April 2010 - Haskell-Cafe

Ocaml for Haskellers tutorial
by José Iborra 17 Apr '10

17 Apr '10

As a Haskeller who is forced to switch to Ocaml by the job market (or at least by the job interview market), I could really use a "Ocaml for Haskellers" tutorial. I was wondering if there are others who have been in this situation recently and can share a few pointers, or better yet, write a blog entry pointing to good sources of information and highlighting the pitfalls. For instance, I have heard that Ocaml is only fast when one uses loops instead of folds, but I wonder if this is an overstatement. Of course due to the strict semantics of Ocaml many of the assumptions that come from laziness will no longer be valid, but I wonder how different is idiomatic Ocaml from idiomatic Haskell in real world. Thanks, pepe

9 8

ANN: HaskellTorrent/Combinatorrent v0.2.0
by Jesper Louis Andersen 17 Apr '10

17 Apr '10

Goodbye HaskellTorrent, hello Combinatorrent! Due to the internationally acclaimed Boing!'er Shae Erisson, HaskellTorrent has been renamed into Combinatorrent. We thus present Combinatorrent v0.2, Easter eggs release. This release marks yet another milestone in getting a decent bittorrent client for Haskell. There are still some performance regressions to be eliminated, but overall it looks far better than in the v0.1.1 release. Also, the system leaks memory, albeit very slowly (or so sayeth the great heap profiler). This is the focus for the next version. We might also look into integrating the FAST extension, since it plugs a series of problematic spots in the original bittorrent protocol specification. If you want to hack on Combinatorrent, there are a lot of interesting things to do. I try, deliberately, to keep tasks around which can be used to gain foothold in the source code. Any level of skill is needed and I am happy to help a newbie if he or she gets stuck on something. The source code is at http://github.com/jlouis/combinatorrent should you want to tinker with it. Any kind of patch is welcome, be it fixes or improvements to the goat teleporter. So come help us avoiding success! Apart from numerous small optimizations and bug fixes, we can present the following big whats-new thingies (with an imaginary neon-red blinking what's new sign): NEW STUFF: * Directory Watching: Every 5 seconds, we wake up to check a directory for new torrents. A .torrent file dropped into the directory will subsequently be downloaded automatically. * Support multiple torrents at once: If given multiple torrents on the command-line the client will now download all of them. * Download rarest piece first: Use the psqueue package for getting access to a priority search queue and use it to maintain a histogram over rare pieces. Always try to get the rarest piece first. * Change all use of CML to STM: Overall, it looks like this change improved the code. There are still some need for cleanup after the bomb was thrown, but it does look like it will turn out positively. If CML put you off the track for hacking on Combinatorrent, you can now do it with STM! * Introduce RTS benchmarks: See http://jlouis.github.com/combinatorrent where we plot SVG-based sparklines for various key RTS parameters over time (works best in Opera or Chrome). A slow-changing regression has a much harder time hiding now. * Fix issue #3: A grave bug in the Piece Manager made the assertions fail when endgaming. Fix. -- J.

2 2

ghc package problem
by Phyx 17 Apr '10

17 Apr '10

Hi all, I'm having a rather weird problem, Ghc-pkg list shows a package installed but doing a ghc -make will give an error saying that It can't find a module that's supposed to be in the installed package. And cabal install always reinstall all the dependencies on every change. It doesn't seem to see that they're already installed.

4 15

GHC Api typechecking
by Phyx 17 Apr '10

17 Apr '10

Hi all, I was wondering if someone knows how to do the following: Im looking to typecheck a string using the GHC Api, where I run into problems is that I need to construct a Target, but the TargetId only seem to reference physical files. Ofcourse I can write the string to a file and typecheck that file, but I would like to do it all in memory and avoid IO if possible. Does anyone know if this is possible? For the record Im trying to create the target as follows createTarget :: String -> IO Target createTarget content = do clock <- getClockTime buffer <- stringToStringBuffer content return $ Target { targetId = TargetModule (mkModuleName "string:internal") ß problem , targetAllowObjCode = True , targetContents = Just (buffer,clock) } typeCheckStringOnly :: String -> IO (ApiResults Bool) typeCheckStringOnly contents = handleSourceError processErrors $ runGhc (Just libdir) $ do dflags <- getSessionDynFlags setSessionDynFlags dflags target <- liftIO $ createTarget contents addTarget target load LoadAllTargets let modName = mkModuleName "string:internal" ß problem again, dont know how to create the dependency graph then. graph <- depanal [modName] True (\a->setSession $ a { hsc_mod_graph = graph }) =<< getSession value <- fmap typecheckedSource (typeCheck modName) return $ ApiOk True Cheers, Phyx

1 0

Re: [Haskell] Monads Terminology Question
by Conor McBride 17 Apr '10

17 Apr '10

Hi (Redirecting to cafe, for general chat.) On 12 Apr 2010, at 01:39, Mark Snyder wrote: > Hello, > > I'm wondering what the correct terminology is for the extra > functions that we define with monads. For instance, State has get > and put, Reader has ask and local, etc. Is there a good name for > these? Yes. Indeed, quite a lot of energy has been expended on the matter. It's worth checking out work by Plotkin and others on "Algebraic Effects" (often transmitted around Haskell-land by helpful citizens like Dan Piponi and Heinrich Apfelmus). This work distinguishes two kinds of "extra function": operations (e.g. get, put, ask, throwError, etc) and control operators (local, catchError, etc). *Operations* have types like s1 -> ... sn -> M t where the s's and t are thought of as "value" types, and M is yer monad. You can think of M as describing an "impure capability", permitting impure functions on values. You might even imagine specifying M's collection of operations by a signature, with this made up notation. sig M where f s1 ... sn :: t Note that I'm careful to mark with :: where the inputs stop and the outputs start, as higher-order functions make this ambiguous. For example sig State s where get :: s put s :: () sig Reader r where ask :: r sig Maybe where throw :: a Many popular monads can be characterized exactly by the signature of their operations and the equational theory those operations must obey (e.g. laws like put s >> get >>= f == put s >> f s). The point of these monads is to deliver the capability specified by the operations and equations. The similiarity between the signatures above and the typeclasses often declared to support monadic functionality is no coincidence. Note that every (set of) signature(s) induces a datatype of command-response trees whose nodes are labelled with a choice of operation and inputs, whose edges are labelled with outputs, and whose leaves carry return values. Such a tree represents a "client strategy" for interacting with a server which offers the capability, at each step selecting an operation to perform and explaining how to continue as a function of the value returned. The equational theory of the operations induces an equivalence on strategies. Command-response trees up to operational equivalence give the most general implementation of the specified monad: return makes leaves, >>= pastes trees together, and each operation creates a node. The monad comes from its operations! But what of local, catchError, and other such things? These are *control operators*, and they operate on "computations", with types often involving resembling a -> (b -> M c) -> M d Typically, the job of a control operator is to make local changes to the meaning of the operations in M's signature. A case in point is "local", whose job is to change the meaning of "ask". It's really shadowing one reader capability with another. Similarly, catchError can be thought of as offering a local exception. Old LISPheads (like me!) might think of operations as EXPRs and control operators as FEXPRs. Haskell does a neat job of hiding the distinction between the two, but it may be conceptually helpful to dig it out a bit. Control operators don't give rise to nodes in command-response trees; rather, they act as tree transformers, building new strategies from old. I could start a pantomime about why operations are heroes and control operators are villains, but I won't. But I will suggest that characterising monads in terms of the operations and/or control operators they support is a useful (and increasingly modular) way to manage effects in programming. After all, most end-user applications effectively equip a bunch of user-operations with a semantics in terms of system-operations. All the best Conor

5 7

Weird behaviour with positional parameters in HDBC-mysql
by Martijn van Steenbergen 17 Apr '10

17 Apr '10

Dear café (CC John and Chris), I'm having some trouble with getting positional parameters in HDBC-mysql to work. Most of the time they work fine, but sometimes (and often enough that it's a serious bother) the parameters don't reach the server correctly. Let me first describe my setup: * Ubuntu 8.10 intrepid (inside VMWare Fusion on Mac OS X Snow Leopard) * mysql Ver 14.12 Distrib 5.0.67, for debian-linux-gnu (i486) using readline 5.2 * ghc 6.10.3 * HDBC-2.2.4 * HDBC-mysql-0.6.1 I have a database with a single, empty table: > CREATE TABLE `foo` ( > `id` int(11) NOT NULL auto_increment, > PRIMARY KEY (`id`) > ) ENGINE=MyISAM AUTO_INCREMENT=1031 DEFAULT CHARSET=latin1 And I have the following Haskell program: > import Database.HDBC > import Database.HDBC.MySQL > import Data.List > import Data.Traversable > import Control.Monad > > main :: IO () > main = do > conn <- connectMySQL defaultMySQLConnectInfo > { mysqlUnixSocket = "/var/run/mysqld/mysqld.sock" > , mysqlUser = "root" > , mysqlDatabase = "db" > } > for [1..1000] $ \n -> do > let sql = "select id from foo where id in (" > ++ intercalate "," (replicate (fromInteger n) "?") ++ ")" > stmt <- prepare conn sql > execute stmt (map SqlInteger [1..n]) > finish stmt > return () The program produces no output. All 1000 queries that are sent to the database are expected to return 0 results, since table foo is empty. This is all fine. The problem is in the queries that the server receives. If I turn on full logging, I expect all queries to have a "where id in (...)" where there is an increasing list of integers within the parentheses. But when examining the logs, some are small negative numbers, NULLs or just other numbers: > 506 Execute [980] select id from foo where id in > (1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26, > 27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,824,825,826,827,828,829, > 830,831,832,833,834,835,836,837,838,839,840,841,842,843,844,845,846,847, > 848,849,850,851,852,853,854,855,856,857,858,859,860,861,862,863,864,865, > 866,867,868,869,870,871,872,873,874,875,876,877,878,879,880,881,882,883, > 884,885,886,887,888,889,890,891,892,893,894,895,896,897,898,899,900,901, > 902,903,904,905,906,907,908,0,0,34359738368,-5210566582833489576,NULL, > NULL,0,34359738368,-5210565758199768552,NULL,NULL,0,34359738368, > -5210564933566047528,NULL,NULL,0,34359738368,-5210564108932326504,NULL, > NULL,0,34359738368,-5210563284298605480,NULL,NULL,0,34359738368, > -5210562459664884456,NULL,NULL,158,159,160,161,162,(snip) The server then faithfully executes the statement and returns the results expected for the query it reports to have received, which (in a table with actual data) is not what the Haskell program was expecting. If I compile the program once, the results vary per run. If I don't use positional parameters but instead render the full SQL query myself, there is no problem. Is this a problem in HDBC or in HDBC-mysql? What's going on here? How do I fix or work around this? Many thanks, Martijn.

3 4

Simple binary-protocol through network test
by Yves Parès 16 Apr '10

16 Apr '10

Hello, I'm trying to use the packages Network and Control.Monad.BinaryProtocol together, with a very simple program in which a client sends an operation to the server, which computes the result and sends it back. But the server holds when trying to receive (Server.hs, line 22), whereas the client has actually sent the data (Client.hs, line 17). The server stops to hold only when the client is killed with a Ctrl-C. This should be rather simple, and I can't get to know why it doesn't work. The files are here: http://old.nabble.com/file/p28157883/Client.hs Client.hs http://old.nabble.com/file/p28157883/Server.hs Server.hs http://old.nabble.com/file/p28157883/SharedData.hs SharedData.hs ----- Yves Parès Live long and prosper -- View this message in context: http://old.nabble.com/Simple-binary-protocol-through-network-test-tp2815788… Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com.

4 16

statistics build error
by Keith Sheppard 16 Apr '10

16 Apr '10

Hello, I ran into this error while trying to install statistics. Does this indicate that I need to upgrade my GHC before I can install? kss@CATSKILL:~/projects/> cabal update Downloading the latest package list from hackage.haskell.org kss@CATSKILL:~/projects/> cabal install statistics Resolving dependencies... Downloading erf-1.0.0.0... Configuring erf-1.0.0.0... Preprocessing library erf-1.0.0.0... Building erf-1.0.0.0... [1 of 1] Compiling Data.Number.Erf ( src/Data/Number/Erf.hs, dist/build/Data/Number/Erf.o ) ar: creating archive dist/build/libHSerf-1.0.0.0.a Installing library in /Users/kss/.cabal/lib/erf-1.0.0.0/ghc-6.10.1 Registering erf-1.0.0.0... Reading package info from "dist/installed-pkg-config" ... done. Writing new package config file... done. Downloading primitive-0.2.1... Configuring primitive-0.2.1... Preprocessing library primitive-0.2.1... Building primitive-0.2.1... [1 of 7] Compiling Data.Primitive.MachDeps ( Data/Primitive/MachDeps.hs, dist/build/Data/Primitive/MachDeps.o ) [2 of 7] Compiling Control.Monad.Primitive ( Control/Monad/Primitive.hs, dist/build/Control/Monad/Primitive.o ) [3 of 7] Compiling Data.Primitive.Types ( Data/Primitive/Types.hs, dist/build/Data/Primitive/Types.o ) [4 of 7] Compiling Data.Primitive.Array ( Data/Primitive/Array.hs, dist/build/Data/Primitive/Array.o ) [5 of 7] Compiling Data.Primitive.ByteArray ( Data/Primitive/ByteArray.hs, dist/build/Data/Primitive/ByteArray.o ) Data/Primitive/ByteArray.hs:57:27: Not in scope: `newAlignedPinnedByteArray#' cabal: Error: some packages failed to install: mwc-random-0.5.0.0 depends on primitive-0.2.1 which failed to install. primitive-0.2.1 failed during the building phase. The exception was: exit: ExitFailure 1 statistics-0.5.1.0 depends on primitive-0.2.1 which failed to install. vector-0.5 depends on primitive-0.2.1 which failed to install. vector-algorithms-0.3.1 depends on primitive-0.2.1 which failed to install. Thanks, Keith -- keithsheppard.name

2 2

Re: [Haskell-cafe] Floyd Warshall performance (again)
by John Lato 16 Apr '10

16 Apr '10

> From: Mathieu Boespflug <mboes(a)tweag.net> > > Dear haskell-cafe, > > I implemented the Floyd Warshall algorithm for finding the shortest > path in a dense graph in Haskell, but noted the performance was > extremely poor compared to C. Even using mutable unboxed arrays it was > running about 30 times slower. I rewrote the program several times, to > arrive at the following definition: > My results are basically the same as Max's, but if I change > #define SIZE 1500 to > sIZE = 1500 and all references from "SIZE" to "sIZE", something ... changes. A lot. MacBook-1:Desktop johnlato$ time ./FW Allocating ... Allocating ... done real 0m0.027s user 0m0.013s sys 0m0.014s I can't figure this out at all. Can anyone else confirm, or offer a good explanation? John

4 4

Floyd Warshall performance (again)
by Mathieu Boespflug 16 Apr '10

16 Apr '10

Dear haskell-cafe, I implemented the Floyd Warshall algorithm for finding the shortest path in a dense graph in Haskell, but noted the performance was extremely poor compared to C. Even using mutable unboxed arrays it was running about 30 times slower. I rewrote the program several times, to arrive at the following definition: module Main where import Data.Array.Base import Data.Array.ST import Data.Array.Unboxed import Control.Monad.ST import Control.Monad import System.IO import GHC.Exts #define SIZE 1500 -- Actual graph is immaterial for this benchmark. graph = [] shortestPath :: [(Int, Int, Int)] -> UArray Int Int shortestPath g = runSTUArray $ do let mnew = newArray (0, SIZE * SIZE) 1 mread arr i j = unsafeRead arr (i * SIZE + j) mwrite arr i j x = unsafeWrite arr (i * SIZE + j) x unsafeIOToST $ hSetBuffering stdout LineBuffering unsafeIOToST $ putStrLn "Allocating ..." pm <- mnew unsafeIOToST $ putStrLn "Allocating ... done" let loop1 SIZE = return () loop1 k = let loop2 SIZE = return () loop2 i = let loop3 SIZE = return () loop3 j = do xij <- mread pm i j xik <- mread pm i k xkj <- mread pm k j mwrite pm i j (min xij (xik + xkj)) loop3 (j + 1) in loop3 0 >> loop2 (i + 1) in loop2 0 >> loop1 (k + 1) loop1 0 return pm main = shortestPath graph `seq` return () These 3 nested loops run a lot faster than generating a list of IO actions and then running them using sequence_. But still this program runs 3 times slower than it's C counterpart: $ time ./FloydWarshall Allocating ... Allocating ... done ./FloydWarshall 45,75s user 0,09s system 97% cpu 46,815 total $ time ./a.out ./a.out 14,92s user 0,03s system 98% cpu 15,230 total The source code for the C program is given below. Using the vector package instead of GHC's inbuilt ST arrays doesn't improve the situation any, neither does compiling via C. Any idea why the rather unidiomatic Haskell program is still 3 times slower, despite using unchecked indexing and the fact that the three nested tail recursive functions above shoud behave exactly as the 3 nested for loops in the C program? Best regards, Mathieu

4 5