[Haskell-cafe] Why distinct tyvars in instance declarations?

[Josh Hoyt]

On 6/27/05, robert dockins <robdockins at fastmail.fm> wrote:
>
> I think that you are mistaken.  The OP listed:
> 
>  > class Foo a
>  > instance Foo (Either b b)
> 
> Without further information, Haskell compilers will assume that the
> type(s) in a class declaration has/have kind * (Report section 4.6).
> Either b b does have kind *, so that's not the problem.
>

Indeed.

> class Foo a
>
> instance Foo (Either b c)

*is* accepted, so this is not a kind error.

Josh Hoyt