X-Git-Url: https://fleuret.org/cgi-bin/gitweb/gitweb.cgi?p=finddup.git;a=blobdiff_plain;f=finddup.c;h=8f2b6c9e07396f279d8fd343b52c216da7acf911;hp=20673723dbec75ae20ede7ad3240fcbe015938e2;hb=ab7b6e26f35ac1dfc88d9bf1e09dd289a30ea782;hpb=e41b7a6614371af4f9ac5e988a6b468d1d1a0e60

diff --git a/finddup.c b/finddup.c
index 2067372..8f2b6c9 100644
--- a/finddup.c
+++ b/finddup.c
@@ -1,8 +1,8 @@
 
 /*
- *  finddup is a simple utility find duplicated files, files common to
- *  several directories, or files present in one directory and not in
- *  another one.
+ *  finddup is a simple utility to find duplicated files, files common
+ *  to several directories, or files present in one directory and not
+ *  in another one.
  *
  *  Copyright (c) 2010 Francois Fleuret
  *  Written by Francois Fleuret <francois@fleuret.org>
@@ -23,7 +23,7 @@
  *
  */
 
-#define VERSION_NUMBER "0.6"
+#define VERSION_NUMBER "0.9"
 
 #define _BSD_SOURCE
 
@@ -41,8 +41,10 @@
 #include <getopt.h>
 #include <fcntl.h>
 
-#define BUFFER_SIZE 4096
-#define LARGE_BUFFER_SIZE 65536
+/* 1M really helps compared to 64k */
+#define READ_BUFFER_SIZE (1024 * 1024)
+
+#define PROGRESS_BUFFER_SIZE 1024
 
 typedef int64_t size_sum_t;
 
@@ -65,8 +67,9 @@ int show_hits = 1; /* 1 means to show the files from dir2
 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 same_inodes_are_different = 0; /* 1 means that comparison between
+                                      two files with same inode will
+                                      always be false */
 
 /********************************************************************/
 
@@ -87,45 +90,31 @@ int ignore_entry(const char *name) {
   return
     strcmp(name, ".") == 0 ||
     strcmp(name, "..") == 0 ||
-    (ignore_dotfiles && name[0] == '.');
-}
-
-void print_size_sum(size_sum_t s) {
-  char tmp[100];
-  char *a = tmp + sizeof(tmp)/sizeof(char);
-  *(--a) = '\0';
-  if(s) {
-    while(s) {
-      *(--a) = s%10 + '0';
-      s /= 10;
-    }
-  } else {
-    *(--a) = '0';
-  }
-  printf(a);
+    (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;
+  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++;
@@ -136,17 +125,17 @@ int file_list_length(struct file_with_size *head) {
 
 /**********************************************************************/
 
-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;
-  char buffer1[LARGE_BUFFER_SIZE], buffer2[LARGE_BUFFER_SIZE];
 
-  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) {
-      s1 = read(fd1, buffer1, LARGE_BUFFER_SIZE);
-      s2 = read(fd2, buffer2, LARGE_BUFFER_SIZE);
+      s1 = read(fd1, buffer1, READ_BUFFER_SIZE);
+      s2 = read(fd2, buffer2, READ_BUFFER_SIZE);
 
       if(s1 < 0 || s2 < 0) {
         close(fd1);
@@ -161,6 +150,7 @@ int same_content(struct file_with_size *f1, struct file_with_size *f2) {
           return 1;
         } else {
           if(memcmp(buffer1, buffer2, s1)) {
+            /* printf("size_to_read = %d\n", size_to_read); */
             close(fd1);
             close(fd2);
             return 0;
@@ -177,35 +167,37 @@ int same_content(struct file_with_size *f1, struct file_with_size *f2) {
     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) {
-  if(ignore_same_inode && f1->inode == f2->inode) {
+int same_files(struct file_node *f1, struct file_node *f2,
+               char *buffer1, char *buffer2) {
+  if(same_inodes_are_different && f1->inode == f2->inode) {
     return 0;
   }
-  return f1->size == f2->size && same_content(f1, f2);
+
+  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) {
@@ -231,12 +223,13 @@ struct file_with_size *scan_directory(struct file_with_size *tail,
     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;
+          tmp->dir_id = -1;
           tail = tmp;
         }
       }
@@ -246,44 +239,55 @@ struct file_with_size *scan_directory(struct file_with_size *tail,
   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(show_groups) {
-      realpath(node->filename, tmp);
-      printf("%d %s\n", node->group_id, tmp);
+    if(realpath(node->name, tmp)) {
+      if(show_groups) {
+        printf("%d %s\n", node->group_id, tmp);
+      } else {
+        printf("%s\n", tmp);
+      }
     } else {
-      realpath(node->filename, tmp);
-      printf("%s\n", tmp);
+      printf("Can not get the realpath of \"%s\": %s\n",
+             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;
   } else if((*f1)->group_id > (*f2)->group_id) {
     return 1;
   } 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 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;
@@ -297,7 +301,7 @@ void print_result(struct file_with_size *list1, struct file_with_size *list2) {
     }
   }
 
-  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) {
@@ -314,103 +318,166 @@ void print_result(struct file_with_size *list1, struct file_with_size *list2) {
     }
   }
 
-  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) {
+      printf("\n");
+    }
     print_file(nodes[n]);
   }
 
   free(nodes);
 }
 
+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);
+      }
+    }
+  }
+}
+
 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 k, p, pp, l1, n;
+  int nb_groups, nb_nodes;
+  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;
+      /* groups are not computed in the not: mode */
+      show_groups = 0;
+      dirname2 += 4;
+    } else if(strncmp(dirname2, "and:", 4) == 0) {
       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");
   }
 
-  k = 0;
-  pp = -1;
-  n = 0;
-  l1 = file_list_length(list1);
+  nb_groups = 0;
+  previous_progress = -1;
+  nb_nodes = 0;
+
+  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) {
-      if(show_progress) {
-        p = (100 * n)/l1;
-        if(p > pp) {
-          fprintf(stderr, "%d%%\n", p);
-          pp = p;
-        }
-        n++;
-      }
+      progress_state.value = nb_nodes;
+      print_progress(&progress_state);
+      nb_nodes++;
 
       found = 0;
 
       for(node2 = list2; !found && node2; node2 = node2->next) {
-        if(same_files(node1, node2)) {
+        if(same_files(node1, node2, buffer1, buffer2)) {
           found = 1;
         }
       }
 
       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) {
-      if(show_progress) {
-        p = (100 * n)/l1;
-        if(p > pp) {
-          fprintf(stderr, "%d%%\n", p);
-          pp = p;
-        }
-        n++;
-      }
+      progress_state.value = nb_nodes;
+      print_progress(&progress_state);
+      nb_nodes++;
 
       for(node2 = list2; node2; node2 = node2->next) {
         if(node1->group_id < 0 || node2->group_id < 0) {
-          if(same_files(node1, node2)) {
+          if(same_files(node1, node2, buffer1, buffer2)) {
             if(node1->group_id < 0) {
               if(node2->group_id >= 0) {
                 node1->group_id = node2->group_id;
               } else {
-                node1->group_id = k;
-                k++;
+                node1->group_id = nb_groups;
+                node1->dir_id = 1;
+                nb_groups++;
               }
             }
             if(node2->group_id < 0) {
               node2->group_id = node1->group_id;
+              node2->dir_id = 2;
             }
           }
         }
@@ -418,6 +485,10 @@ void start(const char *dirname1, const char *dirname2) {
     }
   }
 
+  if(show_progress) {
+    fprintf(stderr, "\n");
+  }
+
   if(dirname2) {
     print_result(list1, list2);
     file_list_delete(list1);
@@ -426,44 +497,56 @@ void start(const char *dirname1, const char *dirname2) {
     print_result(list1, 0);
     file_list_delete(list1);
   }
+
+  free(buffer1);
+  free(buffer2);
 }
 
-void print_help(FILE *out) {
-  fprintf(out, "Usage: finddup [OPTION]... DIR1 [[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.\n");
+  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, "\n");
-  fprintf(out, "   -h   show this help\n");
-  fprintf(out, "   -d   ignore dot files and directories\n");
-  fprintf(out, "   -0   ignore empty files\n");
-  fprintf(out, "   -c   do not show which files in DIR2 corresponds to those in DIR1\n");
-  fprintf(out, "   -g   do not show the file groups\n");
-  fprintf(out, "   -p   show progress\n");
-  fprintf(out, "   -r   show the real file paths\n");
-  fprintf(out, "   -i   consider files with same inode as different\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, "\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;
-
-  root = 0;
 
   setlocale (LC_ALL, "");
 
-  while (1) {
-    c = getopt(argc, argv, "hrcgd0p");
-    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;
@@ -481,7 +564,7 @@ int main(int argc, char **argv) {
       break;
 
     case 'i':
-      ignore_same_inode = 1;
+      same_inodes_are_different = 1;
       break;
 
     case 'g':
@@ -497,17 +580,22 @@ int main(int argc, char **argv) {
       break;
 
     default:
+      usage(stderr);
       exit(EXIT_FAILURE);
     }
   }
 
+  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 {
-    print_help(stderr);
+    usage(stderr);
     exit(EXIT_FAILURE);
   }