3 * finddup is a simple utility to find duplicated files, files common
4 * to several directories, or files present in one directory and not
7 * Copyright (c) 2010 Francois Fleuret
8 * Written by Francois Fleuret <francois@fleuret.org>
10 * This file is part of finddup.
12 * finddup is free software: you can redistribute it and/or modify it
13 * under the terms of the GNU General Public License version 3 as
14 * published by the Free Software Foundation.
16 * finddup is distributed in the hope that it will be useful, but WITHOUT
17 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
18 * or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public
19 * License for more details.
21 * You should have received a copy of the GNU General Public License
22 * along with finddup. If not, see <http://www.gnu.org/licenses/>.
26 #define VERSION_NUMBER "0.6"
30 #include <sys/types.h>
32 #include <sys/param.h>
39 #include <sys/ioctl.h>
44 /* 1M really helps compared to 64k */
45 #define READ_BUFFER_SIZE (1024 * 1024)
47 typedef int64_t size_sum_t;
49 /* Yeah, global variables! */
51 int ignore_dotfiles = 0; /* 1 means ignore files and directories
52 starting with a dot */
54 int ignore_empty_files = 0; /* 1 means ignore empty files */
56 int show_realpaths = 0; /* 1 means ignore files and directories
57 starting with a dot */
59 int show_progress = 0; /* 1 means show a progress bar when we are in a
62 int show_hits = 1; /* 1 means to show the files from dir2
63 corresponding to the ones from dir1 */
65 int show_groups = 1; /* 1 means to show the group IDs when printing
68 int ignore_same_inode = 0; /* 1 means that comparison between two file
69 with same inode will always be false */
71 int tty_width = -1; /* Positive value means what width to use to show
74 /********************************************************************/
76 /* malloc with error checking. */
78 void *safe_malloc(size_t n) {
81 printf("Can not allocate memory: %s\n", strerror(errno));
87 /********************************************************************/
89 int ignore_entry(const char *name) {
91 strcmp(name, ".") == 0 ||
92 strcmp(name, "..") == 0 ||
93 (ignore_dotfiles && name[0] == '.');
96 void print_size_sum(size_sum_t s) {
98 char *a = tmp + sizeof(tmp)/sizeof(char);
111 /**********************************************************************/
113 struct file_with_size {
117 struct file_with_size *next;
118 int group_id, dir_id;
121 void file_list_delete(struct file_with_size *head) {
122 struct file_with_size *next;
125 free(head->filename);
131 int file_list_length(struct file_with_size *head) {
140 /**********************************************************************/
142 int same_content(struct file_with_size *f1, struct file_with_size *f2,
143 char *buffer1, char *buffer2) {
144 int fd1, fd2, s1, s2;
146 fd1 = open(f1->filename, O_RDONLY);
147 fd2 = open(f2->filename, O_RDONLY);
149 if(fd1 >= 0 && fd2 >= 0) {
151 s1 = read(fd1, buffer1, READ_BUFFER_SIZE);
152 s2 = read(fd2, buffer2, READ_BUFFER_SIZE);
154 if(s1 < 0 || s2 < 0) {
166 if(memcmp(buffer1, buffer2, s1)) {
174 "Different read size without error on files of same size.\n");
182 "Can not open \"%s\" error: %s\n",
189 "Can not open \"%s\" error: %s\n",
197 int same_files(struct file_with_size *f1, struct file_with_size *f2,
198 char *buffer1, char *buffer2) {
199 if(ignore_same_inode && f1->inode == f2->inode) {
202 return f1->size == f2->size && same_content(f1, f2, buffer1, buffer2);
205 /**********************************************************************/
207 struct file_with_size *scan_directory(struct file_with_size *tail,
210 struct dirent *dir_e;
212 struct file_with_size *tmp;
213 char subname[PATH_MAX + 1];
215 if(lstat(name, &sb) != 0) {
216 fprintf(stderr, "Can not stat \"%s\": %s\n", name, strerror(errno));
220 if(S_ISLNK(sb.st_mode)) {
227 while((dir_e = readdir(dir))) {
228 if(!ignore_entry(dir_e->d_name)) {
229 snprintf(subname, PATH_MAX, "%s/%s", name, dir_e->d_name);
230 tail = scan_directory(tail, subname);
235 if(S_ISREG(sb.st_mode)) {
236 if(!ignore_entry(name)) {
237 if(!ignore_empty_files || sb.st_size > 0) {
238 tmp = safe_malloc(sizeof(struct file_with_size));
240 tmp->filename = strdup(name);
241 tmp->size = sb.st_size;
242 tmp->inode = sb.st_ino;
254 void print_file(struct file_with_size *node) {
255 char tmp[PATH_MAX + 1];
258 realpath(node->filename, tmp);
259 printf("%d %s\n", node->group_id, tmp);
261 realpath(node->filename, tmp);
266 printf("%d %s\n", node->group_id, node->filename);
268 printf("%s\n", node->filename);
273 int compare_nodes(const void *x1, const void *x2) {
274 const struct file_with_size **f1, **f2;
276 f1 = (const struct file_with_size **) x1;
277 f2 = (const struct file_with_size **) x2;
279 if((*f1)->group_id < (*f2)->group_id) {
281 } else if((*f1)->group_id > (*f2)->group_id) {
284 if((*f1)->dir_id < (*f2)->dir_id) {
286 } else if((*f1)->dir_id > (*f2)->dir_id) {
295 void print_result(struct file_with_size *list1, struct file_with_size *list2) {
296 struct file_with_size *node1, *node2;
297 struct file_with_size **nodes;
301 for(node1 = list1; node1; node1 = node1->next) {
302 if(node1->group_id >= 0) { nb++; }
306 for(node2 = list2; node2; node2 = node2->next) {
307 if(node2->group_id >= 0) { nb++; }
311 nodes = safe_malloc(nb * sizeof(struct file_with_size *));
314 for(node1 = list1; node1; node1 = node1->next) {
315 if(node1->group_id >= 0) {
321 for(node2 = list2; node2; node2 = node2->next) {
322 if(node2->group_id >= 0) {
328 qsort(nodes, nb, sizeof(struct file_with_size *), compare_nodes);
330 for(n = 0; n < nb; n++) {
331 if(!show_groups && n > 0 && nodes[n]->group_id != nodes[n-1]->group_id) {
334 print_file(nodes[n]);
340 void print_progress(int max, int n, int *pp) {
343 if(show_progress && tty_width > 0) {
344 width = tty_width - 7;
345 p = (width * n) / max;
347 for(k = 0; k < p; k++) {
348 fprintf(stderr, "+");
350 for(; k < width; k++) {
351 fprintf(stderr, "-");
355 fprintf(stderr, " [% 3d%%]\r", p);
360 void start(const char *dirname1, const char *dirname2) {
361 struct file_with_size *list1, *list2;
362 struct file_with_size *node1, *node2;
366 char *buffer1 = safe_malloc(sizeof(char) * READ_BUFFER_SIZE);
367 char *buffer2 = safe_malloc(sizeof(char) * READ_BUFFER_SIZE);
372 fprintf(stderr, "Scanning %s ... ", dirname1);
375 list1 = scan_directory(0, dirname1);
378 if(strncmp(dirname2, "not:", 4) == 0) {
380 /* We should show groups even in that mode. However they are not
381 properly calculated for now, so we force it off. */
386 fprintf(stderr, "%s ... ", dirname2);
388 list2 = scan_directory(0, dirname2);
394 fprintf(stderr, "done.\n");
400 l1 = file_list_length(list1);
403 for(node1 = list1; node1; node1 = node1->next) {
404 print_progress(l1, n, &pp);
409 for(node2 = list2; !found && node2; node2 = node2->next) {
410 if(same_files(node1, node2, buffer1, buffer2)) {
417 printf("%s\n", realpath(node1->filename, 0));
419 printf("%s\n", node1->filename);
425 for(node1 = list1; node1; node1 = node1->next) {
426 print_progress(l1, n, &pp);
429 for(node2 = list2; node2; node2 = node2->next) {
430 if(node1->group_id < 0 || node2->group_id < 0) {
431 if(same_files(node1, node2, buffer1, buffer2)) {
432 if(node1->group_id < 0) {
433 if(node2->group_id >= 0) {
434 node1->group_id = node2->group_id;
441 if(node2->group_id < 0) {
442 node2->group_id = node1->group_id;
452 fprintf(stderr, "\n");
456 print_result(list1, list2);
457 file_list_delete(list1);
458 file_list_delete(list2);
460 print_result(list1, 0);
461 file_list_delete(list1);
468 void print_help(FILE *out) {
469 fprintf(out, "Usage: finddup [OPTION]... DIR1 [[not:]DIR2]\n");
470 fprintf(out, "Version %s (%s)\n", VERSION_NUMBER, UNAME);
471 fprintf(out, "Without DIR2, lists duplicated files found in DIR1. With DIR2, lists files common to both directories. With the not: prefix, lists files found in DIR1 which do not exist in DIR2.\n");
473 fprintf(out, " -h show this help\n");
474 fprintf(out, " -d ignore dot files and directories\n");
475 fprintf(out, " -0 ignore empty files\n");
476 fprintf(out, " -c do not show which files in DIR2 corresponds to those in DIR1\n");
477 fprintf(out, " -g do not show the file groups\n");
478 fprintf(out, " -p show progress\n");
479 fprintf(out, " -r show the real file paths\n");
480 fprintf(out, " -i consider files with same inode as different\n");
482 fprintf(out, "Report bugs and comments to <francois@fleuret.org>\n");
485 /**********************************************************************/
487 int main(int argc, char **argv) {
489 struct file_with_size *root;
494 setlocale (LC_ALL, "");
497 c = getopt(argc, argv, "hircgd0p");
514 ignore_empty_files = 1;
522 ignore_same_inode = 1;
543 isatty(STDOUT_FILENO) &&
544 !ioctl (STDOUT_FILENO, TIOCGWINSZ, (char *) &win)) {
545 tty_width = win.ws_col;
548 if(optind + 2 == argc) {
549 start(argv[optind], argv[optind + 1]);
550 } else if(optind + 1 == argc) {
551 ignore_same_inode = 1;
552 start(argv[optind], 0);