fix: Unexpected trait bound not satisfied in HRTB and Associated Type

[LYF1999]

fix #103563

[rustbot]

r? @davidtwco

(rustbot has picked a reviewer for you, use r? to override)

[compiler-errors]

@LYF1999, please provide a more thorough description in the PR for why you think this fixes the issue, since the change is in some pretty fundamental trait selection code.

[LYF1999]

there is code can produce this bug. And it's not about GAT.

use std::fmt::Debug;

fn main() {
    let mut log_service = LogService { inner: Inner };
    log_service.call(Struct1 {});
}

pub trait Service<Request> {
    type Response;

    fn call(&mut self, req: Request) -> Self::Response;
}

pub struct LogService<S> {
    inner: S,
}

impl<T, S, U> Service<T> for LogService<S>
where
    S: Service<T, Response = U>,
    U: for<'a> Extension<'a>,
    for<'a> <U as Extension<'a>>::Item: std::fmt::Debug,
{
    type Response = S::Response;

    fn call(&mut self, req: T) -> Self::Response {
        self.inner.call(req)
    }
}

pub struct Inner;

impl Service<Struct1> for Inner {
    type Response = Resp;

    fn call(&mut self, req: Struct1) -> Self::Response {
        Resp::A(req)
    }
}

pub trait Extension<'a> {
    type Item;
}

pub trait E2
where
    Self: for<'a> Extension<'a>,
{
    fn touch<F>(self, f: F) -> Self
    where
        F: for<'b> Fn(<Self as Extension<'b>>::Item);
}

#[derive(Debug)]
pub struct Struct1 {}

pub enum Resp {
    A(Struct1),
}

impl<'a> Extension<'a> for Resp {
    type Item = RespItem<'a>;
}

impl E2 for Resp {
    fn touch<F>(self, f: F) -> Self
    where
        F: for<'b> Fn(<Self as Extension<'b>>::Item),
    {
        match &self {
            Resp::A(a) => f(RespItem::A(a)),
        }
        self
    }
}

pub enum RespItem<'a> {
    A(&'a Struct1),
}

impl Debug for RespItem<'_> {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        match self {
            Self::A(arg0) => f.debug_tuple("A").field(arg0).finish(),
        }
    }
}

compiler try to evaluate predicate for<'a> <U as Extension<'a>>::Item: std::fmt::Debug.
in this context, fn evaluate_trait_predicate_recursively gets a obligation <Resp<'a> as Extension<'a>>::Item: Debug.
but fn candidate_from_obligation assemble 0 candidate for <Resp<'a> as Extension<'a>>::Item: Debug.
So this code failed.
there are some ways to resolve that:

1. normalize <Resp<'a> as Extension<'a>>::Item: Debug to RespItem<'a>: Debug and evaluate it.
2. normalize <Resp<'a> as Extension<'a>>::Item: Debug in candidate_from_obligation and assemble 1 candidate ImplCandidate. but fn rematch_impl will check ImplCandidate.

@compiler-errors

[davidtwco]

r? rust-lang/types

[jackh726]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 125e9aae4db14716e2ac438806fd4b036ff01872 with merge b84d21ad268a5a9a3962aae7e6d9543ed1a2ac90...

[bors]

☀️ Try build successful - checks-actions
Build commit: b84d21ad268a5a9a3962aae7e6d9543ed1a2ac90 (b84d21ad268a5a9a3962aae7e6d9543ed1a2ac90)

[rust-timer]

Finished benchmarking commit (b84d21ad268a5a9a3962aae7e6d9543ed1a2ac90): comparison URL.

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

@bors rollup=never
@rustbot label: +S-waiting-on-review -S-waiting-on-perf +perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	1.3%	[1.2%, 1.4%]	2
Regressions ❌ (secondary)	3.1%	[0.4%, 4.1%]	7
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.7%	[-0.9%, -0.6%]	4
All ❌✅ (primary)	1.3%	[1.2%, 1.4%]	2

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	1.5%	[1.5%, 1.5%]	1
Regressions ❌ (secondary)	6.4%	[6.4%, 6.4%]	1
Improvements ✅ (primary)	-0.7%	[-1.2%, -0.5%]	3
Improvements ✅ (secondary)	-2.2%	[-2.3%, -2.0%]	3
All ❌✅ (primary)	-0.2%	[-1.2%, 1.5%]	4

Cycles

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	0.5%	[0.5%, 0.5%]	1
Regressions ❌ (secondary)	2.1%	[1.8%, 2.4%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-5.0%	[-5.0%, -5.0%]	2
All ❌✅ (primary)	0.5%	[0.5%, 0.5%]	1

[LYF1999]

@bors try @rust-timer queue

[rfcbot]

The final comment period, with a disposition to merge, as per the review above, is now complete.

As the automated representative of the governance process, I would like to thank the author for their work and everyone else who contributed.

This will be merged soon.

[lcnr]

r=me after the final change

[lcnr]

r? @lcnr

[lcnr]

@bors r+

[bors]

📌 Commit d164448 has been approved by lcnr

It is now in the queue for this repository.

[bors]

⌛ Testing commit d164448 with merge 9bb6e60...

[bors]

☀️ Test successful - checks-actions
Approved by: lcnr
Pushing 9bb6e60 to master...

[rust-timer]

Finished benchmarking commit (9bb6e60): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

This benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.5%	[-2.5%, -2.5%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.5%	[-2.5%, -2.5%]	1

Cycles

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.4%	[-2.4%, -2.4%]	1
All ❌✅ (primary)	-	-	0

[lcnr]

afaict this uses the stack which already includes the current goal, so we get a coinductive cycle in #108897. Have to pop the element pushed in line 988 before recursing so that the stack isn't invalid.

[LYF1999]

thanks~