TST: strictly xfail #38901
TST: strictly xfail #38901
Conversation
rhshadrach
added
the
Testing
|
In #38881 you used request.applymarker which i guess is equivalent to request.node.add_marker. is that an Officially Supported way of doing this? What we've been doing works, but AFAICT isn't something pytest advertises |
|
@jbrockmendel Yes, they are equivalent. Under the hood, Docs for applymarker: https://docs.pytest.org/en/stable/reference.html#pytest.FixtureRequest.applymarker
I take this to mean we are using it in the intended way. |
| @@ -942,29 +956,37 @@ def test_fillna_datetime64tz(self, index_or_series, fill_val, fill_dtype): | |||
| ) | |||
| self._assert_fillna_conversion(obj, fill_val, exp, fill_dtype) | |||
|
|
|||
| @pytest.mark.xfail(reason="Test not implemented") | |||
There was a problem hiding this comment.
we should just parameterize these (followon)
|
great. conflict if you can merge master. also do we have an issue to check for not using |
…ail_strict � Conflicts: � pandas/tests/io/excel/test_readers.py
|
thanks @rhshadrach |
black pandasgit diff upstream/master -u -- "*.py" | flake8 --diffContinuation of #38881. This handles cases where replacing pytest.xfail is clear, opened #38902 for the handful of remaining cases.
In the case where
pytest.xfailis used for a test which has been yet to be implemented, I've addedassert Falseto maintain the current behavior of counting it as an xfail. Maybe there is a better alternative here.raise NotImplementedError.