`-Znext-solver` Eager normalization outside of solver by adwinwhite · Pull Request #155767 · rust-lang/rust

adwinwhite · 2026-04-25T06:48:59Z

This PR adds a normalization routine for the next solver that behaves the same as the normalization in the old solver.
The new routine is used to normalize eagerly outside of the next solver.

This is part 2 of modifying the next solver to support eager normalization.

Those test changes are mostly wording changes, duplicating some errors or reducing some duplicates. But I could have overlooked something.
Notably it fixes the first, third and fourth examples in rust-lang/trait-system-refactor-initiative#191, but not the second variant.

It's probably easier to review commit by commit.

Fixes #151308
Fixes #101557
Fixes #119692
Fixes #136859

matthiaskrgr · 2026-04-25T07:06:45Z

does this fix #151308 ?

adwinwhite · 2026-04-25T07:33:54Z

does this fix #151308 ?

Yes, it probably fixes a bunch of issues. And unfix some issues fixed by the next solver? Unsure about that.
Seems tedious to re-triage those labelled fixed-by-the-next-solver 🤔

adwinwhite · 2026-04-25T10:27:22Z

@bors try @rust-timer queue

`-Znext-solver` Eager normalization outside of solver

rust-bors · 2026-04-25T12:40:01Z

☀️ Try build successful (CI)
Build commit: 59cc905 (59cc90511d0283377de9ba5a516f11853e3c95a9, parent: 7e0430fafc66bd0fdfaff94e65148a1f207438cd)

rust-timer · 2026-04-25T13:20:33Z

Finished benchmarking commit (59cc905): comparison URL.

Overall result: ❌✅ regressions and improvements - please read:

Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf.

Next, please: If you can, justify the regressions found in this try perf run in writing along with @rustbot label: +perf-regression-triaged. If not, fix the regressions and do another perf run. Neutral or positive results will clear the label automatically.

@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
Regressions ❌ (secondary)	2.2%	[0.5%, 7.3%]	12
Improvements ✅ (primary)	-0.3%	[-0.3%, -0.3%]	2
Improvements ✅ (secondary)	-2.4%	[-5.0%, -0.2%]	3
All ❌✅ (primary)	-0.3%	[-0.3%, -0.3%]	2

Max RSS (memory usage)

Results (primary -2.5%, secondary -1.6%)

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)	-	-	0
Improvements ✅ (primary)	-2.5%	[-2.8%, -2.2%]	2
Improvements ✅ (secondary)	-1.6%	[-2.2%, -1.1%]	2
All ❌✅ (primary)	-2.5%	[-2.8%, -2.2%]	2

Cycles

Results (secondary 3.9%)

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)	3.9%	[3.4%, 4.3%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 491.487s -> 483.564s (-1.61%)
Artifact size: 394.23 MiB -> 393.40 MiB (-0.21%)

adwinwhite · 2026-04-27T08:26:33Z

Better get some quick feedback before I go too far in wrong direction/details.

r? lcnr

adwinwhite · 2026-04-27T08:29:02Z

+pub struct Normalized<'tcx, T> {
+    pub value: T,
+    pub obligations: PredicateObligations<'tcx>,
+    pub has_ambig_hr_alias: bool,


Currently unused. Unsure how to propagate it out to be accessed. Many normalization routines have T -> T signature.

View changes since the review

lcnr

has_ambig_hr_alias does not feel useful to me, I think we can avoid it entirely
given that we allow for infer vars in our type, we need to handle both ty.has_aliases and ty.has_infer and need to shallow resolve at the start of fold_ty and fold_const

I would very much like to merge deeply_normalize and this one. The fact that deeply_normalize eagerly errors feels like an impl detail to me.

Instead of doing evaluate_all_error_on_ambiguity_stall_coroutine_predicates at every normalize_alias_term, we should be able to just use try_evaluate_obligations there, and then have the normalize_error_on_ambiguity which takes all the currently pending goals. puts em into another fulfillment context, and does the current evaluate_all_error_on_ambiguity_stall_coroutine_predicates for all normalization goals at once.

View changes since this review

rustbot · 2026-05-02T13:35:25Z

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.

adwinwhite · 2026-05-02T15:18:32Z

@bors r=lcnr rollup=never

rust-bors · 2026-05-02T15:18:35Z

📌 Commit 03cabe7 has been approved by lcnr

It is now in the queue for this repository.

rust-bors · 2026-05-03T00:09:20Z

☀️ Test successful - CI
Approved by: lcnr
Duration: 3h 12m 30s
Pushing a3e96d8 to main...

github-actions · 2026-05-03T00:12:36Z

What is this?

This is an experimental post-merge analysis report that shows differences in test outcomes between the merged PR and its parent PR.

Comparing 20de910 (parent) -> a3e96d8 (this PR)

Test differences

Show 96 test diffs

Stage 1

[crashes] tests/crashes/101557.rs: pass -> [missing] (J1)
[crashes] tests/crashes/119692.rs: pass -> [missing] (J1)
[crashes] tests/crashes/136859.rs: pass -> [missing] (J1)
[crashes] tests/crashes/155761-1.rs: [missing] -> pass (J1)
[ui] tests/ui/traits/next-solver/normalize/dont-ice-on-normalization-failure.rs: [missing] -> pass (J1)
[ui] tests/ui/traits/next-solver/object-projection-with-unsatisfied-bound-2.rs#current: pass -> [missing] (J1)
[ui] tests/ui/traits/next-solver/object-projection-with-unsatisfied-bound-2.rs#next: pass -> [missing] (J1)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-1.rs#next: [missing] -> pass (J1)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-1.rs#old: [missing] -> pass (J1)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-3.rs#next: [missing] -> pass (J1)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-3.rs#old: [missing] -> pass (J1)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-4.rs#next: [missing] -> pass (J1)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-4.rs#old: [missing] -> pass (J1)

Stage 2

[crashes] tests/crashes/101557.rs: pass -> [missing] (J0)
[crashes] tests/crashes/119692.rs: pass -> [missing] (J0)
[crashes] tests/crashes/136859.rs: pass -> [missing] (J0)
[crashes] tests/crashes/155761-1.rs: [missing] -> pass (J0)
[ui] tests/ui/traits/next-solver/normalize/dont-ice-on-normalization-failure.rs: [missing] -> pass (J2)
[ui] tests/ui/traits/next-solver/object-projection-with-unsatisfied-bound-2.rs#current: pass -> [missing] (J2)
[ui] tests/ui/traits/next-solver/object-projection-with-unsatisfied-bound-2.rs#next: pass -> [missing] (J2)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-1.rs#next: [missing] -> pass (J2)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-1.rs#old: [missing] -> pass (J2)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-3.rs#next: [missing] -> pass (J2)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-3.rs#old: [missing] -> pass (J2)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-4.rs#next: [missing] -> pass (J2)
[ui] tests/ui/traits/next-solver/self-referential-closure-sig-4.rs#old: [missing] -> pass (J2)

Additionally, 70 doctest diffs were found. These are ignored, as they are noisy.

Job group index

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard a3e96d8cb8a420022828d0a595277ad3671928a7 --output-dir test-dashboard

And then open test-dashboard/index.html in your browser to see an overview of all executed tests.

Job duration changes

x86_64-rust-for-linux: 40m 15s -> 51m 37s (+28.3%)
x86_64-gnu-debug: 2h 7m -> 1h 31m (-27.9%)
test-various: 1h 38m -> 2h 4m (+27.3%)
i686-gnu-nopt-1: 2h 23m -> 1h 46m (-26.2%)
armhf-gnu: 1h 11m -> 1h 29m (+25.3%)
dist-i686-msvc: 2h 23m -> 1h 50m (-23.1%)
dist-i686-linux: 1h 47m -> 1h 26m (-20.2%)
aarch64-apple: 3h 6m -> 2h 28m (-20.1%)
x86_64-msvc-ext2: 1h 47m -> 2h 4m (+15.6%)
x86_64-msvc-1: 2h 30m -> 2h 46m (+10.6%)

How to interpret the job duration changes?

Job durations can vary a lot, based on the actual runner instance
that executed the job, system noise, invalidated caches, etc. The table above is provided
mostly for t-infra members, for simpler debugging of potential CI slow-downs.

rust-log-analyzer · 2026-05-03T00:46:19Z

A job failed! Check out the build log: (web) (plain enhanced) (plain)

Click to see the possible cause of the failure (guessed by this bot)

  COMMIT_MESSAGE: cargo update 
##[endgroup]
Downloading single artifact
Preparing to download the following artifacts:
- Cargo-lock (ID: 6766890992, Size: 50828, Expected Digest: sha256:1560167d1a16d67a2f1416b177526966a86eced41c3b9ee920d34b06122aae13)
Redirecting to blob download url: https://productionresultssa1.blob.core.windows.net/actions-results/20adfd56-3407-4d52-8a39-1ee962432e4a/workflow-job-run-05aa044e-4ebb-5f79-ac30-3264d268ef08/artifacts/858175980ba8726bab9f963fd06060d5199d874bbd735e523e2d990d354f8c0b.zip
Starting download of artifact to: /home/runner/work/rust/rust
(node:2283) [DEP0005] DeprecationWarning: Buffer() is deprecated due to security and usability issues. Please use the Buffer.alloc(), Buffer.allocUnsafe(), or Buffer.from() methods instead.
(Use `node --trace-deprecation ...` to show where the warning was created)
SHA256 digest of downloaded artifact is 1560167d1a16d67a2f1416b177526966a86eced41c3b9ee920d34b06122aae13
Artifact download completed successfully.
Total of 1 artifact(s) downloaded
Download artifact has finished successfully
---
  COMMIT_MESSAGE: cargo update 
##[endgroup]
Downloading single artifact
Preparing to download the following artifacts:
- cargo-updates (ID: 6766891073, Size: 2920, Expected Digest: sha256:79e37795c23ea623d26e2efdd8aaa26674070d10fa45319c637d381547655d5f)
Redirecting to blob download url: https://productionresultssa1.blob.core.windows.net/actions-results/20adfd56-3407-4d52-8a39-1ee962432e4a/workflow-job-run-05aa044e-4ebb-5f79-ac30-3264d268ef08/artifacts/ce9034d4568b3430f906da1368ae1cf3e14ef1542859eb2e8d635c0876371f53.zip
Starting download of artifact to: /home/runner/work/rust/rust
(node:2294) [DEP0005] DeprecationWarning: Buffer() is deprecated due to security and usability issues. Please use the Buffer.alloc(), Buffer.allocUnsafe(), or Buffer.from() methods instead.
(Use `node --trace-deprecation ...` to show where the warning was created)
SHA256 digest of downloaded artifact is 79e37795c23ea623d26e2efdd8aaa26674070d10fa45319c637d381547655d5f
Artifact download completed successfully.
Total of 1 artifact(s) downloaded
Download artifact has finished successfully
##[group]Run echo "${COMMIT_MESSAGE}" > commit.txt
echo "${COMMIT_MESSAGE}" > commit.txt
cat cargo_update.log >> commit.txt

echo "${PR_MESSAGE}" > body.md
echo '```txt' >> body.md
cat cargo_update.log >> body.md
echo '```' >> body.md
shell: /usr/bin/bash --noprofile --norc -e -o pipefail {0}
---
dep-bumps
following is the output from `cargo update`:
  COMMIT_MESSAGE: cargo update 
##[endgroup]
##[group]Run git config user.name github-actions
git config user.name github-actions
git config user.email [email protected]
git switch --force-create cargo_update
git add ./Cargo.lock ./library/Cargo.lock ./src/tools/rustbook/Cargo.lock
git commit --no-verify --file=commit.txt
shell: /usr/bin/bash --noprofile --norc -e -o pipefail {0}
---
dep-bumps
following is the output from `cargo update`:
  COMMIT_MESSAGE: cargo update 
##[endgroup]
remote: error: GH013: Repository rule violations found for refs/heads/cargo_update.        
remote: Review all repository rules at https://github.com/rust-lang/rust/rules?ref=refs%2Fheads%2Fcargo_update        
remote: 
remote: - Cannot update this protected ref.        
remote: 
remote: - Changes must be made through a pull request.        
remote: 
remote: - Cannot force-push to this branch        
remote: 
To https://github.com/rust-lang/rust
 ! [remote rejected]   cargo_update -> cargo_update (push declined due to repository rule violations)
error: failed to push some refs to 'https://github.com/rust-lang/rust'
##[error]Process completed with exit code 1.
Post job cleanup.

rust-timer · 2026-05-03T00:53:30Z

Finished benchmarking commit (a3e96d8): comparison URL.

Overall result: ❌✅ regressions and improvements - please read:

Our benchmarks found a performance regression caused by this PR.
This might be an actual regression, but it can also be just noise.

Next Steps:

If the regression was expected or you think it can be justified,
please write a comment with sufficient written justification, and add
@rustbot label: +perf-regression-triaged to it, to mark the regression as triaged.
If you think that you know of a way to resolve the regression, try to create
a new PR with a fix for the regression.
If you do not understand the regression or you think that it is just noise,
you can ask the @rust-lang/wg-compiler-performance working group for help (members of this group
were already notified of this PR).

@rustbot label: +perf-regression
cc @rust-lang/wg-compiler-performance

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%]	2
Regressions ❌ (secondary)	4.6%	[0.1%, 11.4%]	8
Improvements ✅ (primary)	-0.4%	[-0.5%, -0.3%]	7
Improvements ✅ (secondary)	-1.6%	[-6.4%, -0.0%]	18
All ❌✅ (primary)	-0.2%	[-0.5%, 0.3%]	9

Max RSS (memory usage)

Results (primary -2.7%, secondary -0.1%)

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)	1.9%	[0.9%, 3.0%]	9
Improvements ✅ (primary)	-2.7%	[-2.7%, -2.7%]	1
Improvements ✅ (secondary)	-2.8%	[-6.1%, -0.5%]	7
All ❌✅ (primary)	-2.7%	[-2.7%, -2.7%]	1

Cycles

Results (secondary 1.1%)

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)	5.3%	[0.6%, 11.5%]	11
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-4.6%	[-8.1%, -0.6%]	8
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 482.455s -> 498.188s (3.26%)
Artifact size: 391.12 MiB -> 394.46 MiB (0.85%)

lcnr · 2026-05-04T08:02:46Z

non-next solver is mostly noise, the typenum regression is acceptable

panstromek · 2026-05-05T09:53:35Z

perf triage:

Marking as triaged based on the comment above. Note that new solver is not enabled by default. Diesel regression looks maybe real to me, but it's very tiny so either way it seems like an acceptable trade-off for a step forward in this work.

@rustbot label: +perf-regression-triaged

move generalization test The forth test of rust-lang/trait-system-refactor-initiative#191 (comment) isn't actually related to closure signature inference. closes rust-lang/trait-system-refactor-initiative#191, which has already been fixed by rust-lang#155767 r? types

This comment has been minimized.

Sign in to view

adwinwhite force-pushed the eager-norm-outside branch from 34677fc to e4b9746 Compare April 25, 2026 07:48

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Apr 25, 2026

Auto merge of #155767 - adwinwhite:eager-norm-outside, r=<try>

59cc905

`-Znext-solver` Eager normalization outside of solver

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Apr 25, 2026

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 Apr 25, 2026

adwinwhite commented Apr 27, 2026

View reviewed changes

Comment thread compiler/rustc_trait_selection/src/solve/normalize.rs Outdated

adwinwhite commented Apr 27, 2026

View reviewed changes

rustbot assigned lcnr Apr 27, 2026

lcnr reviewed Apr 27, 2026

View reviewed changes

Comment thread compiler/rustc_hir_analysis/src/check/check.rs Outdated

lcnr reviewed Apr 27, 2026

View reviewed changes

Comment thread compiler/rustc_next_trait_solver/src/solve/assembly/structural_traits.rs Outdated

Comment thread compiler/rustc_trait_selection/src/traits/coherence.rs Outdated

Comment thread compiler/rustc_hir_analysis/src/check/check.rs Outdated

lcnr reviewed Apr 27, 2026

View reviewed changes

Comment thread compiler/rustc_trait_selection/src/solve/normalize.rs Outdated

lcnr reviewed Apr 27, 2026

View reviewed changes

adwinwhite force-pushed the eager-norm-outside branch from e4b9746 to 6bd4abb Compare April 28, 2026 07:35

This comment has been minimized.

Sign in to view

adwinwhite force-pushed the eager-norm-outside branch from 6bd4abb to ffa59ce Compare April 28, 2026 08:37

lcnr reviewed Apr 28, 2026

View reviewed changes

Comment thread compiler/rustc_hir_analysis/src/autoderef.rs Outdated

lcnr reviewed Apr 28, 2026

View reviewed changes

Comment thread compiler/rustc_trait_selection/src/error_reporting/traits/fulfillment_errors.rs Outdated

adwinwhite added 2 commits May 2, 2026 21:07

bless diagnostics change

1c1372c

bless more than merely diagnostics change

03cabe7

adwinwhite force-pushed the eager-norm-outside branch from 2488143 to 03cabe7 Compare May 2, 2026 13:35

rust-bors Bot added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels May 2, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot added merged-by-bors This PR was explicitly merged by bors. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels May 3, 2026

rust-bors Bot merged commit a3e96d8 into rust-lang:main May 3, 2026
12 checks passed

rustbot added this to the 1.97.0 milestone May 3, 2026

khyperia mentioned this pull request May 4, 2026

generic_const_args: allow paths to non type consts #155341

Open

This was referenced May 5, 2026

we never try to normalize unused alias definitions rust-lang/trait-system-refactor-initiative#139

Open

move generalization test #156205

Open

Uh oh!

Conversation

adwinwhite commented Apr 25, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

matthiaskrgr commented Apr 25, 2026

Uh oh!

This comment has been minimized.

adwinwhite commented Apr 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

adwinwhite commented Apr 25, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented Apr 25, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Apr 25, 2026

Overall result: ❌✅ regressions and improvements - please read:

Uh oh!

Uh oh!

adwinwhite commented Apr 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

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

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

lcnr left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This comment has been minimized.

Uh oh!

Uh oh!

rustbot commented May 2, 2026

Uh oh!

adwinwhite commented May 2, 2026

Uh oh!

rust-bors Bot commented May 2, 2026

Uh oh!

This comment has been minimized.

rust-bors Bot commented May 3, 2026

Uh oh!

Uh oh!

github-actions Bot commented May 3, 2026

Test differences

Stage 1

Stage 2

Job duration changes

Uh oh!

rust-log-analyzer commented May 3, 2026

Uh oh!

rust-timer commented May 3, 2026

Overall result: ❌✅ regressions and improvements - please read:

Uh oh!

lcnr commented May 4, 2026

Uh oh!

panstromek commented May 5, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

adwinwhite commented Apr 25, 2026 •

edited by rustbot

Loading

adwinwhite commented Apr 25, 2026 •

edited

Loading

adwinwhite commented Apr 27, 2026 •

edited

Loading

adwinwhite Apr 27, 2026 •

edited by rustbot

Loading

lcnr left a comment •

edited by rustbot

Loading