match-trees.c

   1 #include "cache.h"
   2 #include "tree.h"
   3 #include "tree-walk.h"
   4 #include "object-store.h"
   5
   6 static int score_missing(unsigned mode, const char *path)
   7 {
   8         int score;
   9
  10         if (S_ISDIR(mode))
  11                 score = -1000;
  12         else if (S_ISLNK(mode))
  13                 score = -500;
  14         else
  15                 score = -50;
  16         return score;
  17 }
  18
  19 static int score_differs(unsigned mode1, unsigned mode2, const char *path)
  20 {
  21         int score;
  22
  23         if (S_ISDIR(mode1) != S_ISDIR(mode2))
  24                 score = -100;
  25         else if (S_ISLNK(mode1) != S_ISLNK(mode2))
  26                 score = -50;
  27         else
  28                 score = -5;
  29         return score;
  30 }
  31
  32 static int score_matches(unsigned mode1, unsigned mode2, const char *path)
  33 {
  34         int score;
  35
  36         /* Heh, we found SHA-1 collisions between different kind of objects */
  37         if (S_ISDIR(mode1) != S_ISDIR(mode2))
  38                 score = -100;
  39         else if (S_ISLNK(mode1) != S_ISLNK(mode2))
  40                 score = -50;
  41
  42         else if (S_ISDIR(mode1))
  43                 score = 1000;
  44         else if (S_ISLNK(mode1))
  45                 score = 500;
  46         else
  47                 score = 250;
  48         return score;
  49 }
  50
  51 static void *fill_tree_desc_strict(struct tree_desc *desc,
  52                                    const struct object_id *hash)
  53 {
  54         void *buffer;
  55         enum object_type type;
  56         unsigned long size;
  57
  58         buffer = read_object_file(hash, &type, &size);
  59         if (!buffer)
  60                 die("unable to read tree (%s)", oid_to_hex(hash));
  61         if (type != OBJ_TREE)
  62                 die("%s is not a tree", oid_to_hex(hash));
  63         init_tree_desc(desc, buffer, size);
  64         return buffer;
  65 }
  66
  67 static int base_name_entries_compare(const struct name_entry *a,
  68                                      const struct name_entry *b)
  69 {
  70         return base_name_compare(a->path, tree_entry_len(a), a->mode,
  71                                  b->path, tree_entry_len(b), b->mode);
  72 }
  73
  74 /*
  75  * Inspect two trees, and give a score that tells how similar they are.
  76  */
  77 static int score_trees(const struct object_id *hash1, const struct object_id *hash2)
  78 {
  79         struct tree_desc one;
  80         struct tree_desc two;
  81         void *one_buf = fill_tree_desc_strict(&one, hash1);
  82         void *two_buf = fill_tree_desc_strict(&two, hash2);
  83         int score = 0;
  84
  85         for (;;) {
  86                 int cmp;
  87
  88                 if (one.size && two.size)
  89                         cmp = base_name_entries_compare(&one.entry, &two.entry);
  90                 else if (one.size)
  91                         /* two lacks this entry */
  92                         cmp = -1;
  93                 else if (two.size)
  94                         /* two has more entries */
  95                         cmp = 1;
  96                 else
  97                         break;
  98
  99                 if (cmp < 0) {
 100                         /* path1 does not appear in two */
 101                         score += score_missing(one.entry.mode, one.entry.path);
 102                         update_tree_entry(&one);
 103                 } else if (cmp > 0) {
 104                         /* path2 does not appear in one */
 105                         score += score_missing(two.entry.mode, two.entry.path);
 106                         update_tree_entry(&two);
 107                 } else {
 108                         /* path appears in both */
 109                         if (!oideq(&one.entry.oid, &two.entry.oid)) {
 110                                 /* they are different */
 111                                 score += score_differs(one.entry.mode,
 112                                                        two.entry.mode,
 113                                                        one.entry.path);
 114                         } else {
 115                                 /* same subtree or blob */
 116                                 score += score_matches(one.entry.mode,
 117                                                        two.entry.mode,
 118                                                        one.entry.path);
 119                         }
 120                         update_tree_entry(&one);
 121                         update_tree_entry(&two);
 122                 }
 123         }
 124         free(one_buf);
 125         free(two_buf);
 126         return score;
 127 }
 128
 129 /*
 130  * Match one itself and its subtrees with two and pick the best match.
 131  */
 132 static void match_trees(const struct object_id *hash1,
 133                         const struct object_id *hash2,
 134                         int *best_score,
 135                         char **best_match,
 136                         const char *base,
 137                         int recurse_limit)
 138 {
 139         struct tree_desc one;
 140         void *one_buf = fill_tree_desc_strict(&one, hash1);
 141
 142         while (one.size) {
 143                 const char *path;
 144                 const struct object_id *elem;
 145                 unsigned mode;
 146                 int score;
 147
 148                 elem = tree_entry_extract(&one, &path, &mode);
 149                 if (!S_ISDIR(mode))
 150                         goto next;
 151                 score = score_trees(elem, hash2);
 152                 if (*best_score < score) {
 153                         free(*best_match);
 154                         *best_match = xstrfmt("%s%s", base, path);
 155                         *best_score = score;
 156                 }
 157                 if (recurse_limit) {
 158                         char *newbase = xstrfmt("%s%s/", base, path);
 159                         match_trees(elem, hash2, best_score, best_match,
 160                                     newbase, recurse_limit - 1);
 161                         free(newbase);
 162                 }
 163
 164         next:
 165                 update_tree_entry(&one);
 166         }
 167         free(one_buf);
 168 }
 169
 170 /*
 171  * A tree "oid1" has a subdirectory at "prefix".  Come up with a tree object by
 172  * replacing it with another tree "oid2".
 173  */
 174 static int splice_tree(const struct object_id *oid1, const char *prefix,
 175                        const struct object_id *oid2, struct object_id *result)
 176 {
 177         char *subpath;
 178         int toplen;
 179         char *buf;
 180         unsigned long sz;
 181         struct tree_desc desc;
 182         unsigned char *rewrite_here;
 183         const struct object_id *rewrite_with;
 184         struct object_id subtree;
 185         enum object_type type;
 186         int status;
 187
 188         subpath = strchrnul(prefix, '/');
 189         toplen = subpath - prefix;
 190         if (*subpath)
 191                 subpath++;
 192
 193         buf = read_object_file(oid1, &type, &sz);
 194         if (!buf)
 195                 die("cannot read tree %s", oid_to_hex(oid1));
 196         init_tree_desc(&desc, buf, sz);
 197
 198         rewrite_here = NULL;
 199         while (desc.size) {
 200                 const char *name;
 201                 unsigned mode;
 202
 203                 tree_entry_extract(&desc, &name, &mode);
 204                 if (strlen(name) == toplen &&
 205                     !memcmp(name, prefix, toplen)) {
 206                         if (!S_ISDIR(mode))
 207                                 die("entry %s in tree %s is not a tree", name,
 208                                     oid_to_hex(oid1));
 209
 210                         /*
 211                          * We cast here for two reasons:
 212                          *
 213                          *   - to flip the "char *" (for the path) to "unsigned
 214                          *     char *" (for the hash stored after it)
 215                          *
 216                          *   - to discard the "const"; this is OK because we
 217                          *     know it points into our non-const "buf"
 218                          */
 219                         rewrite_here = (unsigned char *)(desc.entry.path +
 220                                                          strlen(desc.entry.path) +
 221                                                          1);
 222                         break;
 223                 }
 224                 update_tree_entry(&desc);
 225         }
 226         if (!rewrite_here)
 227                 die("entry %.*s not found in tree %s", toplen, prefix,
 228                     oid_to_hex(oid1));
 229         if (*subpath) {
 230                 struct object_id tree_oid;
 231                 hashcpy(tree_oid.hash, rewrite_here);
 232                 status = splice_tree(&tree_oid, subpath, oid2, &subtree);
 233                 if (status)
 234                         return status;
 235                 rewrite_with = &subtree;
 236         } else {
 237                 rewrite_with = oid2;
 238         }
 239         hashcpy(rewrite_here, rewrite_with->hash);
 240         status = write_object_file(buf, sz, tree_type, result);
 241         free(buf);
 242         return status;
 243 }
 244
 245 /*
 246  * We are trying to come up with a merge between one and two that
 247  * results in a tree shape similar to one.  The tree two might
 248  * correspond to a subtree of one, in which case it needs to be
 249  * shifted down by prefixing otherwise empty directories.  On the
 250  * other hand, it could cover tree one and we might need to pick a
 251  * subtree of it.
 252  */
 253 void shift_tree(const struct object_id *hash1,
 254                 const struct object_id *hash2,
 255                 struct object_id *shifted,
 256                 int depth_limit)
 257 {
 258         char *add_prefix;
 259         char *del_prefix;
 260         int add_score, del_score;
 261
 262         /*
 263          * NEEDSWORK: this limits the recursion depth to hardcoded
 264          * value '2' to avoid excessive overhead.
 265          */
 266         if (!depth_limit)
 267                 depth_limit = 2;
 268
 269         add_score = del_score = score_trees(hash1, hash2);
 270         add_prefix = xcalloc(1, 1);
 271         del_prefix = xcalloc(1, 1);
 272
 273         /*
 274          * See if one's subtree resembles two; if so we need to prefix
 275          * two with a few fake trees to match the prefix.
 276          */
 277         match_trees(hash1, hash2, &add_score, &add_prefix, "", depth_limit);
 278
 279         /*
 280          * See if two's subtree resembles one; if so we need to
 281          * pick only subtree of two.
 282          */
 283         match_trees(hash2, hash1, &del_score, &del_prefix, "", depth_limit);
 284
 285         /* Assume we do not have to do any shifting */
 286         oidcpy(shifted, hash2);
 287
 288         if (add_score < del_score) {
 289                 /* We need to pick a subtree of two */
 290                 unsigned mode;
 291
 292                 if (!*del_prefix)
 293                         return;
 294
 295                 if (get_tree_entry(hash2, del_prefix, shifted, &mode))
 296                         die("cannot find path %s in tree %s",
 297                             del_prefix, oid_to_hex(hash2));
 298                 return;
 299         }
 300
 301         if (!*add_prefix)
 302                 return;
 303
 304         splice_tree(hash1, add_prefix, hash2, shifted);
 305 }
 306
 307 /*
 308  * The user says the trees will be shifted by this much.
 309  * Unfortunately we cannot fundamentally tell which one to
 310  * be prefixed, as recursive merge can work in either direction.
 311  */
 312 void shift_tree_by(const struct object_id *hash1,
 313                    const struct object_id *hash2,
 314                    struct object_id *shifted,
 315                    const char *shift_prefix)
 316 {
 317         struct object_id sub1, sub2;
 318         unsigned mode1, mode2;
 319         unsigned candidate = 0;
 320
 321         /* Can hash2 be a tree at shift_prefix in tree hash1? */
 322         if (!get_tree_entry(hash1, shift_prefix, &sub1, &mode1) &&
 323             S_ISDIR(mode1))
 324                 candidate |= 1;
 325
 326         /* Can hash1 be a tree at shift_prefix in tree hash2? */
 327         if (!get_tree_entry(hash2, shift_prefix, &sub2, &mode2) &&
 328             S_ISDIR(mode2))
 329                 candidate |= 2;
 330
 331         if (candidate == 3) {
 332                 /* Both are plausible -- we need to evaluate the score */
 333                 int best_score = score_trees(hash1, hash2);
 334                 int score;
 335
 336                 candidate = 0;
 337                 score = score_trees(&sub1, hash2);
 338                 if (score > best_score) {
 339                         candidate = 1;
 340                         best_score = score;
 341                 }
 342                 score = score_trees(&sub2, hash1);
 343                 if (score > best_score)
 344                         candidate = 2;
 345         }
 346
 347         if (!candidate) {
 348                 /* Neither is plausible -- do not shift */
 349                 oidcpy(shifted, hash2);
 350                 return;
 351         }
 352
 353         if (candidate == 1)
 354                 /*
 355                  * shift tree2 down by adding shift_prefix above it
 356                  * to match tree1.
 357                  */
 358                 splice_tree(hash1, shift_prefix, hash2, shifted);
 359         else
 360                 /*
 361                  * shift tree2 up by removing shift_prefix from it
 362                  * to match tree1.
 363                  */
 364                 oidcpy(shifted, &sub2);
 365 }