Skippable lints: account for lint capping and use in more places by joshtriplett · Pull Request #160126 · rust-lang/rust

joshtriplett · 2026-07-29T08:27:02Z

skippable_lints allows the compiler to skip doing some work if it only supports a lint that the compiler won't emit. However, skippable_lints didn't take --cap-lints into account. Since dependencies build with --cap-lints allow, it's taking that into account to allow skipping more code when building dependencies.

The rest of this PR adds some additional uses of skippable_lints to skip some potentially expensive operations.

All code in this PR is 100% human-written by me. I used an AI to search for potential performance issues when building an exceptionally large crate (aws-sdk-ec2).

rustbot · 2026-07-29T08:27:05Z

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

rustbot · 2026-07-29T08:27:07Z

r? @wesleywiser

rustbot has assigned @wesleywiser.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

Owners of files modified in this PR: compiler
compiler expanded to 74 candidates
Random selection from 16 candidates

joshtriplett · 2026-07-29T08:38:40Z

@bors try @rust-timer queue

Skippable lints: account for lint capping and use in more places

rust-bors · 2026-07-29T14:39:36Z

💥 Test timed out after 21600s

Dependencies get built with `--cap-lints allow`, so there's value in allowing capped lints to be skipped, even if the crate itself doesn't allow them.

`late_lint_mod` constructs `LateContext` immediately, before figuring out if the whole pass will be skipped. Move its construction into `late_lint_mod_inner`, so it doesn't get called if the pass gets skipped.

These lints will be skippable when building a crate with `--cap-lints allow`, which occurs for every dependency crate. In such cases, we can skip the whole query.

`-Zfuture-incompat-test` forces non-allow lints to report as future-compat. Unfortunately, rather than doing so by modifying the lints themselves, it requires checking separately. Check it and prevent skipping lints if enabled.

rustbot · 2026-07-31T00:30:50Z

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.

joshtriplett · 2026-07-31T00:31:43Z

I reverted the skip of check_mod_deathness due to the test failures; I think there may be a better fix that would allow that skip, but that can happen in a separate PR.

joshtriplett · 2026-07-31T02:24:58Z

@bors try @rust-timer queue

Skippable lints: account for lint capping and use in more places

rust-bors · 2026-07-31T04:34:11Z

☀️ Try build successful (CI)
Build commit: 9162c22 (9162c22b7da6084bf42d5c5306e048e574093367)
Base parent: 8ab9fdf (8ab9fdff5a91b9f2b5ed57fb0275452d9a0d0280)

rust-timer · 2026-07-31T05:15:54Z

Finished benchmarking commit (9162c22): 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 rustc-perf
@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)	0.2%	[0.1%, 0.2%]	2
Regressions ❌ (secondary)	0.2%	[0.1%, 0.3%]	6
Improvements ✅ (primary)	-0.3%	[-0.3%, -0.3%]	2
Improvements ✅ (secondary)	-0.8%	[-1.2%, -0.2%]	7
All ❌✅ (primary)	-0.1%	[-0.3%, 0.2%]	4

Max RSS (memory usage)

Results (primary -2.4%, secondary 1.3%)

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)	6.4%	[6.4%, 6.4%]	1
Improvements ✅ (primary)	-2.4%	[-2.4%, -2.4%]	1
Improvements ✅ (secondary)	-1.3%	[-1.9%, -0.7%]	2
All ❌✅ (primary)	-2.4%	[-2.4%, -2.4%]	1

Cycles

Results (primary -0.3%, secondary -0.6%)

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

	mean	range	count
Regressions ❌ (primary)	0.6%	[0.5%, 0.7%]	5
Regressions ❌ (secondary)	1.6%	[0.6%, 3.2%]	6
Improvements ✅ (primary)	-0.9%	[-2.3%, -0.6%]	6
Improvements ✅ (secondary)	-2.0%	[-4.5%, -0.5%]	9
All ❌✅ (primary)	-0.3%	[-2.3%, 0.7%]	11

Binary size

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

Bootstrap: 492.09s -> 489.291s (-0.57%)
Artifact size: 390.63 MiB -> 390.56 MiB (-0.02%)

joshtriplett · 2026-07-31T06:58:53Z

Perf looks somewhere between neutral and win to me.

The bigger win is when building dependencies, which uses --cap-lints allow. (Unfortunately, perf doesn't test that.) This change saved a couple of seconds on a build of a crate with a dependency on aws-sdk-ec2, because that crate has a massive number of items, and this allows skipping some walks on all those items.

Accounting for cap-lints also (for instance) allows crates that explicitly ask for warnings about missing docs to not do the work to produce those warnings when built as a dependency, and various other cases.

Kobzol · 2026-07-31T07:46:04Z

I tried something like this a few times, and my conclusion was that the remaining lints to be skipped are so fast that it's not worth it. But that of course depends on the exact crates that I tested it on.

lqd · 2026-07-31T08:20:29Z

I've also tried things like this (if you remember, Jakub), and similarly it didn't make much of a difference on many crates then, and only on clean builds of course. We may need to look for outliers on crates.io to see if there are others.

rust-bors · 2026-08-01T08:02:44Z

☔ The latest upstream changes (presumably #160310) made this pull request unmergeable. Please resolve the merge conflicts by rebasing.

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 Jul 29, 2026

rustbot assigned wesleywiser Jul 29, 2026

This comment has been minimized.

Sign in to view

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jul 29, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Jul 29, 2026

Auto merge of #160126 - joshtriplett:skippable-lints, r=<try>

b603093

Skippable lints: account for lint capping and use in more places

This comment has been minimized.

Sign in to view

rust-bors Bot added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Jul 29, 2026

rust-bors Bot removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Jul 29, 2026

This comment has been minimized.

Sign in to view

joshtriplett added 5 commits July 30, 2026 16:58

skippable_lints: Account for lint capping

84007dd

Dependencies get built with `--cap-lints allow`, so there's value in allowing capped lints to be skipped, even if the crate itself doesn't allow them.

late_lint_mod: Don't construct LateContext if skipping

d21e6f6

`late_lint_mod` constructs `LateContext` immediately, before figuring out if the whole pass will be skipped. Move its construction into `late_lint_mod_inner`, so it doesn't get called if the pass gets skipped.

liveness: use lints to avoid repeating their paths

2f02c41

check_liveness: Skip based on skippable_lints

cff5427

These lints will be skippable when building a crate with `--cap-lints allow`, which occurs for every dependency crate. In such cases, we can skip the whole query.

skippable_lints: Handle -Zfuture-incompat-test

da92bde

`-Zfuture-incompat-test` forces non-allow lints to report as future-compat. Unfortunately, rather than doing so by modifying the lints themselves, it requires checking separately. Check it and prevent skipping lints if enabled.

joshtriplett force-pushed the skippable-lints branch from a0268f6 to da92bde Compare July 31, 2026 00:30

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Jul 31, 2026

Auto merge of #160126 - joshtriplett:skippable-lints, r=<try>

9162c22

Skippable lints: account for lint capping and use in more places

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 Jul 31, 2026

joshtriplett mentioned this pull request Jul 31, 2026

[DRAFT] Skip check_mod_deathness when not needed #160252

Draft

Uh oh!

Uh oh!

Conversation

joshtriplett commented Jul 29, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Jul 29, 2026

Uh oh!

rustbot commented Jul 29, 2026

Uh oh!

joshtriplett commented Jul 29, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented Jul 29, 2026

Uh oh!

This comment has been minimized.

rustbot commented Jul 31, 2026

Uh oh!

joshtriplett commented Jul 31, 2026

Uh oh!

joshtriplett commented Jul 31, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented Jul 31, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Jul 31, 2026

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

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

joshtriplett commented Jul 31, 2026

Uh oh!

Kobzol commented Jul 31, 2026

Uh oh!

lqd commented Jul 31, 2026

Uh oh!

rust-bors Bot commented Aug 1, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

joshtriplett commented Jul 29, 2026 •

edited by rustbot

Loading