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
16 #include "parse-options.h"
17 #include "run-command.h"
19 #include "argv-array.h"
22 #define FAILED_RUN "failed to run %s"
24 static const char * const builtin_gc_usage
[] = {
25 N_("git gc [<options>]"),
29 static int pack_refs
= 1;
30 static int prune_reflogs
= 1;
31 static int aggressive_depth
= 50;
32 static int aggressive_window
= 250;
33 static int gc_auto_threshold
= 6700;
34 static int gc_auto_pack_limit
= 50;
35 static int detach_auto
= 1;
36 static unsigned long gc_log_expire_time
;
37 static const char *gc_log_expire
= "1.day.ago";
38 static const char *prune_expire
= "2.weeks.ago";
39 static const char *prune_worktrees_expire
= "3.months.ago";
41 static struct argv_array pack_refs_cmd
= ARGV_ARRAY_INIT
;
42 static struct argv_array reflog
= ARGV_ARRAY_INIT
;
43 static struct argv_array repack
= ARGV_ARRAY_INIT
;
44 static struct argv_array prune
= ARGV_ARRAY_INIT
;
45 static struct argv_array prune_worktrees
= ARGV_ARRAY_INIT
;
46 static struct argv_array rerere
= ARGV_ARRAY_INIT
;
48 static struct tempfile pidfile
;
49 static struct lock_file log_lock
;
51 static struct string_list pack_garbage
= STRING_LIST_INIT_DUP
;
53 static void clean_pack_garbage(void)
56 for (i
= 0; i
< pack_garbage
.nr
; i
++)
57 unlink_or_warn(pack_garbage
.items
[i
].string
);
58 string_list_clear(&pack_garbage
, 0);
61 static void report_pack_garbage(unsigned seen_bits
, const char *path
)
63 if (seen_bits
== PACKDIR_FILE_IDX
)
64 string_list_append(&pack_garbage
, path
);
67 static void process_log_file(void)
70 if (fstat(get_lock_file_fd(&log_lock
), &st
)) {
72 * Perhaps there was an i/o error or another
73 * unlikely situation. Try to make a note of
74 * this in gc.log along with any existing
77 int saved_errno
= errno
;
78 fprintf(stderr
, _("Failed to fstat %s: %s"),
79 get_tempfile_path(&log_lock
.tempfile
),
80 strerror(saved_errno
));
82 commit_lock_file(&log_lock
);
84 } else if (st
.st_size
) {
85 /* There was some error recorded in the lock file */
86 commit_lock_file(&log_lock
);
88 /* No error, clean up any old gc.log */
89 unlink(git_path("gc.log"));
90 rollback_lock_file(&log_lock
);
94 static void process_log_file_at_exit(void)
100 static void process_log_file_on_signal(int signo
)
107 static void gc_config(void)
111 if (!git_config_get_value("gc.packrefs", &value
)) {
112 if (value
&& !strcmp(value
, "notbare"))
115 pack_refs
= git_config_bool("gc.packrefs", value
);
118 git_config_get_int("gc.aggressivewindow", &aggressive_window
);
119 git_config_get_int("gc.aggressivedepth", &aggressive_depth
);
120 git_config_get_int("gc.auto", &gc_auto_threshold
);
121 git_config_get_int("gc.autopacklimit", &gc_auto_pack_limit
);
122 git_config_get_bool("gc.autodetach", &detach_auto
);
123 git_config_get_expiry("gc.pruneexpire", &prune_expire
);
124 git_config_get_expiry("gc.worktreepruneexpire", &prune_worktrees_expire
);
125 git_config_get_expiry("gc.logexpiry", &gc_log_expire
);
127 git_config(git_default_config
, NULL
);
130 static int too_many_loose_objects(void)
133 * Quickly check if a "gc" is needed, by estimating how
134 * many loose objects there are. Because SHA-1 is evenly
135 * distributed, we can check only one and get a reasonable
144 if (gc_auto_threshold
<= 0)
147 dir
= opendir(git_path("objects/17"));
151 auto_threshold
= (gc_auto_threshold
+ 255) / 256;
152 while ((ent
= readdir(dir
)) != NULL
) {
153 if (strspn(ent
->d_name
, "0123456789abcdef") != 38 ||
154 ent
->d_name
[38] != '\0')
156 if (++num_loose
> auto_threshold
) {
165 static int too_many_packs(void)
167 struct packed_git
*p
;
170 if (gc_auto_pack_limit
<= 0)
173 prepare_packed_git();
174 for (cnt
= 0, p
= packed_git
; p
; p
= p
->next
) {
180 * Perhaps check the size of the pack and count only
181 * very small ones here?
185 return gc_auto_pack_limit
< cnt
;
188 static void add_repack_all_option(void)
190 if (prune_expire
&& !strcmp(prune_expire
, "now"))
191 argv_array_push(&repack
, "-a");
193 argv_array_push(&repack
, "-A");
195 argv_array_pushf(&repack
, "--unpack-unreachable=%s", prune_expire
);
199 static void add_repack_incremental_option(void)
201 argv_array_push(&repack
, "--no-write-bitmap-index");
204 static int need_to_gc(void)
207 * Setting gc.auto to 0 or negative can disable the
210 if (gc_auto_threshold
<= 0)
214 * If there are too many loose objects, but not too many
215 * packs, we run "repack -d -l". If there are too many packs,
216 * we run "repack -A -d -l". Otherwise we tell the caller
219 if (too_many_packs())
220 add_repack_all_option();
221 else if (too_many_loose_objects())
222 add_repack_incremental_option();
226 if (run_hook_le(NULL
, "pre-auto-gc", NULL
))
231 /* return NULL on success, else hostname running the gc */
232 static const char *lock_repo_for_gc(int force
, pid_t
* ret_pid
)
234 static struct lock_file lock
;
236 struct strbuf sb
= STRBUF_INIT
;
243 if (is_tempfile_active(&pidfile
))
247 if (gethostname(my_host
, sizeof(my_host
)))
248 xsnprintf(my_host
, sizeof(my_host
), "unknown");
250 pidfile_path
= git_pathdup("gc.pid");
251 fd
= hold_lock_file_for_update(&lock
, pidfile_path
,
254 static char locking_host
[128];
256 fp
= fopen(pidfile_path
, "r");
257 memset(locking_host
, 0, sizeof(locking_host
));
260 !fstat(fileno(fp
), &st
) &&
262 * 12 hour limit is very generous as gc should
263 * never take that long. On the other hand we
264 * don't really need a strict limit here,
265 * running gc --auto one day late is not a big
266 * problem. --force can be used in manual gc
267 * after the user verifies that no gc is
270 time(NULL
) - st
.st_mtime
<= 12 * 3600 &&
271 fscanf(fp
, "%"SCNuMAX
" %127c", &pid
, locking_host
) == 2 &&
272 /* be gentle to concurrent "gc" on remote hosts */
273 (strcmp(locking_host
, my_host
) || !kill(pid
, 0) || errno
== EPERM
);
278 rollback_lock_file(&lock
);
285 strbuf_addf(&sb
, "%"PRIuMAX
" %s",
286 (uintmax_t) getpid(), my_host
);
287 write_in_full(fd
, sb
.buf
, sb
.len
);
289 commit_lock_file(&lock
);
290 register_tempfile(&pidfile
, pidfile_path
);
295 static int report_last_gc_error(void)
297 struct strbuf sb
= STRBUF_INIT
;
300 char *gc_log_path
= git_pathdup("gc.log");
302 if (stat(gc_log_path
, &st
)) {
306 ret
= error_errno(_("Can't stat %s"), gc_log_path
);
310 if (st
.st_mtime
< gc_log_expire_time
)
313 ret
= strbuf_read_file(&sb
, gc_log_path
, 0);
315 ret
= error(_("The last gc run reported the following. "
316 "Please correct the root cause\n"
318 "Automatic cleanup will not be performed "
319 "until the file is removed.\n\n"
321 gc_log_path
, sb
.buf
);
328 static int gc_before_repack(void)
330 if (pack_refs
&& run_command_v_opt(pack_refs_cmd
.argv
, RUN_GIT_CMD
))
331 return error(FAILED_RUN
, pack_refs_cmd
.argv
[0]);
333 if (prune_reflogs
&& run_command_v_opt(reflog
.argv
, RUN_GIT_CMD
))
334 return error(FAILED_RUN
, reflog
.argv
[0]);
341 int cmd_gc(int argc
, const char **argv
, const char *prefix
)
351 struct option builtin_gc_options
[] = {
352 OPT__QUIET(&quiet
, N_("suppress progress reporting")),
353 { OPTION_STRING
, 0, "prune", &prune_expire
, N_("date"),
354 N_("prune unreferenced objects"),
355 PARSE_OPT_OPTARG
, NULL
, (intptr_t)prune_expire
},
356 OPT_BOOL(0, "aggressive", &aggressive
, N_("be more thorough (increased runtime)")),
357 OPT_BOOL(0, "auto", &auto_gc
, N_("enable auto-gc mode")),
358 OPT_BOOL(0, "force", &force
, N_("force running gc even if there may be another gc running")),
362 if (argc
== 2 && !strcmp(argv
[1], "-h"))
363 usage_with_options(builtin_gc_usage
, builtin_gc_options
);
365 argv_array_pushl(&pack_refs_cmd
, "pack-refs", "--all", "--prune", NULL
);
366 argv_array_pushl(&reflog
, "reflog", "expire", "--all", NULL
);
367 argv_array_pushl(&repack
, "repack", "-d", "-l", NULL
);
368 argv_array_pushl(&prune
, "prune", "--expire", NULL
);
369 argv_array_pushl(&prune_worktrees
, "worktree", "prune", "--expire", NULL
);
370 argv_array_pushl(&rerere
, "rerere", "gc", NULL
);
372 /* default expiry time, overwritten in gc_config */
374 if (parse_expiry_date(gc_log_expire
, &gc_log_expire_time
))
375 die(_("Failed to parse gc.logexpiry value %s"), gc_log_expire
);
378 pack_refs
= !is_bare_repository();
380 argc
= parse_options(argc
, argv
, prefix
, builtin_gc_options
,
381 builtin_gc_usage
, 0);
383 usage_with_options(builtin_gc_usage
, builtin_gc_options
);
386 argv_array_push(&repack
, "-f");
387 if (aggressive_depth
> 0)
388 argv_array_pushf(&repack
, "--depth=%d", aggressive_depth
);
389 if (aggressive_window
> 0)
390 argv_array_pushf(&repack
, "--window=%d", aggressive_window
);
393 argv_array_push(&repack
, "-q");
397 * Auto-gc should be least intrusive as possible.
403 fprintf(stderr
, _("Auto packing the repository in background for optimum performance.\n"));
405 fprintf(stderr
, _("Auto packing the repository for optimum performance.\n"));
406 fprintf(stderr
, _("See \"git help gc\" for manual housekeeping.\n"));
409 if (report_last_gc_error())
412 if (gc_before_repack())
415 * failure to daemonize is ok, we'll continue
418 daemonized
= !daemonize();
421 add_repack_all_option();
423 name
= lock_repo_for_gc(force
, &pid
);
426 return 0; /* be quiet on --auto */
427 die(_("gc is already running on machine '%s' pid %"PRIuMAX
" (use --force if not)"),
428 name
, (uintmax_t)pid
);
432 hold_lock_file_for_update(&log_lock
,
435 dup2(get_lock_file_fd(&log_lock
), 2);
436 sigchain_push_common(process_log_file_on_signal
);
437 atexit(process_log_file_at_exit
);
440 if (gc_before_repack())
443 if (!repository_format_precious_objects
) {
444 if (run_command_v_opt(repack
.argv
, RUN_GIT_CMD
))
445 return error(FAILED_RUN
, repack
.argv
[0]);
448 argv_array_push(&prune
, prune_expire
);
450 argv_array_push(&prune
, "--no-progress");
451 if (run_command_v_opt(prune
.argv
, RUN_GIT_CMD
))
452 return error(FAILED_RUN
, prune
.argv
[0]);
456 if (prune_worktrees_expire
) {
457 argv_array_push(&prune_worktrees
, prune_worktrees_expire
);
458 if (run_command_v_opt(prune_worktrees
.argv
, RUN_GIT_CMD
))
459 return error(FAILED_RUN
, prune_worktrees
.argv
[0]);
462 if (run_command_v_opt(rerere
.argv
, RUN_GIT_CMD
))
463 return error(FAILED_RUN
, rerere
.argv
[0]);
465 report_garbage
= report_pack_garbage
;
466 reprepare_packed_git();
467 if (pack_garbage
.nr
> 0)
468 clean_pack_garbage();
470 if (auto_gc
&& too_many_loose_objects())
471 warning(_("There are too many unreachable loose objects; "
472 "run 'git prune' to remove them."));
475 unlink(git_path("gc.log"));