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_largest_active_pack(struct pack_geometry *geometry)
 448 {
 449         if (!geometry) {
 450                 /*
 451                  * No geometry means either an all-into-one repack (in which
 452                  * case there is only one pack left and it is the largest) or an
 453                  * incremental one.
 454                  *
 455                  * If repacking incrementally, then we could check the size of
 456                  * all packs to determine which should be preferred, but leave
 457                  * this for later.
 458                  */
 459                 return NULL;
 460         }
 461         if (geometry->split == geometry->pack_nr)
 462                 return NULL;
 463         return geometry->pack[geometry->pack_nr - 1];
 464 }
 465
 466 static void clear_pack_geometry(struct pack_geometry *geometry)
 467 {
 468         if (!geometry)
 469                 return;
 470
 471         free(geometry->pack);
 472         geometry->pack_nr = 0;
 473         geometry->pack_alloc = 0;
 474         geometry->split = 0;
 475 }
 476
 477 struct midx_snapshot_ref_data {
 478         struct tempfile *f;
 479         struct oidset seen;
 480         int preferred;
 481 };
 482
 483 static int midx_snapshot_ref_one(const char *refname UNUSED,
 484                                  const struct object_id *oid,
 485                                  int flag UNUSED, void *_data)
 486 {
 487         struct midx_snapshot_ref_data *data = _data;
 488         struct object_id peeled;
 489
 490         if (!peel_iterated_oid(oid, &peeled))
 491                 oid = &peeled;
 492
 493         if (oidset_insert(&data->seen, oid))
 494                 return 0; /* already seen */
 495
 496         if (oid_object_info(the_repository, oid, NULL) != OBJ_COMMIT)
 497                 return 0;
 498
 499         fprintf(data->f->fp, "%s%s\n", data->preferred ? "+" : "",
 500                 oid_to_hex(oid));
 501
 502         return 0;
 503 }
 504
 505 static void midx_snapshot_refs(struct tempfile *f)
 506 {
 507         struct midx_snapshot_ref_data data;
 508         const struct string_list *preferred = bitmap_preferred_tips(the_repository);
 509
 510         data.f = f;
 511         data.preferred = 0;
 512         oidset_init(&data.seen, 0);
 513
 514         if (!fdopen_tempfile(f, "w"))
 515                  die(_("could not open tempfile %s for writing"),
 516                      get_tempfile_path(f));
 517
 518         if (preferred) {
 519                 struct string_list_item *item;
 520
 521                 data.preferred = 1;
 522                 for_each_string_list_item(item, preferred)
 523                         for_each_ref_in(item->string, midx_snapshot_ref_one, &data);
 524                 data.preferred = 0;
 525         }
 526
 527         for_each_ref(midx_snapshot_ref_one, &data);
 528
 529         if (close_tempfile_gently(f)) {
 530                 int save_errno = errno;
 531                 delete_tempfile(&f);
 532                 errno = save_errno;
 533                 die_errno(_("could not close refs snapshot tempfile"));
 534         }
 535
 536         oidset_clear(&data.seen);
 537 }
 538
 539 static void midx_included_packs(struct string_list *include,
 540                                 struct string_list *existing_nonkept_packs,
 541                                 struct string_list *existing_kept_packs,
 542                                 struct string_list *names,
 543                                 struct pack_geometry *geometry)
 544 {
 545         struct string_list_item *item;
 546
 547         for_each_string_list_item(item, existing_kept_packs)
 548                 string_list_insert(include, xstrfmt("%s.idx", item->string));
 549         for_each_string_list_item(item, names)
 550                 string_list_insert(include, xstrfmt("pack-%s.idx", item->string));
 551         if (geometry) {
 552                 struct strbuf buf = STRBUF_INIT;
 553                 uint32_t i;
 554                 for (i = geometry->split; i < geometry->pack_nr; i++) {
 555                         struct packed_git *p = geometry->pack[i];
 556
 557                         strbuf_addstr(&buf, pack_basename(p));
 558                         strbuf_strip_suffix(&buf, ".pack");
 559                         strbuf_addstr(&buf, ".idx");
 560
 561                         string_list_insert(include, strbuf_detach(&buf, NULL));
 562                 }
 563
 564                 for_each_string_list_item(item, existing_nonkept_packs) {
 565                         if (!((uintptr_t)item->util & CRUFT_PACK)) {
 566                                 /*
 567                                  * no need to check DELETE_PACK, since we're not
 568                                  * doing an ALL_INTO_ONE repack
 569                                  */
 570                                 continue;
 571                         }
 572                         string_list_insert(include, xstrfmt("%s.idx", item->string));
 573                 }
 574         } else {
 575                 for_each_string_list_item(item, existing_nonkept_packs) {
 576                         if ((uintptr_t)item->util & DELETE_PACK)
 577                                 continue;
 578                         string_list_insert(include, xstrfmt("%s.idx", item->string));
 579                 }
 580         }
 581 }
 582
 583 static int write_midx_included_packs(struct string_list *include,
 584                                      struct pack_geometry *geometry,
 585                                      const char *refs_snapshot,
 586                                      int show_progress, int write_bitmaps)
 587 {
 588         struct child_process cmd = CHILD_PROCESS_INIT;
 589         struct string_list_item *item;
 590         struct packed_git *largest = get_largest_active_pack(geometry);
 591         FILE *in;
 592         int ret;
 593
 594         if (!include->nr)
 595                 return 0;
 596
 597         cmd.in = -1;
 598         cmd.git_cmd = 1;
 599
 600         strvec_push(&cmd.args, "multi-pack-index");
 601         strvec_pushl(&cmd.args, "write", "--stdin-packs", NULL);
 602
 603         if (show_progress)
 604                 strvec_push(&cmd.args, "--progress");
 605         else
 606                 strvec_push(&cmd.args, "--no-progress");
 607
 608         if (write_bitmaps)
 609                 strvec_push(&cmd.args, "--bitmap");
 610
 611         if (largest)
 612                 strvec_pushf(&cmd.args, "--preferred-pack=%s",
 613                              pack_basename(largest));
 614
 615         if (refs_snapshot)
 616                 strvec_pushf(&cmd.args, "--refs-snapshot=%s", refs_snapshot);
 617
 618         ret = start_command(&cmd);
 619         if (ret)
 620                 return ret;
 621
 622         in = xfdopen(cmd.in, "w");
 623         for_each_string_list_item(item, include)
 624                 fprintf(in, "%s\n", item->string);
 625         fclose(in);
 626
 627         return finish_command(&cmd);
 628 }
 629
 630 static void remove_redundant_bitmaps(struct string_list *include,
 631                                      const char *packdir)
 632 {
 633         struct strbuf path = STRBUF_INIT;
 634         struct string_list_item *item;
 635         size_t packdir_len;
 636
 637         strbuf_addstr(&path, packdir);
 638         strbuf_addch(&path, '/');
 639         packdir_len = path.len;
 640
 641         /*
 642          * Remove any pack bitmaps corresponding to packs which are now
 643          * included in the MIDX.
 644          */
 645         for_each_string_list_item(item, include) {
 646                 strbuf_addstr(&path, item->string);
 647                 strbuf_strip_suffix(&path, ".idx");
 648                 strbuf_addstr(&path, ".bitmap");
 649
 650                 if (unlink(path.buf) && errno != ENOENT)
 651                         warning_errno(_("could not remove stale bitmap: %s"),
 652                                       path.buf);
 653
 654                 strbuf_setlen(&path, packdir_len);
 655         }
 656         strbuf_release(&path);
 657 }
 658
 659 static int write_cruft_pack(const struct pack_objects_args *args,
 660                             const char *destination,
 661                             const char *pack_prefix,
 662                             const char *cruft_expiration,
 663                             struct string_list *names,
 664                             struct string_list *existing_packs,
 665                             struct string_list *existing_kept_packs)
 666 {
 667         struct child_process cmd = CHILD_PROCESS_INIT;
 668         struct strbuf line = STRBUF_INIT;
 669         struct string_list_item *item;
 670         FILE *in, *out;
 671         int ret;
 672         const char *scratch;
 673         int local = skip_prefix(destination, packdir, &scratch);
 674
 675         prepare_pack_objects(&cmd, args, destination);
 676
 677         strvec_push(&cmd.args, "--cruft");
 678         if (cruft_expiration)
 679                 strvec_pushf(&cmd.args, "--cruft-expiration=%s",
 680                              cruft_expiration);
 681
 682         strvec_push(&cmd.args, "--honor-pack-keep");
 683         strvec_push(&cmd.args, "--non-empty");
 684         strvec_push(&cmd.args, "--max-pack-size=0");
 685
 686         cmd.in = -1;
 687
 688         ret = start_command(&cmd);
 689         if (ret)
 690                 return ret;
 691
 692         /*
 693          * names has a confusing double use: it both provides the list
 694          * of just-written new packs, and accepts the name of the cruft
 695          * pack we are writing.
 696          *
 697          * By the time it is read here, it contains only the pack(s)
 698          * that were just written, which is exactly the set of packs we
 699          * want to consider kept.
 700          *
 701          * If `--expire-to` is given, the double-use served by `names`
 702          * ensures that the pack written to `--expire-to` excludes any
 703          * objects contained in the cruft pack.
 704          */
 705         in = xfdopen(cmd.in, "w");
 706         for_each_string_list_item(item, names)
 707                 fprintf(in, "%s-%s.pack\n", pack_prefix, item->string);
 708         for_each_string_list_item(item, existing_packs)
 709                 fprintf(in, "-%s.pack\n", item->string);
 710         for_each_string_list_item(item, existing_kept_packs)
 711                 fprintf(in, "%s.pack\n", item->string);
 712         fclose(in);
 713
 714         out = xfdopen(cmd.out, "r");
 715         while (strbuf_getline_lf(&line, out) != EOF) {
 716                 struct string_list_item *item;
 717
 718                 if (line.len != the_hash_algo->hexsz)
 719                         die(_("repack: Expecting full hex object ID lines only "
 720                               "from pack-objects."));
 721                 /*
 722                  * avoid putting packs written outside of the repository in the
 723                  * list of names
 724                  */
 725                 if (local) {
 726                         item = string_list_append(names, line.buf);
 727                         item->util = populate_pack_exts(line.buf);
 728                 }
 729         }
 730         fclose(out);
 731
 732         strbuf_release(&line);
 733
 734         return finish_command(&cmd);
 735 }
 736
 737 int cmd_repack(int argc, const char **argv, const char *prefix)
 738 {
 739         struct child_process cmd = CHILD_PROCESS_INIT;
 740         struct string_list_item *item;
 741         struct string_list names = STRING_LIST_INIT_DUP;
 742         struct string_list existing_nonkept_packs = STRING_LIST_INIT_DUP;
 743         struct string_list existing_kept_packs = STRING_LIST_INIT_DUP;
 744         struct pack_geometry *geometry = NULL;
 745         struct strbuf line = STRBUF_INIT;
 746         struct tempfile *refs_snapshot = NULL;
 747         int i, ext, ret;
 748         FILE *out;
 749         int show_progress;
 750
 751         /* variables to be filled by option parsing */
 752         int delete_redundant = 0;
 753         const char *unpack_unreachable = NULL;
 754         int keep_unreachable = 0;
 755         struct string_list keep_pack_list = STRING_LIST_INIT_NODUP;
 756         struct pack_objects_args po_args = {NULL};
 757         struct pack_objects_args cruft_po_args = {NULL};
 758         int geometric_factor = 0;
 759         int write_midx = 0;
 760         const char *cruft_expiration = NULL;
 761         const char *expire_to = NULL;
 762
 763         struct option builtin_repack_options[] = {
 764                 OPT_BIT('a', NULL, &pack_everything,
 765                                 N_("pack everything in a single pack"), ALL_INTO_ONE),
 766                 OPT_BIT('A', NULL, &pack_everything,
 767                                 N_("same as -a, and turn unreachable objects loose"),
 768                                    LOOSEN_UNREACHABLE | ALL_INTO_ONE),
 769                 OPT_BIT(0, "cruft", &pack_everything,
 770                                 N_("same as -a, pack unreachable cruft objects separately"),
 771                                    PACK_CRUFT),
 772                 OPT_STRING(0, "cruft-expiration", &cruft_expiration, N_("approxidate"),
 773                                 N_("with -C, expire objects older than this")),
 774                 OPT_BOOL('d', NULL, &delete_redundant,
 775                                 N_("remove redundant packs, and run git-prune-packed")),
 776                 OPT_BOOL('f', NULL, &po_args.no_reuse_delta,
 777                                 N_("pass --no-reuse-delta to git-pack-objects")),
 778                 OPT_BOOL('F', NULL, &po_args.no_reuse_object,
 779                                 N_("pass --no-reuse-object to git-pack-objects")),
 780                 OPT_NEGBIT('n', NULL, &run_update_server_info,
 781                                 N_("do not run git-update-server-info"), 1),
 782                 OPT__QUIET(&po_args.quiet, N_("be quiet")),
 783                 OPT_BOOL('l', "local", &po_args.local,
 784                                 N_("pass --local to git-pack-objects")),
 785                 OPT_BOOL('b', "write-bitmap-index", &write_bitmaps,
 786                                 N_("write bitmap index")),
 787                 OPT_BOOL('i', "delta-islands", &use_delta_islands,
 788                                 N_("pass --delta-islands to git-pack-objects")),
 789                 OPT_STRING(0, "unpack-unreachable", &unpack_unreachable, N_("approxidate"),
 790                                 N_("with -A, do not loosen objects older than this")),
 791                 OPT_BOOL('k', "keep-unreachable", &keep_unreachable,
 792                                 N_("with -a, repack unreachable objects")),
 793                 OPT_STRING(0, "window", &po_args.window, N_("n"),
 794                                 N_("size of the window used for delta compression")),
 795                 OPT_STRING(0, "window-memory", &po_args.window_memory, N_("bytes"),
 796                                 N_("same as the above, but limit memory size instead of entries count")),
 797                 OPT_STRING(0, "depth", &po_args.depth, N_("n"),
 798                                 N_("limits the maximum delta depth")),
 799                 OPT_STRING(0, "threads", &po_args.threads, N_("n"),
 800                                 N_("limits the maximum number of threads")),
 801                 OPT_STRING(0, "max-pack-size", &po_args.max_pack_size, N_("bytes"),
 802                                 N_("maximum size of each packfile")),
 803                 OPT_BOOL(0, "pack-kept-objects", &pack_kept_objects,
 804                                 N_("repack objects in packs marked with .keep")),
 805                 OPT_STRING_LIST(0, "keep-pack", &keep_pack_list, N_("name"),
 806                                 N_("do not repack this pack")),
 807                 OPT_INTEGER('g', "geometric", &geometric_factor,
 808                             N_("find a geometric progression with factor <N>")),
 809                 OPT_BOOL('m', "write-midx", &write_midx,
 810                            N_("write a multi-pack index of the resulting packs")),
 811                 OPT_STRING(0, "expire-to", &expire_to, N_("dir"),
 812                            N_("pack prefix to store a pack containing pruned objects")),
 813                 OPT_END()
 814         };
 815
 816         git_config(repack_config, &cruft_po_args);
 817
 818         argc = parse_options(argc, argv, prefix, builtin_repack_options,
 819                                 git_repack_usage, 0);
 820
 821         if (delete_redundant && repository_format_precious_objects)
 822                 die(_("cannot delete packs in a precious-objects repo"));
 823
 824         if (keep_unreachable &&
 825             (unpack_unreachable || (pack_everything & LOOSEN_UNREACHABLE)))
 826                 die(_("options '%s' and '%s' cannot be used together"), "--keep-unreachable", "-A");
 827
 828         if (pack_everything & PACK_CRUFT) {
 829                 pack_everything |= ALL_INTO_ONE;
 830
 831                 if (unpack_unreachable || (pack_everything & LOOSEN_UNREACHABLE))
 832                         die(_("options '%s' and '%s' cannot be used together"), "--cruft", "-A");
 833                 if (keep_unreachable)
 834                         die(_("options '%s' and '%s' cannot be used together"), "--cruft", "-k");
 835         }
 836
 837         if (write_bitmaps < 0) {
 838                 if (!write_midx &&
 839                     (!(pack_everything & ALL_INTO_ONE) || !is_bare_repository()))
 840                         write_bitmaps = 0;
 841         } else if (write_bitmaps &&
 842                    git_env_bool(GIT_TEST_MULTI_PACK_INDEX, 0) &&
 843                    git_env_bool(GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP, 0)) {
 844                 write_bitmaps = 0;
 845         }
 846         if (pack_kept_objects < 0)
 847                 pack_kept_objects = write_bitmaps > 0 && !write_midx;
 848
 849         if (write_bitmaps && !(pack_everything & ALL_INTO_ONE) && !write_midx)
 850                 die(_(incremental_bitmap_conflict_error));
 851
 852         if (write_midx && write_bitmaps) {
 853                 struct strbuf path = STRBUF_INIT;
 854
 855                 strbuf_addf(&path, "%s/%s_XXXXXX", get_object_directory(),
 856                             "bitmap-ref-tips");
 857
 858                 refs_snapshot = xmks_tempfile(path.buf);
 859                 midx_snapshot_refs(refs_snapshot);
 860
 861                 strbuf_release(&path);
 862         }
 863
 864         packdir = mkpathdup("%s/pack", get_object_directory());
 865         packtmp_name = xstrfmt(".tmp-%d-pack", (int)getpid());
 866         packtmp = mkpathdup("%s/%s", packdir, packtmp_name);
 867
 868         collect_pack_filenames(&existing_nonkept_packs, &existing_kept_packs,
 869                                &keep_pack_list);
 870
 871         if (geometric_factor) {
 872                 if (pack_everything)
 873                         die(_("options '%s' and '%s' cannot be used together"), "--geometric", "-A/-a");
 874                 init_pack_geometry(&geometry, &existing_kept_packs);
 875                 split_pack_geometry(geometry, geometric_factor);
 876         }
 877
 878         prepare_pack_objects(&cmd, &po_args, packtmp);
 879
 880         show_progress = !po_args.quiet && isatty(2);
 881
 882         strvec_push(&cmd.args, "--keep-true-parents");
 883         if (!pack_kept_objects)
 884                 strvec_push(&cmd.args, "--honor-pack-keep");
 885         for (i = 0; i < keep_pack_list.nr; i++)
 886                 strvec_pushf(&cmd.args, "--keep-pack=%s",
 887                              keep_pack_list.items[i].string);
 888         strvec_push(&cmd.args, "--non-empty");
 889         if (!geometry) {
 890                 /*
 891                  * We need to grab all reachable objects, including those that
 892                  * are reachable from reflogs and the index.
 893                  *
 894                  * When repacking into a geometric progression of packs,
 895                  * however, we ask 'git pack-objects --stdin-packs', and it is
 896                  * not about packing objects based on reachability but about
 897                  * repacking all the objects in specified packs and loose ones
 898                  * (indeed, --stdin-packs is incompatible with these options).
 899                  */
 900                 strvec_push(&cmd.args, "--all");
 901                 strvec_push(&cmd.args, "--reflog");
 902                 strvec_push(&cmd.args, "--indexed-objects");
 903         }
 904         if (has_promisor_remote())
 905                 strvec_push(&cmd.args, "--exclude-promisor-objects");
 906         if (!write_midx) {
 907                 if (write_bitmaps > 0)
 908                         strvec_push(&cmd.args, "--write-bitmap-index");
 909                 else if (write_bitmaps < 0)
 910                         strvec_push(&cmd.args, "--write-bitmap-index-quiet");
 911         }
 912         if (use_delta_islands)
 913                 strvec_push(&cmd.args, "--delta-islands");
 914
 915         if (pack_everything & ALL_INTO_ONE) {
 916                 repack_promisor_objects(&po_args, &names);
 917
 918                 if (existing_nonkept_packs.nr && delete_redundant &&
 919                     !(pack_everything & PACK_CRUFT)) {
 920                         for_each_string_list_item(item, &names) {
 921                                 strvec_pushf(&cmd.args, "--keep-pack=%s-%s.pack",
 922                                              packtmp_name, item->string);
 923                         }
 924                         if (unpack_unreachable) {
 925                                 strvec_pushf(&cmd.args,
 926                                              "--unpack-unreachable=%s",
 927                                              unpack_unreachable);
 928                         } else if (pack_everything & LOOSEN_UNREACHABLE) {
 929                                 strvec_push(&cmd.args,
 930                                             "--unpack-unreachable");
 931                         } else if (keep_unreachable) {
 932                                 strvec_push(&cmd.args, "--keep-unreachable");
 933                                 strvec_push(&cmd.args, "--pack-loose-unreachable");
 934                         }
 935                 }
 936         } else if (geometry) {
 937                 strvec_push(&cmd.args, "--stdin-packs");
 938                 strvec_push(&cmd.args, "--unpacked");
 939         } else {
 940                 strvec_push(&cmd.args, "--unpacked");
 941                 strvec_push(&cmd.args, "--incremental");
 942         }
 943
 944         if (geometry)
 945                 cmd.in = -1;
 946         else
 947                 cmd.no_stdin = 1;
 948
 949         ret = start_command(&cmd);
 950         if (ret)
 951                 goto cleanup;
 952
 953         if (geometry) {
 954                 FILE *in = xfdopen(cmd.in, "w");
 955                 /*
 956                  * The resulting pack should contain all objects in packs that
 957                  * are going to be rolled up, but exclude objects in packs which
 958                  * are being left alone.
 959                  */
 960                 for (i = 0; i < geometry->split; i++)
 961                         fprintf(in, "%s\n", pack_basename(geometry->pack[i]));
 962                 for (i = geometry->split; i < geometry->pack_nr; i++)
 963                         fprintf(in, "^%s\n", pack_basename(geometry->pack[i]));
 964                 fclose(in);
 965         }
 966
 967         out = xfdopen(cmd.out, "r");
 968         while (strbuf_getline_lf(&line, out) != EOF) {
 969                 struct string_list_item *item;
 970
 971                 if (line.len != the_hash_algo->hexsz)
 972                         die(_("repack: Expecting full hex object ID lines only from pack-objects."));
 973                 item = string_list_append(&names, line.buf);
 974                 item->util = populate_pack_exts(item->string);
 975         }
 976         strbuf_release(&line);
 977         fclose(out);
 978         ret = finish_command(&cmd);
 979         if (ret)
 980                 goto cleanup;
 981
 982         if (!names.nr && !po_args.quiet)
 983                 printf_ln(_("Nothing new to pack."));
 984
 985         if (pack_everything & PACK_CRUFT) {
 986                 const char *pack_prefix;
 987                 if (!skip_prefix(packtmp, packdir, &pack_prefix))
 988                         die(_("pack prefix %s does not begin with objdir %s"),
 989                             packtmp, packdir);
 990                 if (*pack_prefix == '/')
 991                         pack_prefix++;
 992
 993                 if (!cruft_po_args.window)
 994                         cruft_po_args.window = po_args.window;
 995                 if (!cruft_po_args.window_memory)
 996                         cruft_po_args.window_memory = po_args.window_memory;
 997                 if (!cruft_po_args.depth)
 998                         cruft_po_args.depth = po_args.depth;
 999                 if (!cruft_po_args.threads)
1000                         cruft_po_args.threads = po_args.threads;
1001
1002                 cruft_po_args.local = po_args.local;
1003                 cruft_po_args.quiet = po_args.quiet;
1004
1005                 ret = write_cruft_pack(&cruft_po_args, packtmp, pack_prefix,
1006                                        cruft_expiration, &names,
1007                                        &existing_nonkept_packs,
1008                                        &existing_kept_packs);
1009                 if (ret)
1010                         goto cleanup;
1011
1012                 if (delete_redundant && expire_to) {
1013                         /*
1014                          * If `--expire-to` is given with `-d`, it's possible
1015                          * that we're about to prune some objects. With cruft
1016                          * packs, pruning is implicit: any objects from existing
1017                          * packs that weren't picked up by new packs are removed
1018                          * when their packs are deleted.
1019                          *
1020                          * Generate an additional cruft pack, with one twist:
1021                          * `names` now includes the name of the cruft pack
1022                          * written in the previous step. So the contents of
1023                          * _this_ cruft pack exclude everything contained in the
1024                          * existing cruft pack (that is, all of the unreachable
1025                          * objects which are no older than
1026                          * `--cruft-expiration`).
1027                          *
1028                          * To make this work, cruft_expiration must become NULL
1029                          * so that this cruft pack doesn't actually prune any
1030                          * objects. If it were non-NULL, this call would always
1031                          * generate an empty pack (since every object not in the
1032                          * cruft pack generated above will have an mtime older
1033                          * than the expiration).
1034                          */
1035                         ret = write_cruft_pack(&cruft_po_args, expire_to,
1036                                                pack_prefix,
1037                                                NULL,
1038                                                &names,
1039                                                &existing_nonkept_packs,
1040                                                &existing_kept_packs);
1041                         if (ret)
1042                                 goto cleanup;
1043                 }
1044         }
1045
1046         string_list_sort(&names);
1047
1048         close_object_store(the_repository->objects);
1049
1050         /*
1051          * Ok we have prepared all new packfiles.
1052          */
1053         for_each_string_list_item(item, &names) {
1054                 struct generated_pack_data *data = item->util;
1055
1056                 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
1057                         char *fname;
1058
1059                         fname = mkpathdup("%s/pack-%s%s",
1060                                         packdir, item->string, exts[ext].name);
1061
1062                         if (data->tempfiles[ext]) {
1063                                 const char *fname_old = get_tempfile_path(data->tempfiles[ext]);
1064                                 struct stat statbuffer;
1065
1066                                 if (!stat(fname_old, &statbuffer)) {
1067                                         statbuffer.st_mode &= ~(S_IWUSR | S_IWGRP | S_IWOTH);
1068                                         chmod(fname_old, statbuffer.st_mode);
1069                                 }
1070
1071                                 if (rename_tempfile(&data->tempfiles[ext], fname))
1072                                         die_errno(_("renaming pack to '%s' failed"), fname);
1073                         } else if (!exts[ext].optional)
1074                                 die(_("pack-objects did not write a '%s' file for pack %s-%s"),
1075                                     exts[ext].name, packtmp, item->string);
1076                         else if (unlink(fname) < 0 && errno != ENOENT)
1077                                 die_errno(_("could not unlink: %s"), fname);
1078
1079                         free(fname);
1080                 }
1081         }
1082         /* End of pack replacement. */
1083
1084         if (delete_redundant && pack_everything & ALL_INTO_ONE) {
1085                 const int hexsz = the_hash_algo->hexsz;
1086                 for_each_string_list_item(item, &existing_nonkept_packs) {
1087                         char *sha1;
1088                         size_t len = strlen(item->string);
1089                         if (len < hexsz)
1090                                 continue;
1091                         sha1 = item->string + len - hexsz;
1092                         /*
1093                          * Mark this pack for deletion, which ensures that this
1094                          * pack won't be included in a MIDX (if `--write-midx`
1095                          * was given) and that we will actually delete this pack
1096                          * (if `-d` was given).
1097                          */
1098                         if (!string_list_has_string(&names, sha1))
1099                                 item->util = (void*)(uintptr_t)((size_t)item->util | DELETE_PACK);
1100                 }
1101         }
1102
1103         if (write_midx) {
1104                 struct string_list include = STRING_LIST_INIT_NODUP;
1105                 midx_included_packs(&include, &existing_nonkept_packs,
1106                                     &existing_kept_packs, &names, geometry);
1107
1108                 ret = write_midx_included_packs(&include, geometry,
1109                                                 refs_snapshot ? get_tempfile_path(refs_snapshot) : NULL,
1110                                                 show_progress, write_bitmaps > 0);
1111
1112                 if (!ret && write_bitmaps)
1113                         remove_redundant_bitmaps(&include, packdir);
1114
1115                 string_list_clear(&include, 0);
1116
1117                 if (ret)
1118                         goto cleanup;
1119         }
1120
1121         reprepare_packed_git(the_repository);
1122
1123         if (delete_redundant) {
1124                 int opts = 0;
1125                 for_each_string_list_item(item, &existing_nonkept_packs) {
1126                         if (!((uintptr_t)item->util & DELETE_PACK))
1127                                 continue;
1128                         remove_redundant_pack(packdir, item->string);
1129                 }
1130
1131                 if (geometry) {
1132                         struct strbuf buf = STRBUF_INIT;
1133
1134                         uint32_t i;
1135                         for (i = 0; i < geometry->split; i++) {
1136                                 struct packed_git *p = geometry->pack[i];
1137                                 if (string_list_has_string(&names,
1138                                                            hash_to_hex(p->hash)))
1139                                         continue;
1140
1141                                 strbuf_reset(&buf);
1142                                 strbuf_addstr(&buf, pack_basename(p));
1143                                 strbuf_strip_suffix(&buf, ".pack");
1144
1145                                 if ((p->pack_keep) ||
1146                                     (string_list_has_string(&existing_kept_packs,
1147                                                             buf.buf)))
1148                                         continue;
1149
1150                                 remove_redundant_pack(packdir, buf.buf);
1151                         }
1152                         strbuf_release(&buf);
1153                 }
1154                 if (show_progress)
1155                         opts |= PRUNE_PACKED_VERBOSE;
1156                 prune_packed_objects(opts);
1157
1158                 if (!keep_unreachable &&
1159                     (!(pack_everything & LOOSEN_UNREACHABLE) ||
1160                      unpack_unreachable) &&
1161                     is_repository_shallow(the_repository))
1162                         prune_shallow(PRUNE_QUICK);
1163         }
1164
1165         if (run_update_server_info)
1166                 update_server_info(0);
1167
1168         if (git_env_bool(GIT_TEST_MULTI_PACK_INDEX, 0)) {
1169                 unsigned flags = 0;
1170                 if (git_env_bool(GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP, 0))
1171                         flags |= MIDX_WRITE_BITMAP | MIDX_WRITE_REV_INDEX;
1172                 write_midx_file(get_object_directory(), NULL, NULL, flags);
1173         }
1174
1175 cleanup:
1176         string_list_clear(&names, 1);
1177         string_list_clear(&existing_nonkept_packs, 0);
1178         string_list_clear(&existing_kept_packs, 0);
1179         clear_pack_geometry(geometry);
1180
1181         return ret;
1182 }