Avoid incorrect lifetime suggestion involving sym::anon

[estebank]

Given

#!/usr/bin/env -S cargo +nightly -Zscript -q
---
packag.edition = "2024"

[dependencies]
rama = "0.2.0"
---

fn main() {}

struct Foo;

impl<S, Request: 'static> rama::Service<S, Request> for Foo {
    type Response = ();
    type Error = ();

    fn serve(
        &self,
        _ctx: rama::Context<S>,
        _req: Request,
    ) -> impl Future<Output = Result<Self::Response, Self::Error>> {
        async { todo!() }
    }
}

emit

error[E0311]: the parameter type `S` may not live long enough
  --> f20.rs:20:10
   |
20 |     ) -> impl Future<Output = Result<Self::Response, Self::Error>> {
   |          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ...so that the type `impl Future<Output = Result<<Foo as Service<S, Request>>::Response, <Foo as Service<S, Request>>::Error>>` will meet its required lifetime bounds...
   |
  ::: /home/gh-estebank/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/rama-core-0.2.0/src/service/svc.rs:19:9
   |
19 |         &self,
   |         - the parameter type `S` must be valid for the anonymous lifetime as defined here...
   |
note: ...that is required by this bound
  --> /home/gh-estebank/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/rama-core-0.2.0/src/service/svc.rs:22:77
   |
22 |     ) -> impl Future<Output = Result<Self::Response, Self::Error>> + Send + '_;
   |                                                                             ^^

without the incorrect suggestion

help: consider adding an explicit lifetime bound
   |
22 |     ) -> impl Future<Output = Result<Self::Response, Self::Error>> where S: anon {
   |                                                                    +++++++++++++

Address #142887. Needs a repro test.

[rustbot]

r? @nnethercote

rustbot has assigned @nnethercote.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[nnethercote]

Needs a repro test.

Yes!

View changes since this review

[JayanAXHF]

@estebank Hi Triage here, is there anything blocking this PRs implementation? Are you still working on this?

[estebank]

@JayanAXHF I need to come up with a minimized repro case that's suitable for the test suite. I could do with an extra hand on that.

[JayanAXHF]

Ive added the E-needs-mcve label, hopefully the CFP works!

[rust-bors]

☔ The latest upstream changes (presumably #152839) made this pull request unmergeable. Please resolve the merge conflicts.