builtin-for-each-ref.c

   1 #include "builtin.h"
   2 #include "cache.h"
   3 #include "refs.h"
   4 #include "object.h"
   5 #include "tag.h"
   6 #include "commit.h"
   7 #include "tree.h"
   8 #include "blob.h"
   9 #include "quote.h"
  10 #include "parse-options.h"
  11
  12 /* Quoting styles */
  13 #define QUOTE_NONE 0
  14 #define QUOTE_SHELL 1
  15 #define QUOTE_PERL 2
  16 #define QUOTE_PYTHON 4
  17 #define QUOTE_TCL 8
  18
  19 typedef enum { FIELD_STR, FIELD_ULONG, FIELD_TIME } cmp_type;
  20
  21 struct atom_value {
  22         const char *s;
  23         unsigned long ul; /* used for sorting when not FIELD_STR */
  24 };
  25
  26 struct ref_sort {
  27         struct ref_sort *next;
  28         int atom; /* index into used_atom array */
  29         unsigned reverse : 1;
  30 };
  31
  32 struct refinfo {
  33         char *refname;
  34         unsigned char objectname[20];
  35         struct atom_value *value;
  36 };
  37
  38 static struct {
  39         const char *name;
  40         cmp_type cmp_type;
  41 } valid_atom[] = {
  42         { "refname" },
  43         { "objecttype" },
  44         { "objectsize", FIELD_ULONG },
  45         { "objectname" },
  46         { "tree" },
  47         { "parent" },
  48         { "numparent", FIELD_ULONG },
  49         { "object" },
  50         { "type" },
  51         { "tag" },
  52         { "author" },
  53         { "authorname" },
  54         { "authoremail" },
  55         { "authordate", FIELD_TIME },
  56         { "committer" },
  57         { "committername" },
  58         { "committeremail" },
  59         { "committerdate", FIELD_TIME },
  60         { "tagger" },
  61         { "taggername" },
  62         { "taggeremail" },
  63         { "taggerdate", FIELD_TIME },
  64         { "creator" },
  65         { "creatordate", FIELD_TIME },
  66         { "subject" },
  67         { "body" },
  68         { "contents" },
  69 };
  70
  71 /*
  72  * An atom is a valid field atom listed above, possibly prefixed with
  73  * a "*" to denote deref_tag().
  74  *
  75  * We parse given format string and sort specifiers, and make a list
  76  * of properties that we need to extract out of objects.  refinfo
  77  * structure will hold an array of values extracted that can be
  78  * indexed with the "atom number", which is an index into this
  79  * array.
  80  */
  81 static const char **used_atom;
  82 static cmp_type *used_atom_type;
  83 static int used_atom_cnt, sort_atom_limit, need_tagged;
  84
  85 /*
  86  * Used to parse format string and sort specifiers
  87  */
  88 static int parse_atom(const char *atom, const char *ep)
  89 {
  90         const char *sp;
  91         int i, at;
  92
  93         sp = atom;
  94         if (*sp == '*' && sp < ep)
  95                 sp++; /* deref */
  96         if (ep <= sp)
  97                 die("malformed field name: %.*s", (int)(ep-atom), atom);
  98
  99         /* Do we have the atom already used elsewhere? */
 100         for (i = 0; i < used_atom_cnt; i++) {
 101                 int len = strlen(used_atom[i]);
 102                 if (len == ep - atom && !memcmp(used_atom[i], atom, len))
 103                         return i;
 104         }
 105
 106         /* Is the atom a valid one? */
 107         for (i = 0; i < ARRAY_SIZE(valid_atom); i++) {
 108                 int len = strlen(valid_atom[i].name);
 109                 /*
 110                  * If the atom name has a colon, strip it and everything after
 111                  * it off - it specifies the format for this entry, and
 112                  * shouldn't be used for checking against the valid_atom
 113                  * table.
 114                  */
 115                 const char *formatp = strchr(sp, ':');
 116                 if (!formatp || ep < formatp)
 117                         formatp = ep;
 118                 if (len == formatp - sp && !memcmp(valid_atom[i].name, sp, len))
 119                         break;
 120         }
 121
 122         if (ARRAY_SIZE(valid_atom) <= i)
 123                 die("unknown field name: %.*s", (int)(ep-atom), atom);
 124
 125         /* Add it in, including the deref prefix */
 126         at = used_atom_cnt;
 127         used_atom_cnt++;
 128         used_atom = xrealloc(used_atom,
 129                              (sizeof *used_atom) * used_atom_cnt);
 130         used_atom_type = xrealloc(used_atom_type,
 131                                   (sizeof(*used_atom_type) * used_atom_cnt));
 132         used_atom[at] = xmemdupz(atom, ep - atom);
 133         used_atom_type[at] = valid_atom[i].cmp_type;
 134         return at;
 135 }
 136
 137 /*
 138  * In a format string, find the next occurrence of %(atom).
 139  */
 140 static const char *find_next(const char *cp)
 141 {
 142         while (*cp) {
 143                 if (*cp == '%') {
 144                         /* %( is the start of an atom;
 145                          * %% is a quoted per-cent.
 146                          */
 147                         if (cp[1] == '(')
 148                                 return cp;
 149                         else if (cp[1] == '%')
 150                                 cp++; /* skip over two % */
 151                         /* otherwise this is a singleton, literal % */
 152                 }
 153                 cp++;
 154         }
 155         return NULL;
 156 }
 157
 158 /*
 159  * Make sure the format string is well formed, and parse out
 160  * the used atoms.
 161  */
 162 static int verify_format(const char *format)
 163 {
 164         const char *cp, *sp;
 165         for (cp = format; *cp && (sp = find_next(cp)); ) {
 166                 const char *ep = strchr(sp, ')');
 167                 if (!ep)
 168                         return error("malformed format string %s", sp);
 169                 /* sp points at "%(" and ep points at the closing ")" */
 170                 parse_atom(sp + 2, ep);
 171                 cp = ep + 1;
 172         }
 173         return 0;
 174 }
 175
 176 /*
 177  * Given an object name, read the object data and size, and return a
 178  * "struct object".  If the object data we are returning is also borrowed
 179  * by the "struct object" representation, set *eaten as well---it is a
 180  * signal from parse_object_buffer to us not to free the buffer.
 181  */
 182 static void *get_obj(const unsigned char *sha1, struct object **obj, unsigned long *sz, int *eaten)
 183 {
 184         enum object_type type;
 185         void *buf = read_sha1_file(sha1, &type, sz);
 186
 187         if (buf)
 188                 *obj = parse_object_buffer(sha1, type, *sz, buf, eaten);
 189         else
 190                 *obj = NULL;
 191         return buf;
 192 }
 193
 194 /* See grab_values */
 195 static void grab_common_values(struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
 196 {
 197         int i;
 198
 199         for (i = 0; i < used_atom_cnt; i++) {
 200                 const char *name = used_atom[i];
 201                 struct atom_value *v = &val[i];
 202                 if (!!deref != (*name == '*'))
 203                         continue;
 204                 if (deref)
 205                         name++;
 206                 if (!strcmp(name, "objecttype"))
 207                         v->s = typename(obj->type);
 208                 else if (!strcmp(name, "objectsize")) {
 209                         char *s = xmalloc(40);
 210                         sprintf(s, "%lu", sz);
 211                         v->ul = sz;
 212                         v->s = s;
 213                 }
 214                 else if (!strcmp(name, "objectname")) {
 215                         char *s = xmalloc(41);
 216                         strcpy(s, sha1_to_hex(obj->sha1));
 217                         v->s = s;
 218                 }
 219         }
 220 }
 221
 222 /* See grab_values */
 223 static void grab_tag_values(struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
 224 {
 225         int i;
 226         struct tag *tag = (struct tag *) obj;
 227
 228         for (i = 0; i < used_atom_cnt; i++) {
 229                 const char *name = used_atom[i];
 230                 struct atom_value *v = &val[i];
 231                 if (!!deref != (*name == '*'))
 232                         continue;
 233                 if (deref)
 234                         name++;
 235                 if (!strcmp(name, "tag"))
 236                         v->s = tag->tag;
 237                 else if (!strcmp(name, "type") && tag->tagged)
 238                         v->s = typename(tag->tagged->type);
 239                 else if (!strcmp(name, "object") && tag->tagged) {
 240                         char *s = xmalloc(41);
 241                         strcpy(s, sha1_to_hex(tag->tagged->sha1));
 242                         v->s = s;
 243                 }
 244         }
 245 }
 246
 247 static int num_parents(struct commit *commit)
 248 {
 249         struct commit_list *parents;
 250         int i;
 251
 252         for (i = 0, parents = commit->parents;
 253              parents;
 254              parents = parents->next)
 255                 i++;
 256         return i;
 257 }
 258
 259 /* See grab_values */
 260 static void grab_commit_values(struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
 261 {
 262         int i;
 263         struct commit *commit = (struct commit *) obj;
 264
 265         for (i = 0; i < used_atom_cnt; i++) {
 266                 const char *name = used_atom[i];
 267                 struct atom_value *v = &val[i];
 268                 if (!!deref != (*name == '*'))
 269                         continue;
 270                 if (deref)
 271                         name++;
 272                 if (!strcmp(name, "tree")) {
 273                         char *s = xmalloc(41);
 274                         strcpy(s, sha1_to_hex(commit->tree->object.sha1));
 275                         v->s = s;
 276                 }
 277                 if (!strcmp(name, "numparent")) {
 278                         char *s = xmalloc(40);
 279                         v->ul = num_parents(commit);
 280                         sprintf(s, "%lu", v->ul);
 281                         v->s = s;
 282                 }
 283                 else if (!strcmp(name, "parent")) {
 284                         int num = num_parents(commit);
 285                         int i;
 286                         struct commit_list *parents;
 287                         char *s = xmalloc(41 * num + 1);
 288                         v->s = s;
 289                         for (i = 0, parents = commit->parents;
 290                              parents;
 291                              parents = parents->next, i = i + 41) {
 292                                 struct commit *parent = parents->item;
 293                                 strcpy(s+i, sha1_to_hex(parent->object.sha1));
 294                                 if (parents->next)
 295                                         s[i+40] = ' ';
 296                         }
 297                         if (!i)
 298                                 *s = '\0';
 299                 }
 300         }
 301 }
 302
 303 static const char *find_wholine(const char *who, int wholen, const char *buf, unsigned long sz)
 304 {
 305         const char *eol;
 306         while (*buf) {
 307                 if (!strncmp(buf, who, wholen) &&
 308                     buf[wholen] == ' ')
 309                         return buf + wholen + 1;
 310                 eol = strchr(buf, '\n');
 311                 if (!eol)
 312                         return "";
 313                 eol++;
 314                 if (*eol == '\n')
 315                         return ""; /* end of header */
 316                 buf = eol;
 317         }
 318         return "";
 319 }
 320
 321 static const char *copy_line(const char *buf)
 322 {
 323         const char *eol = strchrnul(buf, '\n');
 324         return xmemdupz(buf, eol - buf);
 325 }
 326
 327 static const char *copy_name(const char *buf)
 328 {
 329         const char *cp;
 330         for (cp = buf; *cp && *cp != '\n'; cp++) {
 331                 if (!strncmp(cp, " <", 2))
 332                         return xmemdupz(buf, cp - buf);
 333         }
 334         return "";
 335 }
 336
 337 static const char *copy_email(const char *buf)
 338 {
 339         const char *email = strchr(buf, '<');
 340         const char *eoemail = strchr(email, '>');
 341         if (!email || !eoemail)
 342                 return "";
 343         return xmemdupz(email, eoemail + 1 - email);
 344 }
 345
 346 static void grab_date(const char *buf, struct atom_value *v, const char *atomname)
 347 {
 348         const char *eoemail = strstr(buf, "> ");
 349         char *zone;
 350         unsigned long timestamp;
 351         long tz;
 352         enum date_mode date_mode = DATE_NORMAL;
 353         const char *formatp;
 354
 355         /*
 356          * We got here because atomname ends in "date" or "date<something>";
 357          * it's not possible that <something> is not ":<format>" because
 358          * parse_atom() wouldn't have allowed it, so we can assume that no
 359          * ":" means no format is specified, and use the default.
 360          */
 361         formatp = strchr(atomname, ':');
 362         if (formatp != NULL) {
 363                 formatp++;
 364                 date_mode = parse_date_format(formatp);
 365         }
 366
 367         if (!eoemail)
 368                 goto bad;
 369         timestamp = strtoul(eoemail + 2, &zone, 10);
 370         if (timestamp == ULONG_MAX)
 371                 goto bad;
 372         tz = strtol(zone, NULL, 10);
 373         if ((tz == LONG_MIN || tz == LONG_MAX) && errno == ERANGE)
 374                 goto bad;
 375         v->s = xstrdup(show_date(timestamp, tz, date_mode));
 376         v->ul = timestamp;
 377         return;
 378  bad:
 379         v->s = "";
 380         v->ul = 0;
 381 }
 382
 383 /* See grab_values */
 384 static void grab_person(const char *who, struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
 385 {
 386         int i;
 387         int wholen = strlen(who);
 388         const char *wholine = NULL;
 389
 390         for (i = 0; i < used_atom_cnt; i++) {
 391                 const char *name = used_atom[i];
 392                 struct atom_value *v = &val[i];
 393                 if (!!deref != (*name == '*'))
 394                         continue;
 395                 if (deref)
 396                         name++;
 397                 if (strncmp(who, name, wholen))
 398                         continue;
 399                 if (name[wholen] != 0 &&
 400                     strcmp(name + wholen, "name") &&
 401                     strcmp(name + wholen, "email") &&
 402                     prefixcmp(name + wholen, "date"))
 403                         continue;
 404                 if (!wholine)
 405                         wholine = find_wholine(who, wholen, buf, sz);
 406                 if (!wholine)
 407                         return; /* no point looking for it */
 408                 if (name[wholen] == 0)
 409                         v->s = copy_line(wholine);
 410                 else if (!strcmp(name + wholen, "name"))
 411                         v->s = copy_name(wholine);
 412                 else if (!strcmp(name + wholen, "email"))
 413                         v->s = copy_email(wholine);
 414                 else if (!prefixcmp(name + wholen, "date"))
 415                         grab_date(wholine, v, name);
 416         }
 417
 418         /* For a tag or a commit object, if "creator" or "creatordate" is
 419          * requested, do something special.
 420          */
 421         if (strcmp(who, "tagger") && strcmp(who, "committer"))
 422                 return; /* "author" for commit object is not wanted */
 423         if (!wholine)
 424                 wholine = find_wholine(who, wholen, buf, sz);
 425         if (!wholine)
 426                 return;
 427         for (i = 0; i < used_atom_cnt; i++) {
 428                 const char *name = used_atom[i];
 429                 struct atom_value *v = &val[i];
 430                 if (!!deref != (*name == '*'))
 431                         continue;
 432                 if (deref)
 433                         name++;
 434
 435                 if (!prefixcmp(name, "creatordate"))
 436                         grab_date(wholine, v, name);
 437                 else if (!strcmp(name, "creator"))
 438                         v->s = copy_line(wholine);
 439         }
 440 }
 441
 442 static void find_subpos(const char *buf, unsigned long sz, const char **sub, const char **body)
 443 {
 444         while (*buf) {
 445                 const char *eol = strchr(buf, '\n');
 446                 if (!eol)
 447                         return;
 448                 if (eol[1] == '\n') {
 449                         buf = eol + 1;
 450                         break; /* found end of header */
 451                 }
 452                 buf = eol + 1;
 453         }
 454         while (*buf == '\n')
 455                 buf++;
 456         if (!*buf)
 457                 return;
 458         *sub = buf; /* first non-empty line */
 459         buf = strchr(buf, '\n');
 460         if (!buf) {
 461                 *body = "";
 462                 return; /* no body */
 463         }
 464         while (*buf == '\n')
 465                 buf++; /* skip blank between subject and body */
 466         *body = buf;
 467 }
 468
 469 /* See grab_values */
 470 static void grab_sub_body_contents(struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
 471 {
 472         int i;
 473         const char *subpos = NULL, *bodypos = NULL;
 474
 475         for (i = 0; i < used_atom_cnt; i++) {
 476                 const char *name = used_atom[i];
 477                 struct atom_value *v = &val[i];
 478                 if (!!deref != (*name == '*'))
 479                         continue;
 480                 if (deref)
 481                         name++;
 482                 if (strcmp(name, "subject") &&
 483                     strcmp(name, "body") &&
 484                     strcmp(name, "contents"))
 485                         continue;
 486                 if (!subpos)
 487                         find_subpos(buf, sz, &subpos, &bodypos);
 488                 if (!subpos)
 489                         return;
 490
 491                 if (!strcmp(name, "subject"))
 492                         v->s = copy_line(subpos);
 493                 else if (!strcmp(name, "body"))
 494                         v->s = xstrdup(bodypos);
 495                 else if (!strcmp(name, "contents"))
 496                         v->s = xstrdup(subpos);
 497         }
 498 }
 499
 500 /* We want to have empty print-string for field requests
 501  * that do not apply (e.g. "authordate" for a tag object)
 502  */
 503 static void fill_missing_values(struct atom_value *val)
 504 {
 505         int i;
 506         for (i = 0; i < used_atom_cnt; i++) {
 507                 struct atom_value *v = &val[i];
 508                 if (v->s == NULL)
 509                         v->s = "";
 510         }
 511 }
 512
 513 /*
 514  * val is a list of atom_value to hold returned values.  Extract
 515  * the values for atoms in used_atom array out of (obj, buf, sz).
 516  * when deref is false, (obj, buf, sz) is the object that is
 517  * pointed at by the ref itself; otherwise it is the object the
 518  * ref (which is a tag) refers to.
 519  */
 520 static void grab_values(struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
 521 {
 522         grab_common_values(val, deref, obj, buf, sz);
 523         switch (obj->type) {
 524         case OBJ_TAG:
 525                 grab_tag_values(val, deref, obj, buf, sz);
 526                 grab_sub_body_contents(val, deref, obj, buf, sz);
 527                 grab_person("tagger", val, deref, obj, buf, sz);
 528                 break;
 529         case OBJ_COMMIT:
 530                 grab_commit_values(val, deref, obj, buf, sz);
 531                 grab_sub_body_contents(val, deref, obj, buf, sz);
 532                 grab_person("author", val, deref, obj, buf, sz);
 533                 grab_person("committer", val, deref, obj, buf, sz);
 534                 break;
 535         case OBJ_TREE:
 536                 // grab_tree_values(val, deref, obj, buf, sz);
 537                 break;
 538         case OBJ_BLOB:
 539                 // grab_blob_values(val, deref, obj, buf, sz);
 540                 break;
 541         default:
 542                 die("Eh?  Object of type %d?", obj->type);
 543         }
 544 }
 545
 546 /*
 547  * generate a format suitable for scanf from a ref_rev_parse_rules
 548  * rule, that is replace the "%.*s" spec with a "%s" spec
 549  */
 550 static void gen_scanf_fmt(char *scanf_fmt, const char *rule)
 551 {
 552         char *spec;
 553
 554         spec = strstr(rule, "%.*s");
 555         if (!spec || strstr(spec + 4, "%.*s"))
 556                 die("invalid rule in ref_rev_parse_rules: %s", rule);
 557
 558         /* copy all until spec */
 559         strncpy(scanf_fmt, rule, spec - rule);
 560         scanf_fmt[spec - rule] = '\0';
 561         /* copy new spec */
 562         strcat(scanf_fmt, "%s");
 563         /* copy remaining rule */
 564         strcat(scanf_fmt, spec + 4);
 565
 566         return;
 567 }
 568
 569 /*
 570  * Shorten the refname to an non-ambiguous form
 571  */
 572 static char *get_short_ref(struct refinfo *ref)
 573 {
 574         int i;
 575         static char **scanf_fmts;
 576         static int nr_rules;
 577         char *short_name;
 578
 579         /* pre generate scanf formats from ref_rev_parse_rules[] */
 580         if (!nr_rules) {
 581                 size_t total_len = 0;
 582
 583                 /* the rule list is NULL terminated, count them first */
 584                 for (; ref_rev_parse_rules[nr_rules]; nr_rules++)
 585                         /* no +1 because strlen("%s") < strlen("%.*s") */
 586                         total_len += strlen(ref_rev_parse_rules[nr_rules]);
 587
 588                 scanf_fmts = xmalloc(nr_rules * sizeof(char *) + total_len);
 589
 590                 total_len = 0;
 591                 for (i = 0; i < nr_rules; i++) {
 592                         scanf_fmts[i] = (char *)&scanf_fmts[nr_rules]
 593                                         + total_len;
 594                         gen_scanf_fmt(scanf_fmts[i], ref_rev_parse_rules[i]);
 595                         total_len += strlen(ref_rev_parse_rules[i]);
 596                 }
 597         }
 598
 599         /* bail out if there are no rules */
 600         if (!nr_rules)
 601                 return ref->refname;
 602
 603         /* buffer for scanf result, at most ref->refname must fit */
 604         short_name = xstrdup(ref->refname);
 605
 606         /* skip first rule, it will always match */
 607         for (i = nr_rules - 1; i > 0 ; --i) {
 608                 int j;
 609                 int short_name_len;
 610
 611                 if (1 != sscanf(ref->refname, scanf_fmts[i], short_name))
 612                         continue;
 613
 614                 short_name_len = strlen(short_name);
 615
 616                 /*
 617                  * check if the short name resolves to a valid ref,
 618                  * but use only rules prior to the matched one
 619                  */
 620                 for (j = 0; j < i; j++) {
 621                         const char *rule = ref_rev_parse_rules[j];
 622                         unsigned char short_objectname[20];
 623                         char refname[PATH_MAX];
 624
 625                         /*
 626                          * the short name is ambiguous, if it resolves
 627                          * (with this previous rule) to a valid ref
 628                          * read_ref() returns 0 on success
 629                          */
 630                         mksnpath(refname, sizeof(refname),
 631                                  rule, short_name_len, short_name);
 632                         if (!read_ref(refname, short_objectname))
 633                                 break;
 634                 }
 635
 636                 /*
 637                  * short name is non-ambiguous if all previous rules
 638                  * haven't resolved to a valid ref
 639                  */
 640                 if (j == i)
 641                         return short_name;
 642         }
 643
 644         free(short_name);
 645         return ref->refname;
 646 }
 647
 648
 649 /*
 650  * Parse the object referred by ref, and grab needed value.
 651  */
 652 static void populate_value(struct refinfo *ref)
 653 {
 654         void *buf;
 655         struct object *obj;
 656         int eaten, i;
 657         unsigned long size;
 658         const unsigned char *tagged;
 659
 660         ref->value = xcalloc(sizeof(struct atom_value), used_atom_cnt);
 661
 662         buf = get_obj(ref->objectname, &obj, &size, &eaten);
 663         if (!buf)
 664                 die("missing object %s for %s",
 665                     sha1_to_hex(ref->objectname), ref->refname);
 666         if (!obj)
 667                 die("parse_object_buffer failed on %s for %s",
 668                     sha1_to_hex(ref->objectname), ref->refname);
 669
 670         /* Fill in specials first */
 671         for (i = 0; i < used_atom_cnt; i++) {
 672                 const char *name = used_atom[i];
 673                 struct atom_value *v = &ref->value[i];
 674                 int deref = 0;
 675                 if (*name == '*') {
 676                         deref = 1;
 677                         name++;
 678                 }
 679                 if (!prefixcmp(name, "refname")) {
 680                         const char *formatp = strchr(name, ':');
 681                         const char *refname = ref->refname;
 682
 683                         /* look for "short" refname format */
 684                         if (formatp) {
 685                                 formatp++;
 686                                 if (!strcmp(formatp, "short"))
 687                                         refname = get_short_ref(ref);
 688                                 else
 689                                         die("unknown refname format %s",
 690                                             formatp);
 691                         }
 692
 693                         if (!deref)
 694                                 v->s = refname;
 695                         else {
 696                                 int len = strlen(refname);
 697                                 char *s = xmalloc(len + 4);
 698                                 sprintf(s, "%s^{}", refname);
 699                                 v->s = s;
 700                         }
 701                 }
 702         }
 703
 704         grab_values(ref->value, 0, obj, buf, size);
 705         if (!eaten)
 706                 free(buf);
 707
 708         /* If there is no atom that wants to know about tagged
 709          * object, we are done.
 710          */
 711         if (!need_tagged || (obj->type != OBJ_TAG))
 712                 return;
 713
 714         /* If it is a tag object, see if we use a value that derefs
 715          * the object, and if we do grab the object it refers to.
 716          */
 717         tagged = ((struct tag *)obj)->tagged->sha1;
 718
 719         /* NEEDSWORK: This derefs tag only once, which
 720          * is good to deal with chains of trust, but
 721          * is not consistent with what deref_tag() does
 722          * which peels the onion to the core.
 723          */
 724         buf = get_obj(tagged, &obj, &size, &eaten);
 725         if (!buf)
 726                 die("missing object %s for %s",
 727                     sha1_to_hex(tagged), ref->refname);
 728         if (!obj)
 729                 die("parse_object_buffer failed on %s for %s",
 730                     sha1_to_hex(tagged), ref->refname);
 731         grab_values(ref->value, 1, obj, buf, size);
 732         if (!eaten)
 733                 free(buf);
 734 }
 735
 736 /*
 737  * Given a ref, return the value for the atom.  This lazily gets value
 738  * out of the object by calling populate value.
 739  */
 740 static void get_value(struct refinfo *ref, int atom, struct atom_value **v)
 741 {
 742         if (!ref->value) {
 743                 populate_value(ref);
 744                 fill_missing_values(ref->value);
 745         }
 746         *v = &ref->value[atom];
 747 }
 748
 749 struct grab_ref_cbdata {
 750         struct refinfo **grab_array;
 751         const char **grab_pattern;
 752         int grab_cnt;
 753 };
 754
 755 /*
 756  * A call-back given to for_each_ref().  It is unfortunate that we
 757  * need to use global variables to pass extra information to this
 758  * function.
 759  */
 760 static int grab_single_ref(const char *refname, const unsigned char *sha1, int flag, void *cb_data)
 761 {
 762         struct grab_ref_cbdata *cb = cb_data;
 763         struct refinfo *ref;
 764         int cnt;
 765
 766         if (*cb->grab_pattern) {
 767                 const char **pattern;
 768                 int namelen = strlen(refname);
 769                 for (pattern = cb->grab_pattern; *pattern; pattern++) {
 770                         const char *p = *pattern;
 771                         int plen = strlen(p);
 772
 773                         if ((plen <= namelen) &&
 774                             !strncmp(refname, p, plen) &&
 775                             (refname[plen] == '\0' ||
 776                              refname[plen] == '/' ||
 777                              p[plen-1] == '/'))
 778                                 break;
 779                         if (!fnmatch(p, refname, FNM_PATHNAME))
 780                                 break;
 781                 }
 782                 if (!*pattern)
 783                         return 0;
 784         }
 785
 786         /* We do not open the object yet; sort may only need refname
 787          * to do its job and the resulting list may yet to be pruned
 788          * by maxcount logic.
 789          */
 790         ref = xcalloc(1, sizeof(*ref));
 791         ref->refname = xstrdup(refname);
 792         hashcpy(ref->objectname, sha1);
 793
 794         cnt = cb->grab_cnt;
 795         cb->grab_array = xrealloc(cb->grab_array,
 796                                   sizeof(*cb->grab_array) * (cnt + 1));
 797         cb->grab_array[cnt++] = ref;
 798         cb->grab_cnt = cnt;
 799         return 0;
 800 }
 801
 802 static int cmp_ref_sort(struct ref_sort *s, struct refinfo *a, struct refinfo *b)
 803 {
 804         struct atom_value *va, *vb;
 805         int cmp;
 806         cmp_type cmp_type = used_atom_type[s->atom];
 807
 808         get_value(a, s->atom, &va);
 809         get_value(b, s->atom, &vb);
 810         switch (cmp_type) {
 811         case FIELD_STR:
 812                 cmp = strcmp(va->s, vb->s);
 813                 break;
 814         default:
 815                 if (va->ul < vb->ul)
 816                         cmp = -1;
 817                 else if (va->ul == vb->ul)
 818                         cmp = 0;
 819                 else
 820                         cmp = 1;
 821                 break;
 822         }
 823         return (s->reverse) ? -cmp : cmp;
 824 }
 825
 826 static struct ref_sort *ref_sort;
 827 static int compare_refs(const void *a_, const void *b_)
 828 {
 829         struct refinfo *a = *((struct refinfo **)a_);
 830         struct refinfo *b = *((struct refinfo **)b_);
 831         struct ref_sort *s;
 832
 833         for (s = ref_sort; s; s = s->next) {
 834                 int cmp = cmp_ref_sort(s, a, b);
 835                 if (cmp)
 836                         return cmp;
 837         }
 838         return 0;
 839 }
 840
 841 static void sort_refs(struct ref_sort *sort, struct refinfo **refs, int num_refs)
 842 {
 843         ref_sort = sort;
 844         qsort(refs, num_refs, sizeof(struct refinfo *), compare_refs);
 845 }
 846
 847 static void print_value(struct refinfo *ref, int atom, int quote_style)
 848 {
 849         struct atom_value *v;
 850         get_value(ref, atom, &v);
 851         switch (quote_style) {
 852         case QUOTE_NONE:
 853                 fputs(v->s, stdout);
 854                 break;
 855         case QUOTE_SHELL:
 856                 sq_quote_print(stdout, v->s);
 857                 break;
 858         case QUOTE_PERL:
 859                 perl_quote_print(stdout, v->s);
 860                 break;
 861         case QUOTE_PYTHON:
 862                 python_quote_print(stdout, v->s);
 863                 break;
 864         case QUOTE_TCL:
 865                 tcl_quote_print(stdout, v->s);
 866                 break;
 867         }
 868 }
 869
 870 static int hex1(char ch)
 871 {
 872         if ('0' <= ch && ch <= '9')
 873                 return ch - '0';
 874         else if ('a' <= ch && ch <= 'f')
 875                 return ch - 'a' + 10;
 876         else if ('A' <= ch && ch <= 'F')
 877                 return ch - 'A' + 10;
 878         return -1;
 879 }
 880 static int hex2(const char *cp)
 881 {
 882         if (cp[0] && cp[1])
 883                 return (hex1(cp[0]) << 4) | hex1(cp[1]);
 884         else
 885                 return -1;
 886 }
 887
 888 static void emit(const char *cp, const char *ep)
 889 {
 890         while (*cp && (!ep || cp < ep)) {
 891                 if (*cp == '%') {
 892                         if (cp[1] == '%')
 893                                 cp++;
 894                         else {
 895                                 int ch = hex2(cp + 1);
 896                                 if (0 <= ch) {
 897                                         putchar(ch);
 898                                         cp += 3;
 899                                         continue;
 900                                 }
 901                         }
 902                 }
 903                 putchar(*cp);
 904                 cp++;
 905         }
 906 }
 907
 908 static void show_ref(struct refinfo *info, const char *format, int quote_style)
 909 {
 910         const char *cp, *sp, *ep;
 911
 912         for (cp = format; *cp && (sp = find_next(cp)); cp = ep + 1) {
 913                 ep = strchr(sp, ')');
 914                 if (cp < sp)
 915                         emit(cp, sp);
 916                 print_value(info, parse_atom(sp + 2, ep), quote_style);
 917         }
 918         if (*cp) {
 919                 sp = cp + strlen(cp);
 920                 emit(cp, sp);
 921         }
 922         putchar('\n');
 923 }
 924
 925 static struct ref_sort *default_sort(void)
 926 {
 927         static const char cstr_name[] = "refname";
 928
 929         struct ref_sort *sort = xcalloc(1, sizeof(*sort));
 930
 931         sort->next = NULL;
 932         sort->atom = parse_atom(cstr_name, cstr_name + strlen(cstr_name));
 933         return sort;
 934 }
 935
 936 static int opt_parse_sort(const struct option *opt, const char *arg, int unset)
 937 {
 938         struct ref_sort **sort_tail = opt->value;
 939         struct ref_sort *s;
 940         int len;
 941
 942         if (!arg) /* should --no-sort void the list ? */
 943                 return -1;
 944
 945         *sort_tail = s = xcalloc(1, sizeof(*s));
 946         sort_tail = &s->next;
 947
 948         if (*arg == '-') {
 949                 s->reverse = 1;
 950                 arg++;
 951         }
 952         len = strlen(arg);
 953         s->atom = parse_atom(arg, arg+len);
 954         return 0;
 955 }
 956
 957 static char const * const for_each_ref_usage[] = {
 958         "git for-each-ref [options] [<pattern>]",
 959         NULL
 960 };
 961
 962 int cmd_for_each_ref(int argc, const char **argv, const char *prefix)
 963 {
 964         int i, num_refs;
 965         const char *format = "%(objectname) %(objecttype)\t%(refname)";
 966         struct ref_sort *sort = NULL, **sort_tail = &sort;
 967         int maxcount = 0, quote_style = 0;
 968         struct refinfo **refs;
 969         struct grab_ref_cbdata cbdata;
 970
 971         struct option opts[] = {
 972                 OPT_BIT('s', "shell", &quote_style,
 973                         "quote placeholders suitably for shells", QUOTE_SHELL),
 974                 OPT_BIT('p', "perl",  &quote_style,
 975                         "quote placeholders suitably for perl", QUOTE_PERL),
 976                 OPT_BIT(0 , "python", &quote_style,
 977                         "quote placeholders suitably for python", QUOTE_PYTHON),
 978                 OPT_BIT(0 , "tcl",  &quote_style,
 979                         "quote placeholders suitably for tcl", QUOTE_TCL),
 980
 981                 OPT_GROUP(""),
 982                 OPT_INTEGER( 0 , "count", &maxcount, "show only <n> matched refs"),
 983                 OPT_STRING(  0 , "format", &format, "format", "format to use for the output"),
 984                 OPT_CALLBACK(0 , "sort", sort_tail, "key",
 985                             "field name to sort on", &opt_parse_sort),
 986                 OPT_END(),
 987         };
 988
 989         parse_options(argc, argv, opts, for_each_ref_usage, 0);
 990         if (maxcount < 0) {
 991                 error("invalid --count argument: `%d'", maxcount);
 992                 usage_with_options(for_each_ref_usage, opts);
 993         }
 994         if (HAS_MULTI_BITS(quote_style)) {
 995                 error("more than one quoting style?");
 996                 usage_with_options(for_each_ref_usage, opts);
 997         }
 998         if (verify_format(format))
 999                 usage_with_options(for_each_ref_usage, opts);
1000
1001         if (!sort)
1002                 sort = default_sort();
1003         sort_atom_limit = used_atom_cnt;
1004
1005         memset(&cbdata, 0, sizeof(cbdata));
1006         cbdata.grab_pattern = argv;
1007         for_each_ref(grab_single_ref, &cbdata);
1008         refs = cbdata.grab_array;
1009         num_refs = cbdata.grab_cnt;
1010
1011         for (i = 0; i < used_atom_cnt; i++) {
1012                 if (used_atom[i][0] == '*') {
1013                         need_tagged = 1;
1014                         break;
1015                 }
1016         }
1017
1018         sort_refs(sort, refs, num_refs);
1019
1020         if (!maxcount || num_refs < maxcount)
1021                 maxcount = num_refs;
1022         for (i = 0; i < maxcount; i++)
1023                 show_ref(refs[i], format, quote_style);
1024         return 0;
1025 }