Leo White

> If I understand correctly, your PRs are essentially refining the leaves (and nodes?) of this tree That's right. > should mesh well with this work Yeah, it sounds like...

Presumably would be addressed by #8531

The printing of the type of `g` is correct with the "specification" of `-short-types` so there is nothing to change there. Also note that the implementation of short-paths in Merlin...

> Both of these talk about paths, but here the transformation is from a one-element path into another one-element path, right? In such cases I believe that both implementations try...

> I wonder if a version that uses the original path when the shortened one has the same length would make sense. I should say that I do think that...

See #8900

I think that [this section](https://github.com/ocaml/ocaml/blob/trunk/HACKING.adoc#your-first-compiler-modification) is the one @goldfirere used and it has a couple of short-comings: 1. It says: > 3\. Try the newly built compiler binaries ocamlc, ocamlopt...

Why is the syntax not the one used in the paper and prototype? The additional `module` keyword is unnecessary and seems cumbersome. The parser has changed to use menhir since...

> . I plan to get it sorted, but I thought the typechecker part would need more scrutiny, so best get it out for review sooner. Fair enough. Thanks for...

> This gives us some extra descriptive power for polymorphism that we didn't have before, As described in the implicits paper, this extension doesn't actually give any new power. As...