std::codecvt_mode (3) - Linux Manuals

std::codecvt_mode: std::codecvt_mode

NAME

std::codecvt_mode - std::codecvt_mode

Synopsis

Defined in header <codecvt>
enum codecvt_mode {
consume_header = 4, (since C++11)
generate_header = 2, (deprecated in C++17)
little_endian = 1
};

The facets std::codecvt_utf8, std::codecvt_utf16, and std::codecvt_utf8_utf16 accept an optional value of type std::codecvt_mode as a template argument, which specifies optional features of the unicode string conversion.

Constants

Defined in header <locale>
Value Meaning
little_endian assume the input is in little-endian byte order (applies to UTF-16 input only, the default is big-endian)
consume_header consume the byte order mark, if present at the start of input sequence, and (in case of UTF-16), rely on the byte order it specifies for decoding the rest of the input
generate_header output the byte order mark at the start of the output sequence

The recognized byte order marks are:

0xfe 0xff UTF-16 big-endian
0xff 0xfe UTF-16 little-endian
0xef 0xbb 0xbf UTF-8 (no effect on endianness)

If std::consume_header is not selected when reading a file beginning with byte order mark, the Unicode character U+FEFF (Zero width non-breaking space) will be read as the first character of the string content.

Example

The following example demonstrates consuming the UTF-8 BOM
// Run this code

  #include <fstream>
  #include <iostream>
  #include <string>
  #include <locale>
  #include <codecvt>

  int main()
  {
      // UTF-8 data with BOM
      std::ofstream("text.txt") << u8"\ufeffz\u6c34\U0001d10b";
      // read the UTF8 file, skipping the BOM
      std::wifstream fin("text.txt");
      fin.imbue(std::locale(fin.getloc(),
                            new std::codecvt_utf8<wchar_t, 0x10ffff, std::consume_header>));
      for (wchar_t c; fin.get(c); )
          std::cout << std::hex << std::showbase << c << '\n';
  }

Output:

  0x7a
  0x6c34
  0x1d10b