StateT performance benchmarks?

[tomjaguarpaw]

@Liamzy

It’s nice to know that at least StateT.Strict is safe, but that calls to mind, why isn’t StateT.Strict the default on Control.Monad.Trans.State? You can’t really hide modify to get people to use the “correct” modify’ first, but at least pointing Control.Monad.Trans.State to StateT.Strict would make life simpler and Haskell less footgunny; same applies to the mtl library version.

Agreed, but that would potentially be a silent breaking change so is rather dangerous. We should, however, try to discourage people from using the Lazy versions and from using modify.