In the current Report draft, the idea is that a hiding clause hides both qualified and unqualified names
import Foo hiding( f ) hides Foo.f as well as unqualified f.
I think we all agreed that is a good change. But we (ahem, I) forgot to propagate that change to the sections about the Prelude.
That's odd, because in my paper copy of the Report, I had already (many months ago) scored out the following sentence in section 5.6.1: "The Prelude module is always available as a qualified import: an implicit `import qualified Prelude' is part of every module and names prefixed by `Prelude.' can always be used to refer to entities in the Prelude." So I think you must have declared this as a definite change at some point in the past, although I agree that the errata list as of today does not seem to specify its removal. Maybe a CVS conflict that got resolved the wrong way or something?
The fact that this makes the Report behave like GHC is purely coincidental :-)
And the fact that I have now (last week actually) fixed nhc98 to match ghc's behaviour, based on my already-amended paper copy of the Report, is just confirmation that this is the correct course of action. :-) Regards, Malcolm