*
*/
-#define VERSION_NUMBER "0.7"
+#define VERSION_NUMBER "0.9"
#define _BSD_SOURCE
/* 1M really helps compared to 64k */
#define READ_BUFFER_SIZE (1024 * 1024)
+#define PROGRESS_BUFFER_SIZE 1024
+
typedef int64_t size_sum_t;
/* Yeah, global variables! */
int show_groups = 1; /* 1 means to show the group IDs when printing
file names */
-int ignore_same_inode = 0; /* 1 means that comparison between two file
- with same inode will always be false */
-
-int tty_width = -1; /* Positive value means what width to use to show
- the progress bar */
+int same_inodes_are_different = 0; /* 1 means that comparison between
+ two files with same inode will
+ always be false */
/********************************************************************/
return
strcmp(name, ".") == 0 ||
strcmp(name, "..") == 0 ||
- (ignore_dotfiles && name[0] == '.');
+ (ignore_dotfiles && name[0] == '.' && name[1] != '/');
}
/**********************************************************************/
-struct file_with_size {
- char *filename;
+struct file_node {
+ struct file_node *next;
+ char *name;
size_t size;
ino_t inode;
- struct file_with_size *next;
- int group_id, dir_id;
+ int group_id; /* one per identical file content */
+ int dir_id; /* 1 for DIR1, and 2 for DIR2 */
};
-void file_list_delete(struct file_with_size *head) {
- struct file_with_size *next;
+void file_list_delete(struct file_node *head) {
+ struct file_node *next;
while(head) {
next = head->next;
- free(head->filename);
+ free(head->name);
free(head);
head = next;
}
}
-int file_list_length(struct file_with_size *head) {
+int file_list_length(struct file_node *head) {
int l = 0;
while(head) {
l++;
/**********************************************************************/
-int same_content(struct file_with_size *f1, struct file_with_size *f2,
+int same_content(struct file_node *f1, struct file_node *f2,
char *buffer1, char *buffer2) {
int fd1, fd2, s1, s2;
- fd1 = open(f1->filename, O_RDONLY);
- fd2 = open(f2->filename, O_RDONLY);
+ fd1 = open(f1->name, O_RDONLY);
+ fd2 = open(f2->name, O_RDONLY);
if(fd1 >= 0 && fd2 >= 0) {
while(1) {
return 1;
} else {
if(memcmp(buffer1, buffer2, s1)) {
+ /* printf("size_to_read = %d\n", size_to_read); */
close(fd1);
close(fd2);
return 0;
if(fd1 < 0) {
fprintf(stderr,
"Can not open \"%s\" error: %s\n",
- f1->filename,
+ f1->name,
strerror(errno));
}
if(fd2 < 0) {
fprintf(stderr,
"Can not open \"%s\" error: %s\n",
- f2->filename,
+ f2->name,
strerror(errno));
}
+
exit(EXIT_FAILURE);
}
}
-int same_files(struct file_with_size *f1, struct file_with_size *f2,
+int same_files(struct file_node *f1, struct file_node *f2,
char *buffer1, char *buffer2) {
- if(ignore_same_inode && f1->inode == f2->inode) {
+ if(same_inodes_are_different && f1->inode == f2->inode) {
return 0;
}
+
return f1->size == f2->size && same_content(f1, f2, buffer1, buffer2);
}
/**********************************************************************/
-struct file_with_size *scan_directory(struct file_with_size *tail,
- const char *name) {
+struct file_node *scan_directory(struct file_node *tail, const char *name) {
DIR *dir;
struct dirent *dir_e;
struct stat sb;
- struct file_with_size *tmp;
+ struct file_node *tmp;
char subname[PATH_MAX + 1];
if(lstat(name, &sb) != 0) {
if(S_ISREG(sb.st_mode)) {
if(!ignore_entry(name)) {
if(!ignore_empty_files || sb.st_size > 0) {
- tmp = safe_malloc(sizeof(struct file_with_size));
+ tmp = safe_malloc(sizeof(struct file_node));
tmp->next = tail;
- tmp->filename = strdup(name);
+ tmp->name = strdup(name);
tmp->size = sb.st_size;
tmp->inode = sb.st_ino;
tmp->group_id = -1;
return tail;
}
-void print_file(struct file_with_size *node) {
+void print_file(struct file_node *node) {
char tmp[PATH_MAX + 1];
if(show_realpaths) {
- if(realpath(node->filename, tmp)) {
+ if(realpath(node->name, tmp)) {
if(show_groups) {
printf("%d %s\n", node->group_id, tmp);
} else {
}
} else {
printf("Can not get the realpath of \"%s\": %s\n",
- node->filename,
+ node->name,
strerror(errno));
exit(EXIT_FAILURE);
}
} else {
if(show_groups) {
- printf("%d %s\n", node->group_id, node->filename);
+ printf("%d %s\n", node->group_id, node->name);
} else {
- printf("%s\n", node->filename);
+ printf("%s\n", node->name);
}
}
}
int compare_nodes(const void *x1, const void *x2) {
- const struct file_with_size **f1, **f2;
+ const struct file_node **f1, **f2;
- f1 = (const struct file_with_size **) x1;
- f2 = (const struct file_with_size **) x2;
+ f1 = (const struct file_node **) x1;
+ f2 = (const struct file_node **) x2;
if((*f1)->group_id < (*f2)->group_id) {
return -1;
}
-void print_result(struct file_with_size *list1, struct file_with_size *list2) {
- struct file_with_size *node1, *node2;
- struct file_with_size **nodes;
+void print_result(struct file_node *list1, struct file_node *list2) {
+ struct file_node *node1, *node2;
+ struct file_node **nodes;
int nb, n;
nb = 0;
}
}
- nodes = safe_malloc(nb * sizeof(struct file_with_size *));
+ nodes = safe_malloc(nb * sizeof(struct file_node *));
n = 0;
for(node1 = list1; node1; node1 = node1->next) {
}
}
- qsort(nodes, nb, sizeof(struct file_with_size *), compare_nodes);
+ 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) {
free(nodes);
}
-void print_progress(int max, int n, int *pp) {
- int p, k;
- int width;
- if(show_progress && tty_width > 0) {
- width = tty_width - 7;
- p = (width * n) / (max - 1);
- if(p > *pp) {
- for(k = 0; k < p; k++) {
- fprintf(stderr, "+");
- }
- for(; k < width; k++) {
- fprintf(stderr, "-");
+struct progress_state {
+ int bar_width;
+ int nb_values, value;
+ int last_position;
+};
+
+void print_progress(struct progress_state *state) {
+ int position, k;
+ struct winsize win;
+ char buffer[PROGRESS_BUFFER_SIZE];
+ char *s;
+
+ 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);
+ 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 */
+ if(win.ws_col >= PROGRESS_BUFFER_SIZE - 3) {
+ state->bar_width = PROGRESS_BUFFER_SIZE - 10;
+ } else {
+ state->bar_width = win.ws_col - 7;
+ }
+ position = (state->bar_width * state->value) / (state->nb_values - 1);
+ state->last_position = position;
+ s = buffer;
+ for(k = 0; k < position; k++) {
+ *(s++) = '+';
+ }
+ for(; k < state->bar_width; k++) {
+ *(s++) = '-';
+ }
+
+ /* We need four % because of the fprintf that follows */
+ sprintf(s, " [%3d%%%%]\r",
+ (100 * state->value) / (state->nb_values - 1));
+
+ fprintf(stderr, buffer);
}
- *pp = p;
- p = (100 * n) / (max - 1);
- fprintf(stderr, " [%3d%%]\r", p);
}
}
}
void start(const char *dirname1, const char *dirname2) {
- struct file_with_size *list1, *list2;
- struct file_with_size *node1, *node2;
+ struct file_node *list1, *list2;
+ struct file_node *node1, *node2;
+ struct progress_state progress_state;
int not_in, found;
int nb_groups, nb_nodes;
- int list1_length, previous_progress;
+ int list1_length, list2_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);
+ fprintf(stderr, "Scanning '%s' ... ", dirname1);
}
list1 = scan_directory(0, dirname1);
+ list1_length = file_list_length(list1);
+
+ if(show_progress) {
+ fprintf(stderr, "done (%d file%s).\n",
+ list1_length, (list1_length > 1 ? "s" : ""));
+ }
+
if(dirname2) {
if(strncmp(dirname2, "not:", 4) == 0) {
not_in = 1;
dirname2 += 4;
}
if(show_progress) {
- fprintf(stderr, "%s ... ", dirname2);
+ fprintf(stderr, "Scanning '%s' ... ", dirname2);
}
list2 = scan_directory(0, dirname2);
+ if(show_progress) {
+ list2_length = file_list_length(list2);
+ fprintf(stderr, "done (%d file%s).\n",
+ list2_length, (list2_length > 1 ? "s" : ""));
+ }
} else {
list2 = list1;
}
if(show_progress) {
- fprintf(stderr, "done.\n");
+ fprintf(stderr, "Now looking for identical files (this may take a while).\n");
}
nb_groups = 0;
previous_progress = -1;
nb_nodes = 0;
- list1_length = file_list_length(list1);
+
+ progress_state.bar_width = -1;
+ progress_state.last_position = -1;
+ progress_state.nb_values = list1_length;
if(not_in) {
for(node1 = list1; node1; node1 = node1->next) {
- print_progress(list1_length, nb_nodes, &previous_progress);
+ progress_state.value = nb_nodes;
+ print_progress(&progress_state);
nb_nodes++;
found = 0;
if(!found) {
if(show_realpaths) {
- printf("%s\n", realpath(node1->filename, 0));
+ printf("%s\n", realpath(node1->name, 0));
} else {
- printf("%s\n", node1->filename);
+ printf("%s\n", node1->name);
}
}
}
} else {
for(node1 = list1; node1; node1 = node1->next) {
- print_progress(list1_length, nb_nodes, &previous_progress);
+ progress_state.value = nb_nodes;
+ print_progress(&progress_state);
nb_nodes++;
for(node2 = list2; node2; node2 = node2->next) {
free(buffer2);
}
-void print_help(FILE *out) {
- fprintf(out, "Usage: finddup [OPTION]... DIR1 [[and:|not:]DIR2]\n");
+void usage(FILE *out) {
+ 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");
- fprintf(out, " -h, --help\n");
- fprintf(out, " show this help\n");
- fprintf(out, " -d, --ignore-dots\n");
- fprintf(out, " ignore dot files and directories\n");
- fprintf(out, " -0, --ignore-empty\n");
- fprintf(out, " ignore empty files\n");
- fprintf(out, " -c, --hide-matchings\n");
- fprintf(out, " do not show which files in DIR2 corresponds to those in DIR1\n");
- fprintf(out, " -g, --no-group-ids\n");
- fprintf(out, " do not show the file groups\n");
- fprintf(out, " -p, --show-progress\n");
- fprintf(out, " show progress\n");
- fprintf(out, " -r, --real-paths\n");
- fprintf(out, " show the real file paths\n");
+ /* 01234567890123456789012345678901234567890123456789012345678901234567890123456789*/
+ 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, " -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");
+ fprintf(out, " consider files with same inode as different\n");
fprintf(out, "\n");
- fprintf(out, "Report bugs and comments to <francois@fleuret.org>\n");
+ fprintf(out, "Report bugs and comments to <francois@fleuret.org>.\n");
}
/**********************************************************************/
+static struct option long_options[] = {
+ { "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' },
+ { "ignore-dots", no_argument, 0, 'd' },
+ { "ignore-empty", no_argument, 0, '0' },
+ { "show-progress", no_argument, 0, 'p' },
+ { 0, 0, 0, 0 }
+};
+
int main(int argc, char **argv) {
int c;
- struct file_with_size *root;
- struct winsize win;
-
- root = 0;
setlocale (LC_ALL, "");
- while (1) {
- int option_index = 0;
- static struct option long_options[] = {
- { "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' },
- { "ignore-dots", no_argument, 0, 'd' },
- { "ignore-empty", no_argument, 0, '0' },
- { "show-progress", no_argument, 0, 'p' },
- { 0, 0, 0, 0 }
- };
-
- c = getopt_long(argc, argv, "hircgd0p",
- long_options, &option_index);
- if (c == -1)
- break;
-
+ while ((c = getopt_long(argc, argv, "hircgd0pm",
+ long_options, NULL)) != -1) {
switch (c) {
case 'h':
- print_help(stdout);
+ usage(stdout);
exit(EXIT_SUCCESS);
break;
break;
case 'i':
- ignore_same_inode = 1;
+ same_inodes_are_different = 1;
break;
case 'g':
break;
default:
+ usage(stderr);
exit(EXIT_FAILURE);
}
}
- if(show_progress &&
- isatty(STDOUT_FILENO) &&
- !ioctl (STDOUT_FILENO, TIOCGWINSZ, (char *) &win)) {
- tty_width = win.ws_col;
+ if(!isatty(STDERR_FILENO)) {
+ show_progress = 0;
}
if(optind + 2 == argc) {
start(argv[optind], argv[optind + 1]);
} else if(optind + 1 == argc) {
- ignore_same_inode = 1;
+ same_inodes_are_different = 1;
start(argv[optind], 0);
+ } else if(optind == argc) {
+ same_inodes_are_different = 1;
+ start(".", 0);
} else {
- print_help(stderr);
+ usage(stderr);
exit(EXIT_FAILURE);
}