Let us consider the following code snippet:
blah :: a -> b -> a blah x y = ble x where ble :: b -> b ble x = x
This compiles fine under GHC, which essentially means that b
from the 3rd line is something different than b
from the first line.
My question is simple: is there a way to somehow relate in the type declaration of ble
to a type used in an outer context, i.e. the type declaration of blah
?
Obviously, this is just an example and not a real-world use-case for type declarations.
This is possible with the ScopedTypeVariables extension. You need to use explicit forall's to bring the type variables into scope.
blah :: forall a b. a -> b -> a blah x y = ble x where ble :: b -> b ble x = x
Trying to load this definition with ScopedTypeVariables enabled gives:
foo.hs:2:16: Couldn't match type `a' with `b' `a' is a rigid type variable bound by the type signature for blah :: a -> b -> a at foo.hs:2:1 `b' is a rigid type variable bound by the type signature for blah :: a -> b -> a at foo.hs:2:1 In the first argument of `ble', namely `x' In the expression: ble x In an equation for `blah': blah x y = ble x where ble :: b -> b ble x = x
You can tell that GHC interprets the two b
s as the same type because the error says that a
and b
are bound on the same line.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With