I have a Vec of futures which I want to execute concurrently (but not necessarily in parallel). Basically, I'm looking for some kind of select function that is similar to tokio::select! but takes a collection of futures, or, conversely, a function that is similar to futures::join_all but returns once the first future is done.
An additional requirement is that once a future finished I might want to add a new future to the Vec.
With such a function, my code would roughly look like this:
use std::future::Future;
use std::time::Duration;
use tokio::time::sleep;
async fn wait(millis: u64) -> u64 {
sleep(Duration::from_millis(millis)).await;
millis
}
// This pseudo-implementation simply removes the last
// future and awaits it. I'm looking for something that
// instead polls all futures until one is finished, then
// removes that future from the Vec and returns it.
async fn select<F, O>(futures: &mut Vec<F>) -> O
where
F: Future<Output=O>
{
let future = futures.pop().unwrap();
future.await
}
#[tokio::main]
async fn main() {
let mut futures = vec![
wait(500),
wait(300),
wait(100),
wait(200),
];
while !futures.is_empty() {
let finished = select(&mut futures).await;
println!("Waited {}ms", finished);
if some_condition() {
futures.push(wait(200));
}
}
}
This is exactly what futures::stream::FuturesUnordered is for (which I've found by looking through the source of StreamExt::for_each_concurrent):
use futures::{stream::FuturesUnordered, StreamExt};
use std::time::Duration;
use tokio::time::{sleep, Instant};
async fn wait(millis: u64) -> u64 {
sleep(Duration::from_millis(millis)).await;
millis
}
#[tokio::main]
async fn main() {
let mut futures = FuturesUnordered::new();
futures.push(wait(500));
futures.push(wait(300));
futures.push(wait(100));
futures.push(wait(200));
let start_time = Instant::now();
let mut num_added = 0;
while let Some(wait_time) = futures.next().await {
println!("Waited {}ms", wait_time);
if num_added < 3 {
num_added += 1;
futures.push(wait(200));
}
}
println!("Completed all work in {}ms", start_time.elapsed().as_millis());
}
(playground)
Here's a working prototype based on streams and StreamExt::for_each_concurrent, as Martin Gallagher has suggested in a comment:
use std::time::Duration;
use tokio::sync::RwLock;
use tokio::time::sleep;
use futures::stream::{self, StreamExt};
use futures::{channel::mpsc, sink::SinkExt};
async fn wait(millis: u64) -> u64 {
sleep(Duration::from_millis(millis)).await;
millis
}
#[tokio::main]
async fn main() {
let (mut sink, futures_stream) = mpsc::unbounded();
let start_futures = vec![wait(500), wait(300), wait(100), wait(200)];
let num_futures = RwLock::new(start_futures.len());
sink.send_all(&mut stream::iter(start_futures.into_iter().map(Ok)))
.await
.unwrap();
let sink_lock = RwLock::new(sink);
futures_stream
.for_each_concurrent(None, |fut| async {
let wait_time = fut.await;
println!("Waited {}", wait_time);
if some_condition() {
println!("Adding new future");
let mut sink = sink_lock.write().await;
sink.send(wait(100)).await.unwrap();
} else {
let mut num_futures = num_futures.write().await;
*num_futures -= 1;
}
let num_futures = num_futures.read().await;
if *num_futures <= 0 {
// Close the sink to exit the for_each_concurrent
sink_lock.write().await.close().await.unwrap();
}
})
.await;
}
While this approach works it has the drawback that we need to maintain a separate counter of remaining futures so that we can close the sink -- there's no Vec of futures for which we can check whether it's empty. Closing the sink requires another lock.
Given that I'm fairly new to Rust I wouldn't be surprised if this approach could be made more elegant.
Related
I am writing an echo server that is able to listen on multiple ports. My working server below relies on select to accept connections from 2 different listeners
However, instead of defining the listeners as individual variables, is it possible to define select branches based on a Vec<TcpListener> ?
use tokio::{io, net, select, spawn};
#[tokio::main]
async fn main() {
let listener1 = net::TcpListener::bind("127.0.0.1:8001").await.unwrap();
let listener2 = net::TcpListener::bind("127.0.0.1:8002").await.unwrap();
loop {
let (conn, _) = select! {
v = listener1.accept() => v.unwrap(),
v = listener2.accept() => v.unwrap(),
};
spawn(handle(conn));
}
}
async fn handle(mut conn: net::TcpStream) {
let (mut read, mut write) = conn.split();
io::copy(&mut read, &mut write).await.unwrap();
}
While futures::future::select_all() works, it is not very elegant (IMHO) and it creates an allocation for each round. A better solution is to use streams (note this also allocates on each round, but this allocates much less):
use tokio::{io, net, spawn};
use tokio_stream::wrappers::TcpListenerStream;
use futures::stream::{StreamExt, SelectAll};
#[tokio::main]
async fn main() {
let mut listeners = SelectAll::new();
listeners.push(TcpListenerStream::new(net::TcpListener::bind("127.0.0.1:8001").await.unwrap()));
listeners.push(TcpListenerStream::new(net::TcpListener::bind("127.0.0.1:8002").await.unwrap()));
while let Some(conn) = listeners.next().await {
let conn = conn.unwrap();
spawn(handle(conn));
}
}
You can use the select_all function from the futures crate, which takes an iterator of futures and awaits any of them (instead of all of them, like join_all does):
use futures::{future::select_all, FutureExt};
use tokio::{io, net, select, spawn};
#[tokio::main]
async fn main() {
let mut listeners = [
net::TcpListener::bind("127.0.0.1:8001").await.unwrap(),
net::TcpListener::bind("127.0.0.1:8002").await.unwrap(),
];
loop {
let (result, index, _) = select_all(
listeners
.iter_mut()
// note: `FutureExt::boxed` is called here because `select_all`
// requires the futures to be pinned
.map(|listener| listener.accept().boxed()),
)
.await;
let (conn, _) = result.unwrap();
spawn(handle(conn));
}
}
I have created a very simplified example of the code I am having an issue with:
use core::time;
use std::thread;
use tokio::sync::{mpsc::Receiver, RwLock};
struct MyStruct {
counter: Arc<RwLock<i32>>,
rx: RwLock<Receiver<i32>>,
}
impl MyStruct {
async fn start_here(&self) { // <--------- Lifetime error here on self
while let Some(message) = self.rx.write().await.recv().await {
tokio::spawn(self.do_some_work_then_update_counter());
}
}
async fn do_some_work_then_update_counter(&self) {
let dur = time::Duration::from_millis(10000);
thread::sleep(dur);
let mut counter = self.counter.write().await;
*counter += 1;
}
}
There is a receiver that is receiving messages from another part of the program, and I want to be able to process each message in its own task to prevent blocking the next message from being processed.
As you can imagine it's a lifetime error since the task could outlast self in this case.
One solution I have done is this:
impl MyStruct {
async fn start_here(&self) {
while let Some(message) = self.rx.write().await.recv().await {
let counter = self.counter.clone();
tokio::spawn(do_some_work_then_update_counter(counter));
}
}
}
async fn do_some_work_then_update_counter(counter: Arc<RwLock<i32>>) {
let dur = time::Duration::from_millis(10000);
thread::sleep(dur);
let mut counter = counter.write().await;
*counter += 1;
}
This just doesn't seem like a good option, I want to keep do_some_work_then_update_counter as an impl of MyStruct instead of a free function since it is modifying data on MyStruct.
I am wondering if there is a better solution to this?
You can if you'll return impl Future directly instead of being async fn:
fn do_some_work_then_update_counter(&self) -> impl Future<Output = ()> {
let counter = Arc::clone(&self.counter);
async move {
let dur = time::Duration::from_millis(10000);
thread::sleep(dur);
let mut counter = counter.write().await;
*counter += 1;
}
}
I'm trying to make a Stream that would wait until a specific character is in buffer. I know there's read_until() on BufRead but I actually need a custom solution, as this is a stepping stone to implement waiting until a specific string in in buffer (or, for example, a regexp match happens).
In my project where I first encountered the problem, problem was that future processing just hanged when I get a Ready(_) from inner future and return NotReady from my function. I discovered I shouldn't do that per docs (last paragraph). However, what I didn't get, is what's the actual alternative that is promised in that paragraph. I read all the published documentation on the Tokio site and it doesn't make sense for me at the moment.
So following is my current code. Unfortunately I couldn't make it simpler and smaller as it's already broken. Current result is this:
Err(Custom { kind: Other, error: Error(Shutdown) })
Err(Custom { kind: Other, error: Error(Shutdown) })
Err(Custom { kind: Other, error: Error(Shutdown) })
<ad infinum>
Expected result is getting some Ok(Ready(_)) out of it, while printing W and W', and waiting for specific character in buffer.
extern crate futures;
extern crate tokio_core;
extern crate tokio_io;
extern crate tokio_io_timeout;
extern crate tokio_process;
use futures::stream::poll_fn;
use futures::{Async, Poll, Stream};
use tokio_core::reactor::Core;
use tokio_io::AsyncRead;
use tokio_io_timeout::TimeoutReader;
use tokio_process::CommandExt;
use std::process::{Command, Stdio};
use std::sync::{Arc, Mutex};
use std::thread;
use std::time::Duration;
struct Process {
child: tokio_process::Child,
stdout: Arc<Mutex<tokio_io_timeout::TimeoutReader<tokio_process::ChildStdout>>>,
}
impl Process {
fn new(
command: &str,
reader_timeout: Option<Duration>,
core: &tokio_core::reactor::Core,
) -> Self {
let mut cmd = Command::new(command);
let cat = cmd.stdout(Stdio::piped());
let mut child = cat.spawn_async(&core.handle()).unwrap();
let stdout = child.stdout().take().unwrap();
let mut timeout_reader = TimeoutReader::new(stdout);
timeout_reader.set_timeout(reader_timeout);
let timeout_reader = Arc::new(Mutex::new(timeout_reader));
Self {
child,
stdout: timeout_reader,
}
}
}
fn work() -> Result<(), ()> {
let window = Arc::new(Mutex::new(Vec::new()));
let mut core = Core::new().unwrap();
let process = Process::new("cat", Some(Duration::from_secs(20)), &core);
let mark = Arc::new(Mutex::new(b'c'));
let read_until_stream = poll_fn({
let window = window.clone();
let timeout_reader = process.stdout.clone();
move || -> Poll<Option<u8>, std::io::Error> {
let mut buf = [0; 8];
let poll;
{
let mut timeout_reader = timeout_reader.lock().unwrap();
poll = timeout_reader.poll_read(&mut buf);
}
match poll {
Ok(Async::Ready(0)) => Ok(Async::Ready(None)),
Ok(Async::Ready(x)) => {
{
let mut window = window.lock().unwrap();
println!("W: {:?}", *window);
println!("buf: {:?}", &buf[0..x]);
window.extend(buf[0..x].into_iter().map(|x| *x));
println!("W': {:?}", *window);
if let Some(_) = window.iter().find(|c| **c == *mark.lock().unwrap()) {
Ok(Async::Ready(Some(1)))
} else {
Ok(Async::NotReady)
}
}
}
Ok(Async::NotReady) => Ok(Async::NotReady),
Err(e) => Err(e),
}
}
});
let _stream_thread = thread::spawn(move || {
for o in read_until_stream.wait() {
println!("{:?}", o);
}
});
match core.run(process.child) {
Ok(_) => {}
Err(e) => {
println!("Child error: {:?}", e);
}
}
Ok(())
}
fn main() {
work().unwrap();
}
This is complete example project.
If you need more data you need to call poll_read again until you either find what you were looking for or poll_read returns NotReady.
You might want to avoid looping in one task for too long, so you can build yourself a yield_task function to call instead if poll_read didn't return NotReady; it makes sure your task gets called again ASAP after other pending tasks were run.
To use it just run return yield_task();.
fn yield_inner() {
use futures::task;
task::current().notify();
}
#[inline(always)]
pub fn yield_task<T, E>() -> Poll<T, E> {
yield_inner();
Ok(Async::NotReady)
}
Also see futures-rs#354: Handle long-running, always-ready futures fairly #354.
With the new async/await API futures::task::current is gone; instead you'll need a std::task::Context reference, which is provided as parameter to the new std::future::Future::poll trait method.
If you're already manually implementing the std::future::Future trait you can simply insert:
context.waker().wake_by_ref();
return std::task::Poll::Pending;
Or build yourself a Future-implementing type that yields exactly once:
pub struct Yield {
ready: bool,
}
impl core::future::Future for Yield {
type Output = ();
fn poll(self: core::pin::Pin<&mut Self>, cx: &mut core::task::Context<'_>) -> core::task::Poll<Self::Output> {
let this = self.get_mut();
if this.ready {
core::task::Poll::Ready(())
} else {
cx.waker().wake_by_ref();
this.ready = true; // ready next round
core::task::Poll::Pending
}
}
}
pub fn yield_task() -> Yield {
Yield { ready: false }
}
And then use it in async code like this:
yield_task().await;
I have a struct that sends messages to a channel as well as updating some of its own fields. How do I implement a monitoring thread that looks (read only) at its internal fields periodically?
I can write it using a Arc<Mutex<T>> wrapper, but I feel it is not that efficient as A::x could have been i32 which is stored and updated on the stack. Is there any better way to do it without the locks?
use std::sync::{Arc, Mutex};
use std::sync::mpsc::{channel, Sender};
use std::{thread, time};
struct A {
x: Arc<Mutex<i32>>,
y: Sender<i32>,
}
impl A {
fn do_some_loop(&mut self) {
let sleep_time = time::Duration::from_millis(200);
// This is a long running thread.
for x in 1..1000000 {
*self.x.lock().unwrap() = x;
self.y.send(x);
thread::sleep(sleep_time);
}
}
}
fn test() {
let (sender, recever) = channel();
let x = Arc::new(Mutex::new(1));
let mut a = A { x: x.clone(), y: sender };
thread::spawn(move || {
// Monitor every 10 secs.
let sleep_time = time::Duration::from_millis(10000);
loop {
thread::sleep(sleep_time);
println!("{}", *x.lock().unwrap());
}
});
a.do_some_loop();
}
I have two parts of code that I want to run in a loop. Sometimes I need to make the loop 'sleep', making each iteration skip the second part. The loop should stop sleeping after a set amount of time (for example using a thread with a call to thread::sleep). How do I accomplish this?
use std::thread;
let mut sleeping = false;
let mut handle = thread::spawn(|| {});
loop {
part_1();
if sleeping {
continue;
}
part_2();
if some_condition {
sleeping = true;
handle = thread::spawn(|| thread::sleep_ms(100));
}
}
In this example, if the condition is met, the part_2 call would be skipped for some amount of iterations. My use case is continuing to run graphical updates in a game, while freezing the game's logic (such as counting down timers).
There is no need for the overhead of threads or even the need to sleep. Simply track the time that you should delay executing code until:
use std::time::{Duration, Instant};
fn part_1() {}
fn part_2() {}
fn some_condition() -> bool {
false
}
fn main() {
let mut sleep_until = None;
loop {
part_1();
if let Some(until) = sleep_until {
if until > Instant::now() {
continue;
}
}
part_2();
if some_condition() {
let now = Instant::now();
let until = now + Duration::from_millis(500);
sleep_until = Some(until);
}
}
}
Although I'd probably avoid the use of continue here, and instead embed the logic within:
use std::time::{Duration, Instant};
fn perform_physics_calculation() {}
fn perform_graphics_render() {}
fn main() {
let mut next_graphics_update = Instant::now();
let graphics_delay = Duration::from_millis(500);
loop {
let now = Instant::now();
perform_physics_calculation();
if next_graphics_update <= now {
perform_graphics_render();
next_graphics_update = now + graphics_delay;
}
}
}
Note in one case I use an Option<Instant> and in the other I just use an Instant; both cases can make sense.
Turn your sleeping variable into a reference-counted atomic boolean so that you can reset it on the sleeping thread.
use std::sync::Arc;
use std::sync::atomic::{AtomicBool, Ordering};
use std::thread;
use std::time::Duration;
fn part_1() {}
fn part_2() {}
fn some_condition() -> bool { false }
fn main() {
let sleeping = Arc::new(AtomicBool::new(false));
let mut handle = None;
loop {
part_1();
if sleeping.load(Ordering::Acquire) {
continue;
}
part_2();
if some_condition() {
sleeping.store(true, Ordering::Release);
let sleeping_clone = sleeping.clone();
handle = Some(thread::spawn(move || {
thread::sleep(Duration::from_millis(100));
sleeping_clone.store(false, Ordering::Release);
}));
}
}
}