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