render: fix rounding error when calculating background color with alpha

We use pre-multiplied alpha color channels, but were having bad
rounding errors due to the alpha divider being truncated to an
integer.

The algorithm for pre-multiplying a color channel is:

  alpha_divider = 0xffff / alpha
  pre_mult_color = color / alpha_divider

In order to fix the rounding errors, we could turn ‘alpha_divider’
into a double.

That however would introduce a performance penalty since now we’d need
to do floating point math for each cell.

The algorithm can be trivially converted to:

  pre_mult_color = color * alpha / 0xffff

Since both color and alpa values are < 65536, the multiplication is
“safe”; it will not overflow an uint32_t.

Closes #249

CHANGELOG.md

@@ -28,6 +28,8 @@
 
 * Missing dependencies in meson, causing heavily parallelized builds
   to fail.
+* Background color when alpha < 1.0 being wrong
+  (https://codeberg.org/dnkl/foot/issues/249).
 
 
 ### Security

render.c

@@ -213,14 +213,10 @@ attrs_to_font(const struct terminal *term, const struct attributes *attrs)
 static inline pixman_color_t
 color_hex_to_pixman_with_alpha(uint32_t color, uint16_t alpha)
 {
-    if (alpha == 0)
-        return (pixman_color_t){0, 0, 0, 0};
-
-    int alpha_div = 0xffff / alpha;
     return (pixman_color_t){
-        .red =   ((color >> 16 & 0xff) | (color >> 8 & 0xff00)) / alpha_div,
-        .green = ((color >>  8 & 0xff) | (color >> 0 & 0xff00)) / alpha_div,
-        .blue =  ((color >>  0 & 0xff) | (color << 8 & 0xff00)) / alpha_div,
+        .red =   ((color >> 16 & 0xff) | (color >> 8 & 0xff00)) * alpha / 0xffff,
+        .green = ((color >>  8 & 0xff) | (color >> 0 & 0xff00)) * alpha / 0xffff,
+        .blue =  ((color >>  0 & 0xff) | (color << 8 & 0xff00)) * alpha / 0xffff,
         .alpha = alpha,
     };
 }