tree-walk.c

   1 #define USE_THE_REPOSITORY_VARIABLE
   2
   3 #include "git-compat-util.h"
   4 #include "tree-walk.h"
   5 #include "dir.h"
   6 #include "gettext.h"
   7 #include "hex.h"
   8 #include "object-file.h"
   9 #include "odb.h"
  10 #include "trace2.h"
  11 #include "tree.h"
  12 #include "pathspec.h"
  13 #include "json-writer.h"
  14 #include "environment.h"
  15
  16 static int decode_tree_entry(struct tree_desc *desc, const char *buf, unsigned long size, struct strbuf *err)
  17 {
  18         const char *path;
  19         unsigned int len;
  20         uint16_t mode;
  21         const unsigned hashsz = desc->algo->rawsz;
  22
  23         if (size < hashsz + 3 || buf[size - (hashsz + 1)]) {
  24                 strbuf_addstr(err, _("too-short tree object"));
  25                 return -1;
  26         }
  27
  28         path = parse_mode(buf, &mode);
  29         if (!path) {
  30                 strbuf_addstr(err, _("malformed mode in tree entry"));
  31                 return -1;
  32         }
  33         if (!*path) {
  34                 strbuf_addstr(err, _("empty filename in tree entry"));
  35                 return -1;
  36         }
  37         len = strlen(path) + 1;
  38
  39         /* Initialize the descriptor entry */
  40         desc->entry.path = path;
  41         desc->entry.mode = (desc->flags & TREE_DESC_RAW_MODES) ? mode : canon_mode(mode);
  42         desc->entry.pathlen = len - 1;
  43         oidread(&desc->entry.oid, (const unsigned char *)path + len,
  44                 desc->algo);
  45
  46         return 0;
  47 }
  48
  49 static int init_tree_desc_internal(struct tree_desc *desc,
  50                                    const struct object_id *oid,
  51                                    const void *buffer, unsigned long size,
  52                                    struct strbuf *err,
  53                                    enum tree_desc_flags flags)
  54 {
  55         desc->algo = (oid && oid->algo) ? &hash_algos[oid->algo] : the_hash_algo;
  56         desc->buffer = buffer;
  57         desc->size = size;
  58         desc->flags = flags;
  59         if (size)
  60                 return decode_tree_entry(desc, buffer, size, err);
  61         return 0;
  62 }
  63
  64 void init_tree_desc(struct tree_desc *desc, const struct object_id *tree_oid,
  65                     const void *buffer, unsigned long size)
  66 {
  67         struct strbuf err = STRBUF_INIT;
  68         if (init_tree_desc_internal(desc, tree_oid, buffer, size, &err, 0))
  69                 die("%s", err.buf);
  70         strbuf_release(&err);
  71 }
  72
  73 int init_tree_desc_gently(struct tree_desc *desc, const struct object_id *oid,
  74                           const void *buffer, unsigned long size,
  75                           enum tree_desc_flags flags)
  76 {
  77         struct strbuf err = STRBUF_INIT;
  78         int result = init_tree_desc_internal(desc, oid, buffer, size, &err, flags);
  79         if (result)
  80                 error("%s", err.buf);
  81         strbuf_release(&err);
  82         return result;
  83 }
  84
  85 void *fill_tree_descriptor(struct repository *r,
  86                            struct tree_desc *desc,
  87                            const struct object_id *oid)
  88 {
  89         unsigned long size = 0;
  90         void *buf = NULL;
  91
  92         if (oid) {
  93                 buf = odb_read_object_peeled(r->objects, oid, OBJ_TREE, &size, NULL);
  94                 if (!buf)
  95                         die(_("unable to read tree (%s)"), oid_to_hex(oid));
  96         }
  97         init_tree_desc(desc, oid, buf, size);
  98         return buf;
  99 }
 100
 101 static void entry_clear(struct name_entry *a)
 102 {
 103         memset(a, 0, sizeof(*a));
 104 }
 105
 106 static void entry_extract(struct tree_desc *t, struct name_entry *a)
 107 {
 108         *a = t->entry;
 109 }
 110
 111 static int update_tree_entry_internal(struct tree_desc *desc, struct strbuf *err)
 112 {
 113         const void *buf = desc->buffer;
 114         const unsigned char *end = (const unsigned char *)desc->entry.path + desc->entry.pathlen + 1 + desc->algo->rawsz;
 115         unsigned long size = desc->size;
 116         unsigned long len = end - (const unsigned char *)buf;
 117
 118         if (size < len)
 119                 die(_("too-short tree file"));
 120         buf = end;
 121         size -= len;
 122         desc->buffer = buf;
 123         desc->size = size;
 124         if (size)
 125                 return decode_tree_entry(desc, buf, size, err);
 126         return 0;
 127 }
 128
 129 void update_tree_entry(struct tree_desc *desc)
 130 {
 131         struct strbuf err = STRBUF_INIT;
 132         if (update_tree_entry_internal(desc, &err))
 133                 die("%s", err.buf);
 134         strbuf_release(&err);
 135 }
 136
 137 int update_tree_entry_gently(struct tree_desc *desc)
 138 {
 139         struct strbuf err = STRBUF_INIT;
 140         if (update_tree_entry_internal(desc, &err)) {
 141                 error("%s", err.buf);
 142                 strbuf_release(&err);
 143                 /* Stop processing this tree after error */
 144                 desc->size = 0;
 145                 return -1;
 146         }
 147         strbuf_release(&err);
 148         return 0;
 149 }
 150
 151 int tree_entry(struct tree_desc *desc, struct name_entry *entry)
 152 {
 153         if (!desc->size)
 154                 return 0;
 155
 156         *entry = desc->entry;
 157         update_tree_entry(desc);
 158         return 1;
 159 }
 160
 161 int tree_entry_gently(struct tree_desc *desc, struct name_entry *entry)
 162 {
 163         if (!desc->size)
 164                 return 0;
 165
 166         *entry = desc->entry;
 167         if (update_tree_entry_gently(desc))
 168                 return 0;
 169         return 1;
 170 }
 171
 172 static int traverse_trees_atexit_registered;
 173 static int traverse_trees_count;
 174 static int traverse_trees_cur_depth;
 175 static int traverse_trees_max_depth;
 176
 177 static void trace2_traverse_trees_statistics_atexit(void)
 178 {
 179         struct json_writer jw = JSON_WRITER_INIT;
 180
 181         jw_object_begin(&jw, 0);
 182         jw_object_intmax(&jw, "traverse_trees_count", traverse_trees_count);
 183         jw_object_intmax(&jw, "traverse_trees_max_depth", traverse_trees_max_depth);
 184         jw_end(&jw);
 185
 186         trace2_data_json("traverse_trees", the_repository, "statistics", &jw);
 187
 188         jw_release(&jw);
 189 }
 190
 191 void setup_traverse_info(struct traverse_info *info, const char *base)
 192 {
 193         size_t pathlen = strlen(base);
 194         static struct traverse_info dummy;
 195
 196         memset(info, 0, sizeof(*info));
 197         if (pathlen && base[pathlen-1] == '/')
 198                 pathlen--;
 199         info->pathlen = pathlen ? pathlen + 1 : 0;
 200         info->name = base;
 201         info->namelen = pathlen;
 202         if (pathlen)
 203                 info->prev = &dummy;
 204
 205         if (trace2_is_enabled() && !traverse_trees_atexit_registered) {
 206                 atexit(trace2_traverse_trees_statistics_atexit);
 207                 traverse_trees_atexit_registered = 1;
 208         }
 209 }
 210
 211 char *make_traverse_path(char *path, size_t pathlen,
 212                          const struct traverse_info *info,
 213                          const char *name, size_t namelen)
 214 {
 215         /* Always points to the end of the name we're about to add */
 216         size_t pos = st_add(info->pathlen, namelen);
 217
 218         if (pos >= pathlen)
 219                 BUG("too small buffer passed to make_traverse_path");
 220
 221         path[pos] = 0;
 222         for (;;) {
 223                 if (pos < namelen)
 224                         BUG("traverse_info pathlen does not match strings");
 225                 pos -= namelen;
 226                 memcpy(path + pos, name, namelen);
 227
 228                 if (!pos)
 229                         break;
 230                 path[--pos] = '/';
 231
 232                 if (!info)
 233                         BUG("traverse_info ran out of list items");
 234                 name = info->name;
 235                 namelen = info->namelen;
 236                 info = info->prev;
 237         }
 238         return path;
 239 }
 240
 241 void strbuf_make_traverse_path(struct strbuf *out,
 242                                const struct traverse_info *info,
 243                                const char *name, size_t namelen)
 244 {
 245         size_t len = traverse_path_len(info, namelen);
 246
 247         strbuf_grow(out, len);
 248         make_traverse_path(out->buf + out->len, out->alloc - out->len,
 249                            info, name, namelen);
 250         strbuf_setlen(out, out->len + len);
 251 }
 252
 253 struct tree_desc_skip {
 254         struct tree_desc_skip *prev;
 255         const void *ptr;
 256 };
 257
 258 struct tree_desc_x {
 259         struct tree_desc d;
 260         struct tree_desc_skip *skip;
 261 };
 262
 263 static int check_entry_match(const char *a, int a_len, const char *b, int b_len)
 264 {
 265         /*
 266          * The caller wants to pick *a* from a tree or nothing.
 267          * We are looking at *b* in a tree.
 268          *
 269          * (0) If a and b are the same name, we are trivially happy.
 270          *
 271          * There are three possibilities where *a* could be hiding
 272          * behind *b*.
 273          *
 274          * (1) *a* == "t",   *b* == "ab"  i.e. *b* sorts earlier than *a* no
 275          *                                matter what.
 276          * (2) *a* == "t",   *b* == "t-2" and "t" is a subtree in the tree;
 277          * (3) *a* == "t-2", *b* == "t"   and "t-2" is a blob in the tree.
 278          *
 279          * Otherwise we know *a* won't appear in the tree without
 280          * scanning further.
 281          */
 282
 283         int cmp = name_compare(a, a_len, b, b_len);
 284
 285         /* Most common case first -- reading sync'd trees */
 286         if (!cmp)
 287                 return cmp;
 288
 289         if (0 < cmp) {
 290                 /* a comes after b; it does not matter if it is case (3)
 291                 if (b_len < a_len && !memcmp(a, b, b_len) && a[b_len] < '/')
 292                         return 1;
 293                 */
 294                 return 1; /* keep looking */
 295         }
 296
 297         /* b comes after a; are we looking at case (2)? */
 298         if (a_len < b_len && !memcmp(a, b, a_len) && b[a_len] < '/')
 299                 return 1; /* keep looking */
 300
 301         return -1; /* a cannot appear in the tree */
 302 }
 303
 304 /*
 305  * From the extended tree_desc, extract the first name entry, while
 306  * paying attention to the candidate "first" name.  Most importantly,
 307  * when looking for an entry, if there are entries that sorts earlier
 308  * in the tree object representation than that name, skip them and
 309  * process the named entry first.  We will remember that we haven't
 310  * processed the first entry yet, and in the later call skip the
 311  * entry we processed early when update_extended_entry() is called.
 312  *
 313  * E.g. if the underlying tree object has these entries:
 314  *
 315  *    blob    "t-1"
 316  *    blob    "t-2"
 317  *    tree    "t"
 318  *    blob    "t=1"
 319  *
 320  * and the "first" asks for "t", remember that we still need to
 321  * process "t-1" and "t-2" but extract "t".  After processing the
 322  * entry "t" from this call, the caller will let us know by calling
 323  * update_extended_entry() that we can remember "t" has been processed
 324  * already.
 325  */
 326
 327 static void extended_entry_extract(struct tree_desc_x *t,
 328                                    struct name_entry *a,
 329                                    const char *first,
 330                                    int first_len)
 331 {
 332         const char *path;
 333         int len;
 334         struct tree_desc probe;
 335         struct tree_desc_skip *skip;
 336
 337         /*
 338          * Extract the first entry from the tree_desc, but skip the
 339          * ones that we already returned in earlier rounds.
 340          */
 341         while (1) {
 342                 if (!t->d.size) {
 343                         entry_clear(a);
 344                         break; /* not found */
 345                 }
 346                 entry_extract(&t->d, a);
 347                 for (skip = t->skip; skip; skip = skip->prev)
 348                         if (a->path == skip->ptr)
 349                                 break; /* found */
 350                 if (!skip)
 351                         break;
 352                 /* We have processed this entry already. */
 353                 update_tree_entry(&t->d);
 354         }
 355
 356         if (!first || !a->path)
 357                 return;
 358
 359         /*
 360          * The caller wants "first" from this tree, or nothing.
 361          */
 362         path = a->path;
 363         len = tree_entry_len(a);
 364         switch (check_entry_match(first, first_len, path, len)) {
 365         case -1:
 366                 entry_clear(a);
 367         case 0:
 368                 return;
 369         default:
 370                 break;
 371         }
 372
 373         /*
 374          * We need to look-ahead -- we suspect that a subtree whose
 375          * name is "first" may be hiding behind the current entry "path".
 376          */
 377         probe = t->d;
 378         while (probe.size) {
 379                 entry_extract(&probe, a);
 380                 path = a->path;
 381                 len = tree_entry_len(a);
 382                 switch (check_entry_match(first, first_len, path, len)) {
 383                 case -1:
 384                         entry_clear(a);
 385                 case 0:
 386                         return;
 387                 default:
 388                         update_tree_entry(&probe);
 389                         break;
 390                 }
 391                 /* keep looking */
 392         }
 393         entry_clear(a);
 394 }
 395
 396 static void update_extended_entry(struct tree_desc_x *t, struct name_entry *a)
 397 {
 398         if (t->d.entry.path == a->path) {
 399                 update_tree_entry(&t->d);
 400         } else {
 401                 /* we have returned this entry early */
 402                 struct tree_desc_skip *skip = xmalloc(sizeof(*skip));
 403                 skip->ptr = a->path;
 404                 skip->prev = t->skip;
 405                 t->skip = skip;
 406         }
 407 }
 408
 409 static void free_extended_entry(struct tree_desc_x *t)
 410 {
 411         struct tree_desc_skip *p, *s;
 412
 413         for (s = t->skip; s; s = p) {
 414                 p = s->prev;
 415                 free(s);
 416         }
 417 }
 418
 419 static inline int prune_traversal(struct index_state *istate,
 420                                   struct name_entry *e,
 421                                   struct traverse_info *info,
 422                                   struct strbuf *base,
 423                                   int still_interesting)
 424 {
 425         if (!info->pathspec || still_interesting == 2)
 426                 return 2;
 427         if (still_interesting < 0)
 428                 return still_interesting;
 429         return tree_entry_interesting(istate, e, base,
 430                                       info->pathspec);
 431 }
 432
 433 int traverse_trees(struct index_state *istate,
 434                    int n, struct tree_desc *t,
 435                    struct traverse_info *info)
 436 {
 437         int ret = 0;
 438         struct name_entry *entry;
 439         int i;
 440         struct tree_desc_x *tx;
 441         struct strbuf base = STRBUF_INIT;
 442         int interesting = 1;
 443         char *traverse_path;
 444
 445         if (traverse_trees_cur_depth > max_allowed_tree_depth)
 446                 return error("exceeded maximum allowed tree depth");
 447
 448         traverse_trees_count++;
 449         traverse_trees_cur_depth++;
 450
 451         if (traverse_trees_cur_depth > traverse_trees_max_depth)
 452                 traverse_trees_max_depth = traverse_trees_cur_depth;
 453
 454         ALLOC_ARRAY(entry, n);
 455         ALLOC_ARRAY(tx, n);
 456
 457         for (i = 0; i < n; i++) {
 458                 tx[i].d = t[i];
 459                 tx[i].skip = NULL;
 460         }
 461
 462         if (info->prev) {
 463                 strbuf_make_traverse_path(&base, info->prev,
 464                                           info->name, info->namelen);
 465                 strbuf_addch(&base, '/');
 466                 traverse_path = xstrndup(base.buf, base.len);
 467         } else {
 468                 traverse_path = xstrndup(info->name, info->pathlen);
 469         }
 470         info->traverse_path = traverse_path;
 471         for (;;) {
 472                 int trees_used;
 473                 unsigned long mask, dirmask;
 474                 const char *first = NULL;
 475                 int first_len = 0;
 476                 struct name_entry *e = NULL;
 477                 int len;
 478
 479                 for (i = 0; i < n; i++) {
 480                         e = entry + i;
 481                         extended_entry_extract(tx + i, e, NULL, 0);
 482                 }
 483
 484                 /*
 485                  * A tree may have "t-2" at the current location even
 486                  * though it may have "t" that is a subtree behind it,
 487                  * and another tree may return "t".  We want to grab
 488                  * all "t" from all trees to match in such a case.
 489                  */
 490                 for (i = 0; i < n; i++) {
 491                         e = entry + i;
 492                         if (!e->path)
 493                                 continue;
 494                         len = tree_entry_len(e);
 495                         if (!first) {
 496                                 first = e->path;
 497                                 first_len = len;
 498                                 continue;
 499                         }
 500                         if (name_compare(e->path, len, first, first_len) < 0) {
 501                                 first = e->path;
 502                                 first_len = len;
 503                         }
 504                 }
 505
 506                 if (first) {
 507                         for (i = 0; i < n; i++) {
 508                                 e = entry + i;
 509                                 extended_entry_extract(tx + i, e, first, first_len);
 510                                 /* Cull the ones that are not the earliest */
 511                                 if (!e->path)
 512                                         continue;
 513                                 len = tree_entry_len(e);
 514                                 if (name_compare(e->path, len, first, first_len))
 515                                         entry_clear(e);
 516                         }
 517                 }
 518
 519                 /* Now we have in entry[i] the earliest name from the trees */
 520                 mask = 0;
 521                 dirmask = 0;
 522                 for (i = 0; i < n; i++) {
 523                         if (!entry[i].path)
 524                                 continue;
 525                         mask |= 1ul << i;
 526                         if (S_ISDIR(entry[i].mode))
 527                                 dirmask |= 1ul << i;
 528                         e = &entry[i];
 529                 }
 530                 if (!mask)
 531                         break;
 532                 interesting = prune_traversal(istate, e, info, &base, interesting);
 533                 if (interesting < 0)
 534                         break;
 535                 if (interesting) {
 536                         trees_used = info->fn(n, mask, dirmask, entry, info);
 537                         if (trees_used < 0) {
 538                                 ret = trees_used;
 539                                 if (!info->show_all_errors)
 540                                         break;
 541                         }
 542                         mask &= trees_used;
 543                 }
 544                 for (i = 0; i < n; i++)
 545                         if (mask & (1ul << i))
 546                                 update_extended_entry(tx + i, entry + i);
 547         }
 548         for (i = 0; i < n; i++)
 549                 free_extended_entry(tx + i);
 550         free(tx);
 551         free(entry);
 552         free(traverse_path);
 553         info->traverse_path = NULL;
 554         strbuf_release(&base);
 555
 556         traverse_trees_cur_depth--;
 557         return ret;
 558 }
 559
 560 struct dir_state {
 561         void *tree;
 562         unsigned long size;
 563         struct object_id oid;
 564 };
 565
 566 static int find_tree_entry(struct repository *r, struct tree_desc *t,
 567                            const char *name, struct object_id *result,
 568                            unsigned short *mode)
 569 {
 570         int namelen = strlen(name);
 571         while (t->size) {
 572                 const char *entry;
 573                 struct object_id oid;
 574                 int entrylen, cmp;
 575
 576                 oidcpy(&oid, tree_entry_extract(t, &entry, mode));
 577                 entrylen = tree_entry_len(&t->entry);
 578                 update_tree_entry(t);
 579                 if (entrylen > namelen)
 580                         continue;
 581                 cmp = memcmp(name, entry, entrylen);
 582                 if (cmp > 0)
 583                         continue;
 584                 if (cmp < 0)
 585                         break;
 586                 if (entrylen == namelen) {
 587                         oidcpy(result, &oid);
 588                         return 0;
 589                 }
 590                 if (name[entrylen] != '/')
 591                         continue;
 592                 if (!S_ISDIR(*mode))
 593                         break;
 594                 if (++entrylen == namelen) {
 595                         oidcpy(result, &oid);
 596                         return 0;
 597                 }
 598                 return get_tree_entry(r, &oid, name + entrylen, result, mode);
 599         }
 600         return -1;
 601 }
 602
 603 int get_tree_entry(struct repository *r,
 604                    const struct object_id *tree_oid,
 605                    const char *name,
 606                    struct object_id *oid,
 607                    unsigned short *mode)
 608 {
 609         int retval;
 610         void *tree;
 611         unsigned long size;
 612         struct object_id root;
 613
 614         tree = odb_read_object_peeled(r->objects, tree_oid, OBJ_TREE, &size, &root);
 615         if (!tree)
 616                 return -1;
 617
 618         if (name[0] == '\0') {
 619                 oidcpy(oid, &root);
 620                 free(tree);
 621                 return 0;
 622         }
 623
 624         if (!size) {
 625                 retval = -1;
 626         } else {
 627                 struct tree_desc t;
 628                 init_tree_desc(&t, tree_oid, tree, size);
 629                 retval = find_tree_entry(r, &t, name, oid, mode);
 630         }
 631         free(tree);
 632         return retval;
 633 }
 634
 635 /*
 636  * This is Linux's built-in max for the number of symlinks to follow.
 637  * That limit, of course, does not affect git, but it's a reasonable
 638  * choice.
 639  */
 640 #define GET_TREE_ENTRY_FOLLOW_SYMLINKS_MAX_LINKS 40
 641
 642 /**
 643  * Find a tree entry by following symlinks in tree_sha (which is
 644  * assumed to be the root of the repository).  In the event that a
 645  * symlink points outside the repository (e.g. a link to /foo or a
 646  * root-level link to ../foo), the portion of the link which is
 647  * outside the repository will be returned in result_path, and *mode
 648  * will be set to 0.  It is assumed that result_path is uninitialized.
 649  * If there are no symlinks, or the end result of the symlink chain
 650  * points to an object inside the repository, result will be filled in
 651  * with the sha1 of the found object, and *mode will hold the mode of
 652  * the object.
 653  *
 654  * See the code for enum get_oid_result for a description of
 655  * the return values.
 656  */
 657 enum get_oid_result get_tree_entry_follow_symlinks(struct repository *r,
 658                 struct object_id *tree_oid, const char *name,
 659                 struct object_id *result, struct strbuf *result_path,
 660                 unsigned short *mode)
 661 {
 662         int retval = MISSING_OBJECT;
 663         struct dir_state *parents = NULL;
 664         size_t parents_alloc = 0;
 665         size_t i, parents_nr = 0;
 666         struct object_id current_tree_oid;
 667         struct strbuf namebuf = STRBUF_INIT;
 668         struct tree_desc t;
 669         int follows_remaining = GET_TREE_ENTRY_FOLLOW_SYMLINKS_MAX_LINKS;
 670
 671         init_tree_desc(&t, NULL, NULL, 0UL);
 672         strbuf_addstr(&namebuf, name);
 673         oidcpy(&current_tree_oid, tree_oid);
 674
 675         while (1) {
 676                 int find_result;
 677                 char *first_slash;
 678                 char *remainder = NULL;
 679
 680                 if (!t.buffer) {
 681                         void *tree;
 682                         struct object_id root;
 683                         unsigned long size;
 684                         tree = odb_read_object_peeled(r->objects, &current_tree_oid,
 685                                                       OBJ_TREE, &size, &root);
 686                         if (!tree)
 687                                 goto done;
 688
 689                         ALLOC_GROW(parents, parents_nr + 1, parents_alloc);
 690                         parents[parents_nr].tree = tree;
 691                         parents[parents_nr].size = size;
 692                         oidcpy(&parents[parents_nr].oid, &root);
 693                         parents_nr++;
 694
 695                         if (namebuf.buf[0] == '\0') {
 696                                 oidcpy(result, &root);
 697                                 retval = FOUND;
 698                                 goto done;
 699                         }
 700
 701                         if (!size)
 702                                 goto done;
 703
 704                         /* descend */
 705                         init_tree_desc(&t, &current_tree_oid, tree, size);
 706                 }
 707
 708                 /* Handle symlinks to e.g. a//b by removing leading slashes */
 709                 while (namebuf.buf[0] == '/') {
 710                         strbuf_remove(&namebuf, 0, 1);
 711                 }
 712
 713                 /* Split namebuf into a first component and a remainder */
 714                 if ((first_slash = strchr(namebuf.buf, '/'))) {
 715                         *first_slash = 0;
 716                         remainder = first_slash + 1;
 717                 }
 718
 719                 if (!strcmp(namebuf.buf, "..")) {
 720                         struct dir_state *parent;
 721                         /*
 722                          * We could end up with .. in the namebuf if it
 723                          * appears in a symlink.
 724                          */
 725
 726                         if (parents_nr == 1) {
 727                                 if (remainder)
 728                                         *first_slash = '/';
 729                                 strbuf_add(result_path, namebuf.buf,
 730                                            namebuf.len);
 731                                 *mode = 0;
 732                                 retval = FOUND;
 733                                 goto done;
 734                         }
 735                         parent = &parents[parents_nr - 1];
 736                         free(parent->tree);
 737                         parents_nr--;
 738                         parent = &parents[parents_nr - 1];
 739                         init_tree_desc(&t, &parent->oid, parent->tree, parent->size);
 740                         strbuf_remove(&namebuf, 0, remainder ? 3 : 2);
 741                         continue;
 742                 }
 743
 744                 /* We could end up here via a symlink to dir/.. */
 745                 if (namebuf.buf[0] == '\0') {
 746                         oidcpy(result, &parents[parents_nr - 1].oid);
 747                         retval = FOUND;
 748                         goto done;
 749                 }
 750
 751                 /* Look up the first (or only) path component in the tree. */
 752                 find_result = find_tree_entry(r, &t, namebuf.buf,
 753                                               &current_tree_oid, mode);
 754                 if (find_result) {
 755                         goto done;
 756                 }
 757
 758                 if (S_ISDIR(*mode)) {
 759                         if (!remainder) {
 760                                 oidcpy(result, &current_tree_oid);
 761                                 retval = FOUND;
 762                                 goto done;
 763                         }
 764                         /* Descend the tree */
 765                         t.buffer = NULL;
 766                         strbuf_remove(&namebuf, 0,
 767                                       1 + first_slash - namebuf.buf);
 768                 } else if (S_ISREG(*mode)) {
 769                         if (!remainder) {
 770                                 oidcpy(result, &current_tree_oid);
 771                                 retval = FOUND;
 772                         } else {
 773                                 retval = NOT_DIR;
 774                         }
 775                         goto done;
 776                 } else if (S_ISLNK(*mode)) {
 777                         /* Follow a symlink */
 778                         unsigned long link_len;
 779                         size_t len;
 780                         char *contents, *contents_start;
 781                         struct dir_state *parent;
 782                         enum object_type type;
 783
 784                         if (follows_remaining-- == 0) {
 785                                 /* Too many symlinks followed */
 786                                 retval = SYMLINK_LOOP;
 787                                 goto done;
 788                         }
 789
 790                         /*
 791                          * At this point, we have followed at a least
 792                          * one symlink, so on error we need to report this.
 793                          */
 794                         retval = DANGLING_SYMLINK;
 795
 796                         contents = odb_read_object(r->objects,
 797                                                    &current_tree_oid, &type,
 798                                                    &link_len);
 799
 800                         if (!contents)
 801                                 goto done;
 802
 803                         if (contents[0] == '/') {
 804                                 strbuf_addstr(result_path, contents);
 805                                 free(contents);
 806                                 *mode = 0;
 807                                 retval = FOUND;
 808                                 goto done;
 809                         }
 810
 811                         if (remainder)
 812                                 len = first_slash - namebuf.buf;
 813                         else
 814                                 len = namebuf.len;
 815
 816                         contents_start = contents;
 817
 818                         parent = &parents[parents_nr - 1];
 819                         init_tree_desc(&t, &parent->oid, parent->tree, parent->size);
 820                         strbuf_splice(&namebuf, 0, len,
 821                                       contents_start, link_len);
 822                         if (remainder)
 823                                 namebuf.buf[link_len] = '/';
 824                         free(contents);
 825                 }
 826         }
 827 done:
 828         for (i = 0; i < parents_nr; i++)
 829                 free(parents[i].tree);
 830         free(parents);
 831
 832         strbuf_release(&namebuf);
 833         return retval;
 834 }
 835
 836 static int match_entry(const struct pathspec_item *item,
 837                        const struct name_entry *entry, int pathlen,
 838                        const char *match, int matchlen,
 839                        enum interesting *never_interesting)
 840 {
 841         int m = -1; /* signals that we haven't called strncmp() */
 842
 843         if (item->magic & PATHSPEC_ICASE)
 844                 /*
 845                  * "Never interesting" trick requires exact
 846                  * matching. We could do something clever with inexact
 847                  * matching, but it's trickier (and not to forget that
 848                  * strcasecmp is locale-dependent, at least in
 849                  * glibc). Just disable it for now. It can't be worse
 850                  * than the wildcard's codepath of '[Tt][Hi][Is][Ss]'
 851                  * pattern.
 852                  */
 853                 *never_interesting = entry_not_interesting;
 854         else if (*never_interesting != entry_not_interesting) {
 855                 /*
 856                  * We have not seen any match that sorts later
 857                  * than the current path.
 858                  */
 859
 860                 /*
 861                  * Does match sort strictly earlier than path
 862                  * with their common parts?
 863                  */
 864                 m = strncmp(match, entry->path,
 865                             (matchlen < pathlen) ? matchlen : pathlen);
 866                 if (m < 0)
 867                         return 0;
 868
 869                 /*
 870                  * If we come here even once, that means there is at
 871                  * least one pathspec that would sort equal to or
 872                  * later than the path we are currently looking at.
 873                  * In other words, if we have never reached this point
 874                  * after iterating all pathspecs, it means all
 875                  * pathspecs are either outside of base, or inside the
 876                  * base but sorts strictly earlier than the current
 877                  * one.  In either case, they will never match the
 878                  * subsequent entries.  In such a case, we initialized
 879                  * the variable to -1 and that is what will be
 880                  * returned, allowing the caller to terminate early.
 881                  */
 882                 *never_interesting = entry_not_interesting;
 883         }
 884
 885         if (pathlen > matchlen)
 886                 return 0;
 887
 888         if (matchlen > pathlen) {
 889                 if (match[pathlen] != '/')
 890                         return 0;
 891                 /*
 892                  * Reject non-directories as partial pathnames, except
 893                  * when match is a submodule with a trailing slash and
 894                  * nothing else (to handle 'submod/' and 'submod'
 895                  * uniformly).
 896                  */
 897                 if (!S_ISDIR(entry->mode) &&
 898                     (!S_ISGITLINK(entry->mode) || matchlen > pathlen + 1))
 899                         return 0;
 900         }
 901
 902         if (m == -1)
 903                 /*
 904                  * we cheated and did not do strncmp(), so we do
 905                  * that here.
 906                  */
 907                 m = ps_strncmp(item, match, entry->path, pathlen);
 908
 909         /*
 910          * If common part matched earlier then it is a hit,
 911          * because we rejected the case where path is not a
 912          * leading directory and is shorter than match.
 913          */
 914         if (!m)
 915                 /*
 916                  * match_entry does not check if the prefix part is
 917                  * matched case-sensitively. If the entry is a
 918                  * directory and part of prefix, it'll be rematched
 919                  * eventually by basecmp with special treatment for
 920                  * the prefix.
 921                  */
 922                 return 1;
 923
 924         return 0;
 925 }
 926
 927 /* :(icase)-aware string compare */
 928 static int basecmp(const struct pathspec_item *item,
 929                    const char *base, const char *match, int len)
 930 {
 931         if (item->magic & PATHSPEC_ICASE) {
 932                 int ret, n = len > item->prefix ? item->prefix : len;
 933                 ret = strncmp(base, match, n);
 934                 if (ret)
 935                         return ret;
 936                 base += n;
 937                 match += n;
 938                 len -= n;
 939         }
 940         return ps_strncmp(item, base, match, len);
 941 }
 942
 943 static int match_dir_prefix(const struct pathspec_item *item,
 944                             const char *base,
 945                             const char *match, int matchlen)
 946 {
 947         if (basecmp(item, base, match, matchlen))
 948                 return 0;
 949
 950         /*
 951          * If the base is a subdirectory of a path which
 952          * was specified, all of them are interesting.
 953          */
 954         if (!matchlen ||
 955             base[matchlen] == '/' ||
 956             match[matchlen - 1] == '/')
 957                 return 1;
 958
 959         /* Just a random prefix match */
 960         return 0;
 961 }
 962
 963 /*
 964  * Perform matching on the leading non-wildcard part of
 965  * pathspec. item->nowildcard_len must be greater than zero. Return
 966  * non-zero if base is matched.
 967  */
 968 static int match_wildcard_base(const struct pathspec_item *item,
 969                                const char *base, int baselen,
 970                                int *matched)
 971 {
 972         const char *match = item->match;
 973         /* the wildcard part is not considered in this function */
 974         int matchlen = item->nowildcard_len;
 975
 976         if (baselen) {
 977                 int dirlen;
 978                 /*
 979                  * Return early if base is longer than the
 980                  * non-wildcard part but it does not match.
 981                  */
 982                 if (baselen >= matchlen) {
 983                         *matched = matchlen;
 984                         return !basecmp(item, base, match, matchlen);
 985                 }
 986
 987                 dirlen = matchlen;
 988                 while (dirlen && match[dirlen - 1] != '/')
 989                         dirlen--;
 990
 991                 /*
 992                  * Return early if base is shorter than the
 993                  * non-wildcard part but it does not match. Note that
 994                  * base ends with '/' so we are sure it really matches
 995                  * directory
 996                  */
 997                 if (basecmp(item, base, match, baselen))
 998                         return 0;
 999                 *matched = baselen;
1000         } else
1001                 *matched = 0;
1002         /*
1003          * we could have checked entry against the non-wildcard part
1004          * that is not in base and does similar never_interesting
1005          * optimization as in match_entry. For now just be happy with
1006          * base comparison.
1007          */
1008         return entry_interesting;
1009 }
1010
1011 /*
1012  * Is a tree entry interesting given the pathspec we have?
1013  *
1014  * Pre-condition: either baselen == 0 (i.e. empty path)
1015  * or base[baselen-1] == '/' (i.e. with trailing slash).
1016  */
1017 static enum interesting do_match(struct index_state *istate,
1018                                  const struct name_entry *entry,
1019                                  struct strbuf *base,
1020                                  const struct pathspec *ps,
1021                                  int exclude)
1022 {
1023         int i;
1024         int pathlen, baselen = base->len;
1025         enum interesting never_interesting = ps->has_wildcard ?
1026                 entry_not_interesting : all_entries_not_interesting;
1027
1028         GUARD_PATHSPEC(ps,
1029                        PATHSPEC_FROMTOP |
1030                        PATHSPEC_MAXDEPTH |
1031                        PATHSPEC_LITERAL |
1032                        PATHSPEC_GLOB |
1033                        PATHSPEC_ICASE |
1034                        PATHSPEC_EXCLUDE |
1035                        PATHSPEC_ATTR);
1036
1037         if (!ps->nr) {
1038                 if (!ps->recursive ||
1039                     !(ps->magic & PATHSPEC_MAXDEPTH) ||
1040                     ps->max_depth == -1)
1041                         return all_entries_interesting;
1042                 return within_depth(base->buf, baselen,
1043                                     !!S_ISDIR(entry->mode),
1044                                     ps->max_depth) ?
1045                         entry_interesting : entry_not_interesting;
1046         }
1047
1048         pathlen = tree_entry_len(entry);
1049
1050         for (i = ps->nr - 1; i >= 0; i--) {
1051                 const struct pathspec_item *item = ps->items+i;
1052                 const char *match = item->match;
1053                 const char *base_str = base->buf;
1054                 int matchlen = item->len, matched = 0;
1055
1056                 if ((!exclude &&   item->magic & PATHSPEC_EXCLUDE) ||
1057                     ( exclude && !(item->magic & PATHSPEC_EXCLUDE)))
1058                         continue;
1059
1060                 if (baselen >= matchlen) {
1061                         /* If it doesn't match, move along... */
1062                         if (!match_dir_prefix(item, base_str, match, matchlen))
1063                                 goto match_wildcards;
1064
1065                         if (!ps->recursive ||
1066                             !(ps->magic & PATHSPEC_MAXDEPTH) ||
1067                             ps->max_depth == -1) {
1068                                 if (!item->attr_match_nr)
1069                                         return all_entries_interesting;
1070                                 else
1071                                         goto interesting;
1072                         }
1073
1074                         if (within_depth(base_str + matchlen + 1,
1075                                          baselen - matchlen - 1,
1076                                          !!S_ISDIR(entry->mode),
1077                                          ps->max_depth))
1078                                 goto interesting;
1079                         else
1080                                 return entry_not_interesting;
1081                 }
1082
1083                 /* Either there must be no base, or the base must match. */
1084                 if (baselen == 0 || !basecmp(item, base_str, match, baselen)) {
1085                         if (match_entry(item, entry, pathlen,
1086                                         match + baselen, matchlen - baselen,
1087                                         &never_interesting))
1088                                 goto interesting;
1089
1090                         if (item->nowildcard_len < item->len) {
1091                                 if (!git_fnmatch(item, match + baselen, entry->path,
1092                                                  item->nowildcard_len - baselen))
1093                                         goto interesting;
1094
1095                                 /*
1096                                  * Match all directories. We'll try to
1097                                  * match files later on.
1098                                  */
1099                                 if (ps->recursive && S_ISDIR(entry->mode))
1100                                         return entry_interesting;
1101
1102                                 /*
1103                                  * When matching against submodules with
1104                                  * wildcard characters, ensure that the entry
1105                                  * at least matches up to the first wild
1106                                  * character.  More accurate matching can then
1107                                  * be performed in the submodule itself.
1108                                  */
1109                                 if (ps->recurse_submodules &&
1110                                     S_ISGITLINK(entry->mode) &&
1111                                     !ps_strncmp(item, match + baselen,
1112                                                 entry->path,
1113                                                 item->nowildcard_len - baselen))
1114                                         goto interesting;
1115                         }
1116
1117                         continue;
1118                 }
1119
1120 match_wildcards:
1121                 if (item->nowildcard_len == item->len)
1122                         continue;
1123
1124                 if (item->nowildcard_len &&
1125                     !match_wildcard_base(item, base_str, baselen, &matched))
1126                         continue;
1127
1128                 /*
1129                  * Concatenate base and entry->path into one and do
1130                  * fnmatch() on it.
1131                  *
1132                  * While we could avoid concatenation in certain cases
1133                  * [1], which saves a memcpy and potentially a
1134                  * realloc, it turns out not worth it. Measurement on
1135                  * linux-2.6 does not show any clear improvements,
1136                  * partly because of the nowildcard_len optimization
1137                  * in git_fnmatch(). Avoid micro-optimizations here.
1138                  *
1139                  * [1] if match_wildcard_base() says the base
1140                  * directory is already matched, we only need to match
1141                  * the rest, which is shorter so _in theory_ faster.
1142                  */
1143
1144                 strbuf_add(base, entry->path, pathlen);
1145
1146                 if (!git_fnmatch(item, match, base->buf,
1147                                  item->nowildcard_len)) {
1148                         strbuf_setlen(base, baselen);
1149                         goto interesting;
1150                 }
1151
1152                 /*
1153                  * When matching against submodules with
1154                  * wildcard characters, ensure that the entry
1155                  * at least matches up to the first wild
1156                  * character.  More accurate matching can then
1157                  * be performed in the submodule itself.
1158                  */
1159                 if (ps->recurse_submodules && S_ISGITLINK(entry->mode) &&
1160                     !ps_strncmp(item, match, base->buf,
1161                                 item->nowildcard_len)) {
1162                         strbuf_setlen(base, baselen);
1163                         goto interesting;
1164                 }
1165
1166                 strbuf_setlen(base, baselen);
1167
1168                 /*
1169                  * Match all directories. We'll try to match files
1170                  * later on.
1171                  * max_depth is ignored but we may consider support it
1172                  * in future, see
1173                  * https://lore.kernel.org/git/7vmxo5l2g4.fsf@alter.siamese.dyndns.org/
1174                  */
1175                 if (ps->recursive && S_ISDIR(entry->mode))
1176                         return entry_interesting;
1177                 continue;
1178 interesting:
1179                 if (item->attr_match_nr) {
1180                         int ret;
1181
1182                         /*
1183                          * Must not return all_entries_not_interesting
1184                          * prematurely. We do not know if all entries do not
1185                          * match some attributes with current attr API.
1186                          */
1187                         never_interesting = entry_not_interesting;
1188
1189                         /*
1190                          * Consider all directories interesting (because some
1191                          * of those files inside may match some attributes
1192                          * even though the parent dir does not)
1193                          *
1194                          * FIXME: attributes _can_ match directories and we
1195                          * can probably return all_entries_interesting or
1196                          * all_entries_not_interesting here if matched.
1197                          */
1198                         if (S_ISDIR(entry->mode))
1199                                 return entry_interesting;
1200
1201                         strbuf_add(base, entry->path, pathlen);
1202                         ret = match_pathspec_attrs(istate, base->buf,
1203                                                    base->len, item);
1204                         strbuf_setlen(base, baselen);
1205                         if (!ret)
1206                                 continue;
1207                 }
1208                 return entry_interesting;
1209         }
1210         return never_interesting; /* No matches */
1211 }
1212
1213 /*
1214  * Is a tree entry interesting given the pathspec we have?
1215  *
1216  * Pre-condition: either baselen == 0 (i.e. empty path)
1217  * or base[baselen-1] == '/' (i.e. with trailing slash).
1218  */
1219 enum interesting tree_entry_interesting(struct index_state *istate,
1220                                         const struct name_entry *entry,
1221                                         struct strbuf *base,
1222                                         const struct pathspec *ps)
1223 {
1224         enum interesting positive, negative;
1225         positive = do_match(istate, entry, base, ps, 0);
1226
1227         /*
1228          * case | entry | positive | negative | result
1229          * -----+-------+----------+----------+-------
1230          *   1  |  file |   -1     |  -1..2   |  -1
1231          *   2  |  file |    0     |  -1..2   |   0
1232          *   3  |  file |    1     |   -1     |   1
1233          *   4  |  file |    1     |    0     |   1
1234          *   5  |  file |    1     |    1     |   0
1235          *   6  |  file |    1     |    2     |   0
1236          *   7  |  file |    2     |   -1     |   2
1237          *   8  |  file |    2     |    0     |   1
1238          *   9  |  file |    2     |    1     |   0
1239          *  10  |  file |    2     |    2     |  -1
1240          * -----+-------+----------+----------+-------
1241          *  11  |  dir  |   -1     |  -1..2   |  -1
1242          *  12  |  dir  |    0     |  -1..2   |   0
1243          *  13  |  dir  |    1     |   -1     |   1
1244          *  14  |  dir  |    1     |    0     |   1
1245          *  15  |  dir  |    1     |    1     |   1 (*)
1246          *  16  |  dir  |    1     |    2     |   0
1247          *  17  |  dir  |    2     |   -1     |   2
1248          *  18  |  dir  |    2     |    0     |   1
1249          *  19  |  dir  |    2     |    1     |   1 (*)
1250          *  20  |  dir  |    2     |    2     |  -1
1251          *
1252          * (*) An exclude pattern interested in a directory does not
1253          * necessarily mean it will exclude all of the directory. In
1254          * wildcard case, it can't decide until looking at individual
1255          * files inside. So don't write such directories off yet.
1256          */
1257
1258         if (!(ps->magic & PATHSPEC_EXCLUDE) ||
1259             positive <= entry_not_interesting) /* #1, #2, #11, #12 */
1260                 return positive;
1261
1262         negative = do_match(istate, entry, base, ps, 1);
1263
1264         /* #8, #18 */
1265         if (positive == all_entries_interesting &&
1266             negative == entry_not_interesting)
1267                 return entry_interesting;
1268
1269         /* #3, #4, #7, #13, #14, #17 */
1270         if (negative <= entry_not_interesting)
1271                 return positive;
1272
1273         /* #15, #19 */
1274         if (S_ISDIR(entry->mode) &&
1275             positive >= entry_interesting &&
1276             negative == entry_interesting)
1277                 return entry_interesting;
1278
1279         if ((positive == entry_interesting &&
1280              negative >= entry_interesting) || /* #5, #6, #16 */
1281             (positive == all_entries_interesting &&
1282              negative == entry_interesting)) /* #9 */
1283                 return entry_not_interesting;
1284
1285         return all_entries_not_interesting; /* #10, #20 */
1286 }