[permissions] Add screen sharing permissions (#45)

This was discussed at the WebRTC/Media Capture meeting.  

The `PermissionDescriptor` will look very similar to that for `getUserMedia`, so this should await the resolution of #10.

The [screen sharing API](https://w3c.github.io/mediacapture-screen-share) has an additional state that the working group wanted to be able to examine using `query()`.  That additional state is needed to indicate that some additional user action is required before prompting the user is even possible.  Understanding why that is necessary depends on understanding the security problems associated with screen sharing.

This will sometimes depend on the parameters passed in the descriptor.  If the JS requests access to a screen rather than a window, this will always be blocked unless the user has taken an action.  In that case, the meaning of "prompt" might be OK.  However, for window sharing, the user prompt will contain a subset of the options unless the extra steps have been taken.  Applications will need/want to know about that.

Queries for screen sharing will never produce a state of "allow", but I expect that that won't be an issue.

---
Reply to this email directly or view it on GitHub:
https://github.com/w3c/permissions/issues/45