Fix empty PyTorch uniform device selection

[FlorianPfaff]

What changed

• prefer an existing non-CPU tensor device when constructing the empty-result fallback for PyTorch uniform
• retain the first tensor device when all bounds are on CPU
• add a regression covering CPU low and meta-device high bounds

Root cause

The empty descending-interval compatibility path selected low.device whenever low was a tensor. With a CPU low tensor and a non-CPU high tensor, it therefore attempted to coerce the non-CPU endpoint onto CPU. This was inconsistent with the backend's binary-device selection contract and could fail for non-materializable devices such as meta.

Impact

Zero-sized PyTorch uniform draws now preserve an existing accelerator or meta device regardless of endpoint order.

Validation

• branch is 2 commits ahead of main and 0 behind
• diff is limited to the empty-uniform compatibility helper and focused regression coverage
• full local tests were unavailable because this environment has no GitHub network checkout or gh binary; GitHub Actions should validate the branch

[github-actions]

✅MegaLinter analysis: Success

Descriptor	Linter	Files	Fixed	Errors	Warnings	Elapsed time
✅ COPYPASTE	jscpd	yes		no	no	20.2s
✅ JSON	prettier	7	0	0	0	1.47s
✅ JSON	v8r	7		0	0	3.85s
✅ MARKDOWN	markdownlint	68	0	0	0	2.02s
✅ MARKDOWN	markdown-table-formatter	68	0	0	0	0.84s
✅ PYTHON	black	1585	220	0	0	88.38s
✅ PYTHON	isort	1585	400	0	0	2.64s
✅ REPOSITORY	betterleaks	yes		no	no	2.13s
✅ REPOSITORY	checkov	yes		no	no	50.9s
✅ REPOSITORY	gitleaks	yes		no	no	14.33s
✅ REPOSITORY	git_diff	yes		no	no	0.36s
✅ REPOSITORY	secretlint	yes		no	no	61.58s
✅ REPOSITORY	syft	yes		no	no	5.79s
✅ REPOSITORY	trivy-sbom	yes		no	no	6.67s
✅ REPOSITORY	trufflehog	yes		no	no	29.81s
✅ YAML	prettier	11	0	0	0	0.71s
✅ YAML	v8r	11		0	0	11.82s
✅ YAML	yamllint	11		0	0	0.52s

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