builtin/repack.c

   1 #include "builtin.h"
   2 #include "cache.h"
   3 #include "config.h"
   4 #include "dir.h"
   5 #include "parse-options.h"
   6 #include "run-command.h"
   7 #include "sigchain.h"
   8 #include "strbuf.h"
   9 #include "string-list.h"
  10 #include "strvec.h"
  11 #include "midx.h"
  12 #include "packfile.h"
  13 #include "prune-packed.h"
  14 #include "object-store.h"
  15 #include "promisor-remote.h"
  16 #include "shallow.h"
  17 #include "pack.h"
  18 #include "pack-bitmap.h"
  19 #include "refs.h"
  20
  21 #define ALL_INTO_ONE 1
  22 #define LOOSEN_UNREACHABLE 2
  23 #define PACK_CRUFT 4
  24
  25 #define DELETE_PACK 1
  26 #define CRUFT_PACK 2
  27
  28 static int pack_everything;
  29 static int delta_base_offset = 1;
  30 static int pack_kept_objects = -1;
  31 static int write_bitmaps = -1;
  32 static int use_delta_islands;
  33 static int run_update_server_info = 1;
  34 static char *packdir, *packtmp_name, *packtmp;
  35 static char *cruft_expiration;
  36
  37 static const char *const git_repack_usage[] = {
  38         N_("git repack [<options>]"),
  39         NULL
  40 };
  41
  42 static const char incremental_bitmap_conflict_error[] = N_(
  43 "Incremental repacks are incompatible with bitmap indexes.  Use\n"
  44 "--no-write-bitmap-index or disable the pack.writeBitmaps configuration."
  45 );
  46
  47 struct pack_objects_args {
  48         const char *window;
  49         const char *window_memory;
  50         const char *depth;
  51         const char *threads;
  52         const char *max_pack_size;
  53         int no_reuse_delta;
  54         int no_reuse_object;
  55         int quiet;
  56         int local;
  57 };
  58
  59 static int repack_config(const char *var, const char *value, void *cb)
  60 {
  61         struct pack_objects_args *cruft_po_args = cb;
  62         if (!strcmp(var, "repack.usedeltabaseoffset")) {
  63                 delta_base_offset = git_config_bool(var, value);
  64                 return 0;
  65         }
  66         if (!strcmp(var, "repack.packkeptobjects")) {
  67                 pack_kept_objects = git_config_bool(var, value);
  68                 return 0;
  69         }
  70         if (!strcmp(var, "repack.writebitmaps") ||
  71             !strcmp(var, "pack.writebitmaps")) {
  72                 write_bitmaps = git_config_bool(var, value);
  73                 return 0;
  74         }
  75         if (!strcmp(var, "repack.usedeltaislands")) {
  76                 use_delta_islands = git_config_bool(var, value);
  77                 return 0;
  78         }
  79         if (strcmp(var, "repack.updateserverinfo") == 0) {
  80                 run_update_server_info = git_config_bool(var, value);
  81                 return 0;
  82         }
  83         if (!strcmp(var, "repack.cruftwindow"))
  84                 return git_config_string(&cruft_po_args->window, var, value);
  85         if (!strcmp(var, "repack.cruftwindowmemory"))
  86                 return git_config_string(&cruft_po_args->window_memory, var, value);
  87         if (!strcmp(var, "repack.cruftdepth"))
  88                 return git_config_string(&cruft_po_args->depth, var, value);
  89         if (!strcmp(var, "repack.cruftthreads"))
  90                 return git_config_string(&cruft_po_args->threads, var, value);
  91         return git_default_config(var, value, cb);
  92 }
  93
  94 /*
  95  * Remove temporary $GIT_OBJECT_DIRECTORY/pack/.tmp-$$-pack-* files.
  96  */
  97 static void remove_temporary_files(void)
  98 {
  99         struct strbuf buf = STRBUF_INIT;
 100         size_t dirlen, prefixlen;
 101         DIR *dir;
 102         struct dirent *e;
 103
 104         dir = opendir(packdir);
 105         if (!dir)
 106                 return;
 107
 108         /* Point at the slash at the end of ".../objects/pack/" */
 109         dirlen = strlen(packdir) + 1;
 110         strbuf_addstr(&buf, packtmp);
 111         /* Hold the length of  ".tmp-%d-pack-" */
 112         prefixlen = buf.len - dirlen;
 113
 114         while ((e = readdir(dir))) {
 115                 if (strncmp(e->d_name, buf.buf + dirlen, prefixlen))
 116                         continue;
 117                 strbuf_setlen(&buf, dirlen);
 118                 strbuf_addstr(&buf, e->d_name);
 119                 unlink(buf.buf);
 120         }
 121         closedir(dir);
 122         strbuf_release(&buf);
 123 }
 124
 125 static void remove_pack_on_signal(int signo)
 126 {
 127         remove_temporary_files();
 128         sigchain_pop(signo);
 129         raise(signo);
 130 }
 131
 132 /*
 133  * Adds all packs hex strings to either fname_nonkept_list or
 134  * fname_kept_list based on whether each pack has a corresponding
 135  * .keep file or not.  Packs without a .keep file are not to be kept
 136  * if we are going to pack everything into one file.
 137  */
 138 static void collect_pack_filenames(struct string_list *fname_nonkept_list,
 139                                    struct string_list *fname_kept_list,
 140                                    const struct string_list *extra_keep)
 141 {
 142         DIR *dir;
 143         struct dirent *e;
 144         char *fname;
 145
 146         if (!(dir = opendir(packdir)))
 147                 return;
 148
 149         while ((e = readdir(dir)) != NULL) {
 150                 size_t len;
 151                 int i;
 152
 153                 if (!strip_suffix(e->d_name, ".pack", &len))
 154                         continue;
 155
 156                 for (i = 0; i < extra_keep->nr; i++)
 157                         if (!fspathcmp(e->d_name, extra_keep->items[i].string))
 158                                 break;
 159
 160                 fname = xmemdupz(e->d_name, len);
 161
 162                 if ((extra_keep->nr > 0 && i < extra_keep->nr) ||
 163                     (file_exists(mkpath("%s/%s.keep", packdir, fname)))) {
 164                         string_list_append_nodup(fname_kept_list, fname);
 165                 } else {
 166                         struct string_list_item *item;
 167                         item = string_list_append_nodup(fname_nonkept_list,
 168                                                         fname);
 169                         if (file_exists(mkpath("%s/%s.mtimes", packdir, fname)))
 170                                 item->util = (void*)(uintptr_t)CRUFT_PACK;
 171                 }
 172         }
 173         closedir(dir);
 174
 175         string_list_sort(fname_kept_list);
 176 }
 177
 178 static void remove_redundant_pack(const char *dir_name, const char *base_name)
 179 {
 180         struct strbuf buf = STRBUF_INIT;
 181         struct multi_pack_index *m = get_local_multi_pack_index(the_repository);
 182         strbuf_addf(&buf, "%s.pack", base_name);
 183         if (m && midx_contains_pack(m, buf.buf))
 184                 clear_midx_file(the_repository);
 185         strbuf_insertf(&buf, 0, "%s/", dir_name);
 186         unlink_pack_path(buf.buf, 1);
 187         strbuf_release(&buf);
 188 }
 189
 190 static void prepare_pack_objects(struct child_process *cmd,
 191                                  const struct pack_objects_args *args)
 192 {
 193         strvec_push(&cmd->args, "pack-objects");
 194         if (args->window)
 195                 strvec_pushf(&cmd->args, "--window=%s", args->window);
 196         if (args->window_memory)
 197                 strvec_pushf(&cmd->args, "--window-memory=%s", args->window_memory);
 198         if (args->depth)
 199                 strvec_pushf(&cmd->args, "--depth=%s", args->depth);
 200         if (args->threads)
 201                 strvec_pushf(&cmd->args, "--threads=%s", args->threads);
 202         if (args->max_pack_size)
 203                 strvec_pushf(&cmd->args, "--max-pack-size=%s", args->max_pack_size);
 204         if (args->no_reuse_delta)
 205                 strvec_pushf(&cmd->args, "--no-reuse-delta");
 206         if (args->no_reuse_object)
 207                 strvec_pushf(&cmd->args, "--no-reuse-object");
 208         if (args->local)
 209                 strvec_push(&cmd->args,  "--local");
 210         if (args->quiet)
 211                 strvec_push(&cmd->args,  "--quiet");
 212         if (delta_base_offset)
 213                 strvec_push(&cmd->args,  "--delta-base-offset");
 214         strvec_push(&cmd->args, packtmp);
 215         cmd->git_cmd = 1;
 216         cmd->out = -1;
 217 }
 218
 219 /*
 220  * Write oid to the given struct child_process's stdin, starting it first if
 221  * necessary.
 222  */
 223 static int write_oid(const struct object_id *oid, struct packed_git *pack,
 224                      uint32_t pos, void *data)
 225 {
 226         struct child_process *cmd = data;
 227
 228         if (cmd->in == -1) {
 229                 if (start_command(cmd))
 230                         die(_("could not start pack-objects to repack promisor objects"));
 231         }
 232
 233         xwrite(cmd->in, oid_to_hex(oid), the_hash_algo->hexsz);
 234         xwrite(cmd->in, "\n", 1);
 235         return 0;
 236 }
 237
 238 static struct {
 239         const char *name;
 240         unsigned optional:1;
 241 } exts[] = {
 242         {".pack"},
 243         {".rev", 1},
 244         {".mtimes", 1},
 245         {".bitmap", 1},
 246         {".promisor", 1},
 247         {".idx"},
 248 };
 249
 250 static unsigned populate_pack_exts(char *name)
 251 {
 252         struct stat statbuf;
 253         struct strbuf path = STRBUF_INIT;
 254         unsigned ret = 0;
 255         int i;
 256
 257         for (i = 0; i < ARRAY_SIZE(exts); i++) {
 258                 strbuf_reset(&path);
 259                 strbuf_addf(&path, "%s-%s%s", packtmp, name, exts[i].name);
 260
 261                 if (stat(path.buf, &statbuf))
 262                         continue;
 263
 264                 ret |= (1 << i);
 265         }
 266
 267         strbuf_release(&path);
 268         return ret;
 269 }
 270
 271 static void repack_promisor_objects(const struct pack_objects_args *args,
 272                                     struct string_list *names)
 273 {
 274         struct child_process cmd = CHILD_PROCESS_INIT;
 275         FILE *out;
 276         struct strbuf line = STRBUF_INIT;
 277
 278         prepare_pack_objects(&cmd, args);
 279         cmd.in = -1;
 280
 281         /*
 282          * NEEDSWORK: Giving pack-objects only the OIDs without any ordering
 283          * hints may result in suboptimal deltas in the resulting pack. See if
 284          * the OIDs can be sent with fake paths such that pack-objects can use a
 285          * {type -> existing pack order} ordering when computing deltas instead
 286          * of a {type -> size} ordering, which may produce better deltas.
 287          */
 288         for_each_packed_object(write_oid, &cmd,
 289                                FOR_EACH_OBJECT_PROMISOR_ONLY);
 290
 291         if (cmd.in == -1) {
 292                 /* No packed objects; cmd was never started */
 293                 child_process_clear(&cmd);
 294                 return;
 295         }
 296
 297         close(cmd.in);
 298
 299         out = xfdopen(cmd.out, "r");
 300         while (strbuf_getline_lf(&line, out) != EOF) {
 301                 struct string_list_item *item;
 302                 char *promisor_name;
 303
 304                 if (line.len != the_hash_algo->hexsz)
 305                         die(_("repack: Expecting full hex object ID lines only from pack-objects."));
 306                 item = string_list_append(names, line.buf);
 307
 308                 /*
 309                  * pack-objects creates the .pack and .idx files, but not the
 310                  * .promisor file. Create the .promisor file, which is empty.
 311                  *
 312                  * NEEDSWORK: fetch-pack sometimes generates non-empty
 313                  * .promisor files containing the ref names and associated
 314                  * hashes at the point of generation of the corresponding
 315                  * packfile, but this would not preserve their contents. Maybe
 316                  * concatenate the contents of all .promisor files instead of
 317                  * just creating a new empty file.
 318                  */
 319                 promisor_name = mkpathdup("%s-%s.promisor", packtmp,
 320                                           line.buf);
 321                 write_promisor_file(promisor_name, NULL, 0);
 322
 323                 item->util = (void *)(uintptr_t)populate_pack_exts(item->string);
 324
 325                 free(promisor_name);
 326         }
 327         fclose(out);
 328         if (finish_command(&cmd))
 329                 die(_("could not finish pack-objects to repack promisor objects"));
 330 }
 331
 332 struct pack_geometry {
 333         struct packed_git **pack;
 334         uint32_t pack_nr, pack_alloc;
 335         uint32_t split;
 336 };
 337
 338 static uint32_t geometry_pack_weight(struct packed_git *p)
 339 {
 340         if (open_pack_index(p))
 341                 die(_("cannot open index for %s"), p->pack_name);
 342         return p->num_objects;
 343 }
 344
 345 static int geometry_cmp(const void *va, const void *vb)
 346 {
 347         uint32_t aw = geometry_pack_weight(*(struct packed_git **)va),
 348                  bw = geometry_pack_weight(*(struct packed_git **)vb);
 349
 350         if (aw < bw)
 351                 return -1;
 352         if (aw > bw)
 353                 return 1;
 354         return 0;
 355 }
 356
 357 static void init_pack_geometry(struct pack_geometry **geometry_p,
 358                                struct string_list *existing_kept_packs)
 359 {
 360         struct packed_git *p;
 361         struct pack_geometry *geometry;
 362         struct strbuf buf = STRBUF_INIT;
 363
 364         *geometry_p = xcalloc(1, sizeof(struct pack_geometry));
 365         geometry = *geometry_p;
 366
 367         for (p = get_all_packs(the_repository); p; p = p->next) {
 368                 if (!pack_kept_objects) {
 369                         /*
 370                          * Any pack that has its pack_keep bit set will appear
 371                          * in existing_kept_packs below, but this saves us from
 372                          * doing a more expensive check.
 373                          */
 374                         if (p->pack_keep)
 375                                 continue;
 376
 377                         /*
 378                          * The pack may be kept via the --keep-pack option;
 379                          * check 'existing_kept_packs' to determine whether to
 380                          * ignore it.
 381                          */
 382                         strbuf_reset(&buf);
 383                         strbuf_addstr(&buf, pack_basename(p));
 384                         strbuf_strip_suffix(&buf, ".pack");
 385
 386                         if (string_list_has_string(existing_kept_packs, buf.buf))
 387                                 continue;
 388                 }
 389                 if (p->is_cruft)
 390                         continue;
 391
 392                 ALLOC_GROW(geometry->pack,
 393                            geometry->pack_nr + 1,
 394                            geometry->pack_alloc);
 395
 396                 geometry->pack[geometry->pack_nr] = p;
 397                 geometry->pack_nr++;
 398         }
 399
 400         QSORT(geometry->pack, geometry->pack_nr, geometry_cmp);
 401         strbuf_release(&buf);
 402 }
 403
 404 static void split_pack_geometry(struct pack_geometry *geometry, int factor)
 405 {
 406         uint32_t i;
 407         uint32_t split;
 408         off_t total_size = 0;
 409
 410         if (!geometry->pack_nr) {
 411                 geometry->split = geometry->pack_nr;
 412                 return;
 413         }
 414
 415         /*
 416          * First, count the number of packs (in descending order of size) which
 417          * already form a geometric progression.
 418          */
 419         for (i = geometry->pack_nr - 1; i > 0; i--) {
 420                 struct packed_git *ours = geometry->pack[i];
 421                 struct packed_git *prev = geometry->pack[i - 1];
 422
 423                 if (unsigned_mult_overflows(factor, geometry_pack_weight(prev)))
 424                         die(_("pack %s too large to consider in geometric "
 425                               "progression"),
 426                             prev->pack_name);
 427
 428                 if (geometry_pack_weight(ours) < factor * geometry_pack_weight(prev))
 429                         break;
 430         }
 431
 432         split = i;
 433
 434         if (split) {
 435                 /*
 436                  * Move the split one to the right, since the top element in the
 437                  * last-compared pair can't be in the progression. Only do this
 438                  * when we split in the middle of the array (otherwise if we got
 439                  * to the end, then the split is in the right place).
 440                  */
 441                 split++;
 442         }
 443
 444         /*
 445          * Then, anything to the left of 'split' must be in a new pack. But,
 446          * creating that new pack may cause packs in the heavy half to no longer
 447          * form a geometric progression.
 448          *
 449          * Compute an expected size of the new pack, and then determine how many
 450          * packs in the heavy half need to be joined into it (if any) to restore
 451          * the geometric progression.
 452          */
 453         for (i = 0; i < split; i++) {
 454                 struct packed_git *p = geometry->pack[i];
 455
 456                 if (unsigned_add_overflows(total_size, geometry_pack_weight(p)))
 457                         die(_("pack %s too large to roll up"), p->pack_name);
 458                 total_size += geometry_pack_weight(p);
 459         }
 460         for (i = split; i < geometry->pack_nr; i++) {
 461                 struct packed_git *ours = geometry->pack[i];
 462
 463                 if (unsigned_mult_overflows(factor, total_size))
 464                         die(_("pack %s too large to roll up"), ours->pack_name);
 465
 466                 if (geometry_pack_weight(ours) < factor * total_size) {
 467                         if (unsigned_add_overflows(total_size,
 468                                                    geometry_pack_weight(ours)))
 469                                 die(_("pack %s too large to roll up"),
 470                                     ours->pack_name);
 471
 472                         split++;
 473                         total_size += geometry_pack_weight(ours);
 474                 } else
 475                         break;
 476         }
 477
 478         geometry->split = split;
 479 }
 480
 481 static struct packed_git *get_largest_active_pack(struct pack_geometry *geometry)
 482 {
 483         if (!geometry) {
 484                 /*
 485                  * No geometry means either an all-into-one repack (in which
 486                  * case there is only one pack left and it is the largest) or an
 487                  * incremental one.
 488                  *
 489                  * If repacking incrementally, then we could check the size of
 490                  * all packs to determine which should be preferred, but leave
 491                  * this for later.
 492                  */
 493                 return NULL;
 494         }
 495         if (geometry->split == geometry->pack_nr)
 496                 return NULL;
 497         return geometry->pack[geometry->pack_nr - 1];
 498 }
 499
 500 static void clear_pack_geometry(struct pack_geometry *geometry)
 501 {
 502         if (!geometry)
 503                 return;
 504
 505         free(geometry->pack);
 506         geometry->pack_nr = 0;
 507         geometry->pack_alloc = 0;
 508         geometry->split = 0;
 509 }
 510
 511 struct midx_snapshot_ref_data {
 512         struct tempfile *f;
 513         struct oidset seen;
 514         int preferred;
 515 };
 516
 517 static int midx_snapshot_ref_one(const char *refname,
 518                                  const struct object_id *oid,
 519                                  int flag, void *_data)
 520 {
 521         struct midx_snapshot_ref_data *data = _data;
 522         struct object_id peeled;
 523
 524         if (!peel_iterated_oid(oid, &peeled))
 525                 oid = &peeled;
 526
 527         if (oidset_insert(&data->seen, oid))
 528                 return 0; /* already seen */
 529
 530         if (oid_object_info(the_repository, oid, NULL) != OBJ_COMMIT)
 531                 return 0;
 532
 533         fprintf(data->f->fp, "%s%s\n", data->preferred ? "+" : "",
 534                 oid_to_hex(oid));
 535
 536         return 0;
 537 }
 538
 539 static void midx_snapshot_refs(struct tempfile *f)
 540 {
 541         struct midx_snapshot_ref_data data;
 542         const struct string_list *preferred = bitmap_preferred_tips(the_repository);
 543
 544         data.f = f;
 545         data.preferred = 0;
 546         oidset_init(&data.seen, 0);
 547
 548         if (!fdopen_tempfile(f, "w"))
 549                  die(_("could not open tempfile %s for writing"),
 550                      get_tempfile_path(f));
 551
 552         if (preferred) {
 553                 struct string_list_item *item;
 554
 555                 data.preferred = 1;
 556                 for_each_string_list_item(item, preferred)
 557                         for_each_ref_in(item->string, midx_snapshot_ref_one, &data);
 558                 data.preferred = 0;
 559         }
 560
 561         for_each_ref(midx_snapshot_ref_one, &data);
 562
 563         if (close_tempfile_gently(f)) {
 564                 int save_errno = errno;
 565                 delete_tempfile(&f);
 566                 errno = save_errno;
 567                 die_errno(_("could not close refs snapshot tempfile"));
 568         }
 569
 570         oidset_clear(&data.seen);
 571 }
 572
 573 static void midx_included_packs(struct string_list *include,
 574                                 struct string_list *existing_nonkept_packs,
 575                                 struct string_list *existing_kept_packs,
 576                                 struct string_list *names,
 577                                 struct pack_geometry *geometry)
 578 {
 579         struct string_list_item *item;
 580
 581         for_each_string_list_item(item, existing_kept_packs)
 582                 string_list_insert(include, xstrfmt("%s.idx", item->string));
 583         for_each_string_list_item(item, names)
 584                 string_list_insert(include, xstrfmt("pack-%s.idx", item->string));
 585         if (geometry) {
 586                 struct strbuf buf = STRBUF_INIT;
 587                 uint32_t i;
 588                 for (i = geometry->split; i < geometry->pack_nr; i++) {
 589                         struct packed_git *p = geometry->pack[i];
 590
 591                         strbuf_addstr(&buf, pack_basename(p));
 592                         strbuf_strip_suffix(&buf, ".pack");
 593                         strbuf_addstr(&buf, ".idx");
 594
 595                         string_list_insert(include, strbuf_detach(&buf, NULL));
 596                 }
 597
 598                 for_each_string_list_item(item, existing_nonkept_packs) {
 599                         if (!((uintptr_t)item->util & CRUFT_PACK)) {
 600                                 /*
 601                                  * no need to check DELETE_PACK, since we're not
 602                                  * doing an ALL_INTO_ONE repack
 603                                  */
 604                                 continue;
 605                         }
 606                         string_list_insert(include, xstrfmt("%s.idx", item->string));
 607                 }
 608         } else {
 609                 for_each_string_list_item(item, existing_nonkept_packs) {
 610                         if ((uintptr_t)item->util & DELETE_PACK)
 611                                 continue;
 612                         string_list_insert(include, xstrfmt("%s.idx", item->string));
 613                 }
 614         }
 615 }
 616
 617 static int write_midx_included_packs(struct string_list *include,
 618                                      struct pack_geometry *geometry,
 619                                      const char *refs_snapshot,
 620                                      int show_progress, int write_bitmaps)
 621 {
 622         struct child_process cmd = CHILD_PROCESS_INIT;
 623         struct string_list_item *item;
 624         struct packed_git *largest = get_largest_active_pack(geometry);
 625         FILE *in;
 626         int ret;
 627
 628         if (!include->nr)
 629                 return 0;
 630
 631         cmd.in = -1;
 632         cmd.git_cmd = 1;
 633
 634         strvec_push(&cmd.args, "multi-pack-index");
 635         strvec_pushl(&cmd.args, "write", "--stdin-packs", NULL);
 636
 637         if (show_progress)
 638                 strvec_push(&cmd.args, "--progress");
 639         else
 640                 strvec_push(&cmd.args, "--no-progress");
 641
 642         if (write_bitmaps)
 643                 strvec_push(&cmd.args, "--bitmap");
 644
 645         if (largest)
 646                 strvec_pushf(&cmd.args, "--preferred-pack=%s",
 647                              pack_basename(largest));
 648
 649         if (refs_snapshot)
 650                 strvec_pushf(&cmd.args, "--refs-snapshot=%s", refs_snapshot);
 651
 652         ret = start_command(&cmd);
 653         if (ret)
 654                 return ret;
 655
 656         in = xfdopen(cmd.in, "w");
 657         for_each_string_list_item(item, include)
 658                 fprintf(in, "%s\n", item->string);
 659         fclose(in);
 660
 661         return finish_command(&cmd);
 662 }
 663
 664 static int write_cruft_pack(const struct pack_objects_args *args,
 665                             const char *pack_prefix,
 666                             struct string_list *names,
 667                             struct string_list *existing_packs,
 668                             struct string_list *existing_kept_packs)
 669 {
 670         struct child_process cmd = CHILD_PROCESS_INIT;
 671         struct strbuf line = STRBUF_INIT;
 672         struct string_list_item *item;
 673         FILE *in, *out;
 674         int ret;
 675
 676         prepare_pack_objects(&cmd, args);
 677
 678         strvec_push(&cmd.args, "--cruft");
 679         if (cruft_expiration)
 680                 strvec_pushf(&cmd.args, "--cruft-expiration=%s",
 681                              cruft_expiration);
 682
 683         strvec_push(&cmd.args, "--honor-pack-keep");
 684         strvec_push(&cmd.args, "--non-empty");
 685         strvec_push(&cmd.args, "--max-pack-size=0");
 686
 687         cmd.in = -1;
 688
 689         ret = start_command(&cmd);
 690         if (ret)
 691                 return ret;
 692
 693         /*
 694          * names has a confusing double use: it both provides the list
 695          * of just-written new packs, and accepts the name of the cruft
 696          * pack we are writing.
 697          *
 698          * By the time it is read here, it contains only the pack(s)
 699          * that were just written, which is exactly the set of packs we
 700          * want to consider kept.
 701          */
 702         in = xfdopen(cmd.in, "w");
 703         for_each_string_list_item(item, names)
 704                 fprintf(in, "%s-%s.pack\n", pack_prefix, item->string);
 705         for_each_string_list_item(item, existing_packs)
 706                 fprintf(in, "-%s.pack\n", item->string);
 707         for_each_string_list_item(item, existing_kept_packs)
 708                 fprintf(in, "%s.pack\n", item->string);
 709         fclose(in);
 710
 711         out = xfdopen(cmd.out, "r");
 712         while (strbuf_getline_lf(&line, out) != EOF) {
 713                 if (line.len != the_hash_algo->hexsz)
 714                         die(_("repack: Expecting full hex object ID lines only "
 715                               "from pack-objects."));
 716                 string_list_append(names, line.buf);
 717         }
 718         fclose(out);
 719
 720         strbuf_release(&line);
 721
 722         return finish_command(&cmd);
 723 }
 724
 725 int cmd_repack(int argc, const char **argv, const char *prefix)
 726 {
 727         struct child_process cmd = CHILD_PROCESS_INIT;
 728         struct string_list_item *item;
 729         struct string_list names = STRING_LIST_INIT_DUP;
 730         struct string_list existing_nonkept_packs = STRING_LIST_INIT_DUP;
 731         struct string_list existing_kept_packs = STRING_LIST_INIT_DUP;
 732         struct pack_geometry *geometry = NULL;
 733         struct strbuf line = STRBUF_INIT;
 734         struct tempfile *refs_snapshot = NULL;
 735         int i, ext, ret;
 736         FILE *out;
 737         int show_progress;
 738
 739         /* variables to be filled by option parsing */
 740         int delete_redundant = 0;
 741         const char *unpack_unreachable = NULL;
 742         int keep_unreachable = 0;
 743         struct string_list keep_pack_list = STRING_LIST_INIT_NODUP;
 744         struct pack_objects_args po_args = {NULL};
 745         struct pack_objects_args cruft_po_args = {NULL};
 746         int geometric_factor = 0;
 747         int write_midx = 0;
 748
 749         struct option builtin_repack_options[] = {
 750                 OPT_BIT('a', NULL, &pack_everything,
 751                                 N_("pack everything in a single pack"), ALL_INTO_ONE),
 752                 OPT_BIT('A', NULL, &pack_everything,
 753                                 N_("same as -a, and turn unreachable objects loose"),
 754                                    LOOSEN_UNREACHABLE | ALL_INTO_ONE),
 755                 OPT_BIT(0, "cruft", &pack_everything,
 756                                 N_("same as -a, pack unreachable cruft objects separately"),
 757                                    PACK_CRUFT),
 758                 OPT_STRING(0, "cruft-expiration", &cruft_expiration, N_("approxidate"),
 759                                 N_("with -C, expire objects older than this")),
 760                 OPT_BOOL('d', NULL, &delete_redundant,
 761                                 N_("remove redundant packs, and run git-prune-packed")),
 762                 OPT_BOOL('f', NULL, &po_args.no_reuse_delta,
 763                                 N_("pass --no-reuse-delta to git-pack-objects")),
 764                 OPT_BOOL('F', NULL, &po_args.no_reuse_object,
 765                                 N_("pass --no-reuse-object to git-pack-objects")),
 766                 OPT_NEGBIT('n', NULL, &run_update_server_info,
 767                                 N_("do not run git-update-server-info"), 1),
 768                 OPT__QUIET(&po_args.quiet, N_("be quiet")),
 769                 OPT_BOOL('l', "local", &po_args.local,
 770                                 N_("pass --local to git-pack-objects")),
 771                 OPT_BOOL('b', "write-bitmap-index", &write_bitmaps,
 772                                 N_("write bitmap index")),
 773                 OPT_BOOL('i', "delta-islands", &use_delta_islands,
 774                                 N_("pass --delta-islands to git-pack-objects")),
 775                 OPT_STRING(0, "unpack-unreachable", &unpack_unreachable, N_("approxidate"),
 776                                 N_("with -A, do not loosen objects older than this")),
 777                 OPT_BOOL('k', "keep-unreachable", &keep_unreachable,
 778                                 N_("with -a, repack unreachable objects")),
 779                 OPT_STRING(0, "window", &po_args.window, N_("n"),
 780                                 N_("size of the window used for delta compression")),
 781                 OPT_STRING(0, "window-memory", &po_args.window_memory, N_("bytes"),
 782                                 N_("same as the above, but limit memory size instead of entries count")),
 783                 OPT_STRING(0, "depth", &po_args.depth, N_("n"),
 784                                 N_("limits the maximum delta depth")),
 785                 OPT_STRING(0, "threads", &po_args.threads, N_("n"),
 786                                 N_("limits the maximum number of threads")),
 787                 OPT_STRING(0, "max-pack-size", &po_args.max_pack_size, N_("bytes"),
 788                                 N_("maximum size of each packfile")),
 789                 OPT_BOOL(0, "pack-kept-objects", &pack_kept_objects,
 790                                 N_("repack objects in packs marked with .keep")),
 791                 OPT_STRING_LIST(0, "keep-pack", &keep_pack_list, N_("name"),
 792                                 N_("do not repack this pack")),
 793                 OPT_INTEGER('g', "geometric", &geometric_factor,
 794                             N_("find a geometric progression with factor <N>")),
 795                 OPT_BOOL('m', "write-midx", &write_midx,
 796                            N_("write a multi-pack index of the resulting packs")),
 797                 OPT_END()
 798         };
 799
 800         git_config(repack_config, &cruft_po_args);
 801
 802         argc = parse_options(argc, argv, prefix, builtin_repack_options,
 803                                 git_repack_usage, 0);
 804
 805         if (delete_redundant && repository_format_precious_objects)
 806                 die(_("cannot delete packs in a precious-objects repo"));
 807
 808         if (keep_unreachable &&
 809             (unpack_unreachable || (pack_everything & LOOSEN_UNREACHABLE)))
 810                 die(_("options '%s' and '%s' cannot be used together"), "--keep-unreachable", "-A");
 811
 812         if (pack_everything & PACK_CRUFT) {
 813                 pack_everything |= ALL_INTO_ONE;
 814
 815                 if (unpack_unreachable || (pack_everything & LOOSEN_UNREACHABLE))
 816                         die(_("options '%s' and '%s' cannot be used together"), "--cruft", "-A");
 817                 if (keep_unreachable)
 818                         die(_("options '%s' and '%s' cannot be used together"), "--cruft", "-k");
 819         }
 820
 821         if (write_bitmaps < 0) {
 822                 if (!write_midx &&
 823                     (!(pack_everything & ALL_INTO_ONE) || !is_bare_repository()))
 824                         write_bitmaps = 0;
 825         } else if (write_bitmaps &&
 826                    git_env_bool(GIT_TEST_MULTI_PACK_INDEX, 0) &&
 827                    git_env_bool(GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP, 0)) {
 828                 write_bitmaps = 0;
 829         }
 830         if (pack_kept_objects < 0)
 831                 pack_kept_objects = write_bitmaps > 0 && !write_midx;
 832
 833         if (write_bitmaps && !(pack_everything & ALL_INTO_ONE) && !write_midx)
 834                 die(_(incremental_bitmap_conflict_error));
 835
 836         if (write_midx && write_bitmaps) {
 837                 struct strbuf path = STRBUF_INIT;
 838
 839                 strbuf_addf(&path, "%s/%s_XXXXXX", get_object_directory(),
 840                             "bitmap-ref-tips");
 841
 842                 refs_snapshot = xmks_tempfile(path.buf);
 843                 midx_snapshot_refs(refs_snapshot);
 844
 845                 strbuf_release(&path);
 846         }
 847
 848         packdir = mkpathdup("%s/pack", get_object_directory());
 849         packtmp_name = xstrfmt(".tmp-%d-pack", (int)getpid());
 850         packtmp = mkpathdup("%s/%s", packdir, packtmp_name);
 851
 852         collect_pack_filenames(&existing_nonkept_packs, &existing_kept_packs,
 853                                &keep_pack_list);
 854
 855         if (geometric_factor) {
 856                 if (pack_everything)
 857                         die(_("options '%s' and '%s' cannot be used together"), "--geometric", "-A/-a");
 858                 init_pack_geometry(&geometry, &existing_kept_packs);
 859                 split_pack_geometry(geometry, geometric_factor);
 860         }
 861
 862         sigchain_push_common(remove_pack_on_signal);
 863
 864         prepare_pack_objects(&cmd, &po_args);
 865
 866         show_progress = !po_args.quiet && isatty(2);
 867
 868         strvec_push(&cmd.args, "--keep-true-parents");
 869         if (!pack_kept_objects)
 870                 strvec_push(&cmd.args, "--honor-pack-keep");
 871         for (i = 0; i < keep_pack_list.nr; i++)
 872                 strvec_pushf(&cmd.args, "--keep-pack=%s",
 873                              keep_pack_list.items[i].string);
 874         strvec_push(&cmd.args, "--non-empty");
 875         if (!geometry) {
 876                 /*
 877                  * We need to grab all reachable objects, including those that
 878                  * are reachable from reflogs and the index.
 879                  *
 880                  * When repacking into a geometric progression of packs,
 881                  * however, we ask 'git pack-objects --stdin-packs', and it is
 882                  * not about packing objects based on reachability but about
 883                  * repacking all the objects in specified packs and loose ones
 884                  * (indeed, --stdin-packs is incompatible with these options).
 885                  */
 886                 strvec_push(&cmd.args, "--all");
 887                 strvec_push(&cmd.args, "--reflog");
 888                 strvec_push(&cmd.args, "--indexed-objects");
 889         }
 890         if (has_promisor_remote())
 891                 strvec_push(&cmd.args, "--exclude-promisor-objects");
 892         if (!write_midx) {
 893                 if (write_bitmaps > 0)
 894                         strvec_push(&cmd.args, "--write-bitmap-index");
 895                 else if (write_bitmaps < 0)
 896                         strvec_push(&cmd.args, "--write-bitmap-index-quiet");
 897         }
 898         if (use_delta_islands)
 899                 strvec_push(&cmd.args, "--delta-islands");
 900
 901         if (pack_everything & ALL_INTO_ONE) {
 902                 repack_promisor_objects(&po_args, &names);
 903
 904                 if (existing_nonkept_packs.nr && delete_redundant &&
 905                     !(pack_everything & PACK_CRUFT)) {
 906                         for_each_string_list_item(item, &names) {
 907                                 strvec_pushf(&cmd.args, "--keep-pack=%s-%s.pack",
 908                                              packtmp_name, item->string);
 909                         }
 910                         if (unpack_unreachable) {
 911                                 strvec_pushf(&cmd.args,
 912                                              "--unpack-unreachable=%s",
 913                                              unpack_unreachable);
 914                         } else if (pack_everything & LOOSEN_UNREACHABLE) {
 915                                 strvec_push(&cmd.args,
 916                                             "--unpack-unreachable");
 917                         } else if (keep_unreachable) {
 918                                 strvec_push(&cmd.args, "--keep-unreachable");
 919                                 strvec_push(&cmd.args, "--pack-loose-unreachable");
 920                         }
 921                 }
 922         } else if (geometry) {
 923                 strvec_push(&cmd.args, "--stdin-packs");
 924                 strvec_push(&cmd.args, "--unpacked");
 925         } else {
 926                 strvec_push(&cmd.args, "--unpacked");
 927                 strvec_push(&cmd.args, "--incremental");
 928         }
 929
 930         if (geometry)
 931                 cmd.in = -1;
 932         else
 933                 cmd.no_stdin = 1;
 934
 935         ret = start_command(&cmd);
 936         if (ret)
 937                 return ret;
 938
 939         if (geometry) {
 940                 FILE *in = xfdopen(cmd.in, "w");
 941                 /*
 942                  * The resulting pack should contain all objects in packs that
 943                  * are going to be rolled up, but exclude objects in packs which
 944                  * are being left alone.
 945                  */
 946                 for (i = 0; i < geometry->split; i++)
 947                         fprintf(in, "%s\n", pack_basename(geometry->pack[i]));
 948                 for (i = geometry->split; i < geometry->pack_nr; i++)
 949                         fprintf(in, "^%s\n", pack_basename(geometry->pack[i]));
 950                 fclose(in);
 951         }
 952
 953         out = xfdopen(cmd.out, "r");
 954         while (strbuf_getline_lf(&line, out) != EOF) {
 955                 if (line.len != the_hash_algo->hexsz)
 956                         die(_("repack: Expecting full hex object ID lines only from pack-objects."));
 957                 string_list_append(&names, line.buf);
 958         }
 959         fclose(out);
 960         ret = finish_command(&cmd);
 961         if (ret)
 962                 return ret;
 963
 964         if (!names.nr && !po_args.quiet)
 965                 printf_ln(_("Nothing new to pack."));
 966
 967         if (pack_everything & PACK_CRUFT) {
 968                 const char *pack_prefix;
 969                 if (!skip_prefix(packtmp, packdir, &pack_prefix))
 970                         die(_("pack prefix %s does not begin with objdir %s"),
 971                             packtmp, packdir);
 972                 if (*pack_prefix == '/')
 973                         pack_prefix++;
 974
 975                 if (!cruft_po_args.window)
 976                         cruft_po_args.window = po_args.window;
 977                 if (!cruft_po_args.window_memory)
 978                         cruft_po_args.window_memory = po_args.window_memory;
 979                 if (!cruft_po_args.depth)
 980                         cruft_po_args.depth = po_args.depth;
 981                 if (!cruft_po_args.threads)
 982                         cruft_po_args.threads = po_args.threads;
 983
 984                 cruft_po_args.local = po_args.local;
 985                 cruft_po_args.quiet = po_args.quiet;
 986
 987                 ret = write_cruft_pack(&cruft_po_args, pack_prefix, &names,
 988                                        &existing_nonkept_packs,
 989                                        &existing_kept_packs);
 990                 if (ret)
 991                         return ret;
 992         }
 993
 994         string_list_sort(&names);
 995
 996         for_each_string_list_item(item, &names) {
 997                 item->util = (void *)(uintptr_t)populate_pack_exts(item->string);
 998         }
 999
1000         close_object_store(the_repository->objects);
1001
1002         /*
1003          * Ok we have prepared all new packfiles.
1004          */
1005         for_each_string_list_item(item, &names) {
1006                 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
1007                         char *fname, *fname_old;
1008
1009                         fname = mkpathdup("%s/pack-%s%s",
1010                                         packdir, item->string, exts[ext].name);
1011                         fname_old = mkpathdup("%s-%s%s",
1012                                         packtmp, item->string, exts[ext].name);
1013
1014                         if (((uintptr_t)item->util) & ((uintptr_t)1 << ext)) {
1015                                 struct stat statbuffer;
1016                                 if (!stat(fname_old, &statbuffer)) {
1017                                         statbuffer.st_mode &= ~(S_IWUSR | S_IWGRP | S_IWOTH);
1018                                         chmod(fname_old, statbuffer.st_mode);
1019                                 }
1020
1021                                 if (rename(fname_old, fname))
1022                                         die_errno(_("renaming '%s' failed"), fname_old);
1023                         } else if (!exts[ext].optional)
1024                                 die(_("missing required file: %s"), fname_old);
1025                         else if (unlink(fname) < 0 && errno != ENOENT)
1026                                 die_errno(_("could not unlink: %s"), fname);
1027
1028                         free(fname);
1029                         free(fname_old);
1030                 }
1031         }
1032         /* End of pack replacement. */
1033
1034         if (delete_redundant && pack_everything & ALL_INTO_ONE) {
1035                 const int hexsz = the_hash_algo->hexsz;
1036                 for_each_string_list_item(item, &existing_nonkept_packs) {
1037                         char *sha1;
1038                         size_t len = strlen(item->string);
1039                         if (len < hexsz)
1040                                 continue;
1041                         sha1 = item->string + len - hexsz;
1042                         /*
1043                          * Mark this pack for deletion, which ensures that this
1044                          * pack won't be included in a MIDX (if `--write-midx`
1045                          * was given) and that we will actually delete this pack
1046                          * (if `-d` was given).
1047                          */
1048                         if (!string_list_has_string(&names, sha1))
1049                                 item->util = (void*)(uintptr_t)((size_t)item->util | DELETE_PACK);
1050                 }
1051         }
1052
1053         if (write_midx) {
1054                 struct string_list include = STRING_LIST_INIT_NODUP;
1055                 midx_included_packs(&include, &existing_nonkept_packs,
1056                                     &existing_kept_packs, &names, geometry);
1057
1058                 ret = write_midx_included_packs(&include, geometry,
1059                                                 refs_snapshot ? get_tempfile_path(refs_snapshot) : NULL,
1060                                                 show_progress, write_bitmaps > 0);
1061
1062                 string_list_clear(&include, 0);
1063
1064                 if (ret)
1065                         return ret;
1066         }
1067
1068         reprepare_packed_git(the_repository);
1069
1070         if (delete_redundant) {
1071                 int opts = 0;
1072                 for_each_string_list_item(item, &existing_nonkept_packs) {
1073                         if (!((uintptr_t)item->util & DELETE_PACK))
1074                                 continue;
1075                         remove_redundant_pack(packdir, item->string);
1076                 }
1077
1078                 if (geometry) {
1079                         struct strbuf buf = STRBUF_INIT;
1080
1081                         uint32_t i;
1082                         for (i = 0; i < geometry->split; i++) {
1083                                 struct packed_git *p = geometry->pack[i];
1084                                 if (string_list_has_string(&names,
1085                                                            hash_to_hex(p->hash)))
1086                                         continue;
1087
1088                                 strbuf_reset(&buf);
1089                                 strbuf_addstr(&buf, pack_basename(p));
1090                                 strbuf_strip_suffix(&buf, ".pack");
1091
1092                                 remove_redundant_pack(packdir, buf.buf);
1093                         }
1094                         strbuf_release(&buf);
1095                 }
1096                 if (show_progress)
1097                         opts |= PRUNE_PACKED_VERBOSE;
1098                 prune_packed_objects(opts);
1099
1100                 if (!keep_unreachable &&
1101                     (!(pack_everything & LOOSEN_UNREACHABLE) ||
1102                      unpack_unreachable) &&
1103                     is_repository_shallow(the_repository))
1104                         prune_shallow(PRUNE_QUICK);
1105         }
1106
1107         if (run_update_server_info)
1108                 update_server_info(0);
1109         remove_temporary_files();
1110
1111         if (git_env_bool(GIT_TEST_MULTI_PACK_INDEX, 0)) {
1112                 unsigned flags = 0;
1113                 if (git_env_bool(GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP, 0))
1114                         flags |= MIDX_WRITE_BITMAP | MIDX_WRITE_REV_INDEX;
1115                 write_midx_file(get_object_directory(), NULL, NULL, flags);
1116         }
1117
1118         string_list_clear(&names, 0);
1119         string_list_clear(&existing_nonkept_packs, 0);
1120         string_list_clear(&existing_kept_packs, 0);
1121         clear_pack_geometry(geometry);
1122         strbuf_release(&line);
1123
1124         return 0;
1125 }