I can confirm the problem. We'll implement a workaround in the next update.