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";
37 static struct argv_array pack_refs_cmd
= ARGV_ARRAY_INIT
;
38 static struct argv_array reflog
= ARGV_ARRAY_INIT
;
39 static struct argv_array repack
= ARGV_ARRAY_INIT
;
40 static struct argv_array prune
= ARGV_ARRAY_INIT
;
41 static struct argv_array rerere
= ARGV_ARRAY_INIT
;
45 static void remove_pidfile(void)
51 static void remove_pidfile_on_signal(int signo
)
58 static void git_config_date_string(const char *key
, const char **output
)
60 if (git_config_get_string_const(key
, output
))
62 if (strcmp(*output
, "now")) {
63 unsigned long now
= approxidate("now");
64 if (approxidate(*output
) >= now
)
65 git_die_config(key
, _("Invalid %s: '%s'"), key
, *output
);
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_date_string("gc.pruneexpire", &prune_expire
);
86 git_config(git_default_config
, NULL
);
89 static int too_many_loose_objects(void)
92 * Quickly check if a "gc" is needed, by estimating how
93 * many loose objects there are. Because SHA-1 is evenly
94 * distributed, we can check only one and get a reasonable
98 const char *objdir
= get_object_directory();
105 if (gc_auto_threshold
<= 0)
108 if (sizeof(path
) <= snprintf(path
, sizeof(path
), "%s/17", objdir
)) {
109 warning(_("insanely long object directory %.*s"), 50, objdir
);
116 auto_threshold
= (gc_auto_threshold
+ 255) / 256;
117 while ((ent
= readdir(dir
)) != NULL
) {
118 if (strspn(ent
->d_name
, "0123456789abcdef") != 38 ||
119 ent
->d_name
[38] != '\0')
121 if (++num_loose
> auto_threshold
) {
130 static int too_many_packs(void)
132 struct packed_git
*p
;
135 if (gc_auto_pack_limit
<= 0)
138 prepare_packed_git();
139 for (cnt
= 0, p
= packed_git
; p
; p
= p
->next
) {
145 * Perhaps check the size of the pack and count only
146 * very small ones here?
150 return gc_auto_pack_limit
<= cnt
;
153 static void add_repack_all_option(void)
155 if (prune_expire
&& !strcmp(prune_expire
, "now"))
156 argv_array_push(&repack
, "-a");
158 argv_array_push(&repack
, "-A");
160 argv_array_pushf(&repack
, "--unpack-unreachable=%s", prune_expire
);
164 static int need_to_gc(void)
167 * Setting gc.auto to 0 or negative can disable the
170 if (gc_auto_threshold
<= 0)
174 * If there are too many loose objects, but not too many
175 * packs, we run "repack -d -l". If there are too many packs,
176 * we run "repack -A -d -l". Otherwise we tell the caller
179 if (too_many_packs())
180 add_repack_all_option();
181 else if (!too_many_loose_objects())
184 if (run_hook_le(NULL
, "pre-auto-gc", NULL
))
189 /* return NULL on success, else hostname running the gc */
190 static const char *lock_repo_for_gc(int force
, pid_t
* ret_pid
)
192 static struct lock_file lock
;
194 struct strbuf sb
= STRBUF_INIT
;
204 if (gethostname(my_host
, sizeof(my_host
)))
205 strcpy(my_host
, "unknown");
207 fd
= hold_lock_file_for_update(&lock
, git_path("gc.pid"),
210 static char locking_host
[128];
212 fp
= fopen(git_path("gc.pid"), "r");
213 memset(locking_host
, 0, sizeof(locking_host
));
216 !fstat(fileno(fp
), &st
) &&
218 * 12 hour limit is very generous as gc should
219 * never take that long. On the other hand we
220 * don't really need a strict limit here,
221 * running gc --auto one day late is not a big
222 * problem. --force can be used in manual gc
223 * after the user verifies that no gc is
226 time(NULL
) - st
.st_mtime
<= 12 * 3600 &&
227 fscanf(fp
, "%"PRIuMAX
" %127c", &pid
, locking_host
) == 2 &&
228 /* be gentle to concurrent "gc" on remote hosts */
229 (strcmp(locking_host
, my_host
) || !kill(pid
, 0) || errno
== EPERM
);
234 rollback_lock_file(&lock
);
240 strbuf_addf(&sb
, "%"PRIuMAX
" %s",
241 (uintmax_t) getpid(), my_host
);
242 write_in_full(fd
, sb
.buf
, sb
.len
);
244 commit_lock_file(&lock
);
246 pidfile
= git_pathdup("gc.pid");
247 sigchain_push_common(remove_pidfile_on_signal
);
248 atexit(remove_pidfile
);
253 static int gc_before_repack(void)
255 if (pack_refs
&& run_command_v_opt(pack_refs_cmd
.argv
, RUN_GIT_CMD
))
256 return error(FAILED_RUN
, pack_refs_cmd
.argv
[0]);
258 if (prune_reflogs
&& run_command_v_opt(reflog
.argv
, RUN_GIT_CMD
))
259 return error(FAILED_RUN
, reflog
.argv
[0]);
266 int cmd_gc(int argc
, const char **argv
, const char *prefix
)
275 struct option builtin_gc_options
[] = {
276 OPT__QUIET(&quiet
, N_("suppress progress reporting")),
277 { OPTION_STRING
, 0, "prune", &prune_expire
, N_("date"),
278 N_("prune unreferenced objects"),
279 PARSE_OPT_OPTARG
, NULL
, (intptr_t)prune_expire
},
280 OPT_BOOL(0, "aggressive", &aggressive
, N_("be more thorough (increased runtime)")),
281 OPT_BOOL(0, "auto", &auto_gc
, N_("enable auto-gc mode")),
282 OPT_BOOL(0, "force", &force
, N_("force running gc even if there may be another gc running")),
286 if (argc
== 2 && !strcmp(argv
[1], "-h"))
287 usage_with_options(builtin_gc_usage
, builtin_gc_options
);
289 argv_array_pushl(&pack_refs_cmd
, "pack-refs", "--all", "--prune", NULL
);
290 argv_array_pushl(&reflog
, "reflog", "expire", "--all", NULL
);
291 argv_array_pushl(&repack
, "repack", "-d", "-l", NULL
);
292 argv_array_pushl(&prune
, "prune", "--expire", NULL
);
293 argv_array_pushl(&rerere
, "rerere", "gc", NULL
);
298 pack_refs
= !is_bare_repository();
300 argc
= parse_options(argc
, argv
, prefix
, builtin_gc_options
,
301 builtin_gc_usage
, 0);
303 usage_with_options(builtin_gc_usage
, builtin_gc_options
);
306 argv_array_push(&repack
, "-f");
307 if (aggressive_depth
> 0)
308 argv_array_pushf(&repack
, "--depth=%d", aggressive_depth
);
309 if (aggressive_window
> 0)
310 argv_array_pushf(&repack
, "--window=%d", aggressive_window
);
313 argv_array_push(&repack
, "-q");
317 * Auto-gc should be least intrusive as possible.
323 fprintf(stderr
, _("Auto packing the repository in background for optimum performance.\n"));
325 fprintf(stderr
, _("Auto packing the repository for optimum performance.\n"));
326 fprintf(stderr
, _("See \"git help gc\" for manual housekeeping.\n"));
329 if (gc_before_repack())
332 * failure to daemonize is ok, we'll continue
338 add_repack_all_option();
340 name
= lock_repo_for_gc(force
, &pid
);
343 return 0; /* be quiet on --auto */
344 die(_("gc is already running on machine '%s' pid %"PRIuMAX
" (use --force if not)"),
345 name
, (uintmax_t)pid
);
348 if (gc_before_repack())
351 if (run_command_v_opt(repack
.argv
, RUN_GIT_CMD
))
352 return error(FAILED_RUN
, repack
.argv
[0]);
355 argv_array_push(&prune
, prune_expire
);
357 argv_array_push(&prune
, "--no-progress");
358 if (run_command_v_opt(prune
.argv
, RUN_GIT_CMD
))
359 return error(FAILED_RUN
, prune
.argv
[0]);
362 if (run_command_v_opt(rerere
.argv
, RUN_GIT_CMD
))
363 return error(FAILED_RUN
, rerere
.argv
[0]);
365 if (auto_gc
&& too_many_loose_objects())
366 warning(_("There are too many unreachable loose objects; "
367 "run 'git prune' to remove them."));