[kandinsky/font] Add comment about UTF-8

kandinsky/include/kandinsky/font.h

@@ -8,6 +8,21 @@
 #include <kandinsky/unicode/codepoint.h>
 #include "palette.h"
 
+/* We use UTF-8 encoding. This means that a character is encoded as a code point
+ * that uses between 1 and 4 bytes. Code points can be "combining", in which
+ * case their glyph should be superimposed to the glyph of the previous code
+ * point in the string. This is for instance used to print accents: the string
+ * for the glyph 'è' is composed of the code point for 'e' followed by the
+ * combining code point for '`'.
+ * ASCII characters have the same encoding in ASCII and in UTF-8.
+ *
+ * We do not provide a glyph for each of the 1,112,064 valid UTF-8 code points.
+ * We thus have a table of the glyphs we can draw (uint32_t Codepoints[] in
+ * kandinsky/fonts/codepoints.h). To easily compute the index of a code point in
+ * the Codepoints table, we use the m_table matching table: it contains the
+ * CodepointIndexPairs of the first code point of each series of consecutive
+ * code points in the Codepoints table. */
+
 class KDFont {
 private:
   static constexpr int k_bitsPerPixel = 4; // TODO: Should be generated by the rasterizer