Let Fl_Text_Editor and Fl_Input handle gracefully composed unicode characters.

FL/fl_utf8.h

@@ -1,7 +1,7 @@
 /*
  * Author: Jean-Marc Lienher ( http://oksid.ch )
  * Copyright 2000-2010 by O'ksi'D.
- * Copyright 2016-2021 by Bill Spitzak and others.
+ * Copyright 2016-2026 by Bill Spitzak and others.
  *
  * This library is free software. Distribution and use rights are outlined in
  * the file "COPYING" which should have been included with this file.  If this
@@ -205,6 +205,9 @@ FL_EXPORT void fl_make_path_for_file( const char *path );
 /* OD: recursively create a path in the file system */
 FL_EXPORT char fl_make_path( const char *path );
 
+FL_EXPORT const char *fl_utf8_next_composed_char(const char *from, const char *end);
+
+FL_EXPORT const char *fl_utf8_previous_composed_char(const char *from, const char *begin);
 
 /** @} */
 

src/Fl_Input.cxx

@@ -1,7 +1,7 @@
 //
 // Input widget for the Fast Light Tool Kit (FLTK).
 //
-// Copyright 1998-2024 by Bill Spitzak and others.
+// Copyright 1998-2026 by Bill Spitzak and others.
 //
 // This library is free software. Distribution and use rights are outlined in
 // the file "COPYING" which should have been included with this file.  If this
@@ -190,14 +190,20 @@ int Fl_Input::kf_delete_eol() {
 int Fl_Input::kf_delete_char_right() {
   if (readonly()) { fl_beep(); return 1; }
   if (mark() != insert_position()) cut();
-  else cut(1);
+  else {
+    const char *next = fl_utf8_next_composed_char(value() + insert_position(), value() + size());
+    replace(insert_position(), next - value(), 0);
+  }
   return 1;
 }
 
 int Fl_Input::kf_delete_char_left() {
   if (readonly()) { fl_beep(); return 1; }
   if (mark() != insert_position()) cut();
-  else cut(-1);
+  else {
+    const char *before = fl_utf8_previous_composed_char(value() + insert_position(), value());
+    replace(insert_position(), before - value(), 0);
+  }
   return 1;
 }
 
@@ -225,7 +231,8 @@ int Fl_Input::kf_clear_eol() {
 //    If OPTION_ARROW_FOCUS is disabled, return 1 to prevent focus navigation.
 //
 int Fl_Input::kf_move_char_left() {
-  int i = shift_position(insert_position()-1) + NORMAL_INPUT_MOVE;
+  const char *before = fl_utf8_previous_composed_char(value() + insert_position(), value());
+  int i = shift_position(before - value()) + NORMAL_INPUT_MOVE;
   return Fl::option(Fl::OPTION_ARROW_FOCUS) ? i : 1;
 }
 
@@ -233,7 +240,8 @@ int Fl_Input::kf_move_char_left() {
 //    If OPTION_ARROW_FOCUS is disabled, return 1 to prevent focus navigation.
 //
 int Fl_Input::kf_move_char_right() {
-  int i = shift_position(insert_position()+1) + NORMAL_INPUT_MOVE;
+  const char *next = fl_utf8_next_composed_char(value() + insert_position(), value() + size());
+  int i = shift_position(next - value()) + NORMAL_INPUT_MOVE;
   return Fl::option(Fl::OPTION_ARROW_FOCUS) ? i : 1;
 }
 

src/Fl_Input_.cxx

@@ -1,7 +1,7 @@
 //
 // Common input widget routines for the Fast Light Tool Kit (FLTK).
 //
-// Copyright 1998-2011 by Bill Spitzak and others.
+// Copyright 1998-2026 by Bill Spitzak and others.
 //
 // This library is free software. Distribution and use rights are outlined in
 // the file "COPYING" which should have been included with this file.  If this
@@ -658,8 +658,7 @@ void Fl_Input_::handle_mouse(int X, int Y, int /*W*/, int /*H*/, int drag) {
   const char *l, *r, *t; double f0 = Fl::event_x()-X+xscroll_;
   for (l = p, r = e; l<r; ) {
     double f;
-    int cw = fl_utf8len((char)l[0]);
+    int cw = fl_utf8_next_composed_char(l, value() + size()) - l;
-    if (cw < 1) cw = 1;
     t = l+cw;
     f = X-xscroll_+expandpos(p, t, buf, 0);
     if (f <= Fl::event_x()) {l = t; f0 = Fl::event_x()-f;}
@@ -667,7 +666,7 @@ void Fl_Input_::handle_mouse(int X, int Y, int /*W*/, int /*H*/, int drag) {
   }
   if (l < e) { // see if closer to character on right:
     double f1;
-    int cw = fl_utf8len((char)l[0]);
+    int cw = fl_utf8_next_composed_char(l, value() + size()) - l;
     if (cw > 0) {
       f1 = X-xscroll_+expandpos(p, l + cw, buf, 0) - Fl::event_x();
       if (f1 < f0) l = l+cw;

src/Fl_Text_Buffer.cxx

@@ -1,5 +1,5 @@
 //
-// Copyright 2001-2023 by Bill Spitzak and others.
+// Copyright 2001-2026 by Bill Spitzak and others.
 // Original code Copyright Mark Edel.  Permission to distribute under
 // the LGPL for the FLTK library granted by Mark Edel.
 //
@@ -2090,14 +2090,8 @@ int Fl_Text_Buffer::prev_char_clipped(int pos) const
     return 0;
 
   IS_UTF8_ALIGNED2(this, (pos))
-
+  const char *previous = fl_utf8_previous_composed_char(address(0) + pos, address(0));
-  char c;
+  pos = previous - address(0);
-  do {
-    pos--;
-    if (pos==0)
-      return 0;
-    c = byte_at(pos);
-  } while ( (c&0xc0) == 0x80);
 
   IS_UTF8_ALIGNED2(this, (pos))
   return pos;
@@ -2122,8 +2116,8 @@ int Fl_Text_Buffer::prev_char(int pos) const
 int Fl_Text_Buffer::next_char(int pos) const
 {
   IS_UTF8_ALIGNED2(this, (pos))
-  int n = fl_utf8len1(byte_at(pos));
+  const char *next = fl_utf8_next_composed_char(address(0) + pos, address(0) + mLength);
-  pos += n;
+  pos = next - address(0);
   if (pos>=mLength)
     return mLength;
   IS_UTF8_ALIGNED2(this, (pos))

src/Fl_Text_Display.cxx

@@ -1,5 +1,5 @@
 //
-// Copyright 2001-2022 by Bill Spitzak and others.
+// Copyright 2001-2026 by Bill Spitzak and others.
 // Original code Copyright Mark Edel.  Permission to distribute under
 // the LGPL for the FLTK library granted by Mark Edel.
 //
@@ -2263,7 +2263,8 @@ int Fl_Text_Display::find_x(const char *s, int len, int style, int x) const {
   int i = 0;
   int last_w = 0;       // STR #2788
   while (i<len) {
-    int cl = fl_utf8len1(s[i]);
+    const char *next = fl_utf8_next_composed_char(s + i, s + len);
+    int cl = next - (s+i);
     int w = int( string_width(s, i+cl, style) );
     if (w>x) {
       if (cursor_pos && (w-x < x-last_w)) return i+cl; // STR #2788

src/fl_utf8.cxx

@@ -1629,4 +1629,64 @@ unsigned fl_utf8from_mb(char* dst, unsigned dstlen, const char* src, unsigned sr
   return Fl::system_driver()->utf8from_mb(dst, dstlen, src, srclen);
 }
 
+
+/**
+ Returns pointer to beginning of next unicode character after potentially composed character.
+ Some unicode characters (example: 👩‍✈️ "woman pilot") are composed of several unicode points. They may pair two successive
+ codepoints with U+200D (zero-width joiner) and may qualify any component with variation selectors or Fitzpatrick emoji modifiers.
+ \param from points to a location within a UTF8 string. If this location is inside the UTF8
+ encoding of a codepoint or is an invalid byte, this function returns \p from + 1.
+ \param end points past last codepoint of the string.
+ \return pointer to start of first codepoint after potentially composed character beginning at \p from.
+ */
+const char *fl_utf8_next_composed_char(const char *from, const char *end) {
+  int skip = fl_utf8len(*from);
+  if (skip == -1) return from + 1;
+  from += skip; // skip 1st codepoint
+  while (from < end) {
+    unsigned u = fl_utf8decode(from, end, NULL);
+    if (u == 0x200D) { // zero-width joiner
+      from += fl_utf8len(*from); // skip joiner
+      from += fl_utf8len(*from); // skip joined codepoint
+    } else if (u >= 0xFE00 && u <= 0xFE0F) { // a variation selector
+      from += fl_utf8len(*from); // skip variation selector
+    } else if (u >= 0x1F3FB && u <= 0x1F3FF) { // EMOJI MODIFIER FITZPATRICK
+      from += fl_utf8len(*from); // skip modifier
+    } else break;
+  }
+  return from;
+}
+
+
+/**
+ Returns pointer to beginning of previous potentially composed character before given unicode character.
+ See fl_utf8_next_composed_char() for a hint about what is a composed unicode character.
+ \param from points to a location within a UTF8 string. If this location is inside the UTF8
+ encoding of a codepoint or is an invalid byte, this function returns \p from - 1.
+ \param begin points to start of first codepoint of the string.
+ \return pointer to start of first potentially composed character before the codepoint beginning at \p from.
+ */
+const char *fl_utf8_previous_composed_char(const char *from, const char *begin) {
+  if (from <= begin || fl_utf8len(*from) == -1) return from - 1;
+  const char *keep = from;
+  from = fl_utf8back(from - 1, begin, NULL);
+  while (from >= begin) {
+    unsigned u = fl_utf8decode(from, keep, NULL);
+    if (u >= 0xFE00 && u <= 0xFE0F) { // a variation selector
+      from = fl_utf8back(from - 1, begin, NULL);
+    } else if (u >= 0x1F3FB && u <= 0x1F3FF) { // EMOJI MODIFIER FITZPATRICK
+      from = fl_utf8back(from - 1, begin, NULL);
+    } else if (from > begin) {
+      keep = fl_utf8back(from - 1, begin, NULL);
+      u = fl_utf8decode(keep, from, NULL);
+      if (u == 0x200D) { // zero-width joiner
+        from = fl_utf8back(keep - 1, begin, NULL);
+        continue;
+      }
+      return from;
+    } else break;
+  }
+  return from;
+}
+
 /** @} */