Re-enable assertions on macOS alt builds #146513

madsmtm · 2025-09-13T15:35:00Z

These were previously disabled, in part for performance reasons, in part due to needing availability symbols __isPlatformVersionAtLeast and __isOSVersionAtLeast that compiler-builtins did not provide, see #62592 (comment) and #134275 (comment) for failed checks.

Since #138944 though, std now provides these symbols, so we should be able to re-enable LLVM assertions, debug assertions and overflow checks.

Fixes #59637.

try-job: *apple*

madsmtm · 2025-09-13T15:35:08Z

@bors try

Re-enable assertions on macOS try-job: `*apple*`

rust-bors · 2025-09-13T17:35:51Z

💔 Test for beeae75 failed: CI. Failed jobs:

try - dist-apple-various-alt (web logs, extended logs)

madsmtm · 2025-09-13T19:55:20Z

Hmm, might need a stage0 bump, waiting for that:
@rustbot blocked

EDIT: #146636

madsmtm · 2025-09-27T09:07:15Z

@rustbot author
@bors try

rustbot · 2025-09-27T09:07:19Z

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

Re-enable assertions on macOS try-job: `*apple*`

madsmtm · 2025-09-27T09:08:15Z

@bors try cancel
I need to rebase first probably

rust-bors · 2025-09-27T09:08:18Z

Try build cancelled. Cancelled workflows:

https://github.com/rust-lang/rust/actions/runs/18057828350

madsmtm · 2025-09-27T09:10:38Z

@bors try

Re-enable assertions on macOS try-job: `*apple*`

rust-bors · 2025-09-27T12:06:25Z

💔 Test for 56e5725 failed: CI. Failed jobs:

try - aarch64-apple-alt (web logs, extended logs)

This wasn't caught by CI, because debug assertions aren't enabled there.

rust-bors · 2025-09-27T18:12:10Z

☀️ Try build successful (CI)
Build commit: 2ec5766 (2ec5766d21d2e578d3ba452d7d12814b21e8140b, parent: ade84871f718ea20a6460d28e82290353b4bf3d2)

madsmtm · 2025-09-27T18:31:12Z

So, I think this works!
@rustbot ready

r? @Mark-Simulacrum since you're on t-infra and were assigned to #59637 in the past.

Also CC @jieyouxu, since you've tried to do this in the past.

madsmtm · 2025-09-27T18:31:21Z

I'm unsure if it has a performance impact, comparing it against the auto build on 959b450 gives:

	Previously	This PR
CI Run	Link	Link
`dist-x86_64-apple`	1h 40m	1h 56m
`dist-apple-various`	1h 5m	1h 5m
`dist-aarch64-apple`	2h 2m	1h 37m
`aarch64-apple`	1h 49m	2h 3m

Which feels like wayy too high a variance to really be able to say anything conclusive?

rustbot · 2025-09-27T18:31:25Z

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.

nikic

I'm a bit confused. This appears to enable assertions on a bunch of dist targets, but distributed toolchains are not supposed to enable assertions.

Should this change be limited to only aarch64-apple, which is a test target and as such should enable assertions?

View changes since this review

madsmtm · 2025-09-27T19:31:32Z

I'm a bit confused. This appears to enable assertions on a bunch of dist targets, but distributed toolchains are not supposed to enable assertions.

Right, the dist- builds set DEPLOY. Only the -alt builds get debug assertions, overflow checks and LLVM assertions in here. So this PR as-is only actually affects aarch64-apple, the rest still have all the assertions disabled.

That said, it this will allow others to enable -alt if they find they need to test with debug assertions. I tested an -alt build in this CI run, it currently passes.

Mark-Simulacrum · 2025-09-28T14:26:42Z

@bors r+ rollup=never

This seems reasonable to me. Thanks! I agree that the CI times don't seem affected (at least within noise reasonable to achieve from a single PR's measurements).

bors · 2025-09-28T14:26:45Z

📌 Commit 9878be7 has been approved by Mark-Simulacrum

It is now in the queue for this repository.

bors · 2025-09-28T21:40:07Z

⌛ Testing commit 9878be7 with merge f957826...

bors · 2025-09-29T00:43:51Z

☀️ Test successful - checks-actions
Approved by: Mark-Simulacrum
Pushing f957826 to master...

github-actions · 2025-09-29T00:46:40Z

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 c8905ea (parent) -> f957826 (this PR)

Test differences

Show 35 test diffs

Stage 2

[codegen] tests/codegen-llvm/binary-heap-peek-mut-pop-no-panic.rs: pass -> ignore (ignored when std is built with debug assertions) (J0)
[codegen] tests/codegen-llvm/mem-replace-big-type.rs: pass -> ignore (ignored when std is built with debug assertions) (J0)
[codegen] tests/codegen-llvm/mem-replace-simple-type.rs: ignore (only executed when the architecture is x86_64 ((to not worry about usize differing))) -> ignore (ignored when std is built with debug assertions) (J0)
[codegen] tests/codegen-llvm/range-loop.rs: pass -> ignore (ignored when std is built with debug assertions) (J0)
[codegen] tests/codegen-llvm/slice-reverse.rs: ignore (only executed when the architecture is x86_64) -> ignore (ignored when std is built with debug assertions ((debug assertions prevent generating shufflevector))) (J0)
[codegen] tests/codegen-llvm/swap-small-types.rs: ignore (only executed when the architecture is x86_64) -> ignore (ignored when std is built with debug assertions ((ptr::swap_nonoverlapping has one which blocks some optimizations))) (J0)
[codegen] tests/codegen-llvm/vec-in-place.rs: pass -> ignore (ignored when std is built with debug assertions ((FIXME: checks for call detect scoped noalias metadata))) (J0)
[codegen] tests/codegen-llvm/vec-shrink-panik.rs: pass -> ignore (ignored when std is built with debug assertions ((plain old debug assertions))) (J0)
[codegen] tests/codegen-llvm/vec-with-capacity.rs: pass -> ignore (ignored when std is built with debug assertions) (J0)
[codegen] tests/codegen-llvm/vecdeque-drain.rs: pass -> ignore (ignored when std is built with debug assertions ((FIXME: checks for call detect scoped noalias metadata))) (J0)
[codegen] tests/codegen-llvm/vecdeque_no_panic.rs: pass -> ignore (ignored when std is built with debug assertions ((plain old debug assertions))) (J0)
[crashes] tests/crashes/108428.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[crashes] tests/crashes/118778.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[crashes] tests/crashes/118784.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[crashes] tests/crashes/120175.rs: ignore (ignored if rustc wasn't built with debug assertions) -> ignore (ignored when the target vendor is Apple ((raw-dylib doesn't work on Apple targets yet))) (J0)
[crashes] tests/crashes/123862.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[crashes] tests/crashes/130395.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[crashes] tests/crashes/133965.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[crashes] tests/crashes/134061.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[crashes] tests/crashes/137514.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[crashes] tests/crashes/139381.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[crashes] tests/crashes/139387.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[mir-opt] tests/mir-opt/pre-codegen/mem_replace.rs: pass -> ignore (ignored when std is built with debug assertions) (J0)
[mir-opt] tests/mir-opt/pre-codegen/ptr_offset.rs: pass -> ignore (ignored when std is built with debug assertions ((precondition checks are under cfg(debug_assertions)))) (J0)
[mir-opt] tests/mir-opt/pre-codegen/slice_iter.rs: pass -> ignore (ignored when std is built with debug assertions ((precondition checks on ptr::add are under cfg(debug_assertions)))) (J0)
[ui] tests/ui/associated-consts/issue-93775.rs#current: pass -> ignore (ignored when rustc is built with debug assertions) (J0)
[ui] tests/ui/associated-consts/issue-93775.rs#next: pass -> ignore (ignored when rustc is built with debug assertions) (J0)
[ui] tests/ui/errors/emitter-overflow-bad-whitespace.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[ui] tests/ui/mir/static-by-value-dyn.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[ui] tests/ui/mir/unreachable-loop-jump-threading.rs: ignore (ignored if rustc wasn't built with debug assertions) -> pass (J0)
[ui] tests/ui/print_type_sizes/niche-filling.rs: pass -> ignore (ignored when std is built with debug assertions ((debug assertions will print more types))) (J0)
registry::reach_max_unpack_size: [missing] -> pass (J0)

Additionally, 3 doctest diffs were found. These are ignored, as they are noisy.

Job group index

J0: aarch64-apple

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard f957826bff7a68b267ce75b1ea56352aed0cca0a --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

aarch64-apple: 5057.5s -> 9296.2s (83.8%)
dist-armhf-linux: 4979.7s -> 7862.3s (57.9%)
aarch64-gnu-llvm-20-2: 2690.6s -> 2166.5s (-19.5%)
pr-check-1: 1757.8s -> 1434.7s (-18.4%)
x86_64-gnu-llvm-20: 2792.4s -> 2384.4s (-14.6%)
x86_64-rust-for-linux: 3103.8s -> 2723.7s (-12.2%)
i686-gnu-2: 6382.0s -> 5607.5s (-12.1%)
x86_64-gnu-llvm-20-2: 5503.0s -> 6158.6s (11.9%)
pr-check-2: 2633.3s -> 2334.5s (-11.3%)
i686-gnu-1: 8167.3s -> 7354.5s (-10.0%)

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 · 2025-09-29T01:55:07Z

Finished benchmarking commit (f957826): comparison URL.

Overall result: ❌ regressions - no action needed

@rustbot label: -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)	0.0%	[0.0%, 0.0%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary -1.6%)

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

	mean	range	count
Regressions ❌ (primary)	0.5%	[0.5%, 0.5%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-3.6%	[-3.6%, -3.6%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-1.6%	[-3.6%, 0.5%]	2

Cycles

Results (primary 2.0%, secondary 2.4%)

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)	2.4%	[2.4%, 2.4%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.0%	[2.0%, 2.0%]	1

Binary size

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

Bootstrap: 470.874s -> 470.189s (-0.15%)
Artifact size: 387.66 MiB -> 387.63 MiB (-0.01%)

madsmtm · 2025-09-29T10:35:04Z

Okay, so this did change the timings fairly significantly:

aarch64-apple: 5057.5s -> 9296.2s (83.8%)

Is this within an acceptable threshold? Or should we revert this? Or maybe split the aarch64-apple CI step into two?

nikic · 2025-09-29T10:37:46Z

That result is probably not representative, as this change will have triggered an uncached LLVM rebuild.

madsmtm · 2025-09-29T10:38:39Z

Re perf regression, I think that's spurious, this is only affecting the macOS runners, which aren't perf-tested, and it shouldn't affect the dist- builds (which are the ones that perf would test, and those are still built without assertions).

madsmtm added O-macos Operating system: macOS T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. T-infra Relevant to the infrastructure team, which will review and decide on the PR/issue. A-CI Area: Our Github Actions CI labels Sep 13, 2025

rustbot added A-testsuite Area: The testsuite used to check the correctness of rustc S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Sep 13, 2025

rust-bors bot added a commit that referenced this pull request Sep 13, 2025

Auto merge of #146513 - madsmtm:apple-reenable-assertions, r=<try>

beeae75

Re-enable assertions on macOS try-job: `*apple*`

This comment has been minimized.

Sign in to view

rustbot added S-blocked Status: Blocked on something else such as an RFC or other implementation work. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Sep 16, 2025

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-blocked Status: Blocked on something else such as an RFC or other implementation work. labels Sep 27, 2025

This comment has been minimized.

Sign in to view

rust-bors bot added a commit that referenced this pull request Sep 27, 2025

Auto merge of #146513 - madsmtm:apple-reenable-assertions, r=<try>

772e7fa

Re-enable assertions on macOS try-job: `*apple*`

madsmtm force-pushed the apple-reenable-assertions branch from 2bdfca2 to b2c7b6b Compare September 27, 2025 09:09

rust-bors bot added a commit that referenced this pull request Sep 27, 2025

Auto merge of #146513 - madsmtm:apple-reenable-assertions, r=<try>

56e5725

Re-enable assertions on macOS try-job: `*apple*`

This comment has been minimized.

Sign in to view

Ignore crash test that doesn't crash on Apple platforms

07ed247

This wasn't caught by CI, because debug assertions aren't enabled there.

madsmtm force-pushed the apple-reenable-assertions branch from b2c7b6b to 25bedcd Compare September 27, 2025 12:41

madsmtm removed A-testsuite Area: The testsuite used to check the correctness of rustc T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Sep 27, 2025

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 27, 2025

rustbot assigned Mark-Simulacrum Sep 27, 2025

madsmtm marked this pull request as ready for review September 27, 2025 18:31

nikic reviewed Sep 27, 2025

View reviewed changes

Mark-Simulacrum changed the title ~~Re-enable assertions on macOS~~ Re-enable assertions on macOS alt builds Sep 28, 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 Sep 28, 2025

bors added the merged-by-bors This PR was explicitly merged by bors. label Sep 29, 2025

bors merged commit f957826 into rust-lang:master Sep 29, 2025
12 checks passed

rustbot added this to the 1.92.0 milestone Sep 29, 2025

Zalathar mentioned this pull request Sep 29, 2025

Fix some crash-test directives #147122

Merged

bors mentioned this pull request Sep 29, 2025

Make macOS dist build configuration match where reasonable #145883

Merged

madsmtm deleted the apple-reenable-assertions branch September 29, 2025 10:35

Re-enable assertions on macOS alt builds #146513

Re-enable assertions on macOS alt builds #146513

Uh oh!

Conversation

madsmtm commented Sep 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

madsmtm commented Sep 13, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors bot commented Sep 13, 2025

Uh oh!

madsmtm commented Sep 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

madsmtm commented Sep 27, 2025

Uh oh!

rustbot commented Sep 27, 2025

Uh oh!

This comment has been minimized.

madsmtm commented Sep 27, 2025

Uh oh!

rust-bors bot commented Sep 27, 2025

Uh oh!

madsmtm commented Sep 27, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors bot commented Sep 27, 2025

Uh oh!

rust-bors bot commented Sep 27, 2025

Uh oh!

madsmtm commented Sep 27, 2025

Uh oh!

madsmtm commented Sep 27, 2025

Uh oh!

rustbot commented Sep 27, 2025

Uh oh!

nikic 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!

madsmtm commented Sep 27, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Mark-Simulacrum commented Sep 28, 2025

Uh oh!

bors commented Sep 28, 2025

Uh oh!

bors commented Sep 28, 2025

Uh oh!

bors commented Sep 29, 2025

Uh oh!

Uh oh!

github-actions bot commented Sep 29, 2025

Test differences

Stage 2

Job duration changes

Uh oh!

rust-timer commented Sep 29, 2025

Overall result: ❌ regressions - no action needed

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

madsmtm commented Sep 29, 2025

Uh oh!

nikic commented Sep 29, 2025

Uh oh!

madsmtm commented Sep 29, 2025

Uh oh!

Uh oh!

madsmtm commented Sep 13, 2025 •

edited

Loading

madsmtm commented Sep 13, 2025 •

edited

Loading

nikic left a comment •

edited by rustbot

Loading

madsmtm commented Sep 27, 2025 •

edited

Loading