Interesting, we’ve recently received a very similar report here:
Can you try the same steps described there?
We’re tracking this as the following issue: