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 int git_config_date_string(const char **output
,
61 const char *var
, const char *value
)
63 if (value
&& strcmp(value
, "now")) {
64 unsigned long now
= approxidate("now");
65 if (approxidate(value
) >= now
)
66 return error(_("Invalid %s: '%s'"), var
, value
);
68 return git_config_string(output
, var
, value
);
71 static int gc_config(const char *var
, const char *value
, void *cb
)
73 if (!strcmp(var
, "gc.packrefs")) {
74 if (value
&& !strcmp(value
, "notbare"))
77 pack_refs
= git_config_bool(var
, value
);
80 if (!strcmp(var
, "gc.aggressivewindow")) {
81 aggressive_window
= git_config_int(var
, value
);
84 if (!strcmp(var
, "gc.aggressivedepth")) {
85 aggressive_depth
= git_config_int(var
, value
);
88 if (!strcmp(var
, "gc.auto")) {
89 gc_auto_threshold
= git_config_int(var
, value
);
92 if (!strcmp(var
, "gc.autopacklimit")) {
93 gc_auto_pack_limit
= git_config_int(var
, value
);
96 if (!strcmp(var
, "gc.autodetach")) {
97 detach_auto
= git_config_bool(var
, value
);
100 if (!strcmp(var
, "gc.pruneexpire"))
101 return git_config_date_string(&prune_expire
, var
, value
);
102 if (!strcmp(var
, "gc.prunereposexpire"))
103 return git_config_date_string(&prune_repos_expire
, var
, value
);
104 return git_default_config(var
, value
, cb
);
107 static int too_many_loose_objects(void)
110 * Quickly check if a "gc" is needed, by estimating how
111 * many loose objects there are. Because SHA-1 is evenly
112 * distributed, we can check only one and get a reasonable
116 const char *objdir
= get_object_directory();
123 if (gc_auto_threshold
<= 0)
126 if (sizeof(path
) <= snprintf(path
, sizeof(path
), "%s/17", objdir
)) {
127 warning(_("insanely long object directory %.*s"), 50, objdir
);
134 auto_threshold
= (gc_auto_threshold
+ 255) / 256;
135 while ((ent
= readdir(dir
)) != NULL
) {
136 if (strspn(ent
->d_name
, "0123456789abcdef") != 38 ||
137 ent
->d_name
[38] != '\0')
139 if (++num_loose
> auto_threshold
) {
148 static int too_many_packs(void)
150 struct packed_git
*p
;
153 if (gc_auto_pack_limit
<= 0)
156 prepare_packed_git();
157 for (cnt
= 0, p
= packed_git
; p
; p
= p
->next
) {
163 * Perhaps check the size of the pack and count only
164 * very small ones here?
168 return gc_auto_pack_limit
<= cnt
;
171 static void add_repack_all_option(void)
173 if (prune_expire
&& !strcmp(prune_expire
, "now"))
174 argv_array_push(&repack
, "-a");
176 argv_array_push(&repack
, "-A");
178 argv_array_pushf(&repack
, "--unpack-unreachable=%s", prune_expire
);
182 static int need_to_gc(void)
185 * Setting gc.auto to 0 or negative can disable the
188 if (gc_auto_threshold
<= 0)
192 * If there are too many loose objects, but not too many
193 * packs, we run "repack -d -l". If there are too many packs,
194 * we run "repack -A -d -l". Otherwise we tell the caller
197 if (too_many_packs())
198 add_repack_all_option();
199 else if (!too_many_loose_objects())
202 if (run_hook_le(NULL
, "pre-auto-gc", NULL
))
207 /* return NULL on success, else hostname running the gc */
208 static const char *lock_repo_for_gc(int force
, pid_t
* ret_pid
)
210 static struct lock_file lock
;
212 struct strbuf sb
= STRBUF_INIT
;
222 if (gethostname(my_host
, sizeof(my_host
)))
223 strcpy(my_host
, "unknown");
225 fd
= hold_lock_file_for_update(&lock
, git_path("gc.pid"),
228 static char locking_host
[128];
230 fp
= fopen(git_path("gc.pid"), "r");
231 memset(locking_host
, 0, sizeof(locking_host
));
234 !fstat(fileno(fp
), &st
) &&
236 * 12 hour limit is very generous as gc should
237 * never take that long. On the other hand we
238 * don't really need a strict limit here,
239 * running gc --auto one day late is not a big
240 * problem. --force can be used in manual gc
241 * after the user verifies that no gc is
244 time(NULL
) - st
.st_mtime
<= 12 * 3600 &&
245 fscanf(fp
, "%"PRIuMAX
" %127c", &pid
, locking_host
) == 2 &&
246 /* be gentle to concurrent "gc" on remote hosts */
247 (strcmp(locking_host
, my_host
) || !kill(pid
, 0) || errno
== EPERM
);
252 rollback_lock_file(&lock
);
258 strbuf_addf(&sb
, "%"PRIuMAX
" %s",
259 (uintmax_t) getpid(), my_host
);
260 write_in_full(fd
, sb
.buf
, sb
.len
);
262 commit_lock_file(&lock
);
264 pidfile
= git_pathdup("gc.pid");
265 sigchain_push_common(remove_pidfile_on_signal
);
266 atexit(remove_pidfile
);
271 static int gc_before_repack(void)
273 if (pack_refs
&& run_command_v_opt(pack_refs_cmd
.argv
, RUN_GIT_CMD
))
274 return error(FAILED_RUN
, pack_refs_cmd
.argv
[0]);
276 if (prune_reflogs
&& run_command_v_opt(reflog
.argv
, RUN_GIT_CMD
))
277 return error(FAILED_RUN
, reflog
.argv
[0]);
284 int cmd_gc(int argc
, const char **argv
, const char *prefix
)
293 struct option builtin_gc_options
[] = {
294 OPT__QUIET(&quiet
, N_("suppress progress reporting")),
295 { OPTION_STRING
, 0, "prune", &prune_expire
, N_("date"),
296 N_("prune unreferenced objects"),
297 PARSE_OPT_OPTARG
, NULL
, (intptr_t)prune_expire
},
298 OPT_BOOL(0, "aggressive", &aggressive
, N_("be more thorough (increased runtime)")),
299 OPT_BOOL(0, "auto", &auto_gc
, N_("enable auto-gc mode")),
300 OPT_BOOL(0, "force", &force
, N_("force running gc even if there may be another gc running")),
304 if (argc
== 2 && !strcmp(argv
[1], "-h"))
305 usage_with_options(builtin_gc_usage
, builtin_gc_options
);
307 argv_array_pushl(&pack_refs_cmd
, "pack-refs", "--all", "--prune", NULL
);
308 argv_array_pushl(&reflog
, "reflog", "expire", "--all", NULL
);
309 argv_array_pushl(&repack
, "repack", "-d", "-l", NULL
);
310 argv_array_pushl(&prune
, "prune", "--expire", NULL
);
311 argv_array_pushl(&prune_repos
, "prune", "--repos", "--expire", NULL
);
312 argv_array_pushl(&rerere
, "rerere", "gc", NULL
);
314 git_config(gc_config
, NULL
);
317 pack_refs
= !is_bare_repository();
319 argc
= parse_options(argc
, argv
, prefix
, builtin_gc_options
,
320 builtin_gc_usage
, 0);
322 usage_with_options(builtin_gc_usage
, builtin_gc_options
);
325 argv_array_push(&repack
, "-f");
326 if (aggressive_depth
> 0)
327 argv_array_pushf(&repack
, "--depth=%d", aggressive_depth
);
328 if (aggressive_window
> 0)
329 argv_array_pushf(&repack
, "--window=%d", aggressive_window
);
332 argv_array_push(&repack
, "-q");
336 * Auto-gc should be least intrusive as possible.
342 fprintf(stderr
, _("Auto packing the repository in background for optimum performance.\n"));
344 fprintf(stderr
, _("Auto packing the repository for optimum performance.\n"));
345 fprintf(stderr
, _("See \"git help gc\" for manual housekeeping.\n"));
348 if (gc_before_repack())
351 * failure to daemonize is ok, we'll continue
357 add_repack_all_option();
359 name
= lock_repo_for_gc(force
, &pid
);
362 return 0; /* be quiet on --auto */
363 die(_("gc is already running on machine '%s' pid %"PRIuMAX
" (use --force if not)"),
364 name
, (uintmax_t)pid
);
367 if (gc_before_repack())
370 if (run_command_v_opt(repack
.argv
, RUN_GIT_CMD
))
371 return error(FAILED_RUN
, repack
.argv
[0]);
374 argv_array_push(&prune
, prune_expire
);
376 argv_array_push(&prune
, "--no-progress");
377 if (run_command_v_opt(prune
.argv
, RUN_GIT_CMD
))
378 return error(FAILED_RUN
, prune
.argv
[0]);
381 if (prune_repos_expire
) {
382 argv_array_push(&prune_repos
, prune_repos_expire
);
383 if (run_command_v_opt(prune_repos
.argv
, RUN_GIT_CMD
))
384 return error(FAILED_RUN
, prune_repos
.argv
[0]);
387 if (run_command_v_opt(rerere
.argv
, RUN_GIT_CMD
))
388 return error(FAILED_RUN
, rerere
.argv
[0]);
390 if (auto_gc
&& too_many_loose_objects())
391 warning(_("There are too many unreachable loose objects; "
392 "run 'git prune' to remove them."));