*
*/
-#define VERSION_NUMBER "0.9"
+#define VERSION_NUMBER "1.1"
#define _BSD_SOURCE
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/param.h>
+#include <sys/wait.h>
#include <dirent.h>
#include <stdlib.h>
#include <stdio.h>
#include <locale.h>
#include <getopt.h>
#include <fcntl.h>
-#ifdef WITH_MD5
-#include <openssl/md5.h>
-#endif
/* 1M really helps compared to 64k */
#define READ_BUFFER_SIZE (1024 * 1024)
two files with same inode will
always be false */
-#ifdef WITH_MD5
-int use_md5 = 0; /* 1 means we keep an MD5 signature for each file */
-#endif
+int sort_by_time = 0; /* 1 means to sort files in each group according
+ to the modification time */
+
+char *command_to_exec = 0; /* the name of the command to exec for each
+ group of identical files */
/********************************************************************/
void *safe_malloc(size_t n) {
void *p = malloc(n);
if (!p && n != 0) {
- printf("Can not allocate memory: %s\n", strerror(errno));
+ fprintf(stderr,
+ "finddup: Can not allocate memory: %s\n", strerror(errno));
exit(EXIT_FAILURE);
}
return p;
struct file_node *next;
char *name;
size_t size;
+ time_t atime, mtime, ctime;
ino_t inode;
int group_id; /* one per identical file content */
int dir_id; /* 1 for DIR1, and 2 for DIR2 */
-#ifdef WITH_MD5
- int md5_computed;
- unsigned char md5[MD5_DIGEST_LENGTH];
-#endif
};
void file_list_delete(struct file_node *head) {
char *buffer1, char *buffer2) {
int fd1, fd2, s1, s2;
-#ifdef WITH_MD5
- MD5_CTX c1, c2;
-
- if(use_md5) {
- if(f1->md5_computed && f2->md5_computed) {
- if(!memcmp(f1->md5, f2->md5, MD5_DIGEST_LENGTH)) {
- return 0;
- }
- } else {
- if(!f1->md5_computed) {
- MD5_Init(&c1);
- }
- if(!f2->md5_computed) {
- MD5_Init(&c2);
- }
- }
- }
-#endif
-
fd1 = open(f1->name, O_RDONLY);
fd2 = open(f2->name, O_RDONLY);
if(s1 == 0) {
close(fd1);
close(fd2);
-#ifdef WITH_MD5
- if(use_md5) {
- if(!f1->md5_computed) {
- MD5_Final(f1->md5, &c1);
- f1->md5_computed = 1;
- }
- if(!f2->md5_computed) {
- MD5_Final(f2->md5, &c2);
- f2->md5_computed = 1;
- }
- }
-#endif
return 1;
} else {
if(memcmp(buffer1, buffer2, s1)) {
+ /* printf("size_to_read = %d\n", size_to_read); */
close(fd1);
close(fd2);
return 0;
}
-#ifdef WITH_MD5
- if(use_md5) {
- if(!f1->md5_computed) {
- MD5_Update(&c1, buffer1, s1);
- }
- if(!f2->md5_computed) {
- MD5_Update(&c2, buffer2, s2);
- }
- }
-#endif
}
} else {
fprintf(stderr,
- "Different read size without error on files of same size.\n");
+ "finddup: Different read size without error on files of same size.\n");
exit(EXIT_FAILURE);
}
}
if(fd1 < 0) {
fprintf(stderr,
- "Can not open \"%s\" error: %s\n",
+ "finddup: Can not open \"%s\" error: %s\n",
f1->name,
strerror(errno));
}
if(fd2 < 0) {
fprintf(stderr,
- "Can not open \"%s\" error: %s\n",
+ "finddup: Can not open \"%s\" error: %s\n",
f2->name,
strerror(errno));
}
/**********************************************************************/
-struct file_node *scan_directory(struct file_node *tail, const char *name) {
+struct file_node *scan_directory_rec(struct file_node *tail, const char *name) {
DIR *dir;
struct dirent *dir_e;
struct stat sb;
char subname[PATH_MAX + 1];
if(lstat(name, &sb) != 0) {
- fprintf(stderr, "Can not stat \"%s\": %s\n", name, strerror(errno));
+ fprintf(stderr, "finddup: Can not stat \"%s\": %s\n", name, strerror(errno));
exit(EXIT_FAILURE);
}
while((dir_e = readdir(dir))) {
if(!ignore_entry(dir_e->d_name)) {
snprintf(subname, PATH_MAX, "%s/%s", name, dir_e->d_name);
- tail = scan_directory(tail, subname);
+ tail = scan_directory_rec(tail, subname);
}
}
closedir(dir);
tmp->next = tail;
tmp->name = strdup(name);
tmp->size = sb.st_size;
+ tmp->atime = sb.st_atime;
+ tmp->mtime = sb.st_mtime;
+ tmp->ctime = sb.st_ctime;
tmp->inode = sb.st_ino;
tmp->group_id = -1;
tmp->dir_id = -1;
-#ifdef WITH_MD5
- tmp->md5_computed = 0;
-#endif
tail = tmp;
}
}
return tail;
}
+struct file_node *scan_directory(struct file_node *tail, const char *name) {
+ struct file_node *result;
+ int length;
+
+ if(show_progress) {
+ fprintf(stderr, "Scanning '%s' ... ", name);
+ }
+
+ result = scan_directory_rec(tail, name);
+ length = file_list_length(result);
+
+ if(show_progress) {
+ fprintf(stderr, "done (%d file%s).\n",
+ length, (length > 1 ? "s" : ""));
+ }
+
+
+ return result;
+}
+
void print_file(struct file_node *node) {
char tmp[PATH_MAX + 1];
if(show_realpaths) {
printf("%s\n", tmp);
}
} else {
- printf("Can not get the realpath of \"%s\": %s\n",
- node->name,
- strerror(errno));
+ fprintf(stderr,
+ "finddup: Can not get the realpath of \"%s\": %s\n",
+ node->name,
+ strerror(errno));
exit(EXIT_FAILURE);
}
} else {
} else if((*f1)->group_id > (*f2)->group_id) {
return 1;
} else {
- if((*f1)->dir_id < (*f2)->dir_id) {
- return -1;
- } else if((*f1)->dir_id > (*f2)->dir_id) {
- return 1;
+ if(sort_by_time) {
+ if((*f1)->mtime < (*f2)->mtime) {
+ return -1;
+ } else if((*f1)->mtime > (*f2)->mtime) {
+ return 1;
+ } else {
+ return 0;
+ }
} else {
- return 0;
+ if((*f1)->dir_id < (*f2)->dir_id) {
+ return -1;
+ } else if((*f1)->dir_id > (*f2)->dir_id) {
+ return 1;
+ } else {
+ return 0;
+ }
}
}
}
+void exec_command(int nb, struct file_node **nodes) {
+ char **args;
+ int max_group_size = 0, group_size, m, n, status;
+ pid_t pid;
+
+ for(n = 0; n < nb; n++) {
+ if(n == 0 || nodes[n]->group_id != nodes[n-1]->group_id) {
+ group_size = 0;
+ }
+ group_size++;
+ if(group_size > max_group_size) {
+ max_group_size = group_size;
+ }
+ }
+
+ args = safe_malloc((max_group_size + 2) * sizeof(char *));
+ args[0] = command_to_exec;
+
+ n = 0;
+ while(n < nb) {
+ m = n;
+ while(n < nb && nodes[n]->group_id == nodes[m]->group_id) {
+ args[n - m + 1] = nodes[n]->name;
+ n++;
+ }
+ args[n - m + 1] = 0;
+ pid = fork();
+ if(pid < 0) {
+ } else if(pid == 0) {
+ if(execvp(command_to_exec, args) < 0) {
+ perror("execvp");
+ exit(EXIT_FAILURE);
+ }
+ } else {
+ while(wait(&status) != pid);
+ if(status > 0) {
+ exit(EXIT_FAILURE);
+ }
+ }
+ }
+
+ free(args);
+}
void print_result(struct file_node *list1, struct file_node *list2) {
struct file_node *node1, *node2;
qsort(nodes, nb, sizeof(struct file_node *), compare_nodes);
- for(n = 0; n < nb; n++) {
- if(!show_groups && n > 0 && nodes[n]->group_id != nodes[n-1]->group_id) {
- printf("\n");
+ if(command_to_exec) {
+ exec_command(nb, nodes);
+ } else {
+ for(n = 0; n < nb; n++) {
+ if(!show_groups && n > 0 && nodes[n]->group_id != nodes[n-1]->group_id) {
+ printf("\n");
+ }
+ print_file(nodes[n]);
}
- print_file(nodes[n]);
}
free(nodes);
};
void print_progress(struct progress_state *state) {
- int position, k;
+ int position, k, normalizer;
struct winsize win;
char buffer[PROGRESS_BUFFER_SIZE];
char *s;
+ normalizer = (state->nb_values > 1 ? state->nb_values - 1 : 1);
+
if(show_progress) {
/* We use the previous bar_width to compute the position, so that
we avoid doing too many ioctls */
- position = (state->bar_width * state->value) / (state->nb_values - 1);
+ position = (state->bar_width * state->value) / normalizer;
if(state->bar_width <= 0 || position != state->last_position) {
if(!ioctl (STDERR_FILENO, TIOCGWINSZ, (char *) &win)) {
/* Something weird is going on if the previous test is wrong */
} else {
state->bar_width = win.ws_col - 7;
}
- position = (state->bar_width * state->value) / (state->nb_values - 1);
+ position = (state->bar_width * state->value) / normalizer;
state->last_position = position;
s = buffer;
for(k = 0; k < position; k++) {
/* We need four % because of the fprintf that follows */
sprintf(s, " [%3d%%%%]\r",
- (100 * state->value) / (state->nb_values - 1));
+ (100 * state->value) / normalizer);
fprintf(stderr, buffer);
}
struct progress_state progress_state;
int not_in, found;
int nb_groups, nb_nodes;
- int list1_length, list2_length, previous_progress;
+ int list1_length, previous_progress;
char *buffer1 = safe_malloc(sizeof(char) * READ_BUFFER_SIZE);
char *buffer2 = safe_malloc(sizeof(char) * READ_BUFFER_SIZE);
not_in = 0;
- if(show_progress) {
- fprintf(stderr, "Scanning %s ... ", dirname1);
- }
-
list1 = scan_directory(0, dirname1);
-
list1_length = file_list_length(list1);
if(dirname2) {
} else if(strncmp(dirname2, "and:", 4) == 0) {
dirname2 += 4;
}
- if(show_progress) {
- fprintf(stderr, "%s ... ", dirname2);
- }
list2 = scan_directory(0, dirname2);
} else {
list2 = list1;
}
if(show_progress) {
- fprintf(stderr, "done.\n");
fprintf(stderr,
- "%s: %d file%s.\n",
- dirname1, list1_length, (list1_length > 1 ? "s" : ""));
- if(dirname2) {
- list2_length = file_list_length(list2);
- fprintf(stderr,
- "%s: %d file%s.\n",
- dirname2, list2_length, (list2_length > 1 ? "s" : ""));
- }
- fprintf(stderr, "Now looking for identical files.\n");
+ "Now looking for identical files (this may take a while).\n");
}
nb_groups = 0;
}
void usage(FILE *out) {
- fprintf(out, "Usage: finddup [OPTION]... DIR1 [[and:|not:]DIR2]\n");
+ fprintf(out, "Usage: finddup [OPTION]... [DIR1 [[and:|not:]DIR2]]\n");
fprintf(out, "Version %s (%s)\n", VERSION_NUMBER, UNAME);
- 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. The and: prefix is the default and should be used only if you have a directory starting with 'not:'\n");
+ fprintf(out, "Without DIR2, lists duplicated files found in DIR1, or the current directory if DIR1 is not provided. With DIR2, lists files common to both directories. With the not: prefix, lists files found in DIR1 which do not exist in DIR2. The and: prefix is the default and should be used only if you have a directory starting with 'not:'\n");
fprintf(out, "\n");
/* 01234567890123456789012345678901234567890123456789012345678901234567890123456789*/
+ fprintf(out, " -v, --version prints the version number and exit\n");
fprintf(out, " -h, --help show this help\n");
fprintf(out, " -d, --ignore-dots ignore dot files and directories\n");
fprintf(out, " -0, --ignore-empty ignore empty files\n");
fprintf(out, " -c, --hide-matchings do not show which files in DIR2 corresponds to\n");
fprintf(out, " those in DIR1\n");
fprintf(out, " -g, --no-group-ids do not show the file groups\n");
+ fprintf(out, " -t, --time-sort sort according to modification time in each group\n");
fprintf(out, " -p, --show-progress show progress\n");
fprintf(out, " -r, --real-paths show the real file paths\n");
fprintf(out, " -i, --same-inodes-are-different\n");
fprintf(out, " consider files with same inode as different\n");
-#ifdef WITH_MD5
- fprintf(out, " -m, --md5 use MD5 hashing\n");
-#endif
+ fprintf(out, " -e <command>, --exec <command>\n");
+ fprintf(out, " execute the provided command for each group of\n");
+ fprintf(out, " identical files, with their names as arguments\n");
fprintf(out, "\n");
fprintf(out, "Report bugs and comments to <francois@fleuret.org>.\n");
}
/**********************************************************************/
static struct option long_options[] = {
+ { "version", no_argument, 0, 'v' },
{ "help", no_argument, 0, 'h' },
{ "same-inodes-are-different", no_argument, 0, 'i' },
{ "real-paths", no_argument, 0, 'r' },
{ "hide-matchings", no_argument, 0, 'c' },
{ "no-group-ids", no_argument, 0, 'g' },
+ { "time-sort", no_argument, 0, 't' },
{ "ignore-dots", no_argument, 0, 'd' },
{ "ignore-empty", no_argument, 0, '0' },
{ "show-progress", no_argument, 0, 'p' },
- { "md5", no_argument, 0, 'm' },
+ { "exec", 1, 0, 'e' },
{ 0, 0, 0, 0 }
};
setlocale (LC_ALL, "");
- while ((c = getopt_long(argc, argv, "hircgd0pm",
+ while ((c = getopt_long(argc, argv, "vhircgtd0pme:",
long_options, NULL)) != -1) {
switch (c) {
+ case 'v':
+ printf("finddup version %s (%s)\n", VERSION_NUMBER, UNAME);
+ exit(EXIT_SUCCESS);
+ break;
+
case 'h':
usage(stdout);
exit(EXIT_SUCCESS);
show_groups = 0;
break;
+ case 't':
+ sort_by_time = 1;
+ break;
+
case 'p':
show_progress = 1;
break;
show_hits = 0;
break;
- case 'm':
-#ifdef WITH_MD5
- use_md5 = 1;
-#else
- fprintf(stderr,
- "finddup has not been compiled with MD5 hashing.\n");
- usage(stderr);
- exit(EXIT_FAILURE);
-#endif
+ case 'e':
+ if(command_to_exec != 0) {
+ free(command_to_exec);
+ }
+ command_to_exec = strdup(optarg);
break;
default:
} else if(optind + 1 == argc) {
same_inodes_are_different = 1;
start(argv[optind], 0);
+ } else if(optind == argc) {
+ same_inodes_are_different = 1;
+ start(".", 0);
} else {
usage(stderr);
exit(EXIT_FAILURE);