Skip to content

Latest commit

 

History

History
209 lines (156 loc) · 7.45 KB

README.md

File metadata and controls

209 lines (156 loc) · 7.45 KB
Raw

Meriyah

100% compliant, self-hosted javascript parser with high focus on both performance and stability. Stable and already used in production.

Meriyah NPM Node.js CI License


Meriyah Playground

Features

  • Conforms to the standard ECMAScript® 2024 (ECMA-262 15th Edition) language specification
  • Support some TC39 stage 3 proposals via option "next"
  • Support for additional ECMAScript features for Web Browsers (Annex B)
  • JSX support via option "jsx"
  • Does NOT support TypeScript or Flow syntax
  • Track syntactic node locations with option "ranges" or "loc"
  • Emits an ESTree-compatible abstract syntax tree
  • No backtracking
  • Low memory usage

ESNext Stage 3 features

Supported stage 3 features:

These features need to be enabled with the next option.

Not yet supported stage 3 features:

RegExp support

Meriyah doesn't parse RegExp internal syntax, ESTree spec didn't require internal structure of RegExp. Meriyah does use JavaScript runtime to validate the RegExp literal. That means Meriyah's RegExp support is only as good as JavaScript runtime's RegExp support.

As of Auguest 2024, some latest RegExp features are not supported due to missing implementation in general JavaScript runtime.

In addition, RegExp v flag (unicodeSets) only works on Nodejs v20+ and latest browsers.

Installation

npm install meriyah --save-dev

API

Meriyah generates AST according to ESTree AST format, and can be used to perform syntactic analysis (parsing) of a JavaScript program, and with ES2015 and later a JavaScript program can be either a script or a module.

The parse method exposed by meriyah takes an optional options object which allows you to specify whether to parse in script mode (the default) or in module mode.

// There are also "parseScript" and "parseModule" exported.
import { parse } from 'meriyah';
const result = parse('let some = "code";', { ranges: true });

The available options:

{
  // The flag to allow module code
  module: false;

  // The flag to enable stage 3 support (ESNext)
  next: false;

  // The flag to enable start, end offsets and range: [start, end] to each node
  ranges: false;

  // Enable web compatibility
  webcompat: false;

  // The flag to enable line/column location information to each node
  loc: false;

  // The flag to attach raw property to each literal and identifier node
  raw: false;

  // The flag to allow return in the global scope
  globalReturn: false;

  // The flag to enable implied strict mode
  impliedStrict: false;

  // Allows comment extraction. Accepts either a function or array
  onComment: [];

  // Allows detection of automatic semicolon insertion. Accepts a callback function that will be passed the charater offset where the semicolon was inserted
  onInsertedSemicolon: (pos) => {};

  // Allows token extraction. Accepts either a function or array
  onToken: [];

  // Enable non-standard parenthesized expression node
  preserveParens: false;

  // Enable lexical binding and scope tracking
  lexical: false;

  // Adds a source attribute in every node’s loc object when the locations option is `true`
  source: undefined; // Set to source: 'source-file.js'

  // Enable React JSX parsing
  jsx: false;
}

onComment and onToken

If an array is supplied, comments/tokens will be pushed to the array, the item in the array contains start/end/range information when ranges flag is true, it will also contain loc information when loc flag is true.

If a function callback is supplied, the signature must be

declare function onComment(type: string, value: string, start: number, end: number, loc: SourceLocation): void;

declare function onToken(token: string, start: number, end: number, loc: SourceLocation): void;

Note the start/end/loc information are provided to the function callback regardless of the settings on ranges and loc flags. onComment callback has one extra argument value: string for the body string of the comment.

onInsertedSemicolon

If a function callback is supplied, the signature must be

declare function onInsertedSemicolon(position: number): void;

Example usage

import { parseScript } from './meriyah';

parseScript('({x: [y] = 0} = 1)');

This will return when serialized in json:

{
    type: "Program",
    sourceType: "script",
    body: [
        {
            type: "ExpressionStatement",
            expression: {
                type: "AssignmentExpression",
                left: {
                    type: "ObjectPattern",
                    properties: [
                        {
                            type: "Property",
                            key: {
                                type: "Identifier",
                                name: "x"
                            },
                            value: {
                                type: "AssignmentPattern",
                                left: {
                                    type: "ArrayPattern",
                                    elements: [
                                        {
                                            "type": "Identifier",
                                            "name": "y"
                                        }
                                    ]
                                },
                                right: {
                                    type: "Literal",
                                    value: 0
                                }
                            },
                            kind: "init",
                            computed: false,
                            method: false,
                            shorthand: false
                        }
                    ]
                },
                operator: "=",
                right: {
                    type: "Literal",
                    value: 1
                }
            }
        }
    ]
}