not sure if this is because we don’t bundle up these events and send them one at a time so there’s a backlog we need to work through for scrolling, or if rendering is a bottleneck.
Rendering seems to be quick though, so I’m fairly certain we just need to bundle up scroll events and send one big scroll + scrolled_by_x_lines event every 16ms (60 fps) or so.
The alternative would be to change the implementation of
vximto actually be immediate mode and calling aframefunction every 16ms or so with any events since last frame similar to dvui. That would likely make the issue go away, and might end up being a better design. It would be more complicated though, and we'd likely want to bunch up the scroll events anyway, so it seems like it might be overkill anyway.Might be neat to explore though.