Reduce type errors in async fns #67025
Labels
A-async-await
Area: Async & Await
A-diagnostics
Area: Messages for errors, warnings, and lints
AsyncAwait-Triaged
Async-await issues that have been triaged during a working group meeting.
C-enhancement
Category: An issue proposing an enhancement or a PR with one.
D-verbose
Diagnostics: Too much output caused by a single piece of incorrect code.
T-compiler
Relevant to the compiler team, which will review and decide on the PR/issue.
Comments
|
@birkenfeld: Can you link to your repository, and re-fry on the latest nightly? I think more recent builds of |
|
Didn't see a change with latest nightly (7afe6d9 2019-12-03). Unfortunately I can't link to the repo, but if needed I can try to extract a minimal example which produces similar results. |
|
This looks very similar to an error that I'm getting around async code using mpsc. I can provide a minimal example: use futures::executor::ThreadPool;
use std::sync::mpsc;
use std::time::Duration;
async fn wait(d: Duration) -> f32 {
d.as_secs_f32()
}
async fn add_task(sender: mpsc::Sender<f32>, i: u64) {
// let v = wait(Duration::from_millis(i)).await;
// sender.send(v).unwrap();
sender.send(wait(Duration::from_millis(i)).await).unwrap();
}
fn main() {
let (sender, receiver) = mpsc::channel();
let thread_pool = ThreadPool::new().unwrap();
for i in (0..10).rev() {
thread_pool.spawn_ok(add_task(sender.clone(), i * 100));
}
drop(sender);
for v in receiver {
println!("{}", v);
}
}While using these dependencies: [dependencies]
futures = { version = "0.3", features = [ "thread-pool" ] }results in this error: It's worth noting the the commented version works. Using |
|
Minimized: struct NotSync(*const ());
async fn async_not_sync() {
let a = NotSync(0 as *const ());
// Uncomment this line to see a better error
//let b: *const ();
// Force 'a' to be live across a yield point
async fn dummy() {}
dummy().await;
}
fn require_sync<T: Sync>(val: T) {}
fn main() {
require_sync(async_not_sync());
}It seems that the better error message is only being triggered by certain |
|
Possibly helped by #65345 as well |
|
Marking as "triaged". Not promoting to "on deck" for now -- it seems likely though that existing PRs cover it. We should land those already! |
|
#65345 produces the following error for @Aaron1011's minimized example: |
|
Close as errors reduced. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I will admit that this specific
async fnis a bit long, but I'd like the compiler to tell me in a different way. (Actual reason is addition of one line borrowing the wrong type across an.awaitboundary, of course)Like this, we're coming out ahead of C++ 😄
The text was updated successfully, but these errors were encountered: