[Pkg-rust-maintainers] Bug#1105189: Bug#1105189: rust-rustix - autopkgtest failure on i386.

[Fabian Grünbichler]

On May 13, 2025 5:08:42 AM GMT+02:00, Peter Green <plugwash at debian.org> wrote:
>Package: rust-rustix
>Version: 0.38.37-2
>Severity: serious
>
>The autopkgtests for rust-rustix are failing on i386.
>
>> 1337s ---- backend::vdso::test_vdso stdout ----
>> 1337s 1337s thread 'backend::vdso::test_vdso' panicked at src/backend/linux_raw/vdso.rs:428:28:
>> 1337s called `Option::unwrap()` on a `None` value
>> 1337s stack backtrace:
>> 1337s    0: rust_begin_unwind
>> 1337s              at /usr/src/rustc-1.85.0/library/std/src/panicking.rs:692:5
>> 1337s    1: core::panicking::panic_fmt
>> 1337s              at /usr/src/rustc-1.85.0/library/core/src/panicking.rs:75:14
>> 1337s    2: core::panicking::panic
>> 1337s              at /usr/src/rustc-1.85.0/library/core/src/panicking.rs:145:5
>> 1337s    3: core::option::unwrap_failed
>> 1337s              at /usr/src/rustc-1.85.0/library/core/src/option.rs:2015:5
>> 1337s    4: core::option::Option<T>::unwrap
>> 1337s              at /usr/src/rustc-1.85.0/library/core/src/option.rs:978:21
>> 1337s    5: rustix::backend::vdso::test_vdso
>> 1337s              at ./src/backend/linux_raw/vdso.rs:428:16
>> 1337s    6: rustix::backend::vdso::test_vdso::{{closure}}
>> 1337s              at ./src/backend/linux_raw/vdso.rs:427:15
>> 1337s    7: core::ops::function::FnOnce::call_once
>> 1337s              at /usr/src/rustc-1.85.0/library/core/src/ops/function.rs:250:5
>> 1337s    8: core::ops::function::FnOnce::call_once
>> 1337s              at /usr/src/rustc-1.85.0/library/core/src/ops/function.rs:250:5
>> 1337s note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
>> 1337s 1337s 1337s failures:
>> 1337s     backend::vdso::test_vdso
>
>It looks like the test in question was newly added with the vdso patches.

I already started looking into this yesterday..

The test passes upstream on i386, I'll spend some more time
today seeing what the diff is and either fix it up or skip the test
on i386.