Hyeseong Kim

Personally, I don't prefer to call the "targeted module format" config as just "type" And it might be confusing with the standard `"type"` attribute. ```res @module({ type_: "esmodule", with: {...

> Ok, and what about `format`? I think that it's good if its mental model is the format of the `from` source, not to tailor the compilation output.

Maybe it's better to merge #6979 first, then I'll rebase

Actually, it was a bit improved like https://github.com/rescript-lang/rescript-compiler/pull/6984/commits/03fe31086dccfe98ca91fe66e92983df95bb26b6#diff-ef282bde01036b4b8ce35133e81733df0530c2e4192992862d2ef200c114e5e0 Having dedicated variants rather than just Pccall makes analyzing side effects /specialization much easier.

> I tested against one of our projects and found that the array indexing now does not return `option` anymore, even if Core is open. > > I.e. this compiles:...

Added compatibility/deprecations for `Pervasives`, `Array`, `List`, `Char`, and `String`. Additionally, I modified the `ninja.js` to make `runtime` dir include only primitives and `others` dir include only stdlib. `Pervasives` now belong...

I didn't make all functions compatible. I removed functions and OCaml primitives that were already deprecated. AFAIK we've been seeing messages like `"use Belt.Int.toString instead"` for a long time now....

> Core will not be optional, as in it will always be available and on by default. This is because one of the big goals with Core is to provide...

> Could you just bring back the "cargo clean" in the Makefile in some form or other a discussed above? Then good to go from my point of view. Added...

Hmm, it seems something in the ninja script is broken. I need to make sure the logic isn't changed in this PR