tokio

Module process

source
Available on crate feature process only.
Expand description

An implementation of asynchronous process management for Tokio.

This module provides a Command struct that imitates the interface of the std::process::Command type in the standard library, but provides asynchronous versions of functions that create processes. These functions (spawn, status, output and their variants) return “future aware” types that interoperate with Tokio. The asynchronous process support is provided through signal handling on Unix and system APIs on Windows.

§Examples

Here’s an example program which will spawn echo hello world and then wait for it complete.

use tokio::process::Command;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    // The usage is similar as with the standard library's `Command` type
    let mut child = Command::new("echo")
        .arg("hello")
        .arg("world")
        .spawn()
        .expect("failed to spawn");

    // Await until the command completes
    let status = child.wait().await?;
    println!("the command exited with: {}", status);
    Ok(())
}

Next, let’s take a look at an example where we not only spawn echo hello world but we also capture its output.

use tokio::process::Command;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Like above, but use `output` which returns a future instead of
    // immediately returning the `Child`.
    let output = Command::new("echo").arg("hello").arg("world")
                        .output();

    let output = output.await?;

    assert!(output.status.success());
    assert_eq!(output.stdout, b"hello world\n");
    Ok(())
}

We can also read input line by line.

use tokio::io::{BufReader, AsyncBufReadExt};
use tokio::process::Command;

use std::process::Stdio;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut cmd = Command::new("cat");

    // Specify that we want the command's standard output piped back to us.
    // By default, standard input/output/error will be inherited from the
    // current process (for example, this means that standard input will
    // come from the keyboard and standard output/error will go directly to
    // the terminal if this process is invoked from the command line).
    cmd.stdout(Stdio::piped());

    let mut child = cmd.spawn()
        .expect("failed to spawn command");

    let stdout = child.stdout.take()
        .expect("child did not have a handle to stdout");

    let mut reader = BufReader::new(stdout).lines();

    // Ensure the child process is spawned in the runtime so it can
    // make progress on its own while we await for any output.
    tokio::spawn(async move {
        let status = child.wait().await
            .expect("child process encountered an error");

        println!("child status was: {}", status);
    });

    while let Some(line) = reader.next_line().await? {
        println!("Line: {}", line);
    }

    Ok(())
}

Here is another example using sort writing into the child process standard input, capturing the output of the sorted text.

use tokio::io::AsyncWriteExt;
use tokio::process::Command;

use std::process::Stdio;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut cmd = Command::new("sort");

    // Specifying that we want pipe both the output and the input.
    // Similarly to capturing the output, by configuring the pipe
    // to stdin it can now be used as an asynchronous writer.
    cmd.stdout(Stdio::piped());
    cmd.stdin(Stdio::piped());

    let mut child = cmd.spawn().expect("failed to spawn command");

    // These are the animals we want to sort
    let animals: &[&str] = &["dog", "bird", "frog", "cat", "fish"];

    let mut stdin = child
        .stdin
        .take()
        .expect("child did not have a handle to stdin");

    // Write our animals to the child process
    // Note that the behavior of `sort` is to buffer _all input_ before writing any output.
    // In the general sense, it is recommended to write to the child in a separate task as
    // awaiting its exit (or output) to avoid deadlocks (for example, the child tries to write
    // some output but gets stuck waiting on the parent to read from it, meanwhile the parent
    // is stuck waiting to write its input completely before reading the output).
    stdin
        .write(animals.join("\n").as_bytes())
        .await
        .expect("could not write to stdin");

    // We drop the handle here which signals EOF to the child process.
    // This tells the child process that it there is no more data on the pipe.
    drop(stdin);

    let op = child.wait_with_output().await?;

    // Results should come back in sorted order
    assert_eq!(op.stdout, "bird\ncat\ndog\nfish\nfrog\n".as_bytes());

    Ok(())
}

With some coordination, we can also pipe the output of one command into another.

use tokio::join;
use tokio::process::Command;
use std::process::Stdio;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut echo = Command::new("echo")
        .arg("hello world!")
        .stdout(Stdio::piped())
        .spawn()
        .expect("failed to spawn echo");

    let tr_stdin: Stdio = echo
        .stdout
        .take()
        .unwrap()
        .try_into()
        .expect("failed to convert to Stdio");

    let tr = Command::new("tr")
        .arg("a-z")
        .arg("A-Z")
        .stdin(tr_stdin)
        .stdout(Stdio::piped())
        .spawn()
        .expect("failed to spawn tr");

    let (echo_result, tr_output) = join!(echo.wait(), tr.wait_with_output());

    assert!(echo_result.unwrap().success());

    let tr_output = tr_output.expect("failed to await tr");
    assert!(tr_output.status.success());

    assert_eq!(tr_output.stdout, b"HELLO WORLD!\n");

    Ok(())
}

§Caveats

§Dropping/Cancellation

Similar to the behavior to the standard library, and unlike the futures paradigm of dropping-implies-cancellation, a spawned process will, by default, continue to execute even after the Child handle has been dropped.

The Command::kill_on_drop method can be used to modify this behavior and kill the child process if the Child wrapper is dropped before it has exited.

§Unix Processes

On Unix platforms processes must be “reaped” by their parent process after they have exited in order to release all OS resources. A child process which has exited, but has not yet been reaped by its parent is considered a “zombie” process. Such processes continue to count against limits imposed by the system, and having too many zombie processes present can prevent additional processes from being spawned.

The tokio runtime will, on a best-effort basis, attempt to reap and clean up any process which it has spawned. No additional guarantees are made with regard to how quickly or how often this procedure will take place.

It is recommended to avoid dropping a Child process handle before it has been fully awaited if stricter cleanup guarantees are required.

Structs§

  • Representation of a child process spawned onto an event loop.
  • The standard error stream for spawned children.
  • The standard input stream for spawned children.
  • The standard output stream for spawned children.
  • This structure mimics the API of std::process::Command found in the standard library, but replaces functions that create a process with an asynchronous variant. The main provided asynchronous functions are spawn, status, and output.