Code review comment for lp:~vanvugt/mir/target

You're right, this is just "sampling" input in the most efficient way. Not "re-sampling". I'll adjust the wording...

Indeed, not all apps can work this way. Those that have not been written to handle input efficiently will need the built-in input resampling. Also those that benefit from touch motion prediction will want that. But in theory any app/toolkit could be given optimal input handling by:

  1. Only wake up on input events.
  2. When woken, schedule a frame if none scheduled already.
  3. At the start of each frame, /get/ the latest input state (i.e. current cursor position).
  4. Use the latest input state to render your frame.

This is all stuff I remember being taught in 1998. It's kind of embarrassing we forget it so often.