Use if in _CCCL_TRY_CUDA_API

[Jacobfaib]

Description

Use a simple if in _CCCL_TRY_CUDA_API instead of the switch abomination.

Also get rid of unused include, terminate() is called by the throw macros but they already include the header for us.

Checklist

• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 95335840-a127-4c90-a69d-cb21d9d9efbe

📥 Commits

Reviewing files that changed from the base of the PR and between 05d669b and aa97154.

📒 Files selected for processing (1)

• libcudacxx/include/cuda/__runtime/api_wrapper.h

---

📝 Walkthrough

Summary by CodeRabbit

• Bug Fixes
  • Improved handling of CUDA API errors, ensuring failures consistently clear the CUDA error state and report the appropriate error.
• Refactor
  • Simplified internal CUDA status checking without changing the public API.

Walkthrough

The CUDA API wrapper replaces switch-based status handling with an explicit cudaSuccess check, while preserving CUDA error-state clearing and exception construction on failure.

Changes

CUDA API error handling

Layer / File(s)	Summary
Refactor CUDA status checking libcudacxx/include/cuda/__runtime/api_wrapper.h	_CCCL_TRY_CUDA_API uses an if condition for non-success statuses and retains the existing cudaGetLastError() and cuda_error throw behavior.

---

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

[fbusato]

looks good. Probably the initial intention was to handle more cases

[github-actions]

😬 CI Workflow Results

🟥 Finished in 16h 22m: Pass: 99%/120 | Total: 3d 18h | Max: 2h 26m | Hits: 55%/689556

See results here.