Update RGBDS master documentation

Author: GitHub Action

docs/gbz80.7.pdf

@@ -467,6 +467,24 @@ <h3 class="Ss" id="**_operator_associativity"><a class="permalink" href="#**_ope
 <p class="Pp">Instead, now we have &#x2018;<code class="Li">p ** q ** r == p **
     (q ** r)</code>&#x2019;.</p>
 </section>
+<section class="Ss">
+<h3 class="Ss" id="8-bit_and_5-bit_color_conversion"><a class="permalink" href="#8-bit_and_5-bit_color_conversion">8-bit
+  and 5-bit color conversion</a></h3>
+<p class="Pp">Changed in 1.0.0.</p>
+<p class="Pp">RGBGFX takes 8-bit RGB colors as its PNG input, and outputs 5-bit
+    GBC colors. Its &#x2018;<code class="Li">-r/--reverse</code>&#x2019; mode
+    does the opposite 5-bit to 8-bit conversion. Instead of the previous
+    inaccurate conversions, we now do accurate rounding to the nearest
+    equivalent.</p>
+<p class="Pp">Previously to convert an 8-bit color channel to 5-bit, we
+    truncated it as &#x2018;<code class="Li">c &gt;&gt; 3</code>&#x2019;; and to
+    reverse a 5-bit color channel to 8-bit, we extended it as
+    &#x2018;<code class="Li">(c &lt;&lt; 3) | (c &gt;&gt; 2)</code>&#x2019;.</p>
+<p class="Pp">Instead, now we round 8-bit to 5-bit as
+    &#x2018;<code class="Li">(c * 31 + 127) / 255</code>&#x2019;, and round
+    5-bit to 8-bit as &#x2018;<code class="Li">(c * 255 + 15) /
+    31</code>&#x2019;.</p>
+</section>
 </section>
 <section class="Sh">
 <h2 class="Sh" id="BUGS"><a class="permalink" href="#BUGS">BUGS</a></h2>

docs/rgbasm-old.5.html

@@ -215,6 +215,11 @@ export const toc = [
 	"id": "**_operator_associativity",
 	"level": 3,
 },
+{
+	"value": "8-bit and 5-bit color conversion",
+	"id": "8-bit_and_5-bit_color_conversion",
+	"level": 3,
+},
 {
 	"value": "BUGS",
 	"id": "BUGS",