| What I don't understand, however, is why adding that extra | constraint helps in any way? Shouldn't the addition of new | things in the context only make more options available? Why | does it make some of the existing, amgibuous options go away?
With the extra constraint, the instance decl can build a C (T a) dictionary from a (C a, C [a]) dictionary, by choosing the expl method from the C [a] dictionary it is passed (ignoring the C a one). That defers the choice of which of the overlapping instance decls we are going to use. Perhaps it'll be postponed to a point at which 'a' is known, in which case the choice is easy. The point is that we don't want to have to choose between the C String and C [a] instance decls until we know enough about 'a' to choose the right one.
Yes, but isn't that an implementation problem surfacing at the language level? All the dictionaries needed to delay the decision to the point of use could also be made available when compiling the original program, no? After all, that's the reason why there's an ambiguity in the first place. Not to mention that in the case for which there is an overlap, the String instance will always be chosen as the more specific one.. Claus
| data T a ý [a] | | class C t where | expl :: t -> String | expl x ÿdefault" | | instance C String where expl s ÿString" | instance C a ÿC [a] where expl l ÿ[a]" | | instance (C a {- ,C [a] -} ) ÿC (T a) where | expl (D xs) þxpl xs | | main ÿutStrLn $ expl "hi" | | ------------ | | As is, both ghc and hugs reject the program, whereas | both accept it with the extra constraint in the C (T a) | instance.. Now, I think I can see how the right-hand-side | expl could come either from the C String or from the C [a] | instance - hence ghc's message: | | $ ghc --make Tst.hs | c:\ghc\ghc-5.04\bin\ghc.exe: chasing modules from: Tst.hs | Compiling Main ( Tst.hs, ./Tst.o ) | | Tst.hs:15: | Could not unambiguously deduce (C [a]) | from the context (C (T a), C a) | The choice of (overlapping) instance declaration | depends on the instantiation of `a' | Probable fix: | Add (C [a]) to the class or instance method `expl' | Or add an instance declaration for (C [a]) | arising from use of `expl' at Tst.hs:15 | In the definition of `expl': expl xs | | | Confused, | Claus | | PS. Perhaps related, but why does Hugs seem to ignore the | C a constraint in the context of the original version? | | $ hugs -98 Tst.hs | __ __ __ __ ____ ___ | _________________________________________ | || || || || || || ||__ Hugs 98: Based on the Haskell 98 | standard | ||___|| ||__|| ||__|| __|| Copyright (c) 1994-2001 | ||---|| ___|| World Wide Web: | http://haskell.org/hugs | || || Report bugs to: | hugs-bugs@haskell.org | || || Version: December 2001 | _________________________________________ | | Hugs mode: Restart with command line option +98 for Haskell 98 mode | | Reading file "c:\Program Files\Hugs98\\lib\Prelude.hs": | Reading file "Tst.hs": | Type checking | ERROR "Tst.hs":15 - Cannot justify constraints in instance member | binding | *** Expression : expl | *** Type : C (T a) ÿT a -> String | *** Given context : C (T a) | *** Constraints : C [a] | | Prelude> | _______________________________________________ | Glasgow-haskell-bugs mailing list | Glasgow-haskell-bugs@haskell.org | http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs