P2314R0: Character sets and encodings

Introduction

• Replace the term "execution character set" with "literal encoding". For purposes of this specification, the actual set of characters is not relevant, but the sequence of code points (i.e. the encoding) specified by a given character or string literal are.
• Switch C++ to a modified "model B" approach for universal-character-names as described in the C99 Rationale v5.10, section 5.2.1.
• (Not a wording change) Do not attempt to treat all string literals the same; their treatment depends on (phase 7) context.

Terminology changes

• translation character set: the abstract character set used during translation; can represent the character equivalent of all valid universal-character-names
• basic character set: minimum character set needed to express C++ program source
• basic literal character set: minimum set of characters expressible by literals
• ordinary / wide literal encoding: compile-time encoding used for initializing string literal objects

Behavior changes

This changes the specified behavior of the stringizing preprocessor operator [cpp.stringize] as follows

C++20	this paper
#define S(x) # x const char * s1 = S(Köppe); // "K\\u00f6ppe" const char * s2 = S(K\u00f6ppe); // "K\\u00f6ppe"	#define S(x) # x const char * s1 = S(Köppe); // "Köppe" const char * s2 = S(K\u00f6ppe); // "Köppe"

However, it turns out that all major implementations already implement what this paper specifies, i.e. no implementation provides an escaped UCN.

Not all string-literals are the same

Context	Destination
asm-declaration	build environment
#include "fn" or #include <fn>	file name
language linkage	translation
operator "" [over.literal]	translation
#line directive	diagnostic
argument for [[nodiscard]] and [[deprecated]]	diagnostic
#error, static_assert	diagnostic
__FILE__, __func__	literal encoding
std::typeinfo::name()	literal encoding
character-literal or string-literal appearing elsewhere	literal encoding
user-defined-literal	literal encoding

The destinations have the following meaning:

• build environment: A string likely passed as text (program input) to another component in the build environment.
• file name: A file name suitable for the build environment.
• translation: No use outside the compiler.
• diagnostic: Diagnostic text; kept in the translation character set until needed for output; then treated the same as e.g. identifiers appearing in diagnostic messages
• literal encoding: The implementation-defined encoding for the runtime environment, used for string-literals that appear in usual program text.

Wording changes

multibyte character

sequence of one or more bytes representing a member of the extended character set of either the source or the execution environment the code unit sequence for an encoded character

[Note 1 to entry: The extended character set is a superset of the basic character set (5.3). — end note]

1. Physical source file characters are mapped, in an implementation-defined manner, to the basic source translation character set (introducing new-line characters for end-of-line indicators) if necessary. The set of physical source file characters accepted is implementation-defined. Any source file character not in the basic source character set (5.3 [lex.charset]) is replaced by the universal-character-name that designates that character. An implementation may use any internal encoding, so long as an actual extended character encountered in the source file, and the same extended character expressed in the source file as a universal-character-name (e.g., using the \uXXXX notation), are handled equivalently except where this replacement is reverted (5.4 [lex.pptoken]) in a raw string literal.

...

3. The source file is decomposed into preprocessing tokens (5.4 [lex.pptoken]) and sequences of white-space characters (including comments). A source file shall not end in a partial preprocessing token or in a partial comment. [ Footnote: ... ] Each comment is replaced by one space character. New-line characters are retained. Whether each nonempty sequence of white-space characters other than new-line is retained or replaced by one space character is unspecified. Each universal-character-name outside of a character or string literal is replaced by the designated element of the translation character set ([lex.charset]). The process of dividing a source file’s characters into preprocessing tokens is context-dependent. [Example: See the handling of < within a #include preprocessing directive. — end example]

4. Preprocessing directives are executed, macro invocations are expanded, and _Pragma unary operator expressions are executed. If a character sequence that matches the syntax of a universal-character-name is produced by token concatenation (15.6.3 [lex.concat]), the behavior is undefined. A #include preprocessing directive causes the named header or source file to be processed from phase 1 through phase 4, recursively. All preprocessing directives are then deleted.

5. For a sequence of two or more adjacent string-literal tokens, a common encoding-prefix is determined as specified in 5.13.5 [lex.string]. Each such string-literal token is then considered to have that common encoding-prefix. Each basic source character set member in a character-literal or a string-literal, as well as each escape sequence and universal-character-name in a character-literal or a non-raw string literal, is converted to the corresponding member of the execution character set (5.13.3, 5.13.5); if there is no corresponding member, it is converted to an implementation-defined member other than the null (wide) character. [ Footnote: An implementation need not convert all non-corresponding source characters to the same execution character. ]

6. Adjacent string literal string-literal tokens are concatenated (5.13.5 [lex.string]).

1 The translation character set is a set of abstract characters consisting of the following elements:

• each Unicode character as identified by its unique UCS scalar value and
• a distinct abstract character for each UCS scalar value where no Unicode character is assigned.

[ Note: ISO/IEC 10646 code points are integers in the range [0, 10FFFF] (hexadecimal). A surrogate code point is a value in the range [D800, DFFF] (hexadecimal). A UCS scalar value is any code point that is not a surrogate code point. -- end note ]

2 The basic character set is a subset of the translation character set, consisting of 96 characters as specified in table X. [ Note: Unicode short names are given only as a means to identifying the character; the numerical value has no other meaning in this context. -- end note ]

U+0009	CHARACTER TABULATION
U+000B	LINE TABULATION
U+000C	FORM FEED (FF)
U+0020	SPACE
U+000A	LINE FEED (LF)	new-line
U+0021	EXCLAMATION MARK	!
U+0022	QUOTATION MARK	"
U+0023	NUMBER SIGN	#
U+0025	PERCENT SIGN	%
U+0026	AMPERSAND	&
U+0027	APOSTROPHE	'
U+0028	LEFT PARENTHESIS	(
U+0029	RIGHT PARENTHESIS	)
U+002A	ASTERISK	*
U+002B	PLUS SIGN	+
U+002C	COMMA	,
U+002D	HYPHEN-MINUS	-
U+002E	FULL STOP	.
U+002F	SOLIDUS	/
U+0030 .. U+0039	DIGIT ZERO .. NINE	0 1 2 3 4 5 6 7 8 9
U+003A	COLON	:
U+003B	SEMICOLON	;
U+003C	LESS-THAN SIGN	<
U+003D	EQUALS SIGN	=
U+003E	GREATER-THAN SIGN	>
U+003F	QUESTION MARK	?
U+0041 .. U+005A	LATIN CAPITAL LETTER A .. Z	A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
U+005B	LEFT SQUARE BRACKET	[
U+005C	REVERSE SOLIDUS	\
U+005D	RIGHT SQUARE BRACKET	]
U+005E	CIRCUMFLEX ACCENT	^
U+005F	LOW LINE	_
U+0061 .. U+007A	LATIN SMALL LETTER A .. Z	a b c d e f g h i j k l m n o p q r s t u v w x y z
U+007B	LEFT CURLY BRACKET	{
U+007C	VERTICAL LINE	|
U+007D	RIGHT CURLY BRACKET	}
U+007E	TILDE	~

The universal-character-name construct provides a way to name other characters.

hex-quad :
    hexadecimal-digit hexadecimal-digit hexadecimal-digit hexadecimal-digit
  
universal-character-name :
    \u hex-quad
    \U hex-quad hex-quad

A universal-character-name designates the abstract character in ISO/IEC 10646 (if any) the translation character set whose Unicode code point UCS scalar value is the hexadecimal number represented by the sequence of hexadecimal-digits in the universal-character-name. The program is ill-formed if that number is not a Unicode code point or if it is a surrogate code point UCS scalar value. Noncharacter code points and reserved code points are considered to designate separate characters distinct from any ISO/IEC 10646 character. If a universal-character-name outside the c-char-sequence, s-char-sequence, or r-char-sequence of a character-literal or string-literal (in either case, including within a user-defined-literal) corresponds to a control character or to a character in the basic source character set, the program is ill-formed. [ Footnote: Note: A sequence of characters resembling a universal-character-name in an r-char-sequence (5.13.5) does not form a universal-character-name. ] [Note: ISO/IEC 10646 code points are integers in the range [0, 10FFFF] (hexadecimal). A surrogate code point is a value in the range [D800, DFFF] (hexadecimal). A control character is a character whose code point is in either of the ranges [0, 1F] or [7F, 9F] (hexadecimal). — end note]

The basic literal character set consists of all characters of the basic character set, plus the control characters specified in table Y.

U+0000	NULL
U+0007	BELL
U+0008	BACKSPACE
U+000D	CARRIAGE RETURN (CR)

A code unit is an integer value of character type (6.8.1 [basic.fundamental]). Characters in a character-literal other than a multicharacter or non-encodable character literal or in a string-literal are encoded as a sequence of one or more code units, as determined by the encoding-prefix ([lex.ccon], [lex.string]); this is termed the respective literal encoding. The ordinary literal encoding is the encoding applied to an ordinary character or string literal. The wide literal encoding is the encoding applied to a wide character or string literal.

A literal encoding encodes each element of the basic literal character set as a single distinct code unit with non-negative value. The null character is encoded as the value 0. No other element of the translation character set is encoded with a code unit of value 0. The code unit value of each decimal digit character after the digit 0 (U+0030) shall be one greater than the value of the previous. The ordinary and wide literal encodings are otherwise implementation-defined. [ Note: The encodings of UTF-8, UTF-16, and UTF-32 literals are specified in [lex.ccon] and [lex.string]). -- end note ]

The basic execution character set and the basic execution wide-character set shall each contain all the members of the basic source character set, plus control characters representing alert, backspace, and carriage return, plus a null character (respectively, null wide character), whose value is 0. For each basic execution character set, the values of the members shall be non-negative and distinct from one another. In both the source and execution basic character sets, the value of each character after 0 in the above list of decimal digits shall be one greater than the value of the previous. The execution character set and the execution wide-character set are implementation-defined supersets of the basic execution character set and the basic execution wide-character set, respectively. The values of the members of the execution character sets and the sets of additional members are locale-specific.

basic-c-char:
    any member of the basic source translation character set except the single-quote ’, backslash \, or new-line character
...
conditional-escape-sequence-char:
    any member of the basic source character set that is not an octal-digit, a simple-escape-sequence-char, or
    the characters u, U, or x

[Note 1 : The associated character encoding for ordinary and wide character literals determines encodability, but does not determine the value of non-encodable ordinary or wide character literals or ordinary or wide multicharacter literals. The examples in Table 9 for non-encodable ordinary and wide character literals assume that the specified character lacks representation in the execution character set ordinary literal encoding or execution wide-character set wide literal encoding, respectively, or that encoding it would require more than one code unit. — end note]

Encoding prefix	...	Associated character encoding
none	...	encoding of the execution character set ordinary literal encoding
L	...	encoding of the execution wide-character set wide literal encoding

basic-s-char:
    any member of the basic source translation character set except the double-quote ", backslash \, or new-line character
...
r-char:
    any member of the source translation character set, except a right parenthesis ) followed by
    the initial d-char-sequence (which may be empty) followed by a double quote ".

Encoding prefix	...	Associated character encoding
none	...	encoding of the execution character set ordinary literal encoding
L	...	encoding of the execution widecharacter set wide literal encoding

- 7 - In translation phase 6 (5.2 [lex.phases]), adjacent string-literals are concatenated. The common encoding-prefix for a sequence of adjacent string-literals is determined pairwise as follows: If both two string-literals have the same encoding-prefix, the resulting concatenated string-literal has common encoding-prefix is that encoding-prefix. If one string-literal has no encoding-prefix, it is treated as a string-literal of the same encoding-prefix as the common encoding-prefix is that of the other operand string-literal. If a UTF-8 string literal token is adjacent to a wide string literal token, the program is ill-formed. Any other concatenations combinations are conditionally-supported with implementation-defined behavior. [Note: This concatenation is an interpretation, not a conversion. Because the interpretation happens in translation phase 6 (after each character from a string-literal has been translated into a value from the appropriate character set), a A string-literal’s initial rawness has no effect on the interpretation or well-formedness of the concatenation determination of the common encoding-prefix. -- end note]

Table 13 has some examples of valid concatenations.

- 8 - In translation phase 6 (5.2 [lex.phases]), adjacent string-literals are concatenated. The lexical structure of the contents of the individual string-literals is retained. Characters in concatenated strings are kept distinct. [Example:

"\xA" "B"

contains the two characters represents the code unit ’\xA’ and the character ’B’ after concatenation (and not the single hexadecimal character code unit ’\xAB’). Similarly,

R"(\u00)" "41"

represents six characters, starting with a backslash and ending with the digit 1 (and not the single character "A" specified by a universal-character-name).

Table 13 has some examples of valid concatenations. — end example]

In translation phase 6 (5.2), after adjacent string-literals are concatenated, a null character is appended to the result.

String literal objects are initialized with the sequence of code unit values corresponding to the string-literal's sequence of s-chars (for a non-raw string literal) and r-chars (for a raw string literal), plus a terminating null character, in order as follows:

[ Note: The sequence c₁ c₂ ...c_k can only contain characters from the basic source character set. — end note]

[ Note: The sequence c₁ c₂ ...c_k can only contain characters from the basic source character set. — end note]

The fundamental storage unit in the memory model is the byte. A byte is at least large enough to contain any member the ordinary literal encoding of any element of the basic execution literal character set (5.3) and the eight-bit code units of the Unicode UTF-8 encoding form and is composed of a contiguous sequence of bits, [ Footnote: ... ] the number of which is implementation-defined.

Type char is a distinct type that has an implementation-defined choice of “signed char” or “unsigned char” as its underlying type. The values of type char can represent distinct codes for all members of the implementation’s basic character set. ...

Change in 6.8.2 [basic.fundamental] paragraph 8:

Type wchar_t is a distinct type that has an implementation-defined signed or unsigned integer type as its underlying type. The values of type wchar_t can represent distinct codes for all members of the largest extended any character set specified among the supported locales (28.3.1).

The types char, wchar_t, char8_t, char16_t, char32_t are collectively called character types. The character types, Types bool, char, wchar_t, char8_t, char16_t, char32_t, and the signed and unsigned integer types are collectively called integral types. A synonym for integral type is integer type. [Note: Enumerations (9.7.1) are not integral; however, unscoped enumerations can be promoted to integral types as specified in 7.3.6. — end note]

A literal is a primary expression. The type of a literal is determined based on its form as specified in 5.13 [lex.literal]. A string-literal is an lvalue designating the corresponding string literal object ([lex.string]), a user-defined-literal has the same value category as the corresponding operator call expression described in 5.13.8 [lex.ext], and any other literal is a prvalue.

The resulting tokens comprise the controlling constant expression which is evaluated according to the rules of 7.7 using arithmetic that has at least the ranges specified in 17.3. For the purposes of this token conversion and evaluation all signed and unsigned integer types act as if they have the same representation as, respectively, intmax_t or uintmax_t (17.4). [Note: ... -- end note] This includes interpreting character-literals, which may involve converting escape sequences into execution character set members interpreting escape-sequences and universal-character-names (5.13.3 [lex.ccon]). Whether the numeric value for these character-literals matches the value obtained when an identical character-literal occurs in an expression (other than within a #if or #elif directive) is implementation-defined. [Note: ... -- end note] Also, whether a single-character character-literal may have a negative value is implementation-defined. Each subexpression with type bool is subjected to integral promotion before processing continues.

For both object-like and function-like macro invocations, before the replacement list is reexamined for more macro names to replace, each instance of a ## preprocessing token in the replacement list (not from an argument) is deleted and the preceding preprocessing token is concatenated with the following preprocessing token. Placemarker preprocessing tokens are handled specially: concatenation of two placemarkers results in a single placemarker preprocessing token, and concatenation of a placemarker with a non-placemarker preprocessing token results in the non-placemarker preprocessing token. If the result is not a valid preprocessing token, the behavior is undefined. If the result matches the syntax of a universal-character-name, the behavior is undefined. The resulting token is available for further macro replacement. The order of evaluation of ## operators is unspecified.

...

• A letter is any of the 26 lowercase or 26 uppercase letters in the basic execution character set (5.3 [lex.charset]).
• ...

A null-terminated multibyte string, or ntmbs, is an ntbs that constitutes a sequence of valid multibyte characters, beginning and ending in the initial shift state. [ Footnote: An NTBS that contains characters only from the basic execution literal character set is also an NTMBS. Each multibyte character then consists of a single byte. ]

%Z	The time zone abbreviation or name. A single word is parsed. This word can only contain characters from the basic source character set (5.3 [lex.charset]) that are alphanumeric, or one of ’_’, ’/’, ’-’, or ’+’.

The only characters for which unique transformations are required are those in the basic source character set (5.3 [lex.charset]).

[...]

For any character c in the basic source character set (5.3 [lex.charset]) the transformation is such that

do_widen(do_narrow(c, 0)) == c