- Haskell-Cafe - Haskell.org

ANNOUNCE: Haskell Communities and Activities Report (20th ed., May 2011)
by Janis Voigtländer 18 May '11

18 May '11

On behalf of all the contributors, I am pleased to announce that the Haskell Communities and Activities Report (20th edition, May 2011) is now available in PDF and HTML formats: http://haskell.org/communities/05-2011/report.pdf http://haskell.org/communities/05-2011/html/report.html Many thanks go to all the people that contributed to this report, both directly, by sending in descriptions, and indirectly, by doing all the interesting things that are reported. I hope you will find it as interesting a read as I did. If you have not encountered the Haskell Communities and Activities Reports before, you may like to know that the first of these reports was published in November 2001. Their goal is to improve the communication between the increasingly diverse groups, projects, and individuals working on, with, or inspired by Haskell. The idea behind these reports is simple: Every six months, a call goes out to all of you enjoying Haskell to contribute brief summaries of your own area of work. Many of you respond (eagerly, unprompted, and sometimes in time for the actual deadline ;-) to the call. The editor collects all the contributions into a single report and feeds that back to the community. When I try for the next update, six months from now, you might want to report on your own work, project, research area or group as well. So, please put the following into your diaries now: ========================================== End of October 2011: target deadline for contributions to the November 2011 edition of the HC&A Report ========================================== Unfortunately, many Haskellers working on interesting projects are so busy with their work that they seem to have lost the time to follow the Haskell related mailing lists and newsgroups, and have trouble even finding time to report on their work. If you are a member, user or friend of a project so burdened, please find someone willing to make time to report and ask them to "register" with the editor for a simple e-mail reminder in October (you could point me to them as well, and I can then politely ask if they want to contribute, but it might work better if you do the initial asking). Of course, they will still have to find the ten to fifteen minutes to draw up their report, but maybe we can increase our coverage of all that is going on in the community. Feel free to circulate this announcement further in order to reach people who might otherwise not see it. Enjoy! Janis Voigtlaender <hcar at haskell.org>

1 0

GPL, LGPL, and GHC Linking
by Jon Kristensen 18 May '11

18 May '11

Hello everyone! I just want to let you know that I have written a blog post about my thoughts about the issue of GHC linking and the General Public Licenses. I welcome any feedback or discussions around it. You can find the post at http://www.jonkristensen.com/gpl-lgpl-and-ghc-linking/. Warm regards, Jon Kristensen

1 0

Open CV or alternate image processing library for Haskell on windows?
by Gregory Guthrie 18 May '11

18 May '11

I wanted to look into using Haskell for an introductory Image Processing class, but the main package used for such things (OpenCV) does not appear to be available for windows systems. Is there some other good option for image processing in Haskell, or has anyone ported openCV to a windows Leksah environment? -------------------------------------------

5 9

Re: [Haskell-cafe] ANNOUNCE: iterIO-0.1 - iteratee-based IO with pipe operators
by Simon Marlow 18 May '11

18 May '11

On 12/05/2011 16:04, David Mazieres expires 2011-08-10 PDT wrote: > At Thu, 12 May 2011 09:57:13 +0100, > Simon Marlow wrote: >> >>> So to answer my own question from earlier, I did a bit of >>> benchmarking, and it seems that on my machine (a 2.4 GHz Intel Xeon >>> 3060, running linux 2.6.38), I get the following costs: >>> >>> 9 ns - return () :: IO () -- baseline (meaningless in itself) >>> 13 ns - unsafeUnmask $ return () -- with interrupts enabled >>> 18 ns - unsafeUnmask $ return () -- inside a mask_ >>> >>> 13 ns - ffi -- a null FFI call (getpid cached by libc) >>> 18 ns - unsafeUnmask ffi -- with interrupts enabled >>> 22 ns - unsafeUnmask ffi -- inside a mask_ >> >> Those are lower than I was expecting, but look plausible. There's room >> for improvement too (by inlining some or all of unsafeUnmask#). > > Do you mean inline unsafeUnmask, or unmaskAsyncExceptions#? I tried > inlining unsafeUnmask by writing my own version and giving it the > INLINE pragma, and it didn't affect performance at all. Right, I meant inlining unmaskAsyncExceptions#, which would require compiler support. >> However, the general case of unsafeUnmask E, where E is something more >> complex than return (), will be more expensive because a new closure for >> E has to be created. e.g. try "return x" instead of "return ()", and >> try to make sure that the closure has to be created once per >> unsafeUnmask, not lifted out and shared. > > Okay. I'm surprised by getpid example wasn't already stressing this, > but, indeed, I see a tiny difference with the following code: > > ffi>>= return . (1 +) -- where ffi calls getpid > > 13 ns - no unmasking > 20 ns - unsafeUnmask when not inside _mask > 25 ns - unsafeUnmask when benchmark loop in inside one big _mask > > So now we're talking about 28 cycles or something instead of 22. > Still not a huge deal. Ok, sounds reasonable. >> There are no locks here, thanks to the message-passing implementation we >> use for throwTo between processors. > > Okay, that sounds good. So then there is no guarantee about ordering > of throwTo exceptions? That seems like a good thing since there are > other mechanisms for synchronization. What kind of ordering guarantee did you have in mind? We do guarantee that in throwTo t e1 throwTo t e2 Thread t will receive e1 before e2 (obviously, because throwTo is synchronous and only returns when the exception has been raised). Pending exceptions are processed in LIFO order (for no good reason other than performance), so there's no kind of fairness guarantee of the kind you get with MVars. One thread doing throwTo can be starved by others. I don't think that's a serious problem. Cheers, Simon

4 12

Re: [Haskell-cafe] Is fusion overrated?
by Facundo Domínguez 18 May '11

18 May '11

> which is capable of producing elements one-by-one. So the whole thing > probably should run in constant space as well. Besides reducing the amount of GC calls, performance would also improve because the GC calls that remain are cheaper. The original program may run in constant space, but the fused program may use even a smaller constant space. Which in turn means that whenever the GC needs to make a pass, it is faster. Facundo

1 0

Is fusion overrated?
by Roman Cheplyaka 18 May '11

18 May '11

If one thinks about Haskell data structures as of ordinary data structures, fusion seems a great deal -- instead of producing intermediate lists and possibly running out of memory, we just run a loop and use constant amount of space. But Haskell data structures are quite different -- they are produced as demanded. Consider the example from the Stream Fusion paper[1]: f :: Int → Int f n = sum [ k ∗ m | k ← [1..n], m ← [1..k ] ] Assuming the sum is a strict left fold, it consumes elements of lists one-by-one and runs in constant space. The list part can be transformed to foldr (++) [] $ map (\k -> map (\m -> k*m) [1..k]) [1..n] which is capable of producing elements one-by-one. So the whole thing probably should run in constant space as well. Of course I don't claim that fusion is useless -- just trying to understand the problem it solves. Are we saving a few closures and cons cells here? [1] Stream Fusion. From Lists to Streams to Nothing at All. Duncan Coutts, Roman Leshchinskiy, Don Stewart. -- Roman I. Cheplyaka :: http://ro-che.info/ Don't worry what people think, they don't do it very often.

6 5

Haskell lib in non-Haskell program
by Eric Y. Kow 18 May '11

18 May '11

Hi all, I'd like to share my experiences packaging a Haskell program so that it can be used as a library by a non-Haskell application. I hope somebody might be able to (A) confirm that I'm doing the right thing or better still (B) suggest some corrections and/or simplifications. My objectives are: 1. For the non-Haskell application to use my library 2. ... without any Haskell infrastructure on the machine I think this means I want a shared library [1,2] although I'm a little bit fuzzy on the issue. I am not particularly concerned about the size of my executable or libraries. Including the Haskell RTS and my package deps is acceptable if that's what it takes to meet my primary and secondary objectives. I am targeting Windows and MacOS X. For Windows, my approach is to build a giant DLL with all the Haskell stuff statically linked in. For MacOS X I am using the `-dynamic` flag to compile all my dependencies (actually, I just set shared: True in cabal config) as well as the `-shared` flag. I am also now using GHC 7.0.3 with the latest (at the time of this writing) Haskell Platform (2011.2.0.1) [1] http://hackage.haskell.org/trac/ghc/wiki/SharedLibraries [2] http://www.haskell.org/ghc/docs/latest/html/users_guide/using-shared-libs.h… The whole thing ---------------------------------------------------------------------- You can see my efforts so far with darcs get --lazy http://code.haskell.org/GenI cd GenI cabal install cd geniwrapper make MacOS X users should be able to download http://erickow.com/tmp/MinimalGenI-OSX.tar.gz and run a test-mac.sh script which compiles a program with GCC using GenI, runs it and outputs some JSON string. Exposing some C functions and using them ---------------------------------------------------------------------- So far, I think I know how to expose some of library as C functions and to compile a little program written in C that uses this library. ghc --make -fvia-C MinimalGenI # provides some C functions via FFI exports ghc -c StartEnd.c # http://www.haskell.org/ghc/docs/latest/html/users_guide/win32-dlls.html ghc test-c.c MinimalGenI.o MinimalGenI_stub.o StartEnd.o\ -package GenI -package utf8-string\ -o test-c.o I notice that -fvia-C is going away, but I'm going to ignore this fact for now and assume/hope that doing without it will be easy Building a Windows shared library ---------------------------------------------------------------------- As I understand it, it is possible to build a Windows DLL, albeit one containing the Haskell RTS and all my Haskell libraries statically linked in. Users can download the DLL and link it to their Visual Basic programs without having to touch GHC. The extra work involved looks like this: ghc MinimalGenI.o MinimalGenI_stub.o StartEnd.o\ -package GenI -package utf8-string\ -shared -o MinimalGenI.dll ghc test-c.c MinimalGenI.dll -o test-c2 Note that building test-c.c with ghc does not meet my second objective of being able to combine Haskell lib with non-Haskell program sans Haskell infrastructure. However, I think I know how to do it for MacOS X and that the procedure would be more or less the same. I have not looked into it. [3] http://www.haskell.org/ghc/docs/latest/html/users_guide/win32-dlls.html Building a MacOS X shared library ---------------------------------------------------------------------- For MacOS X, I would almost like to generate a gigantic file like the Windows DLL. But maybe that's the wrong thing to want. I do almost exactly the same thing, except that instead of distributing a single giant file, I track down a whole bunch of dependencies and copy over the dylib files for them. Note that as a prerequisite, this requires reinstalling a lot of packages with --enable-shared (I just edit ~/.cabal/config and set shared to True). ghc MinimalGenI.o MinimalGenI_stub.o StartEnd.o\ -package GenI -package utf8-string\ -dynamic -shared -o MinimalGenI.dylib ghc test-c.c MinimalGenI.dylib -o test-c2 As an alternative to building with GHC, I can compile test-c.c and link with MinimalGenI.dylib directly with GCC. It's a little bit hairy. What I did was to run ghc with -v3 to see what gcc commands it was running, clean them up and package them in a script (test-mac.sh attached). It's a little bit voodoo-ish, a lot of flags that I don't really understand the significance of, and some libraries I'm not 100% clear on why I need the static versions for. Also I find it slightly odd that I can't seem to just merge the first two steps of the process, producing/assembling assembly code for test.c. In any case, it seems to work... Packaging the MacOS X shared library ---------------------------------------------------------------------- I'm trying to distribute my Haskell library in a way that does not require the user to install anything beyond XCode. To do this, I track down the dynamic libraries for all the packages I'm using and their dependencies. Unfortunately, the way I did it was basically trial and error (find all dylib and hand remove the ones that seem totally irrelevant; get a friend to test it on a raw machine). Rather than the dumb approach used in the attached dist-mac.sh, it should be relatively easy to write a small Haskell script to recurse down the package dependency tree and tell me exactly which files I need to copy. Maybe such a tool exists already. If this procedure is right, and if somebody should write such a script (maybe already part of the cabal-macosx package?) then I bet creating app bundles for your MacOS X Haskell GUI tools should be quite straightforward as well. Thanks for reading! -- Eric Kow <http://erickow.com>

1 0

Random thoughts about typeclasses
by Robert Clausecker 18 May '11

18 May '11

Hello! I found out, that GHC implements typeclasses as an extra argument, a record that stores all functions of the typeclass. So I was wondering, is there a way (apart from using newtype) to pass a custom record as the typeclass record, to modify the behavior of the typeclass? I thought about something like this: f :: Show a => [a] -> String f = (>>= show) -- So, f becomes something like this? __f :: ClassShow a -> [a] -> String __f (ClassShow __show) x = x >>= __show -- And if I call the function, it looks somewhat like this: g :: [Int] -> String g = f __g = __f instanceShowInt -- But is it possible to do something like this? g2 = __f (ClassShow (return . fromEnum)) Tis is just a random thought, some compilers like JHC implement them by another way. But would this theoretically be possible? Yours, FUZxxl

4 3

Exception for NaN
by Grigory Sarnitskiy 18 May '11

18 May '11

How do I make my program stop whenever it gets somewhere NaN as a result during a calculation? If there is no appropriate flag for ghc maybe there exist flags for C to use in optc. I don't want NaN to propagate, it is merely stupid, it should be terminated.

13 19

Re: [Haskell-cafe] why doesn't ghc give you a type signature that works ?
by Ivan Lazar Miljenovic 18 May '11

18 May '11

(Re-CC-ing the list) On 18 May 2011 15:16, <briand(a)aracnet.com> wrote: > On Wed, 18 May 2011 13:52:46 +1000 > Ivan Lazar Miljenovic <ivan.miljenovic(a)gmail.com> wrote: > >> What's happening is that the type involves the FFTWReal class. >> However, this isn't immediately visible to ghci as it isn't in scope, >> but it can figure out that it's found in Math.FFT.Base. As such, you >> need to import that module as well (as opposed to just the Math.FFT >> class you have already imported). >> > > Well now, that makes perfect sense doesn't it ? > > I think my problem is that I think if I import the top-level module then the sub-modules should also be visible, provided I use the full name, or in the case of a qualified name, something link FFT.Base.FFTWReal. > > Clearly this is not the case, but why not ? Is it just the convention that haskell chose, or is there a reason for it ? Consider Data.ByteString vs Data.ByteString.Lazy; the have the exact same API but for completely different types. Also, it's quite often for package foo to have module Data.Foo, and then module bar extends it with module Data.Foo.Bar; should importing Data.Foo also bring in Data.Foo.Bar? What happens if bar isn't installed? -- Ivan Lazar Miljenovic Ivan.Miljenovic(a)gmail.com IvanMiljenovic.wordpress.com

1 0