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)
   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)
  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)
  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);
 102                         update_tree_entry(&one);
 103                 } else if (cmp > 0) {
 104                         /* path2 does not appear in one */
 105                         score += score_missing(two.entry.mode);
 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                         } else {
 114                                 /* same subtree or blob */
 115                                 score += score_matches(one.entry.mode,
 116                                                        two.entry.mode);
 117                         }
 118                         update_tree_entry(&one);
 119                         update_tree_entry(&two);
 120                 }
 121         }
 122         free(one_buf);
 123         free(two_buf);
 124         return score;
 125 }
 126
 127 /*
 128  * Match one itself and its subtrees with two and pick the best match.
 129  */
 130 static void match_trees(const struct object_id *hash1,
 131                         const struct object_id *hash2,
 132                         int *best_score,
 133                         char **best_match,
 134                         const char *base,
 135                         int recurse_limit)
 136 {
 137         struct tree_desc one;
 138         void *one_buf = fill_tree_desc_strict(&one, hash1);
 139
 140         while (one.size) {
 141                 const char *path;
 142                 const struct object_id *elem;
 143                 unsigned mode;
 144                 int score;
 145
 146                 elem = tree_entry_extract(&one, &path, &mode);
 147                 if (!S_ISDIR(mode))
 148                         goto next;
 149                 score = score_trees(elem, hash2);
 150                 if (*best_score < score) {
 151                         free(*best_match);
 152                         *best_match = xstrfmt("%s%s", base, path);
 153                         *best_score = score;
 154                 }
 155                 if (recurse_limit) {
 156                         char *newbase = xstrfmt("%s%s/", base, path);
 157                         match_trees(elem, hash2, best_score, best_match,
 158                                     newbase, recurse_limit - 1);
 159                         free(newbase);
 160                 }
 161
 162         next:
 163                 update_tree_entry(&one);
 164         }
 165         free(one_buf);
 166 }
 167
 168 /*
 169  * A tree "oid1" has a subdirectory at "prefix".  Come up with a tree object by
 170  * replacing it with another tree "oid2".
 171  */
 172 static int splice_tree(const struct object_id *oid1, const char *prefix,
 173                        const struct object_id *oid2, struct object_id *result)
 174 {
 175         char *subpath;
 176         int toplen;
 177         char *buf;
 178         unsigned long sz;
 179         struct tree_desc desc;
 180         struct object_id *rewrite_here;
 181         const struct object_id *rewrite_with;
 182         struct object_id subtree;
 183         enum object_type type;
 184         int status;
 185
 186         subpath = strchrnul(prefix, '/');
 187         toplen = subpath - prefix;
 188         if (*subpath)
 189                 subpath++;
 190
 191         buf = read_object_file(oid1, &type, &sz);
 192         if (!buf)
 193                 die("cannot read tree %s", oid_to_hex(oid1));
 194         init_tree_desc(&desc, buf, sz);
 195
 196         rewrite_here = NULL;
 197         while (desc.size) {
 198                 const char *name;
 199                 unsigned mode;
 200                 const struct object_id *oid;
 201
 202                 oid = tree_entry_extract(&desc, &name, &mode);
 203                 if (strlen(name) == toplen &&
 204                     !memcmp(name, prefix, toplen)) {
 205                         if (!S_ISDIR(mode))
 206                                 die("entry %s in tree %s is not a tree", name,
 207                                     oid_to_hex(oid1));
 208                         rewrite_here = (struct object_id *)oid;
 209                         break;
 210                 }
 211                 update_tree_entry(&desc);
 212         }
 213         if (!rewrite_here)
 214                 die("entry %.*s not found in tree %s", toplen, prefix,
 215                     oid_to_hex(oid1));
 216         if (*subpath) {
 217                 status = splice_tree(rewrite_here, subpath, oid2, &subtree);
 218                 if (status)
 219                         return status;
 220                 rewrite_with = &subtree;
 221         } else {
 222                 rewrite_with = oid2;
 223         }
 224         oidcpy(rewrite_here, rewrite_with);
 225         status = write_object_file(buf, sz, tree_type, result);
 226         free(buf);
 227         return status;
 228 }
 229
 230 /*
 231  * We are trying to come up with a merge between one and two that
 232  * results in a tree shape similar to one.  The tree two might
 233  * correspond to a subtree of one, in which case it needs to be
 234  * shifted down by prefixing otherwise empty directories.  On the
 235  * other hand, it could cover tree one and we might need to pick a
 236  * subtree of it.
 237  */
 238 void shift_tree(const struct object_id *hash1,
 239                 const struct object_id *hash2,
 240                 struct object_id *shifted,
 241                 int depth_limit)
 242 {
 243         char *add_prefix;
 244         char *del_prefix;
 245         int add_score, del_score;
 246
 247         /*
 248          * NEEDSWORK: this limits the recursion depth to hardcoded
 249          * value '2' to avoid excessive overhead.
 250          */
 251         if (!depth_limit)
 252                 depth_limit = 2;
 253
 254         add_score = del_score = score_trees(hash1, hash2);
 255         add_prefix = xcalloc(1, 1);
 256         del_prefix = xcalloc(1, 1);
 257
 258         /*
 259          * See if one's subtree resembles two; if so we need to prefix
 260          * two with a few fake trees to match the prefix.
 261          */
 262         match_trees(hash1, hash2, &add_score, &add_prefix, "", depth_limit);
 263
 264         /*
 265          * See if two's subtree resembles one; if so we need to
 266          * pick only subtree of two.
 267          */
 268         match_trees(hash2, hash1, &del_score, &del_prefix, "", depth_limit);
 269
 270         /* Assume we do not have to do any shifting */
 271         oidcpy(shifted, hash2);
 272
 273         if (add_score < del_score) {
 274                 /* We need to pick a subtree of two */
 275                 unsigned mode;
 276
 277                 if (!*del_prefix)
 278                         return;
 279
 280                 if (get_tree_entry(hash2, del_prefix, shifted, &mode))
 281                         die("cannot find path %s in tree %s",
 282                             del_prefix, oid_to_hex(hash2));
 283                 return;
 284         }
 285
 286         if (!*add_prefix)
 287                 return;
 288
 289         splice_tree(hash1, add_prefix, hash2, shifted);
 290 }
 291
 292 /*
 293  * The user says the trees will be shifted by this much.
 294  * Unfortunately we cannot fundamentally tell which one to
 295  * be prefixed, as recursive merge can work in either direction.
 296  */
 297 void shift_tree_by(const struct object_id *hash1,
 298                    const struct object_id *hash2,
 299                    struct object_id *shifted,
 300                    const char *shift_prefix)
 301 {
 302         struct object_id sub1, sub2;
 303         unsigned mode1, mode2;
 304         unsigned candidate = 0;
 305
 306         /* Can hash2 be a tree at shift_prefix in tree hash1? */
 307         if (!get_tree_entry(hash1, shift_prefix, &sub1, &mode1) &&
 308             S_ISDIR(mode1))
 309                 candidate |= 1;
 310
 311         /* Can hash1 be a tree at shift_prefix in tree hash2? */
 312         if (!get_tree_entry(hash2, shift_prefix, &sub2, &mode2) &&
 313             S_ISDIR(mode2))
 314                 candidate |= 2;
 315
 316         if (candidate == 3) {
 317                 /* Both are plausible -- we need to evaluate the score */
 318                 int best_score = score_trees(hash1, hash2);
 319                 int score;
 320
 321                 candidate = 0;
 322                 score = score_trees(&sub1, hash2);
 323                 if (score > best_score) {
 324                         candidate = 1;
 325                         best_score = score;
 326                 }
 327                 score = score_trees(&sub2, hash1);
 328                 if (score > best_score)
 329                         candidate = 2;
 330         }
 331
 332         if (!candidate) {
 333                 /* Neither is plausible -- do not shift */
 334                 oidcpy(shifted, hash2);
 335                 return;
 336         }
 337
 338         if (candidate == 1)
 339                 /*
 340                  * shift tree2 down by adding shift_prefix above it
 341                  * to match tree1.
 342                  */
 343                 splice_tree(hash1, shift_prefix, hash2, shifted);
 344         else
 345                 /*
 346                  * shift tree2 up by removing shift_prefix from it
 347                  * to match tree1.
 348                  */
 349                 oidcpy(shifted, &sub2);
 350 }