I am trying to make a struct in Rust using tokio::select! which allows me to send and receive message from different parties.
pub Destination {
Alice,
Bob
}
struct ConnectionsManager {
rx_from_alice: UnboundedReceiver<Message>,
tx_to_alice: UnboundedSender<Message>,
rx_from_bob: UnboundedReceiver<Message>,
tx_to_bob: UnboundedSender<Message>,
}
impl ConnectionsManager {
async fn listen(&mut self) -> Result<Message, anyhow::Error> {
select! {
received_message = self.rx_from_alice.next() => {
return self.process_received_message(received_message).await
}
received_message = self.rx_from_bob.next() => {
return self.process_received_message(received_message).await
}
}
}
async fn send(&mut self, message: Message, dest: Destination) -> Result<(), anyhow::Error> {
match message {
Destination::Alice => self.tx_to_alice.send(message).await,
Destination::Bob => self.tx_to_bob.send(message).await,
}
}
async fn process_received_message(&mut self, received_message: Option<Message>) -> Result<Message, anyhow::Error> {
// Some further processing of the message
}
}
The Stream implementation of mpsc channels guarantees that there will be no lost messages,so it looks fine so far.
The problem is I need to also use a select! in the code that calls the ConnectionsManager::listen():
async fn listen_to_a_lot_of_stuff(&mut self) -> Result<Message, anyhow::Error> {
let conections_manager = ConnectionsManager::new(/* setting up...*/);
loop {
select! {
processed_message = self.connections_manager.listen() => {
// .......
}
some_other_result = some_other_future => {
// .......
}
}
}
}
In that case, when some_other_future resolves while ConnectionsManager::listen() already received message from a channel and calls ConectionsManager::process_received_message()` on it, the message may be lost due to cancelling of the future.
Is there any pattern that allows to not lose any message in that case?
Some ideas that have failed so far:
Saving connections_manager.listen() future or making it a task does not work, because I need to use ConnectionsManager tx channels while it is not awaited by select!. Storing the listen() Future makes calling ConnectionsManager::send() not possible in that case.
Giving TX channels to Bob and Alice, and have just one RX channel in ConnectionsManager could work, but ConnectionsManager struct needs to know whether either Bob's or Alice's TX channel was closed, and it will not receive that information (as None in RX) when one of them closes the TX channel, but the other one is still open.
Related
I have two async functions: get_message and get_event. I'd like to perform an action whenever a message arrives or an event comes and do that forever in an infinite loop.
The simplified setup looks like this:
use futures::{future::select, future::Either, pin_mut};
impl MsgReceiver {
async fn get_message(&mut self) -> Message { /* ... */ }
}
impl EventListener {
async fn get_event(&mut self) -> Event { /* ... */ }
}
async fn eternal_task(receiver: MsgReceiver, listener: EventListener) -> ! {
let get_msg_fut = receiver.get_message();
pin_mut!(get_msg_fut);
loop {
let get_event_fut = listener.get_event();
pin_mut!(get_event_fut);
match select(get_event_fut, get_msg_fut).await {
Either::Left((ev, r_get_msg_fut)) => {
/* react to the event */
// r_get_msg_fut is not done, how to reuse it in the next iteration?
}
Either::Right((msg, r_get_event_fut)) => {
/* react to the message */
// it's fine to drop get_event_fut here
// the following line causes a double-mut-borrow error on receiver,
// despite receiver isn't borrowed anymore (the old future is completed and dropped)
let new_future = receiver.get_message();
}
};
}
}
I have three major questions here:
When an event comes first, how to tell rust that I want to reuse the incomplete get_message future on the next loop iteration?
When a message comes first, how to construct a new future without a borrow error?
When (2) is solved, how to put the new future into the same pinned memory location and use it on the next loop iteration?
I had success using this, but could not get rid of the Box::pin
use futures::{future::select, future::Either, pin_mut};
use std::sync::Mutex;
#[derive(Debug)]
struct MsgReceiver;
#[derive(Debug)]
struct EventListener;
#[derive(Debug)]
struct Message;
#[derive(Debug)]
struct Event;
impl MsgReceiver {
async fn get_message(&mut self) -> Message {
Message
}
}
impl EventListener {
async fn get_event(&mut self) -> Event {
Event
}
}
async fn eternal_task(receiver: MsgReceiver, mut listener: EventListener) -> ! {
let receiver = Mutex::new(receiver);
let mut f = None;
loop {
let get_msg_fut = match f.take() {
None => {
let mut l = receiver.lock();
Box::pin(async move {
l.get_message().await
})
}
Some(f) => f,
};
let get_event_fut = listener.get_event();
pin_mut!(get_event_fut);
match select(get_event_fut, get_msg_fut).await {
Either::Left((ev, r_get_msg_fut)) => {
/* react to the event */
// store the future for next iteration
f = Some(r_get_msg_fut);
}
Either::Right((msg, r_get_event_fut)) => {
/* react to the message */
}
};
}
}
#[tokio::main]
async fn main() {
eternal_task(MsgReceiver, EventListener).await;
}
I think this is tricky to get right, even with unsafe which would probably be needed to accomplish this. Persisting and reusing the same variables isn't too hard, its actually #2 that's the hardest (at least with the current borrow checker).
I found a solution that totally circumvents the problem by using the async-stream crate to provide an intermediary:
async fn eternal_task(mut receiver: MsgReceiver, mut listener: EventListener) -> ! {
let combined = futures::stream::select(
stream! { loop { yield Either::Left(receiver.get_message().await); } },
stream! { loop { yield Either::Right(listener.get_event().await); } },
);
pin_mut!(combined);
while let Some(msg_or_evt) = combined.next().await {
match msg_or_evt {
Either::Left(msg) => {
// do something with msg
}
Either::Right(evt) => {
// do something with evt
}
};
}
unreachable!()
}
It uses the stream! macro to generate a type that continuously calls and yields values from .get_message() and .get_event(). It then uses futures::stream::select and Either to combine them. And then its just a matter of looping over the results. It works in #![no_std].
For threaded applications, the Rust standard library provides std::sync::mpsc::sync_channel, a buffered channel which blocks on the reading end when the buffer is empty and blocks on the writing end when the buffer is full. In particular, if you set the buffer size to 0, then any read or write will block until there is a matching write or read.
For async code, there is futures::channel::mpsc::channel, but this does not have the same behavior. Here the minimum capacity is the number of senders on the channel, which is greater than 0. The sending end can still block (because it implements Sink, so we can use SinkExt::send and await it), but only after there's at least one thing already in the buffer.
I took a look to see if there were any packages that provide the functionality I'm looking for, but I could not find anything. Tokio provides lots of nice async synchronization primitives, but none of them did quite what I'm looking for. Plus, my program is going to run in the browser, so I don't think I'm able to use a runtime like Tokio. Does anyone know of a package that fits my use case? I would try to implement this myself, since this almost feels like the most minimal use case for the Sink and Stream traits, but even a minimal implementation of these traits seems like it would be really complicated. Thoughts?
Edit: here's a minimal example of what I mean:
fn main() {
let (tx, rx) = blocking_channel();
let ft = async move {
tx.send(3).await;
println!("message sent and received");
}
let fr = async move {
let msg = rx.recv().await;
println!("received {}", msg);
}
block_on(async { join!(ft, fr) });
}
In this example, whichever future runs first will yield to the other, and only print after both rx.recv and tx.send have been called. Obviously, the receiving end can only progress after tx.send has been called, but I want the less obvious behavior of the transmitting end also having to wait.
Interesting question. I don't think something like that already exists.
Here is a quick proof-of-concept prototype I wrote for that. It's not the prettiest, but it seems to work. There might be a better struct layout than just wrapping everything in a RefCell<Option<...>>, though. And I don't particularly like the sender_dropped and receiver_dropped variables.
Be sure to unittest it properly if used in production!!!
extern crate alloc;
use alloc::rc::Rc;
use core::cell::RefCell;
use core::pin::Pin;
use core::task::{Poll, Waker};
use futures::SinkExt;
use futures::StreamExt;
struct Pipe<T> {
send_waker: RefCell<Option<Waker>>,
receive_waker: RefCell<Option<Waker>>,
value: RefCell<Option<T>>,
sender_dropped: RefCell<bool>,
receiver_dropped: RefCell<bool>,
}
impl<T> Pipe<T> {
fn new() -> Rc<Pipe<T>> {
Rc::new(Self {
value: RefCell::new(None),
send_waker: RefCell::new(None),
receive_waker: RefCell::new(None),
sender_dropped: RefCell::new(false),
receiver_dropped: RefCell::new(false),
})
}
}
impl<T> Pipe<T> {
fn wake_sender(&self) {
if let Some(waker) = self.send_waker.replace(None) {
waker.wake();
}
}
fn wake_receiver(&self) {
if let Some(waker) = self.receive_waker.replace(None) {
waker.wake();
}
}
}
pub struct PipeSender<T> {
pipe: Rc<Pipe<T>>,
}
pub struct PipeReceiver<T> {
pipe: Rc<Pipe<T>>,
}
pub fn create_pipe<T>() -> (PipeSender<T>, PipeReceiver<T>) {
let pipe = Pipe::new();
(PipeSender { pipe: pipe.clone() }, PipeReceiver { pipe })
}
impl<T> futures::Sink<T> for PipeSender<T> {
type Error = ();
fn poll_ready(
self: Pin<&mut Self>,
cx: &mut std::task::Context<'_>,
) -> Poll<Result<(), Self::Error>> {
let result = if *self.pipe.receiver_dropped.borrow() {
Poll::Ready(Err(()))
} else if self.pipe.receive_waker.borrow().is_some() && self.pipe.value.borrow().is_none() {
Poll::Ready(Ok(()))
} else {
self.pipe.send_waker.replace(Some(cx.waker().clone()));
Poll::Pending
};
// Wake potential receiver
self.pipe.wake_receiver();
result
}
fn start_send(self: Pin<&mut Self>, item: T) -> Result<(), Self::Error> {
let prev = self.pipe.value.replace(Some(item));
assert!(prev.is_none(), "A value got lost in the pipe.");
Ok(())
}
fn poll_flush(
self: Pin<&mut Self>,
_: &mut futures::task::Context<'_>,
) -> Poll<Result<(), Self::Error>> {
// Noop, start_send already completes the send
Poll::Ready(Ok(()))
}
fn poll_close(
self: Pin<&mut Self>,
_: &mut std::task::Context<'_>,
) -> Poll<Result<(), Self::Error>> {
// Noop, start_send already completes the send
Poll::Ready(Ok(()))
}
}
impl<T> futures::Stream for PipeReceiver<T> {
type Item = T;
fn poll_next(
self: Pin<&mut Self>,
cx: &mut std::task::Context<'_>,
) -> std::task::Poll<Option<Self::Item>> {
let result = {
let value = self.pipe.value.replace(None);
if let Some(value) = value {
Poll::Ready(Some(value))
} else if *self.pipe.sender_dropped.borrow() {
Poll::Ready(None)
} else {
self.pipe.receive_waker.replace(Some(cx.waker().clone()));
Poll::Pending
}
};
// Wake potential sender
self.pipe.wake_sender();
result
}
}
impl<T> Drop for PipeSender<T> {
fn drop(&mut self) {
self.pipe.sender_dropped.replace(true);
self.pipe.wake_receiver();
}
}
impl<T> Drop for PipeReceiver<T> {
fn drop(&mut self) {
self.pipe.receiver_dropped.replace(true);
self.pipe.wake_sender();
}
}
#[tokio::main]
async fn main() {
use std::time::Duration;
let (mut sender, mut receiver) = create_pipe();
tokio::join!(
async move {
for i in 0..5u32 {
println!("Sending {i} ...");
if let Err(_) = sender.send(i).await {
println!("Stream closed.");
break;
}
println!("Sent {i}.");
}
println!("Sender closed.");
},
async move {
println!("Attempting to receive ...");
while let Some(val) = receiver.next().await {
println!("Received: {val}");
println!("\n=== Waiting ... ===\n");
tokio::time::sleep(Duration::from_secs(1)).await;
println!("Attempting to receive ...");
}
println!("Receiver closed.");
}
);
}
Sending 0 ...
Attempting to receive ...
Sent 0.
Sending 1 ...
Received: 0
=== Waiting ... ===
Attempting to receive ...
Sent 1.
Sending 2 ...
Received: 1
=== Waiting ... ===
Attempting to receive ...
Sent 2.
Sending 3 ...
Received: 2
=== Waiting ... ===
Attempting to receive ...
Sent 3.
Sending 4 ...
Received: 3
=== Waiting ... ===
Attempting to receive ...
Sent 4.
Sender closed.
Received: 4
=== Waiting ... ===
Attempting to receive ...
Receiver closed.
I want to send Events between the game client and server and I already got it working, but I do not know how to do it with bevy.
I am dependent to use tokios async TcpStream, because I have to be able to split the stream into a OwnedWriteHalf and OwnedReadhalf using stream.into_split().
My first idea was to just spawn a thread that handles the connection and then send the received events to a queue using mpsc::channel
Then I include this queue into a bevy resource using app.insert_resource(Queue) and pull events from it in the game loop.
the Queue:
use tokio::sync::mpsc;
pub enum Instruction {
Push(GameEvent),
Pull(mpsc::Sender<Option<GameEvent>>),
}
#[derive(Clone, Debug)]
pub struct Queue {
sender: mpsc::Sender<Instruction>,
}
impl Queue {
pub fn init() -> Self {
let (tx, rx) = mpsc::channel(1024);
init(rx);
Self{sender: tx}
}
pub async fn send(&self, event: GameEvent) {
self.sender.send(Instruction::Push(event)).await.unwrap();
}
pub async fn pull(&self) -> Option<GameEvent> {
println!("new pull");
let (tx, mut rx) = mpsc::channel(1);
self.sender.send(Instruction::Pull(tx)).await.unwrap();
rx.recv().await.unwrap()
}
}
fn init(mut rx: mpsc::Receiver<Instruction>) {
tokio::spawn(async move {
let mut queue: Vec<GameEvent> = Vec::new();
loop {
match rx.recv().await.unwrap() {
Instruction::Push(ev) => {
queue.push(ev);
}
Instruction::Pull(sender) => {
sender.send(queue.pop()).await.unwrap();
}
}
}
});
}
But because all this has to be async I have block the pull() function in the sync game loop.
I do this using the futures-lite crate:
fn event_pull(
communication: Res<Communication>
) {
let ev = future::block_on(communication.event_queue.pull());
println!("got event: {:?}", ev);
}
And this works fine, BUT after around 5 seconds the whole program just halts and does not receive any more events.
It seems like that future::block_on() does block indefinitely.
Having the main function, in which bevy::prelude::App gets built and run, to be the async tokio::main function might also be a problem here.
It would probably be best to wrap the async TcpStream initialisation and tokio::sync::mpsc::Sender and thus also Queue.pull into synchronous functions, but I do not know how to do this.
Can anyone help?
How to reproduce
The repo can be found here
Just compile both server and client and then run both in the same order.
I got it to work by just replacing every tokio::sync::mpsc with crossbeam::channel, which might be a problem, as it does block
and manually initializing the tokio runtime.
so the init code looks like this:
pub struct Communicator {
pub event_bridge: bridge::Bridge,
pub event_queue: event_queue::Queue,
_runtime: Runtime,
}
impl Communicator {
pub fn init(ip: &str) -> Self {
let rt = tokio::runtime::Builder::new_multi_thread()
.enable_io()
.build()
.unwrap();
let (bridge, queue, game_rx) = rt.block_on(async move {
let socket = TcpStream::connect(ip).await.unwrap();
let (read, write) = socket.into_split();
let reader = TcpReader::new(read);
let writer = TcpWriter::new(write);
let (bridge, tcp_rx, game_rx) = bridge::Bridge::init();
reader::init(bridge.clone(), reader);
writer::init(tcp_rx, writer);
let event_queue = event_queue::Queue::init();
return (bridge, event_queue, game_rx);
});
// game of game_rx events to queue for game loop
let eq_clone = queue.clone();
rt.spawn(async move {
loop {
let event = game_rx.recv().unwrap();
eq_clone.send(event);
}
});
Self {
event_bridge: bridge,
event_queue: queue,
_runtime: rt,
}
}
}
And main.rs looks like this:
fn main() {
let communicator = communication::Communicator::init("0.0.0.0:8000");
communicator.event_bridge.push_tcp(TcpEvent::Register{name: String::from("luca")});
App::new()
.insert_resource(communicator)
.add_system(event_pull)
.add_plugins(DefaultPlugins)
.run();
}
fn event_pull(
communication: Res<communication::Communicator>
) {
let ev = communication.event_queue.pull();
if let Some(ev) = ev {
println!("ev");
}
}
Perhaps there might be a better solution.
I want to print "Hello" once a second.
Quoting the doc:
Futures use a poll based model. The consumer of a future repeatedly calls the poll function. The future then attempts to complete. If the future is able to complete, it returns Async::Ready(value). If the future is unable to complete due to being blocked on an internal resource (such as a TCP socket), it returns Async::NotReady.
My poll function returns NotReady if Delays return is NotReady, but nothing is printed to stdout.
use futures::{Async, Future, Stream}; // 0.1.25
use std::time::{Duration, Instant};
use tokio::timer::Delay; // 0.1.15
struct SomeStream;
impl Stream for SomeStream {
type Item = String;
type Error = ();
fn poll(&mut self) -> Result<Async<Option<Self::Item>>, Self::Error> {
let when = Instant::now() + Duration::from_millis(1000);
let mut task = Delay::new(when).map_err(|e| eprintln!("{:?}", e));
match task.poll() {
Ok(Async::Ready(value)) => {}
Ok(Async::NotReady) => return Ok(Async::NotReady),
Err(err) => return Err(()),
}
Ok(Async::Ready(Some("Hello".to_string())))
}
}
fn main() {
let s = SomeStream;
let future = s
.for_each(|item| {
println!("{:?}", item);
Ok(())
})
.map_err(|e| {});
tokio::run(future);
}
The main issue here is that state management is missing. You are creating a new Delay future every time the stream is polled, rather than holding on to it until it's resolved.
This would lead to never seeing any items coming out of the stream, since these futures are only being polled once, likely yielding NotReady each time.
You need to keep track of the delay future in your type SomeStream. In this case, one can use an option, so as to also identify whether we need to create a new delay.
#[derive(Debug, Default)]
struct SomeStream {
delay: Option<Delay>,
}
The subsequent code for SomeStream::poll, with better error handling and more idiomatic constructs, would become something like this:
impl Stream for SomeStream {
type Item = String;
type Error = Box<dyn std::error::Error + Send + Sync>; // generic error
fn poll(&mut self) -> Result<Async<Option<Self::Item>>, Self::Error> {
let delay = self.delay.get_or_insert_with(|| {
let when = Instant::now() + Duration::from_millis(1000);
Delay::new(when)
});
match delay.poll() {
Ok(Async::Ready(value)) => {
self.delay = None;
Ok(Async::Ready(Some("Hello".to_string())))
},
Ok(Async::NotReady) => Ok(Async::NotReady),
Err(err) => Err(err.into()),
}
}
}
Or, even better, using the try_ready! macro, which makes the return of errors and NotReady signals with less boilerplate.
fn poll(&mut self) -> Result<Async<Option<Self::Item>>, Self::Error> {
let delay = self.delay.get_or_insert_with(|| {
let when = Instant::now() + Duration::from_millis(1000);
Delay::new(when)
});
try_ready!(delay.poll());
// tick!
self.delay = None;
Ok(Async::Ready(Some("Hello".to_string())))
}
(Playground)
impl A {
fn new() -> (A, std::sync::mpsc::Receiver<Data>) {
let (sender, receiver) = std::sync::mpsc::channel();
let objA = A { sender: sender, }; // A spawns threads, clones and uses sender etc
(objA, receiver)
}
}
impl B {
fn new() -> B {
let (objA, receiver) = A::new();
B {
a: objA,
join_handle: Some(std::thread::spwan(move || {
loop {
match receiver.recv() {
Ok(data) => /* Do Something, inform main thread etc */,
Err(_) => break,
}
}
})),
}
}
}
impl Drop for B {
fn drop(&mut self) {
// Want to do something like "sender.close()/receiver.close()" etc so that the following
// thread joins. But there is no such function. How do i break the following thread ?
self.join_handle().take().unwrap().join().unwrap();
}
}
Is there a way to cleanly exit under such a circumstance ? The thing is that when either receiver or sender is dropped the other sniffs this and gives an error. In case of receiver it will be woken up and will yield an error in which case i am breaking out of the infinite and blocking loop above. However how do i do that explicitly using this very property of channels, without resorting to other flags in conjunction with try_recv()etc., and cleanly exit my thread deterministically?
Why not sending a specific message to shut this thread? I do not know what is your data but most of the time it may be an enum and adding a enum variant like 'MyData::Shutdown' in your receive you can simply break out of the loop.
You can wrap the a field of your B type in an Option. This way in the Drop::drop method you can do drop(self.a.take()) which will replace the field with a None and drop the sender. This closes the channel and your thread can now be properly joined.
You can create a new channel and swap your actual sender out with the dummy-sender. Then you can drop your sender and therefor join the thread:
impl Drop for B {
fn drop(&mut self) {
let (s, _) = channel();
drop(replace(&mut self.a.sender, s));
self.join_handle.take().unwrap().join().unwrap();
}
}
Try it out in the playpen: http://is.gd/y7A9L0
I don't know what the overhead of creating and immediately dropping a channel is, but it's not free and unlikely to be optimized out (There's an Arc in there).
on a side-note, Your infinite loop with a match on receiver.recv() could be replaced by a for loop using the Receiver::iter method:
for _ in receiver.iter() {
// do something with the value
}