- Glasgow-haskell-users

Re: monomorphism/hugs98/ghc-5.02.3
by Matt Fairtlough 31 May '02

31 May '02

Thanks a lot for the solution! And of course the ghci behaviour is much more sensible than the hugs one. And for pointing me to the mail archive. It encourages me to submit my next problem: stack overflow with simple Show instances...more later I wonder if there is a list of Hugs to GHCI "gotchas"? Matt. --------------------------------------------------------------------------- Matt Fairtlough m.fairtlough(a)dcs.shef.ac.uk Verification and Testing Group Room 115 Department of Computer Science University of Sheffield Regent Court, 211 Portobello Street Sheffield S1 4DP Tel: (0)114-22-21826 UK Fax: (0)114-22-21810 WWW: http://www.dcs.shef.ac.uk/~matt/ ----------------------------------------------------------------------------

3 2

Re: monomorphism/hugs98/ghc-5.02.3
by Matt Fairtlough 31 May '02

31 May '02

Hmm. It might be related but I somehow doubt it. The buggy program is ------------------------------------------ module Bug2 where data Bug2 = Bug2 instance Show Bug2 where showsPrec _ Bug2 str = showChar '?' str ------------------------------------------ (stack overflow on Bug2> Bug2) and the fix is ------------------------------------------ module Bug2 where data Bug2 = Bug2 instance Show Bug2 where showsPrec _ Bug2 str = showChar '?' str ------------------------------------------ -- an indentation problem. Quite why globally redefining showsPrec produces a stack overflow when showing Bug2 values I'll have to leave to the experts. --------------------------------------------------------------------------- Matt Fairtlough m.fairtlough(a)dcs.shef.ac.uk Verification and Testing Group Room 115 Department of Computer Science University of Sheffield Regent Court, 211 Portobello Street Sheffield S1 4DP Tel: (0)114-22-21826 UK Fax: (0)114-22-21810 WWW: http://www.dcs.shef.ac.uk/~matt/ ----------------------------------------------------------------------------

1 0

instance Show (Ptr a)
by Hal Daume III 31 May '02

31 May '02

The GHC docs claim Ptr a is an instance of Eq, Ord and Show, but it doesn't actually seem to have a Show instance. At least the docs should reflect the fact that there is not instance; preferably, could someone actually make an instance? It would be nice (primarily for debugging purposes) to be able to see where the allocations from malloc, newArray, etc took place. - Hal -- Hal Daume III "Computer science is no more about computers | hdaume(a)isi.edu than astronomy is about telescopes." -Dijkstra | www.isi.edu/~hdaume

2 1

monomorphism/hugs98/ghc-5.02.3
by Matt Fairtlough 31 May '02

31 May '02

Hello Haskellers, I've just joined this email group because I am teaching Haskell to our 2nd year students. I'm not a Haskell expert but I do enjoy functional programming. I don't know if anyone can help with this; I've checked the FAQs and the bug-reports on SourceForge and found nothing but I'd have thought it might be a common problem for folks migrating from hugs to ghci. I'm used to using hugs in haskell 98 mode but it seems that functions need not be monomorphic in where clauses in that case (is this a departure from the standard?). However, in ghc-5.02.3 for Haskell 98 (without glasgow extensions) it seems that is not the case. I _guessed_ that I need to use the flag -fno-monomorphism-restriction but I don't know whether I've failed to turn it on or if the problem is to be found somewhere else. A related problem I've found is that if I try to specify that flag in a pragma, I get an error message stating that static flags may not be specified in pragmas, and yet the flag is supposed to be dynamic. Here is a simple instance of a program that runs under Hugs but not ghc(i)-5.02.3 : module Test where class Bounded a => Testit a where test1 :: a data Test a = Test {f1::[a], f2::[Char]} class Super a where sup1 :: a -> Maybe a instance (Eq a, Testit a) => Super (Test a) where sup1 (Test{f1=x, f2=y}) = Just Test{f1=x, f2=y} where y' = top y x' = top x top :: [a] -> a top (s:st) = s is it possible to compile this program in ghc-5.02.3? Any advice much appreciated, Matt. --------------------------------------------------------------------------- Matt Fairtlough m.fairtlough(a)dcs.shef.ac.uk Verification and Testing Group Room 115 Department of Computer Science University of Sheffield Regent Court, 211 Portobello Street Sheffield S1 4DP Tel: (0)114-22-21826 UK Fax: (0)114-22-21810 WWW: http://www.dcs.shef.ac.uk/~matt/ ----------------------------------------------------------------------------

2 1

monomorphism/haskell98
by Matt Fairtlough 30 May '02

30 May '02

I don't know if anyone can help with this; I've checked the FAQs and the bug-reports on SourceForge and found nothing but I'd have thought it might be a common problem for folks migrating from hugs to ghci. I'm used to using hugs in haskell 98 mode but it seems that functions need not be monomorphic in where clauses in that case (is this a departure from the standard?). However, in ghc-5.02.3 for Haskell 98 (without glasgow extensions) it seems that is not the case. I _guessed_ that I need to use the flag -fno-monomorphism-restriction but I don't know whether I've failed to turn it on or if the problem is to be found somewhere else. A related problem I've found is that if I try to specify that flag in a pragma, I get an error message stating that static flags may not be specified in pragmas, and yet the flag is supposed to be dynamic. Here is a simple instance of a program that runs under Hugs but not ghc(i)-5.02.3 : module Test where class Bounded a => Testit a where test1 :: a data Test a = Test {f1::[a], f2::[Char]} class Super a where sup1 :: a -> Maybe a instance (Eq a, Testit a) => Super (Test a) where sup1 (Test{f1=x, f2=y}) = Just Test{f1=x, f2=y} where y' = top y x' = top x top :: [a] -> a top (s:st) = s is it possible to compile this program in ghc-5.02.3? Any advice much appreciated, Matt. -- --------------------------------------------------------------------------- Matt Fairtlough m.fairtlough(a)dcs.shef.ac.uk Verification and Testing Group Room 115 Department of Computer Science University of Sheffield Regent Court, 211 Portobello Street Sheffield S1 4DP Tel: (0)114-22-21826 UK Fax: (0)114-22-21810 WWW: http://www.dcs.shef.ac.uk/~matt/ ----------------------------------------------------------------------------

1 0

instance Ord FiniteMap
by Hal Daume III 30 May '02

30 May '02

Is there any particular reason FiniteMap (and hence Set) aren't instances of Ord? I realize it's "weird" to define a map to be ordered, but even if the Ord definition were in some sense "nonsensical", being able to have, for instance, Sets of Sets of things would be really nice. -- Hal Daume III "Computer science is no more about computers | hdaume(a)isi.edu than astronomy is about telescopes." -Dijkstra | www.isi.edu/~hdaume

3 5

specifying common 'dis's in greencard
by Hal Daume III 30 May '02

30 May '02

Hi, I have three modules which all use the same (or some of the same) %dis directives; I'd like to pull these out into their own file (a module called "MyDIS" or something), but I can't figure out how to get green-card to look at those definitions (I've `import`ed it in the modules that need it and put it's directory in the search path with -i). Unfortunately, even so it complains that it doesn't know how to marshall my datatypes. Any suggestions? - Hal -- Hal Daume III "Computer science is no more about computers | hdaume(a)isi.edu than astronomy is about telescopes." -Dijkstra | www.isi.edu/~hdaume

3 4

RE: instance Ord FiniteMap
by Simon Marlow 29 May '02

29 May '02

I'll happily incorporate the code if someone sends me a patch... Cheers, Simon > -----Original Message----- > From: Hal Daume III [mailto:hdaume@ISI.EDU] > Sent: 29 May 2002 15:03 > To: Johannes Waldmann > Cc: GHC Users Mailing List > Subject: Re: instance Ord FiniteMap > > > I agree; the problem is that I fear that making my own > instance by using > setToList will be very inefficient (or at least much more so than an > instance which actually looks at the tree structure). > > -- > Hal Daume III > > "Computer science is no more about computers | hdaume(a)isi.edu > than astronomy is about telescopes." -Dijkstra | www.isi.edu/~hdaume > > On Wed, 29 May 2002, Johannes Waldmann wrote: > > > > for instance, Sets of Sets of things would be really nice. > > > > Sure. One could simply use lexicographic ordering > > (i. e. s1 `compare` s2 = setToList s1 `compare` setToList s2) > > or length-lexicographic ordering (for efficiency) > > ... = (cardinality s1, setToList s1) `compare` > (cardinality s2, setToList s2) > > > > As you write, there seems to be no reason not to do this. > > An Ord instance should be a linear ordering, and the above are. > > -- > > -- Johannes Waldmann ---- > http://www.informatik.uni-leipzig.de/~joe/ -- > > -- > joe(a)informatik.uni-leipzig.de -- phone/fax (+49) 341 9732 204/252 -- > > > > _______________________________________________ > Glasgow-haskell-users mailing list > Glasgow-haskell-users(a)haskell.org > http://www.haskell.org/mailman/listinfo/glasgow-haskell-users >

3 2

Re: GHC ebuild done!
by Sven Moritz Hallberg 29 May '02

29 May '02

On Wednesday 29. May 2002 17:07, you wrote: > > With the help of the excellent new guide on porting GHC > > (Thank you Simon!) > > I have just created an ebuild for GHC. I thought I'd let you > > know, in case > > you're still working on it. I don't bootstrap GHC 5 directly > > from HC files, > > but GHC 4.08.3 (as some suggested) and go on from there. I've > > submitted the > > ebuild yesterday as Bug #3092 to the Gentoo Bugzilla. Feel > > free to comment if > > you want to look at it. > > Does this mean we can claim that Gentoo Linux is supported? :) I hope it will mean that in the very near future. That is, once the ebuild is in Gentoo CVS. I'm currently waiting for that to happen, but they're probably still waiting for their compile to finish... ;^) > Is there a > useful link I can add from the GHC download pages? Hm. No. Ebuilds are the fundament of Portage, the central Gentoo Linux package system, which is directly available to anyone running Gentoo Linux. Maybe you can add a comment saying that all Gentoo users with a recent Portage tree can now simply 'emerge ghc' for much Haskell fun. But, wait with that until the ebuild enters CVS. I'll make a seperate announcement to this liste once that's happened. So long, Sven Moritz

1 0

RE: GMP from haskell
by Simon Marlow 29 May '02

29 May '02

> I should have been more specific :). I was referring to the more > complicated things, such as Lucas numbers, Binomial coefficients, etc. There isn't a full GMP binding as such. However, we do make use of more than just the simple arithmetic in GMP - for example, Prelude.gcd on Integers uses GMP's gcd, and similarly for lcm. Cheers, Simon

1 0