Address libunwind.a inconsistency issues in the bootstrap program #144473

zeroomega · 2025-07-25T23:44:50Z

We noticed when building rustc multiple time in a roll, some files will not be consistent across the build despite the fact that they are built from same source under the same environment. This patch addresses the inconsistency issue we found on libunwind.a, by sorting the order of the files passed to the linker.

rustbot · 2025-07-25T23:44:54Z

r? @Mark-Simulacrum

rustbot has assigned @Mark-Simulacrum.
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

rustbot · 2025-07-25T23:44:56Z

This PR changes how LLVM is built. Consider updating src/bootstrap/download-ci-llvm-stamp.

src/bootstrap/src/core/build_steps/compile.rs

Mark-Simulacrum · 2025-07-27T23:08:16Z

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

+        let source_date_epoch_env = "SOURCE_DATE_EPOCH";
+        if let Ok(var) = std::env::var(&source_date_epoch_env) {
+            cargo.env(&source_date_epoch_env, &var);
+        }


Are we clearing the environment somewhere? I'd expect that passing through an env variable unchanged is the default?

I am not sure. But with out this passthrough, the SOURCE_DATE_EPOCH didn't appear in env arg list when building cargo when running SOURCE_DATE_EPOCH=timestamp python3 x.py install

I did a rebuild and removed the "SOURCE_DATE_EPOCH" pass through code I have. The build command is SOURCE_DATE_EPOCH=1735689600 python3 x.py install --config config4.toml --verbose 2>&1 |tee build4.log I. So the bootstrap program didn't copy its env args to the cargo. See the verbose log:

Building tool cargo (stage1 -> stage2, x86_64-unknown-linux-gnu) running: cd "/mnt/nvme_crypt/SRC/rust" && env -u MAKEFLAGS -u MFLAGS AR_x86_64_unknown_linux_gnu="/mnt/nvme_crypt/SRC/TQ/prebuilt/third_party/clang/linux-x64/bin/llvm-ar" CARGO_INCREMENTAL="0" CARGO_LOG="cargo::core::compiler::fingerprint=info" CARGO_PROFILE_RELEASE_CODEGEN_UNITS="1" CARGO_PROFILE_RELEASE_DEBUG="0" CARGO_PROFILE_RELEASE_DEBUG_ASSERTIONS="false" CARGO_PROFILE_RELEASE_OVERFLOW_CHECKS="false" CARGO_PROFILE_RELEASE_STRIP="false" CARGO_TARGET_DIR="/mnt/nvme_crypt/SRC/rust/build4/x86_64-unknown-linux-gnu/stage1-tools" CARGO_TARGET_X86_64_UNKNOWN_LINUX_GNU_LINKER="/mnt/nvme_crypt/SRC/TQ/prebuilt/third_party/clang/linux-x64/bin/clang++" CC_x86_64_unknown_linux_gnu="/mnt/nvme_crypt/SRC/TQ/prebuilt/third_party/clang/linux-x64/bin/clang" CFG_COMMIT_DATE="2025-07-09" CFG_COMMIT_HASH="eabb4cd923deb73e714f7ad3f5234d68ca284dbe" CFG_COMPILER_BUILD_TRIPLE="x86_64-unknown-linux-gnu" CFG_COMPILER_HOST_TRIPLE="x86_64-unknown-linux-gnu" CFG_RELEASE="1.90.0-nightly" CFG_RELEASE_CHANNEL="nightly" CFG_RELEASE_NUM="1.90.0" CFG_SHORT_COMMIT_HASH="eabb4cd92" CFG_VERSION="1.90.0-nightly (1c6de2150 2025-07-16)" CFG_VER_DATE="2025-07-16" CFG_VER_HASH="1c6de215099bbe33668de762f9591187f6c25eef" CFLAGS_x86_64_unknown_linux_gnu="" CXXFLAGS_x86_64_unknown_linux_gnu="" CXX_x86_64_unknown_linux_gnu="/mnt/nvme_crypt/SRC/TQ/prebuilt/third_party/clang/linux-x64/bin/clang++" DOC_RUST_LANG_ORG_CHANNEL="https://doc.rust-lang.org/nightly" LIBC_CHECK_CFG="1" LIBZ_SYS_STATIC="1" LZMA_API_STATIC="1" RANLIB_x86_64_unknown_linux_gnu="/mnt/nvme_crypt/SRC/TQ/prebuilt/third_party/clang/linux-x64/bin/llvm-ar s" REAL_LIBRARY_PATH_VAR="LD_LIBRARY_PATH" RUSTBUILD_NATIVE_DIR="/mnt/nvme_crypt/SRC/rust/build4/x86_64-unknown-linux-gnu/native" RUSTC="/mnt/nvme_crypt/SRC/rust/build4/bootstrap/debug/rustc" RUSTC_BOOTSTRAP="1" RUSTC_BREAK_ON_ICE="1" RUSTC_ERROR_METADATA_DST="/mnt/nvme_crypt/SRC/rust/build4/tmp/extended-error-metadata" RUSTC_HOST_FLAGS="--check-cfg=cfg(bootstrap) -Clinker=/mnt/nvme_crypt/SRC/TQ/prebuilt/third_party/clang/linux-x64/bin/clang++" RUSTC_INSTALL_BINDIR="bin" RUSTC_LIBDIR="/mnt/nvme_crypt/SRC/rust/build4/x86_64-unknown-linux-gnu/stage1/lib" RUSTC_LINK_STD_INTO_RUSTC_DRIVER="1" RUSTC_LINT_FLAGS="" RUSTC_REAL="/mnt/nvme_crypt/SRC/rust/build4/x86_64-unknown-linux-gnu/stage1/bin/rustc" RUSTC_SNAPSHOT="/mnt/nvme_crypt/SRC/rust/build4/x86_64-unknown-linux-gnu/stage1/bin/rustc" RUSTC_SNAPSHOT_LIBDIR="/mnt/nvme_crypt/SRC/rust/build4/x86_64-unknown-linux-gnu/stage1/lib" RUSTC_STAGE="1" RUSTC_SYSROOT="/mnt/nvme_crypt/SRC/rust/build4/x86_64-unknown-linux-gnu/stage1" RUSTC_TLS_MODEL_INITIAL_EXEC="1" RUSTC_VERBOSE="1" RUSTC_WRAPPER="/mnt/nvme_crypt/SRC/rust/build4/bootstrap/debug/rustc" RUSTDOC="/mnt/nvme_crypt/SRC/rust/build4/bootstrap/debug/rustdoc" RUSTDOCFLAGS="--cfg=windows_raw_dylib -Csymbol-mangling-version=v0 -Zunstable-options --check-cfg=cfg(bootstrap) --check-cfg=cfg(llvm_enzyme) --check-cfg=cfg(rust_analyzer) --crate-version 1.90.0-nightly\t(1c6de2150\t2025-07-16) -Clinker=/mnt/nvme_crypt/SRC/TQ/prebuilt/third_party/clang/linux-x64/bin/clang++" RUSTDOC_REAL="/path/to/nowhere/rustdoc/not/required" RUSTFLAGS="--cfg=windows_raw_dylib -Csymbol-mangling-version=v0 -Zunstable-options --check-cfg=cfg(bootstrap) --check-cfg=cfg(llvm_enzyme) --check-cfg=cfg(rust_analyzer) -Zmacro-backtrace -Csplit-debuginfo=off -Clink-arg=-L/mnt/nvme_crypt/SRC/rust/build4/x86_64-unknown-linux-gnu/llvm/lib\n -Clink-args=-Wl,-z,origin -Clink-args=-Wl,-rpath,$ORIGIN/../lib -Alinker-messages -Zunstable-options" RUST_TEST_THREADS="128" SYSROOT="/mnt/nvme_crypt/SRC/rust/build4/x86_64-unknown-linux-gnu/stage1" __CARGO_DEFAULT_LIB_METADATA="nightlytool-rustc1.90.0" "/mnt/nvme_crypt/SRC/rust/build4/x86_64-unknown-linux-gnu/stage0/bin/cargo" "build" "--target" "x86_64-unknown-linux-gnu" "-Zbinary-dep-depinfo" "-j" "128" "-Zroot-dir=/mnt/nvme_crypt/SRC/rust" "-v" "--release" "--manifest-path" "/mnt/nvme_crypt/SRC/rust/src/tools/cargo/Cargo.toml" "--features" "all-static" "--message-format" "json-render-diagnostics" (failure_mode=Exit)

I took a closer look. The cargo process to build cargo did inherited the SOURCE_DATE_EPOCH variable. It is the debug output didn't print it because it is not explicitly passed to the child process. Maybe worth improve the debug output to include the env args the child process inherited implicitly. It would make the debugging easier.
Anyway, I removed the environment variable handling in the latest patch set. Please take a look.

rustbot · 2025-07-29T21:55:18Z

Some changes occurred in src/tools/cargo

cc @ehuss

We noticed when building rustc multiple time in a roll, some files will not be consistent across the build despite the fact that they are built from same source under the same environment. This patch addresses the inconsistency issue we found on libunwind.a by sorting the order of the files passed to the linker.

zeroomega · 2025-07-31T18:12:06Z

@rustbot ready

Mark-Simulacrum · 2025-08-03T23:23:57Z

Can you fix the PR description + title to reflect the current state of the PR? Those both get persisted into the commit history so good for them to be accurate.

r=me with that done.

zeroomega · 2025-08-04T20:21:33Z

Done.
r=Mark-Simulacrum

Mark-Simulacrum · 2025-08-04T22:17:27Z

@bors r+

bors · 2025-08-04T22:17:29Z

📌 Commit b3f369d has been approved by Mark-Simulacrum

It is now in the queue for this repository.

…ark-Simulacrum Address libunwind.a inconsistency issues in the bootstrap program We noticed when building rustc multiple time in a roll, some files will not be consistent across the build despite the fact that they are built from same source under the same environment. This patch addresses the inconsistency issue we found on libunwind.a, by sorting the order of the files passed to the linker.

Rollup of 16 pull requests Successful merges: - #143857 (Port #[macro_export] to the new attribute parsing infrastructure) - #144133 (Stabilize const TypeId::of) - #144369 (Upgrade semicolon_in_expressions_from_macros from warn to deny) - #144443 (Make target pointer width in target json an integer) - #144473 (Address libunwind.a inconsistency issues in the bootstrap program) - #144659 (bootstrap: refactor mingw dist and fix gnullvm) - #144682 (Stabilize `strict_overflow_ops`) - #144794 (Port `#[coroutine]` to the new attribute system) - #144835 (Anonymize binders in tail call sig) - #144836 (Change visibility of Args new function) - #144900 (Stabilize `unsigned_signed_diff` feature) - #144903 (Rename `begin_panic_handler` to `panic_handler`) - #144910 (Add regression tests for seemingly fixed issues) - #144926 (Correct the use of `must_use` on btree::IterMut) - #144928 (Drop `rust-version` from `rustc_thread_pool`) - #144931 ([win][arm64ec] Fix msvc-wholearchive for Arm64EC) Failed merges: - #144914 (Add support for `ty::Instance` path shortening in diagnostics) r? `@ghost` `@rustbot` modify labels: rollup

…ark-Simulacrum Address libunwind.a inconsistency issues in the bootstrap program We noticed when building rustc multiple time in a roll, some files will not be consistent across the build despite the fact that they are built from same source under the same environment. This patch addresses the inconsistency issue we found on libunwind.a, by sorting the order of the files passed to the linker.

Rollup of 15 pull requests Successful merges: - #143857 (Port #[macro_export] to the new attribute parsing infrastructure) - #144133 (Stabilize const TypeId::of) - #144369 (Upgrade semicolon_in_expressions_from_macros from warn to deny) - #144473 (Address libunwind.a inconsistency issues in the bootstrap program) - #144659 (bootstrap: refactor mingw dist and fix gnullvm) - #144682 (Stabilize `strict_overflow_ops`) - #144794 (Port `#[coroutine]` to the new attribute system) - #144835 (Anonymize binders in tail call sig) - #144836 (Change visibility of Args new function) - #144900 (Stabilize `unsigned_signed_diff` feature) - #144903 (Rename `begin_panic_handler` to `panic_handler`) - #144910 (Add regression tests for seemingly fixed issues) - #144926 (Correct the use of `must_use` on btree::IterMut) - #144928 (Drop `rust-version` from `rustc_thread_pool`) - #144931 ([win][arm64ec] Fix msvc-wholearchive for Arm64EC) r? `@ghost` `@rustbot` modify labels: rollup

rustbot assigned Mark-Simulacrum Jul 25, 2025

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-bootstrap Relevant to the bootstrap subteam: Rust's build system (x.py and src/bootstrap) labels Jul 25, 2025

This comment has been minimized.

Sign in to view

Mark-Simulacrum reviewed Jul 27, 2025

View reviewed changes

zeroomega force-pushed the rustc_inconsistency branch from 2285c8c to 7515619 Compare July 29, 2025 21:55

This comment has been minimized.

Sign in to view

zeroomega force-pushed the rustc_inconsistency branch from 7515619 to 52bbc67 Compare July 29, 2025 22:01

This comment has been minimized.

Sign in to view

zeroomega force-pushed the rustc_inconsistency branch from 52bbc67 to 0792f7d Compare July 29, 2025 22:40

zeroomega force-pushed the rustc_inconsistency branch from 0792f7d to b3f369d Compare July 30, 2025 21:28

zeroomega changed the title ~~Address some rustc inconsistency issues~~ Address libunwind.a inconsistency issues in the bootstrap program Aug 4, 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 Aug 4, 2025

Zalathar mentioned this pull request Aug 5, 2025

Rollup of 16 pull requests #144942

Closed

Kobzol mentioned this pull request Aug 5, 2025

Rollup of 15 pull requests #144943

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Address libunwind.a inconsistency issues in the bootstrap program #144473

Address libunwind.a inconsistency issues in the bootstrap program #144473

zeroomega commented Jul 25, 2025 •

edited

Loading

Uh oh!

rustbot commented Jul 25, 2025

Uh oh!

rustbot commented Jul 25, 2025

Uh oh!

This comment has been minimized.

Uh oh!

Mark-Simulacrum Jul 27, 2025

Uh oh!

zeroomega Jul 28, 2025

Uh oh!

zeroomega Jul 29, 2025

Uh oh!

zeroomega Jul 30, 2025

Uh oh!

rustbot commented Jul 29, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

zeroomega commented Jul 31, 2025

Uh oh!

Mark-Simulacrum commented Aug 3, 2025

Uh oh!

zeroomega commented Aug 4, 2025

Uh oh!

Mark-Simulacrum commented Aug 4, 2025

Uh oh!

bors commented Aug 4, 2025

Uh oh!

Uh oh!

Address libunwind.a inconsistency issues in the bootstrap program #144473

Are you sure you want to change the base?

Address libunwind.a inconsistency issues in the bootstrap program #144473

Conversation

zeroomega commented Jul 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Jul 25, 2025

Uh oh!

rustbot commented Jul 25, 2025

Uh oh!

This comment has been minimized.

Uh oh!

Mark-Simulacrum Jul 27, 2025

Choose a reason for hiding this comment

Uh oh!

zeroomega Jul 28, 2025

Choose a reason for hiding this comment

Uh oh!

zeroomega Jul 29, 2025

Choose a reason for hiding this comment

Uh oh!

zeroomega Jul 30, 2025

Choose a reason for hiding this comment

Uh oh!

rustbot commented Jul 29, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

zeroomega commented Jul 31, 2025

Uh oh!

Mark-Simulacrum commented Aug 3, 2025

Uh oh!

zeroomega commented Aug 4, 2025

Uh oh!

Mark-Simulacrum commented Aug 4, 2025

Uh oh!

bors commented Aug 4, 2025

Uh oh!

Uh oh!

zeroomega commented Jul 25, 2025 •

edited

Loading