5.6 KiB

Executable File

Raw Blame History

ES Module Lexer

A JS module syntax lexer used in es-module-shims.

Outputs the list of exports and locations of import specifiers, including dynamic import and import meta handling.

A very small single JS file (4KiB gzipped) that includes inlined Web Assembly for very fast source analysis of ECMAScript module syntax only.

For an example of the performance, Angular 1 (720KiB) is fully parsed in 5ms, in comparison to the fastest JS parser, Acorn which takes over 100ms.

Comprehensively handles the JS language grammar while remaining small and fast. - ~10ms per MB of JS cold and ~5ms per MB of JS warm, see benchmarks for more info.

Usage

npm install es-module-lexer

For use in CommonJS:

const { init, parse } = require('es-module-lexer');

(async () => {
  // either await init, or call parse asynchronously
  // this is necessary for the Web Assembly boot
  await init;

  const [imports, exports] = parse('export var p = 5');
  exports[0] === 'p';
})();

An ES module version is also available from dist/lexer.js:
Note: This version will be automatically used in rollup/es-dev-server/node (if an es-module project)

import { init, parse } from 'es-module-lexer/dist/lexer.js';

(async () => {
  await init;

  const source = `
    import { a } from 'asdf';
    export var p = 5;
    export function q () {

    };

    // Comments provided to demonstrate edge cases
    import /*comment!*/ ('asdf');
    import /*comment!*/.meta.asdf;
  `;

  const [imports, exports] = parse(source, 'optional-sourcename');

  // Returns "asdf"
  imports[0].n
  source.substring(imports[0].s, imports[0].e);
  // "s" is shorthand for "start"
  // "e" is shorthand for "end"

  // Returns "import { a } from 'asdf';"
  source.substring(imports[0].ss, imports[0].se);
  // "ss" is shorthand for "statement start"
  // "se" is shorthand for "statement end"

  // Returns "p,q"
  exports.toString();

  // Dynamic imports are indicated by imports[1].d > -1
  // In this case the "d" index is the start of the dynamic import
  // Returns true
  imports[1].d > -1;

  // Returns "asdf"
  imports[1].n
  source.substring(imports[1].s, imports[1].e);
  // Returns "import /*comment!*/ ("
  source.substring(imports[1].d, imports[1].s);
  // Returns "import /*comment!*/ ('asdf')"
  source.substring(imports[1].d, imports[1].e + 1);
  // imports[1].ss and imports[1].se is not meaningful 
  // because dynamic import is not a statement

  // import.meta is indicated by imports[2].d === -2
  // Returns true
  imports[2].d === -2;
  // Returns "import /*comment!*/.meta"
  source.substring(imports[2].s, imports[2].e);
})();

Escape Sequences

To handle escape sequences in specifier strings, the .n field of imported specifiers will be provided where possible.

For dynamic import expressions, this field will be empty if not a valid JS string.

Facade Detection

Facade modules that only use import / export syntax can be detected via the third return value:

const [,, facade] = parse(`
  export * from 'external';
  import * as ns from 'external2';
  export { a as b } from 'external3';
  export { ns };
`);
facade === true;

Environment Support

Node.js 10+, and all browsers with Web Assembly support.

Grammar Support

Token state parses all line comments, block comments, strings, template strings, blocks, parens and punctuators.
Division operator / regex token ambiguity is handled via backtracking checks against punctuator prefixes, including closing brace or paren backtracking.
Always correctly parses valid JS source, but may parse invalid JS source without errors.

Limitations

The lexing approach is designed to deal with the full language grammar including RegEx / division operator ambiguity through backtracking and paren / brace tracking.

The only limitation to the reduced parser is that the "exports" list may not correctly gather all export identifiers in the following edge cases:

// Only "a" is detected as an export, "q" isn't
export var a = 'asdf', q = z;

// "b" is not detected as an export
export var { a: b } = asdf;

The above cases are handled gracefully in that the lexer will keep going fine, it will just not properly detect the export names above.

Benchmarks

Benchmarks can be run with npm run bench.

Current results:

Module load time
> 7ms
Cold Run, All Samples
test/samples/*.js (3057 KiB)
> 33ms

Warm Runs (average of 25 runs)
test/samples/angular.js (719 KiB)
> 4.08ms
test/samples/angular.min.js (188 KiB)
> 2.08ms
test/samples/d3.js (491 KiB)
> 4.72ms
test/samples/d3.min.js (274 KiB)
> 3ms
test/samples/magic-string.js (34 KiB)
> 0.04ms
test/samples/magic-string.min.js (20 KiB)
> 0ms
test/samples/rollup.js (902 KiB)
> 8.16ms
test/samples/rollup.min.js (429 KiB)
> 4.28ms

Warm Runs, All Samples (average of 25 runs)
test/samples/*.js (3057 KiB)
> 25.68ms

Building

To build download the WASI SDK from https://github.com/WebAssembly/wasi-sdk/releases.

The Makefile assumes the existence of "wasi-sdk-11.0" and "wabt" (optional) as sibling folders to this project.

The build through the Makefile is then run via make lib/lexer.wasm, which can also be triggered via npm run build-wasm to create dist/lexer.js.

On Windows it may be preferable to use the Linux subsystem.

After the Web Assembly build, the CJS build can be triggered via npm run build.

License

MIT

5.6 KiB Executable File Raw Blame History