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
15 #include "parse-options.h"
16 #include "run-command.h"
18 #include "argv-array.h"
21 #define FAILED_RUN "failed to run %s"
23 static const char * const builtin_gc_usage
[] = {
24 N_("git gc [options]"),
28 static int pack_refs
= 1;
29 static int prune_reflogs
= 1;
30 static int aggressive_depth
= 250;
31 static int aggressive_window
= 250;
32 static int gc_auto_threshold
= 6700;
33 static int gc_auto_pack_limit
= 50;
34 static int detach_auto
= 1;
35 static const char *prune_expire
= "2.weeks.ago";
36 static const char *prune_repos_expire
= "3.months.ago";
38 static struct argv_array pack_refs_cmd
= ARGV_ARRAY_INIT
;
39 static struct argv_array reflog
= ARGV_ARRAY_INIT
;
40 static struct argv_array repack
= ARGV_ARRAY_INIT
;
41 static struct argv_array prune
= ARGV_ARRAY_INIT
;
42 static struct argv_array prune_repos
= ARGV_ARRAY_INIT
;
43 static struct argv_array rerere
= ARGV_ARRAY_INIT
;
47 static void remove_pidfile(void)
53 static void remove_pidfile_on_signal(int signo
)
60 static void git_config_get_date_string_const(const char *var
, const char **val
)
62 if (!git_config_get_string_const(var
, val
) && *val
&& strcmp(*val
, "now")) {
63 unsigned long now
= approxidate("now");
64 if (approxidate(*val
) >= now
)
65 git_die_config(var
, _("Invalid %s: '%s'"), var
, *val
);
69 static void gc_config(void)
73 if (!git_config_get_value("gc.packrefs", &value
)) {
74 if (value
&& !strcmp(value
, "notbare"))
77 pack_refs
= git_config_bool("gc.packrefs", value
);
80 git_config_get_int("gc.aggressivewindow", &aggressive_window
);
81 git_config_get_int("gc.aggressivedepth", &aggressive_depth
);
82 git_config_get_int("gc.auto", &gc_auto_threshold
);
83 git_config_get_int("gc.autopacklimit", &gc_auto_pack_limit
);
84 git_config_get_bool("gc.autodetach", &detach_auto
);
85 git_config_get_date_string_const("gc.pruneexpire", &prune_expire
);
86 git_config_get_date_string_const("gc.prunereposexpire", &prune_repos_expire
);
87 git_config(git_default_config
, NULL
);
90 static int too_many_loose_objects(void)
93 * Quickly check if a "gc" is needed, by estimating how
94 * many loose objects there are. Because SHA-1 is evenly
95 * distributed, we can check only one and get a reasonable
99 const char *objdir
= get_object_directory();
106 if (gc_auto_threshold
<= 0)
109 if (sizeof(path
) <= snprintf(path
, sizeof(path
), "%s/17", objdir
)) {
110 warning(_("insanely long object directory %.*s"), 50, objdir
);
117 auto_threshold
= (gc_auto_threshold
+ 255) / 256;
118 while ((ent
= readdir(dir
)) != NULL
) {
119 if (strspn(ent
->d_name
, "0123456789abcdef") != 38 ||
120 ent
->d_name
[38] != '\0')
122 if (++num_loose
> auto_threshold
) {
131 static int too_many_packs(void)
133 struct packed_git
*p
;
136 if (gc_auto_pack_limit
<= 0)
139 prepare_packed_git();
140 for (cnt
= 0, p
= packed_git
; p
; p
= p
->next
) {
146 * Perhaps check the size of the pack and count only
147 * very small ones here?
151 return gc_auto_pack_limit
<= cnt
;
154 static void add_repack_all_option(void)
156 if (prune_expire
&& !strcmp(prune_expire
, "now"))
157 argv_array_push(&repack
, "-a");
159 argv_array_push(&repack
, "-A");
161 argv_array_pushf(&repack
, "--unpack-unreachable=%s", prune_expire
);
165 static int need_to_gc(void)
168 * Setting gc.auto to 0 or negative can disable the
171 if (gc_auto_threshold
<= 0)
175 * If there are too many loose objects, but not too many
176 * packs, we run "repack -d -l". If there are too many packs,
177 * we run "repack -A -d -l". Otherwise we tell the caller
180 if (too_many_packs())
181 add_repack_all_option();
182 else if (!too_many_loose_objects())
185 if (run_hook_le(NULL
, "pre-auto-gc", NULL
))
190 /* return NULL on success, else hostname running the gc */
191 static const char *lock_repo_for_gc(int force
, pid_t
* ret_pid
)
193 static struct lock_file lock
;
195 struct strbuf sb
= STRBUF_INIT
;
205 if (gethostname(my_host
, sizeof(my_host
)))
206 strcpy(my_host
, "unknown");
208 fd
= hold_lock_file_for_update(&lock
, git_path("gc.pid"),
211 static char locking_host
[128];
213 fp
= fopen(git_path("gc.pid"), "r");
214 memset(locking_host
, 0, sizeof(locking_host
));
217 !fstat(fileno(fp
), &st
) &&
219 * 12 hour limit is very generous as gc should
220 * never take that long. On the other hand we
221 * don't really need a strict limit here,
222 * running gc --auto one day late is not a big
223 * problem. --force can be used in manual gc
224 * after the user verifies that no gc is
227 time(NULL
) - st
.st_mtime
<= 12 * 3600 &&
228 fscanf(fp
, "%"PRIuMAX
" %127c", &pid
, locking_host
) == 2 &&
229 /* be gentle to concurrent "gc" on remote hosts */
230 (strcmp(locking_host
, my_host
) || !kill(pid
, 0) || errno
== EPERM
);
235 rollback_lock_file(&lock
);
241 strbuf_addf(&sb
, "%"PRIuMAX
" %s",
242 (uintmax_t) getpid(), my_host
);
243 write_in_full(fd
, sb
.buf
, sb
.len
);
245 commit_lock_file(&lock
);
247 pidfile
= git_pathdup("gc.pid");
248 sigchain_push_common(remove_pidfile_on_signal
);
249 atexit(remove_pidfile
);
254 static int gc_before_repack(void)
256 if (pack_refs
&& run_command_v_opt(pack_refs_cmd
.argv
, RUN_GIT_CMD
))
257 return error(FAILED_RUN
, pack_refs_cmd
.argv
[0]);
259 if (prune_reflogs
&& run_command_v_opt(reflog
.argv
, RUN_GIT_CMD
))
260 return error(FAILED_RUN
, reflog
.argv
[0]);
267 int cmd_gc(int argc
, const char **argv
, const char *prefix
)
276 struct option builtin_gc_options
[] = {
277 OPT__QUIET(&quiet
, N_("suppress progress reporting")),
278 { OPTION_STRING
, 0, "prune", &prune_expire
, N_("date"),
279 N_("prune unreferenced objects"),
280 PARSE_OPT_OPTARG
, NULL
, (intptr_t)prune_expire
},
281 OPT_BOOL(0, "aggressive", &aggressive
, N_("be more thorough (increased runtime)")),
282 OPT_BOOL(0, "auto", &auto_gc
, N_("enable auto-gc mode")),
283 OPT_BOOL(0, "force", &force
, N_("force running gc even if there may be another gc running")),
287 if (argc
== 2 && !strcmp(argv
[1], "-h"))
288 usage_with_options(builtin_gc_usage
, builtin_gc_options
);
290 argv_array_pushl(&pack_refs_cmd
, "pack-refs", "--all", "--prune", NULL
);
291 argv_array_pushl(&reflog
, "reflog", "expire", "--all", NULL
);
292 argv_array_pushl(&repack
, "repack", "-d", "-l", NULL
);
293 argv_array_pushl(&prune
, "prune", "--expire", NULL
);
294 argv_array_pushl(&prune_repos
, "prune", "--repos", "--expire", NULL
);
295 argv_array_pushl(&rerere
, "rerere", "gc", NULL
);
300 pack_refs
= !is_bare_repository();
302 argc
= parse_options(argc
, argv
, prefix
, builtin_gc_options
,
303 builtin_gc_usage
, 0);
305 usage_with_options(builtin_gc_usage
, builtin_gc_options
);
308 argv_array_push(&repack
, "-f");
309 if (aggressive_depth
> 0)
310 argv_array_pushf(&repack
, "--depth=%d", aggressive_depth
);
311 if (aggressive_window
> 0)
312 argv_array_pushf(&repack
, "--window=%d", aggressive_window
);
315 argv_array_push(&repack
, "-q");
319 * Auto-gc should be least intrusive as possible.
325 fprintf(stderr
, _("Auto packing the repository in background for optimum performance.\n"));
327 fprintf(stderr
, _("Auto packing the repository for optimum performance.\n"));
328 fprintf(stderr
, _("See \"git help gc\" for manual housekeeping.\n"));
331 if (gc_before_repack())
334 * failure to daemonize is ok, we'll continue
340 add_repack_all_option();
342 name
= lock_repo_for_gc(force
, &pid
);
345 return 0; /* be quiet on --auto */
346 die(_("gc is already running on machine '%s' pid %"PRIuMAX
" (use --force if not)"),
347 name
, (uintmax_t)pid
);
350 if (gc_before_repack())
353 if (run_command_v_opt(repack
.argv
, RUN_GIT_CMD
))
354 return error(FAILED_RUN
, repack
.argv
[0]);
357 argv_array_push(&prune
, prune_expire
);
359 argv_array_push(&prune
, "--no-progress");
360 if (run_command_v_opt(prune
.argv
, RUN_GIT_CMD
))
361 return error(FAILED_RUN
, prune
.argv
[0]);
364 if (prune_repos_expire
) {
365 argv_array_push(&prune_repos
, prune_repos_expire
);
366 if (run_command_v_opt(prune_repos
.argv
, RUN_GIT_CMD
))
367 return error(FAILED_RUN
, prune_repos
.argv
[0]);
370 if (run_command_v_opt(rerere
.argv
, RUN_GIT_CMD
))
371 return error(FAILED_RUN
, rerere
.argv
[0]);
373 if (auto_gc
&& too_many_loose_objects())
374 warning(_("There are too many unreachable loose objects; "
375 "run 'git prune' to remove them."));