Use of require(lib) versus <script> in Electron apps

微笑、不失礼 提交于 2019-11-28 14:18:48

Some libraries only expose their variables through a CommonJS interface. Others, like jQuery, will also expose them as global variables.

The reason you can't just do <script src="..."></script> for a library that only exposes through CommonJS is that it won't bind to the global space.

Binding with CommonJS

module.exports = myLibrary;

Bindings to the global scope

window.myLibrary = myLibrary;

If a library only does the former then you will not be able to access the value without using require. If a library only does the latter, then you will not be able to access it with require in the sense of const myLibrary = require('my-library')

In general, it's a better idea to use CommonJS over global variables. Adding variables to the global scope can lead to name collisions and directly loading in your dependencies makes it easier for the next person to tell where that dependency came from. Not to mention, CommonJS allows static analysis tools to work better so you're more likely to get relevant code completions and type definitions.

Using the example of jQuery, it would be better to use it like this.

// main.js
const $ = require('./js/jquery-2.2.4.min.js');
// could also be done like this if you install it as a Node dependency
// const $ = require('jquery');

$(document).ready(...);

<!-- index.html -->
...
<script src="main.js"></script>

TL;DR

Use require('my-library') when possible, load them as globals when it is not.

Before module bundlers, libraries would have to be imported either via a <script> tag or via module loaders such as RequireJS.

Now it's easier to assume a CommonJS environment and get everything through a module bundler which will expose a require function for you in a browser context.

All of this is not necessary in the context of an Electron app:

In normal browsers, web pages usually run in a sandboxed environment and are not allowed access to native resources. Electron users, however, have the power to use Node.js APIs in web pages allowing lower level operating system interactions.

Cf. renderer process

That means that the native Node.js require function (amongst other things) is available in your renderer process!

Here's a simple Electron app to prove it:

My package.json:

{
  "name": "foobar",
  "version": "1.0.0",
  "main": "main.js",
  "scripts": {
    "start": "electron ."
  },
  "devDependencies": {
    "electron": "^3.0.7"
  },
  "dependencies": {
    "the-answer": "^1.0.0"
  }
}

My main.js: (the main process)

const {app, BrowserWindow} = require('electron');

let mainWindow;

function createWindow () {
  mainWindow = new BrowserWindow({width: 800, height: 600})
  mainWindow.loadFile('index.html');
}

app.on('ready', createWindow);

My index.html: (the renderer process)

<!DOCTYPE html>
<html>
  <body>
    <script>
      const os = require('os'); // Standard Node.js module
      const answer= require('the-answer'); // An NPM package that returns the answer to everything.
    </script>
    <button onclick="alert(os.platform())">YOUR PLATFORM</button>
    <button onclick="alert(answer)">THE ANSWER</button>
  </body>
</html>

So which method should you use?

Electron exposes the native Node.js require function. It would be a shame not to leverage this: you would be able to require packages by their names and split your code into reusable modules as you would do in any Node.js apps.

易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!