X-Git-Url: https://www.fleuret.org/cgi-bin/gitweb/gitweb.cgi?p=finddup.git;a=blobdiff_plain;f=finddup.c;h=150279166903355e63d76c9f95a4e2a115aedb66;hp=d3e5a12bd9d04472657912c582449ee4646acdb8;hb=5d880a48149f85f6d5ac7942a696248eac0b8acc;hpb=85eb33af87f8e0cfdb08cb888adbd1e5cc02a540

diff --git a/finddup.c b/finddup.c
index d3e5a12..1502791 100644
--- a/finddup.c
+++ b/finddup.c
@@ -1,7 +1,8 @@
 
 /*
- *  finddup is a simple utility to display the files and directories
- *  according to their total disk occupancy.
+ *  finddup is a simple utility 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>
@@ -22,10 +23,13 @@
  *
  */
 
+#define VERSION_NUMBER "0.6"
+
 #define _BSD_SOURCE
 
 #include <sys/types.h>
 #include <sys/stat.h>
+#include <sys/param.h>
 #include <dirent.h>
 #include <stdlib.h>
 #include <stdio.h>
@@ -38,6 +42,7 @@
 #include <fcntl.h>
 
 #define BUFFER_SIZE 4096
+#define LARGE_BUFFER_SIZE 65536
 
 typedef int64_t size_sum_t;
 
@@ -46,19 +51,22 @@ typedef int64_t size_sum_t;
 int ignore_dotfiles = 0; /* 1 means ignore files and directories
                             starting with a dot */
 
-int forced_width = 0; /* -1 means no width limit, strictly positive
-                         means limit, 0 means not active */
+int ignore_empty_files = 0; /* 1 means ignore empty files */
+
+int show_realpaths = 0; /* 1 means ignore files and directories
+                            starting with a dot */
 
-int forced_height = 0; /* -1 means no height limit, strictly positive
-                           means limit, 0 means not active */
+int show_progress = 0; /* 1 means show a progress bar when we are in a
+                          tty */
 
-int fancy_size_display = 0; /* 1 means to use floating values with K, M and G
-                               as units */
+int show_hits = 1; /* 1 means to show the files from dir2
+                      corresponding to the ones from dir1 */
 
-int reverse_sorting = 0; /* 1 means to show the large ones first */
+int show_groups = 1; /* 1 means to show the group IDs when printing
+                        file names */
 
-int show_top = 0; /* 1 means to show the top of the sorted list
-                     instead of the bottom */
+int ignore_same_inode = 0; /* 1 means that comparison between two file
+                              with same inode will always be false */
 
 /********************************************************************/
 
@@ -97,148 +105,115 @@ void print_size_sum(size_sum_t s) {
   printf(a);
 }
 
-size_sum_t file_or_dir_size(const char *name) {
-  DIR *dir;
-  struct dirent *dir_e;
-  struct stat dummy;
-  size_sum_t result;
-  char subname[BUFFER_SIZE];
-
-  result = 0;
-
-  if(lstat(name, &dummy) != 0) {
-    printf("Can not stat %s: %s\n", name, strerror(errno));
-    exit(EXIT_FAILURE);
-  }
-
-  if(S_ISLNK(dummy.st_mode)) {
-    return 0;
-  }
-
-  dir = opendir(name);
-
-  if(dir) {
-    while((dir_e = readdir(dir))) {
-      if(!ignore_entry(dir_e->d_name)) {
-        snprintf(subname, BUFFER_SIZE, "%s/%s", name, dir_e->d_name);
-        result += file_or_dir_size(subname);
-      }
-    }
-    closedir(dir);
-  } else {
-    if(S_ISREG(dummy.st_mode)) {
-      result += dummy.st_size;
-    }
-  }
-
-  return result;
-}
-
 /**********************************************************************/
 
 struct file_with_size {
   char *filename;
   size_t size;
+  ino_t inode;
   struct file_with_size *next;
-  int error;
-};
-
-struct file_list {
-  struct file_with_size *head, *tail;
+  int group_id, dir_id;
 };
 
-void create(struct file_list *list, const char *name) {
-  list->head = safe_malloc(sizeof(struct file_with_size));
-  list->tail = 0;
-  list->head->filename = strdup(name);
-  list->head->size = file_or_dir_size(name);
-  list->head->next = 0;
+void file_list_delete(struct file_with_size *head) {
+  struct file_with_size *next;
+  while(head) {
+    next = head->next;
+    free(head->filename);
+    free(head);
+    head = next;
+  }
 }
 
-void destroy(struct file_with_size *node) {
-  struct file_with_size *next;
-  while(node) {
-    next = node->next;
-    free(node->filename);
-    free(node);
-    node = next;
+int file_list_length(struct file_with_size *head) {
+  int l = 0;
+  while(head) {
+    l++;
+    head = head->next;
   }
+  return l;
 }
 
 /**********************************************************************/
 
-int same_size(struct file_with_size *f1, struct file_with_size *f2) {
-  return f1->size == f2->size;
-}
-
 int same_content(struct file_with_size *f1, struct file_with_size *f2) {
-  int fd1, fd2, eq, s1, s2;
-  char buffer1[BUFFER_SIZE], buffer2[BUFFER_SIZE];
+  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);
-  eq = 1;
 
   if(fd1 >= 0 && fd2 >= 0) {
-    while(eq) {
-      s1 = read(fd1, buffer1, BUFFER_SIZE);
-      if(s1 < 0) {
-        f1->error = 1;
-        eq = 0;
-      }
-      s2 = read(fd2, buffer2, BUFFER_SIZE);
-      if(s2 < 0) {
-        f2->error = 1;
-        eq = 0;
+    while(1) {
+      s1 = read(fd1, buffer1, LARGE_BUFFER_SIZE);
+      s2 = read(fd2, buffer2, LARGE_BUFFER_SIZE);
+
+      if(s1 < 0 || s2 < 0) {
+        close(fd1);
+        close(fd2);
+        return 0;
       }
-      if(s1 > 0 && s2 > 0) {
-        if(s1 == s2) {
-          eq = (strncmp(buffer1, buffer2, s1) == 0);
+
+      if(s1 == s2) {
+        if(s1 == 0) {
+          close(fd1);
+          close(fd2);
+          return 1;
         } else {
-          fprintf(stderr,
-                  "Different read size without error on files of same size.\n");
-          exit(EXIT_FAILURE);
+          if(memcmp(buffer1, buffer2, s1)) {
+            close(fd1);
+            close(fd2);
+            return 0;
+          }
         }
+      } else {
+        fprintf(stderr,
+                "Different read size without error on files of same size.\n");
+        exit(EXIT_FAILURE);
       }
     }
-  }
-
-  if(fd1 >= 0) {
-    close(fd1);
   } else {
-    f1->error = 1;
-    eq = 0;
-  }
 
-  if(fd2 >= 0) {
-    close(fd2);
-  } else {
-    f2->error = 1;
-    eq = 0;
-  }
+    if(fd1 < 0) {
+      fprintf(stderr,
+              "Can not open \"%s\" error: %s\n",
+              f1->filename,
+              strerror(errno));
+    }
 
-  return eq;
+    if(fd2 < 0) {
+      fprintf(stderr,
+              "Can not open \"%s\" error: %s\n",
+              f2->filename,
+              strerror(errno));
+    }
+    exit(EXIT_FAILURE);
+  }
 }
 
 int same_files(struct file_with_size *f1, struct file_with_size *f2) {
-  return same_size(f1, f2) && same_content(f1, f2);
+  if(ignore_same_inode && f1->inode == f2->inode) {
+    return 0;
+  }
+  return f1->size == f2->size && same_content(f1, f2);
 }
 
 /**********************************************************************/
 
-struct file_with_size *scan_directory(struct file_with_size *tail, const char *name) {
+struct file_with_size *scan_directory(struct file_with_size *tail,
+                                      const char *name) {
   DIR *dir;
   struct dirent *dir_e;
-  struct stat dummy;
+  struct stat sb;
   struct file_with_size *tmp;
-  char subname[BUFFER_SIZE];
+  char subname[PATH_MAX + 1];
 
-  if(lstat(name, &dummy) != 0) {
-    fprintf(stderr, "Can not stat %s: %s\n", name, strerror(errno));
+  if(lstat(name, &sb) != 0) {
+    fprintf(stderr, "Can not stat \"%s\": %s\n", name, strerror(errno));
     exit(EXIT_FAILURE);
   }
 
-  if(S_ISLNK(dummy.st_mode)) {
+  if(S_ISLNK(sb.st_mode)) {
     return tail;
   }
 
@@ -247,40 +222,239 @@ struct file_with_size *scan_directory(struct file_with_size *tail, const char *n
   if(dir) {
     while((dir_e = readdir(dir))) {
       if(!ignore_entry(dir_e->d_name)) {
-        snprintf(subname, BUFFER_SIZE, "%s/%s", name, dir_e->d_name);
+        snprintf(subname, PATH_MAX, "%s/%s", name, dir_e->d_name);
         tail = scan_directory(tail, subname);
       }
     }
     closedir(dir);
   } else {
-    if(S_ISREG(dummy.st_mode)) {
-      tmp = safe_malloc(sizeof(struct file_with_size));
-      tmp->next = tail;
-      tmp->filename = strdup(name);
-      tmp->size = file_or_dir_size(name);
-      tail = tmp;
+    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->next = tail;
+          tmp->filename = strdup(name);
+          tmp->size = sb.st_size;
+          tmp->inode = sb.st_ino;
+          tmp->group_id = -1;
+          tmp->dir_id = -1;
+          tail = tmp;
+        }
+      }
     }
   }
 
   return tail;
 }
 
+void print_file(struct file_with_size *node) {
+  char tmp[PATH_MAX + 1];
+  if(show_realpaths) {
+    if(show_groups) {
+      realpath(node->filename, tmp);
+      printf("%d %s\n", node->group_id, tmp);
+    } else {
+      realpath(node->filename, tmp);
+      printf("%s\n", tmp);
+    }
+  } else {
+    if(show_groups) {
+      printf("%d %s\n", node->group_id, node->filename);
+    } else {
+      printf("%s\n", node->filename);
+    }
+  }
+}
+
+int compare_nodes(const void *x1, const void *x2) {
+  const struct file_with_size **f1, **f2;
+
+  f1 = (const struct file_with_size **) x1;
+  f2 = (const struct file_with_size **) x2;
+
+  if((*f1)->group_id < (*f2)->group_id) {
+    return -1;
+  } 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;
+    } 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;
+  int nb, n;
+
+  nb = 0;
+  for(node1 = list1; node1; node1 = node1->next) {
+    if(node1->group_id >= 0) { nb++; }
+  }
+
+  if(list2) {
+    for(node2 = list2; node2; node2 = node2->next) {
+      if(node2->group_id >= 0) { nb++; }
+    }
+  }
+
+  nodes = safe_malloc(nb * sizeof(struct file_with_size *));
+
+  n = 0;
+  for(node1 = list1; node1; node1 = node1->next) {
+    if(node1->group_id >= 0) {
+      nodes[n++] = node1;
+    }
+  }
+
+  if(list2) {
+    for(node2 = list2; node2; node2 = node2->next) {
+      if(node2->group_id >= 0) {
+        nodes[n++] = node2;
+      }
+    }
+  }
+
+  qsort(nodes, nb, sizeof(struct file_with_size *), 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);
+}
+
 void start(const char *dirname1, const char *dirname2) {
   struct file_with_size *list1, *list2;
-  struct file_with_size *node;
+  struct file_with_size *node1, *node2;
+  int not_in, found;
+  int k, p, pp, l1, n;
+
+  not_in = 0;
+
+  if(show_progress) {
+    fprintf(stderr, "Scanning %s ... ", dirname1);
+  }
+
   list1 = scan_directory(0, dirname1);
-  list2 = scan_directory(0, dirname2);
 
-  for(node = list1; node; node = node->next) {
-    printf("list1: %s\n", node->filename);
+  if(dirname2) {
+    if(strncmp(dirname2, "not:", 4) == 0) {
+      not_in = 1;
+      dirname2 += 4;
+    }
+    if(show_progress) {
+      fprintf(stderr, "%s ... ", dirname2);
+    }
+    list2 = scan_directory(0, dirname2);
+  } else {
+    list2 = list1;
+  }
+
+  if(show_progress) {
+    fprintf(stderr, "done.\n");
+  }
+
+  k = 0;
+  pp = -1;
+  n = 0;
+  l1 = file_list_length(list1);
+
+  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++;
+      }
+
+      found = 0;
+
+      for(node2 = list2; !found && node2; node2 = node2->next) {
+        if(same_files(node1, node2)) {
+          found = 1;
+        }
+      }
+
+      if(!found) {
+        if(show_realpaths) {
+          printf("%s\n", realpath(node1->filename, 0));
+        } else {
+          printf("%s\n", node1->filename);
+        }
+      }
+    }
+
+  } 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++;
+      }
+
+      for(node2 = list2; node2; node2 = node2->next) {
+        if(node1->group_id < 0 || node2->group_id < 0) {
+          if(same_files(node1, node2)) {
+            if(node1->group_id < 0) {
+              if(node2->group_id >= 0) {
+                node1->group_id = node2->group_id;
+              } else {
+                node1->group_id = k;
+                node1->dir_id = 1;
+                k++;
+              }
+            }
+            if(node2->group_id < 0) {
+              node2->group_id = node1->group_id;
+              node2->dir_id = 2;
+            }
+          }
+        }
+      }
+    }
   }
 
-  for(node = list2; node; node = node->next) {
-    printf("list2: %s\n", node->filename);
+  if(dirname2) {
+    print_result(list1, list2);
+    file_list_delete(list1);
+    file_list_delete(list2);
+  } else {
+    print_result(list1, 0);
+    file_list_delete(list1);
   }
+}
 
-  destroy(list1);
-  destroy(list2);
+void print_help(FILE *out) {
+  fprintf(out, "Usage: finddup [OPTION]... DIR1 [[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, "\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");
+  fprintf(out, "\n");
+  fprintf(out, "Report bugs and comments to <francois@fleuret.org>\n");
 }
 
 /**********************************************************************/
@@ -294,28 +468,58 @@ int main(int argc, char **argv) {
   setlocale (LC_ALL, "");
 
   while (1) {
-    c = getopt(argc, argv, "h");
+    c = getopt(argc, argv, "hircgd0p");
     if (c == -1)
       break;
 
     switch (c) {
 
     case 'h':
-      printf("Usage: finddup [OPTION]... [FILE]...\n");
-      printf("Report bugs and comments to <francois@fleuret.org>\n");
+      print_help(stdout);
       exit(EXIT_SUCCESS);
 
       break;
 
+    case 'd':
+      ignore_dotfiles = 1;
+      break;
+
+    case '0':
+      ignore_empty_files = 1;
+      break;
+
+    case 'r':
+      show_realpaths = 1;
+      break;
+
+    case 'i':
+      ignore_same_inode = 1;
+      break;
+
+    case 'g':
+      show_groups = 0;
+      break;
+
+    case 'p':
+      show_progress = 1;
+      break;
+
+    case 'c':
+      show_hits = 0;
+      break;
+
     default:
       exit(EXIT_FAILURE);
     }
   }
 
-  if(optind + 1 < argc) {
+  if(optind + 2 == argc) {
     start(argv[optind], argv[optind + 1]);
+  } else if(optind + 1 == argc) {
+    ignore_same_inode = 1;
+    start(argv[optind], 0);
   } else {
-    fprintf(stderr, "%s [OPTIONS] <dir1> <dir2>\n", argv[0]);
+    print_help(stderr);
     exit(EXIT_FAILURE);
   }