tree-walk.c

   1 #include "cache.h"
   2 #include "tree-walk.h"
   3 #include "unpack-trees.h"
   4 #include "dir.h"
   5 #include "tree.h"
   6 #include "pathspec.h"
   7
   8 static const char *get_mode(const char *str, unsigned int *modep)
   9 {
  10         unsigned char c;
  11         unsigned int mode = 0;
  12
  13         if (*str == ' ')
  14                 return NULL;
  15
  16         while ((c = *str++) != ' ') {
  17                 if (c < '0' || c > '7')
  18                         return NULL;
  19                 mode = (mode << 3) + (c - '0');
  20         }
  21         *modep = mode;
  22         return str;
  23 }
  24
  25 static void decode_tree_entry(struct tree_desc *desc, const char *buf, unsigned long size)
  26 {
  27         const char *path;
  28         unsigned int mode, len;
  29
  30         if (size < 24 || buf[size - 21])
  31                 die("corrupt tree file");
  32
  33         path = get_mode(buf, &mode);
  34         if (!path || !*path)
  35                 die("corrupt tree file");
  36         len = strlen(path) + 1;
  37
  38         /* Initialize the descriptor entry */
  39         desc->entry.path = path;
  40         desc->entry.mode = canon_mode(mode);
  41         desc->entry.oid  = (const struct object_id *)(path + len);
  42 }
  43
  44 void init_tree_desc(struct tree_desc *desc, const void *buffer, unsigned long size)
  45 {
  46         desc->buffer = buffer;
  47         desc->size = size;
  48         if (size)
  49                 decode_tree_entry(desc, buffer, size);
  50 }
  51
  52 void *fill_tree_descriptor(struct tree_desc *desc, const unsigned char *sha1)
  53 {
  54         unsigned long size = 0;
  55         void *buf = NULL;
  56
  57         if (sha1) {
  58                 buf = read_object_with_reference(sha1, tree_type, &size, NULL);
  59                 if (!buf)
  60                         die("unable to read tree %s", sha1_to_hex(sha1));
  61         }
  62         init_tree_desc(desc, buf, size);
  63         return buf;
  64 }
  65
  66 static void entry_clear(struct name_entry *a)
  67 {
  68         memset(a, 0, sizeof(*a));
  69 }
  70
  71 static void entry_extract(struct tree_desc *t, struct name_entry *a)
  72 {
  73         *a = t->entry;
  74 }
  75
  76 void update_tree_entry(struct tree_desc *desc)
  77 {
  78         const void *buf = desc->buffer;
  79         const unsigned char *end = desc->entry.oid->hash + 20;
  80         unsigned long size = desc->size;
  81         unsigned long len = end - (const unsigned char *)buf;
  82
  83         if (size < len)
  84                 die("corrupt tree file");
  85         buf = end;
  86         size -= len;
  87         desc->buffer = buf;
  88         desc->size = size;
  89         if (size)
  90                 decode_tree_entry(desc, buf, size);
  91 }
  92
  93 int tree_entry(struct tree_desc *desc, struct name_entry *entry)
  94 {
  95         if (!desc->size)
  96                 return 0;
  97
  98         *entry = desc->entry;
  99         update_tree_entry(desc);
 100         return 1;
 101 }
 102
 103 void setup_traverse_info(struct traverse_info *info, const char *base)
 104 {
 105         int pathlen = strlen(base);
 106         static struct traverse_info dummy;
 107
 108         memset(info, 0, sizeof(*info));
 109         if (pathlen && base[pathlen-1] == '/')
 110                 pathlen--;
 111         info->pathlen = pathlen ? pathlen + 1 : 0;
 112         info->name.path = base;
 113         info->name.oid = (void *)(base + pathlen + 1);
 114         if (pathlen)
 115                 info->prev = &dummy;
 116 }
 117
 118 char *make_traverse_path(char *path, const struct traverse_info *info, const struct name_entry *n)
 119 {
 120         int len = tree_entry_len(n);
 121         int pathlen = info->pathlen;
 122
 123         path[pathlen + len] = 0;
 124         for (;;) {
 125                 memcpy(path + pathlen, n->path, len);
 126                 if (!pathlen)
 127                         break;
 128                 path[--pathlen] = '/';
 129                 n = &info->name;
 130                 len = tree_entry_len(n);
 131                 info = info->prev;
 132                 pathlen -= len;
 133         }
 134         return path;
 135 }
 136
 137 struct tree_desc_skip {
 138         struct tree_desc_skip *prev;
 139         const void *ptr;
 140 };
 141
 142 struct tree_desc_x {
 143         struct tree_desc d;
 144         struct tree_desc_skip *skip;
 145 };
 146
 147 static int check_entry_match(const char *a, int a_len, const char *b, int b_len)
 148 {
 149         /*
 150          * The caller wants to pick *a* from a tree or nothing.
 151          * We are looking at *b* in a tree.
 152          *
 153          * (0) If a and b are the same name, we are trivially happy.
 154          *
 155          * There are three possibilities where *a* could be hiding
 156          * behind *b*.
 157          *
 158          * (1) *a* == "t",   *b* == "ab"  i.e. *b* sorts earlier than *a* no
 159          *                                matter what.
 160          * (2) *a* == "t",   *b* == "t-2" and "t" is a subtree in the tree;
 161          * (3) *a* == "t-2", *b* == "t"   and "t-2" is a blob in the tree.
 162          *
 163          * Otherwise we know *a* won't appear in the tree without
 164          * scanning further.
 165          */
 166
 167         int cmp = name_compare(a, a_len, b, b_len);
 168
 169         /* Most common case first -- reading sync'd trees */
 170         if (!cmp)
 171                 return cmp;
 172
 173         if (0 < cmp) {
 174                 /* a comes after b; it does not matter if it is case (3)
 175                 if (b_len < a_len && !memcmp(a, b, b_len) && a[b_len] < '/')
 176                         return 1;
 177                 */
 178                 return 1; /* keep looking */
 179         }
 180
 181         /* b comes after a; are we looking at case (2)? */
 182         if (a_len < b_len && !memcmp(a, b, a_len) && b[a_len] < '/')
 183                 return 1; /* keep looking */
 184
 185         return -1; /* a cannot appear in the tree */
 186 }
 187
 188 /*
 189  * From the extended tree_desc, extract the first name entry, while
 190  * paying attention to the candidate "first" name.  Most importantly,
 191  * when looking for an entry, if there are entries that sorts earlier
 192  * in the tree object representation than that name, skip them and
 193  * process the named entry first.  We will remember that we haven't
 194  * processed the first entry yet, and in the later call skip the
 195  * entry we processed early when update_extended_entry() is called.
 196  *
 197  * E.g. if the underlying tree object has these entries:
 198  *
 199  *    blob    "t-1"
 200  *    blob    "t-2"
 201  *    tree    "t"
 202  *    blob    "t=1"
 203  *
 204  * and the "first" asks for "t", remember that we still need to
 205  * process "t-1" and "t-2" but extract "t".  After processing the
 206  * entry "t" from this call, the caller will let us know by calling
 207  * update_extended_entry() that we can remember "t" has been processed
 208  * already.
 209  */
 210
 211 static void extended_entry_extract(struct tree_desc_x *t,
 212                                    struct name_entry *a,
 213                                    const char *first,
 214                                    int first_len)
 215 {
 216         const char *path;
 217         int len;
 218         struct tree_desc probe;
 219         struct tree_desc_skip *skip;
 220
 221         /*
 222          * Extract the first entry from the tree_desc, but skip the
 223          * ones that we already returned in earlier rounds.
 224          */
 225         while (1) {
 226                 if (!t->d.size) {
 227                         entry_clear(a);
 228                         break; /* not found */
 229                 }
 230                 entry_extract(&t->d, a);
 231                 for (skip = t->skip; skip; skip = skip->prev)
 232                         if (a->path == skip->ptr)
 233                                 break; /* found */
 234                 if (!skip)
 235                         break;
 236                 /* We have processed this entry already. */
 237                 update_tree_entry(&t->d);
 238         }
 239
 240         if (!first || !a->path)
 241                 return;
 242
 243         /*
 244          * The caller wants "first" from this tree, or nothing.
 245          */
 246         path = a->path;
 247         len = tree_entry_len(a);
 248         switch (check_entry_match(first, first_len, path, len)) {
 249         case -1:
 250                 entry_clear(a);
 251         case 0:
 252                 return;
 253         default:
 254                 break;
 255         }
 256
 257         /*
 258          * We need to look-ahead -- we suspect that a subtree whose
 259          * name is "first" may be hiding behind the current entry "path".
 260          */
 261         probe = t->d;
 262         while (probe.size) {
 263                 entry_extract(&probe, a);
 264                 path = a->path;
 265                 len = tree_entry_len(a);
 266                 switch (check_entry_match(first, first_len, path, len)) {
 267                 case -1:
 268                         entry_clear(a);
 269                 case 0:
 270                         return;
 271                 default:
 272                         update_tree_entry(&probe);
 273                         break;
 274                 }
 275                 /* keep looking */
 276         }
 277         entry_clear(a);
 278 }
 279
 280 static void update_extended_entry(struct tree_desc_x *t, struct name_entry *a)
 281 {
 282         if (t->d.entry.path == a->path) {
 283                 update_tree_entry(&t->d);
 284         } else {
 285                 /* we have returned this entry early */
 286                 struct tree_desc_skip *skip = xmalloc(sizeof(*skip));
 287                 skip->ptr = a->path;
 288                 skip->prev = t->skip;
 289                 t->skip = skip;
 290         }
 291 }
 292
 293 static void free_extended_entry(struct tree_desc_x *t)
 294 {
 295         struct tree_desc_skip *p, *s;
 296
 297         for (s = t->skip; s; s = p) {
 298                 p = s->prev;
 299                 free(s);
 300         }
 301 }
 302
 303 static inline int prune_traversal(struct name_entry *e,
 304                                   struct traverse_info *info,
 305                                   struct strbuf *base,
 306                                   int still_interesting)
 307 {
 308         if (!info->pathspec || still_interesting == 2)
 309                 return 2;
 310         if (still_interesting < 0)
 311                 return still_interesting;
 312         return tree_entry_interesting(e, base, 0, info->pathspec);
 313 }
 314
 315 int traverse_trees(int n, struct tree_desc *t, struct traverse_info *info)
 316 {
 317         int error = 0;
 318         struct name_entry *entry = xmalloc(n*sizeof(*entry));
 319         int i;
 320         struct tree_desc_x *tx = xcalloc(n, sizeof(*tx));
 321         struct strbuf base = STRBUF_INIT;
 322         int interesting = 1;
 323         char *traverse_path;
 324
 325         for (i = 0; i < n; i++)
 326                 tx[i].d = t[i];
 327
 328         if (info->prev) {
 329                 strbuf_grow(&base, info->pathlen);
 330                 make_traverse_path(base.buf, info->prev, &info->name);
 331                 base.buf[info->pathlen-1] = '/';
 332                 strbuf_setlen(&base, info->pathlen);
 333                 traverse_path = xstrndup(base.buf, info->pathlen);
 334         } else {
 335                 traverse_path = xstrndup(info->name.path, info->pathlen);
 336         }
 337         info->traverse_path = traverse_path;
 338         for (;;) {
 339                 int trees_used;
 340                 unsigned long mask, dirmask;
 341                 const char *first = NULL;
 342                 int first_len = 0;
 343                 struct name_entry *e = NULL;
 344                 int len;
 345
 346                 for (i = 0; i < n; i++) {
 347                         e = entry + i;
 348                         extended_entry_extract(tx + i, e, NULL, 0);
 349                 }
 350
 351                 /*
 352                  * A tree may have "t-2" at the current location even
 353                  * though it may have "t" that is a subtree behind it,
 354                  * and another tree may return "t".  We want to grab
 355                  * all "t" from all trees to match in such a case.
 356                  */
 357                 for (i = 0; i < n; i++) {
 358                         e = entry + i;
 359                         if (!e->path)
 360                                 continue;
 361                         len = tree_entry_len(e);
 362                         if (!first) {
 363                                 first = e->path;
 364                                 first_len = len;
 365                                 continue;
 366                         }
 367                         if (name_compare(e->path, len, first, first_len) < 0) {
 368                                 first = e->path;
 369                                 first_len = len;
 370                         }
 371                 }
 372
 373                 if (first) {
 374                         for (i = 0; i < n; i++) {
 375                                 e = entry + i;
 376                                 extended_entry_extract(tx + i, e, first, first_len);
 377                                 /* Cull the ones that are not the earliest */
 378                                 if (!e->path)
 379                                         continue;
 380                                 len = tree_entry_len(e);
 381                                 if (name_compare(e->path, len, first, first_len))
 382                                         entry_clear(e);
 383                         }
 384                 }
 385
 386                 /* Now we have in entry[i] the earliest name from the trees */
 387                 mask = 0;
 388                 dirmask = 0;
 389                 for (i = 0; i < n; i++) {
 390                         if (!entry[i].path)
 391                                 continue;
 392                         mask |= 1ul << i;
 393                         if (S_ISDIR(entry[i].mode))
 394                                 dirmask |= 1ul << i;
 395                         e = &entry[i];
 396                 }
 397                 if (!mask)
 398                         break;
 399                 interesting = prune_traversal(e, info, &base, interesting);
 400                 if (interesting < 0)
 401                         break;
 402                 if (interesting) {
 403                         trees_used = info->fn(n, mask, dirmask, entry, info);
 404                         if (trees_used < 0) {
 405                                 error = trees_used;
 406                                 if (!info->show_all_errors)
 407                                         break;
 408                         }
 409                         mask &= trees_used;
 410                 }
 411                 for (i = 0; i < n; i++)
 412                         if (mask & (1ul << i))
 413                                 update_extended_entry(tx + i, entry + i);
 414         }
 415         free(entry);
 416         for (i = 0; i < n; i++)
 417                 free_extended_entry(tx + i);
 418         free(tx);
 419         free(traverse_path);
 420         info->traverse_path = NULL;
 421         strbuf_release(&base);
 422         return error;
 423 }
 424
 425 struct dir_state {
 426         void *tree;
 427         unsigned long size;
 428         unsigned char sha1[20];
 429 };
 430
 431 static int find_tree_entry(struct tree_desc *t, const char *name, unsigned char *result, unsigned *mode)
 432 {
 433         int namelen = strlen(name);
 434         while (t->size) {
 435                 const char *entry;
 436                 const struct object_id *oid;
 437                 int entrylen, cmp;
 438
 439                 oid = tree_entry_extract(t, &entry, mode);
 440                 entrylen = tree_entry_len(&t->entry);
 441                 update_tree_entry(t);
 442                 if (entrylen > namelen)
 443                         continue;
 444                 cmp = memcmp(name, entry, entrylen);
 445                 if (cmp > 0)
 446                         continue;
 447                 if (cmp < 0)
 448                         break;
 449                 if (entrylen == namelen) {
 450                         hashcpy(result, oid->hash);
 451                         return 0;
 452                 }
 453                 if (name[entrylen] != '/')
 454                         continue;
 455                 if (!S_ISDIR(*mode))
 456                         break;
 457                 if (++entrylen == namelen) {
 458                         hashcpy(result, oid->hash);
 459                         return 0;
 460                 }
 461                 return get_tree_entry(oid->hash, name + entrylen, result, mode);
 462         }
 463         return -1;
 464 }
 465
 466 int get_tree_entry(const unsigned char *tree_sha1, const char *name, unsigned char *sha1, unsigned *mode)
 467 {
 468         int retval;
 469         void *tree;
 470         unsigned long size;
 471         unsigned char root[20];
 472
 473         tree = read_object_with_reference(tree_sha1, tree_type, &size, root);
 474         if (!tree)
 475                 return -1;
 476
 477         if (name[0] == '\0') {
 478                 hashcpy(sha1, root);
 479                 free(tree);
 480                 return 0;
 481         }
 482
 483         if (!size) {
 484                 retval = -1;
 485         } else {
 486                 struct tree_desc t;
 487                 init_tree_desc(&t, tree, size);
 488                 retval = find_tree_entry(&t, name, sha1, mode);
 489         }
 490         free(tree);
 491         return retval;
 492 }
 493
 494 /*
 495  * This is Linux's built-in max for the number of symlinks to follow.
 496  * That limit, of course, does not affect git, but it's a reasonable
 497  * choice.
 498  */
 499 #define GET_TREE_ENTRY_FOLLOW_SYMLINKS_MAX_LINKS 40
 500
 501 /**
 502  * Find a tree entry by following symlinks in tree_sha (which is
 503  * assumed to be the root of the repository).  In the event that a
 504  * symlink points outside the repository (e.g. a link to /foo or a
 505  * root-level link to ../foo), the portion of the link which is
 506  * outside the repository will be returned in result_path, and *mode
 507  * will be set to 0.  It is assumed that result_path is uninitialized.
 508  * If there are no symlinks, or the end result of the symlink chain
 509  * points to an object inside the repository, result will be filled in
 510  * with the sha1 of the found object, and *mode will hold the mode of
 511  * the object.
 512  *
 513  * See the code for enum follow_symlink_result for a description of
 514  * the return values.
 515  */
 516 enum follow_symlinks_result get_tree_entry_follow_symlinks(unsigned char *tree_sha1, const char *name, unsigned char *result, struct strbuf *result_path, unsigned *mode)
 517 {
 518         int retval = MISSING_OBJECT;
 519         struct dir_state *parents = NULL;
 520         size_t parents_alloc = 0;
 521         ssize_t parents_nr = 0;
 522         unsigned char current_tree_sha1[20];
 523         struct strbuf namebuf = STRBUF_INIT;
 524         struct tree_desc t;
 525         int follows_remaining = GET_TREE_ENTRY_FOLLOW_SYMLINKS_MAX_LINKS;
 526         int i;
 527
 528         init_tree_desc(&t, NULL, 0UL);
 529         strbuf_init(result_path, 0);
 530         strbuf_addstr(&namebuf, name);
 531         hashcpy(current_tree_sha1, tree_sha1);
 532
 533         while (1) {
 534                 int find_result;
 535                 char *first_slash;
 536                 char *remainder = NULL;
 537
 538                 if (!t.buffer) {
 539                         void *tree;
 540                         unsigned char root[20];
 541                         unsigned long size;
 542                         tree = read_object_with_reference(current_tree_sha1,
 543                                                           tree_type, &size,
 544                                                           root);
 545                         if (!tree)
 546                                 goto done;
 547
 548                         ALLOC_GROW(parents, parents_nr + 1, parents_alloc);
 549                         parents[parents_nr].tree = tree;
 550                         parents[parents_nr].size = size;
 551                         hashcpy(parents[parents_nr].sha1, root);
 552                         parents_nr++;
 553
 554                         if (namebuf.buf[0] == '\0') {
 555                                 hashcpy(result, root);
 556                                 retval = FOUND;
 557                                 goto done;
 558                         }
 559
 560                         if (!size)
 561                                 goto done;
 562
 563                         /* descend */
 564                         init_tree_desc(&t, tree, size);
 565                 }
 566
 567                 /* Handle symlinks to e.g. a//b by removing leading slashes */
 568                 while (namebuf.buf[0] == '/') {
 569                         strbuf_remove(&namebuf, 0, 1);
 570                 }
 571
 572                 /* Split namebuf into a first component and a remainder */
 573                 if ((first_slash = strchr(namebuf.buf, '/'))) {
 574                         *first_slash = 0;
 575                         remainder = first_slash + 1;
 576                 }
 577
 578                 if (!strcmp(namebuf.buf, "..")) {
 579                         struct dir_state *parent;
 580                         /*
 581                          * We could end up with .. in the namebuf if it
 582                          * appears in a symlink.
 583                          */
 584
 585                         if (parents_nr == 1) {
 586                                 if (remainder)
 587                                         *first_slash = '/';
 588                                 strbuf_add(result_path, namebuf.buf,
 589                                            namebuf.len);
 590                                 *mode = 0;
 591                                 retval = FOUND;
 592                                 goto done;
 593                         }
 594                         parent = &parents[parents_nr - 1];
 595                         free(parent->tree);
 596                         parents_nr--;
 597                         parent = &parents[parents_nr - 1];
 598                         init_tree_desc(&t, parent->tree, parent->size);
 599                         strbuf_remove(&namebuf, 0, remainder ? 3 : 2);
 600                         continue;
 601                 }
 602
 603                 /* We could end up here via a symlink to dir/.. */
 604                 if (namebuf.buf[0] == '\0') {
 605                         hashcpy(result, parents[parents_nr - 1].sha1);
 606                         retval = FOUND;
 607                         goto done;
 608                 }
 609
 610                 /* Look up the first (or only) path component in the tree. */
 611                 find_result = find_tree_entry(&t, namebuf.buf,
 612                                               current_tree_sha1, mode);
 613                 if (find_result) {
 614                         goto done;
 615                 }
 616
 617                 if (S_ISDIR(*mode)) {
 618                         if (!remainder) {
 619                                 hashcpy(result, current_tree_sha1);
 620                                 retval = FOUND;
 621                                 goto done;
 622                         }
 623                         /* Descend the tree */
 624                         t.buffer = NULL;
 625                         strbuf_remove(&namebuf, 0,
 626                                       1 + first_slash - namebuf.buf);
 627                 } else if (S_ISREG(*mode)) {
 628                         if (!remainder) {
 629                                 hashcpy(result, current_tree_sha1);
 630                                 retval = FOUND;
 631                         } else {
 632                                 retval = NOT_DIR;
 633                         }
 634                         goto done;
 635                 } else if (S_ISLNK(*mode)) {
 636                         /* Follow a symlink */
 637                         unsigned long link_len;
 638                         size_t len;
 639                         char *contents, *contents_start;
 640                         struct dir_state *parent;
 641                         enum object_type type;
 642
 643                         if (follows_remaining-- == 0) {
 644                                 /* Too many symlinks followed */
 645                                 retval = SYMLINK_LOOP;
 646                                 goto done;
 647                         }
 648
 649                         /*
 650                          * At this point, we have followed at a least
 651                          * one symlink, so on error we need to report this.
 652                          */
 653                         retval = DANGLING_SYMLINK;
 654
 655                         contents = read_sha1_file(current_tree_sha1, &type,
 656                                                   &link_len);
 657
 658                         if (!contents)
 659                                 goto done;
 660
 661                         if (contents[0] == '/') {
 662                                 strbuf_addstr(result_path, contents);
 663                                 free(contents);
 664                                 *mode = 0;
 665                                 retval = FOUND;
 666                                 goto done;
 667                         }
 668
 669                         if (remainder)
 670                                 len = first_slash - namebuf.buf;
 671                         else
 672                                 len = namebuf.len;
 673
 674                         contents_start = contents;
 675
 676                         parent = &parents[parents_nr - 1];
 677                         init_tree_desc(&t, parent->tree, parent->size);
 678                         strbuf_splice(&namebuf, 0, len,
 679                                       contents_start, link_len);
 680                         if (remainder)
 681                                 namebuf.buf[link_len] = '/';
 682                         free(contents);
 683                 }
 684         }
 685 done:
 686         for (i = 0; i < parents_nr; i++)
 687                 free(parents[i].tree);
 688         free(parents);
 689
 690         strbuf_release(&namebuf);
 691         return retval;
 692 }
 693
 694 static int match_entry(const struct pathspec_item *item,
 695                        const struct name_entry *entry, int pathlen,
 696                        const char *match, int matchlen,
 697                        enum interesting *never_interesting)
 698 {
 699         int m = -1; /* signals that we haven't called strncmp() */
 700
 701         if (item->magic & PATHSPEC_ICASE)
 702                 /*
 703                  * "Never interesting" trick requires exact
 704                  * matching. We could do something clever with inexact
 705                  * matching, but it's trickier (and not to forget that
 706                  * strcasecmp is locale-dependent, at least in
 707                  * glibc). Just disable it for now. It can't be worse
 708                  * than the wildcard's codepath of '[Tt][Hi][Is][Ss]'
 709                  * pattern.
 710                  */
 711                 *never_interesting = entry_not_interesting;
 712         else if (*never_interesting != entry_not_interesting) {
 713                 /*
 714                  * We have not seen any match that sorts later
 715                  * than the current path.
 716                  */
 717
 718                 /*
 719                  * Does match sort strictly earlier than path
 720                  * with their common parts?
 721                  */
 722                 m = strncmp(match, entry->path,
 723                             (matchlen < pathlen) ? matchlen : pathlen);
 724                 if (m < 0)
 725                         return 0;
 726
 727                 /*
 728                  * If we come here even once, that means there is at
 729                  * least one pathspec that would sort equal to or
 730                  * later than the path we are currently looking at.
 731                  * In other words, if we have never reached this point
 732                  * after iterating all pathspecs, it means all
 733                  * pathspecs are either outside of base, or inside the
 734                  * base but sorts strictly earlier than the current
 735                  * one.  In either case, they will never match the
 736                  * subsequent entries.  In such a case, we initialized
 737                  * the variable to -1 and that is what will be
 738                  * returned, allowing the caller to terminate early.
 739                  */
 740                 *never_interesting = entry_not_interesting;
 741         }
 742
 743         if (pathlen > matchlen)
 744                 return 0;
 745
 746         if (matchlen > pathlen) {
 747                 if (match[pathlen] != '/')
 748                         return 0;
 749                 if (!S_ISDIR(entry->mode) && !S_ISGITLINK(entry->mode))
 750                         return 0;
 751         }
 752
 753         if (m == -1)
 754                 /*
 755                  * we cheated and did not do strncmp(), so we do
 756                  * that here.
 757                  */
 758                 m = ps_strncmp(item, match, entry->path, pathlen);
 759
 760         /*
 761          * If common part matched earlier then it is a hit,
 762          * because we rejected the case where path is not a
 763          * leading directory and is shorter than match.
 764          */
 765         if (!m)
 766                 /*
 767                  * match_entry does not check if the prefix part is
 768                  * matched case-sensitively. If the entry is a
 769                  * directory and part of prefix, it'll be rematched
 770                  * eventually by basecmp with special treatment for
 771                  * the prefix.
 772                  */
 773                 return 1;
 774
 775         return 0;
 776 }
 777
 778 /* :(icase)-aware string compare */
 779 static int basecmp(const struct pathspec_item *item,
 780                    const char *base, const char *match, int len)
 781 {
 782         if (item->magic & PATHSPEC_ICASE) {
 783                 int ret, n = len > item->prefix ? item->prefix : len;
 784                 ret = strncmp(base, match, n);
 785                 if (ret)
 786                         return ret;
 787                 base += n;
 788                 match += n;
 789                 len -= n;
 790         }
 791         return ps_strncmp(item, base, match, len);
 792 }
 793
 794 static int match_dir_prefix(const struct pathspec_item *item,
 795                             const char *base,
 796                             const char *match, int matchlen)
 797 {
 798         if (basecmp(item, base, match, matchlen))
 799                 return 0;
 800
 801         /*
 802          * If the base is a subdirectory of a path which
 803          * was specified, all of them are interesting.
 804          */
 805         if (!matchlen ||
 806             base[matchlen] == '/' ||
 807             match[matchlen - 1] == '/')
 808                 return 1;
 809
 810         /* Just a random prefix match */
 811         return 0;
 812 }
 813
 814 /*
 815  * Perform matching on the leading non-wildcard part of
 816  * pathspec. item->nowildcard_len must be greater than zero. Return
 817  * non-zero if base is matched.
 818  */
 819 static int match_wildcard_base(const struct pathspec_item *item,
 820                                const char *base, int baselen,
 821                                int *matched)
 822 {
 823         const char *match = item->match;
 824         /* the wildcard part is not considered in this function */
 825         int matchlen = item->nowildcard_len;
 826
 827         if (baselen) {
 828                 int dirlen;
 829                 /*
 830                  * Return early if base is longer than the
 831                  * non-wildcard part but it does not match.
 832                  */
 833                 if (baselen >= matchlen) {
 834                         *matched = matchlen;
 835                         return !basecmp(item, base, match, matchlen);
 836                 }
 837
 838                 dirlen = matchlen;
 839                 while (dirlen && match[dirlen - 1] != '/')
 840                         dirlen--;
 841
 842                 /*
 843                  * Return early if base is shorter than the
 844                  * non-wildcard part but it does not match. Note that
 845                  * base ends with '/' so we are sure it really matches
 846                  * directory
 847                  */
 848                 if (basecmp(item, base, match, baselen))
 849                         return 0;
 850                 *matched = baselen;
 851         } else
 852                 *matched = 0;
 853         /*
 854          * we could have checked entry against the non-wildcard part
 855          * that is not in base and does similar never_interesting
 856          * optimization as in match_entry. For now just be happy with
 857          * base comparison.
 858          */
 859         return entry_interesting;
 860 }
 861
 862 /*
 863  * Is a tree entry interesting given the pathspec we have?
 864  *
 865  * Pre-condition: either baselen == base_offset (i.e. empty path)
 866  * or base[baselen-1] == '/' (i.e. with trailing slash).
 867  */
 868 static enum interesting do_match(const struct name_entry *entry,
 869                                  struct strbuf *base, int base_offset,
 870                                  const struct pathspec *ps,
 871                                  int exclude)
 872 {
 873         int i;
 874         int pathlen, baselen = base->len - base_offset;
 875         enum interesting never_interesting = ps->has_wildcard ?
 876                 entry_not_interesting : all_entries_not_interesting;
 877
 878         GUARD_PATHSPEC(ps,
 879                        PATHSPEC_FROMTOP |
 880                        PATHSPEC_MAXDEPTH |
 881                        PATHSPEC_LITERAL |
 882                        PATHSPEC_GLOB |
 883                        PATHSPEC_ICASE |
 884                        PATHSPEC_EXCLUDE);
 885
 886         if (!ps->nr) {
 887                 if (!ps->recursive ||
 888                     !(ps->magic & PATHSPEC_MAXDEPTH) ||
 889                     ps->max_depth == -1)
 890                         return all_entries_interesting;
 891                 return within_depth(base->buf + base_offset, baselen,
 892                                     !!S_ISDIR(entry->mode),
 893                                     ps->max_depth) ?
 894                         entry_interesting : entry_not_interesting;
 895         }
 896
 897         pathlen = tree_entry_len(entry);
 898
 899         for (i = ps->nr - 1; i >= 0; i--) {
 900                 const struct pathspec_item *item = ps->items+i;
 901                 const char *match = item->match;
 902                 const char *base_str = base->buf + base_offset;
 903                 int matchlen = item->len, matched = 0;
 904
 905                 if ((!exclude &&   item->magic & PATHSPEC_EXCLUDE) ||
 906                     ( exclude && !(item->magic & PATHSPEC_EXCLUDE)))
 907                         continue;
 908
 909                 if (baselen >= matchlen) {
 910                         /* If it doesn't match, move along... */
 911                         if (!match_dir_prefix(item, base_str, match, matchlen))
 912                                 goto match_wildcards;
 913
 914                         if (!ps->recursive ||
 915                             !(ps->magic & PATHSPEC_MAXDEPTH) ||
 916                             ps->max_depth == -1)
 917                                 return all_entries_interesting;
 918
 919                         return within_depth(base_str + matchlen + 1,
 920                                             baselen - matchlen - 1,
 921                                             !!S_ISDIR(entry->mode),
 922                                             ps->max_depth) ?
 923                                 entry_interesting : entry_not_interesting;
 924                 }
 925
 926                 /* Either there must be no base, or the base must match. */
 927                 if (baselen == 0 || !basecmp(item, base_str, match, baselen)) {
 928                         if (match_entry(item, entry, pathlen,
 929                                         match + baselen, matchlen - baselen,
 930                                         &never_interesting))
 931                                 return entry_interesting;
 932
 933                         if (item->nowildcard_len < item->len) {
 934                                 if (!git_fnmatch(item, match + baselen, entry->path,
 935                                                  item->nowildcard_len - baselen))
 936                                         return entry_interesting;
 937
 938                                 /*
 939                                  * Match all directories. We'll try to
 940                                  * match files later on.
 941                                  */
 942                                 if (ps->recursive && S_ISDIR(entry->mode))
 943                                         return entry_interesting;
 944                         }
 945
 946                         continue;
 947                 }
 948
 949 match_wildcards:
 950                 if (item->nowildcard_len == item->len)
 951                         continue;
 952
 953                 if (item->nowildcard_len &&
 954                     !match_wildcard_base(item, base_str, baselen, &matched))
 955                         continue;
 956
 957                 /*
 958                  * Concatenate base and entry->path into one and do
 959                  * fnmatch() on it.
 960                  *
 961                  * While we could avoid concatenation in certain cases
 962                  * [1], which saves a memcpy and potentially a
 963                  * realloc, it turns out not worth it. Measurement on
 964                  * linux-2.6 does not show any clear improvements,
 965                  * partly because of the nowildcard_len optimization
 966                  * in git_fnmatch(). Avoid micro-optimizations here.
 967                  *
 968                  * [1] if match_wildcard_base() says the base
 969                  * directory is already matched, we only need to match
 970                  * the rest, which is shorter so _in theory_ faster.
 971                  */
 972
 973                 strbuf_add(base, entry->path, pathlen);
 974
 975                 if (!git_fnmatch(item, match, base->buf + base_offset,
 976                                  item->nowildcard_len)) {
 977                         strbuf_setlen(base, base_offset + baselen);
 978                         return entry_interesting;
 979                 }
 980                 strbuf_setlen(base, base_offset + baselen);
 981
 982                 /*
 983                  * Match all directories. We'll try to match files
 984                  * later on.
 985                  * max_depth is ignored but we may consider support it
 986                  * in future, see
 987                  * http://thread.gmane.org/gmane.comp.version-control.git/163757/focus=163840
 988                  */
 989                 if (ps->recursive && S_ISDIR(entry->mode))
 990                         return entry_interesting;
 991         }
 992         return never_interesting; /* No matches */
 993 }
 994
 995 /*
 996  * Is a tree entry interesting given the pathspec we have?
 997  *
 998  * Pre-condition: either baselen == base_offset (i.e. empty path)
 999  * or base[baselen-1] == '/' (i.e. with trailing slash).
1000  */
1001 enum interesting tree_entry_interesting(const struct name_entry *entry,
1002                                         struct strbuf *base, int base_offset,
1003                                         const struct pathspec *ps)
1004 {
1005         enum interesting positive, negative;
1006         positive = do_match(entry, base, base_offset, ps, 0);
1007
1008         /*
1009          * case | entry | positive | negative | result
1010          * -----+-------+----------+----------+-------
1011          *   1  |  file |   -1     |  -1..2   |  -1
1012          *   2  |  file |    0     |  -1..2   |   0
1013          *   3  |  file |    1     |   -1     |   1
1014          *   4  |  file |    1     |    0     |   1
1015          *   5  |  file |    1     |    1     |   0
1016          *   6  |  file |    1     |    2     |   0
1017          *   7  |  file |    2     |   -1     |   2
1018          *   8  |  file |    2     |    0     |   2
1019          *   9  |  file |    2     |    1     |   0
1020          *  10  |  file |    2     |    2     |  -1
1021          * -----+-------+----------+----------+-------
1022          *  11  |  dir  |   -1     |  -1..2   |  -1
1023          *  12  |  dir  |    0     |  -1..2   |   0
1024          *  13  |  dir  |    1     |   -1     |   1
1025          *  14  |  dir  |    1     |    0     |   1
1026          *  15  |  dir  |    1     |    1     |   1 (*)
1027          *  16  |  dir  |    1     |    2     |   0
1028          *  17  |  dir  |    2     |   -1     |   2
1029          *  18  |  dir  |    2     |    0     |   2
1030          *  19  |  dir  |    2     |    1     |   1 (*)
1031          *  20  |  dir  |    2     |    2     |  -1
1032          *
1033          * (*) An exclude pattern interested in a directory does not
1034          * necessarily mean it will exclude all of the directory. In
1035          * wildcard case, it can't decide until looking at individual
1036          * files inside. So don't write such directories off yet.
1037          */
1038
1039         if (!(ps->magic & PATHSPEC_EXCLUDE) ||
1040             positive <= entry_not_interesting) /* #1, #2, #11, #12 */
1041                 return positive;
1042
1043         negative = do_match(entry, base, base_offset, ps, 1);
1044
1045         /* #3, #4, #7, #8, #13, #14, #17, #18 */
1046         if (negative <= entry_not_interesting)
1047                 return positive;
1048
1049         /* #15, #19 */
1050         if (S_ISDIR(entry->mode) &&
1051             positive >= entry_interesting &&
1052             negative == entry_interesting)
1053                 return entry_interesting;
1054
1055         if ((positive == entry_interesting &&
1056              negative >= entry_interesting) || /* #5, #6, #16 */
1057             (positive == all_entries_interesting &&
1058              negative == entry_interesting)) /* #9 */
1059                 return entry_not_interesting;
1060
1061         return all_entries_not_interesting; /* #10, #20 */
1062 }