Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

New CI failures in sparse with nightly numpy #17033

Closed
WarrenWeckesser opened this issue Sep 16, 2022 · 1 comment · Fixed by #17035
Closed

New CI failures in sparse with nightly numpy #17033

WarrenWeckesser opened this issue Sep 16, 2022 · 1 comment · Fixed by #17035
Labels
defect A clear bug or issue that prevents SciPy from being installed or used as expected scipy.sparse
Milestone
1.10.0

Comments

@WarrenWeckesser
Copy link
Member

WarrenWeckesser commented Sep 16, 2022 •
edited

If you have submitted a pull request recently, you probably noticed some new failures in the Azure job prerelease_deps_coverage_64bit_blas that are unrelated to your changes:

=========================== short test summary info ============================
FAILED scipy/sparse/tests/test_base.py::TestCOO::test_reshape_copy - Assertio...
FAILED scipy/sparse/tests/test_base.py::TestCOONonCanonical::test_reshape_copy
FAILED scipy/sparse/tests/test_base.py::Test64Bit::test_resiliency_limit_10[TestCOO-test_reshape_copy]
FAILED scipy/sparse/tests/test_base.py::Test64Bit::test_no_64[TestCOO-test_reshape_copy]
FAILED scipy/sparse/tests/test_base.py::Test64Bit::test_resiliency_random[TestCOO-test_reshape_copy]
FAILED scipy/sparse/tests/test_base.py::Test64Bit::test_resiliency_all_32[TestCOO-test_reshape_copy]
FAILED scipy/sparse/tests/test_base.py::Test64Bit::test_resiliency_all_64[TestCOO-test_reshape_copy]

This job is the one that uses a "nightly" NumPy build. The failures are the result of a recent change in NumPy; see numpy/numpy#21995 (comment) for a description of how the NumPy change breaks the tests in scipy.sparse.

Based on the follow-up comment and the tests in that PR that were approved, it looks like this is an intentional change that is not considered a backwards compatibility break. It changes behavior, but it is apparently behavior that was never guaranteed to always remain the same.

I think we'll be able to work-around the NumPy change pretty easily.
@WarrenWeckesser WarrenWeckesser added defect A clear bug or issue that prevents SciPy from being installed or used as expected scipy.sparse labels Sep 16, 2022
@mdhaber
Copy link
Contributor

mdhaber commented Sep 16, 2022

There are also harmless failures in stats that should be fixed by gh-17034.

WarrenWeckesser added a commit to WarrenWeckesser/scipy that referenced this issue Sep 16, 2022
@WarrenWeckesser
BUG: sparse: Avoid creating a view when ensuring that an array has na…
1c290ef 
…tive byte order.

In NumPy 1.24+, the call

    np.asarray(A, dtype=A.dtype.newbyteorder('native'))

will *always* create a new view of A, even if
A.dtype == A.dtype.newbyteorder('native').
Avoid this by only making the call if A.dtype is
not already native order.

Closes scipygh-17033.
@WarrenWeckesser WarrenWeckesser mentioned this issue Sep 16, 2022
Merged
This was referenced Sep 16, 2022
Merged
Merged
@mdhaber mdhaber mentioned this issue Sep 19, 2022
Merged
@mdhaber mdhaber added this to the 1.10.0 milestone Nov 29, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
defect A clear bug or issue that prevents SciPy from being installed or used as expected scipy.sparse
Projects
None yet
Milestone
1.10.0
2 participants
@WarrenWeckesser @mdhaber