Meetup.js – An introduction to Generators



Meetup.js – An introduction to Generators

0 0


meetup-js-generators

:bulb: Introduction to ES6 Generators

On Github codekult / meetup-js-generators

Meetup.js

An introduction to Generators

Hi, I'm Diego Calderón

I work @ MercadoLibre, in MercadoShops dpt.

We gonna talk about generatorsin ES6 EcmaScript2015.

So let’s start…

JS’ run-to-completion behaviour

JavaScript is single threaded.

And have what is called a run-to-completion behavior.

Or in a more figurative way…

But there is a new kid on the block:

Generators

Generators are a new type of function which can be paused and resumed. Breaking the run-to-completion behavior.

Can we use generators right now?

http://kangax.github.io/compat-table/es6/#generators

There are some features money can’t buy.

For everything else there’s Babel.

Show me the syntax

Generators are declared as regular functions but in the following manner:

function* myGen() { ... }
// Or the same `function *myGen() { ... }`

We can argue about how to write a generator signature, but only when we finished the tabs vs spaces talk.

Also the yield keyword is introduced, and we’ll use as a mark for pause/restart the generator and as a way to send and receive data.

function *myGen() {
   console.log('Hi');
   yield;
   console.log('Bye');
}

var objGen = myGen();
objGen.next(); // 'Hi'
objGen.next(); // 'Bye'

Error handling

A benefit of generators is its synchronous like semantics, so we can use ol’ try { ... } catch (err) {} to handle errors.

function *myGen () {
     try {
         var number = 3;
         yield;
         console.log(number.toUpperCase());
     } catch (err) {
         console.error(err);
     }
 }

 var objGen = myGen();
 objGen.next();
 objGen.next();

Online example

Delegate generators

We can include a generator inside another generator using a yield variation: yield*. The nesting levels are infinite.

function *innerGen() {
    yield console.log(2);
    yield console.log(3);
    yield console.log(4);
}

function *outerGen() {
    yield console.log(1);
    yield* innerGen();
    // Or the same yield *innerGen();
    yield console.log(5);
}

Online example

Understanding iterators and iterables

Every time we’re using a generator like this:

// Given a ´myGen´ generator function
var objGen = myGen();

We’re creating a new iterator, which we assign to the variable objGen, to control *myGen() generator.

Iterator's interface counts with the next method.

Iterables are those data structures that comes with an built-in iterator in their prototypes.

A benefit of iterables is that them can be consumed by data consumers as for…of loops, the spread operator and can be used in array destructuring.

In ES6 some iterable data sources are:

Calling .next on an iterable gives us an object containing a value, or an indication that there are no further values.

var myArr = [1,2,3],
    myArrIt = myArr[Symbol.iterator]();

myArrIt.next(); // Object {value: 1, done: false}
myArrIt.next(); // Object {value: 2, done: false}
myArrIt.next(); // Object {value: 3, done: false}
myArrIt.next(); // Object {value: undefined, done: true}

More on the iteration protocol:

Generator’s roles

Generators as data producers

Each yield can return a value via next().

Which means that generators can produce sequences of values via loops and recursion.

Return values with yield

yield can return a value to the generator caller.

function *myGen (x) {
    var single = yield x,
        double = yield (x * 2),
        triple = yield (x * 3);
}

var objGen = myGen(4);
console.log(objGen.next().value); // 4
console.log(objGen.next().value); // 8
console.log(objGen.next().value); // 12

Online Example

Using return inside generators

We can achieve the later example using return.

function *myGen (x) {
    var single = yield x,
        double = yield (x * 2),
        triple = x * 3;
    return triple;
}

var objGen = myGen(4);
console.log(objGen.next().value); // 4
console.log(objGen.next().value); // 8
console.log(objGen.next().value); // 12

Online example

Also we can use return with delegated generators.

Let’s see an Online example

Cycling through generators

Given a generator as the following:

function *myGen() {
    yield 1;
    yield 2;
    yield 3;
}

Using for…of loops:

for (let n of myGen()) {
    console.log(n); // 1, 2, 3
}

Online example

Using spread operator:

let myArr = [...myGen()]; // [1,2,3]

Online example

Destructuring :

let [a,b, ...rest] = myGen(); // [1,2,3]

Online example

Generators as data consumers

yield can also receive a value from next().

That means that generator become a data consumer that pause itself until a new value is pushed into it.

There are three ways of sending data to the generator: via next(), return(), or throw().

Sending data via next():

function *myGen() {
    console.log('Started…');
    var double = (yield) * 2,
        triple = (yield) * 3;
    console.log(double);
    console.log(triple);
}

var objGen = myGen();
objGen.next(); // Start generator
objGen.next(5); // 10
objGen.next(3); // 9

Terminating the generator via return():

function *myGen() {
    console.log('Started…');
    yield; // Inserts a `return` at this point
    console.log('Never gets here');
}

var objGen = myGen();
objGen.next();
console.log(objGen.return('…This is the end.').value);

Online example

Throwing errors via throw():

function *myGen() {
    try {
        console.log('Started…');
        yield; // Throws error at this point
        console.log('Never gets here');
    } catch (err) {
        console.error(`Error: ${err}`);
    }
}

var objGen = myGen();
objGen.next();
objGen.throw('An error ocurred');

Online example

Generators as data producers/consumers

We’ve seen how generators can act as data producers or consumers.

Now we can apply these behaviours in a combined way to create new patterns, as cooperative tasks, mimic coroutines or applying complex techniques as CSP.

Also we can apply this pattern to asynchronous operations to get the maximun benefit from generators capabilities.

GeneratorsmeetPromises

We can use yield for waiting for a promise, and making that the promise itself resume the generator flow when it’s resolved.

This allow us to write more readable and maintenable code, which looks synchonous like, hiding away the async implementations.

Let’s see an example of this power.

The async future

Async functions

The previous pattern is very useful.

And is included in the ES7 EcmaScript2016 specification as async functions.

This new feature is merely sugar for the implementation that we saw earlier.

We can use an async function as follow:

async function() {
    try {
        let req = await request(http://someurl.com/getContent); // Promise aware function
        // Do someting with `req`
    } catch (err) {
        console.error(err);
    }
}

push/async generators and observables

We got generator functions

and async functions…

So we we can haveasync generator functions.

However, async generators has beendeprecated in favor of Observables.

And we have an alternative but similar feature: push generators.

Both push generators and observables are just proposals at the moment.

Credits

Q&A

Thanks!

codekult.github.io

Twitter: @codekult

Meetup.js An introduction to Generators