int Q_UTF8ToUChar32( const char *pUTF8_, unsigned int &uValueOut, bool &bErrorOut )

{

  const unsigned char *pUTF8 = (const unsigned char*)pUTF8_;

  int nBytes = 1;

  unsigned int uValue = pUTF8[0];

  unsigned int uMinValue = 0;

  // 0....... single byte

  if( uValue < 0x80 )

    goto decodeFinishedNoCheck;

  // Expecting at least a two-byte sequence with 0xC0 <= first <= 0xF7 (110...... and 11110...)

  if( ( uValue - 0xC0u ) > 0x37u || ( pUTF8[1] & 0xC0 ) != 0x80 )

    goto decodeError;

  uValue = ( uValue << 6 ) - ( 0xC0 << 6 ) + pUTF8[1] - 0x80;

  nBytes = 2;

  uMinValue = 0x80;

  // 110..... two-byte lead byte

  if( !( uValue & ( 0x20 << 6 ) ) )

    goto decodeFinished;

  // Expecting at least a three-byte sequence

  if( ( pUTF8[2] & 0xC0 ) != 0x80 )

    goto decodeError;

  uValue = ( uValue << 6 ) - ( 0x20 << 12 ) + pUTF8[2] - 0x80;

  nBytes = 3;

  uMinValue = 0x800;

  // 1110.... three-byte lead byte

decodeFinished:

  if( uValue >= uMinValue && Q_IsValidUChar32( uValue ) )

  {

  decodeFinishedNoCheck:

    uValueOut = uValue;

    bErrorOut = false;

    return nBytes;

  }

decodeError:

  uValueOut = '?';

  bErrorOut = true;

  return nBytes;

decodeFinishedMaybeCESU8:

  // Do we have a full UTF-16 surrogate pair that's been UTF-8 encoded afterwards?

  // That is, do we have 0xD800-0xDBFF followed by 0xDC00-0xDFFF? If so, decode it all.

  if( ( uValue - 0xD800u ) < 0x400u && pUTF8[3] == 0xED && (unsigned char)( pUTF8[4] - 0xB0 ) < 0x10 && ( pUTF8[5] & 0xC0 ) == 0x80 )

  {

    uValue = 0x10000 + ( ( uValue - 0xD800u ) << 10 ) + ( (unsigned char)( pUTF8[4] - 0xB0 ) << 6 ) + pUTF8[5] - 0x80;

    nBytes = 6;

    uMinValue = 0x10000;

  }

  goto decodeFinished;

}