Defer evaluating type system constants when they use infers or params #140553

BoxyUwU · 2025-05-01T15:05:28Z

Split out of #137972, the parts necessary for associated const equality and min generic const args to make progress and have correct semantics around when CTFE is invoked. According to a previous perf run of adding the new const_arg_kind query we should expect minor regressions here.

I think this is acceptable as we should be able to remove this query relatively soon once mgca is more complete as we'll then be able to implement GCE in terms of mgca and rip out GCEConst at which point it's trivial to determine what kind of anon const we're dealing with (either it has generics and is a repeat expr hack, or it doesnt and is a normal anon const).

This should only affect unstable code as we handle repeat exprs specially and those are the only kinds of type system consts that are allowed to make use of generic parameters.

Fixes #133066
Fixes #133199
Fixes #136894
Fixes #137813

r? compiler-errors

rustbot · 2025-05-01T15:05:34Z

This PR changes a file inside tests/crashes. If a crash was fixed, please move into the corresponding ui subdir and add 'Fixes #' to the PR description to autoclose the issue upon merge.

compiler/rustc_trait_selection/src/traits/mod.rs

compiler/rustc_middle/src/ty/consts.rs

compiler/rustc_middle/src/query/mod.rs

bors · 2025-05-06T14:52:06Z

☔ The latest upstream changes (presumably #140561) made this pull request unmergeable. Please resolve the merge conflicts.

compiler-errors

:>

compiler-errors · 2025-05-22T09:52:31Z

cool

@bors r+ rollup=never

bors · 2025-05-22T09:52:34Z

📌 Commit bb7e27a has been approved by compiler-errors

It is now in the queue for this repository.

BoxyUwU · 2025-05-22T11:16:35Z

@bors r- :<
I just remembered that we smuggle const patterns into ty::Const so I don't think:

This should only affect unstable code as we handle repeat exprs specially and those are the only kinds of type system consts that are allowed to make use of generic parameters.

is true. at the very least I want to make sure I understand why this doesnt break anything.. because I would probably expect it to

BoxyUwU · 2025-05-22T11:19:02Z

fn unevaluated_to_pat(
...
        // try to resolve e.g. associated constants to their definition on an impl, and then
        // evaluate the const.
        let valtree = match self.tcx.const_eval_resolve_for_typeck(typing_env, uv, self.span) {

We don't invoke type system norm on these constants though we just directly call into const eval resolve so this is fine. I.e. we won't break the following code:

trait Trait { const ASSOC: usize; }

impl<T> Trait for T {
    const ASSOC: usize = 10;
}

fn bar<T>(a: usize) {
    match a {
        <T as Trait>::ASSOC => (),
        _ => (),
    }
}

Where we try to evaluate <T as Trait>::ASSOC which this PR would reject if going through evaluate_const as it syntactically makes use of a generic parameter.

BoxyUwU · 2025-05-22T11:39:59Z

I think tests/ui/const-generics/const_eval_unchecked_doesnt_fire_patterns.rs would fail if we did go through type system norm instead of manually calling into ctfe so I don't need to add a test here 🤔

BoxyUwU · 2025-05-22T11:54:19Z

Added a test anyway because it seems nice to have one specifically for this case instead of only having a test that happens to fail but was meant for something else. @rustbot ready

(I checked locally what tests fail if I change us to use type system norm to evaluate patterns and also confirmed that the test I added fails if we do that)

compiler-errors · 2025-05-22T13:18:39Z

@bors r+

bors · 2025-05-22T13:18:41Z

📌 Commit fdccb42 has been approved by compiler-errors

It is now in the queue for this repository.

bors · 2025-05-23T05:30:49Z

⌛ Testing commit fdccb42 with merge 52bf0cf...

bors · 2025-05-23T08:43:30Z

☀️ Test successful - checks-actions
Approved by: compiler-errors
Pushing 52bf0cf to master...

github-actions · 2025-05-23T08:46:47Z

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 e7f4317 (parent) -> 52bf0cf (this PR)

Test differences

Show 16 test diffs

Stage 1

[crashes] tests/crashes/133199.rs: pass -> [missing] (J1)
[crashes] tests/crashes/136894.rs: pass -> [missing] (J1)
[crashes] tests/crashes/137813.rs: pass -> [missing] (J1)
[ui] tests/ui/const-generics/associated_const_equality/equality_bound_with_infer.rs: [missing] -> pass (J1)
[ui] tests/ui/const-generics/associated_const_equality/unconstrained_impl_param.rs: [missing] -> pass (J1)
[ui] tests/ui/const-generics/generic_const_exprs/cross-crate-2.rs: [missing] -> pass (J1)
[ui] tests/ui/const-generics/generic_const_exprs/serializing_error_guaranteed.rs: [missing] -> pass (J1)
[ui] tests/ui/pattern/unused-parameters-const-pattern.rs: [missing] -> pass (J1)

Stage 2

[ui] tests/ui/const-generics/associated_const_equality/equality_bound_with_infer.rs: [missing] -> pass (J0)
[ui] tests/ui/const-generics/associated_const_equality/unconstrained_impl_param.rs: [missing] -> pass (J0)
[ui] tests/ui/const-generics/generic_const_exprs/cross-crate-2.rs: [missing] -> pass (J0)
[ui] tests/ui/const-generics/generic_const_exprs/serializing_error_guaranteed.rs: [missing] -> pass (J0)
[ui] tests/ui/pattern/unused-parameters-const-pattern.rs: [missing] -> pass (J0)
[crashes] tests/crashes/133199.rs: pass -> [missing] (J2)
[crashes] tests/crashes/136894.rs: pass -> [missing] (J2)
[crashes] tests/crashes/137813.rs: pass -> [missing] (J2)

Job group index

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard 52bf0cf795dfecc8b929ebb1c1e2545c3f41d4c9 --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

dist-apple-various: 8665.5s -> 7002.1s (-19.2%)
dist-aarch64-msvc: 8288.7s -> 9669.4s (16.7%)
x86_64-apple-1: 7578.5s -> 6400.5s (-15.5%)
dist-i686-linux: 5886.9s -> 6326.4s (7.5%)
dist-aarch64-apple: 4998.1s -> 5338.5s (6.8%)
x86_64-mingw-2: 7090.8s -> 7553.1s (6.5%)
x86_64-gnu-llvm-19-3: 7313.1s -> 6873.7s (-6.0%)
x86_64-rust-for-linux: 2627.5s -> 2772.6s (5.5%)
x86_64-msvc-ext1: 7440.2s -> 7080.8s (-4.8%)
dist-ohos-aarch64: 4468.9s -> 4659.3s (4.3%)

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-timer · 2025-05-23T10:14:02Z

Finished benchmarking commit (52bf0cf): comparison URL.

Overall result: ❌ regressions - please read the text below

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

This is the most reliable metric that we have; it was used to determine the overall result at the top of this comment. However, even this metric can sometimes exhibit noise.

	mean	range	count
Regressions ❌ (primary)	0.3%	[0.1%, 0.5%]	15
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.3%	[0.1%, 0.5%]	15

Max RSS (memory usage)

Results (primary 0.2%, secondary 4.2%)

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)	1.1%	[0.7%, 2.0%]	5
Regressions ❌ (secondary)	4.2%	[2.3%, 7.8%]	3
Improvements ✅ (primary)	-1.3%	[-1.8%, -0.9%]	3
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.2%	[-1.8%, 2.0%]	8

Cycles

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

Binary size

Results (primary 0.4%, secondary 0.4%)

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)	0.4%	[0.0%, 1.1%]	120
Regressions ❌ (secondary)	0.4%	[0.0%, 1.1%]	22
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.4%	[0.0%, 1.1%]	120

Bootstrap: 778.581s -> 777.568s (-0.13%)
Artifact size: 365.44 MiB -> 365.56 MiB (0.03%)

Kobzol · 2025-05-23T10:18:28Z

The binary size regressions correspond to metadata getting larger, not actual binary size increases.

…mpiler-errors Defer evaluating type system constants when they use infers or params Split out of rust-lang#137972, the parts necessary for associated const equality and min generic const args to make progress and have correct semantics around when CTFE is invoked. According to a [previous perf run](https://perf.rust-lang.org/compare.html?start=93257e2d20809d82d1bc0fcc1942480d1a66d7cd&end=01b4cbf0f47c3f782330db88fa5ba199bba1f8a2&stat=instructions:u) of adding the new `const_arg_kind` query we should expect minor regressions here. I think this is acceptable as we should be able to remove this query relatively soon once mgca is more complete as we'll then be able to implement GCE in terms of mgca and rip out `GCEConst` at which point it's trivial to determine what kind of anon const we're dealing with (either it has generics and is a repeat expr hack, or it doesnt and is a normal anon const). This should only affect unstable code as we handle repeat exprs specially and those are the only kinds of type system consts that are allowed to make use of generic parameters. Fixes rust-lang#133066 Fixes rust-lang#133199 Fixes rust-lang#136894 Fixes rust-lang#137813 r? compiler-errors

rylev · 2025-05-27T13:41:39Z

@BoxyUwU @compiler-errors looking at some of the regressions in detail it does look like we're spending more time in generate_crate_metadata.

I see that perf regressions were expected. Are we fine with the magnitude seen here?

compiler-errors · 2025-05-27T13:42:50Z

Yeah, this is a consequence of encoding a new query into metadata. I believe that this perf regression is okay.

rustbot assigned compiler-errors May 1, 2025

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. labels May 1, 2025

compiler-errors reviewed May 1, 2025

View reviewed changes

compiler/rustc_trait_selection/src/traits/mod.rs Outdated Show resolved Hide resolved

compiler-errors reviewed May 1, 2025

View reviewed changes

compiler/rustc_middle/src/ty/consts.rs Show resolved Hide resolved

compiler/rustc_middle/src/ty/consts.rs Outdated Show resolved Hide resolved

compiler/rustc_middle/src/query/mod.rs Outdated Show resolved Hide resolved

This comment has been minimized.

Sign in to view

BoxyUwU force-pushed the defer_type_system_ctfe branch from e276dd1 to 42f9d9a Compare May 1, 2025 15:44

BoxyUwU mentioned this pull request May 1, 2025

Set groundwork for proper const normalization #140549

Merged

This comment has been minimized.

Sign in to view

BoxyUwU force-pushed the defer_type_system_ctfe branch from 42f9d9a to de3acb8 Compare May 1, 2025 16:21

fmease mentioned this pull request May 2, 2025

GCI: At their def site, actually wfcheck the where-clause & always eval free lifetime-generic constants #136429

Open

BoxyUwU force-pushed the defer_type_system_ctfe branch from de3acb8 to 2784c3c Compare May 6, 2025 15:36

BoxyUwU added 3 commits May 21, 2025 20:20

Introduce tcx.anon_const_kind query

996a185

Don't evaluate constants depending on infers or params

b4079c6

Fix tests

b1774b8

BoxyUwU force-pushed the defer_type_system_ctfe branch from 2784c3c to bb7e27a Compare May 21, 2025 19:28

compiler-errors approved these changes May 22, 2025

View reviewed changes

bors 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 May 22, 2025

bors 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-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels May 22, 2025

BoxyUwU force-pushed the defer_type_system_ctfe branch from bb7e27a to 4f2e93a Compare May 22, 2025 11:27

Review Comments

217c4ad

Add test/comment about const patterns with unused params

fdccb42

BoxyUwU force-pushed the defer_type_system_ctfe branch from 4f2e93a to fdccb42 Compare May 22, 2025 11:53

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 May 22, 2025

bors 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 May 22, 2025

bors added the merged-by-bors This PR was explicitly merged by bors. label May 23, 2025

bors merged commit 52bf0cf into rust-lang:master May 23, 2025
7 checks passed

rustbot added this to the 1.89.0 milestone May 23, 2025

rustbot added the perf-regression Performance regression. label May 23, 2025

tamird mentioned this pull request May 24, 2025

rustdoc query cycle regression #141492

Open

moxian mentioned this pull request May 26, 2025

ICE+errors: associated_const_equality: normalize_erasing_regions.rs: Failed to normalize Alias(Projection, _) #141489

Open

Defer evaluating type system constants when they use infers or params #140553

Defer evaluating type system constants when they use infers or params #140553

Uh oh!

Conversation

BoxyUwU commented May 1, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented May 1, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

This comment has been minimized.

This comment has been minimized.

bors commented May 6, 2025

Uh oh!

compiler-errors left a comment

Choose a reason for hiding this comment

Uh oh!

compiler-errors commented May 22, 2025

Uh oh!

bors commented May 22, 2025

Uh oh!

BoxyUwU commented May 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

BoxyUwU commented May 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

BoxyUwU commented May 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

BoxyUwU commented May 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

compiler-errors commented May 22, 2025

Uh oh!

bors commented May 22, 2025

Uh oh!

bors commented May 23, 2025

Uh oh!

bors commented May 23, 2025

Uh oh!

Uh oh!

github-actions bot commented May 23, 2025

Test differences

Stage 1

Stage 2

Job duration changes

Uh oh!

rust-timer commented May 23, 2025

Overall result: ❌ regressions - please read the text below

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

Kobzol commented May 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rylev commented May 27, 2025

Uh oh!

compiler-errors commented May 27, 2025

Uh oh!

Uh oh!

BoxyUwU commented May 1, 2025 •

edited

Loading

BoxyUwU commented May 22, 2025 •

edited

Loading

BoxyUwU commented May 22, 2025 •

edited

Loading

BoxyUwU commented May 22, 2025 •

edited

Loading

BoxyUwU commented May 22, 2025 •

edited

Loading

Kobzol commented May 23, 2025 •

edited

Loading