coroutines
The tracking issue for this feature is: #43122
The coroutines
feature gate in Rust allows you to define coroutine or
coroutine literals. A coroutine is a "resumable function" that syntactically
resembles a closure but compiles to much different semantics in the compiler
itself. The primary feature of a coroutine is that it can be suspended during
execution to be resumed at a later date. Coroutines use the yield
keyword to
"return", and then the caller can resume
a coroutine to resume execution just
after the yield
keyword.
Coroutines are an extra-unstable feature in the compiler right now. Added in RFC 2033 they're mostly intended right now as a information/constraint gathering phase. The intent is that experimentation can happen on the nightly compiler before actual stabilization. A further RFC will be required to stabilize coroutines and will likely contain at least a few small tweaks to the overall design.
A syntactical example of a coroutine is:
#![feature(coroutines, coroutine_trait, stmt_expr_attributes)] use std::ops::{Coroutine, CoroutineState}; use std::pin::Pin; fn main() { let mut coroutine = #[coroutine] || { yield 1; return "foo" }; match Pin::new(&mut coroutine).resume(()) { CoroutineState::Yielded(1) => {} _ => panic!("unexpected value from resume"), } match Pin::new(&mut coroutine).resume(()) { CoroutineState::Complete("foo") => {} _ => panic!("unexpected value from resume"), } }
Coroutines are closure-like literals which are annotated with #[coroutine]
and can contain a yield
statement. The
yield
statement takes an optional expression of a value to yield out of the
coroutine. All coroutine literals implement the Coroutine
trait in the
std::ops
module. The Coroutine
trait has one main method, resume
, which
resumes execution of the coroutine at the previous suspension point.
An example of the control flow of coroutines is that the following example prints all numbers in order:
#![feature(coroutines, coroutine_trait, stmt_expr_attributes)] use std::ops::Coroutine; use std::pin::Pin; fn main() { let mut coroutine = #[coroutine] || { println!("2"); yield; println!("4"); }; println!("1"); Pin::new(&mut coroutine).resume(()); println!("3"); Pin::new(&mut coroutine).resume(()); println!("5"); }
At this time the main use case of coroutines is an implementation
primitive for async
/await
and gen
syntax, but coroutines
will likely be extended to other primitives in the future.
Feedback on the design and usage is always appreciated!
The Coroutine
trait
The Coroutine
trait in std::ops
currently looks like:
#![allow(unused)] fn main() { #![feature(arbitrary_self_types, coroutine_trait)] use std::ops::CoroutineState; use std::pin::Pin; pub trait Coroutine<R = ()> { type Yield; type Return; fn resume(self: Pin<&mut Self>, resume: R) -> CoroutineState<Self::Yield, Self::Return>; } }
The Coroutine::Yield
type is the type of values that can be yielded with the
yield
statement. The Coroutine::Return
type is the returned type of the
coroutine. This is typically the last expression in a coroutine's definition or
any value passed to return
in a coroutine. The resume
function is the entry
point for executing the Coroutine
itself.
The return value of resume
, CoroutineState
, looks like:
#![allow(unused)] fn main() { pub enum CoroutineState<Y, R> { Yielded(Y), Complete(R), } }
The Yielded
variant indicates that the coroutine can later be resumed. This
corresponds to a yield
point in a coroutine. The Complete
variant indicates
that the coroutine is complete and cannot be resumed again. Calling resume
after a coroutine has returned Complete
will likely result in a panic of the
program.
Closure-like semantics
The closure-like syntax for coroutines alludes to the fact that they also have closure-like semantics. Namely:
-
When created, a coroutine executes no code. A closure literal does not actually execute any of the closure's code on construction, and similarly a coroutine literal does not execute any code inside the coroutine when constructed.
-
Coroutines can capture outer variables by reference or by move, and this can be tweaked with the
move
keyword at the beginning of the closure. Like closures all coroutines will have an implicit environment which is inferred by the compiler. Outer variables can be moved into a coroutine for use as the coroutine progresses. -
Coroutine literals produce a value with a unique type which implements the
std::ops::Coroutine
trait. This allows actual execution of the coroutine through theCoroutine::resume
method as well as also naming it in return types and such. -
Traits like
Send
andSync
are automatically implemented for aCoroutine
depending on the captured variables of the environment. Unlike closures, coroutines also depend on variables live across suspension points. This means that although the ambient environment may beSend
orSync
, the coroutine itself may not be due to internal variables live acrossyield
points being not-Send
or not-Sync
. Note that coroutines do not implement traits likeCopy
orClone
automatically. -
Whenever a coroutine is dropped it will drop all captured environment variables.
Coroutines as state machines
In the compiler, coroutines are currently compiled as state machines. Each
yield
expression will correspond to a different state that stores all live
variables over that suspension point. Resumption of a coroutine will dispatch on
the current state and then execute internally until a yield
is reached, at
which point all state is saved off in the coroutine and a value is returned.
Let's take a look at an example to see what's going on here:
#![feature(coroutines, coroutine_trait, stmt_expr_attributes)] use std::ops::Coroutine; use std::pin::Pin; fn main() { let ret = "foo"; let mut coroutine = #[coroutine] move || { yield 1; return ret }; Pin::new(&mut coroutine).resume(()); Pin::new(&mut coroutine).resume(()); }
This coroutine literal will compile down to something similar to:
#![feature(arbitrary_self_types, coroutine_trait)] use std::ops::{Coroutine, CoroutineState}; use std::pin::Pin; fn main() { let ret = "foo"; let mut coroutine = { enum __Coroutine { Start(&'static str), Yield1(&'static str), Done, } impl Coroutine for __Coroutine { type Yield = i32; type Return = &'static str; fn resume(mut self: Pin<&mut Self>, resume: ()) -> CoroutineState<i32, &'static str> { use std::mem; match mem::replace(&mut *self, __Coroutine::Done) { __Coroutine::Start(s) => { *self = __Coroutine::Yield1(s); CoroutineState::Yielded(1) } __Coroutine::Yield1(s) => { *self = __Coroutine::Done; CoroutineState::Complete(s) } __Coroutine::Done => { panic!("coroutine resumed after completion") } } } } __Coroutine::Start(ret) }; Pin::new(&mut coroutine).resume(()); Pin::new(&mut coroutine).resume(()); }
Notably here we can see that the compiler is generating a fresh type,
__Coroutine
in this case. This type has a number of states (represented here
as an enum
) corresponding to each of the conceptual states of the coroutine.
At the beginning we're closing over our outer variable foo
and then that
variable is also live over the yield
point, so it's stored in both states.
When the coroutine starts it'll immediately yield 1, but it saves off its state
just before it does so indicating that it has reached the yield point. Upon
resuming again we'll execute the return ret
which returns the Complete
state.
Here we can also note that the Done
state, if resumed, panics immediately as
it's invalid to resume a completed coroutine. It's also worth noting that this
is just a rough desugaring, not a normative specification for what the compiler
does.