unicode.cpp

#include <string>
#include <locale>

namespace unicode {
enum class encoding_form {
  utf_8,
  utf_16,
  utf_32,
};

// code unit
// code point
// text element

enum class char8_t : unsigned char {lulz};
enum class codepoint : uint32_t {adena};
enum class scalar_value : uint32_t {adena};
}

using namespace unicode;

int main(int argc, const char **argv) {
  std::string s(argv[1]);
  std::locale l("en-US");
  
  converting_iterator<system, encoding_form::utf_8>(std::begin(s));
  converting_iterator<system, encoding_form::utf_8>(std::end(s));

  // char is assumed to be the system encoding. char8_t is explicitly UTF-8.
  std::u16string s16(make_conversion_range<encoding_form::utf_16>(s));
  std::u8string s8(make_conversion_range<encoding_form::utf_8, abort_handler>(s));
  
  // In place. What argument type...
  to_upper(s8, l);
  
  make_to_upper_range(s8, l);
  
  // No to upper iterator as it's not a 1 to 1 mapping.
  
  // Error. system encoding can only be used for encoding change.
  // auto cp = make_codepoint_range(s);
  auto cp = make_codepoint_range(s8);
  make_to_upper_range(cp, l); // Also takes code points.
}