JavaScript: de devoluciones de llamada a async / await

Programación

JavaScript es sincrónico. Esto significa que ejecutará su bloque de código por orden después de la elevación. Antes se ejecuta el código, vary functionlas declaraciones son “izan” a la parte superior de su alcance.

Este es un ejemplo de un código síncrono:

console.log('1') console.log('2') console.log('3')

Este código registrará de forma fiable "1 2 3".

Las solicitudes asincrónicas esperarán a que finalice un temporizador o una solicitud para responder mientras el resto del código continúa ejecutándose. Luego, cuando sea el momento adecuado, una devolución de llamada hará que estas solicitudes asincrónicas entren en acción.

Este es un ejemplo de un código asincrónico:

console.log('1') setTimeout(function afterTwoSeconds() { console.log('2') }, 2000) console.log('3')

En realidad, esto registrará "1 3 2", ya que el "2" está en un setTimeoutque solo se ejecutará, en este ejemplo, después de dos segundos. Su aplicación no se cuelga esperando que finalicen los dos segundos. En cambio, sigue ejecutando el resto del código y cuando finaliza el tiempo de espera vuelve a afterTwoSeconds.

Puede preguntar "¿Por qué es útil?" o "¿Cómo puedo sincronizar mi código asincrónico?". Ojalá pueda mostrarte las respuestas.

"El problema"

Digamos que nuestro objetivo es buscar un usuario de GitHub y obtener todos los repositorios de ese usuario. Lo que pasa es que no sabemos el nombre exacto del usuario. Entonces tenemos que listar todos los usuarios con nombre similar y sus respectivos repositorios.

No necesita ser súper elegante, algo como esto

En estos ejemplos, el código de solicitud utilizará XHR (XMLHttpRequest). Puede reemplazarlo con jQuery $.ajaxo el enfoque nativo más reciente llamado fetch. Ambos le darán el enfoque de promesas desde la puerta.

Se modificará ligeramente según su enfoque, pero como punto de partida:

// url argument can be something like '//api.github.com/users/daspinola/repos' function request(url) { const xhr = new XMLHttpRequest(); xhr.timeout = 2000; xhr.onreadystatechange = function(e) { if (xhr.readyState === 4) { if (xhr.status === 200) { // Code here for the server answer when successful } else { // Code here for the server answer when not successful } } } xhr.ontimeout = function () { // Well, it took to long do some code here to handle that } xhr.open('get', url, true) xhr.send(); }

Recuerde que en estos ejemplos lo importante no es cuál es el resultado final del código. En cambio, su objetivo debe ser comprender las diferencias de los enfoques y cómo puede aprovecharlos para su desarrollo.

Llamar de vuelta

Puede guardar una referencia de una función en una variable cuando usa JavaScript. Luego, puede usarlos como argumentos de otra función para ejecutar más tarde. Esta es nuestra "devolución de llamada".

Un ejemplo sería:

// Execute the function "doThis" with another function as parameter, in this case "andThenThis". doThis will execute whatever code it has and when it finishes it should have "andThenThis" being executed. doThis(andThenThis) // Inside of "doThis" it's referenced as "callback" which is just a variable that is holding the reference to this function function andThenThis() { console.log('and then this') } // You can name it whatever you want, "callback" is common approach function doThis(callback) { console.log('this first') // the '()' is when you are telling your code to execute the function reference else it will just log the reference callback() }

Usar el callbackpara resolver nuestro problema nos permite hacer algo como esto con la requestfunción que definimos anteriormente:

function request(url, callback) { const xhr = new XMLHttpRequest(); xhr.timeout = 2000; xhr.onreadystatechange = function(e) { if (xhr.readyState === 4) { if (xhr.status === 200) { callback(null, xhr.response) } else { callback(xhr.status, null) } } } xhr.ontimeout = function () { console.log('Timeout') } xhr.open('get', url, true) xhr.send(); }

Nuestra función para la solicitud ahora aceptará a, de callbackmodo que cuando requestse haga se llamará en caso de error y en caso de éxito.

const userGet = `//api.github.com/search/users?page=1&q=daspinola&type=Users` request(userGet, function handleUsersList(error, users) { if (error) throw error const list = JSON.parse(users).items list.forEach(function(user) { request(user.repos_url, function handleReposList(err, repos) { if (err) throw err // Handle the repositories list here }) }) })

Desglosando esto:

  • Hacemos una solicitud para obtener los repositorios de un usuario.
  • Una vez completada la solicitud, utilizamos la devolución de llamada handleUsersList
  • Si no hay ningún error, analizamos la respuesta de nuestro servidor en un objeto usando JSON.parse
  • Luego iteramos nuestra lista de usuarios, ya que puede tener más de uno.

    Para cada usuario solicitamos su lista de repositorios.

    Usaremos la URL que devolvió por usuario en nuestra primera respuesta.

    Llamamos repos_urlcomo la URL para nuestras próximas solicitudes o desde la primera respuesta

  • Cuando la solicitud haya completado la devolución de llamada, llamaremos

    Esto manejará su error o la respuesta con la lista de repositorios para ese usuario

Nota : Enviar el error primero como parámetro es una práctica común, especialmente cuando se usa Node.js.

Un enfoque más "completo" y legible sería tener algún manejo de errores. Mantendríamos la devolución de llamada separada de la ejecución de la solicitud.

Algo como esto:

try { request(userGet, handleUsersList) } catch (e) { console.error('Request boom! ', e) } function handleUsersList(error, users) { if (error) throw error const list = JSON.parse(users).items list.forEach(function(user) { request(user.repos_url, handleReposList) }) } function handleReposList(err, repos) { if (err) throw err // Handle the repositories list here console.log('My very few repos', repos) }

Esto termina teniendo problemas como carreras y problemas de manejo de errores. Las carreras ocurren cuando no controlas qué usuario obtendrás primero. Estamos solicitando la información de todos ellos en caso de que haya más de uno. No estamos teniendo en cuenta un pedido. Por ejemplo, el usuario 10 puede ser el primero y el usuario 2 el último. Tenemos una posible solución más adelante en el artículo.

The main problem with callbacks is that maintenance and readability can become a pain. It sort of already is and the code does hardly anything. This is known as callback hell which can be avoided with our next approach.

Promises

Promises you can make your code more readable. A new developer can come to the code base and see a clear order of execution to your code.

To create a promise you can use:

const myPromise = new Promise(function(resolve, reject) { // code here if (codeIsFine) { resolve('fine') } else { reject('error') } }) myPromise .then(function whenOk(response) { console.log(response) return response }) .catch(function notOk(err) { console.error(err) })

Let us decompose it:

  • A promise is initialized with a function that has resolve and reject statements
  • Make your async code inside the Promise function

    resolve when everything happens as desired

    Otherwise reject

  • When a resolve is found the .then method will execute for that Promise

    When a reject is found the .catch will be triggered

Things to bear in mind:

  • resolve and reject only accept one parameter

    resolve(‘yey’, ‘works’) will only send ‘yey’ to the .then callback function

  • If you chain multiple .then

    Add a return if you want the next .then value not to be undefined

  • When a reject is caught with .catch if you have a .then chained to it

    It will still execute that .then

    You can see the .then as an “always executes” and you can check an example in this comment

  • With a chain on .then if an error happens on the first one

    It will skip subsequent .then until it finds a .catch

  • A promise has three states

    pending

  • When waiting for a resolve or reject to happen

    resolved

    rejected

  • Once it’s in a resolved or rejected state

    It cannot be changed

Note: You can create promises without the function at the moment of declarations. The way that I’m showing it is only a common way of doing it.

“Theory, theory, theory…I’m confused” you may say.

Let’s use our request example with a promise to try to clear things up:

function request(url) { return new Promise(function (resolve, reject) { const xhr = new XMLHttpRequest(); xhr.timeout = 2000; xhr.onreadystatechange = function(e) { if (xhr.readyState === 4) { if (xhr.status === 200) { resolve(xhr.response) } else { reject(xhr.status) } } } xhr.ontimeout = function () { reject('timeout') } xhr.open('get', url, true) xhr.send(); }) }

In this scenario when you execute request it will return something like this:

const userGet = `//api.github.com/search/users?page=1&q=daspinola&type=Users` const myPromise = request(userGet) console.log('will be pending when logged', myPromise) myPromise .then(function handleUsersList(users) { console.log('when resolve is found it comes here with the response, in this case users ', users) const list = JSON.parse(users).items return Promise.all(list.map(function(user) { return request(user.repos_url) })) }) .then(function handleReposList(repos) { console.log('All users repos in an array', repos) }) .catch(function handleErrors(error) { console.log('when a reject is executed it will come here ignoring the then statement ', error) })

This is how we solve racing and some of the error handling problems. The code is still a bit convoluted. But its a way to show you that this approach can also create readability problems.

A quick fix would be to separate the callbacks like so:

const userGet = `//api.github.com/search/users?page=1&q=daspinola&type=Users` const userRequest = request(userGet) // Just by reading this part out loud you have a good idea of what the code does userRequest .then(handleUsersList) .then(repoRequest) .then(handleReposList) .catch(handleErrors) function handleUsersList(users) { return JSON.parse(users).items } function repoRequest(users) { return Promise.all(users.map(function(user) { return request(user.repos_url) })) } function handleReposList(repos) { console.log('All users repos in an array', repos) } function handleErrors(error) { console.error('Something went wrong ', error) }

By looking at what userRequest is waiting in order with the .then you can get a sense of what we expect of this code block. Everything is more or less separated by responsibility.

This is “scratching the surface” of what Promises are. To have a great insight on how they work I cannot recommend enough this article.

Generators

Another approach is to use the generators. This is a bit more advance so if you are starting out feel free to jump to the next topic.

One use for generators is that they allow you to have async code looking like sync.

They are represented by a * in a function and look something like:

function* foo() { yield 1 const args = yield 2 console.log(args) } var fooIterator = foo() console.log(fooIterator.next().value) // will log 1 console.log(fooIterator.next().value) // will log 2 fooIterator.next('aParam') // will log the console.log inside the generator 'aParam'

Instead of returning with a return, generators have a yield statement. It stops the function execution until a .next is made for that function iteration. It is similar to .then promise that only executes when resolved comes back.

Our request function would look like this:

function request(url) { return function(callback) { const xhr = new XMLHttpRequest(); xhr.onreadystatechange = function(e) { if (xhr.readyState === 4) { if (xhr.status === 200) { callback(null, xhr.response) } else { callback(xhr.status, null) } } } xhr.ontimeout = function () { console.log('timeout') } xhr.open('get', url, true) xhr.send() } }

We want to have the url as an argument. But instead of executing the request out of the gate we want it only when we have a callback to handle the response.

Our generator would be something like:

function* list() { const userGet = `//api.github.com/search/users?page=1&q=daspinola&type=Users` const users = yield request(userGet) yield for (let i = 0; i<=users.length; i++) { yield request(users[i].repos_url) } }

It will:

  • Wait until the first request is prepared
  • Return a function reference expecting a callback for the first request

    Our request function accepts a url

    and returns a function that expects a callback

  • Expect a users to be sent in the next .next
  • Iterate over users
  • Wait for a .next for each of the users
  • Return their respective callback function

So an execution of this would be:

try { const iterator = list() iterator.next().value(function handleUsersList(err, users) { if (err) throw err const list = JSON.parse(users).items // send the list of users for the iterator iterator.next(list) list.forEach(function(user) { iterator.next().value(function userRepos(error, repos) { if (error) throw repos // Handle each individual user repo here console.log(user, JSON.parse(repos)) }) }) }) } catch (e) { console.error(e) }

We could separate the callback functions like we did previously. You get the deal by now, a takeaway is that we now can handle each individual user repository list individually.

I have mixed felling about generators. On one hand I can get a grasp of what is expected of the code by looking at the generator.

But its execution ends up having similar problems to the callback hell.

Like async/await, a compiler is recommended. This is because it isn’t supported in older browser versions.

Also it isn’t that common in my experience. So it may generate confusing in codebases maintained by various developers.

An awesome insight of how generators work can be found in this article. And here is another great resource.

Async/Await

This method seems like a mix of generators with promises. You just have to tell your code what functions are to be async. And what part of the code will have to await for that promise to finish.

sumTwentyAfterTwoSeconds(10) .then(result => console.log('after 2 seconds', result)) async function sumTwentyAfterTwoSeconds(value) { const remainder = afterTwoSeconds(20) return value + await remainder } function afterTwoSeconds(value) { return new Promise(resolve => { setTimeout(() => { resolve(value) }, 2000); }); }

In this scenario:

  • We have sumTwentyAfterTwoSeconds as being an async function
  • We tell our code to wait for the resolve or reject for our promise function afterTwoSeconds
  • It will only end up in the .then when the await operations finish

    In this case there is only one

Applying this to our request we leave it as a promise as seen earlier:

function request(url) { return new Promise(function(resolve, reject) { const xhr = new XMLHttpRequest(); xhr.onreadystatechange = function(e) { if (xhr.readyState === 4) { if (xhr.status === 200) { resolve(xhr.response) } else { reject(xhr.status) } } } xhr.ontimeout = function () { reject('timeout') } xhr.open('get', url, true) xhr.send() }) }

We create our async function with the needed awaits like so:

async function list() { const userGet = `//api.github.com/search/users?page=1&q=daspinola&type=Users` const users = await request(userGet) const usersList = JSON.parse(users).items usersList.forEach(async function (user) { const repos = await request(user.repos_url) handleRepoList(user, repos) }) } function handleRepoList(user, repos) { const userRepos = JSON.parse(repos) // Handle each individual user repo here console.log(user, userRepos) }

So now we have an async list function that will handle the requests. Another async is needed in the forEach so that we have the list of repos for each user to manipulate.

We call it as:

list() .catch(e => console.error(e))

This and the promises approach are my favorites since the code is easy to read and change. You can read about async/await more in depth here.

A downside of using async/await is that it isn’t supported in the front-end by older browsers or in the back-end. You have to use the Node 8.

You can use a compiler like babel to help solve that.

“Solution”

You can see the end code accomplishing our initial goal using async/await in this snippet.

A good thing to do is to try it yourself in the various forms referenced in this article.

Conclusion

Depending on the scenario you might find yourself using:

  • async/await
  • callbacks
  • mix

It’s up to you what fits your purposes. And what lets you maintain the code so that it is understandable to others and your future self.

Note: Any of the approaches become slightly less verbose when using the alternatives for requests like $.ajax and fetch.

Let me know what you would do different and different ways you found to make each approach more readable.

This is Article 11 of 30. It is part of a project for publishing an article at least once a week, from idle thoughts to tutorials. Leave a comment, follow me on Diogo Spínola and then go back to your brilliant project!