`sys::process::windows::tests::test_thread_handle` spuriously failed on aarch64-pc-windows-msvc

[jieyouxu]

Failed in #145997 (comment)

failures:

---- sys::process::windows::tests::test_thread_handle stdout ----

thread 'sys::process::windows::tests::test_thread_handle' (2448) panicked at library\std\src\sys\process\windows\tests.rs:47:5:
assertion failed: res.unwrap().is_some()
stack backtrace:
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
---- sys::process::windows::tests::test_thread_handle stdout end ----

failures:
    sys::process::windows::tests::test_thread_handle

[jieyouxu]

Could this be a genuine test failure, even though not very reproducible? 🤔

cc target maintainers:

@ChrisDenton @dpaoliello @lambdageek @sivadeilra @wesleywiser

[ChrisDenton]

Hm, yeah, that test seems like it's asking for spurious fails. It assumes the spawned process will exit within 100 milliseconds. Which is usually a safe assumption! But not if there's low resources. high contention etc (such as when running lots of tests in parallel in CI).

A simple fix would be to increase the sleep time. But I'd be more incline to question why try_wait is being used instead of wait. Maybe to avoid blocking indefinitely if something goes wrong. But there are better ways to guard against that.