fix(samples): waive CUDA-GL interop on non-NVIDIA GL contexts

[yimoj]

Description

Follow-up to the sample migration in #2959; related OpenGL availability handling is tracked in #2767 (this PR does not close that issue).

The migration replaced the old example wrapper's unconditional GL-sample skip with actual sample execution. A window can now open successfully on a non-NVIDIA GL context, such as Microsoft/D3D12, but these samples still attempt CUDA resource registration. The existing graphics tests already reject such contexts.

Reuse gl_context_not_on_nvidia_gpu_reason() in glInteropFluid, glInteropMipmapLod, and glInteropPlasma after creating the GL window and before registering CUDA resources. Ineligible contexts report the reason, close the window, and exit with the negotiated sample-waiver code. NVIDIA GL contexts continue normally, and genuine CUDA errors remain failures.

Only these three samples change. No new tests are added: the existing GL-helper tests cover vendor eligibility, and the sample harness already runs all three examples for three frames.

Validation

• Ruff checks and formatting checks pass for the three changed samples.
• python -m compileall -q passes for the three changed samples.
• git diff --check passes.
• Targeted existing sample tests could not collect locally because the environment has no importable CUDA package. WSL and native NVIDIA GL runtime validation remain outstanding; no hardware-backed success is claimed.

Checklist

• New or existing tests cover these changes.
• The documentation is up to date with these changes (no documentation changes are required).

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: Repository: NVIDIA/cuda-python/.coderabbit.yaml
• Review profile: CHILL
• Plan: Enterprise
• Run ID: 7a51acf4-4eb9-4bd5-9bc4-c8b733f1efbc

📥 Commits

Reviewing files that changed from the base of the PR and between 28b188f and 03fe8b8.

📒 Files selected for processing: 3

• samples/cuda_core/glInteropFluid/glInteropFluid.py
• samples/cuda_core/glInteropMipmapLod/glInteropMipmapLod.py
• samples/cuda_core/glInteropPlasma/glInteropPlasma.py

Included review availability: This review used your included allowance. Your plan provides up to 12 included reviews per hour; 11 remain after this review.

📝 Summary

Summary by CodeRabbit

• Bug Fixes
  • The fluid, mipmap LOD, and plasma graphics interop samples now check whether the active OpenGL context is running on an NVIDIA GPU. If the context is unsupported, the sample displays the reason, closes its window, and exits as waived rather than attempting to run. With a supported context, the samples continue to run as usual.

The Fluid, MipmapLod, and Plasma samples now check whether their OpenGL context is on an NVIDIA GPU. If the helper reports an unsupported context, each sample prints the reason, closes the window, and exits with EXIT_WAIVED.

Changes: OpenGL context waivers

Layer / File(s)	Summary
Load the helper and check the GL context samples/cuda_core/glInteropFluid/glInteropFluid.py, samples/cuda_core/glInteropMipmapLod/glInteropMipmapLod.py, samples/cuda_core/glInteropPlasma/glInteropPlasma.py	Each sample imports the shared context-check helper. After window creation, each sample checks the context and exits with EXIT_WAIVED when the helper returns a reason.

Suggested reviewers: danielfrg

Priority: ⬇️ Low

Change: Bug fix

Merge Risk: ⚪ Minimal · up to 03fe8

These samples now waive before CUDA registration when GL reports a non-NVIDIA vendor, avoiding the unsupported interop path. No concrete source-level merge blocker is established; hardware-backed validation can remain a follow-up.

✨ Finishing Touches: 💡 1

🛠️ Fix failing CI checks: 💡

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

Doc Preview CI
Preview removed because the pull request was closed or merged.

[lijinf2]

👍

[lijinf2]

The PR needs an additional label "Bug" besides cuda.core to get CI green. I just added the label.

Thank you for the PR @yimoj !

[yimoj]

@lijinf2 while you're here, please help set assignee too, which seems to be required by CI as well.