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_window
= 250;
31 static int gc_auto_threshold
= 6700;
32 static int gc_auto_pack_limit
= 50;
33 static int detach_auto
= 1;
34 static const char *prune_expire
= "2.weeks.ago";
36 static struct argv_array pack_refs_cmd
= ARGV_ARRAY_INIT
;
37 static struct argv_array reflog
= ARGV_ARRAY_INIT
;
38 static struct argv_array repack
= ARGV_ARRAY_INIT
;
39 static struct argv_array prune
= ARGV_ARRAY_INIT
;
40 static struct argv_array rerere
= ARGV_ARRAY_INIT
;
44 static void remove_pidfile(void)
50 static void remove_pidfile_on_signal(int signo
)
57 static int gc_config(const char *var
, const char *value
, void *cb
)
59 if (!strcmp(var
, "gc.packrefs")) {
60 if (value
&& !strcmp(value
, "notbare"))
63 pack_refs
= git_config_bool(var
, value
);
66 if (!strcmp(var
, "gc.aggressivewindow")) {
67 aggressive_window
= git_config_int(var
, value
);
70 if (!strcmp(var
, "gc.auto")) {
71 gc_auto_threshold
= git_config_int(var
, value
);
74 if (!strcmp(var
, "gc.autopacklimit")) {
75 gc_auto_pack_limit
= git_config_int(var
, value
);
78 if (!strcmp(var
, "gc.autodetach")) {
79 detach_auto
= git_config_bool(var
, value
);
82 if (!strcmp(var
, "gc.pruneexpire")) {
83 if (value
&& strcmp(value
, "now")) {
84 unsigned long now
= approxidate("now");
85 if (approxidate(value
) >= now
)
86 return error(_("Invalid %s: '%s'"), var
, value
);
88 return git_config_string(&prune_expire
, var
, value
);
90 return git_default_config(var
, value
, cb
);
93 static int too_many_loose_objects(void)
96 * Quickly check if a "gc" is needed, by estimating how
97 * many loose objects there are. Because SHA-1 is evenly
98 * distributed, we can check only one and get a reasonable
102 const char *objdir
= get_object_directory();
109 if (gc_auto_threshold
<= 0)
112 if (sizeof(path
) <= snprintf(path
, sizeof(path
), "%s/17", objdir
)) {
113 warning(_("insanely long object directory %.*s"), 50, objdir
);
120 auto_threshold
= (gc_auto_threshold
+ 255) / 256;
121 while ((ent
= readdir(dir
)) != NULL
) {
122 if (strspn(ent
->d_name
, "0123456789abcdef") != 38 ||
123 ent
->d_name
[38] != '\0')
125 if (++num_loose
> auto_threshold
) {
134 static int too_many_packs(void)
136 struct packed_git
*p
;
139 if (gc_auto_pack_limit
<= 0)
142 prepare_packed_git();
143 for (cnt
= 0, p
= packed_git
; p
; p
= p
->next
) {
149 * Perhaps check the size of the pack and count only
150 * very small ones here?
154 return gc_auto_pack_limit
<= cnt
;
157 static void add_repack_all_option(void)
159 if (prune_expire
&& !strcmp(prune_expire
, "now"))
160 argv_array_push(&repack
, "-a");
162 argv_array_push(&repack
, "-A");
164 argv_array_pushf(&repack
, "--unpack-unreachable=%s", prune_expire
);
168 static int need_to_gc(void)
171 * Setting gc.auto to 0 or negative can disable the
174 if (gc_auto_threshold
<= 0)
178 * If there are too many loose objects, but not too many
179 * packs, we run "repack -d -l". If there are too many packs,
180 * we run "repack -A -d -l". Otherwise we tell the caller
183 if (too_many_packs())
184 add_repack_all_option();
185 else if (!too_many_loose_objects())
188 if (run_hook(NULL
, "pre-auto-gc", NULL
))
193 /* return NULL on success, else hostname running the gc */
194 static const char *lock_repo_for_gc(int force
, pid_t
* ret_pid
)
196 static struct lock_file lock
;
197 static char locking_host
[128];
199 struct strbuf sb
= STRBUF_INIT
;
209 if (gethostname(my_host
, sizeof(my_host
)))
210 strcpy(my_host
, "unknown");
212 fd
= hold_lock_file_for_update(&lock
, git_path("gc.pid"),
215 fp
= fopen(git_path("gc.pid"), "r");
216 memset(locking_host
, 0, sizeof(locking_host
));
219 !fstat(fileno(fp
), &st
) &&
221 * 12 hour limit is very generous as gc should
222 * never take that long. On the other hand we
223 * don't really need a strict limit here,
224 * running gc --auto one day late is not a big
225 * problem. --force can be used in manual gc
226 * after the user verifies that no gc is
229 time(NULL
) - st
.st_mtime
<= 12 * 3600 &&
230 fscanf(fp
, "%"PRIuMAX
" %127c", &pid
, locking_host
) == 2 &&
231 /* be gentle to concurrent "gc" on remote hosts */
232 (strcmp(locking_host
, my_host
) || !kill(pid
, 0) || errno
== EPERM
);
237 rollback_lock_file(&lock
);
243 strbuf_addf(&sb
, "%"PRIuMAX
" %s",
244 (uintmax_t) getpid(), my_host
);
245 write_in_full(fd
, sb
.buf
, sb
.len
);
247 commit_lock_file(&lock
);
249 pidfile
= git_pathdup("gc.pid");
250 sigchain_push_common(remove_pidfile_on_signal
);
251 atexit(remove_pidfile
);
256 static int gc_before_repack(void)
258 if (pack_refs
&& run_command_v_opt(pack_refs_cmd
.argv
, RUN_GIT_CMD
))
259 return error(FAILED_RUN
, pack_refs_cmd
.argv
[0]);
261 if (prune_reflogs
&& run_command_v_opt(reflog
.argv
, RUN_GIT_CMD
))
262 return error(FAILED_RUN
, reflog
.argv
[0]);
269 int cmd_gc(int argc
, const char **argv
, const char *prefix
)
278 struct option builtin_gc_options
[] = {
279 OPT__QUIET(&quiet
, N_("suppress progress reporting")),
280 { OPTION_STRING
, 0, "prune", &prune_expire
, N_("date"),
281 N_("prune unreferenced objects"),
282 PARSE_OPT_OPTARG
, NULL
, (intptr_t)prune_expire
},
283 OPT_BOOL(0, "aggressive", &aggressive
, N_("be more thorough (increased runtime)")),
284 OPT_BOOL(0, "auto", &auto_gc
, N_("enable auto-gc mode")),
285 OPT_BOOL(0, "force", &force
, N_("force running gc even if there may be another gc running")),
289 if (argc
== 2 && !strcmp(argv
[1], "-h"))
290 usage_with_options(builtin_gc_usage
, builtin_gc_options
);
292 argv_array_pushl(&pack_refs_cmd
, "pack-refs", "--all", "--prune", NULL
);
293 argv_array_pushl(&reflog
, "reflog", "expire", "--all", NULL
);
294 argv_array_pushl(&repack
, "repack", "-d", "-l", NULL
);
295 argv_array_pushl(&prune
, "prune", "--expire", NULL
);
296 argv_array_pushl(&rerere
, "rerere", "gc", NULL
);
298 git_config(gc_config
, NULL
);
301 pack_refs
= !is_bare_repository();
303 argc
= parse_options(argc
, argv
, prefix
, builtin_gc_options
,
304 builtin_gc_usage
, 0);
306 usage_with_options(builtin_gc_usage
, builtin_gc_options
);
309 argv_array_push(&repack
, "-f");
310 argv_array_push(&repack
, "--depth=250");
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 (run_command_v_opt(rerere
.argv
, RUN_GIT_CMD
))
365 return error(FAILED_RUN
, rerere
.argv
[0]);
367 if (auto_gc
&& too_many_loose_objects())
368 warning(_("There are too many unreachable loose objects; "
369 "run 'git prune' to remove them."));