[Bug]: regression for module block parse

[fisker]

💻

• Would you like to work on a fix?

How are you using Babel?

Programmatic API (babel.transform, babel.parse)

Input code

> require('@babel/parser').parse('module {/* comment */}', {plugins: ['moduleBlocks']}).program.body[0].expression.body
Node {
  type: 'Program',
  start: 21,   // <---------------------
  end: 8,  // <---------------------
  loc: SourceLocation {
    start: Position { line: 1, column: 21, index: 21 },
    end: Position { line: 1, column: 8, index: 8 },
    filename: undefined,
    identifierName: undefined
  },
  sourceType: 'module',
  interpreter: null,
  body: [],
  directives: [],
  trailingComments: [
    {
      type: 'CommentBlock',
      value: ' comment ',
      start: 8,
      end: 21,
      loc: [SourceLocation]
    }
  ]
}

Configuration file name

No response

Configuration

N/A

Current and expected behavior

Range should be 8 ~ 21

Environment

N/A

Possible solution

No response

Additional context

Possible introduced in #14980 @JLHwung

[babel-bot]

Hey @fisker! We really appreciate you taking the time to report an issue. The collaborators on this project attempt to help as many people as possible, but we're a limited number of volunteers, so it's possible this won't be addressed swiftly.

If you need any help, or just have general Babel or JavaScript questions, we have a vibrant Slack community that typically always has someone willing to help. You can sign-up here for an invite.

[nicolo-ribaudo]

If anyone has trouble understanding the problem: start and end are swapped.

[liuxingbaoyu]

It's weird, escaping our testing and even end-to-end testing.

[JLHwung]

In Babel-parser 7.18.0, the program node starts at 21 and ends with 22:

https://astexplorer.net/#/gist/1b193f48d92b558bcd78e1516bef9c71/5a9bf187ab80c21af33bbbff591361d5f33eb533

After #14980 it starts at 21 but ends at 8, which is worse. The previous situation is not ideal, the Program node should start at comment's start (8) and end at 21, like how we parse a comment-only plain program /* comment */.

[fisker]

like how we parse a comment-only plain program /* comment */

But we start at 0 for <SPACE>/* comment */, should this module block Program start from index after {?

[JLHwung]

But we start at 0 for <SPACE>/* comment */, should this module block Program start from index after {?

Yes, the module block Program should start immediately after { and end immediately before the matching }. I may have to rewind the tokenizer to the position immediately before }.

[fisker]

Is it possible this

module {
  #!/usr/bin/env node
}

became a thing? then shebang is not on first line. 😄

[JLHwung]

@fisker Good question! Per spec:

The InputElementHashbangOrRegExp goal is used at the start of a Script or Module.

Because a module block can only contains optional ModuleBody, the hashbang comment is not allowed in the module block. Currently Babel already disallows it but I will check if we have such a test case.

[fisker]

Maybe too earily to consider about the estree plugin, but Eslint(Espree)'s Program starts from first token.