Support TypeScript 4.0

[sosukesuzuki]

TypeScript 4.0 beta has been released!

https://devblogs.microsoft.com/typescript/announcing-typescript-4-0-beta/

We need to support new syntaxes:

• Variadic Tuple Types
• Labeled Tuple Elements
• Short-Circuiting Assignment Operators
  • This syntax is already supported via Babel (https://babeljs.io/docs/en/next/babel-plugin-proposal-logical-assignment-operators.html)

note: We should wait typescript-eslint/typescript-eslint#2260 for default TS parser and babel/babel#11760 for babel-ts parser.

[orta]

Also a catch clause can legally have a type now (only : any and : unknown are allowed) - today Prettier drops it entirely

Playground

[bradennapier]

Yep and it may be useful to drop if its anything other than any or unknown but may be getting more into a linting-like situation there which @typescript-eslint should instead implement.

Def looking forawrd to this for the variadic tuples once prettier supports it :-P

another one to look out for that isn't obvious at first:

type ReduceNextElement<
  T extends readonly unknown[]
  // eslint-disable-next-line prettier/prettier
> = T extends readonly [infer V, ...infer R] ? [V, R] : never

Prettier will currently change this to

type ReduceNextElement<T extends readonly unknown[]> = T extends readonly [
  infer V,
  ...(infer R)
]
  ? [V, R]
  : never;

which produces a ts error:

Where we expect the following:

type Check = ReduceNextElement<[1, 2, 3]>; // [1, [2, 3]]

[fisker]

Type annotations on catch clauses #8805

[wclr]

Labeled tuples not supported.

[fisker]

Labeled tuples for babel-ts #8885

[liamcarter111]

Since Typescript official 4.0.2 was released to the masses yesterday, any update on when #8982 will be released?

[sosukesuzuki]

We are working for release 2.1(#8882), please wait.

[papb]

Hello @sosukesuzuki, I am still experiencing the issue reported by @bradennapier above, in which ...infer T is incorrectly replaced with ...(infer T). I am using Prettier 2.1.0. I will work around this with // prettier-ignore for now.

[bradennapier]

Figured it'd get missed ;) subtle difference but completely breaks inference.

[fisker]

Can you raise another issue? I think we missed that

[papb]

@fisker Done! #9041