Convert UTF-16 to UTF-8

Question

I am current using VC++ 2008 MFC. Due to PostgreSQL doesn't support UTF-16 (Encoding used by Windows for Unicode), I need to convert string from UTF-16 to UTF-8, before store it.

Here is my code snippet.

// demo.cpp : Defines the entry point for the console application.
//

#include "stdafx.h"
#include "demo.h"
#include "Utils.h"
#include <iostream>

#ifdef _DEBUG
#define new DEBUG_NEW
#endif


// The one and only application object

CWinApp theApp;

using namespace std;

int _tmain(int argc, TCHAR* argv[], TCHAR* envp[])
{
    int nRetCode = 0;

    // initialize MFC and print and error on failure
    if (!AfxWinInit(::GetModuleHandle(NULL), NULL, ::GetCommandLine(), 0))
    {
        // TODO: change error code to suit your needs
        _tprintf(_T("Fatal Error: MFC initialization failed\n"));
        nRetCode = 1;
    }
    else
    {
        // TODO: code your application's behavior here.
    }

    CString utf16 = _T("Hello");
    std::cout << utf16.GetLength() << std::endl;
    CStringA utf8 = UTF8Util::ConvertUTF16ToUTF8(utf16);
    std::cout << utf8.GetLength() << std::endl;
    getchar();
    return nRetCode;
}

and the conversion functions.

namespace UTF8Util
{
//----------------------------------------------------------------------------
// FUNCTION: ConvertUTF8ToUTF16
// DESC: Converts Unicode UTF-8 text to Unicode UTF-16 (Windows default).
//----------------------------------------------------------------------------
CStringW ConvertUTF8ToUTF16( __in const CHAR * pszTextUTF8 )
{
    //
    // Special case of NULL or empty input string
    //
    if ( (pszTextUTF8 == NULL) || (*pszTextUTF8 == '\0') )
    {
        // Return empty string
        return L"";
    }

    //
    // Consider CHAR's count corresponding to total input string length,
    // including end-of-string (\0) character
    //
    const size_t cchUTF8Max = INT_MAX - 1;
    size_t cchUTF8;
    HRESULT hr = ::StringCchLengthA( pszTextUTF8, cchUTF8Max, &cchUTF8 );

    if ( FAILED( hr ) )
    {
        AtlThrow( hr );
    }

    // Consider also terminating \0
    ++cchUTF8;

    // Convert to 'int' for use with MultiByteToWideChar API
    int cbUTF8 = static_cast<int>( cchUTF8 );

    //
    // Get size of destination UTF-16 buffer, in WCHAR's
    //
    int cchUTF16 = ::MultiByteToWideChar(
        CP_UTF8,                // convert from UTF-8
        MB_ERR_INVALID_CHARS,   // error on invalid chars
        pszTextUTF8,            // source UTF-8 string
        cbUTF8,                 // total length of source UTF-8 string,
                                // in CHAR's (= bytes), including end-of-string \0
        NULL,                   // unused - no conversion done in this step
        0                       // request size of destination buffer, in WCHAR's
        );

    ATLASSERT( cchUTF16 != 0 );

    if ( cchUTF16 == 0 )
    {
        AtlThrowLastWin32();
    }

    //
    // Allocate destination buffer to store UTF-16 string
    //
    CStringW strUTF16;
    WCHAR * pszUTF16 = strUTF16.GetBuffer( cchUTF16 );

    //
    // Do the conversion from UTF-8 to UTF-16
    //

    int result = ::MultiByteToWideChar(
        CP_UTF8,                // convert from UTF-8
        MB_ERR_INVALID_CHARS,   // error on invalid chars
        pszTextUTF8,            // source UTF-8 string
        cbUTF8,                 // total length of source UTF-8 string,
                                // in CHAR's (= bytes), including end-of-string \0
        pszUTF16,               // destination buffer
        cchUTF16                // size of destination buffer, in WCHAR's
        );

    ATLASSERT( result != 0 );

    if ( result == 0 )
    {
        AtlThrowLastWin32();
    }

    // Release internal CString buffer
    strUTF16.ReleaseBuffer();

    // Return resulting UTF16 string
    return strUTF16;
}

//----------------------------------------------------------------------------
// FUNCTION: ConvertUTF16ToUTF8
// DESC: Converts Unicode UTF-16 (Windows default) text to Unicode UTF-8.
//----------------------------------------------------------------------------
CStringA ConvertUTF16ToUTF8( __in const WCHAR * pszTextUTF16 )
{
    //
    // Special case of NULL or empty input string
    //
    if ( (pszTextUTF16 == NULL) || (*pszTextUTF16 == L'\0') )
    {
        // Return empty string
        return "";
    }

    //
    // Consider WCHAR's count corresponding to total input string length,
    // including end-of-string (L'\0') character.
    //
    const size_t cchUTF16Max = INT_MAX - 1;
    size_t cchUTF16;
    HRESULT hr = ::StringCchLengthW( pszTextUTF16, cchUTF16Max, &cchUTF16 );

    if ( FAILED( hr ) )
    {
        AtlThrow( hr );
    }

    // Consider also terminating \0
    ++cchUTF16;

    //
    // WC_ERR_INVALID_CHARS flag is set to fail if invalid input character
    // is encountered.
    // This flag is supported on Windows Vista and later.
    // Don't use it on Windows XP and previous.
    //

#if (WINVER >= 0x0600)
    DWORD dwConversionFlags = WC_ERR_INVALID_CHARS;
#else
    DWORD dwConversionFlags = 0;
#endif

    //
    // Get size of destination UTF-8 buffer, in CHAR's (= bytes)
    //
    int cbUTF8 = ::WideCharToMultiByte(
        CP_UTF8,                // convert to UTF-8
        dwConversionFlags,      // specify conversion behavior
        pszTextUTF16,           // source UTF-16 string
        static_cast<int>( cchUTF16 ),   // total source string length, in WCHAR's,
                                        // including end-of-string \0
        NULL,                   // unused - no conversion required in this step
        0,                      // request buffer size
        NULL, NULL              // unused
        );

    ATLASSERT( cbUTF8 != 0 );

    if ( cbUTF8 == 0 )
    {
        AtlThrowLastWin32();
    }

    //
    // Allocate destination buffer for UTF-8 string
    //
    CStringA strUTF8;
    int cchUTF8 = cbUTF8; // sizeof(CHAR) = 1 byte
    CHAR * pszUTF8 = strUTF8.GetBuffer( cchUTF8 );

    //
    // Do the conversion from UTF-16 to UTF-8
    //
    int result = ::WideCharToMultiByte(
        CP_UTF8,                // convert to UTF-8
        dwConversionFlags,      // specify conversion behavior
        pszTextUTF16,           // source UTF-16 string
        static_cast<int>( cchUTF16 ),   // total source string length, in WCHAR's,
                                        // including end-of-string \0
        pszUTF8,                // destination buffer
        cbUTF8,                 // destination buffer size, in bytes
        NULL, NULL              // unused
        ); 

    ATLASSERT( result != 0 );

    if ( result == 0 )
    {
        AtlThrowLastWin32();
    }

    // Release internal CString buffer
    strUTF8.ReleaseBuffer();

    // Return resulting UTF-8 string
    return strUTF8;
}

} // namespace UTF8Util

However, during runtime, I get the exception at

ATLASSERT( cbUTF8 != 0 );

while trying to get size of destination UTF-8 buffer

1. What thing I had missed out?
2. If I am testing using a Chinese characters, How can I verify the resultant UTF-8 string is correct?

Answer 1

The problem is you specified the WC_ERR_INVALID_CHARS flag:

Windows Vista and later: Fail if an invalid input character is encountered. If this flag is not set, the function silently drops illegal code points. A call to GetLastError returns ERROR_NO_UNICODE_TRANSLATION. Note that this flag only applies when CodePage is specified as CP_UTF8 or 54936 (for Windows Vista and later). It cannot be used with other code page values.

Your conversion function seems quite long. How does this one work for you?

//----------------------------------------------------------------------------
// FUNCTION: ConvertUTF16ToUTF8
// DESC: Converts Unicode UTF-16 (Windows default) text to Unicode UTF-8.
//----------------------------------------------------------------------------
CStringA ConvertUTF16ToUTF8( __in LPCWSTR pszTextUTF16 ) {
    if (pszTextUTF16 == NULL) return "";

    int utf16len = wcslen(pszTextUTF16);
    int utf8len = WideCharToMultiByte(CP_UTF8, 0, pszTextUTF16, utf16len, 
        NULL, 0, NULL, NULL );

    CArray<CHAR> buffer;
    buffer.SetSize(utf8len+1);
    buffer.SetAt(utf8len, '\0');

    WideCharToMultiByte(CP_UTF8, 0, pszTextUTF16, utf16len, 
        buffer.GetData(), utf8len, 0, 0 );

    return buffer.GetData();
}

I see you use a function called StringCchLengthW to get the required length of the output buffer. Most of the places I look recommend using the WideCharToMultiByte function itself to tell you how many CHARs it wants.

Edit:
As Rob pointed out, you can use CW2A with the CP_UTF8 code page:

CStringA str = CW2A(wStr, CP_UTF8);

While I'm editing, I can answer your second question:

How can I verify the resultant UTF-8 string is correct?

Write it to a text file, then open it in Mozilla Firefox or an equivillant program. In the View menu, you can go to Character Encoding and switch manually to UTF-8 (assuming Firefox didn't guess it correctly to begin with). Compare it with a UTF-16 document with the same text and see if there are any differences.

Answer 2

You can also use the ATL text conversion macros - to convert from UTF-16 to UTF-8 use CW2A and pass CP_UTF8 as the code page, e.g.:

CW2A utf8(buffer, CP_UTF8);
const char* data = utf8.m_psz;