match-trees.c

   1 #include "cache.h"
   2 #include "tree.h"
   3 #include "tree-walk.h"
   4
   5 static int score_missing(unsigned mode, const char *path)
   6 {
   7         int score;
   8
   9         if (S_ISDIR(mode))
  10                 score = -1000;
  11         else if (S_ISLNK(mode))
  12                 score = -500;
  13         else
  14                 score = -50;
  15         return score;
  16 }
  17
  18 static int score_differs(unsigned mode1, unsigned mode2, const char *path)
  19 {
  20         int score;
  21
  22         if (S_ISDIR(mode1) != S_ISDIR(mode2))
  23                 score = -100;
  24         else if (S_ISLNK(mode1) != S_ISLNK(mode2))
  25                 score = -50;
  26         else
  27                 score = -5;
  28         return score;
  29 }
  30
  31 static int score_matches(unsigned mode1, unsigned mode2, const char *path)
  32 {
  33         int score;
  34
  35         /* Heh, we found SHA-1 collisions between different kind of objects */
  36         if (S_ISDIR(mode1) != S_ISDIR(mode2))
  37                 score = -100;
  38         else if (S_ISLNK(mode1) != S_ISLNK(mode2))
  39                 score = -50;
  40
  41         else if (S_ISDIR(mode1))
  42                 score = 1000;
  43         else if (S_ISLNK(mode1))
  44                 score = 500;
  45         else
  46                 score = 250;
  47         return score;
  48 }
  49
  50 static void *fill_tree_desc_strict(struct tree_desc *desc,
  51                                    const unsigned char *hash)
  52 {
  53         void *buffer;
  54         enum object_type type;
  55         unsigned long size;
  56
  57         buffer = read_sha1_file(hash, &type, &size);
  58         if (!buffer)
  59                 die("unable to read tree (%s)", sha1_to_hex(hash));
  60         if (type != OBJ_TREE)
  61                 die("%s is not a tree", sha1_to_hex(hash));
  62         init_tree_desc(desc, buffer, size);
  63         return buffer;
  64 }
  65
  66 static int base_name_entries_compare(const struct name_entry *a,
  67                                      const struct name_entry *b)
  68 {
  69         return base_name_compare(a->path, tree_entry_len(a), a->mode,
  70                                  b->path, tree_entry_len(b), b->mode);
  71 }
  72
  73 /*
  74  * Inspect two trees, and give a score that tells how similar they are.
  75  */
  76 static int score_trees(const unsigned char *hash1, const unsigned char *hash2)
  77 {
  78         struct tree_desc one;
  79         struct tree_desc two;
  80         void *one_buf = fill_tree_desc_strict(&one, hash1);
  81         void *two_buf = fill_tree_desc_strict(&two, hash2);
  82         int score = 0;
  83
  84         for (;;) {
  85                 struct name_entry e1, e2;
  86                 int got_entry_from_one = tree_entry(&one, &e1);
  87                 int got_entry_from_two = tree_entry(&two, &e2);
  88                 int cmp;
  89
  90                 if (got_entry_from_one && got_entry_from_two)
  91                         cmp = base_name_entries_compare(&e1, &e2);
  92                 else if (got_entry_from_one)
  93                         /* two lacks this entry */
  94                         cmp = -1;
  95                 else if (got_entry_from_two)
  96                         /* two has more entries */
  97                         cmp = 1;
  98                 else
  99                         break;
 100
 101                 if (cmp < 0)
 102                         /* path1 does not appear in two */
 103                         score += score_missing(e1.mode, e1.path);
 104                 else if (cmp > 0)
 105                         /* path2 does not appear in one */
 106                         score += score_missing(e2.mode, e2.path);
 107                 else if (hashcmp(e1.sha1, e2.sha1))
 108                         /* they are different */
 109                         score += score_differs(e1.mode, e2.mode, e1.path);
 110                 else
 111                         /* same subtree or blob */
 112                         score += score_matches(e1.mode, e2.mode, e1.path);
 113         }
 114         free(one_buf);
 115         free(two_buf);
 116         return score;
 117 }
 118
 119 /*
 120  * Match one itself and its subtrees with two and pick the best match.
 121  */
 122 static void match_trees(const unsigned char *hash1,
 123                         const unsigned char *hash2,
 124                         int *best_score,
 125                         char **best_match,
 126                         const char *base,
 127                         int recurse_limit)
 128 {
 129         struct tree_desc one;
 130         void *one_buf = fill_tree_desc_strict(&one, hash1);
 131
 132         while (one.size) {
 133                 const char *path;
 134                 const unsigned char *elem;
 135                 unsigned mode;
 136                 int score;
 137
 138                 elem = tree_entry_extract(&one, &path, &mode);
 139                 if (!S_ISDIR(mode))
 140                         goto next;
 141                 score = score_trees(elem, hash2);
 142                 if (*best_score < score) {
 143                         char *newpath;
 144                         newpath = xmalloc(strlen(base) + strlen(path) + 1);
 145                         sprintf(newpath, "%s%s", base, path);
 146                         free(*best_match);
 147                         *best_match = newpath;
 148                         *best_score = score;
 149                 }
 150                 if (recurse_limit) {
 151                         char *newbase;
 152                         newbase = xmalloc(strlen(base) + strlen(path) + 2);
 153                         sprintf(newbase, "%s%s/", base, path);
 154                         match_trees(elem, hash2, best_score, best_match,
 155                                     newbase, recurse_limit - 1);
 156                         free(newbase);
 157                 }
 158
 159         next:
 160                 update_tree_entry(&one);
 161         }
 162         free(one_buf);
 163 }
 164
 165 /*
 166  * A tree "hash1" has a subdirectory at "prefix".  Come up with a
 167  * tree object by replacing it with another tree "hash2".
 168  */
 169 static int splice_tree(const unsigned char *hash1,
 170                        const char *prefix,
 171                        const unsigned char *hash2,
 172                        unsigned char *result)
 173 {
 174         char *subpath;
 175         int toplen;
 176         char *buf;
 177         unsigned long sz;
 178         struct tree_desc desc;
 179         unsigned char *rewrite_here;
 180         const unsigned char *rewrite_with;
 181         unsigned char subtree[20];
 182         enum object_type type;
 183         int status;
 184
 185         subpath = strchr(prefix, '/');
 186         if (!subpath)
 187                 toplen = strlen(prefix);
 188         else {
 189                 toplen = subpath - prefix;
 190                 subpath++;
 191         }
 192
 193         buf = read_sha1_file(hash1, &type, &sz);
 194         if (!buf)
 195                 die("cannot read tree %s", sha1_to_hex(hash1));
 196         init_tree_desc(&desc, buf, sz);
 197
 198         rewrite_here = NULL;
 199         while (desc.size) {
 200                 const char *name;
 201                 unsigned mode;
 202                 const unsigned char *sha1;
 203
 204                 sha1 = tree_entry_extract(&desc, &name, &mode);
 205                 if (strlen(name) == toplen &&
 206                     !memcmp(name, prefix, toplen)) {
 207                         if (!S_ISDIR(mode))
 208                                 die("entry %s in tree %s is not a tree",
 209                                     name, sha1_to_hex(hash1));
 210                         rewrite_here = (unsigned char *) sha1;
 211                         break;
 212                 }
 213                 update_tree_entry(&desc);
 214         }
 215         if (!rewrite_here)
 216                 die("entry %.*s not found in tree %s",
 217                     toplen, prefix, sha1_to_hex(hash1));
 218         if (subpath) {
 219                 status = splice_tree(rewrite_here, subpath, hash2, subtree);
 220                 if (status)
 221                         return status;
 222                 rewrite_with = subtree;
 223         }
 224         else
 225                 rewrite_with = hash2;
 226         hashcpy(rewrite_here, rewrite_with);
 227         status = write_sha1_file(buf, sz, tree_type, result);
 228         free(buf);
 229         return status;
 230 }
 231
 232 /*
 233  * We are trying to come up with a merge between one and two that
 234  * results in a tree shape similar to one.  The tree two might
 235  * correspond to a subtree of one, in which case it needs to be
 236  * shifted down by prefixing otherwise empty directories.  On the
 237  * other hand, it could cover tree one and we might need to pick a
 238  * subtree of it.
 239  */
 240 void shift_tree(const unsigned char *hash1,
 241                 const unsigned char *hash2,
 242                 unsigned char *shifted,
 243                 int depth_limit)
 244 {
 245         char *add_prefix;
 246         char *del_prefix;
 247         int add_score, del_score;
 248
 249         /*
 250          * NEEDSWORK: this limits the recursion depth to hardcoded
 251          * value '2' to avoid excessive overhead.
 252          */
 253         if (!depth_limit)
 254                 depth_limit = 2;
 255
 256         add_score = del_score = score_trees(hash1, hash2);
 257         add_prefix = xcalloc(1, 1);
 258         del_prefix = xcalloc(1, 1);
 259
 260         /*
 261          * See if one's subtree resembles two; if so we need to prefix
 262          * two with a few fake trees to match the prefix.
 263          */
 264         match_trees(hash1, hash2, &add_score, &add_prefix, "", depth_limit);
 265
 266         /*
 267          * See if two's subtree resembles one; if so we need to
 268          * pick only subtree of two.
 269          */
 270         match_trees(hash2, hash1, &del_score, &del_prefix, "", depth_limit);
 271
 272         /* Assume we do not have to do any shifting */
 273         hashcpy(shifted, hash2);
 274
 275         if (add_score < del_score) {
 276                 /* We need to pick a subtree of two */
 277                 unsigned mode;
 278
 279                 if (!*del_prefix)
 280                         return;
 281
 282                 if (get_tree_entry(hash2, del_prefix, shifted, &mode))
 283                         die("cannot find path %s in tree %s",
 284                             del_prefix, sha1_to_hex(hash2));
 285                 return;
 286         }
 287
 288         if (!*add_prefix)
 289                 return;
 290
 291         splice_tree(hash1, add_prefix, hash2, shifted);
 292 }
 293
 294 /*
 295  * The user says the trees will be shifted by this much.
 296  * Unfortunately we cannot fundamentally tell which one to
 297  * be prefixed, as recursive merge can work in either direction.
 298  */
 299 void shift_tree_by(const unsigned char *hash1,
 300                    const unsigned char *hash2,
 301                    unsigned char *shifted,
 302                    const char *shift_prefix)
 303 {
 304         unsigned char sub1[20], sub2[20];
 305         unsigned mode1, mode2;
 306         unsigned candidate = 0;
 307
 308         /* Can hash2 be a tree at shift_prefix in tree hash1? */
 309         if (!get_tree_entry(hash1, shift_prefix, sub1, &mode1) &&
 310             S_ISDIR(mode1))
 311                 candidate |= 1;
 312
 313         /* Can hash1 be a tree at shift_prefix in tree hash2? */
 314         if (!get_tree_entry(hash2, shift_prefix, sub2, &mode2) &&
 315             S_ISDIR(mode2))
 316                 candidate |= 2;
 317
 318         if (candidate == 3) {
 319                 /* Both are plausible -- we need to evaluate the score */
 320                 int best_score = score_trees(hash1, hash2);
 321                 int score;
 322
 323                 candidate = 0;
 324                 score = score_trees(sub1, hash2);
 325                 if (score > best_score) {
 326                         candidate = 1;
 327                         best_score = score;
 328                 }
 329                 score = score_trees(sub2, hash1);
 330                 if (score > best_score)
 331                         candidate = 2;
 332         }
 333
 334         if (!candidate) {
 335                 /* Neither is plausible -- do not shift */
 336                 hashcpy(shifted, hash2);
 337                 return;
 338         }
 339
 340         if (candidate == 1)
 341                 /*
 342                  * shift tree2 down by adding shift_prefix above it
 343                  * to match tree1.
 344                  */
 345                 splice_tree(hash1, shift_prefix, hash2, shifted);
 346         else
 347                 /*
 348                  * shift tree2 up by removing shift_prefix from it
 349                  * to match tree1.
 350                  */
 351                 hashcpy(shifted, sub2);
 352 }