builtin/cat-file.c

   1 /*
   2  * GIT - The information manager from hell
   3  *
   4  * Copyright (C) Linus Torvalds, 2005
   5  */
   6 #define USE_THE_INDEX_COMPATIBILITY_MACROS
   7 #include "cache.h"
   8 #include "config.h"
   9 #include "builtin.h"
  10 #include "diff.h"
  11 #include "parse-options.h"
  12 #include "userdiff.h"
  13 #include "streaming.h"
  14 #include "tree-walk.h"
  15 #include "oid-array.h"
  16 #include "packfile.h"
  17 #include "object-store.h"
  18 #include "promisor-remote.h"
  19
  20 enum batch_mode {
  21         BATCH_MODE_CONTENTS,
  22         BATCH_MODE_INFO,
  23         BATCH_MODE_QUEUE_AND_DISPATCH,
  24 };
  25
  26 struct batch_options {
  27         int enabled;
  28         int follow_symlinks;
  29         enum batch_mode batch_mode;
  30         int buffer_output;
  31         int all_objects;
  32         int unordered;
  33         int transform_mode; /* may be 'w' or 'c' for --filters or --textconv */
  34         const char *format;
  35 };
  36
  37 static const char *force_path;
  38
  39 static int filter_object(const char *path, unsigned mode,
  40                          const struct object_id *oid,
  41                          char **buf, unsigned long *size)
  42 {
  43         enum object_type type;
  44
  45         *buf = read_object_file(oid, &type, size);
  46         if (!*buf)
  47                 return error(_("cannot read object %s '%s'"),
  48                              oid_to_hex(oid), path);
  49         if ((type == OBJ_BLOB) && S_ISREG(mode)) {
  50                 struct strbuf strbuf = STRBUF_INIT;
  51                 struct checkout_metadata meta;
  52
  53                 init_checkout_metadata(&meta, NULL, NULL, oid);
  54                 if (convert_to_working_tree(&the_index, path, *buf, *size, &strbuf, &meta)) {
  55                         free(*buf);
  56                         *size = strbuf.len;
  57                         *buf = strbuf_detach(&strbuf, NULL);
  58                 }
  59         }
  60
  61         return 0;
  62 }
  63
  64 static int stream_blob(const struct object_id *oid)
  65 {
  66         if (stream_blob_to_fd(1, oid, NULL, 0))
  67                 die("unable to stream %s to stdout", oid_to_hex(oid));
  68         return 0;
  69 }
  70
  71 static int cat_one_file(int opt, const char *exp_type, const char *obj_name,
  72                         int unknown_type)
  73 {
  74         struct object_id oid;
  75         enum object_type type;
  76         char *buf;
  77         unsigned long size;
  78         struct object_context obj_context;
  79         struct object_info oi = OBJECT_INFO_INIT;
  80         struct strbuf sb = STRBUF_INIT;
  81         unsigned flags = OBJECT_INFO_LOOKUP_REPLACE;
  82         unsigned get_oid_flags = GET_OID_RECORD_PATH | GET_OID_ONLY_TO_DIE;
  83         const char *path = force_path;
  84         const int opt_cw = (opt == 'c' || opt == 'w');
  85         if (!path && opt_cw)
  86                 get_oid_flags |= GET_OID_REQUIRE_PATH;
  87
  88         if (unknown_type)
  89                 flags |= OBJECT_INFO_ALLOW_UNKNOWN_TYPE;
  90
  91         if (get_oid_with_context(the_repository, obj_name, get_oid_flags, &oid,
  92                                  &obj_context))
  93                 die("Not a valid object name %s", obj_name);
  94
  95         if (!path)
  96                 path = obj_context.path;
  97         if (obj_context.mode == S_IFINVALID)
  98                 obj_context.mode = 0100644;
  99
 100         buf = NULL;
 101         switch (opt) {
 102         case 't':
 103                 oi.type_name = &sb;
 104                 if (oid_object_info_extended(the_repository, &oid, &oi, flags) < 0)
 105                         die("git cat-file: could not get object info");
 106                 if (sb.len) {
 107                         printf("%s\n", sb.buf);
 108                         strbuf_release(&sb);
 109                         return 0;
 110                 }
 111                 break;
 112
 113         case 's':
 114                 oi.sizep = &size;
 115                 if (oid_object_info_extended(the_repository, &oid, &oi, flags) < 0)
 116                         die("git cat-file: could not get object info");
 117                 printf("%"PRIuMAX"\n", (uintmax_t)size);
 118                 return 0;
 119
 120         case 'e':
 121                 return !has_object_file(&oid);
 122
 123         case 'w':
 124
 125                 if (filter_object(path, obj_context.mode,
 126                                   &oid, &buf, &size))
 127                         return -1;
 128                 break;
 129
 130         case 'c':
 131                 if (textconv_object(the_repository, path, obj_context.mode,
 132                                     &oid, 1, &buf, &size))
 133                         break;
 134                 /* else fallthrough */
 135
 136         case 'p':
 137                 type = oid_object_info(the_repository, &oid, NULL);
 138                 if (type < 0)
 139                         die("Not a valid object name %s", obj_name);
 140
 141                 /* custom pretty-print here */
 142                 if (type == OBJ_TREE) {
 143                         const char *ls_args[3] = { NULL };
 144                         ls_args[0] =  "ls-tree";
 145                         ls_args[1] =  obj_name;
 146                         return cmd_ls_tree(2, ls_args, NULL);
 147                 }
 148
 149                 if (type == OBJ_BLOB)
 150                         return stream_blob(&oid);
 151                 buf = read_object_file(&oid, &type, &size);
 152                 if (!buf)
 153                         die("Cannot read object %s", obj_name);
 154
 155                 /* otherwise just spit out the data */
 156                 break;
 157
 158         case 0:
 159         {
 160                 enum object_type exp_type_id = type_from_string(exp_type);
 161
 162                 if (exp_type_id == OBJ_BLOB) {
 163                         struct object_id blob_oid;
 164                         if (oid_object_info(the_repository, &oid, NULL) == OBJ_TAG) {
 165                                 char *buffer = read_object_file(&oid, &type,
 166                                                                 &size);
 167                                 const char *target;
 168                                 if (!skip_prefix(buffer, "object ", &target) ||
 169                                     get_oid_hex(target, &blob_oid))
 170                                         die("%s not a valid tag", oid_to_hex(&oid));
 171                                 free(buffer);
 172                         } else
 173                                 oidcpy(&blob_oid, &oid);
 174
 175                         if (oid_object_info(the_repository, &blob_oid, NULL) == OBJ_BLOB)
 176                                 return stream_blob(&blob_oid);
 177                         /*
 178                          * we attempted to dereference a tag to a blob
 179                          * and failed; there may be new dereference
 180                          * mechanisms this code is not aware of.
 181                          * fall-back to the usual case.
 182                          */
 183                 }
 184                 buf = read_object_with_reference(the_repository, &oid,
 185                                                  exp_type_id, &size, NULL);
 186                 break;
 187         }
 188         default:
 189                 die("git cat-file: unknown option: %s", exp_type);
 190         }
 191
 192         if (!buf)
 193                 die("git cat-file %s: bad file", obj_name);
 194
 195         write_or_die(1, buf, size);
 196         free(buf);
 197         free(obj_context.path);
 198         return 0;
 199 }
 200
 201 struct expand_data {
 202         struct object_id oid;
 203         enum object_type type;
 204         unsigned long size;
 205         off_t disk_size;
 206         const char *rest;
 207         struct object_id delta_base_oid;
 208
 209         /*
 210          * If mark_query is true, we do not expand anything, but rather
 211          * just mark the object_info with items we wish to query.
 212          */
 213         int mark_query;
 214
 215         /*
 216          * Whether to split the input on whitespace before feeding it to
 217          * get_sha1; this is decided during the mark_query phase based on
 218          * whether we have a %(rest) token in our format.
 219          */
 220         int split_on_whitespace;
 221
 222         /*
 223          * After a mark_query run, this object_info is set up to be
 224          * passed to oid_object_info_extended. It will point to the data
 225          * elements above, so you can retrieve the response from there.
 226          */
 227         struct object_info info;
 228
 229         /*
 230          * This flag will be true if the requested batch format and options
 231          * don't require us to call oid_object_info, which can then be
 232          * optimized out.
 233          */
 234         unsigned skip_object_info : 1;
 235 };
 236
 237 static int is_atom(const char *atom, const char *s, int slen)
 238 {
 239         int alen = strlen(atom);
 240         return alen == slen && !memcmp(atom, s, alen);
 241 }
 242
 243 static void expand_atom(struct strbuf *sb, const char *atom, int len,
 244                         void *vdata)
 245 {
 246         struct expand_data *data = vdata;
 247
 248         if (is_atom("objectname", atom, len)) {
 249                 if (!data->mark_query)
 250                         strbuf_addstr(sb, oid_to_hex(&data->oid));
 251         } else if (is_atom("objecttype", atom, len)) {
 252                 if (data->mark_query)
 253                         data->info.typep = &data->type;
 254                 else
 255                         strbuf_addstr(sb, type_name(data->type));
 256         } else if (is_atom("objectsize", atom, len)) {
 257                 if (data->mark_query)
 258                         data->info.sizep = &data->size;
 259                 else
 260                         strbuf_addf(sb, "%"PRIuMAX , (uintmax_t)data->size);
 261         } else if (is_atom("objectsize:disk", atom, len)) {
 262                 if (data->mark_query)
 263                         data->info.disk_sizep = &data->disk_size;
 264                 else
 265                         strbuf_addf(sb, "%"PRIuMAX, (uintmax_t)data->disk_size);
 266         } else if (is_atom("rest", atom, len)) {
 267                 if (data->mark_query)
 268                         data->split_on_whitespace = 1;
 269                 else if (data->rest)
 270                         strbuf_addstr(sb, data->rest);
 271         } else if (is_atom("deltabase", atom, len)) {
 272                 if (data->mark_query)
 273                         data->info.delta_base_oid = &data->delta_base_oid;
 274                 else
 275                         strbuf_addstr(sb,
 276                                       oid_to_hex(&data->delta_base_oid));
 277         } else
 278                 die("unknown format element: %.*s", len, atom);
 279 }
 280
 281 static size_t expand_format(struct strbuf *sb, const char *start, void *data)
 282 {
 283         const char *end;
 284
 285         if (*start != '(')
 286                 return 0;
 287         end = strchr(start + 1, ')');
 288         if (!end)
 289                 die("format element '%s' does not end in ')'", start);
 290
 291         expand_atom(sb, start + 1, end - start - 1, data);
 292
 293         return end - start + 1;
 294 }
 295
 296 static void batch_write(struct batch_options *opt, const void *data, int len)
 297 {
 298         if (opt->buffer_output) {
 299                 if (fwrite(data, 1, len, stdout) != len)
 300                         die_errno("unable to write to stdout");
 301         } else
 302                 write_or_die(1, data, len);
 303 }
 304
 305 static void print_object_or_die(struct batch_options *opt, struct expand_data *data)
 306 {
 307         const struct object_id *oid = &data->oid;
 308
 309         assert(data->info.typep);
 310
 311         if (data->type == OBJ_BLOB) {
 312                 if (opt->buffer_output)
 313                         fflush(stdout);
 314                 if (opt->transform_mode) {
 315                         char *contents;
 316                         unsigned long size;
 317
 318                         if (!data->rest)
 319                                 die("missing path for '%s'", oid_to_hex(oid));
 320
 321                         if (opt->transform_mode == 'w') {
 322                                 if (filter_object(data->rest, 0100644, oid,
 323                                                   &contents, &size))
 324                                         die("could not convert '%s' %s",
 325                                             oid_to_hex(oid), data->rest);
 326                         } else if (opt->transform_mode == 'c') {
 327                                 enum object_type type;
 328                                 if (!textconv_object(the_repository,
 329                                                      data->rest, 0100644, oid,
 330                                                      1, &contents, &size))
 331                                         contents = read_object_file(oid,
 332                                                                     &type,
 333                                                                     &size);
 334                                 if (!contents)
 335                                         die("could not convert '%s' %s",
 336                                             oid_to_hex(oid), data->rest);
 337                         } else
 338                                 BUG("invalid transform_mode: %c", opt->transform_mode);
 339                         batch_write(opt, contents, size);
 340                         free(contents);
 341                 } else {
 342                         stream_blob(oid);
 343                 }
 344         }
 345         else {
 346                 enum object_type type;
 347                 unsigned long size;
 348                 void *contents;
 349
 350                 contents = read_object_file(oid, &type, &size);
 351                 if (!contents)
 352                         die("object %s disappeared", oid_to_hex(oid));
 353                 if (type != data->type)
 354                         die("object %s changed type!?", oid_to_hex(oid));
 355                 if (data->info.sizep && size != data->size)
 356                         die("object %s changed size!?", oid_to_hex(oid));
 357
 358                 batch_write(opt, contents, size);
 359                 free(contents);
 360         }
 361 }
 362
 363 static void print_default_format(struct strbuf *scratch, struct expand_data *data)
 364 {
 365         strbuf_addf(scratch, "%s %s %"PRIuMAX"\n", oid_to_hex(&data->oid),
 366                     type_name(data->type),
 367                     (uintmax_t)data->size);
 368 }
 369
 370 /*
 371  * If "pack" is non-NULL, then "offset" is the byte offset within the pack from
 372  * which the object may be accessed (though note that we may also rely on
 373  * data->oid, too). If "pack" is NULL, then offset is ignored.
 374  */
 375 static void batch_object_write(const char *obj_name,
 376                                struct strbuf *scratch,
 377                                struct batch_options *opt,
 378                                struct expand_data *data,
 379                                struct packed_git *pack,
 380                                off_t offset)
 381 {
 382         if (!data->skip_object_info) {
 383                 int ret;
 384
 385                 if (pack)
 386                         ret = packed_object_info(the_repository, pack, offset,
 387                                                  &data->info);
 388                 else
 389                         ret = oid_object_info_extended(the_repository,
 390                                                        &data->oid, &data->info,
 391                                                        OBJECT_INFO_LOOKUP_REPLACE);
 392                 if (ret < 0) {
 393                         printf("%s missing\n",
 394                                obj_name ? obj_name : oid_to_hex(&data->oid));
 395                         fflush(stdout);
 396                         return;
 397                 }
 398         }
 399
 400         strbuf_reset(scratch);
 401
 402         if (!opt->format) {
 403                 print_default_format(scratch, data);
 404         } else {
 405                 strbuf_expand(scratch, opt->format, expand_format, data);
 406                 strbuf_addch(scratch, '\n');
 407         }
 408
 409         batch_write(opt, scratch->buf, scratch->len);
 410
 411         if (opt->batch_mode == BATCH_MODE_CONTENTS) {
 412                 print_object_or_die(opt, data);
 413                 batch_write(opt, "\n", 1);
 414         }
 415 }
 416
 417 static void batch_one_object(const char *obj_name,
 418                              struct strbuf *scratch,
 419                              struct batch_options *opt,
 420                              struct expand_data *data)
 421 {
 422         struct object_context ctx;
 423         int flags = opt->follow_symlinks ? GET_OID_FOLLOW_SYMLINKS : 0;
 424         enum get_oid_result result;
 425
 426         result = get_oid_with_context(the_repository, obj_name,
 427                                       flags, &data->oid, &ctx);
 428         if (result != FOUND) {
 429                 switch (result) {
 430                 case MISSING_OBJECT:
 431                         printf("%s missing\n", obj_name);
 432                         break;
 433                 case SHORT_NAME_AMBIGUOUS:
 434                         printf("%s ambiguous\n", obj_name);
 435                         break;
 436                 case DANGLING_SYMLINK:
 437                         printf("dangling %"PRIuMAX"\n%s\n",
 438                                (uintmax_t)strlen(obj_name), obj_name);
 439                         break;
 440                 case SYMLINK_LOOP:
 441                         printf("loop %"PRIuMAX"\n%s\n",
 442                                (uintmax_t)strlen(obj_name), obj_name);
 443                         break;
 444                 case NOT_DIR:
 445                         printf("notdir %"PRIuMAX"\n%s\n",
 446                                (uintmax_t)strlen(obj_name), obj_name);
 447                         break;
 448                 default:
 449                         BUG("unknown get_sha1_with_context result %d\n",
 450                                result);
 451                         break;
 452                 }
 453                 fflush(stdout);
 454                 return;
 455         }
 456
 457         if (ctx.mode == 0) {
 458                 printf("symlink %"PRIuMAX"\n%s\n",
 459                        (uintmax_t)ctx.symlink_path.len,
 460                        ctx.symlink_path.buf);
 461                 fflush(stdout);
 462                 return;
 463         }
 464
 465         batch_object_write(obj_name, scratch, opt, data, NULL, 0);
 466 }
 467
 468 struct object_cb_data {
 469         struct batch_options *opt;
 470         struct expand_data *expand;
 471         struct oidset *seen;
 472         struct strbuf *scratch;
 473 };
 474
 475 static int batch_object_cb(const struct object_id *oid, void *vdata)
 476 {
 477         struct object_cb_data *data = vdata;
 478         oidcpy(&data->expand->oid, oid);
 479         batch_object_write(NULL, data->scratch, data->opt, data->expand,
 480                            NULL, 0);
 481         return 0;
 482 }
 483
 484 static int collect_loose_object(const struct object_id *oid,
 485                                 const char *path,
 486                                 void *data)
 487 {
 488         oid_array_append(data, oid);
 489         return 0;
 490 }
 491
 492 static int collect_packed_object(const struct object_id *oid,
 493                                  struct packed_git *pack,
 494                                  uint32_t pos,
 495                                  void *data)
 496 {
 497         oid_array_append(data, oid);
 498         return 0;
 499 }
 500
 501 static int batch_unordered_object(const struct object_id *oid,
 502                                   struct packed_git *pack, off_t offset,
 503                                   void *vdata)
 504 {
 505         struct object_cb_data *data = vdata;
 506
 507         if (oidset_insert(data->seen, oid))
 508                 return 0;
 509
 510         oidcpy(&data->expand->oid, oid);
 511         batch_object_write(NULL, data->scratch, data->opt, data->expand,
 512                            pack, offset);
 513         return 0;
 514 }
 515
 516 static int batch_unordered_loose(const struct object_id *oid,
 517                                  const char *path,
 518                                  void *data)
 519 {
 520         return batch_unordered_object(oid, NULL, 0, data);
 521 }
 522
 523 static int batch_unordered_packed(const struct object_id *oid,
 524                                   struct packed_git *pack,
 525                                   uint32_t pos,
 526                                   void *data)
 527 {
 528         return batch_unordered_object(oid, pack,
 529                                       nth_packed_object_offset(pack, pos),
 530                                       data);
 531 }
 532
 533 typedef void (*parse_cmd_fn_t)(struct batch_options *, const char *,
 534                                struct strbuf *, struct expand_data *);
 535
 536 struct queued_cmd {
 537         parse_cmd_fn_t fn;
 538         char *line;
 539 };
 540
 541 static void parse_cmd_contents(struct batch_options *opt,
 542                              const char *line,
 543                              struct strbuf *output,
 544                              struct expand_data *data)
 545 {
 546         opt->batch_mode = BATCH_MODE_CONTENTS;
 547         batch_one_object(line, output, opt, data);
 548 }
 549
 550 static void parse_cmd_info(struct batch_options *opt,
 551                            const char *line,
 552                            struct strbuf *output,
 553                            struct expand_data *data)
 554 {
 555         opt->batch_mode = BATCH_MODE_INFO;
 556         batch_one_object(line, output, opt, data);
 557 }
 558
 559 static void dispatch_calls(struct batch_options *opt,
 560                 struct strbuf *output,
 561                 struct expand_data *data,
 562                 struct queued_cmd *cmd,
 563                 int nr)
 564 {
 565         int i;
 566
 567         if (!opt->buffer_output)
 568                 die(_("flush is only for --buffer mode"));
 569
 570         for (i = 0; i < nr; i++)
 571                 cmd[i].fn(opt, cmd[i].line, output, data);
 572
 573         fflush(stdout);
 574 }
 575
 576 static void free_cmds(struct queued_cmd *cmd, size_t *nr)
 577 {
 578         size_t i;
 579
 580         for (i = 0; i < *nr; i++)
 581                 FREE_AND_NULL(cmd[i].line);
 582
 583         *nr = 0;
 584 }
 585
 586
 587 static const struct parse_cmd {
 588         const char *name;
 589         parse_cmd_fn_t fn;
 590         unsigned takes_args;
 591 } commands[] = {
 592         { "contents", parse_cmd_contents, 1},
 593         { "info", parse_cmd_info, 1},
 594         { "flush", NULL, 0},
 595 };
 596
 597 static void batch_objects_command(struct batch_options *opt,
 598                                     struct strbuf *output,
 599                                     struct expand_data *data)
 600 {
 601         struct strbuf input = STRBUF_INIT;
 602         struct queued_cmd *queued_cmd = NULL;
 603         size_t alloc = 0, nr = 0;
 604
 605         while (!strbuf_getline(&input, stdin)) {
 606                 int i;
 607                 const struct parse_cmd *cmd = NULL;
 608                 const char *p = NULL, *cmd_end;
 609                 struct queued_cmd call = {0};
 610
 611                 if (!input.len)
 612                         die(_("empty command in input"));
 613                 if (isspace(*input.buf))
 614                         die(_("whitespace before command: '%s'"), input.buf);
 615
 616                 for (i = 0; i < ARRAY_SIZE(commands); i++) {
 617                         if (!skip_prefix(input.buf, commands[i].name, &cmd_end))
 618                                 continue;
 619
 620                         cmd = &commands[i];
 621                         if (cmd->takes_args) {
 622                                 if (*cmd_end != ' ')
 623                                         die(_("%s requires arguments"),
 624                                             commands[i].name);
 625
 626                                 p = cmd_end + 1;
 627                         } else if (*cmd_end) {
 628                                 die(_("%s takes no arguments"),
 629                                     commands[i].name);
 630                         }
 631
 632                         break;
 633                 }
 634
 635                 if (!cmd)
 636                         die(_("unknown command: '%s'"), input.buf);
 637
 638                 if (!strcmp(cmd->name, "flush")) {
 639                         dispatch_calls(opt, output, data, queued_cmd, nr);
 640                         free_cmds(queued_cmd, &nr);
 641                 } else if (!opt->buffer_output) {
 642                         cmd->fn(opt, p, output, data);
 643                 } else {
 644                         ALLOC_GROW(queued_cmd, nr + 1, alloc);
 645                         call.fn = cmd->fn;
 646                         call.line = xstrdup_or_null(p);
 647                         queued_cmd[nr++] = call;
 648                 }
 649         }
 650
 651         if (opt->buffer_output &&
 652             nr &&
 653             !git_env_bool("GIT_TEST_CAT_FILE_NO_FLUSH_ON_EXIT", 0)) {
 654                 dispatch_calls(opt, output, data, queued_cmd, nr);
 655                 free_cmds(queued_cmd, &nr);
 656         }
 657
 658         free(queued_cmd);
 659         strbuf_release(&input);
 660 }
 661
 662 #define DEFAULT_FORMAT "%(objectname) %(objecttype) %(objectsize)"
 663
 664 static int batch_objects(struct batch_options *opt)
 665 {
 666         struct strbuf input = STRBUF_INIT;
 667         struct strbuf output = STRBUF_INIT;
 668         struct expand_data data;
 669         int save_warning;
 670         int retval = 0;
 671
 672         /*
 673          * Expand once with our special mark_query flag, which will prime the
 674          * object_info to be handed to oid_object_info_extended for each
 675          * object.
 676          */
 677         memset(&data, 0, sizeof(data));
 678         data.mark_query = 1;
 679         strbuf_expand(&output,
 680                       opt->format ? opt->format : DEFAULT_FORMAT,
 681                       expand_format,
 682                       &data);
 683         data.mark_query = 0;
 684         strbuf_release(&output);
 685         if (opt->transform_mode)
 686                 data.split_on_whitespace = 1;
 687
 688         if (opt->format && !strcmp(opt->format, DEFAULT_FORMAT))
 689                 opt->format = NULL;
 690         /*
 691          * If we are printing out the object, then always fill in the type,
 692          * since we will want to decide whether or not to stream.
 693          */
 694         if (opt->batch_mode == BATCH_MODE_CONTENTS)
 695                 data.info.typep = &data.type;
 696
 697         if (opt->all_objects) {
 698                 struct object_cb_data cb;
 699                 struct object_info empty = OBJECT_INFO_INIT;
 700
 701                 if (!memcmp(&data.info, &empty, sizeof(empty)))
 702                         data.skip_object_info = 1;
 703
 704                 if (has_promisor_remote())
 705                         warning("This repository uses promisor remotes. Some objects may not be loaded.");
 706
 707                 read_replace_refs = 0;
 708
 709                 cb.opt = opt;
 710                 cb.expand = &data;
 711                 cb.scratch = &output;
 712
 713                 if (opt->unordered) {
 714                         struct oidset seen = OIDSET_INIT;
 715
 716                         cb.seen = &seen;
 717
 718                         for_each_loose_object(batch_unordered_loose, &cb, 0);
 719                         for_each_packed_object(batch_unordered_packed, &cb,
 720                                                FOR_EACH_OBJECT_PACK_ORDER);
 721
 722                         oidset_clear(&seen);
 723                 } else {
 724                         struct oid_array sa = OID_ARRAY_INIT;
 725
 726                         for_each_loose_object(collect_loose_object, &sa, 0);
 727                         for_each_packed_object(collect_packed_object, &sa, 0);
 728
 729                         oid_array_for_each_unique(&sa, batch_object_cb, &cb);
 730
 731                         oid_array_clear(&sa);
 732                 }
 733
 734                 strbuf_release(&output);
 735                 return 0;
 736         }
 737
 738         /*
 739          * We are going to call get_sha1 on a potentially very large number of
 740          * objects. In most large cases, these will be actual object sha1s. The
 741          * cost to double-check that each one is not also a ref (just so we can
 742          * warn) ends up dwarfing the actual cost of the object lookups
 743          * themselves. We can work around it by just turning off the warning.
 744          */
 745         save_warning = warn_on_object_refname_ambiguity;
 746         warn_on_object_refname_ambiguity = 0;
 747
 748         if (opt->batch_mode == BATCH_MODE_QUEUE_AND_DISPATCH) {
 749                 batch_objects_command(opt, &output, &data);
 750                 goto cleanup;
 751         }
 752
 753         while (strbuf_getline(&input, stdin) != EOF) {
 754                 if (data.split_on_whitespace) {
 755                         /*
 756                          * Split at first whitespace, tying off the beginning
 757                          * of the string and saving the remainder (or NULL) in
 758                          * data.rest.
 759                          */
 760                         char *p = strpbrk(input.buf, " \t");
 761                         if (p) {
 762                                 while (*p && strchr(" \t", *p))
 763                                         *p++ = '\0';
 764                         }
 765                         data.rest = p;
 766                 }
 767
 768                 batch_one_object(input.buf, &output, opt, &data);
 769         }
 770
 771  cleanup:
 772         strbuf_release(&input);
 773         strbuf_release(&output);
 774         warn_on_object_refname_ambiguity = save_warning;
 775         return retval;
 776 }
 777
 778 static int git_cat_file_config(const char *var, const char *value, void *cb)
 779 {
 780         if (userdiff_config(var, value) < 0)
 781                 return -1;
 782
 783         return git_default_config(var, value, cb);
 784 }
 785
 786 static int batch_option_callback(const struct option *opt,
 787                                  const char *arg,
 788                                  int unset)
 789 {
 790         struct batch_options *bo = opt->value;
 791
 792         BUG_ON_OPT_NEG(unset);
 793
 794         if (bo->enabled) {
 795                 return error(_("only one batch option may be specified"));
 796         }
 797
 798         bo->enabled = 1;
 799
 800         if (!strcmp(opt->long_name, "batch"))
 801                 bo->batch_mode = BATCH_MODE_CONTENTS;
 802         else if (!strcmp(opt->long_name, "batch-check"))
 803                 bo->batch_mode = BATCH_MODE_INFO;
 804         else if (!strcmp(opt->long_name, "batch-command"))
 805                 bo->batch_mode = BATCH_MODE_QUEUE_AND_DISPATCH;
 806         else
 807                 BUG("%s given to batch-option-callback", opt->long_name);
 808
 809         bo->format = arg;
 810
 811         return 0;
 812 }
 813
 814 int cmd_cat_file(int argc, const char **argv, const char *prefix)
 815 {
 816         int opt = 0;
 817         int opt_cw = 0;
 818         int opt_epts = 0;
 819         const char *exp_type = NULL, *obj_name = NULL;
 820         struct batch_options batch = {0};
 821         int unknown_type = 0;
 822
 823         const char * const usage[] = {
 824                 N_("git cat-file <type> <object>"),
 825                 N_("git cat-file (-e | -p) <object>"),
 826                 N_("git cat-file (-t | -s) [--allow-unknown-type] <object>"),
 827                 N_("git cat-file (--batch | --batch-check | --batch-command) [--batch-all-objects]\n"
 828                    "             [--buffer] [--follow-symlinks] [--unordered]\n"
 829                    "             [--textconv | --filters]"),
 830                 N_("git cat-file (--textconv | --filters)\n"
 831                    "             [<rev>:<path|tree-ish> | --path=<path|tree-ish> <rev>]"),
 832                 NULL
 833         };
 834         const struct option options[] = {
 835                 /* Simple queries */
 836                 OPT_GROUP(N_("Check object existence or emit object contents")),
 837                 OPT_CMDMODE('e', NULL, &opt,
 838                             N_("check if <object> exists"), 'e'),
 839                 OPT_CMDMODE('p', NULL, &opt, N_("pretty-print <object> content"), 'p'),
 840
 841                 OPT_GROUP(N_("Emit [broken] object attributes")),
 842                 OPT_CMDMODE('t', NULL, &opt, N_("show object type (one of 'blob', 'tree', 'commit', 'tag', ...)"), 't'),
 843                 OPT_CMDMODE('s', NULL, &opt, N_("show object size"), 's'),
 844                 OPT_BOOL(0, "allow-unknown-type", &unknown_type,
 845                           N_("allow -s and -t to work with broken/corrupt objects")),
 846                 /* Batch mode */
 847                 OPT_GROUP(N_("Batch objects requested on stdin (or --batch-all-objects)")),
 848                 OPT_CALLBACK_F(0, "batch", &batch, N_("format"),
 849                         N_("show full <object> or <rev> contents"),
 850                         PARSE_OPT_OPTARG | PARSE_OPT_NONEG,
 851                         batch_option_callback),
 852                 OPT_CALLBACK_F(0, "batch-check", &batch, N_("format"),
 853                         N_("like --batch, but don't emit <contents>"),
 854                         PARSE_OPT_OPTARG | PARSE_OPT_NONEG,
 855                         batch_option_callback),
 856                 OPT_CALLBACK_F(0, "batch-command", &batch, N_("format"),
 857                         N_("read commands from stdin"),
 858                         PARSE_OPT_OPTARG | PARSE_OPT_NONEG,
 859                         batch_option_callback),
 860                 OPT_CMDMODE(0, "batch-all-objects", &opt,
 861                             N_("with --batch[-check]: ignores stdin, batches all known objects"), 'b'),
 862                 /* Batch-specific options */
 863                 OPT_GROUP(N_("Change or optimize batch output")),
 864                 OPT_BOOL(0, "buffer", &batch.buffer_output, N_("buffer --batch output")),
 865                 OPT_BOOL(0, "follow-symlinks", &batch.follow_symlinks,
 866                          N_("follow in-tree symlinks")),
 867                 OPT_BOOL(0, "unordered", &batch.unordered,
 868                          N_("do not order objects before emitting them")),
 869                 /* Textconv options, stand-ole*/
 870                 OPT_GROUP(N_("Emit object (blob or tree) with conversion or filter (stand-alone, or with batch)")),
 871                 OPT_CMDMODE(0, "textconv", &opt,
 872                             N_("run textconv on object's content"), 'c'),
 873                 OPT_CMDMODE(0, "filters", &opt,
 874                             N_("run filters on object's content"), 'w'),
 875                 OPT_STRING(0, "path", &force_path, N_("blob|tree"),
 876                            N_("use a <path> for (--textconv | --filters); Not with 'batch'")),
 877                 OPT_END()
 878         };
 879
 880         git_config(git_cat_file_config, NULL);
 881
 882         batch.buffer_output = -1;
 883
 884         argc = parse_options(argc, argv, prefix, options, usage, 0);
 885         opt_cw = (opt == 'c' || opt == 'w');
 886         opt_epts = (opt == 'e' || opt == 'p' || opt == 't' || opt == 's');
 887
 888         /* --batch-all-objects? */
 889         if (opt == 'b')
 890                 batch.all_objects = 1;
 891
 892         /* Option compatibility */
 893         if (force_path && !opt_cw)
 894                 usage_msg_optf(_("'%s=<%s>' needs '%s' or '%s'"),
 895                                usage, options,
 896                                "--path", _("path|tree-ish"), "--filters",
 897                                "--textconv");
 898
 899         /* Option compatibility with batch mode */
 900         if (batch.enabled)
 901                 ;
 902         else if (batch.follow_symlinks)
 903                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
 904                                "--follow-symlinks");
 905         else if (batch.buffer_output >= 0)
 906                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
 907                                "--buffer");
 908         else if (batch.all_objects)
 909                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
 910                                "--batch-all-objects");
 911
 912         /* Batch defaults */
 913         if (batch.buffer_output < 0)
 914                 batch.buffer_output = batch.all_objects;
 915
 916         /* Return early if we're in batch mode? */
 917         if (batch.enabled) {
 918                 if (opt_cw)
 919                         batch.transform_mode = opt;
 920                 else if (opt && opt != 'b')
 921                         usage_msg_optf(_("'-%c' is incompatible with batch mode"),
 922                                        usage, options, opt);
 923                 else if (argc)
 924                         usage_msg_opt(_("batch modes take no arguments"), usage,
 925                                       options);
 926
 927                 return batch_objects(&batch);
 928         }
 929
 930         if (opt) {
 931                 if (!argc && opt == 'c')
 932                         usage_msg_optf(_("<rev> required with '%s'"),
 933                                        usage, options, "--textconv");
 934                 else if (!argc && opt == 'w')
 935                         usage_msg_optf(_("<rev> required with '%s'"),
 936                                        usage, options, "--filters");
 937                 else if (!argc && opt_epts)
 938                         usage_msg_optf(_("<object> required with '-%c'"),
 939                                        usage, options, opt);
 940                 else if (argc == 1)
 941                         obj_name = argv[0];
 942                 else
 943                         usage_msg_opt(_("too many arguments"), usage, options);
 944         } else if (!argc) {
 945                 usage_with_options(usage, options);
 946         } else if (argc != 2) {
 947                 usage_msg_optf(_("only two arguments allowed in <type> <object> mode, not %d"),
 948                               usage, options, argc);
 949         } else if (argc) {
 950                 exp_type = argv[0];
 951                 obj_name = argv[1];
 952         }
 953
 954         if (unknown_type && opt != 't' && opt != 's')
 955                 die("git cat-file --allow-unknown-type: use with -s or -t");
 956         return cat_one_file(opt, exp_type, obj_name, unknown_type);
 957 }