setClusterState should be preserved to handle cluster states.

The implementation lacks some checks around the handling of
setClusterState, so add a TODO about this so we can get rid of former
XXX comment.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1692 71dcc9de-d417-0410-9af5-da40c76e7ee4

neo/master/handlers/administration.py

@@ -35,7 +35,8 @@ class AdministrationHandler(MasterHandler):
         conn.answer(Packets.AnswerPrimary(app.uuid, []))
 
     def setClusterState(self, conn, state):
-        # XXX: dedicate a packet to start the cluster
+        # TODO: check which cluster state transition we are told to follow,
+        # and refuse to follow meaningless ones.
         if state == ClusterStates.VERIFYING:
             self.app._startup_allowed = True
         else: