Add special handling for scripts when flushed from cache to avoid duplicate ExtendedLoopingCalls.

evennia/scripts/scripts.py

@@ -18,6 +18,17 @@ from future.utils import with_metaclass
 __all__ = ["DefaultScript", "DoNothing", "Store"]
 
 
+FLUSHING_INSTANCES = False  # whether we're in the process of flushing scripts from the cache
+SCRIPT_FLUSH_TIMERS = {}  # stores timers for scripts that are currently being flushed
+
+
+def restart_scripts_after_flush():
+    """After instances are flushed, validate scripts so they're not dead for a long period of time"""
+    global FLUSHING_INSTANCES
+    ScriptDB.objects.validate()
+    FLUSHING_INSTANCES = False
+
+
 class ExtendedLoopingCall(LoopingCall):
     """
     LoopingCall that can start at a delay different
@@ -278,6 +289,27 @@ class DefaultScript(ScriptBase):
             return max(0, self.db_repeats - task.callcount)
         return None
 
+    def at_idmapper_flush(self):
+        """If we're flushing this object, make sure the LoopingCall is gone too"""
+        ret = super(DefaultScript, self).at_idmapper_flush()
+        if ret:
+            try:
+                from twisted.internet import reactor
+                global FLUSHING_INSTANCES
+                # store the current timers for the _task and stop it to avoid duplicates after cache flush
+                paused_time = self.ndb._task.next_call_time()
+                callcount = self.ndb._task.callcount
+                self._stop_task()
+                SCRIPT_FLUSH_TIMERS[self.id] = (paused_time, callcount)
+                # here we ensure that the restart call only happens once, not once per script
+                if not FLUSHING_INSTANCES:
+                    FLUSHING_INSTANCES = True
+                    reactor.callLater(2, restart_scripts_after_flush)
+            except Exception:
+                import traceback
+                traceback.print_exc()
+        return ret
+
     def start(self, force_restart=False):
         """
         Called every time the script is started (for persistent
@@ -294,9 +326,19 @@ class DefaultScript(ScriptBase):
                 started or not. Used in counting.
 
         """
-
         if self.is_active and not force_restart:
-            # script already runs and should not be restarted.
+            # The script is already running, but make sure we have a _task if this is after a cache flush
+            if not self.ndb._task:
+                self.ndb._task = ExtendedLoopingCall(self._step_task)
+                try:
+                    start_delay, callcount = SCRIPT_FLUSH_TIMERS[self.id]
+                    del SCRIPT_FLUSH_TIMERS[self.id]
+                    now = False
+                except (KeyError, ValueError, TypeError):
+                    now = not self.db_start_delay
+                    start_delay = None
+                    callcount = 0
+                self.ndb._task.start(self.db_interval, now=now, start_delay=start_delay, count_start=callcount)
             return 0
 
         obj = self.obj