haskell world here. 2) Sun Microsystem's planned support of hardware transactional memory http://www.theregister.co.uk/2007/08/21/sun_transactional_memory_rock/ I have fished around and collected some Sun papers and slides. If any anybody wants I can post the URLs or send to the interested Regards, Vasys

3 3

ANN: Hoogle 3.1
by Dan Weston 01 Mar '08

01 Mar '08

I have a question about the ranking algorithm of Hoogle 3.1: The top match of hoogle "(a -> b) -> b" is the inexact match Control.Monad.State.Class.gets :: MonadState s m => (s -> a) -> m a (which cannot be made to unify with (a -> b) -> b) instead of Control.Monad.Cont.runCont (undefined :: Cont r a) :: (a -> r) -> r which does. I would have thought all exact matches would come first (after considering all partial applications), especially as the increase in complexity is at most a small constant factor (~ 2 or 3) for considering all possible partial applications (type tails) of known library functions. Dan Neil Mitchell wrote: > Hi, > > I am pleased to announce Hoogle 3.1. Hoogle is a search engine for > Haskell functions, that allows the user to search by name or > approximate type signature. Hoogle can be used without installing > anything, by simply visiting the online interface, or can be installed > locally. > > * Online interface: http://www.haskell.org/hoogle/ > * Cabal release: > http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hoogle-3.1 > > The Cabal release contains a command line version of the Hoogle tool, > which you can install with the standard Cabal commands. > > Hoogle 3.1 has a number of new features. > > Command line version: > > * A proper cabal release, including proper installing of the search database. > * cabal-install compatible > * No longer requires haskell-source-extensions (trhsx) to be installed > > Web interface: > > * Better entity escaping in some bits of the web interface > * Keyword links work again > * Lambdabot suggestions work > * Update to the most recent version of base > * Security fixes, specifically possible code injection bugs > (http://en.wikipedia.org/wiki/Code_injection) - none of which were > ever exploited, or could have been exploited (as far as I can tell) > > I intend to write a more detailed account of the security fixes in a > blog post, which will be available from http://planet.haskell.org/ . > Running a custom version of the web interface is not really supported, > but if you are, please upgrade to the latest darcs version. > > There are still several known bugs in the Hoogle 3.1 release, but > these are more difficult to fix, and will probably be solved in a > forthcoming Hoogle 4.0. > > Thanks > > Neil > _______________________________________________ > Haskell mailing list > Haskell(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell > >

5 10

cabal errors
by Kristofer Buffington 01 Mar '08

01 Mar '08

Hi I installed ghc 6.8 from source and I've been installing packages from hackage. I'm not sure when the problem started, but I've been getting this error trying to install any cabal package.. accept apparently, Cabal itself. Setup.hs:2:0: Warning: Deprecated use of `defaultUserHooks' (imported from Distribution.Simple): Use simpleUserHooks or autoconfUserHooks Setup: /usr/local/lib/Cabal-1.3.6/ghc-6.8.2.20080225/HSCabal-1.3.6.o: unknown symbol `directoryzm1zi0zi0zi0_SystemziDirectory_a9_closure' Setup: exception :: GhcException Any ideas? -- -kb

3 2

Trouble with numbers
by Lloyd Smith 29 Feb '08

29 Feb '08

I have a question about how numeric classes and type checking works. I have two functions that I think should behave the same but don't. -- I want to split n things up using a given list of fractions -- for example >allocate' 100 [1/3,1/3,1/3] -- [33,33,33] allocate' n fs = vs where vs = map (floor . (*n)) fs -- I want to find anything left over eventually I will want to -- return what is unallocated as well but for now allocated -- and unallocated are not used! allocate n fs = vs where vs = map (floor . (*n)) fs allocated = sum vs unallocated = n - allocated When I load these function in the top level everything looks good [1 of 1] Compiling Main ( allocate.hs, interpreted ) Ok, modules loaded: Main.main *Main> allocate' 100 [1/3,1/3,1/3] [33,33,33] *Main> allocate 100 [1/3,1/3,1/3] <interactive>:1:0: Ambiguous type variable `t' in the constraints: `Integral t' arising from a use of `allocate' at <interactive>:1:0-25 `RealFrac t' arising from a use of `allocate' at <interactive>:1:0-25 Probable fix: add a type signature that fixes these type variable(s) *Main> I mixed up my types when finding the allocated and unallocated, but I am not sure why it produces an error when unallocated and allocated are never used? Shouldn't the two functions be compiled down to the same thing? Suggestions on how to do this more elegantly as well as pointers for understanding numeric type classes would be appreciated. TIA Lloyd

3 2

Error I haven't seen before
by Jefferson Heard 29 Feb '08

29 Feb '08

I'm getting this error when I try to render the text contained in a record structure in the Protein module. Does this mean that the thunk that could calculate go_terms is being evaluated for the first time in the program? go_terms is calculated by a Data.Map.! lookup operation. Main: No match in record selector Protein.go_terms

2 2

Wadler's blog post on the expression problem
by Ryan Ingram 29 Feb '08

29 Feb '08

I was reading Philip Wadler's post regarding the expression problem and Wouter Swierstra's paper, "Data Types a la Carte". Philip asks: The instances of the injection function for subtypes are assymmetric. Is there a symmetric solution in Haskell? Here is a solution which works under GHC 6.8.2, heavily inspired by the HList type computation work, available from http://www.okmij.org/ftp/Haskell/types.html#HList There are three keys to this trick, each of which builds on the previous. 1) TypEq: for any two types a, b, there is an instance for TypEq a b HTrue if and only if a == b, and there is an instance for TypEq a b HFalse otherwise. 2) This allows the definition of IsTreeMember which gives HTrue if a type occurs somewhere within a tree and otherwise HFalse. 3) We can then use IsTreeMember as a predicate to determine whether to use the left branch or right branch of a (:+:) constructor. Source code follows: {-# OPTIONS_GHC -fglasgow-exts -fallow-undecidable-instances -fallow-overlapping-instances #-} module Expr where newtype Expr f = In (f (Expr f)) instance Show (f (Expr f)) => Show (Expr f) where showsPrec _ (In x) = showParen True (showString "In " . showsPrec 11 x) out :: Expr f -> f (Expr f) out (In x) = x data (f :+: g) e = Inl (f e) | Inr (g e) deriving Show instance (Functor f, Functor g) => Functor (f :+: g) where fmap h (Inl f) = Inl (fmap h f) fmap h (Inr g) = Inr (fmap h g) class (Functor sub, Functor sup) => (:<:) sub sup where inj :: sub a -> sup a instance TypTree sub sup => (:<:) sub sup where inj = treeInj data Val e = Val Int deriving Show data Add e = Add e e deriving Show data Mul e = Mul e e deriving Show instance Functor Val where fmap f (Val x) = Val x instance Functor Add where fmap f (Add l r) = Add (f l) (f r) instance Functor Mul where fmap f (Mul l r) = Mul (f l) (f r) foldExpr :: Functor f => (f a -> a) -> Expr f -> a foldExpr f (In t) = f (fmap (foldExpr f) t) class Functor f => Eval f where evalA :: f Int -> Int instance Eval Val where evalA (Val x) = x instance Eval Add where evalA (Add x y) = x + y instance Eval Mul where evalA (Mul x y) = x * y instance (Eval f, Eval g) => Eval (f :+: g) where evalA (Inl f) = evalA f evalA (Inr g) = evalA g eval :: Eval f => Expr f -> Int eval expr = foldExpr evalA expr val :: (Val :<: e) => Int -> Expr e val x = In $ inj $ Val x add :: (Add :<: e) => Expr e -> Expr e -> Expr e add x y = In $ inj $ Add x y mul :: (Mul :<: e) => Expr e -> Expr e -> Expr e mul x y = In $ inj $ Mul x y test :: Expr (Val :+: Add) test = In (Inr (Add (val 118) (val 1219))) test2 :: Expr (Add :+: Val) test2 = val 1 test3 :: Expr ((Add :+: Val) :+: Mul) test3 = add (mul (val 1) (val 2)) (val 3) test4 :: Expr (Add :+: (Val :+: Mul)) test4 = add (mul (val 1) (val 2)) (val 3) -- our typtree selection prefers left injection test5 :: Expr ((Val :+: Val) :+: (Val :+: Val)) test5 = val 1 -- -- TypTree. This is basically the same as (:<:). -- I kept it a separate class during development. This also allows the use of -- additional constraints on (:<:) which improve the error messages when you try, -- for example: -- testBroken :: Expr (Mul :+: Add) -- testBroken = val 3 -- class (Functor sub, Functor sup) => TypTree sub sup where treeInj :: sub a -> sup a instance Functor x => TypTree x x where treeInj = id -- -- The magic all happens here -- We use "IsTreeMember" to determine if a type is part of a tree with leaves -- of various types and internal nodes of type (l :+: r). -- class IsTreeMember (sub :: * -> *) (sup :: * -> *) b | sub sup -> b instance TypEq x y b => IsTreeMember x y b instance (IsTreeMember x l bl, IsTreeMember x r br, TypOr bl br b) => IsTreeMember x (l :+: r) b class (Functor sub, Functor l, Functor r) => TypTree' b sub l r where treeInj' :: b -> sub a -> (l :+: r) a -- -- We can then use this result to decide whether to select from the left or the right. -- instance (TypTree x l, Functor r) => TypTree' HTrue x l r where treeInj' _ = Inl . treeInj instance (TypTree x r, Functor l) => TypTree' HFalse x l r where treeInj' _ = Inr . treeInj -- -- Finally, this allows us to select which treeInj' to use based on the -- type passed in. -- instance (IsTreeMember x l b, TypTree' b x l r) => TypTree x (l :+: r) where treeInj = treeInj' (undefined :: b) class TypOr b1 b2 res | b1 b2 -> res instance TypOr HFalse HFalse HFalse instance TypOr HFalse HTrue HTrue instance TypOr HTrue HFalse HTrue instance TypOr HTrue HTrue HTrue -- Type equality, semi-lifted from the hlist paper; this only works in GHC. -- -- You can avoid the reliance on GHC6.8 type equality constraints -- by using TypeCast from the HList library instead. -- -- see http://www.okmij.org/ftp/Haskell/types.html#HList -- for the source of this idea. data HFalse data HTrue class TypEq (x :: * -> *) (y :: * -> *) b | x y -> b instance TypEq x x HTrue instance (b ~ HFalse) => TypEq x y b

1 1

ANNOUNCE: Sessions 2008 2 28 - RFC
by Matthew Sackman 29 Feb '08

29 Feb '08

Hi, So for those of you that watch the hackage what's-new page, you will probably be aware of this. For the last few months, there's been significant work on my implementation of Session Types in Haskell.[0] The reason for this particular announcement is that it should now be at a point where functionality is verging on useful. There are a few things that are still missing which should appear over the next few weeks but if you're at all interested in message-passing as a tool in the box for dealing with concurrency, I'd love to know what you think. To that end, I would *not* encourage you to try to look at the implementation too closely as you may suddenly discover a powerful urge to stick knives in your eyes and/or eat your computer. However, the Tests.hs[1] file should give you a good idea of the kinds of things that are possible and they do cover all the functionality currently implemented. Note that ghci may well take about 5 mins to type check Tests.hs. Or longer if you're still running a BBC B Micro. I would almost certainly not advise anyone to look at the haddock[2]. I really can't quite believe how horrible the type sigs end up looking under haddock. Some of my comments may almost pass the barrier of being useful. Nicely formatted type sigs exist in the source. If however, you are as mad as I, and enjoy pushing the GHC type system to the limit, then you may enjoy looking at the implementation in all its glory. base10 numbers, lists, associative maps and a whole lot of other madness works correctly at the type-level in ghc 6.8.2. Anyway, as the subject says, comments gratefully received. [0] http://hackage.haskell.org/cgi-bin/hackage-scripts/package/sessions-2008.2.… [1] http://wellquite.org/non-blog/sessions-browseable/Control/Concurrent/Sessio… [2] http://wellquite.org/non-blog/sessions-browseable/doc/html/sessions/ [3] http://wellquite.org/non-blog/sessions-browseable/ Matthew -- Matthew Sackman http://www.wellquite.org/

2 1

Best practice for embedding files in a GHC-compiled tool?
by Dave Bayer 29 Feb '08

29 Feb '08

What is the best way to embed an arbitrary file in a Haskell program? I would like to use GHC to compile command-line tools to be used with OS X. I want the tool to be a single file, not a package or a directory, that makes no assumptions about what else is present. For example, it should be able to run as part of an OS X install disk. I want this tool to be "self reproducing" in the sense that one of its options causes it to output its own documentation and source code. I want this data to be stored as a compressed file within the tool binary itself. The distribution model I'm imagining here is where one writes a program anonymously, that isn't hosted anywhere but is passed from user to user because it is useful, and eventually reaches another user who wants to modify the code. Assume that intermediate users will care less about this, and will try to delete anything that they can. That rules out storing the data in a separate file. Think of the M.I.T. game "Core Wars" from the dawn of the computer age. I'm looking for a strategy here that will be evolutionarily successful in a fairly hostile environment. In other words, I want to be able to write in Haskell, without losing the automatic distribution of source code enjoyed by interpreted languages. No one deletes the source code for a Perl script, because by doing so they're killing the program. There must be some library I'm overlooking that would make this very easy. All I can think of is to use Template Haskell to read a compressed tar file into a Haskell variable. Is that what one does, or is there a better way? Thanks in advance, Dave

4 3

ArrowChoice for simulation arrow (Sim)
by Steve Lihn 29 Feb '08

29 Feb '08

Hi, In John Hughes\' paper \"Programming with arrows\" [1] Section 5, the simulation arrow is defined (Sim). All the examples went well. However, when I want to use it beyond circuit simulation and I added a \"if-then-else\" line, GHC is asking to define ArrowChoice for Sim. Does anybody have the ArrowChoice definition for Sim ? Thanks, Steve [1] http://www.cs.chalmers.se/~rjmh/afp-arrows.pdf

1 0