RFC: Let cluster config/init bubble up exception groups

tests/conftest.py

@@ -25,7 +25,7 @@
 from zigpy.quirks import get_device
 import zigpy.types
 from zigpy.zcl.clusters.general import Basic, Groups
-from zigpy.zcl.foundation import Status
+from zigpy.zcl.foundation import GENERAL_COMMANDS, GeneralCommand, Status
 import zigpy.zdo.types as zdo_t
 
 from tests import common
@@ -450,7 +450,6 @@ def _mock_dev(
             endpoint = device.add_endpoint(epid)
             endpoint.device_type = ep[SIG_EP_TYPE]
             endpoint.profile_id = ep.get(SIG_EP_PROFILE)
-            endpoint.request = AsyncMock(return_value=[0])
 
             for cluster_id in ep.get(SIG_EP_INPUT, []):
                 endpoint.add_input_cluster(cluster_id)
@@ -463,9 +462,38 @@ def _mock_dev(
         else:
             device = get_device(device)
 
+        async def mock_request(
+            cluster: zigpy.types.ClusterId,
+            sequence: zigpy.types.uint8_t,
+            data: bytes,
+            expect_reply: bool = True,
+            command_id: GeneralCommand | zigpy.types.uint8_t = 0x00,
+        ):
+            # if isinstance(command_id, (int, GeneralCommand)):
+            # Some commands can't handle default response, and will
+            # fail with a non list element as first element.
+            if command_id in (
+                GeneralCommand.Read_Reporting_Configuration,
+                GeneralCommand.Write_Attributes,
+            ):
+                return [[]]
+
+            if command_id in GeneralCommand.__members__:
+                return GENERAL_COMMANDS[GeneralCommand.Default_Response].schema(
+                    command_id=command_id, status=Status.UNSUP_GENERAL_COMMAND
+                )
+
+            return [0]
+
+        # add request mock after device creation since quirks may have added endpoints
+        for epid, endpoint in device.endpoints.items():
+            if epid:
+                endpoint.request = AsyncMock(side_effect=mock_request)
+            else:
+                endpoint.request = AsyncMock(return_value=[0])
+
         if patch_cluster:
             for endpoint in (ep for epid, ep in device.endpoints.items() if epid):
-                endpoint.request = AsyncMock(return_value=[0])
                 for cluster in itertools.chain(
                     endpoint.in_clusters.values(), endpoint.out_clusters.values()
                 ):

tests/test_cluster_handlers.py

@@ -837,8 +837,10 @@ async def test_ep_cluster_handlers_configure(cluster_handler) -> None:
         mock.patch.dict(endpoint.claimed_cluster_handlers, claimed, clear=True),
         mock.patch.dict(endpoint.client_cluster_handlers, client_handlers, clear=True),
     ):
-        await endpoint.async_configure()
-        await endpoint.async_initialize(mock.sentinel.from_cache)
+        with pytest.raises(ExceptionGroup):
+            await endpoint.async_configure()
+        with pytest.raises(ExceptionGroup):
+            await endpoint.async_initialize(mock.sentinel.from_cache)
 
     for ch in [*claimed.values(), *client_handlers.values()]:
         assert ch.async_initialize.call_count == 1
@@ -847,9 +849,6 @@ async def test_ep_cluster_handlers_configure(cluster_handler) -> None:
         assert ch.async_configure.call_count == 1
         assert ch.async_configure.await_count == 1
 
-    assert ch_3.debug.call_count == 2
-    assert ch_5.debug.call_count == 2
-
 
 async def test_poll_control_configure(
     poll_control_ch: PollControlClusterHandler,

tests/test_device.py

@@ -616,7 +616,6 @@ async def test_async_add_to_group_remove_from_group(
 async def test_async_bind_to_group(
     device_joined: Callable[[ZigpyDevice], Awaitable[Device]],
     zigpy_device: Callable[..., ZigpyDevice],  # pylint: disable=redefined-outer-name
-    caplog: pytest.LogCaptureFixture,
 ) -> None:
     """Test async_bind_to_group method."""
     zigpy_dev = zigpy_device(with_basic_cluster_handler=True)
@@ -640,19 +639,12 @@ async def test_async_bind_to_group(
         group.group_id,
         [ClusterBinding(name="on_off", type=CLUSTER_TYPE_OUT, id=6, endpoint_id=3)],
     )
-    assert (
-        "0xb79c: Bind_req 00:0d:7f:00:0a:90:69:e8, ep: 3, cluster: 6 to group: 0x1001 completed: [<Status.SUCCESS: 0>]"
-        in caplog.text
-    )
 
     await zha_device_remote.async_unbind_from_group(
         group.group_id,
         [ClusterBinding(name="on_off", type=CLUSTER_TYPE_OUT, id=6, endpoint_id=3)],
     )
 
-    m1 = "0xb79c: Unbind_req 00:0d:7f:00:0a:90:69:e8, ep: 3, cluster: 6"
-    assert f"{m1} to group: 0x1001 completed: [<Status.SUCCESS: 0>]" in caplog.text
-
 
 async def test_device_automation_triggers(
     device_joined: Callable[[ZigpyDevice], Awaitable[Device]],

zha/application/gateway.py

@@ -23,6 +23,7 @@
 )
 import zigpy.device
 import zigpy.endpoint
+from zigpy.exceptions import DeliveryError
 import zigpy.group
 from zigpy.state import State
 from zigpy.types.named import EUI64
@@ -618,7 +619,11 @@
         )
         # we don't have to do this on a nwk swap
         # but we don't have a way to tell currently
-        await zha_device.async_configure()
+        try:
+            await zha_device.async_configure()
+        except* (TimeoutError, DeliveryError):
+            zha_device.debug("ignoring error %s during rejoin", exc_info=True)
+
         device_info = ExtendedDeviceInfoWithPairingStatus(
             pairing_status=DevicePairingStatus.CONFIGURED,
             **zha_device.extended_device_info.__dict__,

zha/zigbee/cluster_handlers/__init__.py

@@ -374,7 +374,7 @@ def _configure_reporting_status(
         event_data: dict[str, dict[str, Any]],
     ) -> None:
         """Parse configure reporting result."""
-        if isinstance(res, (Exception, ConfigureReportingResponseRecord)):
+        if isinstance(res, (Exception, ConfigureReportingResponseRecord, int)):
             # assume default response
             self.debug(
                 "attr reporting for '%s' on '%s': %s",

zha/zigbee/cluster_handlers/lightlink.py

@@ -2,7 +2,6 @@
 
 import zigpy.exceptions
 from zigpy.zcl.clusters.lightlink import LightLink
-from zigpy.zcl.foundation import GENERAL_COMMANDS, GeneralCommand
 
 from zha.zigbee.cluster_handlers import ClusterHandler, ClusterHandlerStatus, registries
 
@@ -34,14 +33,16 @@ async def async_configure(self) -> None:
             self.warning("Couldn't get list of groups: %s", str(exc))
             return
 
-        if isinstance(rsp, GENERAL_COMMANDS[GeneralCommand.Default_Response].schema):
-            groups = []
-        else:
+        if isinstance(
+            rsp, LightLink.ClientCommandDefs.get_group_identifiers_rsp.schema
+        ):
             groups = rsp.group_info_records
+        else:
+            groups = []
 
         if groups:
             for group in groups:
                 self.debug("Adding coordinator to 0x%04x group id", group.group_id)
                 await coordinator.add_to_group(group.group_id)
         else:
-            await coordinator.add_to_group(0x0000, name="Default Lightlink Group")
+            await coordinator.add_to_group(0x0000, name="Lightlink Group")

zha/zigbee/endpoint.py

@@ -192,7 +192,16 @@ async def _execute_handler_tasks(
             *self.claimed_cluster_handlers.values(),
             *self.client_cluster_handlers.values(),
         ]
-        tasks = [getattr(ch, func_name)(*args) for ch in cluster_handlers]
+
+        faults: list[Exception] = []
+
+        async def caller(ch: ClusterHandler):
+            try:
+                await getattr(ch, func_name)(*args)
+            except Exception as outcome:
+                faults.append(outcome)
+
+        tasks = [caller(ch) for ch in cluster_handlers]
 
         gather: Callable[..., Awaitable]
 
@@ -201,14 +210,9 @@ async def _execute_handler_tasks(
         else:
             gather = functools.partial(gather_with_limited_concurrency, max_concurrency)
 
-        results = await gather(*tasks, return_exceptions=True)
-        for cluster_handler, outcome in zip(cluster_handlers, results):
-            if isinstance(outcome, Exception):
-                cluster_handler.debug(
-                    "'%s' stage failed: %s", func_name, str(outcome), exc_info=outcome
-                )
-            else:
-                cluster_handler.debug("'%s' stage succeeded", func_name)
+        await gather(*tasks)
+        if faults:
+            raise ExceptionGroup(f"{func_name}: some clusters failed", faults)
 
     def async_new_entity(
         self,