Crate futures_concurrency

source ·
Expand description

Performant, portable, structured concurrency operations for async Rust. It works with any runtime, does not erase lifetimes, always handles cancellation, and always returns output to the caller.

futures-concurrency provides concurrency operations for both groups of futures and streams. Both for bounded and unbounded sets of futures and streams. In both cases performance should be on par with, if not exceed conventional executor implementations.

§Examples

Await multiple futures of different types

use futures_concurrency::prelude::*;
use std::future;

let a = future::ready(1u8);
let b = future::ready("hello");
let c = future::ready(3u16);
assert_eq!((a, b, c).join().await, (1, "hello", 3));

Concurrently process items in a collection

use futures_concurrency::prelude::*;

let v: Vec<_> = vec!["chashu", "nori"]
    .into_co_stream()
    .map(|msg| async move { format!("hello {msg}") })
    .collect()
    .await;

assert_eq!(v, &["hello chashu", "hello nori"]);

Access stack data outside the futures’ scope

Adapted from std::thread::scope.

use futures_concurrency::prelude::*;

let mut container = vec![1, 2, 3];
let mut num = 0;

let a = async {
    println!("hello from the first future");
    dbg!(&container);
};

let b = async {
    println!("hello from the second future");
    num += container[0] + container[2];
};

println!("hello from the main future");
let _ = (a, b).join().await;
container.push(4);
assert_eq!(num, container.len());

§Operations

§Futures

For futures which return a regular type T only the join and race operations are available. join waits for all futures to complete, while race will wait for the first future to complete. However for futures which return a Try<Output = T> two additional operations are available. The following table describes the behavior of concurrency operations for fallible futures:

Wait for all outputsWait for first output
Continue on errorFuture::joinFuture::race_ok
Short-circuit on errorFuture::try_joinFuture::race

The following futures implementations are provided by futures-concurrency:

§Streams

Streams yield outputs one-by-one, which means that deciding to stop iterating is the same for fallible and infallible streams. The operations provided for streams can be categorized based on whether their inputs can be concurrently evaluated, and whether their outputs can be concurrently processed.

Specifically in the case of merge, it takes N streams in, and yields items one-by-one as soon as any are available. This enables the output of individual streams to be concurrently processed by further operations later on.

Sequential output processingConcurrent output processing
Sequential input evaluationStream::chainnot yet available
Concurrent input evaluationStream::zipStream::merge

‡: This could be addressed by a hypothetical Stream::unzip operation, however because we aspire for semantic compatibility with std::iter::Iterator in our operations, the path to adding it is currently unclear.

The following streams implementations are provided by futures-concurrency:

§Runtime Support

futures-concurrency does not depend on any runtime executor being present. This enables it to work out of the box with any async runtime, including: tokio, async-std, smol, glommio, and monoio. It also supports #[no_std] environments, allowing it to be used with embedded async runtimes such as embassy.

§Feature Flags

The std feature flag is enabled by default. To target alloc or no_std environments, you can enable the following configuration:

[dependencies]
# no_std
futures-concurrency = { version = "7.5.0", default-features = false }

# alloc
futures-concurrency = { version = "7.5.0", default-features = false, features = ["alloc"] }

§Further Reading

futures-concurrency has been developed over the span of several years. It is primarily maintained by Yosh Wuyts, a member of the Rust Async WG. You can read more about the development and ideas behind futures-concurrency here:

Modules§

  • Helper functions and types for fixed-length arrays.
  • Concurrent execution of streams
  • Asynchronous basic functionality.
  • The futures concurrency prelude.
  • Composable asynchronous iteration.
  • Parallel iterator types for vectors (Vec<T>)