give next-solver's `evaluate_const` its own `try_evaluate_const` logic by zedddie · Pull Request #161418 · rust-lang/rust

zedddie · 2026-08-20T18:22:40Z

Currently evaluating type-level constants is convoluted as code is split in a lot of scattered functions across compiler which makes it hard to reason about.

This PR refactors next-solver's CTFE entry point, only commit changing behavior is 8c2dc29, by replacing evaluate_const call to normalization goal which also regressed some diagnostics.

for context see #project-const-generics > evaluate_const is scary

r? BoxyUwU

rustbot · 2026-08-20T18:22:45Z

BoxyUwU is currently at their maximum review capacity.
They may take a while to respond.

BoxyUwU · 2026-09-08T11:37:10Z

+        typing_env: Self::TypingEnv,
+        ct: ty::AliasConst<'tcx>,
+        span: Self::Span,
+    ) -> Result<Result<Self::ValTree, Self::Ty>, Option<Self::ErrorGuaranteed>> {


@khyperia's right that this is a scary return type :3 I think we would ideally make this a bespoke enum and maybe even change the original const_eval_resolve_for_typeck query to return it 🤔

I think that can be a future PR tho

View changes since the review

BoxyUwU · 2026-09-08T11:38:45Z

@rustbot author

very coolio :3

rustbot · 2026-09-08T11:38:49Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

BoxyUwU · 2026-10-26T14:41:54Z

-                    self.evaluate_added_goals_and_make_canonical_response(Certainty::AMBIGUOUS)
-                }
+            ty::ConstKind::Alias(ty::IsRigid::No, _) => {
+                self.normalize(param_env, ty::Unnormalized::new(ct))?;


this PR so wonderfully aligns the implementation of ConstEvaluatable with my mental model of it which is just "it makes sure we always normalize const args"

View changes since the review

BoxyUwU

wondies :3 thanks zedddie

View changes since this review

BoxyUwU · 2026-10-26T14:44:14Z

r? khyperia

for a final review, but otherwise LGTM

khyperia · 2026-10-28T09:45:20Z

lovely! unforchies this duplicates so much code, very much looking forward to nuking the old solver and gce and all this duplicated stuff, otherwise this would be a bit spooky to me~

@bors r=khyperia,BoxyUwU,bit-aloo rollup

rust-bors · 2026-10-28T09:45:24Z

📌 Commit 3228bae has been approved by khyperia,BoxyUwU,bit-aloo

It is now in the queue for this repository.

…yperia,BoxyUwU,bit-aloo give next-solver's `evaluate_const` its own `try_evaluate_const` logic Currently evaluating type-level constants is convoluted as code is split in a lot of scattered functions across compiler which makes it hard to reason about. This PR refactors next-solver's CTFE entry point, only commit changing behavior is rust-lang@8c2dc29, by replacing `evaluate_const` call to normalization goal which also regressed some diagnostics. for context see [#project-const-generics > evaluate_const is scary](https://rust-lang.zulipchat.com/#narrow/channel/260443-project-const-generics/topic/.60evaluate_const.60.20is.20scary) r? BoxyUwU

…uwer Rollup of 22 pull requests Successful merges: - #165032 (stdarch subtree update) - #165072 (Submodule update of rustc-perf) - #144090 (Make `IoSlice` and `IoSliceMut` methods unstably const) - #162342 (More stdarch regression tests) - #163909 (std: Test the exit behavior of `abort` and `abort_immediate`) - #164766 (Unstably introduce MVP edition migration lint `unmarked_unchecked_type_aliases` together with bikeshed attribute `#[rustc_unchecked_type_alias]`) - #164916 (CI: rfl: move job forward to Linux v7.3) - #164924 (Remove `TypeKind` from `rustc_codegen_ssa`) - #164969 (perf: avoid allocating HashSet in is_distinct check on small SwitchInts) - #165083 (Gate `z` inline asm template modifier) - #165087 (Fix flaky CI test `named_temp_file_new_creates_file`) - #161131 (Check associated const binding types) - #161418 (give next-solver's `evaluate_const` its own `try_evaluate_const` logic) - #164315 (make powerpc `Complex` ABI GCC-compatible) - #164476 (Avoid ICE on None from prepare_vtable_segments) - #164981 (Reflection add TypeId::name) - #165003 (follow up for global-alloc-reentry test) - #165020 (Bless mir-opt panic-abort tests) - #165066 (skip max_repr_align in Layout::eq_abi) - #165080 (Syntactically allow C-variadics in parenthesized argument lists) - #165081 (move `#[rustc_allow_const_fn_unstable]` attr check to attr parsing stage) - #165082 (Remove stability note from `rust_splat`)

…yperia,BoxyUwU,bit-aloo give next-solver's `evaluate_const` its own `try_evaluate_const` logic Currently evaluating type-level constants is convoluted as code is split in a lot of scattered functions across compiler which makes it hard to reason about. This PR refactors next-solver's CTFE entry point, only commit changing behavior is rust-lang@8c2dc29, by replacing `evaluate_const` call to normalization goal which also regressed some diagnostics. for context see [#project-const-generics > evaluate_const is scary](https://rust-lang.zulipchat.com/#narrow/channel/260443-project-const-generics/topic/.60evaluate_const.60.20is.20scary) r? BoxyUwU

…uwer Rollup of 23 pull requests Successful merges: - #165032 (stdarch subtree update) - #165072 (Submodule update of rustc-perf) - #144090 (Make `IoSlice` and `IoSliceMut` methods unstably const) - #162342 (More stdarch regression tests) - #163909 (std: Test the exit behavior of `abort` and `abort_immediate`) - #164766 (Unstably introduce MVP edition migration lint `unmarked_unchecked_type_aliases` together with bikeshed attribute `#[rustc_unchecked_type_alias]`) - #164916 (CI: rfl: move job forward to Linux v7.3) - #164924 (Remove `TypeKind` from `rustc_codegen_ssa`) - #164969 (perf: avoid allocating HashSet in is_distinct check on small SwitchInts) - #165083 (Gate `z` inline asm template modifier) - #165087 (Fix flaky CI test `named_temp_file_new_creates_file`) - #165123 (Temporarily drop macos-26 jobs) - #161131 (Check associated const binding types) - #161418 (give next-solver's `evaluate_const` its own `try_evaluate_const` logic) - #164315 (make powerpc `Complex` ABI GCC-compatible) - #164476 (Avoid ICE on None from prepare_vtable_segments) - #164981 (Reflection add TypeId::name) - #165003 (follow up for global-alloc-reentry test) - #165020 (Bless mir-opt panic-abort tests) - #165066 (skip max_repr_align in Layout::eq_abi) - #165080 (Syntactically allow C-variadics in parenthesized argument lists) - #165081 (move `#[rustc_allow_const_fn_unstable]` attr check to attr parsing stage) - #165082 (Remove stability note from `rust_splat`)

…uwer Rollup of 22 pull requests Successful merges: - #165032 (stdarch subtree update) - #165123 (Temporarily drop macos-26 jobs) - #165072 (Submodule update of rustc-perf) - #144090 (Make `IoSlice` and `IoSliceMut` methods unstably const) - #162342 (More stdarch regression tests) - #163909 (std: Test the exit behavior of `abort` and `abort_immediate`) - #164766 (Unstably introduce MVP edition migration lint `unmarked_unchecked_type_aliases` together with bikeshed attribute `#[rustc_unchecked_type_alias]`) - #164916 (CI: rfl: move job forward to Linux v7.3) - #164924 (Remove `TypeKind` from `rustc_codegen_ssa`) - #164969 (perf: avoid allocating HashSet in is_distinct check on small SwitchInts) - #165083 (Gate `z` inline asm template modifier) - #165087 (Fix flaky CI test `named_temp_file_new_creates_file`) - #161131 (Check associated const binding types) - #161418 (give next-solver's `evaluate_const` its own `try_evaluate_const` logic) - #164315 (make powerpc `Complex` ABI GCC-compatible) - #164476 (Avoid ICE on None from prepare_vtable_segments) - #165003 (follow up for global-alloc-reentry test) - #165020 (Bless mir-opt panic-abort tests) - #165066 (skip max_repr_align in Layout::eq_abi) - #165080 (Syntactically allow C-variadics in parenthesized argument lists) - #165081 (move `#[rustc_allow_const_fn_unstable]` attr check to attr parsing stage) - #165082 (Remove stability note from `rust_splat`)

Rollup merge of #161418 - zedddie:evaluate_const-stuff, r=khyperia,BoxyUwU,bit-aloo give next-solver's `evaluate_const` its own `try_evaluate_const` logic Currently evaluating type-level constants is convoluted as code is split in a lot of scattered functions across compiler which makes it hard to reason about. This PR refactors next-solver's CTFE entry point, only commit changing behavior is 8c2dc29, by replacing `evaluate_const` call to normalization goal which also regressed some diagnostics. for context see [#project-const-generics > evaluate_const is scary](https://rust-lang.zulipchat.com/#narrow/channel/260443-project-const-generics/topic/.60evaluate_const.60.20is.20scary) r? BoxyUwU

rust-timer · 2026-10-30T02:31:38Z

Note

This PR was benchmarked as part of triage of its containing rollup: triage URL.

Finished benchmarking commit (876a209): 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%	[0.1%, 0.2%]	5
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.2%	[-0.3%, -0.2%]	16
All ❌✅ (primary)	0.2%	[0.1%, 0.2%]	5

Max RSS (memory usage)

Results (primary -2.2%)

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)	-2.2%	[-2.3%, -2.2%]	2
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.2%	[-2.3%, -2.2%]	2

Cycles

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

Binary size

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

Artifact size: 406.68 MiB -> 406.82 MiB (0.03%)

…uwer Rollup of 22 pull requests Successful merges: - rust-lang/rust#165032 (stdarch subtree update) - rust-lang/rust#165123 (Temporarily drop macos-26 jobs) - rust-lang/rust#165072 (Submodule update of rustc-perf) - rust-lang/rust#144090 (Make `IoSlice` and `IoSliceMut` methods unstably const) - rust-lang/rust#162342 (More stdarch regression tests) - rust-lang/rust#163909 (std: Test the exit behavior of `abort` and `abort_immediate`) - rust-lang/rust#164766 (Unstably introduce MVP edition migration lint `unmarked_unchecked_type_aliases` together with bikeshed attribute `#[rustc_unchecked_type_alias]`) - rust-lang/rust#164916 (CI: rfl: move job forward to Linux v7.3) - rust-lang/rust#164924 (Remove `TypeKind` from `rustc_codegen_ssa`) - rust-lang/rust#164969 (perf: avoid allocating HashSet in is_distinct check on small SwitchInts) - rust-lang/rust#165083 (Gate `z` inline asm template modifier) - rust-lang/rust#165087 (Fix flaky CI test `named_temp_file_new_creates_file`) - rust-lang/rust#161131 (Check associated const binding types) - rust-lang/rust#161418 (give next-solver's `evaluate_const` its own `try_evaluate_const` logic) - rust-lang/rust#164315 (make powerpc `Complex` ABI GCC-compatible) - rust-lang/rust#164476 (Avoid ICE on None from prepare_vtable_segments) - rust-lang/rust#165003 (follow up for global-alloc-reentry test) - rust-lang/rust#165020 (Bless mir-opt panic-abort tests) - rust-lang/rust#165066 (skip max_repr_align in Layout::eq_abi) - rust-lang/rust#165080 (Syntactically allow C-variadics in parenthesized argument lists) - rust-lang/rust#165081 (move `#[rustc_allow_const_fn_unstable]` attr check to attr parsing stage) - rust-lang/rust#165082 (Remove stability note from `rust_splat`)

JonathanBrouwer · 2026-11-03T10:36:22Z

@rustbot label: +perf-regression-triaged
Noise

rustbot assigned BoxyUwU Aug 20, 2026

This comment has been minimized.

Sign in to view

BoxyUwU mentioned this pull request Aug 21, 2026

check wf before const-evaluating #160237

Open

This comment has been minimized.

Sign in to view

BoxyUwU mentioned this pull request Aug 22, 2026

Full Const Generics rust-lang/goals#100

Open

zedddie force-pushed the evaluate_const-stuff branch from 00a3838 to 92b2dea Compare August 24, 2026 17:15

This comment has been minimized.

Sign in to view

zedddie force-pushed the evaluate_const-stuff branch from 92b2dea to 5a6ab62 Compare August 27, 2026 09:49

This comment has been minimized.

Sign in to view

zedddie force-pushed the evaluate_const-stuff branch from 5a6ab62 to c40c0cd Compare August 28, 2026 20:32

This comment has been minimized.

Sign in to view

zedddie force-pushed the evaluate_const-stuff branch from c40c0cd to 67b25fa Compare August 29, 2026 13:52

zedddie force-pushed the evaluate_const-stuff branch from 67b25fa to c814483 Compare September 8, 2026 03:49

zedddie changed the title ~~wip try to make evaluate_const less scary~~ give next-solver's evaluate_const its own try_evaluate_const logic Sep 8, 2026

zedddie marked this pull request as ready for review September 8, 2026 04:15

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 Sep 8, 2026

BoxyUwU reviewed Sep 8, 2026

View reviewed changes

Comment thread compiler/rustc_next_trait_solver/src/solve/eval_ctxt/mod.rs Resolved

BoxyUwU reviewed Sep 8, 2026

View reviewed changes

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Sep 8, 2026

This comment has been minimized.

Sign in to view

BoxyUwU mentioned this pull request Oct 2, 2026

prevent next-solver ICE when evaluating ill-formed associated consts #161537

Closed

zedddie force-pushed the evaluate_const-stuff branch from 2916ada to 3228bae Compare October 26, 2026 12:30

BoxyUwU reviewed Oct 26, 2026

View reviewed changes

BoxyUwU approved these changes Oct 26, 2026

View reviewed changes

rustbot assigned khyperia and unassigned BoxyUwU Oct 26, 2026

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-review Status: Awaiting review from the assignee but also interested parties. labels Oct 28, 2026

JonathanBrouwer mentioned this pull request Oct 28, 2026

Rollup of 22 pull requests #165095

Closed

JonathanBrouwer mentioned this pull request Oct 29, 2026

Rollup of 23 pull requests #165130

Closed

JonathanBrouwer mentioned this pull request Oct 29, 2026

Rollup of 22 pull requests #165151

Merged

rust-bors Bot merged commit f25bc4c into rust-lang:main Oct 29, 2026
14 checks passed

rustbot added this to the 1.101.0 milestone Oct 29, 2026

rustbot added the perf-regression Performance regression. label Oct 30, 2026

rustbot added the perf-regression-triaged The performance regression has been triaged. label Nov 3, 2026

Conversation

zedddie commented Aug 20, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Aug 20, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

Uh oh!

BoxyUwU Sep 8, 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!

BoxyUwU commented Sep 8, 2026

Uh oh!

rustbot commented Sep 8, 2026

Uh oh!

This comment has been minimized.

BoxyUwU Oct 26, 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!

BoxyUwU 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!

BoxyUwU commented Oct 26, 2026

Uh oh!

khyperia commented Oct 28, 2026

Uh oh!

rust-bors Bot commented Oct 28, 2026

Uh oh!

Uh oh!

rust-timer commented Oct 30, 2026

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

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

JonathanBrouwer commented Nov 3, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

zedddie commented Aug 20, 2026 •

edited by rustbot

Loading

BoxyUwU Sep 8, 2026 •

edited by rustbot

Loading

BoxyUwU Oct 26, 2026 •

edited by rustbot

Loading

BoxyUwU left a comment •

edited by rustbot

Loading