Some cleanups around fatal errors and delayed bugs by bjorn3 · Pull Request #162057 · rust-lang/rust

bjorn3 · 2026-08-31T11:24:40Z

With the aim of making it easier to move some compiler passes to different locations. And in particular to allow moving them across has_errors_or_delayed_bugs calls. This is a prerequisite for a local change that improves the effectiveness of cargo build pipelining.

A fatal error that doesn't actually abort is indistinguishable from a regular error. And the only places where emit_almost_fatal is called, the produced FatalError is ignored.

This avoids a delayed bug if compilation is aborted between checking function ABIs and codegening all functions.

And remove the encoded metadata if there are any errors or delayed bugs after encoding.

rustbot · 2026-08-31T11:24:43Z

rustc_codegen_gcc is developed in its own repository. If possible, consider making this change to rust-lang/rustc_codegen_gcc instead.

cc @antoyo, @GuillaumeGomez

rustbot · 2026-08-31T11:24:47Z

r? @mu001999

rustbot has assigned @mu001999.
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 75 candidates
Random selection from 18 candidates

mejrs · 2026-08-31T12:01:41Z

+error: values of the type `[&usize; usize::MAX]` are too big for the target architecture
+  --> $SRC_DIR/alloc/src/boxed.rs:LL:COL
+
 error[E0080]: values of the type `[&usize; usize::MAX]` are too big for the target architecture
  --> $SRC_DIR/core/src/mem/mod.rs:LL:COL


Only one of these errors has the E0080 code attached to it, can you make it consistent?

View changes since the review

E0080 indicates that a const failed to evaluate: https://doc.rust-lang.org/error_codes/E0080.html

A constant value failed to get evaluated.

Erroneous code example:

enum Enum { X = (1 << 500), Y = (1 / 0), }

This error indicates that the compiler was unable to sensibly evaluate a constant expression that had to be evaluated. Attempting to divide by 0 or causing an integer overflow are two ways to induce this error.

Ensure that the expressions given can be evaluated as the desired integer type.

See the Discriminants section of the Reference for more information about setting custom integer types on enums using the repr attribute.

The first error however doesn't happen during const eval at all. It happens when determining the function ABI.

mejrs · 2026-08-31T12:07:00Z

+                    if let Ok(data) = self.tcx.eval_static_initializer(def_id) {
+                        record!(self.tables.eval_static_initializer[def_id] <- data);
+                    }


Is removing this unwrap sound? ( alternatively, one could delay a bug in the err case)

View changes since the review

eval_static_initializer returns Result<_, ErrorGuranteed>.

It returns Result<_, ErrorHandled> and ErrorHandled has a variant without an ErrorGuaranteed. I don't know whether that could happen here.

Right, ErrorHandled::TooGeneric can't happen here as statics are never generic. But will change it to a delayed bug anyway just in case.

Thanks! I do occasionally see proposals for generic statics... so this delayed bug would be nice to have.

rust-log-analyzer · 2026-08-31T12:20:42Z

The job x86_64-gnu-llvm-21 failed! Check out the build log: (web) (plain enhanced) (plain)

Click to see the possible cause of the failure (guessed by this bot)

---- [ui] tests/ui/limits/issue-17913.rs stdout ----
Saved the actual 32bit.stderr to `/checkout/obj/build/x86_64-unknown-linux-gnu/test/ui/limits/issue-17913/issue-17913.32bit.stderr`
diff of 32bit.stderr:

+ error: values of the type `[&usize; usize::MAX]` are too big for the target architecture
+   --> $SRC_DIR/alloc/src/boxed.rs:LL:COL
+ 
1 error[E0080]: values of the type `[&usize; usize::MAX]` are too big for the target architecture
2   --> $SRC_DIR/core/src/mem/mod.rs:LL:COL
3    |

9 LL |     let a: Box<_> = Box::new([&n; SIZE]);
10    |                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
11 
- error: aborting due to 1 previous error
+ error: aborting due to 2 previous errors
13 
14 For more information about this error, try `rustc --explain E0080`.
15 

Note: some mismatched output was normalized before being compared
-   --> /rustc/FAKE_PREFIX/library/alloc/src/boxed.rs:290:4
+ error: values of the type `[&usize; usize::MAX]` are too big for the target architecture
+   --> $SRC_DIR/alloc/src/boxed.rs:LL:COL
+ 
+ error: aborting due to 2 previous errors


The actual 32bit.stderr differed from the expected 32bit.stderr
To update references, rerun the tests and pass the `--bless` flag
To only update this specific test, also pass `--test-args limits/issue-17913.rs`

error: 1 errors occurred comparing output.
status: exit status: 1
command: env -u RUSTC_LOG_COLOR RUSTC_ICE="0" RUST_BACKTRACE="short" "/checkout/obj/build/x86_64-unknown-linux-gnu/stage2/bin/rustc" "/checkout/tests/ui/limits/issue-17913.rs" "-Zsimulate-remapped-rust-src-base=/rustc/FAKE_PREFIX" "-Ztranslate-remapped-path-to-local-path=no" "-Z" "ignore-directory-in-diagnostics-source-blocks=/cargo" "-Z" "ignore-directory-in-diagnostics-source-blocks=/checkout/vendor" "--sysroot" "/checkout/obj/build/x86_64-unknown-linux-gnu/stage2" "--target=i686-unknown-linux-gnu" "--check-cfg" "cfg(test,FALSE)" "--error-format" "json" "--json" "future-incompat" "-Ccodegen-units=1" "-Zui-testing" "-Zdeduplicate-diagnostics=no" "-Zwrite-long-types-to-disk=no" "-Cstrip=debuginfo" "-C" "prefer-dynamic" "--out-dir" "/checkout/obj/build/x86_64-unknown-linux-gnu/test/ui/limits/issue-17913" "-Znext-solver=coherence" "-A" "unused" "-W" "unused_attributes" "-A" "internal_features" "-A" "incomplete_features" "-A" "unused_parens" "-A" "unused_braces" "-Crpath" "-Cdebuginfo=0" "-Lnative=/checkout/obj/build/i686-unknown-linux-gnu/native/rust-test-helpers" "-Clinker=x86_64-linux-gnu-gcc"
stdout: none
--- stderr -------------------------------
error: values of the type `[&usize; usize::MAX]` are too big for the target architecture
##[error]  --> /rustc/FAKE_PREFIX/library/alloc/src/boxed.rs:290:4

error[E0080]: values of the type `[&usize; usize::MAX]` are too big for the target architecture
##[error]  --> /rustc/FAKE_PREFIX/library/core/src/mem/mod.rs:1463:24
   |
   = note: evaluation of `<[&usize; usize::MAX] as std::mem::SizedTypeProperties>::SIZE` failed here

note: the above error was encountered while instantiating `fn Box::<[&usize; usize::MAX]>::new`
  --> /checkout/tests/ui/limits/issue-17913.rs:16:21
   |
LL |     let a: Box<_> = Box::new([&n; 0xFFFFFFFF_usize]);
   |                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

error: aborting due to 2 previous errors

For more information about this error, try `rustc --explain E0080`.

mu001999 · 2026-08-31T14:08:52Z

r? mejrs

mejrs

r=me with delayed bug added and ci green

View changes since this review

rust-bors · 2026-09-04T10:57:08Z

☔ The latest upstream changes (presumably #162284) made this pull request unmergeable. Please resolve the merge conflicts by rebasing.

bjorn3 added 4 commits August 31, 2026 12:16

Remove emit_almost_fatal

fa5fb0e

A fatal error that doesn't actually abort is indistinguishable from a regular error. And the only places where emit_almost_fatal is called, the produced FatalError is ignored.

Avoid fatal errors in raw-dylib handling

64041d6

Don't rely on codegen emitting fn_abi_of_* errors in abi checks

01b761d

This avoids a delayed bug if compilation is aborted between checking function ABIs and codegening all functions.

Avoid panic when metadata encoder is called with errors

8ae4868

And remove the encoded metadata if there are any errors or delayed bugs after encoding.

rustbot assigned mu001999 Aug 31, 2026

mejrs reviewed Aug 31, 2026

View reviewed changes

rustbot assigned mejrs and unassigned mu001999 Aug 31, 2026

mejrs approved these changes Sep 3, 2026

View reviewed changes

Uh oh!

Uh oh!

Conversation

bjorn3 commented Aug 31, 2026

Uh oh!

rustbot commented Aug 31, 2026

Uh oh!

rustbot commented Aug 31, 2026

Uh oh!

mejrs Aug 31, 2026 • 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!

bjorn3 Aug 31, 2026

Choose a reason for hiding this comment

Uh oh!

mejrs Aug 31, 2026 • 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!

bjorn3 Aug 31, 2026

Choose a reason for hiding this comment

Uh oh!

mejrs Aug 31, 2026

Choose a reason for hiding this comment

Uh oh!

bjorn3 Aug 31, 2026

Choose a reason for hiding this comment

Uh oh!

mejrs Sep 3, 2026

Choose a reason for hiding this comment

Uh oh!

rust-log-analyzer commented Aug 31, 2026

Uh oh!

mu001999 commented Aug 31, 2026

Uh oh!

mejrs 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!

rust-bors Bot commented Sep 4, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

mejrs Aug 31, 2026 •

edited by rustbot

Loading

mejrs Aug 31, 2026 •

edited by rustbot

Loading

mejrs left a comment •

edited by rustbot

Loading