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

[Robert Pluim]

>>>>> On Tue, 21 Sep 2021 14:31:10 +0300, Eli Zaretskii <eliz@gnu.org> said:

    >> Date: Tue, 21 Sep 2021 13:54:02 +0300
    >> From: Eli Zaretskii <eliz@gnu.org>
    >> Cc: rgm@gnu.org, 39799@debbugs.gnu.org, mfabian@redhat.com
    >> 
    >> > I think this means you'd have to add the Variation Selectors to the
    >> > emoji script
    >> 
    >> Yes, of course.

    Eli> Btw, we could recognize VS-16 explicitly in font_range, and avoid
    Eli> putting them into the 'emoji' script.  But that would be too kludgey,
    Eli> I guess.

FE0F is already in script-representative-chars for emoji :-)

Eli, is this the kind of thing you were thinking of? Seems to work so
far (with a small addition to blocks.awk).

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

diff --git a/src/composite.c b/src/composite.c
index e97f8e2b4c..85485e9358 100644
--- a/src/composite.c
+++ b/src/composite.c
@@ -882,14 +882,15 @@ fill_gstring_body (Lisp_Object gstring)
 /* Try to compose the characters at CHARPOS according to composition
    rule RULE ([PATTERN PREV-CHARS FUNC]).  LIMIT limits the characters
    to compose.  STRING, if not nil, is a target string.  WIN is a
-   window where the characters are being displayed.  If characters are
+   window where the characters are being displayed.  TRIGGER is the
+   character that triggered the composition check.  If characters are
    successfully composed, return the composition as a glyph-string
    object.  Otherwise return nil.  */
 
 static Lisp_Object
 autocmp_chars (Lisp_Object rule, ptrdiff_t charpos, ptrdiff_t bytepos,
               ptrdiff_t limit, struct window *win, struct face *face,
-              Lisp_Object string, Lisp_Object direction)
+              Lisp_Object string, Lisp_Object direction, int trigger)
 {
   ptrdiff_t count = SPECPDL_INDEX ();
   Lisp_Object pos = make_fixnum (charpos);
@@ -920,7 +921,7 @@ autocmp_chars (Lisp_Object rule, ptrdiff_t charpos, 
ptrdiff_t bytepos,
   struct frame *f = XFRAME (font_object);
   if (FRAME_WINDOW_P (f))
     {
-      font_object = font_range (charpos, bytepos, &to, win, face, string);
+      font_object = font_range (charpos, bytepos, &to, win, face, string, 
trigger);
       if (! FONT_OBJECT_P (font_object)
          || (! NILP (re)
              && to < limit
@@ -1269,7 +1270,7 @@ composition_reseat_it (struct composition_it *cmp_it, 
ptrdiff_t charpos,
              if (XFIXNAT (AREF (elt, 1)) != cmp_it->lookback)
                goto no_composition;
              lgstring = autocmp_chars (elt, charpos, bytepos, endpos,
-                                       w, face, string, direction);
+                                       w, face, string, direction, cmp_it->ch);
              if (composition_gstring_p (lgstring))
                break;
              lgstring = Qnil;
@@ -1307,7 +1308,7 @@ composition_reseat_it (struct composition_it *cmp_it, 
ptrdiff_t charpos,
          else
            direction = QR2L;
          lgstring = autocmp_chars (elt, cpos, bpos, charpos + 1, w, face,
-                                   string, direction);
+                                   string, direction, cmp_it->ch);
          if (! composition_gstring_p (lgstring)
              || cpos + LGSTRING_CHAR_LEN (lgstring) - 1 != charpos)
            /* Composition failed or didn't cover the current
@@ -1676,7 +1677,7 @@ find_automatic_composition (ptrdiff_t pos, ptrdiff_t 
limit, ptrdiff_t backlim,
                  for (check = cur; check_pos < check.pos; )
                    BACKWARD_CHAR (check, stop);
                  *gstring = autocmp_chars (elt, check.pos, check.pos_byte,
-                                           tail, w, NULL, string, Qnil);
+                                           tail, w, NULL, string, Qnil, c);
                  need_adjustment = 1;
                  if (NILP (*gstring))
                    {
diff --git a/src/font.c b/src/font.c
index e043ef8d01..74a1214b38 100644
--- a/src/font.c
+++ b/src/font.c
@@ -3866,6 +3866,9 @@ font_at (int c, ptrdiff_t pos, struct face *face, struct 
window *w,
    If STRING is not nil, it is the string to check instead of the current
    buffer.  In that case, FACE must be not NULL.
 
+   TRIGGER is the character that actually caused the composition
+   process to start, it may be different from the character at POS.
+
    The return value is the font-object for the character at POS.
    *LIMIT is set to the position where that font can't be used.
 
@@ -3873,15 +3876,16 @@ font_at (int c, ptrdiff_t pos, struct face *face, 
struct window *w,
 
 Lisp_Object
 font_range (ptrdiff_t pos, ptrdiff_t pos_byte, ptrdiff_t *limit,
-           struct window *w, struct face *face, Lisp_Object string)
+           struct window *w, struct face *face, Lisp_Object string,
+           int trigger)
 {
   ptrdiff_t ignore;
   int c;
   Lisp_Object font_object = Qnil;
+  struct frame *f = XFRAME (w->frame);
 
   if (!face)
     {
-      struct frame *f = XFRAME (w->frame);
       int face_id;
 
       if (NILP (string))
@@ -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;
@@ -5423,6 +5444,7 @@ syms_of_font (void)
   DEFSYM (Qiso8859_1, "iso8859-1");
   DEFSYM (Qiso10646_1, "iso10646-1");
   DEFSYM (Qunicode_bmp, "unicode-bmp");
+  DEFSYM (Qemoji, "emoji");
 
   /* Symbols representing keys of font extra info.  */
   DEFSYM (QCotf, ":otf");
diff --git a/src/font.h b/src/font.h
index d3e1530642..1da72cca07 100644
--- a/src/font.h
+++ b/src/font.h
@@ -885,7 +885,7 @@ valid_font_driver (struct font_driver const *d)
 extern Lisp_Object font_update_drivers (struct frame *f, Lisp_Object list);
 extern Lisp_Object font_range (ptrdiff_t, ptrdiff_t, ptrdiff_t *,
                               struct window *, struct face *,
-                              Lisp_Object);
+                              Lisp_Object, int);
 extern void font_fill_lglyph_metrics (Lisp_Object, struct font *, unsigned 
int);
 
 extern Lisp_Object font_put_extra (Lisp_Object font, Lisp_Object prop,