builtin/gc.c

   1 /*
   2  * git gc builtin command
   3  *
   4  * Cleanup unreachable files and optimize the repository.
   5  *
   6  * Copyright (c) 2007 James Bowes
   7  *
   8  * Based on git-gc.sh, which is
   9  *
  10  * Copyright (c) 2006 Shawn O. Pearce
  11  */
  12
  13 #include "builtin.h"
  14 #include "cache.h"
  15 #include "parse-options.h"
  16 #include "run-command.h"
  17 #include "sigchain.h"
  18 #include "argv-array.h"
  19
  20 #define FAILED_RUN "failed to run %s"
  21
  22 static const char * const builtin_gc_usage[] = {
  23         N_("git gc [options]"),
  24         NULL
  25 };
  26
  27 static int pack_refs = 1;
  28 static int aggressive_window = 250;
  29 static int gc_auto_threshold = 6700;
  30 static int gc_auto_pack_limit = 50;
  31 static const char *prune_expire = "2.weeks.ago";
  32
  33 static struct argv_array pack_refs_cmd = ARGV_ARRAY_INIT;
  34 static struct argv_array reflog = ARGV_ARRAY_INIT;
  35 static struct argv_array repack = ARGV_ARRAY_INIT;
  36 static struct argv_array prune = ARGV_ARRAY_INIT;
  37 static struct argv_array rerere = ARGV_ARRAY_INIT;
  38
  39 static char *pidfile;
  40
  41 static void remove_pidfile(void)
  42 {
  43         if (pidfile)
  44                 unlink(pidfile);
  45 }
  46
  47 static void remove_pidfile_on_signal(int signo)
  48 {
  49         remove_pidfile();
  50         sigchain_pop(signo);
  51         raise(signo);
  52 }
  53
  54 static int gc_config(const char *var, const char *value, void *cb)
  55 {
  56         if (!strcmp(var, "gc.packrefs")) {
  57                 if (value && !strcmp(value, "notbare"))
  58                         pack_refs = -1;
  59                 else
  60                         pack_refs = git_config_bool(var, value);
  61                 return 0;
  62         }
  63         if (!strcmp(var, "gc.aggressivewindow")) {
  64                 aggressive_window = git_config_int(var, value);
  65                 return 0;
  66         }
  67         if (!strcmp(var, "gc.auto")) {
  68                 gc_auto_threshold = git_config_int(var, value);
  69                 return 0;
  70         }
  71         if (!strcmp(var, "gc.autopacklimit")) {
  72                 gc_auto_pack_limit = git_config_int(var, value);
  73                 return 0;
  74         }
  75         if (!strcmp(var, "gc.pruneexpire")) {
  76                 if (value && strcmp(value, "now")) {
  77                         unsigned long now = approxidate("now");
  78                         if (approxidate(value) >= now)
  79                                 return error(_("Invalid %s: '%s'"), var, value);
  80                 }
  81                 return git_config_string(&prune_expire, var, value);
  82         }
  83         return git_default_config(var, value, cb);
  84 }
  85
  86 static int too_many_loose_objects(void)
  87 {
  88         /*
  89          * Quickly check if a "gc" is needed, by estimating how
  90          * many loose objects there are.  Because SHA-1 is evenly
  91          * distributed, we can check only one and get a reasonable
  92          * estimate.
  93          */
  94         char path[PATH_MAX];
  95         const char *objdir = get_object_directory();
  96         DIR *dir;
  97         struct dirent *ent;
  98         int auto_threshold;
  99         int num_loose = 0;
 100         int needed = 0;
 101
 102         if (gc_auto_threshold <= 0)
 103                 return 0;
 104
 105         if (sizeof(path) <= snprintf(path, sizeof(path), "%s/17", objdir)) {
 106                 warning(_("insanely long object directory %.*s"), 50, objdir);
 107                 return 0;
 108         }
 109         dir = opendir(path);
 110         if (!dir)
 111                 return 0;
 112
 113         auto_threshold = (gc_auto_threshold + 255) / 256;
 114         while ((ent = readdir(dir)) != NULL) {
 115                 if (strspn(ent->d_name, "0123456789abcdef") != 38 ||
 116                     ent->d_name[38] != '\0')
 117                         continue;
 118                 if (++num_loose > auto_threshold) {
 119                         needed = 1;
 120                         break;
 121                 }
 122         }
 123         closedir(dir);
 124         return needed;
 125 }
 126
 127 static int too_many_packs(void)
 128 {
 129         struct packed_git *p;
 130         int cnt;
 131
 132         if (gc_auto_pack_limit <= 0)
 133                 return 0;
 134
 135         prepare_packed_git();
 136         for (cnt = 0, p = packed_git; p; p = p->next) {
 137                 if (!p->pack_local)
 138                         continue;
 139                 if (p->pack_keep)
 140                         continue;
 141                 /*
 142                  * Perhaps check the size of the pack and count only
 143                  * very small ones here?
 144                  */
 145                 cnt++;
 146         }
 147         return gc_auto_pack_limit <= cnt;
 148 }
 149
 150 static void add_repack_all_option(void)
 151 {
 152         if (prune_expire && !strcmp(prune_expire, "now"))
 153                 argv_array_push(&repack, "-a");
 154         else {
 155                 argv_array_push(&repack, "-A");
 156                 if (prune_expire)
 157                         argv_array_pushf(&repack, "--unpack-unreachable=%s", prune_expire);
 158         }
 159 }
 160
 161 static int need_to_gc(void)
 162 {
 163         /*
 164          * Setting gc.auto to 0 or negative can disable the
 165          * automatic gc.
 166          */
 167         if (gc_auto_threshold <= 0)
 168                 return 0;
 169
 170         /*
 171          * If there are too many loose objects, but not too many
 172          * packs, we run "repack -d -l".  If there are too many packs,
 173          * we run "repack -A -d -l".  Otherwise we tell the caller
 174          * there is no need.
 175          */
 176         if (too_many_packs())
 177                 add_repack_all_option();
 178         else if (!too_many_loose_objects())
 179                 return 0;
 180
 181         if (run_hook(NULL, "pre-auto-gc", NULL))
 182                 return 0;
 183         return 1;
 184 }
 185
 186 /* return NULL on success, else hostname running the gc */
 187 static const char *lock_repo_for_gc(int force, pid_t* ret_pid)
 188 {
 189         static struct lock_file lock;
 190         static char locking_host[128];
 191         char my_host[128];
 192         struct strbuf sb = STRBUF_INIT;
 193         struct stat st;
 194         uintmax_t pid;
 195         FILE *fp;
 196         int fd, should_exit;
 197
 198         if (pidfile)
 199                 /* already locked */
 200                 return NULL;
 201
 202         if (gethostname(my_host, sizeof(my_host)))
 203                 strcpy(my_host, "unknown");
 204
 205         fd = hold_lock_file_for_update(&lock, git_path("gc.pid"),
 206                                        LOCK_DIE_ON_ERROR);
 207         if (!force) {
 208                 fp = fopen(git_path("gc.pid"), "r");
 209                 memset(locking_host, 0, sizeof(locking_host));
 210                 should_exit =
 211                         fp != NULL &&
 212                         !fstat(fileno(fp), &st) &&
 213                         /*
 214                          * 12 hour limit is very generous as gc should
 215                          * never take that long. On the other hand we
 216                          * don't really need a strict limit here,
 217                          * running gc --auto one day late is not a big
 218                          * problem. --force can be used in manual gc
 219                          * after the user verifies that no gc is
 220                          * running.
 221                          */
 222                         time(NULL) - st.st_mtime <= 12 * 3600 &&
 223                         fscanf(fp, "%"PRIuMAX" %127c", &pid, locking_host) == 2 &&
 224                         /* be gentle to concurrent "gc" on remote hosts */
 225                         (strcmp(locking_host, my_host) || !kill(pid, 0));
 226                 if (fp != NULL)
 227                         fclose(fp);
 228                 if (should_exit) {
 229                         if (fd >= 0)
 230                                 rollback_lock_file(&lock);
 231                         *ret_pid = pid;
 232                         return locking_host;
 233                 }
 234         }
 235
 236         strbuf_addf(&sb, "%"PRIuMAX" %s",
 237                     (uintmax_t) getpid(), my_host);
 238         write_in_full(fd, sb.buf, sb.len);
 239         strbuf_release(&sb);
 240         commit_lock_file(&lock);
 241
 242         pidfile = git_pathdup("gc.pid");
 243         sigchain_push_common(remove_pidfile_on_signal);
 244         atexit(remove_pidfile);
 245
 246         return NULL;
 247 }
 248
 249 int cmd_gc(int argc, const char **argv, const char *prefix)
 250 {
 251         int aggressive = 0;
 252         int auto_gc = 0;
 253         int quiet = 0;
 254         int force = 0;
 255         const char *name;
 256         pid_t pid;
 257
 258         struct option builtin_gc_options[] = {
 259                 OPT__QUIET(&quiet, N_("suppress progress reporting")),
 260                 { OPTION_STRING, 0, "prune", &prune_expire, N_("date"),
 261                         N_("prune unreferenced objects"),
 262                         PARSE_OPT_OPTARG, NULL, (intptr_t)prune_expire },
 263                 OPT_BOOL(0, "aggressive", &aggressive, N_("be more thorough (increased runtime)")),
 264                 OPT_BOOL(0, "auto", &auto_gc, N_("enable auto-gc mode")),
 265                 OPT_BOOL(0, "force", &force, N_("force running gc even if there may be another gc running")),
 266                 OPT_END()
 267         };
 268
 269         if (argc == 2 && !strcmp(argv[1], "-h"))
 270                 usage_with_options(builtin_gc_usage, builtin_gc_options);
 271
 272         argv_array_pushl(&pack_refs_cmd, "pack-refs", "--all", "--prune", NULL);
 273         argv_array_pushl(&reflog, "reflog", "expire", "--all", NULL);
 274         argv_array_pushl(&repack, "repack", "-d", "-l", NULL);
 275         argv_array_pushl(&prune, "prune", "--expire", NULL );
 276         argv_array_pushl(&rerere, "rerere", "gc", NULL);
 277
 278         git_config(gc_config, NULL);
 279
 280         if (pack_refs < 0)
 281                 pack_refs = !is_bare_repository();
 282
 283         argc = parse_options(argc, argv, prefix, builtin_gc_options,
 284                              builtin_gc_usage, 0);
 285         if (argc > 0)
 286                 usage_with_options(builtin_gc_usage, builtin_gc_options);
 287
 288         if (aggressive) {
 289                 argv_array_push(&repack, "-f");
 290                 argv_array_push(&repack, "--depth=250");
 291                 if (aggressive_window > 0)
 292                         argv_array_pushf(&repack, "--window=%d", aggressive_window);
 293         }
 294         if (quiet)
 295                 argv_array_push(&repack, "-q");
 296
 297         if (auto_gc) {
 298                 /*
 299                  * Auto-gc should be least intrusive as possible.
 300                  */
 301                 if (!need_to_gc())
 302                         return 0;
 303                 if (!quiet)
 304                         fprintf(stderr,
 305                                         _("Auto packing the repository for optimum performance. You may also\n"
 306                                         "run \"git gc\" manually. See "
 307                                         "\"git help gc\" for more information.\n"));
 308         } else
 309                 add_repack_all_option();
 310
 311         name = lock_repo_for_gc(force, &pid);
 312         if (name) {
 313                 if (auto_gc)
 314                         return 0; /* be quiet on --auto */
 315                 die(_("gc is already running on machine '%s' pid %"PRIuMAX" (use --force if not)"),
 316                     name, (uintmax_t)pid);
 317         }
 318
 319         if (pack_refs && run_command_v_opt(pack_refs_cmd.argv, RUN_GIT_CMD))
 320                 return error(FAILED_RUN, pack_refs_cmd.argv[0]);
 321
 322         if (run_command_v_opt(reflog.argv, RUN_GIT_CMD))
 323                 return error(FAILED_RUN, reflog.argv[0]);
 324
 325         if (run_command_v_opt(repack.argv, RUN_GIT_CMD))
 326                 return error(FAILED_RUN, repack.argv[0]);
 327
 328         if (prune_expire) {
 329                 argv_array_push(&prune, prune_expire);
 330                 if (quiet)
 331                         argv_array_push(&prune, "--no-progress");
 332                 if (run_command_v_opt(prune.argv, RUN_GIT_CMD))
 333                         return error(FAILED_RUN, prune.argv[0]);
 334         }
 335
 336         if (run_command_v_opt(rerere.argv, RUN_GIT_CMD))
 337                 return error(FAILED_RUN, rerere.argv[0]);
 338
 339         if (auto_gc && too_many_loose_objects())
 340                 warning(_("There are too many unreachable loose objects; "
 341                         "run 'git prune' to remove them."));
 342
 343         return 0;
 344 }