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                         free(*best_match);
 144                         *best_match = xstrfmt("%s%s", base, path);
 145                         *best_score = score;
 146                 }
 147                 if (recurse_limit) {
 148                         char *newbase = xstrfmt("%s%s/", base, path);
 149                         match_trees(elem, hash2, best_score, best_match,
 150                                     newbase, recurse_limit - 1);
 151                         free(newbase);
 152                 }
 153
 154         next:
 155                 update_tree_entry(&one);
 156         }
 157         free(one_buf);
 158 }
 159
 160 /*
 161  * A tree "hash1" has a subdirectory at "prefix".  Come up with a
 162  * tree object by replacing it with another tree "hash2".
 163  */
 164 static int splice_tree(const unsigned char *hash1,
 165                        const char *prefix,
 166                        const unsigned char *hash2,
 167                        unsigned char *result)
 168 {
 169         char *subpath;
 170         int toplen;
 171         char *buf;
 172         unsigned long sz;
 173         struct tree_desc desc;
 174         unsigned char *rewrite_here;
 175         const unsigned char *rewrite_with;
 176         unsigned char subtree[20];
 177         enum object_type type;
 178         int status;
 179
 180         subpath = strchrnul(prefix, '/');
 181         toplen = subpath - prefix;
 182         if (*subpath)
 183                 subpath++;
 184
 185         buf = read_sha1_file(hash1, &type, &sz);
 186         if (!buf)
 187                 die("cannot read tree %s", sha1_to_hex(hash1));
 188         init_tree_desc(&desc, buf, sz);
 189
 190         rewrite_here = NULL;
 191         while (desc.size) {
 192                 const char *name;
 193                 unsigned mode;
 194                 const unsigned char *sha1;
 195
 196                 sha1 = tree_entry_extract(&desc, &name, &mode);
 197                 if (strlen(name) == toplen &&
 198                     !memcmp(name, prefix, toplen)) {
 199                         if (!S_ISDIR(mode))
 200                                 die("entry %s in tree %s is not a tree",
 201                                     name, sha1_to_hex(hash1));
 202                         rewrite_here = (unsigned char *) sha1;
 203                         break;
 204                 }
 205                 update_tree_entry(&desc);
 206         }
 207         if (!rewrite_here)
 208                 die("entry %.*s not found in tree %s",
 209                     toplen, prefix, sha1_to_hex(hash1));
 210         if (*subpath) {
 211                 status = splice_tree(rewrite_here, subpath, hash2, subtree);
 212                 if (status)
 213                         return status;
 214                 rewrite_with = subtree;
 215         }
 216         else
 217                 rewrite_with = hash2;
 218         hashcpy(rewrite_here, rewrite_with);
 219         status = write_sha1_file(buf, sz, tree_type, result);
 220         free(buf);
 221         return status;
 222 }
 223
 224 /*
 225  * We are trying to come up with a merge between one and two that
 226  * results in a tree shape similar to one.  The tree two might
 227  * correspond to a subtree of one, in which case it needs to be
 228  * shifted down by prefixing otherwise empty directories.  On the
 229  * other hand, it could cover tree one and we might need to pick a
 230  * subtree of it.
 231  */
 232 void shift_tree(const unsigned char *hash1,
 233                 const unsigned char *hash2,
 234                 unsigned char *shifted,
 235                 int depth_limit)
 236 {
 237         char *add_prefix;
 238         char *del_prefix;
 239         int add_score, del_score;
 240
 241         /*
 242          * NEEDSWORK: this limits the recursion depth to hardcoded
 243          * value '2' to avoid excessive overhead.
 244          */
 245         if (!depth_limit)
 246                 depth_limit = 2;
 247
 248         add_score = del_score = score_trees(hash1, hash2);
 249         add_prefix = xcalloc(1, 1);
 250         del_prefix = xcalloc(1, 1);
 251
 252         /*
 253          * See if one's subtree resembles two; if so we need to prefix
 254          * two with a few fake trees to match the prefix.
 255          */
 256         match_trees(hash1, hash2, &add_score, &add_prefix, "", depth_limit);
 257
 258         /*
 259          * See if two's subtree resembles one; if so we need to
 260          * pick only subtree of two.
 261          */
 262         match_trees(hash2, hash1, &del_score, &del_prefix, "", depth_limit);
 263
 264         /* Assume we do not have to do any shifting */
 265         hashcpy(shifted, hash2);
 266
 267         if (add_score < del_score) {
 268                 /* We need to pick a subtree of two */
 269                 unsigned mode;
 270
 271                 if (!*del_prefix)
 272                         return;
 273
 274                 if (get_tree_entry(hash2, del_prefix, shifted, &mode))
 275                         die("cannot find path %s in tree %s",
 276                             del_prefix, sha1_to_hex(hash2));
 277                 return;
 278         }
 279
 280         if (!*add_prefix)
 281                 return;
 282
 283         splice_tree(hash1, add_prefix, hash2, shifted);
 284 }
 285
 286 /*
 287  * The user says the trees will be shifted by this much.
 288  * Unfortunately we cannot fundamentally tell which one to
 289  * be prefixed, as recursive merge can work in either direction.
 290  */
 291 void shift_tree_by(const unsigned char *hash1,
 292                    const unsigned char *hash2,
 293                    unsigned char *shifted,
 294                    const char *shift_prefix)
 295 {
 296         unsigned char sub1[20], sub2[20];
 297         unsigned mode1, mode2;
 298         unsigned candidate = 0;
 299
 300         /* Can hash2 be a tree at shift_prefix in tree hash1? */
 301         if (!get_tree_entry(hash1, shift_prefix, sub1, &mode1) &&
 302             S_ISDIR(mode1))
 303                 candidate |= 1;
 304
 305         /* Can hash1 be a tree at shift_prefix in tree hash2? */
 306         if (!get_tree_entry(hash2, shift_prefix, sub2, &mode2) &&
 307             S_ISDIR(mode2))
 308                 candidate |= 2;
 309
 310         if (candidate == 3) {
 311                 /* Both are plausible -- we need to evaluate the score */
 312                 int best_score = score_trees(hash1, hash2);
 313                 int score;
 314
 315                 candidate = 0;
 316                 score = score_trees(sub1, hash2);
 317                 if (score > best_score) {
 318                         candidate = 1;
 319                         best_score = score;
 320                 }
 321                 score = score_trees(sub2, hash1);
 322                 if (score > best_score)
 323                         candidate = 2;
 324         }
 325
 326         if (!candidate) {
 327                 /* Neither is plausible -- do not shift */
 328                 hashcpy(shifted, hash2);
 329                 return;
 330         }
 331
 332         if (candidate == 1)
 333                 /*
 334                  * shift tree2 down by adding shift_prefix above it
 335                  * to match tree1.
 336                  */
 337                 splice_tree(hash1, shift_prefix, hash2, shifted);
 338         else
 339                 /*
 340                  * shift tree2 up by removing shift_prefix from it
 341                  * to match tree1.
 342                  */
 343                 hashcpy(shifted, sub2);
 344 }