sha1_name.c

   1 #include "cache.h"
   2 #include "tag.h"
   3 #include "commit.h"
   4 #include "tree.h"
   5 #include "blob.h"
   6 #include "tree-walk.h"
   7 #include "refs.h"
   8 #include "remote.h"
   9
  10 static int get_sha1_oneline(const char *, unsigned char *, struct commit_list *);
  11
  12 typedef int (*disambiguate_hint_fn)(const unsigned char *, void *);
  13
  14 struct disambiguate_state {
  15         disambiguate_hint_fn fn;
  16         void *cb_data;
  17         unsigned char candidate[20];
  18         unsigned candidate_exists:1;
  19         unsigned candidate_checked:1;
  20         unsigned candidate_ok:1;
  21         unsigned disambiguate_fn_used:1;
  22         unsigned ambiguous:1;
  23         unsigned always_call_fn:1;
  24 };
  25
  26 static void update_candidates(struct disambiguate_state *ds, const unsigned char *current)
  27 {
  28         if (ds->always_call_fn) {
  29                 ds->ambiguous = ds->fn(current, ds->cb_data) ? 1 : 0;
  30                 return;
  31         }
  32         if (!ds->candidate_exists) {
  33                 /* this is the first candidate */
  34                 hashcpy(ds->candidate, current);
  35                 ds->candidate_exists = 1;
  36                 return;
  37         } else if (!hashcmp(ds->candidate, current)) {
  38                 /* the same as what we already have seen */
  39                 return;
  40         }
  41
  42         if (!ds->fn) {
  43                 /* cannot disambiguate between ds->candidate and current */
  44                 ds->ambiguous = 1;
  45                 return;
  46         }
  47
  48         if (!ds->candidate_checked) {
  49                 ds->candidate_ok = ds->fn(ds->candidate, ds->cb_data);
  50                 ds->disambiguate_fn_used = 1;
  51                 ds->candidate_checked = 1;
  52         }
  53
  54         if (!ds->candidate_ok) {
  55                 /* discard the candidate; we know it does not satisify fn */
  56                 hashcpy(ds->candidate, current);
  57                 ds->candidate_checked = 0;
  58                 return;
  59         }
  60
  61         /* if we reach this point, we know ds->candidate satisfies fn */
  62         if (ds->fn(current, ds->cb_data)) {
  63                 /*
  64                  * if both current and candidate satisfy fn, we cannot
  65                  * disambiguate.
  66                  */
  67                 ds->candidate_ok = 0;
  68                 ds->ambiguous = 1;
  69         }
  70
  71         /* otherwise, current can be discarded and candidate is still good */
  72 }
  73
  74 static void find_short_object_filename(int len, const char *hex_pfx, struct disambiguate_state *ds)
  75 {
  76         struct alternate_object_database *alt;
  77         char hex[40];
  78         static struct alternate_object_database *fakeent;
  79
  80         if (!fakeent) {
  81                 /*
  82                  * Create a "fake" alternate object database that
  83                  * points to our own object database, to make it
  84                  * easier to get a temporary working space in
  85                  * alt->name/alt->base while iterating over the
  86                  * object databases including our own.
  87                  */
  88                 const char *objdir = get_object_directory();
  89                 int objdir_len = strlen(objdir);
  90                 int entlen = objdir_len + 43;
  91                 fakeent = xmalloc(sizeof(*fakeent) + entlen);
  92                 memcpy(fakeent->base, objdir, objdir_len);
  93                 fakeent->name = fakeent->base + objdir_len + 1;
  94                 fakeent->name[-1] = '/';
  95         }
  96         fakeent->next = alt_odb_list;
  97
  98         sprintf(hex, "%.2s", hex_pfx);
  99         for (alt = fakeent; alt && !ds->ambiguous; alt = alt->next) {
 100                 struct dirent *de;
 101                 DIR *dir;
 102                 sprintf(alt->name, "%.2s/", hex_pfx);
 103                 dir = opendir(alt->base);
 104                 if (!dir)
 105                         continue;
 106
 107                 while (!ds->ambiguous && (de = readdir(dir)) != NULL) {
 108                         unsigned char sha1[20];
 109
 110                         if (strlen(de->d_name) != 38)
 111                                 continue;
 112                         if (memcmp(de->d_name, hex_pfx + 2, len - 2))
 113                                 continue;
 114                         memcpy(hex + 2, de->d_name, 38);
 115                         if (!get_sha1_hex(hex, sha1))
 116                                 update_candidates(ds, sha1);
 117                 }
 118                 closedir(dir);
 119         }
 120 }
 121
 122 static int match_sha(unsigned len, const unsigned char *a, const unsigned char *b)
 123 {
 124         do {
 125                 if (*a != *b)
 126                         return 0;
 127                 a++;
 128                 b++;
 129                 len -= 2;
 130         } while (len > 1);
 131         if (len)
 132                 if ((*a ^ *b) & 0xf0)
 133                         return 0;
 134         return 1;
 135 }
 136
 137 static void unique_in_pack(int len,
 138                           const unsigned char *bin_pfx,
 139                            struct packed_git *p,
 140                            struct disambiguate_state *ds)
 141 {
 142         uint32_t num, last, i, first = 0;
 143         const unsigned char *current = NULL;
 144
 145         open_pack_index(p);
 146         num = p->num_objects;
 147         last = num;
 148         while (first < last) {
 149                 uint32_t mid = (first + last) / 2;
 150                 const unsigned char *current;
 151                 int cmp;
 152
 153                 current = nth_packed_object_sha1(p, mid);
 154                 cmp = hashcmp(bin_pfx, current);
 155                 if (!cmp) {
 156                         first = mid;
 157                         break;
 158                 }
 159                 if (cmp > 0) {
 160                         first = mid+1;
 161                         continue;
 162                 }
 163                 last = mid;
 164         }
 165
 166         /*
 167          * At this point, "first" is the location of the lowest object
 168          * with an object name that could match "bin_pfx".  See if we have
 169          * 0, 1 or more objects that actually match(es).
 170          */
 171         for (i = first; i < num && !ds->ambiguous; i++) {
 172                 current = nth_packed_object_sha1(p, i);
 173                 if (!match_sha(len, bin_pfx, current))
 174                         break;
 175                 update_candidates(ds, current);
 176         }
 177 }
 178
 179 static void find_short_packed_object(int len, const unsigned char *bin_pfx,
 180                                      struct disambiguate_state *ds)
 181 {
 182         struct packed_git *p;
 183
 184         prepare_packed_git();
 185         for (p = packed_git; p && !ds->ambiguous; p = p->next)
 186                 unique_in_pack(len, bin_pfx, p, ds);
 187 }
 188
 189 #define SHORT_NAME_NOT_FOUND (-1)
 190 #define SHORT_NAME_AMBIGUOUS (-2)
 191
 192 static int finish_object_disambiguation(struct disambiguate_state *ds,
 193                                         unsigned char *sha1)
 194 {
 195         if (ds->ambiguous)
 196                 return SHORT_NAME_AMBIGUOUS;
 197
 198         if (!ds->candidate_exists)
 199                 return SHORT_NAME_NOT_FOUND;
 200
 201         if (!ds->candidate_checked)
 202                 /*
 203                  * If this is the only candidate, there is no point
 204                  * calling the disambiguation hint callback.
 205                  *
 206                  * On the other hand, if the current candidate
 207                  * replaced an earlier candidate that did _not_ pass
 208                  * the disambiguation hint callback, then we do have
 209                  * more than one objects that match the short name
 210                  * given, so we should make sure this one matches;
 211                  * otherwise, if we discovered this one and the one
 212                  * that we previously discarded in the reverse order,
 213                  * we would end up showing different results in the
 214                  * same repository!
 215                  */
 216                 ds->candidate_ok = (!ds->disambiguate_fn_used ||
 217                                     ds->fn(ds->candidate, ds->cb_data));
 218
 219         if (!ds->candidate_ok)
 220                 return SHORT_NAME_AMBIGUOUS;
 221
 222         hashcpy(sha1, ds->candidate);
 223         return 0;
 224 }
 225
 226 static int disambiguate_commit_only(const unsigned char *sha1, void *cb_data_unused)
 227 {
 228         int kind = sha1_object_info(sha1, NULL);
 229         return kind == OBJ_COMMIT;
 230 }
 231
 232 static int disambiguate_committish_only(const unsigned char *sha1, void *cb_data_unused)
 233 {
 234         struct object *obj;
 235         int kind;
 236
 237         kind = sha1_object_info(sha1, NULL);
 238         if (kind == OBJ_COMMIT)
 239                 return 1;
 240         if (kind != OBJ_TAG)
 241                 return 0;
 242
 243         /* We need to do this the hard way... */
 244         obj = deref_tag(lookup_object(sha1), NULL, 0);
 245         if (obj && obj->type == OBJ_COMMIT)
 246                 return 1;
 247         return 0;
 248 }
 249
 250 static int disambiguate_tree_only(const unsigned char *sha1, void *cb_data_unused)
 251 {
 252         int kind = sha1_object_info(sha1, NULL);
 253         return kind == OBJ_TREE;
 254 }
 255
 256 static int disambiguate_treeish_only(const unsigned char *sha1, void *cb_data_unused)
 257 {
 258         struct object *obj;
 259         int kind;
 260
 261         kind = sha1_object_info(sha1, NULL);
 262         if (kind == OBJ_TREE || kind == OBJ_COMMIT)
 263                 return 1;
 264         if (kind != OBJ_TAG)
 265                 return 0;
 266
 267         /* We need to do this the hard way... */
 268         obj = deref_tag(lookup_object(sha1), NULL, 0);
 269         if (obj && (obj->type == OBJ_TREE || obj->type == OBJ_COMMIT))
 270                 return 1;
 271         return 0;
 272 }
 273
 274 static int disambiguate_blob_only(const unsigned char *sha1, void *cb_data_unused)
 275 {
 276         int kind = sha1_object_info(sha1, NULL);
 277         return kind == OBJ_BLOB;
 278 }
 279
 280 static int prepare_prefixes(const char *name, int len,
 281                             unsigned char *bin_pfx,
 282                             char *hex_pfx)
 283 {
 284         int i;
 285
 286         hashclr(bin_pfx);
 287         memset(hex_pfx, 'x', 40);
 288         for (i = 0; i < len ;i++) {
 289                 unsigned char c = name[i];
 290                 unsigned char val;
 291                 if (c >= '0' && c <= '9')
 292                         val = c - '0';
 293                 else if (c >= 'a' && c <= 'f')
 294                         val = c - 'a' + 10;
 295                 else if (c >= 'A' && c <='F') {
 296                         val = c - 'A' + 10;
 297                         c -= 'A' - 'a';
 298                 }
 299                 else
 300                         return -1;
 301                 hex_pfx[i] = c;
 302                 if (!(i & 1))
 303                         val <<= 4;
 304                 bin_pfx[i >> 1] |= val;
 305         }
 306         return 0;
 307 }
 308
 309 static int get_short_sha1(const char *name, int len, unsigned char *sha1,
 310                           unsigned flags)
 311 {
 312         int status;
 313         char hex_pfx[40];
 314         unsigned char bin_pfx[20];
 315         struct disambiguate_state ds;
 316         int quietly = !!(flags & GET_SHA1_QUIETLY);
 317
 318         if (len < MINIMUM_ABBREV || len > 40)
 319                 return -1;
 320         if (prepare_prefixes(name, len, bin_pfx, hex_pfx) < 0)
 321                 return -1;
 322
 323         prepare_alt_odb();
 324
 325         memset(&ds, 0, sizeof(ds));
 326         if (flags & GET_SHA1_COMMIT)
 327                 ds.fn = disambiguate_commit_only;
 328         else if (flags & GET_SHA1_COMMITTISH)
 329                 ds.fn = disambiguate_committish_only;
 330         else if (flags & GET_SHA1_TREE)
 331                 ds.fn = disambiguate_tree_only;
 332         else if (flags & GET_SHA1_TREEISH)
 333                 ds.fn = disambiguate_treeish_only;
 334         else if (flags & GET_SHA1_BLOB)
 335                 ds.fn = disambiguate_blob_only;
 336
 337         find_short_object_filename(len, hex_pfx, &ds);
 338         find_short_packed_object(len, bin_pfx, &ds);
 339         status = finish_object_disambiguation(&ds, sha1);
 340
 341         if (!quietly && (status == SHORT_NAME_AMBIGUOUS))
 342                 return error("short SHA1 %.*s is ambiguous.", len, hex_pfx);
 343         return status;
 344 }
 345
 346
 347 int for_each_abbrev(const char *prefix, each_abbrev_fn fn, void *cb_data)
 348 {
 349         char hex_pfx[40];
 350         unsigned char bin_pfx[20];
 351         struct disambiguate_state ds;
 352         int len = strlen(prefix);
 353
 354         if (len < MINIMUM_ABBREV || len > 40)
 355                 return -1;
 356         if (prepare_prefixes(prefix, len, bin_pfx, hex_pfx) < 0)
 357                 return -1;
 358
 359         prepare_alt_odb();
 360
 361         memset(&ds, 0, sizeof(ds));
 362         ds.always_call_fn = 1;
 363         ds.cb_data = cb_data;
 364         ds.fn = fn;
 365
 366         find_short_object_filename(len, hex_pfx, &ds);
 367         find_short_packed_object(len, bin_pfx, &ds);
 368         return ds.ambiguous;
 369 }
 370
 371 const char *find_unique_abbrev(const unsigned char *sha1, int len)
 372 {
 373         int status, exists;
 374         static char hex[41];
 375
 376         exists = has_sha1_file(sha1);
 377         memcpy(hex, sha1_to_hex(sha1), 40);
 378         if (len == 40 || !len)
 379                 return hex;
 380         while (len < 40) {
 381                 unsigned char sha1_ret[20];
 382                 status = get_short_sha1(hex, len, sha1_ret, GET_SHA1_QUIETLY);
 383                 if (exists
 384                     ? !status
 385                     : status == SHORT_NAME_NOT_FOUND) {
 386                         hex[len] = 0;
 387                         return hex;
 388                 }
 389                 len++;
 390         }
 391         return hex;
 392 }
 393
 394 static int ambiguous_path(const char *path, int len)
 395 {
 396         int slash = 1;
 397         int cnt;
 398
 399         for (cnt = 0; cnt < len; cnt++) {
 400                 switch (*path++) {
 401                 case '\0':
 402                         break;
 403                 case '/':
 404                         if (slash)
 405                                 break;
 406                         slash = 1;
 407                         continue;
 408                 case '.':
 409                         continue;
 410                 default:
 411                         slash = 0;
 412                         continue;
 413                 }
 414                 break;
 415         }
 416         return slash;
 417 }
 418
 419 static inline int upstream_mark(const char *string, int len)
 420 {
 421         const char *suffix[] = { "@{upstream}", "@{u}" };
 422         int i;
 423
 424         for (i = 0; i < ARRAY_SIZE(suffix); i++) {
 425                 int suffix_len = strlen(suffix[i]);
 426                 if (suffix_len <= len
 427                     && !memcmp(string, suffix[i], suffix_len))
 428                         return suffix_len;
 429         }
 430         return 0;
 431 }
 432
 433 static int get_sha1_1(const char *name, int len, unsigned char *sha1, unsigned lookup_flags);
 434 static int interpret_nth_prior_checkout(const char *name, struct strbuf *buf);
 435
 436 static int get_sha1_basic(const char *str, int len, unsigned char *sha1)
 437 {
 438         static const char *warn_msg = "refname '%.*s' is ambiguous.";
 439         char *real_ref = NULL;
 440         int refs_found = 0;
 441         int at, reflog_len, nth_prior = 0;
 442
 443         if (len == 40 && !get_sha1_hex(str, sha1))
 444                 return 0;
 445
 446         /* basic@{time or number or -number} format to query ref-log */
 447         reflog_len = at = 0;
 448         if (len && str[len-1] == '}') {
 449                 for (at = len-4; at >= 0; at--) {
 450                         if (str[at] == '@' && str[at+1] == '{') {
 451                                 if (str[at+2] == '-') {
 452                                         if (at != 0)
 453                                                 /* @{-N} not at start */
 454                                                 return -1;
 455                                         nth_prior = 1;
 456                                         continue;
 457                                 }
 458                                 if (!upstream_mark(str + at, len - at)) {
 459                                         reflog_len = (len-1) - (at+2);
 460                                         len = at;
 461                                 }
 462                                 break;
 463                         }
 464                 }
 465         }
 466
 467         /* Accept only unambiguous ref paths. */
 468         if (len && ambiguous_path(str, len))
 469                 return -1;
 470
 471         if (nth_prior) {
 472                 struct strbuf buf = STRBUF_INIT;
 473                 int detached;
 474
 475                 if (interpret_nth_prior_checkout(str, &buf) > 0) {
 476                         detached = (buf.len == 40 && !get_sha1_hex(buf.buf, sha1));
 477                         strbuf_release(&buf);
 478                         if (detached)
 479                                 return 0;
 480                 }
 481         }
 482
 483         if (!len && reflog_len)
 484                 /* allow "@{...}" to mean the current branch reflog */
 485                 refs_found = dwim_ref("HEAD", 4, sha1, &real_ref);
 486         else if (reflog_len)
 487                 refs_found = dwim_log(str, len, sha1, &real_ref);
 488         else
 489                 refs_found = dwim_ref(str, len, sha1, &real_ref);
 490
 491         if (!refs_found)
 492                 return -1;
 493
 494         if (warn_ambiguous_refs && refs_found > 1)
 495                 warning(warn_msg, len, str);
 496
 497         if (reflog_len) {
 498                 int nth, i;
 499                 unsigned long at_time;
 500                 unsigned long co_time;
 501                 int co_tz, co_cnt;
 502
 503                 /* Is it asking for N-th entry, or approxidate? */
 504                 for (i = nth = 0; 0 <= nth && i < reflog_len; i++) {
 505                         char ch = str[at+2+i];
 506                         if ('0' <= ch && ch <= '9')
 507                                 nth = nth * 10 + ch - '0';
 508                         else
 509                                 nth = -1;
 510                 }
 511                 if (100000000 <= nth) {
 512                         at_time = nth;
 513                         nth = -1;
 514                 } else if (0 <= nth)
 515                         at_time = 0;
 516                 else {
 517                         int errors = 0;
 518                         char *tmp = xstrndup(str + at + 2, reflog_len);
 519                         at_time = approxidate_careful(tmp, &errors);
 520                         free(tmp);
 521                         if (errors)
 522                                 return -1;
 523                 }
 524                 if (read_ref_at(real_ref, at_time, nth, sha1, NULL,
 525                                 &co_time, &co_tz, &co_cnt)) {
 526                         if (at_time)
 527                                 warning("Log for '%.*s' only goes "
 528                                         "back to %s.", len, str,
 529                                         show_date(co_time, co_tz, DATE_RFC2822));
 530                         else {
 531                                 free(real_ref);
 532                                 die("Log for '%.*s' only has %d entries.",
 533                                     len, str, co_cnt);
 534                         }
 535                 }
 536         }
 537
 538         free(real_ref);
 539         return 0;
 540 }
 541
 542 static int get_parent(const char *name, int len,
 543                       unsigned char *result, int idx)
 544 {
 545         unsigned char sha1[20];
 546         int ret = get_sha1_1(name, len, sha1, GET_SHA1_COMMITTISH);
 547         struct commit *commit;
 548         struct commit_list *p;
 549
 550         if (ret)
 551                 return ret;
 552         commit = lookup_commit_reference(sha1);
 553         if (!commit)
 554                 return -1;
 555         if (parse_commit(commit))
 556                 return -1;
 557         if (!idx) {
 558                 hashcpy(result, commit->object.sha1);
 559                 return 0;
 560         }
 561         p = commit->parents;
 562         while (p) {
 563                 if (!--idx) {
 564                         hashcpy(result, p->item->object.sha1);
 565                         return 0;
 566                 }
 567                 p = p->next;
 568         }
 569         return -1;
 570 }
 571
 572 static int get_nth_ancestor(const char *name, int len,
 573                             unsigned char *result, int generation)
 574 {
 575         unsigned char sha1[20];
 576         struct commit *commit;
 577         int ret;
 578
 579         ret = get_sha1_1(name, len, sha1, GET_SHA1_COMMITTISH);
 580         if (ret)
 581                 return ret;
 582         commit = lookup_commit_reference(sha1);
 583         if (!commit)
 584                 return -1;
 585
 586         while (generation--) {
 587                 if (parse_commit(commit) || !commit->parents)
 588                         return -1;
 589                 commit = commit->parents->item;
 590         }
 591         hashcpy(result, commit->object.sha1);
 592         return 0;
 593 }
 594
 595 struct object *peel_to_type(const char *name, int namelen,
 596                             struct object *o, enum object_type expected_type)
 597 {
 598         if (name && !namelen)
 599                 namelen = strlen(name);
 600         while (1) {
 601                 if (!o || (!o->parsed && !parse_object(o->sha1)))
 602                         return NULL;
 603                 if (expected_type == OBJ_ANY || o->type == expected_type)
 604                         return o;
 605                 if (o->type == OBJ_TAG)
 606                         o = ((struct tag*) o)->tagged;
 607                 else if (o->type == OBJ_COMMIT)
 608                         o = &(((struct commit *) o)->tree->object);
 609                 else {
 610                         if (name)
 611                                 error("%.*s: expected %s type, but the object "
 612                                       "dereferences to %s type",
 613                                       namelen, name, typename(expected_type),
 614                                       typename(o->type));
 615                         return NULL;
 616                 }
 617         }
 618 }
 619
 620 static int peel_onion(const char *name, int len, unsigned char *sha1)
 621 {
 622         unsigned char outer[20];
 623         const char *sp;
 624         unsigned int expected_type = 0;
 625         unsigned lookup_flags = 0;
 626         struct object *o;
 627
 628         /*
 629          * "ref^{type}" dereferences ref repeatedly until you cannot
 630          * dereference anymore, or you get an object of given type,
 631          * whichever comes first.  "ref^{}" means just dereference
 632          * tags until you get a non-tag.  "ref^0" is a shorthand for
 633          * "ref^{commit}".  "commit^{tree}" could be used to find the
 634          * top-level tree of the given commit.
 635          */
 636         if (len < 4 || name[len-1] != '}')
 637                 return -1;
 638
 639         for (sp = name + len - 1; name <= sp; sp--) {
 640                 int ch = *sp;
 641                 if (ch == '{' && name < sp && sp[-1] == '^')
 642                         break;
 643         }
 644         if (sp <= name)
 645                 return -1;
 646
 647         sp++; /* beginning of type name, or closing brace for empty */
 648         if (!strncmp(commit_type, sp, 6) && sp[6] == '}')
 649                 expected_type = OBJ_COMMIT;
 650         else if (!strncmp(tree_type, sp, 4) && sp[4] == '}')
 651                 expected_type = OBJ_TREE;
 652         else if (!strncmp(blob_type, sp, 4) && sp[4] == '}')
 653                 expected_type = OBJ_BLOB;
 654         else if (!prefixcmp(sp, "object}"))
 655                 expected_type = OBJ_ANY;
 656         else if (sp[0] == '}')
 657                 expected_type = OBJ_NONE;
 658         else if (sp[0] == '/')
 659                 expected_type = OBJ_COMMIT;
 660         else
 661                 return -1;
 662
 663         if (expected_type == OBJ_COMMIT)
 664                 lookup_flags = GET_SHA1_COMMITTISH;
 665         else if (expected_type == OBJ_TREE)
 666                 lookup_flags = GET_SHA1_TREEISH;
 667
 668         if (get_sha1_1(name, sp - name - 2, outer, lookup_flags))
 669                 return -1;
 670
 671         o = parse_object(outer);
 672         if (!o)
 673                 return -1;
 674         if (!expected_type) {
 675                 o = deref_tag(o, name, sp - name - 2);
 676                 if (!o || (!o->parsed && !parse_object(o->sha1)))
 677                         return -1;
 678                 hashcpy(sha1, o->sha1);
 679                 return 0;
 680         }
 681
 682         /*
 683          * At this point, the syntax look correct, so
 684          * if we do not get the needed object, we should
 685          * barf.
 686          */
 687         o = peel_to_type(name, len, o, expected_type);
 688         if (!o)
 689                 return -1;
 690
 691         hashcpy(sha1, o->sha1);
 692         if (sp[0] == '/') {
 693                 /* "$commit^{/foo}" */
 694                 char *prefix;
 695                 int ret;
 696                 struct commit_list *list = NULL;
 697
 698                 /*
 699                  * $commit^{/}. Some regex implementation may reject.
 700                  * We don't need regex anyway. '' pattern always matches.
 701                  */
 702                 if (sp[1] == '}')
 703                         return 0;
 704
 705                 prefix = xstrndup(sp + 1, name + len - 1 - (sp + 1));
 706                 commit_list_insert((struct commit *)o, &list);
 707                 ret = get_sha1_oneline(prefix, sha1, list);
 708                 free(prefix);
 709                 return ret;
 710         }
 711         return 0;
 712 }
 713
 714 static int get_describe_name(const char *name, int len, unsigned char *sha1)
 715 {
 716         const char *cp;
 717         unsigned flags = GET_SHA1_QUIETLY | GET_SHA1_COMMIT;
 718
 719         for (cp = name + len - 1; name + 2 <= cp; cp--) {
 720                 char ch = *cp;
 721                 if (hexval(ch) & ~0377) {
 722                         /* We must be looking at g in "SOMETHING-g"
 723                          * for it to be describe output.
 724                          */
 725                         if (ch == 'g' && cp[-1] == '-') {
 726                                 cp++;
 727                                 len -= cp - name;
 728                                 return get_short_sha1(cp, len, sha1, flags);
 729                         }
 730                 }
 731         }
 732         return -1;
 733 }
 734
 735 static int get_sha1_1(const char *name, int len, unsigned char *sha1, unsigned lookup_flags)
 736 {
 737         int ret, has_suffix;
 738         const char *cp;
 739
 740         /*
 741          * "name~3" is "name^^^", "name~" is "name~1", and "name^" is "name^1".
 742          */
 743         has_suffix = 0;
 744         for (cp = name + len - 1; name <= cp; cp--) {
 745                 int ch = *cp;
 746                 if ('0' <= ch && ch <= '9')
 747                         continue;
 748                 if (ch == '~' || ch == '^')
 749                         has_suffix = ch;
 750                 break;
 751         }
 752
 753         if (has_suffix) {
 754                 int num = 0;
 755                 int len1 = cp - name;
 756                 cp++;
 757                 while (cp < name + len)
 758                         num = num * 10 + *cp++ - '0';
 759                 if (!num && len1 == len - 1)
 760                         num = 1;
 761                 if (has_suffix == '^')
 762                         return get_parent(name, len1, sha1, num);
 763                 /* else if (has_suffix == '~') -- goes without saying */
 764                 return get_nth_ancestor(name, len1, sha1, num);
 765         }
 766
 767         ret = peel_onion(name, len, sha1);
 768         if (!ret)
 769                 return 0;
 770
 771         ret = get_sha1_basic(name, len, sha1);
 772         if (!ret)
 773                 return 0;
 774
 775         /* It could be describe output that is "SOMETHING-gXXXX" */
 776         ret = get_describe_name(name, len, sha1);
 777         if (!ret)
 778                 return 0;
 779
 780         return get_short_sha1(name, len, sha1, lookup_flags);
 781 }
 782
 783 /*
 784  * This interprets names like ':/Initial revision of "git"' by searching
 785  * through history and returning the first commit whose message starts
 786  * the given regular expression.
 787  *
 788  * For future extension, ':/!' is reserved. If you want to match a message
 789  * beginning with a '!', you have to repeat the exclamation mark.
 790  */
 791 #define ONELINE_SEEN (1u<<20)
 792
 793 static int handle_one_ref(const char *path,
 794                 const unsigned char *sha1, int flag, void *cb_data)
 795 {
 796         struct commit_list **list = cb_data;
 797         struct object *object = parse_object(sha1);
 798         if (!object)
 799                 return 0;
 800         if (object->type == OBJ_TAG) {
 801                 object = deref_tag(object, path, strlen(path));
 802                 if (!object)
 803                         return 0;
 804         }
 805         if (object->type != OBJ_COMMIT)
 806                 return 0;
 807         commit_list_insert_by_date((struct commit *)object, list);
 808         return 0;
 809 }
 810
 811 static int get_sha1_oneline(const char *prefix, unsigned char *sha1,
 812                             struct commit_list *list)
 813 {
 814         struct commit_list *backup = NULL, *l;
 815         int found = 0;
 816         regex_t regex;
 817
 818         if (prefix[0] == '!') {
 819                 if (prefix[1] != '!')
 820                         die ("Invalid search pattern: %s", prefix);
 821                 prefix++;
 822         }
 823
 824         if (regcomp(&regex, prefix, REG_EXTENDED))
 825                 die("Invalid search pattern: %s", prefix);
 826
 827         for (l = list; l; l = l->next) {
 828                 l->item->object.flags |= ONELINE_SEEN;
 829                 commit_list_insert(l->item, &backup);
 830         }
 831         while (list) {
 832                 char *p, *to_free = NULL;
 833                 struct commit *commit;
 834                 enum object_type type;
 835                 unsigned long size;
 836                 int matches;
 837
 838                 commit = pop_most_recent_commit(&list, ONELINE_SEEN);
 839                 if (!parse_object(commit->object.sha1))
 840                         continue;
 841                 if (commit->buffer)
 842                         p = commit->buffer;
 843                 else {
 844                         p = read_sha1_file(commit->object.sha1, &type, &size);
 845                         if (!p)
 846                                 continue;
 847                         to_free = p;
 848                 }
 849
 850                 p = strstr(p, "\n\n");
 851                 matches = p && !regexec(&regex, p + 2, 0, NULL, 0);
 852                 free(to_free);
 853
 854                 if (matches) {
 855                         hashcpy(sha1, commit->object.sha1);
 856                         found = 1;
 857                         break;
 858                 }
 859         }
 860         regfree(&regex);
 861         free_commit_list(list);
 862         for (l = backup; l; l = l->next)
 863                 clear_commit_marks(l->item, ONELINE_SEEN);
 864         free_commit_list(backup);
 865         return found ? 0 : -1;
 866 }
 867
 868 struct grab_nth_branch_switch_cbdata {
 869         int remaining;
 870         struct strbuf buf;
 871 };
 872
 873 static int grab_nth_branch_switch(unsigned char *osha1, unsigned char *nsha1,
 874                                   const char *email, unsigned long timestamp, int tz,
 875                                   const char *message, void *cb_data)
 876 {
 877         struct grab_nth_branch_switch_cbdata *cb = cb_data;
 878         const char *match = NULL, *target = NULL;
 879         size_t len;
 880
 881         if (!prefixcmp(message, "checkout: moving from ")) {
 882                 match = message + strlen("checkout: moving from ");
 883                 target = strstr(match, " to ");
 884         }
 885
 886         if (!match || !target)
 887                 return 0;
 888         if (--(cb->remaining) == 0) {
 889                 len = target - match;
 890                 strbuf_reset(&cb->buf);
 891                 strbuf_add(&cb->buf, match, len);
 892                 return 1; /* we are done */
 893         }
 894         return 0;
 895 }
 896
 897 /*
 898  * Parse @{-N} syntax, return the number of characters parsed
 899  * if successful; otherwise signal an error with negative value.
 900  */
 901 static int interpret_nth_prior_checkout(const char *name, struct strbuf *buf)
 902 {
 903         long nth;
 904         int retval;
 905         struct grab_nth_branch_switch_cbdata cb;
 906         const char *brace;
 907         char *num_end;
 908
 909         if (name[0] != '@' || name[1] != '{' || name[2] != '-')
 910                 return -1;
 911         brace = strchr(name, '}');
 912         if (!brace)
 913                 return -1;
 914         nth = strtol(name + 3, &num_end, 10);
 915         if (num_end != brace)
 916                 return -1;
 917         if (nth <= 0)
 918                 return -1;
 919         cb.remaining = nth;
 920         strbuf_init(&cb.buf, 20);
 921
 922         retval = 0;
 923         if (0 < for_each_reflog_ent_reverse("HEAD", grab_nth_branch_switch, &cb)) {
 924                 strbuf_reset(buf);
 925                 strbuf_add(buf, cb.buf.buf, cb.buf.len);
 926                 retval = brace - name + 1;
 927         }
 928
 929         strbuf_release(&cb.buf);
 930         return retval;
 931 }
 932
 933 int get_sha1_mb(const char *name, unsigned char *sha1)
 934 {
 935         struct commit *one, *two;
 936         struct commit_list *mbs;
 937         unsigned char sha1_tmp[20];
 938         const char *dots;
 939         int st;
 940
 941         dots = strstr(name, "...");
 942         if (!dots)
 943                 return get_sha1(name, sha1);
 944         if (dots == name)
 945                 st = get_sha1("HEAD", sha1_tmp);
 946         else {
 947                 struct strbuf sb;
 948                 strbuf_init(&sb, dots - name);
 949                 strbuf_add(&sb, name, dots - name);
 950                 st = get_sha1_committish(sb.buf, sha1_tmp);
 951                 strbuf_release(&sb);
 952         }
 953         if (st)
 954                 return st;
 955         one = lookup_commit_reference_gently(sha1_tmp, 0);
 956         if (!one)
 957                 return -1;
 958
 959         if (get_sha1_committish(dots[3] ? (dots + 3) : "HEAD", sha1_tmp))
 960                 return -1;
 961         two = lookup_commit_reference_gently(sha1_tmp, 0);
 962         if (!two)
 963                 return -1;
 964         mbs = get_merge_bases(one, two, 1);
 965         if (!mbs || mbs->next)
 966                 st = -1;
 967         else {
 968                 st = 0;
 969                 hashcpy(sha1, mbs->item->object.sha1);
 970         }
 971         free_commit_list(mbs);
 972         return st;
 973 }
 974
 975 /* parse @something syntax, when 'something' is not {.*} */
 976 static int interpret_empty_at(const char *name, int namelen, int len, struct strbuf *buf)
 977 {
 978         if (len || name[1] == '{')
 979                 return -1;
 980
 981         strbuf_reset(buf);
 982         strbuf_add(buf, "HEAD", 4);
 983         return 1;
 984 }
 985
 986 static int reinterpret(const char *name, int namelen, int len, struct strbuf *buf)
 987 {
 988         /* we have extra data, which might need further processing */
 989         struct strbuf tmp = STRBUF_INIT;
 990         int used = buf->len;
 991         int ret;
 992
 993         strbuf_add(buf, name + len, namelen - len);
 994         ret = interpret_branch_name(buf->buf, &tmp);
 995         /* that data was not interpreted, remove our cruft */
 996         if (ret < 0) {
 997                 strbuf_setlen(buf, used);
 998                 return len;
 999         }
1000         strbuf_reset(buf);
1001         strbuf_addbuf(buf, &tmp);
1002         strbuf_release(&tmp);
1003         /* tweak for size of {-N} versus expanded ref name */
1004         return ret - used + len;
1005 }
1006
1007 /*
1008  * This reads short-hand syntax that not only evaluates to a commit
1009  * object name, but also can act as if the end user spelled the name
1010  * of the branch from the command line.
1011  *
1012  * - "@{-N}" finds the name of the Nth previous branch we were on, and
1013  *   places the name of the branch in the given buf and returns the
1014  *   number of characters parsed if successful.
1015  *
1016  * - "<branch>@{upstream}" finds the name of the other ref that
1017  *   <branch> is configured to merge with (missing <branch> defaults
1018  *   to the current branch), and places the name of the branch in the
1019  *   given buf and returns the number of characters parsed if
1020  *   successful.
1021  *
1022  * If the input is not of the accepted format, it returns a negative
1023  * number to signal an error.
1024  *
1025  * If the input was ok but there are not N branch switches in the
1026  * reflog, it returns 0.
1027  */
1028 int interpret_branch_name(const char *name, struct strbuf *buf)
1029 {
1030         char *cp;
1031         struct branch *upstream;
1032         int namelen = strlen(name);
1033         int len = interpret_nth_prior_checkout(name, buf);
1034         int tmp_len;
1035
1036         if (!len)
1037                 return len; /* syntax Ok, not enough switches */
1038         if (0 < len && len == namelen)
1039                 return len; /* consumed all */
1040         else if (0 < len)
1041                 return reinterpret(name, namelen, len, buf);
1042
1043         cp = strchr(name, '@');
1044         if (!cp)
1045                 return -1;
1046
1047         len = interpret_empty_at(name, namelen, cp - name, buf);
1048         if (len > 0)
1049                 return reinterpret(name, namelen, len, buf);
1050
1051         tmp_len = upstream_mark(cp, namelen - (cp - name));
1052         if (!tmp_len)
1053                 return -1;
1054
1055         len = cp + tmp_len - name;
1056         cp = xstrndup(name, cp - name);
1057         upstream = branch_get(*cp ? cp : NULL);
1058         /*
1059          * Upstream can be NULL only if cp refers to HEAD and HEAD
1060          * points to something different than a branch.
1061          */
1062         if (!upstream)
1063                 return error(_("HEAD does not point to a branch"));
1064         if (!upstream->merge || !upstream->merge[0]->dst) {
1065                 if (!ref_exists(upstream->refname))
1066                         return error(_("No such branch: '%s'"), cp);
1067                 if (!upstream->merge)
1068                         return error(_("No upstream configured for branch '%s'"),
1069                                      upstream->name);
1070                 return error(
1071                         _("Upstream branch '%s' not stored as a remote-tracking branch"),
1072                         upstream->merge[0]->src);
1073         }
1074         free(cp);
1075         cp = shorten_unambiguous_ref(upstream->merge[0]->dst, 0);
1076         strbuf_reset(buf);
1077         strbuf_addstr(buf, cp);
1078         free(cp);
1079         return len;
1080 }
1081
1082 int strbuf_branchname(struct strbuf *sb, const char *name)
1083 {
1084         int len = strlen(name);
1085         if (interpret_branch_name(name, sb) == len)
1086                 return 0;
1087         strbuf_add(sb, name, len);
1088         return len;
1089 }
1090
1091 int strbuf_check_branch_ref(struct strbuf *sb, const char *name)
1092 {
1093         strbuf_branchname(sb, name);
1094         if (name[0] == '-')
1095                 return -1;
1096         strbuf_splice(sb, 0, 0, "refs/heads/", 11);
1097         return check_refname_format(sb->buf, 0);
1098 }
1099
1100 /*
1101  * This is like "get_sha1_basic()", except it allows "sha1 expressions",
1102  * notably "xyz^" for "parent of xyz"
1103  */
1104 int get_sha1(const char *name, unsigned char *sha1)
1105 {
1106         struct object_context unused;
1107         return get_sha1_with_context(name, 0, sha1, &unused);
1108 }
1109
1110 /*
1111  * Many callers know that the user meant to name a committish by
1112  * syntactical positions where the object name appears.  Calling this
1113  * function allows the machinery to disambiguate shorter-than-unique
1114  * abbreviated object names between committish and others.
1115  *
1116  * Note that this does NOT error out when the named object is not a
1117  * committish. It is merely to give a hint to the disambiguation
1118  * machinery.
1119  */
1120 int get_sha1_committish(const char *name, unsigned char *sha1)
1121 {
1122         struct object_context unused;
1123         return get_sha1_with_context(name, GET_SHA1_COMMITTISH,
1124                                      sha1, &unused);
1125 }
1126
1127 int get_sha1_treeish(const char *name, unsigned char *sha1)
1128 {
1129         struct object_context unused;
1130         return get_sha1_with_context(name, GET_SHA1_TREEISH,
1131                                      sha1, &unused);
1132 }
1133
1134 int get_sha1_commit(const char *name, unsigned char *sha1)
1135 {
1136         struct object_context unused;
1137         return get_sha1_with_context(name, GET_SHA1_COMMIT,
1138                                      sha1, &unused);
1139 }
1140
1141 int get_sha1_tree(const char *name, unsigned char *sha1)
1142 {
1143         struct object_context unused;
1144         return get_sha1_with_context(name, GET_SHA1_TREE,
1145                                      sha1, &unused);
1146 }
1147
1148 int get_sha1_blob(const char *name, unsigned char *sha1)
1149 {
1150         struct object_context unused;
1151         return get_sha1_with_context(name, GET_SHA1_BLOB,
1152                                      sha1, &unused);
1153 }
1154
1155 /* Must be called only when object_name:filename doesn't exist. */
1156 static void diagnose_invalid_sha1_path(const char *prefix,
1157                                        const char *filename,
1158                                        const unsigned char *tree_sha1,
1159                                        const char *object_name,
1160                                        int object_name_len)
1161 {
1162         struct stat st;
1163         unsigned char sha1[20];
1164         unsigned mode;
1165
1166         if (!prefix)
1167                 prefix = "";
1168
1169         if (!lstat(filename, &st))
1170                 die("Path '%s' exists on disk, but not in '%.*s'.",
1171                     filename, object_name_len, object_name);
1172         if (errno == ENOENT || errno == ENOTDIR) {
1173                 char *fullname = xmalloc(strlen(filename)
1174                                              + strlen(prefix) + 1);
1175                 strcpy(fullname, prefix);
1176                 strcat(fullname, filename);
1177
1178                 if (!get_tree_entry(tree_sha1, fullname,
1179                                     sha1, &mode)) {
1180                         die("Path '%s' exists, but not '%s'.\n"
1181                             "Did you mean '%.*s:%s' aka '%.*s:./%s'?",
1182                             fullname,
1183                             filename,
1184                             object_name_len, object_name,
1185                             fullname,
1186                             object_name_len, object_name,
1187                             filename);
1188                 }
1189                 die("Path '%s' does not exist in '%.*s'",
1190                     filename, object_name_len, object_name);
1191         }
1192 }
1193
1194 /* Must be called only when :stage:filename doesn't exist. */
1195 static void diagnose_invalid_index_path(int stage,
1196                                         const char *prefix,
1197                                         const char *filename)
1198 {
1199         struct stat st;
1200         struct cache_entry *ce;
1201         int pos;
1202         unsigned namelen = strlen(filename);
1203         unsigned fullnamelen;
1204         char *fullname;
1205
1206         if (!prefix)
1207                 prefix = "";
1208
1209         /* Wrong stage number? */
1210         pos = cache_name_pos(filename, namelen);
1211         if (pos < 0)
1212                 pos = -pos - 1;
1213         if (pos < active_nr) {
1214                 ce = active_cache[pos];
1215                 if (ce_namelen(ce) == namelen &&
1216                     !memcmp(ce->name, filename, namelen))
1217                         die("Path '%s' is in the index, but not at stage %d.\n"
1218                             "Did you mean ':%d:%s'?",
1219                             filename, stage,
1220                             ce_stage(ce), filename);
1221         }
1222
1223         /* Confusion between relative and absolute filenames? */
1224         fullnamelen = namelen + strlen(prefix);
1225         fullname = xmalloc(fullnamelen + 1);
1226         strcpy(fullname, prefix);
1227         strcat(fullname, filename);
1228         pos = cache_name_pos(fullname, fullnamelen);
1229         if (pos < 0)
1230                 pos = -pos - 1;
1231         if (pos < active_nr) {
1232                 ce = active_cache[pos];
1233                 if (ce_namelen(ce) == fullnamelen &&
1234                     !memcmp(ce->name, fullname, fullnamelen))
1235                         die("Path '%s' is in the index, but not '%s'.\n"
1236                             "Did you mean ':%d:%s' aka ':%d:./%s'?",
1237                             fullname, filename,
1238                             ce_stage(ce), fullname,
1239                             ce_stage(ce), filename);
1240         }
1241
1242         if (!lstat(filename, &st))
1243                 die("Path '%s' exists on disk, but not in the index.", filename);
1244         if (errno == ENOENT || errno == ENOTDIR)
1245                 die("Path '%s' does not exist (neither on disk nor in the index).",
1246                     filename);
1247
1248         free(fullname);
1249 }
1250
1251
1252 static char *resolve_relative_path(const char *rel)
1253 {
1254         if (prefixcmp(rel, "./") && prefixcmp(rel, "../"))
1255                 return NULL;
1256
1257         if (!startup_info)
1258                 die("BUG: startup_info struct is not initialized.");
1259
1260         if (!is_inside_work_tree())
1261                 die("relative path syntax can't be used outside working tree.");
1262
1263         /* die() inside prefix_path() if resolved path is outside worktree */
1264         return prefix_path(startup_info->prefix,
1265                            startup_info->prefix ? strlen(startup_info->prefix) : 0,
1266                            rel);
1267 }
1268
1269 static int get_sha1_with_context_1(const char *name,
1270                                    unsigned flags,
1271                                    const char *prefix,
1272                                    unsigned char *sha1,
1273                                    struct object_context *oc)
1274 {
1275         int ret, bracket_depth;
1276         int namelen = strlen(name);
1277         const char *cp;
1278         int only_to_die = flags & GET_SHA1_ONLY_TO_DIE;
1279
1280         memset(oc, 0, sizeof(*oc));
1281         oc->mode = S_IFINVALID;
1282         ret = get_sha1_1(name, namelen, sha1, flags);
1283         if (!ret)
1284                 return ret;
1285         /*
1286          * sha1:path --> object name of path in ent sha1
1287          * :path -> object name of absolute path in index
1288          * :./path -> object name of path relative to cwd in index
1289          * :[0-3]:path -> object name of path in index at stage
1290          * :/foo -> recent commit matching foo
1291          */
1292         if (name[0] == ':') {
1293                 int stage = 0;
1294                 struct cache_entry *ce;
1295                 char *new_path = NULL;
1296                 int pos;
1297                 if (!only_to_die && namelen > 2 && name[1] == '/') {
1298                         struct commit_list *list = NULL;
1299                         for_each_ref(handle_one_ref, &list);
1300                         return get_sha1_oneline(name + 2, sha1, list);
1301                 }
1302                 if (namelen < 3 ||
1303                     name[2] != ':' ||
1304                     name[1] < '0' || '3' < name[1])
1305                         cp = name + 1;
1306                 else {
1307                         stage = name[1] - '0';
1308                         cp = name + 3;
1309                 }
1310                 new_path = resolve_relative_path(cp);
1311                 if (!new_path) {
1312                         namelen = namelen - (cp - name);
1313                 } else {
1314                         cp = new_path;
1315                         namelen = strlen(cp);
1316                 }
1317
1318                 strncpy(oc->path, cp,
1319                         sizeof(oc->path));
1320                 oc->path[sizeof(oc->path)-1] = '\0';
1321
1322                 if (!active_cache)
1323                         read_cache();
1324                 pos = cache_name_pos(cp, namelen);
1325                 if (pos < 0)
1326                         pos = -pos - 1;
1327                 while (pos < active_nr) {
1328                         ce = active_cache[pos];
1329                         if (ce_namelen(ce) != namelen ||
1330                             memcmp(ce->name, cp, namelen))
1331                                 break;
1332                         if (ce_stage(ce) == stage) {
1333                                 hashcpy(sha1, ce->sha1);
1334                                 oc->mode = ce->ce_mode;
1335                                 free(new_path);
1336                                 return 0;
1337                         }
1338                         pos++;
1339                 }
1340                 if (only_to_die && name[1] && name[1] != '/')
1341                         diagnose_invalid_index_path(stage, prefix, cp);
1342                 free(new_path);
1343                 return -1;
1344         }
1345         for (cp = name, bracket_depth = 0; *cp; cp++) {
1346                 if (*cp == '{')
1347                         bracket_depth++;
1348                 else if (bracket_depth && *cp == '}')
1349                         bracket_depth--;
1350                 else if (!bracket_depth && *cp == ':')
1351                         break;
1352         }
1353         if (*cp == ':') {
1354                 unsigned char tree_sha1[20];
1355                 int len = cp - name;
1356                 if (!get_sha1_1(name, len, tree_sha1, GET_SHA1_TREEISH)) {
1357                         const char *filename = cp+1;
1358                         char *new_filename = NULL;
1359
1360                         new_filename = resolve_relative_path(filename);
1361                         if (new_filename)
1362                                 filename = new_filename;
1363                         ret = get_tree_entry(tree_sha1, filename, sha1, &oc->mode);
1364                         if (ret && only_to_die) {
1365                                 diagnose_invalid_sha1_path(prefix, filename,
1366                                                            tree_sha1,
1367                                                            name, len);
1368                         }
1369                         hashcpy(oc->tree, tree_sha1);
1370                         strncpy(oc->path, filename,
1371                                 sizeof(oc->path));
1372                         oc->path[sizeof(oc->path)-1] = '\0';
1373
1374                         free(new_filename);
1375                         return ret;
1376                 } else {
1377                         if (only_to_die)
1378                                 die("Invalid object name '%.*s'.", len, name);
1379                 }
1380         }
1381         return ret;
1382 }
1383
1384 /*
1385  * Call this function when you know "name" given by the end user must
1386  * name an object but it doesn't; the function _may_ die with a better
1387  * diagnostic message than "no such object 'name'", e.g. "Path 'doc' does not
1388  * exist in 'HEAD'" when given "HEAD:doc", or it may return in which case
1389  * you have a chance to diagnose the error further.
1390  */
1391 void maybe_die_on_misspelt_object_name(const char *name, const char *prefix)
1392 {
1393         struct object_context oc;
1394         unsigned char sha1[20];
1395         get_sha1_with_context_1(name, GET_SHA1_ONLY_TO_DIE, prefix, sha1, &oc);
1396 }
1397
1398 int get_sha1_with_context(const char *str, unsigned flags, unsigned char *sha1, struct object_context *orc)
1399 {
1400         return get_sha1_with_context_1(str, flags, NULL, sha1, orc);
1401 }