Struct tokio::sync::mpsc::Sender

pub struct Sender<T> { /* fields omitted */ }

This is supported on feature="sync" only.

Send values to the associated Receiver.

Instances are created by the channel function.

Implementations

impl<T> Sender<T>

pub async fn send<'_>(&'_ mut self, value: T) -> Result<(), SendError<T>>

This is supported on feature="sync" only.

Sends a value, waiting until there is capacity.

A successful send occurs when it is determined that the other end of the channel has not hung up already. An unsuccessful send would be one where the corresponding receiver has already been closed. Note that a return value of Err means that the data will never be received, but a return value of Ok does not mean that the data will be received. It is possible for the corresponding receiver to hang up immediately after this function returns Ok.

Errors

If the receive half of the channel is closed, either due to close being called or the Receiver handle dropping, the function returns an error. The error includes the value passed to send.

Examples

In the following example, each call to send will block until the previously sent value was received.

use tokio::sync::mpsc;

#[tokio::main]
async fn main() {
    let (mut tx, mut rx) = mpsc::channel(1);

    tokio::spawn(async move {
        for i in 0..10 {
            if let Err(_) = tx.send(i).await {
                println!("receiver dropped");
                return;
            }
        }
    });

    while let Some(i) = rx.recv().await {
        println!("got = {}", i);
    }
}

pub fn try_send(&mut self, message: T) -> Result<(), TrySendError<T>>

This is supported on feature="sync" only.

Attempts to immediately send a message on this Sender

This method differs from send by returning immediately if the channel's buffer is full or no receiver is waiting to acquire some data. Compared with send, this function has two failure cases instead of one (one for disconnection, one for a full buffer).

This function may be paired with poll_ready in order to wait for channel capacity before trying to send a value.

Errors

If the channel capacity has been reached, i.e., the channel has n buffered values where n is the argument passed to channel, then an error is returned.

If the receive half of the channel is closed, either due to close being called or the Receiver handle dropping, the function returns an error. The error includes the value passed to send.

Examples

use tokio::sync::mpsc;

#[tokio::main]
async fn main() {
    // Create a channel with buffer size 1
    let (mut tx1, mut rx) = mpsc::channel(1);
    let mut tx2 = tx1.clone();

    tokio::spawn(async move {
        tx1.send(1).await.unwrap();
        tx1.send(2).await.unwrap();
        // task waits until the receiver receives a value.
    });

    tokio::spawn(async move {
        // This will return an error and send
        // no message if the buffer is full
        let _ = tx2.try_send(3);
    });

    let mut msg;
    msg = rx.recv().await.unwrap();
    println!("message {} received", msg);

    msg = rx.recv().await.unwrap();
    println!("message {} received", msg);

    // Third message may have never been sent
    match rx.recv().await {
        Some(msg) => println!("message {} received", msg),
        None => println!("the third message was never sent"),
    }
}

pub async fn send_timeout<'_>(
    &'_ mut self,
    value: T,
    timeout: Duration
) -> Result<(), SendTimeoutError<T>>

This is supported on feature="sync" and feature="time" only.

Sends a value, waiting until there is capacity, but only for a limited time.

Shares the same success and error conditions as send, adding one more condition for an unsuccessful send, which is when the provided timeout has elapsed, and there is no capacity available.

Errors

If the receive half of the channel is closed, either due to close being called or the Receiver having been dropped, the function returns an error. The error includes the value passed to send.

Examples

In the following example, each call to send_timeout will block until the previously sent value was received, unless the timeout has elapsed.

use tokio::sync::mpsc;
use tokio::time::{delay_for, Duration};

#[tokio::main]
async fn main() {
    let (mut tx, mut rx) = mpsc::channel(1);

    tokio::spawn(async move {
        for i in 0..10 {
            if let Err(e) = tx.send_timeout(i, Duration::from_millis(100)).await {
                println!("send error: #{:?}", e);
                return;
            }
        }
    });

    while let Some(i) = rx.recv().await {
        println!("got = {}", i);
        delay_for(Duration::from_millis(200)).await;
    }
}

pub fn poll_ready(&mut self, cx: &mut Context) -> Poll<Result<(), ClosedError>>

This is supported on feature="sync" only.

Returns Poll::Ready(Ok(())) when the channel is able to accept another item.

If the channel is full, then Poll::Pending is returned and the task is notified when a slot becomes available.

Once poll_ready returns Poll::Ready(Ok(())), a call to try_send will succeed unless the channel has since been closed. To provide this guarantee, the channel reserves one slot in the channel for the coming send. This reserved slot is not available to other Sender instances, so you need to be careful to not end up with deadlocks by blocking after calling poll_ready but before sending an element.

If, after poll_ready succeeds, you decide you do not wish to send an item after all, you can use disarm to release the reserved slot.

Until an item is sent or disarm is called, repeated calls to poll_ready will return either Poll::Ready(Ok(())) or Poll::Ready(Err(_)) if channel is closed.

pub fn disarm(&mut self) -> bool

This is supported on feature="sync" only.

Undo a successful call to poll_ready.

Once a call to poll_ready returns Poll::Ready(Ok(())), it holds up one slot in the channel to make room for the coming send. disarm allows you to give up that slot if you decide you do not wish to send an item after all. After calling disarm, you must call poll_ready until it returns Poll::Ready(Ok(())) before attempting to send again.

Returns false if no slot is reserved for this sender (usually because poll_ready was not previously called, or did not succeed).

Motivation

Since poll_ready takes up one of the finite number of slots in a bounded channel, callers need to send an item shortly after poll_ready succeeds. If they do not, idle senders may take up all the slots of the channel, and prevent active senders from getting any requests through. Consider this code that forwards from one channel to another:

loop {
  ready!(tx.poll_ready(cx))?;
  if let Some(item) = ready!(rx.poll_recv(cx)) {
    tx.try_send(item)?;
  } else {
    break;
  }
}

If many such forwarders exist, and they all forward into a single (cloned) Sender, then any number of forwarders may be waiting for rx.poll_recv at the same time. While they do, they are effectively each reducing the channel's capacity by 1. If enough of these forwarders are idle, forwarders whose rx do have elements will be unable to find a spot for them through poll_ready, and the system will deadlock.

disarm solves this problem by allowing you to give up the reserved slot if you find that you have to block. We can then fix the code above by writing:

loop {
  ready!(tx.poll_ready(cx))?;
  let item = rx.poll_recv(cx);
  if let Poll::Ready(Ok(_)) = item {
    // we're going to send the item below, so don't disarm
  } else {
    // give up our send slot, we won't need it for a while
    tx.disarm();
  }
  if let Some(item) = ready!(item) {
    tx.try_send(item)?;
  } else {
    break;
  }
}

Trait Implementations

impl<T> Clone for Sender<T>

fn clone(&self) -> Self

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<T> Debug for Sender<T>

fn fmt(&self, fmt: &mut Formatter) -> Result

Formats the value using the given formatter.