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