Converting CStringW to CString resulting question mark - c++

Question

My vc++ project's property "character set" is set to "Use Multi-Byte Character", I am converting CStringW (Chinese language string) into CString and it's converting the Chinese characters into question marks.
CStringA utf8;
int cc=0;
// get length (cc) of the new multibyte string excluding the \0 terminator first
if ((cc = WideCharToMultiByte(CP_UTF8, 0, stringw, -1, NULL, 0, 0, 0) - 1) > 0)
{
// convert
char *buf = utf8.GetBuffer(cc);
if (buf) WideCharToMultiByte(CP_UTF8, 0, stringw, -1, buf, cc, 0, 0);
utf8.ReleaseBuffer();
Please help, I'm new in c++.

Answer 1

The conversion is working correctly, you can send the UTF-8 result to another program which requires UTF-8, or test it in a web browser, or test it in Notepad with UTF-8 encoding.
The only problem you may face is that Windows API cannot directly display UTF-8 string, so it may look as if it failed. For example, ::MessageBoxA(0, utf8, 0, 0) will show the wrong result.
To see if the conversion worked, you can convert back to UTF-16. For example:
#include <atlstr.h>
...
CStringW utf16 = L"汉字 / 漢字";
CStringA utf8 = CW2A(utf16, CP_UTF8);
CStringW copy = CA2W(utf8, CP_UTF8);
if (copy == utf16)
::MessageBoxW(0, copy, L"success", 0);
else
::MessageBoxW(0, copy, L"failed", 0);
New Windows programs should be in Unicode (UTF-16). Note that Multi-byte and ANSI are deprecated.
To use Multi-Byte in an old program, set code page to CP_ACP or another language. For example, WideCharToMultiByte(CP_ACP, ...). However this depends on codepage for the current thread and is not reliable.
#include <iostream>
#include <atlstr.h>
#include <Windows.h>
int main()
{
//ANSI codepage for Greek language:
int codepage = 1253;
CStringW utf16 = L"ελληνική";
CStringA ansi = CW2A(utf16, codepage);
SetConsoleOutputCP(codepage);
//This should appear correctly
std::cout << ansi << "\n";
//This won't show correctly in English language computer
MessageBoxA(0, ansi, 0, 0);
return 0;
}
Lastly, if project is not Unicode, and for some reason you cannot change it to Unicode, you still have access to Unicode API's (for example MessageBoxW) so you can patch the old program. For example, instead of using CTooltipCtrl::AddTool, use the equivalent WinAPI code which exposes Unicode functions:
CWnd *button = GetDlgItem(IDOK);
if (button)
{
TOOLINFOW ti;
memset(&ti, 0, sizeof(ti));
ti.cbSize = sizeof(ti);
ti.hwnd = ::GetParent(button->m_hWnd);
ti.uFlags = TTF_IDISHWND;
ti.uId = (UINT_PTR)button->m_hWnd;
ti.lpszText = L"汉字 / 漢字";
ToolTip->SendMessage(TTM_ADDTOOLW, 0, (LPARAM)&ti);
}