Josh Berdine

This seems like a good idea to me. I think calling it "refactor-friendly" is a tangent: this proposal reduces indentation in a way that does not cause ambiguity, so good....

This is due to a current deficiency in the internal implementation. The core parenthesization logic was designed without taking attributes fully into account. In particular, there are more forms of...

I don't know the best implementation strategy. The issues I am trying to resolve are, I think, similar in nature to the following example. Formatting: ```ocaml val trms : t...

Hmm, I am literally comparing this PR with master. I'll dig some more to see if I can understand what is happening.

Ugh, this is an instance of https://github.com/ocaml-ppx/ocamlformat/issues/1675. If you start with master and format ``` val trms : t -> trm iter (** [trms a] enumerates the indeterminate terms appearing...

Hmm, this feels like another case of odoc just doing something different from what is in the ocaml manual, so I don't know. To my understanding, `{[...]}` are described as...

I don't think that there is an odoc issue about this. I'm not sure of the constraints on the odoc side, and on the interop between ocamlformat and odoc. We...

In the original report I was mistaken to bring up utf length computation, that turned out not to be the issue. The issue in the [comment](https://github.com/ocaml-ppx/ocamlformat/pull/1736#issuecomment-884488828) above remains though, and...

FWIW, almost from the beginning I have wanted to refactor to combine the format stream monad hiding inside Fmt with the state monad implicit in the comments tables. I wonder...

This code is formatted by `Fmt_ast.fmt_value_binding` which itself produces the code from `let` to `=`, and then uses `fmt_expression` to produce from `fun` onward. To achieve what you suggest, I...