[PATCH] Update fsck-cache (take 2)
[git.git] / fsck-cache.c
index abdec92..eae73cc 100644 (file)
@@ -12,6 +12,8 @@
 static int show_root = 0;
 static int show_tags = 0;
 static int show_unreachable = 0;
+static int standalone = 0;
+static int check_full = 0;
 static int keep_cache_objects = 0; 
 static unsigned char head_sha1[20];
 
@@ -25,12 +27,17 @@ static void check_connectivity(void)
                struct object_list *refs;
 
                if (!obj->parsed) {
-                       printf("missing %s %s\n", obj->type, sha1_to_hex(obj->sha1));
+                       if (!standalone && has_sha1_file(obj->sha1))
+                               ; /* it is in pack */
+                       else
+                               printf("missing %s %s\n",
+                                      obj->type, sha1_to_hex(obj->sha1));
                        continue;
                }
 
                for (refs = obj->refs; refs; refs = refs->next) {
-                       if (refs->item->parsed)
+                       if (refs->item->parsed ||
+                           (!standalone && has_sha1_file(refs->item->sha1)))
                                continue;
                        printf("broken link from %7s %s\n",
                               obj->type, sha1_to_hex(obj->sha1));
@@ -38,12 +45,9 @@ static void check_connectivity(void)
                               refs->item->type, sha1_to_hex(refs->item->sha1));
                }
 
-               /* Don't bother with tag reachability. */
-               if (obj->type == tag_type)
-                       continue;
-
                if (show_unreachable && !(obj->flags & REACHABLE)) {
-                       printf("unreachable %s %s\n", obj->type, sha1_to_hex(obj->sha1));
+                       printf("unreachable %s %s\n",
+                              obj->type, sha1_to_hex(obj->sha1));
                        continue;
                }
 
@@ -62,6 +66,9 @@ static void check_connectivity(void)
  * So a directory called "a" is ordered _after_ a file
  * called "a.c", because "a/" sorts after "a.c".
  */
+#define TREE_UNORDERED (-1)
+#define TREE_HAS_DUPS  (-2)
+
 static int verify_ordered(struct tree_entry_list *a, struct tree_entry_list *b)
 {
        int len1 = strlen(a->name);
@@ -74,7 +81,7 @@ static int verify_ordered(struct tree_entry_list *a, struct tree_entry_list *b)
        if (cmp < 0)
                return 0;
        if (cmp > 0)
-               return -1;
+               return TREE_UNORDERED;
 
        /*
         * Ok, the first <len> characters are the same.
@@ -83,11 +90,18 @@ static int verify_ordered(struct tree_entry_list *a, struct tree_entry_list *b)
         */
        c1 = a->name[len];
        c2 = b->name[len];
+       if (!c1 && !c2)
+               /*
+                * git-write-tree used to write out a nonsense tree that has
+                * entries with the same name, one blob and one tree.  Make
+                * sure we do not have duplicate entries.
+                */
+               return TREE_HAS_DUPS;
        if (!c1 && a->directory)
                c1 = '/';
        if (!c2 && b->directory)
                c2 = '/';
-       return c1 < c2 ? 0 : -1;
+       return c1 < c2 ? 0 : TREE_UNORDERED;
 }
 
 static int fsck_tree(struct tree *item)
@@ -123,10 +137,18 @@ static int fsck_tree(struct tree *item)
                }
 
                if (last) {
-                       if (verify_ordered(last, entry) < 0) {
+                       switch (verify_ordered(last, entry)) {
+                       case TREE_UNORDERED:
                                fprintf(stderr, "tree %s not ordered\n",
                                        sha1_to_hex(item->object.sha1));
                                return -1;
+                       case TREE_HAS_DUPS:
+                               fprintf(stderr, "tree %s has duplicate entries for '%s'\n",
+                                       sha1_to_hex(item->object.sha1),
+                                       entry->name);
+                               return -1;
+                       default:
+                               break;
                        }
                }
 
@@ -134,7 +156,7 @@ static int fsck_tree(struct tree *item)
        }
 
        if (has_full_path) {
-               fprintf(stderr, "warning: fsck-cache: tree %s "
+               fprintf(stderr, "warning: git-fsck-cache: tree %s "
                        "has full pathnames in it\n", 
                        sha1_to_hex(item->object.sha1));
        }
@@ -144,6 +166,8 @@ static int fsck_tree(struct tree *item)
 
 static int fsck_commit(struct commit *commit)
 {
+       free(commit->buffer);
+       commit->buffer = NULL;
        if (!commit->tree)
                return -1;
        if (!commit->parents && show_root)
@@ -278,10 +302,93 @@ static int fsck_dir(int i, char *path)
        return 0;
 }
 
+static int read_sha1_reference(const char *path)
+{
+       char hexname[60];
+       unsigned char sha1[20];
+       int fd = open(path, O_RDONLY), len;
+       struct object *obj;
+
+       if (fd < 0)
+               return -1;
+
+       len = read(fd, hexname, sizeof(hexname));
+       close(fd);
+       if (len < 40)
+               return -1;
+
+       if (get_sha1_hex(hexname, sha1) < 0)
+               return -1;
+
+       obj = lookup_object(sha1);
+       if (!obj) {
+               if (!standalone && has_sha1_file(sha1))
+                       return 0; /* it is in pack */
+               return error("%s: invalid sha1 pointer %.40s", path, hexname);
+       }
+
+       obj->used = 1;
+       mark_reachable(obj, REACHABLE);
+       return 0;
+}
+
+static int find_file_objects(const char *base, const char *name)
+{
+       int baselen = strlen(base);
+       int namelen = strlen(name);
+       char *path = xmalloc(baselen + namelen + 2);
+       struct stat st;
+
+       memcpy(path, base, baselen);
+       path[baselen] = '/';
+       memcpy(path + baselen + 1, name, namelen+1);
+       if (stat(path, &st) < 0)
+               return 0;
+
+       /*
+        * Recurse into directories
+        */
+       if (S_ISDIR(st.st_mode)) {
+               int count = 0;
+               DIR *dir = opendir(path);
+               if (dir) {
+                       struct dirent *de;
+                       while ((de = readdir(dir)) != NULL) {
+                               if (de->d_name[0] == '.')
+                                       continue;
+                               count += find_file_objects(path, de->d_name);
+                       }
+                       closedir(dir);
+               }
+               return count;
+       }
+       if (S_ISREG(st.st_mode))
+               return read_sha1_reference(path) == 0;
+       return 0;
+}
+
+static void get_default_heads(void)
+{
+       char *git_dir = gitenv(GIT_DIR_ENVIRONMENT) ? : DEFAULT_GIT_DIR_ENVIRONMENT;
+       int count = find_file_objects(git_dir, "refs");
+       if (!count)
+               die("No default references");
+}
+
+static void fsck_object_dir(const char *path)
+{
+       int i;
+       for (i = 0; i < 256; i++) {
+               static char dir[4096];
+               sprintf(dir, "%s/%02x", path, i);
+               fsck_dir(i, dir);
+       }
+       fsck_sha1_list();
+}
+
 int main(int argc, char **argv)
 {
        int i, heads;
-       char *sha1_dir;
 
        for (i = 1; i < argc; i++) {
                const char *arg = argv[i];
@@ -302,17 +409,45 @@ int main(int argc, char **argv)
                        keep_cache_objects = 1;
                        continue;
                }
+               if (!strcmp(arg, "--standalone")) {
+                       standalone = 1;
+                       continue;
+               }
+               if (!strcmp(arg, "--full")) {
+                       check_full = 1;
+                       continue;
+               }
                if (*arg == '-')
-                       usage("fsck-cache [--tags] [[--unreachable] [--cache] <head-sha1>*]");
+                       usage("git-fsck-cache [--tags] [[--unreachable] [--cache] [--standalone | --full] <head-sha1>*]");
        }
 
-       sha1_dir = getenv(DB_ENVIRONMENT) ? : DEFAULT_DB_ENVIRONMENT;
-       for (i = 0; i < 256; i++) {
-               static char dir[4096];
-               sprintf(dir, "%s/%02x", sha1_dir, i);
-               fsck_dir(i, dir);
+       if (standalone && check_full)
+               die("Only one of --standalone or --full can be used.");
+       if (standalone)
+               unsetenv("GIT_ALTERNATE_OBJECT_DIRECTORIES");
+
+       fsck_object_dir(get_object_directory());
+       if (check_full) {
+               int j;
+               struct packed_git *p;
+               prepare_alt_odb();
+               for (j = 0; alt_odb[j].base; j++) {
+                       alt_odb[j].name[-1] = 0; /* was slash */
+                       fsck_object_dir(alt_odb[j].base);
+                       alt_odb[j].name[-1] = '/';
+               }
+               prepare_packed_git();
+               for (p = packed_git; p; p = p->next) {
+                       int num = num_packed_objects(p);
+                       for (i = 0; i < num; i++) {
+                               unsigned char sha1[20];
+                               nth_packed_object_sha1(p, i, sha1);
+                               if (fsck_sha1(sha1) < 0)
+                                       fprintf(stderr, "bad sha1 entry '%s'\n", sha1_to_hex(sha1));
+
+                       }
+               }
        }
-       fsck_sha1_list();
 
        heads = 0;
        for (i = 1; i < argc; i++) {
@@ -336,6 +471,16 @@ int main(int argc, char **argv)
                error("expected sha1, got %s", arg);
        }
 
+       /*
+        * If we've not been given any explicit head information, do the
+        * default ones from .git/refs. We also consider the index file
+        * in this case (ie this implies --cache).
+        */
+       if (!heads) {
+               get_default_heads();
+               keep_cache_objects = 1;
+       }
+
        if (keep_cache_objects) {
                int i;
                read_cache();
@@ -350,15 +495,6 @@ int main(int argc, char **argv)
                }
        }
 
-       if (!heads && !keep_cache_objects) {
-               if (show_unreachable) {
-                       fprintf(stderr, "unable to do reachability without a head nor --cache\n");
-                       show_unreachable = 0; 
-               }
-               if (!heads)
-                       fprintf(stderr, "expect dangling commits - potential heads - due to lack of head information\n");
-       }
-
        check_connectivity();
        return 0;
 }