Eagerly normalize in generalize by jdonszelmann · Pull Request #151746 · rust-lang/rust

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please do so in sufficient writing along with @rustbot label: +perf-regression-triaged. If not, please fix the regressions and do another perf run. If its results are neutral or positive, the label will be automatically removed.

@bors rollup=never
@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)	0.2%	[0.2%, 0.3%]	5
Regressions ❌ (secondary)	0.3%	[0.1%, 0.4%]	11
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.3%	[-0.3%, -0.3%]	1
All ❌✅ (primary)	0.2%	[0.2%, 0.3%]	5

Max RSS (memory usage)

Results (primary 2.6%, secondary 0.7%)

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

	mean	range	count
Regressions ❌ (primary)	2.6%	[2.6%, 2.6%]	1
Regressions ❌ (secondary)	3.2%	[0.8%, 5.7%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-4.5%	[-4.5%, -4.5%]	1
All ❌✅ (primary)	2.6%	[2.6%, 2.6%]	1

Cycles

Results (primary 2.6%)

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

	mean	range	count
Regressions ❌ (primary)	2.6%	[2.6%, 2.7%]	2
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.6%	[2.6%, 2.7%]	2

Binary size

Results (primary 0.0%, secondary 0.0%)

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

	mean	range	count
Regressions ❌ (primary)	0.0%	[0.0%, 0.0%]	4
Regressions ❌ (secondary)	0.0%	[0.0%, 0.1%]	14
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.0%	[0.0%, 0.0%]	4

Bootstrap: 474.937s -> 474.9s (-0.01%)
Artifact size: 397.89 MiB -> 398.08 MiB (0.05%)

compiler/rustc_trait_selection/src/solve.rs

rustbot · 2026-02-26T13:57:25Z

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.

lcnr · 2026-02-26T17:53:43Z

compiler/rustc_infer/src/infer/relate/generalize.rs

    /// `<T as Bar<<?0 as Foo>::Assoc>::Assoc == ?0`. This equality can
    /// hold by either normalizing the outer or the inner associated type.
-    in_alias: bool,
+    state: GeneralizerState,


outdated comment, refer to the docs of GeneralizerState

lcnr · 2026-02-26T17:54:31Z

compiler/rustc_infer/src/infer/relate/generalize.rs

+#[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
+enum GeneralizerState {
+    /// Treat aliases as potentially normalizable.
+    Default,
+    IncompletelyRelateHigherRankedAlias,
+    /// Only one layer
+    ShallowStructurallyRelateAliases,
+    StructurallyRelateAliases,
+}


please add more context here

when are we in which state

how do they impact behavior

or maybe just refer to handle_alias_ty and rely on the actual control-flow in that function

lcnr · 2026-02-26T17:56:48Z

compiler/rustc_infer/src/infer/relate/generalize.rs

+
+                            match self.state {
+                                GeneralizerState::IncompletelyRelateHigherRankedAlias => {
+                                    if self.infcx.next_trait_solver()


i feel like it'd be easier for us to only ever be in IncompletelyRelateHigherRankedAlias if we're in the new solver, so that we don't have to check this. In the old solver Default and IncompletelyRelateHigherRankedAlias are equivalent, are they not?

lcnr · 2026-02-26T17:57:50Z

compiler/rustc_trait_selection/src/solve.rs

+    span: Span,
+    alias: ty::AliasTy<'tcx>,
+) -> Ty<'tcx> {
+    let infcx = unsafe {


do you want to encapsulate these transmute in inhernet functions on InferCtxt and add a safety comment there?

lcnr · 2026-02-26T18:02:53Z

tests/ui/impl-trait/unsized_coercion.rs

@@ -12,7 +13,6 @@ impl Trait for u32 {}
 fn hello() -> Box<impl Trait> {
    if true {
        let x = hello();
-        //[next]~^ ERROR: the size for values of type `dyn Trait` cannot be known at compilation time
        let y: Box<dyn Trait> = x;
    }


hmm, I don't fully get that 🤔

What does "implicitly contains the + Sized bound" mean? For coerce to consider that Sized bound, it has to be in the root fulfillment context I think?

How does that Sized bound get there. I am also surprised that x has to be equal to the return type of hello() instead of a subtype, how does that happen? ✨

rustbot assigned lcnr Jan 27, 2026

This comment has been minimized.

Sign in to view

rust-bors bot pushed a commit that referenced this pull request Jan 29, 2026

Auto merge of #151746 - jdonszelmann:eagerly-normalize-in-generalize,…

92bb985

… r=<try> Eagerly normalize in generalize

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jan 29, 2026

This comment has been minimized.

Sign in to view

jdonszelmann force-pushed the eagerly-normalize-in-generalize branch from 5bd20bb to d0cdaae Compare February 2, 2026 18:38

This comment has been minimized.

Sign in to view

jdonszelmann force-pushed the eagerly-normalize-in-generalize branch from d0cdaae to fb5d5cf Compare February 3, 2026 15:04

This comment has been minimized.

Sign in to view

jdonszelmann force-pushed the eagerly-normalize-in-generalize branch from fb5d5cf to d96cab0 Compare February 3, 2026 15:51

This comment has been minimized.

Sign in to view

lcnr reviewed Feb 4, 2026

View reviewed changes

compiler/rustc_trait_selection/src/solve.rs Outdated Show resolved Hide resolved

jdonszelmann force-pushed the eagerly-normalize-in-generalize branch from d96cab0 to 5611a89 Compare February 5, 2026 10:47

jdonszelmann marked this pull request as ready for review February 5, 2026 10:47

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Feb 5, 2026

jdonszelmann force-pushed the eagerly-normalize-in-generalize branch from 5611a89 to 3f90d08 Compare February 5, 2026 10:50

This comment has been minimized.

Sign in to view

rustbot added perf-regression Performance regression. and removed S-waiting-on-perf Status: Waiting on a perf run to be completed. labels Feb 6, 2026

jdonszelmann mentioned this pull request Feb 9, 2026

don't use env with infer vars #151871

Merged

jdonszelmann force-pushed the eagerly-normalize-in-generalize branch from 3f90d08 to fd20cab Compare February 9, 2026 11:45

This comment has been minimized.

Sign in to view

lcnr reviewed Feb 9, 2026

View reviewed changes

compiler/rustc_trait_selection/src/solve.rs Outdated Show resolved Hide resolved

jdonszelmann force-pushed the eagerly-normalize-in-generalize branch from fd20cab to 3bcdac1 Compare February 19, 2026 16:12

This comment has been minimized.

Sign in to view

jdonszelmann force-pushed the eagerly-normalize-in-generalize branch from 3bcdac1 to a335a41 Compare February 20, 2026 09:49

This comment has been minimized.

Sign in to view

jdonszelmann force-pushed the eagerly-normalize-in-generalize branch from a335a41 to 29ae136 Compare February 24, 2026 10:40

This comment has been minimized.

Sign in to view

jdonszelmann and others added 12 commits February 26, 2026 11:44

remove debug requirement from hooks

f033651

Regression test for trait-system-refactor#262

244b6d5

eagerly normalize during generalization

5d0863b

implement eager normalization in a fresh context during typeck

04f2c01

merge generalizer state and structurally relate aliases

917713b

explicitly provide type in transmute

7a123e8

normalize at the start of generalize if we can

f50591f

try generalizing if normalization isn't a tyvar

e8e3544

fixup span in obligation cause

2d411a0

bless some tests

6edf58f

inline into

9aa065b

address review

1adea68

jdonszelmann force-pushed the eagerly-normalize-in-generalize branch from 29ae136 to 1adea68 Compare February 26, 2026 13:57

lcnr reviewed Feb 26, 2026

View reviewed changes

Uh oh!

Conversation

jdonszelmann commented Jan 27, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment has been minimized.

This comment has been minimized.

lcnr commented Jan 29, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

rust-bors bot commented Jan 29, 2026

Uh oh!

lcnr commented Jan 30, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

Uh oh!

rustbot commented Feb 5, 2026

Uh oh!

jdonszelmann commented Feb 5, 2026

Uh oh!

This comment has been minimized.

jdonszelmann commented Feb 5, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Feb 6, 2026

Overall result: ❌✅ regressions and improvements - please read the text below

Uh oh!

This comment has been minimized.

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

rustbot commented Feb 26, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

jdonszelmann commented Jan 27, 2026 •

edited by rustbot

Loading