bug#39799: 28.0.50; Most emoji sequences don’t render correctly

[Eli Zaretskii]

> From: Robert Pluim <rpluim@gmail.com>
> Cc: rgm@gnu.org,  39799@debbugs.gnu.org,  mfabian@redhat.com
> Date: Tue, 21 Sep 2021 19:43:05 +0200
> 
> Eli, is this the kind of thing you were thinking of? Seems to work so
> far (with a small addition to blocks.awk).

Yes, with a minor comment below.

> We'll need to find a better name for the new arg than 'trigger'
> though.

How about just 'ch'?  We use such names all over the place, so
describing what it is in the comment should be enough.

> @@ -3912,6 +3916,23 @@ font_range (ptrdiff_t pos, ptrdiff_t pos_byte, 
> ptrdiff_t *limit,
>       continue;
>        if (NILP (font_object))
>       {
> +       if (EQ (CHAR_TABLE_REF (Vchar_script_table, trigger),
> +               Qemoji))
> +         {
> +           Lisp_Object val = assq_no_quit (Qemoji, 
> Vscript_representative_chars);
> +           if (CONSP (val))
> +             {
> +               int face_id;
> +               val = XCDR (val);
> +               if (CONSP (val))
> +                 val = XCAR (val);
> +               else if (VECTORP (val))
> +                 val = AREF (val, 0);
> +               c = XFIXNAT (val);
> +               face_id = FACE_FOR_CHAR (f, face, c, pos - 1, string);
> +               face = FACE_FROM_ID (f, face_id);
> +             }
> +         }
>         font_object = font_for_char (face, c, pos - 1, string);
>         if (NILP (font_object))
>           return Qnil;

For backward compatibility, I'd prefer here, if font_for_char returns
nil for the representative Emoji character, to call font_for_char
again with the face and codepoint for the original character.

Thanks.