fix(parser): correctly parse record and tuple tokens

[fedeci]

Q	A
Fixed Issues?	Fixes #13417
Patch: Bug Fix?	Y
Major: Breaking Change?
Minor: New Feature?
Tests Added + Pass?	Yes
Documentation PR Link
Any Dependency Changes?
License	MIT

#{ and #[ tokens now hold the correct location.
|] and |} are now finished as tokens and not operators.

[babel-bot]

Build successful! You can test your changes in the REPL here: https://babeljs.io/repl/build/46719/

[fedeci]

I noted that we don't have token specific fixtures, should we add those?
Moreover #13417 also mentions that value for the most of the tokens is undefined. We may do something like this in the Token constructor:

constructor(state: State) {
    this.type = state.type;
    if (state.value !== undefined) {
      this.value = state.value;
    }
    this.start = state.start;
    this.end = state.end;
    this.loc = new SourceLocation(state.startLoc, state.endLoc);
  }

but no tests are throwing, since they are not writing value in the output if undefined.

[fisker]

Thanks for the quick fix, have you checked |} token as well?

[nicolo-ribaudo]

I noted that we don't have token specific fixtures, should we add those?

@fedeci You can add a new babel-parser/test/tokens.js file with normal describe-it-expect Jest tests, or a new subfolder of test/fixtures.

[JLHwung]

We need similar fix for [| and {|.

[KFlash]

That token constructor code is super slow :) Why there is a need for a new SourceLocation. Can't this be a plain object?
And regarding the "missing" value. All object should keep the same shape to avoid hidden classes issue and perf regression in V8.

[codesandbox-ci]

This pull request is automatically built and testable in CodeSandbox.

To see build info of the built libraries, click here or the icon next to each commit SHA.

Latest deployment of this branch, based on commit 3d15e07:

Sandbox	Source
babel-repl-custom-plugin	Configuration
babel-plugin-multi-config	Configuration

[KFlash]

@JLHwung No plans to fix the performance issues?

[JLHwung]

@KFlash I am aware of the performance issue here, but fixing the performance falls out of the scope of this PR and thus should be addressed in another PR.