utf-8 の encode / decode

decode

int decode() {
  const unsigned char mask[4] = {0x7f, 0x1f, 0x0f, 0x07};
  unsigned char c = 0;
  scanf("%c", &c);
  int type = 0;
  for (; type < 4 && c >= 0xff - mask[type]; type++) ;
  int result = (c & mask[type]) << type * 6;
  for (int i = 1; i <= type; i++) {
    scanf("%c", &c);
    result += (c & 0x3f) << (6 * (type - i));
  }
  return result;
}

encode

void encode(int x) {
  unsigned char str[6] = {};
  if (x < 0x80) str[0] = x;
  else {
    int n = 3;
    char mask1 = 0xfd, mask2 = 0x01;
    if (x < 0x800) {
      n = 1; mask1 = 0xd0; mask2 = 0x1f;
    } else if (x < 0x10000) {
      n = 2; mask1 = 0xe0; mask2 = 0x0f;
    } 
    str[0] = mask1 | ((x >> (6 * n)) & mask2);
    for (int i = 1; i <= n; i++) 
      str[i] = 0x80 | ((x >> 6 * (n - i)) & 0x3f);    
  }
  printf("%s", str);
}

意外とwikipedia( https://ja.wikipedia.org/wiki/UTF-8 ) がわかりやすくて便利だった