compute monitoring deltas on each poll

this means this monitoring code does not need to know if a cache-refresh happened or not

subtleties: if someone modifies the poller dict elsewhere, this won't catch that becuase its sharing that mutable old state
and so that mutator needs to send a monitoring notification themselves.

that's how things are already i think: pollitem.scale_out and pollitem.scale_in already do that monitoring change

this is probably broken - get some better testing. or perhaps redo how monitoring deltas happen entirely?

Author: benclifford

parsl/jobs/job_status_poller.py

@@ -23,10 +23,16 @@ def __init__(self, executor: BlockProviderExecutor, monitoring: Optional["parsl.
 
     def poll(self) -> None:
         now = time.time()
+        previous_status = self.executor._poller_mutable_status
+
         if now >= self._last_poll_time + self._executor.status_polling_interval:
-            previous_status = self._executor._poller_mutable_status
             self._executor._poller_mutable_status = self._executor.status()
             self._last_poll_time = now
+
+        if previous_status != self.executor._poller_mutable_status:
+            # short circuit the case where the two objects are identical so
+            # delta_status must end up empty.
+
             delta_status = {}
             for block_id in self._executor._poller_mutable_status:
                 if block_id not in previous_status \