TST: Skip ctypes dependent test that fails on Python < 2.7.7. #11159
Conversation
numpy/core/tests/test_multiarray.py
Outdated
| @@ -6517,6 +6517,7 @@ def test_max_dims(self): | |||
| a = np.empty((1,) * 32) | |||
| self._check_roundtrip(a) | |||
|
|
|||
| #pytest.mark.skipif(sys.version_info < (2, 7, 7), reason="ctypes issue") | |||
There was a problem hiding this comment.
Should be @. May as well link to the bpo bug in the reason.
| @@ -6517,6 +6517,7 @@ def test_max_dims(self): | |||
| a = np.empty((1,) * 32) | |||
| self._check_roundtrip(a) | |||
|
|
|||
| @pytest.mark.skipif(sys.version_info < (2, 7, 7), reason="See gh-11115") | |||
There was a problem hiding this comment.
I'd prefer either:
relies on bpo-10744 being fixedrelies on https://bugs.python.org/issue10744 being fixed
At any rate, we can just refer to the python issue here - the github one is not interesting
There was a problem hiding this comment.
@eric-wieser The Python issue seems a bit less discoverable and the numpy reference has a fair amount of useful discussion. Given that we will drop 2.7 support at the end of the year, probably not worth spending much time on this. If nothing else, git blame will lead back here.
|
Again, the commit message is a little vague here. Possibly better would be something like |
Not really - only related in that the two independent regressions were introduced at the same time |
Fixes #11115 related to #11150
Python before 2.7.7 had a broken ctypes memoryview format for multi-demensional arrays. Skip the test