ES6 Generators

Agenda

  • What is a generator?
  • Generators vs Regular functions
  • What does the syntax of generator look like?
  • How do you call a generator?
  • Generators calling generators
  • What would you use this for?

Learning Goals

  • You should be able to make your own generator
  • You should be able to iterate over your generator
  • You should know how to pass values to the generator iterator

Vocab

  • Generator
  • Generator iterator
  • Generator object

What is an ES6 Generator Function?

Generators are functions that can be exited, and then re-entered, and save context (variable bindings) across re-entrances. Let’s start with an example of familiar JS syntax.

function cheer() {
  console.log('skol Vikings')
  console.log('WOOOOOO!')
}

cheer()

What do you expect the output to be?

=> skol Vikings!
=> WOOOOO!

This seems legit.

Typically, JavaScript functions will run each line of code one after the other in sequential order without stopping until the return statement. Once a function is called it will “run to completion” before any other JS code can be run.

This is because JavaScript is “single-threaded”. Meaning only one command can be executed at any given time.

Wouldn’t it be awesome if I could execute one line, do something else, then execute the next line?

function cheer() {   
  console.log('skol Vikings!')
  doSomethingImportant()
  console.log('WOOOO!')
}

function doSomethingImportant() {
	console.log('getting a snack');
}

cheer()

We get:

=> skol Vikings!
=> getting a snack
=> WOOOOO!

Great! Everything happened a little fast, but it did what I wanted it to do.

What about this:

function cheer() {
  console.log('skol Vikings!')
  doSomethingImportant()
  console.log('WOOOO!')
}

function doSomethingImportant() {
	setTimeout(() => console.log('getting a snack'), 1000)
}

cheer()

What do you expect to happen here? In a perfect world, I still want to see skol Vikings!, then I want to see getting a snack, then WOOOO!. Instead, the first line doesn’t wait for the second line to execute so we get:

=> skol Vikings!
=> WOOOO!
=> getting a snack

This is where Generators come in to save the day.

Generators are functions that can be exited, and then re-entered, and save context (variable bindings) across re-entrances.

Context: The async/await pattern in ES7 is built on top of the pattern of generators. They were introduced to the language primarily as a way to avoid the ‘callback hell’ that was prevelant in JavaScript before the introduction of Promises.

Let’s take a look at a simple example:

function* doSomething() {
  yield 'hello world'  
  yield 'something else'
}

There are a few things to note here. First, generator functions are indicated with an asterisk.

Note: The asterisk can be next to either the word function* or the function name *doSomething.

Second, the content within the function starts with a yield expression. This is what tells the generator to pause. Once the generator function is called, it will only execute the code up until it encounters the special word yield. This tells the generator function to return whatever is to the right of the yield, and then pause until told to continue.

After executing the code to the right of the first yield, it cannot continue until you tell it to, which is done using a Generator Iterator. When you tell it to continue, it will only execute as far as the next yield statement.

It’s also important to note that we cannot use the arrow syntax to define a generator. We need to use the keyword function.

Generator Iterators

Take a look at our first example of a generator function.

function* doSomething() {
  yield 'hello world'  
  yield 'other stuff'
}

Stop and think: What do you expect to see if you run doSomething()? What do we actually get?

doSomething()
=> //....radio silence

This is because doSomething() is no longer a vanilla JS function, it has been transformed by the asterisk into a Generator. Check it out:

function* doSomething() {
  yield 'hello world'  
  yield 'other stuff'
}

function normalFunction() {
  return 'huzzah!'
}

console.log(doSomething())
console.log(normalFunction())

=> Generator {}
=> huzzah!

Simply calling a generator function (ie: just calling doSomething()) doesn’t actually execute it’s contents, instead it returns a Generator Object. We need something more. Luckily, there’s this thing called a Generator Iterator. To everyone’s surprise, it…iterates…over a generator. Let’s see what that actually looks like:

First off, we need to construct the iterator. This is done by first saving your generator function to a variable for easy access.

const words = doSomething()

To command the generator to execute the function and read the next line of code you must use the method .next().

function* doSomething() {
  yield 'hello world'
  yield 'HAHAHAH'
}

const words = doSomething()

console.log(words.next())

Once again, what do you expect to see?

The answer is super interesting! You might expect the return value to be 'hello world'. In fact, the Generator returns a Generator Object.

=> Object {value: 'hello world', done: false}

The object tells us two things:

  1. The return value from the yield expression.
  2. A boolean indicating if the function is fully executed.

In order to get the value, simply chain .value onto the call.

console.log(words.next().value)

Practical Use Cases

But when do you actually want to use this type of thing?

One use case is to iterate over a large function, executing other things on the way without breaking the flow of the function.

function *numbers() {
  yield 1
  yield 2
  yield 3
  yield 4
  yield 5
}

const generator = numbers()

If you were to call generator.next() 5 times, you would get 5 generator objects. Seems legit. But like always, calling a function 5 times to do the same thing defeats the purpose of being a lazy programmer.

Instead, we can call a for of loop (which, shockingly, is also ES6).

for(let num of generator){
	console.log(num)
}

Or collect the data in an external variable.

const list = []
for(let num of generator){
  list.push(num)
}
console.log(list)

Or use a spread operator:

const [...numbers] = foo();

This seems kind of useless…we might as well just make an array. But let’s check out an example where instead of simply printing numbers we pop out of our function and do something else.

function *numbers() {
  yield 1
  yield 2
  yield* moreNums()
  yield 6
  yield 7
}

function *moreNums() {
  yield 3
  yield 4
  yield 5
}

const generator = numbers()

for (let num of generator ){
  console.log(num)
}

With the ability to step through a function, you are now able to control when your function executes each line of code, as well as implement a kind of 2-way communication.

With normal functions, you can pass in parameters at the beginning, and you get a return value at the end. With generators, you send messages out with each yield, and you can pass messages back in each time you restart the function.

For example:

function *adding(){
  const result = 1 + 1
  return 20 + (yield result)
}

const sum = adding()

Think about what you expect to see on the first sum.next() call.

sum.next()
=> Object { value: 2, done: false}

This is because the function will kick off and execute until it encounters the first yield expression. At that point it will pause and return anything to the right of that expression, in this case result, which is 1+1, which is 2.

Let’s do it again. What do you expect?

sum.next()
=> Object {value: NaN, done: true}

Why did this happen? In this situation, the generator has paused on the yield expression and is waiting for input. What input you provide will replace the entire (yield result) chunk of code.

sum.next(10)
=> Object{value: 30, done:true}

Here is another example taken from this blog post (which also has a monty python bridge reference…which is cool…):

function* crossBridge() {
  const reply = yield 'What is your favorite color?';
  console.log(reply);
  if (reply !== 'yellow') return 'Wrong!'
  return 'You may pass.';
}

{
  const iter = crossBridge();
  const q = iter.next().value; // Iterator yields question
  console.log(q);
  const a = iter.next('blue').value; // Pass reply back into generator
  console.log(a);
}

Let’s look at another example to demonstrate how pieces of the function get replaced with arguments passed in.

function* yellStuff(){
  var firstYell = yield 'AUGHH KELLY CLARKSON'
  var secondYell = yield firstYell + "...HAHAHA!!!"
  return firstYell + "..." + secondYell + "...LOUD NOISES!!!"
}

const shouting = yellStuff()

console.log(shouting.next())
=> Object {value: "AUGHH KELLY CLARKSON", done: false}

So our first execution returns what is to the right of the first yield expression, in this case AUGHH KELLY CLARKSON. Then it pauses. Because we saved our yield expression to a variable that we reference in our next line, our generator expects an input to replace that value.

If I give it nothing:

console.log(shouting.next())
=> Object {value: "AUGHH KELLY CLARKSON", done: false}

console.log(shouting.next())
=> Object {value: 'undefined...HAHAHA!!', done: false}

Vs

console.log(shouting.next())
=> Object {value: "AUGHH KELLY CLARKSON", done: false}

console.log(shouting.next('OTHER LOUD THINGS'))
=> Object {value: 'OTHER LOUD THINGS...HAHAHA!!', done: false}

And our last execution needs another input to replace yield firstYell statement, and save to the variable secondYell.

function* yellStuff(){
  var firstYell = yield 'AUGHH KELLY CLARKSON'
  var secondYell = yield firstYell + "...HAHAHA!!!"
  return firstYell + "..." + secondYell + "...LOUD NOISES"
}

const shouting = yellStuff()

console.log(shouting.next())
// Pause and wait for input. Input will replace the entire chunk of code to the right of the first =.
=> Object {value: "AUGHH KELLY CLARKSON", done: false}

console.log(shouting.next('OTHER LOUD THINGS'))
// firstYell = 'OTHER LOUD THINGS'
// yield pauses the function and waits for another input which will replace everything to the right of second `=`
=> Object {value: "OTHER LOUD THINGS...HAHAHA!!!", done: false}

console.log(shouting.next('HAM SANDWICH'))
// secondYell = 'HAM SANDWICH'

=> Object {value: "OTHER LOUD THINGS...HAM SANDWICH...LOUD NOISES", done: true}

One final example with asynchronous functions.

function* makeAPizza() {
    var toppings = yield getToppings();
    var pizza = yield getPizza(toppings)
};

function getToppings(){
	console.log("in toppings function")
  return 'What are your fav pizza toppings?'
}

function getPizza(toppings){
  console.log("making pizza")
  return "Here is your " + [...toppings] + " pizza!"
}

var doStuff = makeAPizza();

console.log(doStuff.next())
console.log(doStuff.next(['pineapple', 'pepperoni']).value)

Generators vs Callbacks

Generators help you avoid the callback pyramid of hell. Take the following example:

function doAllTheThings(callback) {
  //....
  thing1(function(val) {
    //....
    thing2(val, function(val) {
      //....
      thing3(val, callback);
    });
  });
}

This is pretty gross. So what about Promises?

Generators vs Promises

Promises are great! The code above using Promises would look something like this:

function doAllTheThings() {
  return Promise.resolve()
    .then(function() {
        //...
        return thing1(val)
    })
    .then(function(val) {
        //... some code 2
        return thing2(val)
    })
    .then(function(val) {
        //... some code 3
        return thing3(val);
    });
}

This is much more readable, but it still requires you to write the function asynchronously and provide callbacks.

Using Generators that code now looks like this:

function* doAllTheThings() {
    //... some code 1
    var val1 = yield task1();
    //... some code 2
    var val2 = yield task2(val1);
    //... some code 3
    var val3 = yield task3(val2);

    return val3;
};

Cool Examples

function* fibonacci() { // a generator function
  let [prev, curr] = [1, 1];
  while (true) {
    [prev, curr] = [curr, prev + curr];
    yield curr;
  }
}

for (let n of fibonacci()) {
  console.log(n);
  // truncate the sequence at 1000
  if (n >= 1000) {
    break;
  }
}

Read

MDN Docs on Generators
The Basics of ES6 Generators

Lesson Search Results

Showing top 10 results