Obtaining the current working directory的一个Linux特性

   一般的，如果要获得当前进程的当前工作目录(CWD)的话，显然要用到如下函数：

#include

char * getcwd (char *buf, size_t size);

这个函数的使用时显而易见的就是传递一个size长度的缓冲区，然后函数调用会填充这个缓冲区，出错情况下返回NULL,这里不讨论这个函数的一般使用，而是要说一点关于这个函数在Linux下的一个特殊使用方法：

POSIX dictates that the behavior of getcwd( ) is undefined if buf is NULL. Linux’s C

library, in this case, will allocate a buffer of length size bytes, and store the current

working directory there. If size is 0, the C library will allocate a buffer sufficiently

large to store the current working directory. It is then the application’s responsibility

to free the buffer, via free( ), when it’s done with it. Because this behavior is Linux-

specific, applications that value portability or a strict adherence to POSIX should not

rely on this functionality. This feature, however, does make usage very simple!

Here’s an example:

char *cwd;
cwd = getcwd (NULL, 0);
if (!cwd) {
        perror ("getcwd");
        exit (EXIT_FAILURE);
}
printf ("cwd = %s\n", cwd);
free (cwd);

Linux’s C library also provides a get_current_dir_name( ) fu

same behavior as getcwd( ) when passed a NULL buf and a size 

#define _GNU_SOURCE

#include

char * get_current_dir_name (void);

Thus, this snippet behaves the same as the previous one:

char *cwd;
cwd = get_current_dir_name ( );
if (!cwd) {
        perror ("get_current_dir_name");
        exit (EXIT_FAILURE);
}
printf ("cwd = %s\n", cwd);
free (cwd);