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