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_VARIABLE
   7 #include "cache.h"
   8 #include "alloc.h"
   9 #include "config.h"
  10 #include "convert.h"
  11 #include "builtin.h"
  12 #include "diff.h"
  13 #include "environment.h"
  14 #include "gettext.h"
  15 #include "hex.h"
  16 #include "ident.h"
  17 #include "parse-options.h"
  18 #include "userdiff.h"
  19 #include "streaming.h"
  20 #include "tree-walk.h"
  21 #include "oid-array.h"
  22 #include "packfile.h"
  23 #include "object-file.h"
  24 #include "object-name.h"
  25 #include "object-store.h"
  26 #include "replace-object.h"
  27 #include "promisor-remote.h"
  28 #include "mailmap.h"
  29 #include "write-or-die.h"
  30
  31 enum batch_mode {
  32         BATCH_MODE_CONTENTS,
  33         BATCH_MODE_INFO,
  34         BATCH_MODE_QUEUE_AND_DISPATCH,
  35 };
  36
  37 struct batch_options {
  38         int enabled;
  39         int follow_symlinks;
  40         enum batch_mode batch_mode;
  41         int buffer_output;
  42         int all_objects;
  43         int unordered;
  44         int transform_mode; /* may be 'w' or 'c' for --filters or --textconv */
  45         int nul_terminated;
  46         const char *format;
  47 };
  48
  49 static const char *force_path;
  50
  51 static struct string_list mailmap = STRING_LIST_INIT_NODUP;
  52 static int use_mailmap;
  53
  54 static char *replace_idents_using_mailmap(char *, size_t *);
  55
  56 static char *replace_idents_using_mailmap(char *object_buf, size_t *size)
  57 {
  58         struct strbuf sb = STRBUF_INIT;
  59         const char *headers[] = { "author ", "committer ", "tagger ", NULL };
  60
  61         strbuf_attach(&sb, object_buf, *size, *size + 1);
  62         apply_mailmap_to_header(&sb, headers, &mailmap);
  63         *size = sb.len;
  64         return strbuf_detach(&sb, NULL);
  65 }
  66
  67 static int filter_object(const char *path, unsigned mode,
  68                          const struct object_id *oid,
  69                          char **buf, unsigned long *size)
  70 {
  71         enum object_type type;
  72
  73         *buf = repo_read_object_file(the_repository, oid, &type, size);
  74         if (!*buf)
  75                 return error(_("cannot read object %s '%s'"),
  76                              oid_to_hex(oid), path);
  77         if ((type == OBJ_BLOB) && S_ISREG(mode)) {
  78                 struct strbuf strbuf = STRBUF_INIT;
  79                 struct checkout_metadata meta;
  80
  81                 init_checkout_metadata(&meta, NULL, NULL, oid);
  82                 if (convert_to_working_tree(&the_index, path, *buf, *size, &strbuf, &meta)) {
  83                         free(*buf);
  84                         *size = strbuf.len;
  85                         *buf = strbuf_detach(&strbuf, NULL);
  86                 }
  87         }
  88
  89         return 0;
  90 }
  91
  92 static int stream_blob(const struct object_id *oid)
  93 {
  94         if (stream_blob_to_fd(1, oid, NULL, 0))
  95                 die("unable to stream %s to stdout", oid_to_hex(oid));
  96         return 0;
  97 }
  98
  99 static int cat_one_file(int opt, const char *exp_type, const char *obj_name,
 100                         int unknown_type)
 101 {
 102         int ret;
 103         struct object_id oid;
 104         enum object_type type;
 105         char *buf;
 106         unsigned long size;
 107         struct object_context obj_context;
 108         struct object_info oi = OBJECT_INFO_INIT;
 109         struct strbuf sb = STRBUF_INIT;
 110         unsigned flags = OBJECT_INFO_LOOKUP_REPLACE;
 111         unsigned get_oid_flags = GET_OID_RECORD_PATH | GET_OID_ONLY_TO_DIE;
 112         const char *path = force_path;
 113         const int opt_cw = (opt == 'c' || opt == 'w');
 114         if (!path && opt_cw)
 115                 get_oid_flags |= GET_OID_REQUIRE_PATH;
 116
 117         if (unknown_type)
 118                 flags |= OBJECT_INFO_ALLOW_UNKNOWN_TYPE;
 119
 120         if (get_oid_with_context(the_repository, obj_name, get_oid_flags, &oid,
 121                                  &obj_context))
 122                 die("Not a valid object name %s", obj_name);
 123
 124         if (!path)
 125                 path = obj_context.path;
 126         if (obj_context.mode == S_IFINVALID)
 127                 obj_context.mode = 0100644;
 128
 129         buf = NULL;
 130         switch (opt) {
 131         case 't':
 132                 oi.type_name = &sb;
 133                 if (oid_object_info_extended(the_repository, &oid, &oi, flags) < 0)
 134                         die("git cat-file: could not get object info");
 135                 if (sb.len) {
 136                         printf("%s\n", sb.buf);
 137                         strbuf_release(&sb);
 138                         ret = 0;
 139                         goto cleanup;
 140                 }
 141                 break;
 142
 143         case 's':
 144                 oi.sizep = &size;
 145
 146                 if (use_mailmap) {
 147                         oi.typep = &type;
 148                         oi.contentp = (void**)&buf;
 149                 }
 150
 151                 if (oid_object_info_extended(the_repository, &oid, &oi, flags) < 0)
 152                         die("git cat-file: could not get object info");
 153
 154                 if (use_mailmap && (type == OBJ_COMMIT || type == OBJ_TAG)) {
 155                         size_t s = size;
 156                         buf = replace_idents_using_mailmap(buf, &s);
 157                         size = cast_size_t_to_ulong(s);
 158                 }
 159
 160                 printf("%"PRIuMAX"\n", (uintmax_t)size);
 161                 ret = 0;
 162                 goto cleanup;
 163
 164         case 'e':
 165                 return !repo_has_object_file(the_repository, &oid);
 166
 167         case 'w':
 168
 169                 if (filter_object(path, obj_context.mode,
 170                                   &oid, &buf, &size)) {
 171                         ret = -1;
 172                         goto cleanup;
 173                 }
 174                 break;
 175
 176         case 'c':
 177                 if (textconv_object(the_repository, path, obj_context.mode,
 178                                     &oid, 1, &buf, &size))
 179                         break;
 180                 /* else fallthrough */
 181
 182         case 'p':
 183                 type = oid_object_info(the_repository, &oid, NULL);
 184                 if (type < 0)
 185                         die("Not a valid object name %s", obj_name);
 186
 187                 /* custom pretty-print here */
 188                 if (type == OBJ_TREE) {
 189                         const char *ls_args[3] = { NULL };
 190                         ls_args[0] =  "ls-tree";
 191                         ls_args[1] =  obj_name;
 192                         ret = cmd_ls_tree(2, ls_args, NULL);
 193                         goto cleanup;
 194                 }
 195
 196                 if (type == OBJ_BLOB) {
 197                         ret = stream_blob(&oid);
 198                         goto cleanup;
 199                 }
 200                 buf = repo_read_object_file(the_repository, &oid, &type,
 201                                             &size);
 202                 if (!buf)
 203                         die("Cannot read object %s", obj_name);
 204
 205                 if (use_mailmap) {
 206                         size_t s = size;
 207                         buf = replace_idents_using_mailmap(buf, &s);
 208                         size = cast_size_t_to_ulong(s);
 209                 }
 210
 211                 /* otherwise just spit out the data */
 212                 break;
 213
 214         case 0:
 215         {
 216                 enum object_type exp_type_id = type_from_string(exp_type);
 217
 218                 if (exp_type_id == OBJ_BLOB) {
 219                         struct object_id blob_oid;
 220                         if (oid_object_info(the_repository, &oid, NULL) == OBJ_TAG) {
 221                                 char *buffer = repo_read_object_file(the_repository,
 222                                                                      &oid,
 223                                                                      &type,
 224                                                                      &size);
 225                                 const char *target;
 226                                 if (!skip_prefix(buffer, "object ", &target) ||
 227                                     get_oid_hex(target, &blob_oid))
 228                                         die("%s not a valid tag", oid_to_hex(&oid));
 229                                 free(buffer);
 230                         } else
 231                                 oidcpy(&blob_oid, &oid);
 232
 233                         if (oid_object_info(the_repository, &blob_oid, NULL) == OBJ_BLOB) {
 234                                 ret = stream_blob(&blob_oid);
 235                                 goto cleanup;
 236                         }
 237                         /*
 238                          * we attempted to dereference a tag to a blob
 239                          * and failed; there may be new dereference
 240                          * mechanisms this code is not aware of.
 241                          * fall-back to the usual case.
 242                          */
 243                 }
 244                 buf = read_object_with_reference(the_repository, &oid,
 245                                                  exp_type_id, &size, NULL);
 246
 247                 if (use_mailmap) {
 248                         size_t s = size;
 249                         buf = replace_idents_using_mailmap(buf, &s);
 250                         size = cast_size_t_to_ulong(s);
 251                 }
 252                 break;
 253         }
 254         default:
 255                 die("git cat-file: unknown option: %s", exp_type);
 256         }
 257
 258         if (!buf)
 259                 die("git cat-file %s: bad file", obj_name);
 260
 261         write_or_die(1, buf, size);
 262         ret = 0;
 263 cleanup:
 264         free(buf);
 265         free(obj_context.path);
 266         return ret;
 267 }
 268
 269 struct expand_data {
 270         struct object_id oid;
 271         enum object_type type;
 272         unsigned long size;
 273         off_t disk_size;
 274         const char *rest;
 275         struct object_id delta_base_oid;
 276
 277         /*
 278          * If mark_query is true, we do not expand anything, but rather
 279          * just mark the object_info with items we wish to query.
 280          */
 281         int mark_query;
 282
 283         /*
 284          * Whether to split the input on whitespace before feeding it to
 285          * get_sha1; this is decided during the mark_query phase based on
 286          * whether we have a %(rest) token in our format.
 287          */
 288         int split_on_whitespace;
 289
 290         /*
 291          * After a mark_query run, this object_info is set up to be
 292          * passed to oid_object_info_extended. It will point to the data
 293          * elements above, so you can retrieve the response from there.
 294          */
 295         struct object_info info;
 296
 297         /*
 298          * This flag will be true if the requested batch format and options
 299          * don't require us to call oid_object_info, which can then be
 300          * optimized out.
 301          */
 302         unsigned skip_object_info : 1;
 303 };
 304
 305 static int is_atom(const char *atom, const char *s, int slen)
 306 {
 307         int alen = strlen(atom);
 308         return alen == slen && !memcmp(atom, s, alen);
 309 }
 310
 311 static void expand_atom(struct strbuf *sb, const char *atom, int len,
 312                         void *vdata)
 313 {
 314         struct expand_data *data = vdata;
 315
 316         if (is_atom("objectname", atom, len)) {
 317                 if (!data->mark_query)
 318                         strbuf_addstr(sb, oid_to_hex(&data->oid));
 319         } else if (is_atom("objecttype", atom, len)) {
 320                 if (data->mark_query)
 321                         data->info.typep = &data->type;
 322                 else
 323                         strbuf_addstr(sb, type_name(data->type));
 324         } else if (is_atom("objectsize", atom, len)) {
 325                 if (data->mark_query)
 326                         data->info.sizep = &data->size;
 327                 else
 328                         strbuf_addf(sb, "%"PRIuMAX , (uintmax_t)data->size);
 329         } else if (is_atom("objectsize:disk", atom, len)) {
 330                 if (data->mark_query)
 331                         data->info.disk_sizep = &data->disk_size;
 332                 else
 333                         strbuf_addf(sb, "%"PRIuMAX, (uintmax_t)data->disk_size);
 334         } else if (is_atom("rest", atom, len)) {
 335                 if (data->mark_query)
 336                         data->split_on_whitespace = 1;
 337                 else if (data->rest)
 338                         strbuf_addstr(sb, data->rest);
 339         } else if (is_atom("deltabase", atom, len)) {
 340                 if (data->mark_query)
 341                         data->info.delta_base_oid = &data->delta_base_oid;
 342                 else
 343                         strbuf_addstr(sb,
 344                                       oid_to_hex(&data->delta_base_oid));
 345         } else
 346                 die("unknown format element: %.*s", len, atom);
 347 }
 348
 349 static size_t expand_format(struct strbuf *sb, const char *start, void *data)
 350 {
 351         const char *end;
 352
 353         if (*start != '(')
 354                 return 0;
 355         end = strchr(start + 1, ')');
 356         if (!end)
 357                 die("format element '%s' does not end in ')'", start);
 358
 359         expand_atom(sb, start + 1, end - start - 1, data);
 360
 361         return end - start + 1;
 362 }
 363
 364 static void batch_write(struct batch_options *opt, const void *data, int len)
 365 {
 366         if (opt->buffer_output) {
 367                 if (fwrite(data, 1, len, stdout) != len)
 368                         die_errno("unable to write to stdout");
 369         } else
 370                 write_or_die(1, data, len);
 371 }
 372
 373 static void print_object_or_die(struct batch_options *opt, struct expand_data *data)
 374 {
 375         const struct object_id *oid = &data->oid;
 376
 377         assert(data->info.typep);
 378
 379         if (data->type == OBJ_BLOB) {
 380                 if (opt->buffer_output)
 381                         fflush(stdout);
 382                 if (opt->transform_mode) {
 383                         char *contents;
 384                         unsigned long size;
 385
 386                         if (!data->rest)
 387                                 die("missing path for '%s'", oid_to_hex(oid));
 388
 389                         if (opt->transform_mode == 'w') {
 390                                 if (filter_object(data->rest, 0100644, oid,
 391                                                   &contents, &size))
 392                                         die("could not convert '%s' %s",
 393                                             oid_to_hex(oid), data->rest);
 394                         } else if (opt->transform_mode == 'c') {
 395                                 enum object_type type;
 396                                 if (!textconv_object(the_repository,
 397                                                      data->rest, 0100644, oid,
 398                                                      1, &contents, &size))
 399                                         contents = repo_read_object_file(the_repository,
 400                                                                          oid,
 401                                                                          &type,
 402                                                                          &size);
 403                                 if (!contents)
 404                                         die("could not convert '%s' %s",
 405                                             oid_to_hex(oid), data->rest);
 406                         } else
 407                                 BUG("invalid transform_mode: %c", opt->transform_mode);
 408                         batch_write(opt, contents, size);
 409                         free(contents);
 410                 } else {
 411                         stream_blob(oid);
 412                 }
 413         }
 414         else {
 415                 enum object_type type;
 416                 unsigned long size;
 417                 void *contents;
 418
 419                 contents = repo_read_object_file(the_repository, oid, &type,
 420                                                  &size);
 421
 422                 if (use_mailmap) {
 423                         size_t s = size;
 424                         contents = replace_idents_using_mailmap(contents, &s);
 425                         size = cast_size_t_to_ulong(s);
 426                 }
 427
 428                 if (!contents)
 429                         die("object %s disappeared", oid_to_hex(oid));
 430                 if (type != data->type)
 431                         die("object %s changed type!?", oid_to_hex(oid));
 432                 if (data->info.sizep && size != data->size && !use_mailmap)
 433                         die("object %s changed size!?", oid_to_hex(oid));
 434
 435                 batch_write(opt, contents, size);
 436                 free(contents);
 437         }
 438 }
 439
 440 static void print_default_format(struct strbuf *scratch, struct expand_data *data)
 441 {
 442         strbuf_addf(scratch, "%s %s %"PRIuMAX"\n", oid_to_hex(&data->oid),
 443                     type_name(data->type),
 444                     (uintmax_t)data->size);
 445 }
 446
 447 /*
 448  * If "pack" is non-NULL, then "offset" is the byte offset within the pack from
 449  * which the object may be accessed (though note that we may also rely on
 450  * data->oid, too). If "pack" is NULL, then offset is ignored.
 451  */
 452 static void batch_object_write(const char *obj_name,
 453                                struct strbuf *scratch,
 454                                struct batch_options *opt,
 455                                struct expand_data *data,
 456                                struct packed_git *pack,
 457                                off_t offset)
 458 {
 459         if (!data->skip_object_info) {
 460                 int ret;
 461
 462                 if (use_mailmap)
 463                         data->info.typep = &data->type;
 464
 465                 if (pack)
 466                         ret = packed_object_info(the_repository, pack, offset,
 467                                                  &data->info);
 468                 else
 469                         ret = oid_object_info_extended(the_repository,
 470                                                        &data->oid, &data->info,
 471                                                        OBJECT_INFO_LOOKUP_REPLACE);
 472                 if (ret < 0) {
 473                         printf("%s missing\n",
 474                                obj_name ? obj_name : oid_to_hex(&data->oid));
 475                         fflush(stdout);
 476                         return;
 477                 }
 478
 479                 if (use_mailmap && (data->type == OBJ_COMMIT || data->type == OBJ_TAG)) {
 480                         size_t s = data->size;
 481                         char *buf = NULL;
 482
 483                         buf = repo_read_object_file(the_repository, &data->oid, &data->type,
 484                                                     &data->size);
 485                         buf = replace_idents_using_mailmap(buf, &s);
 486                         data->size = cast_size_t_to_ulong(s);
 487
 488                         free(buf);
 489                 }
 490         }
 491
 492         strbuf_reset(scratch);
 493
 494         if (!opt->format) {
 495                 print_default_format(scratch, data);
 496         } else {
 497                 strbuf_expand(scratch, opt->format, expand_format, data);
 498                 strbuf_addch(scratch, '\n');
 499         }
 500
 501         batch_write(opt, scratch->buf, scratch->len);
 502
 503         if (opt->batch_mode == BATCH_MODE_CONTENTS) {
 504                 print_object_or_die(opt, data);
 505                 batch_write(opt, "\n", 1);
 506         }
 507 }
 508
 509 static void batch_one_object(const char *obj_name,
 510                              struct strbuf *scratch,
 511                              struct batch_options *opt,
 512                              struct expand_data *data)
 513 {
 514         struct object_context ctx;
 515         int flags = opt->follow_symlinks ? GET_OID_FOLLOW_SYMLINKS : 0;
 516         enum get_oid_result result;
 517
 518         result = get_oid_with_context(the_repository, obj_name,
 519                                       flags, &data->oid, &ctx);
 520         if (result != FOUND) {
 521                 switch (result) {
 522                 case MISSING_OBJECT:
 523                         printf("%s missing\n", obj_name);
 524                         break;
 525                 case SHORT_NAME_AMBIGUOUS:
 526                         printf("%s ambiguous\n", obj_name);
 527                         break;
 528                 case DANGLING_SYMLINK:
 529                         printf("dangling %"PRIuMAX"\n%s\n",
 530                                (uintmax_t)strlen(obj_name), obj_name);
 531                         break;
 532                 case SYMLINK_LOOP:
 533                         printf("loop %"PRIuMAX"\n%s\n",
 534                                (uintmax_t)strlen(obj_name), obj_name);
 535                         break;
 536                 case NOT_DIR:
 537                         printf("notdir %"PRIuMAX"\n%s\n",
 538                                (uintmax_t)strlen(obj_name), obj_name);
 539                         break;
 540                 default:
 541                         BUG("unknown get_sha1_with_context result %d\n",
 542                                result);
 543                         break;
 544                 }
 545                 fflush(stdout);
 546                 return;
 547         }
 548
 549         if (ctx.mode == 0) {
 550                 printf("symlink %"PRIuMAX"\n%s\n",
 551                        (uintmax_t)ctx.symlink_path.len,
 552                        ctx.symlink_path.buf);
 553                 fflush(stdout);
 554                 return;
 555         }
 556
 557         batch_object_write(obj_name, scratch, opt, data, NULL, 0);
 558 }
 559
 560 struct object_cb_data {
 561         struct batch_options *opt;
 562         struct expand_data *expand;
 563         struct oidset *seen;
 564         struct strbuf *scratch;
 565 };
 566
 567 static int batch_object_cb(const struct object_id *oid, void *vdata)
 568 {
 569         struct object_cb_data *data = vdata;
 570         oidcpy(&data->expand->oid, oid);
 571         batch_object_write(NULL, data->scratch, data->opt, data->expand,
 572                            NULL, 0);
 573         return 0;
 574 }
 575
 576 static int collect_loose_object(const struct object_id *oid,
 577                                 const char *path UNUSED,
 578                                 void *data)
 579 {
 580         oid_array_append(data, oid);
 581         return 0;
 582 }
 583
 584 static int collect_packed_object(const struct object_id *oid,
 585                                  struct packed_git *pack UNUSED,
 586                                  uint32_t pos UNUSED,
 587                                  void *data)
 588 {
 589         oid_array_append(data, oid);
 590         return 0;
 591 }
 592
 593 static int batch_unordered_object(const struct object_id *oid,
 594                                   struct packed_git *pack, off_t offset,
 595                                   void *vdata)
 596 {
 597         struct object_cb_data *data = vdata;
 598
 599         if (oidset_insert(data->seen, oid))
 600                 return 0;
 601
 602         oidcpy(&data->expand->oid, oid);
 603         batch_object_write(NULL, data->scratch, data->opt, data->expand,
 604                            pack, offset);
 605         return 0;
 606 }
 607
 608 static int batch_unordered_loose(const struct object_id *oid,
 609                                  const char *path UNUSED,
 610                                  void *data)
 611 {
 612         return batch_unordered_object(oid, NULL, 0, data);
 613 }
 614
 615 static int batch_unordered_packed(const struct object_id *oid,
 616                                   struct packed_git *pack,
 617                                   uint32_t pos,
 618                                   void *data)
 619 {
 620         return batch_unordered_object(oid, pack,
 621                                       nth_packed_object_offset(pack, pos),
 622                                       data);
 623 }
 624
 625 typedef void (*parse_cmd_fn_t)(struct batch_options *, const char *,
 626                                struct strbuf *, struct expand_data *);
 627
 628 struct queued_cmd {
 629         parse_cmd_fn_t fn;
 630         char *line;
 631 };
 632
 633 static void parse_cmd_contents(struct batch_options *opt,
 634                              const char *line,
 635                              struct strbuf *output,
 636                              struct expand_data *data)
 637 {
 638         opt->batch_mode = BATCH_MODE_CONTENTS;
 639         batch_one_object(line, output, opt, data);
 640 }
 641
 642 static void parse_cmd_info(struct batch_options *opt,
 643                            const char *line,
 644                            struct strbuf *output,
 645                            struct expand_data *data)
 646 {
 647         opt->batch_mode = BATCH_MODE_INFO;
 648         batch_one_object(line, output, opt, data);
 649 }
 650
 651 static void dispatch_calls(struct batch_options *opt,
 652                 struct strbuf *output,
 653                 struct expand_data *data,
 654                 struct queued_cmd *cmd,
 655                 int nr)
 656 {
 657         int i;
 658
 659         if (!opt->buffer_output)
 660                 die(_("flush is only for --buffer mode"));
 661
 662         for (i = 0; i < nr; i++)
 663                 cmd[i].fn(opt, cmd[i].line, output, data);
 664
 665         fflush(stdout);
 666 }
 667
 668 static void free_cmds(struct queued_cmd *cmd, size_t *nr)
 669 {
 670         size_t i;
 671
 672         for (i = 0; i < *nr; i++)
 673                 FREE_AND_NULL(cmd[i].line);
 674
 675         *nr = 0;
 676 }
 677
 678
 679 static const struct parse_cmd {
 680         const char *name;
 681         parse_cmd_fn_t fn;
 682         unsigned takes_args;
 683 } commands[] = {
 684         { "contents", parse_cmd_contents, 1},
 685         { "info", parse_cmd_info, 1},
 686         { "flush", NULL, 0},
 687 };
 688
 689 static void batch_objects_command(struct batch_options *opt,
 690                                     struct strbuf *output,
 691                                     struct expand_data *data)
 692 {
 693         struct strbuf input = STRBUF_INIT;
 694         struct queued_cmd *queued_cmd = NULL;
 695         size_t alloc = 0, nr = 0;
 696
 697         while (1) {
 698                 int i, ret;
 699                 const struct parse_cmd *cmd = NULL;
 700                 const char *p = NULL, *cmd_end;
 701                 struct queued_cmd call = {0};
 702
 703                 if (opt->nul_terminated)
 704                         ret = strbuf_getline_nul(&input, stdin);
 705                 else
 706                         ret = strbuf_getline(&input, stdin);
 707
 708                 if (ret)
 709                         break;
 710
 711                 if (!input.len)
 712                         die(_("empty command in input"));
 713                 if (isspace(*input.buf))
 714                         die(_("whitespace before command: '%s'"), input.buf);
 715
 716                 for (i = 0; i < ARRAY_SIZE(commands); i++) {
 717                         if (!skip_prefix(input.buf, commands[i].name, &cmd_end))
 718                                 continue;
 719
 720                         cmd = &commands[i];
 721                         if (cmd->takes_args) {
 722                                 if (*cmd_end != ' ')
 723                                         die(_("%s requires arguments"),
 724                                             commands[i].name);
 725
 726                                 p = cmd_end + 1;
 727                         } else if (*cmd_end) {
 728                                 die(_("%s takes no arguments"),
 729                                     commands[i].name);
 730                         }
 731
 732                         break;
 733                 }
 734
 735                 if (!cmd)
 736                         die(_("unknown command: '%s'"), input.buf);
 737
 738                 if (!strcmp(cmd->name, "flush")) {
 739                         dispatch_calls(opt, output, data, queued_cmd, nr);
 740                         free_cmds(queued_cmd, &nr);
 741                 } else if (!opt->buffer_output) {
 742                         cmd->fn(opt, p, output, data);
 743                 } else {
 744                         ALLOC_GROW(queued_cmd, nr + 1, alloc);
 745                         call.fn = cmd->fn;
 746                         call.line = xstrdup_or_null(p);
 747                         queued_cmd[nr++] = call;
 748                 }
 749         }
 750
 751         if (opt->buffer_output &&
 752             nr &&
 753             !git_env_bool("GIT_TEST_CAT_FILE_NO_FLUSH_ON_EXIT", 0)) {
 754                 dispatch_calls(opt, output, data, queued_cmd, nr);
 755                 free_cmds(queued_cmd, &nr);
 756         }
 757
 758         free_cmds(queued_cmd, &nr);
 759         free(queued_cmd);
 760         strbuf_release(&input);
 761 }
 762
 763 #define DEFAULT_FORMAT "%(objectname) %(objecttype) %(objectsize)"
 764
 765 static int batch_objects(struct batch_options *opt)
 766 {
 767         struct strbuf input = STRBUF_INIT;
 768         struct strbuf output = STRBUF_INIT;
 769         struct expand_data data;
 770         int save_warning;
 771         int retval = 0;
 772
 773         /*
 774          * Expand once with our special mark_query flag, which will prime the
 775          * object_info to be handed to oid_object_info_extended for each
 776          * object.
 777          */
 778         memset(&data, 0, sizeof(data));
 779         data.mark_query = 1;
 780         strbuf_expand(&output,
 781                       opt->format ? opt->format : DEFAULT_FORMAT,
 782                       expand_format,
 783                       &data);
 784         data.mark_query = 0;
 785         strbuf_release(&output);
 786         if (opt->transform_mode)
 787                 data.split_on_whitespace = 1;
 788
 789         if (opt->format && !strcmp(opt->format, DEFAULT_FORMAT))
 790                 opt->format = NULL;
 791         /*
 792          * If we are printing out the object, then always fill in the type,
 793          * since we will want to decide whether or not to stream.
 794          */
 795         if (opt->batch_mode == BATCH_MODE_CONTENTS)
 796                 data.info.typep = &data.type;
 797
 798         if (opt->all_objects) {
 799                 struct object_cb_data cb;
 800                 struct object_info empty = OBJECT_INFO_INIT;
 801
 802                 if (!memcmp(&data.info, &empty, sizeof(empty)))
 803                         data.skip_object_info = 1;
 804
 805                 if (repo_has_promisor_remote(the_repository))
 806                         warning("This repository uses promisor remotes. Some objects may not be loaded.");
 807
 808                 read_replace_refs = 0;
 809
 810                 cb.opt = opt;
 811                 cb.expand = &data;
 812                 cb.scratch = &output;
 813
 814                 if (opt->unordered) {
 815                         struct oidset seen = OIDSET_INIT;
 816
 817                         cb.seen = &seen;
 818
 819                         for_each_loose_object(batch_unordered_loose, &cb, 0);
 820                         for_each_packed_object(batch_unordered_packed, &cb,
 821                                                FOR_EACH_OBJECT_PACK_ORDER);
 822
 823                         oidset_clear(&seen);
 824                 } else {
 825                         struct oid_array sa = OID_ARRAY_INIT;
 826
 827                         for_each_loose_object(collect_loose_object, &sa, 0);
 828                         for_each_packed_object(collect_packed_object, &sa, 0);
 829
 830                         oid_array_for_each_unique(&sa, batch_object_cb, &cb);
 831
 832                         oid_array_clear(&sa);
 833                 }
 834
 835                 strbuf_release(&output);
 836                 return 0;
 837         }
 838
 839         /*
 840          * We are going to call get_sha1 on a potentially very large number of
 841          * objects. In most large cases, these will be actual object sha1s. The
 842          * cost to double-check that each one is not also a ref (just so we can
 843          * warn) ends up dwarfing the actual cost of the object lookups
 844          * themselves. We can work around it by just turning off the warning.
 845          */
 846         save_warning = warn_on_object_refname_ambiguity;
 847         warn_on_object_refname_ambiguity = 0;
 848
 849         if (opt->batch_mode == BATCH_MODE_QUEUE_AND_DISPATCH) {
 850                 batch_objects_command(opt, &output, &data);
 851                 goto cleanup;
 852         }
 853
 854         while (1) {
 855                 int ret;
 856                 if (opt->nul_terminated)
 857                         ret = strbuf_getline_nul(&input, stdin);
 858                 else
 859                         ret = strbuf_getline(&input, stdin);
 860
 861                 if (ret == EOF)
 862                         break;
 863
 864                 if (data.split_on_whitespace) {
 865                         /*
 866                          * Split at first whitespace, tying off the beginning
 867                          * of the string and saving the remainder (or NULL) in
 868                          * data.rest.
 869                          */
 870                         char *p = strpbrk(input.buf, " \t");
 871                         if (p) {
 872                                 while (*p && strchr(" \t", *p))
 873                                         *p++ = '\0';
 874                         }
 875                         data.rest = p;
 876                 }
 877
 878                 batch_one_object(input.buf, &output, opt, &data);
 879         }
 880
 881  cleanup:
 882         strbuf_release(&input);
 883         strbuf_release(&output);
 884         warn_on_object_refname_ambiguity = save_warning;
 885         return retval;
 886 }
 887
 888 static int git_cat_file_config(const char *var, const char *value, void *cb)
 889 {
 890         if (userdiff_config(var, value) < 0)
 891                 return -1;
 892
 893         return git_default_config(var, value, cb);
 894 }
 895
 896 static int batch_option_callback(const struct option *opt,
 897                                  const char *arg,
 898                                  int unset)
 899 {
 900         struct batch_options *bo = opt->value;
 901
 902         BUG_ON_OPT_NEG(unset);
 903
 904         if (bo->enabled) {
 905                 return error(_("only one batch option may be specified"));
 906         }
 907
 908         bo->enabled = 1;
 909
 910         if (!strcmp(opt->long_name, "batch"))
 911                 bo->batch_mode = BATCH_MODE_CONTENTS;
 912         else if (!strcmp(opt->long_name, "batch-check"))
 913                 bo->batch_mode = BATCH_MODE_INFO;
 914         else if (!strcmp(opt->long_name, "batch-command"))
 915                 bo->batch_mode = BATCH_MODE_QUEUE_AND_DISPATCH;
 916         else
 917                 BUG("%s given to batch-option-callback", opt->long_name);
 918
 919         bo->format = arg;
 920
 921         return 0;
 922 }
 923
 924 int cmd_cat_file(int argc, const char **argv, const char *prefix)
 925 {
 926         int opt = 0;
 927         int opt_cw = 0;
 928         int opt_epts = 0;
 929         const char *exp_type = NULL, *obj_name = NULL;
 930         struct batch_options batch = {0};
 931         int unknown_type = 0;
 932
 933         const char * const usage[] = {
 934                 N_("git cat-file <type> <object>"),
 935                 N_("git cat-file (-e | -p) <object>"),
 936                 N_("git cat-file (-t | -s) [--allow-unknown-type] <object>"),
 937                 N_("git cat-file (--batch | --batch-check | --batch-command) [--batch-all-objects]\n"
 938                    "             [--buffer] [--follow-symlinks] [--unordered]\n"
 939                    "             [--textconv | --filters] [-z]"),
 940                 N_("git cat-file (--textconv | --filters)\n"
 941                    "             [<rev>:<path|tree-ish> | --path=<path|tree-ish> <rev>]"),
 942                 NULL
 943         };
 944         const struct option options[] = {
 945                 /* Simple queries */
 946                 OPT_GROUP(N_("Check object existence or emit object contents")),
 947                 OPT_CMDMODE('e', NULL, &opt,
 948                             N_("check if <object> exists"), 'e'),
 949                 OPT_CMDMODE('p', NULL, &opt, N_("pretty-print <object> content"), 'p'),
 950
 951                 OPT_GROUP(N_("Emit [broken] object attributes")),
 952                 OPT_CMDMODE('t', NULL, &opt, N_("show object type (one of 'blob', 'tree', 'commit', 'tag', ...)"), 't'),
 953                 OPT_CMDMODE('s', NULL, &opt, N_("show object size"), 's'),
 954                 OPT_BOOL(0, "allow-unknown-type", &unknown_type,
 955                           N_("allow -s and -t to work with broken/corrupt objects")),
 956                 OPT_BOOL(0, "use-mailmap", &use_mailmap, N_("use mail map file")),
 957                 OPT_ALIAS(0, "mailmap", "use-mailmap"),
 958                 /* Batch mode */
 959                 OPT_GROUP(N_("Batch objects requested on stdin (or --batch-all-objects)")),
 960                 OPT_CALLBACK_F(0, "batch", &batch, N_("format"),
 961                         N_("show full <object> or <rev> contents"),
 962                         PARSE_OPT_OPTARG | PARSE_OPT_NONEG,
 963                         batch_option_callback),
 964                 OPT_CALLBACK_F(0, "batch-check", &batch, N_("format"),
 965                         N_("like --batch, but don't emit <contents>"),
 966                         PARSE_OPT_OPTARG | PARSE_OPT_NONEG,
 967                         batch_option_callback),
 968                 OPT_BOOL('z', NULL, &batch.nul_terminated, N_("stdin is NUL-terminated")),
 969                 OPT_CALLBACK_F(0, "batch-command", &batch, N_("format"),
 970                         N_("read commands from stdin"),
 971                         PARSE_OPT_OPTARG | PARSE_OPT_NONEG,
 972                         batch_option_callback),
 973                 OPT_CMDMODE(0, "batch-all-objects", &opt,
 974                             N_("with --batch[-check]: ignores stdin, batches all known objects"), 'b'),
 975                 /* Batch-specific options */
 976                 OPT_GROUP(N_("Change or optimize batch output")),
 977                 OPT_BOOL(0, "buffer", &batch.buffer_output, N_("buffer --batch output")),
 978                 OPT_BOOL(0, "follow-symlinks", &batch.follow_symlinks,
 979                          N_("follow in-tree symlinks")),
 980                 OPT_BOOL(0, "unordered", &batch.unordered,
 981                          N_("do not order objects before emitting them")),
 982                 /* Textconv options, stand-ole*/
 983                 OPT_GROUP(N_("Emit object (blob or tree) with conversion or filter (stand-alone, or with batch)")),
 984                 OPT_CMDMODE(0, "textconv", &opt,
 985                             N_("run textconv on object's content"), 'c'),
 986                 OPT_CMDMODE(0, "filters", &opt,
 987                             N_("run filters on object's content"), 'w'),
 988                 OPT_STRING(0, "path", &force_path, N_("blob|tree"),
 989                            N_("use a <path> for (--textconv | --filters); Not with 'batch'")),
 990                 OPT_END()
 991         };
 992
 993         git_config(git_cat_file_config, NULL);
 994
 995         batch.buffer_output = -1;
 996
 997         argc = parse_options(argc, argv, prefix, options, usage, 0);
 998         opt_cw = (opt == 'c' || opt == 'w');
 999         opt_epts = (opt == 'e' || opt == 'p' || opt == 't' || opt == 's');
1000
1001         if (use_mailmap)
1002                 read_mailmap(&mailmap);
1003
1004         /* --batch-all-objects? */
1005         if (opt == 'b')
1006                 batch.all_objects = 1;
1007
1008         /* Option compatibility */
1009         if (force_path && !opt_cw)
1010                 usage_msg_optf(_("'%s=<%s>' needs '%s' or '%s'"),
1011                                usage, options,
1012                                "--path", _("path|tree-ish"), "--filters",
1013                                "--textconv");
1014
1015         /* Option compatibility with batch mode */
1016         if (batch.enabled)
1017                 ;
1018         else if (batch.follow_symlinks)
1019                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
1020                                "--follow-symlinks");
1021         else if (batch.buffer_output >= 0)
1022                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
1023                                "--buffer");
1024         else if (batch.all_objects)
1025                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
1026                                "--batch-all-objects");
1027         else if (batch.nul_terminated)
1028                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
1029                                "-z");
1030
1031         /* Batch defaults */
1032         if (batch.buffer_output < 0)
1033                 batch.buffer_output = batch.all_objects;
1034
1035         /* Return early if we're in batch mode? */
1036         if (batch.enabled) {
1037                 if (opt_cw)
1038                         batch.transform_mode = opt;
1039                 else if (opt && opt != 'b')
1040                         usage_msg_optf(_("'-%c' is incompatible with batch mode"),
1041                                        usage, options, opt);
1042                 else if (argc)
1043                         usage_msg_opt(_("batch modes take no arguments"), usage,
1044                                       options);
1045
1046                 return batch_objects(&batch);
1047         }
1048
1049         if (opt) {
1050                 if (!argc && opt == 'c')
1051                         usage_msg_optf(_("<rev> required with '%s'"),
1052                                        usage, options, "--textconv");
1053                 else if (!argc && opt == 'w')
1054                         usage_msg_optf(_("<rev> required with '%s'"),
1055                                        usage, options, "--filters");
1056                 else if (!argc && opt_epts)
1057                         usage_msg_optf(_("<object> required with '-%c'"),
1058                                        usage, options, opt);
1059                 else if (argc == 1)
1060                         obj_name = argv[0];
1061                 else
1062                         usage_msg_opt(_("too many arguments"), usage, options);
1063         } else if (!argc) {
1064                 usage_with_options(usage, options);
1065         } else if (argc != 2) {
1066                 usage_msg_optf(_("only two arguments allowed in <type> <object> mode, not %d"),
1067                               usage, options, argc);
1068         } else if (argc) {
1069                 exp_type = argv[0];
1070                 obj_name = argv[1];
1071         }
1072
1073         if (unknown_type && opt != 't' && opt != 's')
1074                 die("git cat-file --allow-unknown-type: use with -s or -t");
1075         return cat_one_file(opt, exp_type, obj_name, unknown_type);
1076 }