You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I noticed in both #271 and #268 that test_basic.py::test_math_util_inner1d SKIPPED happens whenever numpy 2.x is present. This implies that math_util C-extension is failing to build with numpy 2.x and spherical-geometry is silently falling back to Python implementations, if available.
I think you have to change your build requirement to use numpy 2.0 (up to you if you want to use RC2 or wait for stable release). Its ABI would be backward compatible with numpy 1.x up to a certain point. This means you can build with numpy 2.0 while still let users choose to use this package with numpy 1.x at runtime.
The text was updated successfully, but these errors were encountered:
BTW you don't see the build failure in tox log because the log hides a lot of stuff. You can opt to turn on verbosity but that might pull in a lot of noise. The easier way might be to install numpy 2.0.rc2 from PyPI locally and then try to build this package locally outside of tox to see what is crashing.
I noticed in both #271 and #268 that
test_basic.py::test_math_util_inner1d SKIPPED
happens whenever numpy 2.x is present. This implies thatmath_util
C-extension is failing to build with numpy 2.x and spherical-geometry is silently falling back to Python implementations, if available.I think you have to change your build requirement to use numpy 2.0 (up to you if you want to use RC2 or wait for stable release). Its ABI would be backward compatible with numpy 1.x up to a certain point. This means you can build with numpy 2.0 while still let users choose to use this package with numpy 1.x at runtime.
The text was updated successfully, but these errors were encountered: