GHC support for the new "record" package

[Christopher Done]

On 29 January 2015 at 00:40, Edward Kmett ekmett at gmail.com wrote:

There is a problem with the old TRex syntax.

In a world with kind signatures and rank-2 types, it would appear that

type Point2D = Rec ( x :: Coord, y :: Coord )

is ambiguous.

The kind-signature resemblance had occurred to me, but I’d assumed
Hugs treated it as syntactical sugar like [record|{ x :: Coord, y ::
Coord }|]. Apparently not.