GHC 8.2 bug where Coercible is not symmetric (on GHC 8.0, 8.2, somehow fixed on 8.4)

S.hs

-- Coercible not symmetric on GHC 8.0, 8.2, somehow fixed on 8.4

{-# LANGUAGE
    AllowAmbiguousTypes,
    FlexibleContexts,
    ScopedTypeVariables,
    TypeFamilies,
    TypeApplications
    #-}

module S where

import GHC.Generics
import Data.Coerce

type family F a :: * -> *

-- Type checks
e :: forall a b x. (Coercible (F a x) (F b x)) => F a x -> F b x
e = coerce @(F a x) @(F b x)

-- Doesn't type check
f :: forall a b x. (Coercible (F b x) (F a x)) => F a x -> F b x
f = coerce @(F a x) @(F b x)

{-
S.hs:24:5: error:
    • Could not deduce: Coercible (F a x) (F b x)
        arising from a use of ‘coerce’
      from the context: Coercible (F b x) (F a x)
        bound by the type signature for:
                   f :: Coercible (F b x) (F a x) => F a x -> F b x
        at S.hs:23:1-64
      NB: ‘F’ is a type function, and may not be injective
    • In the expression: coerce @(F a x) @(F b x)
      In an equation for ‘f’: f = coerce @(F a x) @(F b x)
    • Relevant bindings include
        f :: F a x -> F b x (bound at S.hs:24:1)
-}