lemene

說明

big endian和little endian表示如何存放多字節數據。前者低位字節排放在內存的高端，后者相反。將unsigned long數據強制轉換成unsigned char*數據，則它們在兩種模式下的對應關系如下：

big endian：

ul = (uc[0]<< 24) + (uc[1]<<16) + (uc[2]<<8) + uc[3];

little endian:

ul = (uc[3]<<24) + (uc[2]<<16) + (uc[1]<<8) + uc[0];

實驗代碼

/**
 * @file little_big_endian.cpp
 * @brief 測試大小端字節序
 * @copyright public domain
 */
#include <iostream>
static bool is_little_endian() {
union {
long l;
char cs[4];
} t;
t.l = 1;
return t.cs[0] == 1;
}
int main() {
unsigned long ul = 0x12345678;
unsigned char* uc = (unsigned char*)&ul;
if (is_little_endian()) {
bool r = (uc[0] + (uc[1]<<8) + (uc[2]<<16) + (uc[3]<<24)) == ul;
std::cout << "little: (uc[0] + (uc[1]<<8) + (uc[2]<<16) + (uc[3]<<24)) == ul is " << (r ? "true" : "false") << std::endl;
} else {
bool r = (uc[3] + (uc[2]<<8) + (uc[1]<<16) + (uc[0]<<24)) == ul;
std::cout << "little: (uc[3] + (uc[2]<<8) + (uc[1]<<16) + (uc[0]<<24)) == ul is " << (r ? "true" : "false") << std::endl;
}
return 0;
}

運行及結果

$ g++ little_big_endian.cpp 
$ ./a.out 
little: (uc[0] + (uc[1]<<8) + (uc[2]<<16) + (uc[3]<<24)) == ul is true

常見問題

字節序的問題容易出現在不同電腦交互數據的時候，因此當數據輸出時——保存成文件或在網絡上傳輸——就應該考慮字節序。