Merge three separate "fetch refs" functions
[gitweb.git] / fsck-cache.c
index 7dda9c3eb790ef57ee14c479a8730f0312dea744..8e21bf1327a60840273973c08216c1387aac95b8 100644 (file)
@@ -6,12 +6,17 @@
 #include "tree.h"
 #include "blob.h"
 #include "tag.h"
+#include "refs.h"
+#include "pack.h"
 
 #define REACHABLE 0x0001
 
 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];
 
 static void check_connectivity(void)
@@ -24,25 +29,27 @@ 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));
                        printf("              to %7s %s\n",
-                              obj->type, sha1_to_hex(refs->item->sha1));
+                              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;
                }
 
@@ -61,6 +68,9 @@ 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);
@@ -73,7 +83,7 @@ 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.
@@ -82,11 +92,18 @@ 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)
@@ -99,11 +116,41 @@ static int fsck_tree(struct tree *item)
                if (strchr(entry->name, '/'))
                        has_full_path = 1;
 
+               switch (entry->mode) {
+               /*
+                * Standard modes.. 
+                */
+               case S_IFREG | 0755:
+               case S_IFREG | 0644:
+               case S_IFLNK:
+               case S_IFDIR:
+                       break;
+               /*
+                * This is nonstandard, but we had a few of these
+                * early on when we honored the full set of mode
+                * bits..
+                */
+               case S_IFREG | 0664:
+                       break;
+               default:
+                       printf("tree %s has entry %o %s\n",
+                               sha1_to_hex(item->object.sha1),
+                               entry->mode, entry->name);
+               }
+
                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;
                        }
                }
 
@@ -111,7 +158,7 @@ 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));
        }
@@ -121,6 +168,8 @@ 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)
@@ -133,14 +182,17 @@ static int fsck_commit(struct commit *commit)
 
 static int fsck_tag(struct tag *tag)
 {
+       struct object *tagged = tag->tagged;
+
+       if (!tagged) {
+               printf("bad object in tag %s\n", sha1_to_hex(tag->object.sha1));
+               return -1;
+       }
        if (!show_tags)
                return 0;
 
-       printf("tagged %s %s",
-              tag->tagged->type,
-              sha1_to_hex(tag->tagged->sha1));
-       printf(" (%s) in %s\n",
-              tag->tag, sha1_to_hex(tag->object.sha1));
+       printf("tagged %s %s", tagged->type, sha1_to_hex(tagged->sha1));
+       printf(" (%s) in %s\n", tag->tag, sha1_to_hex(tag->object.sha1));
        return 0;
 }
 
@@ -252,10 +304,74 @@ static int fsck_dir(int i, char *path)
        return 0;
 }
 
+static int default_refs = 0;
+
+static int fsck_handle_ref(const char *refname, const unsigned char *sha1)
+{
+       struct object *obj;
+
+       obj = lookup_object(sha1);
+       if (!obj) {
+               if (!standalone && has_sha1_file(sha1)) {
+                       default_refs++;
+                       return 0; /* it is in a pack */
+               }
+               error("%s: invalid sha1 pointer %s", refname, sha1_to_hex(sha1));
+               /* We'll continue with the rest despite the error.. */
+               return 0;
+       }
+       default_refs++;
+       obj->used = 1;
+       mark_reachable(obj, REACHABLE);
+       return 0;
+}
+
+static void get_default_heads(void)
+{
+       for_each_ref(fsck_handle_ref);
+       if (!default_refs)
+               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();
+}
+
+static int fsck_head_link(void)
+{
+       int fd, count;
+       char hex[40];
+       unsigned char sha1[20];
+       static char path[PATH_MAX], link[PATH_MAX];
+       const char *git_dir = gitenv(GIT_DIR_ENVIRONMENT) ? : DEFAULT_GIT_DIR_ENVIRONMENT;
+
+       snprintf(path, sizeof(path), "%s/HEAD", git_dir);
+       if (readlink(path, link, sizeof(link)) < 0)
+               return error("HEAD is not a symlink");
+       if (strncmp("refs/heads/", link, 11))
+               return error("HEAD points to something strange (%s)", link);
+       fd = open(path, O_RDONLY);
+       if (fd < 0)
+               return error("HEAD: %s", strerror(errno));
+       count = read(fd, hex, sizeof(hex));
+       close(fd);
+       if (count < 0)
+               return error("HEAD: %s", strerror(errno));
+       if (count < 40 || get_sha1_hex(hex, sha1))
+               return error("HEAD: not a valid git pointer");
+       return 0;
+}
+
 int main(int argc, char **argv)
 {
        int i, heads;
-       char *sha1_dir;
 
        for (i = 1; i < argc; i++) {
                const char *arg = argv[i];
@@ -272,17 +388,56 @@ int main(int argc, char **argv)
                        show_root = 1;
                        continue;
                }
+               if (!strcmp(arg, "--cache")) {
+                       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] <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_head_link();
+       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++) {
+                       char namebuf[PATH_MAX];
+                       int namelen = alt_odb[j].name - alt_odb[j].base;
+                       memcpy(namebuf, alt_odb[j].base, namelen);
+                       namebuf[namelen - 1] = 0;
+                       fsck_object_dir(namebuf);
+               }
+               prepare_packed_git();
+               for (p = packed_git; p; p = p->next)
+                       /* verify gives error messages itself */
+                       verify_pack(p, 0);
+
+               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++) {
@@ -292,14 +447,12 @@ int main(int argc, char **argv)
                        continue;
 
                if (!get_sha1(arg, head_sha1)) {
-                       struct commit *commit = lookup_commit(head_sha1);
-                       struct object *obj;
+                       struct object *obj = lookup_object(head_sha1);
 
-                       /* Error is printed by lookup_commit(). */
-                       if (!commit)
+                       /* Error is printed by lookup_object(). */
+                       if (!obj)
                                continue;
 
-                       obj = &commit->object;
                        obj->used = 1;
                        mark_reachable(obj, REACHABLE);
                        heads++;
@@ -308,12 +461,28 @@ 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) {
-               if (show_unreachable) {
-                       fprintf(stderr, "unable to do reachability without a head\n");
-                       show_unreachable = 0; 
+               get_default_heads();
+               keep_cache_objects = 1;
+       }
+
+       if (keep_cache_objects) {
+               int i;
+               read_cache();
+               for (i = 0; i < active_nr; i++) {
+                       struct blob *blob = lookup_blob(active_cache[i]->sha1);
+                       struct object *obj;
+                       if (!blob)
+                               continue;
+                       obj = &blob->object;
+                       obj->used = 1;
+                       mark_reachable(obj, REACHABLE);
                }
-               fprintf(stderr, "expect dangling commits - potential heads - due to lack of head information\n");
        }
 
        check_connectivity();