Avoid floating point arithmetic in RGB→8-bit ANSI approximation

src/terminal.rs

@@ -13,14 +13,10 @@ fn rgb2ansi(r: u8, g: u8, b: u8) -> u8 {
         } else if r > 248 {
             WHITE
         } else {
-            let fr = f32::from(r);
-            (((fr - 8.) / 247.) * 24.) as u8 + 232
+            ((r - 8) as u16 * 24 / 247) as u8 + 232
         }
     } else {
-        let fr = f32::from(r);
-        let fg = f32::from(g);
-        let fb = f32::from(b);
-        16 + (36 * (fr / 255. * 5.) as u8) + (6 * (fg / 255. * 5.) as u8) + (fb / 255. * 5.) as u8
+        36 * (r / 51) + 6 * (g / 51) + (b / 51) + 16
     }
 }