Normalize region obligations before regionck by adwinwhite · Pull Request #159649 · rust-lang/rust

adwinwhite · 2026-07-21T09:52:33Z

We now normalize TypeOutlives predicates when evaluating them in the next solver.
So we don't need to normalize in regionck anymore.

r? @ShoyuVanilla
cc @lcnr

rustbot · 2026-07-21T09:52:37Z

Some changes occurred to the core trait solver

cc @rust-lang/initiative-trait-system-refactor

adwinwhite · 2026-07-21T09:56:29Z

    /// This function assumes that all infer variables are already constrained.
+    ///
+    /// FIXME: this can probably be moved back to `rustc_infer` now that normalization is
+    /// on longer required. These two extention traits won't be needed then.


I didn't do this in this PR because a lot of stuff would be moved and I'm uncertain about some of them.

View changes since the review

want to put FIXME(#155345) here so that we do it as part of that refactor

lcnr · 2026-07-21T09:56:46Z

                // `TypeOutlives` is structural, so we should try to opportunistically resolve all
                // region vids before processing regions, so we have a better chance to match clauses
                // in our param-env.
+                let sup_type = self.shallow_resolve(sup_type);


is that necessary/could we just use eager_resolve_vars here?

Hmm, also, shallow_resolve feels wrong? because we could have ?x which resolves to Vec<?y> where ?y can be resolved to u32 later on. I feel like this should have otherwise been resolve_vars_if_possible to recur into the type

View changes since the review

Without it we'll have unresolved infer var error later.
Yeah, it should be resolve_vars_if_possible. We had a resolve_vars_if_possible somewhere before.

ye, my thoughts are that we could use

let (sup_type, sub_region) = eager_resolve_vars(self, (sup_type, sub_region));

instead of

let sup_type = self.shallow_resolve(sup_type); let (sup_type, sub_region) = (sup_type, sub_region).fold_with(&mut OpportunisticRegionResolver::new(self));

lcnr · 2026-07-21T09:56:57Z

        }

-        self.evaluate_added_goals_and_make_canonical_response(Certainty::Yes)
+        if ty.has_non_region_infer() || ty.has_non_rigid_aliases() {


comment please :>

View changes since the review

lcnr · 2026-07-21T09:57:34Z

-                        }
-                    }
-                }
+            if let Some(type_outlives) = bound.as_type_outlives_clause() {


debug_assert: if next-solver && non-igid aliases => ICE

View changes since the review

lcnr · 2026-07-21T10:00:31Z

there's also normalize_and_add_type_outlives_constraints which can be removed now I think

ShoyuVanilla · 2026-07-22T09:08:37Z

    /// This function assumes that all infer variables are already constrained.
+    ///
+    /// FIXME: this can probably be moved back to `rustc_infer` now that normalization is
+    /// on longer required. These two extention traits won't be needed then.


Suggested change

/// on longer required. These two extention traits won't be needed then.

/// no longer required. These two extension traits won't be needed then.

View changes since the review

rustbot · 2026-07-22T12:19:23Z

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

adwinwhite · 2026-07-22T12:25:19Z

@bors try @rust-timer queue

Normalize region obligations before regionck

adwinwhite · 2026-07-22T12:53:38Z

@bors try cancel

rust-bors · 2026-07-22T12:53:42Z

Try build cancelled. Cancelled workflows:

https://github.com/rust-lang/rust/actions/runs/29919558340

adwinwhite · 2026-07-23T05:57:25Z

@bors try @rust-timer queue

Normalize region obligations before regionck

rust-bors · 2026-07-23T08:02:50Z

☀️ Try build successful (CI)
Build commit: d85ff7e (d85ff7ea0db7e79c08713eec21d583acd432bbdc)
Base parent: 390279b (390279b302ca98ae270f434100ae3730531d1246)

rust-timer · 2026-07-23T08:43:23Z

Finished benchmarking commit (d85ff7e): comparison URL.

Overall result: ❌✅ regressions and improvements - no action needed

Benchmarking means the PR may be perf-sensitive. Consider adding rollup=never if this change is not fit for rolling up.

@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	1.2%	[1.2%, 1.2%]	3
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.3%	[-0.3%, -0.3%]	2
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (secondary 6.3%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	6.3%	[6.3%, 6.3%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary 3.6%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	3.6%	[3.6%, 3.6%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

This perf run didn't have relevant results for this metric.

Bootstrap: 486.724s -> 486.349s (-0.08%)
Artifact size: 387.61 MiB -> 390.01 MiB (0.62%)

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. WG-trait-system-refactor The Rustc Trait System Refactor Initiative (-Znext-solver) labels Jul 21, 2026

rustbot assigned ShoyuVanilla Jul 21, 2026

adwinwhite commented Jul 21, 2026

View reviewed changes

lcnr reviewed Jul 21, 2026

View reviewed changes