Unicode: full Unicode Support (6 squashed commits) (#2541, #2538)

fix build for WideCharToMultiByte
[3181ff1e] Full Unicode Support
[6c9e73ac] Fix ImTextCountUtf8BytesFromChar and ImTextCharToUtf8, these APIs assume the input is an unicode code point, not UTF-16
[ba85665b] Add AddInputCharacterUTF16 for windows backend to handle WM_CHAR
[fafdcaf0] Use Windows API to convert UTF-16 for ImFileOpen
[dc7d5925] Use windows API to convert UTF-16 for clipboard

examples/imgui_impl_win32.cpp

@@ -324,7 +324,8 @@ IMGUI_IMPL_API LRESULT ImGui_ImplWin32_WndProcHandler(HWND hwnd, UINT msg, WPARA
         return 0;
     case WM_CHAR:
         // You can also use ToAscii()+GetKeyboardState() to retrieve characters.
-        io.AddInputCharacter((unsigned int)wParam);
+        if (wParam > 0 && wParam < 0x10000)
+            io.AddInputCharacterUTF16((unsigned short)wParam);
         return 0;
     case WM_SETCURSOR:
         if (LOWORD(lParam) == HTCLIENT && ImGui_ImplWin32_UpdateMouseCursor())

imconfig.h

@@ -77,6 +77,9 @@
 // Read about ImGuiBackendFlags_RendererHasVtxOffset for details.
 //#define ImDrawIdx unsigned int
 
+//---- Use 32-bit for ImWchar (default is 16-bit) to support full unicode code points.
+//#define ImWchar ImWchar32
+
 //---- Override ImDrawCallback signature (will need to modify renderer back-ends accordingly)
 //struct ImDrawList;
 //struct ImDrawCmd;

imgui.cpp

@@ -1098,13 +1098,35 @@ void ImGuiIO::AddInputCharacter(unsigned int c)
         InputQueueCharacters.push_back((ImWchar)c);
 }
 
+// UTF16 string use Surrogate to encode unicode > 0x10000, so we should save the Surrogate.
+void ImGuiIO::AddInputCharacterUTF16(ImWchar16 c)
+{
+    if (c >= 0xD800 && c <= 0xDBFF)
+    {
+        Surrogate = c;
+    }
+    else
+    {
+        ImWchar cp = c;
+        if (c >= 0xDC00 && c <= 0xDFFF)
+        {
+            if (sizeof(ImWchar) == 2)
+                cp = IM_UNICODE_CODEPOINT_INVALID;
+            else
+                cp = ((ImWchar)(Surrogate - 0xD800) << 10) + (c - 0xDC00) + 0x10000;
+            Surrogate = 0;
+        }
+        InputQueueCharacters.push_back(cp);
+    }
+}
+
 void ImGuiIO::AddInputCharactersUTF8(const char* utf8_chars)
 {
     while (*utf8_chars != 0)
     {
         unsigned int c = 0;
         utf8_chars += ImTextCharFromUtf8(&c, utf8_chars, NULL);
-        if (c > 0 && c <= IM_UNICODE_CODEPOINT_MAX)
+        if (c > 0)
             InputQueueCharacters.push_back((ImWchar)c);
     }
 }
@@ -1488,13 +1510,13 @@ ImFileHandle ImFileOpen(const char* filename, const char* mode)
 {
 #if defined(_WIN32) && !defined(IMGUI_DISABLE_WIN32_FUNCTIONS) && !defined(__CYGWIN__) && !defined(__GNUC__)
     // We need a fopen() wrapper because MSVC/Windows fopen doesn't handle UTF-8 filenames.
-    const int filename_wsize = ImTextCountCharsFromUtf8(filename, NULL) + 1;
-    const int mode_wsize = ImTextCountCharsFromUtf8(mode, NULL) + 1;
+    const int filename_wsize = ::MultiByteToWideChar(CP_UTF8, 0, filename, -1, NULL, 0);
+    const int mode_wsize = ::MultiByteToWideChar(CP_UTF8, 0, mode, -1, NULL, 0);
     ImVector<ImWchar> buf;
     buf.resize(filename_wsize + mode_wsize);
-    ImTextStrFromUtf8(&buf[0], filename_wsize, filename, NULL);
-    ImTextStrFromUtf8(&buf[filename_wsize], mode_wsize, mode, NULL);
-    return _wfopen((wchar_t*)&buf[0], (wchar_t*)&buf[filename_wsize]);
+    ::MultiByteToWideChar(CP_UTF8, 0, filename, -1, &buf[0], filename_wsize);
+    ::MultiByteToWideChar(CP_UTF8, 0, mode, -1,&buf[filename_wsize], mode_wsize);
+    return _wfopen(&buf[0], &buf[filename_wsize]);
 #else
     return fopen(filename, mode);
 #endif
@@ -1606,6 +1628,8 @@ int ImTextCharFromUtf8(unsigned int* out_char, const char* in_text, const char*
         c += (*str++ & 0x3f);
         // utf-8 encodings of values used in surrogate pairs are invalid
         if ((c & 0xFFFFF800) == 0xD800) return 4;
+        // If ImWchar is 16bit, use replacement character U+FFFD instead
+        if (sizeof(ImWchar) == 2 && c >= 0x10000) c = IM_UNICODE_CODEPOINT_INVALID;
         *out_char = c;
         return 4;
     }
@@ -1623,7 +1647,6 @@ int ImTextStrFromUtf8(ImWchar* buf, int buf_size, const char* in_text, const cha
         in_text += ImTextCharFromUtf8(&c, in_text, in_text_end);
         if (c == 0)
             break;
-        if (c <= IM_UNICODE_CODEPOINT_MAX)    // FIXME: Losing characters that don't fit in 2 bytes
         *buf_out++ = (ImWchar)c;
     }
     *buf_out = 0;
@@ -1641,7 +1664,6 @@ int ImTextCountCharsFromUtf8(const char* in_text, const char* in_text_end)
         in_text += ImTextCharFromUtf8(&c, in_text, in_text_end);
         if (c == 0)
             break;
-        if (c <= IM_UNICODE_CODEPOINT_MAX)
         char_count++;
     }
     return char_count;
@@ -1662,11 +1684,15 @@ static inline int ImTextCharToUtf8(char* buf, int buf_size, unsigned int c)
         buf[1] = (char)(0x80 + (c & 0x3f));
         return 2;
     }
-    if (c >= 0xdc00 && c < 0xe000)
+    if (c < 0x10000)
     {
-        return 0;
+        if (buf_size < 3) return 0;
+        buf[0] = (char)(0xe0 + (c >> 12));
+        buf[1] = (char)(0x80 + ((c>> 6) & 0x3f));
+        buf[2] = (char)(0x80 + ((c ) & 0x3f));
+        return 3;
     }
-    if (c >= 0xd800 && c < 0xdc00)
+    if (c <= 0x10FFFF)
     {
         if (buf_size < 4) return 0;
         buf[0] = (char)(0xf0 + (c >> 18));
@@ -1675,14 +1701,8 @@ static inline int ImTextCharToUtf8(char* buf, int buf_size, unsigned int c)
         buf[3] = (char)(0x80 + ((c ) & 0x3f));
         return 4;
     }
-    //else if (c < 0x10000)
-    {
-        if (buf_size < 3) return 0;
-        buf[0] = (char)(0xe0 + (c >> 12));
-        buf[1] = (char)(0x80 + ((c>> 6) & 0x3f));
-        buf[2] = (char)(0x80 + ((c ) & 0x3f));
-        return 3;
-    }
+    // Invalid code point, the max unicode is 0x10FFFF
+    return 0;
 }
 
 // Not optimal but we very rarely use this function.
@@ -1696,8 +1716,8 @@ static inline int ImTextCountUtf8BytesFromChar(unsigned int c)
 {
     if (c < 0x80) return 1;
     if (c < 0x800) return 2;
-    if (c >= 0xdc00 && c < 0xe000) return 0;
-    if (c >= 0xd800 && c < 0xdc00) return 4;
+    if (c < 0x10000) return 3;
+    if (c <= 0x10FFFF) return 4;
     return 3;
 }
 
@@ -9748,6 +9768,7 @@ static void WindowSettingsHandler_WriteAll(ImGuiContext* ctx, ImGuiSettingsHandl
 #else
 #include <windows.h>
 #endif
+#include <stringapiset.h>
 #if defined(WINAPI_FAMILY) && (WINAPI_FAMILY == WINAPI_FAMILY_APP) // UWP doesn't have Win32 functions
 #define IMGUI_DISABLE_WIN32_DEFAULT_CLIPBOARD_FUNCTIONS
 #define IMGUI_DISABLE_WIN32_DEFAULT_IME_FUNCTIONS
@@ -9760,6 +9781,7 @@ static void WindowSettingsHandler_WriteAll(ImGuiContext* ctx, ImGuiSettingsHandl
 
 #ifdef _MSC_VER
 #pragma comment(lib, "user32")
+#pragma comment(lib, "kernel32")
 #endif
 
 // Win32 clipboard implementation
@@ -9775,11 +9797,11 @@ static const char* GetClipboardTextFn_DefaultImpl(void*)
         ::CloseClipboard();
         return NULL;
     }
-    if (ImWchar* wbuf_global = (ImWchar*)::GlobalLock(wbuf_handle))
+    if (const WCHAR* wbuf_global = (const WCHAR*)::GlobalLock(wbuf_handle))
     {
-        int buf_len = ImTextCountUtf8BytesFromStr(wbuf_global, NULL) + 1;
+        int buf_len = ::WideCharToMultiByte(CP_UTF8, 0, wbuf_global, -1, NULL, 0, NULL, NULL);
         buf_local.resize(buf_len);
-        ImTextStrToUtf8(buf_local.Data, buf_len, wbuf_global, NULL);
+        ::WideCharToMultiByte(CP_UTF8, 0, wbuf_global, -1, buf_local.Data, buf_len, NULL, NULL);
     }
     ::GlobalUnlock(wbuf_handle);
     ::CloseClipboard();
@@ -9790,15 +9812,15 @@ static void SetClipboardTextFn_DefaultImpl(void*, const char* text)
 {
     if (!::OpenClipboard(NULL))
         return;
-    const int wbuf_length = ImTextCountCharsFromUtf8(text, NULL) + 1;
-    HGLOBAL wbuf_handle = ::GlobalAlloc(GMEM_MOVEABLE, (SIZE_T)wbuf_length * sizeof(ImWchar));
+    const int wbuf_length = ::MultiByteToWideChar(CP_UTF8, 0, text, -1, NULL, 0);
+    HGLOBAL wbuf_handle = ::GlobalAlloc(GMEM_MOVEABLE, (SIZE_T)wbuf_length * sizeof(WCHAR));
     if (wbuf_handle == NULL)
     {
         ::CloseClipboard();
         return;
     }
-    ImWchar* wbuf_global = (ImWchar*)::GlobalLock(wbuf_handle);
-    ImTextStrFromUtf8(wbuf_global, wbuf_length, text, NULL);
+    WCHAR* wbuf_global = (WCHAR*)::GlobalLock(wbuf_handle);
+    ::MultiByteToWideChar(CP_UTF8, 0, text, -1, wbuf_global, wbuf_length);
     ::GlobalUnlock(wbuf_handle);
     ::EmptyClipboard();
     if (::SetClipboardData(CF_UNICODETEXT, wbuf_handle) == NULL)

imgui.h

@@ -143,7 +143,11 @@ struct ImGuiTextFilter;             // Helper to parse and apply text filters (e
 typedef void* ImTextureID;          // User data to identify a texture (this is whatever to you want it to be! read the FAQ about ImTextureID in imgui.cpp)
 #endif
 typedef unsigned int ImGuiID;       // Unique ID used by widgets (typically hashed from a stack of string)
-typedef unsigned short ImWchar;     // A single U16 character for keyboard input/display. We encode them as multi bytes UTF-8 when used in strings.
+#ifndef ImWchar
+#define ImWchar ImWchar16
+#endif
+typedef unsigned short ImWchar16;   // A single U16 character for keyboard input/display. We encode them as multi bytes UTF-8 when used in strings.
+typedef int ImWchar32;              // A single 32bit character for keyboard input/display, define ImWchar to ImWchar32 to use it. See imconfig.h .
 typedef int ImGuiCol;               // -> enum ImGuiCol_             // Enum: A color identifier for styling
 typedef int ImGuiCond;              // -> enum ImGuiCond_            // Enum: A condition for many Set*() functions
 typedef int ImGuiDataType;          // -> enum ImGuiDataType_        // Enum: A primary data type
@@ -1465,6 +1469,7 @@ struct ImGuiIO
 
     // Functions
     IMGUI_API void  AddInputCharacter(unsigned int c);          // Queue new character input
+    IMGUI_API void  AddInputCharacterUTF16(ImWchar16 c);        // Queue new character input from an UTF-16 character, it can be a surrogate
     IMGUI_API void  AddInputCharactersUTF8(const char* str);    // Queue new characters input from an UTF-8 string
     IMGUI_API void  ClearInputCharacters();                     // Clear the text input buffer manually
 
@@ -1507,6 +1512,7 @@ struct ImGuiIO
     float       KeysDownDurationPrev[512];      // Previous duration the key has been down
     float       NavInputsDownDuration[ImGuiNavInput_COUNT];
     float       NavInputsDownDurationPrev[ImGuiNavInput_COUNT];
+    ImWchar16   Surrogate;                      // For AddInputCharacterUTF16
     ImVector<ImWchar> InputQueueCharacters;     // Queue of _characters_ input (obtained by platform back-end). Fill using AddInputCharacter() helper.
 
     IMGUI_API   ImGuiIO();
@@ -2092,7 +2098,11 @@ struct ImFontGlyphRangesBuilder
     ImVector<ImU32> UsedChars;            // Store 1-bit per Unicode code point (0=unused, 1=used)
 
     ImFontGlyphRangesBuilder()          { Clear(); }
-    inline void     Clear()             { int size_in_bytes = (IM_UNICODE_CODEPOINT_MAX+1) / 8; UsedChars.resize(size_in_bytes / (int)sizeof(ImU32)); memset(UsedChars.Data, 0, (size_t)size_in_bytes); }
+    inline void     Clear()             
+    {
+        int MaxUnicode = sizeof(ImWchar) == 2 ? 0x10000 : 0x110000;
+        UsedChars.resize(MaxUnicode / sizeof(int)); memset(UsedChars.Data, 0, MaxUnicode / sizeof(int));
+    }
     inline bool     GetBit(int n) const { int off = (n >> 5); ImU32 mask = 1u << (n & 31); return (UsedChars[off] & mask) != 0; }  // Get bit n in the array
     inline void     SetBit(int n)       { int off = (n >> 5); ImU32 mask = 1u << (n & 31); UsedChars[off] |= mask; }               // Set bit n in the array
     inline void     AddChar(ImWchar c)  { SetBit(c); }                          // Add character

imgui_draw.cpp

@@ -2550,7 +2550,6 @@ void ImFontGlyphRangesBuilder::AddText(const char* text, const char* text_end)
         text += c_len;
         if (c_len == 0)
             break;
-        if (c <= IM_UNICODE_CODEPOINT_MAX)
         AddChar((ImWchar)c);
     }
 }