0
0
mirror of https://github.com/nodejs/node.git synced 2024-12-01 16:10:02 +01:00
nodejs/doc/api/esm.md
Guy Bedford 04028aac3a
doc: update conditional exports recommendations
Co-Authored-By: Geoffrey Booth <GeoffreyBooth@users.noreply.github.com>
PR-URL: https://github.com/nodejs/node/pull/32098
Reviewed-By: Geoffrey Booth <webmaster@geoffreybooth.com>
2020-03-13 10:08:37 +01:00

60 KiB
Raw Blame History

ECMAScript Modules

Stability: 1 - Experimental

Introduction

ECMAScript modules are the official standard format to package JavaScript code for reuse. Modules are defined using a variety of import and export statements.

Node.js fully supports ECMAScript modules as they are currently specified and provides limited interoperability between them and the existing module format, CommonJS.

Node.js contains support for ES Modules based upon the Node.js EP for ES Modules and the ECMAScript-modules implementation.

Expect major changes in the implementation including interoperability support, specifier resolution, and default behavior.

Enabling

Experimental support for ECMAScript modules is enabled by default. Node.js will treat the following as ES modules when passed to node as the initial input, or when referenced by import statements within ES module code:

  • Files ending in .mjs.

  • Files ending in .js when the nearest parent package.json file contains a top-level field "type" with a value of "module".

  • Strings passed in as an argument to --eval or --print, or piped to node via STDIN, with the flag --input-type=module.

Node.js will treat as CommonJS all other forms of input, such as .js files where the nearest parent package.json file contains no top-level "type" field, or string input without the flag --input-type. This behavior is to preserve backward compatibility. However, now that Node.js supports both CommonJS and ES modules, it is best to be explicit whenever possible. Node.js will treat the following as CommonJS when passed to node as the initial input, or when referenced by import statements within ES module code:

  • Files ending in .cjs.

  • Files ending in .js when the nearest parent package.json file contains a top-level field "type" with a value of "commonjs".

  • Strings passed in as an argument to --eval or --print, or piped to node via STDIN, with the flag --input-type=commonjs.

package.json "type" field

Files ending with .js will be loaded as ES modules when the nearest parent package.json file contains a top-level field "type" with a value of "module".

The nearest parent package.json is defined as the first package.json found when searching in the current folder, that folders parent, and so on up until the root of the volume is reached.

// package.json
{
  "type": "module"
}
# In same folder as above package.json
node my-app.js # Runs as ES module

If the nearest parent package.json lacks a "type" field, or contains "type": "commonjs", .js files are treated as CommonJS. If the volume root is reached and no package.json is found, Node.js defers to the default, a package.json with no "type" field.

import statements of .js files are treated as ES modules if the nearest parent package.json contains "type": "module".

// my-app.js, part of the same example as above
import './startup.js'; // Loaded as ES module because of package.json

Package authors should include the "type" field, even in packages where all sources are CommonJS. Being explicit about the type of the package will future-proof the package in case the default type of Node.js ever changes, and it will also make things easier for build tools and loaders to determine how the files in the package should be interpreted.

Regardless of the value of the "type" field, .mjs files are always treated as ES modules and .cjs files are always treated as CommonJS.

Package Scope and File Extensions

A folder containing a package.json file, and all subfolders below that folder down until the next folder containing another package.json, is considered a package scope. The "type" field defines how .js files should be treated within a particular package.json files package scope. Every package in a projects node_modules folder contains its own package.json file, so each projects dependencies have their own package scopes. A package.json lacking a "type" field is treated as if it contained "type": "commonjs".

The package scope applies not only to initial entry points (node my-app.js) but also to files referenced by import statements and import() expressions.

// my-app.js, in an ES module package scope because there is a package.json
// file in the same folder with "type": "module".

import './startup/init.js';
// Loaded as ES module since ./startup contains no package.json file,
// and therefore inherits the ES module package scope from one level up.

import 'commonjs-package';
// Loaded as CommonJS since ./node_modules/commonjs-package/package.json
// lacks a "type" field or contains "type": "commonjs".

import './node_modules/commonjs-package/index.js';
// Loaded as CommonJS since ./node_modules/commonjs-package/package.json
// lacks a "type" field or contains "type": "commonjs".

Files ending with .mjs are always loaded as ES modules regardless of package scope.

Files ending with .cjs are always loaded as CommonJS regardless of package scope.

import './legacy-file.cjs';
// Loaded as CommonJS since .cjs is always loaded as CommonJS.

import 'commonjs-package/src/index.mjs';
// Loaded as ES module since .mjs is always loaded as ES module.

The .mjs and .cjs extensions may be used to mix types within the same package scope:

  • Within a "type": "module" package scope, Node.js can be instructed to interpret a particular file as CommonJS by naming it with a .cjs extension (since both .js and .mjs files are treated as ES modules within a "module" package scope).

  • Within a "type": "commonjs" package scope, Node.js can be instructed to interpret a particular file as an ES module by naming it with an .mjs extension (since both .js and .cjs files are treated as CommonJS within a "commonjs" package scope).

--input-type flag

Strings passed in as an argument to --eval or --print (or -e or -p), or piped to node via STDIN, will be treated as ES modules when the --input-type=module flag is set.

node --input-type=module --eval "import { sep } from 'path'; console.log(sep);"

echo "import { sep } from 'path'; console.log(sep);" | node --input-type=module

For completeness there is also --input-type=commonjs, for explicitly running string input as CommonJS. This is the default behavior if --input-type is unspecified.

Packages

Package Entry Points

In a packages package.json file, two fields can define entry points for a package: "main" and "exports". The "main" field is supported in all versions of Node.js, but its capabilities are limited: it only defines the main entry point of the package.

The "exports" field provides an alternative to "main" where the package main entry point can be defined while also encapsulating the package, preventing any other entry points besides those defined in "exports". If package entry points are defined in both "main" and "exports", the latter takes precedence in versions of Node.js that support "exports". Conditional Exports can also be used within "exports" to define different package entry points per environment, including whether the package is referenced via require or via import.

If both "exports" and "main" are defined, the "exports" field takes precedence over "main".

Both "main" and "exports" entry points are not specific to ES modules or CommonJS; "main" will be overridden by "exports" in a require so it is not a CommonJS fallback.

This is important with regard to require, since require of ES module files throws an error in all versions of Node.js. To create a package that works both in modern Node.js via import and require and also legacy Node.js versions, see the dual CommonJS/ES module packages section.

Main Entry Point Export

To set the main entry point for a package, it is advisable to define both "exports" and "main" in the packages package.json file:

{
  "main": "./main.js",
  "exports": "./main.js"
}

The benefit of doing this is that when using the "exports" field all subpaths of the package will no longer be available to importers under require('pkg/subpath.js'), and instead they will get a new error, ERR_PACKAGE_PATH_NOT_EXPORTED.

This encapsulation of exports provides more reliable guarantees about package interfaces for tools and when handling semver upgrades for a package. It is not a strong encapsulation since a direct require of any absolute subpath of the package such as require('/path/to/node_modules/pkg/subpath.js') will still load subpath.js.

Subpath Exports

When using the "exports" field, custom subpaths can be defined along with the main entry point by treating the main entry point as the "." subpath:

{
  "main": "./main.js",
  "exports": {
    ".": "./main.js",
    "./submodule": "./src/submodule.js"
  }
}

Now only the defined subpath in "exports" can be imported by a consumer:

import submodule from 'es-module-package/submodule';
// Loads ./node_modules/es-module-package/src/submodule.js

While other subpaths will error:

import submodule from 'es-module-package/private-module.js';
// Throws ERR_PACKAGE_PATH_NOT_EXPORTED

Entire folders can also be mapped with package exports:

// ./node_modules/es-module-package/package.json
{
  "exports": {
    "./features/": "./src/features/"
  }
}

With the above, all modules within the ./src/features/ folder are exposed deeply to import and require:

import feature from 'es-module-package/features/x.js';
// Loads ./node_modules/es-module-package/src/features/x.js

When using folder mappings, ensure that you do want to expose every module inside the subfolder. Any modules which are not public should be moved to another folder to retain the encapsulation benefits of exports.

Package Exports Fallbacks

For possible new specifier support in future, array fallbacks are supported for all invalid specifiers:

{
  "exports": {
    "./submodule": ["not:valid", "./submodule.js"]
  }
}

Since "not:valid" is not a valid specifier, "./submodule.js" is used instead as the fallback, as if it were the only target.

Exports Sugar

If the "." export is the only export, the "exports" field provides sugar for this case being the direct "exports" field value.

If the "." export has a fallback array or string value, then the "exports" field can be set to this value directly.

{
  "exports": {
    ".": "./main.js"
  }
}

can be written:

{
  "exports": "./main.js"
}

Conditional Exports

Conditional exports provide a way to map to different paths depending on certain conditions. They are supported for both CommonJS and ES module imports.

For example, a package that wants to provide different ES module exports for require() and import can be written:

// package.json
{
  "main": "./main-require.cjs",
  "exports": {
    "import": "./main-module.js",
    "require": "./main-require.cjs"
  },
  "type": "module"
}

Node.js supports the following conditions:

  • "import" - matched when the package is loaded via import or import(). Can reference either an ES module or CommonJS file, as both import and import() can load either ES module or CommonJS sources.
  • "require" - matched when the package is loaded via require(). As require() only supports CommonJS, the referenced file must be CommonJS.
  • "node" - matched for any Node.js environment. Can be a CommonJS or ES module file. This condition should always come after "import" or "require".
  • "default" - the generic fallback that will always match. Can be a CommonJS or ES module file. This condition should always come last.

Condition matching is applied in object order from first to last within the "exports" object. The general rule is that conditions should be used from most specific to least specific in object order.

Other conditions such as "browser", "electron", "deno", "react-native", etc. are ignored by Node.js but may be used by other runtimes or tools. Further restrictions, definitions or guidance on condition names may be provided in the future.

Using the "import" and "require" conditions can lead to some hazards, which are explained further in the dual CommonJS/ES module packages section.

Conditional exports can also be extended to exports subpaths, for example:

{
  "main": "./main.js",
  "exports": {
    ".": "./main.js",
    "./feature": {
      "browser": "./feature-browser.js",
      "default": "./feature.js"
    }
  }
}

Defines a package where require('pkg/feature') and import 'pkg/feature' could provide different implementations between the browser and Node.js, given third-party tool support for a "browser" condition.

Nested conditions

In addition to direct mappings, Node.js also supports nested condition objects.

For example, to define a package that only has dual mode entry points for use in Node.js but not the browser:

{
  "main": "./main.js",
  "exports": {
    "browser": "./feature-browser.mjs",
    "node": {
      "import": "./feature-node.mjs",
      "require": "./feature-node.cjs"
    }
  }
}

Conditions continue to be matched in order as with flat conditions. If a nested conditional does not have any mapping it will continue checking the remaining conditions of the parent condition. In this way nested conditions behave analogously to nested JavaScript if statements.

Self-referencing a package using its name

Within a package, the values defined in the packages package.json "exports" field can be referenced via the packages name. For example, assuming the package.json is:

// package.json
{
  "name": "a-package",
  "exports": {
    ".": "./main.mjs",
    "./foo": "./foo.js"
  }
}

Then any module in that package can reference an export in the package itself:

// ./a-module.mjs
import { something } from 'a-package'; // Imports "something" from ./main.mjs.

Self-referencing is available only if package.json has exports, and will allow importing only what that exports (in the package.json) allows. So the code below, given the package above, will generate a runtime error:

// ./another-module.mjs

// Imports "another" from ./m.mjs. Fails because
// the "package.json" "exports" field
// does not provide an export named "./m.mjs".
import { another } from 'a-package/m.mjs';

Self-referencing is also available when using require, both in an ES module, and in a CommonJS one. For example, this code will also work:

// ./a-module.js
const { something } = require('a-package/foo'); // Loads from ./foo.js.

Dual CommonJS/ES Module Packages

Prior to the introduction of support for ES modules in Node.js, it was a common pattern for package authors to include both CommonJS and ES module JavaScript sources in their package, with package.json "main" specifying the CommonJS entry point and package.json "module" specifying the ES module entry point. This enabled Node.js to run the CommonJS entry point while build tools such as bundlers used the ES module entry point, since Node.js ignored (and still ignores) the top-level "module" field.

Node.js can now run ES module entry points, and a package can contain both CommonJS and ES module entry points (either via separate specifiers such as 'pkg' and 'pkg/es-module', or both at the same specifier via Conditional Exports). Unlike in the scenario where "module" is only used by bundlers, or ES module files are transpiled into CommonJS on the fly before evaluation by Node.js, the files referenced by the ES module entry point are evaluated as ES modules.

Dual Package Hazard

When an application is using a package that provides both CommonJS and ES module sources, there is a risk of certain bugs if both versions of the package get loaded. This potential comes from the fact that the pkgInstance created by const pkgInstance = require('pkg') is not the same as the pkgInstance created by import pkgInstance from 'pkg' (or an alternative main path like 'pkg/module'). This is the “dual package hazard,” where two versions of the same package can be loaded within the same runtime environment. While it is unlikely that an application or package would intentionally load both versions directly, it is common for an application to load one version while a dependency of the application loads the other version. This hazard can happen because Node.js supports intermixing CommonJS and ES modules, and can lead to unexpected behavior.

If the package main export is a constructor, an instanceof comparison of instances created by the two versions returns false, and if the export is an object, properties added to one (like pkgInstance.foo = 3) are not present on the other. This differs from how import and require statements work in all-CommonJS or all-ES module environments, respectively, and therefore is surprising to users. It also differs from the behavior users are familiar with when using transpilation via tools like Babel or esm.

Writing Dual Packages While Avoiding or Minimizing Hazards

First, the hazard described in the previous section occurs when a package contains both CommonJS and ES module sources and both sources are provided for use in Node.js, either via separate main entry points or exported paths. A package could instead be written where any version of Node.js receives only CommonJS sources, and any separate ES module sources the package may contain could be intended only for other environments such as browsers. Such a package would be usable by any version of Node.js, since import can refer to CommonJS files; but it would not provide any of the advantages of using ES module syntax.

A package could also switch from CommonJS to ES module syntax in a breaking change version bump. This has the obvious disadvantage that the newest version of the package would only be usable in ES module-supporting versions of Node.js.

Every pattern has tradeoffs, but there are two broad approaches that satisfy the following conditions:

  1. The package is usable via both require and import.
  2. The package is usable in both current Node.js and older versions of Node.js that lack support for ES modules.
  3. The package main entry point, e.g. 'pkg' can be used by both require to resolve to a CommonJS file and by import to resolve to an ES module file. (And likewise for exported paths, e.g. 'pkg/feature'.)
  4. The package provides named exports, e.g. import { name } from 'pkg' rather than import pkg from 'pkg'; pkg.name.
  5. The package is potentially usable in other ES module environments such as browsers.
  6. The hazards described in the previous section are avoided or minimized.
Approach #1: Use an ES Module Wrapper

Write the package in CommonJS or transpile ES module sources into CommonJS, and create an ES module wrapper file that defines the named exports. Using Conditional Exports, the ES module wrapper is used for import and the CommonJS entry point for require.

// ./node_modules/pkg/package.json
{
  "type": "module",
  "main": "./index.cjs",
  "exports": {
    "import": "./wrapper.mjs",
    "require": "./index.cjs"
  }
}
// ./node_modules/pkg/index.cjs
exports.name = 'value';
// ./node_modules/pkg/wrapper.mjs
import cjsModule from './index.cjs';
export const name = cjsModule.name;

In this example, the name from import { name } from 'pkg' is the same singleton as the name from const { name } = require('pkg'). Therefore === returns true when comparing the two names and the divergent specifier hazard is avoided.

If the module is not simply a list of named exports, but rather contains a unique function or object export like module.exports = function () { ... }, or if support in the wrapper for the import pkg from 'pkg' pattern is desired, then the wrapper would instead be written to export the default optionally along with any named exports as well:

import cjsModule from './index.cjs';
export const name = cjsModule.name;
export default cjsModule;

This approach is appropriate for any of the following use cases:

  • The package is currently written in CommonJS and the author would prefer not to refactor it into ES module syntax, but wishes to provide named exports for ES module consumers.
  • The package has other packages that depend on it, and the end user might install both this package and those other packages. For example a utilities package is used directly in an application, and a utilities-plus package adds a few more functions to utilities. Because the wrapper exports underlying CommonJS files, it doesnt matter if utilities-plus is written in CommonJS or ES module syntax; it will work either way.
  • The package stores internal state, and the package author would prefer not to refactor the package to isolate its state management. See the next section.

A variant of this approach not requiring conditional exports for consumers could be to add an export, e.g. "./module", to point to an all-ES module-syntax version of the package. This could be used via import 'pkg/module' by users who are certain that the CommonJS version will not be loaded anywhere in the application, such as by dependencies; or if the CommonJS version can be loaded but doesnt affect the ES module version (for example, because the package is stateless):

// ./node_modules/pkg/package.json
{
  "type": "module",
  "main": "./index.cjs",
  "exports": {
    ".": "./index.cjs",
    "./module": "./wrapper.mjs"
  }
}
Approach #2: Isolate State

The most straightforward package.json would be one that defines the separate CommonJS and ES module entry points directly:

// ./node_modules/pkg/package.json
{
  "type": "module",
  "main": "./index.cjs",
  "exports": {
    "import": "./index.mjs",
    "require": "./index.cjs"
  }
}

This can be done if both the CommonJS and ES module versions of the package are equivalent, for example because one is the transpiled output of the other; and the packages management of state is carefully isolated (or the package is stateless).

The reason that state is an issue is because both the CommonJS and ES module versions of the package may get used within an application; for example, the users application code could import the ES module version while a dependency requires the CommonJS version. If that were to occur, two copies of the package would be loaded in memory and therefore two separate states would be present. This would likely cause hard-to-troubleshoot bugs.

Aside from writing a stateless package (if JavaScripts Math were a package, for example, it would be stateless as all of its methods are static), there are some ways to isolate state so that its shared between the potentially loaded CommonJS and ES module instances of the package:

  1. If possible, contain all state within an instantiated object. JavaScripts Date, for example, needs to be instantiated to contain state; if it were a package, it would be used like this:

    import Date from 'date';
    const someDate = new Date();
    // someDate contains state; Date does not
    

    The new keyword isnt required; a packages function can return a new object, or modify a passed-in object, to keep the state external to the package.

  2. Isolate the state in one or more CommonJS files that are shared between the CommonJS and ES module versions of the package. For example, if the CommonJS and ES module entry points are index.cjs and index.mjs, respectively:

    // ./node_modules/pkg/index.cjs
    const state = require('./state.cjs');
    module.exports.state = state;
    
    // ./node_modules/pkg/index.mjs
    import state from './state.cjs';
    export {
      state
    };
    

    Even if pkg is used via both require and import in an application (for example, via import in application code and via require by a dependency) each reference of pkg will contain the same state; and modifying that state from either module system will apply to both.

Any plugins that attach to the packages singleton would need to separately attach to both the CommonJS and ES module singletons.

This approach is appropriate for any of the following use cases:

  • The package is currently written in ES module syntax and the package author wants that version to be used wherever such syntax is supported.
  • The package is stateless or its state can be isolated without too much difficulty.
  • The package is unlikely to have other public packages that depend on it, or if it does, the package is stateless or has state that need not be shared between dependencies or with the overall application.

Even with isolated state, there is still the cost of possible extra code execution between the CommonJS and ES module versions of a package.

As with the previous approach, a variant of this approach not requiring conditional exports for consumers could be to add an export, e.g. "./module", to point to an all-ES module-syntax version of the package:

// ./node_modules/pkg/package.json
{
  "type": "module",
  "main": "./index.cjs",
  "exports": {
    ".": "./index.cjs",
    "./module": "./index.mjs"
  }
}

import Specifiers

Terminology

The specifier of an import statement is the string after the from keyword, e.g. 'path' in import { sep } from 'path'. Specifiers are also used in export from statements, and as the argument to an import() expression.

There are four types of specifiers:

  • Bare specifiers like 'some-package'. They refer to an entry point of a package by the package name.

  • Deep import specifiers like 'some-package/lib/shuffle.mjs'. They refer to a path within a package prefixed by the package name.

  • Relative specifiers like './startup.js' or '../config.mjs'. They refer to a path relative to the location of the importing file.

  • Absolute specifiers like 'file:///opt/nodejs/config.js'. They refer directly and explicitly to a full path.

Bare specifiers, and the bare specifier portion of deep import specifiers, are strings; but everything else in a specifier is a URL.

Only file: and data: URLs are supported. A specifier like 'https://example.com/app.js' may be supported by browsers but it is not supported in Node.js.

Specifiers may not begin with / or //. These are reserved for potential future use. The root of the current volume may be referenced via file:///.

data: Imports

data: URLs are supported for importing with the following MIME types:

  • text/javascript for ES Modules
  • application/json for JSON
  • application/wasm for WASM.

data: URLs only resolve Bare specifiers for builtin modules and Absolute specifiers. Resolving Relative specifiers will not work because data: is not a special scheme. For example, attempting to load ./foo from data:text/javascript,import "./foo"; will fail to resolve since there is no concept of relative resolution for data: URLs. An example of a data: URLs being used is:

import 'data:text/javascript,console.log("hello!");';
import _ from 'data:application/json,"world!"';

import.meta

  • {Object}

The import.meta metaproperty is an Object that contains the following property:

  • url {string} The absolute file: URL of the module.

Differences Between ES Modules and CommonJS

Mandatory file extensions

A file extension must be provided when using the import keyword. Directory indexes (e.g. './startup/index.js') must also be fully specified.

This behavior matches how import behaves in browser environments, assuming a typically configured server.

No NODE_PATH

NODE_PATH is not part of resolving import specifiers. Please use symlinks if this behavior is desired.

No require, exports, module.exports, __filename, __dirname

These CommonJS variables are not available in ES modules.

require can be imported into an ES module using module.createRequire().

Equivalents of __filename and __dirname can be created inside of each file via import.meta.url.

import { fileURLToPath } from 'url';
import { dirname } from 'path';

const __filename = fileURLToPath(import.meta.url);
const __dirname = dirname(__filename);

No require.resolve

Former use cases relying on require.resolve to determine the resolved path of a module can be supported via import.meta.resolve, which is experimental and supported via the --experimental-import-meta-resolve flag:

(async () => {
  const dependencyAsset = await import.meta.resolve('component-lib/asset.css');
})();

import.meta.resolve also accepts a second argument which is the parent module from which to resolve from:

(async () => {
  // Equivalent to import.meta.resolve('./dep')
  await import.meta.resolve('./dep', import.meta.url);
})();

This function is asynchronous since the ES module resolver in Node.js is asynchronous. With the introduction of Top-Level Await, these use cases will be easier as they won't require an async function wrapper.

No require.extensions

require.extensions is not used by import. The expectation is that loader hooks can provide this workflow in the future.

No require.cache

require.cache is not used by import. It has a separate cache.

URL-based paths

ES modules are resolved and cached based upon URL semantics. This means that files containing special characters such as # and ? need to be escaped.

Modules will be loaded multiple times if the import specifier used to resolve them have a different query or fragment.

import './foo.mjs?query=1'; // loads ./foo.mjs with query of "?query=1"
import './foo.mjs?query=2'; // loads ./foo.mjs with query of "?query=2"

For now, only modules using the file: protocol can be loaded.

Interoperability with CommonJS

require

require always treats the files it references as CommonJS. This applies whether require is used the traditional way within a CommonJS environment, or in an ES module environment using module.createRequire().

To include an ES module into CommonJS, use import().

import statements

An import statement can reference an ES module or a CommonJS module. Other file types such as JSON or Native modules are not supported. For those, use module.createRequire().

import statements are permitted only in ES modules. For similar functionality in CommonJS, see import().

The specifier of an import statement (the string after the from keyword) can either be an URL-style relative path like './file.mjs' or a package name like 'fs'.

Like in CommonJS, files within packages can be accessed by appending a path to the package name; unless the packages package.json contains an "exports" field, in which case files within packages need to be accessed via the path defined in "exports".

import { sin, cos } from 'geometry/trigonometry-functions.mjs';

Only the “default export” is supported for CommonJS files or packages:

import packageMain from 'commonjs-package'; // Works

import { method } from 'commonjs-package'; // Errors

import() expressions

Dynamic import() is supported in both CommonJS and ES modules. It can be used to include ES module files from CommonJS code.

(async () => {
  await import('./my-app.mjs');
})();

CommonJS, JSON, and Native Modules

CommonJS, JSON, and Native modules can be used with module.createRequire().

// cjs.cjs
module.exports = 'cjs';

// esm.mjs
import { createRequire } from 'module';

const require = createRequire(import.meta.url);

const cjs = require('./cjs.cjs');
cjs === 'cjs'; // true

Builtin modules

Builtin modules will provide named exports of their public API. A default export is also provided which is the value of the CommonJS exports. The default export can be used for, among other things, modifying the named exports. Named exports of builtin modules are updated only by calling module.syncBuiltinESMExports().

import EventEmitter from 'events';
const e = new EventEmitter();
import { readFile } from 'fs';
readFile('./foo.txt', (err, source) => {
  if (err) {
    console.error(err);
  } else {
    console.log(source);
  }
});
import fs, { readFileSync } from 'fs';
import { syncBuiltinESMExports } from 'module';

fs.readFileSync = () => Buffer.from('Hello, ESM');
syncBuiltinESMExports();

fs.readFileSync === readFileSync;

Experimental JSON Modules

Currently importing JSON modules are only supported in the commonjs mode and are loaded using the CJS loader. WHATWG JSON modules specification are still being standardized, and are experimentally supported by including the additional flag --experimental-json-modules when running Node.js.

When the --experimental-json-modules flag is included both the commonjs and module mode will use the new experimental JSON loader. The imported JSON only exposes a default, there is no support for named exports. A cache entry is created in the CommonJS cache, to avoid duplication. The same object will be returned in CommonJS if the JSON module has already been imported from the same path.

Assuming an index.mjs with

import packageConfig from './package.json';

The --experimental-json-modules flag is needed for the module to work.

node index.mjs # fails
node --experimental-json-modules index.mjs # works

Experimental Wasm Modules

Importing Web Assembly modules is supported under the --experimental-wasm-modules flag, allowing any .wasm files to be imported as normal modules while also supporting their module imports.

This integration is in line with the ES Module Integration Proposal for Web Assembly.

For example, an index.mjs containing:

import * as M from './module.wasm';
console.log(M);

executed under:

node --experimental-wasm-modules index.mjs

would provide the exports interface for the instantiation of module.wasm.

Experimental Loaders

Note: This API is currently being redesigned and will still change.

To customize the default module resolution, loader hooks can optionally be provided via a --experimental-loader ./loader-name.mjs argument to Node.js.

When hooks are used they only apply to ES module loading and not to any CommonJS modules loaded.

Hooks

resolve hook

Note: The loaders API is being redesigned. This hook may disappear or its signature may change. Do not rely on the API described below.

The resolve hook returns the resolved file URL for a given module specifier and parent URL. The module specifier is the string in an import statement or import() expression, and the parent URL is the URL of the module that imported this one, or undefined if this is the main entry point for the application.

/**
 * @param {string} specifier
 * @param {object} context
 * @param {string} context.parentURL
 * @param {function} defaultResolve
 * @returns {object} response
 * @returns {string} response.url
 */
export async function resolve(specifier, context, defaultResolve) {
  const { parentURL = null } = context;
  if (someCondition) {
    // For some or all specifiers, do some custom logic for resolving.
    // Always return an object of the form {url: <string>}
    return {
      url: (parentURL) ?
        new URL(specifier, parentURL).href : new URL(specifier).href
    };
  }
  // Defer to Node.js for all other specifiers.
  return defaultResolve(specifier, context, defaultResolve);
}

getFormat hook

Note: The loaders API is being redesigned. This hook may disappear or its signature may change. Do not rely on the API described below.

The getFormat hook provides a way to define a custom method of determining how a URL should be interpreted. This can be one of the following:

format Description
'builtin' Load a Node.js builtin module
'commonjs' Load a Node.js CommonJS module
'dynamic' Use a dynamic instantiate hook
'json' Load a JSON file
'module' Load a standard JavaScript module (ES module)
'wasm' Load a WebAssembly module
/**
 * @param {string} url
 * @param {object} context (currently empty)
 * @param {function} defaultGetFormat
 * @returns {object} response
 * @returns {string} response.format
 */
export async function getFormat(url, context, defaultGetFormat) {
  if (someCondition) {
    // For some or all URLs, do some custom logic for determining format.
    // Always return an object of the form {format: <string>}, where the
    // format is one of the strings in the table above.
    return {
      format: 'module'
    };
  }
  // Defer to Node.js for all other URLs.
  return defaultGetFormat(url, context, defaultGetFormat);
}

getSource hook

Note: The loaders API is being redesigned. This hook may disappear or its signature may change. Do not rely on the API described below.

The getSource hook provides a way to define a custom method for retrieving the source code of an ES module specifier. This would allow a loader to potentially avoid reading files from disk.

/**
 * @param {string} url
 * @param {object} context
 * @param {string} context.format
 * @param {function} defaultGetSource
 * @returns {object} response
 * @returns {string|buffer} response.source
 */
export async function getSource(url, context, defaultGetSource) {
  const { format } = context;
  if (someCondition) {
    // For some or all URLs, do some custom logic for retrieving the source.
    // Always return an object of the form {source: <string|buffer>}.
    return {
      source: '...'
    };
  }
  // Defer to Node.js for all other URLs.
  return defaultGetSource(url, context, defaultGetSource);
}

transformSource hook

Note: The loaders API is being redesigned. This hook may disappear or its signature may change. Do not rely on the API described below.

The transformSource hook provides a way to modify the source code of a loaded ES module file after the source string has been loaded but before Node.js has done anything with it.

If this hook is used to convert unknown-to-Node.js file types into executable JavaScript, a resolve hook is also necessary in order to register any unknown-to-Node.js file extensions. See the transpiler loader example below.

/**
 * @param {string|buffer} source
 * @param {object} context
 * @param {string} context.url
 * @param {string} context.format
 * @param {function} defaultTransformSource
 * @returns {object} response
 * @returns {string|buffer} response.source
 */
export async function transformSource(source,
                                      context,
                                      defaultTransformSource) {
  const { url, format } = context;
  if (someCondition) {
    // For some or all URLs, do some custom logic for modifying the source.
    // Always return an object of the form {source: <string|buffer>}.
    return {
      source: '...'
    };
  }
  // Defer to Node.js for all other sources.
  return defaultTransformSource(
    source, context, defaultTransformSource);
}

dynamicInstantiate hook

Note: The loaders API is being redesigned. This hook may disappear or its signature may change. Do not rely on the API described below.

To create a custom dynamic module that doesn't correspond to one of the existing format interpretations, the dynamicInstantiate hook can be used. This hook is called only for modules that return format: 'dynamic' from the getFormat hook.

/**
 * @param {string} url
 * @returns {object} response
 * @returns {array} response.exports
 * @returns {function} response.execute
 */
export async function dynamicInstantiate(url) {
  return {
    exports: ['customExportName'],
    execute: (exports) => {
      // Get and set functions provided for pre-allocated export names
      exports.customExportName.set('value');
    }
  };
}

With the list of module exports provided upfront, the execute function will then be called at the exact point of module evaluation order for that module in the import tree.

Examples

The various loader hooks can be used together to accomplish wide-ranging customizations of Node.js code loading and evaluation behaviors.

HTTPS loader

In current Node.js, specifiers starting with https:// are unsupported. The loader below registers hooks to enable rudimentary support for such specifiers. While this may seem like a significant improvement to Node.js core functionality, there are substantial downsides to actually using this loader: performance is much slower than loading files from disk, there is no caching, and there is no security.

// https-loader.mjs
import { get } from 'https';

export function resolve(specifier, context, defaultResolve) {
  const { parentURL = null } = context;

  // Normally Node.js would error on specifiers starting with 'https://', so
  // this hook intercepts them and converts them into absolute URLs to be
  // passed along to the later hooks below.
  if (specifier.startsWith('https://')) {
    return {
      url: specifier
    };
  } else if (parentURL && parentURL.startsWith('https://')) {
    return {
      url: new URL(specifier, parentURL).href
    };
  }

  // Let Node.js handle all other specifiers.
  return defaultResolve(specifier, context, defaultResolve);
}

export function getFormat(url, context, defaultGetFormat) {
  // This loader assumes all network-provided JavaScript is ES module code.
  if (url.startsWith('https://')) {
    return {
      format: 'module'
    };
  }

  // Let Node.js handle all other URLs.
  return defaultGetFormat(url, context, defaultGetFormat);
}

export function getSource(url, context, defaultGetSource) {
  // For JavaScript to be loaded over the network, we need to fetch and
  // return it.
  if (url.startsWith('https://')) {
    return new Promise((resolve, reject) => {
      get(url, (res) => {
        let data = '';
        res.on('data', (chunk) => data += chunk);
        res.on('end', () => resolve({ source: data }));
      }).on('error', (err) => reject(err));
    });
  }

  // Let Node.js handle all other URLs.
  return defaultGetSource(url, context, defaultGetSource);
}
// main.mjs
import { VERSION } from 'https://coffeescript.org/browser-compiler-modern/coffeescript.js';

console.log(VERSION);

With this loader, running:

node --experimental-loader ./https-loader.mjs ./main.js

Will print the current version of CoffeeScript per the module at the URL in main.mjs.

Transpiler loader

Sources that are in formats Node.js doesnt understand can be converted into JavaScript using the transformSource hook. Before that hook gets called, however, other hooks need to tell Node.js not to throw an error on unknown file types; and to tell Node.js how to load this new file type.

This is obviously less performant than transpiling source files before running Node.js; a transpiler loader should only be used for development and testing purposes.

// coffeescript-loader.mjs
import { URL, pathToFileURL } from 'url';
import CoffeeScript from 'coffeescript';

const baseURL = pathToFileURL(`${process.cwd()}/`).href;

// CoffeeScript files end in .coffee, .litcoffee or .coffee.md.
const extensionsRegex = /\.coffee$|\.litcoffee$|\.coffee\.md$/;

export function resolve(specifier, context, defaultResolve) {
  const { parentURL = baseURL } = context;

  // Node.js normally errors on unknown file extensions, so return a URL for
  // specifiers ending in the CoffeeScript file extensions.
  if (extensionsRegex.test(specifier)) {
    return {
      url: new URL(specifier, parentURL).href
    };
  }

  // Let Node.js handle all other specifiers.
  return defaultResolve(specifier, context, defaultResolve);
}

export function getFormat(url, context, defaultGetFormat) {
  // Now that we patched resolve to let CoffeeScript URLs through, we need to
  // tell Node.js what format such URLs should be interpreted as. For the
  // purposes of this loader, all CoffeeScript URLs are ES modules.
  if (extensionsRegex.test(url)) {
    return {
      format: 'module'
    };
  }

  // Let Node.js handle all other URLs.
  return defaultGetFormat(url, context, defaultGetFormat);
}

export function transformSource(source, context, defaultTransformSource) {
  const { url, format } = context;

  if (extensionsRegex.test(url)) {
    return {
      source: CoffeeScript.compile(source, { bare: true })
    };
  }

  // Let Node.js handle all other sources.
  return defaultTransformSource(source, context, defaultTransformSource);
}
# main.coffee
import { scream } from './scream.coffee'
console.log scream 'hello, world'

import { version } from 'process'
console.log "Brought to you by Node.js version #{version}"
# scream.coffee
export scream = (str) -> str.toUpperCase()

With this loader, running:

node --experimental-loader ./coffeescript-loader.mjs main.coffee

Will cause main.coffee to be turned into JavaScript after its source code is loaded from disk but before Node.js executes it; and so on for any .coffee, .litcoffee or .coffee.md files referenced via import statements of any loaded file.

Resolution Algorithm

Features

The resolver has the following properties:

  • FileURL-based resolution as is used by ES modules
  • Support for builtin module loading
  • Relative and absolute URL resolution
  • No default extensions
  • No folder mains
  • Bare specifier package resolution lookup through node_modules

Resolver Algorithm

The algorithm to load an ES module specifier is given through the ESM_RESOLVE method below. It returns the resolved URL for a module specifier relative to a parentURL.

The algorithm to determine the module format of a resolved URL is provided by ESM_FORMAT, which returns the unique module format for any file. The "module" format is returned for an ECMAScript Module, while the "commonjs" format is used to indicate loading through the legacy CommonJS loader. Additional formats such as "addon" can be extended in future updates.

In the following algorithms, all subroutine errors are propagated as errors of these top-level routines unless stated otherwise.

defaultEnv is the conditional environment name priority array, ["node", "import"].

The resolver can throw the following errors:

  • Invalid Module Specifier: Module specifier is an invalid URL, package name or package subpath specifier.
  • Invalid Package Configuration: package.json configuration is invalid or contains an invalid configuration.
  • Invalid Package Target: Package exports define a target module within the package that is an invalid type or string target.
  • Package Path Not Exported: Package exports do not define or permit a target subpath in the package for the given module.
  • Module Not Found: The package or module requested does not exist.
Resolver algorithm specification

ESM_RESOLVE(specifier, parentURL)

  1. Let resolvedURL be undefined.
  2. If specifier is a valid URL, then
    1. Set resolvedURL to the result of parsing and reserializing specifier as a URL.
  3. Otherwise, if specifier starts with "/", then
    1. Throw an Invalid Module Specifier error.
  4. Otherwise, if specifier starts with "./" or "../", then
    1. Set resolvedURL to the URL resolution of specifier relative to parentURL.
  5. Otherwise,
    1. Note: specifier is now a bare specifier.
    2. Set resolvedURL the result of PACKAGE_RESOLVE(specifier, parentURL).
  6. If resolvedURL contains any percent encodings of "/" or "\" ("%2f" and "%5C" respectively), then
    1. Throw an Invalid Module Specifier error.
  7. If resolvedURL does not end with a trailing "/" and the file at resolvedURL does not exist, then
    1. Throw a Module Not Found error.
  8. Set resolvedURL to the real path of resolvedURL.
  9. Let format be the result of ESM_FORMAT(resolvedURL).
  10. Load resolvedURL as module format, format.
  11. Return resolvedURL.

PACKAGE_RESOLVE(packageSpecifier, parentURL)

  1. Let packageName be undefined.
  2. Let packageSubpath be undefined.
  3. If packageSpecifier is an empty string, then
    1. Throw an Invalid Module Specifier error.
  4. Otherwise,
    1. If packageSpecifier does not contain a "/" separator, then
      1. Throw an Invalid Module Specifier error.
    2. Set packageName to the substring of packageSpecifier until the second "/" separator or the end of the string.
  5. If packageName starts with "." or contains "\" or "%", then
    1. Throw an Invalid Module Specifier error.
  6. Let packageSubpath be undefined.
  7. If the length of packageSpecifier is greater than the length of packageName, then
    1. Set packageSubpath to "." concatenated with the substring of packageSpecifier from the position at the length of packageName.
  8. If packageSubpath contains any "." or ".." segments or percent encoded strings for "/" or "\", then
    1. Throw an Invalid Module Specifier error.
  9. Set selfUrl to the result of SELF_REFERENCE_RESOLVE(packageName, packageSubpath, parentURL).
  10. If selfUrl isn't empty, return selfUrl.
  11. If packageSubpath is undefined and packageName is a Node.js builtin module, then
    1. Return the string "nodejs:" concatenated with packageSpecifier.
  12. While parentURL is not the file system root,
    1. Let packageURL be the URL resolution of "node_modules/" concatenated with packageSpecifier, relative to parentURL.
    2. Set parentURL to the parent folder URL of parentURL.
    3. If the folder at packageURL does not exist, then
      1. Set parentURL to the parent URL path of parentURL.
      2. Continue the next loop iteration.
    4. Let pjson be the result of READ_PACKAGE_JSON(packageURL).
    5. If packageSubpath is equal to "./", then
      1. Return packageURL + "/".
    6. If packageSubpath is undefined_, then
      1. Return the result of PACKAGE_MAIN_RESOLVE(packageURL, pjson).
    7. Otherwise,
      1. If pjson is not null and pjson has an "exports" key, then
        1. Let exports be pjson.exports.
        2. If exports is not null or undefined, then
          1. Return PACKAGE_EXPORTS_RESOLVE(packageURL, packageSubpath, pjson.exports).
      2. Return the URL resolution of packageSubpath in packageURL.
  13. Throw a Module Not Found error.

SELF_REFERENCE_RESOLVE(packageName, packageSubpath, parentURL)

  1. Let packageURL be the result of READ_PACKAGE_SCOPE(parentURL).
  2. If packageURL is null, then
    1. Return undefined.
  3. Let pjson be the result of READ_PACKAGE_JSON(packageURL).
  4. If pjson does not include an "exports" property, then
    1. Return undefined.
  5. If pjson.name is equal to packageName, then
    1. If packageSubpath is equal to "./", then
      1. Return packageURL + "/".
    2. If packageSubpath is undefined, then
      1. Return the result of PACKAGE_MAIN_RESOLVE(packageURL, pjson).
    3. Otherwise,
      1. If pjson is not null and pjson has an "exports" key, then
        1. Let exports be pjson.exports.
        2. If exports is not null or undefined, then
          1. Return PACKAGE_EXPORTS_RESOLVE(packageURL, subpath, pjson.exports).
      2. Return the URL resolution of subpath in packageURL.
  6. Otherwise, return undefined.

PACKAGE_MAIN_RESOLVE(packageURL, pjson)

  1. If pjson is null, then
    1. Throw a Module Not Found error.
  2. If pjson.exports is not null or undefined, then
    1. If exports is an Object with both a key starting with "." and a key not starting with ".", throw an Invalid Package Configuration error.
    2. If pjson.exports is a String or Array, or an Object containing no keys starting with ".", then
      1. Return PACKAGE_EXPORTS_TARGET_RESOLVE(packageURL, pjson.exports, "").
    3. If pjson.exports is an Object containing a "." property, then
      1. Let mainExport be the "." property in pjson.exports.
      2. Return PACKAGE_EXPORTS_TARGET_RESOLVE(packageURL, mainExport, "").
    4. Throw a Package Path Not Exported error.
  3. If pjson.main is a String, then
    1. Let resolvedMain be the URL resolution of packageURL, "/", and pjson.main.
    2. If the file at resolvedMain exists, then
      1. Return resolvedMain.
  4. If pjson.type is equal to "module", then
    1. Throw a Module Not Found error.
  5. Let legacyMainURL be the result applying the legacy LOAD_AS_DIRECTORY CommonJS resolver to packageURL, throwing a Module Not Found error for no resolution.
  6. Return legacyMainURL.

PACKAGE_EXPORTS_RESOLVE(packageURL, packagePath, exports)

  1. If exports is an Object with both a key starting with "." and a key not starting with ".", throw an Invalid Package Configuration error.
  2. If exports is an Object and all keys of exports start with ".", then
    1. Set packagePath to "./" concatenated with packagePath.
    2. If packagePath is a key of exports, then
      1. Let target be the value of exports[packagePath].
      2. Return PACKAGE_EXPORTS_TARGET_RESOLVE(packageURL, target, "", defaultEnv).
    3. Let directoryKeys be the list of keys of exports ending in "/", sorted by length descending.
    4. For each key directory in directoryKeys, do
      1. If packagePath starts with directory, then
        1. Let target be the value of exports[directory].
        2. Let subpath be the substring of target starting at the index of the length of directory.
        3. Return PACKAGE_EXPORTS_TARGET_RESOLVE(packageURL, target, subpath, defaultEnv).
  3. Throw a Package Path Not Exported error.

PACKAGE_EXPORTS_TARGET_RESOLVE(packageURL, target, subpath, env)

  1. If target is a String, then
    1. If target does not start with "./" or contains any "node_modules" segments including "node_modules" percent-encoding, throw an Invalid Package Target error.
    2. Let resolvedTarget be the URL resolution of the concatenation of packageURL and target.
    3. If resolvedTarget is not contained in packageURL, throw an Invalid Package Target error.
    4. If subpath has non-zero length and target does not end with "/", throw an Invalid Module Specifier error.
    5. Let resolved be the URL resolution of the concatenation of subpath and resolvedTarget.
    6. If resolved is not contained in resolvedTarget, throw an Invalid Module Specifier error.
    7. Return resolved.
  2. Otherwise, if target is a non-null Object, then
    1. If exports contains any index property keys, as defined in ECMA-262 6.1.7 Array Index, throw an Invalid Package Configuration error.
    2. For each property p of target, in object insertion order as,
      1. If env contains an entry for p, then
        1. Let targetValue be the value of the p property in target.
        2. Return the result of PACKAGE_EXPORTS_TARGET_RESOLVE( packageURL, targetValue, subpath, env), continuing the loop on any Package Path Not Exported error.
    3. Throw a Package Path Not Exported error.
  3. Otherwise, if target is an Array, then
    1. If _target.length is zero, throw an Invalid Package Target error.
    2. For each item targetValue in target, do
      1. If targetValue is an Array, continue the loop.
      2. Return the result of PACKAGE_EXPORTS_TARGET_RESOLVE(packageURL, targetValue, subpath, env), continuing the loop on any Package Path Not Exported or Invalid Package Target error.
    3. Throw the last fallback resolution error.
  4. Otherwise throw an Invalid Package Target error.

ESM_FORMAT(url)

  1. Assert: url corresponds to an existing file.
  2. Let pjson be the result of READ_PACKAGE_SCOPE(url).
  3. If url ends in ".mjs", then
    1. Return "module".
  4. If url ends in ".cjs", then
    1. Return "commonjs".
  5. If pjson?.type exists and is "module", then
    1. If url ends in ".js", then
      1. Return "module".
    2. Throw an Unsupported File Extension error.
  6. Otherwise,
    1. Throw an Unsupported File Extension error.

READ_PACKAGE_SCOPE(url)

  1. Let scopeURL be url.
  2. While scopeURL is not the file system root,
    1. If scopeURL ends in a "node_modules" path segment, return null.
    2. Let pjson be the result of READ_PACKAGE_JSON(scopeURL).
    3. If pjson is not null, then
      1. Return pjson.
    4. Set scopeURL to the parent URL of scopeURL.
  3. Return null.

READ_PACKAGE_JSON(packageURL)

  1. Let pjsonURL be the resolution of "package.json" within packageURL.
  2. If the file at pjsonURL does not exist, then
    1. Return null.
  3. If the file at packageURL does not parse as valid JSON, then
    1. Throw an Invalid Package Configuration error.
  4. Return the parsed JSON source of the file at pjsonURL.

Customizing ESM specifier resolution algorithm

The current specifier resolution does not support all default behavior of the CommonJS loader. One of the behavior differences is automatic resolution of file extensions and the ability to import directories that have an index file.

The --experimental-specifier-resolution=[mode] flag can be used to customize the extension resolution algorithm. The default mode is explicit, which requires the full path to a module be provided to the loader. To enable the automatic extension resolution and importing from directories that include an index file use the node mode.

$ node index.mjs
success!
$ node index # Failure!
Error: Cannot find module
$ node --experimental-specifier-resolution=node index
success!