Check if SAVE_LIMIT is bigger than 0 before running any queries.

On MS SQL Server, the select_for_update() query below will take a
ROWLOCK,UPDLOCK on every row in the Task table. This happens because it locks
every row it needs to look at to return the result of the query.
Since there is no index on the default sorting column "stopped", it has
to do a full table scan, and hence it will take rowlocks on each row
in the table.
On other databases, it will still do a full table scan, but without the
excessive locking.
If SAVE_LIMIT is 0 the result of this query is unused, so by moving the
check a level higher we avoid running this expensive query.

django_q/cluster.py

@@ -474,10 +474,11 @@ def save_task(task, broker: Broker):
     # SAVE LIMIT > 0: Prune database, SAVE_LIMIT 0: No pruning
     close_old_django_connections()
     try:
-        with db.transaction.atomic():
-            last = Success.objects.select_for_update().last()
-            if task["success"] and 0 < Conf.SAVE_LIMIT <= Success.objects.count():
-                last.delete()
+        if Conf.SAVE_LIMIT > 0:
+            with db.transaction.atomic():
+                last = Success.objects.select_for_update().last()
+                if task["success"] and Conf.SAVE_LIMIT <= Success.objects.count():
+                    last.delete()
         # check if this task has previous results
         if Task.objects.filter(id=task["id"], name=task["name"]).exists():
             existing_task = Task.objects.get(id=task["id"], name=task["name"])