Numpy 2.0 arrives with breaking changes

The numpy team has issued this statement

Basically you shall temporarily add requirements to pinned numpy<2 version, then create a CI based on numpy2 and correct all relevant errors before migrating.

If you already have a pinned numpy version in Slicer builds, then you have no issues until next release, but users must be warned not to upgrade numpy in any way. Therefore, you shall block numpy upgrades in the pip_install function.

Since an issue has been created for this topic, I would suggest to continue discussion there:

1 Like