Unable to run any test in VSCode

rust-analyzer version: rust-analyzer version: 0.4.2270-standalone

rustc version: rustc 1.86.0-nightly (6067b3631 2025-01-17)

editor or extension: vscode extension 0.4.2270

relevant settings: set in settings.json:

{
    "terminal.integrated.env.linux": {
        "PATH": "/workspaces/pyo3/venv/bin:/home/codespace/.cargo/bin"
    },
    "debug.allowBreakpointsEverywhere": true,
    "rust-analyzer.server.extraEnv": {
        "CARGO": "/home/codespace/.cargo/bin/cargo",
        "RUSTC": "/home/codespace/.cargo/bin/rustc"
    },
    "rust-analyzer.cargo.extraEnv": {"PATH": "/home/codespace/.cargo/bin"}
}

repository link (if public, optional): clin1234/pyo3

code snippet to reproduce:

#[test]
#[cfg(not(Py_LIMITED_API))]
fn ascii_object_bitfield() {
    let ob_base: PyObject = unsafe { std::mem::zeroed() };

    let mut o = PyASCIIObject {
        ob_base,
        length: 0,
        #[cfg(not(PyPy))]
        hash: 0,
        state: 0u32,
        #[cfg(not(Py_3_12))]
        wstr: std::ptr::null_mut() as *mut wchar_t,
    };

    unsafe {
        assert_eq!(o.interned(), 0);
        assert_eq!(o.kind(), 0);
        assert_eq!(o.compact(), 0);
        assert_eq!(o.ascii(), 0);
        #[cfg(not(Py_3_12))]
        assert_eq!(o.ready(), 0);

        let interned_count = if cfg!(Py_3_12) { 2 } else { 4 };

        for i in 0..interned_count {
            o.set_interned(i);
            assert_eq!(o.interned(), i);
        }

        for i in 0..8 {
            o.set_kind(i);
            assert_eq!(o.kind(), i);
        }

        o.set_compact(1);
        assert_eq!(o.compact(), 1);

        o.set_ascii(1);
        assert_eq!(o.ascii(), 1);

        #[cfg(not(Py_3_12))]
        o.set_ready(1);
        #[cfg(not(Py_3_12))]
        assert_eq!(o.ready(), 1);
    }
}

Start by reproducing the provided ascii_object_bitfield test in clin1234/pyo3 with the listed rust-analyzer, rustc, VS Code extension, and settings. Then trace rust-analyzer's test-running entry point and relevant logs to determine why the test cannot launch; done means the test runs from VS Code and the failure has regression coverage.