I Think there might be a test bug where events at the edge of the polling range get lost because the filter's range might be exclusive in ganache (although it should not be).
since the model de-dupes events well, I suggest we overlap the query range by one block.
I Think there might be a test bug where events at the edge of the polling range get lost because the filter's range might be exclusive in ganache (although it should not be).
since the model de-dupes events well, I suggest we overlap the query range by one block.