- Haskell - Haskell.org

framework for composing monads?
by Elke Kasimir 18 Feb '01

18 Feb '01

Does someone like to comment on this? I'm planning a new cli/odbc-based database connectivity library for Haskell 98 and want to manage hidden state (various management information) on the Haskell side. Some libs. i.e. for gui, extend the IO monad for this using some "start" function: main :: IO () main = start prog prog :: GUI () ... I could do the same with my library, but then it is difficult to combine, for example, Gui operations with Db operations. I found a better solution in Jeuring/Meijer 1995 in Mark P. Jones' contribution which consists of using monad transformers and lifting monad operations to class level like this: type M a = DB (GUI IO) a startM = start . start main = startM prog prog :: (Io m, Gui m, Db m) => m () prog = ... I hope that this approach not only solves the problem mentioned above but also supports a more modular approach to handling state etc. in general and encourages the programmer to add his own monads (instead of packing all into IO or passing around state etc. explicitly). However, to be really useful, it would be best if there could be established a "standard framework" for such libraries. It would be nice if there were at least agreement on how to lift IO operations to class level: 'Io.putStr' (which provokes name clashes with the standard prelude),'Io.putStrC' (C for class level), or some other prefix/suffix. In addition to Mark P. Jones' examples in the mentioned article, - Class Io must be provided, - all specific monad operations should be encapsulated in ADTs, - it must be possible to map 'lift' into the ADT In the attachements I have worked out a scenario, where DB and GUI are state monad transformers and Io and Gui are defined as follows: class Monad m => Io m where io :: IO a -> m a class Monad m => Db m where db :: DBImpl m a DBImpl is an ADT which encapsulates a set of characteristic monad operations, in this case the 'modify' operation of a state transformer. For such ADTs, a special class Liftable l where mapLift :: (MonadT t, Monad m) => l m -> l (t m) has been introduced to supply a unified symbol for applying lift to the encapsulated monad transformation(s). (I don't like the class name either but didn't find a better up to now...) However, there remains a problem with the type system. Ideally, it would be possible to require: class (Monad m, Monad (t m)) => MonadT t where lift = ... such that MonadT applications become automatically members of class Monad, and declare instance (Db m, MonadT t) => Db (t m) where db = mapLift db (same for class Gui etc.) with the obvious effects. Both seems to be not even possible in the extended type systems of Hugs and ghc. I want to create a library for Haskell 98, so this is of minor importance, but if there was a way to make life easier for those who use type system extensions without complicating the situation for Haskell 98 users, I would of course prefer such a solution. Until then, again, I think that "standard framework" would make it easier for users to create those instance declarations by themselves needed to combine monad transformers from different sources. Elke. --- "If you have nothing to say, don't do it here..." Elke Kasimir Skalitzer Str. 79 10997 Berlin (Germany) fon: +49 (030) 612 852 16 mail: elke.kasimir(a)catmint.de> see: <http://www.catmint.de/elke> for pgp public key see: <http://www.catmint.de/elke/pgp_signature.html>

3 4

Re: Literate Programming in Haskell?
by andrew＠andrewcooke.free-online.co.uk 18 Feb '01

18 Feb '01

Hi, I'm no expert, but I've found a solution that works for me. I use the \begin{code} ... \end{code} delimiters (rather than >). Since these aren't standard Latex, I threw together a style file that called the listings package to handle this. The result is a file that ghc and hugs can compile directly, but which latex can also process. No need for weaving and tangling, while line numbers in error messages are correct. On the other hand, I'm not doing anything very sophisticated - no labelling of blocks or anything like that. I needed to update to the version of tetex in the "testing" Debian release to get this to work - the listing package in the standard distribution doesn't support Haskell (not that I really use that support), so if you get an error message about being unable to load something, that may be the problem. (Warning: I wouldn't recommend moving to Debian testing at the moment; when I last tried the X support had changed significantly and failed to install, so don't upgraed anything apart from tetex). I've included the style file. I don't have a public example yet (hopefully later this week), but I can email you a postscript file directly if you want. Haskell is the first language where I've managed to work happily with literate programming - partly because of the direct integration and partly because the order isn't critical and the code is compact. Cheers, Andrew PS The style file includes lots of stuff unrelated to literate programming - delete everything but the first block, if you like. On Sun, Feb 18, 2001 at 01:53:41AM -0500, Tom Moertel wrote: > In the Haskell community is there a generally accepted best way to > approach Literate Programming? The language has support for literate > comments, but it seems that many common LP tools don't respect it. > > For example, in order to convert some .lhs code into LaTeX via the noweb > LP tools, I had to write a preprocessor to convert the ">" code blocks > into something that noweb would respect. (The preprocessor actually > does a bit more and, in conjunection with noweb, gives pretty good > results for little effort. For an example, see: > > http://www.ellium.com/~thor/hangman/cheating-hangman.lhs > http://www.ellium.com/~thor/hangman/cheating-hangman.pdf > ) > > Yet somehow, I don't think that my homebrew approach is optimal. Can > anybody recommend a particularly elegant LP setup for Haskell > programming? Or if you have an approach that works well for you, would > you mind sharing it? > > Cheers, > Tom > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe > -- http://www.andrewcooke.free-online.co.uk/index.html

1 0

Lifting to the limit
by Saswat Anand 17 Feb '01

17 Feb '01

Dear All, I have this following simplified problem. Please help. type Formula a b = a -> b type Env = (Integer,Char) fun :: Formula Integer Integer fun n = n+1 sun :: Formula Char Integer sun c = toInteger.ord c I want to be able to write (sun + fun) (2,'a') (fun + sun) (2,'a') both of which should be equal to (sun 'a' + fun 2) So I did this: lift2 :: (a -> b -> c) -> Formula d a -> Formula e b -> Formula Env c lift2 f a1 a2 = \env -> f ( (lift a1) n ) ( (lift a2) n ) -- Here hugs gives error class Lift a where lift :: Formula a b -> Formula Env b instance Lift Char where lift f = \(i,c) -> f c instance Lift Integer where lift f =\(i,c) -> f i With this, hugs (with extensions enabled) gives following error: Cannot justify constraints in explicitly typed binding ***Expression : lift2 ***Type : (a -> b -> c) -> Formula d a -> Formula e b -> Formula Env c ***Given Constraint: () ***Constraints : (Lift d, Lift e) Any help is very much appreciated. Thanks, Saswat

2 2

Re: [newbie] Lazy >>= ?!
by andrew＠andrewcooke.free-online.co.uk 17 Feb '01

17 Feb '01

Thanks, that was very helpful, but it does raise two further questions (neither of which is critical, to my program, but if anyone cares to comment I'm interested). 1. After digesting what you wrote I managed to make a lazy list of IO monads containing random numbers, but couldn't make an IO monad that contained a lazy list of random numbers. Is this intentional, me being stupid, or just chance? My code for producing [IO ..] is: listParams :: [IO (Point, Double)] listParams = entry : listParams entry :: IO (Point, Double) entry = do x <- randomIO y <- randomIO t <- randomIO return ((x,y),t) Also, should I be worried about having more than one IO monad - it seems odd encapsulating the "outside world" more than once. 2. Why does the following break finite lists? Wouldn't they just become lazy lists that evaluate to finite lists once map or length or whatever is applied? > Now, if this were changed to > ~(x:xs) >>= f = f x ++ (xs >>= f) > (a lazy pattern match) then your listList2 would work, but finite > lists would stop working. Thanks again, Andrew On Fri, Feb 16, 2001 at 05:51:08PM -0500, Scott Turner wrote: [...] > listList1 works because it has a different meaning from listList2 > and listList3. Note that the length of listList1 is explicitly 1, > but the length of the result list in listList2 and listList3 cannot > be determined. > > Thus, the following "works" because the length of listList5 is > expressed apart from evaluating listList5. > listList5 = do rest <- [head listList5] > return (1:rest) > > > >From another point of view, the (1:rest) _is_ a lazy construct. > However, rest <- listList2 demands to know, at a minimum, > what's the length of listList2. See the definition of the list type > as a monad in Prelude.hs, > (x:xs) >>= f = f x ++ (xs >>= f) > [] >>= f = [] > > Now, if this were changed to > ~(x:xs) >>= f = f x ++ (xs >>= f) > (a lazy pattern match) then your listList2 would work, but finite > lists would stop working. > -- Scott > > >Thanks, Andrew (code follows) > > > >listList1 :: [[Integer]] > >listList1 = [x] where x = 1:x > > > >listList2 :: [[Integer]] > >listList2 = do rest <- listList2 > > return (1:rest) > > > >listList3 :: [[Integer]] > >listList3 = listList3 >>= (\x -> return (1:x)) -- http://www.andrewcooke.free-online.co.uk/index.html

1 0

[newbie] Lazy >>= ?!
by andrew＠andrewcooke.free-online.co.uk 16 Feb '01

16 Feb '01

There seems to be something funny going on here - I thought all the following were equivalent, but only listList1 works with > liftM (take 2) listListn Is there any way to write listList2 or 3 so that they are lazy (what I'm really trying to do is get the same kind of code to work on an IO Monad that contains an infinite list, but when that failed I wondered why it worked for a list monad containing an infinite list and then managed to get that not to work either...) Thanks, Andrew (code follows) listList1 :: [[Integer]] listList1 = [x] where x = 1:x listList2 :: [[Integer]] listList2 = do rest <- listList2 return (1:rest) listList3 :: [[Integer]] listList3 = listList3 >>= (\x -> return (1:x)) -- http://www.andrewcooke.free-online.co.uk/index.html

1 0

ANNOUNCE: nhc98 version 1.02
by malcolm-nhc＠cs.york.ac.uk 16 Feb '01

16 Feb '01

We are pleased to announce the new release 1.02 of nhc98, a compiler for Haskell'98, written in Haskell'98. http://www.cs.york.ac.uk/fp/nhc98/ It is available as a source package (which can be built via ghc, hbc, an existing nhc98 installation, or simply with a C compiler if you don't already have a Haskell compiler installed). Binary packages are also currently available for ix86-Linux and sparc-solaris2. Binary packages for other architectures can be added to the list if users contribute them. What's new ---------- This is an interim release, mainly fixing bugs in nhc98-1.00. The tracing and debugging tool, Hat, is about to undergo some significant changes, so release 1.02 is a stable snapshot before we start to break too many things! Amongst the many bugfixes, notable ones are as follows: * Fixed a space-leak in the compiler. Compilation now requires approximately half the previous maximum heap, and as a result, compile times are 5-10% faster. * The raw speed of input and output for compiled programs is now much faster, although you will probably only notice a difference if your program is severely I/O-bound. * Local infix declarations now work correctly. * `newtype T a = T a' now works correctly. * @-pattern-bindings are fixed. This means that Happy-generated parsers now compile correctly. * We now accept {-# pragmas #-} in any source position, and understand LINE pragmas. Recent bugfixes and new features in Hat include: * When a traced program is interrupted or terminates with an error, you get a virtual stack trace of the computation "for free", without having to start up a browser. * Foreign imports (via the new common primitive FFI) with an I/O result type are now also traced (previously only pure foreign functions were traced). * Interface and object files for tracing now have distinct file suffixes: .T.hi and .T.o. This improves matters enormously when switching between normal and tracing versions of a program. It also fixes some dependency bugs when building the tracing version of the compiler. Both the compiler and hmake have knowledge of the new suffixes. * Tracing versions of the following libraries have been added: IO, Array, FFI, IOExtras. (Still missing: Directory, System, Time, Locale, CPUTime, Random.) Regards, Malcolm

2 1

RE: Haskell threads & pipes & UNIX processes
by Julian Seward (Intl Vendor) 16 Feb '01

16 Feb '01

| So what's going on? How can the goal be achieved? I don't know, but here's a different suggestion, using bzip2 (you could do the same with zlib for .gz files): use the foreign import mechanism to make BZ2_bzopen, BZ2_bzwrite and BZ2_bzclose available in your program. Then: do bz2 <- bzopen "output.bz2" "w" let chars_to_write = ... :: String bzwrite bz2 (packString chars_to_write) (length chars_to_write) -- the above as many times as you like bzclose bz2 and link your program with libbz2.a or libbz2.so. Almost the same trick works with the zlib library for directly reading/writing .gz files. This strikes me as far more convenient than messing round with pipes, etc. J

3 3

Postdoctoral Position
by Elke Tetzner 16 Feb '01

16 Feb '01

Please pass on to interested Post doctorants. Apologies for multiple copies. ---------------------------------------------------------------------------- New Postdoctoral Position (DEADLINE 1 April 2001) Department of Computer Science, University of Rostock Germany Vacancy for an assistant professor Responsibilities: education and research in programming languages and compiler construction research in implementation techniques in different areas of computer science cooperation in academic self-administration Qualifications: PhD in computer science interests in programming languages and compiler constructions priorities: declarative programming, attribute grammars and applications, semantics of programming languages and compiler construction further interests: aspect-oriented programming, specification technics, real-time programming and other educational skills a good knowledge of german language engagement: 1st April 2001 (permanent) salary: BAT-O Ib (special in Germany), full time further information: Prof. Dr. G. Riedewald (http://www.informatik.uni-rostock.de/FB/Praktik/psuet/griedew.html) Dr. W. Mahrhold phone: (+49 381)498-3397 Qualified women are emphatically invited for application. Severely handicapped person with same qualification is prefered.

1 0

Haskell Graphics Lib problem
by Don Wakefield 16 Feb '01

16 Feb '01

I just bought 'The Haskell School of Expression', so I downloaded the latest version of Hugs98, February 2001, and the latest version of the Hugs Graphics Library, 2.0.3. Since the documentation mentioned building on Linux, I built and installed hugs98, ran some tests and felt it was okay. This is on a Redhat Linux 6.0 installation. Then I built the Hugs Graphics Library, and tried to run the first test suggested: ======================================================================= <1122:~/tmp/haskell> runhugs -Pgraphics-2.0.3/lib/x11: graphics-2.0.3/demos/HelloWorld.hs runhugs: Error occurred Reading file "graphics-2.0.3/demos/HelloWorld.hs": Reading file "graphics-2.0.3/lib/x11/GraphicsUtils.hs": Reading file "graphics-2.0.3/lib/x11/GraphicsCore.hs": Reading file "graphics-2.0.3/lib/x11/GraphicsEvent.hs": Reading file "graphics-2.0.3/lib/x11/GraphicsWindow.hs": Reading file "graphics-2.0.3/lib/x11/Xlib.hs": Reading file "graphics-2.0.3/lib/x11/StdDIS.hs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/Int.hs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/Bits.hs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/Int.hs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/Word.hs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/Addr.hs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/Monad.hs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/Foreign.hs": Reading file "graphics-2.0.3/lib/x11/StdDIS.hs": Reading file "graphics-2.0.3/lib/x11/X.hs": Reading file "graphics-2.0.3/lib/x11/Xlib.hs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/Concurrent.lhs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/ChannelVar.lhs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/ConcBase.hs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/ChannelVar.lhs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/Channel.lhs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/Semaphore.lhs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/SampleVar.lhs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/exts/Concurrent.lhs": Reading file "/user/donw/tmp/haskell/Hugs.ixl/share/hugs/lib/System.hs": Reading file "graphics-2.0.3/lib/x11/Graphics_Utilities.hs": Dependency analysis ERROR graphics-2.0.3/lib/x11/Graphics_Utilities.hs:73 - Illegal type "IOResult a" in constructor application <1123:~/tmp/haskell> pushd graphics-2.0.3 ~/tmp/haskell/graphics-2.0.3 ~/tmp/haskell <1125:~/tmp/haskell/graphics-2.0.3> find . -name '*hs' -print | xargs grep IOResult ./lib/x11/Graphics_Utilities.hs: mkErr :: Maybe (IOResult a) -> IOError ./lib/win32/GraphicsUtilities.hs: mkErr :: Maybe (IOResult a) -> IOError ======================================================================= I couldn't find a discussion of IOResult in the html versions of the Haskell 98 report or library report, but it is in the prelude: ======================================================================= ./lib/Prelude.hs:data IOResult ======================================================================= I'm still learning all the ins and outs of Haskell, so I don't really know what the problem is. Any helpful hints? -- Don Wakefield Mentor Graphics Corporation (503) 685-1262 8005 S.W. Boeckman Road don_wakefield(a)mentorg.com Wilsonville, OR 97070-7777

2 1

Stack depth
by Sengan 16 Feb '01

16 Feb '01

I have a program which works fine, except that for certain (large) input files I have to give it a larger than normal Stack (eg 4M in ghc). GHC allows one to profile heap, but I'd like to know how to profile stack contents. Is there a means with ghc/hugs/or something else? Does anyone have any suggestions on how to find out where stack's being used? Thanks Sengan

1 0