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 "argv-array.h"
  11 #include "midx.h"
  12 #include "packfile.h"
  13 #include "object-store.h"
  14
  15 static int delta_base_offset = 1;
  16 static int pack_kept_objects = -1;
  17 static int write_bitmaps = -1;
  18 static int use_delta_islands;
  19 static char *packdir, *packtmp;
  20
  21 static const char *const git_repack_usage[] = {
  22         N_("git repack [<options>]"),
  23         NULL
  24 };
  25
  26 static const char incremental_bitmap_conflict_error[] = N_(
  27 "Incremental repacks are incompatible with bitmap indexes.  Use\n"
  28 "--no-write-bitmap-index or disable the pack.writebitmaps configuration."
  29 );
  30
  31
  32 static int repack_config(const char *var, const char *value, void *cb)
  33 {
  34         if (!strcmp(var, "repack.usedeltabaseoffset")) {
  35                 delta_base_offset = git_config_bool(var, value);
  36                 return 0;
  37         }
  38         if (!strcmp(var, "repack.packkeptobjects")) {
  39                 pack_kept_objects = git_config_bool(var, value);
  40                 return 0;
  41         }
  42         if (!strcmp(var, "repack.writebitmaps") ||
  43             !strcmp(var, "pack.writebitmaps")) {
  44                 write_bitmaps = git_config_bool(var, value);
  45                 return 0;
  46         }
  47         if (!strcmp(var, "repack.usedeltaislands")) {
  48                 use_delta_islands = git_config_bool(var, value);
  49                 return 0;
  50         }
  51         return git_default_config(var, value, cb);
  52 }
  53
  54 /*
  55  * Remove temporary $GIT_OBJECT_DIRECTORY/pack/.tmp-$$-pack-* files.
  56  */
  57 static void remove_temporary_files(void)
  58 {
  59         struct strbuf buf = STRBUF_INIT;
  60         size_t dirlen, prefixlen;
  61         DIR *dir;
  62         struct dirent *e;
  63
  64         dir = opendir(packdir);
  65         if (!dir)
  66                 return;
  67
  68         /* Point at the slash at the end of ".../objects/pack/" */
  69         dirlen = strlen(packdir) + 1;
  70         strbuf_addstr(&buf, packtmp);
  71         /* Hold the length of  ".tmp-%d-pack-" */
  72         prefixlen = buf.len - dirlen;
  73
  74         while ((e = readdir(dir))) {
  75                 if (strncmp(e->d_name, buf.buf + dirlen, prefixlen))
  76                         continue;
  77                 strbuf_setlen(&buf, dirlen);
  78                 strbuf_addstr(&buf, e->d_name);
  79                 unlink(buf.buf);
  80         }
  81         closedir(dir);
  82         strbuf_release(&buf);
  83 }
  84
  85 static void remove_pack_on_signal(int signo)
  86 {
  87         remove_temporary_files();
  88         sigchain_pop(signo);
  89         raise(signo);
  90 }
  91
  92 static int has_pack_keep_file(void)
  93 {
  94         struct packed_git *p;
  95
  96         for (p = get_all_packs(the_repository); p; p = p->next) {
  97                 if (p->pack_keep)
  98                         return 1;
  99         }
 100         return 0;
 101 }
 102
 103 /*
 104  * Adds all packs hex strings to the fname list, which do not
 105  * have a corresponding .keep file. These packs are not to
 106  * be kept if we are going to pack everything into one file.
 107  */
 108 static void get_non_kept_pack_filenames(struct string_list *fname_list,
 109                                         const struct string_list *extra_keep)
 110 {
 111         DIR *dir;
 112         struct dirent *e;
 113         char *fname;
 114
 115         if (!(dir = opendir(packdir)))
 116                 return;
 117
 118         while ((e = readdir(dir)) != NULL) {
 119                 size_t len;
 120                 int i;
 121
 122                 for (i = 0; i < extra_keep->nr; i++)
 123                         if (!fspathcmp(e->d_name, extra_keep->items[i].string))
 124                                 break;
 125                 if (extra_keep->nr > 0 && i < extra_keep->nr)
 126                         continue;
 127
 128                 if (!strip_suffix(e->d_name, ".pack", &len))
 129                         continue;
 130
 131                 fname = xmemdupz(e->d_name, len);
 132
 133                 if (!file_exists(mkpath("%s/%s.keep", packdir, fname)))
 134                         string_list_append_nodup(fname_list, fname);
 135                 else
 136                         free(fname);
 137         }
 138         closedir(dir);
 139 }
 140
 141 static void remove_redundant_pack(const char *dir_name, const char *base_name)
 142 {
 143         struct strbuf buf = STRBUF_INIT;
 144         strbuf_addf(&buf, "%s/%s.pack", dir_name, base_name);
 145         unlink_pack_path(buf.buf, 1);
 146         strbuf_release(&buf);
 147 }
 148
 149 struct pack_objects_args {
 150         const char *window;
 151         const char *window_memory;
 152         const char *depth;
 153         const char *threads;
 154         const char *max_pack_size;
 155         int no_reuse_delta;
 156         int no_reuse_object;
 157         int quiet;
 158         int local;
 159 };
 160
 161 static void prepare_pack_objects(struct child_process *cmd,
 162                                  const struct pack_objects_args *args)
 163 {
 164         argv_array_push(&cmd->args, "pack-objects");
 165         if (args->window)
 166                 argv_array_pushf(&cmd->args, "--window=%s", args->window);
 167         if (args->window_memory)
 168                 argv_array_pushf(&cmd->args, "--window-memory=%s", args->window_memory);
 169         if (args->depth)
 170                 argv_array_pushf(&cmd->args, "--depth=%s", args->depth);
 171         if (args->threads)
 172                 argv_array_pushf(&cmd->args, "--threads=%s", args->threads);
 173         if (args->max_pack_size)
 174                 argv_array_pushf(&cmd->args, "--max-pack-size=%s", args->max_pack_size);
 175         if (args->no_reuse_delta)
 176                 argv_array_pushf(&cmd->args, "--no-reuse-delta");
 177         if (args->no_reuse_object)
 178                 argv_array_pushf(&cmd->args, "--no-reuse-object");
 179         if (args->local)
 180                 argv_array_push(&cmd->args,  "--local");
 181         if (args->quiet)
 182                 argv_array_push(&cmd->args,  "--quiet");
 183         if (delta_base_offset)
 184                 argv_array_push(&cmd->args,  "--delta-base-offset");
 185         argv_array_push(&cmd->args, packtmp);
 186         cmd->git_cmd = 1;
 187         cmd->out = -1;
 188 }
 189
 190 /*
 191  * Write oid to the given struct child_process's stdin, starting it first if
 192  * necessary.
 193  */
 194 static int write_oid(const struct object_id *oid, struct packed_git *pack,
 195                      uint32_t pos, void *data)
 196 {
 197         struct child_process *cmd = data;
 198
 199         if (cmd->in == -1) {
 200                 if (start_command(cmd))
 201                         die(_("could not start pack-objects to repack promisor objects"));
 202         }
 203
 204         xwrite(cmd->in, oid_to_hex(oid), GIT_SHA1_HEXSZ);
 205         xwrite(cmd->in, "\n", 1);
 206         return 0;
 207 }
 208
 209 static void repack_promisor_objects(const struct pack_objects_args *args,
 210                                     struct string_list *names)
 211 {
 212         struct child_process cmd = CHILD_PROCESS_INIT;
 213         FILE *out;
 214         struct strbuf line = STRBUF_INIT;
 215
 216         prepare_pack_objects(&cmd, args);
 217         cmd.in = -1;
 218
 219         /*
 220          * NEEDSWORK: Giving pack-objects only the OIDs without any ordering
 221          * hints may result in suboptimal deltas in the resulting pack. See if
 222          * the OIDs can be sent with fake paths such that pack-objects can use a
 223          * {type -> existing pack order} ordering when computing deltas instead
 224          * of a {type -> size} ordering, which may produce better deltas.
 225          */
 226         for_each_packed_object(write_oid, &cmd,
 227                                FOR_EACH_OBJECT_PROMISOR_ONLY);
 228
 229         if (cmd.in == -1)
 230                 /* No packed objects; cmd was never started */
 231                 return;
 232
 233         close(cmd.in);
 234
 235         out = xfdopen(cmd.out, "r");
 236         while (strbuf_getline_lf(&line, out) != EOF) {
 237                 char *promisor_name;
 238                 int fd;
 239                 if (line.len != the_hash_algo->hexsz)
 240                         die(_("repack: Expecting full hex object ID lines only from pack-objects."));
 241                 string_list_append(names, line.buf);
 242
 243                 /*
 244                  * pack-objects creates the .pack and .idx files, but not the
 245                  * .promisor file. Create the .promisor file, which is empty.
 246                  */
 247                 promisor_name = mkpathdup("%s-%s.promisor", packtmp,
 248                                           line.buf);
 249                 fd = open(promisor_name, O_CREAT|O_EXCL|O_WRONLY, 0600);
 250                 if (fd < 0)
 251                         die_errno(_("unable to create '%s'"), promisor_name);
 252                 close(fd);
 253                 free(promisor_name);
 254         }
 255         fclose(out);
 256         if (finish_command(&cmd))
 257                 die(_("could not finish pack-objects to repack promisor objects"));
 258 }
 259
 260 #define ALL_INTO_ONE 1
 261 #define LOOSEN_UNREACHABLE 2
 262
 263 int cmd_repack(int argc, const char **argv, const char *prefix)
 264 {
 265         struct {
 266                 const char *name;
 267                 unsigned optional:1;
 268         } exts[] = {
 269                 {".pack"},
 270                 {".idx"},
 271                 {".bitmap", 1},
 272                 {".promisor", 1},
 273         };
 274         struct child_process cmd = CHILD_PROCESS_INIT;
 275         struct string_list_item *item;
 276         struct string_list names = STRING_LIST_INIT_DUP;
 277         struct string_list rollback = STRING_LIST_INIT_NODUP;
 278         struct string_list existing_packs = STRING_LIST_INIT_DUP;
 279         struct strbuf line = STRBUF_INIT;
 280         int i, ext, ret, failed;
 281         FILE *out;
 282
 283         /* variables to be filled by option parsing */
 284         int pack_everything = 0;
 285         int delete_redundant = 0;
 286         const char *unpack_unreachable = NULL;
 287         int keep_unreachable = 0;
 288         struct string_list keep_pack_list = STRING_LIST_INIT_NODUP;
 289         int no_update_server_info = 0;
 290         int midx_cleared = 0;
 291         struct pack_objects_args po_args = {NULL};
 292
 293         struct option builtin_repack_options[] = {
 294                 OPT_BIT('a', NULL, &pack_everything,
 295                                 N_("pack everything in a single pack"), ALL_INTO_ONE),
 296                 OPT_BIT('A', NULL, &pack_everything,
 297                                 N_("same as -a, and turn unreachable objects loose"),
 298                                    LOOSEN_UNREACHABLE | ALL_INTO_ONE),
 299                 OPT_BOOL('d', NULL, &delete_redundant,
 300                                 N_("remove redundant packs, and run git-prune-packed")),
 301                 OPT_BOOL('f', NULL, &po_args.no_reuse_delta,
 302                                 N_("pass --no-reuse-delta to git-pack-objects")),
 303                 OPT_BOOL('F', NULL, &po_args.no_reuse_object,
 304                                 N_("pass --no-reuse-object to git-pack-objects")),
 305                 OPT_BOOL('n', NULL, &no_update_server_info,
 306                                 N_("do not run git-update-server-info")),
 307                 OPT__QUIET(&po_args.quiet, N_("be quiet")),
 308                 OPT_BOOL('l', "local", &po_args.local,
 309                                 N_("pass --local to git-pack-objects")),
 310                 OPT_BOOL('b', "write-bitmap-index", &write_bitmaps,
 311                                 N_("write bitmap index")),
 312                 OPT_BOOL('i', "delta-islands", &use_delta_islands,
 313                                 N_("pass --delta-islands to git-pack-objects")),
 314                 OPT_STRING(0, "unpack-unreachable", &unpack_unreachable, N_("approxidate"),
 315                                 N_("with -A, do not loosen objects older than this")),
 316                 OPT_BOOL('k', "keep-unreachable", &keep_unreachable,
 317                                 N_("with -a, repack unreachable objects")),
 318                 OPT_STRING(0, "window", &po_args.window, N_("n"),
 319                                 N_("size of the window used for delta compression")),
 320                 OPT_STRING(0, "window-memory", &po_args.window_memory, N_("bytes"),
 321                                 N_("same as the above, but limit memory size instead of entries count")),
 322                 OPT_STRING(0, "depth", &po_args.depth, N_("n"),
 323                                 N_("limits the maximum delta depth")),
 324                 OPT_STRING(0, "threads", &po_args.threads, N_("n"),
 325                                 N_("limits the maximum number of threads")),
 326                 OPT_STRING(0, "max-pack-size", &po_args.max_pack_size, N_("bytes"),
 327                                 N_("maximum size of each packfile")),
 328                 OPT_BOOL(0, "pack-kept-objects", &pack_kept_objects,
 329                                 N_("repack objects in packs marked with .keep")),
 330                 OPT_STRING_LIST(0, "keep-pack", &keep_pack_list, N_("name"),
 331                                 N_("do not repack this pack")),
 332                 OPT_END()
 333         };
 334
 335         git_config(repack_config, NULL);
 336
 337         argc = parse_options(argc, argv, prefix, builtin_repack_options,
 338                                 git_repack_usage, 0);
 339
 340         if (delete_redundant && repository_format_precious_objects)
 341                 die(_("cannot delete packs in a precious-objects repo"));
 342
 343         if (keep_unreachable &&
 344             (unpack_unreachable || (pack_everything & LOOSEN_UNREACHABLE)))
 345                 die(_("--keep-unreachable and -A are incompatible"));
 346
 347         if (write_bitmaps < 0) {
 348                 if (!(pack_everything & ALL_INTO_ONE) ||
 349                     !is_bare_repository() ||
 350                     keep_pack_list.nr != 0 ||
 351                     has_pack_keep_file())
 352                         write_bitmaps = 0;
 353         }
 354         if (pack_kept_objects < 0)
 355                 pack_kept_objects = !!write_bitmaps;
 356
 357         if (write_bitmaps && !(pack_everything & ALL_INTO_ONE))
 358                 die(_(incremental_bitmap_conflict_error));
 359
 360         packdir = mkpathdup("%s/pack", get_object_directory());
 361         packtmp = mkpathdup("%s/.tmp-%d-pack", packdir, (int)getpid());
 362
 363         sigchain_push_common(remove_pack_on_signal);
 364
 365         prepare_pack_objects(&cmd, &po_args);
 366
 367         argv_array_push(&cmd.args, "--keep-true-parents");
 368         if (!pack_kept_objects)
 369                 argv_array_push(&cmd.args, "--honor-pack-keep");
 370         for (i = 0; i < keep_pack_list.nr; i++)
 371                 argv_array_pushf(&cmd.args, "--keep-pack=%s",
 372                                  keep_pack_list.items[i].string);
 373         argv_array_push(&cmd.args, "--non-empty");
 374         argv_array_push(&cmd.args, "--all");
 375         argv_array_push(&cmd.args, "--reflog");
 376         argv_array_push(&cmd.args, "--indexed-objects");
 377         if (repository_format_partial_clone)
 378                 argv_array_push(&cmd.args, "--exclude-promisor-objects");
 379         if (write_bitmaps > 0)
 380                 argv_array_push(&cmd.args, "--write-bitmap-index");
 381         else if (write_bitmaps < 0)
 382                 argv_array_push(&cmd.args, "--write-bitmap-index-quiet");
 383         if (use_delta_islands)
 384                 argv_array_push(&cmd.args, "--delta-islands");
 385
 386         if (pack_everything & ALL_INTO_ONE) {
 387                 get_non_kept_pack_filenames(&existing_packs, &keep_pack_list);
 388
 389                 repack_promisor_objects(&po_args, &names);
 390
 391                 if (existing_packs.nr && delete_redundant) {
 392                         if (unpack_unreachable) {
 393                                 argv_array_pushf(&cmd.args,
 394                                                 "--unpack-unreachable=%s",
 395                                                 unpack_unreachable);
 396                                 argv_array_push(&cmd.env_array, "GIT_REF_PARANOIA=1");
 397                         } else if (pack_everything & LOOSEN_UNREACHABLE) {
 398                                 argv_array_push(&cmd.args,
 399                                                 "--unpack-unreachable");
 400                         } else if (keep_unreachable) {
 401                                 argv_array_push(&cmd.args, "--keep-unreachable");
 402                                 argv_array_push(&cmd.args, "--pack-loose-unreachable");
 403                         } else {
 404                                 argv_array_push(&cmd.env_array, "GIT_REF_PARANOIA=1");
 405                         }
 406                 }
 407         } else {
 408                 argv_array_push(&cmd.args, "--unpacked");
 409                 argv_array_push(&cmd.args, "--incremental");
 410         }
 411
 412         cmd.no_stdin = 1;
 413
 414         ret = start_command(&cmd);
 415         if (ret)
 416                 return ret;
 417
 418         out = xfdopen(cmd.out, "r");
 419         while (strbuf_getline_lf(&line, out) != EOF) {
 420                 if (line.len != the_hash_algo->hexsz)
 421                         die(_("repack: Expecting full hex object ID lines only from pack-objects."));
 422                 string_list_append(&names, line.buf);
 423         }
 424         fclose(out);
 425         ret = finish_command(&cmd);
 426         if (ret)
 427                 return ret;
 428
 429         if (!names.nr && !po_args.quiet)
 430                 printf_ln(_("Nothing new to pack."));
 431
 432         close_object_store(the_repository->objects);
 433
 434         /*
 435          * Ok we have prepared all new packfiles.
 436          * First see if there are packs of the same name and if so
 437          * if we can move them out of the way (this can happen if we
 438          * repacked immediately after packing fully.
 439          */
 440         failed = 0;
 441         for_each_string_list_item(item, &names) {
 442                 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
 443                         char *fname, *fname_old;
 444
 445                         if (!midx_cleared) {
 446                                 clear_midx_file(the_repository);
 447                                 midx_cleared = 1;
 448                         }
 449
 450                         fname = mkpathdup("%s/pack-%s%s", packdir,
 451                                                 item->string, exts[ext].name);
 452                         if (!file_exists(fname)) {
 453                                 free(fname);
 454                                 continue;
 455                         }
 456
 457                         fname_old = mkpathdup("%s/old-%s%s", packdir,
 458                                                 item->string, exts[ext].name);
 459                         if (file_exists(fname_old))
 460                                 if (unlink(fname_old))
 461                                         failed = 1;
 462
 463                         if (!failed && rename(fname, fname_old)) {
 464                                 free(fname);
 465                                 free(fname_old);
 466                                 failed = 1;
 467                                 break;
 468                         } else {
 469                                 string_list_append(&rollback, fname);
 470                                 free(fname_old);
 471                         }
 472                 }
 473                 if (failed)
 474                         break;
 475         }
 476         if (failed) {
 477                 struct string_list rollback_failure = STRING_LIST_INIT_DUP;
 478                 for_each_string_list_item(item, &rollback) {
 479                         char *fname, *fname_old;
 480                         fname = mkpathdup("%s/%s", packdir, item->string);
 481                         fname_old = mkpathdup("%s/old-%s", packdir, item->string);
 482                         if (rename(fname_old, fname))
 483                                 string_list_append(&rollback_failure, fname);
 484                         free(fname);
 485                         free(fname_old);
 486                 }
 487
 488                 if (rollback_failure.nr) {
 489                         int i;
 490                         fprintf(stderr,
 491                                 _("WARNING: Some packs in use have been renamed by\n"
 492                                   "WARNING: prefixing old- to their name, in order to\n"
 493                                   "WARNING: replace them with the new version of the\n"
 494                                   "WARNING: file.  But the operation failed, and the\n"
 495                                   "WARNING: attempt to rename them back to their\n"
 496                                   "WARNING: original names also failed.\n"
 497                                   "WARNING: Please rename them in %s manually:\n"), packdir);
 498                         for (i = 0; i < rollback_failure.nr; i++)
 499                                 fprintf(stderr, "WARNING:   old-%s -> %s\n",
 500                                         rollback_failure.items[i].string,
 501                                         rollback_failure.items[i].string);
 502                 }
 503                 exit(1);
 504         }
 505
 506         /* Now the ones with the same name are out of the way... */
 507         for_each_string_list_item(item, &names) {
 508                 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
 509                         char *fname, *fname_old;
 510                         struct stat statbuffer;
 511                         int exists = 0;
 512                         fname = mkpathdup("%s/pack-%s%s",
 513                                         packdir, item->string, exts[ext].name);
 514                         fname_old = mkpathdup("%s-%s%s",
 515                                         packtmp, item->string, exts[ext].name);
 516                         if (!stat(fname_old, &statbuffer)) {
 517                                 statbuffer.st_mode &= ~(S_IWUSR | S_IWGRP | S_IWOTH);
 518                                 chmod(fname_old, statbuffer.st_mode);
 519                                 exists = 1;
 520                         }
 521                         if (exists || !exts[ext].optional) {
 522                                 if (rename(fname_old, fname))
 523                                         die_errno(_("renaming '%s' failed"), fname_old);
 524                         }
 525                         free(fname);
 526                         free(fname_old);
 527                 }
 528         }
 529
 530         /* Remove the "old-" files */
 531         for_each_string_list_item(item, &names) {
 532                 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
 533                         char *fname;
 534                         fname = mkpathdup("%s/old-%s%s",
 535                                           packdir,
 536                                           item->string,
 537                                           exts[ext].name);
 538                         if (remove_path(fname))
 539                                 warning(_("failed to remove '%s'"), fname);
 540                         free(fname);
 541                 }
 542         }
 543
 544         /* End of pack replacement. */
 545
 546         reprepare_packed_git(the_repository);
 547
 548         if (delete_redundant) {
 549                 const int hexsz = the_hash_algo->hexsz;
 550                 int opts = 0;
 551                 string_list_sort(&names);
 552                 for_each_string_list_item(item, &existing_packs) {
 553                         char *sha1;
 554                         size_t len = strlen(item->string);
 555                         if (len < hexsz)
 556                                 continue;
 557                         sha1 = item->string + len - hexsz;
 558                         if (!string_list_has_string(&names, sha1))
 559                                 remove_redundant_pack(packdir, item->string);
 560                 }
 561                 if (!po_args.quiet && isatty(2))
 562                         opts |= PRUNE_PACKED_VERBOSE;
 563                 prune_packed_objects(opts);
 564
 565                 if (!keep_unreachable &&
 566                     (!(pack_everything & LOOSEN_UNREACHABLE) ||
 567                      unpack_unreachable) &&
 568                     is_repository_shallow(the_repository))
 569                         prune_shallow(PRUNE_QUICK);
 570         }
 571
 572         if (!no_update_server_info)
 573                 update_server_info(0);
 574         remove_temporary_files();
 575
 576         if (git_env_bool(GIT_TEST_MULTI_PACK_INDEX, 0))
 577                 write_midx_file(get_object_directory());
 578
 579         string_list_clear(&names, 0);
 580         string_list_clear(&rollback, 0);
 581         string_list_clear(&existing_packs, 0);
 582         strbuf_release(&line);
 583
 584         return 0;
 585 }