Merge branch 'rs/no-mode-to-open-when-appending' into maint
[git.git] / builtin / gc.c
blobac60662619e9d14ab5100712c2964a5b6eee8619
1 /*
2 * git gc builtin command
4 * Cleanup unreachable files and optimize the repository.
6 * Copyright (c) 2007 James Bowes
8 * Based on git-gc.sh, which is
10 * Copyright (c) 2006 Shawn O. Pearce
13 #include "builtin.h"
14 #include "repository.h"
15 #include "config.h"
16 #include "tempfile.h"
17 #include "lockfile.h"
18 #include "parse-options.h"
19 #include "run-command.h"
20 #include "sigchain.h"
21 #include "strvec.h"
22 #include "commit.h"
23 #include "commit-graph.h"
24 #include "packfile.h"
25 #include "object-store.h"
26 #include "pack.h"
27 #include "pack-objects.h"
28 #include "blob.h"
29 #include "tree.h"
30 #include "promisor-remote.h"
31 #include "refs.h"
32 #include "remote.h"
33 #include "object-store.h"
34 #include "exec-cmd.h"
36 #define FAILED_RUN "failed to run %s"
38 static const char * const builtin_gc_usage[] = {
39 N_("git gc [<options>]"),
40 NULL
43 static int pack_refs = 1;
44 static int prune_reflogs = 1;
45 static int aggressive_depth = 50;
46 static int aggressive_window = 250;
47 static int gc_auto_threshold = 6700;
48 static int gc_auto_pack_limit = 50;
49 static int detach_auto = 1;
50 static timestamp_t gc_log_expire_time;
51 static const char *gc_log_expire = "1.day.ago";
52 static const char *prune_expire = "2.weeks.ago";
53 static const char *prune_worktrees_expire = "3.months.ago";
54 static unsigned long big_pack_threshold;
55 static unsigned long max_delta_cache_size = DEFAULT_DELTA_CACHE_SIZE;
57 static struct strvec reflog = STRVEC_INIT;
58 static struct strvec repack = STRVEC_INIT;
59 static struct strvec prune = STRVEC_INIT;
60 static struct strvec prune_worktrees = STRVEC_INIT;
61 static struct strvec rerere = STRVEC_INIT;
63 static struct tempfile *pidfile;
64 static struct lock_file log_lock;
66 static struct string_list pack_garbage = STRING_LIST_INIT_DUP;
68 static void clean_pack_garbage(void)
70 int i;
71 for (i = 0; i < pack_garbage.nr; i++)
72 unlink_or_warn(pack_garbage.items[i].string);
73 string_list_clear(&pack_garbage, 0);
76 static void report_pack_garbage(unsigned seen_bits, const char *path)
78 if (seen_bits == PACKDIR_FILE_IDX)
79 string_list_append(&pack_garbage, path);
82 static void process_log_file(void)
84 struct stat st;
85 if (fstat(get_lock_file_fd(&log_lock), &st)) {
87 * Perhaps there was an i/o error or another
88 * unlikely situation. Try to make a note of
89 * this in gc.log along with any existing
90 * messages.
92 int saved_errno = errno;
93 fprintf(stderr, _("Failed to fstat %s: %s"),
94 get_lock_file_path(&log_lock),
95 strerror(saved_errno));
96 fflush(stderr);
97 commit_lock_file(&log_lock);
98 errno = saved_errno;
99 } else if (st.st_size) {
100 /* There was some error recorded in the lock file */
101 commit_lock_file(&log_lock);
102 } else {
103 /* No error, clean up any old gc.log */
104 unlink(git_path("gc.log"));
105 rollback_lock_file(&log_lock);
109 static void process_log_file_at_exit(void)
111 fflush(stderr);
112 process_log_file();
115 static void process_log_file_on_signal(int signo)
117 process_log_file();
118 sigchain_pop(signo);
119 raise(signo);
122 static int gc_config_is_timestamp_never(const char *var)
124 const char *value;
125 timestamp_t expire;
127 if (!git_config_get_value(var, &value) && value) {
128 if (parse_expiry_date(value, &expire))
129 die(_("failed to parse '%s' value '%s'"), var, value);
130 return expire == 0;
132 return 0;
135 static void gc_config(void)
137 const char *value;
139 if (!git_config_get_value("gc.packrefs", &value)) {
140 if (value && !strcmp(value, "notbare"))
141 pack_refs = -1;
142 else
143 pack_refs = git_config_bool("gc.packrefs", value);
146 if (gc_config_is_timestamp_never("gc.reflogexpire") &&
147 gc_config_is_timestamp_never("gc.reflogexpireunreachable"))
148 prune_reflogs = 0;
150 git_config_get_int("gc.aggressivewindow", &aggressive_window);
151 git_config_get_int("gc.aggressivedepth", &aggressive_depth);
152 git_config_get_int("gc.auto", &gc_auto_threshold);
153 git_config_get_int("gc.autopacklimit", &gc_auto_pack_limit);
154 git_config_get_bool("gc.autodetach", &detach_auto);
155 git_config_get_expiry("gc.pruneexpire", &prune_expire);
156 git_config_get_expiry("gc.worktreepruneexpire", &prune_worktrees_expire);
157 git_config_get_expiry("gc.logexpiry", &gc_log_expire);
159 git_config_get_ulong("gc.bigpackthreshold", &big_pack_threshold);
160 git_config_get_ulong("pack.deltacachesize", &max_delta_cache_size);
162 git_config(git_default_config, NULL);
165 struct maintenance_run_opts;
166 static int maintenance_task_pack_refs(MAYBE_UNUSED struct maintenance_run_opts *opts)
168 struct strvec pack_refs_cmd = STRVEC_INIT;
169 strvec_pushl(&pack_refs_cmd, "pack-refs", "--all", "--prune", NULL);
171 return run_command_v_opt(pack_refs_cmd.v, RUN_GIT_CMD);
174 static int too_many_loose_objects(void)
177 * Quickly check if a "gc" is needed, by estimating how
178 * many loose objects there are. Because SHA-1 is evenly
179 * distributed, we can check only one and get a reasonable
180 * estimate.
182 DIR *dir;
183 struct dirent *ent;
184 int auto_threshold;
185 int num_loose = 0;
186 int needed = 0;
187 const unsigned hexsz_loose = the_hash_algo->hexsz - 2;
189 dir = opendir(git_path("objects/17"));
190 if (!dir)
191 return 0;
193 auto_threshold = DIV_ROUND_UP(gc_auto_threshold, 256);
194 while ((ent = readdir(dir)) != NULL) {
195 if (strspn(ent->d_name, "0123456789abcdef") != hexsz_loose ||
196 ent->d_name[hexsz_loose] != '\0')
197 continue;
198 if (++num_loose > auto_threshold) {
199 needed = 1;
200 break;
203 closedir(dir);
204 return needed;
207 static struct packed_git *find_base_packs(struct string_list *packs,
208 unsigned long limit)
210 struct packed_git *p, *base = NULL;
212 for (p = get_all_packs(the_repository); p; p = p->next) {
213 if (!p->pack_local)
214 continue;
215 if (limit) {
216 if (p->pack_size >= limit)
217 string_list_append(packs, p->pack_name);
218 } else if (!base || base->pack_size < p->pack_size) {
219 base = p;
223 if (base)
224 string_list_append(packs, base->pack_name);
226 return base;
229 static int too_many_packs(void)
231 struct packed_git *p;
232 int cnt;
234 if (gc_auto_pack_limit <= 0)
235 return 0;
237 for (cnt = 0, p = get_all_packs(the_repository); p; p = p->next) {
238 if (!p->pack_local)
239 continue;
240 if (p->pack_keep)
241 continue;
243 * Perhaps check the size of the pack and count only
244 * very small ones here?
246 cnt++;
248 return gc_auto_pack_limit < cnt;
251 static uint64_t total_ram(void)
253 #if defined(HAVE_SYSINFO)
254 struct sysinfo si;
256 if (!sysinfo(&si))
257 return si.totalram;
258 #elif defined(HAVE_BSD_SYSCTL) && (defined(HW_MEMSIZE) || defined(HW_PHYSMEM))
259 int64_t physical_memory;
260 int mib[2];
261 size_t length;
263 mib[0] = CTL_HW;
264 # if defined(HW_MEMSIZE)
265 mib[1] = HW_MEMSIZE;
266 # else
267 mib[1] = HW_PHYSMEM;
268 # endif
269 length = sizeof(int64_t);
270 if (!sysctl(mib, 2, &physical_memory, &length, NULL, 0))
271 return physical_memory;
272 #elif defined(GIT_WINDOWS_NATIVE)
273 MEMORYSTATUSEX memInfo;
275 memInfo.dwLength = sizeof(MEMORYSTATUSEX);
276 if (GlobalMemoryStatusEx(&memInfo))
277 return memInfo.ullTotalPhys;
278 #endif
279 return 0;
282 static uint64_t estimate_repack_memory(struct packed_git *pack)
284 unsigned long nr_objects = approximate_object_count();
285 size_t os_cache, heap;
287 if (!pack || !nr_objects)
288 return 0;
291 * First we have to scan through at least one pack.
292 * Assume enough room in OS file cache to keep the entire pack
293 * or we may accidentally evict data of other processes from
294 * the cache.
296 os_cache = pack->pack_size + pack->index_size;
297 /* then pack-objects needs lots more for book keeping */
298 heap = sizeof(struct object_entry) * nr_objects;
300 * internal rev-list --all --objects takes up some memory too,
301 * let's say half of it is for blobs
303 heap += sizeof(struct blob) * nr_objects / 2;
305 * and the other half is for trees (commits and tags are
306 * usually insignificant)
308 heap += sizeof(struct tree) * nr_objects / 2;
309 /* and then obj_hash[], underestimated in fact */
310 heap += sizeof(struct object *) * nr_objects;
311 /* revindex is used also */
312 heap += (sizeof(off_t) + sizeof(uint32_t)) * nr_objects;
314 * read_sha1_file() (either at delta calculation phase, or
315 * writing phase) also fills up the delta base cache
317 heap += delta_base_cache_limit;
318 /* and of course pack-objects has its own delta cache */
319 heap += max_delta_cache_size;
321 return os_cache + heap;
324 static int keep_one_pack(struct string_list_item *item, void *data)
326 strvec_pushf(&repack, "--keep-pack=%s", basename(item->string));
327 return 0;
330 static void add_repack_all_option(struct string_list *keep_pack)
332 if (prune_expire && !strcmp(prune_expire, "now"))
333 strvec_push(&repack, "-a");
334 else {
335 strvec_push(&repack, "-A");
336 if (prune_expire)
337 strvec_pushf(&repack, "--unpack-unreachable=%s", prune_expire);
340 if (keep_pack)
341 for_each_string_list(keep_pack, keep_one_pack, NULL);
344 static void add_repack_incremental_option(void)
346 strvec_push(&repack, "--no-write-bitmap-index");
349 static int need_to_gc(void)
352 * Setting gc.auto to 0 or negative can disable the
353 * automatic gc.
355 if (gc_auto_threshold <= 0)
356 return 0;
359 * If there are too many loose objects, but not too many
360 * packs, we run "repack -d -l". If there are too many packs,
361 * we run "repack -A -d -l". Otherwise we tell the caller
362 * there is no need.
364 if (too_many_packs()) {
365 struct string_list keep_pack = STRING_LIST_INIT_NODUP;
367 if (big_pack_threshold) {
368 find_base_packs(&keep_pack, big_pack_threshold);
369 if (keep_pack.nr >= gc_auto_pack_limit) {
370 big_pack_threshold = 0;
371 string_list_clear(&keep_pack, 0);
372 find_base_packs(&keep_pack, 0);
374 } else {
375 struct packed_git *p = find_base_packs(&keep_pack, 0);
376 uint64_t mem_have, mem_want;
378 mem_have = total_ram();
379 mem_want = estimate_repack_memory(p);
382 * Only allow 1/2 of memory for pack-objects, leave
383 * the rest for the OS and other processes in the
384 * system.
386 if (!mem_have || mem_want < mem_have / 2)
387 string_list_clear(&keep_pack, 0);
390 add_repack_all_option(&keep_pack);
391 string_list_clear(&keep_pack, 0);
392 } else if (too_many_loose_objects())
393 add_repack_incremental_option();
394 else
395 return 0;
397 if (run_hook_le(NULL, "pre-auto-gc", NULL))
398 return 0;
399 return 1;
402 /* return NULL on success, else hostname running the gc */
403 static const char *lock_repo_for_gc(int force, pid_t* ret_pid)
405 struct lock_file lock = LOCK_INIT;
406 char my_host[HOST_NAME_MAX + 1];
407 struct strbuf sb = STRBUF_INIT;
408 struct stat st;
409 uintmax_t pid;
410 FILE *fp;
411 int fd;
412 char *pidfile_path;
414 if (is_tempfile_active(pidfile))
415 /* already locked */
416 return NULL;
418 if (xgethostname(my_host, sizeof(my_host)))
419 xsnprintf(my_host, sizeof(my_host), "unknown");
421 pidfile_path = git_pathdup("gc.pid");
422 fd = hold_lock_file_for_update(&lock, pidfile_path,
423 LOCK_DIE_ON_ERROR);
424 if (!force) {
425 static char locking_host[HOST_NAME_MAX + 1];
426 static char *scan_fmt;
427 int should_exit;
429 if (!scan_fmt)
430 scan_fmt = xstrfmt("%s %%%ds", "%"SCNuMAX, HOST_NAME_MAX);
431 fp = fopen(pidfile_path, "r");
432 memset(locking_host, 0, sizeof(locking_host));
433 should_exit =
434 fp != NULL &&
435 !fstat(fileno(fp), &st) &&
437 * 12 hour limit is very generous as gc should
438 * never take that long. On the other hand we
439 * don't really need a strict limit here,
440 * running gc --auto one day late is not a big
441 * problem. --force can be used in manual gc
442 * after the user verifies that no gc is
443 * running.
445 time(NULL) - st.st_mtime <= 12 * 3600 &&
446 fscanf(fp, scan_fmt, &pid, locking_host) == 2 &&
447 /* be gentle to concurrent "gc" on remote hosts */
448 (strcmp(locking_host, my_host) || !kill(pid, 0) || errno == EPERM);
449 if (fp != NULL)
450 fclose(fp);
451 if (should_exit) {
452 if (fd >= 0)
453 rollback_lock_file(&lock);
454 *ret_pid = pid;
455 free(pidfile_path);
456 return locking_host;
460 strbuf_addf(&sb, "%"PRIuMAX" %s",
461 (uintmax_t) getpid(), my_host);
462 write_in_full(fd, sb.buf, sb.len);
463 strbuf_release(&sb);
464 commit_lock_file(&lock);
465 pidfile = register_tempfile(pidfile_path);
466 free(pidfile_path);
467 return NULL;
471 * Returns 0 if there was no previous error and gc can proceed, 1 if
472 * gc should not proceed due to an error in the last run. Prints a
473 * message and returns -1 if an error occurred while reading gc.log
475 static int report_last_gc_error(void)
477 struct strbuf sb = STRBUF_INIT;
478 int ret = 0;
479 ssize_t len;
480 struct stat st;
481 char *gc_log_path = git_pathdup("gc.log");
483 if (stat(gc_log_path, &st)) {
484 if (errno == ENOENT)
485 goto done;
487 ret = error_errno(_("cannot stat '%s'"), gc_log_path);
488 goto done;
491 if (st.st_mtime < gc_log_expire_time)
492 goto done;
494 len = strbuf_read_file(&sb, gc_log_path, 0);
495 if (len < 0)
496 ret = error_errno(_("cannot read '%s'"), gc_log_path);
497 else if (len > 0) {
499 * A previous gc failed. Report the error, and don't
500 * bother with an automatic gc run since it is likely
501 * to fail in the same way.
503 warning(_("The last gc run reported the following. "
504 "Please correct the root cause\n"
505 "and remove %s\n"
506 "Automatic cleanup will not be performed "
507 "until the file is removed.\n\n"
508 "%s"),
509 gc_log_path, sb.buf);
510 ret = 1;
512 strbuf_release(&sb);
513 done:
514 free(gc_log_path);
515 return ret;
518 static void gc_before_repack(void)
521 * We may be called twice, as both the pre- and
522 * post-daemonized phases will call us, but running these
523 * commands more than once is pointless and wasteful.
525 static int done = 0;
526 if (done++)
527 return;
529 if (pack_refs && maintenance_task_pack_refs(NULL))
530 die(FAILED_RUN, "pack-refs");
532 if (prune_reflogs && run_command_v_opt(reflog.v, RUN_GIT_CMD))
533 die(FAILED_RUN, reflog.v[0]);
536 int cmd_gc(int argc, const char **argv, const char *prefix)
538 int aggressive = 0;
539 int auto_gc = 0;
540 int quiet = 0;
541 int force = 0;
542 const char *name;
543 pid_t pid;
544 int daemonized = 0;
545 int keep_largest_pack = -1;
546 timestamp_t dummy;
548 struct option builtin_gc_options[] = {
549 OPT__QUIET(&quiet, N_("suppress progress reporting")),
550 { OPTION_STRING, 0, "prune", &prune_expire, N_("date"),
551 N_("prune unreferenced objects"),
552 PARSE_OPT_OPTARG, NULL, (intptr_t)prune_expire },
553 OPT_BOOL(0, "aggressive", &aggressive, N_("be more thorough (increased runtime)")),
554 OPT_BOOL_F(0, "auto", &auto_gc, N_("enable auto-gc mode"),
555 PARSE_OPT_NOCOMPLETE),
556 OPT_BOOL_F(0, "force", &force,
557 N_("force running gc even if there may be another gc running"),
558 PARSE_OPT_NOCOMPLETE),
559 OPT_BOOL(0, "keep-largest-pack", &keep_largest_pack,
560 N_("repack all other packs except the largest pack")),
561 OPT_END()
564 if (argc == 2 && !strcmp(argv[1], "-h"))
565 usage_with_options(builtin_gc_usage, builtin_gc_options);
567 strvec_pushl(&reflog, "reflog", "expire", "--all", NULL);
568 strvec_pushl(&repack, "repack", "-d", "-l", NULL);
569 strvec_pushl(&prune, "prune", "--expire", NULL);
570 strvec_pushl(&prune_worktrees, "worktree", "prune", "--expire", NULL);
571 strvec_pushl(&rerere, "rerere", "gc", NULL);
573 /* default expiry time, overwritten in gc_config */
574 gc_config();
575 if (parse_expiry_date(gc_log_expire, &gc_log_expire_time))
576 die(_("failed to parse gc.logexpiry value %s"), gc_log_expire);
578 if (pack_refs < 0)
579 pack_refs = !is_bare_repository();
581 argc = parse_options(argc, argv, prefix, builtin_gc_options,
582 builtin_gc_usage, 0);
583 if (argc > 0)
584 usage_with_options(builtin_gc_usage, builtin_gc_options);
586 if (prune_expire && parse_expiry_date(prune_expire, &dummy))
587 die(_("failed to parse prune expiry value %s"), prune_expire);
589 if (aggressive) {
590 strvec_push(&repack, "-f");
591 if (aggressive_depth > 0)
592 strvec_pushf(&repack, "--depth=%d", aggressive_depth);
593 if (aggressive_window > 0)
594 strvec_pushf(&repack, "--window=%d", aggressive_window);
596 if (quiet)
597 strvec_push(&repack, "-q");
599 if (auto_gc) {
601 * Auto-gc should be least intrusive as possible.
603 if (!need_to_gc())
604 return 0;
605 if (!quiet) {
606 if (detach_auto)
607 fprintf(stderr, _("Auto packing the repository in background for optimum performance.\n"));
608 else
609 fprintf(stderr, _("Auto packing the repository for optimum performance.\n"));
610 fprintf(stderr, _("See \"git help gc\" for manual housekeeping.\n"));
612 if (detach_auto) {
613 int ret = report_last_gc_error();
614 if (ret < 0)
615 /* an I/O error occurred, already reported */
616 exit(128);
617 if (ret == 1)
618 /* Last gc --auto failed. Skip this one. */
619 return 0;
621 if (lock_repo_for_gc(force, &pid))
622 return 0;
623 gc_before_repack(); /* dies on failure */
624 delete_tempfile(&pidfile);
627 * failure to daemonize is ok, we'll continue
628 * in foreground
630 daemonized = !daemonize();
632 } else {
633 struct string_list keep_pack = STRING_LIST_INIT_NODUP;
635 if (keep_largest_pack != -1) {
636 if (keep_largest_pack)
637 find_base_packs(&keep_pack, 0);
638 } else if (big_pack_threshold) {
639 find_base_packs(&keep_pack, big_pack_threshold);
642 add_repack_all_option(&keep_pack);
643 string_list_clear(&keep_pack, 0);
646 name = lock_repo_for_gc(force, &pid);
647 if (name) {
648 if (auto_gc)
649 return 0; /* be quiet on --auto */
650 die(_("gc is already running on machine '%s' pid %"PRIuMAX" (use --force if not)"),
651 name, (uintmax_t)pid);
654 if (daemonized) {
655 hold_lock_file_for_update(&log_lock,
656 git_path("gc.log"),
657 LOCK_DIE_ON_ERROR);
658 dup2(get_lock_file_fd(&log_lock), 2);
659 sigchain_push_common(process_log_file_on_signal);
660 atexit(process_log_file_at_exit);
663 gc_before_repack();
665 if (!repository_format_precious_objects) {
666 close_object_store(the_repository->objects);
667 if (run_command_v_opt(repack.v, RUN_GIT_CMD))
668 die(FAILED_RUN, repack.v[0]);
670 if (prune_expire) {
671 strvec_push(&prune, prune_expire);
672 if (quiet)
673 strvec_push(&prune, "--no-progress");
674 if (has_promisor_remote())
675 strvec_push(&prune,
676 "--exclude-promisor-objects");
677 if (run_command_v_opt(prune.v, RUN_GIT_CMD))
678 die(FAILED_RUN, prune.v[0]);
682 if (prune_worktrees_expire) {
683 strvec_push(&prune_worktrees, prune_worktrees_expire);
684 if (run_command_v_opt(prune_worktrees.v, RUN_GIT_CMD))
685 die(FAILED_RUN, prune_worktrees.v[0]);
688 if (run_command_v_opt(rerere.v, RUN_GIT_CMD))
689 die(FAILED_RUN, rerere.v[0]);
691 report_garbage = report_pack_garbage;
692 reprepare_packed_git(the_repository);
693 if (pack_garbage.nr > 0) {
694 close_object_store(the_repository->objects);
695 clean_pack_garbage();
698 prepare_repo_settings(the_repository);
699 if (the_repository->settings.gc_write_commit_graph == 1)
700 write_commit_graph_reachable(the_repository->objects->odb,
701 !quiet && !daemonized ? COMMIT_GRAPH_WRITE_PROGRESS : 0,
702 NULL);
704 if (auto_gc && too_many_loose_objects())
705 warning(_("There are too many unreachable loose objects; "
706 "run 'git prune' to remove them."));
708 if (!daemonized)
709 unlink(git_path("gc.log"));
711 return 0;
714 static const char *const builtin_maintenance_run_usage[] = {
715 N_("git maintenance run [--auto] [--[no-]quiet] [--task=<task>] [--schedule]"),
716 NULL
719 enum schedule_priority {
720 SCHEDULE_NONE = 0,
721 SCHEDULE_WEEKLY = 1,
722 SCHEDULE_DAILY = 2,
723 SCHEDULE_HOURLY = 3,
726 static enum schedule_priority parse_schedule(const char *value)
728 if (!value)
729 return SCHEDULE_NONE;
730 if (!strcasecmp(value, "hourly"))
731 return SCHEDULE_HOURLY;
732 if (!strcasecmp(value, "daily"))
733 return SCHEDULE_DAILY;
734 if (!strcasecmp(value, "weekly"))
735 return SCHEDULE_WEEKLY;
736 return SCHEDULE_NONE;
739 static int maintenance_opt_schedule(const struct option *opt, const char *arg,
740 int unset)
742 enum schedule_priority *priority = opt->value;
744 if (unset)
745 die(_("--no-schedule is not allowed"));
747 *priority = parse_schedule(arg);
749 if (!*priority)
750 die(_("unrecognized --schedule argument '%s'"), arg);
752 return 0;
755 struct maintenance_run_opts {
756 int auto_flag;
757 int quiet;
758 enum schedule_priority schedule;
761 /* Remember to update object flag allocation in object.h */
762 #define SEEN (1u<<0)
764 struct cg_auto_data {
765 int num_not_in_graph;
766 int limit;
769 static int dfs_on_ref(const char *refname,
770 const struct object_id *oid, int flags,
771 void *cb_data)
773 struct cg_auto_data *data = (struct cg_auto_data *)cb_data;
774 int result = 0;
775 struct object_id peeled;
776 struct commit_list *stack = NULL;
777 struct commit *commit;
779 if (!peel_iterated_oid(oid, &peeled))
780 oid = &peeled;
781 if (oid_object_info(the_repository, oid, NULL) != OBJ_COMMIT)
782 return 0;
784 commit = lookup_commit(the_repository, oid);
785 if (!commit)
786 return 0;
787 if (parse_commit(commit) ||
788 commit_graph_position(commit) != COMMIT_NOT_FROM_GRAPH)
789 return 0;
791 data->num_not_in_graph++;
793 if (data->num_not_in_graph >= data->limit)
794 return 1;
796 commit_list_append(commit, &stack);
798 while (!result && stack) {
799 struct commit_list *parent;
801 commit = pop_commit(&stack);
803 for (parent = commit->parents; parent; parent = parent->next) {
804 if (parse_commit(parent->item) ||
805 commit_graph_position(parent->item) != COMMIT_NOT_FROM_GRAPH ||
806 parent->item->object.flags & SEEN)
807 continue;
809 parent->item->object.flags |= SEEN;
810 data->num_not_in_graph++;
812 if (data->num_not_in_graph >= data->limit) {
813 result = 1;
814 break;
817 commit_list_append(parent->item, &stack);
821 free_commit_list(stack);
822 return result;
825 static int should_write_commit_graph(void)
827 int result;
828 struct cg_auto_data data;
830 data.num_not_in_graph = 0;
831 data.limit = 100;
832 git_config_get_int("maintenance.commit-graph.auto",
833 &data.limit);
835 if (!data.limit)
836 return 0;
837 if (data.limit < 0)
838 return 1;
840 result = for_each_ref(dfs_on_ref, &data);
842 repo_clear_commit_marks(the_repository, SEEN);
844 return result;
847 static int run_write_commit_graph(struct maintenance_run_opts *opts)
849 struct child_process child = CHILD_PROCESS_INIT;
851 child.git_cmd = 1;
852 strvec_pushl(&child.args, "commit-graph", "write",
853 "--split", "--reachable", NULL);
855 if (opts->quiet)
856 strvec_push(&child.args, "--no-progress");
858 return !!run_command(&child);
861 static int maintenance_task_commit_graph(struct maintenance_run_opts *opts)
863 prepare_repo_settings(the_repository);
864 if (!the_repository->settings.core_commit_graph)
865 return 0;
867 close_object_store(the_repository->objects);
868 if (run_write_commit_graph(opts)) {
869 error(_("failed to write commit-graph"));
870 return 1;
873 return 0;
876 static int fetch_remote(struct remote *remote, void *cbdata)
878 struct maintenance_run_opts *opts = cbdata;
879 struct child_process child = CHILD_PROCESS_INIT;
881 if (remote->skip_default_update)
882 return 0;
884 child.git_cmd = 1;
885 strvec_pushl(&child.args, "fetch", remote->name,
886 "--prefetch", "--prune", "--no-tags",
887 "--no-write-fetch-head", "--recurse-submodules=no",
888 NULL);
890 if (opts->quiet)
891 strvec_push(&child.args, "--quiet");
893 return !!run_command(&child);
896 static int maintenance_task_prefetch(struct maintenance_run_opts *opts)
898 git_config_set_multivar_gently("log.excludedecoration",
899 "refs/prefetch/",
900 "refs/prefetch/",
901 CONFIG_FLAGS_FIXED_VALUE |
902 CONFIG_FLAGS_MULTI_REPLACE);
904 if (for_each_remote(fetch_remote, opts)) {
905 error(_("failed to prefetch remotes"));
906 return 1;
909 return 0;
912 static int maintenance_task_gc(struct maintenance_run_opts *opts)
914 struct child_process child = CHILD_PROCESS_INIT;
916 child.git_cmd = 1;
917 strvec_push(&child.args, "gc");
919 if (opts->auto_flag)
920 strvec_push(&child.args, "--auto");
921 if (opts->quiet)
922 strvec_push(&child.args, "--quiet");
923 else
924 strvec_push(&child.args, "--no-quiet");
926 close_object_store(the_repository->objects);
927 return run_command(&child);
930 static int prune_packed(struct maintenance_run_opts *opts)
932 struct child_process child = CHILD_PROCESS_INIT;
934 child.git_cmd = 1;
935 strvec_push(&child.args, "prune-packed");
937 if (opts->quiet)
938 strvec_push(&child.args, "--quiet");
940 return !!run_command(&child);
943 struct write_loose_object_data {
944 FILE *in;
945 int count;
946 int batch_size;
949 static int loose_object_auto_limit = 100;
951 static int loose_object_count(const struct object_id *oid,
952 const char *path,
953 void *data)
955 int *count = (int*)data;
956 if (++(*count) >= loose_object_auto_limit)
957 return 1;
958 return 0;
961 static int loose_object_auto_condition(void)
963 int count = 0;
965 git_config_get_int("maintenance.loose-objects.auto",
966 &loose_object_auto_limit);
968 if (!loose_object_auto_limit)
969 return 0;
970 if (loose_object_auto_limit < 0)
971 return 1;
973 return for_each_loose_file_in_objdir(the_repository->objects->odb->path,
974 loose_object_count,
975 NULL, NULL, &count);
978 static int bail_on_loose(const struct object_id *oid,
979 const char *path,
980 void *data)
982 return 1;
985 static int write_loose_object_to_stdin(const struct object_id *oid,
986 const char *path,
987 void *data)
989 struct write_loose_object_data *d = (struct write_loose_object_data *)data;
991 fprintf(d->in, "%s\n", oid_to_hex(oid));
993 return ++(d->count) > d->batch_size;
996 static int pack_loose(struct maintenance_run_opts *opts)
998 struct repository *r = the_repository;
999 int result = 0;
1000 struct write_loose_object_data data;
1001 struct child_process pack_proc = CHILD_PROCESS_INIT;
1004 * Do not start pack-objects process
1005 * if there are no loose objects.
1007 if (!for_each_loose_file_in_objdir(r->objects->odb->path,
1008 bail_on_loose,
1009 NULL, NULL, NULL))
1010 return 0;
1012 pack_proc.git_cmd = 1;
1014 strvec_push(&pack_proc.args, "pack-objects");
1015 if (opts->quiet)
1016 strvec_push(&pack_proc.args, "--quiet");
1017 strvec_pushf(&pack_proc.args, "%s/pack/loose", r->objects->odb->path);
1019 pack_proc.in = -1;
1021 if (start_command(&pack_proc)) {
1022 error(_("failed to start 'git pack-objects' process"));
1023 return 1;
1026 data.in = xfdopen(pack_proc.in, "w");
1027 data.count = 0;
1028 data.batch_size = 50000;
1030 for_each_loose_file_in_objdir(r->objects->odb->path,
1031 write_loose_object_to_stdin,
1032 NULL,
1033 NULL,
1034 &data);
1036 fclose(data.in);
1038 if (finish_command(&pack_proc)) {
1039 error(_("failed to finish 'git pack-objects' process"));
1040 result = 1;
1043 return result;
1046 static int maintenance_task_loose_objects(struct maintenance_run_opts *opts)
1048 return prune_packed(opts) || pack_loose(opts);
1051 static int incremental_repack_auto_condition(void)
1053 struct packed_git *p;
1054 int enabled;
1055 int incremental_repack_auto_limit = 10;
1056 int count = 0;
1058 if (git_config_get_bool("core.multiPackIndex", &enabled) ||
1059 !enabled)
1060 return 0;
1062 git_config_get_int("maintenance.incremental-repack.auto",
1063 &incremental_repack_auto_limit);
1065 if (!incremental_repack_auto_limit)
1066 return 0;
1067 if (incremental_repack_auto_limit < 0)
1068 return 1;
1070 for (p = get_packed_git(the_repository);
1071 count < incremental_repack_auto_limit && p;
1072 p = p->next) {
1073 if (!p->multi_pack_index)
1074 count++;
1077 return count >= incremental_repack_auto_limit;
1080 static int multi_pack_index_write(struct maintenance_run_opts *opts)
1082 struct child_process child = CHILD_PROCESS_INIT;
1084 child.git_cmd = 1;
1085 strvec_pushl(&child.args, "multi-pack-index", "write", NULL);
1087 if (opts->quiet)
1088 strvec_push(&child.args, "--no-progress");
1090 if (run_command(&child))
1091 return error(_("failed to write multi-pack-index"));
1093 return 0;
1096 static int multi_pack_index_expire(struct maintenance_run_opts *opts)
1098 struct child_process child = CHILD_PROCESS_INIT;
1100 child.git_cmd = 1;
1101 strvec_pushl(&child.args, "multi-pack-index", "expire", NULL);
1103 if (opts->quiet)
1104 strvec_push(&child.args, "--no-progress");
1106 close_object_store(the_repository->objects);
1108 if (run_command(&child))
1109 return error(_("'git multi-pack-index expire' failed"));
1111 return 0;
1114 #define TWO_GIGABYTES (INT32_MAX)
1116 static off_t get_auto_pack_size(void)
1119 * The "auto" value is special: we optimize for
1120 * one large pack-file (i.e. from a clone) and
1121 * expect the rest to be small and they can be
1122 * repacked quickly.
1124 * The strategy we select here is to select a
1125 * size that is one more than the second largest
1126 * pack-file. This ensures that we will repack
1127 * at least two packs if there are three or more
1128 * packs.
1130 off_t max_size = 0;
1131 off_t second_largest_size = 0;
1132 off_t result_size;
1133 struct packed_git *p;
1134 struct repository *r = the_repository;
1136 reprepare_packed_git(r);
1137 for (p = get_all_packs(r); p; p = p->next) {
1138 if (p->pack_size > max_size) {
1139 second_largest_size = max_size;
1140 max_size = p->pack_size;
1141 } else if (p->pack_size > second_largest_size)
1142 second_largest_size = p->pack_size;
1145 result_size = second_largest_size + 1;
1147 /* But limit ourselves to a batch size of 2g */
1148 if (result_size > TWO_GIGABYTES)
1149 result_size = TWO_GIGABYTES;
1151 return result_size;
1154 static int multi_pack_index_repack(struct maintenance_run_opts *opts)
1156 struct child_process child = CHILD_PROCESS_INIT;
1158 child.git_cmd = 1;
1159 strvec_pushl(&child.args, "multi-pack-index", "repack", NULL);
1161 if (opts->quiet)
1162 strvec_push(&child.args, "--no-progress");
1164 strvec_pushf(&child.args, "--batch-size=%"PRIuMAX,
1165 (uintmax_t)get_auto_pack_size());
1167 close_object_store(the_repository->objects);
1169 if (run_command(&child))
1170 return error(_("'git multi-pack-index repack' failed"));
1172 return 0;
1175 static int maintenance_task_incremental_repack(struct maintenance_run_opts *opts)
1177 prepare_repo_settings(the_repository);
1178 if (!the_repository->settings.core_multi_pack_index) {
1179 warning(_("skipping incremental-repack task because core.multiPackIndex is disabled"));
1180 return 0;
1183 if (multi_pack_index_write(opts))
1184 return 1;
1185 if (multi_pack_index_expire(opts))
1186 return 1;
1187 if (multi_pack_index_repack(opts))
1188 return 1;
1189 return 0;
1192 typedef int maintenance_task_fn(struct maintenance_run_opts *opts);
1195 * An auto condition function returns 1 if the task should run
1196 * and 0 if the task should NOT run. See needs_to_gc() for an
1197 * example.
1199 typedef int maintenance_auto_fn(void);
1201 struct maintenance_task {
1202 const char *name;
1203 maintenance_task_fn *fn;
1204 maintenance_auto_fn *auto_condition;
1205 unsigned enabled:1;
1207 enum schedule_priority schedule;
1209 /* -1 if not selected. */
1210 int selected_order;
1213 enum maintenance_task_label {
1214 TASK_PREFETCH,
1215 TASK_LOOSE_OBJECTS,
1216 TASK_INCREMENTAL_REPACK,
1217 TASK_GC,
1218 TASK_COMMIT_GRAPH,
1219 TASK_PACK_REFS,
1221 /* Leave as final value */
1222 TASK__COUNT
1225 static struct maintenance_task tasks[] = {
1226 [TASK_PREFETCH] = {
1227 "prefetch",
1228 maintenance_task_prefetch,
1230 [TASK_LOOSE_OBJECTS] = {
1231 "loose-objects",
1232 maintenance_task_loose_objects,
1233 loose_object_auto_condition,
1235 [TASK_INCREMENTAL_REPACK] = {
1236 "incremental-repack",
1237 maintenance_task_incremental_repack,
1238 incremental_repack_auto_condition,
1240 [TASK_GC] = {
1241 "gc",
1242 maintenance_task_gc,
1243 need_to_gc,
1246 [TASK_COMMIT_GRAPH] = {
1247 "commit-graph",
1248 maintenance_task_commit_graph,
1249 should_write_commit_graph,
1251 [TASK_PACK_REFS] = {
1252 "pack-refs",
1253 maintenance_task_pack_refs,
1254 NULL,
1258 static int compare_tasks_by_selection(const void *a_, const void *b_)
1260 const struct maintenance_task *a = a_;
1261 const struct maintenance_task *b = b_;
1263 return b->selected_order - a->selected_order;
1266 static int maintenance_run_tasks(struct maintenance_run_opts *opts)
1268 int i, found_selected = 0;
1269 int result = 0;
1270 struct lock_file lk;
1271 struct repository *r = the_repository;
1272 char *lock_path = xstrfmt("%s/maintenance", r->objects->odb->path);
1274 if (hold_lock_file_for_update(&lk, lock_path, LOCK_NO_DEREF) < 0) {
1276 * Another maintenance command is running.
1278 * If --auto was provided, then it is likely due to a
1279 * recursive process stack. Do not report an error in
1280 * that case.
1282 if (!opts->auto_flag && !opts->quiet)
1283 warning(_("lock file '%s' exists, skipping maintenance"),
1284 lock_path);
1285 free(lock_path);
1286 return 0;
1288 free(lock_path);
1290 for (i = 0; !found_selected && i < TASK__COUNT; i++)
1291 found_selected = tasks[i].selected_order >= 0;
1293 if (found_selected)
1294 QSORT(tasks, TASK__COUNT, compare_tasks_by_selection);
1296 for (i = 0; i < TASK__COUNT; i++) {
1297 if (found_selected && tasks[i].selected_order < 0)
1298 continue;
1300 if (!found_selected && !tasks[i].enabled)
1301 continue;
1303 if (opts->auto_flag &&
1304 (!tasks[i].auto_condition ||
1305 !tasks[i].auto_condition()))
1306 continue;
1308 if (opts->schedule && tasks[i].schedule < opts->schedule)
1309 continue;
1311 trace2_region_enter("maintenance", tasks[i].name, r);
1312 if (tasks[i].fn(opts)) {
1313 error(_("task '%s' failed"), tasks[i].name);
1314 result = 1;
1316 trace2_region_leave("maintenance", tasks[i].name, r);
1319 rollback_lock_file(&lk);
1320 return result;
1323 static void initialize_maintenance_strategy(void)
1325 char *config_str;
1327 if (git_config_get_string("maintenance.strategy", &config_str))
1328 return;
1330 if (!strcasecmp(config_str, "incremental")) {
1331 tasks[TASK_GC].schedule = SCHEDULE_NONE;
1332 tasks[TASK_COMMIT_GRAPH].enabled = 1;
1333 tasks[TASK_COMMIT_GRAPH].schedule = SCHEDULE_HOURLY;
1334 tasks[TASK_PREFETCH].enabled = 1;
1335 tasks[TASK_PREFETCH].schedule = SCHEDULE_HOURLY;
1336 tasks[TASK_INCREMENTAL_REPACK].enabled = 1;
1337 tasks[TASK_INCREMENTAL_REPACK].schedule = SCHEDULE_DAILY;
1338 tasks[TASK_LOOSE_OBJECTS].enabled = 1;
1339 tasks[TASK_LOOSE_OBJECTS].schedule = SCHEDULE_DAILY;
1340 tasks[TASK_PACK_REFS].enabled = 1;
1341 tasks[TASK_PACK_REFS].schedule = SCHEDULE_WEEKLY;
1345 static void initialize_task_config(int schedule)
1347 int i;
1348 struct strbuf config_name = STRBUF_INIT;
1349 gc_config();
1351 if (schedule)
1352 initialize_maintenance_strategy();
1354 for (i = 0; i < TASK__COUNT; i++) {
1355 int config_value;
1356 char *config_str;
1358 strbuf_reset(&config_name);
1359 strbuf_addf(&config_name, "maintenance.%s.enabled",
1360 tasks[i].name);
1362 if (!git_config_get_bool(config_name.buf, &config_value))
1363 tasks[i].enabled = config_value;
1365 strbuf_reset(&config_name);
1366 strbuf_addf(&config_name, "maintenance.%s.schedule",
1367 tasks[i].name);
1369 if (!git_config_get_string(config_name.buf, &config_str)) {
1370 tasks[i].schedule = parse_schedule(config_str);
1371 free(config_str);
1375 strbuf_release(&config_name);
1378 static int task_option_parse(const struct option *opt,
1379 const char *arg, int unset)
1381 int i, num_selected = 0;
1382 struct maintenance_task *task = NULL;
1384 BUG_ON_OPT_NEG(unset);
1386 for (i = 0; i < TASK__COUNT; i++) {
1387 if (tasks[i].selected_order >= 0)
1388 num_selected++;
1389 if (!strcasecmp(tasks[i].name, arg)) {
1390 task = &tasks[i];
1394 if (!task) {
1395 error(_("'%s' is not a valid task"), arg);
1396 return 1;
1399 if (task->selected_order >= 0) {
1400 error(_("task '%s' cannot be selected multiple times"), arg);
1401 return 1;
1404 task->selected_order = num_selected + 1;
1406 return 0;
1409 static int maintenance_run(int argc, const char **argv, const char *prefix)
1411 int i;
1412 struct maintenance_run_opts opts;
1413 struct option builtin_maintenance_run_options[] = {
1414 OPT_BOOL(0, "auto", &opts.auto_flag,
1415 N_("run tasks based on the state of the repository")),
1416 OPT_CALLBACK(0, "schedule", &opts.schedule, N_("frequency"),
1417 N_("run tasks based on frequency"),
1418 maintenance_opt_schedule),
1419 OPT_BOOL(0, "quiet", &opts.quiet,
1420 N_("do not report progress or other information over stderr")),
1421 OPT_CALLBACK_F(0, "task", NULL, N_("task"),
1422 N_("run a specific task"),
1423 PARSE_OPT_NONEG, task_option_parse),
1424 OPT_END()
1426 memset(&opts, 0, sizeof(opts));
1428 opts.quiet = !isatty(2);
1430 for (i = 0; i < TASK__COUNT; i++)
1431 tasks[i].selected_order = -1;
1433 argc = parse_options(argc, argv, prefix,
1434 builtin_maintenance_run_options,
1435 builtin_maintenance_run_usage,
1436 PARSE_OPT_STOP_AT_NON_OPTION);
1438 if (opts.auto_flag && opts.schedule)
1439 die(_("use at most one of --auto and --schedule=<frequency>"));
1441 initialize_task_config(opts.schedule);
1443 if (argc != 0)
1444 usage_with_options(builtin_maintenance_run_usage,
1445 builtin_maintenance_run_options);
1446 return maintenance_run_tasks(&opts);
1449 static char *get_maintpath(void)
1451 struct strbuf sb = STRBUF_INIT;
1452 const char *p = the_repository->worktree ?
1453 the_repository->worktree : the_repository->gitdir;
1455 strbuf_realpath(&sb, p, 1);
1456 return strbuf_detach(&sb, NULL);
1459 static int maintenance_register(void)
1461 int rc;
1462 char *config_value;
1463 struct child_process config_set = CHILD_PROCESS_INIT;
1464 struct child_process config_get = CHILD_PROCESS_INIT;
1465 char *maintpath = get_maintpath();
1467 /* Disable foreground maintenance */
1468 git_config_set("maintenance.auto", "false");
1470 /* Set maintenance strategy, if unset */
1471 if (!git_config_get_string("maintenance.strategy", &config_value))
1472 free(config_value);
1473 else
1474 git_config_set("maintenance.strategy", "incremental");
1476 config_get.git_cmd = 1;
1477 strvec_pushl(&config_get.args, "config", "--global", "--get",
1478 "--fixed-value", "maintenance.repo", maintpath, NULL);
1479 config_get.out = -1;
1481 if (start_command(&config_get)) {
1482 rc = error(_("failed to run 'git config'"));
1483 goto done;
1486 /* We already have this value in our config! */
1487 if (!finish_command(&config_get)) {
1488 rc = 0;
1489 goto done;
1492 config_set.git_cmd = 1;
1493 strvec_pushl(&config_set.args, "config", "--add", "--global", "maintenance.repo",
1494 maintpath, NULL);
1496 rc = run_command(&config_set);
1498 done:
1499 free(maintpath);
1500 return rc;
1503 static int maintenance_unregister(void)
1505 int rc;
1506 struct child_process config_unset = CHILD_PROCESS_INIT;
1507 char *maintpath = get_maintpath();
1509 config_unset.git_cmd = 1;
1510 strvec_pushl(&config_unset.args, "config", "--global", "--unset",
1511 "--fixed-value", "maintenance.repo", maintpath, NULL);
1513 rc = run_command(&config_unset);
1514 free(maintpath);
1515 return rc;
1518 static const char *get_frequency(enum schedule_priority schedule)
1520 switch (schedule) {
1521 case SCHEDULE_HOURLY:
1522 return "hourly";
1523 case SCHEDULE_DAILY:
1524 return "daily";
1525 case SCHEDULE_WEEKLY:
1526 return "weekly";
1527 default:
1528 BUG("invalid schedule %d", schedule);
1532 static char *launchctl_service_name(const char *frequency)
1534 struct strbuf label = STRBUF_INIT;
1535 strbuf_addf(&label, "org.git-scm.git.%s", frequency);
1536 return strbuf_detach(&label, NULL);
1539 static char *launchctl_service_filename(const char *name)
1541 char *expanded;
1542 struct strbuf filename = STRBUF_INIT;
1543 strbuf_addf(&filename, "~/Library/LaunchAgents/%s.plist", name);
1545 expanded = expand_user_path(filename.buf, 1);
1546 if (!expanded)
1547 die(_("failed to expand path '%s'"), filename.buf);
1549 strbuf_release(&filename);
1550 return expanded;
1553 static char *launchctl_get_uid(void)
1555 return xstrfmt("gui/%d", getuid());
1558 static int launchctl_boot_plist(int enable, const char *filename, const char *cmd)
1560 int result;
1561 struct child_process child = CHILD_PROCESS_INIT;
1562 char *uid = launchctl_get_uid();
1564 strvec_split(&child.args, cmd);
1565 if (enable)
1566 strvec_push(&child.args, "bootstrap");
1567 else
1568 strvec_push(&child.args, "bootout");
1569 strvec_push(&child.args, uid);
1570 strvec_push(&child.args, filename);
1572 child.no_stderr = 1;
1573 child.no_stdout = 1;
1575 if (start_command(&child))
1576 die(_("failed to start launchctl"));
1578 result = finish_command(&child);
1580 free(uid);
1581 return result;
1584 static int launchctl_remove_plist(enum schedule_priority schedule, const char *cmd)
1586 const char *frequency = get_frequency(schedule);
1587 char *name = launchctl_service_name(frequency);
1588 char *filename = launchctl_service_filename(name);
1589 int result = launchctl_boot_plist(0, filename, cmd);
1590 unlink(filename);
1591 free(filename);
1592 free(name);
1593 return result;
1596 static int launchctl_remove_plists(const char *cmd)
1598 return launchctl_remove_plist(SCHEDULE_HOURLY, cmd) ||
1599 launchctl_remove_plist(SCHEDULE_DAILY, cmd) ||
1600 launchctl_remove_plist(SCHEDULE_WEEKLY, cmd);
1603 static int launchctl_list_contains_plist(const char *name, const char *cmd)
1605 int result;
1606 struct child_process child = CHILD_PROCESS_INIT;
1607 char *uid = launchctl_get_uid();
1609 strvec_split(&child.args, cmd);
1610 strvec_pushl(&child.args, "list", name, NULL);
1612 child.no_stderr = 1;
1613 child.no_stdout = 1;
1615 if (start_command(&child))
1616 die(_("failed to start launchctl"));
1618 result = finish_command(&child);
1620 free(uid);
1622 /* Returns failure if 'name' doesn't exist. */
1623 return !result;
1626 static int launchctl_schedule_plist(const char *exec_path, enum schedule_priority schedule, const char *cmd)
1628 int i, fd;
1629 const char *preamble, *repeat;
1630 const char *frequency = get_frequency(schedule);
1631 char *name = launchctl_service_name(frequency);
1632 char *filename = launchctl_service_filename(name);
1633 struct lock_file lk = LOCK_INIT;
1634 static unsigned long lock_file_timeout_ms = ULONG_MAX;
1635 struct strbuf plist = STRBUF_INIT, plist2 = STRBUF_INIT;
1636 struct stat st;
1638 preamble = "<?xml version=\"1.0\"?>\n"
1639 "<!DOCTYPE plist PUBLIC \"-//Apple//DTD PLIST 1.0//EN\" \"http://www.apple.com/DTDs/PropertyList-1.0.dtd\">\n"
1640 "<plist version=\"1.0\">"
1641 "<dict>\n"
1642 "<key>Label</key><string>%s</string>\n"
1643 "<key>ProgramArguments</key>\n"
1644 "<array>\n"
1645 "<string>%s/git</string>\n"
1646 "<string>--exec-path=%s</string>\n"
1647 "<string>for-each-repo</string>\n"
1648 "<string>--config=maintenance.repo</string>\n"
1649 "<string>maintenance</string>\n"
1650 "<string>run</string>\n"
1651 "<string>--schedule=%s</string>\n"
1652 "</array>\n"
1653 "<key>StartCalendarInterval</key>\n"
1654 "<array>\n";
1655 strbuf_addf(&plist, preamble, name, exec_path, exec_path, frequency);
1657 switch (schedule) {
1658 case SCHEDULE_HOURLY:
1659 repeat = "<dict>\n"
1660 "<key>Hour</key><integer>%d</integer>\n"
1661 "<key>Minute</key><integer>0</integer>\n"
1662 "</dict>\n";
1663 for (i = 1; i <= 23; i++)
1664 strbuf_addf(&plist, repeat, i);
1665 break;
1667 case SCHEDULE_DAILY:
1668 repeat = "<dict>\n"
1669 "<key>Day</key><integer>%d</integer>\n"
1670 "<key>Hour</key><integer>0</integer>\n"
1671 "<key>Minute</key><integer>0</integer>\n"
1672 "</dict>\n";
1673 for (i = 1; i <= 6; i++)
1674 strbuf_addf(&plist, repeat, i);
1675 break;
1677 case SCHEDULE_WEEKLY:
1678 strbuf_addstr(&plist,
1679 "<dict>\n"
1680 "<key>Day</key><integer>0</integer>\n"
1681 "<key>Hour</key><integer>0</integer>\n"
1682 "<key>Minute</key><integer>0</integer>\n"
1683 "</dict>\n");
1684 break;
1686 default:
1687 /* unreachable */
1688 break;
1690 strbuf_addstr(&plist, "</array>\n</dict>\n</plist>\n");
1692 if (safe_create_leading_directories(filename))
1693 die(_("failed to create directories for '%s'"), filename);
1695 if ((long)lock_file_timeout_ms < 0 &&
1696 git_config_get_ulong("gc.launchctlplistlocktimeoutms",
1697 &lock_file_timeout_ms))
1698 lock_file_timeout_ms = 150;
1700 fd = hold_lock_file_for_update_timeout(&lk, filename, LOCK_DIE_ON_ERROR,
1701 lock_file_timeout_ms);
1704 * Does this file already exist? With the intended contents? Is it
1705 * registered already? Then it does not need to be re-registered.
1707 if (!stat(filename, &st) && st.st_size == plist.len &&
1708 strbuf_read_file(&plist2, filename, plist.len) == plist.len &&
1709 !strbuf_cmp(&plist, &plist2) &&
1710 launchctl_list_contains_plist(name, cmd))
1711 rollback_lock_file(&lk);
1712 else {
1713 if (write_in_full(fd, plist.buf, plist.len) < 0 ||
1714 commit_lock_file(&lk))
1715 die_errno(_("could not write '%s'"), filename);
1717 /* bootout might fail if not already running, so ignore */
1718 launchctl_boot_plist(0, filename, cmd);
1719 if (launchctl_boot_plist(1, filename, cmd))
1720 die(_("failed to bootstrap service %s"), filename);
1723 free(filename);
1724 free(name);
1725 strbuf_release(&plist);
1726 strbuf_release(&plist2);
1727 return 0;
1730 static int launchctl_add_plists(const char *cmd)
1732 const char *exec_path = git_exec_path();
1734 return launchctl_schedule_plist(exec_path, SCHEDULE_HOURLY, cmd) ||
1735 launchctl_schedule_plist(exec_path, SCHEDULE_DAILY, cmd) ||
1736 launchctl_schedule_plist(exec_path, SCHEDULE_WEEKLY, cmd);
1739 static int launchctl_update_schedule(int run_maintenance, int fd, const char *cmd)
1741 if (run_maintenance)
1742 return launchctl_add_plists(cmd);
1743 else
1744 return launchctl_remove_plists(cmd);
1747 static char *schtasks_task_name(const char *frequency)
1749 struct strbuf label = STRBUF_INIT;
1750 strbuf_addf(&label, "Git Maintenance (%s)", frequency);
1751 return strbuf_detach(&label, NULL);
1754 static int schtasks_remove_task(enum schedule_priority schedule, const char *cmd)
1756 int result;
1757 struct strvec args = STRVEC_INIT;
1758 const char *frequency = get_frequency(schedule);
1759 char *name = schtasks_task_name(frequency);
1761 strvec_split(&args, cmd);
1762 strvec_pushl(&args, "/delete", "/tn", name, "/f", NULL);
1764 result = run_command_v_opt(args.v, 0);
1766 strvec_clear(&args);
1767 free(name);
1768 return result;
1771 static int schtasks_remove_tasks(const char *cmd)
1773 return schtasks_remove_task(SCHEDULE_HOURLY, cmd) ||
1774 schtasks_remove_task(SCHEDULE_DAILY, cmd) ||
1775 schtasks_remove_task(SCHEDULE_WEEKLY, cmd);
1778 static int schtasks_schedule_task(const char *exec_path, enum schedule_priority schedule, const char *cmd)
1780 int result;
1781 struct child_process child = CHILD_PROCESS_INIT;
1782 const char *xml;
1783 struct tempfile *tfile;
1784 const char *frequency = get_frequency(schedule);
1785 char *name = schtasks_task_name(frequency);
1786 struct strbuf tfilename = STRBUF_INIT;
1788 strbuf_addf(&tfilename, "%s/schedule_%s_XXXXXX",
1789 get_git_common_dir(), frequency);
1790 tfile = xmks_tempfile(tfilename.buf);
1791 strbuf_release(&tfilename);
1793 if (!fdopen_tempfile(tfile, "w"))
1794 die(_("failed to create temp xml file"));
1796 xml = "<?xml version=\"1.0\" ?>\n"
1797 "<Task version=\"1.4\" xmlns=\"http://schemas.microsoft.com/windows/2004/02/mit/task\">\n"
1798 "<Triggers>\n"
1799 "<CalendarTrigger>\n";
1800 fputs(xml, tfile->fp);
1802 switch (schedule) {
1803 case SCHEDULE_HOURLY:
1804 fprintf(tfile->fp,
1805 "<StartBoundary>2020-01-01T01:00:00</StartBoundary>\n"
1806 "<Enabled>true</Enabled>\n"
1807 "<ScheduleByDay>\n"
1808 "<DaysInterval>1</DaysInterval>\n"
1809 "</ScheduleByDay>\n"
1810 "<Repetition>\n"
1811 "<Interval>PT1H</Interval>\n"
1812 "<Duration>PT23H</Duration>\n"
1813 "<StopAtDurationEnd>false</StopAtDurationEnd>\n"
1814 "</Repetition>\n");
1815 break;
1817 case SCHEDULE_DAILY:
1818 fprintf(tfile->fp,
1819 "<StartBoundary>2020-01-01T00:00:00</StartBoundary>\n"
1820 "<Enabled>true</Enabled>\n"
1821 "<ScheduleByWeek>\n"
1822 "<DaysOfWeek>\n"
1823 "<Monday />\n"
1824 "<Tuesday />\n"
1825 "<Wednesday />\n"
1826 "<Thursday />\n"
1827 "<Friday />\n"
1828 "<Saturday />\n"
1829 "</DaysOfWeek>\n"
1830 "<WeeksInterval>1</WeeksInterval>\n"
1831 "</ScheduleByWeek>\n");
1832 break;
1834 case SCHEDULE_WEEKLY:
1835 fprintf(tfile->fp,
1836 "<StartBoundary>2020-01-01T00:00:00</StartBoundary>\n"
1837 "<Enabled>true</Enabled>\n"
1838 "<ScheduleByWeek>\n"
1839 "<DaysOfWeek>\n"
1840 "<Sunday />\n"
1841 "</DaysOfWeek>\n"
1842 "<WeeksInterval>1</WeeksInterval>\n"
1843 "</ScheduleByWeek>\n");
1844 break;
1846 default:
1847 break;
1850 xml = "</CalendarTrigger>\n"
1851 "</Triggers>\n"
1852 "<Principals>\n"
1853 "<Principal id=\"Author\">\n"
1854 "<LogonType>InteractiveToken</LogonType>\n"
1855 "<RunLevel>LeastPrivilege</RunLevel>\n"
1856 "</Principal>\n"
1857 "</Principals>\n"
1858 "<Settings>\n"
1859 "<MultipleInstancesPolicy>IgnoreNew</MultipleInstancesPolicy>\n"
1860 "<Enabled>true</Enabled>\n"
1861 "<Hidden>true</Hidden>\n"
1862 "<UseUnifiedSchedulingEngine>true</UseUnifiedSchedulingEngine>\n"
1863 "<WakeToRun>false</WakeToRun>\n"
1864 "<ExecutionTimeLimit>PT72H</ExecutionTimeLimit>\n"
1865 "<Priority>7</Priority>\n"
1866 "</Settings>\n"
1867 "<Actions Context=\"Author\">\n"
1868 "<Exec>\n"
1869 "<Command>\"%s\\git.exe\"</Command>\n"
1870 "<Arguments>--exec-path=\"%s\" for-each-repo --config=maintenance.repo maintenance run --schedule=%s</Arguments>\n"
1871 "</Exec>\n"
1872 "</Actions>\n"
1873 "</Task>\n";
1874 fprintf(tfile->fp, xml, exec_path, exec_path, frequency);
1875 strvec_split(&child.args, cmd);
1876 strvec_pushl(&child.args, "/create", "/tn", name, "/f", "/xml",
1877 get_tempfile_path(tfile), NULL);
1878 close_tempfile_gently(tfile);
1880 child.no_stdout = 1;
1881 child.no_stderr = 1;
1883 if (start_command(&child))
1884 die(_("failed to start schtasks"));
1885 result = finish_command(&child);
1887 delete_tempfile(&tfile);
1888 free(name);
1889 return result;
1892 static int schtasks_schedule_tasks(const char *cmd)
1894 const char *exec_path = git_exec_path();
1896 return schtasks_schedule_task(exec_path, SCHEDULE_HOURLY, cmd) ||
1897 schtasks_schedule_task(exec_path, SCHEDULE_DAILY, cmd) ||
1898 schtasks_schedule_task(exec_path, SCHEDULE_WEEKLY, cmd);
1901 static int schtasks_update_schedule(int run_maintenance, int fd, const char *cmd)
1903 if (run_maintenance)
1904 return schtasks_schedule_tasks(cmd);
1905 else
1906 return schtasks_remove_tasks(cmd);
1909 #define BEGIN_LINE "# BEGIN GIT MAINTENANCE SCHEDULE"
1910 #define END_LINE "# END GIT MAINTENANCE SCHEDULE"
1912 static int crontab_update_schedule(int run_maintenance, int fd, const char *cmd)
1914 int result = 0;
1915 int in_old_region = 0;
1916 struct child_process crontab_list = CHILD_PROCESS_INIT;
1917 struct child_process crontab_edit = CHILD_PROCESS_INIT;
1918 FILE *cron_list, *cron_in;
1919 struct strbuf line = STRBUF_INIT;
1921 strvec_split(&crontab_list.args, cmd);
1922 strvec_push(&crontab_list.args, "-l");
1923 crontab_list.in = -1;
1924 crontab_list.out = dup(fd);
1925 crontab_list.git_cmd = 0;
1927 if (start_command(&crontab_list))
1928 return error(_("failed to run 'crontab -l'; your system might not support 'cron'"));
1930 /* Ignore exit code, as an empty crontab will return error. */
1931 finish_command(&crontab_list);
1934 * Read from the .lock file, filtering out the old
1935 * schedule while appending the new schedule.
1937 cron_list = fdopen(fd, "r");
1938 rewind(cron_list);
1940 strvec_split(&crontab_edit.args, cmd);
1941 crontab_edit.in = -1;
1942 crontab_edit.git_cmd = 0;
1944 if (start_command(&crontab_edit))
1945 return error(_("failed to run 'crontab'; your system might not support 'cron'"));
1947 cron_in = fdopen(crontab_edit.in, "w");
1948 if (!cron_in) {
1949 result = error(_("failed to open stdin of 'crontab'"));
1950 goto done_editing;
1953 while (!strbuf_getline_lf(&line, cron_list)) {
1954 if (!in_old_region && !strcmp(line.buf, BEGIN_LINE))
1955 in_old_region = 1;
1956 else if (in_old_region && !strcmp(line.buf, END_LINE))
1957 in_old_region = 0;
1958 else if (!in_old_region)
1959 fprintf(cron_in, "%s\n", line.buf);
1961 strbuf_release(&line);
1963 if (run_maintenance) {
1964 struct strbuf line_format = STRBUF_INIT;
1965 const char *exec_path = git_exec_path();
1967 fprintf(cron_in, "%s\n", BEGIN_LINE);
1968 fprintf(cron_in,
1969 "# The following schedule was created by Git\n");
1970 fprintf(cron_in, "# Any edits made in this region might be\n");
1971 fprintf(cron_in,
1972 "# replaced in the future by a Git command.\n\n");
1974 strbuf_addf(&line_format,
1975 "%%s %%s * * %%s \"%s/git\" --exec-path=\"%s\" for-each-repo --config=maintenance.repo maintenance run --schedule=%%s\n",
1976 exec_path, exec_path);
1977 fprintf(cron_in, line_format.buf, "0", "1-23", "*", "hourly");
1978 fprintf(cron_in, line_format.buf, "0", "0", "1-6", "daily");
1979 fprintf(cron_in, line_format.buf, "0", "0", "0", "weekly");
1980 strbuf_release(&line_format);
1982 fprintf(cron_in, "\n%s\n", END_LINE);
1985 fflush(cron_in);
1986 fclose(cron_in);
1987 close(crontab_edit.in);
1989 done_editing:
1990 if (finish_command(&crontab_edit))
1991 result = error(_("'crontab' died"));
1992 else
1993 fclose(cron_list);
1994 return result;
1997 #if defined(__APPLE__)
1998 static const char platform_scheduler[] = "launchctl";
1999 #elif defined(GIT_WINDOWS_NATIVE)
2000 static const char platform_scheduler[] = "schtasks";
2001 #else
2002 static const char platform_scheduler[] = "crontab";
2003 #endif
2005 static int update_background_schedule(int enable)
2007 int result;
2008 const char *scheduler = platform_scheduler;
2009 const char *cmd = scheduler;
2010 char *testing;
2011 struct lock_file lk;
2012 char *lock_path = xstrfmt("%s/schedule", the_repository->objects->odb->path);
2014 testing = xstrdup_or_null(getenv("GIT_TEST_MAINT_SCHEDULER"));
2015 if (testing) {
2016 char *sep = strchr(testing, ':');
2017 if (!sep)
2018 die("GIT_TEST_MAINT_SCHEDULER unparseable: %s", testing);
2019 *sep = '\0';
2020 scheduler = testing;
2021 cmd = sep + 1;
2024 if (hold_lock_file_for_update(&lk, lock_path, LOCK_NO_DEREF) < 0) {
2025 result = error(_("another process is scheduling background maintenance"));
2026 goto cleanup;
2029 if (!strcmp(scheduler, "launchctl"))
2030 result = launchctl_update_schedule(enable, get_lock_file_fd(&lk), cmd);
2031 else if (!strcmp(scheduler, "schtasks"))
2032 result = schtasks_update_schedule(enable, get_lock_file_fd(&lk), cmd);
2033 else if (!strcmp(scheduler, "crontab"))
2034 result = crontab_update_schedule(enable, get_lock_file_fd(&lk), cmd);
2035 else
2036 die("unknown background scheduler: %s", scheduler);
2038 rollback_lock_file(&lk);
2040 cleanup:
2041 free(lock_path);
2042 free(testing);
2043 return result;
2046 static int maintenance_start(void)
2048 if (maintenance_register())
2049 warning(_("failed to add repo to global config"));
2051 return update_background_schedule(1);
2054 static int maintenance_stop(void)
2056 return update_background_schedule(0);
2059 static const char builtin_maintenance_usage[] = N_("git maintenance <subcommand> [<options>]");
2061 int cmd_maintenance(int argc, const char **argv, const char *prefix)
2063 if (argc < 2 ||
2064 (argc == 2 && !strcmp(argv[1], "-h")))
2065 usage(builtin_maintenance_usage);
2067 if (!strcmp(argv[1], "run"))
2068 return maintenance_run(argc - 1, argv + 1, prefix);
2069 if (!strcmp(argv[1], "start"))
2070 return maintenance_start();
2071 if (!strcmp(argv[1], "stop"))
2072 return maintenance_stop();
2073 if (!strcmp(argv[1], "register"))
2074 return maintenance_register();
2075 if (!strcmp(argv[1], "unregister"))
2076 return maintenance_unregister();
2078 die(_("invalid subcommand: %s"), argv[1]);