Re: [GHC] #14451: Need proper SCC analysis of type declarations, taking account of CUSKs (was: Type synonym of type family causes error, error jumps to (fairly) unrelated source location)