fix(tests+mutator): unbreak the docker-shadow test env + let mutator delete from active

Two related fixes that came out of running the W5 tests locally:

1. tests/__init__.py — empty file, makes 'tests/' a package so pytest
   stops inserting it into sys.path.  Without it, 'tests/docker/'
   (the docker-image test category) shadowed the installed docker SDK
   on every engine-touching test in the repo:

     module 'docker' has no attribute 'DockerClient'

   Pytest's default --import-mode=prepend was the culprit; making
   tests/ a package is the cheapest fix and doesn't change
   --import-mode for the whole tree.

2. delete_topology_decky / delete_topology_edge / delete_lan grow an
   'enforce_pending: bool = True' kwarg.  Default preserves the HTTP
   CRUD guard (api_decky_crud / api_edge_crud / api_lan_crud get the
   409 for free).  apply_remove_decky / apply_detach_decky /
   apply_remove_lan now pass enforce_pending=False — the mutator
   queue is the live-editing surface and has its own active-topology
   gating; the repo's pending-only guard was for design-time CRUD
   that mustn't bypass it.  Without this, apply_remove_decky was
   silently broken on active topologies pre-W5; W5's new test
   surfaced it on first run.

10/10 new W5 tests pass; 58/58 across mutator + topology suites.
This commit is contained in:
2026-04-29 00:24:17 -04:00
parent 98c929894c
commit a27e3f5e0f
6 changed files with 52 additions and 11 deletions

View File

@@ -558,7 +558,11 @@ async def apply_remove_lan(
f"{d['decky_config']['name']!r}; remove the decky first"
)
lan_name = lan["name"]
await repo.delete_lan(lan["id"])
# enforce_pending=False: the mutator queue is the live-editing
# surface, gated on topology status by us before we got here. The
# repo's pending-only guard is for HTTP CRUD callers that mustn't
# bypass it.
await repo.delete_lan(lan["id"], enforce_pending=False)
# Live materialisation symmetric to apply_add_lan: tear down the
# docker bridge and re-render compose so a future redeploy doesn't
@@ -757,7 +761,7 @@ async def apply_detach_decky(
await repo.update_topology_decky(
decky["uuid"], {"decky_config": new_cfg}
)
await repo.delete_topology_edge(edge["id"])
await repo.delete_topology_edge(edge["id"], enforce_pending=False)
# Live materialisation: SDK network.disconnect on the base
# container. Service containers automatically lose visibility into
# the LAN because they share the base's netns.
@@ -778,7 +782,7 @@ async def apply_remove_decky(
raise MutationError(f"decky {payload['decky']!r} not found")
decky_name = decky["decky_config"]["name"]
services_list = list(decky.get("services") or [])
await repo.delete_topology_decky(decky["uuid"])
await repo.delete_topology_decky(decky["uuid"], enforce_pending=False)
# Live materialisation: stop + rm -f the decky's containers. We
# capture decky_name + services BEFORE the delete so the helper
# has the targets even though the row is gone.

View File

@@ -787,17 +787,29 @@ class BaseRepository(ABC):
# -------------------- pre-deploy (pending-only) mutations --------------------
async def delete_lan(
self, lan_id: str, *, expected_version: Optional[int] = None
self,
lan_id: str,
*,
expected_version: Optional[int] = None,
enforce_pending: bool = True,
) -> None:
raise NotImplementedError
async def delete_topology_decky(
self, decky_uuid: str, *, expected_version: Optional[int] = None
self,
decky_uuid: str,
*,
expected_version: Optional[int] = None,
enforce_pending: bool = True,
) -> None:
raise NotImplementedError
async def delete_topology_edge(
self, edge_id: str, *, expected_version: Optional[int] = None
self,
edge_id: str,
*,
expected_version: Optional[int] = None,
enforce_pending: bool = True,
) -> None:
raise NotImplementedError

View File

@@ -72,8 +72,19 @@ class TopologyDeckiesMixin:
decky_uuid: str,
*,
expected_version: Optional[int] = None,
enforce_pending: bool = True,
) -> None:
"""Cascade-delete a decky + all its edges from a pending topology."""
"""Cascade-delete a decky + all its edges from a topology.
Defaults to ``enforce_pending=True`` so HTTP CRUD callers
(api_decky_crud.py) get the existing 409 guard for free. The
mutator's ``apply_remove_decky`` is the only path that's
legitimately allowed to delete from an active topology — it
passes ``enforce_pending=False`` after dequeuing the mutation
through its own active-topology gating (the queue is the live
editing surface; the repo's CRUD guard is for the design-time
endpoints that mustn't bypass it).
"""
async with self._session() as session:
result = await session.execute(
select(TopologyDecky).where(TopologyDecky.uuid == decky_uuid)
@@ -81,7 +92,8 @@ class TopologyDeckiesMixin:
d = result.scalar_one_or_none()
if d is None:
return
await self._assert_pending(session, d.topology_id)
if enforce_pending:
await self._assert_pending(session, d.topology_id)
if expected_version is not None:
await self._check_and_bump_version(
session, d.topology_id, expected_version

View File

@@ -33,7 +33,12 @@ class TopologyEdgesMixin:
edge_id: str,
*,
expected_version: Optional[int] = None,
enforce_pending: bool = True,
) -> None:
"""Delete one edge. ``enforce_pending=True`` by default — the
mutator's ``apply_detach_decky`` opts out, same rationale as
``delete_topology_decky``.
"""
async with self._session() as session:
result = await session.execute(
select(TopologyEdge).where(TopologyEdge.id == edge_id)
@@ -41,7 +46,8 @@ class TopologyEdgesMixin:
edge = result.scalar_one_or_none()
if edge is None:
return
await self._assert_pending(session, edge.topology_id)
if enforce_pending:
await self._assert_pending(session, edge.topology_id)
if expected_version is not None:
await self._check_and_bump_version(
session, edge.topology_id, expected_version

View File

@@ -61,12 +61,18 @@ class LansMixin:
lan_id: str,
*,
expected_version: Optional[int] = None,
enforce_pending: bool = True,
) -> None:
"""Cascade-delete a LAN from a pending topology.
"""Cascade-delete a LAN.
Rejects if any decky declares this LAN as its home (i.e. has a
non-bridge edge to it — the only LAN that decky lives in). The
caller must delete or reassign the home-deckies first.
``enforce_pending=True`` by default keeps the HTTP CRUD guard
intact; the mutator's ``apply_remove_lan`` opts out (it has
already gated on topology status and the live-LAN docker
materialisation runs after).
"""
from decnet.topology.status import TopologyNotEditable # noqa: F401
@@ -75,7 +81,8 @@ class LansMixin:
lan = result.scalar_one_or_none()
if lan is None:
return
await self._assert_pending(session, lan.topology_id)
if enforce_pending:
await self._assert_pending(session, lan.topology_id)
# Home-decky check: any decky whose only edge lands here?
edges_result = await session.execute(

0
tests/__init__.py Normal file
View File