core-js-builder

I highly recommend reading this: So, what's next?

For some cases could be useful to exclude some core-js features or generate a polyfill for target engines. This API helps conditionally include or exclude certain parts of core-js and build for targets. modules, exclude and targets options are specified in the core-js-compat format.

import builder from 'core-js-builder';

const bundle = await builder({
  // entry / module / namespace / an array of them, by default - all `core-js` modules
  modules: ['core-js/actual', /^esnext\.reflect\./],
  // a blacklist of entries / modules / namespaces, by default - empty list
  exclude: [/^es\.math\./, 'es.number.constructor'],
  // optional browserslist or core-js-compat format query
  targets: '> 0.5%, not dead, ie 9-11',
  // shows summary for the bundle, disabled by default
  summary: {
    // in the console, you could specify required parts or set `true` for enable all of them
    console: { size: true, modules: false },
    // in the comment in the target file, similarly to `summary.console`
    comment: { size: false, modules: true },
  },
  // output format, 'bundle' by default, can be 'cjs' or 'esm', and in this case
  // the result will not be bundled and will contain imports of required modules
  format: 'bundle',
  // optional target filename, if it's missed a file will not be created
  filename: PATH_TO_MY_COREJS_BUNDLE,
});

ℹ️ When using TypeScript, make sure to set esModuleInterop to true.

Changelog

3.40.0 - 2025.01.08

• Changes v3.39.0...v3.40.0 (130 commits)
• Added Error.isError stage 3 proposal:
  • Added built-ins:
    • Error.isError
  • We have no bulletproof way to polyfill this method / check if the object is an error, so it's an enough naive implementation that is marked as .sham
• Explicit Resource Management stage 3 proposal:
  • Updated the way async disposing of only sync disposable resources, tc39/proposal-explicit-resource-management/218
• Iterator sequencing stage 2.7 proposal:
  • Reuse IteratorResult objects when possible, tc39/proposal-iterator-sequencing/17, tc39/proposal-iterator-sequencing/18, December 2024 TC39 meeting
• Added a fix of V8 < 12.8 / NodeJS < 22.10 bug with handling infinite length of set-like objects in Set methods
• Optimized DataView.prototype.{ getFloat16, setFloat16 } performance, #1379, thanks @LeviPesin
• Dropped unneeded feature detection of non-standard %TypedArray%.prototype.toSpliced
• Dropped possible re-usage of some non-standard / early stage features (like Math.scale) available on global
• Some other minor improvements
• Compat data improvements:
  • RegExp.escape marked as shipped from Safari 18.2
  • Promise.try marked as shipped from Safari 18.2
  • Math.f16round and DataView.prototype.{ getFloat16, setFloat16 } marked as shipped from Safari 18.2
  • Uint8Array to / from base64 and hex proposal methods marked as shipped from Safari 18.2
  • JSON.parse source text access proposal features marked as shipped from FF135
  • RegExp.escape marked as shipped from FF134
  • Promise.try marked as shipped from FF134
  • Symbol.dispose, Symbol.asyncDispose and Iterator.prototype[@@dispose] marked as shipped from FF135
  • JSON.parse source text access proposal features marked as shipped from Bun 1.1.43
  • Fixed NodeJS version where URL.parse was added - 22.1 instead of 22.0
  • Added Deno 2.1 compat data mapping
  • Added Rhino 1.8.0 compat data with significant number of modern features
  • Added Electron 35 compat data mapping
  • Updated Opera 115+ compat data mapping
  • Added Opera Android 86 and 87 compat data mapping