September 2009 - Haskell-Cafe

F# mailing list?
by Günther Schmidt 28 Sep '09

28 Sep '09

Hi guys, is there a mailing list for haskellers that defected to F#? Not that I was I going to, just asking, absolutely hypothetically. Uhm. Günther

3 3

Edinburgh Hackathon (Hac7) summary
by Dougal Stanton 27 Sep '09

27 Sep '09

It's been nearly a month now since the Edinburgh Hackathon (29 & 30 August), organised to coincide with the convergence of Haskellers arriving for the ICFP. My apologies for this very late summary! Eric Kow was the driving force behind this Hackathon, though I agreed to help out locally because I live in Edinburgh. We had trouble organising a good venue through a combination of university holidays and Edinburgh International Festival. We spent the first day camping out in a cafe. On the second day Graeme Hutton arranged for us to use an empty room in the bowels of the ICFP venue. If I remember well the first day was heavily Darcs-oriented, and there were interesting conversations on version control representation, conflicts, and the future of Darcs over breakfast. The staff at Henderson's Restaurant didn't seem to mind our impromptu geeky enclave, and we kept buying coffee to pay for our wifi usage. I'll let the other Haskellers explain in their own words what they were up to for the rest of the time. Maybe some of them have updates on where these projects have gone since then? "At the hackathon I started working on a small extension to the GHCi debugger. The goal of the extension is to allow the debugger to show the term representation of thunks where possible. Presently, to find out the value of a thunk in the debugger you must force its evaluation. However, this is not always desirable, especially if the thunk contains breakpoints, or if it diverges. In some cases it ought to be possible to extract source code information from a thunk and show that to the user. I didn't get very far in the implementation because I spent most of the time re-learning how the term printer in the debugger works (and chatting to other people)." -- Bernie Pope "I was continuing to work on my camp correctness proofs. In particular, investigating how best to handle symmetric lemmas, and refactoring the existing definitions and proofs." -- Ian Lynagh "I worked on fixing Solaris/SPARC port of GHC. It turned out to be a tedious header file related issue, not very interesting." -- Ben Lippmeier "I worked on the formlets (again), I released a version that doesn't add any functionality, it only solves a bug." -- Chris Eidhof "I worked on a feature to let darcs users more conveniently upgrade to the hashed repository format. I also had lots of interesting discussions with darcs users and hackers :-)" -- Eric Kow "I worked on analysing some darcs performance problems, fixing a couple of GHC bugs, and some discussion about packages." -- Simon Marlow "I (along with Benedikt Huber) worked on adding a 'cabal init' command which automates much of the work of cabalising a project (creating a .cabal file, a Setup.hs file, a LICENSE file, etc). It's not quite done yet but should be finding its way into the cabal-install HEAD soon! -- Brent Yorgey "I worked on the new hackage-server a bit. darcs get http://code.haskell.org/hackage-server/ live instance at http://sparky.haskell.org:8080/" -- Duncan Coutts Thanks everyone who took part and who supplied an elegant summary of their work that weekend. I hope you all had a good time! Cheers, D -- Dougal Stanton dougal(a)dougalstanton.net // http://www.dougalstanton.net

1 0

(mostly OT) Strange patterns of commas
by Matthias Kilian 27 Sep '09

27 Sep '09

Fibonacci numbers are always surprising. Try this on a terminal that's 159 characters wide (using a properly sized xterm(1) may be a good idea): let f = 1 : 1 : zipWith (+) f (tail f) in take 780 f Ciao, Kili -- There are two ways of constructing a software design: One way is to make it so simple that there are obviously no deficiencies, and the other way is to make it so complicated that there are no obvious deficiencies. The first method is far more difficult. -- C.A.R. Hoare

2 1

ANN: rss2irc 0.4 released
by Simon Michael 27 Sep '09

27 Sep '09

I have released rss2irc 0.4, with some improvements from the field: - fix a problem connecting with irc.quakenet.org (Radoslav Dorcik) - feed polling now recovers from transient failures - can poll a local file: uri as well as remote uris - more robust new item detection, with some alternate strategies available - --idle option, waits for channel inactivity before announcing - limit individual field sizes, preventing too many irc messages per item - misc. output improvements - logging improvements - --delay now takes minutes, not seconds See hackage for more detail. Note the new contributor! Thanks Radoslav. home: http://hackage.haskell.org/package/rss2irc darcs repo: http://joyful.com/darcsweb/darcsweb.cgi?r=rss2irc I am currently using this to run hackagebot, darcscommitbot and other announce-bots on freenode. A tip: if you experiment, avoid joining the irc server too frequently; once per minute may be ok. Feedback and patches welcome. Best, -Simon

1 1

Re: [Haskell-cafe] 16 bit floating point data in Haskell?
by Olex P 27 Sep '09

27 Sep '09

Hi, Yes, I mean "sizeOf 2". It's useful not only on GPUs but also in "normal" software. Think of huge data sets in computer graphics (particle clouds, volumetric data, images etc.) Some data (normals, density, temperature and so on) can be easily represented as float 16 making files 200 GB instead of 300 GB. Good benefits. Cheers, Oleksandr. On Sun, Sep 27, 2009 at 9:19 PM, Peter Verswyvelen <bugfact(a)gmail.com>wrote: > He meant 16-bit floats, which have sizeOf 2 > On GPUs this is common and implemented in hardware (at least on the old > GPUs). > > On DPSs you commonly had 24-bit floats too. > > But these days I guess 32-bit is the minimum one would want to use? Most of > the time I just use double anyway :) > > On Sun, Sep 27, 2009 at 9:47 PM, Ross Mellgren <rmm-haskell(a)z.odi.ac>wrote: > >> What about the built-in Float type? >> >> Prelude Foreign.Storable> sizeOf (undefined :: Float) >> 4 >> Prelude Foreign.Storable> sizeOf (undefined :: Double) >> 8 >> >> Or maybe you mean something that can be used with FFI calls to C, in which >> case Foreign.C.Types (CFloat). >> >> Both instance the Floating, RealFloat, RealFrac, etc, classes so should >> operate largely the same as (modulo precision) a Double. >> >> -Ross >> >> >> On Sep 27, 2009, at 2:42 PM, Olex P wrote: >> >> Hi guys, >>> >>> Do we have anything like half precision floats in Haskell? Maybe in some >>> non standard libraries? Or I have to use FFI + OpenEXR library to achieve >>> this? >>> >>> Cheers, >>> Oleksandr. >>> _______________________________________________ >>> Haskell-Cafe mailing list >>> Haskell-Cafe(a)haskell.org >>> http://www.haskell.org/mailman/listinfo/haskell-cafe >>> >> >> _______________________________________________ >> Haskell-Cafe mailing list >> Haskell-Cafe(a)haskell.org >> http://www.haskell.org/mailman/listinfo/haskell-cafe >> > >

2 2

Processing EXIF data with Haskell
by Colin Paul Adams 27 Sep '09

27 Sep '09

I'm writing web software for a photograph gallery. I want to be able to display selective fields from the EXIF data. I thought I would be able to use the GD and exif libraries from Hackage to do this. However: Calling Graphics.GD.loadJpegByteString followed by Graphics.GD.saveJpegFile does not preserve the EXIF fields of the uploaded file, but instead writes some new ones (many fewer fields than the original, and nothing of interest). Graphics.Exif.allTags shows an interesting subset of the original EXIF data, but nothing from the version saved by GD. Can anyone shed any light on any of this? -- Colin Adams Preston Lancashire

2 2

yi-editor: Duplicate instance declarations
by Marcelo Sousa 27 Sep '09

27 Sep '09

Hey guys, I'm trying to install yi using cabal but I got this error. Any ideas how to solve it?! I'm using ghc-6.10.1 and cabal-install version 0.6.2 using version 1.6.0.2 of the Cabal library. Thanks, Marcelo -- Code -- $ cabal install yi Resolving dependencies... Configuring yi-0.6.1... Preprocessing library yi-0.6.1... Preprocessing executables for yi-0.6.1... Building yi-0.6.1... [ 1 of 120] Compiling System.FriendlyPath ( System/FriendlyPath.hs, dist/build/System/FriendlyPath.o ) [ 2 of 120] Compiling Shim.ProjectContent ( Shim/ProjectContent.hs, dist/build/Shim/ProjectContent.o ) [ 3 of 120] Compiling Parser.Incremental ( Parser/Incremental.hs, dist/build/Parser/Incremental.o ) [ 4 of 120] Compiling Data.Trie ( Data/Trie.hs, dist/build/Data/Trie.o ) [ 5 of 120] Compiling Data.DelayList ( Data/DelayList.hs, dist/build/Data/DelayList.o ) [ 6 of 120] Compiling Data.Rope ( Data/Rope.hs, dist/build/Data/Rope.o ) [ 7 of 120] Compiling Data.Prototype ( Data/Prototype.hs, dist/build/Data/Prototype.o ) [ 8 of 120] Compiling HConf.Utils ( HConf/Utils.hs, dist/build/HConf/Utils.o ) [ 9 of 120] Compiling HConf.Paths ( HConf/Paths.hs, dist/build/HConf/Paths.o ) [ 10 of 120] Compiling Paths_yi ( dist/build/autogen/Paths_yi.hs, dist/build/Paths_yi.o ) [ 11 of 120] Compiling HConf ( HConf.hs, dist/build/HConf.o ) [ 12 of 120] Compiling Yi.Char.Unicode ( Yi/Char/Unicode.hs, dist/build/Yi/Char/Unicode.o ) [ 13 of 120] Compiling Yi.UI.Common[boot] ( Yi/UI/Common.hs-boot, dist/build/Yi/UI/Common.o-boot ) [ 14 of 120] Compiling Yi.String ( Yi/String.hs, dist/build/Yi/String.o ) [ 15 of 120] Compiling Yi.Monad ( Yi/Monad.hs, dist/build/Yi/Monad.o ) [ 16 of 120] Compiling Yi.Keymap.Completion ( Yi/Keymap/Completion.hs, dist/build/Yi/Keymap/Completion.o ) [ 17 of 120] Compiling Yi.Editor[boot] ( Yi/Editor.hs-boot, dist/build/Yi/Editor.o-boot ) [ 18 of 120] Compiling Yi.Debug ( Yi/Debug.hs, dist/build/Yi/Debug.o ) [ 19 of 120] Compiling Yi.Prelude ( Yi/Prelude.hs, dist/build/Yi/Prelude.o ) Yi/Prelude.hs:182:9: Duplicate instance declarations: instance Category Accessor.T -- Defined at Yi/Prelude.hs:182:9-38 instance Category Accessor.T -- Defined in data-accessor-0.2.1:Data.Accessor.Private cabal: Error: some packages failed to install: yi-0.6.1 failed during the building phase. The exception was: exit: ExitFailure 1

3 2

Quadratic complexity though use of STArrays
by Dan Rosén 27 Sep '09

27 Sep '09

Dear haskell-cafe users, I am constructing a shuffle function: given an StdGen and a list, return the list permuted, with all permutations of equal probability. There is the simlpe recursive definition: generate a number from 1 to length list, take this element out from the list, call the function recursively on the remaining list and then cons the element on the shuffled list. A more imperative approach is to make the list an array, and traverse the array in reverse, swapping the iterated element with an arbitrary element less than or equal to the iterator. These functions are implemented as shuffleRec and shuffleArr, respectively. What complexity does these functions have? I argue that the shuffleArr function should be O(n), since it only contains one loop of n, where each loop does actions that are O(1): generating a random number and swapping two elements in an array. I argue that the shuffleRec function should be O(n^2), since for each call, it creates a new list in O(n), with the drop and take calls, and calls itself recursively. This yields O(n^2). However, they both have the same runnig time (roughly), and through looking at the plot it _very_ much looks quadratic. I am compiling with GHC and I guess there is something in the lazy semantics that confuses me about the complexities, and maybe I have misunderstood how STArrays work. Any pointers to what's going in is greatly appreciated! Best regards, Dan Rosén, Sweden Here is the code: module Main where import Control.Monad import Control.Monad.ST import Data.Array.ST import Data.STRef import System.Random import Time import CPUTime shuffleArr :: StdGen -> [a] -> [a] shuffleArr g list = runST $ do let n = length list gref <- newSTRef g arr <- listToArray list forM_ [n,n-1..2] $ \p -> do m <- rand (1,p) gref swap arr m p getElems arr where rand range gref = do g <- readSTRef gref let (v,g') = randomR range g writeSTRef gref g' return v swap a n m = do [n',m'] <- mapM (readArray a) [n,m] mapM (uncurry $ writeArray a) [(m,n'),(n,m')] listToArray :: [a] -> ST s (STArray s Int a) listToArray list = let n = length list in newListArray (1,n) list shuffleRec :: StdGen -> [a] -> [a] shuffleRec g list = x:shuffleArr g' xs where (n,g') = randomR (0,length list-1) g (x:xs') = drop n list xs = take n list ++ xs' -- A somewhat lame attempt to derive the complexities through testing, -- prints the times for the different functions in a table main :: IO () main = do let times = take 30 $ iterate (+30000) 10000 answers <- mapM test times sequence_ [ putStrLn $ concatMap ((++ "\t"). show) [toInteger t,arr,rec] | (t,(arr,rec)) <- zip times answers ] -- Perform a test of size n, and return the cycles it took for the different -- algorithms in a pair. Evaluation is enforced by seq on length of the list. test :: Int -> IO (Integer,Integer) test n = do let list = [1..n] [g1,g2] <- replicateM 2 newStdGen length list `seq` do s <- doTime ("shuffleArr " ++ show n) $ (length $ shuffleArr g1 list) `seq` return () s' <- doTime ("shuffleRec " ++ show n) $ (length $ shuffleRec g2 list) `seq` return () return (s,s') -- This is taken from GenUtil from the JHC creator's homepage doTime :: String -> IO a -> IO Integer doTime str action = do start <- getCPUTime x <- action end <- getCPUTime let time = (end - start) `div` 1000000 -- `div` cpuTimePrecision -- putStrLn $ "Timing: " ++ str ++ " " ++ show time return time

5 5

Hurray! I hit my 1st true lazy IO problem
by Günther Schmidt 27 Sep '09

27 Sep '09

and now I'll really have to figure out Iteratee / Enumerator to solve it. Can I call myself a Haskeller now? Günther. PS: The problem domain is crawling over a website.

4 3

Snow Leopard Breaks GHC
by David Leimbach 27 Sep '09

27 Sep '09

Just thought I'd point out that my old GHC install is now broken by the update to Snow Leopard. Dave

9 21