To-do list for 4.0.0

[not-an-aardvark]

There are a lot of issues where we have discussed changing behavior in the next major version, but we haven't really kept track of these. This issue is intended to be a place to put breaking changes that we are evaluating, so that we don't forget make those changes when the release happens.

---

Accepted changes:

• Core:
  • Fix scoped plugin names (Unmatched rule names of scoped plugins. #6362)
  • Calculate leading/trailing comments (Calculate leading/trailing comments #6724)
  • Fix ignore-path relative paths (When running eslint from another directory and using --ignore-path, paths are not relative to .eslintignore location #6759)
  • Traverse into type annotations (Traverse into type annotations #7129)
• eslint:recommended changes:
  • (none yet)
• Other rule changes:
  • (none yet)

[not-an-aardvark]

Here are a few changes I think we should make:

• Rules to add to eslint:recommended:
  • (discussion moved to Update eslint:recommended for 4.0.0 #8236)
• Changes to rule defaults:
  • For indent, set {MemberExpression: 1} by default. (already done in rewrite)
  • For valid-typeof, set {requireStringLiterals: true} by default.
  • For no-extra-parens, set {nestedBinaryExpressions: false} by default (to avoid conflicting with no-mixed-operators)
  • For space-before-function-paren, set {asyncArrow: "always"} by default (it is currently "ignore" for backwards-compatibility, but it's inconsistent with the other options).
• Other:
  • Hopefully get the indent rewrite in (meta-issue: indent rule discussion #7274). The indent rewrite currently checks a lot of indentation cases that the current rule ignores. I think this is a good thing (ideally, no cases would be ignored by default), but it's possible it will cause a lot of builds to fail. We should make sure to put it in 4.0.0 if it's not already in by then.

[kaicataldo]

I'm hoping that the comment attachment changes also get in

[not-an-aardvark]

As discussed in #7689, we should consider reporting an error for for ObjectExpression and ArrayExpression errors/rejection reasons in no-throw-literal and no-reject-literal.

[platinumazure]

Possible addition: Refactor public-facing APIs to use ES6 classes, where feasible (this can't be done outside of semver-major, for sure).

[JamesHenry]

@nzakas Will we able to get eslint/eslint-scope into v4 do you think?

[nzakas]

Yeah, we should work to get eslint-scope into 4.0.0. We should probably open an issue on babel-eslint to let them know of the change ahead of time.

[not-an-aardvark]

I'm going to close this issue in favor of using the v4.0.0 project for tracking instead. The changes that have been suggested are all already on the tracking project in some form.