Rocket/site/guide/3-overview.md

11 KiB

Overview

Rocket provides primitives to build web servers and applications with Rust: Rocket provides routing, pre-processing of requests, and post-processing of responses; the rest is up to you. Your application code instructs Rocket on what to pre-process and post-process and fills the gaps between pre-processing and post-processing.

Lifecycle

Rocket's main task is to listen for incoming web requests, dispatch the request to the application code, and return a response to the client. We call the process that goes from request to response the "lifecycle". We summarize the lifecycle as the following sequence of steps:

  1. Routing

    Rocket parses an incoming HTTP request into native structures that your code operates on indirectly. Rocket determines which request handler to invoke by matching against route attributes declared in your application.

  2. Validation

    Rocket validates the incoming request against types and guards present in the matched route. If validation fails, Rocket forwards the request to the next matching route or calls an error handler.

  3. Processing

    The request handler associated with the route is invoked with validated arguments. This is the main business logic of an application. Processing completes by returning a Response.

  4. Response

    The returned Response is processed. Rocket generates the appropriate HTTP response and sends it to the client. This completes the lifecycle. Rocket continues listening for requests, restarting the lifecycle for each incoming request.

The remainder of this section details the routing phase as well as additional components needed for Rocket to begin dispatching requests to request handlers. The sections following describe the request and response phases as well as other components of Rocket.

Routing

Rocket applications are centered around routes and handlers. A route is a combination of:

  • A set of parameters to match an incoming request against.
  • A handler to process the request and return a response.

A handler is simply a function that takes an arbitrary number of arguments and returns any arbitrary type.

The parameters to match against include static paths, dynamic paths, path segments, forms, query strings, request format specifiers, and body data. Rocket uses attributes, which look like function decorators in other languages, to make declaring routes easy. Routes are declared by annotating a function, the handler, with the set of parameters to match against. A complete route declaration looks like this:

# #[macro_use] extern crate rocket;

#[get("/world")]              // <- route attribute
fn world() -> &'static str {  // <- request handler
    "hello, world!"
}

This declares the world route to match against the static path "/world" on incoming GET requests. Instead of #[get], we could have used #[post] or #[put] for other HTTP methods, or #[catch] for serving custom error pages. Additionally, other route parameters may be necessary when building more interesting applications. The Requests chapter, which follows this one, has further details on routing and error handling.

! note: We prefer #[macro_use], but you may prefer explicit imports.

Throughout this guide and the majority of Rocket's documentation, we import rocket explicitly with #[macro_use] even though the Rust 2018 edition makes explicitly importing crates optional. However, explicitly importing with #[macro_use] imports macros globally, allowing you to use Rocket's macros anywhere in your application without importing them explicitly.

You may instead prefer to import macros explicitly or refer to them with absolute paths: use rocket::get; or #[rocket::get]. The hello_2018 example showcases this alternative.

Mounting

Before Rocket can dispatch requests to a route, the route needs to be mounted:

# #[macro_use] extern crate rocket;

# #[get("/world")]
# fn world() -> &'static str {
#     "hello, world!"
# }

rocket::ignite().mount("/hello", routes![world]);

The mount method takes as input:

  1. A base path to namespace a list of routes under, here, /hello.
  2. A list of routes via the routes! macro: here, routes![world], with multiple routes: routes![a, b, c].

This creates a new Rocket instance via the ignite function and mounts the world route to the /hello base path, making Rocket aware of the route. GET requests to /hello/world will be directed to the world function.

The mount method, like all other builder methods on Rocket, can be chained any number of times, and routes can be reused by mount points:

# #[macro_use] extern crate rocket;

# #[get("/world")]
# fn world() -> &'static str {
#     "hello, world!"
# }

rocket::ignite()
    .mount("/hello", routes![world])
    .mount("/hi", routes![world]);

By mounting world to both /hello and /hi, requests to "/hello/world" and "/hi/world" will be directed to the world function.

! note: In many cases, the base path will simply be "/".

Launching

Rocket begins serving requests after being launched, which starts a multi-threaded asynchronous server and dispatches requests to matching routes as they arrive.

There are two mechnisms by which a Rocket can be launched. The first and preferred approach is via the #[launch] route attribute, which generates a main function that sets up an async runtime and starts the server. With #[launch], our complete Hello, world! application looks like:

#[macro_use] extern crate rocket;

#[get("/world")]
fn world() -> &'static str {
    "Hello, world!"
}

#[launch]
fn rocket() -> rocket::Rocket {
    rocket::ignite().mount("/hello", routes![world])
}

Running the application, the console shows:

> cargo run
🔧 Configured for debug.
    => address: 127.0.0.1
    => port: 8000
    => workers: 64
    => log level: normal
    => secret key: [zero]
    => limits: forms = 32KiB
    => cli colors: true
    => keep-alive: 5s
    => tls: disabled
🛰  Mounting /hello:
    => GET /hello/world (world)
🚀 Rocket has launched from http://127.0.0.1:8000

! tip: You can also return _ from a #[launch] function!

If you find it more pleasing, #[launch] can infer the return type of Rocket for you by using _ as the return type:

#[launch] fn rocket() -> _ { /* ... */ }

If we visit http://127.0.0.1:8000/hello/world, we see Hello, world!, exactly as we expected.

! note: This and other examples are on GitHub.

A version of this example's complete crate, ready to cargo run, can be found on GitHub. You can find dozens of other complete examples, spanning all of Rocket's features, in the GitHub examples directory.

The second approach uses the #[rocket::main] route attribute. #[rocket::main] also generates a main function that sets up an async runtime but unlike #[launch], allows you to start the server:

# #[macro_use] extern crate rocket;
#
# #[get("/world")]
# fn world() -> &'static str {
#     "Hello, world!"
# }

#[rocket::main]
async fn main() {
    rocket::ignite()
        .mount("/hello", routes![world])
        .launch()
        .await;
}

#[rocket::main] is useful when a handle to the Future returned by launch() is desired, or when the return value of launch() is to be inspected. The errors example for instance, inspects the return value.

Futures and Async

Rocket uses Rust Futures for concurrency. Asynchronous programming with Futures and async/await allows route handlers to perform wait-heavy I/O such as filesystem and network access while still allowing other requests to be make progress. For an overview of Rust Futures, see Asynchronous Programming in Rust.

In general, you should prefer to use async-ready libraries instead of synchronous equivalents inside Rocket applications.

async appears in several places in Rocket:

  • Routes and Error Catchers can be async fns. Inside an async fn, you can .await Futures from Rocket or other libraries.
  • Several of Rocket's traits, such as FromData and FromRequest, have methods that return Futures.
  • Data and DataStream, incoming request data, and Response and Body, outgoing response data, are based on tokio::io::AsyncRead instead of std::io::Read.

You can find async-ready libraries on crates.io with the async tag.

! note

Rocket master uses the tokio runtime. The runtime is started for you if you use #[launch] or #[rocket::main], but you can still launch() a Rocket instance on a custom-built runtime by not using either attribute.

Async Routes

Rocket makes it easy to use async/await in routes.

# #[macro_use] extern crate rocket;
use rocket::tokio::time::{sleep, Duration};

#[get("/delay/<seconds>")]
async fn delay(seconds: u64) -> String {
    sleep(Duration::from_secs(seconds)).await;
    format!("Waited for {} seconds", seconds)
}

First, notice that the route function is an async fn. This enables the use of await inside the handler. sleep is an asynchronous function, so we must await it.

Multitasking

Rust's Futures are a form of cooperative multitasking. In general, Futures and async fns should only .await on operations and never block. Some common examples of blocking include locking non-async mutexes, joining threads, or using non-async library functions (including those in std) that perform I/O.

If a Future or async fn blocks the thread, inefficient resource usage, stalls, or sometimes even deadlocks can occur.

Sometimes there is no good async alternative for a library or operation. If necessary, you can convert a synchronous operation to an async one with tokio::task::spawn_blocking:

# #[macro_use] extern crate rocket;
use std::io;
use rocket::tokio::task::spawn_blocking;
use rocket::response::Debug;

#[get("/blocking_task")]
async fn blocking_task() -> Result<Vec<u8>, Debug<io::Error>> {
    // In a real app, use rocket::response::NamedFile or tokio::fs::File.
    let vec = spawn_blocking(|| std::fs::read("data.txt")).await
        .map_err(|e| io::Error::new(io::ErrorKind::Interrupted, e))??;

    Ok(vec)
}