English 中文

Why the conversion between C++ string/wstring/u8string/u16string/u32string so painful?

You can convert strings through the following API:

• ext::convert<std::string>
• ext::convert<std::u8string>
• ext::convert<std::u16string>
• ext::convert<std::u32string>
• ext::convert<std::wstring>

#include <iostream>
#include <ext/convert.hpp>

using std::string_view_literals::operator""sv;

int main() {
  std::cout << ext::convert<std::string>(U"Hello world!"sv) << std::endl;
  return 0;
}

You can customize namespace and function name:

#include <iostream>
#define PAINFUL_CPP_STRING_CONVERSION_NAMESPACE utf
#define PAINFUL_CPP_STRING_CONVERSION_FUNCTION  as
#include <ext/convert.hpp>

using std::string_view_literals::operator""sv;

int main() {
  std::cout << utf::as<std::string>(U"Hello world!"sv) << std::endl;
  return 0;
}

• ✅ OK.
• ⚠️ Assuming that std::string always stores text in UTF-8 encoding.
• std::wstring can be considered as std::u16string on Windows.

• Supports span or span-like types as parameter
• Supports simdutf as a high-speed conversion implementation