new trait solver regression in `amaru-pure-stage` library

[rkuhn]

Compiling amaru-pure-stage fails with the following error message on nightly-2026-08-22:

error[E0275]: overflow evaluating the requirement `impl futures_util::Future<Output = ()>: std::marker::Send`
   --> crates/amaru-pure-stage/src/tokio.rs:515:47
    |
515 |                     let handle = tokio::spawn(stage);
    |                                  ------------ ^^^^^
    |                                  |
    |                                  required by a bound introduced by this call
    |
note: required by a bound in `tokio::spawn`
   --> /Users/rkuhn/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/tokio-1.53.1/src/task/spawn.rs:176:21
    |
174 |     pub fn spawn<F>(future: F) -> JoinHandle<F::Output>
    |            ----- required by a bound in this function
175 |     where
176 |         F: Future + Send + 'static,
    |                     ^^^^ required by this bound in `spawn`

It does not fail with RUSTFLAGS=-Znext-solver=coherence or on nightly-2026-08-21 (no matter which solver).

[apiraino]

@matthiaskrgr could this be another duplicate of #159680?

[lcnr]

seems to be rust-lang/trait-system-refactor-initiative#209 🤔

use std::future::Future;

fn assert_send(_: impl Send) {}

async fn run_stage() {
    interpreter().await;
}

fn interpreter<'a>() -> impl Future + Send + use<'a> {
    async { assert_send(run_stage()) }
}

fn main() {}

[lcnr]

@rkuhn fixing this issue is quite hard. It's impact is fairly small and the reason that this currently works is brittle. Myy current intent is to likely accept this breakage for the initial stabilization of the new trait solver and fix this as part of the larger region handling rework https://rust-lang.github.io/goals/2026/assumptions_on_binders.html

Opening a PR with works around this issue in pragma-org/amaru#1282

[rkuhn]

Many thanks for the quick response and resolution! Unfortunately our CI setup requires that we “redo” the PR, hope you don’t mind.