Don't process events we've already processed. Remember to process state events

synapse/handlers/federation.py

@@ -250,6 +250,7 @@ class FederationHandler(BaseHandler):
         # For each edge get the current state.
 
         auth_events = {}
+        state_events = {}
         events_to_state = {}
         for e_id in edges:
             state, auth = yield self.replication_layer.get_state_for_room(
@@ -258,8 +259,13 @@ class FederationHandler(BaseHandler):
                 event_id=e_id
             )
             auth_events.update({a.event_id: a for a in auth})
+            state_events.update({s.event_id: s for s in state})
             events_to_state[e_id] = state
 
+        seen_events = yield self.store.have_events(
+            set(auth_events.keys()) | set(state_events.keys())
+        )
+
         yield defer.gatherResults(
             [
                 self._handle_new_event(
@@ -270,6 +276,22 @@ class FederationHandler(BaseHandler):
                     }
                 )
                 for a in auth_events.values()
+                if a.event_id not in seen_events
+            ],
+            consumeErrors=True,
+        ).addErrback(unwrapFirstError)
+
+        yield defer.gatherResults(
+            [
+                self._handle_new_event(
+                    dest, s,
+                    auth_events={
+                        (e.type, e.state_key): e for e in auth_events
+                        if e.event_id in [a_id for a_id, _ in s.auth_events]
+                    }
+                )
+                for s in state_events.values()
+                if s.event_id not in seen_events
             ],
             consumeErrors=True,
         ).addErrback(unwrapFirstError)