We do not need to install separate event listeners for every callback.
Just install one listener and keep track of the callbacks ourself.

This is clearer, and also determines the order in which callbacks are
called.  (Although, we don't rely on that currently.)

This also adds a tests.

- Refactor the three visual-mode modes.
- Use the key-handling framework from #2022.
- Strip some legacy edit-mode code.
- Rename the file (the old file name was misleading).
- Add "aw" and "as", previously we had the code for this from edit mode.

This previous file name was chosen when we (I) had the intention of
implementing edit mode too.

That initiative has been abandoned, so the file name is inappropriate.

Renaming now in preparation for a significant refactoring of visual
mode.

Previously, the key-handling logic (keyQueue, etc) was and the backend
whereas passKeys were handled in the content scripts - so they were a
long way apart.

Now that they're in the same place, it makes more sense to integrate
passKey handling into the regular key handling, because they depend upon
the same data structures.

... and fix two bugs:

- not suppressing keyup event after keyChar matched in keydown.
- we cannot check the passKeys keyChar in keyup because the key state
  has changed; so we track what the next keyup response should be.

Previously, we blocked keyboard events for a fixed 200ms.

With this PR, we continue blocking keyboard events until 150ms after the
last `keydown` or `keypress` event.  So, we wait until we think the user
has stopped typing.

Fixes #1842.

When we exit visual mode with "y"...

"y" is a command for visual mode.  Currently it clears the selection.
Because it's executed as a command, scrollIntoView is called after
running the command.  Because the selection is cleared, scrollIntoView
found nothing to scroll, and was creating an error.

So we need to check whether there's anything to scroll into view, before
trying to scroll it.

Also fix mis-named file in the tests.