Understanding Callbacks in JavaScript
A common pattern in JavaScript is passing an inline function as a parameter to another function. For example:
[1, 2, 3].forEach(function callback(v) {
console.log(v); // 1, 2, 3
});
fetch('https://google.com').then(function callback(res) {
console.log(res);
});
new EventEmitter().on('event', function callback(ev) {
console.log(ev);
});In the most generic sense, a callback in JavaScript is a function that will
be called for you by someone else. The "someone else" may be a built-in JavaScript
function like setTimeout(), a framework like Express, a library like Axios, or any other bit of
JavaScript code. Callbacks are reactive as opposed to imperative: you define a function that should be called by another section of code, but you are not responsible for calling that function.
// The JavaScript runtime is responsible for calling `callback`
// in 5 seconds
setTimeout(function callback() {
console.log('Hello, World!');
}, 5000);
const app = require('express')();
// Express is responsible for calling `callback` with the
// correct parameters
app.get('/', function callback(req, res) {
res.send('Hello, World!');
});Node-style Callbacks
Promises are a relatively new addition to JavaScript, they were first added
in 2015. Before promises and async/await, the de facto paradigm for async
operations in Node.js was passing a callback function that took 2 parameters,
an error err and a result res. When someone talks about "callback hell",
they're usually referring to using Node-style callbacks.
For example, the Node.js SDK for Amazon AWS' S3 currently only supports callbacks. For the AWS SDK to notify you when your file upload is complete, you need to pass a callback function.
const AWS = require('aws-sdk');
const fs = require('fs');
AWS.config.update({
accessKeyId: process.env.AWS_ACCESS_KEY_ID,
secretAccessKey: process.env.AWS_SECRET_ACCESS_KEY
});
const s3 = new AWS.S3();
const options = {
Bucket: process.env.AWS_BUCKET,
Body: fs.createReadStream('./package.json'),
Key: 'package.json'
};
// Pass a `callback` that takes 2 params: `err` and `res`. If `err` is
// null or undefined, you can assume the upload succeeded.
s3.upload(options, function callback(err, res) {
if (err != null) {
// If `err` is set, you should handle the error
}
});The primary purpose of Node-style callbacks is to tell you whether an asynchronous operation succeeded or failed, and, if it succeeded, what the result was. Promises have largely replaced Node-style callbacks for this purpose, because nesting callbacks is syntactically painful, and because Node-style callbacks were never standardized.
Callbacks to Promises
Node.js has a built-in promisify() function that converts callback-based APIs into functions that return promises. Here's how you can use promisify() to convert a Node-callback-based function to a function that returns a promise:
function myCallbackBasedFunction(callback) {
setTimeout(() => callback(null, 'Hello, World!'), 25);
}
const util = require('util');
const myPromiseBasedFunction =
util.promisify(myCallbackBasedFunction);
await myPromiseBasedFunction(); // 'Hello, World!'