From 45039bd62167a5339e3c377485e85944dae358d1 Mon Sep 17 00:00:00 2001
From: anti <samuel@securejump.cl>
Date: Fri, 17 Apr 2026 16:23:00 -0400
Subject: [PATCH] fix(cache): lazy-init TTL cache locks to survive event-loop
 turnover

A module-level asyncio.Lock binds to the loop it was first awaited on.
Under pytest-anyio (and xdist) each test spins up a new loop; any later
test that hit /health or /config would wait on a lock owned by a dead
loop and the whole worker would hang.

Create the lock on first use and drop it in the test-reset helpers so a
fresh loop always gets a fresh lock.
---
 decnet/web/router/config/api_get_config.py |  3 +++
 decnet/web/router/health/api_get_health.py | 11 ++++++++---
 2 files changed, 11 insertions(+), 3 deletions(-)

diff --git a/decnet/web/router/config/api_get_config.py b/decnet/web/router/config/api_get_config.py
index e47cceb..3e751aa 100644
--- a/decnet/web/router/config/api_get_config.py
+++ b/decnet/web/router/config/api_get_config.py
@@ -24,6 +24,9 @@ _state_locks: dict[str, asyncio.Lock] = {}
 def _reset_state_cache() -> None:
     """Reset cached config state — used by tests."""
     _state_cache.clear()
+    # Drop any locks bound to the previous event loop — reusing one from
+    # a dead loop deadlocks the next test.
+    _state_locks.clear()
 
 
 async def _get_state_cached(name: str) -> Optional[dict[str, Any]]:
diff --git a/decnet/web/router/health/api_get_health.py b/decnet/web/router/health/api_get_health.py
index f9bac90..ad39d76 100644
--- a/decnet/web/router/health/api_get_health.py
+++ b/decnet/web/router/health/api_get_health.py
@@ -24,7 +24,9 @@ _DOCKER_CHECK_INTERVAL = 5.0  # seconds between actual Docker pings
 # repo.get_total_logs() and filling the aiosqlite queue.
 _db_component: Optional[ComponentHealth] = None
 _db_last_check: float = 0.0
-_db_lock = asyncio.Lock()
+# Lazy-init — an asyncio.Lock bound to a dead event loop deadlocks any
+# later test running under a fresh loop.  Create on first use.
+_db_lock: Optional[asyncio.Lock] = None
 _DB_CHECK_INTERVAL = 1.0  # seconds
 
 
@@ -39,16 +41,19 @@ def _reset_docker_cache() -> None:
 
 def _reset_db_cache() -> None:
     """Reset cached DB liveness — used by tests."""
-    global _db_component, _db_last_check
+    global _db_component, _db_last_check, _db_lock
     _db_component = None
     _db_last_check = 0.0
+    _db_lock = None
 
 
 async def _check_database_cached() -> ComponentHealth:
-    global _db_component, _db_last_check
+    global _db_component, _db_last_check, _db_lock
     now = time.monotonic()
     if _db_component is not None and now - _db_last_check < _DB_CHECK_INTERVAL:
         return _db_component
+    if _db_lock is None:
+        _db_lock = asyncio.Lock()
     async with _db_lock:
         now = time.monotonic()
         if _db_component is not None and now - _db_last_check < _DB_CHECK_INTERVAL: