- Libraries - Haskell.org

RE: Combinators for ReadP
by Koen Claessen 28 Jul '04

28 Jul '04

Simon Peyton-Jones said: > Koen, not me! I know that one can get away with the local quantification by adding an extra type parameter to the type constructor: newtype ReadP r a = ... But that wouldn't do much good, would it? Regards, /Koen

2 2

RE: Prelude function suggestions
by Graham Klyne 28 Jul '04

28 Jul '04

At 10:16 28/07/04 +0100, Simon Marlow wrote: >Does anyone object to calling it 'split'? Not I. 'split' is a name I have used on occasion for my own definition. #g ------------ Graham Klyne For email: http://www.ninebynine.org/#Contact

2 1

Prelude function suggestions
by Henning Thielemann 28 Jul '04

28 Jul '04

There are some functions that I regularly use which are so general that I wonder if they should be in the Prelude. Maybe I have overlooked them or some generalizations of them. Here are the ones that I most frequently use: {- | Compositional power of a function, i.e. apply the function n times to a value. -} nest :: Int -> (a -> a) -> a -> a nest 0 f x = x nest n f x = f (nest (n-1) f x) {- | Split the list at the occurrences of a separator into sub-list. This is a generalization of 'words'. -} chop :: (a -> Bool) -> [a] -> [[a]] chop p s = let (l, s') = break p s in l : case s' of [] -> [] (_:rest) -> chop p rest {- | Returns 'Just' if the precondition is fulfilled. -} toMaybe :: Bool -> a -> Maybe a toMaybe False _ = Nothing toMaybe True x = Just x

4 3

RE: Combinators for ReadP
by Simon Peyton-Jones 28 Jul '04

28 Jul '04

| >| Maybe I should write a test suite too, while I'm at it. Is there any | >| specific location in the cvs tree where test suites should be put? | > | >Yes please! The right place is in | > testsuite/tests/ghc-regress/lib/ | >you can make a ReadP sub-directory. | | Er, doesn't that make it more difficult to create a library+test suite that | can be shipped with other compilers? | | E.g., for my XML parser work, I created a separate test suite subdirectory | in the source tree for that library. It may not be perfect, but that's the way the CVS source tree is currently arranged. Of course, a separate package can adopt whatever conventions it likes, but Martin was asking about the CVS source tree that contains the libs that come with GHC/Hugs etc. S

1 0

RE: Combinators for ReadP
by Simon Marlow 27 Jul '04

27 Jul '04

On 27 July 2004 15:16, Martin Sjögren wrote: > On Tue, 27 Jul 2004 14:31:21 +0100, Simon Peyton-Jones > <simonpj(a)microsoft.com> wrote: >> You could do this, I believe, but it'd make ReadP less covenient for >> most users, because they'd have to add these redundant (and >> inexplicable) 'r' parameters. By "inexplicable" I mean that given a >> value p::ReadP r a we can say that p is a parser that parses >> some input, returning a value of type 'a'... but what's this 'r' >> thing? >> >> Wrapping up a polymorphic fn in a newtype is very standard, even >> though it is not H98. >> >> I'd be pretty reluctant to make this change > > Oh I agree. Having to pass around the extra parameter sucks. > Unfortunately, nhc98 doesn't support forall. :) Incedentally, universal quantification in constructor fields is one of the easier extensions to implement. Much easier than multi-parameter type classes, for example. You don't have to go the whole hog and implement rank-N polymorphism ala GHC in order to get most of the benefit. Cheers, Simon

2 1

RE: Combinators for ReadP
by Simon Peyton-Jones 27 Jul '04

27 Jul '04

You could do this, I believe, but it'd make ReadP less covenient for most users, because they'd have to add these redundant (and inexplicable) 'r' parameters. By "inexplicable" I mean that given a value p::ReadP r a we can say that p is a parser that parses some input, returning a value of type 'a'... but what's this 'r' thing? Wrapping up a polymorphic fn in a newtype is very standard, even though it is not H98. I'd be pretty reluctant to make this change Simon | -----Original Message----- | From: libraries-bounces(a)haskell.org [mailto:libraries-bounces@haskell.org] On Behalf Of Martin | Sjögren | Sent: 26 July 2004 12:18 | To: libraries(a)haskell.org | Subject: Re: Combinators for ReadP | | On Mon, 26 Jul 2004 11:00:31 +0100, Malcolm Wallace | <malcolm.wallace(a)cs.york.ac.uk> wrote: | > > While we're discussing changes, is there an overwhelmingly good reason | > > for the use of local quantification beyond making the type signatures | > > simpler? Writing | > > newtype ReadP r a = R ((a -> P r) -> P r) | > > (like in ContT) would make ReadP completely Haskell98 as far as I can tell. | > | > I think you'll find that without the local universal quantifier, | > you cannot make ReadP a (useful) instance of Monad, because your | > 'r' parameter would be fixed across >>=, whereas it really needs to | > be variable. | | Uhh. But it works for ContT, so why not for ReadP? The only point time | you have to fix the r is when you want to "run" the parser, all other | parsers will be polymorphic in r: | | parseFoo :: ReadP r Foo | parseBar :: ReadP r Bar | | readP_to_S :: ReadP a a -> ReadS a | | I do exactly this in the package description parsing code of Cabal | (using a locally hacked ReadP, Cabal really must be H98...), and it | works just fine. | | | /Martin | _______________________________________________ | Libraries mailing list | Libraries(a)haskell.org | http://www.haskell.org/mailman/listinfo/libraries

2 1

RE: Haskell Extensions in various compilers (Cabal)
by Simon Marlow 26 Jul '04

26 Jul '04

On 13 July 2004 04:15, Isaac Jones wrote: > - Do any of Hugs, GHC, and NHC disallow Long.Hierarchical.Module.Names > by default? Can they all turn them off? GHC includes them by default, and they cannot be turned off. This is a bug (but we're unlikely to fix it). My only comment is that some of the extensions you've formalised are probably a bit too experimental or GHC-specific to be included in the general Extension type. I'd leave out InlinePhase for one. PatternTypeAnnotations should probably be called ScopedTypeVariables. I'm not sure what UnsafeOverlappingInstances does compared with AllowOverlappingInstances and the other instance-related flags. Cheers, Simon

4 6

RE: Combinators for ReadP
by Simon Peyton-Jones 26 Jul '04

26 Jul '04

Koen, not me! Simon | -----Original Message----- | From: libraries-bounces(a)haskell.org [mailto:libraries-bounces@haskell.org] On Behalf Of Malcolm | Wallace | Sent: 26 July 2004 17:04 | To: libraries(a)haskell.org | Subject: Re: Combinators for ReadP | | Martin Sjögren <msjogren(a)gmail.com> writes: | | > > I think you'll find that without the local universal quantifier, | > > you cannot make ReadP a (useful) instance of Monad, because your | > > 'r' parameter would be fixed across >>=, whereas it really needs to | > > be variable. | > | > Uhh. But it works for ContT, so why not for ReadP? The only point time | > you have to fix the r is when you want to "run" the parser, all other | > parsers will be polymorphic in r: | | OK, cool. If you can successfully convert ReadP to Haskell'98, we'd | all be delighted. But Koen and Simon PJ are the people most likely | to be able to give relevant comment on whether any subtle properties | will be lost. | | Regards, | Malcolm | _______________________________________________ | Libraries mailing list | Libraries(a)haskell.org | http://www.haskell.org/mailman/listinfo/libraries

1 0

RE: Combinators for ReadP
by Simon Peyton-Jones 26 Jul '04

26 Jul '04

Martin Thanks. If you were to take the existing ReadP module, add your stuff, Haddockise it with explanatory text, then I'll put the result in place. ReadP would be much more useful if it had some introductory Haddock documentation too, including a reference to Koen Claessen's paper, if you felt up to doing that too. Many thanks for your help. Simon | -----Original Message----- | From: libraries-bounces(a)haskell.org [mailto:libraries-bounces@haskell.org] On Behalf Of Martin | Sjögren | Sent: 22 July 2004 23:42 | To: libraries(a)haskell.org | Subject: Combinators for ReadP | | So I was wondering why there weren't any helpful combinators defined | in the ReadP module and found myself rewriting them over and over. I | put a bunch of combinators in the attached module (based on | corresponding combinators in Parsec.Combinator) and it would be neat | if they could be incorporated in the standard ReadP module in some | form. | | Cheers, | Martin

5 8

Haskell Graphics
by Alejandro Ramiro Ruiz 26 Jul '04

26 Jul '04

1 0