V5 Planned Changes

[bmish]

Starting a list of my desired breaking changes, feedback welcome. Likely this release won't happen until 2022-05 or 2022-06 when we can drop Node 12/17 support, unless we come up with additional breaking changes we want to release sooner.

Breaking changes for v5:

• breaking: drop Node 12/17 support #256
• breaking: drop ESLint v6 support #257
• breaking: strictly define Node API #259
• breaking: add recommended rules prefer-message-ids, prefer-output-null, no-missing-message-ids, no-unused-message-ids #258
• Update existing rules:
  • Update no-missing-placeholders and no-unused-placeholders to handle messageIds #252

Future major version (v6 or later):

• Convert package to ESM or TypeScript?

For reference, V4 was release 2021-10 (CHANGELOG, discussion in #120).

[dimaMachina]

Convert to Typescript is better 🙂

[aladdin-add]

the plugin has supported ts. is there something stopping linting ts code?

[dimaMachina]

I wanted to say convert package to ts

[aladdin-add]

I did not see why devs will import the plugin, so converting to ts seems not compelling to me.

[bmish]

I'm supportive of converting this package to both ESM and TypeScript. Both would improve the internal developer experience (TypeScript especially is great for safely traversing the AST).

Neither of these would really affect consumers since I don't think developers import any utils from this plugin (and they will be prevented from doing so once we add "exports": "./lib/index.js" unless we decide to specifically expose something).

[G-Rath]

@bmish I'm happy to help out with that if you like (I've converted a few plugins to TypeScript now, including eslint-plugin-jest).

[bmish]

@aladdin-add status update:

• I've opened up PRs for all the breaking changes for v5
• I think we should consider any conversion to ESM/TypeScript in the future and not in this major release since we already have enough changes
• I believe this is ready to ship

[aladdin-add]

should we wait for the eslint v9 to be released, to see if need some more breaking changes?

[bmish]

@aladdin-add unless we have a specific need to wait, I recommend releasing now, and here's why:

• ESLint v8 wasn't released until October, so ESLint v9 could be several months away still. If we do end up wanting to make any more changes for ESLint v9, those might just be bug fixes and not necessarily breaking changes. And any such changes would require additional implementation time, further delaying the release.
• Based on the current planned changes for ESLint v9, I don't immediately see anything necessitating a change.
• Releasing would provide users with a version that is fully optimized for ESLint v8.
• Releasing allows us to begin taking advantage of Node 14 features in our PRs.
• The sooner we release, the sooner we begin to receive feedback on the new recommended rules.
• We already have a sizable set of breaking changes lined up, and these releases become unwieldy when the change list grows too big.
• We can perform another release in Q4 if we've gathered worthwhile changes by then.

[aladdin-add]

closing, as v5 has been released: https://github.com/not-an-aardvark/eslint-plugin-eslint-plugin/releases/tag/v5.0.0