December 2009 - Haskell-Cafe

Re: [Haskell-cafe] Re: [Haskell] ANNOUNCE: control-monad-exception 0.5 with monadic call traces
by Henning Thielemann 07 Dec '09

07 Dec '09

klondike schrieb: > Henning Thielemann escribió: > >> It seems again to me, that mixing of (programming) errors and >> exceptions is going on, and I assumed that the purpose of >> control-monad-exception is to separate them in a better way. > You know, could you tell me when using head on an empty list is a > programming error and when it is a exception, I have seen both cases... The case of (head []) is simple: It is a programming error, since the precondition for calling 'head' is that the argument list is non-empty. The caller of 'head' is responsible to check this. If the list comes from user input, then the program part that receives this list from the user is responsible to check for the empty list before calling 'head'. If there is no such check, this is a programming error, and it will not be possible to handle this (like an exception). Before you answer: "What about web servers?", please read on the article I have written recently to sum up the confusion about errors and exceptions: http://www.haskell.org/haskellwiki/Error_vs._Exception

2 2

Do you need Windows USB in Haskell?
by Maurício CA 07 Dec '09

07 Dec '09

Hi, I keep this direct binding to libusb-1.0.x: http://hackage.haskell.org/package/bindings-libusb on top of which Bas maintains a nice USB library: http://hackage.haskell.org/package/usb Work has been done to support libusb-1.0.x in Windows. So, as long as my bindings-libusb works properly with that, Bas' as well as any other library based on it will work too. Problem is: I don't have a Windows machine where I could test this. So, if you need USB in windows, please keep in touch. I wouldn't ask you to write any code, but I need to know what builds and what doesn't. Best, Maurício

2 2

Insert a laziness break into strict ST
by Henning Thielemann 07 Dec '09

07 Dec '09

I have lot of ST actions that shall be bound strictly (they write to a buffer), but somewhere between these actions I like to have a "laziness break". I thought I could do this by temporarily switching to Lazy.ST, but this does not work. It follows a simplified example Prelude> :module Control.Monad.ST.Lazy Prelude Control.Monad.ST.Lazy> runST (Monad.liftM2 (,) (return 'a') (undefined::Monad m => m Char)) ('a',*** Exception: Prelude.undefined Prelude Control.Monad.ST.Lazy> Control.Monad.ST.runST (lazyToStrictST $ Monad.liftM2 (,) (return 'a') (undefined::Monad m => m Char)) *** Exception: Prelude.undefined I hoped to get the first answer also for the second command. It seems that conversion from lazy to strict ST also removes laziness breaks. It seems that I have to stick to unsafeInterleaveIO, but I like to know, why the above method does not work.

2 2

When are undecidables ok?
by Michael Snoyman 07 Dec '09

07 Dec '09

I know this is basically a rewording of a previous e-mail, but I realized this is the question I *really* wanted to ask. We have this language extension UndecidableInstances (not to mention OverlappingInstances), which seem to divide the Haskell camp into two factions: * Hey, GHC said to turn on this flag. Ok! * Undecidables are the devil! I get the feeling the truth lies in the middle. As I understand it (please correct me if I am wrong), the problem with undecidables is that they can create non-terminating instances. However, for certain cases the programmer should be able to prove to him/herself that the instances will terminate. My question is: how can you make such a proof? I've had to cases in particular that made me want undecidables. Both of them, IMO, could be solved by creating new extensions to GHC which would not require undecidables. Nonetheless, for now we only have undecidables at our disposal. The examples are: * Context synonyms (eg, MonadFailure = Monad + Failure). * Subclass function defaulting For an example of the second, a nifty definition of Monad would be: class Applicative m => Monad m where >>= ... return ... pure = return (<*>) = ap fmap = liftM Of course, neither of these is possible in Haskell, so we can use undecidables. How can a programmer prove that a set of instances is, in fact, safe? And if they make a mistake and right a bad set of undecidable/overlapping instances, what's the worst case scenario? Is it a compile-time or run-time error*? Michael * Yes, I mean error.

3 3

What is the rank of a polymorphic type?
by Eugene Kirpichov 06 Dec '09

06 Dec '09

Hello. Consider the type: (forall a . a) -> String. On one hand, it is rank-2 polymorphic, because it abstracts over a rank-1 polymorphic type. On the other hand, it is monomorphic because it isn't actually quantified itself: in my intuitive view, a parametrically polymorphic type has infinitely many instantiations (for example, Int -> Int is an instantiation of forall a . a -> a, and String -> String also is), and this type doesn't have any instantiations at all. Which is correct? Is there really a contradiction? What is the definition of rank of a polymorphic type? -- Eugene Kirpichov Web IR developer, market.yandex.ru

5 8

Haskell-Newbie and Char-Function
by MeAdAstra 06 Dec '09

06 Dec '09

Hi guys, I only started learning Haskell some days ago. Maybe one of you can give me a hint on how to implement a function that needs a character in the range (a,b,...z) and an integer number k and returns the k-next neighbor of the character? For example, fct a 5 would result in output f. Tobias -- View this message in context: http://old.nabble.com/Haskell-Newbie-and-Char-Function-tp26656676p26656676.… Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com.

5 5

ANN: readline-statevar-1.0.1.0
by Krzysztof Skrzętnicki 06 Dec '09

06 Dec '09

Hello I am happy to announce a release of small package called readline-statevar. It's a small wrapping library around readline, which in turn wraps libreadline. The reason I wrote it is because I wasn't happy with the API of readline. It was composed of tons of functions like setX/getX, where there could be just one StateVar instead, yielding OpenGL-kind API. This version aims to match readline-1.0.1.0. (Hence the version number chosen). I wrote it mostly manually, with a little help from my emacs-fu, so there is a chance I missed a thing or two. There is git repo available [1], along with issue tracker [2] and Hackage page [3]. Feel encouraged to: > cabal update > cabal install readline-statevar And give it a try! Best regards Krzysztof Skrzętnicki [1] http://github.com/Tener/haskell-readline-statevar/ [2] http://github.com/Tener/haskell-readline-statevar/issues [3] http://hackage.haskell.org/package/readline-statevar

4 4

ANNOUNCE: error-message
by Gregory Crosswhite 06 Dec '09

06 Dec '09

If there is one thing that we really don't have enough of in Haskell, it is *ways to handle errors*! Thus, I am pleased to announce the release of the "error-message" package to help in filling this, erm, gap. This philosophy behind this package is that it is often better to find out all of the errors that have occured in a computation and report them simultaneously, rather than aborting as soon as the first error is encountered. Towards this end, this package supplies a type of /combinable error messages/ (ErrorMessage in the Data.ErrorMessage module) so that all of the errors from subcomputations can be gathered and presented together. The following provides an example of how these can be used: ================================================== sqrtWithError :: Float -> Either ErrorMessage Float sqrtWithError x | x < 0 = leftErrorMessageText ("Error computing the square root of " ++ (show x) ++ ":") "Square roots cannot be taken of negative numbers." | otherwise = Right (sqrt x) sumWithError :: Either ErrorMessage Float -> Either ErrorMessage Float -> Either ErrorMessage Float sumWithError (Left error1) (Left error2) = Left (error1 `mappend` error2) sumWithError (Left error) _ = Left error sumWithError _ (Left error) = Left error sumWithError (Right value1) (Right value2) = Right (value1 + value2) showSumOrErrorOf :: Float -> Float -> String showSumOrErrorOf x y = case sumWithError (sqrtWithError x) (sqrtWithError y) of Right value -> "The value is " ++ show value Left error -> show . formatErrorMessage $ error ================================================== The result of @showSumOrErrorOf (-1) (-2)@ is the string, Error computing the square root of -1: Square roots cannot be taken of negative numbers. Error computing the square root of -2: Square roots cannot be taken of negative numbers. whereas the result of @showSumOrErrorOf (-1) (-1)@ is the string, Error computing the square root of -1: Square roots cannot be taken of negative numbers. Note how the error message only appears once; this is because the process of combining the error messages automatically eliminates all identical headings under the assumption that they came from the same original computation, as was the case here. Currently, the definition of @sumWithError@ is largely boilerplate. Happily, the Haskell community has done a lot of work to identify patterns such as these and to write libraries that allow us to express them concisely. In particular, a standard trick when working with errors like this is to express the calculation as a 'Monad', such as by using the following definition: ================================================== sumWithError_2 argument1 argument2 = do value1 <- argument1 value2 <- argument2 return (value1 + value2) ================================================== Or, even more concisely: ================================================== sumWithError_3 = liftM2 (+) ================================================== Unfortunately though, neither of these definitions have the same semantics as the original @sumWithError@, as using both we get the following error message for @showSumOrErrorOf (-1) (-2)@: ================================================== Error computing the square root of -1: Square roots cannot be taken of negative numbers. ================================================== That is, we have lost the second of the two error messages. The reason for this is that 'Monad'-style error processing expresses the computation as a sequence, and gives up as soon as it sees any error. In this case of @sumWithError@, however, the evaluation of the second argument can proceed even if there was an error in the first argument. Thus, rather than using a 'Monad' pattern, we use an 'Applicative' pattern: ================================================== sumWithError_4 = liftA2 (+) ================================================== Now both error messages are displayed. Anyway, I hope that someone other than myself finds this pattern to be helpful. :-) Cheers, Greg

5 7

help
by Ezekiel Smithburg 06 Dec '09

06 Dec '09

On Sun, Dec 6, 2009 at 4:53 AM, <haskell-cafe-request(a)haskell.org> wrote: > Send Haskell-Cafe mailing list submissions to > haskell-cafe(a)haskell.org > > To subscribe or unsubscribe via the World Wide Web, visit > http://www.haskell.org/mailman/listinfo/haskell-cafe > or, via email, send a message with subject or body 'help' to > haskell-cafe-request(a)haskell.org > > You can reach the person managing the list at > haskell-cafe-owner(a)haskell.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Haskell-Cafe digest..." > > Today's Topics: > > 1. Re: Low Level Audio - Writing bytes to the sound card? > (Daniel Fischer) > 2. Re: HP + Gtk2hs? (M Xyz) > 3. Re: Low Level Audio - Writing bytes to the sound card? (M Xyz) > 4. Re: Low Level Audio - Writing bytes to the sound card? > (Daniel Fischer) > 5. RE: Low Level Audio - Writing bytes to the sound card? (john lask) > 6. Re: New Hackage category: Error Handling (Henning Thielemann) > 7. Re: Low Level Audio - Writing bytes to the sound card? (M Xyz) > 8. RE: Low Level Audio - Writing bytes to the sound card? (M Xyz) > 9. PortAudio library successfully built on Windows (M Xyz) > 10. When are undecidables ok? (Michael Snoyman) > 11. Re: When are undecidables ok? (Luke Palmer) > 12. Re: When are undecidables ok? (Michael Snoyman) > 13. Re: ANNOUNCE: error-message (Ariel J. Birnbaum) > 14. Re: What is the rank of a polymorphic type? (Eugene Kirpichov) > 15. Re: universal binary version of Haskell Platform? > (Benjamin L.Russell) > 16. Re: What is the rank of a polymorphic type? (Dan Doel) > 17. Re[2]: [Haskell-cafe] Low Level Audio - Writing bytes to the > sound card? (Bulat Ziganshin) > 18. Re: universal binary version of Haskell Platform? > (Heinrich Apfelmus) > 19. Re: Low Level Audio - Writing bytes to the sound card? > (Andrew Coppin) > 20. Re: binding to C libraries on Windows was Low Level Audio - > Writing bytes to the sound card? (Andrew Coppin) > > > ---------- Forwarded message ---------- > From: Daniel Fischer <daniel.is.fischer(a)web.de> > To: haskell-cafe(a)haskell.org > Date: Sun, 6 Dec 2009 01:24:17 +0100 > Subject: Re: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > Am Sonntag 06 Dezember 2009 00:47:38 schrieb M Xyz: > > Daniel, Thank you for your thoughtful reply. I didn't know about those > > flags. The log is fairly long, and as I'm new to Haskell and Cabal it is > > mostly meaningless to me. I see very many incidences of "searching for > ___ > > in path. Cannot find ___ on the path" so maybe this is all as simple as > me > > not setting my environment correctly. > > If you have cpphs, hugs, jhc, greencard etc., it is probably something > about your > environment. If you don't have them, it's clear that they aren't found. > > On the other hand, that doesn't explain > > Using c2hs version 0.16.0 given by user at: C:\Program > Files\Haskell\bin\c2hs.exe > > -- so it finds c2hs, and can apparently run c2hs --version > > ("C:\\Program > Files\\Haskell\\bin\\c2hs.exe",["--include=dist\\build","--cppopts=- > D__GLASGOW_HASKELL__=610","--cppopts=-IC: > > \\A\\install\\programming\\portaudio\\portaudio\\include","--output-dir=dist\\build","-- > output=Sound\\PortAudio\\Base.hs",".\\Sound\\PortAudio\\Base.chs"]) > c2hs.exe: does not exist > C:\Program Files\Haskell\bin\c2hs.exe returned ExitFailure 1 > > -- bang > > To ascertain whether c2hs works at all, can you try to run it manually? > > (cd to an appropriate directory, > > cabal unpack portaudio > > cd portaudio (or whatever, so that Base.chs is found via > .\Sound\PortAudio\Base.chs > > md dist\build > > c2hs.exe --include=dist\build --cppopts=-D__GLASGOW_HASKELL__=610 > --cppopts=-IC: > \A\install\programming\portaudio\portaudio\include --output-dir=dist\build > -- > output=Sound\PortAudio\Base.hs .\Sound\PortAudio\Base.chs > ) > If that works, the problem is somewhere in cabal, otherwise in c2hs, either > way, we'll > know more. > > > > ---------- Forwarded message ---------- > From: M Xyz <functionallyharmonious(a)yahoo.com> > To: haskell-cafe(a)haskell.org, Andrew Coppin <andrewcoppin(a)btinternet.com> > Date: Sat, 5 Dec 2009 16:34:52 -0800 (PST) > Subject: Re: [Haskell-cafe] HP + Gtk2hs? > > > --- On *Sat, 12/5/09, Andrew Coppin <andrewcoppin(a)btinternet.com>* wrote: > > >> Hell, I even followed a C++ guide to Win32 programming and > >> managed to translate an "open a blank window" program to Haskell, and it > >> worked. Maybe somebody just needs to sit down and write a nice binding > >> for doing native GUI stuff under Win32? > > Well there's > http://haskell.org/ghc/docs/latest/html/libraries/Win32/Graphics-Win32.html… opening blank windows. ;) > > This relates to my question a couple days ago about a good cross platform > 2d library. I wonder why there isnt a Haskell library that follows in the > footsteps of Java's Swing - by that I mean, just have a very simple low > level way of opening a window, getting mouse clicks, and an abstracted > drawing layer... and make the GUIs with that. The GUIs are drawn and not OS > widgets. That's exactly why I was looking for something like Cairo. My > interest is in composable fudgets (stealing the word from fudgets). > > This way every platform can share the same GUI library and HP can ship with > the minimal tools to open a window and draw on every system. Plus OS guis > are the wrong thing for Haskell. You need to retain the power to make better > abstractions. > > > > > > ---------- Forwarded message ---------- > From: M Xyz <functionallyharmonious(a)yahoo.com> > To: haskell-cafe(a)haskell.org, Daniel Fischer <daniel.is.fischer(a)web.de> > Date: Sat, 5 Dec 2009 16:49:49 -0800 (PST) > Subject: Re: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > > > --- On *Sat, 12/5/09, Daniel Fischer <daniel.is.fischer(a)web.de>* wrote: > > > > If you have cpphs, hugs, jhc, greencard etc., it is probably something > about your > environment. If you don't have them, it's clear that they aren't found. > > I don't know what those things are. I have nothing but what came with HP > other than installing c2hs today. > > If that works, the problem is somewhere in cabal, otherwise in c2hs, either > way, we'll > know more. > > Alright, I followed the instructions and everything was as you said. I > still get "c2hs.exe does not exist". > > C:\A\install\programming\portaudio\haskell\portaudio-0.0.1>"C:\Program > Files\Has > kell\bin\c2hs.exe" --include=dist\build --cppopts=-D__GLASGOW_HASKELL__=610 > --cp > popts=-IC:\A\install\programming\portaudio\portaudio\include > --output-dir=dist\b > uild --output=Sound\PortAudio\Base.hs .\Sound\PortAudio\Base.chs > c2hs.exe: does not exist > > C:\A\install\programming\portaudio\haskell\portaudio-0.0.1>pause > Press any key to continue . . . > > By the way, I have posted this problem to the c2hs mailing list in case > they have some insight. > > > > > ---------- Forwarded message ---------- > From: Daniel Fischer <daniel.is.fischer(a)web.de> > To: M Xyz <functionallyharmonious(a)yahoo.com> > Date: Sun, 6 Dec 2009 02:34:17 +0100 > Subject: Re: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > Am Sonntag 06 Dezember 2009 01:49:49 schrieb M Xyz: > > I just had another idea. > > dafis@linux-mkk1:~> c2hs -o memyself.hs memyself.chs > c2hs: does not exist > > it's not that c2hs isn't found or something, c2hs doesn't find Base.chs! > > Try installing from the unpacked sources (cd portaudio; cabal install) or > the old- > fashioned way: > > cd portaudio-0.0.1 > > ghc --make Setup > > ./Setup configure --help > (choose your options, prefix, profiling, ...) > > ./Setup configure $OPTIONS > ./Setup build > > if all's well, > > ./Setup haddock > ./Setup install > > (dies for me with > dist/build/Sound/PortAudio/Base.chs.h:1:23: error: portaudio.h: Datei oder > Verzeichnisnicht gefunden > c2hs: Error during preprocessing custom header file > cabal: Error: some packages failed to install: > portaudio-0.0.1 failed during the building phase. The exception was: > exit: ExitFailure 1 > because I don't have portaudio installed) > > > > ---------- Forwarded message ---------- > From: john lask <jvlask(a)hotmail.com> > To: <daniel.is.fischer(a)web.de>, <functionallyharmonious(a)yahoo.com> > Date: Sun, 6 Dec 2009 02:03:32 +0000 > Subject: RE: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > > > I don't know whether this will help you but I just downloaded an built > the haskell portaudio package ... (I had a windows msvc build of > portaudio dll already) the process I used ... ghc 6.10.4, portaudio-19 > > make an import lib for ghc from dll: > pexports libpa19.dll > libpa19.def > dlltool --input-def libpa19.def --output-lib libpa19.a > > edit the .cabal file or use command line flags > > extra-Libraries: pa19 > extra-lib-dirs: c:\portaudio19\lib > include-dirs: c:\portaudio19\include > > runghc setup configure > > make sure you have cpp i.e. the c-preprocessor on your exe path, > otherwise you will get > 'cpp' is not recognized as an internal or external command, > operable program or batch file. > c2hs.exe: Error during preprocessing custom header file > > runghc setup build > > builds ok ...?? > > > > From: daniel.is.fischer(a)web.de > > To: functionallyharmonious(a)yahoo.com > > Subject: Re: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > > Date: Sun, 6 Dec 2009 02:34:17 +0100 > > CC: haskell-cafe(a)haskell.org > > > > Am Sonntag 06 Dezember 2009 01:49:49 schrieb M Xyz: > > > > I just had another idea. > > > > dafis@linux-mkk1:~> c2hs -o memyself.hs memyself.chs > > c2hs: does not exist > > > > it's not that c2hs isn't found or something, c2hs doesn't find Base.chs! > > > > Try installing from the unpacked sources (cd portaudio; cabal install) or > the old- > > fashioned way: > > > > cd portaudio-0.0.1 > > > > ghc --make Setup > > > > ./Setup configure --help > > (choose your options, prefix, profiling, ...) > > > > ./Setup configure $OPTIONS > > ./Setup build > > > > if all's well, > > > > ./Setup haddock > > ./Setup install > > > > (dies for me with > > dist/build/Sound/PortAudio/Base.chs.h:1:23: error: portaudio.h: Datei > oder > > Verzeichnisnicht gefunden > > c2hs: Error during preprocessing custom header file > > cabal: Error: some packages failed to install: > > portaudio-0.0.1 failed during the building phase. The exception was: > > exit: ExitFailure 1 > > because I don't have portaudio installed) > > _______________________________________________ > > Haskell-Cafe mailing list > > Haskell-Cafe(a)haskell.org > > http://www.haskell.org/mailman/listinfo/haskell-cafe > > ------------------------------ > Australia's #1 job site If It Exists, You'll Find it on SEEK<http://clk.atdmt.com/NMN/go/157639755/direct/01/> > > > ---------- Forwarded message ---------- > From: Henning Thielemann <lemming(a)henning-thielemann.de> > To: Michael Snoyman <michael(a)snoyman.com> > Date: Sun, 06 Dec 2009 03:05:30 +0100 (CET) > Subject: Re: [Haskell-cafe] New Hackage category: Error Handling > > On Sat, 5 Dec 2009, Henning Thielemann wrote: > > On Sun, 6 Dec 2009, Michael Snoyman wrote: >> >> I think there are plenty of examples like web servers. A text editor with >>> plugins? I >>> don't want to lose three hours worth of work just because some plugin >>> wasn't written >>> correctly. For many classes of programs, the distinction between error >>> and exception is >>> not only blurred, it's fully irrelevant. Harping on people every time >>> they use error in >>> the "wrong" sense seems unhelpful. >>> >>> Hope my commenting on this subject doesn't become my own form of >>> *pedantry*. >>> >> >> In an earlier thread I have explained that one can consider a software >> architecture as divided into levels. What is an error in one level (text >> editor plugin, web server thread, operating system process) is an exception >> in the next higher level (text editor, web server, shell respectively). This >> doesn't reduce the importance to distinguish between errors and exceptions >> within one level. All approaches so far that I have seen in Haskell just mix >> exceptions and errors in an arbitrary way. >> > > > I have just written more details on this topic: > http://www.haskell.org/haskellwiki/Error_vs._Exception > > > > ---------- Forwarded message ---------- > From: M Xyz <functionallyharmonious(a)yahoo.com> > To: Daniel Fischer <daniel.is.fischer(a)web.de> > Date: Sat, 5 Dec 2009 18:08:07 -0800 (PST) > Subject: Re: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > > > --- On *Sat, 12/5/09, Daniel Fischer <daniel.is.fischer(a)web.de>* wrote: > > > > cd portaudio-0.0.1 > > ghc --make Setup > > ./Setup configure --help > (choose your options, prefix, profiling, ...) > > ./Setup configure $OPTIONS > ./Setup build > > > Everything went well until "Setup build" which yielded our friend "c2hs.exe > does not exist". > > > C:\A\install\programming\portaudio\haskell\portaudio-0.0.1>ghc --make Setup > [1 of 1] Compiling Main ( Setup.hs, Setup.o ) > Linking Setup.exe ... > > C:\A\install\programming\portaudio\haskell\portaudio-0.0.1>Setup configure > --hel > p > Usage: Setup configure [FLAGS] > > Flags for configure: .......(edited out)................. > > C:\A\install\programming\portaudio\haskell\portaudio-0..0.1>Setup configure > --ext > ra-include-dirs="C:\A\install\programming\portaudio\portaudio\include" > --extra-l > > ib-dirs="C:\A\install\programming\portaudio\portaudio\build\msvc\Win32\Release" > Configuring portaudio-0.0.1... > > C:\A\install\programming\portaudio\haskell\portaudio-0.0.1>Setup build > Preprocessing library portaudio-0.0.1... > c2hs.exe: does not exist > > > > > > > ---------- Forwarded message ---------- > From: M Xyz <functionallyharmonious(a)yahoo.com> > To: daniel.is.fischer(a)web.de, john lask <jvlask(a)hotmail.com> > Date: Sat, 5 Dec 2009 18:17:29 -0800 (PST) > Subject: RE: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > I am going to give this a try. Thanks. > Where can I get the pexports and dlltool utilities? > > Google yields: > http://www.emmestech.com/software/pexports-0.43/download_pexports.html > http://sourceware.org/binutils/ > > Are those correct? > > --- On *Sat, 12/5/09, john lask <jvlask(a)hotmail.com>* wrote: > > > From: john lask <jvlask(a)hotmail.com> > Subject: RE: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > To: daniel.is.fischer(a)web.de, functionallyharmonious(a)yahoo.com > Cc: haskell-cafe(a)haskell.org > Date: Saturday, December 5, 2009, 9:03 PM > > > > I don't know whether this will help you but I just downloaded an built > the haskell portaudio package ... (I had a windows msvc build of > portaudio dll already) the process I used ... ghc 6.10.4, portaudio-19 > > make an import lib for ghc from dll: > pexports libpa19.dll > libpa19.def > dlltool --input-def libpa19.def --output-lib libpa19.a > > edit the .cabal file or use command line flags > > extra-Libraries: pa19 > extra-lib-dirs: c:\portaudio19\lib > include-dirs: c:\portaudio19\include > > runghc setup configure > > make sure you have cpp i.e. the c-preprocessor on your exe path, > otherwise you will get > 'cpp' is not recognized as an internal or external command, > operable program or batch file. > c2hs.exe: Error during preprocessing custom header file > > runghc setup build > > builds ok ...?? > > > > From: daniel.is.fischer(a)web.de > > To: functionallyharmonious(a)yahoo.com > > Subject: Re: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > > Date: Sun, 6 Dec 2009 02:34:17 +0100 > > CC: haskell-cafe(a)haskell.org > > > > Am Sonntag 06 Dezember 2009 01:49:49 schrieb M Xyz: > > > > I just had another idea. > > > > dafis@linux-mkk1:~> c2hs -o memyself.hs memyself.chs > > c2hs: does not exist > > > > it's not that c2hs isn't found or something, c2hs doesn't find Base.chs! > > > > Try installing from the unpacked sources (cd portaudio; cabal install) or > the old- > > fashioned way: > > > > cd portaudio-0.0.1 > > > > ghc --make Setup > > > > ./Setup configure --help > > (choose your options, prefix, profiling, ...) > > > > ./Setup configure $OPTIONS > > ./Setup build > > > > if all's well, > > > > ./Setup haddock > > ./Setup install > > > > (dies for me with > > dist/build/Sound/PortAudio/Base.chs.h:1:23: error: portaudio.h: Datei > oder > > Verzeichnisnicht gefunden > > c2hs: Error during preprocessing custom header file > > cabal: Error: some packages failed to install: > > portaudio-0.0.1 failed during the building phase. The exception was: > > exit: ExitFailure 1 > > because I don't have portaudio installed) > > _______________________________________________ > > Haskell-Cafe mailing list > > Haskell-Cafe(a)haskell.org > > http://www.haskell.org/mailman/listinfo/haskell-cafe > > ------------------------------ > Australia's #1 job site If It Exists, You'll Find it on SEEK<http://clk.atdmt.com/NMN/go/157639755/direct/01/> > > > > > ---------- Forwarded message ---------- > From: M Xyz <functionallyharmonious(a)yahoo.com> > To: john lask <jvlask(a)hotmail.com>, vanenkj(a)gmail.com > Date: Sat, 5 Dec 2009 19:47:04 -0800 (PST) > Subject: [Haskell-cafe] PortAudio library successfully built on Windows > > John Lask, > The steps you enumerated below successfully built portaudio for me! Thank > you! > > I wonder though, you said to rename the dll to libpa19.dll from > portaudio_x86.dll > and to change the .cabal entry from "extra-Libraries: portaudio" to > "extra-Libraries: pa19". > Since my .dll name and .cabal file entry were mismatched, was that what > could have been wrong all along? Was the creation of a .a file necessary? > I would delete the .a file and retry with libportaudio.dll but I won't push > my luck. > > Now if I could just get Gtk2hs working now, I could be happy *and* > productive. :) > > --- On *Sat, 12/5/09, john lask <jvlask(a)hotmail.com>* wrote: > > > From: john lask <jvlask(a)hotmail.com> > Subject: RE: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > To: daniel.is.fischer(a)web.de, functionallyharmonious(a)yahoo.com > Cc: haskell-cafe(a)haskell.org > Date: Saturday, December 5, 2009, 9:03 PM > > > > I don't know whether this will help you but I just downloaded an built > the haskell portaudio package ... (I had a windows msvc build of > portaudio dll already) the process I used ... ghc 6.10.4, portaudio-19 > > make an import lib for ghc from dll: > pexports libpa19.dll > libpa19.def > dlltool --input-def libpa19.def --output-lib libpa19.a > > edit the .cabal file or use command line flags > > extra-Libraries: pa19 > extra-lib-dirs: c:\portaudio19\lib > include-dirs: c:\portaudio19\include > > runghc setup configure > > make sure you have cpp i.e. the c-preprocessor on your exe path, > otherwise you will get > 'cpp' is not recognized as an internal or external command, > operable program or batch file. > c2hs.exe: Error during preprocessing custom header file > > runghc setup build > > builds ok ...?? > > > > From: daniel.is.fischer(a)web.de > > To: functionallyharmonious(a)yahoo.com > > Subject: Re: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > > Date: Sun, 6 Dec 2009 02:34:17 +0100 > > CC: haskell-cafe(a)haskell.org > > > > Am Sonntag 06 Dezember 2009 01:49:49 schrieb M Xyz: > > > > I just had another idea. > > > > dafis@linux-mkk1:~> c2hs -o memyself.hs memyself.chs > > c2hs: does not exist > > > > it's not that c2hs isn't found or something, c2hs doesn't find Base.chs! > > > > Try installing from the unpacked sources (cd portaudio; cabal install) or > the old- > > fashioned way: > > > > cd portaudio-0.0.1 > > > > ghc --make Setup > > > > ./Setup configure --help > > (choose your options, prefix, profiling, ...) > > > > ./Setup configure $OPTIONS > > ./Setup build > > > > if all's well, > > > > ./Setup haddock > > ./Setup install > > > > (dies for me with > > dist/build/Sound/PortAudio/Base.chs.h:1:23: error: portaudio.h: Datei > oder > > Verzeichnisnicht gefunden > > c2hs: Error during preprocessing custom header file > > cabal: Error: some packages failed to install: > > portaudio-0.0.1 failed during the building phase. The exception was: > > exit: ExitFailure 1 > > because I don't have portaudio installed) > > _______________________________________________ > > Haskell-Cafe mailing list > > Haskell-Cafe(a)haskell.org > > http://www.haskell.org/mailman/listinfo/haskell-cafe > > ------------------------------ > Australia's #1 job site If It Exists, You'll Find it on SEEK<http://clk.atdmt.com/NMN/go/157639755/direct/01/> > > > > > ---------- Forwarded message ---------- > From: Michael Snoyman <michael(a)snoyman.com> > To: Haskell Cafe <haskell-cafe(a)haskell.org> > Date: Sun, 6 Dec 2009 07:04:46 +0200 > Subject: [Haskell-cafe] When are undecidables ok? > I know this is basically a rewording of a previous e-mail, but I realized > this is the question I *really* wanted to ask. > > We have this language extension UndecidableInstances (not to mention > OverlappingInstances), which seem to divide the Haskell camp into two > factions: > > * Hey, GHC said to turn on this flag. Ok! > * Undecidables are the devil! > > I get the feeling the truth lies in the middle. As I understand it (please > correct me if I am wrong), the problem with undecidables is that they can > create non-terminating instances. However, for certain cases the programmer > should be able to prove to him/herself that the instances will terminate. My > question is: how can you make such a proof? > > I've had to cases in particular that made me want undecidables. Both of > them, IMO, could be solved by creating new extensions to GHC which would not > require undecidables. Nonetheless, for now we only have undecidables at our > disposal. The examples are: > > * Context synonyms (eg, MonadFailure = Monad + Failure). > * Subclass function defaulting > > For an example of the second, a nifty definition of Monad would be: > > class Applicative m => Monad m where > >>= ... > return ... > pure = return > (<*>) = ap > fmap = liftM > > Of course, neither of these is possible in Haskell, so we can use > undecidables. How can a programmer prove that a set of instances is, in > fact, safe? And if they make a mistake and right a bad set of > undecidable/overlapping instances, what's the worst case scenario? Is it a > compile-time or run-time error*? > > Michael > > * Yes, I mean error. > > > ---------- Forwarded message ---------- > From: Luke Palmer <lrpalmer(a)gmail.com> > To: Michael Snoyman <michael(a)snoyman.com> > Date: Sat, 5 Dec 2009 22:36:43 -0700 > Subject: Re: [Haskell-cafe] When are undecidables ok? > On Sat, Dec 5, 2009 at 10:04 PM, Michael Snoyman <michael(a)snoyman.com> > wrote: > > I know this is basically a rewording of a previous e-mail, but I realized > > this is the question I *really* wanted to ask. > > > > We have this language extension UndecidableInstances (not to mention > > OverlappingInstances), which seem to divide the Haskell camp into two > > factions: > > > > * Hey, GHC said to turn on this flag. Ok! > > * Undecidables are the devil! > > > > I get the feeling the truth lies in the middle. As I understand it > (please > > correct me if I am wrong), the problem with undecidables is that they can > > create non-terminating instances. However, for certain cases the > programmer > > should be able to prove to him/herself that the instances will terminate. > My > > question is: how can you make such a proof? > > Well, the reasoning for the "devil" camp (which I admit to being > firmly in[1]) is that such proofs must rely on the algorithm the > compiler uses to resolve instances. You might be able to prove it, > but the proof is necessarily only valid for (possibly current versions > of) GHC. The typeclass resolution algorithm is not in the report, and > there are several conceivable ways of of going about it. > > So it is fine to use them if you are okay with making your code > unportable and future-brittle. I am typically against the mere > existence of code that that is future-brittle, because it encourages > compiler authors not to innovate (and by that token, unportable too, > because it discourages compiler competition). > > Luke > > [1] > http://lukepalmer.wordpress.com/2008/04/08/stop-using-undecidable-instances/ > > > > ---------- Forwarded message ---------- > From: Michael Snoyman <michael(a)snoyman.com> > To: Luke Palmer <lrpalmer(a)gmail.com> > Date: Sun, 6 Dec 2009 08:20:09 +0200 > Subject: Re: [Haskell-cafe] When are undecidables ok? > > > On Sun, Dec 6, 2009 at 7:36 AM, Luke Palmer <lrpalmer(a)gmail.com> wrote: > >> On Sat, Dec 5, 2009 at 10:04 PM, Michael Snoyman <michael(a)snoyman.com> >> wrote: >> > I know this is basically a rewording of a previous e-mail, but I >> realized >> > this is the question I *really* wanted to ask. >> > >> > We have this language extension UndecidableInstances (not to mention >> > OverlappingInstances), which seem to divide the Haskell camp into two >> > factions: >> > >> > * Hey, GHC said to turn on this flag. Ok! >> > * Undecidables are the devil! >> > >> > I get the feeling the truth lies in the middle. As I understand it >> (please >> > correct me if I am wrong), the problem with undecidables is that they >> can >> > create non-terminating instances. However, for certain cases the >> programmer >> > should be able to prove to him/herself that the instances will >> terminate. My >> > question is: how can you make such a proof? >> >> Well, the reasoning for the "devil" camp (which I admit to being >> firmly in[1]) is that such proofs must rely on the algorithm the >> compiler uses to resolve instances. You might be able to prove it, >> but the proof is necessarily only valid for (possibly current versions >> of) GHC. The typeclass resolution algorithm is not in the report, and >> there are several conceivable ways of of going about it. >> >> So it is fine to use them if you are okay with making your code >> unportable and future-brittle. I am typically against the mere >> existence of code that that is future-brittle, because it encourages >> compiler authors not to innovate (and by that token, unportable too, >> because it discourages compiler competition). >> >> Luke >> >> [1] >> http://lukepalmer.wordpress.com/2008/04/08/stop-using-undecidable-instances/ >> > > So in that case, perhaps the compiler authors can give us some ideas as to > when it's safe to use undecidables? Seems like we should go straight to the > horse's mouth. > > Michael > > > > ---------- Forwarded message ---------- > From: "Ariel J. Birnbaum" <valgarv(a)gmx.net> > To: Gregory Crosswhite <gcross(a)phys.washington.edu> > Date: Sun, 06 Dec 2009 08:34:16 +0200 > Subject: Re: [Haskell-cafe] ANNOUNCE: error-message > > In particular, the motivation for this package was that I have written > > a build system, and I wanted to collect as many errors in the build as > > possible and show them all to the user at once. > > YMMV, but at least for me a deluge of errors is less helpful than a > short list I can fix quickly, then try again. Often by fixing one error > from the list a sizeable portion of the rest just vanish (I forget the > English term for that kind of error). > > (In before "you can use sth like quickfix to go through the first few > errors and compile again whenever you feel like it".) > > -- > Ariel J. Birnbaum > > > > > ---------- Forwarded message ---------- > From: Eugene Kirpichov <ekirpichov(a)gmail.com> > To: Stefan Holdermans <stefan(a)cs.uu.nl> > Date: Sun, 6 Dec 2009 09:42:34 +0300 > Subject: Re: [Haskell-cafe] What is the rank of a polymorphic type? > 2009/12/6 Stefan Holdermans <stefan(a)cs.uu.nl>: > > Eugene, > > > >> 1) Does there exist an authoritative source saying the same? Not that > >> I'm doubting, just supposing that the source would have other > >> interesting information, too :) > > > > You may want to have a look at the already mentioned JFP-article by > Peyton > > Jones et al. and perhaps the work of Kfoury and Wells. > > > >> 2) Is it true that rank (forall a . a, forall a . a) == 0 ? > > > > No, for pairs one takes the maximum of the constituent types. So, here > you'd > > get rank 1. > > > > Note that this is an impredicative type, which is yet another extension > of > > the standard Hindley-Milner typing discipline. > > OK, thanks. > However, isn't the type (forall a . a) -> String impredicative because > it instantiates a type variable of the type constructor (->) p q with > p = forall a . a? > > > > > Cheers, > > > > Stefan > >> > >>> > > > > > > -- > Eugene Kirpichov > Web IR developer, market.yandex.ru > > > > ---------- Forwarded message ---------- > From: Benjamin L.Russell <dekudekuplex(a)yahoo.com> > To: haskell-cafe(a)haskell.org > Date: Sun, 06 Dec 2009 16:14:06 +0900 > Subject: [Haskell-cafe] Re: universal binary version of Haskell Platform? > On Thu, 03 Dec 2009 10:52:35 +0000, Duncan Coutts > <duncan.coutts(a)googlemail.com> wrote: > > >[...] > > > >There no binary platform installer for OSX PPC. You'll have to grab > >ghc-6.10.4 for PPC from the ghc download page and then install the > >platform from the generic source tarball. > > Ah, that's too bad. That means that I won't be able to invoke GHC > outside of the Terminal application by default, and that even if I > create a Darwin shell script and alias to invoke it from Aqua, the > icon for that shell script will only be generic by default. I'd > rather click on an application icon in the Dock. > > > > >If you'd like to help us next time to make a platform binary for PPC > >then that'd be great. I don't think we have the setup to make universal > >binaries but it should be possible to make a PPC build if we have a > >volunteer. > > Sure, barring job-related time constraints, I'd be happy to volunteer. > My Mac just went out of service because of a hardware problem with the > memory, so I'm order new replacement RAM this weekend. As soon as > that arrives and I install it, the problem should be resolved. > > What should I do? > > -- Benjamin L. Russell > > > > > ---------- Forwarded message ---------- > From: Dan Doel <dan.doel(a)gmail.com> > To: haskell-cafe(a)haskell.org > Date: Sun, 6 Dec 2009 02:48:09 -0500 > Subject: Re: [Haskell-cafe] What is the rank of a polymorphic type? > On Sunday 06 December 2009 1:42:34 am Eugene Kirpichov wrote: > > OK, thanks. > > However, isn't the type (forall a . a) -> String impredicative because > > it instantiates a type variable of the type constructor (->) p q with > > p = forall a . a? > > There's probably no clear cut answer to this independent of how you think > of > (->). For instance, if we explain the Haskell type system by way of a pure > type system, (->) is a special case of a pi type, which looks like: > > pi x : k. t > > where any xs in t are bound by the pi. We then have: > > p -> q = pi _ : p. q > forall a : k. b = pi a : k. b > > pi types are given types by sets of rules, which look like triples. If > (s,t,u) > is a rule, then: > > G |- k : s G, a : k |- b : t > ------------------------------- > G |- (pi a : k. b) : u > > is the corresponding typing rule. Type systems like Haskell's are commonly > thought of in terms of the lambda cube, which has constant sorts * and [], > with * : []. The rule (*,*,*) gives you ordinary functions. (*,[],[]) gives > you dependent types, so that's out. > > ([],*,*) is an impredicative rule for polymorphism. This says that, for > instance: > > forall a. a -> a = (pi a : *. pi _ : a. a) : * > > because (pi _ : a. a) : * if a : *, by the (*,*,*) rule, and then we apply > the > impredicative rule for the universal quantification. One could also use the > predicative rule ([],*,[]), which would result in forall a. a -> a having > type > []. > > However, Haskell also has arbitrarily higher-order types. This is given by > the > rule ([],[],[]), which allows expressions like: > > (* -> *) -> * = pi _ : (pi _ : *. *). * > > This type system is called F_omega, while just the ([],*,?) rule is known > as > F_2. > > However, the F_omega rule also allows for arbitrary rank polymorphism even > with the predicative universal quantifier rule above (predicative F_2 > allows a > little, but it's very limited*). For instance, the higher rank type: > > forall a. (forall b. b) -> a > > checks thusly: > > (forall b. b) : [] via ([],*,[]) > ((forall b. b) -> a) : [] via ([],*,[]) > (forall a. (forall b. b) -> a) : [] via ([],[],[]) > > Data types, by contrast, have kinds like * -> *, so using say, > > Maybe (forall a. a -> a) > > genuinely relies on the impredicative rule. GHC's type system isn't exactly > set up in this way, but (->) is similarly special in that it somehow isn't > quite just another type constructor with kind * -> * -> * (or even whatever > special kinds GHC uses to support unboxed values and such). > > Hope that wasn't too confusing. :) > > -- Dan > > * Predicative F_2 will essentially allow one universal quantifier somewhere > in > the type. This can be: > > forall a. a -> a > > or it can be: > > (((forall a. a) -> T) -> U) -> V > > for T, U and V of kind * (the only kind in F_2), which is a rank-4 type. It > doesn't allow: > > forall a b. a -> b > > even, because the inner (forall b. a -> b) : [], so adding the forall a > requires the F_omega rule. > > Predicative F_2 and F_w also blow up with quantification on the right of an > arrow, because it looks like the rule for dependent types: > > T -> (forall a. a) > > T : *, (forall a. a) : [] > > so the rule (*,[],[]) would be invoked. > > GHC doesn't have this sort of hierarchy, and so doesn't have these sorts of > weird cases, despite being predicative of a sort. Instead it distinguishes > somehow between monotypes ([Float], String -> Int, a -> b) and polytypes > (forall a. a, ...), although it doesn't really display the difference. > Quantifiers are only supposed to range over kinds that classify monotypes > (or > monotype constructors), which keeps the predicativity (although, even this > gets fudged some: If I have forall a. a -> a, I can instantiate a to the > polytype forall a. a -> a with rank-n polymorphism, because it only seems > to > worry about the validity of the resulting type, and (->) is special; by > contrast, the same cannot be said for forall a. Maybe a, because Maybe > genuinely only accepts monotypes without -XImpredicativeTypes). > > > > ---------- Forwarded message ---------- > From: Bulat Ziganshin <bulat.ziganshin(a)gmail.com> > To: Andrew Coppin <andrewcoppin(a)btinternet.com> > Date: Sun, 6 Dec 2009 11:13:21 +0300 > Subject: Re[2]: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > Hello Andrew, > > Sunday, December 6, 2009, 1:09:18 AM, you wrote: > > > Maybe once I get hired by some financial modelling consultants and get > > paid shedloads of money to write Haskell all day, I'll be able to afford > > a Mac. But until then... > > with such attitude you will never be hired by financial sector. try > instead: > > mac? great! windows? i love it! unix? i've used it since 60's > > > -- > Best regards, > Bulat mailto:Bulat.Ziganshin@gmail.com > > > > > ---------- Forwarded message ---------- > From: Heinrich Apfelmus <apfelmus(a)quantentunnel.de> > To: haskell-cafe(a)haskell.org > Date: Sun, 06 Dec 2009 11:28:25 +0100 > Subject: [Haskell-cafe] Re: universal binary version of Haskell Platform? > Benjamin L.Russell wrote: > > Ah, that's too bad. That means that I won't be able to invoke GHC > > outside of the Terminal application by default, and that even if I > > create a Darwin shell script and alias to invoke it from Aqua, the > > icon for that shell script will only be generic by default. I'd > > rather click on an application icon in the Dock. > > You can write an applescript similar to > > Open Terminal Here > from http://www.entropy.ch/software/applescript/ > > and endow it with a custom icon. > > > Regards, > Heinrich Apfelmus > > -- > http://apfelmus.nfshost.com > > > > > ---------- Forwarded message ---------- > From: Andrew Coppin <andrewcoppin(a)btinternet.com> > To: haskell-cafe(a)haskell.org > Date: Sun, 06 Dec 2009 13:05:19 +0000 > Subject: Re: [Haskell-cafe] Low Level Audio - Writing bytes to the sound > card? > Bulat Ziganshin wrote: > >> Hello Andrew, >> >> Sunday, December 6, 2009, 1:09:18 AM, you wrote: >> >> >> >>> Maybe once I get hired by some financial modelling consultants and get >>> paid shedloads of money to write Haskell all day, I'll be able to afford >>> a Mac. But until then... >>> >>> >> >> with such attitude you will never be hired by financial sector. try >> instead: >> >> mac? great! windows? i love it! unix? i've used it since 60's >> >> > > That's OK. I'm reasonably sure I will never get hired by the financial > sector /anyway/. :-( > > > > > ---------- Forwarded message ---------- > From: Andrew Coppin <andrewcoppin(a)btinternet.com> > To: haskell-cafe(a)haskell.org > Date: Sun, 06 Dec 2009 13:18:59 +0000 > Subject: Re: [Haskell-cafe] binding to C libraries on Windows was Low Level > Audio - Writing bytes to the sound card? > john lask wrote: > >> I think there are some misapprehensions here:- >> >> Many haskell packages binding to c libraries will compile with ghc >> without problems on windows - without cygwin, without mingw/msys system. >> > > OK, well I haven't tried building every C binding on all of Hackage, just a > few of them. So far none of them have worked. (Including wxHaskell, SDL, one > of the database packages, etc.) It's got to the point where I've simply > given up trying. Most such packages just don't seem to work on Windows, and > nobody on the mailing lists or on IRC has any clue why. (I guess because few > people in the Haskell community use Windows and understand how it's supposed > to work.) > > Some such packages build "out of the box" on windows, like the zlib >> package which contains the c source for the c zlib library. GHC is able >> to compile and build this packages without any other c >> compiler/libraries/unix emulators etc because ghc itself contains part >> of the gcc c compiler tool chain and comes with all c standard headers, >> c++ headers and c/c++ runtime libraries. >> > > Yes, I realised long ago that GHC uses GCC and other GNU build tools. > (Presumably this makes porting to Windows far easier.) More recently I > realised that it also includes a surprisingly large set of header files, > seemingly including the entire Win32 API, which is interesting. > > There >> is only one gotch-ya - you need to have a import library for the gcc >> tool chain (thats what ghc uses) i.e. a ".a" library and not the native >> windows ".LIB" import library. >> > > They're different?? o_O > > Oh. Suddenly several things seem clearer... > > If you don't have ".a" import library >> but have the dll then the '.a' import library be built for any dll >> relativley easily. >> > > Any idea how? > > the bigest problem hamperring cleaner builds of haskell packages on >> windows is the lack of any standardised scheme for the installation of >> c-libraries and header files (and of course the availability of a >> suitable build of the library). >> > > Isn't one of Cabal's jobs to figure out where stuff is? Can't we get Cabal > on Windows to say "hey, I need to find foo.h, you know where that is?" Or > something like that? > > Another problem hampering the install of haskell packages on windows is >> the use of the unix autoconf build system (./configure) - package >> writters note! >> > > Heh. I found one Wiki page once describing how to set up a Cabal package. > When it started talking about how to integrate Automake, I hung my head is > dispair. (I believe the page in question is fixed now. But a lot of package > authors seem to assume that everybody just uses Unix of some kind...) > > > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe > >

2 1

ANNOUNCE: Blueprint 0.1 -- PREVIEW
by Gregory Crosswhite 06 Dec '09

06 Dec '09

I earlier posted to this list asking for help in getting Cabal to play with a Fortran compiler. When it turned out that there was no answer to this, I of course did the obvious thing --- I wrote my own build system. :-) It already boasts the following features: * Fully self-hosting --- can both build and install itself. * Implicitly detects and tracks module dependencies. * Builds in parallel as much as possible. * Can scan through a source directory to add all files, saving the trouble of adding them manually. * Supports compiling C and Fortran sources in addition to Haskell sources, and linking everything together. * Utilizes a “tool” system to make it easy to extend its functionality, such as by adding a new language. * Flexible configuration system: tool options (e.g., path to the Fortran compiler) can be detected automatically, set in a configuration file, or specified on the command line. * Provides help messages both to list the targets and to explain the configuration command-line options. * Does not terminate at the first error, but instead tries to do as much as possible and then report all errors. (This is a general philosophy that holds both for the build, the configuration, etc.) * Easily supports multiple targets that can depend on the results from other targets. * Automatically resolves package dependencies, as read from .cabal file. Detects when a module is not in any known dependency, and reports an error that also lists all packages that export this module. What separates this system apart from others is that it uses Haskell’s built-in laziness to handle dependency ordering and parallelism. That is, one expresses the desired build result as a function of the input sources; when the result is demanded, it causes Haskell to go through and execute on-demand all of the necessary build steps. To get parallelism, just tell GHC to use multiple threads with the “+RTS -N# -RTS” option, where # is the number of threads. (All that I had to do to get this to work was put parLists in a couple of places.) This is a big win for me because it means that I get the correct behavior without having to write and test my own complicated graph data structure for tracking the dependencies between build steps. The philosophy of building up a big calculation of which only parts need be lazily demanded can be extended further to the build phases, which I call “targets”. For example, the first line of the code defining the build target is: build = configure >>= \configuration -> ... This demands the configuration target, and then uses the result in the build. So when you run “./Setup configure”, you only demand that Setup compute the value of “configure”, but when you run “./Setup build” you demand both “build” and also “configure”. All of the errors that are reported are “composable” in the sense that they can be combined into a new error message that leaves nothing out and also has no duplicates. As much as possible the system tries not to stop until it has nothing else that it can do, allowing it to uncover as many errors as possible and then compose them into a single message. I have posted Blueprint to Hackage so that people can see what I have done and possibly play with it. I hope to eventually document it and add examples, but I figured it was better to release early and get it out there sooner rather than later so that people can see what I’ve done. Thus, at this point it should be considered a PREVIEW release, not even an alpha release. Having said that, within the domain of the functionality it implements it works well enough that I use it myself as the build system for a Physics simulation code that is a mixture of C, Haskell, and Fortran. To see what Blueprint looks like in action, take a look at Setup.hs. For example, Setup can either build a library (“build” target) or it can build itself (“bootstrap” target); it is smart enough to use different flags and directories in the two cases so that there ends up being in essence two parallel builds. Setup also has a lot of targets: configure, reconfigure (delete cached configuration and then configure), build, rebuild (clean then build), clean, distclean, install, bootstrap (build itself rather than a library), haddock (documentation generation that does not work particularly well yet). Anyway, take a look if any of this sounds interesting to you, and let me know if you have any questions! - Greg PS: You're best bet for browsing through the sources is to use Leo (leo.sourceforge.net) since this makes it easier for you to see how everything is organized.

2 3