Re: [GHC] #14324: Consider deprecating STM invariant mechanism