C23特性简介

2026-06-04T00:00:00+00:00
C23（ISO/IEC 9899:2024）是 C 语言的最新正式修订版，__STDC_VERSION__</code> 定义为 202311L</code>。它引入了许多提升可读性、安全性、与 C++ 兼容性和现代浮点/整数支持的功能，同时移除或废弃了一些过时特性。</p>
核心语言特性</h3>
属性（Attributes）</h4>
C23 采用 C++11 风格的双中括号属性语法 [[...]]</code>，并新增/标准化多个属性。</p>

[[nodiscard]]</code>（带或不带消息）</li>
[[maybe_unused]]</code></li>
[[deprecated]]</code>（带或不带消息）</li>
[[fallthrough]]</code></li>
[[noreturn]]</code></li>
[[unsequenced]]</code> 和 [[reproducible]]</code>（用于函数优化提示）</li>
</ul>
示例代码</strong>：</p>
[[</span>nodiscard</span>(</span>"检查返回值"</span>)]]</span></span>
int</span> create_resource</span>(</span>void</span>);</span></span>
</span>
void</span> process</span>(</span>void</span>) {</span></span>
    [[</span>maybe_unused</span>]]</span> int</span> unused_var </span>=</span> 42</span>;</span></span>
    switch</span> (</span>x</span>) {</span></span>
        case</span> 1</span>:</span></span>
            [[</span>fallthrough</span>]];</span></span>
        case</span> 2</span>:</span></span>
            // ...</span></span>
    }</span></span>
}</span></span></code></pre>二进制整数常量和数字分隔符</h4>
支持 0b</code> / 0B</code> 前缀的二进制字面量，以及 '</code> 作为数字分隔符。</p>
示例代码</strong>：</p>
int</span> flags </span>=</span> 0b</span>1010</span>'</span>1100</span>;</span></span>
long</span> big </span>=</span> 1</span>'</span>000</span>'</span>000</span>'</span>000</span>;</span></span>
double</span> pi_approx </span>=</span> 3.141</span>'</span>592</span>'</span>653</span>;</span></span></code></pre>static_assert</code> 简化与关键字化</h4>
static_assert</code> 现在支持单参数形式（无消息），且成为关键字（兼容宏仍可能存在）。</p>
示例代码</strong>：</p>
static_assert</span>(</span>sizeof</span>(</span>int</span>)</span> >=</span> 4</span>);</span>  // 无消息版本</span></span>
static_assert</span>(</span>CHAR_BIT </span>==</span> 8</span>,</span> "字节必须为 8 位"</span>);</span></span></code></pre>nullptr</code> 常量和 nullptr_t</code> 类型</h4>
引入 nullptr</code> 关键字和对应的 nullptr_t</code> 类型，提升空指针安全性。</p>
示例代码</strong>：</p>
#include</span> <stddef.h></span></span>
</span>
nullptr_t</span> np </span>=</span> nullptr</span>;</span></span>
int</span>*</span> p </span>=</span> nullptr</span>;</span>  // 更清晰</span></span></code></pre>true</code> / false</code> 成为关键字</h4>
bool</code>、true</code>、false</code> 正式标准化（兼容旧 _Bool</code> 等）。</p>
示例代码</strong>：</p>
bool</span> flag </span>=</span> true</span>;</span></span>
if</span> (</span>flag </span>==</span> false</span>) {</span> /* ... */</span> }</span></span></code></pre>类型推断 auto</code>（对象定义）</h4>
auto</code> 用于对象定义时进行类型推断（保留原有存储类说明符语义）。仅限对象定义，不支持函数返回类型或参数推断。</p>
示例代码</strong>：</p>
auto x </span>=</span> 42</span>;</span>           // int</span></span>
auto y </span>=</span> 3.14</span>;</span>         // double</span></span>
auto str </span>=</span> "hello"</span>;</span>    // const char*</span></span>
auto arr </span>=</span> (</span>int[]</span>){</span>1</span>,</span>2</span>,</span>3</span>};</span>  // 复合字面量</span></span></code></pre>typeof</code> 和 typeof_unqual</code></h4>
类型查询运算符，支持 unqual</code> 去除限定符。</p>
示例代码</strong>：</p>
int</span> i </span>=</span> 10</span>;</span></span>
typeof</span>(</span>i</span>)</span> j </span>=</span> i</span>;</span>                    // int j</span></span>
typeof_unqual</span>(</span>const volatile</span> int</span>)</span> k</span>;</span> // int k</span></span></code></pre>位精确整数类型 _BitInt(N)</code></h4>
精确指定位宽的整数类型（_BitInt(N)</code> 和 unsigned _BitInt(N)</code>），支持 wb</code>/uwb</code> 后缀。</p>
示例代码</strong>：</p>
_BitInt</span>(</span>24</span>)</span> precise24 </span>=</span> 0xABCDEFwb</span>;</span></span>
unsigned</span> _BitInt</span>(</span>16</span>)</span> u16 </span>=</span> 65535uwb</span>;</span></span>
</span>
_BitInt</span>(</span>128</span>)</span> big </span>=</span> 1</span>;</span>  // 支持大位宽</span></span></code></pre>空初始化器 {}</code></h4>
支持使用 {}</code> 进行零初始化（包括 VLA）。</p>
示例代码</strong>：</p>
int</span> arr</span>[</span>10</span>]</span> =</span> {};</span>          // 全零</span></span>
struct</span> S s </span>=</span> {};</span></span>
int</span> n </span>=</span> 5</span>;</span></span>
int</span> vla</span>[</span>n</span>]</span> =</span> {};</span>           // VLA 也支持</span></span></code></pre>其他语言改进</h4>

标签可置于声明之前和复合语句末尾。</li>
函数定义中允许未命名参数。</li>
移除无原型函数定义/声明（K&R 风格废弃）。</li>
#elifdef</code> / #elifndef</code>、#warning</code>、__has_include</code>、__has_c_attribute</code>、__VA_OPT__</code> 等预处理器增强。</li>
#embed</code> 指令（嵌入二进制资源）。</li>
constexpr</code> 对象（非函数）。</li>
u8</code> 字符常量和字符串字面量类型改为 char8_t</code>。</li>
十进制浮点类型（_Decimal32</code> 等，可选）。</li>
</ul>
示例代码</strong>（预处理器）：</p>
#if</span> __has_include</span>(</span><</span>stdbit</span>.</span>h</span>></span>)</span></span>
#  include</span> <stdbit.h></span></span>
#else</span></span>
#  warning</span> "stdbit.h not available"</span></span>
#endif</span></span></code></pre>标准库更新</h3>
<limits.h></code> 新增位宽常量</h4>
新增一系列 *_WIDTH</code> 宏，用于标准整数类型的位宽。</p>
示例代码</strong>：</p>
#include</span> <limits.h></span></span>
</span>
printf</span>(</span>"int 位宽: %d</span>\n</span>"</span>,</span> INT_WIDTH</span>);</span></span>
printf</span>(</span>"bool 位宽: %d</span>\n</span>"</span>,</span> BOOL_WIDTH</span>);</span></span>
printf</span>(</span>"long long 位宽: %d</span>\n</span>"</span>,</span> LLONG_WIDTH</span>);</span></span></code></pre><math.h></code> 新增 nextup</code> / nextdown</code> 系列</h4>
nextup</code>、nextupf</code>、nextupl</code>、nextdown</code>、nextdownf</code>、nextdownl</code>：返回大于/小于给定值的下一个可表示浮点值。</p>
示例代码</strong>：</p>
#include</span> <math.h></span></span>
</span>
double</span> x </span>=</span> 1.0</span>;</span></span>
double</span> next </span>=</span> nextup</span>(</span>x</span>);</span>      // 大于 1.0 的下一个可表示值</span></span>
float</span> fnext </span>=</span> nextdownf</span>(</span>0.0f</span>);</span> // 小于 0.0f 的下一个可表示值</span></span></code></pre>其他库特性</h4>

memset_explicit()</code>：安全清除敏感内存（防止优化消除）。</li>
<stdbit.h></code>：位操作工具（如 stdc_count_ones()</code> 等）。</li>
<stdckdint.h></code>：带溢出检查的整数运算宏（ckd_add</code> 等）。</li>
扩展的 printf</code>/scanf</code> 支持（%b</code> 二进制、%wN</code> 宽度修饰符等）。</li>
UTF-8 支持增强（char8_t</code>、mbrtoc8</code> 等）。</li>
更多 IEEE 754 兼容函数和十进制浮点支持。</li>
</ul>
废弃/移除特性</h3>

K&R 无原型函数定义/声明。</li>
三字符序列（trigraphs）。</li>
非二补码有符号整数表示。</li>
某些旧特性测试宏和 <stdnoreturn.h></code> 等。</li>
NAN</code>, INFINITY</code> 等常数的 <math.h></code> 定义废弃，现移动至 <float.h></code>。</li>
</ul>
mokurin000's blog - C

C23特性简介

标准库更新</h3>
