builtin/repack.c

   1 #include "builtin.h"
   2 #include "alloc.h"
   3 #include "config.h"
   4 #include "dir.h"
   5 #include "environment.h"
   6 #include "gettext.h"
   7 #include "hex.h"
   8 #include "parse-options.h"
   9 #include "run-command.h"
  10 #include "sigchain.h"
  11 #include "strbuf.h"
  12 #include "string-list.h"
  13 #include "strvec.h"
  14 #include "midx.h"
  15 #include "packfile.h"
  16 #include "prune-packed.h"
  17 #include "object-store.h"
  18 #include "promisor-remote.h"
  19 #include "shallow.h"
  20 #include "pack.h"
  21 #include "pack-bitmap.h"
  22 #include "refs.h"
  23
  24 #define ALL_INTO_ONE 1
  25 #define LOOSEN_UNREACHABLE 2
  26 #define PACK_CRUFT 4
  27
  28 #define DELETE_PACK 1
  29 #define CRUFT_PACK 2
  30
  31 static int pack_everything;
  32 static int delta_base_offset = 1;
  33 static int pack_kept_objects = -1;
  34 static int write_bitmaps = -1;
  35 static int use_delta_islands;
  36 static int run_update_server_info = 1;
  37 static char *packdir, *packtmp_name, *packtmp;
  38
  39 static const char *const git_repack_usage[] = {
  40         N_("git repack [<options>]"),
  41         NULL
  42 };
  43
  44 static const char incremental_bitmap_conflict_error[] = N_(
  45 "Incremental repacks are incompatible with bitmap indexes.  Use\n"
  46 "--no-write-bitmap-index or disable the pack.writeBitmaps configuration."
  47 );
  48
  49 struct pack_objects_args {
  50         const char *window;
  51         const char *window_memory;
  52         const char *depth;
  53         const char *threads;
  54         const char *max_pack_size;
  55         int no_reuse_delta;
  56         int no_reuse_object;
  57         int quiet;
  58         int local;
  59 };
  60
  61 static int repack_config(const char *var, const char *value, void *cb)
  62 {
  63         struct pack_objects_args *cruft_po_args = cb;
  64         if (!strcmp(var, "repack.usedeltabaseoffset")) {
  65                 delta_base_offset = git_config_bool(var, value);
  66                 return 0;
  67         }
  68         if (!strcmp(var, "repack.packkeptobjects")) {
  69                 pack_kept_objects = git_config_bool(var, value);
  70                 return 0;
  71         }
  72         if (!strcmp(var, "repack.writebitmaps") ||
  73             !strcmp(var, "pack.writebitmaps")) {
  74                 write_bitmaps = git_config_bool(var, value);
  75                 return 0;
  76         }
  77         if (!strcmp(var, "repack.usedeltaislands")) {
  78                 use_delta_islands = git_config_bool(var, value);
  79                 return 0;
  80         }
  81         if (strcmp(var, "repack.updateserverinfo") == 0) {
  82                 run_update_server_info = git_config_bool(var, value);
  83                 return 0;
  84         }
  85         if (!strcmp(var, "repack.cruftwindow"))
  86                 return git_config_string(&cruft_po_args->window, var, value);
  87         if (!strcmp(var, "repack.cruftwindowmemory"))
  88                 return git_config_string(&cruft_po_args->window_memory, var, value);
  89         if (!strcmp(var, "repack.cruftdepth"))
  90                 return git_config_string(&cruft_po_args->depth, var, value);
  91         if (!strcmp(var, "repack.cruftthreads"))
  92                 return git_config_string(&cruft_po_args->threads, var, value);
  93         return git_default_config(var, value, cb);
  94 }
  95
  96 /*
  97  * Adds all packs hex strings to either fname_nonkept_list or
  98  * fname_kept_list based on whether each pack has a corresponding
  99  * .keep file or not.  Packs without a .keep file are not to be kept
 100  * if we are going to pack everything into one file.
 101  */
 102 static void collect_pack_filenames(struct string_list *fname_nonkept_list,
 103                                    struct string_list *fname_kept_list,
 104                                    const struct string_list *extra_keep)
 105 {
 106         DIR *dir;
 107         struct dirent *e;
 108         char *fname;
 109
 110         if (!(dir = opendir(packdir)))
 111                 return;
 112
 113         while ((e = readdir(dir)) != NULL) {
 114                 size_t len;
 115                 int i;
 116
 117                 if (!strip_suffix(e->d_name, ".pack", &len))
 118                         continue;
 119
 120                 for (i = 0; i < extra_keep->nr; i++)
 121                         if (!fspathcmp(e->d_name, extra_keep->items[i].string))
 122                                 break;
 123
 124                 fname = xmemdupz(e->d_name, len);
 125
 126                 if ((extra_keep->nr > 0 && i < extra_keep->nr) ||
 127                     (file_exists(mkpath("%s/%s.keep", packdir, fname)))) {
 128                         string_list_append_nodup(fname_kept_list, fname);
 129                 } else {
 130                         struct string_list_item *item;
 131                         item = string_list_append_nodup(fname_nonkept_list,
 132                                                         fname);
 133                         if (file_exists(mkpath("%s/%s.mtimes", packdir, fname)))
 134                                 item->util = (void*)(uintptr_t)CRUFT_PACK;
 135                 }
 136         }
 137         closedir(dir);
 138
 139         string_list_sort(fname_kept_list);
 140 }
 141
 142 static void remove_redundant_pack(const char *dir_name, const char *base_name)
 143 {
 144         struct strbuf buf = STRBUF_INIT;
 145         struct multi_pack_index *m = get_local_multi_pack_index(the_repository);
 146         strbuf_addf(&buf, "%s.pack", base_name);
 147         if (m && midx_contains_pack(m, buf.buf))
 148                 clear_midx_file(the_repository);
 149         strbuf_insertf(&buf, 0, "%s/", dir_name);
 150         unlink_pack_path(buf.buf, 1);
 151         strbuf_release(&buf);
 152 }
 153
 154 static void prepare_pack_objects(struct child_process *cmd,
 155                                  const struct pack_objects_args *args,
 156                                  const char *out)
 157 {
 158         strvec_push(&cmd->args, "pack-objects");
 159         if (args->window)
 160                 strvec_pushf(&cmd->args, "--window=%s", args->window);
 161         if (args->window_memory)
 162                 strvec_pushf(&cmd->args, "--window-memory=%s", args->window_memory);
 163         if (args->depth)
 164                 strvec_pushf(&cmd->args, "--depth=%s", args->depth);
 165         if (args->threads)
 166                 strvec_pushf(&cmd->args, "--threads=%s", args->threads);
 167         if (args->max_pack_size)
 168                 strvec_pushf(&cmd->args, "--max-pack-size=%s", args->max_pack_size);
 169         if (args->no_reuse_delta)
 170                 strvec_pushf(&cmd->args, "--no-reuse-delta");
 171         if (args->no_reuse_object)
 172                 strvec_pushf(&cmd->args, "--no-reuse-object");
 173         if (args->local)
 174                 strvec_push(&cmd->args,  "--local");
 175         if (args->quiet)
 176                 strvec_push(&cmd->args,  "--quiet");
 177         if (delta_base_offset)
 178                 strvec_push(&cmd->args,  "--delta-base-offset");
 179         strvec_push(&cmd->args, out);
 180         cmd->git_cmd = 1;
 181         cmd->out = -1;
 182 }
 183
 184 /*
 185  * Write oid to the given struct child_process's stdin, starting it first if
 186  * necessary.
 187  */
 188 static int write_oid(const struct object_id *oid,
 189                      struct packed_git *pack UNUSED,
 190                      uint32_t pos UNUSED, void *data)
 191 {
 192         struct child_process *cmd = data;
 193
 194         if (cmd->in == -1) {
 195                 if (start_command(cmd))
 196                         die(_("could not start pack-objects to repack promisor objects"));
 197         }
 198
 199         xwrite(cmd->in, oid_to_hex(oid), the_hash_algo->hexsz);
 200         xwrite(cmd->in, "\n", 1);
 201         return 0;
 202 }
 203
 204 static struct {
 205         const char *name;
 206         unsigned optional:1;
 207 } exts[] = {
 208         {".pack"},
 209         {".rev", 1},
 210         {".mtimes", 1},
 211         {".bitmap", 1},
 212         {".promisor", 1},
 213         {".idx"},
 214 };
 215
 216 struct generated_pack_data {
 217         struct tempfile *tempfiles[ARRAY_SIZE(exts)];
 218 };
 219
 220 static struct generated_pack_data *populate_pack_exts(const char *name)
 221 {
 222         struct stat statbuf;
 223         struct strbuf path = STRBUF_INIT;
 224         struct generated_pack_data *data = xcalloc(1, sizeof(*data));
 225         int i;
 226
 227         for (i = 0; i < ARRAY_SIZE(exts); i++) {
 228                 strbuf_reset(&path);
 229                 strbuf_addf(&path, "%s-%s%s", packtmp, name, exts[i].name);
 230
 231                 if (stat(path.buf, &statbuf))
 232                         continue;
 233
 234                 data->tempfiles[i] = register_tempfile(path.buf);
 235         }
 236
 237         strbuf_release(&path);
 238         return data;
 239 }
 240
 241 static void repack_promisor_objects(const struct pack_objects_args *args,
 242                                     struct string_list *names)
 243 {
 244         struct child_process cmd = CHILD_PROCESS_INIT;
 245         FILE *out;
 246         struct strbuf line = STRBUF_INIT;
 247
 248         prepare_pack_objects(&cmd, args, packtmp);
 249         cmd.in = -1;
 250
 251         /*
 252          * NEEDSWORK: Giving pack-objects only the OIDs without any ordering
 253          * hints may result in suboptimal deltas in the resulting pack. See if
 254          * the OIDs can be sent with fake paths such that pack-objects can use a
 255          * {type -> existing pack order} ordering when computing deltas instead
 256          * of a {type -> size} ordering, which may produce better deltas.
 257          */
 258         for_each_packed_object(write_oid, &cmd,
 259                                FOR_EACH_OBJECT_PROMISOR_ONLY);
 260
 261         if (cmd.in == -1) {
 262                 /* No packed objects; cmd was never started */
 263                 child_process_clear(&cmd);
 264                 return;
 265         }
 266
 267         close(cmd.in);
 268
 269         out = xfdopen(cmd.out, "r");
 270         while (strbuf_getline_lf(&line, out) != EOF) {
 271                 struct string_list_item *item;
 272                 char *promisor_name;
 273
 274                 if (line.len != the_hash_algo->hexsz)
 275                         die(_("repack: Expecting full hex object ID lines only from pack-objects."));
 276                 item = string_list_append(names, line.buf);
 277
 278                 /*
 279                  * pack-objects creates the .pack and .idx files, but not the
 280                  * .promisor file. Create the .promisor file, which is empty.
 281                  *
 282                  * NEEDSWORK: fetch-pack sometimes generates non-empty
 283                  * .promisor files containing the ref names and associated
 284                  * hashes at the point of generation of the corresponding
 285                  * packfile, but this would not preserve their contents. Maybe
 286                  * concatenate the contents of all .promisor files instead of
 287                  * just creating a new empty file.
 288                  */
 289                 promisor_name = mkpathdup("%s-%s.promisor", packtmp,
 290                                           line.buf);
 291                 write_promisor_file(promisor_name, NULL, 0);
 292
 293                 item->util = populate_pack_exts(item->string);
 294
 295                 free(promisor_name);
 296         }
 297         fclose(out);
 298         if (finish_command(&cmd))
 299                 die(_("could not finish pack-objects to repack promisor objects"));
 300 }
 301
 302 struct pack_geometry {
 303         struct packed_git **pack;
 304         uint32_t pack_nr, pack_alloc;
 305         uint32_t split;
 306 };
 307
 308 static uint32_t geometry_pack_weight(struct packed_git *p)
 309 {
 310         if (open_pack_index(p))
 311                 die(_("cannot open index for %s"), p->pack_name);
 312         return p->num_objects;
 313 }
 314
 315 static int geometry_cmp(const void *va, const void *vb)
 316 {
 317         uint32_t aw = geometry_pack_weight(*(struct packed_git **)va),
 318                  bw = geometry_pack_weight(*(struct packed_git **)vb);
 319
 320         if (aw < bw)
 321                 return -1;
 322         if (aw > bw)
 323                 return 1;
 324         return 0;
 325 }
 326
 327 static void init_pack_geometry(struct pack_geometry **geometry_p,
 328                                struct string_list *existing_kept_packs)
 329 {
 330         struct packed_git *p;
 331         struct pack_geometry *geometry;
 332         struct strbuf buf = STRBUF_INIT;
 333
 334         *geometry_p = xcalloc(1, sizeof(struct pack_geometry));
 335         geometry = *geometry_p;
 336
 337         for (p = get_all_packs(the_repository); p; p = p->next) {
 338                 if (!pack_kept_objects) {
 339                         /*
 340                          * Any pack that has its pack_keep bit set will appear
 341                          * in existing_kept_packs below, but this saves us from
 342                          * doing a more expensive check.
 343                          */
 344                         if (p->pack_keep)
 345                                 continue;
 346
 347                         /*
 348                          * The pack may be kept via the --keep-pack option;
 349                          * check 'existing_kept_packs' to determine whether to
 350                          * ignore it.
 351                          */
 352                         strbuf_reset(&buf);
 353                         strbuf_addstr(&buf, pack_basename(p));
 354                         strbuf_strip_suffix(&buf, ".pack");
 355
 356                         if (string_list_has_string(existing_kept_packs, buf.buf))
 357                                 continue;
 358                 }
 359                 if (p->is_cruft)
 360                         continue;
 361
 362                 ALLOC_GROW(geometry->pack,
 363                            geometry->pack_nr + 1,
 364                            geometry->pack_alloc);
 365
 366                 geometry->pack[geometry->pack_nr] = p;
 367                 geometry->pack_nr++;
 368         }
 369
 370         QSORT(geometry->pack, geometry->pack_nr, geometry_cmp);
 371         strbuf_release(&buf);
 372 }
 373
 374 static void split_pack_geometry(struct pack_geometry *geometry, int factor)
 375 {
 376         uint32_t i;
 377         uint32_t split;
 378         off_t total_size = 0;
 379
 380         if (!geometry->pack_nr) {
 381                 geometry->split = geometry->pack_nr;
 382                 return;
 383         }
 384
 385         /*
 386          * First, count the number of packs (in descending order of size) which
 387          * already form a geometric progression.
 388          */
 389         for (i = geometry->pack_nr - 1; i > 0; i--) {
 390                 struct packed_git *ours = geometry->pack[i];
 391                 struct packed_git *prev = geometry->pack[i - 1];
 392
 393                 if (unsigned_mult_overflows(factor, geometry_pack_weight(prev)))
 394                         die(_("pack %s too large to consider in geometric "
 395                               "progression"),
 396                             prev->pack_name);
 397
 398                 if (geometry_pack_weight(ours) < factor * geometry_pack_weight(prev))
 399                         break;
 400         }
 401
 402         split = i;
 403
 404         if (split) {
 405                 /*
 406                  * Move the split one to the right, since the top element in the
 407                  * last-compared pair can't be in the progression. Only do this
 408                  * when we split in the middle of the array (otherwise if we got
 409                  * to the end, then the split is in the right place).
 410                  */
 411                 split++;
 412         }
 413
 414         /*
 415          * Then, anything to the left of 'split' must be in a new pack. But,
 416          * creating that new pack may cause packs in the heavy half to no longer
 417          * form a geometric progression.
 418          *
 419          * Compute an expected size of the new pack, and then determine how many
 420          * packs in the heavy half need to be joined into it (if any) to restore
 421          * the geometric progression.
 422          */
 423         for (i = 0; i < split; i++) {
 424                 struct packed_git *p = geometry->pack[i];
 425
 426                 if (unsigned_add_overflows(total_size, geometry_pack_weight(p)))
 427                         die(_("pack %s too large to roll up"), p->pack_name);
 428                 total_size += geometry_pack_weight(p);
 429         }
 430         for (i = split; i < geometry->pack_nr; i++) {
 431                 struct packed_git *ours = geometry->pack[i];
 432
 433                 if (unsigned_mult_overflows(factor, total_size))
 434                         die(_("pack %s too large to roll up"), ours->pack_name);
 435
 436                 if (geometry_pack_weight(ours) < factor * total_size) {
 437                         if (unsigned_add_overflows(total_size,
 438                                                    geometry_pack_weight(ours)))
 439                                 die(_("pack %s too large to roll up"),
 440                                     ours->pack_name);
 441
 442                         split++;
 443                         total_size += geometry_pack_weight(ours);
 444                 } else
 445                         break;
 446         }
 447
 448         geometry->split = split;
 449 }
 450
 451 static struct packed_git *get_largest_active_pack(struct pack_geometry *geometry)
 452 {
 453         if (!geometry) {
 454                 /*
 455                  * No geometry means either an all-into-one repack (in which
 456                  * case there is only one pack left and it is the largest) or an
 457                  * incremental one.
 458                  *
 459                  * If repacking incrementally, then we could check the size of
 460                  * all packs to determine which should be preferred, but leave
 461                  * this for later.
 462                  */
 463                 return NULL;
 464         }
 465         if (geometry->split == geometry->pack_nr)
 466                 return NULL;
 467         return geometry->pack[geometry->pack_nr - 1];
 468 }
 469
 470 static void clear_pack_geometry(struct pack_geometry *geometry)
 471 {
 472         if (!geometry)
 473                 return;
 474
 475         free(geometry->pack);
 476         geometry->pack_nr = 0;
 477         geometry->pack_alloc = 0;
 478         geometry->split = 0;
 479 }
 480
 481 struct midx_snapshot_ref_data {
 482         struct tempfile *f;
 483         struct oidset seen;
 484         int preferred;
 485 };
 486
 487 static int midx_snapshot_ref_one(const char *refname UNUSED,
 488                                  const struct object_id *oid,
 489                                  int flag UNUSED, void *_data)
 490 {
 491         struct midx_snapshot_ref_data *data = _data;
 492         struct object_id peeled;
 493
 494         if (!peel_iterated_oid(oid, &peeled))
 495                 oid = &peeled;
 496
 497         if (oidset_insert(&data->seen, oid))
 498                 return 0; /* already seen */
 499
 500         if (oid_object_info(the_repository, oid, NULL) != OBJ_COMMIT)
 501                 return 0;
 502
 503         fprintf(data->f->fp, "%s%s\n", data->preferred ? "+" : "",
 504                 oid_to_hex(oid));
 505
 506         return 0;
 507 }
 508
 509 static void midx_snapshot_refs(struct tempfile *f)
 510 {
 511         struct midx_snapshot_ref_data data;
 512         const struct string_list *preferred = bitmap_preferred_tips(the_repository);
 513
 514         data.f = f;
 515         data.preferred = 0;
 516         oidset_init(&data.seen, 0);
 517
 518         if (!fdopen_tempfile(f, "w"))
 519                  die(_("could not open tempfile %s for writing"),
 520                      get_tempfile_path(f));
 521
 522         if (preferred) {
 523                 struct string_list_item *item;
 524
 525                 data.preferred = 1;
 526                 for_each_string_list_item(item, preferred)
 527                         for_each_ref_in(item->string, midx_snapshot_ref_one, &data);
 528                 data.preferred = 0;
 529         }
 530
 531         for_each_ref(midx_snapshot_ref_one, &data);
 532
 533         if (close_tempfile_gently(f)) {
 534                 int save_errno = errno;
 535                 delete_tempfile(&f);
 536                 errno = save_errno;
 537                 die_errno(_("could not close refs snapshot tempfile"));
 538         }
 539
 540         oidset_clear(&data.seen);
 541 }
 542
 543 static void midx_included_packs(struct string_list *include,
 544                                 struct string_list *existing_nonkept_packs,
 545                                 struct string_list *existing_kept_packs,
 546                                 struct string_list *names,
 547                                 struct pack_geometry *geometry)
 548 {
 549         struct string_list_item *item;
 550
 551         for_each_string_list_item(item, existing_kept_packs)
 552                 string_list_insert(include, xstrfmt("%s.idx", item->string));
 553         for_each_string_list_item(item, names)
 554                 string_list_insert(include, xstrfmt("pack-%s.idx", item->string));
 555         if (geometry) {
 556                 struct strbuf buf = STRBUF_INIT;
 557                 uint32_t i;
 558                 for (i = geometry->split; i < geometry->pack_nr; i++) {
 559                         struct packed_git *p = geometry->pack[i];
 560
 561                         strbuf_addstr(&buf, pack_basename(p));
 562                         strbuf_strip_suffix(&buf, ".pack");
 563                         strbuf_addstr(&buf, ".idx");
 564
 565                         string_list_insert(include, strbuf_detach(&buf, NULL));
 566                 }
 567
 568                 for_each_string_list_item(item, existing_nonkept_packs) {
 569                         if (!((uintptr_t)item->util & CRUFT_PACK)) {
 570                                 /*
 571                                  * no need to check DELETE_PACK, since we're not
 572                                  * doing an ALL_INTO_ONE repack
 573                                  */
 574                                 continue;
 575                         }
 576                         string_list_insert(include, xstrfmt("%s.idx", item->string));
 577                 }
 578         } else {
 579                 for_each_string_list_item(item, existing_nonkept_packs) {
 580                         if ((uintptr_t)item->util & DELETE_PACK)
 581                                 continue;
 582                         string_list_insert(include, xstrfmt("%s.idx", item->string));
 583                 }
 584         }
 585 }
 586
 587 static int write_midx_included_packs(struct string_list *include,
 588                                      struct pack_geometry *geometry,
 589                                      const char *refs_snapshot,
 590                                      int show_progress, int write_bitmaps)
 591 {
 592         struct child_process cmd = CHILD_PROCESS_INIT;
 593         struct string_list_item *item;
 594         struct packed_git *largest = get_largest_active_pack(geometry);
 595         FILE *in;
 596         int ret;
 597
 598         if (!include->nr)
 599                 return 0;
 600
 601         cmd.in = -1;
 602         cmd.git_cmd = 1;
 603
 604         strvec_push(&cmd.args, "multi-pack-index");
 605         strvec_pushl(&cmd.args, "write", "--stdin-packs", NULL);
 606
 607         if (show_progress)
 608                 strvec_push(&cmd.args, "--progress");
 609         else
 610                 strvec_push(&cmd.args, "--no-progress");
 611
 612         if (write_bitmaps)
 613                 strvec_push(&cmd.args, "--bitmap");
 614
 615         if (largest)
 616                 strvec_pushf(&cmd.args, "--preferred-pack=%s",
 617                              pack_basename(largest));
 618
 619         if (refs_snapshot)
 620                 strvec_pushf(&cmd.args, "--refs-snapshot=%s", refs_snapshot);
 621
 622         ret = start_command(&cmd);
 623         if (ret)
 624                 return ret;
 625
 626         in = xfdopen(cmd.in, "w");
 627         for_each_string_list_item(item, include)
 628                 fprintf(in, "%s\n", item->string);
 629         fclose(in);
 630
 631         return finish_command(&cmd);
 632 }
 633
 634 static void remove_redundant_bitmaps(struct string_list *include,
 635                                      const char *packdir)
 636 {
 637         struct strbuf path = STRBUF_INIT;
 638         struct string_list_item *item;
 639         size_t packdir_len;
 640
 641         strbuf_addstr(&path, packdir);
 642         strbuf_addch(&path, '/');
 643         packdir_len = path.len;
 644
 645         /*
 646          * Remove any pack bitmaps corresponding to packs which are now
 647          * included in the MIDX.
 648          */
 649         for_each_string_list_item(item, include) {
 650                 strbuf_addstr(&path, item->string);
 651                 strbuf_strip_suffix(&path, ".idx");
 652                 strbuf_addstr(&path, ".bitmap");
 653
 654                 if (unlink(path.buf) && errno != ENOENT)
 655                         warning_errno(_("could not remove stale bitmap: %s"),
 656                                       path.buf);
 657
 658                 strbuf_setlen(&path, packdir_len);
 659         }
 660         strbuf_release(&path);
 661 }
 662
 663 static int write_cruft_pack(const struct pack_objects_args *args,
 664                             const char *destination,
 665                             const char *pack_prefix,
 666                             const char *cruft_expiration,
 667                             struct string_list *names,
 668                             struct string_list *existing_packs,
 669                             struct string_list *existing_kept_packs)
 670 {
 671         struct child_process cmd = CHILD_PROCESS_INIT;
 672         struct strbuf line = STRBUF_INIT;
 673         struct string_list_item *item;
 674         FILE *in, *out;
 675         int ret;
 676         const char *scratch;
 677         int local = skip_prefix(destination, packdir, &scratch);
 678
 679         prepare_pack_objects(&cmd, args, destination);
 680
 681         strvec_push(&cmd.args, "--cruft");
 682         if (cruft_expiration)
 683                 strvec_pushf(&cmd.args, "--cruft-expiration=%s",
 684                              cruft_expiration);
 685
 686         strvec_push(&cmd.args, "--honor-pack-keep");
 687         strvec_push(&cmd.args, "--non-empty");
 688         strvec_push(&cmd.args, "--max-pack-size=0");
 689
 690         cmd.in = -1;
 691
 692         ret = start_command(&cmd);
 693         if (ret)
 694                 return ret;
 695
 696         /*
 697          * names has a confusing double use: it both provides the list
 698          * of just-written new packs, and accepts the name of the cruft
 699          * pack we are writing.
 700          *
 701          * By the time it is read here, it contains only the pack(s)
 702          * that were just written, which is exactly the set of packs we
 703          * want to consider kept.
 704          *
 705          * If `--expire-to` is given, the double-use served by `names`
 706          * ensures that the pack written to `--expire-to` excludes any
 707          * objects contained in the cruft pack.
 708          */
 709         in = xfdopen(cmd.in, "w");
 710         for_each_string_list_item(item, names)
 711                 fprintf(in, "%s-%s.pack\n", pack_prefix, item->string);
 712         for_each_string_list_item(item, existing_packs)
 713                 fprintf(in, "-%s.pack\n", item->string);
 714         for_each_string_list_item(item, existing_kept_packs)
 715                 fprintf(in, "%s.pack\n", item->string);
 716         fclose(in);
 717
 718         out = xfdopen(cmd.out, "r");
 719         while (strbuf_getline_lf(&line, out) != EOF) {
 720                 struct string_list_item *item;
 721
 722                 if (line.len != the_hash_algo->hexsz)
 723                         die(_("repack: Expecting full hex object ID lines only "
 724                               "from pack-objects."));
 725                 /*
 726                  * avoid putting packs written outside of the repository in the
 727                  * list of names
 728                  */
 729                 if (local) {
 730                         item = string_list_append(names, line.buf);
 731                         item->util = populate_pack_exts(line.buf);
 732                 }
 733         }
 734         fclose(out);
 735
 736         strbuf_release(&line);
 737
 738         return finish_command(&cmd);
 739 }
 740
 741 int cmd_repack(int argc, const char **argv, const char *prefix)
 742 {
 743         struct child_process cmd = CHILD_PROCESS_INIT;
 744         struct string_list_item *item;
 745         struct string_list names = STRING_LIST_INIT_DUP;
 746         struct string_list existing_nonkept_packs = STRING_LIST_INIT_DUP;
 747         struct string_list existing_kept_packs = STRING_LIST_INIT_DUP;
 748         struct pack_geometry *geometry = NULL;
 749         struct strbuf line = STRBUF_INIT;
 750         struct tempfile *refs_snapshot = NULL;
 751         int i, ext, ret;
 752         FILE *out;
 753         int show_progress;
 754
 755         /* variables to be filled by option parsing */
 756         int delete_redundant = 0;
 757         const char *unpack_unreachable = NULL;
 758         int keep_unreachable = 0;
 759         struct string_list keep_pack_list = STRING_LIST_INIT_NODUP;
 760         struct pack_objects_args po_args = {NULL};
 761         struct pack_objects_args cruft_po_args = {NULL};
 762         int geometric_factor = 0;
 763         int write_midx = 0;
 764         const char *cruft_expiration = NULL;
 765         const char *expire_to = NULL;
 766
 767         struct option builtin_repack_options[] = {
 768                 OPT_BIT('a', NULL, &pack_everything,
 769                                 N_("pack everything in a single pack"), ALL_INTO_ONE),
 770                 OPT_BIT('A', NULL, &pack_everything,
 771                                 N_("same as -a, and turn unreachable objects loose"),
 772                                    LOOSEN_UNREACHABLE | ALL_INTO_ONE),
 773                 OPT_BIT(0, "cruft", &pack_everything,
 774                                 N_("same as -a, pack unreachable cruft objects separately"),
 775                                    PACK_CRUFT),
 776                 OPT_STRING(0, "cruft-expiration", &cruft_expiration, N_("approxidate"),
 777                                 N_("with -C, expire objects older than this")),
 778                 OPT_BOOL('d', NULL, &delete_redundant,
 779                                 N_("remove redundant packs, and run git-prune-packed")),
 780                 OPT_BOOL('f', NULL, &po_args.no_reuse_delta,
 781                                 N_("pass --no-reuse-delta to git-pack-objects")),
 782                 OPT_BOOL('F', NULL, &po_args.no_reuse_object,
 783                                 N_("pass --no-reuse-object to git-pack-objects")),
 784                 OPT_NEGBIT('n', NULL, &run_update_server_info,
 785                                 N_("do not run git-update-server-info"), 1),
 786                 OPT__QUIET(&po_args.quiet, N_("be quiet")),
 787                 OPT_BOOL('l', "local", &po_args.local,
 788                                 N_("pass --local to git-pack-objects")),
 789                 OPT_BOOL('b', "write-bitmap-index", &write_bitmaps,
 790                                 N_("write bitmap index")),
 791                 OPT_BOOL('i', "delta-islands", &use_delta_islands,
 792                                 N_("pass --delta-islands to git-pack-objects")),
 793                 OPT_STRING(0, "unpack-unreachable", &unpack_unreachable, N_("approxidate"),
 794                                 N_("with -A, do not loosen objects older than this")),
 795                 OPT_BOOL('k', "keep-unreachable", &keep_unreachable,
 796                                 N_("with -a, repack unreachable objects")),
 797                 OPT_STRING(0, "window", &po_args.window, N_("n"),
 798                                 N_("size of the window used for delta compression")),
 799                 OPT_STRING(0, "window-memory", &po_args.window_memory, N_("bytes"),
 800                                 N_("same as the above, but limit memory size instead of entries count")),
 801                 OPT_STRING(0, "depth", &po_args.depth, N_("n"),
 802                                 N_("limits the maximum delta depth")),
 803                 OPT_STRING(0, "threads", &po_args.threads, N_("n"),
 804                                 N_("limits the maximum number of threads")),
 805                 OPT_STRING(0, "max-pack-size", &po_args.max_pack_size, N_("bytes"),
 806                                 N_("maximum size of each packfile")),
 807                 OPT_BOOL(0, "pack-kept-objects", &pack_kept_objects,
 808                                 N_("repack objects in packs marked with .keep")),
 809                 OPT_STRING_LIST(0, "keep-pack", &keep_pack_list, N_("name"),
 810                                 N_("do not repack this pack")),
 811                 OPT_INTEGER('g', "geometric", &geometric_factor,
 812                             N_("find a geometric progression with factor <N>")),
 813                 OPT_BOOL('m', "write-midx", &write_midx,
 814                            N_("write a multi-pack index of the resulting packs")),
 815                 OPT_STRING(0, "expire-to", &expire_to, N_("dir"),
 816                            N_("pack prefix to store a pack containing pruned objects")),
 817                 OPT_END()
 818         };
 819
 820         git_config(repack_config, &cruft_po_args);
 821
 822         argc = parse_options(argc, argv, prefix, builtin_repack_options,
 823                                 git_repack_usage, 0);
 824
 825         if (delete_redundant && repository_format_precious_objects)
 826                 die(_("cannot delete packs in a precious-objects repo"));
 827
 828         if (keep_unreachable &&
 829             (unpack_unreachable || (pack_everything & LOOSEN_UNREACHABLE)))
 830                 die(_("options '%s' and '%s' cannot be used together"), "--keep-unreachable", "-A");
 831
 832         if (pack_everything & PACK_CRUFT) {
 833                 pack_everything |= ALL_INTO_ONE;
 834
 835                 if (unpack_unreachable || (pack_everything & LOOSEN_UNREACHABLE))
 836                         die(_("options '%s' and '%s' cannot be used together"), "--cruft", "-A");
 837                 if (keep_unreachable)
 838                         die(_("options '%s' and '%s' cannot be used together"), "--cruft", "-k");
 839         }
 840
 841         if (write_bitmaps < 0) {
 842                 if (!write_midx &&
 843                     (!(pack_everything & ALL_INTO_ONE) || !is_bare_repository()))
 844                         write_bitmaps = 0;
 845         } else if (write_bitmaps &&
 846                    git_env_bool(GIT_TEST_MULTI_PACK_INDEX, 0) &&
 847                    git_env_bool(GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP, 0)) {
 848                 write_bitmaps = 0;
 849         }
 850         if (pack_kept_objects < 0)
 851                 pack_kept_objects = write_bitmaps > 0 && !write_midx;
 852
 853         if (write_bitmaps && !(pack_everything & ALL_INTO_ONE) && !write_midx)
 854                 die(_(incremental_bitmap_conflict_error));
 855
 856         if (write_midx && write_bitmaps) {
 857                 struct strbuf path = STRBUF_INIT;
 858
 859                 strbuf_addf(&path, "%s/%s_XXXXXX", get_object_directory(),
 860                             "bitmap-ref-tips");
 861
 862                 refs_snapshot = xmks_tempfile(path.buf);
 863                 midx_snapshot_refs(refs_snapshot);
 864
 865                 strbuf_release(&path);
 866         }
 867
 868         packdir = mkpathdup("%s/pack", get_object_directory());
 869         packtmp_name = xstrfmt(".tmp-%d-pack", (int)getpid());
 870         packtmp = mkpathdup("%s/%s", packdir, packtmp_name);
 871
 872         collect_pack_filenames(&existing_nonkept_packs, &existing_kept_packs,
 873                                &keep_pack_list);
 874
 875         if (geometric_factor) {
 876                 if (pack_everything)
 877                         die(_("options '%s' and '%s' cannot be used together"), "--geometric", "-A/-a");
 878                 init_pack_geometry(&geometry, &existing_kept_packs);
 879                 split_pack_geometry(geometry, geometric_factor);
 880         }
 881
 882         prepare_pack_objects(&cmd, &po_args, packtmp);
 883
 884         show_progress = !po_args.quiet && isatty(2);
 885
 886         strvec_push(&cmd.args, "--keep-true-parents");
 887         if (!pack_kept_objects)
 888                 strvec_push(&cmd.args, "--honor-pack-keep");
 889         for (i = 0; i < keep_pack_list.nr; i++)
 890                 strvec_pushf(&cmd.args, "--keep-pack=%s",
 891                              keep_pack_list.items[i].string);
 892         strvec_push(&cmd.args, "--non-empty");
 893         if (!geometry) {
 894                 /*
 895                  * We need to grab all reachable objects, including those that
 896                  * are reachable from reflogs and the index.
 897                  *
 898                  * When repacking into a geometric progression of packs,
 899                  * however, we ask 'git pack-objects --stdin-packs', and it is
 900                  * not about packing objects based on reachability but about
 901                  * repacking all the objects in specified packs and loose ones
 902                  * (indeed, --stdin-packs is incompatible with these options).
 903                  */
 904                 strvec_push(&cmd.args, "--all");
 905                 strvec_push(&cmd.args, "--reflog");
 906                 strvec_push(&cmd.args, "--indexed-objects");
 907         }
 908         if (repo_has_promisor_remote(the_repository))
 909                 strvec_push(&cmd.args, "--exclude-promisor-objects");
 910         if (!write_midx) {
 911                 if (write_bitmaps > 0)
 912                         strvec_push(&cmd.args, "--write-bitmap-index");
 913                 else if (write_bitmaps < 0)
 914                         strvec_push(&cmd.args, "--write-bitmap-index-quiet");
 915         }
 916         if (use_delta_islands)
 917                 strvec_push(&cmd.args, "--delta-islands");
 918
 919         if (pack_everything & ALL_INTO_ONE) {
 920                 repack_promisor_objects(&po_args, &names);
 921
 922                 if (existing_nonkept_packs.nr && delete_redundant &&
 923                     !(pack_everything & PACK_CRUFT)) {
 924                         for_each_string_list_item(item, &names) {
 925                                 strvec_pushf(&cmd.args, "--keep-pack=%s-%s.pack",
 926                                              packtmp_name, item->string);
 927                         }
 928                         if (unpack_unreachable) {
 929                                 strvec_pushf(&cmd.args,
 930                                              "--unpack-unreachable=%s",
 931                                              unpack_unreachable);
 932                         } else if (pack_everything & LOOSEN_UNREACHABLE) {
 933                                 strvec_push(&cmd.args,
 934                                             "--unpack-unreachable");
 935                         } else if (keep_unreachable) {
 936                                 strvec_push(&cmd.args, "--keep-unreachable");
 937                                 strvec_push(&cmd.args, "--pack-loose-unreachable");
 938                         }
 939                 }
 940         } else if (geometry) {
 941                 strvec_push(&cmd.args, "--stdin-packs");
 942                 strvec_push(&cmd.args, "--unpacked");
 943         } else {
 944                 strvec_push(&cmd.args, "--unpacked");
 945                 strvec_push(&cmd.args, "--incremental");
 946         }
 947
 948         if (geometry)
 949                 cmd.in = -1;
 950         else
 951                 cmd.no_stdin = 1;
 952
 953         ret = start_command(&cmd);
 954         if (ret)
 955                 goto cleanup;
 956
 957         if (geometry) {
 958                 FILE *in = xfdopen(cmd.in, "w");
 959                 /*
 960                  * The resulting pack should contain all objects in packs that
 961                  * are going to be rolled up, but exclude objects in packs which
 962                  * are being left alone.
 963                  */
 964                 for (i = 0; i < geometry->split; i++)
 965                         fprintf(in, "%s\n", pack_basename(geometry->pack[i]));
 966                 for (i = geometry->split; i < geometry->pack_nr; i++)
 967                         fprintf(in, "^%s\n", pack_basename(geometry->pack[i]));
 968                 fclose(in);
 969         }
 970
 971         out = xfdopen(cmd.out, "r");
 972         while (strbuf_getline_lf(&line, out) != EOF) {
 973                 struct string_list_item *item;
 974
 975                 if (line.len != the_hash_algo->hexsz)
 976                         die(_("repack: Expecting full hex object ID lines only from pack-objects."));
 977                 item = string_list_append(&names, line.buf);
 978                 item->util = populate_pack_exts(item->string);
 979         }
 980         strbuf_release(&line);
 981         fclose(out);
 982         ret = finish_command(&cmd);
 983         if (ret)
 984                 goto cleanup;
 985
 986         if (!names.nr && !po_args.quiet)
 987                 printf_ln(_("Nothing new to pack."));
 988
 989         if (pack_everything & PACK_CRUFT) {
 990                 const char *pack_prefix;
 991                 if (!skip_prefix(packtmp, packdir, &pack_prefix))
 992                         die(_("pack prefix %s does not begin with objdir %s"),
 993                             packtmp, packdir);
 994                 if (*pack_prefix == '/')
 995                         pack_prefix++;
 996
 997                 if (!cruft_po_args.window)
 998                         cruft_po_args.window = po_args.window;
 999                 if (!cruft_po_args.window_memory)
1000                         cruft_po_args.window_memory = po_args.window_memory;
1001                 if (!cruft_po_args.depth)
1002                         cruft_po_args.depth = po_args.depth;
1003                 if (!cruft_po_args.threads)
1004                         cruft_po_args.threads = po_args.threads;
1005
1006                 cruft_po_args.local = po_args.local;
1007                 cruft_po_args.quiet = po_args.quiet;
1008
1009                 ret = write_cruft_pack(&cruft_po_args, packtmp, pack_prefix,
1010                                        cruft_expiration, &names,
1011                                        &existing_nonkept_packs,
1012                                        &existing_kept_packs);
1013                 if (ret)
1014                         goto cleanup;
1015
1016                 if (delete_redundant && expire_to) {
1017                         /*
1018                          * If `--expire-to` is given with `-d`, it's possible
1019                          * that we're about to prune some objects. With cruft
1020                          * packs, pruning is implicit: any objects from existing
1021                          * packs that weren't picked up by new packs are removed
1022                          * when their packs are deleted.
1023                          *
1024                          * Generate an additional cruft pack, with one twist:
1025                          * `names` now includes the name of the cruft pack
1026                          * written in the previous step. So the contents of
1027                          * _this_ cruft pack exclude everything contained in the
1028                          * existing cruft pack (that is, all of the unreachable
1029                          * objects which are no older than
1030                          * `--cruft-expiration`).
1031                          *
1032                          * To make this work, cruft_expiration must become NULL
1033                          * so that this cruft pack doesn't actually prune any
1034                          * objects. If it were non-NULL, this call would always
1035                          * generate an empty pack (since every object not in the
1036                          * cruft pack generated above will have an mtime older
1037                          * than the expiration).
1038                          */
1039                         ret = write_cruft_pack(&cruft_po_args, expire_to,
1040                                                pack_prefix,
1041                                                NULL,
1042                                                &names,
1043                                                &existing_nonkept_packs,
1044                                                &existing_kept_packs);
1045                         if (ret)
1046                                 goto cleanup;
1047                 }
1048         }
1049
1050         string_list_sort(&names);
1051
1052         close_object_store(the_repository->objects);
1053
1054         /*
1055          * Ok we have prepared all new packfiles.
1056          */
1057         for_each_string_list_item(item, &names) {
1058                 struct generated_pack_data *data = item->util;
1059
1060                 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
1061                         char *fname;
1062
1063                         fname = mkpathdup("%s/pack-%s%s",
1064                                         packdir, item->string, exts[ext].name);
1065
1066                         if (data->tempfiles[ext]) {
1067                                 const char *fname_old = get_tempfile_path(data->tempfiles[ext]);
1068                                 struct stat statbuffer;
1069
1070                                 if (!stat(fname_old, &statbuffer)) {
1071                                         statbuffer.st_mode &= ~(S_IWUSR | S_IWGRP | S_IWOTH);
1072                                         chmod(fname_old, statbuffer.st_mode);
1073                                 }
1074
1075                                 if (rename_tempfile(&data->tempfiles[ext], fname))
1076                                         die_errno(_("renaming pack to '%s' failed"), fname);
1077                         } else if (!exts[ext].optional)
1078                                 die(_("pack-objects did not write a '%s' file for pack %s-%s"),
1079                                     exts[ext].name, packtmp, item->string);
1080                         else if (unlink(fname) < 0 && errno != ENOENT)
1081                                 die_errno(_("could not unlink: %s"), fname);
1082
1083                         free(fname);
1084                 }
1085         }
1086         /* End of pack replacement. */
1087
1088         if (delete_redundant && pack_everything & ALL_INTO_ONE) {
1089                 const int hexsz = the_hash_algo->hexsz;
1090                 for_each_string_list_item(item, &existing_nonkept_packs) {
1091                         char *sha1;
1092                         size_t len = strlen(item->string);
1093                         if (len < hexsz)
1094                                 continue;
1095                         sha1 = item->string + len - hexsz;
1096                         /*
1097                          * Mark this pack for deletion, which ensures that this
1098                          * pack won't be included in a MIDX (if `--write-midx`
1099                          * was given) and that we will actually delete this pack
1100                          * (if `-d` was given).
1101                          */
1102                         if (!string_list_has_string(&names, sha1))
1103                                 item->util = (void*)(uintptr_t)((size_t)item->util | DELETE_PACK);
1104                 }
1105         }
1106
1107         if (write_midx) {
1108                 struct string_list include = STRING_LIST_INIT_NODUP;
1109                 midx_included_packs(&include, &existing_nonkept_packs,
1110                                     &existing_kept_packs, &names, geometry);
1111
1112                 ret = write_midx_included_packs(&include, geometry,
1113                                                 refs_snapshot ? get_tempfile_path(refs_snapshot) : NULL,
1114                                                 show_progress, write_bitmaps > 0);
1115
1116                 if (!ret && write_bitmaps)
1117                         remove_redundant_bitmaps(&include, packdir);
1118
1119                 string_list_clear(&include, 0);
1120
1121                 if (ret)
1122                         goto cleanup;
1123         }
1124
1125         reprepare_packed_git(the_repository);
1126
1127         if (delete_redundant) {
1128                 int opts = 0;
1129                 for_each_string_list_item(item, &existing_nonkept_packs) {
1130                         if (!((uintptr_t)item->util & DELETE_PACK))
1131                                 continue;
1132                         remove_redundant_pack(packdir, item->string);
1133                 }
1134
1135                 if (geometry) {
1136                         struct strbuf buf = STRBUF_INIT;
1137
1138                         uint32_t i;
1139                         for (i = 0; i < geometry->split; i++) {
1140                                 struct packed_git *p = geometry->pack[i];
1141                                 if (string_list_has_string(&names,
1142                                                            hash_to_hex(p->hash)))
1143                                         continue;
1144
1145                                 strbuf_reset(&buf);
1146                                 strbuf_addstr(&buf, pack_basename(p));
1147                                 strbuf_strip_suffix(&buf, ".pack");
1148
1149                                 if ((p->pack_keep) ||
1150                                     (string_list_has_string(&existing_kept_packs,
1151                                                             buf.buf)))
1152                                         continue;
1153
1154                                 remove_redundant_pack(packdir, buf.buf);
1155                         }
1156                         strbuf_release(&buf);
1157                 }
1158                 if (show_progress)
1159                         opts |= PRUNE_PACKED_VERBOSE;
1160                 prune_packed_objects(opts);
1161
1162                 if (!keep_unreachable &&
1163                     (!(pack_everything & LOOSEN_UNREACHABLE) ||
1164                      unpack_unreachable) &&
1165                     is_repository_shallow(the_repository))
1166                         prune_shallow(PRUNE_QUICK);
1167         }
1168
1169         if (run_update_server_info)
1170                 update_server_info(0);
1171
1172         if (git_env_bool(GIT_TEST_MULTI_PACK_INDEX, 0)) {
1173                 unsigned flags = 0;
1174                 if (git_env_bool(GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP, 0))
1175                         flags |= MIDX_WRITE_BITMAP | MIDX_WRITE_REV_INDEX;
1176                 write_midx_file(get_object_directory(), NULL, NULL, flags);
1177         }
1178
1179 cleanup:
1180         string_list_clear(&names, 1);
1181         string_list_clear(&existing_nonkept_packs, 0);
1182         string_list_clear(&existing_kept_packs, 0);
1183         clear_pack_geometry(geometry);
1184
1185         return ret;
1186 }