bundle.c

   1 #define USE_THE_REPOSITORY_VARIABLE
   2 #define DISABLE_SIGN_COMPARE_WARNINGS
   3
   4 #include "git-compat-util.h"
   5 #include "lockfile.h"
   6 #include "bundle.h"
   7 #include "environment.h"
   8 #include "gettext.h"
   9 #include "hex.h"
  10 #include "object-store.h"
  11 #include "repository.h"
  12 #include "object.h"
  13 #include "commit.h"
  14 #include "diff.h"
  15 #include "revision.h"
  16 #include "list-objects.h"
  17 #include "run-command.h"
  18 #include "refs.h"
  19 #include "strvec.h"
  20 #include "list-objects-filter-options.h"
  21 #include "connected.h"
  22 #include "write-or-die.h"
  23
  24 static const char v2_bundle_signature[] = "# v2 git bundle\n";
  25 static const char v3_bundle_signature[] = "# v3 git bundle\n";
  26 static struct {
  27         int version;
  28         const char *signature;
  29 } bundle_sigs[] = {
  30         { 2, v2_bundle_signature },
  31         { 3, v3_bundle_signature },
  32 };
  33
  34 void bundle_header_init(struct bundle_header *header)
  35 {
  36         struct bundle_header blank = BUNDLE_HEADER_INIT;
  37         memcpy(header, &blank, sizeof(*header));
  38 }
  39
  40 void bundle_header_release(struct bundle_header *header)
  41 {
  42         string_list_clear(&header->prerequisites, 1);
  43         string_list_clear(&header->references, 1);
  44         list_objects_filter_release(&header->filter);
  45 }
  46
  47 static int parse_capability(struct bundle_header *header, const char *capability)
  48 {
  49         const char *arg;
  50         if (skip_prefix(capability, "object-format=", &arg)) {
  51                 int algo = hash_algo_by_name(arg);
  52                 if (algo == GIT_HASH_UNKNOWN)
  53                         return error(_("unrecognized bundle hash algorithm: %s"), arg);
  54                 header->hash_algo = &hash_algos[algo];
  55                 return 0;
  56         }
  57         if (skip_prefix(capability, "filter=", &arg)) {
  58                 parse_list_objects_filter(&header->filter, arg);
  59                 return 0;
  60         }
  61         return error(_("unknown capability '%s'"), capability);
  62 }
  63
  64 static int parse_bundle_signature(struct bundle_header *header, const char *line)
  65 {
  66         int i;
  67
  68         for (i = 0; i < ARRAY_SIZE(bundle_sigs); i++) {
  69                 if (!strcmp(line, bundle_sigs[i].signature)) {
  70                         header->version = bundle_sigs[i].version;
  71                         return 0;
  72                 }
  73         }
  74         return -1;
  75 }
  76
  77 int read_bundle_header_fd(int fd, struct bundle_header *header,
  78                           const char *report_path)
  79 {
  80         struct strbuf buf = STRBUF_INIT;
  81         int status = 0;
  82
  83         /* The bundle header begins with the signature */
  84         if (strbuf_getwholeline_fd(&buf, fd, '\n') ||
  85             parse_bundle_signature(header, buf.buf)) {
  86                 if (report_path)
  87                         error(_("'%s' does not look like a v2 or v3 bundle file"),
  88                               report_path);
  89                 status = -1;
  90                 goto abort;
  91         }
  92
  93         /*
  94          * The default hash format for bundles is SHA1, unless told otherwise
  95          * by an "object-format=" capability, which is being handled in
  96          * `parse_capability()`.
  97          */
  98         header->hash_algo = &hash_algos[GIT_HASH_SHA1];
  99
 100         /* The bundle header ends with an empty line */
 101         while (!strbuf_getwholeline_fd(&buf, fd, '\n') &&
 102                buf.len && buf.buf[0] != '\n') {
 103                 struct object_id oid;
 104                 int is_prereq = 0;
 105                 const char *p;
 106
 107                 strbuf_rtrim(&buf);
 108
 109                 if (header->version == 3 && *buf.buf == '@') {
 110                         if (parse_capability(header, buf.buf + 1)) {
 111                                 status = -1;
 112                                 break;
 113                         }
 114                         continue;
 115                 }
 116
 117                 if (*buf.buf == '-') {
 118                         is_prereq = 1;
 119                         strbuf_remove(&buf, 0, 1);
 120                 }
 121
 122                 /*
 123                  * Tip lines have object name, SP, and refname.
 124                  * Prerequisites have object name that is optionally
 125                  * followed by SP and subject line.
 126                  */
 127                 if (parse_oid_hex_algop(buf.buf, &oid, &p, header->hash_algo) ||
 128                     (*p && !isspace(*p)) ||
 129                     (!is_prereq && !*p)) {
 130                         if (report_path)
 131                                 error(_("unrecognized header: %s%s (%d)"),
 132                                       (is_prereq ? "-" : ""), buf.buf, (int)buf.len);
 133                         status = -1;
 134                         break;
 135                 } else {
 136                         struct object_id *dup = oiddup(&oid);
 137                         if (is_prereq)
 138                                 string_list_append(&header->prerequisites, "")->util = dup;
 139                         else
 140                                 string_list_append(&header->references, p + 1)->util = dup;
 141                 }
 142         }
 143
 144  abort:
 145         if (status) {
 146                 close(fd);
 147                 fd = -1;
 148         }
 149         strbuf_release(&buf);
 150         return fd;
 151 }
 152
 153 int read_bundle_header(const char *path, struct bundle_header *header)
 154 {
 155         int fd = open(path, O_RDONLY);
 156
 157         if (fd < 0)
 158                 return error(_("could not open '%s'"), path);
 159         return read_bundle_header_fd(fd, header, path);
 160 }
 161
 162 int is_bundle(const char *path, int quiet)
 163 {
 164         struct bundle_header header = BUNDLE_HEADER_INIT;
 165         int fd = open(path, O_RDONLY);
 166
 167         if (fd < 0)
 168                 return 0;
 169         fd = read_bundle_header_fd(fd, &header, quiet ? NULL : path);
 170         if (fd >= 0)
 171                 close(fd);
 172         bundle_header_release(&header);
 173         return (fd >= 0);
 174 }
 175
 176 static int list_refs(struct string_list *r, int argc, const char **argv)
 177 {
 178         int i;
 179
 180         for (i = 0; i < r->nr; i++) {
 181                 struct object_id *oid;
 182                 const char *name;
 183
 184                 if (argc > 1) {
 185                         int j;
 186                         for (j = 1; j < argc; j++)
 187                                 if (!strcmp(r->items[i].string, argv[j]))
 188                                         break;
 189                         if (j == argc)
 190                                 continue;
 191                 }
 192
 193                 oid = r->items[i].util;
 194                 name = r->items[i].string;
 195                 printf("%s %s\n", oid_to_hex(oid), name);
 196         }
 197         return 0;
 198 }
 199
 200 /* Remember to update object flag allocation in object.h */
 201 #define PREREQ_MARK (1u<<16)
 202
 203 struct string_list_iterator {
 204         struct string_list *list;
 205         size_t cur;
 206 };
 207
 208 static const struct object_id *iterate_ref_map(void *cb_data)
 209 {
 210         struct string_list_iterator *iter = cb_data;
 211
 212         if (iter->cur >= iter->list->nr)
 213                 return NULL;
 214
 215         return iter->list->items[iter->cur++].util;
 216 }
 217
 218 int verify_bundle(struct repository *r,
 219                   struct bundle_header *header,
 220                   enum verify_bundle_flags flags)
 221 {
 222         /*
 223          * Do fast check, then if any prereqs are missing then go line by line
 224          * to be verbose about the errors
 225          */
 226         struct string_list *p = &header->prerequisites;
 227         int i, ret = 0;
 228         const char *message = _("Repository lacks these prerequisite commits:");
 229         struct string_list_iterator iter = {
 230                 .list = p,
 231         };
 232         struct check_connected_options opts = {
 233                 .quiet = 1,
 234         };
 235
 236         if (!r || !r->objects || !r->objects->odb)
 237                 return error(_("need a repository to verify a bundle"));
 238
 239         for (i = 0; i < p->nr; i++) {
 240                 struct string_list_item *e = p->items + i;
 241                 const char *name = e->string;
 242                 struct object_id *oid = e->util;
 243                 struct object *o = parse_object(r, oid);
 244                 if (o)
 245                         continue;
 246                 ret++;
 247                 if (flags & VERIFY_BUNDLE_QUIET)
 248                         continue;
 249                 if (ret == 1)
 250                         error("%s", message);
 251                 error("%s %s", oid_to_hex(oid), name);
 252         }
 253         if (ret)
 254                 goto cleanup;
 255
 256         if ((ret = check_connected(iterate_ref_map, &iter, &opts)))
 257                 error(_("some prerequisite commits exist in the object store, "
 258                         "but are not connected to the repository's history"));
 259
 260         /* TODO: preserve this verbose language. */
 261         if (flags & VERIFY_BUNDLE_VERBOSE) {
 262                 struct string_list *r;
 263
 264                 r = &header->references;
 265                 printf_ln(Q_("The bundle contains this ref:",
 266                              "The bundle contains these %"PRIuMAX" refs:",
 267                              r->nr),
 268                           (uintmax_t)r->nr);
 269                 list_refs(r, 0, NULL);
 270
 271                 r = &header->prerequisites;
 272                 if (!r->nr) {
 273                         printf_ln(_("The bundle records a complete history."));
 274                 } else {
 275                         printf_ln(Q_("The bundle requires this ref:",
 276                                      "The bundle requires these %"PRIuMAX" refs:",
 277                                      r->nr),
 278                                   (uintmax_t)r->nr);
 279                         list_refs(r, 0, NULL);
 280                 }
 281
 282                 printf_ln(_("The bundle uses this hash algorithm: %s"),
 283                           header->hash_algo->name);
 284                 if (header->filter.choice)
 285                         printf_ln(_("The bundle uses this filter: %s"),
 286                                   list_objects_filter_spec(&header->filter));
 287         }
 288 cleanup:
 289         return ret;
 290 }
 291
 292 int list_bundle_refs(struct bundle_header *header, int argc, const char **argv)
 293 {
 294         return list_refs(&header->references, argc, argv);
 295 }
 296
 297 static int is_tag_in_date_range(struct object *tag, struct rev_info *revs)
 298 {
 299         unsigned long size;
 300         enum object_type type;
 301         char *buf = NULL, *line, *lineend;
 302         timestamp_t date;
 303         int result = 1;
 304
 305         if (revs->max_age == -1 && revs->min_age == -1)
 306                 goto out;
 307
 308         buf = repo_read_object_file(the_repository, &tag->oid, &type, &size);
 309         if (!buf)
 310                 goto out;
 311         line = memmem(buf, size, "\ntagger ", 8);
 312         if (!line++)
 313                 goto out;
 314         lineend = memchr(line, '\n', buf + size - line);
 315         line = memchr(line, '>', lineend ? lineend - line : buf + size - line);
 316         if (!line++)
 317                 goto out;
 318         date = parse_timestamp(line, NULL, 10);
 319         result = (revs->max_age == -1 || revs->max_age < date) &&
 320                 (revs->min_age == -1 || revs->min_age > date);
 321 out:
 322         free(buf);
 323         return result;
 324 }
 325
 326
 327 /* Write the pack data to bundle_fd */
 328 static int write_pack_data(int bundle_fd, struct rev_info *revs, struct strvec *pack_options)
 329 {
 330         struct child_process pack_objects = CHILD_PROCESS_INIT;
 331         int i;
 332
 333         strvec_pushl(&pack_objects.args,
 334                      "pack-objects",
 335                      "--stdout", "--thin", "--delta-base-offset",
 336                      NULL);
 337         strvec_pushv(&pack_objects.args, pack_options->v);
 338         if (revs->filter.choice)
 339                 strvec_pushf(&pack_objects.args, "--filter=%s",
 340                              list_objects_filter_spec(&revs->filter));
 341         pack_objects.in = -1;
 342         pack_objects.out = bundle_fd;
 343         pack_objects.git_cmd = 1;
 344
 345         /*
 346          * start_command() will close our descriptor if it's >1. Duplicate it
 347          * to avoid surprising the caller.
 348          */
 349         if (pack_objects.out > 1) {
 350                 pack_objects.out = dup(pack_objects.out);
 351                 if (pack_objects.out < 0) {
 352                         error_errno(_("unable to dup bundle descriptor"));
 353                         child_process_clear(&pack_objects);
 354                         return -1;
 355                 }
 356         }
 357
 358         if (start_command(&pack_objects))
 359                 return error(_("Could not spawn pack-objects"));
 360
 361         for (i = 0; i < revs->pending.nr; i++) {
 362                 struct object *object = revs->pending.objects[i].item;
 363                 if (object->flags & UNINTERESTING)
 364                         write_or_die(pack_objects.in, "^", 1);
 365                 write_or_die(pack_objects.in, oid_to_hex(&object->oid), the_hash_algo->hexsz);
 366                 write_or_die(pack_objects.in, "\n", 1);
 367         }
 368         close(pack_objects.in);
 369         if (finish_command(&pack_objects))
 370                 return error(_("pack-objects died"));
 371         return 0;
 372 }
 373
 374 /*
 375  * Write out bundle refs based on the tips already
 376  * parsed into revs.pending. As a side effect, may
 377  * manipulate revs.pending to include additional
 378  * necessary objects (like tags).
 379  *
 380  * Returns the number of refs written, or negative
 381  * on error.
 382  */
 383 static int write_bundle_refs(int bundle_fd, struct rev_info *revs)
 384 {
 385         int i;
 386         int ref_count = 0;
 387         struct strset objects = STRSET_INIT;
 388
 389         for (i = 0; i < revs->pending.nr; i++) {
 390                 struct object_array_entry *e = revs->pending.objects + i;
 391                 struct object_id oid;
 392                 char *ref;
 393                 const char *display_ref;
 394                 int flag;
 395
 396                 if (e->item->flags & UNINTERESTING)
 397                         continue;
 398                 if (repo_dwim_ref(the_repository, e->name, strlen(e->name),
 399                                   &oid, &ref, 0) != 1)
 400                         goto skip_write_ref;
 401                 if (refs_read_ref_full(get_main_ref_store(the_repository), e->name, RESOLVE_REF_READING, &oid, &flag))
 402                         flag = 0;
 403                 display_ref = (flag & REF_ISSYMREF) ? e->name : ref;
 404
 405                 if (strset_contains(&objects, display_ref))
 406                         goto skip_write_ref;
 407
 408                 if (e->item->type == OBJ_TAG &&
 409                                 !is_tag_in_date_range(e->item, revs)) {
 410                         e->item->flags |= UNINTERESTING;
 411                         goto skip_write_ref;
 412                 }
 413
 414                 /*
 415                  * Make sure the refs we wrote out is correct; --max-count and
 416                  * other limiting options could have prevented all the tips
 417                  * from getting output.
 418                  *
 419                  * Non commit objects such as tags and blobs do not have
 420                  * this issue as they are not affected by those extra
 421                  * constraints.
 422                  */
 423                 if (!(e->item->flags & SHOWN) && e->item->type == OBJ_COMMIT) {
 424                         warning(_("ref '%s' is excluded by the rev-list options"),
 425                                 e->name);
 426                         goto skip_write_ref;
 427                 }
 428
 429                 ref_count++;
 430                 strset_add(&objects, display_ref);
 431                 write_or_die(bundle_fd, oid_to_hex(&e->item->oid), the_hash_algo->hexsz);
 432                 write_or_die(bundle_fd, " ", 1);
 433                 write_or_die(bundle_fd, display_ref, strlen(display_ref));
 434                 write_or_die(bundle_fd, "\n", 1);
 435  skip_write_ref:
 436                 free(ref);
 437         }
 438
 439         strset_clear(&objects);
 440
 441         /* end header */
 442         write_or_die(bundle_fd, "\n", 1);
 443         return ref_count;
 444 }
 445
 446 struct bundle_prerequisites_info {
 447         struct object_array *pending;
 448         int fd;
 449 };
 450
 451 static void write_bundle_prerequisites(struct commit *commit, void *data)
 452 {
 453         struct bundle_prerequisites_info *bpi = data;
 454         struct object *object;
 455         struct pretty_print_context ctx = { 0 };
 456         struct strbuf buf = STRBUF_INIT;
 457
 458         if (!(commit->object.flags & BOUNDARY))
 459                 return;
 460         strbuf_addf(&buf, "-%s ", oid_to_hex(&commit->object.oid));
 461         write_or_die(bpi->fd, buf.buf, buf.len);
 462
 463         ctx.fmt = CMIT_FMT_ONELINE;
 464         ctx.output_encoding = get_log_output_encoding();
 465         strbuf_reset(&buf);
 466         pretty_print_commit(&ctx, commit, &buf);
 467         strbuf_trim(&buf);
 468
 469         object = (struct object *)commit;
 470         object->flags |= UNINTERESTING;
 471         add_object_array_with_path(object, buf.buf, bpi->pending, S_IFINVALID,
 472                                    NULL);
 473         strbuf_addch(&buf, '\n');
 474         write_or_die(bpi->fd, buf.buf, buf.len);
 475         strbuf_release(&buf);
 476 }
 477
 478 int create_bundle(struct repository *r, const char *path,
 479                   int argc, const char **argv, struct strvec *pack_options, int version)
 480 {
 481         struct lock_file lock = LOCK_INIT;
 482         int bundle_fd = -1;
 483         int bundle_to_stdout;
 484         int ref_count = 0;
 485         struct rev_info revs, revs_copy;
 486         int min_version = 2;
 487         struct bundle_prerequisites_info bpi;
 488         int ret;
 489         int i;
 490
 491         /* init revs to list objects for pack-objects later */
 492         save_commit_buffer = 0;
 493         repo_init_revisions(r, &revs, NULL);
 494
 495         /*
 496          * Pre-initialize the '--objects' flag so we can parse a
 497          * --filter option successfully.
 498          */
 499         revs.tree_objects = revs.blob_objects = 1;
 500
 501         argc = setup_revisions(argc, argv, &revs, NULL);
 502
 503         /*
 504          * Reasons to require version 3:
 505          *
 506          * 1. @object-format is required because our hash algorithm is not
 507          *    SHA1.
 508          * 2. @filter is required because we parsed an object filter.
 509          */
 510         if (the_hash_algo != &hash_algos[GIT_HASH_SHA1] || revs.filter.choice)
 511                 min_version = 3;
 512
 513         if (argc > 1) {
 514                 ret = error(_("unrecognized argument: %s"), argv[1]);
 515                 goto out;
 516         }
 517
 518         bundle_to_stdout = !strcmp(path, "-");
 519         if (bundle_to_stdout)
 520                 bundle_fd = 1;
 521         else
 522                 bundle_fd = hold_lock_file_for_update(&lock, path,
 523                                                       LOCK_DIE_ON_ERROR);
 524
 525         if (version == -1)
 526                 version = min_version;
 527
 528         if (version < 2 || version > 3) {
 529                 die(_("unsupported bundle version %d"), version);
 530         } else if (version < min_version) {
 531                 die(_("cannot write bundle version %d with algorithm %s"), version, the_hash_algo->name);
 532         } else if (version == 2) {
 533                 write_or_die(bundle_fd, v2_bundle_signature, strlen(v2_bundle_signature));
 534         } else {
 535                 const char *capability = "@object-format=";
 536                 write_or_die(bundle_fd, v3_bundle_signature, strlen(v3_bundle_signature));
 537                 write_or_die(bundle_fd, capability, strlen(capability));
 538                 write_or_die(bundle_fd, the_hash_algo->name, strlen(the_hash_algo->name));
 539                 write_or_die(bundle_fd, "\n", 1);
 540
 541                 if (revs.filter.choice) {
 542                         const char *value = expand_list_objects_filter_spec(&revs.filter);
 543                         capability = "@filter=";
 544                         write_or_die(bundle_fd, capability, strlen(capability));
 545                         write_or_die(bundle_fd, value, strlen(value));
 546                         write_or_die(bundle_fd, "\n", 1);
 547                 }
 548         }
 549
 550         /* save revs.pending in revs_copy for later use */
 551         memcpy(&revs_copy, &revs, sizeof(revs));
 552         revs_copy.pending.nr = 0;
 553         revs_copy.pending.alloc = 0;
 554         revs_copy.pending.objects = NULL;
 555         for (i = 0; i < revs.pending.nr; i++) {
 556                 struct object_array_entry *e = revs.pending.objects + i;
 557                 if (e)
 558                         add_object_array_with_path(e->item, e->name,
 559                                                    &revs_copy.pending,
 560                                                    e->mode, e->path);
 561         }
 562
 563         /* write prerequisites */
 564         revs.boundary = 1;
 565         if (prepare_revision_walk(&revs))
 566                 die("revision walk setup failed");
 567         bpi.fd = bundle_fd;
 568         bpi.pending = &revs_copy.pending;
 569
 570         /*
 571          * Remove any object walking here. We only care about commits and
 572          * tags here. The revs_copy has the right instances of these values.
 573          */
 574         revs.blob_objects = revs.tree_objects = 0;
 575         traverse_commit_list(&revs, write_bundle_prerequisites, NULL, &bpi);
 576
 577         /* write bundle refs */
 578         ref_count = write_bundle_refs(bundle_fd, &revs_copy);
 579         if (!ref_count) {
 580                 die(_("Refusing to create empty bundle."));
 581         } else if (ref_count < 0) {
 582                 ret = -1;
 583                 goto out;
 584         }
 585
 586         /* write pack */
 587         if (write_pack_data(bundle_fd, &revs_copy, pack_options)) {
 588                 ret = -1;
 589                 goto out;
 590         }
 591
 592         if (!bundle_to_stdout) {
 593                 if (commit_lock_file(&lock))
 594                         die_errno(_("cannot create '%s'"), path);
 595         }
 596
 597         ret = 0;
 598
 599 out:
 600         object_array_clear(&revs_copy.pending);
 601         release_revisions(&revs);
 602         rollback_lock_file(&lock);
 603         return ret;
 604 }
 605
 606 int unbundle(struct repository *r, struct bundle_header *header,
 607              int bundle_fd, struct strvec *extra_index_pack_args,
 608              struct unbundle_opts *opts)
 609 {
 610         struct child_process ip = CHILD_PROCESS_INIT;
 611         struct unbundle_opts opts_fallback = { 0 };
 612
 613         if (!opts)
 614                 opts = &opts_fallback;
 615
 616         if (verify_bundle(r, header, opts->flags)) {
 617                 close(bundle_fd);
 618                 return -1;
 619         }
 620
 621         strvec_pushl(&ip.args, "index-pack", "--fix-thin", "--stdin", NULL);
 622
 623         /* If there is a filter, then we need to create the promisor pack. */
 624         if (header->filter.choice)
 625                 strvec_push(&ip.args, "--promisor=from-bundle");
 626
 627         if (opts->flags & VERIFY_BUNDLE_FSCK)
 628                 strvec_pushf(&ip.args, "--fsck-objects%s",
 629                              opts->fsck_msg_types ? opts->fsck_msg_types : "");
 630
 631         if (extra_index_pack_args)
 632                 strvec_pushv(&ip.args, extra_index_pack_args->v);
 633
 634         ip.in = bundle_fd;
 635         ip.no_stdout = 1;
 636         ip.git_cmd = 1;
 637         if (run_command(&ip))
 638                 return error(_("index-pack died"));
 639         return 0;
 640 }