You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I realized that even the Newtype class only gets us to an approximation of the platonic ideal. In particular, my understanding is that n and n' are really supposed to be the same newtype, perhaps with different type arguments. Can we express that? I bet we can, with ConstraintKinds!
The original concept of this API is to work "over", "under", or
"a la" a *single* newtype, potentially changing its type arguments
but not changing it wholesale. Back in the day, Haskell couldn't
express that idea, but now it can (for the most part). We can take
advantage of that to improve inference and typed hole error messages.
Fixeslove-haskell#15
The original concept of this API is to work "over", "under", or
"a la" a *single* newtype, potentially changing its type arguments
but not changing it wholesale. Back in the day, Haskell couldn't
express that idea, but now it can (for the most part). We can take
advantage of that to improve inference and typed hole error messages.
Fixes#15
Also remove support for older GHC:
`Coercible` seems to be really fragile under GHC versions before
8.4. 8.2 seems especially fussy: even when the package compiles,
use cases may fail mysteriously.
I realized that even the
Newtype
class only gets us to an approximation of the platonic ideal. In particular, my understanding is thatn
andn'
are really supposed to be the same newtype, perhaps with different type arguments. Can we express that? I bet we can, withConstraintKinds
!I haven't tried compiling this, So I don't know if it works or even if it can be made to work, but I think it's a pretty interesting possibility.
The text was updated successfully, but these errors were encountered: