test: make lock-propagation comment more concise

Author: vdusek

tests/integration/test_request_queue.py

@@ -563,13 +563,9 @@ async def test_request_queue_unlock_requests(client: ApifyClient | ApifyClientAs
         assert len(lock_response.items) == 3
         locked_ids = {item.id for item in lock_response.items}
 
-        # Wait for the locks to propagate before unlocking (eventual consistency): the lock-head response
-        # is acknowledged before all three lock writes are visible to subsequent reads, so unlocking
-        # immediately can see fewer locks than were just acquired. Locked requests are excluded from the
-        # queue head, so poll `list_head` until none of the locked requests reappear there, rather than
-        # guessing a fixed sleep. Note this is a best-effort mitigation: the head read and the unlock may
-        # be served by different replicas, so convergence of one does not guarantee the other — but it
-        # narrows the race window enough in practice.
+        # Locks are acknowledged before they are visible to subsequent reads, so unlocking immediately can
+        # see fewer locks than were just acquired. Since locked requests are excluded from the queue head,
+        # poll `list_head` until the locked IDs disappear from it (best-effort mitigation of the race).
         async def all_locks_visible() -> bool:
             head = await maybe_await(rq_client.list_head(limit=5))
             assert isinstance(head, RequestQueueHead)