When using CommonJS modules in Node, you can detect whether a script is being run from the command line using require.main === module.
What is an equivalent way to detect whether a script is being run from the command line when using ES Modules in Node (with the --experimental-modules flag)?
Use
if (import.meta.url === `file://${process.argv[1]}`) {
// module was not imported but called directly
}
See the MDN docs on import.meta for details.
Update Sep 27, 2021
Perhaps more robust, but involving an extra import (via Rich Harris)
import {pathToFileURL} from 'url'
if (import.meta.url === pathToFileURL(process.argv[1]).href) {
// module was not imported but called directly
}
There is none - yet (it's still experimental!). Although the prevailing opinion is that such a check is a bad practice anyway and you should just provide separate scripts for the library and the executable, there is an idea to provide a boolean import.meta.main property for this purpose.
The other answers get close, but will miss the mark for a pretty typical usecase - cli scripts exposed by the bin property in package.json files.
These scripts will be symlinked in the node_modules/.bin folder. These can be invoked through npx or as scripts defined in the scripts-object in package.json. process.argv[1] will in that case be the symlink and not the actual file referenced by import.meta.url
Furthermore, we need to convert the file path to an actual file://-url otherwise it will not work correctly on different platforms.
import { realpathSync } from "fs";
import { pathToFileURL } from "url";
function wasCalledAsScript() {
// We use realpathSync to resolve symlinks, as cli scripts will often
// be executed from symlinks in the `node_modules/.bin`-folder
const realPath = realpathSync(process.argv[1]);
// Convert the file-path to a file-url before comparing it
const realPathAsUrl = pathToFileURL(realPath).href;
return import.meta.url === realPathAsUrl;
}
if (wasCalledAsScript()) {
// module was executed and not imported by another file.
}
I would have posted this as a comment on the accepted answer, but apparently I'm not allowed to comment with a fresh account.
The module global variable will be defined in CommonJS, but won’t exist at
all in an ES module. Yes, there is an inconsistency there, that ES modules are
the things that don’t have module variables.
You can check for an undefined variable by seeing if typeof v is the string
(not value!) 'undefined'.
That turns into:
const inCommonJs = typeof module !== 'undefined';
console.log(`inCommonJs = ${inCommonJs}`);
If we put that exact code into both .cjs and .mjs files, we get the correct answers:
$ node foo.mjs
inCommonJs = false
$ cp foo.mjs foo.cjs
$ node foo.cjs
inCommonJs = true
I like import.meta.url === `file://${process.argv[1]}` , but it does not work in Windows inside bash shell. This is the alternative that is only checking the basename:
const runningAsScript = import.meta.url.endsWith(path.basename(process.argv[1]));
It looks like there is a documented way to do this now:
if (require.main === module) {
console.log('executed directly');
. . .
}
Related
Modules generally start something like this
(function(root, factory)
{
/* globals define */
if (typeof define === 'function' && define.amd)
{
// AMD. Register as an anonymous module.
define([], factory);
}
else if (typeof module === 'object' && typeof exports !== 'undefined')
{
// Node. Does not work with strict CommonJS, but
// only CommonJS-like environments that support module.exports,
// like Node.
module.exports = factory();
}
else
{
// Browser globals (root is window)
root.Papa = factory();
}
I'm trying to implement require to handle node style CommonJS-like modules.
Find the package folder, parse package.json to learn the entrypoint and dependencies, recursively descent with a shared cache to load dependencies... that stuff works.
But having loaded the script for a module, I'm having trouble executing it in such a way as to have it populate module.exports
This will all end up running on Jint (a JS interpreter) so node isn't supplying the usual furniture, I have to build it all myself. There's no step-debug with Jint so I'm using node from VS Code and faking Jint like this.
import * as fs from "fs";
var code = fs.readFileSync("node_modules/papaparse/papaparse.js").toString();
let x = 3;
console.log(eval("x*x"))
let result = eval(`
let module = { exports: "dunno" };
${code}
console.log(module.exports);
`);
This is in a file test.js and package.json nominates this file as main and specifies a type of module. It launches, reads the module code, creates module and runs the code, which promptly complains that it Cannot set properties of undefined (setting 'Papa').
Looking at the snippet above, that tells us it's executing the last else clause, which means it's not seeing module. I thought it might be some sort of scope thing for eval which is where this came from
let x = 3;
console.log(eval("x*x"))
but that duly writes 9 to the console so the question is why module isn't in scope.
This is one of those JavaScript closure weirdnesses; can anyone guide me on how to provide the module object so that the second clause will take effect populating module.exports with the result of factory()?
I know I said it's running in the absence of Node, but I'm debugging it using Node mostly because that's what you get when you launch a js file in VS Code. As already mentioned the production target is Jint. The debug environment is as close an approximation as I can manage, and I'm refraining from using facilities that won't be available in production.
In the debug environment I've created a package.json file that governs whether it's treated as CommonJS or ES6. Experiments show that the production environment behaves like more ES6 than commonjs. The trouble is that most packages expect that they will run either in a browser or in node. They don't consider the possibility of another headless environment. The code above decides it's browser hosted and tries to to install Papa in a DOM that isn't there.
The solution is to wrap the module like so (I'm constructing a string in C#).
string wrapped =
#"(function () {
var module = { exports: { } }; " +
jsSource + #"
var strays = Object.keys(this).filter(itemName => itemName !== 'module');
if (strays.length === 1)
module.exports = this[strays[0]]
else
strays.forEach(itemName => module.exports[itemName] = this[itemName]);
return module.exports;
}).apply({});";
Wrapping it in an anonymous function and using apply to set this for the call allows the "DOM write" to succeed, and a little follow-up code looks for these strays, gathering them up into module exports before returning to normal operation.
In Node.js, is there any way to require file from the same package without using relative paths? For example, here's a snippet of code from ESLint.
const rule = require("../../../lib/rules/accessor-pairs"),
{ RuleTester } = require("../../../lib/rule-tester");
The fact that we have to walk all the way up the tree ../../../ to get to the root is not only annoying. It's also brittle, because I can't move the code without updating all of dependency references.
Yet somehow Node.js developers seem to have lived with it the past 10 years. I can't find anything in the docs or Stack Overflow that solves this problem other than a third-party dependency called require-self. Nor have I been able to find a definitive statement that using relative paths is the only non-hacky way for a file to require another file in the same module.
If there's a way to specify a path relative to the package root in ECMAScript Modules (ESM) but not CommonJS (CJS), or vice versa, I would like to know that as well.
To be clear, I don't think there is a solution to the problem. If there is great. Otherwise, I'm looking for confirmation with an authoritative reference.
Not necessarily the same package - if you are writing libraries this won't be useful, but if you are writing the "final application" - the thing that actually gets run:
One option:
If the NODE_PATH environment variable is set to a colon-delimited list of absolute paths, then Node.js will search those paths for modules if they are not found elsewhere.
So you can do any of:
1.
export NODE_PATH=.
node app.js
NODE_PATH=. node app.js
// app.js (or whatever your entry point is) before *any* require() calls
process.env.NODE_PATH = __dirname;
require('module').Module._initPaths();
Or, another way:
The Module object representing the entry script loaded when the Node.js process launched.
https://nodejs.org/api/modules.html#modules_require_main
So you can just do:
const rule = require.main.require("./lib/rules/accessor-pairs")
anytime you want it to be relative to the root (assuming that is how you have your project structured).
You can use the package name itself as a "symlink" to the package root.
Example - foo package imports bar script relative to the foo package root.
package.json
{
"dependencies": {
"foo": "file:./foo"
}
}
index.js
const foo = require('foo');
console.log(foo.bar); // prints "hello"
foo/index.js
const bar = require('foo/bar'); // import relative to the package root
module.exports = {
bar: bar
}
foo/bar.js
module.exports = 'hello';
If you use vscode then you're in luck - !!!! jsconfig.json in project root handles this masterfully for commonjs, es6, amd, umd, etc
The jsconfig.json file specifies the root files and the options for the features provided by the JavaScript language service.
jsconfig.json
{
"compilerOptions": {
"module": "commonjs",
"baseUrl": ".",
"paths": {
"#rules/*": ["path/to/lib/rules/"]
}
}
}
and then to use the alias:
const rule = require('#rules/accessor-pairs'),
{ RuleTester } = require('#rules/rule-tester');
Read more:
https://code.visualstudio.com/docs/languages/jsconfig
I'm in the process of building an npm package which will be installed globally. Is it possible to have non-code files installed alongside code files that can be referenced from code files?
For example, if my package includes someTextFile.txt and a module.js file (and my package.json includes "bin": {"someCommand":"./module.js"}) can I read the contents of someTextFile.txt into memory in module.js? How would I do that?
The following is an example of a module that loads the contents of a file (string) into the global scope.
core.js : the main module file (entry point of package.json)
//:Understanding: module.exports
module.exports = {
reload:(cb)=>{ console.log("[>] Magick reloading to memory"); ReadSpellBook(cb)}
}
//:Understanding: global object
//the following function is only accesible by the magick module
const ReadSpellBook=(cb)=>{
require('fs').readFile(__dirname+"/spellBook.txt","utf8",(e,theSpells)=>{
if(e){ console.log("[!] The Spell Book is MISSING!\n"); cb(e)}
else{
console.log("[*] Reading Spell Book")
//since we want to make the contents of .txt accesible :
global.SpellBook = theSpells // global.SpellBook is now shared accross all the code (global scope)
cb()//callBack
}
})
}
//·: Initialize :.
console.log("[+] Time for some Magick!")
ReadSpellBook((e)=>e?console.log(e):console.log(SpellBook))
spellBook.txt
ᚠ ᚡ ᚢ ᚣ ᚤ ᚥ ᚦ ᚧ ᚨ ᚩ ᚪ ᚫ ᚬ ᚭ ᚮ ᚯ
ᚰ ᚱ ᚲ ᚳ ᚴ ᚵ ᚶ ᚷ ᚸ ᚹ ᚺ ᚻ ᚼ ᚽ ᚾ ᚿ
ᛀ ᛁ ᛂ ᛃ ᛄ ᛅ ᛆ ᛇ ᛈ ᛉ ᛊ ᛋ ᛌ ᛍ ᛎ ᛏ
ᛐ ᛑ ᛒ ᛓ ᛔ ᛕ ᛖ ᛗ ᛘ ᛙ ᛚ ᛛ ᛜ ᛝ ᛞ ᛟ
ᛠ ᛡ ᛢ ᛣ ᛤ ᛥ ᛦ ᛧ ᛨ ᛩ ᛪ ᛫ ᛬ ᛭ ᛮ ᛯ
If you require it from another piece of code, you will see how it prints to the console and initializes by itself.
If you want to achieve a manual initalization, simply remove the 3 last lines (·: Initialize :.) and use reload() :
const magick = require("core.js")
magick.reload((error)=>{ if(error){throw error}else{
//now you know the SpellBook is loaded
console.log(SpellBook.length)
})
I have built some CLIs which were distributed privately, so I believe I can illuminate a bit here.
Let's say your global modules are installed at a directory called $PATH. When your package will be installed on any machine, it will essentially be extracted at that directory.
When you'll fire up someCommand from any terminal, the module.js will be invoked which was kept at $PATH. If you initially kept the template file in the same directory as your package, then it will be present at that location which is local to module.js.
Assuming you edit the template as a string and then want to write it locally to where the user wished / pwd, you just have to use process.cwd() to get the path to that directory. This totally depends on how you code it out.
In case you want to explicitly include the files only in the npm package, then use files attribute of package.json.
As to particularly answer "how can my code file in the npm package locate the path to the globally installed npm folder in which it is located in a way that is guaranteed to work across OSes and is future proof?", that is very very different from the template thingy you were trying to achieve. Anyway, what you're simply asking here is the global path of npm modules. As a fail safe option, use the path returned by require.main.filename within your code to keep that as a reference.
When you npm publish, it packages everything in the folder, excluding things noted in .npmignore. (If you don't have an .npmignore file, it'll dig into .gitignore. See https://docs.npmjs.com/misc/developers#keeping-files-out-of-your-package) So in short, yes, you can package the text file into your module. Installing the module (locally or globally) will get the text file into place in a way you expect.
How do you find the text file once it's installed? __dirname gives you the path of the current file ... if you ask early enough. See https://nodejs.org/docs/latest/api/globals.html#globals_dirname (If you use __dirname inside a closure, it may be the path of the enclosing function.) For the near-term of "future", this doesn't look like it'll change, and will work as expected in all conditions -- whether the module is installed locally or globally, and whether others depend on the module or it's a direct install.
So let's assume the text file is in the same directory as the currently running script:
var fs = require('fs');
var path = require('path');
var dir = __dirname;
function runIt(cb) {
var fullPath = path.combine(__dirname, 'myfile.txt');
fs.readFile(fullPath, 'utf8' , function (e,content) {
if (e) {
return cb(e);
}
// content now has the contents of the file
cb(content);
}
}
module.exports = runIt;
Sweet!
I'm using the expect.js library with my mocha unit tests. Currently, I'm requiring the library on the first line of each file, like this:
var expect = require('expect.js');
describe('something', function () {
it('should pass', function () {
expect(true).to.be(true); // works
});
});
If possible, I'd like to remove the boilerplate require code from the first line of each file, and have my unit tests magically know about expect. I thought I might be able to do this using the mocha.opts file:
--require ./node_modules/expect.js/index.js
But now I get the following error when running my test:
ReferenceError: expect is not defined
This seems to make sense - how can it know that the reference to expect in my tests refers to what is exported by the expect.js library?
The expect library is definitely getting loaded, as if I change the path to something non-existent then mocha says:
"Error: Cannot find module './does-not-exist.js'"
Is there any way to accomplish what I want? I'm running my tests from a gulp task if perhaps that could help.
You are requiring the module properly but as you figured out, the symbols that the module export won't automatically find themselves into the global space. You can remedy this with your own helper module.
Create test/helper.js:
var expect = require("expect.js")
global.expect = expect;
and set your test/mocha.opts to:
--require test/helper
While Louis's answer is spot on, in the end I solved this with a different approach by using karma and the karma-chai plugin:
Install:
npm install karma-chai --save-dev
Configure:
karma.set({
frameworks: ['mocha', 'chai']
// ...
});
Use:
describe('something', function () {
it('should pass', function () {
expect(true).to.be(true); // works
});
});
Thanks to Louis answer and a bit of fiddling around I sorted out my test environment references using mocha.opts. Here is the complete setup.
My project is a legacy JavaScript application with a lot of "plain" js files which I wish to reference both in an html file using script tags and using require for unit testing with mocha.
I am not certain that this is good practice but I am used to Mocha for unit testing in node project and was eager to use the same tool with minimal adaptation.
I found that exporting is easy:
class Foo{...}
class Bar{...}
if (typeof module !== 'undefined') module.exports = { Foo, Bar };
or
class Buzz{...}
if (typeof module !== 'undefined') module.exports = Buzz;
However, trying to use require in all the files was an issue as the browser would complain about variables being already declared even when enclosed in an if block such as:
if (typeof require !== 'undefined') {
var {Foo,Bar} = require('./foobar.js');
}
So I got rid of the require part in the files and set up a mocha.opts file in my test folder with this content. The paths are relative to the root folder:
--require test/mocha.opts.js
mocha.opts.js content. The paths are relative to the location of the file:
global.assert = require('assert');
global.Foo = require("../foobar.js").Foo;
global.Bar = require("../foobar.js").Bar;
global.Buzz = require("../buzz.js");
What is best pratice for referencing a local commonjs module without using a relative path as below?
var custMod= require(‘customModule’);
custMod(true);
custMod.save(‘foo’);
Is there any reference for building a module like this?
If I wrote module like below, getting undefined when I call custMode.save(12);
module.exports = customModule;function customModule(mode) {
var mode = 1;
return {
save: function (value) {
console.log(mode, value)
},
get: function (id) {
console.log(mode, id)
}
}
You can add a path for require to check using
require.paths.push('/my/path');
or
require.main.paths.push('/my/path');
Depending on your node version.
Then if customModule.js exists at /my/path/customModule.js, you can just use
require('customModule');
Do note though, you'd need to do this on every module that you intend to use this method on.
I wish node made this easier to be honest. One possibility:
project_root
`--node_modules
|--npm-module-1
|--npm-module-2
|--... (etc)
`--lib
|--my-module.js
`--my-other-module.js
With the above you can then type require('lib/my-module') from anywhere in your project. (Just make sure and never install an npm module named lib :) Another possibility:
project_root
|--node_modules
| |--npm-module-1
| |--npm-module-2
| `--... (etc)
`--lib
`--node_modules
|--my-module.js
`--my-other-module.js
With the above you can then type require('my-module'), but only for any files under project_root/lib/.
An advantage of the former approach is that require('lib/my-module') makes it super easy at a glance to tell which modules are local to the project. However the latter is less typing.