Re: unicode

On Tue, 26 Jun 2007 08:04:44 -0700, jraul <jraulinth@yahoo.com> wrote:


Hi,

I do prefer writing Unicode data to files using Unicode UTF-8 encoding
(for several reasons, some of them I wrote in previous posts, like
e.g. not having endiannes problems, etc.)

Your code did not work in my test, too.

But, using the UTF-8 approach, everything is fine, and I can open the
document from both Windows NotePad and Word (note that no BOM mark is
required for UTF-8) and see the three funny symbols :) : the pirate
skull symbol, the Communist symbol, and the yin-yang symbol, is this
correct?

Here's my code:

<CODE>

  // Unicode string (UTF-16)
  std::wstring s = L"\u2620\u262D\u262F\n";

  // Convert from UTF-16 to UTF-8
  CW2U utf8String( s.c_str() );

  // Store UTF-8 string in std::string
  std::string outputString( static_cast<const char *>( utf8String ) );

  // Write Unicode UTF-8 string to file
  std::ofstream fout_utf8("c:\\data_utf8.txt");
  if ( fout_utf8 )
  {
      fout_utf8 << outputString << std::endl;
      fout_utf8.close();
  }

</CODE>

The 'CW2U' is an helper template class I developed (like the ATL
string conversion helpers), here it is its source code:

<CODE>
//----------------------------------------------------------------------------
// Class: CW2UEX
// Descr: Convert from Unicode UTF-16 (WideChars) to Unicode UTF-8
//----------------------------------------------------------------------------
template< int t_nBufferLength = 128 >
class CW2UEX
{
public:
    CW2UEX( LPCWSTR psz ) throw(...) :
        m_psz( m_szBuffer )
    {
        Init( psz );
    }

    ~CW2UEX() throw()
    {
        if( m_psz != m_szBuffer )
        {
            free( m_psz );
        }
    }

    operator LPSTR() const throw()
    {
        return( m_psz );
    }

private:
    void Init( LPCWSTR psz ) throw(...)
    {
        if (psz == NULL)
        {
            m_psz = NULL;
            return;
        }
        int nLengthW = lstrlenW( psz )+1;

        // One Unicode UTF-16 character could be converted
        // up to 4 UTF-8 characters
        int nLengthUtf8 = nLengthW * 4;

        if( nLengthUtf8 > t_nBufferLength )
        {
            m_psz = static_cast< LPSTR >( malloc( nLengthUtf8*
                                          sizeof( char ) ) );
            if (m_psz == NULL)
            {
                AtlThrow( E_OUTOFMEMORY );
            }
        }

        if (::WideCharToMultiByte( CP_UTF8, 0, psz, nLengthW,
                m_psz, nLengthUtf8, NULL, NULL ) == 0)
        {
            AtlThrowLastWin32();
        }
    }

public:
    LPSTR m_psz;
    char m_szBuffer[t_nBufferLength];

private:
    CW2UEX( const CW2UEX& ) throw();
    CW2UEX& operator=( const CW2UEX& ) throw();
};

typedef CW2UEX<> CW2U;

</CODE>

MrAsm