On Tue Oct 14 09:36:30 2025 +0000, Rémi Bernon wrote:
Any reason to use rational numbers over float? And if there's actually some, I think we should try to avoid increasing the size of each coordinate, using 16.16 bit rationals like X11 XInput or 24.8 like Wayland.
AFAIR, Windows uses fixed-point in GDI with 4-bit fractional part.