- Haskell-Cafe - Haskell.org

RE: Library Infrastructure code & help wanted
by Simon Marlow 22 Jan '04

22 Jan '04

> 2) I also would like someone to help out with the package > configuration file IO. Right now, it uses HaXmL, with no generated > code. Someone who has seen HaXmL before could give me advice about > how my code could be better, submit patches, or rewrite it. As > discussed on libraries(a)haskell.org, I'm open to not using XML if > someone wants to redo what I've got there (not a huge task). This > parsing code is in Distribution.Package. I do want something more > flexible than straight-up read/show. I rather think that this should be considered a "sample implementation" of the Distribution.Package API. A consumer of the API should neither know nor care about the on-disk representation of the package database. I think we should specify the form of the library meta-data, at least for the default build system (by meta-data I mean the form of toolInfo in the example in sec. 3.4 of the doc). We can take the approach that we specify a very simple form, which works for most straightforward Haskell-only libraries, or we could define a rich form which will provide enough for all but the most complex library packages. I think we should start with the simple form - complex libraries will probably continue to use their own build systems for a while yet. The idea, as I understand it, is that we're going to provide two main APIs: (b) The high-level API (Distribution.Core in the example). Used by simple Setup.lhs scripts to provide generic build, install, and package registration services. Consumes Distribution.Package internally. (a) The low-level API (Distribution.Package). This is for custom build-systems which provide their own build & install, so that they can register a package. Also used by the command-line tool (haskell-config?). There may also be a mid-level API (Distribution.Build, Distribution.Install), but I'm not sure I can see a clear use for that at the moment. If not, then we don't need to specify these APIs, and they can be internal to the default build system. Also, I really think we should clarify the situation as regards multiple Haskell implementations, which seems to be glossed over at the moment. Having a single package database for all Haskell implementations seems a hard thing to implement: when GHC is installed, it will have to check whether the database and haskell-config already exist, and similarly for the other Haskell implementations. When a Haskell system is deinstalled, it has to decide whether to de-isntall haskell-config and the package database. I'd rather just admit that the Haskell systems are separate, and each provides its own implementation of Distribution.Package. Yes, this means that if you run Setup.lhs with Hugs it is different than running it with GHCi: if I want to install a package for GHC I have to run the Setup.lhs script using GHC. I believe that's the right way to go, and it's much easier from our point of view: all we have to do is agree on the Distribution.* APIs, we don't have to have any clever install/deinstall magic to deal with shared components. Cheers, Simon

2 1

RE: MVar swap
by Simon Marlow 21 Jan '04

21 Jan '04

> I took a look at the implementation of the MVar function swap > and I am > just woundering where the atomicy comes from. > > -- |Swap the contents of an 'MVar' for a new value. > swapMVar :: MVar a -> a -> IO a > swapMVar mvar new = > block $ do > old <- takeMVar mvar > putMVar mvar new > return old > > Is it really just the Exception block, that makes it not > interruptable? swapMVar is only atomic with respect to other well-behaved threads. By well-behaved I mean that they only do takeMVar followed by putMVar on the MVar, never a putMVar on its own. Does that answer your question? Cheers, Simon

1 0

MVar swap
by Georg Martius 20 Jan '04

20 Jan '04

Hi, I took a look at the implementation of the MVar function swap and I am just woundering where the atomicy comes from. -- |Swap the contents of an 'MVar' for a new value. swapMVar :: MVar a -> a -> IO a swapMVar mvar new = block $ do old <- takeMVar mvar putMVar mvar new return old Is it really just the Exception block, that makes it not interruptable? Thanks! Georg

1 0

GHC kit for Windows
by Graham Klyne 20 Jan '04

20 Jan '04

I note that the GHC kit for Windows includes some of the (I presume) MinGW binutil components, but does not include 'ar'. I find that 'ar' is needed for building HaXml. May I suggest that 'ar' also be included in this kit? #g ------------ Graham Klyne For email: http://www.ninebynine.org/#Contact

1 0

Library Infrastructure code & help wanted
by Isaac Jones 20 Jan '04

20 Jan '04

Greetings, I'm looking for some help on the library infrastructure project. The project and its goals are explained here: http://www.haskell.org/libraryInfrastructure/ And the code is here: http://www.haskell.org/libraryInfrastructure/code.html In brief, the Library Infrastructure Project is an effort to provide a framework for developers to more effectively contribute their software to the Haskell community. I'm hoping to not have to implement this entire proposal myself, and I'm very flexible about implementation details. I've made some concrete progress, and I just mirrored my code repository on my web site, so now is the time to get involved! Specifically, at the moment, I want a couple of things: 1) A version parser. I implemented a parser which parses versions that are made of dates, or simple numeric versions. I'd like it to do more kinds of versions, though. Distribution.Version is attached. What is this version stuff used for? Well for one thing, you can use it to say something like: ./haskell-config --user list "MyPackage>1.2" where "MyPackage" is the name of a package. This will list all of the user-installed packages whose versions are greater than 1.2. If you look at the HUnit tests at the bottom of the attached file, you'll get a good idea of what it does right now. It uses Parsec. BTW, I understand that the Debian project has some standard version number parser which might give you some ideas. 2) I also would like someone to help out with the package configuration file IO. Right now, it uses HaXmL, with no generated code. Someone who has seen HaXmL before could give me advice about how my code could be better, submit patches, or rewrite it. As discussed on libraries(a)haskell.org, I'm open to not using XML if someone wants to redo what I've got there (not a huge task). This parsing code is in Distribution.Package. I do want something more flexible than straight-up read/show. 3) Someone to make it work with in nhc & hugs. The HaXmL page says that it doesn't work w/ Hugs yet (can anyone tell me why?), so that's one step. Also, at least for the version of NHC I have on my system, Parsec is different. So maybe we have to move away from Parsec, or maybe a different version thereof? It is absolutely vital for this project to be completely cross-platform and cross-compiler, so if we have to avoid depending on such things, I'm very willing to change. 4) Someone to make it work with Windows. Avoid using a filename that's called "Aux" ;) ABOUT THE CODE: It only works with GHC at the moment (see above) :( Right now what I've got is a first stab at Distribution.Package and haskell-config as explained in the proposal [1], though some things may have changed. It's hosted with Arch[2], a newish version control system. You can branch my code on your own machine and submit patches. Arch has some tools to help with patches. I've also provided a tarball. Arch is cool, though. You should try it. The HUnit tests in some of the modules should be pretty illuminating. The interactive help for haskell-config is pretty OK. Get the code here: http://www.haskell.org/libraryInfrastructure/code.html peace, isaac [1] http://www.haskell.org/libraryInfrastructure/proposal/index.html [2] http://regexps.srparish.net/www/

2 1

LIP and the AAP project
by Isaac Jones 20 Jan '04

20 Jan '04

Greetings, Now that I have at least gotten something started on the Library Infrastructure Project, I'm going to bend some of my energy toward the build/install issues for Haskell modules. The plan at this point, which is summarized on the project web site[1] is to attempt to leverage HMake's usefulness as a non-compiler-specific build system. Since I wrote the proposal, I've been made aware of the AAP project[2]. Does anyone have any experience with this? Does anyone have a sense for how it might fit in with LIP? I know that Peter Simons seems to think it might. I haven't read through all of the documentation, but I suspect it might help us out with the CPAN-style distribution stuff, and perhaps the building stuff. One big objection I have with using it would be that Haskell modules would depend on Python (!), but I thought I'd let everyone know about it anyway. Seems like an interesting project, and we could quite possibly learn something from it. peace, isaac [1] http://www.haskell.org/libraryInfrastructure/ [2] http://www.a-a-p.org/

1 0

Haskell binaries for AIX?
by Dean Herington 16 Jan '04

16 Jan '04

Does anyone out there have a prebuilt Haskell system for AIX? I'm trying to find the easiest route to running a Haskell interpreter on AIX (5.1 and 5.2). Dean

1 0

FFI preprocessor for GHC
by Gour 16 Jan '04

16 Jan '04

Hi! I'd like to create Haskell bindings for swisseph C library for calculating ephemeris. Which preprocessor would be a suitable for ghc compiler with the ability to run the code both on Linux & Win32? Sincerely, Gour -- Gour gour(a)mail.inet.hr Registered Linux User #278493

3 2

RE: ghc optimization of incremental update
by Simon Marlow 13 Jan '04

13 Jan '04

> Does ghc make any attempt to analyze dataflow and eliminate > unnecessary > copying for incremental updates? I compile the code below > with -O, and > based on the time it takes to run I'd guess it's copying the > array with > every update, even though there's no need in this particular > case to do > that. Perhaps there's a compiler flag or pragma to turn on > the analysis? > > module Main where > import Array > main = print (take 1 (reverse (elems (array_neg (setup > 100000))))) > setup i = listArray (1,i) [1..i] > -- the next three lines are the important part. > array_neg array = array_neg2 (length (elems array)) array > array_neg2 0 array = array > array_neg2 i array = array_neg2 (i-1) (array // [(i,-i)]) > > I realize I can force it to stop copying if I use a Monad, and there > are of course more efficient ways to code this contrived testcase. > I want to know if ghc can make it efficient as it stands. GHC doesn't do this kind of optimisation, no. There is the DiffArray type which provides constant-time array update, although it has been reported to have a high constant factor overhead. Your best bet for algorithms using mutable arrays is to encapsulate it in a state thread with the ST monad. Cheers, Simon

1 0

EDI X12
by Christopher Milton 13 Jan '04

13 Jan '04

Has anyone in the Cafe written EDI X12 generator/parser in Haskell? ===== Christopher Milton cmiltonperl(a)yahoo.com

1 0