How to reuse commonjs modules in the browser and server using modulr? - javascript

I'm using modulr for using commonjs modules in the browser.
The goal is to be able to reuse some of those modules also in a server environment.
These "shared" modules need to do something like this:
var _ = _ || require("underscore");
meaning:
if _ exists as a global var (browser environment), use it
else load the "underscore" module (server), and use it instead
Now, since modulr does static analysis in all the code, looking for the require calls in order to generate the final js file, it will fail the build.
Is there a way to work around this problem?
(For example, if modulr supported something like --ignore=<module_list> parameter, everything would run fine.)

Apparently there's no way to fix this in modulr, so I had to create a workaround module named Env which looks like this:
// Env.js
var my = {
modules: undefined,
require: require
};
exports.override = function(modules) {
my.modules = modules;
};
exports.require = function(path) {
if (my.modules && my.modules[path]) {
return my.modules[path];
} else {
// my.require(...) is needed instead of simply require(...)
// because simply require(...) will cause a modulr parsing failure
return my.require(path);
}
};
And at the client side, have a specific initializer that does:
// ClientInitializer.js
Env = require('shared/Env');
Env.override({ underscore: _ });
So, the "shared" modules can do:
// SharedModule.js
var _ = require('shared/Env').require('underscore');
If the "shared" module is running in the server, the normal require function is called.
If it is running in the browser, the Env module will answer with the global _ variable.

Related

Implementing require in the absence of node

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.

Splitting the code(commands/events) in different files [duplicate]

I'm trying to get JavaScript to read/write to a PostgreSQL database. I found this project on GitHub. I was able to get the following sample code to run in Node.
var pg = require('pg'); //native libpq bindings = `var pg = require('pg').native`
var conString = "tcp://postgres:1234#localhost/postgres";
var client = new pg.Client(conString);
client.connect();
//queries are queued and executed one after another once the connection becomes available
client.query("CREATE TEMP TABLE beatles(name varchar(10), height integer, birthday timestamptz)");
client.query("INSERT INTO beatles(name, height, birthday) values($1, $2, $3)", ['Ringo', 67, new Date(1945, 11, 2)]);
client.query("INSERT INTO beatles(name, height, birthday) values($1, $2, $3)", ['John', 68, new Date(1944, 10, 13)]);
//queries can be executed either via text/parameter values passed as individual arguments
//or by passing an options object containing text, (optional) parameter values, and (optional) query name
client.query({
name: 'insert beatle',
text: "INSERT INTO beatles(name, height, birthday) values($1, $2, $3)",
values: ['George', 70, new Date(1946, 02, 14)]
});
//subsequent queries with the same name will be executed without re-parsing the query plan by postgres
client.query({
name: 'insert beatle',
values: ['Paul', 63, new Date(1945, 04, 03)]
});
var query = client.query("SELECT * FROM beatles WHERE name = $1", ['John']);
//can stream row results back 1 at a time
query.on('row', function(row) {
console.log(row);
console.log("Beatle name: %s", row.name); //Beatle name: John
console.log("Beatle birth year: %d", row.birthday.getYear()); //dates are returned as javascript dates
console.log("Beatle height: %d' %d\"", Math.floor(row.height/12), row.height%12); //integers are returned as javascript ints
});
//fired after last row is emitted
query.on('end', function() {
client.end();
});
Next I tried to make it run on a webpage, but nothing seemed to happen. I checked on the JavaScript console and it just says "require not defined".
So what is this "require"? Why does it work in Node but not in a webpage?
Also, before I got it to work in Node, I had to do npm install pg. What's that about? I looked in the directory and didn't find a file pg. Where did it put it, and how does JavaScript find it?
So what is this "require?"
require() is not part of the standard JavaScript API. But in Node.js, it's a built-in function with a special purpose: to load modules.
Modules are a way to split an application into separate files instead of having all of your application in one file. This concept is also present in other languages with minor differences in syntax and behavior, like C's include, Python's import, and so on.
One big difference between Node.js modules and browser JavaScript is how one script's code is accessed from another script's code.
In browser JavaScript, scripts are added via the <script> element. When they execute, they all have direct access to the global scope, a "shared space" among all scripts. Any script can freely define/modify/remove/call anything on the global scope.
In Node.js, each module has its own scope. A module cannot directly access things defined in another module unless it chooses to expose them. To expose things from a module, they must be assigned to exports or module.exports. For a module to access another module's exports or module.exports, it must use require().
In your code, var pg = require('pg'); loads the pg module, a PostgreSQL client for Node.js. This allows your code to access functionality of the PostgreSQL client's APIs via the pg variable.
Why does it work in node but not in a webpage?
require(), module.exports and exports are APIs of a module system that is specific to Node.js. Browsers do not implement this module system.
Also, before I got it to work in node, I had to do npm install pg. What's that about?
NPM is a package repository service that hosts published JavaScript modules. npm install is a command that lets you download packages from their repository.
Where did it put it, and how does Javascript find it?
The npm cli puts all the downloaded modules in a node_modules directory where you ran npm install. Node.js has very detailed documentation on how modules find other modules which includes finding a node_modules directory.
Alright, so let's first start with making the distinction between Javascript in a web browser, and Javascript on a server (CommonJS and Node).
Javascript is a language traditionally confined to a web browser with a limited global context defined mostly by what came to be known as the Document Object Model (DOM) level 0 (the Netscape Navigator Javascript API).
Server-side Javascript eliminates that restriction and allows Javascript to call into various pieces of native code (like the Postgres library) and open sockets.
Now require() is a special function call defined as part of the CommonJS spec. In node, it resolves libraries and modules in the Node search path, now usually defined as node_modules in the same directory (or the directory of the invoked javascript file) or the system-wide search path.
To try to answer the rest of your question, we need to use a proxy between the code running in the the browser and the database server.
Since we are discussing Node and you are already familiar with how to run a query from there, it would make sense to use Node as that proxy.
As a simple example, we're going to make a URL that returns a few facts about a Beatle, given a name, as JSON.
/* your connection code */
var express = require('express');
var app = express.createServer();
app.get('/beatles/:name', function(req, res) {
var name = req.params.name || '';
name = name.replace(/[^a-zA_Z]/, '');
if (!name.length) {
res.send({});
} else {
var query = client.query('SELECT * FROM BEATLES WHERE name =\''+name+'\' LIMIT 1');
var data = {};
query.on('row', function(row) {
data = row;
res.send(data);
});
};
});
app.listen(80, '127.0.0.1');
I noticed that whilst the other answers explained what require is and that it is used to load modules in Node they did not give a full reply on how to load node modules when working in the Browser.
It is quite simple to do. Install your module using npm as you describe, and the module itself will be located in a folder usually called node_modules.
Now the simplest way to load it into your app is to reference it from your html with a script tag which points at this directory. i.e if your node_modules directory is in the root of the project at the same level as your index.html you would write this in your index.html:
<script src="node_modules/ng"></script>
That whole script will now be loaded into the page - so you can access its variables and methods directly.
There are other approaches which are more widely used in larger projects, such as a module loader like require.js. Of the two, I have not used Require myself, but I think it is considered by many people the way to go.
It's used to load modules. Let's use a simple example.
In file circle_object.js:
var Circle = function (radius) {
this.radius = radius
}
Circle.PI = 3.14
Circle.prototype = {
area: function () {
return Circle.PI * this.radius * this.radius;
}
}
We can use this via require, like:
node> require('circle_object')
{}
node> Circle
{ [Function] PI: 3.14 }
node> var c = new Circle(3)
{ radius: 3 }
node> c.area()
The require() method is used to load and cache JavaScript modules. So, if you want to load a local, relative JavaScript module into a Node.js application, you can simply use the require() method.
Example:
var yourModule = require( "your_module_name" ); //.js file extension is optional
Necromancing.
IMHO, the existing answers leave much to be desired.
At first, it's very confusing.
You have a (nowhere defined) function "require", which is used to get modules.
And in said (CommonJS) modules, you can use require, exports and module, WITHOUT THEM EVER BEING DEFINED.
Not that it would be new that you could use undefined variables in JS, but you couldn't use an undefined function.
So it looks a little like magic at first.
But all magic is based on deception.
When you dig a little deeper, it turns out it is really quite simple:
Require is simply a (non-standard) function defined at global scope.
(global scope = window-object in browser, global-object in NodeJS).
Note that by default, the "require function" is only implemented in NodeJS, not in the browser.
Also, note that to add to the confusion, for the browser, there is RequireJS, which, despite the name containing the characters "require", RequireJS absolutely does NOT implement require/CommonJS - instead RequireJS implements AMD, which is something similar, but not the same (aka incompatible).
That last one is just one important thing you have to realize on your way to understanding require.
Now, as such, to answer the question "what is require", we "simply" need to know what this function does.
This is perhaps best explained with code.
Here's a simple implementation by Michele Nasti, the code you can find on his github page.
Let's call our minimalisc implementation of the require function "myRequire":
function myRequire(name)
{
console.log(`Evaluating file ${name}`);
if (!(name in myRequire.cache)) {
console.log(`${name} is not in cache; reading from disk`);
let code = fs.readFileSync(name, 'utf8');
let module = { exports: {} };
myRequire.cache[name] = module;
let wrapper = Function("require, exports, module", code);
wrapper(myRequire, module.exports, module);
}
console.log(`${name} is in cache. Returning it...`);
return myRequire.cache[name].exports;
}
myRequire.cache = Object.create(null);
window.require = myRequire;
const stuff = window.require('./main.js');
console.log(stuff);
Now you notice, the object "fs" is used here.
For simplicity's sake, Michele just imported the NodeJS fs module:
const fs = require('fs');
Which wouldn't be necessary.
So in the browser, you could make a simple implementation of require with a SYNCHRONOUS XmlHttpRequest:
const fs = {
file: `
// module.exports = \"Hello World\";
module.exports = function(){ return 5*3;};
`
, getFile(fileName: string, encoding: string): string
{
// https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest/Synchronous_and_Asynchronous_Requests
let client = new XMLHttpRequest();
// client.setRequestHeader("Content-Type", "text/plain;charset=UTF-8");
// open(method, url, async)
client.open("GET", fileName, false);
client.send();
if (client.status === 200)
return client.responseText;
return null;
}
, readFileSync: function (fileName: string, encoding: string): string
{
// this.getFile(fileName, encoding);
return this.file; // Example, getFile would fetch this file
}
};
Basically, what require thus does, is it downloads a JavaScript-file, evals it in an anonymous namespace (aka Function), with the parameters "require", "exports" and "module", and returns the exports, meaning an object's public functions and properties.
Note that this evaluation is recursive: you require files, which themselfs can require files.
This way, all "global" variables used in your module are variables in the require-wrapper-function namespace, and don't pollute the global scope with unwanted variables.
Also, this way, you can reuse code without depending on namespaces, so you get "modularity" in JavaScript. "modularity" in quotes, because this is not exactly true, though, because you can still write window.bla/global.bla, and hence still pollute the global scope... Also, this establishes a separation between private and public functions, the public functions being the exports.
Now instead of saying
module.exports = function(){ return 5*3;};
You can also say:
function privateSomething()
{
return 42:
}
function privateSomething2()
{
return 21:
}
module.exports = {
getRandomNumber: privateSomething
,getHalfRandomNumber: privateSomething2
};
and return an object.
Also, because your modules get evaluated in a function with parameters
"require", "exports" and "module", your modules can use the undeclared variables "require", "exports" and "module", which might be startling at first. The require parameter there is of course a pointer to the require function saved into a variable.
Cool, right ?
Seen this way, require looses its magic, and becomes simple.
Now, the real require-function will do a few more checks and quirks, of course, but this is the essence of what that boils down to.
Also, in 2020, you should use the ECMA implementations instead of require:
import defaultExport from "module-name";
import * as name from "module-name";
import { export1 } from "module-name";
import { export1 as alias1 } from "module-name";
import { export1 , export2 } from "module-name";
import { foo , bar } from "module-name/path/to/specific/un-exported/file";
import { export1 , export2 as alias2 , [...] } from "module-name";
import defaultExport, { export1 [ , [...] ] } from "module-name";
import defaultExport, * as name from "module-name";
import "module-name";
And if you need a dynamic non-static import (e.g. load a polyfill based on browser-type), there is the ECMA-import function/keyword:
var promise = import("module-name");
note that import is not synchronous like require.
Instead, import is a promise, so
var something = require("something");
becomes
var something = await import("something");
because import returns a promise (asynchronous).
So basically, unlike require, import replaces fs.readFileSync with fs.readFileAsync.
async readFileAsync(fileName, encoding)
{
const textDecoder = new TextDecoder(encoding);
// textDecoder.ignoreBOM = true;
const response = await fetch(fileName);
console.log(response.ok);
console.log(response.status);
console.log(response.statusText);
// let json = await response.json();
// let txt = await response.text();
// let blo:Blob = response.blob();
// let ab:ArrayBuffer = await response.arrayBuffer();
// let fd = await response.formData()
// Read file almost by line
// https://developer.mozilla.org/en-US/docs/Web/API/ReadableStreamDefaultReader/read#Example_2_-_handling_text_line_by_line
let buffer = await response.arrayBuffer();
let file = textDecoder.decode(buffer);
return file;
} // End Function readFileAsync
This of course requires the import-function to be async as well.
"use strict";
async function myRequireAsync(name) {
console.log(`Evaluating file ${name}`);
if (!(name in myRequireAsync.cache)) {
console.log(`${name} is not in cache; reading from disk`);
let code = await fs.readFileAsync(name, 'utf8');
let module = { exports: {} };
myRequireAsync.cache[name] = module;
let wrapper = Function("asyncRequire, exports, module", code);
await wrapper(myRequireAsync, module.exports, module);
}
console.log(`${name} is in cache. Returning it...`);
return myRequireAsync.cache[name].exports;
}
myRequireAsync.cache = Object.create(null);
window.asyncRequire = myRequireAsync;
async () => {
const asyncStuff = await window.asyncRequire('./main.js');
console.log(asyncStuff);
};
Even better, right ?
Well yea, except that there is no ECMA-way to dynamically import synchronously (without promise).
Now, to understand the repercussions, you absolutely might want to read up on promises/async-await here, if you don't know what that is.
But very simply put, if a function returns a promise, it can be "awaited":
"use strict";
function sleep(interval)
{
return new Promise(
function (resolve, reject)
{
let wait = setTimeout(function () {
clearTimeout(wait);
//reject(new Error(`Promise timed out ! (timeout = ${timeout})`));
resolve();
}, interval);
});
}
The promise would then normally be used like this:
function testSleep()
{
sleep(3000).then(function ()
{
console.log("Waited for 3 seconds");
});
}
But when you return a promise, you can also use await, which means we get rid of the callback (sort of - actually, it is being replaced with a state-machine in the compiler/interpreter).
This way, we make asynchronous code feel like synchronous, so we now can use try-catch for error-handling.
Note that if you want to use await in a function, that function must be declared async (hence async-await).
async function testSleep()
{
await sleep(5000);
console.log("i waited 5 seconds");
}
And also please note that in JavaScript, there is no way to call an async function (blockingly) from a synchronous one (the ones you know). So if you want to use await (aka ECMA-import), all your code needs to be async, which most likely is a problem, if everything isn't already async...
An example of where this simplified implementation of require fails, is when you require a file that is not valid JavaScript, e.g. when you require css, html, txt, svg and images or other binary files.
And it's easy to see why:
If you e.g. put HTML into a JavaScript function body, you of course rightfully get
SyntaxError: Unexpected token '<'
because of Function("bla", "<doctype...")
Now, if you wanted to extend this to for example include non-modules, you could just check the downloaded file-contents for code.indexOf("module.exports") == -1, and then e.g. eval("jquery content") instead of Func (which works fine as long as you're in the browser). Since downloads with Fetch/XmlHttpRequests are subject to the same-origin-policy, and integrity is ensured by SSL/TLS, the use of eval here is rather harmless, provided you checked the JS files before you added them to your site, but that much should be standard-operating-procedure.
Note that there are several implementations of require-like functionality:
the CommonJS (CJS) format, used in Node.js, uses a require function and module.exports to define dependencies and modules. The npm ecosystem is built upon this format. (this is what is implemented above)
the Asynchronous Module Definition (AMD) format, used in browsers, uses a define function to define modules. (basically, this is overcomplicated archaic crap that you wouldn't ever want to use). Also, AMD is the format that is implemented by RequireJS (note that despite the name containing the characters "require", AMD absolutely is NOT CommonJS).
the ES Module (ESM) format. As of ES6 (ES2015), JavaScript supports a native module format. It uses an export keyword to export a module’s public API and an import keyword to import it. This is the one you should use if you don't give a flying f*ck about archaic browsers, such as Safari and IE/EdgeHTML.
the System.register format, designed to support ES6 modules within ES5. (the one you should use, if you need support for older browsers (Safari & IE & old versions of Chrome on mobile phones/tablets), because it can load all formats [for some, plugins are required], can handle cyclic-dependencies, and CSS and HTML - don't define your modules as system.register, though - the format is rather complicated, and remember, it can read the other easier formats)
the Universal Module Definition (UMD) format, compatible to all the above mentioned formats (except ECMA), used both in the browser and in Node.js. It’s especially useful if you write modules that can be used in both NodeJS and the browser. It's somewhat flawed, as it doesn't support the latest ECMA modules, though (maybe this will get fixed) - use System.register instead.
Important sidenote on the function argument "exports":
JavaScript uses call-by-value-sharing - meaning objects are passed as a pointer, but the pointer-value itselfs is passed BY VALUE, not by reference. So you can't override exports by assigning it a new object. Instead, if you want to override exports, you need to assign the new object to module.exports - because hey, module is the pointer passed by value, but exports in module.exports is the reference to the original exports pointer.
Important sidenote on module-Scope:
Modules are evaluated ONCE, and then cached by require.
That means all your modules have a Singleton scope.
If you want a non-singleton scope, you have to do something like:
var x = require("foo.js").createInstance();
or simply
var x = require("foo.js")();
with appropriate code returned by your module.
If you need CommonJS-support for the browser (IE5+, Chrome, Firefox),
check out my code in my comment on Michele Nasti's project
You know how when you are running JavaScript in the browser, you have access to variables like "window" or Math? You do not have to declare these variables, they have been written for you to use whenever you want.
Well, when you are running a file in the Node.js environment, there is a variable that you can use. It is called "module" It is an object. It has a property called "exports." And it works like this:
In a file that we will name example.js, you write:
example.js
module.exports = "some code";
Now, you want this string "some code" in another file.
We will name the other file otherFile.js
In this file, you write:
otherFile.js
let str = require('./example.js')
That require() statement goes to the file that you put inside of it, finds whatever data is stored on the module.exports property. The let str = ... part of your code means that whatever that require statement returns is stored to the str variable.
So, in this example, the end-result is that in otherFile.js you now have this:
let string = "some code";
or -
let str = ('./example.js').module.exports
Note:
the file-name that is written inside of the require statement: If it is a local file, it should be the file-path to example.js. Also, the .js extension is added by default, so I didn't have to write it.
You do something similar when requiring node.js libraries, such as Express. In the express.js file, there is an object named 'module', with a property named 'exports'.
So, it looks something like along these lines, under the hood (I am somewhat of a beginner so some of these details might not be exact, but it's to show the concept:
express.js
module.exports = function() {
//It returns an object with all of the server methods
return {
listen: function(port){},
get: function(route, function(req, res){}){}
}
}
If you are requiring a module, it looks like this:
const moduleName = require("module-name");
If you are requiring a local file, it looks like this:
const localFile = require("./path/to/local-file");
(notice the ./ at the beginning of the file name)
Also note that by default, the export is an object .. eg module.exports = {} So, you can write module.exports.myfunction = () => {} before assigning a value to the module.exports. But you can also replace the object by writing module.exports = "I am not an object anymore."
Two flavours of module.exports / require:
(see here)
Flavour 1
export file (misc.js):
var x = 5;
var addX = function(value) {
return value + x;
};
module.exports.x = x;
module.exports.addX = addX;
other file:
var misc = require('./misc');
console.log("Adding %d to 10 gives us %d", misc.x, misc.addX(10));
Flavour 2
export file (user.js):
var User = function(name, email) {
this.name = name;
this.email = email;
};
module.exports = User;
other file:
var user = require('./user');
var u = new user();

Read environment variables and then replace them in client-side JS when using gulp for building prod or dev code

So lets say I have some code in js
const myApiKey = 'id_0001'
But instead of harcoding it I want to put it in some bash script with other env vars and read from it and then replace it in the JS
So lets say for prod I would read from prod-env.sh or for dev I would read them from dev-env.sh and then gulp or some other tool does the magic and replaces MY_API_KEY based on whatever is established inside of prod-env.sh or dev-env.sh.
const myApiKey = MY_API_KEY
Update: I want to add I only care about unix OS, not concerned about windows. In golang there is way to read for example envVars.get('MY_API_KEY'), I'm looking for something similar but for JS in the client side.
If you're using gulp, it sounds like you could use any gulp string replacer, like gulp-replace.
As for writing the gulp task(s). If you are willing to import the environment into your shell first, before running node, you can access the environment via process.env
gulp.task('build', function(){
gulp.src(['example.js'])
.pipe(replace('MY_API_KEY', process.env.MY_API_KEY))
.pipe(gulp.dest('build/'));
});
If you don't want to import the environment files before running node, you can use a library like env2 to read shell environment files.
Another option would be to use js/json to define those environment files, and load them with require.
prod-env.js
{
"MY_API_KEY": "api_key"
}
gulpfile.js
const myEnv = require('./prod-env')
gulp.task('build', function(){
gulp.src(['example.js'])
.pipe(replace('MY_API_KEY', myEnv.MY_API_KEY))
.pipe(gulp.dest('build/'));
});
Also, for a more generic, loopy version of the replace you can do:
gulp.task('build', function () {
stream = gulp.src(['example.js']);
for (const key in process.env) {
stream.pipe('${' + key + '}', process.env[key]);
}
stream.pipe(gulp.dest('build/'));
});
In that last example I added ${} around the environment variable name to make it less prone to accidents. So the source file becomes:
const myApiKey = ${MY_API_KEY}
This answer is an easy way to do this for someone who doesn't want to touch the code they are managing. For example you are on the ops team but not the dev team and need to do what you are describing.
The environment variable NODE_OPTIONS can control many things about the node.js runtime - see https://nodejs.org/api/cli.html#cli_node_options_options
One such option we can set is --require which allows us to run code before anything else is even loaded.
So using this you can create a overwrite.js file to perform this replacement on any non-node_modules script files:
const fs = require('fs');
const original = fs.readFileSync;
// set some custom env variables
// API_KEY_ENV_VAR - the value to set
// API_KEY_TEMPLATE_TOKEN - the token to replace with the value
if (!process.env.API_KEY_TEMPLATE_TOKEN) {
console.error('Please set API_KEY_TEMPLATE_TOKEN');
process.exit(1);
}
if (!process.env.API_KEY_ENV_VAR) {
console.error('Please set API_KEY_ENV_VAR');
process.exit(1);
}
fs.readFileSync = (file, ...args) => {
if (file.includes('node_modules')) {
return original(file, ...args);
}
const fileContents = original(file, ...args).toString(
/* set encoding here, or let it default to utf-8 */
);
return fileContents
.split(process.env.API_KEY_TEMPLATE_TOKEN)
.join(process.env.API_KEY_ENV_VAR);
};
Then use it with a command like this:
export API_KEY_ENV_VAR=123;
export API_KEY_TEMPLATE_TOKEN=TOKEN;
NODE_OPTIONS="--require ./overwrite.js" node target.js
Supposing you had a script target.js
console.log('TOKEN');
It would log 123. You can use this pretty much universally with node, so it should work fine with gulp, grunt, or any others.

Define global variable with webpack

Is it possible to define a global variable with webpack to result something like this:
var myvar = {};
All of the examples I saw were using external file require("imports?$=jquery!./file.js")
There are several way to approach globals:
1. Put your variables in a module.
Webpack evaluates modules only once, so your instance remains global and carries changes through from module to module. So if you create something like a globals.js and export an object of all your globals then you can import './globals' and read/write to these globals. You can import into one module, make changes to the object from a function and import into another module and read those changes in a function. Also remember the order things happen. Webpack will first take all the imports and load them up in order starting in your entry.js. Then it will execute entry.js. So where you read/write to globals is important. Is it from the root scope of a module or in a function called later?
config.js
export default {
FOO: 'bar'
}
somefile.js
import CONFIG from './config.js'
console.log(`FOO: ${CONFIG.FOO}`)
Note: If you want the instance to be new each time, then use an ES6 class. Traditionally in JS you would capitalize classes (as opposed to the lowercase for objects) like
import FooBar from './foo-bar' // <-- Usage: myFooBar = new FooBar()
2. Use Webpack's ProvidePlugin.
Here's how you can do it using Webpack's ProvidePlugin (which makes a module available as a variable in every module and only those modules where you actually use it). This is useful when you don't want to keep typing import Bar from 'foo' again and again. Or you can bring in a package like jQuery or lodash as global here (although you might take a look at Webpack's Externals).
Step 1. Create any module. For example, a global set of utilities would be handy:
utils.js
export function sayHello () {
console.log('hello')
}
Step 2. Alias the module and add to ProvidePlugin:
webpack.config.js
var webpack = require("webpack");
var path = require("path");
// ...
module.exports = {
// ...
resolve: {
extensions: ['', '.js'],
alias: {
'utils': path.resolve(__dirname, './utils') // <-- When you build or restart dev-server, you'll get an error if the path to your utils.js file is incorrect.
}
},
plugins: [
// ...
new webpack.ProvidePlugin({
'utils': 'utils'
})
]
}
Now just call utils.sayHello() in any js file and it should work. Make sure you restart your dev-server if you are using that with Webpack.
Note: Don't forget to tell your linter about the global, so it won't complain. For example, see my answer for ESLint here.
3. Use Webpack's DefinePlugin.
If you just want to use const with string values for your globals, then you can add this plugin to your list of Webpack plugins:
new webpack.DefinePlugin({
PRODUCTION: JSON.stringify(true),
VERSION: JSON.stringify("5fa3b9"),
BROWSER_SUPPORTS_HTML5: true,
TWO: "1+1",
"typeof window": JSON.stringify("object")
})
Use it like:
console.log("Running App version " + VERSION);
if(!BROWSER_SUPPORTS_HTML5) require("html5shiv");
4. Use the global window object (or Node's global).
window.foo = 'bar' // For SPA's, browser environment.
global.foo = 'bar' // Webpack will automatically convert this to window if your project is targeted for web (default), read more here: https://webpack.js.org/configuration/node/
You'll see this commonly used for polyfills, for example: window.Promise = Bluebird
5. Use a package like dotenv.
(For server side projects) The dotenv package will take a local configuration file (which you could add to your .gitignore if there are any keys/credentials) and adds your configuration variables to Node's process.env object.
// As early as possible in your application, require and configure dotenv.
require('dotenv').config()
Create a .env file in the root directory of your project. Add environment-specific variables on new lines in the form of NAME=VALUE. For example:
DB_HOST=localhost
DB_USER=root
DB_PASS=s1mpl3
That's it.
process.env now has the keys and values you defined in your .env file.
var db = require('db')
db.connect({
host: process.env.DB_HOST,
username: process.env.DB_USER,
password: process.env.DB_PASS
})
Notes
Regarding Webpack's Externals, use it if you want to exclude some modules from being included in your built bundle. Webpack will make the module globally available but won't put it in your bundle. This is handy for big libraries like jQuery (because tree shaking external packages doesn't work in Webpack) where you have these loaded on your page already in separate script tags (perhaps from a CDN).
I was about to ask the very same question. After searching a bit further and decyphering part of webpack's documentation I think that what you want is the output.library and output.libraryTarget in the webpack.config.js file.
For example:
js/index.js:
var foo = 3;
var bar = true;
webpack.config.js
module.exports = {
...
entry: './js/index.js',
output: {
path: './www/js/',
filename: 'index.js',
library: 'myLibrary',
libraryTarget: 'var'
...
}
Now if you link the generated www/js/index.js file in a html script tag you can access to myLibrary.foo from anywhere in your other scripts.
Use DefinePlugin.
The DefinePlugin allows you to create global constants which can be
configured at compile time.
new webpack.DefinePlugin(definitions)
Example:
plugins: [
new webpack.DefinePlugin({
PRODUCTION: JSON.stringify(true)
})
//...
]
Usage:
console.log(`Environment is in production: ${PRODUCTION}`);
You can use define window.myvar = {}.
When you want to use it, you can use like window.myvar = 1
DefinePlugin doesn't actually define anything. What it does is replace variables that exist in your bundle code. If the variable doesn't exist in your code, it will do nothing. So it doesn't create global variables.
In order to create a global variable, write it in your code:
window.MyGlobal = MY_GLOBAL;
And use DefinePlugin to replace MY_GLOBAL with some code:
new webpack.DefinePlugin({
'MY_GLOBAL': `'foo'`,
// or
'MY_GLOBAL': `Math.random()`,
}),
Then your output JS will be like this:
window.MyGlobal = 'foo';
// or
window.MyGlobal = Math.random();
But MY_GLOBAL will never actually exist at runtime, because it is never defined. So that's why DefinePlugin has a misleading name.
I solved this issue by setting the global variables as a static properties on the classes to which they are most relevant. In ES5 it looks like this:
var Foo = function(){...};
Foo.globalVar = {};
You may hit this issue, when triing bundle < script > tag js files in some old project.
Do not use webpack for this, it may be even impossible if joining 50+ libraries like jquery and then figuring out all global variables or if they used nested require. I would advice to simply use uglify js instead , which drops all this problems in 2 commands.
npm install uglify-js -g
uglifyjs --compress --mangle --output bundle.js -- js/jquery.js js/silly.js

NodeJS and Javascript (requirejs) dependency injection

I am currently using requirejs to manage module js/css dependencies.
I'd like to discover the possibilities of having node do this via a centralized config file.
So instead of manually doing something like
define([
'jquery'
'lib/somelib'
'views/someview']
within each module.
I'd have node inject the dependencies ie
require('moduleA').setDeps('jquery','lib/somelib','views/someview')
Anyway, I'm interested in any projects looking at dependency injection for node.
thanks
I've come up with a solution for dependency injection. It's called injectr, and it uses node's vm library and replaces the default functionality of require when including a file.
So in your tests, instead of require('libToTest'), use injectr('libToTest' { 'libToMock' : myMock });. I wanted to make the interface as straightforward as possible, with no need to alter the code being tested. I think it works quite well.
It's just worth noting that injectr files are relative to the working directory, unlike require which is relative to the current file, but that shouldn't matter because it's only used in tests.
I've previously toyed with the idea of providing an alternate require to make a form of dependency injection available in Node.js.
Module code
For example, suppose you have following statements in code.js:
fs = require('fs');
console.log(fs.readFileSync('text.txt', 'utf-8'));
If you run this code with node code.js, then it will print out the contents of text.txt.
Injector code
However, suppose you have a test module that wants to abstract away the file system.
Your test file test.js could then look like this:
var origRequire = global.require;
global.require = dependencyLookup;
require('./code.js');
function dependencyLookup (file) {
switch (file) {
case 'fs': return { readFileSync: function () { return "test contents"; } };
default: return origRequire(file);
}
}
If you now run node test.js, it will print out "test contents", even though it includes code.js.
I've also written a module to accomplish this, it's called rewire. Just use npm install rewire and then:
var rewire = require("rewire"),
myModule = rewire("./path/to/myModule.js"); // exactly like require()
// Your module will now export a special setter and getter for private variables.
myModule.__set__("myPrivateVar", 123);
myModule.__get__("myPrivateVar"); // = 123
// This allows you to mock almost everything within the module e.g. the fs-module.
// Just pass the variable name as first parameter and your mock as second.
myModule.__set__("fs", {
readFile: function (path, encoding, cb) {
cb(null, "Success!");
}
});
myModule.readSomethingFromFileSystem(function (err, data) {
console.log(data); // = Success!
});
I've been inspired by Nathan MacInnes's injectr but used a different approach. I don't use vm to eval the test-module, in fact I use node's own require. This way your module behaves exactly like using require() (except your modifications). Also debugging is fully supported.

Categories