[Nir-Az]

Summary

• stop_wrapper could throw when resetting safety_mode back to run, and in finally blocks this exception replaced the original test failure (e.g. FPS assertion errors were masked by RuntimeError: Failed to set the option to value 0)
• Wrapped the cleanup in try/except so failures are logged clearly without hiding the real test error

Test plan

• Run pytest-ah-configurations on D585S — verify FPS assertion errors propagate correctly when cleanup also fails
• Verify stop_wrapper cleanup failure is logged with -E- Cleanup failed: could not set safety_mode back to run: ...

[Copilot]

Pull request overview

This PR prevents stop_wrapper cleanup errors (resetting safety_mode back to run) from masking the original test failure when stop_wrapper is invoked from finally blocks in live tests (e.g., FPS assertion failures on D585S).

Changes:

• Import rspy.log in the wrapper module.
• Wrap D585S safety-mode reset in stop_wrapper with try/except and log cleanup failures instead of letting them replace the original exception.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

stop_wrapper now swallows any exception from the cleanup unconditionally. This prevents the original failure from being masked in a finally, but it also means a test can pass even if the device could not be returned to safety_mode.run (potentially impacting subsequent tests). Consider logging the cleanup error and re-raising it when no other exception is currently being handled (e.g., detect active exception via sys.exc_info()), so cleanup failures still fail the test when they’re the only error.

[Nir-Az]

The cleanup failure is an infrastructure issue, not a test failure — logging it with -E- makes it visible in CI logs. The test result should reflect what the test actually tested, not whether teardown succeeded.