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
  36 static const char *const git_repack_usage[] = {
  37         N_("git repack [<options>]"),
  38         NULL
  39 };
  40
  41 static const char incremental_bitmap_conflict_error[] = N_(
  42 "Incremental repacks are incompatible with bitmap indexes.  Use\n"
  43 "--no-write-bitmap-index or disable the pack.writeBitmaps configuration."
  44 );
  45
  46 struct pack_objects_args {
  47         const char *window;
  48         const char *window_memory;
  49         const char *depth;
  50         const char *threads;
  51         const char *max_pack_size;
  52         int no_reuse_delta;
  53         int no_reuse_object;
  54         int quiet;
  55         int local;
  56 };
  57
  58 static int repack_config(const char *var, const char *value, void *cb)
  59 {
  60         struct pack_objects_args *cruft_po_args = cb;
  61         if (!strcmp(var, "repack.usedeltabaseoffset")) {
  62                 delta_base_offset = git_config_bool(var, value);
  63                 return 0;
  64         }
  65         if (!strcmp(var, "repack.packkeptobjects")) {
  66                 pack_kept_objects = git_config_bool(var, value);
  67                 return 0;
  68         }
  69         if (!strcmp(var, "repack.writebitmaps") ||
  70             !strcmp(var, "pack.writebitmaps")) {
  71                 write_bitmaps = git_config_bool(var, value);
  72                 return 0;
  73         }
  74         if (!strcmp(var, "repack.usedeltaislands")) {
  75                 use_delta_islands = git_config_bool(var, value);
  76                 return 0;
  77         }
  78         if (strcmp(var, "repack.updateserverinfo") == 0) {
  79                 run_update_server_info = git_config_bool(var, value);
  80                 return 0;
  81         }
  82         if (!strcmp(var, "repack.cruftwindow"))
  83                 return git_config_string(&cruft_po_args->window, var, value);
  84         if (!strcmp(var, "repack.cruftwindowmemory"))
  85                 return git_config_string(&cruft_po_args->window_memory, var, value);
  86         if (!strcmp(var, "repack.cruftdepth"))
  87                 return git_config_string(&cruft_po_args->depth, var, value);
  88         if (!strcmp(var, "repack.cruftthreads"))
  89                 return git_config_string(&cruft_po_args->threads, var, value);
  90         return git_default_config(var, value, cb);
  91 }
  92
  93 /*
  94  * Adds all packs hex strings to either fname_nonkept_list or
  95  * fname_kept_list based on whether each pack has a corresponding
  96  * .keep file or not.  Packs without a .keep file are not to be kept
  97  * if we are going to pack everything into one file.
  98  */
  99 static void collect_pack_filenames(struct string_list *fname_nonkept_list,
 100                                    struct string_list *fname_kept_list,
 101                                    const struct string_list *extra_keep)
 102 {
 103         DIR *dir;
 104         struct dirent *e;
 105         char *fname;
 106
 107         if (!(dir = opendir(packdir)))
 108                 return;
 109
 110         while ((e = readdir(dir)) != NULL) {
 111                 size_t len;
 112                 int i;
 113
 114                 if (!strip_suffix(e->d_name, ".pack", &len))
 115                         continue;
 116
 117                 for (i = 0; i < extra_keep->nr; i++)
 118                         if (!fspathcmp(e->d_name, extra_keep->items[i].string))
 119                                 break;
 120
 121                 fname = xmemdupz(e->d_name, len);
 122
 123                 if ((extra_keep->nr > 0 && i < extra_keep->nr) ||
 124                     (file_exists(mkpath("%s/%s.keep", packdir, fname)))) {
 125                         string_list_append_nodup(fname_kept_list, fname);
 126                 } else {
 127                         struct string_list_item *item;
 128                         item = string_list_append_nodup(fname_nonkept_list,
 129                                                         fname);
 130                         if (file_exists(mkpath("%s/%s.mtimes", packdir, fname)))
 131                                 item->util = (void*)(uintptr_t)CRUFT_PACK;
 132                 }
 133         }
 134         closedir(dir);
 135
 136         string_list_sort(fname_kept_list);
 137 }
 138
 139 static void remove_redundant_pack(const char *dir_name, const char *base_name)
 140 {
 141         struct strbuf buf = STRBUF_INIT;
 142         struct multi_pack_index *m = get_local_multi_pack_index(the_repository);
 143         strbuf_addf(&buf, "%s.pack", base_name);
 144         if (m && midx_contains_pack(m, buf.buf))
 145                 clear_midx_file(the_repository);
 146         strbuf_insertf(&buf, 0, "%s/", dir_name);
 147         unlink_pack_path(buf.buf, 1);
 148         strbuf_release(&buf);
 149 }
 150
 151 static void prepare_pack_objects(struct child_process *cmd,
 152                                  const struct pack_objects_args *args,
 153                                  const char *out)
 154 {
 155         strvec_push(&cmd->args, "pack-objects");
 156         if (args->window)
 157                 strvec_pushf(&cmd->args, "--window=%s", args->window);
 158         if (args->window_memory)
 159                 strvec_pushf(&cmd->args, "--window-memory=%s", args->window_memory);
 160         if (args->depth)
 161                 strvec_pushf(&cmd->args, "--depth=%s", args->depth);
 162         if (args->threads)
 163                 strvec_pushf(&cmd->args, "--threads=%s", args->threads);
 164         if (args->max_pack_size)
 165                 strvec_pushf(&cmd->args, "--max-pack-size=%s", args->max_pack_size);
 166         if (args->no_reuse_delta)
 167                 strvec_pushf(&cmd->args, "--no-reuse-delta");
 168         if (args->no_reuse_object)
 169                 strvec_pushf(&cmd->args, "--no-reuse-object");
 170         if (args->local)
 171                 strvec_push(&cmd->args,  "--local");
 172         if (args->quiet)
 173                 strvec_push(&cmd->args,  "--quiet");
 174         if (delta_base_offset)
 175                 strvec_push(&cmd->args,  "--delta-base-offset");
 176         strvec_push(&cmd->args, out);
 177         cmd->git_cmd = 1;
 178         cmd->out = -1;
 179 }
 180
 181 /*
 182  * Write oid to the given struct child_process's stdin, starting it first if
 183  * necessary.
 184  */
 185 static int write_oid(const struct object_id *oid, struct packed_git *pack,
 186                      uint32_t pos, void *data)
 187 {
 188         struct child_process *cmd = data;
 189
 190         if (cmd->in == -1) {
 191                 if (start_command(cmd))
 192                         die(_("could not start pack-objects to repack promisor objects"));
 193         }
 194
 195         xwrite(cmd->in, oid_to_hex(oid), the_hash_algo->hexsz);
 196         xwrite(cmd->in, "\n", 1);
 197         return 0;
 198 }
 199
 200 static struct {
 201         const char *name;
 202         unsigned optional:1;
 203 } exts[] = {
 204         {".pack"},
 205         {".rev", 1},
 206         {".mtimes", 1},
 207         {".bitmap", 1},
 208         {".promisor", 1},
 209         {".idx"},
 210 };
 211
 212 struct generated_pack_data {
 213         struct tempfile *tempfiles[ARRAY_SIZE(exts)];
 214 };
 215
 216 static struct generated_pack_data *populate_pack_exts(const char *name)
 217 {
 218         struct stat statbuf;
 219         struct strbuf path = STRBUF_INIT;
 220         struct generated_pack_data *data = xcalloc(1, sizeof(*data));
 221         int i;
 222
 223         for (i = 0; i < ARRAY_SIZE(exts); i++) {
 224                 strbuf_reset(&path);
 225                 strbuf_addf(&path, "%s-%s%s", packtmp, name, exts[i].name);
 226
 227                 if (stat(path.buf, &statbuf))
 228                         continue;
 229
 230                 data->tempfiles[i] = register_tempfile(path.buf);
 231         }
 232
 233         strbuf_release(&path);
 234         return data;
 235 }
 236
 237 static void repack_promisor_objects(const struct pack_objects_args *args,
 238                                     struct string_list *names)
 239 {
 240         struct child_process cmd = CHILD_PROCESS_INIT;
 241         FILE *out;
 242         struct strbuf line = STRBUF_INIT;
 243
 244         prepare_pack_objects(&cmd, args, packtmp);
 245         cmd.in = -1;
 246
 247         /*
 248          * NEEDSWORK: Giving pack-objects only the OIDs without any ordering
 249          * hints may result in suboptimal deltas in the resulting pack. See if
 250          * the OIDs can be sent with fake paths such that pack-objects can use a
 251          * {type -> existing pack order} ordering when computing deltas instead
 252          * of a {type -> size} ordering, which may produce better deltas.
 253          */
 254         for_each_packed_object(write_oid, &cmd,
 255                                FOR_EACH_OBJECT_PROMISOR_ONLY);
 256
 257         if (cmd.in == -1) {
 258                 /* No packed objects; cmd was never started */
 259                 child_process_clear(&cmd);
 260                 return;
 261         }
 262
 263         close(cmd.in);
 264
 265         out = xfdopen(cmd.out, "r");
 266         while (strbuf_getline_lf(&line, out) != EOF) {
 267                 struct string_list_item *item;
 268                 char *promisor_name;
 269
 270                 if (line.len != the_hash_algo->hexsz)
 271                         die(_("repack: Expecting full hex object ID lines only from pack-objects."));
 272                 item = string_list_append(names, line.buf);
 273
 274                 /*
 275                  * pack-objects creates the .pack and .idx files, but not the
 276                  * .promisor file. Create the .promisor file, which is empty.
 277                  *
 278                  * NEEDSWORK: fetch-pack sometimes generates non-empty
 279                  * .promisor files containing the ref names and associated
 280                  * hashes at the point of generation of the corresponding
 281                  * packfile, but this would not preserve their contents. Maybe
 282                  * concatenate the contents of all .promisor files instead of
 283                  * just creating a new empty file.
 284                  */
 285                 promisor_name = mkpathdup("%s-%s.promisor", packtmp,
 286                                           line.buf);
 287                 write_promisor_file(promisor_name, NULL, 0);
 288
 289                 item->util = populate_pack_exts(item->string);
 290
 291                 free(promisor_name);
 292         }
 293         fclose(out);
 294         if (finish_command(&cmd))
 295                 die(_("could not finish pack-objects to repack promisor objects"));
 296 }
 297
 298 struct pack_geometry {
 299         struct packed_git **pack;
 300         uint32_t pack_nr, pack_alloc;
 301         uint32_t split;
 302 };
 303
 304 static uint32_t geometry_pack_weight(struct packed_git *p)
 305 {
 306         if (open_pack_index(p))
 307                 die(_("cannot open index for %s"), p->pack_name);
 308         return p->num_objects;
 309 }
 310
 311 static int geometry_cmp(const void *va, const void *vb)
 312 {
 313         uint32_t aw = geometry_pack_weight(*(struct packed_git **)va),
 314                  bw = geometry_pack_weight(*(struct packed_git **)vb);
 315
 316         if (aw < bw)
 317                 return -1;
 318         if (aw > bw)
 319                 return 1;
 320         return 0;
 321 }
 322
 323 static void init_pack_geometry(struct pack_geometry **geometry_p,
 324                                struct string_list *existing_kept_packs)
 325 {
 326         struct packed_git *p;
 327         struct pack_geometry *geometry;
 328         struct strbuf buf = STRBUF_INIT;
 329
 330         *geometry_p = xcalloc(1, sizeof(struct pack_geometry));
 331         geometry = *geometry_p;
 332
 333         for (p = get_all_packs(the_repository); p; p = p->next) {
 334                 if (!pack_kept_objects) {
 335                         /*
 336                          * Any pack that has its pack_keep bit set will appear
 337                          * in existing_kept_packs below, but this saves us from
 338                          * doing a more expensive check.
 339                          */
 340                         if (p->pack_keep)
 341                                 continue;
 342
 343                         /*
 344                          * The pack may be kept via the --keep-pack option;
 345                          * check 'existing_kept_packs' to determine whether to
 346                          * ignore it.
 347                          */
 348                         strbuf_reset(&buf);
 349                         strbuf_addstr(&buf, pack_basename(p));
 350                         strbuf_strip_suffix(&buf, ".pack");
 351
 352                         if (string_list_has_string(existing_kept_packs, buf.buf))
 353                                 continue;
 354                 }
 355                 if (p->is_cruft)
 356                         continue;
 357
 358                 ALLOC_GROW(geometry->pack,
 359                            geometry->pack_nr + 1,
 360                            geometry->pack_alloc);
 361
 362                 geometry->pack[geometry->pack_nr] = p;
 363                 geometry->pack_nr++;
 364         }
 365
 366         QSORT(geometry->pack, geometry->pack_nr, geometry_cmp);
 367         strbuf_release(&buf);
 368 }
 369
 370 static void split_pack_geometry(struct pack_geometry *geometry, int factor)
 371 {
 372         uint32_t i;
 373         uint32_t split;
 374         off_t total_size = 0;
 375
 376         if (!geometry->pack_nr) {
 377                 geometry->split = geometry->pack_nr;
 378                 return;
 379         }
 380
 381         /*
 382          * First, count the number of packs (in descending order of size) which
 383          * already form a geometric progression.
 384          */
 385         for (i = geometry->pack_nr - 1; i > 0; i--) {
 386                 struct packed_git *ours = geometry->pack[i];
 387                 struct packed_git *prev = geometry->pack[i - 1];
 388
 389                 if (unsigned_mult_overflows(factor, geometry_pack_weight(prev)))
 390                         die(_("pack %s too large to consider in geometric "
 391                               "progression"),
 392                             prev->pack_name);
 393
 394                 if (geometry_pack_weight(ours) < factor * geometry_pack_weight(prev))
 395                         break;
 396         }
 397
 398         split = i;
 399
 400         if (split) {
 401                 /*
 402                  * Move the split one to the right, since the top element in the
 403                  * last-compared pair can't be in the progression. Only do this
 404                  * when we split in the middle of the array (otherwise if we got
 405                  * to the end, then the split is in the right place).
 406                  */
 407                 split++;
 408         }
 409
 410         /*
 411          * Then, anything to the left of 'split' must be in a new pack. But,
 412          * creating that new pack may cause packs in the heavy half to no longer
 413          * form a geometric progression.
 414          *
 415          * Compute an expected size of the new pack, and then determine how many
 416          * packs in the heavy half need to be joined into it (if any) to restore
 417          * the geometric progression.
 418          */
 419         for (i = 0; i < split; i++) {
 420                 struct packed_git *p = geometry->pack[i];
 421
 422                 if (unsigned_add_overflows(total_size, geometry_pack_weight(p)))
 423                         die(_("pack %s too large to roll up"), p->pack_name);
 424                 total_size += geometry_pack_weight(p);
 425         }
 426         for (i = split; i < geometry->pack_nr; i++) {
 427                 struct packed_git *ours = geometry->pack[i];
 428
 429                 if (unsigned_mult_overflows(factor, total_size))
 430                         die(_("pack %s too large to roll up"), ours->pack_name);
 431
 432                 if (geometry_pack_weight(ours) < factor * total_size) {
 433                         if (unsigned_add_overflows(total_size,
 434                                                    geometry_pack_weight(ours)))
 435                                 die(_("pack %s too large to roll up"),
 436                                     ours->pack_name);
 437
 438                         split++;
 439                         total_size += geometry_pack_weight(ours);
 440                 } else
 441                         break;
 442         }
 443
 444         geometry->split = split;
 445 }
 446
 447 static struct packed_git *get_preferred_pack(struct pack_geometry *geometry)
 448 {
 449         uint32_t i;
 450
 451         if (!geometry) {
 452                 /*
 453                  * No geometry means either an all-into-one repack (in which
 454                  * case there is only one pack left and it is the largest) or an
 455                  * incremental one.
 456                  *
 457                  * If repacking incrementally, then we could check the size of
 458                  * all packs to determine which should be preferred, but leave
 459                  * this for later.
 460                  */
 461                 return NULL;
 462         }
 463         if (geometry->split == geometry->pack_nr)
 464                 return NULL;
 465
 466         /*
 467          * The preferred pack is the largest pack above the split line. In
 468          * other words, it is the largest pack that does not get rolled up in
 469          * the geometric repack.
 470          */
 471         for (i = geometry->pack_nr; i > geometry->split; i--)
 472                 /*
 473                  * A pack that is not local would never be included in a
 474                  * multi-pack index. We thus skip over any non-local packs.
 475                  */
 476                 if (geometry->pack[i - 1]->pack_local)
 477                         return geometry->pack[i - 1];
 478
 479         return NULL;
 480 }
 481
 482 static void clear_pack_geometry(struct pack_geometry *geometry)
 483 {
 484         if (!geometry)
 485                 return;
 486
 487         free(geometry->pack);
 488         geometry->pack_nr = 0;
 489         geometry->pack_alloc = 0;
 490         geometry->split = 0;
 491 }
 492
 493 struct midx_snapshot_ref_data {
 494         struct tempfile *f;
 495         struct oidset seen;
 496         int preferred;
 497 };
 498
 499 static int midx_snapshot_ref_one(const char *refname UNUSED,
 500                                  const struct object_id *oid,
 501                                  int flag UNUSED, void *_data)
 502 {
 503         struct midx_snapshot_ref_data *data = _data;
 504         struct object_id peeled;
 505
 506         if (!peel_iterated_oid(oid, &peeled))
 507                 oid = &peeled;
 508
 509         if (oidset_insert(&data->seen, oid))
 510                 return 0; /* already seen */
 511
 512         if (oid_object_info(the_repository, oid, NULL) != OBJ_COMMIT)
 513                 return 0;
 514
 515         fprintf(data->f->fp, "%s%s\n", data->preferred ? "+" : "",
 516                 oid_to_hex(oid));
 517
 518         return 0;
 519 }
 520
 521 static void midx_snapshot_refs(struct tempfile *f)
 522 {
 523         struct midx_snapshot_ref_data data;
 524         const struct string_list *preferred = bitmap_preferred_tips(the_repository);
 525
 526         data.f = f;
 527         data.preferred = 0;
 528         oidset_init(&data.seen, 0);
 529
 530         if (!fdopen_tempfile(f, "w"))
 531                  die(_("could not open tempfile %s for writing"),
 532                      get_tempfile_path(f));
 533
 534         if (preferred) {
 535                 struct string_list_item *item;
 536
 537                 data.preferred = 1;
 538                 for_each_string_list_item(item, preferred)
 539                         for_each_ref_in(item->string, midx_snapshot_ref_one, &data);
 540                 data.preferred = 0;
 541         }
 542
 543         for_each_ref(midx_snapshot_ref_one, &data);
 544
 545         if (close_tempfile_gently(f)) {
 546                 int save_errno = errno;
 547                 delete_tempfile(&f);
 548                 errno = save_errno;
 549                 die_errno(_("could not close refs snapshot tempfile"));
 550         }
 551
 552         oidset_clear(&data.seen);
 553 }
 554
 555 static void midx_included_packs(struct string_list *include,
 556                                 struct string_list *existing_nonkept_packs,
 557                                 struct string_list *existing_kept_packs,
 558                                 struct string_list *names,
 559                                 struct pack_geometry *geometry)
 560 {
 561         struct string_list_item *item;
 562
 563         for_each_string_list_item(item, existing_kept_packs)
 564                 string_list_insert(include, xstrfmt("%s.idx", item->string));
 565         for_each_string_list_item(item, names)
 566                 string_list_insert(include, xstrfmt("pack-%s.idx", item->string));
 567         if (geometry) {
 568                 struct strbuf buf = STRBUF_INIT;
 569                 uint32_t i;
 570                 for (i = geometry->split; i < geometry->pack_nr; i++) {
 571                         struct packed_git *p = geometry->pack[i];
 572
 573                         /*
 574                          * The multi-pack index never refers to packfiles part
 575                          * of an alternate object database, so we skip these.
 576                          * While git-multi-pack-index(1) would silently ignore
 577                          * them anyway, this allows us to skip executing the
 578                          * command completely when we have only non-local
 579                          * packfiles.
 580                          */
 581                         if (!p->pack_local)
 582                                 continue;
 583
 584                         strbuf_addstr(&buf, pack_basename(p));
 585                         strbuf_strip_suffix(&buf, ".pack");
 586                         strbuf_addstr(&buf, ".idx");
 587
 588                         string_list_insert(include, strbuf_detach(&buf, NULL));
 589                 }
 590
 591                 for_each_string_list_item(item, existing_nonkept_packs) {
 592                         if (!((uintptr_t)item->util & CRUFT_PACK)) {
 593                                 /*
 594                                  * no need to check DELETE_PACK, since we're not
 595                                  * doing an ALL_INTO_ONE repack
 596                                  */
 597                                 continue;
 598                         }
 599                         string_list_insert(include, xstrfmt("%s.idx", item->string));
 600                 }
 601         } else {
 602                 for_each_string_list_item(item, existing_nonkept_packs) {
 603                         if ((uintptr_t)item->util & DELETE_PACK)
 604                                 continue;
 605                         string_list_insert(include, xstrfmt("%s.idx", item->string));
 606                 }
 607         }
 608 }
 609
 610 static int write_midx_included_packs(struct string_list *include,
 611                                      struct pack_geometry *geometry,
 612                                      const char *refs_snapshot,
 613                                      int show_progress, int write_bitmaps)
 614 {
 615         struct child_process cmd = CHILD_PROCESS_INIT;
 616         struct string_list_item *item;
 617         struct packed_git *preferred = get_preferred_pack(geometry);
 618         FILE *in;
 619         int ret;
 620
 621         if (!include->nr)
 622                 return 0;
 623
 624         cmd.in = -1;
 625         cmd.git_cmd = 1;
 626
 627         strvec_push(&cmd.args, "multi-pack-index");
 628         strvec_pushl(&cmd.args, "write", "--stdin-packs", NULL);
 629
 630         if (show_progress)
 631                 strvec_push(&cmd.args, "--progress");
 632         else
 633                 strvec_push(&cmd.args, "--no-progress");
 634
 635         if (write_bitmaps)
 636                 strvec_push(&cmd.args, "--bitmap");
 637
 638         if (preferred)
 639                 strvec_pushf(&cmd.args, "--preferred-pack=%s",
 640                              pack_basename(preferred));
 641
 642         if (refs_snapshot)
 643                 strvec_pushf(&cmd.args, "--refs-snapshot=%s", refs_snapshot);
 644
 645         ret = start_command(&cmd);
 646         if (ret)
 647                 return ret;
 648
 649         in = xfdopen(cmd.in, "w");
 650         for_each_string_list_item(item, include)
 651                 fprintf(in, "%s\n", item->string);
 652         fclose(in);
 653
 654         return finish_command(&cmd);
 655 }
 656
 657 static void remove_redundant_bitmaps(struct string_list *include,
 658                                      const char *packdir)
 659 {
 660         struct strbuf path = STRBUF_INIT;
 661         struct string_list_item *item;
 662         size_t packdir_len;
 663
 664         strbuf_addstr(&path, packdir);
 665         strbuf_addch(&path, '/');
 666         packdir_len = path.len;
 667
 668         /*
 669          * Remove any pack bitmaps corresponding to packs which are now
 670          * included in the MIDX.
 671          */
 672         for_each_string_list_item(item, include) {
 673                 strbuf_addstr(&path, item->string);
 674                 strbuf_strip_suffix(&path, ".idx");
 675                 strbuf_addstr(&path, ".bitmap");
 676
 677                 if (unlink(path.buf) && errno != ENOENT)
 678                         warning_errno(_("could not remove stale bitmap: %s"),
 679                                       path.buf);
 680
 681                 strbuf_setlen(&path, packdir_len);
 682         }
 683         strbuf_release(&path);
 684 }
 685
 686 static int write_cruft_pack(const struct pack_objects_args *args,
 687                             const char *destination,
 688                             const char *pack_prefix,
 689                             const char *cruft_expiration,
 690                             struct string_list *names,
 691                             struct string_list *existing_packs,
 692                             struct string_list *existing_kept_packs)
 693 {
 694         struct child_process cmd = CHILD_PROCESS_INIT;
 695         struct strbuf line = STRBUF_INIT;
 696         struct string_list_item *item;
 697         FILE *in, *out;
 698         int ret;
 699         const char *scratch;
 700         int local = skip_prefix(destination, packdir, &scratch);
 701
 702         prepare_pack_objects(&cmd, args, destination);
 703
 704         strvec_push(&cmd.args, "--cruft");
 705         if (cruft_expiration)
 706                 strvec_pushf(&cmd.args, "--cruft-expiration=%s",
 707                              cruft_expiration);
 708
 709         strvec_push(&cmd.args, "--honor-pack-keep");
 710         strvec_push(&cmd.args, "--non-empty");
 711         strvec_push(&cmd.args, "--max-pack-size=0");
 712
 713         cmd.in = -1;
 714
 715         ret = start_command(&cmd);
 716         if (ret)
 717                 return ret;
 718
 719         /*
 720          * names has a confusing double use: it both provides the list
 721          * of just-written new packs, and accepts the name of the cruft
 722          * pack we are writing.
 723          *
 724          * By the time it is read here, it contains only the pack(s)
 725          * that were just written, which is exactly the set of packs we
 726          * want to consider kept.
 727          *
 728          * If `--expire-to` is given, the double-use served by `names`
 729          * ensures that the pack written to `--expire-to` excludes any
 730          * objects contained in the cruft pack.
 731          */
 732         in = xfdopen(cmd.in, "w");
 733         for_each_string_list_item(item, names)
 734                 fprintf(in, "%s-%s.pack\n", pack_prefix, item->string);
 735         for_each_string_list_item(item, existing_packs)
 736                 fprintf(in, "-%s.pack\n", item->string);
 737         for_each_string_list_item(item, existing_kept_packs)
 738                 fprintf(in, "%s.pack\n", item->string);
 739         fclose(in);
 740
 741         out = xfdopen(cmd.out, "r");
 742         while (strbuf_getline_lf(&line, out) != EOF) {
 743                 struct string_list_item *item;
 744
 745                 if (line.len != the_hash_algo->hexsz)
 746                         die(_("repack: Expecting full hex object ID lines only "
 747                               "from pack-objects."));
 748                 /*
 749                  * avoid putting packs written outside of the repository in the
 750                  * list of names
 751                  */
 752                 if (local) {
 753                         item = string_list_append(names, line.buf);
 754                         item->util = populate_pack_exts(line.buf);
 755                 }
 756         }
 757         fclose(out);
 758
 759         strbuf_release(&line);
 760
 761         return finish_command(&cmd);
 762 }
 763
 764 int cmd_repack(int argc, const char **argv, const char *prefix)
 765 {
 766         struct child_process cmd = CHILD_PROCESS_INIT;
 767         struct string_list_item *item;
 768         struct string_list names = STRING_LIST_INIT_DUP;
 769         struct string_list existing_nonkept_packs = STRING_LIST_INIT_DUP;
 770         struct string_list existing_kept_packs = STRING_LIST_INIT_DUP;
 771         struct pack_geometry *geometry = NULL;
 772         struct strbuf line = STRBUF_INIT;
 773         struct tempfile *refs_snapshot = NULL;
 774         int i, ext, ret;
 775         FILE *out;
 776         int show_progress;
 777
 778         /* variables to be filled by option parsing */
 779         int delete_redundant = 0;
 780         const char *unpack_unreachable = NULL;
 781         int keep_unreachable = 0;
 782         struct string_list keep_pack_list = STRING_LIST_INIT_NODUP;
 783         struct pack_objects_args po_args = {NULL};
 784         struct pack_objects_args cruft_po_args = {NULL};
 785         int geometric_factor = 0;
 786         int write_midx = 0;
 787         const char *cruft_expiration = NULL;
 788         const char *expire_to = NULL;
 789
 790         struct option builtin_repack_options[] = {
 791                 OPT_BIT('a', NULL, &pack_everything,
 792                                 N_("pack everything in a single pack"), ALL_INTO_ONE),
 793                 OPT_BIT('A', NULL, &pack_everything,
 794                                 N_("same as -a, and turn unreachable objects loose"),
 795                                    LOOSEN_UNREACHABLE | ALL_INTO_ONE),
 796                 OPT_BIT(0, "cruft", &pack_everything,
 797                                 N_("same as -a, pack unreachable cruft objects separately"),
 798                                    PACK_CRUFT),
 799                 OPT_STRING(0, "cruft-expiration", &cruft_expiration, N_("approxidate"),
 800                                 N_("with -C, expire objects older than this")),
 801                 OPT_BOOL('d', NULL, &delete_redundant,
 802                                 N_("remove redundant packs, and run git-prune-packed")),
 803                 OPT_BOOL('f', NULL, &po_args.no_reuse_delta,
 804                                 N_("pass --no-reuse-delta to git-pack-objects")),
 805                 OPT_BOOL('F', NULL, &po_args.no_reuse_object,
 806                                 N_("pass --no-reuse-object to git-pack-objects")),
 807                 OPT_NEGBIT('n', NULL, &run_update_server_info,
 808                                 N_("do not run git-update-server-info"), 1),
 809                 OPT__QUIET(&po_args.quiet, N_("be quiet")),
 810                 OPT_BOOL('l', "local", &po_args.local,
 811                                 N_("pass --local to git-pack-objects")),
 812                 OPT_BOOL('b', "write-bitmap-index", &write_bitmaps,
 813                                 N_("write bitmap index")),
 814                 OPT_BOOL('i', "delta-islands", &use_delta_islands,
 815                                 N_("pass --delta-islands to git-pack-objects")),
 816                 OPT_STRING(0, "unpack-unreachable", &unpack_unreachable, N_("approxidate"),
 817                                 N_("with -A, do not loosen objects older than this")),
 818                 OPT_BOOL('k', "keep-unreachable", &keep_unreachable,
 819                                 N_("with -a, repack unreachable objects")),
 820                 OPT_STRING(0, "window", &po_args.window, N_("n"),
 821                                 N_("size of the window used for delta compression")),
 822                 OPT_STRING(0, "window-memory", &po_args.window_memory, N_("bytes"),
 823                                 N_("same as the above, but limit memory size instead of entries count")),
 824                 OPT_STRING(0, "depth", &po_args.depth, N_("n"),
 825                                 N_("limits the maximum delta depth")),
 826                 OPT_STRING(0, "threads", &po_args.threads, N_("n"),
 827                                 N_("limits the maximum number of threads")),
 828                 OPT_STRING(0, "max-pack-size", &po_args.max_pack_size, N_("bytes"),
 829                                 N_("maximum size of each packfile")),
 830                 OPT_BOOL(0, "pack-kept-objects", &pack_kept_objects,
 831                                 N_("repack objects in packs marked with .keep")),
 832                 OPT_STRING_LIST(0, "keep-pack", &keep_pack_list, N_("name"),
 833                                 N_("do not repack this pack")),
 834                 OPT_INTEGER('g', "geometric", &geometric_factor,
 835                             N_("find a geometric progression with factor <N>")),
 836                 OPT_BOOL('m', "write-midx", &write_midx,
 837                            N_("write a multi-pack index of the resulting packs")),
 838                 OPT_STRING(0, "expire-to", &expire_to, N_("dir"),
 839                            N_("pack prefix to store a pack containing pruned objects")),
 840                 OPT_END()
 841         };
 842
 843         git_config(repack_config, &cruft_po_args);
 844
 845         argc = parse_options(argc, argv, prefix, builtin_repack_options,
 846                                 git_repack_usage, 0);
 847
 848         if (delete_redundant && repository_format_precious_objects)
 849                 die(_("cannot delete packs in a precious-objects repo"));
 850
 851         if (keep_unreachable &&
 852             (unpack_unreachable || (pack_everything & LOOSEN_UNREACHABLE)))
 853                 die(_("options '%s' and '%s' cannot be used together"), "--keep-unreachable", "-A");
 854
 855         if (pack_everything & PACK_CRUFT) {
 856                 pack_everything |= ALL_INTO_ONE;
 857
 858                 if (unpack_unreachable || (pack_everything & LOOSEN_UNREACHABLE))
 859                         die(_("options '%s' and '%s' cannot be used together"), "--cruft", "-A");
 860                 if (keep_unreachable)
 861                         die(_("options '%s' and '%s' cannot be used together"), "--cruft", "-k");
 862         }
 863
 864         if (write_bitmaps < 0) {
 865                 if (!write_midx &&
 866                     (!(pack_everything & ALL_INTO_ONE) || !is_bare_repository()))
 867                         write_bitmaps = 0;
 868         } else if (write_bitmaps &&
 869                    git_env_bool(GIT_TEST_MULTI_PACK_INDEX, 0) &&
 870                    git_env_bool(GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP, 0)) {
 871                 write_bitmaps = 0;
 872         }
 873         if (pack_kept_objects < 0)
 874                 pack_kept_objects = write_bitmaps > 0 && !write_midx;
 875
 876         if (write_bitmaps && !(pack_everything & ALL_INTO_ONE) && !write_midx)
 877                 die(_(incremental_bitmap_conflict_error));
 878
 879         if (write_midx && write_bitmaps) {
 880                 struct strbuf path = STRBUF_INIT;
 881
 882                 strbuf_addf(&path, "%s/%s_XXXXXX", get_object_directory(),
 883                             "bitmap-ref-tips");
 884
 885                 refs_snapshot = xmks_tempfile(path.buf);
 886                 midx_snapshot_refs(refs_snapshot);
 887
 888                 strbuf_release(&path);
 889         }
 890
 891         packdir = mkpathdup("%s/pack", get_object_directory());
 892         packtmp_name = xstrfmt(".tmp-%d-pack", (int)getpid());
 893         packtmp = mkpathdup("%s/%s", packdir, packtmp_name);
 894
 895         collect_pack_filenames(&existing_nonkept_packs, &existing_kept_packs,
 896                                &keep_pack_list);
 897
 898         if (geometric_factor) {
 899                 if (pack_everything)
 900                         die(_("options '%s' and '%s' cannot be used together"), "--geometric", "-A/-a");
 901                 init_pack_geometry(&geometry, &existing_kept_packs);
 902                 split_pack_geometry(geometry, geometric_factor);
 903         }
 904
 905         prepare_pack_objects(&cmd, &po_args, packtmp);
 906
 907         show_progress = !po_args.quiet && isatty(2);
 908
 909         strvec_push(&cmd.args, "--keep-true-parents");
 910         if (!pack_kept_objects)
 911                 strvec_push(&cmd.args, "--honor-pack-keep");
 912         for (i = 0; i < keep_pack_list.nr; i++)
 913                 strvec_pushf(&cmd.args, "--keep-pack=%s",
 914                              keep_pack_list.items[i].string);
 915         strvec_push(&cmd.args, "--non-empty");
 916         if (!geometry) {
 917                 /*
 918                  * We need to grab all reachable objects, including those that
 919                  * are reachable from reflogs and the index.
 920                  *
 921                  * When repacking into a geometric progression of packs,
 922                  * however, we ask 'git pack-objects --stdin-packs', and it is
 923                  * not about packing objects based on reachability but about
 924                  * repacking all the objects in specified packs and loose ones
 925                  * (indeed, --stdin-packs is incompatible with these options).
 926                  */
 927                 strvec_push(&cmd.args, "--all");
 928                 strvec_push(&cmd.args, "--reflog");
 929                 strvec_push(&cmd.args, "--indexed-objects");
 930         }
 931         if (has_promisor_remote())
 932                 strvec_push(&cmd.args, "--exclude-promisor-objects");
 933         if (!write_midx) {
 934                 if (write_bitmaps > 0)
 935                         strvec_push(&cmd.args, "--write-bitmap-index");
 936                 else if (write_bitmaps < 0)
 937                         strvec_push(&cmd.args, "--write-bitmap-index-quiet");
 938         }
 939         if (use_delta_islands)
 940                 strvec_push(&cmd.args, "--delta-islands");
 941
 942         if (pack_everything & ALL_INTO_ONE) {
 943                 repack_promisor_objects(&po_args, &names);
 944
 945                 if (existing_nonkept_packs.nr && delete_redundant &&
 946                     !(pack_everything & PACK_CRUFT)) {
 947                         for_each_string_list_item(item, &names) {
 948                                 strvec_pushf(&cmd.args, "--keep-pack=%s-%s.pack",
 949                                              packtmp_name, item->string);
 950                         }
 951                         if (unpack_unreachable) {
 952                                 strvec_pushf(&cmd.args,
 953                                              "--unpack-unreachable=%s",
 954                                              unpack_unreachable);
 955                         } else if (pack_everything & LOOSEN_UNREACHABLE) {
 956                                 strvec_push(&cmd.args,
 957                                             "--unpack-unreachable");
 958                         } else if (keep_unreachable) {
 959                                 strvec_push(&cmd.args, "--keep-unreachable");
 960                                 strvec_push(&cmd.args, "--pack-loose-unreachable");
 961                         }
 962                 }
 963         } else if (geometry) {
 964                 strvec_push(&cmd.args, "--stdin-packs");
 965                 strvec_push(&cmd.args, "--unpacked");
 966         } else {
 967                 strvec_push(&cmd.args, "--unpacked");
 968                 strvec_push(&cmd.args, "--incremental");
 969         }
 970
 971         if (geometry)
 972                 cmd.in = -1;
 973         else
 974                 cmd.no_stdin = 1;
 975
 976         ret = start_command(&cmd);
 977         if (ret)
 978                 goto cleanup;
 979
 980         if (geometry) {
 981                 FILE *in = xfdopen(cmd.in, "w");
 982                 /*
 983                  * The resulting pack should contain all objects in packs that
 984                  * are going to be rolled up, but exclude objects in packs which
 985                  * are being left alone.
 986                  */
 987                 for (i = 0; i < geometry->split; i++)
 988                         fprintf(in, "%s\n", pack_basename(geometry->pack[i]));
 989                 for (i = geometry->split; i < geometry->pack_nr; i++)
 990                         fprintf(in, "^%s\n", pack_basename(geometry->pack[i]));
 991                 fclose(in);
 992         }
 993
 994         out = xfdopen(cmd.out, "r");
 995         while (strbuf_getline_lf(&line, out) != EOF) {
 996                 struct string_list_item *item;
 997
 998                 if (line.len != the_hash_algo->hexsz)
 999                         die(_("repack: Expecting full hex object ID lines only from pack-objects."));
1000                 item = string_list_append(&names, line.buf);
1001                 item->util = populate_pack_exts(item->string);
1002         }
1003         strbuf_release(&line);
1004         fclose(out);
1005         ret = finish_command(&cmd);
1006         if (ret)
1007                 goto cleanup;
1008
1009         if (!names.nr && !po_args.quiet)
1010                 printf_ln(_("Nothing new to pack."));
1011
1012         if (pack_everything & PACK_CRUFT) {
1013                 const char *pack_prefix;
1014                 if (!skip_prefix(packtmp, packdir, &pack_prefix))
1015                         die(_("pack prefix %s does not begin with objdir %s"),
1016                             packtmp, packdir);
1017                 if (*pack_prefix == '/')
1018                         pack_prefix++;
1019
1020                 if (!cruft_po_args.window)
1021                         cruft_po_args.window = po_args.window;
1022                 if (!cruft_po_args.window_memory)
1023                         cruft_po_args.window_memory = po_args.window_memory;
1024                 if (!cruft_po_args.depth)
1025                         cruft_po_args.depth = po_args.depth;
1026                 if (!cruft_po_args.threads)
1027                         cruft_po_args.threads = po_args.threads;
1028
1029                 cruft_po_args.local = po_args.local;
1030                 cruft_po_args.quiet = po_args.quiet;
1031
1032                 ret = write_cruft_pack(&cruft_po_args, packtmp, pack_prefix,
1033                                        cruft_expiration, &names,
1034                                        &existing_nonkept_packs,
1035                                        &existing_kept_packs);
1036                 if (ret)
1037                         goto cleanup;
1038
1039                 if (delete_redundant && expire_to) {
1040                         /*
1041                          * If `--expire-to` is given with `-d`, it's possible
1042                          * that we're about to prune some objects. With cruft
1043                          * packs, pruning is implicit: any objects from existing
1044                          * packs that weren't picked up by new packs are removed
1045                          * when their packs are deleted.
1046                          *
1047                          * Generate an additional cruft pack, with one twist:
1048                          * `names` now includes the name of the cruft pack
1049                          * written in the previous step. So the contents of
1050                          * _this_ cruft pack exclude everything contained in the
1051                          * existing cruft pack (that is, all of the unreachable
1052                          * objects which are no older than
1053                          * `--cruft-expiration`).
1054                          *
1055                          * To make this work, cruft_expiration must become NULL
1056                          * so that this cruft pack doesn't actually prune any
1057                          * objects. If it were non-NULL, this call would always
1058                          * generate an empty pack (since every object not in the
1059                          * cruft pack generated above will have an mtime older
1060                          * than the expiration).
1061                          */
1062                         ret = write_cruft_pack(&cruft_po_args, expire_to,
1063                                                pack_prefix,
1064                                                NULL,
1065                                                &names,
1066                                                &existing_nonkept_packs,
1067                                                &existing_kept_packs);
1068                         if (ret)
1069                                 goto cleanup;
1070                 }
1071         }
1072
1073         string_list_sort(&names);
1074
1075         close_object_store(the_repository->objects);
1076
1077         /*
1078          * Ok we have prepared all new packfiles.
1079          */
1080         for_each_string_list_item(item, &names) {
1081                 struct generated_pack_data *data = item->util;
1082
1083                 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
1084                         char *fname;
1085
1086                         fname = mkpathdup("%s/pack-%s%s",
1087                                         packdir, item->string, exts[ext].name);
1088
1089                         if (data->tempfiles[ext]) {
1090                                 const char *fname_old = get_tempfile_path(data->tempfiles[ext]);
1091                                 struct stat statbuffer;
1092
1093                                 if (!stat(fname_old, &statbuffer)) {
1094                                         statbuffer.st_mode &= ~(S_IWUSR | S_IWGRP | S_IWOTH);
1095                                         chmod(fname_old, statbuffer.st_mode);
1096                                 }
1097
1098                                 if (rename_tempfile(&data->tempfiles[ext], fname))
1099                                         die_errno(_("renaming pack to '%s' failed"), fname);
1100                         } else if (!exts[ext].optional)
1101                                 die(_("pack-objects did not write a '%s' file for pack %s-%s"),
1102                                     exts[ext].name, packtmp, item->string);
1103                         else if (unlink(fname) < 0 && errno != ENOENT)
1104                                 die_errno(_("could not unlink: %s"), fname);
1105
1106                         free(fname);
1107                 }
1108         }
1109         /* End of pack replacement. */
1110
1111         if (delete_redundant && pack_everything & ALL_INTO_ONE) {
1112                 const int hexsz = the_hash_algo->hexsz;
1113                 for_each_string_list_item(item, &existing_nonkept_packs) {
1114                         char *sha1;
1115                         size_t len = strlen(item->string);
1116                         if (len < hexsz)
1117                                 continue;
1118                         sha1 = item->string + len - hexsz;
1119                         /*
1120                          * Mark this pack for deletion, which ensures that this
1121                          * pack won't be included in a MIDX (if `--write-midx`
1122                          * was given) and that we will actually delete this pack
1123                          * (if `-d` was given).
1124                          */
1125                         if (!string_list_has_string(&names, sha1))
1126                                 item->util = (void*)(uintptr_t)((size_t)item->util | DELETE_PACK);
1127                 }
1128         }
1129
1130         if (write_midx) {
1131                 struct string_list include = STRING_LIST_INIT_NODUP;
1132                 midx_included_packs(&include, &existing_nonkept_packs,
1133                                     &existing_kept_packs, &names, geometry);
1134
1135                 ret = write_midx_included_packs(&include, geometry,
1136                                                 refs_snapshot ? get_tempfile_path(refs_snapshot) : NULL,
1137                                                 show_progress, write_bitmaps > 0);
1138
1139                 if (!ret && write_bitmaps)
1140                         remove_redundant_bitmaps(&include, packdir);
1141
1142                 string_list_clear(&include, 0);
1143
1144                 if (ret)
1145                         goto cleanup;
1146         }
1147
1148         reprepare_packed_git(the_repository);
1149
1150         if (delete_redundant) {
1151                 int opts = 0;
1152                 for_each_string_list_item(item, &existing_nonkept_packs) {
1153                         if (!((uintptr_t)item->util & DELETE_PACK))
1154                                 continue;
1155                         remove_redundant_pack(packdir, item->string);
1156                 }
1157
1158                 if (geometry) {
1159                         struct strbuf buf = STRBUF_INIT;
1160
1161                         uint32_t i;
1162                         for (i = 0; i < geometry->split; i++) {
1163                                 struct packed_git *p = geometry->pack[i];
1164                                 if (string_list_has_string(&names,
1165                                                            hash_to_hex(p->hash)))
1166                                         continue;
1167
1168                                 strbuf_reset(&buf);
1169                                 strbuf_addstr(&buf, pack_basename(p));
1170                                 strbuf_strip_suffix(&buf, ".pack");
1171
1172                                 if ((p->pack_keep) ||
1173                                     (string_list_has_string(&existing_kept_packs,
1174                                                             buf.buf)))
1175                                         continue;
1176
1177                                 remove_redundant_pack(packdir, buf.buf);
1178                         }
1179                         strbuf_release(&buf);
1180                 }
1181                 if (show_progress)
1182                         opts |= PRUNE_PACKED_VERBOSE;
1183                 prune_packed_objects(opts);
1184
1185                 if (!keep_unreachable &&
1186                     (!(pack_everything & LOOSEN_UNREACHABLE) ||
1187                      unpack_unreachable) &&
1188                     is_repository_shallow(the_repository))
1189                         prune_shallow(PRUNE_QUICK);
1190         }
1191
1192         if (run_update_server_info)
1193                 update_server_info(0);
1194
1195         if (git_env_bool(GIT_TEST_MULTI_PACK_INDEX, 0)) {
1196                 unsigned flags = 0;
1197                 if (git_env_bool(GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP, 0))
1198                         flags |= MIDX_WRITE_BITMAP | MIDX_WRITE_REV_INDEX;
1199                 write_midx_file(get_object_directory(), NULL, NULL, flags);
1200         }
1201
1202 cleanup:
1203         string_list_clear(&names, 1);
1204         string_list_clear(&existing_nonkept_packs, 0);
1205         string_list_clear(&existing_kept_packs, 0);
1206         clear_pack_geometry(geometry);
1207
1208         return ret;
1209 }