89 lines
3.0 KiB
C
89 lines
3.0 KiB
C
|
#include <c_StrIndexKmp.h>
|
||
|
|
#include <c_Memory.h>
|
||
|
|
|
||
|
|
/**
|
||
|
|
* @brief 私有辅助内部函数：为模式串动态解包并构建 KMP 高速失配跳转跳转表
|
||
|
|
*/
|
||
|
|
C_STATIC_FORCE_INLINE
|
||
|
|
c_err_t c_Kmp_BuildNextTable(const uint8_t* p, c_size_t m, c_size_t* next) {
|
||
|
|
next[0] = 0; // KMP 起始基石
|
||
|
|
c_size_t len = 0; // 前缀的核心长度
|
||
|
|
c_size_t i = 1;
|
||
|
|
|
||
|
|
while (i < m) {
|
||
|
|
if (p[i] == p[len]) {
|
||
|
|
len++;
|
||
|
|
next[i] = len;
|
||
|
|
i++;
|
||
|
|
} else {
|
||
|
|
if (len != 0) {
|
||
|
|
len = next[len - 1]; // 顺着失配位置向左回溯
|
||
|
|
} else {
|
||
|
|
next[i] = 0;
|
||
|
|
i++;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
return C_ERR_OK;
|
||
|
|
}
|
||
|
|
|
||
|
|
c_size_t c_StrIndexKmpEx(const void* text, c_size_t text_len,
|
||
|
|
const void* pattern, c_size_t pattern_len,
|
||
|
|
c_size_t start_offset, c_Allocator_t* allocator) {
|
||
|
|
// 1. 强御级无效边界防御
|
||
|
|
if (!text || !pattern) return C_KMP_NOT_FOUND;
|
||
|
|
|
||
|
|
const uint8_t* t = (const uint8_t*)text;
|
||
|
|
const uint8_t* p = (const uint8_t*)pattern;
|
||
|
|
|
||
|
|
// 如果传入 0 则自适应兼容裸字符串
|
||
|
|
c_size_t n = (text_len == 0) ? (c_size_t)strlen((const char*)text) : text_len;
|
||
|
|
c_size_t m = (pattern_len == 0) ? (c_size_t)strlen((const char*)pattern) : pattern_len;
|
||
|
|
|
||
|
|
if (m == 0 || n == 0 || start_offset >= n || m > (n - start_offset)) {
|
||
|
|
return C_KMP_NOT_FOUND;
|
||
|
|
}
|
||
|
|
|
||
|
|
c_Allocator_t* alloc = (allocator != NULL) ? allocator : &c_DefaultAllocator;
|
||
|
|
|
||
|
|
// 2. 【安全机制】：坚决不使用栈上的局部可变长数组(VLA)，防止被大特征串顶爆引发 Stack Overflow
|
||
|
|
// 从内部绑定的专属多态分配器中动态申领物理块资源
|
||
|
|
c_size_t* next = (c_size_t*)c_Allocator_Alloc(alloc, m * sizeof(c_size_t));
|
||
|
|
if (!next) return C_KMP_NOT_FOUND;
|
||
|
|
|
||
|
|
// 3. 构建 KMP 高速失配加速表
|
||
|
|
c_Kmp_BuildNextTable(p, m, next);
|
||
|
|
|
||
|
|
// 4. 开始执行双指针拉平线性单向匹配检索流 (无回溯 O(M+N))
|
||
|
|
c_size_t i = start_offset; // 文本大字符串的推进指针
|
||
|
|
c_size_t j = 0; // 模式串的局部指针
|
||
|
|
|
||
|
|
while (i < n) {
|
||
|
|
if (t[i] == p[j]) {
|
||
|
|
i++;
|
||
|
|
j++;
|
||
|
|
}
|
||
|
|
|
||
|
|
if (j == m) {
|
||
|
|
// 完美踩中特征指纹，捕获命中！
|
||
|
|
c_size_t found_idx = i - j;
|
||
|
|
|
||
|
|
// 强异常安全性物理火化：归还申领的跳转数组块，杜绝堆碎片跑飞
|
||
|
|
c_Allocator_Free(alloc, next);
|
||
|
|
return found_idx; // 返回命中位置相对于首地址的绝对物理偏移量
|
||
|
|
}
|
||
|
|
else if (i < n && t[i] != p[j]) {
|
||
|
|
// 发生失配抖动，KMP 的精髓：文本指针 i 绝对不后退，只原位等待，而模式串指针 j 顺着失配表向左高速跳转
|
||
|
|
if (j != 0) {
|
||
|
|
j = next[j - 1];
|
||
|
|
} else {
|
||
|
|
i++;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
// 全文扫描结束，未发生特征匹配，体面卸载资源退出
|
||
|
|
c_Allocator_Free(alloc, next);
|
||
|
|
return C_KMP_NOT_FOUND;
|
||
|
|
}
|