Code review comment for lp://staging/~vanvugt/mir/mesa-double

There's an obvious visual difference when dragging things around with this branch. And also tests added for the theory being correct. Not sure what else we can do in terms of measurement.

You do see a very slight amount of lag remaining even with this branch (mouse pointer ahead of the shell rendering). But that's expected. If and when we get to a zero-lag prototype, the visual proof will then be conclusive if the mouse pointer sticks perfectly.