truncate narrow string converted from unicode string

Now we have one Unicode string, and we want to convert it to be narrow string which means multi byte string and truncate it to desired length.
There is one key point: in Unicode string, two bytes represent one charactor; after it is converted into multi byte string, two or one bytes can represent one charactor. Wether the charactor can be displayed correctly without being '?', firstly the correct code page for non-Unicode program should be set; second we need to make sure the Unicode byte after truncating must be in a pair. Below function is for such purpose use.

void truncateFileName(UTxString8& narrStrFileName, unsigned nDestiNum)
 {
  // No need to truncate file name.
  if(narrStrFileName.length() < nDestiNum)
   return;

  // Truncate file name to be desired length.
  narrStrFileName = narrStrFileName.substr(narrStrFileName.length()-nDestiNum, nDestiNum);

  // Truncate one more byte if finding the number of Unicode char byte is odd, which means
  // half of the original Unicode character has been truncated so we need to remove the other
  // half.
  //
  int nCharNum = narrStrFileName.length();
  int nUnicodeCharNum = 0;
  for (int i = 0; i != nCharNum; ++i)
  {
   char cChar = narrStrFileName.at(i);

   // Half of an Unicode character occupies one byte and it can always be translated to be
   // a negative integer.
   // Once reach to a ASCII char, stop the counting because left Unicode chars are definitely
   // pair by pair.
   if(cChar < 0)
    ++nUnicodeCharNum;
   else
    break;
  }

  if(nUnicodeCharNum%2 == 1) // odd number.
   narrStrFileName = narrStrFileName.substr(1, nCharNum-1);

 }