Partial revert of the detection method for negates that I used in the previous fix that was apparently causing the bug I observed. This version should still fix the issues I'd identified with the previous implementation of the negate tracker, but, like the original version, doesn't depend on the apparently unreliable ACTION_RESULT_SILENCED.
Partial revert of the detection method for negates that I used in the previous fix that was apparently causing the bug I observed. This version should still fix the issues I'd identified with the previous implementation of the negate tracker, but, like the original version, doesn't depend on the apparently unreliable ACTION_RESULT_SILENCED.