June 2012 - Haskell-Cafe

AI - machine learning
by miro 30 Jul '12

30 Jul '12

Hi All, recently I started to take a look at haskell, especially at AI. I can see some email addresses of interested people there but not so much of other activity behind. Does it exist some mailing group especially for AI? Basically I'm interested in trying some machine learning algorithms. Start with reinforcement learning and value-based), and go towards AGI (Artificial General Intelligence). Does anybody know about some already existing haskell approaches, or is there anybody working on this? Cheers, m.

4 3

Combining Network Descriptions in Reactive.Banana
by Alexander Foremny 10 Jul '12

10 Jul '12

Hello list, I am currently experimenting with Reactive.Banana and I am trying to write a program that simply displays a String which is supplied by the user. Thus in Main.hs I'd like to create a Behavior t String from some arbitrary sources and pass this to a library function performing the printing. However, I am not able to get the desired behavior of my program. A sample Main.hs is available as Gist [1]. The cause of error appears to be that the type t over which NetworkDescription and Behavior are parameterized are distinct! I tried several variations of the code in [1] using ExistentialQuantification over t and/or giving explicit type signatures. The error is slightly different in each case, but always boils down to the same reason. In case explicit error messages are required I can post them to the list. Any insight into this would be of great help. And additionally I'd be interested if this kind of program is possible with Reactive.Banana at all. [1] https://gist.github.com/3004430

4 4

Odd HDBC connection bug
by William Shackleton 07 Jul '12

07 Jul '12

Hi I'm having issues with HDBC when connecting to a remote MySQL server - certain queries cause the DB connection to be lost. The following program demonstrates this: import Database.HDBC import Database.HDBC.ODBC main = do conn <- connectODBC "DSN=owlro" putStrLn "Connected" quickQuery' conn "SELECT time, power FROM Power LIMIT 100;" [] putStrLn "Finished query 1" quickQuery' conn "SELECT time, power FROM Power ORDER BY time;" [] putStrLn "Finished query 2" disconnect conn The DSN points to a remote 32-bit Ubuntu 12.04 Server. The connection and queries work using mysql and isql on the command line, and the table in question contains about 1.3 million rows and only the columns id, time, power. When I compile and run this on a 64-bit Ubuntu 12.04 PC using HDBC-2.3.1.1 and HDBC-odbc-2.3.1.0 using GHC, I get the following (unwanted) output: $ ghc Test.hs $ ./Test Connected Finished query 1 Test: SqlError {seState = "[\"08S01\"]", seNativeError = -1, seErrorMsg = "execute execute: [\"2013: [MySQL][ODBC 5.1 Driver][mysqld-5.5.24-0ubuntu0.12.04.1]Lost connection to MySQL server during query\"]"} This program finishes (and fails) in less than half a second. When I run it using runhaskell however, the slow second query completes: $ runhaskell Test.hs Connected Finished query 1 Finished query 2 I get the same result from ghci; this method takes about 10 seconds as expected. Basically, simple (short) queries complete, yet long ones crash the connection. I also noticed similar results with HDBC-mysql. Any ideas on what is causing this? Thanks, Will Shackleton digitalsquid.co.uk

1 1

Martin Odersky on "What's wrong with Monads"
by Jonathan Geddes 05 Jul '12

05 Jul '12

Cafe, I was watching a panel on languages[0] recently and Martin Odersky (the creator of Scala) said something about Monads: "What's wrong with Monads is that if you go into a Monad you have to change your whole syntax from scratch. Every single line of your program changes if you get it in or out of a Monad. They're not polymorphic so it's really the old days of Pascal. A monomorphic type system that says 'well that's all I do' ... there's no way to abstract over things. " [0, 53:45] Thoughts? --J Arthur [0] - http://css.dzone.com/articles/you-can-write-large-programs

20 26

Re: [Haskell-cafe] Data.Bitmap, readBitmap exception
by Balazs Komuves 02 Jul '12

02 Jul '12

Hello, Author of Data.Bitmap here. The issue is caused by the fact that Data.Bitmap does *not* support bmp files at all. This should be clear from the documentation. The functions you tried to used use their own file format, which a really very simple format; they are only added for simple experimentation. (again, this is explicit in the docs). The same documentation points to the stb-image [1] package, which supports loading of some common formats (png, jpg, bmp). In general, I suggest that with package-specific question, unless it is a widely used package, mail the maintainer directly; you have better chance to get an answer that way. Regards, Balazs [1] http://hackage.haskell.org/package/stb-image Hello, > > I am using bitmap-0.2 [1] to load a Bitmap file [2]. However, the > following sample GHCi session fails for me. > > > Prelude Data.Bitmap Data.Bitmap.Pure.File> readBitmap "b.bmp" :: IO > (Bitmap Word8) > > *** Exception: Data/Bitmap/Internal.hs:(63,17)-(67,15): Non-exhaustive > patterns in case > > I downloaded the sources of bitmap-0.2 from Hackage and modified the > relevant function by appending the last line. > > > decodeCType :: CInt -> PixelComponentType > > decodeCType k = case k of > > 1 -> PctWord8 > > 2 -> PctWord16 > > 3 -> PctWord32 > > 4 -> PctFloat > > _ -> error $ "decodeCType: unexpected integer (" ++ show k ++ ")" > > With this modification the same GHCi session given above results in > the following error. > > > Prelude Data.Bitmap Data.Bitmap.Pure.File> readBitmap "b.bmp" :: IO > (Bitmap Word8) > > *** Exception: decodeCType: unexpected integer (7077888) > > I don't have any idea where 7077888 comes from. Am I doing something > wrong? Is my bitmap corrupted in some way? The number 7077888 has no > apparent significance except that it is a perfect cube [3]. So I guess > this is not a problem with endianness. > > In fact as far as I can tell `decodeCType` is only called on > `(PixelComponent t => c_type t)` and `c_type t` only yields integer > values in the (inclusive) range of 1 to 4. That is if I am not missing > some orphaned instance. > > Could anyone please enlighten me what is happening and if I am doing > something wrong? > > In case this is relevant: I am using bitmap-0.2 from Hackage with GHC > 7.4.1 on a 3.2.20 kernel i686 architecture Linux. > > Regards, > Alexander Foremny > > [1] http://hackage.haskell.org/package/bitmap > [2] https://www.dropbox.com/s/pyutvni9vx6f6mo/b.bmp > [3] http://www.wolframalpha.com/input/?i=7077888 >

2 1

not enough fusion?
by Johannes Waldmann 02 Jul '12

02 Jul '12

Dear all, while doing some benchmarking (*) I noticed that function s1 is considerably faster than s2 (but I wanted s2 because it looks more natural) (for n = 10000, s1 takes 20 s, s2 takes 13 s; compiled by ghc-7.4.2 -O2) s1 :: Int -> Int s1 n = sum $ do x <- [ 0 .. n-1 ] return $ sum $ do y <- [ 0 .. n-1 ] return $ gcd x y s2 :: Int -> Int s2 n = sum $ do x <- [ 0 .. n-1 ] y <- [ 0 .. n-1 ] return $ gcd x y I was expecting that in both programs, all lists will be fused away (are they?) so the code generator essentially can produce straightforward assembly code (no allocations, no closures, etc.) For reference, I also wrote the equivalent imperative program (two nested loops, one accumulator for the sum) (with the straightforward recursive gcd) and runtimes are (for same input as above) C/gcc: 7.3 s , Java: 7.7 s, C#/Mono: 8.7 s So, they sort of agree with each other, but disagree with ghc. Where does the factor 2 come from? Lists? Laziness? Does ghc turn the tail recursion (in gcd) into a loop? (gcc does). (I am looking at -ddump-asm but can't quite see through it.) (*) benchmarking to show that today's compilers are clever enough such that the choice of paradigm/language does not really matter for this kind of low-level programming.

7 8

ANN: prof2pretty-0.1.0.0
by Claude Heiland-Allen 30 Jun '12

30 Jun '12

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi all, I wanted to try visual-prof[1] but found it was broken by ghc RTS changes[2]. I ended up writing prof2pretty[3]: http://hackage.haskell.org/package/prof2pretty - --8<-- sccpragmabomb adds SCC pragmas encoding source location. prof2pretty extracts those SCCs from the profiling output to annotate the sources using HTML + CSS + JavaScript. - --8<-- [1] http://hackage.haskell.org/package/visual-prof [2] http://hackage.haskell.org/trac/ghc/ticket/4094 [3] prof2pretty started as a fork of visual-prof but soon there was none of the original code left, though the essential concept remains the same Thanks, Claude -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.12 (GNU/Linux) iQEcBAEBAgAGBQJP7wPyAAoJEHZDo4jueIiW7WsH/3Lmk2k7o+2i5u+Vxn+EZXRX 5XRDXZ8mLm1NBb/OSfRFr6RIyZxwou8JeavZJuhqOBL38KAlppf89OnaJyxUB+ap Cp03ui/1LbihqIbMxupNH03pQwkNR7b2MV2S07c5TOvOJHsksE53zEin/EHQQ6jO 6QxrQdNibjSYlXDNtZCzEj0HWVpSBD4gXdPFv9nFS+RKPjAkgC4gmY4yItYDep60 qboOrxKVhYJdqTmhl+OXhe7LAAJ9Nc3xiheGLbeyTfH2owAnT8X4zOdtssOsPPdm AQBgxKa8p5KckZCKS7Elj8pap9uzUHGc8KuINpjx40b7envHP44t65xngEq1TaU= =znB1 -----END PGP SIGNATURE-----

1 0

Maintainer of pngload listening?
by Daniel Trstenjak 30 Jun '12

30 Jun '12

Hi, Marko Lauronens email address from the package is invalid, so I'm trying to reach him this way. pngload doesn't compile with ghc >= 7.2 because of it's dependency to base and haskell98. Greetings, Daniel

3 3

Slowdown with GHC when using multiple CPUs
by David Powell 29 Jun '12

29 Jun '12

I recently ran into a problem writing a program that I hoped I could speed up by running over multiple CPUs. I want non-haskell users to be able to run the tool and take advantage of multiple CPUs. *But* there is a serious slowdown when the RTS is run with -N and some of the CPUs are already busy. This is already mentioned in the GHC docs, but the problem I experienced was more serious than I expected causing a slowdown of around 2x compared to running with a single CPU. I reproduced the problem with the following code from the haskell wiki: {-# LANGUAGE BangPatterns #-} import Data.Digest.Pure.MD5 import qualified Data.ByteString.Lazy as L import System.Environment import Control.Concurrent import Control.Monad (replicateM_) main = do files <- getArgs str <- newEmptyMVar mapM_ (forkIO . hashAndPrint str) files printNrResults (length files) str printNrResults i var = replicateM_ i (takeMVar var >>= putStrLn) hashAndPrint str f = do bs <- L.readFile f let !h = show $ md5 bs putMVar str (f ++ ": " ++ h) When run on 4 idle CPU cores, I get the following wall clock times: ./run +RTS -N1 : 20.4 sec ./run +RTS -N2 : 11.0 sec ./run +RTS -N4 : 6.7 sec When run on the same 4 core machine, but with 2 cores already busy: ./run +RTS -N1 : 23.5 sec ./run +RTS -N2 : 14.1 sec ./run +RTS -N4 : 57.8 sec <---- Blowout... This is quite a problem in practice when running on a shared server. Is there anything that can be done to address this? (I wrote up a few more details here: http://thunking.drp.id.au/2012/06/slowdown-with-ghc-when-using-multiple.html ) Thanks, -- David Powell

2 1

Help requested: naming things in conduit
by Michael Snoyman 29 Jun '12

29 Jun '12

Hi all, I'm just about ready to make the 0.5 release of conduit. And as usual, I'm running up against the hardest thing in programming: naming things. Here's the crux of the matter: in older versions of conduit, functions would have a type signature of Source, Sink, or Conduit. For example: sourceFile :: MonadResource m => FilePath -> Source m ByteString I think most people can guess at what this function does: it produces a stream of ByteStrings, which are read from the given file. Now the trick: Source (and Sink and Conduit) are all type synonyms wrapping around the same type, Pipe. Ideally, we'd like to be able to reuse functions like sourceFile in other contexts, such as producing a Conduit that calls sourceFile[1]. However, the type synonym Source over-specifies some of the type parameters to Pipe, and therefore `sourceFile` can't be used directly to create a Conduit[2]. To get around this whole problem, I've added a number of type synonyms with rank-2 types, that don't over-specify. You can see the type synonyms here[3], and more explanation of the problem here[4]. So my question is: can anyone come up with better names for these synonyms? Just to summarize here: * All of the generalized types start with a G, e.g., Source becomes GSource. * For Sinks and Conduits, if leftovers are generated, there's an L after the G (e.g., GLSink). * For Sinks and Conduits which consume all of their input and then return the upstream result, we tack on an Inf for Infinite (e.g., GInfConduit, GLInfSink). I think these names are relatively descriptive, and certain `GSink ByteString m Int` is easier to follow than `Pipe l ByteString o u m Int`, but I was wondering if anyone had some better recommendations. Michael [1] For example, maybe we want to produce `conduitFiles :: MonadResource m => Conduit FilePath m ByteString` [2] This problem exists to a smaller extent in conduit 0.4. This is the purpose of the sinkToPipe function. [3] https://github.com/snoyberg/conduit/blob/52d7bc0b551b877de92be4c87f933e3ffb… [4] https://github.com/snoyberg/conduit/blob/a853141d7b9eed047c7cc790979f73a346…

2 2