More GND + role inference woes

Richard Eisenberg eir at cis.upenn.edu
Sun Dec 15 00:53:23 UTC 2013


Yes, I believe that's right. As far as I can figure out, these classes really *are* problematic, in that if we allowed GeneralizedNewtypeDeriving for them, there would be a way to subvert the type system. To make these derivable, we would need to be able to restrict various type parameters from taking on values that take a nominal argument. Without the ability to restrict the values in this way, there could be trouble.

Richard

On Dec 14, 2013, at 4:52 PM, Ben Gamari wrote:

> Edward Kmett <ekmett at gmail.com> writes:
> 
>> If this forced me to write those instances by hand, I could accept
>> that as a tax for correctness. It means you can't GND any of the
>> HasFoo dictionaries that lens builds, but meh.
>> 
> Am I correct in assuming that Bind, R1, R2, R3, and R4 are the
> problematic instances in linear? With recent GHC I get the errors below.
> 
> Cheers,
> 
> - Ben
> 
> 
>    src/Linear/Affine.hs:112:34:
>        Could not coerce from ‛f (f a)’ to ‛f (Point f a)’
>          because ‛f (f a)’ and ‛f (Point f a)’ are different types.
>          arising from the coercion of the method ‛join’ from type
>                       ‛forall a. f (f a) -> f a’ to type
>                       ‛forall a. Point f (Point f a) -> Point f a’
>        Possible fix:
>          use a standalone 'deriving instance' declaration,
>            so you can specify the instance context yourself
>        When deriving the instance for (Bind (Point f))
> 
>    src/Linear/Affine.hs:112:58:
>        Could not coerce from ‛g (f x)’ to ‛g (Point f x)’
>          because ‛g (f x)’ and ‛g (Point f x)’ are different types.
>          arising from the coercion of the method ‛core’ from type
>                       ‛forall a.
>                        ((forall (g :: * -> *) x.
>                          Functor g =>
>                          (x -> g x) -> f x -> g (f x))
>                         -> a)
>                        -> f a’
>                       to type
>                       ‛forall a.
>                        ((forall (g :: * -> *) x.
>                          Functor g =>
>                          (x -> g x) -> Point f x -> g (Point f x))
>                         -> a)
>                        -> Point f a’
>        Possible fix:
>          use a standalone 'deriving instance' declaration,
>            so you can specify the instance context yourself
>        When deriving the instance for (Core (Point f))
> 
>    src/Linear/Affine.hs:112:64:
>        Could not coerce from ‛f1 (f a)’ to ‛f1 (Point f a)’
>          because ‛f1 (f a)’ and ‛f1 (Point f a)’ are different types.
>          arising from the coercion of the method ‛_x’ from type
>                       ‛forall a (f :: * -> *). Functor f => (a -> f a) -> f a -> f (f a)’
>                       to type
>                       ‛forall a (f :: * -> *).
>                        Functor f =>
>                        (a -> f a) -> Point f a -> f (Point f a)’
>        Possible fix:
>          use a standalone 'deriving instance' declaration,
>            so you can specify the instance context yourself
>        When deriving the instance for (R1 (Point f))
> 
>    src/Linear/Affine.hs:112:68:
>        Could not coerce from ‛f1 (f a)’ to ‛f1 (Point f a)’
>          because ‛f1 (f a)’ and ‛f1 (Point f a)’ are different types.
>          arising from the coercion of the method ‛_xy’ from type
>                       ‛forall a (f :: * -> *).
>                        Functor f =>
>                        (V2 a -> f (V2 a)) -> f a -> f (f a)’
>                       to type
>                       ‛forall a (f :: * -> *).
>                        Functor f =>
>                        (V2 a -> f (V2 a)) -> Point f a -> f (Point f a)’
>        Possible fix:
>          use a standalone 'deriving instance' declaration,
>            so you can specify the instance context yourself
>        When deriving the instance for (R2 (Point f))
> 
>    src/Linear/Affine.hs:112:68:
>        Could not coerce from ‛f1 (f a)’ to ‛f1 (Point f a)’
>          because ‛f1 (f a)’ and ‛f1 (Point f a)’ are different types.
>          arising from the coercion of the method ‛_y’ from type
>                       ‛forall a (f :: * -> *). Functor f => (a -> f a) -> f a -> f (f a)’
>                       to type
>                       ‛forall a (f :: * -> *).
>                        Functor f =>
>                        (a -> f a) -> Point f a -> f (Point f a)’
>        Possible fix:
>          use a standalone 'deriving instance' declaration,
>            so you can specify the instance context yourself
>        When deriving the instance for (R2 (Point f))
> 
>    src/Linear/Affine.hs:112:72:
>        Could not coerce from ‛f1 (f a)’ to ‛f1 (Point f a)’
>          because ‛f1 (f a)’ and ‛f1 (Point f a)’ are different types.
>          arising from the coercion of the method ‛_xyz’ from type
>                       ‛forall a (f :: * -> *).
>                        Functor f =>
>                        (V3 a -> f (V3 a)) -> f a -> f (f a)’
>                       to type
>                       ‛forall a (f :: * -> *).
>                        Functor f =>
>                        (V3 a -> f (V3 a)) -> Point f a -> f (Point f a)’
>        Possible fix:
>          use a standalone 'deriving instance' declaration,
>            so you can specify the instance context yourself
>        When deriving the instance for (R3 (Point f))
> 
>    src/Linear/Affine.hs:112:72:
>        Could not coerce from ‛f1 (f a)’ to ‛f1 (Point f a)’
>          because ‛f1 (f a)’ and ‛f1 (Point f a)’ are different types.
>          arising from the coercion of the method ‛_z’ from type
>                       ‛forall a (f :: * -> *). Functor f => (a -> f a) -> f a -> f (f a)’
>                       to type
>                       ‛forall a (f :: * -> *).
>                        Functor f =>
>                        (a -> f a) -> Point f a -> f (Point f a)’
>        Possible fix:
>          use a standalone 'deriving instance' declaration,
>            so you can specify the instance context yourself
>        When deriving the instance for (R3 (Point f))
> 
>    src/Linear/Affine.hs:112:76:
>        Could not coerce from ‛f1 (f a)’ to ‛f1 (Point f a)’
>          because ‛f1 (f a)’ and ‛f1 (Point f a)’ are different types.
>          arising from the coercion of the method ‛_xyzw’ from type
>                       ‛forall a (f :: * -> *).
>                        Functor f =>
>                        (V4 a -> f (V4 a)) -> f a -> f (f a)’
>                       to type
>                       ‛forall a (f :: * -> *).
>                        Functor f =>
>                        (V4 a -> f (V4 a)) -> Point f a -> f (Point f a)’
>        Possible fix:
>          use a standalone 'deriving instance' declaration,
>            so you can specify the instance context yourself
>        When deriving the instance for (R4 (Point f))
> 
>    src/Linear/Affine.hs:112:76:
>        Could not coerce from ‛f1 (f a)’ to ‛f1 (Point f a)’
>          because ‛f1 (f a)’ and ‛f1 (Point f a)’ are different types.
>          arising from the coercion of the method ‛_w’ from type
>                       ‛forall a (f :: * -> *). Functor f => (a -> f a) -> f a -> f (f a)’
>                       to type
>                       ‛forall a (f :: * -> *).
>                        Functor f =>
>                        (a -> f a) -> Point f a -> f (Point f a)’
>        Possible fix:
>          use a standalone 'deriving instance' declaration,
>            so you can specify the instance context yourself
>        When deriving the instance for (R4 (Point f))



More information about the ghc-devs mailing list