Proper UTF-8 <--> current system codepage conversion

Author: Xottab-DUTY

src/xrCore/Text/MbHelpers.cpp

@@ -1,5 +1,8 @@
 #include "stdafx.h"
 #include "MbHelpers.h"
+#include <codecvt>
+
+// XXX: use c++11 functions and kill that DUMP_CONVERSION!!!11
 
 #define BITS1_MASK 0x80 // 10000000b
 #define BITS2_MASK 0xC0 // 11000000b
@@ -127,3 +130,21 @@ u16 mbhMulti2Wide(wchar_t* WideStr, wchar_t* WidePos, u16 WideStrSize, const cha
         WideStr[0] = dpos;
     return dpos;
 }
+
+xr_string StringFromUTF8(const char* in, const std::locale& locale /*= std::locale("")*/)
+{
+    using wcvt = std::wstring_convert<std::codecvt_utf8<wchar_t>, wchar_t>;
+    auto wstr = wcvt{}.from_bytes(in);
+    xr_string result(wstr.size(), '\0');
+    std::use_facet<std::ctype<wchar_t>>(locale).narrow(wstr.data(), wstr.data() + wstr.size(), '?', &result[0]);
+    return result;
+}
+
+xr_string StringToUTF8(const char* in, const std::locale& locale /*= std::locale("")*/)
+{
+    using wcvt = std::wstring_convert<std::codecvt_utf8<wchar_t>, wchar_t>;
+    std::wstring wstr(xr_strlen(in), L'\0');
+    std::use_facet<std::ctype<wchar_t>>(locale).widen(in, in + xr_strlen(in), &wstr[0]);
+    std::string result = wcvt{}.to_bytes(wstr.data(), wstr.data() + wstr.size());
+    return result.data();
+}

src/xrCore/Text/MbHelpers.h

@@ -1,6 +1,9 @@
 #pragma once
 #include "xrCore/xrCore.h"
 
+// XXX: rename this file to StringConversion.hpp
+// XXX: use c++11 functions
+
 #define MAX_MB_CHARS 4096
 
 XRCORE_API u16 mbhMulti2Wide(wchar_t* WideStr, wchar_t* WidePos, u16 WideStrSize, const char* MultiStr);
@@ -27,3 +30,6 @@ IC bool IsAlphaCharacter(wchar_t wc)
         ((wc >= 0x0061) && (wc <= 0x007A)) || ((wc >= 0xFF10) && (wc <= 0xFF19)) ||
         ((wc >= 0xFF21) && (wc <= 0xFF3A)) || ((wc >= 0xFF41) && (wc <= 0xFF5A)));
 }
+
+XRCORE_API xr_string StringFromUTF8(const char* string, const std::locale& locale = std::locale(""));
+XRCORE_API xr_string StringToUTF8(const char* string, const std::locale& locale = std::locale(""));

src/xrEngine/edit_actions.cpp

@@ -9,7 +9,7 @@
 #include "edit_actions.h"
 #include "line_edit_control.h"
 #include "xr_input.h"
-#include <codecvt>
+#include "xrCore/Text/MbHelpers.h"
 
 namespace text_editor
 {
@@ -55,17 +55,6 @@ void type_pair::init(int dik, char c, char c_shift, bool b_translate)
     m_char_shift = c_shift;
 }
 
-xr_string utf8_to_string(const char* utf8str, const std::locale& loc)
-{
-    // UTF-8 to wstring
-    std::wstring_convert<std::codecvt_utf8<wchar_t>> wconv;
-    std::wstring wstr = wconv.from_bytes(utf8str);
-    // wstring to string
-    std::vector<char> buf(wstr.size());
-    std::use_facet<std::ctype<wchar_t>>(loc).narrow(wstr.data(), wstr.data() + wstr.size(), '?', buf.data());
-    return xr_string(buf.data(), buf.size());
-}
-
 void type_pair::on_key_press(line_edit_control* const control)
 {
     char c = 0;
@@ -82,7 +71,7 @@ void type_pair::on_key_press(line_edit_control* const control)
             case SDL_TEXTINPUT:
             {
                 const std::locale locale("");
-                auto str = utf8_to_string(event.text.text, locale);
+                auto str = StringFromUTF8(event.text.text, locale);
 
                 if (std::isalpha(str[0], locale) || str[0] == char(-1)) // "я" = -1
                 {

src/xrEngine/xr_input.cpp

@@ -6,6 +6,7 @@
 #include "Include/editor/ide.hpp"
 #include "GameFont.h"
 #include "PerformanceAlert.hpp"
+#include "xrCore/Text/MbHelpers.h"
 
 #ifndef _EDITOR
 #include "xr_input_xinput.h"
@@ -196,14 +197,14 @@ pcstr KeyToMouseButtonName(const int dik)
 
 bool CInput::get_dik_name(int dik, LPSTR dest_str, int dest_sz)
 {
-    pcstr keyname;
-    
+    xr_string keyname;
+
     if (dik < SDL_NUM_SCANCODES)
-        keyname = SDL_GetKeyName(SDL_GetKeyFromScancode((SDL_Scancode)dik));
+        keyname = StringFromUTF8(SDL_GetKeyName(SDL_GetKeyFromScancode((SDL_Scancode)dik)));
     else
         keyname = KeyToMouseButtonName(dik);
 
-    if (0 == strlen(keyname))
+    if (keyname.empty())
     {
         if (dik == SDL_SCANCODE_UNKNOWN)
             keyname = "Unknown";
@@ -214,7 +215,7 @@ bool CInput::get_dik_name(int dik, LPSTR dest_str, int dest_sz)
         }
     }
 
-    xr_strcpy(dest_str, dest_sz, keyname);
+    xr_strcpy(dest_str, dest_sz, keyname.c_str());
     return true;
 }