DerivingVia fails to roundtrip through TH

The following programs works on GHC HEAD, as expected:

{-# LANGUAGE DerivingVia #-}
{-# LANGUAGE ExplicitForAll #-}
{-# LANGUAGE StandaloneDeriving #-}
{-# LANGUAGE TemplateHaskell #-}
module Bug where

newtype List a = MkList [a]
deriving via forall a. [a] instance Eq a => Eq (List a)

However, if you declare the standalone-derived instance using Template Haskell quotes:

$([d| deriving via forall a. [a] instance Eq a => Eq (List a) |])

Then it fails to typecheck:

[1 of 1] Compiling Bug              ( Bug.hs, interpreted )

Bug.hs:8:2: error:
    • Couldn't match representation of type ‘a’ with that of ‘a1’
        arising from a use of ‘GHC.Prim.coerce’
      ‘a’ is a rigid type variable bound by
        the instance declaration
        at Bug.hs:8:2-65
      ‘a1’ is a rigid type variable bound by
        the instance declaration
        at Bug.hs:8:2-65
    • In the expression:
        GHC.Prim.coerce
          @([] a_a3ek -> [] a_a3ek -> Bool) @(List a -> List a -> Bool)
          ((==) @([] a_a3ek))
      In an equation for ‘==’:
          (==)
            = GHC.Prim.coerce
                @([] a_a3ek -> [] a_a3ek -> Bool) @(List a -> List a -> Bool)
                ((==) @([] a_a3ek))
      When typechecking the code for ‘==’
        in a derived instance for ‘Eq (List a)’:
        To see the code I am typechecking, use -ddump-deriv
      In the instance declaration for ‘Eq (List a)’
    • Relevant bindings include
        (==) :: List a -> List a -> Bool (bound at Bug.hs:8:2)
  |
8 | $([d| deriving via forall a. [a] instance Eq a => Eq (List a) |])
  |  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Bug.hs:8:2: error:
    • Couldn't match representation of type ‘a’ with that of ‘a1’
        arising from a use of ‘GHC.Prim.coerce’
      ‘a’ is a rigid type variable bound by
        the instance declaration
        at Bug.hs:8:2-65
      ‘a1’ is a rigid type variable bound by
        the instance declaration
        at Bug.hs:8:2-65
    • In the expression:
        GHC.Prim.coerce
          @([] a_a3ek -> [] a_a3ek -> Bool) @(List a -> List a -> Bool)
          ((/=) @([] a_a3ek))
      In an equation for ‘/=’:
          (/=)
            = GHC.Prim.coerce
                @([] a_a3ek -> [] a_a3ek -> Bool) @(List a -> List a -> Bool)
                ((/=) @([] a_a3ek))
      When typechecking the code for ‘/=’
        in a derived instance for ‘Eq (List a)’:
        To see the code I am typechecking, use -ddump-deriv
      In the instance declaration for ‘Eq (List a)’
    • Relevant bindings include
        (/=) :: List a -> List a -> Bool (bound at Bug.hs:8:2)
  |
8 | $([d| deriving via forall a. [a] instance Eq a => Eq (List a) |])
  |  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Eek. What went wrong? -ddump-splices reveals the answer:

[1 of 1] Compiling Bug              ( Bug.hs, interpreted )
Bug.hs:8:2-65: Splicing declarations
    [d| deriving via forall a_a3g8. [a_a3g8] instance Eq a_a3g8 =>
                                                      Eq (List a_a3g8) |]
  ======>
    deriving via forall a_a3gW. [a_a3gW] instance Eq a_a3g8 =>
                                                  Eq (List a_a3g8)

By the time the declaration has been spliced in, the a in the [a] part is no longer the same a as in the Eq a => Eq (List a) part. Crucially, these must be the same in order for this instance to work.