2dc8f7715a
hermes_cli/dump.py::_gateway_status indexed snapshot.gateway_pids[0] whenever snapshot.running was true. Under s6 the service is up but the process scan is empty (python -c launcher argv, no gateway.pid), so the IndexError was swallowed into "unknown". Same class as the _cmd_status bug: a supervisor-reported service must not be equated with a PID hit. Topology T10 (s6/Docker).
15 lines
733 B
Python
15 lines
733 B
Python
"""``hermes dump`` gateway line for a supervised gateway with no scannable PID (#125390)."""
|
|
|
|
from __future__ import annotations
|
|
|
|
from hermes_cli.dump import _gateway_status
|
|
from hermes_cli.gateway import GatewayRuntimeSnapshot
|
|
|
|
|
|
def test_s6_supervised_gateway_without_pid_is_running_not_unknown(monkeypatch):
|
|
"""s6 service up, empty process scan: ``running (...)`` — not an IndexError swallowed into ``unknown``."""
|
|
snapshot = GatewayRuntimeSnapshot(
|
|
manager="s6 (container supervisor)", service_installed=True, service_running=True, gateway_pids=())
|
|
monkeypatch.setattr("hermes_cli.gateway.get_gateway_runtime_snapshot", lambda: snapshot)
|
|
assert _gateway_status() == "running (s6 (container supervisor))"
|