Reject masked JAX matrix-power exponents

[FlorianPfaff]

Bug

The JAX linalg backend normalized matrix_power exponents with operator.index before checking NumPy masked-array semantics. A masked scalar such as np.ma.array(2, mask=True) therefore exposed its hidden payload and was accepted as exponent 2.

This can turn a missing configuration value into a real matrix operation instead of failing deterministically.

Fix

• reject genuinely masked exponent inputs before integer coercion;
• preserve ordinary Python, NumPy, and JAX integer scalar handling;
• preserve fully unmasked integer-valued MaskedArray scalars;
• add focused subprocess regression coverage on the JAX backend.

Validation

• reproduced that operator.index(np.ma.array(2, mask=True)) returns the hidden value 2;
• focused local JAX validation passed for two masked rejection cases and the unmasked compatibility case;
• branch comparison: 2 commits ahead, 0 behind main;
• final diff is limited to 2 implementation lines and 21 test lines across two files.

GitHub Actions provides the full supported JAX, Python-version, lint, and packaging matrix.

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no	no	25.16s
✅ JSON	prettier	7	0	0	0	0.99s
✅ JSON	v8r	7		0	0	4.74s
✅ MARKDOWN	markdownlint	68	0	0	0	2.29s
✅ MARKDOWN	markdown-table-formatter	68	0	0	0	0.87s
✅ PYTHON	black	1809	56	0	0	93.8s
✅ PYTHON	isort	1809	88	0	0	2.73s
✅ REPOSITORY	betterleaks	yes		no	no	2.58s
✅ REPOSITORY	checkov	yes		no	no	54.3s
✅ REPOSITORY	gitleaks	yes		no	no	17.06s
✅ REPOSITORY	git_diff	yes		no	no	0.35s
✅ REPOSITORY	secretlint	yes		no	no	81.08s
✅ REPOSITORY	syft	yes		no	no	5.6s
✅ REPOSITORY	trivy-sbom	yes		no	no	6.06s
✅ REPOSITORY	trufflehog	yes		no	no	32.22s
✅ YAML	prettier	11	0	0	0	0.74s
✅ YAML	v8r	11		0	0	13.18s
✅ YAML	yamllint	11		0	0	0.71s

Notices

📣 MegaLinter 9.5.0 is out! Discover the new features and security recommendations in the release announcement. (Skip this info by defining SECURITY_SUGGESTIONS: false)

See detailed reports in MegaLinter artifacts

Your project could benefit from a custom flavor, which would allow you to run only the linters you need, and thus improve runtime performances. (Skip this info by defining FLAVOR_SUGGESTIONS: false)

• Documentation: Custom Flavors
• Command: npx mega-linter-runner@9.6.0 --custom-flavor-setup --custom-flavor-linters PYTHON_BLACK,PYTHON_ISORT,COPYPASTE_JSCPD,JSON_V8R,JSON_PRETTIER,MARKDOWN_MARKDOWNLINT,MARKDOWN_MARKDOWN_TABLE_FORMATTER,REPOSITORY_CHECKOV,REPOSITORY_GIT_DIFF,REPOSITORY_GITLEAKS,REPOSITORY_BETTERLEAKS,REPOSITORY_SECRETLINT,REPOSITORY_SYFT,REPOSITORY_TRIVY_SBOM,REPOSITORY_TRUFFLEHOG,YAML_PRETTIER,YAML_YAMLLINT,YAML_V8R

Show us your support by starring ⭐ the repository

[FlorianPfaff]

Closing as low value. This only adds rejection semantics for NumPy masked scalars passed as JAX matrix_power exponents. Masked arrays are outside the normal JAX/PyRecEst parameter contract, so isolated support would add backend-specific complexity without practical benefit.