While discussing #322, the question came up of whether gDM should consume the transient activation which it requires. At first glance, this seems reasonable:

• There might be some potential attacks that could be mitigated here.
• I cannot immediately think of legitimate reasons to allow the activation to be used for anything in addition to gDM. All use cases I can presently imagine would require a new activation after the user finishes their interaction with the gDM-dialog, which is exactly why we produced getDisplayMedia's promise confers transient activation when resolved #322.

My intuition here is that this is at least non-harmful, so let's do it. Can you think of any novel devil's advocate's arguments against making this change, @youennf, @jan-ivar?