[英]How to recursively list directories in C on Linux?
我需要递归列出 C 编程中的所有目录和文件。 我已经研究过 FTW,但它不包含在我使用的 2 个操作系统(Fedora 和 Minix)中。 在过去的几个小时里,我读到的所有不同的东西都让我头疼。
如果有人知道我可以查看的代码片段,那将是惊人的,或者如果有人能给我很好的指导,我将非常感激。
为什么每个人都坚持一次又一次地重新发明轮子?
POSIX.1-2008 标准化了nftw()
函数,也在单一 Unix 规范 v4 (SuSv4) 中定义,并且在 Linux (glibc, man 3 nftw
)、OS X 和大多数当前的 BSD 变体中可用。 这一点也不新鲜。
朴素的基于opendir()
/ readdir()
/ closedir()
的实现几乎从不处理在树遍历期间移动、重命名或删除目录或文件的情况,而nftw()
应该优雅地处理它们。
例如,考虑以下 C 程序,它列出了从当前工作目录开始的目录树,或者在命令行中命名的每个目录,或者只是在命令行中命名的文件:
/* We want POSIX.1-2008 + XSI, i.e. SuSv4, features */
#define _XOPEN_SOURCE 700
/* Added on 2017-06-25:
If the C library can support 64-bit file sizes
and offsets, using the standard names,
these defines tell the C library to do so. */
#define _LARGEFILE64_SOURCE
#define _FILE_OFFSET_BITS 64
#include <stdlib.h>
#include <unistd.h>
#include <ftw.h>
#include <time.h>
#include <stdio.h>
#include <string.h>
#include <errno.h>
/* POSIX.1 says each process has at least 20 file descriptors.
* Three of those belong to the standard streams.
* Here, we use a conservative estimate of 15 available;
* assuming we use at most two for other uses in this program,
* we should never run into any problems.
* Most trees are shallower than that, so it is efficient.
* Deeper trees are traversed fine, just a bit slower.
* (Linux allows typically hundreds to thousands of open files,
* so you'll probably never see any issues even if you used
* a much higher value, say a couple of hundred, but
* 15 is a safe, reasonable value.)
*/
#ifndef USE_FDS
#define USE_FDS 15
#endif
int print_entry(const char *filepath, const struct stat *info,
const int typeflag, struct FTW *pathinfo)
{
/* const char *const filename = filepath + pathinfo->base; */
const double bytes = (double)info->st_size; /* Not exact if large! */
struct tm mtime;
localtime_r(&(info->st_mtime), &mtime);
printf("%04d-%02d-%02d %02d:%02d:%02d",
mtime.tm_year+1900, mtime.tm_mon+1, mtime.tm_mday,
mtime.tm_hour, mtime.tm_min, mtime.tm_sec);
if (bytes >= 1099511627776.0)
printf(" %9.3f TiB", bytes / 1099511627776.0);
else
if (bytes >= 1073741824.0)
printf(" %9.3f GiB", bytes / 1073741824.0);
else
if (bytes >= 1048576.0)
printf(" %9.3f MiB", bytes / 1048576.0);
else
if (bytes >= 1024.0)
printf(" %9.3f KiB", bytes / 1024.0);
else
printf(" %9.0f B ", bytes);
if (typeflag == FTW_SL) {
char *target;
size_t maxlen = 1023;
ssize_t len;
while (1) {
target = malloc(maxlen + 1);
if (target == NULL)
return ENOMEM;
len = readlink(filepath, target, maxlen);
if (len == (ssize_t)-1) {
const int saved_errno = errno;
free(target);
return saved_errno;
}
if (len >= (ssize_t)maxlen) {
free(target);
maxlen += 1024;
continue;
}
target[len] = '\0';
break;
}
printf(" %s -> %s\n", filepath, target);
free(target);
} else
if (typeflag == FTW_SLN)
printf(" %s (dangling symlink)\n", filepath);
else
if (typeflag == FTW_F)
printf(" %s\n", filepath);
else
if (typeflag == FTW_D || typeflag == FTW_DP)
printf(" %s/\n", filepath);
else
if (typeflag == FTW_DNR)
printf(" %s/ (unreadable)\n", filepath);
else
printf(" %s (unknown)\n", filepath);
return 0;
}
int print_directory_tree(const char *const dirpath)
{
int result;
/* Invalid directory path? */
if (dirpath == NULL || *dirpath == '\0')
return errno = EINVAL;
result = nftw(dirpath, print_entry, USE_FDS, FTW_PHYS);
if (result >= 0)
errno = result;
return errno;
}
int main(int argc, char *argv[])
{
int arg;
if (argc < 2) {
if (print_directory_tree(".")) {
fprintf(stderr, "%s.\n", strerror(errno));
return EXIT_FAILURE;
}
} else {
for (arg = 1; arg < argc; arg++) {
if (print_directory_tree(argv[arg])) {
fprintf(stderr, "%s.\n", strerror(errno));
return EXIT_FAILURE;
}
}
}
return EXIT_SUCCESS;
}
上面的大部分代码都在print_entry()
。 它的任务是打印出每个目录条目。 在print_directory_tree()
,我们告诉nftw()
为它看到的每个目录条目调用它。
上面唯一的手动细节是决定应该让nftw()
使用多少个文件描述符。 如果您的程序在文件树遍历期间最多使用两个额外的文件描述符(除了标准流之外),则已知 15 是安全的(在所有具有nftw()
并且大部分符合 POSIX 的系统上)。
在 Linux 中,您可以使用sysconf(_SC_OPEN_MAX)
来查找打开文件的最大数量,并减去您与nftw()
调用同时使用的nftw()
,但我不会打扰(除非我知道该实用程序将主要用于病态的深层目录结构)。 十五描述不限制树的深度; nftw()
只是变得更慢(并且如果从一个目录中遍历一个比 13 个目录更深的目录,则可能无法检测到目录中的更改,尽管系统和 C 库实现之间的权衡和检测更改的一般能力各不相同)。 仅仅使用这样的编译时常量就可以保持代码的可移植性——它不仅可以在 Linux 上运行,还可以在 Mac OS X 和所有当前的 BSD 变体以及大多数其他不太老的 Unix 变体上运行。
在评论中,Ruslan 提到他们不得不切换到nftw64()
因为他们有需要 64 位大小/偏移量的文件系统条目,而nftw()
的“正常”版本失败了errno == EOVERFLOW
。 正确的解决方案是不切换到 GLIBC 特定的 64 位函数,而是定义_LARGEFILE64_SOURCE
和_FILE_OFFSET_BITS 64
。 这些告诉 C 库尽可能切换到 64 位文件大小和偏移量,同时使用标准函数( nftw()
、 fstat()
等)和类型名称( off_t
等)。
这是一个递归版本:
#include <unistd.h>
#include <sys/types.h>
#include <dirent.h>
#include <stdio.h>
#include <string.h>
void listdir(const char *name, int indent)
{
DIR *dir;
struct dirent *entry;
if (!(dir = opendir(name)))
return;
while ((entry = readdir(dir)) != NULL) {
if (entry->d_type == DT_DIR) {
char path[1024];
if (strcmp(entry->d_name, ".") == 0 || strcmp(entry->d_name, "..") == 0)
continue;
snprintf(path, sizeof(path), "%s/%s", name, entry->d_name);
printf("%*s[%s]\n", indent, "", entry->d_name);
listdir(path, indent + 2);
} else {
printf("%*s- %s\n", indent, "", entry->d_name);
}
}
closedir(dir);
}
int main(void) {
listdir(".", 0);
return 0;
}
int is_directory_we_want_to_list(const char *parent, char *name) {
struct stat st_buf;
if (!strcmp(".", name) || !strcmp("..", name))
return 0;
char *path = alloca(strlen(name) + strlen(parent) + 2);
sprintf(path, "%s/%s", parent, name);
stat(path, &st_buf);
return S_ISDIR(st_buf.st_mode);
}
int list(const char *name) {
DIR *dir = opendir(name);
struct dirent *ent;
while (ent = readdir(dir)) {
char *entry_name = ent->d_name;
printf("%s\n", entry_name);
if (is_directory_we_want_to_list(name, entry_name)) {
// You can consider using alloca instead.
char *next = malloc(strlen(name) + strlen(entry_name) + 2);
sprintf(next, "%s/%s", name, entry_name);
list(next);
free(next);
}
}
closedir(dir);
}
在此上下文中值得略读的头文件: stat.h 、 dirent.h 。 请记住,上面的代码不会检查可能发生的任何错误。
ftw.h 中定义的ftw
提供了一种完全不同的方法。
正如我在评论中提到的,我认为递归方法有两个固有的缺陷。
第一个缺陷是对打开文件的限制。 这个限制对深度遍历施加了限制。 如果有足够多的子文件夹,递归方法就会中断。 (请参阅有关堆栈溢出的编辑)
第二个缺陷有点微妙。 递归方法使得很难测试硬链接。 如果文件夹树是循环的(由于硬链接),递归方法将中断(希望没有堆栈溢出)。 (请参阅有关硬链接的编辑)
但是,通过用单个文件描述符和链表替换递归来避免这些问题非常简单。
我认为这不是学校项目,递归是可选的。
这是一个示例应用程序。
使用a.out ./
查看文件夹树。
我为宏和东西道歉......我通常使用内联函数,但我认为如果代码全部在一个函数中会更容易理解。
#include <dirent.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/types.h>
int main(int argc, char const *argv[]) {
/* print use instruction unless a folder name was given */
if (argc < 2)
fprintf(stderr,
"\nuse:\n"
" %s <directory>\n"
"for example:\n"
" %s ./\n\n",
argv[0], argv[0]),
exit(0);
/*************** a small linked list macro implementation ***************/
typedef struct list_s {
struct list_s *next;
struct list_s *prev;
} list_s;
#define LIST_INIT(name) \
{ .next = &name, .prev = &name }
#define LIST_PUSH(dest, node) \
do { \
(node)->next = (dest)->next; \
(node)->prev = (dest); \
(node)->next->prev = (node); \
(dest)->next = (node); \
} while (0);
#define LIST_POP(list, var) \
if ((list)->next == (list)) { \
var = NULL; \
} else { \
var = (list)->next; \
(list)->next = var->next; \
var->next->prev = var->prev; \
}
/*************** a record (file / folder) item type ***************/
typedef struct record_s {
/* this is a flat processing queue. */
list_s queue;
/* this will list all queued and processed folders (cyclic protection) */
list_s folders;
/* this will list all the completed items (siblings and such) */
list_s list;
/* unique ID */
ino_t ino;
/* name length */
size_t len;
/* name string */
char name[];
} record_s;
/* take a list_s pointer and convert it to the record_s pointer */
#define NODE2RECORD(node, list_name) \
((record_s *)(((uintptr_t)(node)) - \
((uintptr_t) & ((record_s *)0)->list_name)))
/* initializes a new record */
#define RECORD_INIT(name) \
(record_s){.queue = LIST_INIT((name).queue), \
.folders = LIST_INIT((name).folders), \
.list = LIST_INIT((name).list)}
/*************** the actual code ***************/
record_s records = RECORD_INIT(records);
record_s *pos, *item;
list_s *tmp;
DIR *dir;
struct dirent *entry;
/* initialize the root folder record and add it to the queue */
pos = malloc(sizeof(*pos) + strlen(argv[1]) + 2);
*pos = RECORD_INIT(*pos);
pos->len = strlen(argv[1]);
memcpy(pos->name, argv[1], pos->len);
if (pos->name[pos->len - 1] != '/')
pos->name[pos->len++] = '/';
pos->name[pos->len] = 0;
/* push to queue, but also push to list (first item processed) */
LIST_PUSH(&records.queue, &pos->queue);
LIST_PUSH(&records.list, &pos->list);
/* as long as the queue has items to be processed, do so */
while (records.queue.next != &records.queue) {
/* pop queued item */
LIST_POP(&records.queue, tmp);
/* collect record to process */
pos = NODE2RECORD(tmp, queue);
/* add record to the processed folder list */
LIST_PUSH(&records.folders, &pos->folders);
/* process the folder and add all folder data to current list */
dir = opendir(pos->name);
if (!dir)
continue;
while ((entry = readdir(dir)) != NULL) {
/* create new item, copying it's path data and unique ID */
item = malloc(sizeof(*item) + pos->len + entry->d_namlen + 2);
*item = RECORD_INIT(*item);
item->len = pos->len + entry->d_namlen;
memcpy(item->name, pos->name, pos->len);
memcpy(item->name + pos->len, entry->d_name, entry->d_namlen);
item->name[item->len] = 0;
item->ino = entry->d_ino;
/* add item to the list, right after the `pos` item */
LIST_PUSH(&pos->list, &item->list);
/* unless it's a folder, we're done. */
if (entry->d_type != DT_DIR)
continue;
/* test for '.' and '..' */
if (entry->d_name[0] == '.' &&
(entry->d_name[1] == 0 ||
(entry->d_name[1] == '.' && entry->d_name[2] == 0)))
continue;
/* add folder marker */
item->name[item->len++] = '/';
item->name[item->len] = 0;
/* test for cyclic processing */
list_s *t = records.folders.next;
while (t != &records.folders) {
if (NODE2RECORD(t, folders)->ino == item->ino) {
/* we already processed this folder! */
break; /* this breaks from the small loop... */
}
t = t->next;
}
if (t != &records.folders)
continue; /* if we broke from the small loop, entry is done */
/* item is a new folder, add to queue */
LIST_PUSH(&records.queue, &item->queue);
}
closedir(dir);
}
/*************** Printing the results and cleaning up ***************/
while (records.list.next != &records.list) {
/* pop list item */
LIST_POP(&records.list, tmp);
/* collect and process record */
pos = NODE2RECORD(tmp, list);
fwrite(pos->name, pos->len, 1, stderr);
fwrite("\n", 1, 1, stderr);
/* free node */
free(pos);
}
return 0;
}
编辑
@Stargateur 在评论中提到递归代码可能会在达到打开文件限制之前溢出堆栈。
尽管我看不出堆栈溢出有什么好处,但只要进程在调用时不接近文件限制,这种评估就可能是正确的。
@Stargateur 在评论中提到的另一点是递归代码的深度受到最大子目录数量(ext4 文件系统上的 64000)的限制,并且硬链接极不可能(因为到文件夹的硬链接不是在 Linux/Unix 上允许)。
如果代码在 Linux 上运行(根据问题是这样),这是个好消息,所以这个问题不是真正的问题(除非在 macOS 或 Windows 上运行代码)......尽管有 64K 子文件夹在递归中可能会彻底打开堆栈。
话虽如此,none recursive 选项仍然具有优势,例如能够轻松地对处理的项目数量添加限制以及能够缓存结果。
聚苯乙烯
根据评论,这是不检查循环层次结构的代码的非递归版本。 它更快并且应该足够安全,可以在不允许硬链接到文件夹的 Linux 机器上使用。
#include <dirent.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/types.h>
int main(int argc, char const *argv[]) {
/* print use instruction unless a folder name was given */
if (argc < 2)
fprintf(stderr,
"\nuse:\n"
" %s <directory>\n"
"for example:\n"
" %s ./\n\n",
argv[0], argv[0]),
exit(0);
/*************** a small linked list macro implementation ***************/
typedef struct list_s {
struct list_s *next;
struct list_s *prev;
} list_s;
#define LIST_INIT(name) \
{ .next = &name, .prev = &name }
#define LIST_PUSH(dest, node) \
do { \
(node)->next = (dest)->next; \
(node)->prev = (dest); \
(node)->next->prev = (node); \
(dest)->next = (node); \
} while (0);
#define LIST_POP(list, var) \
if ((list)->next == (list)) { \
var = NULL; \
} else { \
var = (list)->next; \
(list)->next = var->next; \
var->next->prev = var->prev; \
}
/*************** a record (file / folder) item type ***************/
typedef struct record_s {
/* this is a flat processing queue. */
list_s queue;
/* this will list all the completed items (siblings and such) */
list_s list;
/* unique ID */
ino_t ino;
/* name length */
size_t len;
/* name string */
char name[];
} record_s;
/* take a list_s pointer and convert it to the record_s pointer */
#define NODE2RECORD(node, list_name) \
((record_s *)(((uintptr_t)(node)) - \
((uintptr_t) & ((record_s *)0)->list_name)))
/* initializes a new record */
#define RECORD_INIT(name) \
(record_s){.queue = LIST_INIT((name).queue), .list = LIST_INIT((name).list)}
/*************** the actual code ***************/
record_s records = RECORD_INIT(records);
record_s *pos, *item;
list_s *tmp;
DIR *dir;
struct dirent *entry;
/* initialize the root folder record and add it to the queue */
pos = malloc(sizeof(*pos) + strlen(argv[1]) + 2);
*pos = RECORD_INIT(*pos);
pos->len = strlen(argv[1]);
memcpy(pos->name, argv[1], pos->len);
if (pos->name[pos->len - 1] != '/')
pos->name[pos->len++] = '/';
pos->name[pos->len] = 0;
/* push to queue, but also push to list (first item processed) */
LIST_PUSH(&records.queue, &pos->queue);
LIST_PUSH(&records.list, &pos->list);
/* as long as the queue has items to be processed, do so */
while (records.queue.next != &records.queue) {
/* pop queued item */
LIST_POP(&records.queue, tmp);
/* collect record to process */
pos = NODE2RECORD(tmp, queue);
/* process the folder and add all folder data to current list */
dir = opendir(pos->name);
if (!dir)
continue;
while ((entry = readdir(dir)) != NULL) {
/* create new item, copying it's path data and unique ID */
item = malloc(sizeof(*item) + pos->len + entry->d_namlen + 2);
*item = RECORD_INIT(*item);
item->len = pos->len + entry->d_namlen;
memcpy(item->name, pos->name, pos->len);
memcpy(item->name + pos->len, entry->d_name, entry->d_namlen);
item->name[item->len] = 0;
item->ino = entry->d_ino;
/* add item to the list, right after the `pos` item */
LIST_PUSH(&pos->list, &item->list);
/* unless it's a folder, we're done. */
if (entry->d_type != DT_DIR)
continue;
/* test for '.' and '..' */
if (entry->d_name[0] == '.' &&
(entry->d_name[1] == 0 ||
(entry->d_name[1] == '.' && entry->d_name[2] == 0)))
continue;
/* add folder marker */
item->name[item->len++] = '/';
item->name[item->len] = 0;
/* item is a new folder, add to queue */
LIST_PUSH(&records.queue, &item->queue);
}
closedir(dir);
}
/*************** Printing the results and cleaning up ***************/
while (records.list.next != &records.list) {
/* pop list item */
LIST_POP(&records.list, tmp);
/* collect and process record */
pos = NODE2RECORD(tmp, list);
fwrite(pos->name, pos->len, 1, stderr);
fwrite("\n", 1, 1, stderr);
/* free node */
free(pos);
}
return 0;
}
这是一个递归的简化版本,但使用的堆栈空间要少得多:
#include <errno.h>
#include <stdio.h>
#include <string.h>
#include <sys/types.h>
#include <unistd.h>
#include <dirent.h>
void listdir(char *path, size_t size) {
DIR *dir;
struct dirent *entry;
size_t len = strlen(path);
if (!(dir = opendir(path))) {
fprintf(stderr, "path not found: %s: %s\n",
path, strerror(errno));
return;
}
puts(path);
while ((entry = readdir(dir)) != NULL) {
char *name = entry->d_name;
if (entry->d_type == DT_DIR) {
if (!strcmp(name, ".") || !strcmp(name, ".."))
continue;
if (len + strlen(name) + 2 > size) {
fprintf(stderr, "path too long: %s/%s\n", path, name);
} else {
path[len] = '/';
strcpy(path + len + 1, name);
listdir(path, size);
path[len] = '\0';
}
} else {
printf("%s/%s\n", path, name);
}
}
closedir(dir);
}
int main(void) {
char path[1024] = ".";
listdir(path, sizeof path);
return 0;
}
在我的系统上,它的输出与find .
输出完全相同find .
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.