| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| |\
| |
| | |
BUG,MAINT: Ensure masked elements can be tested against nan and inf.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This brought to light two bugs in tests, which are fixed here, viz.,
that a sample ndarray subclass that tested propagation of an added
parameter was incomplete, in that in propagating the parameter in
__array_wrap__ it assumed it was there on self, but that assumption
could be broken when a view of self was taken (as is done by
x[~flagged] in the test routine), since there was no
__array_finalize__ defined.
The other subclass bug counted, incorrectly, on only needing to provide
one type of comparison, the __lt__ being explicitly tested. But flags
are compared with __eq__ and those flags will have the same subclass.
|
| | |
| |
| |
| |
| |
| |
| | |
The removal of nan and inf from arrays that are compared using
test routines like assert_array_equal treated the two arrays
separately, which for masked arrays meant that some elements would
not be removed when they should have been. This PR corrects this.
|
| | | |
|
| |/
|
|
| |
Since this is now in `np.testing._private`, it's no longer usable by the outside world anyway
|
| |
|
|
|
|
|
| |
The vectorize version of this test was failing consistently on several of the Appveyor builds, ever since a recent pytest upgrade.
Our theory is that by random chance, things changed so that during the call to vectorize(op).__call__, python started running a garbage collection, which perturbed the refcounts that this test is checking. (Specifically this test is doing a weird thing and checking that the refcount of the object 1 doesn't decrease, and it's very plausible that some random bit of garbage was holding a reference to this object.)
Disabling the gc during the test makes this kind of refcount assertion more reliable, and seems to have fixed the appveyor builds, so I guess it's good.
|
| |
|
|
|
| |
The underlying problem is that ma.all() evaluates to masked,
which is falsy, and thus triggers test failures.
|
| |
|
|
| |
It's not always possible to guarantee this, so also adds a test to verify that we don't hang
|
| |
|
|
|
|
|
|
|
| |
An example output for the test added in the previous commit is:
AssertionError: Reference cycles were found when calling make_cycle: 1 objects were collected, of which 1 are shown below:
list object with id=2279664872136:
[<Recursion on list with id=2279664872136>,
<Recursion on list with id=2279664872136>]
|
| |
|
|
|
|
| |
This also means we can now test that our test is actually able to detect the type of failure we expect
Trying to give myself some tools to debug the failure at https://github.com/numpy/numpy/pull/10882/files#r180813166
|
|
|
This is to prepare for the switch to pytest.
* Rename `numpy/testing/nose_tools` to `numpy/testing/_private`.
* Redirect imports as needed.
* Copy `_testutils.py` from scipy to `numpy/testing/_private`.
* Rename `_testutils.py` to `_pytester.py` and remove unneeded bits.
|
