Function utf8ToCodepoint

Utilities/cmjsoncpp/src/lib_json/json_writer.cpp:189–238

}

static unsigned int utf8ToCodepoint(const char*& s, const char* e) {
  const unsigned int REPLACEMENT_CHARACTER = 0xFFFD;

  unsigned int firstByte = static_cast<unsigned char>(*s);

  if (firstByte < 0x80)
    return firstByte;

  if (firstByte < 0xE0) {
    if (e - s < 2)
      return REPLACEMENT_CHARACTER;

    unsigned int calculated =
        ((firstByte & 0x1F) << 6) | (static_cast<unsigned int>(s[1]) & 0x3F);
    s += 1;
    // oversized encoded characters are invalid
    return calculated < 0x80 ? REPLACEMENT_CHARACTER : calculated;
  }

  if (firstByte < 0xF0) {
    if (e - s < 3)
      return REPLACEMENT_CHARACTER;

    unsigned int calculated = ((firstByte & 0x0F) << 12) |
                              ((static_cast<unsigned int>(s[1]) & 0x3F) << 6) |
                              (static_cast<unsigned int>(s[2]) & 0x3F);
    s += 2;
    // surrogates aren't valid codepoints itself
    // shouldn't be UTF-8 encoded
    if (calculated >= 0xD800 && calculated <= 0xDFFF)
      return REPLACEMENT_CHARACTER;
    // oversized encoded characters are invalid
    return calculated < 0x800 ? REPLACEMENT_CHARACTER : calculated;
  }

  if (firstByte < 0xF8) {
    if (e - s < 4)
      return REPLACEMENT_CHARACTER;

    unsigned int calculated = ((firstByte & 0x07) << 18) |
                              ((static_cast<unsigned int>(s[1]) & 0x3F) << 12) |
                              ((static_cast<unsigned int>(s[2]) & 0x3F) << 6) |
                              (static_cast<unsigned int>(s[3]) & 0x3F);
    s += 3;
    // oversized encoded characters are invalid
    return calculated < 0x10000 ? REPLACEMENT_CHARACTER : calculated;
  }

  return REPLACEMENT_CHARACTER;
}

static const char hex2[] = "000102030405060708090a0b0c0d0e0f"
                           "101112131415161718191a1b1c1d1e1f"