TST: scope xfail for test_np_fix to non-numpy-dev to avoid XPASS in dev CI

[antznette1]

Notes:

• test_np_fix now uses [pytest.mark.xfail(condition=not](cci:4://file://pytest.mark.xfail(condition=not:0:0-0:0) is_numpy_dev, reason=...) without strict=True.
• No runtime probing; minimal and idiomatic.

[jbrockmendel]

Most of this PR is not doing what the description says. Did a human write this?

[antznette1]

Most of this PR is not doing what the description says. Did a human write this?

Thanks for the review. I’ve reduced the PR to a single test change:
test_np_fix uses is_numpy_dev in the xfail marker (no strict, no runtime probing). The title & description now reflect the minimal scope. The intent is to avoid numpy-dev/python-dev XPASS(strict) failures while keeping behavior unchanged on non-dev builds. Happy to adjust if you prefer a different guard.

[antznette1]

@jbrockmendel
The test_np_fix uses is_numpy_dev in the xfail marker (no strict, no runtime probing).
pre-commit.ci autofix

[jbrockmendel]

The diff is empty. Is there anything left to do here?

[antznette1]

The diff is empty. Is there anything left to do here?

You’re right; after the change on main to switch this test over to np.trunc and clean up the behavior, this PR no longer has anything useful to add.

The original intent was just to scope an xfail for the old np.fix(Series) behavior in dev CI, but that’s been superseded by the current test layout.

I’m fine closing this as obsolete; thanks for the review.