I suspect he means that with `t1 :: k1` and `t2 :: k2`, `t1 ~ t2` is
permitted as long as `k1 ~ k2` accompanies it. However, this would
already work with homogeneous ~, wouldn't it (because homogeneity is
up to whatever constraints are in scope)?
On Mon, Nov 23, 2015 at 11:05 AM, Richard Eisenberg
Naive question: Would it be "evil" or otherwise complicated to assume that (~) is heterogeneous only in the _presence_ of kind constraint? Or only when the kind can be inferred?
I'm afraid I don't understand. Do you mean that it's heterogeneous only when -XTypeInType is specified? That just might be possible, but I don't personally like it. Others might, though.
Richard
_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs