Skip to content

BUG: incoming deprecation warnings from numpy 2.5 (dev): direct mutations to array shape attributes are deprecated #19126

Closed
0 of 2 issues completed

Activity

  1. added theissue type on Jan 8, 2026
  2. neutrinoceros commented on Jan 12, 2026

    @neutrinoceros
    ContributorAuthor

    It's reached nightlies.
    I get close to 900 tests failing when including scipy and pandas in the env. Only ~150 if I don't.

  3. pllim commented on Jan 12, 2026

    @pllim
    Member

    Yeah, seeing the failures downstream too (e.g., Ginga). Is there something I can help to push the fixes forward? Thanks!

  4. neutrinoceros commented on Jan 12, 2026

    @neutrinoceros
    ContributorAuthor

    At this stage even reading logs is complicated because most errors are not going away with any particular fix, so I'm hoping it'll get easier once the first salve of fixes gets in.

  5. added this to the v7.2.1 milestone on Jan 12, 2026
  6. osasisorae commented on Jan 18, 2026

    @osasisorae

    I've implemented a comprehensive fix for the NumPy 2.5 shape deprecation warnings. Please check out my pull request: #19179

    All .shape = assignments replaced with .resize(new_shape, refcheck=False) calls across 11 source files and 2 test files. Zero deprecation warnings, full backward compatibility maintained.

  7. neutrinoceros commented on Jan 23, 2026

    @neutrinoceros
    ContributorAuthor

    All immediate failures are resolved and CI is back to green. Follow ups will be addressed from sub issues.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Type

    Projects

    No projects

      Milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions