bootstrap: leverage cargo trim-paths by weihanglo · Pull Request #161049 · rust-lang/rust

weihanglo · 2026-08-13T13:08:46Z

What

Leverage Cargo's -Ztrim-paths to remap compiler and library. See rust-lang/cargo#17309. Bootstrap no longer needs custom remap.

Changes in toolchain distributions:

/rustc/<sha> remap does not change for toolchain binaries
/rustc-dev/<sha> remap does not change for toolchain private binaries
/rustc/llvm remap does not change for LLVM we distributed
crates.io dependency remaps becomes /cargo/registry/<hash> from /rust/deps
vendored dependency remaps becomes /cargo/deps from /rust/deps

This has some implicit dependencies:

cc-rs must be 1.3.0+ in rustc_llvm and other packages that depend on cc-rs.
Bootstrap stage0 must be cbae9b4 to get __CARGO_RUSTC_BOOTSTRAP_WS_REMAP support
Remaps only apply to the release and dev Cargo profiles.
It follows cargo's remap rules: https://doc.rust-lang.org/nightly/cargo/reference/unstable.html#profile-trim-paths-option

try-jobs: dist-x86_64-msvc
try-jobs: dist-x86_64-linux
try-jobs: dist-aarch64-apple

🤖 LLM disclosure: I used LLM to do the first pass of bootstrap integration and temporary stage0 bump. The usage was signed off here on Zulip: #llm-reviews > Experiment with Cargo's trim-path in bootstrap. See #161049 (comment) for a more recent disclosure.

weihanglo · 2026-08-13T13:09:50Z

@bors try

[experiment] bootstrap: leverage cargo trim-paths

weihanglo · 2026-08-13T13:21:21Z

@bors try cancel

rust-bors · 2026-08-13T13:21:25Z

Try build cancelled. Cancelled workflows:

https://github.com/rust-lang/rust/actions/runs/31703558440

weihanglo · 2026-08-13T13:21:52Z

@bors try jobs=dist-x86_64-msvc,dist-x86_64-linux,dist-aarch64-apple

[experiment] bootstrap: leverage cargo trim-paths try-job: dist-x86_64-msvc try-job: dist-x86_64-linux try-job: dist-aarch64-apple

rust-bors · 2026-08-13T15:40:30Z

💔 Test for bd743b7 failed: CI. Failed jobs:

try - dist-aarch64-apple (web logs, enhanced plaintext logs)
try - dist-x86_64-linux (web logs, enhanced plaintext logs)
try - dist-x86_64-msvc (web logs, enhanced plaintext logs)

weihanglo · 2026-08-13T18:09:47Z

@bors try jobs=dist-x86_64-msvc,dist-x86_64-linux,dist-aarch64-apple

[experiment] bootstrap: leverage cargo trim-paths try-job: dist-x86_64-msvc try-job: dist-x86_64-linux try-job: dist-aarch64-apple

rust-bors · 2026-08-13T21:14:31Z

☀️ Try build successful (CI)
Build commit: 67c6105 (67c6105718f0714463e9014eca86d60e96ba0b12)
Base parent: ba28ff7 (ba28ff76f353a722f31c4f3dd2ac4e437d36411b)

weihanglo · 2026-08-14T01:46:27Z

@bors try jobs=dist-x86_64-msvc,dist-x86_64-linux,dist-aarch64-apple

[experiment] bootstrap: leverage cargo trim-paths try-job: dist-x86_64-msvc try-job: dist-x86_64-linux try-job: dist-aarch64-apple

Urgau · 2026-09-01T21:51:07Z

Sure, I just installed the toolchain with rustup-toolchain-install-master b54c46dce1c5f83017e8be8f3ab20e07a05c215b --component rustc-dev.

And then tried to compile this dumb test program (rustc +b54c46dce1c5f83017e8be8f3ab20e07a05c215b a.rs):

#![feature(rustc_private)]

extern crate rustc_session;

fn main() {
    let sess = rustc_session::Session::timer(12u32);
}

Urgau · 2026-09-01T21:55:37Z

I manually looked at the debuginfo with llvm-dwarfdump of the standard library rlibs and didn't find any leaked paths for Rust source code.

There are however several leaks in compiler-builtins rlib:

DW_AT_decl_file ("/rustroot/lib/clang/23/include/stdint.h")

It also reproduces on nightly, so it's not a regression of this PR, and I don't think it's problematic, it's not like we are gonna try to un-remap C deps anyway.

Urgau

Looks good to me.

Thanks for all the hard work, in particular on the Cargo side.

I will let @jieyouxu do his review as well, in particular the bootstrap part.

View changes since this review

jieyouxu

Thanks

View changes since this review

jieyouxu · 2026-09-02T09:59:11Z

@bors r=Urgau,jieyouxu rollup=never note="path remapping logic"

rust-bors · 2026-09-02T09:59:14Z

📌 Commit 31e8b4b has been approved by Urgau,jieyouxu

It is now in the queue for this repository.

Replace `RUSTC_DEBUGINFO_MAP` and `RUSTC_CARGO_REGISTRY_SRC_TO_REMAP` with cargo trim-paths and `__CARGO_RUSTC_BOOTSTRAP_WS_REMAP` override (See cargo#17309). The `RUSTC_DEBUGINFO_MAP` into `-ffile-prefix-map` in rustc_llvm will be covered cc@1.3.0+ natively (which inherits and forwards `CARGO_TRIM_PATHS_REMAP`)

The current dwarfdump checks cannot see `.rmeta` leaks: * The compiler keeps unremapped local paths in metadata unless the remap scope is `all` (see issue 159621) * std ships metadata as separate `.rmeta` via `-Zembed-metadata=no` so the leak does not even appear in the rlibs This commit enhances to also check rmeta files.

Registry dependencies are now remapped by cargo trim-paths as `/cargo/registry/{source-hash}/{pkg}-{ver}/`

The `-fdebug-prefix-map` flags in `cc_unhandled_cflags` served three kinds of consumers. * The cc-rs-driven C/C++ builds inside cargo: They now inherit the same remap pairs from cargo trim-paths so passing the flag through `CFLAGS` there is redundant. * The CMake-driven LLVM build: This is the one we need the remaps. * The remaining callers (`compiler_file` probing, cc detection, test fixtures): They never produce distributed artifacts.

rust-bors · 2026-09-02T13:42:45Z

⚠️ A new commit 8d43ad3441784c2be601b5ec97bc9bca18c889b3 was pushed.

This pull request was unapproved.

rustbot · 2026-09-02T13:42:45Z

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.

Urgau · 2026-09-02T14:55:08Z

@bors r=Urgau,jieyouxu rollup=never note="path remapping logic"

rust-bors · 2026-09-02T14:55:11Z

📌 Commit 8d43ad3 has been approved by Urgau,jieyouxu

It is now in the queue for this repository.

rust-bors · 2026-09-02T20:12:22Z

☀️ Test successful - CI
Approved by: Urgau,jieyouxu
Duration: 3h 13m 31s
Pushing fb9a338 to main...

github-actions · 2026-09-02T20:15:39Z

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 824336a (parent) -> fb9a338 (this PR)

Test differences

Show 3 test diffs

Stage 2

[run-make] tests/run-make/compressed-debuginfo-zstd: pass -> ignore (ignored if LLVM wasn't build with zstd for ELF section compression or LLVM is not the default codegen backend) (J0)

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

Job group index

J0: x86_64-gnu-nopt

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard fb9a3389fdeb347c7658a422ed41b79499434452 --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

x86_64-gnu-gcc-core-tests: 14m 13s -> 23m 33s (+65.7%)
x86_64-rust-for-linux: 46m 13s -> 1h 7m (+46.2%)
dist-x86_64-netbsd: 1h 3m -> 1h 32m (+44.6%)
dist-s390x-linux: 1h 4m -> 1h 31m (+42.5%)
x86_64-gnu-aux: 1h 57m -> 2h 45m (+41.4%)
pr-check-2: 41m 26s -> 56m 26s (+36.2%)
x86_64-gnu-llvm-21-1: 43m 50s -> 59m 37s (+36.0%)
test-various: 1h 53m -> 2h 31m (+33.4%)
optional-x86_64-gnu-autodiff: 38m 7s -> 49m 34s (+30.0%)
x86_64-gnu-stdlib-semver-check: 17m 54s -> 23m 2s (+28.7%)

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.

Kobzol · 2026-09-02T21:10:09Z

[run-make] tests/run-make/compressed-debuginfo-zstd: pass -> ignore (ignored if LLVM wasn't build with zstd for ELF section compression or LLVM is not the default codegen backend) (J0)

The test diff is curious. Why would this PR cause compiletest think that LLVM doesn't have zstd or isn't the default codegen backend anymore 🤔

Kobzol · 2026-09-02T21:11:30Z

Ah, maybe that was just because the LLVM stamp was bumped, so the test used a locally built LLVM without zstd (because zstd isn't installed on the test runner), interesting. This is essentially the same issue that we had with GCC and binutils (CC @antoyo just to let you know that we also run into this issue with LLVM).

rust-timer · 2026-09-02T21:14:57Z

Finished benchmarking commit (fb9a338): comparison URL.

Overall result: no relevant changes - BENCHMARK(S) FAILED

@rustbot label: -perf-regression

❗ ❗ ❗ ❗ ❗
Warning ⚠️: The following benchmark(s) failed to build:

Job failure

❗ ❗ ❗ ❗ ❗

cc @rust-lang/wg-compiler-performance

Instruction count

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

Max RSS (memory usage)

Results (primary 4.4%)

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

	mean	range	count
Regressions ❌ (primary)	4.4%	[4.4%, 4.4%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	4.4%	[4.4%, 4.4%]	1

Cycles

Results (primary 2.5%, 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)	2.5%	[2.5%, 2.5%]	1
Regressions ❌ (secondary)	2.9%	[2.6%, 3.1%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.5%	[2.5%, 2.5%]	1

Binary size

Results (primary 0.0%, secondary 0.1%)

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

	mean	range	count
Regressions ❌ (primary)	0.0%	[0.0%, 0.0%]	3
Regressions ❌ (secondary)	0.1%	[0.1%, 0.1%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.0%	[0.0%, 0.0%]	3

Bootstrap: missing data
Artifact size: 401.09 MiB -> 401.19 MiB (0.03%)

Kobzol · 2026-09-02T21:15:45Z

The failure is not caused by this PR.. sigh.

jieyouxu · 2026-09-03T08:08:25Z

Did the merge status on this PR... also get lost?

Kobzol · 2026-09-03T08:09:33Z

What merge status?

jieyouxu · 2026-09-03T08:13:41Z

What merge status?

Normally merged PRs have this:

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Aug 13, 2026

Auto merge of #161049 - weihanglo:bootstrap-trim-paths-exp, r=<try>

e23bdd9

[experiment] bootstrap: leverage cargo trim-paths

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Aug 13, 2026

Auto merge of #161049 - weihanglo:bootstrap-trim-paths-exp, r=<try>

bd743b7

[experiment] bootstrap: leverage cargo trim-paths try-job: dist-x86_64-msvc try-job: dist-x86_64-linux try-job: dist-aarch64-apple

weihanglo mentioned this pull request Aug 13, 2026

Support custom remapping rules with trim-paths for bootstrap rust-lang/cargo#17309

Closed

This comment has been minimized.

Sign in to view

weihanglo force-pushed the bootstrap-trim-paths-exp branch from b0b31e6 to 9981c62 Compare August 13, 2026 16:17

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Aug 13, 2026

Auto merge of #161049 - weihanglo:bootstrap-trim-paths-exp, r=<try>

67c6105

[experiment] bootstrap: leverage cargo trim-paths try-job: dist-x86_64-msvc try-job: dist-x86_64-linux try-job: dist-aarch64-apple

weihanglo force-pushed the bootstrap-trim-paths-exp branch from 9981c62 to 21aed96 Compare August 13, 2026 22:52

This comment has been minimized.

Sign in to view

weihanglo force-pushed the bootstrap-trim-paths-exp branch from 21aed96 to 80d0ecc Compare August 14, 2026 01:45

rustbot added the A-LLVM Area: Code generation parts specific to LLVM. Both correctness bugs and optimization-related issues. label Aug 14, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Aug 14, 2026

Auto merge of #161049 - weihanglo:bootstrap-trim-paths-exp, r=<try>

f892cbb

[experiment] bootstrap: leverage cargo trim-paths try-job: dist-x86_64-msvc try-job: dist-x86_64-linux try-job: dist-aarch64-apple

Urgau approved these changes Sep 1, 2026

View reviewed changes

weihanglo mentioned this pull request Sep 2, 2026

Tracking Issue for trim-paths RFC 3127 rust-lang/cargo#12137

Open

18 tasks

jieyouxu approved these changes Sep 2, 2026

View reviewed changes

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 Sep 2, 2026

Kobzol reviewed Sep 2, 2026

View reviewed changes

Comment thread src/bootstrap/src/core/builder/cargo.rs

weihanglo added 6 commits September 2, 2026 09:42

tests: check $CARGO_HOME leaks

6085af6

tests: switch remap prefix expectation to /cargo/registry

e751b8d

Registry dependencies are now remapped by cargo trim-paths as `/cargo/registry/{source-hash}/{pkg}-{ver}/`

bootstrap(llvm): touch download-ci-llvm-stamp

8d43ad3

This comment has been minimized.

Sign in to view

Uh oh!

Uh oh!

Conversation

weihanglo commented Aug 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What

Uh oh!

weihanglo commented Aug 13, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

weihanglo commented Aug 13, 2026

Uh oh!

rust-bors Bot commented Aug 13, 2026

Uh oh!

weihanglo commented Aug 13, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented Aug 13, 2026

Uh oh!

This comment has been minimized.

weihanglo commented Aug 13, 2026

Uh oh!

This comment has been minimized.

rust-bors Bot commented Aug 13, 2026

Uh oh!

This comment has been minimized.

weihanglo commented Aug 14, 2026

Uh oh!

This comment has been minimized.

Urgau commented Sep 1, 2026

Uh oh!

Urgau commented Sep 1, 2026

Uh oh!

Urgau 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!

jieyouxu 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!

jieyouxu commented Sep 2, 2026

Uh oh!

rust-bors Bot commented Sep 2, 2026

Uh oh!

Uh oh!

rust-bors Bot commented Sep 2, 2026

Uh oh!

rustbot commented Sep 2, 2026

Uh oh!

Urgau commented Sep 2, 2026

Uh oh!

rust-bors Bot commented Sep 2, 2026

Uh oh!

This comment has been minimized.

rust-bors Bot commented Sep 2, 2026

Uh oh!

github-actions Bot commented Sep 2, 2026

Test differences

Stage 2

Job duration changes

Uh oh!

Kobzol commented Sep 2, 2026

Uh oh!

Kobzol commented Sep 2, 2026

Uh oh!

rust-timer commented Sep 2, 2026

Overall result: no relevant changes - BENCHMARK(S) FAILED

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

Kobzol commented Sep 2, 2026

Uh oh!

jieyouxu commented Sep 3, 2026

Uh oh!

Kobzol commented Sep 3, 2026

weihanglo commented Aug 13, 2026 •

edited

Loading

Urgau left a comment •

edited by rustbot

Loading

jieyouxu left a comment •

edited by rustbot

Loading