pub const POLY_SUBPIXEL_SHIFT: u32 = 8;
These constants determine the subpixel accuracy (number of fractional bits). With 8-bit fractional part and 32-bit integers, coordinate capacity is 24 bits.