New releases of pytest-regtest>=2.0.0 on PyPI are currently broken as the released packages don't appear to include the Python code for the module itself. This has been reported upstream and when fixed I will remove the dependency.
Made the pre-commit updates that were automated in #253 where the pytest-regtest issue cropped up
Added no-commit-to-branch check to pre-commit hooks which protects against commits to master locally.
Introduced NPY201 checks under ruff linting to check for forthcoming Numpy2 deprecations.
Numpy2 is due to be released early later this year (first release candidate is scheduled 2024-02-01).
This may introduce some breaking changes, a broad overview is available here.
Mentioned in that article is that scikit-image has issues, which is indirect because the package has an optional dependency on matplotlib (I've discovered this whilst working on skan issues).
A useful summary of how to handle potential problems is provided at Numpy #24300 and broadly suggests pinning numpy<2.0 for releases whilst development proceeds on resolving issues.
New releases of
pytest-regtest>=2.0.0
on PyPI are currently broken as the released packages don't appear to include the Python code for the module itself. This has been reported upstream and when fixed I will remove the dependency.At the same time I've...
pre-commit
updates that were automated in #253 where thepytest-regtest
issue cropped upno-commit-to-branch
check topre-commit
hooks which protects against commits tomaster
locally.NPY201
checks underruff
linting to check for forthcoming Numpy2 deprecations.Numpy2 is due to be released early later this year (first release candidate is scheduled 2024-02-01).
This may introduce some breaking changes, a broad overview is available here.
Mentioned in that article is that
scikit-image
has issues, which is indirect because the package has an optional dependency onmatplotlib
(I've discovered this whilst working onskan
issues).The official NumPy 2.0 migration guide — NumPy v2.0.dev0 Manual suggests that
ruff
can be leveraged to address the mentioned changes under Ruff Plugin.A useful summary of how to handle potential problems is provided at Numpy #24300 and broadly suggests pinning
numpy<2.0
for releases whilst development proceeds on resolving issues.