Add a check for impossible predicates to trivial_const by saethlin · Pull Request #156934 · rust-lang/rust

saethlin · 2026-05-25T19:45:29Z

The problem here is that trivial consts bypass the MIR pass which replaces bodies with unreachable when there are false global bounds. see #147721 (comment).

This fixes the problem, but it is a bit hacky. But maybe all the handling of false global bounds is hacky?

rustbot · 2026-05-25T19:45:33Z

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

rustbot · 2026-05-25T19:45:35Z

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

cjgillot · 2026-05-30T22:24:28Z

@bors try @rust-timer queue

Add a check for impossible predicates to trivial_const

rust-bors · 2026-05-31T00:39:20Z

☀️ Try build successful (CI)
Build commit: 131aad2 (131aad2325ddc63f4d13d56c4362a59aafa966ba, parent: f8a08b688cbe60acc386ed1fbd1b7cbaaf5576b1)

rust-timer · 2026-05-31T01:55:23Z

Finished benchmarking commit (131aad2): comparison URL.

Overall result: ❌ regressions - 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)	0.4%	[0.2%, 0.5%]	11
Regressions ❌ (secondary)	0.4%	[0.0%, 0.7%]	13
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.2%	[-2.2%, -2.2%]	1
All ❌✅ (primary)	0.4%	[0.2%, 0.5%]	11

Max RSS (memory usage)

Results (secondary 2.7%)

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)	2.7%	[2.7%, 2.7%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

Results (primary 2.0%, secondary 2.9%)

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

	mean	range	count
Regressions ❌ (primary)	2.0%	[2.0%, 2.0%]	1
Regressions ❌ (secondary)	3.8%	[2.4%, 5.0%]	7
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.0%	[-3.0%, -3.0%]	1
All ❌✅ (primary)	2.0%	[2.0%, 2.0%]	1

Binary size

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

Bootstrap: 511.354s -> 510.996s (-0.07%)
Artifact size: 400.80 MiB -> 400.70 MiB (-0.03%)

saethlin · 2026-05-31T20:49:25Z

IMO the perf overhead is fine, all that the "regression" means is that the trival_const optimization is now a touch less of an improvement, see this perf run from the initial implementation: #148040 (comment)

BoxyUwU · 2026-06-07T22:04:55Z

false global bounds continue to make me unhappy 😆 in the long term we hopefully can cleanup this special case but for now it seems reasonable enough to me.

note that impossible_bounds uses the new solver internally even on stable. this seems reasonable enough to me and probably doesn't warrant wider sign off? cc @rust-lang/initiative-trait-system-refactor

can you add something to the PR description which links to #147721 specifically helix's comment about trivial consts bypassing the existing jank around false global bounds on const items

saethlin · 2026-06-07T23:51:48Z

Updated. Feel free to just modify the PR description yourself, I don't really understand type system implementation context here I just identified where to insert the right hack to make a test case pass.

BoxyUwU · 2026-06-08T22:36:06Z

        return None;
    }

+    if crate::impossible_predicates::has_impossible_predicates(tcx, def.into()) {


Suggested change

if crate::impossible_predicates::has_impossible_predicates(tcx, def.into()) {

// If there are impossible predicates then MIR passes will replace the body with

// `unreachable` causing const eval errors when trying to evaluate the body. For

// now we avoid using trivial consts for such bodies so that the behaviour doesn't

// change.

if crate::impossible_predicates::has_impossible_predicates(tcx, def.into()) {

though, writing this out, your logic here just straight up looks at the MIR so I'm surprised there'd wind up being a divergence? are you looking at different MIR than what const eval actually executes? if so why :3

are you looking at different MIR than what const eval actually executes? if so why :3

Yes. The entire idea of trivial_const is to bypass the numerous MIR passes and queries that are done on MIR bodies to lower consts, because in the common case all those passes and queries just add compile time.

It would probably make sense for trivial_const to intercept const lowering before MIR is built.

lcnr · 2026-06-09T12:20:12Z

note that impossible_bounds uses the new solver internally even on stable. this seems reasonable enough to me and probably doesn't warrant wider sign off? cc @rust-lang/initiative-trait-system-refactor

yeah, we already use this when building vtables, so relying on this more sgtm

saethlin · 2026-06-09T22:07:18Z

@bors squash

* Add a check for impossible predicates to trivial_const * Improve comment on impossible predicates check Co-authored-by: Boxy <rust@boxyuwu.dev>

rust-bors · 2026-06-09T22:07:40Z

🔨 2 commits were squashed into 8dc9f9a.

BoxyUwU · 2026-06-10T19:49:14Z

@bors r+ rollup=never (perf sensitive)

rust-bors · 2026-06-10T19:49:17Z

📌 Commit 8dc9f9a has been approved by BoxyUwU

It is now in the queue for this repository.

Add a check for impossible predicates to trivial_const The problem here is that trivial consts bypass the MIR pass which replaces bodies with `unreachable` when there are false global bounds. see #147721 (comment). This fixes the problem, but it is a bit hacky. But maybe all the handling of false global bounds is hacky?

JonathanBrouwer · 2026-06-11T07:18:09Z

@bors yield
Stuck again, hasn't printed anything in hours

rust-bors · 2026-06-11T07:18:13Z

Auto build was cancelled. Cancelled workflows:

https://github.com/rust-lang/rust/actions/runs/27320112205

The next pull request likely to be tested is #157739.

Add a check for impossible predicates to trivial_const The problem here is that trivial consts bypass the MIR pass which replaces bodies with `unreachable` when there are false global bounds. see #147721 (comment). This fixes the problem, but it is a bit hacky. But maybe all the handling of false global bounds is hacky?

rust-bors · 2026-06-12T16:04:51Z

💔 Test for 5f9f26e failed: CI. Failed job:

auto - dist-i686-mingw (web logs, enhanced plaintext logs)

JonathanBrouwer · 2026-06-12T16:07:05Z

@bors retry

rust-log-analyzer · 2026-06-12T16:30:45Z

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)

  WIX: /d/a/rust/rust/wix
##[endgroup]
    Updating crates.io index
error: failed to get `simd-adler32` as a dependency of package `miniz_oxide v0.8.8`
    ... which satisfies dependency `miniz_oxide = "^0.8.5"` of package `flate2 v1.1.9`
    ... which satisfies dependency `flate2 = "^1.1.9"` of package `citool v0.1.0 (D:\a\rust\rust\src\ci\citool)`

Caused by:
  failed to load source for dependency `simd-adler32`

Caused by:

JonathanBrouwer · 2026-06-12T17:20:42Z

@bors p=6
Scheduling this before the next rollup

rust-bors · 2026-06-12T20:35:16Z

☀️ Test successful - CI
Approved by: BoxyUwU
Duration: 3h 13m 42s
Pushing 3bdd7f8 to main...

github-actions · 2026-06-12T20:38:08Z

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 a1e52fc (parent) -> 3bdd7f8 (this PR)

Test differences

Show 4 test diffs

Stage 1

[ui] tests/ui/traits/trivial-const-with-impossible-bounds.rs: [missing] -> pass (J0)

Stage 2

[ui] tests/ui/traits/trivial-const-with-impossible-bounds.rs: [missing] -> pass (J1)

Additionally, 2 doctest diffs were found. These are ignored, as they are noisy.

Job group index

J0: x86_64-gnu-llvm-21-3, x86_64-gnu-llvm-22-3
J1: aarch64-apple, aarch64-gnu, aarch64-gnu-llvm-21-1, aarch64-msvc-1, arm-android, armhf-gnu, dist-i586-gnu-i586-i686-musl, i686-gnu-1, i686-gnu-nopt-1, i686-msvc-1, optional-x86_64-gnu-parallel-frontend, test-various, x86_64-gnu, x86_64-gnu-debug, x86_64-gnu-gcc, x86_64-gnu-llvm-21, x86_64-gnu-llvm-21-2, x86_64-gnu-llvm-22-2, x86_64-gnu-nopt, x86_64-gnu-stable, x86_64-mingw-1, x86_64-msvc-1

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard 3bdd7f86fcecc0ea65d79e62e39a70c8a2aabf15 --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

pr-check-2: 40m 49s -> 27m 14s (-33.3%)
dist-apple-various: 1h 29m -> 1h 56m (+30.0%)
i686-msvc-1: 3h 16m -> 2h 19m (-28.6%)
i686-gnu-2: 1h 15m -> 1h 32m (+21.5%)
dist-x86_64-netbsd: 1h 27m -> 1h 8m (-21.3%)
pr-check-1: 27m 28s -> 21m 49s (-20.5%)
arm-android: 1h 21m -> 1h 35m (+17.9%)
aarch64-apple: 3h 25m -> 2h 50m (-17.1%)
x86_64-gnu-nopt: 2h 37m -> 2h 16m (-13.6%)
x86_64-gnu-llvm-21: 1h 12m -> 1h 21m (+13.0%)

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 · 2026-06-12T21:16:57Z

Finished benchmarking commit (3bdd7f8): comparison URL.

Overall result: ❌ regressions - 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)	0.4%	[0.2%, 0.5%]	10
Regressions ❌ (secondary)	0.4%	[0.2%, 0.7%]	10
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.4%	[0.2%, 0.5%]	10

Max RSS (memory usage)

Results (secondary -3.3%)

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)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.3%	[-6.3%, -1.5%]	3
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary 4.2%)

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)	4.2%	[4.2%, 4.2%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 519.113s -> 519.494s (0.07%)
Artifact size: 400.82 MiB -> 400.81 MiB (-0.00%)

panstromek · 2026-06-15T15:44:04Z

perf triage:

This regression was justified per-merge in #156934 (comment)

@rustbot label: +perf-regression-triaged

…BoxyUwU Add a check for impossible predicates to trivial_const The problem here is that trivial consts bypass the MIR pass which replaces bodies with `unreachable` when there are false global bounds. see rust-lang#147721 (comment). This fixes the problem, but it is a bit hacky. But maybe all the handling of false global bounds is hacky?

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 25, 2026

rustbot assigned BoxyUwU May 25, 2026

saethlin mentioned this pull request May 25, 2026

Generic patterns and generic array lengths can call const fn with impossible generics #147721

Open