This is awesome. A nice improvement would be to have the loop continue with the old code when there are errors (rather than pausing iteration).
- Always run from the back buffer
- Always scan the updated file into the front buffer
- If the front buffer is fresh AND the front buffer contains no bugs, bufferswap