March 2009 - Glasgow-haskell-users

Loading plugin with ghc api
by John O'Donnell 23 Mar '09

23 Mar '09

Hi, I would like to use the API for ghc, so that a running program can load a module (Foo.o) and use a function defined in that module. From the documentation available, it seems like thatś possible, but I can´t figure out how to do it. There is an example on the wiki, but the explanation just says the code is equivalent to ghc --make. That example does indeed work, but there is no hint as to whether the running program can then access anything defined in the module it has just loaded. Is there any available documentation on how to load a plugin, or any working example somewhere? Thanks in advance, John O´Donnell

3 2

mkIfThenElse
by Colin Paul Adams 22 Mar '09

22 Mar '09

This function has disappeard from CoreUtils in recent release(s). I am portfing some code which uses this function (just once). What should I be looking to use instead? -- Colin Adams Preston Lancashire

2 3

How do I add new dependencies and source files into the build system?
by Colin Paul Adams 22 Mar '09

22 Mar '09

I have added a subdirectory named verify to the compiler directory, and made a change to simplCore to call functions from it. I have managed to get it to compile by editing ghc.cabal in the compiler directory (after some problems with cabal which are now sorted). But I think this is not the correct way to do it, as ./configure rebuilds this file, it seems. Where should I be doing this? -- Colin Adams Preston Lancashire

2 2

Can't find parsec
by Colin Paul Adams 21 Mar '09

21 Mar '09

I am trying to incorporate Dana Xu's static contract checking code into ghc 6.11. I've run into a problem that puzzles me. I am getting the following error message: verify/SimplIface.lhs:16:7: Could not find module `Text.ParserCombinators.Parsec.Language': Use -v to see a list of the files searched for. But a quick find command shows that ./utils/ext-core/Language/Core/ParsecParser.hs has this very same import, so I figure it must be usable within the compiler. What do I have to do to get it to find this module? -- Colin Adams Preston Lancashire

2 2

Another victim of the state hack
by Jules Bean 20 Mar '09

20 Mar '09

I have the impression that it's useful to you to see how often the state hack trips people up, so I've added some more comments to http://hackage.haskell.org/trac/ghc/ticket/2284 ...about how it bit me. This is a real headache. To be frank, controlling lazy evaluation and sharing to get the right complexity for your code can be hard enough at the best of times, without the compiler sometimes ignoring the sharing you try to ask for...

2 1

Adding DPH to HEAD
by Colin Paul Adams 19 Mar '09

19 Mar '09

I tried following the advice on the DPH wiki: ./sync-all --dph get This wouldn't run because of permissions, so I tried putting sh in front of the command. This produced a lot of error messages: ./sync-all: line 3: use: command not found ./sync-all: line 4: use: command not found ./sync-all: line 8: git: command not found ./sync-all: line 8: my: command not found ./sync-all: line 8: chomp: command not found ./sync-all: line 9: git: command not found ./sync-all: line 9: my: command not found ./sync-all: line 9: chomp: command not found ./sync-all: line 10: git: command not found ./sync-all: line 10: my: command not found ./sync-all: line 10: chomp: command not found ./sync-all: line 12: my: command not found ./sync-all: line 13: my: command not found ./sync-all: line 15: syntax error near unexpected token `{' ./sync-all: line 15: `if ($defaultrepo =~ /^...*:/) {' Following a suggestion from Simon Marlow, I instead tried: ./darcs-all --dph get which was more successful. Does the wiki page need updating? But when I then typed make I got: <snip> --constraint="Cabal == 1.7.0" Configuring installPackage-1.0... cabal-bin: The following installed packages are broken because other packages they depend on are missing. These broken packages must be rebuilt before they can be used. package Cabal-1.7.0 is broken due to missing package array-0.2.0.0, containers-0.2.0.0, process-1.0.1.0 make[2]: *** [with-bootstrapping-compiler] Error 1 make[2]: Leaving directory `/home/colin/ghc/utils/installPackage' make[1]: *** [with-bootstrapping-compiler.installPackage] Error 2 make[1]: Leaving directory `/home/colin/ghc/utils' make: *** [stage1] Error 2 What do I do in such circumstances? I tried a cabal install Cabal --rebuild which seemed to work OK, but when i type make again i get the same error as before. -- Colin Adams Preston Lancashire

3 3

RE: [Haskell] Re: indirectly recursive dictionaries
by Simon Peyton-Jones 19 Mar '09

19 Mar '09

[Redirecting to GHC users.] | Tom Schrijvers wrote: | > The cyclic dictionaries approach is a bit fragile. The problem appears to | > be here that GHC alternates exhaustive phases of constraint reduction and | > functional dependency improvement. The problem is that in your example you | > need both for detecting a cycle. The whole thing relies on "spotting a loop", and that's inherently a bit fragile. I don't know of any formal work on the subject, although I bet there is some. GHC's current algorithm does not run functional dependencies sufficiently aggressively, because it treats fundeps a different kind of thing to class constraints. Our new solver, long promised but still in the works, fully integrates type classes with type equalities (fundeps, type functions etc), and so should do a better job here. Roughly speaking, the idea is to combine our ICFP'08 paper [1] with a type-class solver. Since writing the ICFP'08 paper we have found some very useful simplifications; and we also have a new plan for the solving strategy "OutsideIn" [2]. That said, solving recursive problems is not our primary focus right now -- getting it working is -- so I can't promise that it'll do a better job, but I think it will. | It seems we can convince GHC to do constraint reduction and | improvement in the order we desire. The key is to use the | continuation-passing style -- which forces things to happen in the | right order, both at the run-time and at the compile time. Oleg you are a master at persuading GHC's somewhat ad-hoc implementation to dance to your tune. But it'd be better just to make the implementation more complete in the solutions it finds. That's what we are working on now. Simon [1] http://research.microsoft.com/~simonpj/papers/assoc-types/index.htm [2] http://research.microsoft.com/~simonpj/papers/gadt

2 1

Advice wanted on parallel processing
by Colin Paul Adams 19 Mar '09

19 Mar '09

I've just managed to build ghc 6.11 (Thanks Simon). I did this for two reasons, one of which is I want to try to improve the speed of the AI for the Chu Shogi program I am writing by making use of parallel processing. I have a 4-core Xeon runing Fedora Linux 10 (AMD64). I have a repeatable scenario for timings. With ghc 6.10.1, no threaded runtime, the computer takes 51 or 52 seconds to move (experimental variation up to 1.2 seconds). With ghc 6.11, it takes the same time. If I switch on the threaded runtime, then it rises slightly (perhaps not significantly - I measure 53-54 seconds). With 2, 3 or 4 processors, I measured (one-off, so not reliable) 65, 55 and 58 seconds respectively. Well, that doesn't really matter, provided I can get the time back with interest by exploiting parallelism. My first thought for an "easy" win is the move generator. At present I have this code: -- | Generate all legal moves from state. -- -- The best moves (according to some heuristic) should come first generate_moves :: (Non_interactive_state, Maybe Move) -> [(Non_interactive_state, Maybe Move)] generate_moves (state, _) = let colour = case is_black_to_play state of True -> Black False -> White pieces' = all_pieces_of_colour (board state) colour unsorted = concatMap (generate_moves_for_piece state) pieces' sorted = sortBy best_move unsorted moves = mapMaybe new_move sorted in {-trace (concat (intersperse "\n" (map (print_move . fromJust . snd) moves)))-} moves where new_move mv = case update_from_move (state, Nothing) mv of Nothing -> Nothing Just (st', Nothing) -> Just (st', Just mv) and my idea was to run the call to generate_moves_for_piece in parallel over the pieces' list. I thought I could do that simply by replacing concatMap with parFlatMap, but discovered I need to supply an extra argument - a Strategy. Not knowing what this should be (where's the best place to read up on this?) I tried specifying rwhnf. My timings then are: -N1 83 seconds -N2 57 seconds -N3 58 seconds -N4 60 seconds so a complete failure. Where should I go from here? -- Colin Adams Preston Lancashire

4 8

A patch to fix getNumberOfProcessors() on OS X
by Austin Seipp 18 Mar '09

18 Mar '09

(Tried sending to cvs-ghc, said it was blocked for moderation because it matched a certain pattern...) Hello, I have a patch to fix the runtime system's getNumberOfProcessors() function to work properly on OS X and detect the number of CPUs. However, when I run validate, these are the final results: OVERALL SUMMARY for test run started at Tue Mar 17 21:34:36 CDT 2009 2318 total tests, which gave rise to 8661 test cases, of which 0 caused framework failures 6662 were skipped 1923 expected passes 74 expected failures 1 unexpected passes 1 unexpected failures Unexpected passes: hpc_ghc_ghci(normal) Unexpected failures: break017(ghci) Are these to be expected on OS X, or did my patch really break something? Patch is attached. Austin

1 0

Bootstrapping the compiler
by Colin Paul Adams 18 Mar '09

18 Mar '09

I have just downloaded a darcs snapshot, pulled patches and followed the instructions at http://hackage.haskell.org/trac/ghc/wiki/Building/QuickStart When I got to do make it didn't work. The tail of the output looks like this: [55 of 55] Compiling Main ( cabal-bin.hs, /home/colin/ghc/libraries/bootstrapping/Main.o ) Linking cabal-bin ... echo "[]" > bootstrapping.conf.tmp cd extensible-exceptions && /home/colin/ghc/libraries/cabal-bin ghc --make -fno-warn-unrecognised-pragmas /home/colin/ghc/libraries/bootstrapping.conf 1.7.0 clean --distpref=dist-bootstrapping unrecognised command: /home/colin/ghc/libraries/bootstrapping.conf (try --help) make[1]: [bootstrapping.conf] Error 1 (ignored) cd extensible-exceptions && /home/colin/ghc/libraries/cabal-bin ghc --make -fno-warn-unrecognised-pragmas /home/colin/ghc/libraries/bootstrapping.conf 1.7.0 configure --distpref=dist-bootstrapping --with-compiler=ghc --make -fno-warn-unrecognised-pragmas --with-hc-pkg=/usr/local/bin/ghc-pkg --package-db=/home/colin/ghc/libraries/bootstrapping.conf.tmp unrecognised command: /home/colin/ghc/libraries/bootstrapping.conf (try --help) make[1]: *** [bootstrapping.conf] Error 1 make[1]: Leaving directory `/home/colin/ghc/libraries' make: *** [stage1] Error 2 I then tried the following: [colin@susannah ghc]$ ls libraries/bootstrapping.conf ls: cannot access libraries/bootstrapping.conf: No such file or directory [colin@susannah ghc]$ ls libraries/bootstrapping.conf.tmp libraries/bootstrapping.conf.tmp It looks to me like all I have to do is to rename libraries/bootstrapping.conf.tmp to libraries/bootstrapping.conf but I don't see why this is happening. -- Colin Adams Preston Lancashire

2 4