Emoji input: remove context-dependent unicode points from output of emoji palette.

The character palette allowing to input emojis in text generates in some cases a series
of unicode points to represent a single emoji. These series contain various kinds of
unicode points with context-dependent meaning. This commit prevents such context-
dependent unicodepoints from being inserted in FLTK text because FLTK text edition
mechanism is not ready to handle properly context dependency in edited UTF-8 text.

2 files changed, 28 insertions, 5 deletions

diff --git a/src/Fl_cocoa.mm b/src/Fl_cocoa.mm
index 51eeced51..eeeedf425 100644
--- a/src/Fl_cocoa.mm
+++ b/src/Fl_cocoa.mm
@@ -2954,8 +2954,23 @@ static FLTextInputContext* fltextinputcontext_instance = nil;
   BOOL has_text_key = Fl::e_keysym <= '~' || Fl::e_keysym == FL_Iso_Key ||
                       (Fl::e_keysym >= FL_KP && Fl::e_keysym <= FL_KP_Last && Fl::e_keysym != FL_KP_Enter);
   // insertText sent during handleEvent of a key without text cannot be processed in a single FL_KEYBOARD event.
-  // Occurs with deadkey followed by non-text key
+  // Occurs with deadkey followed by non-text key. Occurs also with emoji palette.
   if (!in_key_event || !has_text_key) {
+    if (fl_utf_nb_char((const uchar*)Fl::e_text, Fl::e_length) > 1) {
+      // Some emojis are expressed by a series of Unicode points
+      const char *p = Fl::e_text, *end = Fl::e_text + Fl::e_length;
+      int len;
+      while (p < end) { // loop over all unicode points of the series
+        unsigned u = fl_utf8decode(p, end, &len); // extract one such unicode point
+        if ((u >= 0xFE00 && u <= 0xFE0F)      // variation selectors
+            || u == 0x200D   // zero-width joiner
+            || (u >= 0x1F3FB && u <= 0x1F3FF) // EMOJI MODIFIERS FITZPATRICK TYPE
+            ) { // remove context-dependent unicode points
+          memmove((void*)p, p + len, (end - (p+len)) + 1);
+          Fl::e_length -= len; end -= len;
+        } else p += len; // keep other unicode points
+      }
+    }
     Fl::handle(FL_KEYBOARD, target);
     Fl::e_length = 0;
     }
diff --git a/src/Fl_win32.cxx b/src/Fl_win32.cxx
index c3770024d..da7f70e4d 100644
--- a/src/Fl_win32.cxx
+++ b/src/Fl_win32.cxx
@@ -1562,8 +1562,8 @@ static LRESULT CALLBACK WndProc(HWND hWnd, UINT uMsg, WPARAM wParam, LPARAM lPar
         if (uMsg == WM_CHAR || uMsg == WM_SYSCHAR) {
           wchar_t u = (wchar_t)wParam;
           // Windows emoji palette triggered with Windows + dot sends 2 or more WM_CHAR messages:
-          // the 2 components of a surrogate pair, or variation selectors, or zero-width stuff,
-          // or extra Unicode points.
+          // the 2 components of a surrogate pair, or variation selectors, or zero-width joiner,
+          // or emoji modifiers FITZPATRICK or extra Unicode points.
           if (u >= 0xD800 && u <= 0xDFFF) { // handle the 2 components of a surrogate pair
             static wchar_t surrogate_pair[2];
             if (IS_HIGH_SURROGATE(u)) {
@@ -1572,10 +1572,18 @@ static LRESULT CALLBACK WndProc(HWND hWnd, UINT uMsg, WPARAM wParam, LPARAM lPar
               return 0; // and wait for next WM_CHAR message that will give the 2nd member
             } else {
               surrogate_pair[1] = u; // memorize the 2nd member of the pair
+              unsigned u32 = // convert surrogate pair to UTF-32
+                0x10000 + ((surrogate_pair[0] & 0x3ff) << 10) + (surrogate_pair[1] & 0x3ff);
+              if (u32 >= 0x1F3FB && u32 <= 0x1F3FF) { // emoji modifiers FITZPATRICK
+                Fl::e_length = 0; // skip them
+                return 0;
+              }
               Fl::e_length = fl_utf8fromwc(buffer, 1024, surrogate_pair, 2); // transform to UTF-8
             }
-          } else if ((u >= 0xFE00 && u <= 0xFE0F) || (u >= 0x200B && u <= 0x200D)) {
-            Fl::e_length = 0; // skip variation selectors and zero-width Unicode points
+          } else if ( (u >= 0xFE00 && u <= 0xFE0F) // variation selectors
+                      || u == 0x200D // zero-width joiner
+                    ) {
+            Fl::e_length = 0; // skip these context-dependent Unicode points
             return 0;
           } else {
             Fl::e_length = fl_utf8fromwc(buffer, 1024, &u, 1); // process regular Unicode point