Code review comment for lp:~kdub/mir/RFC-surface-arrangements

@Chris,
Agreed that it needs some sorting...

So, given that "parent surface" == the /stream/ that is generated by default when the surface is created...

For the "glorious future" nested-passthrough scenario, we'll have a certain set of surfaces that we're passing through, and then one surface that's used as the GL surface nested can optionally draw to. We might want to hide the parent surface's stream, and not swap it. I'm still sketching through how synchronizing the swaps+updates might work, but it makes sense to me to somehow delay the swaps of the selected surfaces until a mir_surface_spec_commit_changes() is called.