# C23

> C++ Atlas: `cppatlas.com/w/c/23` | Source: https://en.cppreference.com/w/c/23 | Domain: Standards & support

ISO/IEC 9899:2024, a.k.a. C23, is the current revision of the C standard.

## Official Specification Summary (cppreference)

**ISO/IEC 9899:2024**, a.k.a. **C23**, is the current revision of the C standard.

- The latest publicly available WD of C23 (n3435, 2025-01-03).
- The latest WD of C23 (n3149, 2023-07-09, a password-protected zip archive, see n3150).

> Note: This section is incomplete Reason: accepted changes included in the final draft

C23 bumps the predefined macro `__STDC_VERSION__` to `202311L`.

### Obsolete features

#### Removed features

- Old-style function declarations and function definitions
- Representations for signed integers other than two's complement
- Permission that `u`/`U`-prefixed character constants and string literals may be not UTF-16/32
- Mixed wide string literal concatenation
- Support for calling `realloc()` with zero size (the behavior becomes undefined)
- `__alignof_is_defined` and `__alignas_is_defined`
- `static_assert` is not provided as a macro defined in `<assert.h>` (becomes a keyword)
- `thread_local` is not provided as a macro defined in `<threads.h>` (becomes a keyword)

#### Deprecated features

- `<stdnoreturn.h>`
- Old feature-test macros
  - `__STDC_IEC_559__`
  - `__STDC_IEC_559_COMPLEX__`
- `_Noreturn` function specifier
- `_Noreturn` attribute token
- `asctime()`
- `ctime()`
- `DECIMAL_DIG` (use the appropriate type-specific macro (`FLT_DECIMAL_DIG`, etc) instead)
- Definition of following numeric limit macros in `<math.h>` (they should be used via `<float.h>`)
  - `INFINITY`
  - `DEC_INFINITY`
  - `NAN`
  - `DEC_NAN`
- `__bool_true_false_are_defined`

### New language features

- Decimal floating-point types (`_Decimal32`, `_Decimal64`, and `_Decimal128`)
- Bit-precise integers (`_BitInt`)
- Binary integer constants
- `u8` character constants
- Type change of `u8` string literals
- Digit separator `'`
- Empty initializer `={}`
- Attributes
  - `[[deprecated]]`
  - `[[fallthrough]]`
  - `[[maybe_unused]]`
  - `[[nodiscard]]`
  - `[[noreturn]]`
  - `[[reproducible]]`
  - `[[unsequenced]]`
- Unnamed parameters in function definitions
- Identical cvr-qualifications for array types and their element types
- Single-argument `_Static_assert`
- `static_assert` becomes a keyword (may be a predefined macro for compatibility reasons)
- `thread_local` becomes a keyword (may be a predefined macro for compatibility reasons)
- Labels followed by declarations and `}`
- `nullptr` constant and the associated `nullptr_t` type
- `true` and `false` become keywords (may be predefined macros for compatibility reasons)
- New preprocessor directives
  - `#elifdef`
  - `#elifndef`
  - `#warning`
  - `#embed`
- Pragmas for rounding direction
  - `STDC` `FENV_ROUND`
  - `STDC` `FENV_DEC_ROUND`

#### Feature test macros for optional features

- `__STDC_IEC_60559_BFP__`
  - Indicates IEEE-754 binary floating-point arithmetic and required math functions are supported.
    This macro supersedes `__STDC_IEC_559__`.
- `__STDC_IEC_60559_DFP__`
  - Indicates IEEE-754 decimal floating-point arithmetic and required math functions are supported.
- `__STDC_IEC_60559_COMPLEX__`
  - Indicates IEEE-754 complex arithmetic and required math functions are supported.
    This macro supersedes `__STDC_IEC_559_COMPLEX__`.

### New library features

#### New headers

- `<stdbit.h>`
- `<stdckdint.h>`

#### Library features

- Extended binary floating-point math functions
  - > Note: This section is incomplete Reason: List TBD
- Decimal floating-point math functions
  - -`dN` variants for existing and new floating-point math functions
  - `quantizedN()`
  - `samequantumdN()`
  - `quantumdN()`
  - `llquantexpdN()`
  - `encodedecdN()`
  - `decodedecdN()`
  - `encodebindN()`
  - `decodebindN()`
- Floating-point formatting functions
- Library support for UTF-8
  - `char8_t` type alias
  - `mbrtoc8()`
  - `c8rtomb()`
  - `atomic_char8_t` type alias
  - `ATOMIC_CHAR8_T_LOCK_FREE` test macro
- `memset_explicit()`
- POSIX functions
  - `memccpy()`
  - `strdup()`
  - `strndup()`
  - `gmtime_r()`
  - `localtime_r()`
  - Extensions for `strftime()` and `wcsftime()`
- Extensions for `fscanf()` and `fprintf()` function families
  - `wN` and `wfN` length modifiers for `[u]intN_t` and `[u]int_fastN_t` respectively
  - `H`, `D`, and `DD` length modifiers for `_Decimal32`, `_Decimal64`, and `_Decimal128` respectively
  - `b` conversion specifier for unsigned integer types
- `timespec_getres()`
- Macro constants for width of integer types
- Additional numeric limit macros for floating-point types
- Library version-test macros
  - `__STDC_VERSION_FENV_H__`
  - `__STDC_VERSION_MATH_H__`
  - `__STDC_VERSION_STDINT_H__`
  - `__STDC_VERSION_STDLIB_H__`
  - `__STDC_VERSION_TGMATH_H__`
  - `__STDC_VERSION_TIME_H__`
  - `__STDC_VERSION_STDCKDINT_H__`
  - `__STDC_VERSION_STDBIT_H__`

### Defect reports

**Defect Reports fixed in C23 (? defects)**

- DR 440
- DR 432
- DR 467
- DR 476
- DR 482
- DR 488
- DR 489
- DR 494
- DR 496
- DR 497
- DR 499
- DR 500
- DR 501

### Compiler support

#### C23 core language features

> Note: This section is incomplete Reason: status for Apple Clang and other compilers supporting C2x

| C23 feature | Paper(s) | GCC | Clang | MSVC | Apple Clang | EDG eccp | Intel C++ | Nvidia HPC C++ (ex PGI)* | Nvidia nvcc | Cray |   |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| `static_assert` with no message | N2265 | 9 | 9 | Yes | Yes | 6.5 | 2021.1.2 (clang based) |   |   |   |   |
| `[[nodiscard]]` | N2267 | 10 | 9 |   | Yes | 6.4 | 2021.1.2 (clang based) |   |   |   |   |
| `[[maybe_unused]]` | N2270 | 10 | 9 |   | Yes | 6.4 | 2021.1.2 (clang based) |   |   |   |   |
| `[[deprecated]]` | N2334 | 10 | 9 |   | Yes | 6.4 | 2021.1.2 (clang based) |   |   |   |   |
| Attributes | N2335 N2554 | 10 | 9 |   | Yes | 6.4 | 2021.1.2 (clang based) |   |   |   |   |
| IEEE 754 decimal floating-point types | N2341 | 4.2 (partial)* 12 |   |   |   |   | 13.0 (partial)* |   |   |   |   |
| `[[fallthrough]]` | N2408 | 10 | 9 |   | Yes | 6.4 | 2021.1.2 (clang based) |   |   |   |   |
| `u8` character constants | N2418 | 10 | 15 |   |   | 6.5 | 2022.2 |   |   |   |   |
| Removal of function definitions without prototype | N2432 | 10 | 15 |   |   |   | 2022.2 |   |   |   |   |
| `[[nodiscard]]` with message | N2448 | 11 | 10 |   | Yes | 6.4 | 2021.1.2 (clang based) |   |   |   |   |
| Unnamed parameters in function definitions | N2480 | 11 | 11 |   | Yes | 6.4 | 2021.1.2 (clang based) |   |   |   |   |
| Labels before declarations and end of blocks | N2508 | 11 | 16 | Partial* |   | 6.5 | 17.0* |   |   |   |   |
| Binary integer constants | N2549 | 4.3* 11 | 2.9* 9 | 19.0 (2015)** | Yes | 6.5 | 11.0* |   |   |   |   |
| `__has_c_attribute` in preprocessor conditionals | N2553 | 11 | 9 |   | Yes | 6.5 | 2021.1.2 (clang based) |   |   |   |   |
| Allow duplicate attributes | N2557 | 11 | 13 |   | Yes | 6.5 | 2021.4 (clang-based |   |   |   |   |
| IEEE 754 interchange and extended types | N2601 | 7 (partial)* 14 | 6 (partial)* |   | Partial* |   |   |   |   |   |   |
| Digit separators | N2626 | 12 | 13 | 19.0 (2015)** | Yes | 6.5 | 18.0* |   |   |   |   |
| `#elifdef` and `#elifndef` | N2645 | 12 | 13 | 19.40* | 13.1.6* | 6.5 | 2021.4 |   |   |   |   |
| Type change of `u8` string literals | N2653 | 13 |   |   |   |   |   |   |   |   |   |
| `[[maybe_unused]]` for labels | N2662 | 11 | 16 |   |   | 6.5 | 2022.2 |   |   |   |   |
| `#warning` | N2686 | Yes | Yes |   | Yes | 6.5 | Yes |   |   |   |   |
| Bit-precise integer types (`_BitInt`) | N2763 | 14 (partial)* | 15 |   |   | 6.5 | 2022.2 |   |   |   |   |
| `[[noreturn]]` | N2764 | 13 | 15 |   |   | 6.5 | 2022.2 |   |   |   |   |
| Suffixes for bit-precise integer constants | N2775 | 14 | 15 |   |   |   | 2022.2 |   |   |   |   |
| `__has_include` in preprocessor conditionals | N2799 | 5 | Yes | 19.11* | Yes | 6.5 | 18.0 |   |   |   |   |
| Identifier Syntax using Unicode Standard Annex 31 | N2836 | 13 | 15 |   |   | 6.5 | 2022.2 |   |   |   |   |
| Removal of function declarations without prototype | N2841 | 13 | 15 |   |   |   | 2022.2 |   |   |   |   |
| Empty initializers | N2900 | Partial* 13 | Partial* |   | Partial* | Partial* | Partial* |   |   |   |   |
| `typeof` and `typeof_unqual` | N2927 N2930 | Partial* 13 | Partial* 16 | 19.39* | Partial* | Partial* | Partial* |   |   | Partial* |   |
| New spelling of keywords | N2934 | 13 | 16 |   |   | 6.5 |   |   |   |   |   |
| Predefined `true` and `false` | N2935 | 13 | 15 |   |   |   | 2022.2 |   |   |   |   |
| `[[unsequenced]]` and `[[reproducible]]` | N2956 | 15 |   |   |   |   |   |   |   |   |   |
| Relax requirements for variadic parameter list | N2975 | 13 | 16 |   |   | 6.5 | 2023.1 |   |   |   |   |
| Type inference in object definitions | N3007 | 13 | 18 |   |   |   |   |   |   |   |   |
| `#embed` | N3017 | 15 | 19 |   |   |   |   |   |   |   |   |
| `constexpr` objects | N3018 | 13 | 19 |   |   |   |   |   |   |   |   |
| Improved Normal Enumerations | N3029 | 13 | 20* |   |   |   |   |   |   |   |   |
| Enumerations with fixed underlying types | N3030 | 13 | 20* |   |   |   |   |   |   |   |   |
| `__VA_OPT__` | N3033 | 8 13 | 12 | 19.39* |   | 6.5 |   |   |   |   |   |
| Storage-class specifiers for compound literals | N3038 | 13 |   |   |   |   |   |   |   |   |   |
| `nullptr` | N3042 | 13 | 16 |   |   |   |   |   |   |   |   |
| C23 feature | Paper(s) | GCC | Clang | MSVC | Apple Clang | EDG eccp | Intel C++ | Nvidia HPC C++ (ex PGI)* | Nvidia nvcc | Cray |   |

#### C23 library features

> Note: This section is incomplete Reason: a different list for C standard libraries

---
Attribution: content from https://en.cppreference.com/w/c/23 (cppreference.com), licensed under CC-BY-SA 3.0 and GFDL. Modified: converted to markdown, study notes added. Offline snapshot of 9 February 2025.