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 (oidcmp(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         struct object_id *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                 const struct object_id *oid;
 203
 204                 oid = 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", name,
 209                                     oid_to_hex(oid1));
 210                         rewrite_here = (struct object_id *)oid;
 211                         break;
 212                 }
 213                 update_tree_entry(&desc);
 214         }
 215         if (!rewrite_here)
 216                 die("entry %.*s not found in tree %s", toplen, prefix,
 217                     oid_to_hex(oid1));
 218         if (*subpath) {
 219                 status = splice_tree(rewrite_here, subpath, oid2, &subtree);
 220                 if (status)
 221                         return status;
 222                 rewrite_with = &subtree;
 223         } else {
 224                 rewrite_with = oid2;
 225         }
 226         oidcpy(rewrite_here, rewrite_with);
 227         status = write_object_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 struct object_id *hash1,
 241                 const struct object_id *hash2,
 242                 struct object_id *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         oidcpy(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, oid_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 struct object_id *hash1,
 300                    const struct object_id *hash2,
 301                    struct object_id *shifted,
 302                    const char *shift_prefix)
 303 {
 304         struct object_id sub1, sub2;
 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                 oidcpy(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                 oidcpy(shifted, &sub2);
 352 }