Representation of Wide-Character Literals

Wide-character literals can be represented in various ways. These representations and how they are handled are described below:

Trigraphs

Trigraph characters used as literals are converted to their corresponding Unicode characters. For example:

wchar_t *wcs = L" ??(";

is equivalent to:

wchar_t wcs[] = {0x0020, 0x005B, 0x0000};

For more information about trigraphs, see the ILE C/C++ Language Reference.

Character Escape Codes (\a, \b, \f, \n, \r, \t, \v, \', \", \?)

Character escape codes are converted to their corresponding Unicode escape codes. For example:

wchar_t *wcs = L" \t \n";

is equivalent to:

wchar_t wcs[] = {0x0020, 0x0009, 0x0020, 0x000A, 0x0000};

Numeric Escape Codes (\xnnnn, \ooo)

Numeric escape codes are not converted to Unicode. Instead, the numeric portion of the literal is preserved and assumed to be a Unicode character represented in hexadecimal or octal form. For example:

wchar_t *wcs = L" \x4145";

is equivalent to:

wchar_t wcs[] = {0x0020, 0x4145, 0x0000};

Specifying \xnn in a wchar_t string literal is equivalent to specifying \x00nn.

Hexadecimal constant values larger than 0xFF are normally considered invalid. Setting the *LOCALEUCS2 option changes this to allow 2–byte hexadecimal initialization of wchar_t types only. For example:

wchar_t wc = L'\x4145';  /* Valid only with *LOCALEUCS2 option, */
                         /* otherwise an out of bounds error    */
                         /* will result.                        */

char     c =  '\x4145';  /* Not valid due to size restriction.  */
                         /* Error will result with or without   */
                         /* specifying *LOCALEUCS2 option.      */

Note: Numeric hexadecimal escape codes are not validated other than to ensure type size-limits are not exceeded.

DBCS Characters

DBCS characters entered as hexadecimal escape sequences are not converted to Unicode. They are stored as received.