mixphix

This is a tough nut to crack. To have documented the biasing principles when the typeclass was initially designed would have been ideal. Changing existing implementations might lead to bugs...

It appears that the MR contains only documentation changes. Per the README: > The primary responsibility of CLC is to manage API changes of base package. The ownership of base...

You want a newtype instance that will *prevent* `NFData` from fully evaluating it?

Strictness annotations are great for enforcing the evaluation order even in an `Applicative` context. ```hs foldMapA :: (Foldable t, Applicative f, Monoid y) => (x -> f y) -> t...

A draft merge request indicates a serious commitment to seeing the changes through. A branch with these changes would easily pass most of the GHC CI (in time). Benchmarks would...

As for the benchmarks of `foldMapA`: ``` All bench foldMapA: OK 6.29 ns ± 508 ps foldlM: OK 19.2 ms ± 1.6 ms foldlM_1: OK 24.0 ms ± 2.3 ms...

> The `foldA` benchmark isn't actually running the state computation. You need an `evalState` in `example`. You're so right! Let's fix that! ```hs main :: IO () main = defaultMain...

> [You're](https://github.com/haskell/core-libraries-committee/issues/255#issuecomment-1944470658) folding over a list ```hs benchFoldlM :: [Int] -> ((Int -> Int -> State () Int) -> Int -> [Int] -> State () Int) -> Int benchFoldlM xs...

@vdukhovni do you wish to proceed with the CLC proposal?

This is a breaking change, as the current type application order for `flip` is: ```hs flip @a @b @c :: (a -> b -> c) -> (b -> a ->...