cargo test fails with fatal runtime error on stable-aarch64-apple-darwin

[segeljakt]

Code

#[test]
fn test() {
    assert!(false);
}

Meta

rustc --version --verbose:

rustc 1.66.0 (69f9c33d7 2022-12-12)
binary: rustc
commit-hash: 69f9c33d71c871fc16ac445211281c6e7a340943
commit-date: 2022-12-12
host: aarch64-apple-darwin
release: 1.66.0
LLVM version: 15.0.2

Error output

helloworld ❯ RUST_BACKTRACE=1 cargo test
    Finished test [unoptimized + debuginfo] target(s) in 0.00s
     Running unittests src/main.rs (target/debug/deps/helloworld-ed0dadb8f7ac71f4)

running 1 test
fatal runtime error: failed to initiate panic, error 5
error: test failed, to rerun pass `--bin helloworld`

Caused by:
  process didn't exit successfully: `/Users/klasseg/Workspace/rust/helloworld/target/debug/deps/helloworld-ed0dadb8f7ac71f4` (signal: 6, SIGABRT: process abort signal)

No backtrace is printed.

The error is solved temporarily if I clean and rerun using:

cargo clean && env -i PATH="$HOME/.cargo/bin:/usr/bin:/bin" cargo test

[skyzh]

Is it related to #104388?

To verify, use rust-lldb to run your program, see if the abort happens in libunwind.

To fix, simply brew install llvm@15 and use llvm-15's lld as the linker.

[Qi-Hu]

Does this error only occur on aarch64-apple-darwin? I tried it on x86_64-unknown-linux-gnu and didn't see the error.

Code

cat tests/test.rs:

#[test]
fn test() {
    assert!(false);
}

Meta

rustc -vV:

rustc 1.72.0-dev
binary: rustc
commit-hash: unknown
commit-date: unknown
host: x86_64-unknown-linux-gnu
release: 1.72.0-dev
LLVM version: 16.0.4

Output

RUST_BACKTRACE=1 cargo test:

   Compiling helloworld v0.1.0 (/home/q00815166/helloworld)
    Finished test [unoptimized + debuginfo] target(s) in 0.24s
     Running unittests src/main.rs (target/debug/deps/helloworld-df59c9ac901802cb)

running 0 tests

test result: ok. 0 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s

     Running tests/test.rs (target/debug/deps/test-15c53ff575d885af)

running 1 test
test test ... FAILED

failures:

---- test stdout ----
thread 'test' panicked at 'assertion failed: false', tests/test.rs:3:5
stack backtrace:
   0: rust_begin_unwind
   1: core::panicking::panic_fmt
   2: core::panicking::panic
   3: test::test
             at ./tests/test.rs:3:5
   4: test::test::{{closure}}
             at ./tests/test.rs:2:11
   5: core::ops::function::FnOnce::call_once
             at /home/q00815166/rust/library/core/src/ops/function.rs:250:5
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.


failures:
    test

test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.02s

error: test failed, to rerun pass `--test test`

The test failed as expected.

[skyzh]

This only happens on macOS and the specific version of rustc.

[memark]

I can reproduce this on 1.71.0 as well as on nightly:

❯ rustc -vV
rustc 1.71.0 (8ede3aae2 2023-07-12)
binary: rustc
commit-hash: 8ede3aae28fe6e4d52b38157d7bfe0d3bceef225
commit-date: 2023-07-12
host: aarch64-apple-darwin
release: 1.71.0
LLVM version: 16.0.5

❯ rustc +nightly -vV
rustc 1.73.0-nightly (d12c6e947 2023-08-01)
binary: rustc
commit-hash: d12c6e947ceacf3b22c154caf9532b390d8dc88a
commit-date: 2023-08-01
host: aarch64-apple-darwin
release: 1.73.0-nightly
LLVM version: 16.0.5

Other info:

• Mac M1 Max
• MacOS Sonoma public beta 2

❯ sw_vers
ProductName:		macOS
ProductVersion:		14.0
BuildVersion:		23A5301h

❯ uname -a
Darwin MEMac.local 23.0.0 Darwin Kernel Version 23.0.0: Tue Jul 18 20:35:35 PDT 2023; root:xnu-10002.0.199.505.1~3/RELEASE_ARM64_T6000 arm64 arm Darwin

---

@skyzh, that fix works, thank you!

Can I help out with rust-lldb as you hinted? I'm not sure how to use it to run my test.

---

For anyone else coming to this thread:

• The llvm fix can be applied to all your rust projects by
• • running brew install llvm@15
• • adding build.rustflags = [ "-Clink-arg=-fuse-ld=/opt/homebrew/opt/llvm@15/bin/ld64.lld" ] to ~/.cargo/config.toml
• the cargo clean && env -i fix suggested above doesn't seem to work (anymore?)
• using cargo-nextest still gives the error, but doesn't crash the test runner and shows the expected info about which tests failed and so on