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 "builtin.h"
   8 #include "config.h"
   9 #include "convert.h"
  10 #include "diff.h"
  11 #include "environment.h"
  12 #include "gettext.h"
  13 #include "hex.h"
  14 #include "ident.h"
  15 #include "parse-options.h"
  16 #include "userdiff.h"
  17 #include "streaming.h"
  18 #include "oid-array.h"
  19 #include "packfile.h"
  20 #include "object-file.h"
  21 #include "object-name.h"
  22 #include "object-store-ll.h"
  23 #include "replace-object.h"
  24 #include "promisor-remote.h"
  25 #include "mailmap.h"
  26 #include "write-or-die.h"
  27
  28 enum batch_mode {
  29         BATCH_MODE_CONTENTS,
  30         BATCH_MODE_INFO,
  31         BATCH_MODE_QUEUE_AND_DISPATCH,
  32 };
  33
  34 struct batch_options {
  35         int enabled;
  36         int follow_symlinks;
  37         enum batch_mode batch_mode;
  38         int buffer_output;
  39         int all_objects;
  40         int unordered;
  41         int transform_mode; /* may be 'w' or 'c' for --filters or --textconv */
  42         char input_delim;
  43         char output_delim;
  44         const char *format;
  45 };
  46
  47 static const char *force_path;
  48
  49 static struct string_list mailmap = STRING_LIST_INIT_NODUP;
  50 static int use_mailmap;
  51
  52 static char *replace_idents_using_mailmap(char *, size_t *);
  53
  54 static char *replace_idents_using_mailmap(char *object_buf, size_t *size)
  55 {
  56         struct strbuf sb = STRBUF_INIT;
  57         const char *headers[] = { "author ", "committer ", "tagger ", NULL };
  58
  59         strbuf_attach(&sb, object_buf, *size, *size + 1);
  60         apply_mailmap_to_header(&sb, headers, &mailmap);
  61         *size = sb.len;
  62         return strbuf_detach(&sb, NULL);
  63 }
  64
  65 static int filter_object(const char *path, unsigned mode,
  66                          const struct object_id *oid,
  67                          char **buf, unsigned long *size)
  68 {
  69         enum object_type type;
  70
  71         *buf = repo_read_object_file(the_repository, oid, &type, size);
  72         if (!*buf)
  73                 return error(_("cannot read object %s '%s'"),
  74                              oid_to_hex(oid), path);
  75         if ((type == OBJ_BLOB) && S_ISREG(mode)) {
  76                 struct strbuf strbuf = STRBUF_INIT;
  77                 struct checkout_metadata meta;
  78
  79                 init_checkout_metadata(&meta, NULL, NULL, oid);
  80                 if (convert_to_working_tree(&the_index, path, *buf, *size, &strbuf, &meta)) {
  81                         free(*buf);
  82                         *size = strbuf.len;
  83                         *buf = strbuf_detach(&strbuf, NULL);
  84                 }
  85         }
  86
  87         return 0;
  88 }
  89
  90 static int stream_blob(const struct object_id *oid)
  91 {
  92         if (stream_blob_to_fd(1, oid, NULL, 0))
  93                 die("unable to stream %s to stdout", oid_to_hex(oid));
  94         return 0;
  95 }
  96
  97 static int cat_one_file(int opt, const char *exp_type, const char *obj_name,
  98                         int unknown_type)
  99 {
 100         int ret;
 101         struct object_id oid;
 102         enum object_type type;
 103         char *buf;
 104         unsigned long size;
 105         struct object_context obj_context;
 106         struct object_info oi = OBJECT_INFO_INIT;
 107         struct strbuf sb = STRBUF_INIT;
 108         unsigned flags = OBJECT_INFO_LOOKUP_REPLACE;
 109         unsigned get_oid_flags = GET_OID_RECORD_PATH | GET_OID_ONLY_TO_DIE;
 110         const char *path = force_path;
 111         const int opt_cw = (opt == 'c' || opt == 'w');
 112         if (!path && opt_cw)
 113                 get_oid_flags |= GET_OID_REQUIRE_PATH;
 114
 115         if (unknown_type)
 116                 flags |= OBJECT_INFO_ALLOW_UNKNOWN_TYPE;
 117
 118         if (get_oid_with_context(the_repository, obj_name, get_oid_flags, &oid,
 119                                  &obj_context))
 120                 die("Not a valid object name %s", obj_name);
 121
 122         if (!path)
 123                 path = obj_context.path;
 124         if (obj_context.mode == S_IFINVALID)
 125                 obj_context.mode = 0100644;
 126
 127         buf = NULL;
 128         switch (opt) {
 129         case 't':
 130                 oi.type_name = &sb;
 131                 if (oid_object_info_extended(the_repository, &oid, &oi, flags) < 0)
 132                         die("git cat-file: could not get object info");
 133                 if (sb.len) {
 134                         printf("%s\n", sb.buf);
 135                         strbuf_release(&sb);
 136                         ret = 0;
 137                         goto cleanup;
 138                 }
 139                 break;
 140
 141         case 's':
 142                 oi.sizep = &size;
 143
 144                 if (use_mailmap) {
 145                         oi.typep = &type;
 146                         oi.contentp = (void**)&buf;
 147                 }
 148
 149                 if (oid_object_info_extended(the_repository, &oid, &oi, flags) < 0)
 150                         die("git cat-file: could not get object info");
 151
 152                 if (use_mailmap && (type == OBJ_COMMIT || type == OBJ_TAG)) {
 153                         size_t s = size;
 154                         buf = replace_idents_using_mailmap(buf, &s);
 155                         size = cast_size_t_to_ulong(s);
 156                 }
 157
 158                 printf("%"PRIuMAX"\n", (uintmax_t)size);
 159                 ret = 0;
 160                 goto cleanup;
 161
 162         case 'e':
 163                 return !repo_has_object_file(the_repository, &oid);
 164
 165         case 'w':
 166
 167                 if (filter_object(path, obj_context.mode,
 168                                   &oid, &buf, &size)) {
 169                         ret = -1;
 170                         goto cleanup;
 171                 }
 172                 break;
 173
 174         case 'c':
 175                 if (textconv_object(the_repository, path, obj_context.mode,
 176                                     &oid, 1, &buf, &size))
 177                         break;
 178                 /* else fallthrough */
 179
 180         case 'p':
 181                 type = oid_object_info(the_repository, &oid, NULL);
 182                 if (type < 0)
 183                         die("Not a valid object name %s", obj_name);
 184
 185                 /* custom pretty-print here */
 186                 if (type == OBJ_TREE) {
 187                         const char *ls_args[3] = { NULL };
 188                         ls_args[0] =  "ls-tree";
 189                         ls_args[1] =  obj_name;
 190                         ret = cmd_ls_tree(2, ls_args, NULL);
 191                         goto cleanup;
 192                 }
 193
 194                 if (type == OBJ_BLOB) {
 195                         ret = stream_blob(&oid);
 196                         goto cleanup;
 197                 }
 198                 buf = repo_read_object_file(the_repository, &oid, &type,
 199                                             &size);
 200                 if (!buf)
 201                         die("Cannot read object %s", obj_name);
 202
 203                 if (use_mailmap) {
 204                         size_t s = size;
 205                         buf = replace_idents_using_mailmap(buf, &s);
 206                         size = cast_size_t_to_ulong(s);
 207                 }
 208
 209                 /* otherwise just spit out the data */
 210                 break;
 211
 212         case 0:
 213         {
 214                 enum object_type exp_type_id = type_from_string(exp_type);
 215
 216                 if (exp_type_id == OBJ_BLOB) {
 217                         struct object_id blob_oid;
 218                         if (oid_object_info(the_repository, &oid, NULL) == OBJ_TAG) {
 219                                 char *buffer = repo_read_object_file(the_repository,
 220                                                                      &oid,
 221                                                                      &type,
 222                                                                      &size);
 223                                 const char *target;
 224
 225                                 if (!buffer)
 226                                         die(_("unable to read %s"), oid_to_hex(&oid));
 227
 228                                 if (!skip_prefix(buffer, "object ", &target) ||
 229                                     get_oid_hex(target, &blob_oid))
 230                                         die("%s not a valid tag", oid_to_hex(&oid));
 231                                 free(buffer);
 232                         } else
 233                                 oidcpy(&blob_oid, &oid);
 234
 235                         if (oid_object_info(the_repository, &blob_oid, NULL) == OBJ_BLOB) {
 236                                 ret = stream_blob(&blob_oid);
 237                                 goto cleanup;
 238                         }
 239                         /*
 240                          * we attempted to dereference a tag to a blob
 241                          * and failed; there may be new dereference
 242                          * mechanisms this code is not aware of.
 243                          * fall-back to the usual case.
 244                          */
 245                 }
 246                 buf = read_object_with_reference(the_repository, &oid,
 247                                                  exp_type_id, &size, NULL);
 248
 249                 if (use_mailmap) {
 250                         size_t s = size;
 251                         buf = replace_idents_using_mailmap(buf, &s);
 252                         size = cast_size_t_to_ulong(s);
 253                 }
 254                 break;
 255         }
 256         default:
 257                 die("git cat-file: unknown option: %s", exp_type);
 258         }
 259
 260         if (!buf)
 261                 die("git cat-file %s: bad file", obj_name);
 262
 263         write_or_die(1, buf, size);
 264         ret = 0;
 265 cleanup:
 266         free(buf);
 267         free(obj_context.path);
 268         return ret;
 269 }
 270
 271 struct expand_data {
 272         struct object_id oid;
 273         enum object_type type;
 274         unsigned long size;
 275         off_t disk_size;
 276         const char *rest;
 277         struct object_id delta_base_oid;
 278
 279         /*
 280          * If mark_query is true, we do not expand anything, but rather
 281          * just mark the object_info with items we wish to query.
 282          */
 283         int mark_query;
 284
 285         /*
 286          * Whether to split the input on whitespace before feeding it to
 287          * get_sha1; this is decided during the mark_query phase based on
 288          * whether we have a %(rest) token in our format.
 289          */
 290         int split_on_whitespace;
 291
 292         /*
 293          * After a mark_query run, this object_info is set up to be
 294          * passed to oid_object_info_extended. It will point to the data
 295          * elements above, so you can retrieve the response from there.
 296          */
 297         struct object_info info;
 298
 299         /*
 300          * This flag will be true if the requested batch format and options
 301          * don't require us to call oid_object_info, which can then be
 302          * optimized out.
 303          */
 304         unsigned skip_object_info : 1;
 305 };
 306
 307 static int is_atom(const char *atom, const char *s, int slen)
 308 {
 309         int alen = strlen(atom);
 310         return alen == slen && !memcmp(atom, s, alen);
 311 }
 312
 313 static int expand_atom(struct strbuf *sb, const char *atom, int len,
 314                        struct expand_data *data)
 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                 return 0;
 347         return 1;
 348 }
 349
 350 static void expand_format(struct strbuf *sb, const char *start,
 351                           struct expand_data *data)
 352 {
 353         while (strbuf_expand_step(sb, &start)) {
 354                 const char *end;
 355
 356                 if (skip_prefix(start, "%", &start) || *start != '(')
 357                         strbuf_addch(sb, '%');
 358                 else if ((end = strchr(start + 1, ')')) &&
 359                          expand_atom(sb, start + 1, end - start - 1, data))
 360                         start = end + 1;
 361                 else
 362                         strbuf_expand_bad_format(start, "cat-file");
 363         }
 364 }
 365
 366 static void batch_write(struct batch_options *opt, const void *data, int len)
 367 {
 368         if (opt->buffer_output) {
 369                 if (fwrite(data, 1, len, stdout) != len)
 370                         die_errno("unable to write to stdout");
 371         } else
 372                 write_or_die(1, data, len);
 373 }
 374
 375 static void print_object_or_die(struct batch_options *opt, struct expand_data *data)
 376 {
 377         const struct object_id *oid = &data->oid;
 378
 379         assert(data->info.typep);
 380
 381         if (data->type == OBJ_BLOB) {
 382                 if (opt->buffer_output)
 383                         fflush(stdout);
 384                 if (opt->transform_mode) {
 385                         char *contents;
 386                         unsigned long size;
 387
 388                         if (!data->rest)
 389                                 die("missing path for '%s'", oid_to_hex(oid));
 390
 391                         if (opt->transform_mode == 'w') {
 392                                 if (filter_object(data->rest, 0100644, oid,
 393                                                   &contents, &size))
 394                                         die("could not convert '%s' %s",
 395                                             oid_to_hex(oid), data->rest);
 396                         } else if (opt->transform_mode == 'c') {
 397                                 enum object_type type;
 398                                 if (!textconv_object(the_repository,
 399                                                      data->rest, 0100644, oid,
 400                                                      1, &contents, &size))
 401                                         contents = repo_read_object_file(the_repository,
 402                                                                          oid,
 403                                                                          &type,
 404                                                                          &size);
 405                                 if (!contents)
 406                                         die("could not convert '%s' %s",
 407                                             oid_to_hex(oid), data->rest);
 408                         } else
 409                                 BUG("invalid transform_mode: %c", opt->transform_mode);
 410                         batch_write(opt, contents, size);
 411                         free(contents);
 412                 } else {
 413                         stream_blob(oid);
 414                 }
 415         }
 416         else {
 417                 enum object_type type;
 418                 unsigned long size;
 419                 void *contents;
 420
 421                 contents = repo_read_object_file(the_repository, oid, &type,
 422                                                  &size);
 423                 if (!contents)
 424                         die("object %s disappeared", oid_to_hex(oid));
 425
 426                 if (use_mailmap) {
 427                         size_t s = size;
 428                         contents = replace_idents_using_mailmap(contents, &s);
 429                         size = cast_size_t_to_ulong(s);
 430                 }
 431
 432                 if (type != data->type)
 433                         die("object %s changed type!?", oid_to_hex(oid));
 434                 if (data->info.sizep && size != data->size && !use_mailmap)
 435                         die("object %s changed size!?", oid_to_hex(oid));
 436
 437                 batch_write(opt, contents, size);
 438                 free(contents);
 439         }
 440 }
 441
 442 static void print_default_format(struct strbuf *scratch, struct expand_data *data,
 443                                  struct batch_options *opt)
 444 {
 445         strbuf_addf(scratch, "%s %s %"PRIuMAX"%c", oid_to_hex(&data->oid),
 446                     type_name(data->type),
 447                     (uintmax_t)data->size, opt->output_delim);
 448 }
 449
 450 /*
 451  * If "pack" is non-NULL, then "offset" is the byte offset within the pack from
 452  * which the object may be accessed (though note that we may also rely on
 453  * data->oid, too). If "pack" is NULL, then offset is ignored.
 454  */
 455 static void batch_object_write(const char *obj_name,
 456                                struct strbuf *scratch,
 457                                struct batch_options *opt,
 458                                struct expand_data *data,
 459                                struct packed_git *pack,
 460                                off_t offset)
 461 {
 462         if (!data->skip_object_info) {
 463                 int ret;
 464
 465                 if (use_mailmap)
 466                         data->info.typep = &data->type;
 467
 468                 if (pack)
 469                         ret = packed_object_info(the_repository, pack, offset,
 470                                                  &data->info);
 471                 else
 472                         ret = oid_object_info_extended(the_repository,
 473                                                        &data->oid, &data->info,
 474                                                        OBJECT_INFO_LOOKUP_REPLACE);
 475                 if (ret < 0) {
 476                         printf("%s missing%c",
 477                                obj_name ? obj_name : oid_to_hex(&data->oid), opt->output_delim);
 478                         fflush(stdout);
 479                         return;
 480                 }
 481
 482                 if (use_mailmap && (data->type == OBJ_COMMIT || data->type == OBJ_TAG)) {
 483                         size_t s = data->size;
 484                         char *buf = NULL;
 485
 486                         buf = repo_read_object_file(the_repository, &data->oid, &data->type,
 487                                                     &data->size);
 488                         if (!buf)
 489                                 die(_("unable to read %s"), oid_to_hex(&data->oid));
 490                         buf = replace_idents_using_mailmap(buf, &s);
 491                         data->size = cast_size_t_to_ulong(s);
 492
 493                         free(buf);
 494                 }
 495         }
 496
 497         strbuf_reset(scratch);
 498
 499         if (!opt->format) {
 500                 print_default_format(scratch, data, opt);
 501         } else {
 502                 expand_format(scratch, opt->format, data);
 503                 strbuf_addch(scratch, opt->output_delim);
 504         }
 505
 506         batch_write(opt, scratch->buf, scratch->len);
 507
 508         if (opt->batch_mode == BATCH_MODE_CONTENTS) {
 509                 print_object_or_die(opt, data);
 510                 batch_write(opt, &opt->output_delim, 1);
 511         }
 512 }
 513
 514 static void batch_one_object(const char *obj_name,
 515                              struct strbuf *scratch,
 516                              struct batch_options *opt,
 517                              struct expand_data *data)
 518 {
 519         struct object_context ctx;
 520         int flags = opt->follow_symlinks ? GET_OID_FOLLOW_SYMLINKS : 0;
 521         enum get_oid_result result;
 522
 523         result = get_oid_with_context(the_repository, obj_name,
 524                                       flags, &data->oid, &ctx);
 525         if (result != FOUND) {
 526                 switch (result) {
 527                 case MISSING_OBJECT:
 528                         printf("%s missing%c", obj_name, opt->output_delim);
 529                         break;
 530                 case SHORT_NAME_AMBIGUOUS:
 531                         printf("%s ambiguous%c", obj_name, opt->output_delim);
 532                         break;
 533                 case DANGLING_SYMLINK:
 534                         printf("dangling %"PRIuMAX"%c%s%c",
 535                                (uintmax_t)strlen(obj_name),
 536                                opt->output_delim, obj_name, opt->output_delim);
 537                         break;
 538                 case SYMLINK_LOOP:
 539                         printf("loop %"PRIuMAX"%c%s%c",
 540                                (uintmax_t)strlen(obj_name),
 541                                opt->output_delim, obj_name, opt->output_delim);
 542                         break;
 543                 case NOT_DIR:
 544                         printf("notdir %"PRIuMAX"%c%s%c",
 545                                (uintmax_t)strlen(obj_name),
 546                                opt->output_delim, obj_name, opt->output_delim);
 547                         break;
 548                 default:
 549                         BUG("unknown get_sha1_with_context result %d\n",
 550                                result);
 551                         break;
 552                 }
 553                 fflush(stdout);
 554                 return;
 555         }
 556
 557         if (ctx.mode == 0) {
 558                 printf("symlink %"PRIuMAX"%c%s%c",
 559                        (uintmax_t)ctx.symlink_path.len,
 560                        opt->output_delim, ctx.symlink_path.buf, opt->output_delim);
 561                 fflush(stdout);
 562                 return;
 563         }
 564
 565         batch_object_write(obj_name, scratch, opt, data, NULL, 0);
 566 }
 567
 568 struct object_cb_data {
 569         struct batch_options *opt;
 570         struct expand_data *expand;
 571         struct oidset *seen;
 572         struct strbuf *scratch;
 573 };
 574
 575 static int batch_object_cb(const struct object_id *oid, void *vdata)
 576 {
 577         struct object_cb_data *data = vdata;
 578         oidcpy(&data->expand->oid, oid);
 579         batch_object_write(NULL, data->scratch, data->opt, data->expand,
 580                            NULL, 0);
 581         return 0;
 582 }
 583
 584 static int collect_loose_object(const struct object_id *oid,
 585                                 const char *path UNUSED,
 586                                 void *data)
 587 {
 588         oid_array_append(data, oid);
 589         return 0;
 590 }
 591
 592 static int collect_packed_object(const struct object_id *oid,
 593                                  struct packed_git *pack UNUSED,
 594                                  uint32_t pos UNUSED,
 595                                  void *data)
 596 {
 597         oid_array_append(data, oid);
 598         return 0;
 599 }
 600
 601 static int batch_unordered_object(const struct object_id *oid,
 602                                   struct packed_git *pack, off_t offset,
 603                                   void *vdata)
 604 {
 605         struct object_cb_data *data = vdata;
 606
 607         if (oidset_insert(data->seen, oid))
 608                 return 0;
 609
 610         oidcpy(&data->expand->oid, oid);
 611         batch_object_write(NULL, data->scratch, data->opt, data->expand,
 612                            pack, offset);
 613         return 0;
 614 }
 615
 616 static int batch_unordered_loose(const struct object_id *oid,
 617                                  const char *path UNUSED,
 618                                  void *data)
 619 {
 620         return batch_unordered_object(oid, NULL, 0, data);
 621 }
 622
 623 static int batch_unordered_packed(const struct object_id *oid,
 624                                   struct packed_git *pack,
 625                                   uint32_t pos,
 626                                   void *data)
 627 {
 628         return batch_unordered_object(oid, pack,
 629                                       nth_packed_object_offset(pack, pos),
 630                                       data);
 631 }
 632
 633 typedef void (*parse_cmd_fn_t)(struct batch_options *, const char *,
 634                                struct strbuf *, struct expand_data *);
 635
 636 struct queued_cmd {
 637         parse_cmd_fn_t fn;
 638         char *line;
 639 };
 640
 641 static void parse_cmd_contents(struct batch_options *opt,
 642                              const char *line,
 643                              struct strbuf *output,
 644                              struct expand_data *data)
 645 {
 646         opt->batch_mode = BATCH_MODE_CONTENTS;
 647         batch_one_object(line, output, opt, data);
 648 }
 649
 650 static void parse_cmd_info(struct batch_options *opt,
 651                            const char *line,
 652                            struct strbuf *output,
 653                            struct expand_data *data)
 654 {
 655         opt->batch_mode = BATCH_MODE_INFO;
 656         batch_one_object(line, output, opt, data);
 657 }
 658
 659 static void dispatch_calls(struct batch_options *opt,
 660                 struct strbuf *output,
 661                 struct expand_data *data,
 662                 struct queued_cmd *cmd,
 663                 int nr)
 664 {
 665         int i;
 666
 667         if (!opt->buffer_output)
 668                 die(_("flush is only for --buffer mode"));
 669
 670         for (i = 0; i < nr; i++)
 671                 cmd[i].fn(opt, cmd[i].line, output, data);
 672
 673         fflush(stdout);
 674 }
 675
 676 static void free_cmds(struct queued_cmd *cmd, size_t *nr)
 677 {
 678         size_t i;
 679
 680         for (i = 0; i < *nr; i++)
 681                 FREE_AND_NULL(cmd[i].line);
 682
 683         *nr = 0;
 684 }
 685
 686
 687 static const struct parse_cmd {
 688         const char *name;
 689         parse_cmd_fn_t fn;
 690         unsigned takes_args;
 691 } commands[] = {
 692         { "contents", parse_cmd_contents, 1},
 693         { "info", parse_cmd_info, 1},
 694         { "flush", NULL, 0},
 695 };
 696
 697 static void batch_objects_command(struct batch_options *opt,
 698                                     struct strbuf *output,
 699                                     struct expand_data *data)
 700 {
 701         struct strbuf input = STRBUF_INIT;
 702         struct queued_cmd *queued_cmd = NULL;
 703         size_t alloc = 0, nr = 0;
 704
 705         while (strbuf_getdelim_strip_crlf(&input, stdin, opt->input_delim) != EOF) {
 706                 int i;
 707                 const struct parse_cmd *cmd = NULL;
 708                 const char *p = NULL, *cmd_end;
 709                 struct queued_cmd call = {0};
 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         expand_format(&output,
 781                       opt->format ? opt->format : DEFAULT_FORMAT,
 782                       &data);
 783         data.mark_query = 0;
 784         strbuf_release(&output);
 785         if (opt->transform_mode)
 786                 data.split_on_whitespace = 1;
 787
 788         if (opt->format && !strcmp(opt->format, DEFAULT_FORMAT))
 789                 opt->format = NULL;
 790         /*
 791          * If we are printing out the object, then always fill in the type,
 792          * since we will want to decide whether or not to stream.
 793          */
 794         if (opt->batch_mode == BATCH_MODE_CONTENTS)
 795                 data.info.typep = &data.type;
 796
 797         if (opt->all_objects) {
 798                 struct object_cb_data cb;
 799                 struct object_info empty = OBJECT_INFO_INIT;
 800
 801                 if (!memcmp(&data.info, &empty, sizeof(empty)))
 802                         data.skip_object_info = 1;
 803
 804                 if (repo_has_promisor_remote(the_repository))
 805                         warning("This repository uses promisor remotes. Some objects may not be loaded.");
 806
 807                 disable_replace_refs();
 808
 809                 cb.opt = opt;
 810                 cb.expand = &data;
 811                 cb.scratch = &output;
 812
 813                 if (opt->unordered) {
 814                         struct oidset seen = OIDSET_INIT;
 815
 816                         cb.seen = &seen;
 817
 818                         for_each_loose_object(batch_unordered_loose, &cb, 0);
 819                         for_each_packed_object(batch_unordered_packed, &cb,
 820                                                FOR_EACH_OBJECT_PACK_ORDER);
 821
 822                         oidset_clear(&seen);
 823                 } else {
 824                         struct oid_array sa = OID_ARRAY_INIT;
 825
 826                         for_each_loose_object(collect_loose_object, &sa, 0);
 827                         for_each_packed_object(collect_packed_object, &sa, 0);
 828
 829                         oid_array_for_each_unique(&sa, batch_object_cb, &cb);
 830
 831                         oid_array_clear(&sa);
 832                 }
 833
 834                 strbuf_release(&output);
 835                 return 0;
 836         }
 837
 838         /*
 839          * We are going to call get_sha1 on a potentially very large number of
 840          * objects. In most large cases, these will be actual object sha1s. The
 841          * cost to double-check that each one is not also a ref (just so we can
 842          * warn) ends up dwarfing the actual cost of the object lookups
 843          * themselves. We can work around it by just turning off the warning.
 844          */
 845         save_warning = warn_on_object_refname_ambiguity;
 846         warn_on_object_refname_ambiguity = 0;
 847
 848         if (opt->batch_mode == BATCH_MODE_QUEUE_AND_DISPATCH) {
 849                 batch_objects_command(opt, &output, &data);
 850                 goto cleanup;
 851         }
 852
 853         while (strbuf_getdelim_strip_crlf(&input, stdin, opt->input_delim) != EOF) {
 854                 if (data.split_on_whitespace) {
 855                         /*
 856                          * Split at first whitespace, tying off the beginning
 857                          * of the string and saving the remainder (or NULL) in
 858                          * data.rest.
 859                          */
 860                         char *p = strpbrk(input.buf, " \t");
 861                         if (p) {
 862                                 while (*p && strchr(" \t", *p))
 863                                         *p++ = '\0';
 864                         }
 865                         data.rest = p;
 866                 }
 867
 868                 batch_one_object(input.buf, &output, opt, &data);
 869         }
 870
 871  cleanup:
 872         strbuf_release(&input);
 873         strbuf_release(&output);
 874         warn_on_object_refname_ambiguity = save_warning;
 875         return retval;
 876 }
 877
 878 static int git_cat_file_config(const char *var, const char *value,
 879                                const struct config_context *ctx, void *cb)
 880 {
 881         if (userdiff_config(var, value) < 0)
 882                 return -1;
 883
 884         return git_default_config(var, value, ctx, cb);
 885 }
 886
 887 static int batch_option_callback(const struct option *opt,
 888                                  const char *arg,
 889                                  int unset)
 890 {
 891         struct batch_options *bo = opt->value;
 892
 893         BUG_ON_OPT_NEG(unset);
 894
 895         if (bo->enabled) {
 896                 return error(_("only one batch option may be specified"));
 897         }
 898
 899         bo->enabled = 1;
 900
 901         if (!strcmp(opt->long_name, "batch"))
 902                 bo->batch_mode = BATCH_MODE_CONTENTS;
 903         else if (!strcmp(opt->long_name, "batch-check"))
 904                 bo->batch_mode = BATCH_MODE_INFO;
 905         else if (!strcmp(opt->long_name, "batch-command"))
 906                 bo->batch_mode = BATCH_MODE_QUEUE_AND_DISPATCH;
 907         else
 908                 BUG("%s given to batch-option-callback", opt->long_name);
 909
 910         bo->format = arg;
 911
 912         return 0;
 913 }
 914
 915 int cmd_cat_file(int argc, const char **argv, const char *prefix)
 916 {
 917         int opt = 0;
 918         int opt_cw = 0;
 919         int opt_epts = 0;
 920         const char *exp_type = NULL, *obj_name = NULL;
 921         struct batch_options batch = {0};
 922         int unknown_type = 0;
 923         int input_nul_terminated = 0;
 924         int nul_terminated = 0;
 925
 926         const char * const usage[] = {
 927                 N_("git cat-file <type> <object>"),
 928                 N_("git cat-file (-e | -p) <object>"),
 929                 N_("git cat-file (-t | -s) [--allow-unknown-type] <object>"),
 930                 N_("git cat-file (--textconv | --filters)\n"
 931                    "             [<rev>:<path|tree-ish> | --path=<path|tree-ish> <rev>]"),
 932                 N_("git cat-file (--batch | --batch-check | --batch-command) [--batch-all-objects]\n"
 933                    "             [--buffer] [--follow-symlinks] [--unordered]\n"
 934                    "             [--textconv | --filters] [-Z]"),
 935                 NULL
 936         };
 937         const struct option options[] = {
 938                 /* Simple queries */
 939                 OPT_GROUP(N_("Check object existence or emit object contents")),
 940                 OPT_CMDMODE('e', NULL, &opt,
 941                             N_("check if <object> exists"), 'e'),
 942                 OPT_CMDMODE('p', NULL, &opt, N_("pretty-print <object> content"), 'p'),
 943
 944                 OPT_GROUP(N_("Emit [broken] object attributes")),
 945                 OPT_CMDMODE('t', NULL, &opt, N_("show object type (one of 'blob', 'tree', 'commit', 'tag', ...)"), 't'),
 946                 OPT_CMDMODE('s', NULL, &opt, N_("show object size"), 's'),
 947                 OPT_BOOL(0, "allow-unknown-type", &unknown_type,
 948                           N_("allow -s and -t to work with broken/corrupt objects")),
 949                 OPT_BOOL(0, "use-mailmap", &use_mailmap, N_("use mail map file")),
 950                 OPT_ALIAS(0, "mailmap", "use-mailmap"),
 951                 /* Batch mode */
 952                 OPT_GROUP(N_("Batch objects requested on stdin (or --batch-all-objects)")),
 953                 OPT_CALLBACK_F(0, "batch", &batch, N_("format"),
 954                         N_("show full <object> or <rev> contents"),
 955                         PARSE_OPT_OPTARG | PARSE_OPT_NONEG,
 956                         batch_option_callback),
 957                 OPT_CALLBACK_F(0, "batch-check", &batch, N_("format"),
 958                         N_("like --batch, but don't emit <contents>"),
 959                         PARSE_OPT_OPTARG | PARSE_OPT_NONEG,
 960                         batch_option_callback),
 961                 OPT_BOOL_F('z', NULL, &input_nul_terminated, N_("stdin is NUL-terminated"),
 962                         PARSE_OPT_HIDDEN),
 963                 OPT_BOOL('Z', NULL, &nul_terminated, N_("stdin and stdout is NUL-terminated")),
 964                 OPT_CALLBACK_F(0, "batch-command", &batch, N_("format"),
 965                         N_("read commands from stdin"),
 966                         PARSE_OPT_OPTARG | PARSE_OPT_NONEG,
 967                         batch_option_callback),
 968                 OPT_CMDMODE(0, "batch-all-objects", &opt,
 969                             N_("with --batch[-check]: ignores stdin, batches all known objects"), 'b'),
 970                 /* Batch-specific options */
 971                 OPT_GROUP(N_("Change or optimize batch output")),
 972                 OPT_BOOL(0, "buffer", &batch.buffer_output, N_("buffer --batch output")),
 973                 OPT_BOOL(0, "follow-symlinks", &batch.follow_symlinks,
 974                          N_("follow in-tree symlinks")),
 975                 OPT_BOOL(0, "unordered", &batch.unordered,
 976                          N_("do not order objects before emitting them")),
 977                 /* Textconv options, stand-ole*/
 978                 OPT_GROUP(N_("Emit object (blob or tree) with conversion or filter (stand-alone, or with batch)")),
 979                 OPT_CMDMODE(0, "textconv", &opt,
 980                             N_("run textconv on object's content"), 'c'),
 981                 OPT_CMDMODE(0, "filters", &opt,
 982                             N_("run filters on object's content"), 'w'),
 983                 OPT_STRING(0, "path", &force_path, N_("blob|tree"),
 984                            N_("use a <path> for (--textconv | --filters); Not with 'batch'")),
 985                 OPT_END()
 986         };
 987
 988         git_config(git_cat_file_config, NULL);
 989
 990         batch.buffer_output = -1;
 991
 992         argc = parse_options(argc, argv, prefix, options, usage, 0);
 993         opt_cw = (opt == 'c' || opt == 'w');
 994         opt_epts = (opt == 'e' || opt == 'p' || opt == 't' || opt == 's');
 995
 996         if (use_mailmap)
 997                 read_mailmap(&mailmap);
 998
 999         /* --batch-all-objects? */
1000         if (opt == 'b')
1001                 batch.all_objects = 1;
1002
1003         /* Option compatibility */
1004         if (force_path && !opt_cw)
1005                 usage_msg_optf(_("'%s=<%s>' needs '%s' or '%s'"),
1006                                usage, options,
1007                                "--path", _("path|tree-ish"), "--filters",
1008                                "--textconv");
1009
1010         /* Option compatibility with batch mode */
1011         if (batch.enabled)
1012                 ;
1013         else if (batch.follow_symlinks)
1014                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
1015                                "--follow-symlinks");
1016         else if (batch.buffer_output >= 0)
1017                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
1018                                "--buffer");
1019         else if (batch.all_objects)
1020                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
1021                                "--batch-all-objects");
1022         else if (input_nul_terminated)
1023                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
1024                                "-z");
1025         else if (nul_terminated)
1026                 usage_msg_optf(_("'%s' requires a batch mode"), usage, options,
1027                                "-Z");
1028
1029         batch.input_delim = batch.output_delim = '\n';
1030         if (input_nul_terminated)
1031                 batch.input_delim = '\0';
1032         if (nul_terminated)
1033                 batch.input_delim = batch.output_delim = '\0';
1034
1035         /* Batch defaults */
1036         if (batch.buffer_output < 0)
1037                 batch.buffer_output = batch.all_objects;
1038
1039         /* Return early if we're in batch mode? */
1040         if (batch.enabled) {
1041                 if (opt_cw)
1042                         batch.transform_mode = opt;
1043                 else if (opt && opt != 'b')
1044                         usage_msg_optf(_("'-%c' is incompatible with batch mode"),
1045                                        usage, options, opt);
1046                 else if (argc)
1047                         usage_msg_opt(_("batch modes take no arguments"), usage,
1048                                       options);
1049
1050                 return batch_objects(&batch);
1051         }
1052
1053         if (opt) {
1054                 if (!argc && opt == 'c')
1055                         usage_msg_optf(_("<rev> required with '%s'"),
1056                                        usage, options, "--textconv");
1057                 else if (!argc && opt == 'w')
1058                         usage_msg_optf(_("<rev> required with '%s'"),
1059                                        usage, options, "--filters");
1060                 else if (!argc && opt_epts)
1061                         usage_msg_optf(_("<object> required with '-%c'"),
1062                                        usage, options, opt);
1063                 else if (argc == 1)
1064                         obj_name = argv[0];
1065                 else
1066                         usage_msg_opt(_("too many arguments"), usage, options);
1067         } else if (!argc) {
1068                 usage_with_options(usage, options);
1069         } else if (argc != 2) {
1070                 usage_msg_optf(_("only two arguments allowed in <type> <object> mode, not %d"),
1071                               usage, options, argc);
1072         } else if (argc) {
1073                 exp_type = argv[0];
1074                 obj_name = argv[1];
1075         }
1076
1077         if (unknown_type && opt != 't' && opt != 's')
1078                 die("git cat-file --allow-unknown-type: use with -s or -t");
1079         return cat_one_file(opt, exp_type, obj_name, unknown_type);
1080 }