Simon Peyton-Jones:
> | - We need to distinguish between open and closed newtypes in a
> number of
> | places, because looking through newtypes doesn't work easily for
> open ones.
>
> I couldn't see where this "bites" except in TcTyDecls when computing
> recursiveness.
There are a few of places where GHC looks through newtypes:
* MkExternalCore.collect_tdefs, the computation of `repclause'
* SimplUtils.mkCase1 (here it actually reconstucts the worker
unfolding)
* Coercion.splitNewTypeRepCo_maybe
> The place where "looking through" newtypes is decided upon is in
> BuildTyCl line 145. I've added some more comments to explain; and also
> in TyCon.
>
> If we make (some, non-recursive) newtypes transparent again, we might
> have to remember *not* to make open ones transparent!
Part of the aim of the patch to which this log messages belongs is to
achieve exactly that. If I got it right, making some newtypes
transparent again should work fine without further tweaks. (Open ones
will always be treated as opaque, ie like recursive ones).
Manuel
> |
> | M ./compiler/basicTypes/DataCon.lhs -3 +5
> | M ./compiler/basicTypes/MkId.lhs -4 +26
> | M ./compiler/coreSyn/MkExternalCore.lhs -1 +1
> | M ./compiler/deSugar/DsUtils.lhs -1 +1
> | M ./compiler/simplCore/SimplUtils.lhs -2 +2
> | M ./compiler/typecheck/TcPat.lhs -1 +10
> | M ./compiler/types/Coercion.lhs -2 +2
> | M ./compiler/types/TyCon.lhs -1 +10
> | M ./compiler/types/Type.lhs -5 +4
> | _______________________________________________
> | Cvs-ghc mailing list
> | [EMAIL PROTECTED]
> | http://www.haskell.org/mailman/listinfo/cvs-ghc
_______________________________________________
Cvs-ghc mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/cvs-ghc
