September 2011 - Haskell-Cafe

SPECIALIZE in the presence of constraints
by Nicu Ionita 25 Sep '11

25 Sep '11

Hallo List, I can't understand how pragma SPECIALIZE works in the presence of constraints. I have 2 modules, one which defines a general search framework, and one which implements it in a concrete context. The general module defines functions like: {-# SPECIALIZE pvQSearch :: Node (Game m) Move Int => Int -> Int -> Int -> Search (Game m) Move Int Int #-} pvQSearch :: Node m e s => s -> s -> Int -> Search m e s s pvQSearch a b c = do ... while the implementation uses concrete data types Move and Int (for e and s respectively) and defines instances and other types (like the polimorphic type Game m, where m is a monad). Node m e s is a class and the instance Node (Game m) Move Int is defined in the implementation module. From outside this construct is used by calling a generic search function, but giving parameters which match the implementation (i.e. with Move and Int for e and s, for example). Now what I don't understand: 1. how can the compiler (here ghc) know which function to expose as the correct generic search function? There must be two search functions generated, one generic and its specialization. Does the module export both and later the linker chooses the correct one, when the client module is known? 2. how can I know that the specialization is really used? When I have constraints, will the specializations be generated in the assumption that the constraint is matched? When will be this match checked? My problem is that, with or without specializations, the performance of the code is the same - so I doubt the specializations are used. Thanks, Nicu

2 1

Is there still at list of active Haskell projects
by Vasili I. Galchin 25 Sep '11

25 Sep '11

Hello, On http://www.haskell.org I didn't see a list of active Haskell projects. ?? Thanks, Bill

3 2

A Missing Issue on Second Generation Strategies
by Burak Ekici 24 Sep '11

24 Sep '11

Dear List, I am trying to parallelize RSA encryption and decryption by using below manner, but when I run executable output file with "+RTS -s -N2" command on Windows 7, output stats say 4 sparks are being created however none of them converted into real OS threads. -- SPARKS :4 (0 converted, 4 pruned) -- I was thinking that the problem could occur due to lack of forcing parallelization but, as far as I know 'rdeepseq' works for that aim. Briefly, I could not solve the issue why parallelization was not being implemented. I would be appreciated if any of you shed a light on the issue that I missed. Here is the mentioned part of code: split4ToEnc :: RSAPublicKey -> [Integer] -> [Integer] split4ToEnc (PUB n e) [] = [] split4ToEnc (PUB n e) (x:xs) = ((ersa (PUB n e) secondPart2) ++ (ersa (PUB n e) secondPart1) ++ (ersa (PUB n e) firstPart2) ++ (ersa (PUB n e) firstPart1)) `using` strategy where firstPart1 = fst (Main.splitAt((length (x:xs)) `div` 4) (fst(Main.splitAt ((length (x:xs)) `div` 2) (x:xs)))) firstPart2 = snd (Main.splitAt((length (x:xs)) `div` 4) (fst(Main.splitAt ((length (x:xs)) `div` 2) (x:xs)))) secondPart1 = fst (Main.splitAt((length (x:xs)) `div` 4) (snd(Main.splitAt ((length (x:xs)) `div` 2) (x:xs)))) secondPart2 = snd (Main.splitAt((length (x:xs)) `div` 4) (snd(Main.splitAt ((length (x:xs)) `div` 2) (x:xs)))) strategy res = do a <- rpar (ersa (PUB n e) (firstPart1) `using` rdeepseq) b <- rpar (ersa (PUB n e) (firstPart2) `using` rdeepseq) c <- rpar (ersa (PUB n e) (secondPart1) `using` rdeepseq) d <- rpar (ersa (PUB n e) (secondPart2) `using` rdeepseq) rdeepseq res Thanks a lot, Burak.

3 4

Fwd: problem with cabal install MissingH-1.1.1.0
by Ozgur Akgun 24 Sep '11

24 Sep '11

Ooops, apparently I forgot to reply to the list. Sorry. ---------- Forwarded message ---------- From: Ozgur Akgun <ozgurakgun(a)gmail.com> Date: 22 September 2011 16:49 Subject: Re: [Haskell-cafe] problem with cabal install MissingH-1.1.1.0 To: Mariano Cortesi <mcortesi(a)gmail.com> Hi Mariano. On 22 September 2011 12:48, Mariano Cortesi <mcortesi(a)gmail.com> wrote: > I'm with mac OS X lion, ghc version 7.2.1 > and when a i try to install MissingH version 1.1.1.0 it fails with > > [24 of 47] Compiling Data.BinPacking ( src/Data/BinPacking.hs, > dist/build/Data/BinPacking.o ) > > src/Data/BinPacking.hs:45:6: > Illegal datatype context (use -XDatatypeContexts): (Num size, > This is weird, because on OS X Lion, and with the following versions, it doesn't fail for me: The Glorious Glasgow Haskell Compilation System, version 7.2.1 cabal-install version 0.10.2 using version 1.12.0 of the Cabal library Nevertheless, you can try the following to do what cabal asks you to do: cabal install --ghc-option=-XDatatypeContexts Which will give a warning, but should succeed anyway. Hope this helps, Ozgur -- Ozgur Akgun

5 4

Where is SNMap for stable names?
by Sean Leather 24 Sep '11

24 Sep '11

There is an abstract type called SNMap for stable names referred to in [1]. This has apparently disappeared from GHC a long time ago. Is it still available somewhere, or is there a suitable replacement for it? Regards, Sean [1] "Stretching the storage manager: weak pointers and stable names in Haskell" - http://research.microsoft.com/apps/pubs/default.aspx?id=67497

4 8

ANN: flexiwrap-0.1.0 etc.
by Iain Alexander 23 Sep '11

23 Sep '11

New versions of the package flexiwrap (supplying flexible wrappers for instance selection) and related packages have been released to Hackage. flexiwrap-0.1.0 data-type-0.1.0 function-combine-0.1.0 flexiwrap-smallcheck-0.0.1 flexiwrap version 0.1.0 supports instances of many additional classes for FlexiWrap, and improved documentation, although much remains to be done - more class instances for FlexiWrap, and other wrappers, and more documentation. data-type version 0.1.0 adds a proxy type Data.Type.Proxy. function-combine version 0.1.0 updates its dependencies to support the new version of data-type. flexiwrap-smallcheck version 0.1.0 is a new package supplying support for instances of Serial for FlexiWrap, for use with the smallcheck package. Support for Serial instances for other wrappers is outstanding. Regards, Iain. -- ia(a)stryx.demon.co.uk

1 0

Standard 'decorator' class?
by Jacques Carette 23 Sep '11

23 Sep '11

Has a class like class Decoration t where type Datum t :: * value :: t -> Datum t been defined in any of the haskell packages on hackage? [There are so many variants on naming, it is rather difficult to search for this] Jacques

1 0

Re: [Haskell-cafe] Fwd: problem with cabal install MissingH-1.1.1.0
by Tom Murphy 23 Sep '11

23 Sep '11

I used to have a lot of cabal problems when running a version of the GHC installed with MacPorts. I notice you're using homebrew. Maybe that's part of the problem? Tom On Sep 22, 2011 8:11 PM, "Mariano Cortesi" <mcortesi(a)gmail.com> wrote:

2 1

[ANNOUNCE] skein-0.1: Skein, a family of cryptographic hash functions. Includes Skein-MAC as well.
by Felipe Almeida Lessa 23 Sep '11

23 Sep '11

Hello! I'm pleased to announce the first version of the skein package [1]! Skein is a family of fast secure cryptographic hash functions [2]. The skein package provides high-level bindings (using crypto-api [3]) to the optimized Skein C library. Currently we support Skein as a hash function and also Skein as a message authentication code (MAC). The code is well-documented with examples and has many known answer tests (KATs). It currently lives on patch-tag [4]. Hackage will be build the docs soon enough. Meanwhile, here's a short example: Prelude> :s -XOverloadedStrings Prelude> :m Crypto.Skein Crypto.Classes Data.ByteString.Char8 Data.Serialize Prelude Crypto.Skein Crypto.Classes Data.ByteString.Char8 Data.Serialize> encode (hash' "Haskell" :: Skein_512_512) "b\227\213t\170\GS\131w\151\240\160\183\205\US\169\138\&6\239\215\156\227\SI\234\142P\132\182\SOH\180t\203b\147\190\DLE{\168\214\DC1\238\232y\168\161s\207\172\216\a\202\130c$\SYN\asD<\151\201D\138A\SUB" Prelude Crypto.Skein Crypto.Classes Data.ByteString.Char8 Data.Serialize> encode (skeinMAC' "secretkey" "Haskell" :: Skein_512_512) "_i>.Z\129\ENQ\217L\t9BZ\220\163c&9#\135\234A{\181\&2*\149\STX\216_\DC43\bF\189\196D\r ;\208\EOT}\ENQY\147\232\244e/;4\212!\224\US\138\138\234q=1\129\244" Enjoy! =D [1] http://hackage.haskell.org/package/skein [2] http://swww.skein-hash.info/ [3] http://hackage.haskell.org/package/crypto-api [4] https://patch-tag.com/r/felipe/skein/ -- Felipe.

3 7

Package versioning policy - Harmless type changes?
by Heinrich Apfelmus 23 Sep '11

23 Sep '11

Dear all, the package versioning policy specifies that changing the type of any entity makes a change of the B number in A.B.C necessary. However, it seems to me that a small selection of type changes don't break dependent code. In particular, consider the following example, where I drop a Typeable class constraint: - foo :: Typeable a => AddHandler a -> NetworkDescription (Event a) + foo :: AddHandler a -> NetworkDescription (Event a) Can removing a type class constraint on a function break dependent code? Should I change the B number or just the C in version A.B.C when introducing this change? Best regards, Heinrich Apfelmus -- http://apfelmus.nfshost.com

3 3