fix(gateway): restart on whatsapp bridge child exit (#2334)

Co-authored-by: Frederico Ribeiro <fr@tecompanytea.com>
This commit is contained in:
Teknium
2026-03-21 09:38:52 -07:00
committed by GitHub
parent e6299960cc
commit 8304a7716d
4 changed files with 160 additions and 5 deletions

View File

@@ -1,3 +1,5 @@
from unittest.mock import AsyncMock
import pytest
from gateway.config import GatewayConfig, Platform, PlatformConfig
@@ -27,6 +29,23 @@ class _FatalAdapter(BasePlatformAdapter):
return {"id": chat_id}
class _RuntimeRetryableAdapter(BasePlatformAdapter):
def __init__(self):
super().__init__(PlatformConfig(enabled=True, token="token"), Platform.WHATSAPP)
async def connect(self) -> bool:
return True
async def disconnect(self) -> None:
self._mark_disconnected()
async def send(self, chat_id, content, reply_to=None, metadata=None):
raise NotImplementedError
async def get_chat_info(self, chat_id):
return {"id": chat_id}
@pytest.mark.asyncio
async def test_runner_requests_clean_exit_for_nonretryable_startup_conflict(monkeypatch, tmp_path):
config = GatewayConfig(
@@ -44,3 +63,31 @@ async def test_runner_requests_clean_exit_for_nonretryable_startup_conflict(monk
assert ok is True
assert runner.should_exit_cleanly is True
assert "already using this Telegram bot token" in runner.exit_reason
@pytest.mark.asyncio
async def test_runner_requests_failure_exit_for_retryable_runtime_fatal(monkeypatch, tmp_path):
config = GatewayConfig(
platforms={
Platform.WHATSAPP: PlatformConfig(enabled=True, token="token")
},
sessions_dir=tmp_path / "sessions",
)
runner = GatewayRunner(config)
adapter = _RuntimeRetryableAdapter()
adapter._set_fatal_error(
"whatsapp_bridge_exited",
"WhatsApp bridge process exited unexpectedly (code 1).",
retryable=True,
)
runner.adapters = {Platform.WHATSAPP: adapter}
runner.delivery_router.adapters = runner.adapters
runner.stop = AsyncMock()
await runner._handle_adapter_fatal_error(adapter)
assert runner.should_exit_cleanly is False
assert runner.should_exit_with_failure is True
assert "exited unexpectedly" in runner.exit_reason
runner.stop.assert_awaited_once()

View File

@@ -53,6 +53,15 @@ def _make_adapter():
adapter._bridge_process = None
adapter._reply_prefix = None
adapter._running = False
adapter._message_handler = None
adapter._fatal_error_code = None
adapter._fatal_error_message = None
adapter._fatal_error_retryable = True
adapter._fatal_error_handler = None
adapter._active_sessions = {}
adapter._pending_messages = {}
adapter._background_tasks = set()
adapter._auto_tts_disabled_chats = set()
adapter._message_queue = asyncio.Queue()
return adapter
@@ -200,6 +209,54 @@ class TestFileHandleClosedOnError:
mock_fh.close.assert_called_once()
assert adapter._bridge_log_fh is None
class TestBridgeRuntimeFailure:
"""Verify runtime bridge death is surfaced as a fatal adapter error."""
@pytest.mark.asyncio
async def test_send_marks_retryable_fatal_when_managed_bridge_exits(self):
adapter = _make_adapter()
fatal_handler = AsyncMock()
adapter.set_fatal_error_handler(fatal_handler)
adapter._running = True
mock_fh = MagicMock()
adapter._bridge_log_fh = mock_fh
mock_proc = MagicMock()
mock_proc.poll.return_value = 7
adapter._bridge_process = mock_proc
result = await adapter.send("chat-123", "hello")
assert result.success is False
assert "exited unexpectedly" in result.error
assert adapter.fatal_error_code == "whatsapp_bridge_exited"
assert adapter.fatal_error_retryable is True
fatal_handler.assert_awaited_once()
mock_fh.close.assert_called_once()
assert adapter._bridge_log_fh is None
@pytest.mark.asyncio
async def test_poll_messages_marks_retryable_fatal_when_managed_bridge_exits(self):
adapter = _make_adapter()
fatal_handler = AsyncMock()
adapter.set_fatal_error_handler(fatal_handler)
adapter._running = True
mock_fh = MagicMock()
adapter._bridge_log_fh = mock_fh
mock_proc = MagicMock()
mock_proc.poll.return_value = 23
adapter._bridge_process = mock_proc
await adapter._poll_messages()
assert adapter.fatal_error_code == "whatsapp_bridge_exited"
assert adapter.fatal_error_retryable is True
fatal_handler.assert_awaited_once()
mock_fh.close.assert_called_once()
assert adapter._bridge_log_fh is None
@pytest.mark.asyncio
async def test_closed_when_http_not_ready(self):
"""Health endpoint never returns 200 within 15 attempts."""