test(memory): hoisted-retriever fixture uses a real tmp db, not :memory:

Review follow-up on the #76142 salvage: MemoryStore path-resolves and
shares one process-wide connection per file, so MemoryStore(":memory:")
creates a literal ./:memory: FILE whose state leaks across test runs —
the second run of the file failed all three spy tests because the
NULL-vector test had permanently wiped hrr_vector in the leaked db.
tmp_path isolates each run; verified two consecutive runs green + full
tests/plugins/memory/ green.

tests/plugins/memory/test_holographic_retrieval.py

@@ -108,10 +108,15 @@ from plugins.memory.holographic import holographic as hrr
 
 
 @pytest.fixture
-def hoisted_retriever():
+def hoisted_retriever(tmp_path):
     """30 facts with HRR vectors, default dim (smaller dims trip an
-    inhomogeneous-shape edge in the fact encoder)."""
-    store = MemoryStore(":memory:")
+    inhomogeneous-shape edge in the fact encoder).
+
+    NOTE: a real tmp_path db, NOT ":memory:" — MemoryStore resolves the
+    path and shares one process-wide connection per file, so ":memory:"
+    becomes a literal ./:memory: file that leaks state across runs (and
+    the NULL-vector test below would permanently corrupt it)."""
+    store = MemoryStore(str(tmp_path / "hoist_store.db"))
     for i in range(30):
         store.add_fact(
             content=f"deploy target {i} setting alpha beta gamma option {i % 7}",