Reliability Python active any

Event-loop ownership stays at application boundaries

rel-python-002

Intent

Avoid nested loop bugs and library helpers that secretly create or run their own event loops.

Applicability

Applies to Python code calling asyncio.run, new_event_loop, set_event_loop, or run_until_complete.

What to inspect

Loop-control APIs and whether they appear in entry points or reusable helpers.

Pass criteria

Only explicit application entry points own loop creation and top-level execution.

Fail criteria

The diff creates or drives an event loop from a reusable helper or library module.

Do not flag

Main modules, CLI entry points, and tests that deliberately own an isolated loop.

Confidence guidance

HIGH when loop creation in a helper is directly visible. MEDIUM when the file role is inferred. LOW when entry-point ownership is unclear.

Remediation

Move event-loop control to the application’s entry point and keep helpers async.

Pass example

if __name__ == "__main__":
    asyncio.run(main())

Fail example

def load_user():
    return asyncio.new_event_loop().run_until_complete(fetch())