Fix command wrappers exiting after successful runs

[ishitadatta]

Summary

• fix the success-path indentation bug in setup_env.py and utils/e2e_benchmark.py
• add focused regression tests for the Python command wrappers

Why this matters

Both helpers currently call sys.exit(1) after a successful subprocess.run(...) when log_step is not provided. In practice this makes the benchmark helper report failure even when the benchmark itself succeeds, and it leaves the setup helper vulnerable to the same control-flow bug on any future no-log-step call site.

This change keeps the existing failure behavior, but stops the scripts from exiting with an error after successful runs.

Validation

• python3 -m py_compile setup_env.py utils/e2e_benchmark.py tests/test_command_wrappers.py
• direct smoke check confirming both wrappers do not call sys.exit(1) on successful subprocess execution
• pytest could not be run in this environment because it is not installed here

[ishitadatta]

@microsoft-github-policy-service agree