match-trees.c

   1 #include "git-compat-util.h"
   2 #include "hex.h"
   3 #include "match-trees.h"
   4 #include "strbuf.h"
   5 #include "tree.h"
   6 #include "tree-walk.h"
   7 #include "object-store-ll.h"
   8
   9 static int score_missing(unsigned mode)
  10 {
  11         int score;
  12
  13         if (S_ISDIR(mode))
  14                 score = -1000;
  15         else if (S_ISLNK(mode))
  16                 score = -500;
  17         else
  18                 score = -50;
  19         return score;
  20 }
  21
  22 static int score_differs(unsigned mode1, unsigned mode2)
  23 {
  24         int score;
  25
  26         if (S_ISDIR(mode1) != S_ISDIR(mode2))
  27                 score = -100;
  28         else if (S_ISLNK(mode1) != S_ISLNK(mode2))
  29                 score = -50;
  30         else
  31                 score = -5;
  32         return score;
  33 }
  34
  35 static int score_matches(unsigned mode1, unsigned mode2)
  36 {
  37         int score;
  38
  39         /* Heh, we found SHA-1 collisions between different kind of objects */
  40         if (S_ISDIR(mode1) != S_ISDIR(mode2))
  41                 score = -100;
  42         else if (S_ISLNK(mode1) != S_ISLNK(mode2))
  43                 score = -50;
  44
  45         else if (S_ISDIR(mode1))
  46                 score = 1000;
  47         else if (S_ISLNK(mode1))
  48                 score = 500;
  49         else
  50                 score = 250;
  51         return score;
  52 }
  53
  54 static void *fill_tree_desc_strict(struct tree_desc *desc,
  55                                    const struct object_id *hash)
  56 {
  57         void *buffer;
  58         enum object_type type;
  59         unsigned long size;
  60
  61         buffer = repo_read_object_file(the_repository, hash, &type, &size);
  62         if (!buffer)
  63                 die("unable to read tree (%s)", oid_to_hex(hash));
  64         if (type != OBJ_TREE)
  65                 die("%s is not a tree", oid_to_hex(hash));
  66         init_tree_desc(desc, buffer, size);
  67         return buffer;
  68 }
  69
  70 static int base_name_entries_compare(const struct name_entry *a,
  71                                      const struct name_entry *b)
  72 {
  73         return base_name_compare(a->path, tree_entry_len(a), a->mode,
  74                                  b->path, tree_entry_len(b), b->mode);
  75 }
  76
  77 /*
  78  * Inspect two trees, and give a score that tells how similar they are.
  79  */
  80 static int score_trees(const struct object_id *hash1, const struct object_id *hash2)
  81 {
  82         struct tree_desc one;
  83         struct tree_desc two;
  84         void *one_buf = fill_tree_desc_strict(&one, hash1);
  85         void *two_buf = fill_tree_desc_strict(&two, hash2);
  86         int score = 0;
  87
  88         for (;;) {
  89                 int cmp;
  90
  91                 if (one.size && two.size)
  92                         cmp = base_name_entries_compare(&one.entry, &two.entry);
  93                 else if (one.size)
  94                         /* two lacks this entry */
  95                         cmp = -1;
  96                 else if (two.size)
  97                         /* two has more entries */
  98                         cmp = 1;
  99                 else
 100                         break;
 101
 102                 if (cmp < 0) {
 103                         /* path1 does not appear in two */
 104                         score += score_missing(one.entry.mode);
 105                         update_tree_entry(&one);
 106                 } else if (cmp > 0) {
 107                         /* path2 does not appear in one */
 108                         score += score_missing(two.entry.mode);
 109                         update_tree_entry(&two);
 110                 } else {
 111                         /* path appears in both */
 112                         if (!oideq(&one.entry.oid, &two.entry.oid)) {
 113                                 /* they are different */
 114                                 score += score_differs(one.entry.mode,
 115                                                        two.entry.mode);
 116                         } else {
 117                                 /* same subtree or blob */
 118                                 score += score_matches(one.entry.mode,
 119                                                        two.entry.mode);
 120                         }
 121                         update_tree_entry(&one);
 122                         update_tree_entry(&two);
 123                 }
 124         }
 125         free(one_buf);
 126         free(two_buf);
 127         return score;
 128 }
 129
 130 /*
 131  * Match one itself and its subtrees with two and pick the best match.
 132  */
 133 static void match_trees(const struct object_id *hash1,
 134                         const struct object_id *hash2,
 135                         int *best_score,
 136                         char **best_match,
 137                         const char *base,
 138                         int recurse_limit)
 139 {
 140         struct tree_desc one;
 141         void *one_buf = fill_tree_desc_strict(&one, hash1);
 142
 143         while (one.size) {
 144                 const char *path;
 145                 const struct object_id *elem;
 146                 unsigned short mode;
 147                 int score;
 148
 149                 elem = tree_entry_extract(&one, &path, &mode);
 150                 if (!S_ISDIR(mode))
 151                         goto next;
 152                 score = score_trees(elem, hash2);
 153                 if (*best_score < score) {
 154                         free(*best_match);
 155                         *best_match = xstrfmt("%s%s", base, path);
 156                         *best_score = score;
 157                 }
 158                 if (recurse_limit) {
 159                         char *newbase = xstrfmt("%s%s/", base, path);
 160                         match_trees(elem, hash2, best_score, best_match,
 161                                     newbase, recurse_limit - 1);
 162                         free(newbase);
 163                 }
 164
 165         next:
 166                 update_tree_entry(&one);
 167         }
 168         free(one_buf);
 169 }
 170
 171 /*
 172  * A tree "oid1" has a subdirectory at "prefix".  Come up with a tree object by
 173  * replacing it with another tree "oid2".
 174  */
 175 static int splice_tree(const struct object_id *oid1, const char *prefix,
 176                        const struct object_id *oid2, struct object_id *result)
 177 {
 178         char *subpath;
 179         int toplen;
 180         char *buf;
 181         unsigned long sz;
 182         struct tree_desc desc;
 183         unsigned char *rewrite_here;
 184         const struct object_id *rewrite_with;
 185         struct object_id subtree;
 186         enum object_type type;
 187         int status;
 188
 189         subpath = strchrnul(prefix, '/');
 190         toplen = subpath - prefix;
 191         if (*subpath)
 192                 subpath++;
 193
 194         buf = repo_read_object_file(the_repository, oid1, &type, &sz);
 195         if (!buf)
 196                 die("cannot read tree %s", oid_to_hex(oid1));
 197         init_tree_desc(&desc, buf, sz);
 198
 199         rewrite_here = NULL;
 200         while (desc.size) {
 201                 const char *name;
 202                 unsigned short mode;
 203
 204                 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
 211                         /*
 212                          * We cast here for two reasons:
 213                          *
 214                          *   - to flip the "char *" (for the path) to "unsigned
 215                          *     char *" (for the hash stored after it)
 216                          *
 217                          *   - to discard the "const"; this is OK because we
 218                          *     know it points into our non-const "buf"
 219                          */
 220                         rewrite_here = (unsigned char *)(desc.entry.path +
 221                                                          strlen(desc.entry.path) +
 222                                                          1);
 223                         break;
 224                 }
 225                 update_tree_entry(&desc);
 226         }
 227         if (!rewrite_here)
 228                 die("entry %.*s not found in tree %s", toplen, prefix,
 229                     oid_to_hex(oid1));
 230         if (*subpath) {
 231                 struct object_id tree_oid;
 232                 oidread(&tree_oid, rewrite_here);
 233                 status = splice_tree(&tree_oid, subpath, oid2, &subtree);
 234                 if (status)
 235                         return status;
 236                 rewrite_with = &subtree;
 237         } else {
 238                 rewrite_with = oid2;
 239         }
 240         hashcpy(rewrite_here, rewrite_with->hash);
 241         status = write_object_file(buf, sz, OBJ_TREE, result);
 242         free(buf);
 243         return status;
 244 }
 245
 246 /*
 247  * We are trying to come up with a merge between one and two that
 248  * results in a tree shape similar to one.  The tree two might
 249  * correspond to a subtree of one, in which case it needs to be
 250  * shifted down by prefixing otherwise empty directories.  On the
 251  * other hand, it could cover tree one and we might need to pick a
 252  * subtree of it.
 253  */
 254 void shift_tree(struct repository *r,
 255                 const struct object_id *hash1,
 256                 const struct object_id *hash2,
 257                 struct object_id *shifted,
 258                 int depth_limit)
 259 {
 260         char *add_prefix;
 261         char *del_prefix;
 262         int add_score, del_score;
 263
 264         /*
 265          * NEEDSWORK: this limits the recursion depth to hardcoded
 266          * value '2' to avoid excessive overhead.
 267          */
 268         if (!depth_limit)
 269                 depth_limit = 2;
 270
 271         add_score = del_score = score_trees(hash1, hash2);
 272         add_prefix = xcalloc(1, 1);
 273         del_prefix = xcalloc(1, 1);
 274
 275         /*
 276          * See if one's subtree resembles two; if so we need to prefix
 277          * two with a few fake trees to match the prefix.
 278          */
 279         match_trees(hash1, hash2, &add_score, &add_prefix, "", depth_limit);
 280
 281         /*
 282          * See if two's subtree resembles one; if so we need to
 283          * pick only subtree of two.
 284          */
 285         match_trees(hash2, hash1, &del_score, &del_prefix, "", depth_limit);
 286
 287         /* Assume we do not have to do any shifting */
 288         oidcpy(shifted, hash2);
 289
 290         if (add_score < del_score) {
 291                 /* We need to pick a subtree of two */
 292                 unsigned short mode;
 293
 294                 if (!*del_prefix)
 295                         return;
 296
 297                 if (get_tree_entry(r, hash2, del_prefix, shifted, &mode))
 298                         die("cannot find path %s in tree %s",
 299                             del_prefix, oid_to_hex(hash2));
 300                 return;
 301         }
 302
 303         if (!*add_prefix)
 304                 return;
 305
 306         splice_tree(hash1, add_prefix, hash2, shifted);
 307 }
 308
 309 /*
 310  * The user says the trees will be shifted by this much.
 311  * Unfortunately we cannot fundamentally tell which one to
 312  * be prefixed, as recursive merge can work in either direction.
 313  */
 314 void shift_tree_by(struct repository *r,
 315                    const struct object_id *hash1,
 316                    const struct object_id *hash2,
 317                    struct object_id *shifted,
 318                    const char *shift_prefix)
 319 {
 320         struct object_id sub1, sub2;
 321         unsigned short mode1, mode2;
 322         unsigned candidate = 0;
 323
 324         /* Can hash2 be a tree at shift_prefix in tree hash1? */
 325         if (!get_tree_entry(r, hash1, shift_prefix, &sub1, &mode1) &&
 326             S_ISDIR(mode1))
 327                 candidate |= 1;
 328
 329         /* Can hash1 be a tree at shift_prefix in tree hash2? */
 330         if (!get_tree_entry(r, hash2, shift_prefix, &sub2, &mode2) &&
 331             S_ISDIR(mode2))
 332                 candidate |= 2;
 333
 334         if (candidate == 3) {
 335                 /* Both are plausible -- we need to evaluate the score */
 336                 int best_score = score_trees(hash1, hash2);
 337                 int score;
 338
 339                 candidate = 0;
 340                 score = score_trees(&sub1, hash2);
 341                 if (score > best_score) {
 342                         candidate = 1;
 343                         best_score = score;
 344                 }
 345                 score = score_trees(&sub2, hash1);
 346                 if (score > best_score)
 347                         candidate = 2;
 348         }
 349
 350         if (!candidate) {
 351                 /* Neither is plausible -- do not shift */
 352                 oidcpy(shifted, hash2);
 353                 return;
 354         }
 355
 356         if (candidate == 1)
 357                 /*
 358                  * shift tree2 down by adding shift_prefix above it
 359                  * to match tree1.
 360                  */
 361                 splice_tree(hash1, shift_prefix, hash2, shifted);
 362         else
 363                 /*
 364                  * shift tree2 up by removing shift_prefix from it
 365                  * to match tree1.
 366                  */
 367                 oidcpy(shifted, &sub2);
 368 }