Skip to content

Conversation

@ev-br
Copy link
Member

@ev-br ev-br commented Jan 17, 2026

Previously, we only tested the matching type/dtype combinations, i.e.

  • float_array, float_scalar
  • int_array, int_scalar
  • bool_array, bool_scalar

So float_array, int_scalar combinations were missing (cross-ref #364)

Tested locally with 10_000 examples on array_api_strict, numpy, cupy, torch and jax, and 1000 examples on dask. No new failures.

TODO

ev-br added 2 commits January 17, 2026 20:11
Get rid of ad hoc mM and positive arguments. Use the same
`min_value`, `max_value` kwargs for scalars and arrays of varying
types/dtypes.
Previously, we only tested the matching type/dtype combinations, i.e.
- float_array, float_scalar
- int_array, int_scalar
- bool_array, bool_scalar
@ev-br
Copy link
Member Author

ev-br commented Jan 17, 2026

Updated test_clip. Caught a failure on JAX with

E           ========== FAILING CODE SNIPPET:
E           xp.clip(Array(0, dtype=uint64), **kw) with kw = {'min': 0, 'max': 9223372036854775808}
E           ====================

Reported upstream at jax-ml/jax#34455

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

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant