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
  18 #define FAILED_RUN "failed to run %s"
  19
  20 static const char * const builtin_gc_usage[] = {
  21         "git gc [options]",
  22         NULL
  23 };
  24
  25 static char do_rev_cache = 0;
  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 #define MAX_ADD 10
  33 static const char *argv_pack_refs[] = {"pack-refs", "--all", "--prune", NULL};
  34 static const char *argv_reflog[] = {"reflog", "expire", "--all", NULL};
  35 static const char *argv_repack[MAX_ADD] = {"repack", "-d", "-l", NULL};
  36 static const char *argv_prune[] = {"prune", "--expire", NULL, NULL};
  37 static const char *argv_rerere[] = {"rerere", "gc", NULL};
  38 static const char *argv_rev_cache[] = {"rev-cache", "fuse", "--all", "--ignore-size", NULL};
  39
  40 static int gc_config(const char *var, const char *value, void *cb)
  41 {
  42         if (!strcmp(var, "gc.revcache")) {
  43                 do_rev_cache = 1;
  44                 return 0;
  45         }
  46         if (!strcmp(var, "gc.packrefs")) {
  47                 if (value && !strcmp(value, "notbare"))
  48                         pack_refs = -1;
  49                 else
  50                         pack_refs = git_config_bool(var, value);
  51                 return 0;
  52         }
  53         if (!strcmp(var, "gc.aggressivewindow")) {
  54                 aggressive_window = git_config_int(var, value);
  55                 return 0;
  56         }
  57         if (!strcmp(var, "gc.auto")) {
  58                 gc_auto_threshold = git_config_int(var, value);
  59                 return 0;
  60         }
  61         if (!strcmp(var, "gc.autopacklimit")) {
  62                 gc_auto_pack_limit = git_config_int(var, value);
  63                 return 0;
  64         }
  65         if (!strcmp(var, "gc.pruneexpire")) {
  66                 if (value && strcmp(value, "now")) {
  67                         unsigned long now = approxidate("now");
  68                         if (approxidate(value) >= now)
  69                                 return error("Invalid %s: '%s'", var, value);
  70                 }
  71                 return git_config_string(&prune_expire, var, value);
  72         }
  73         return git_default_config(var, value, cb);
  74 }
  75
  76 static void append_option(const char **cmd, const char *opt, int max_length)
  77 {
  78         int i;
  79
  80         for (i = 0; cmd[i]; i++)
  81                 ;
  82
  83         if (i + 2 >= max_length)
  84                 die("Too many options specified");
  85         cmd[i++] = opt;
  86         cmd[i] = NULL;
  87 }
  88
  89 static int too_many_loose_objects(void)
  90 {
  91         /*
  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
  95          * estimate.
  96          */
  97         char path[PATH_MAX];
  98         const char *objdir = get_object_directory();
  99         DIR *dir;
 100         struct dirent *ent;
 101         int auto_threshold;
 102         int num_loose = 0;
 103         int needed = 0;
 104
 105         if (gc_auto_threshold <= 0)
 106                 return 0;
 107
 108         if (sizeof(path) <= snprintf(path, sizeof(path), "%s/17", objdir)) {
 109                 warning("insanely long object directory %.*s", 50, objdir);
 110                 return 0;
 111         }
 112         dir = opendir(path);
 113         if (!dir)
 114                 return 0;
 115
 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')
 120                         continue;
 121                 if (++num_loose > auto_threshold) {
 122                         needed = 1;
 123                         break;
 124                 }
 125         }
 126         closedir(dir);
 127         return needed;
 128 }
 129
 130 static int too_many_packs(void)
 131 {
 132         struct packed_git *p;
 133         int cnt;
 134
 135         if (gc_auto_pack_limit <= 0)
 136                 return 0;
 137
 138         prepare_packed_git();
 139         for (cnt = 0, p = packed_git; p; p = p->next) {
 140                 if (!p->pack_local)
 141                         continue;
 142                 if (p->pack_keep)
 143                         continue;
 144                 /*
 145                  * Perhaps check the size of the pack and count only
 146                  * very small ones here?
 147                  */
 148                 cnt++;
 149         }
 150         return gc_auto_pack_limit <= cnt;
 151 }
 152
 153 static int need_to_gc(void)
 154 {
 155         /*
 156          * Setting gc.auto to 0 or negative can disable the
 157          * automatic gc.
 158          */
 159         if (gc_auto_threshold <= 0)
 160                 return 0;
 161
 162         /*
 163          * If there are too many loose objects, but not too many
 164          * packs, we run "repack -d -l".  If there are too many packs,
 165          * we run "repack -A -d -l".  Otherwise we tell the caller
 166          * there is no need.
 167          */
 168         if (too_many_packs())
 169                 append_option(argv_repack,
 170                               prune_expire && !strcmp(prune_expire, "now") ?
 171                               "-a" : "-A",
 172                               MAX_ADD);
 173         else if (!too_many_loose_objects())
 174                 return 0;
 175
 176         if (run_hook(NULL, "pre-auto-gc", NULL))
 177                 return 0;
 178         return 1;
 179 }
 180
 181 int cmd_gc(int argc, const char **argv, const char *prefix)
 182 {
 183         int aggressive = 0;
 184         int auto_gc = 0;
 185         int quiet = 0;
 186         char buf[80];
 187
 188         struct option builtin_gc_options[] = {
 189                 { OPTION_STRING, 0, "prune", &prune_expire, "date",
 190                         "prune unreferenced objects",
 191                         PARSE_OPT_OPTARG, NULL, (intptr_t)prune_expire },
 192                 OPT_BOOLEAN(0, "aggressive", &aggressive, "be more thorough (increased runtime)"),
 193                 OPT_BOOLEAN(0, "auto", &auto_gc, "enable auto-gc mode"),
 194                 OPT_BOOLEAN('q', "quiet", &quiet, "suppress progress reports"),
 195                 OPT_END()
 196         };
 197
 198         git_config(gc_config, NULL);
 199
 200         if (pack_refs < 0)
 201                 pack_refs = !is_bare_repository();
 202
 203         argc = parse_options(argc, argv, prefix, builtin_gc_options,
 204                              builtin_gc_usage, 0);
 205         if (argc > 0)
 206                 usage_with_options(builtin_gc_usage, builtin_gc_options);
 207
 208         if (aggressive) {
 209                 append_option(argv_repack, "-f", MAX_ADD);
 210                 append_option(argv_repack, "--depth=250", MAX_ADD);
 211                 if (aggressive_window > 0) {
 212                         sprintf(buf, "--window=%d", aggressive_window);
 213                         append_option(argv_repack, buf, MAX_ADD);
 214                 }
 215         }
 216         if (quiet)
 217                 append_option(argv_repack, "-q", MAX_ADD);
 218
 219         if (auto_gc) {
 220                 /*
 221                  * Auto-gc should be least intrusive as possible.
 222                  */
 223                 if (!need_to_gc())
 224                         return 0;
 225                 fprintf(stderr, "Auto packing your repository for optimum "
 226                         "performance. You may also\n"
 227                         "run \"git gc\" manually. See "
 228                         "\"git help gc\" for more information.\n");
 229         } else
 230                 append_option(argv_repack,
 231                               prune_expire && !strcmp(prune_expire, "now")
 232                               ? "-a" : "-A",
 233                               MAX_ADD);
 234
 235         if (pack_refs && run_command_v_opt(argv_pack_refs, RUN_GIT_CMD))
 236                 return error(FAILED_RUN, argv_pack_refs[0]);
 237
 238         if (run_command_v_opt(argv_reflog, RUN_GIT_CMD))
 239                 return error(FAILED_RUN, argv_reflog[0]);
 240
 241         if (run_command_v_opt(argv_repack, RUN_GIT_CMD))
 242                 return error(FAILED_RUN, argv_repack[0]);
 243
 244         if (prune_expire) {
 245                 argv_prune[2] = prune_expire;
 246                 if (run_command_v_opt(argv_prune, RUN_GIT_CMD))
 247                         return error(FAILED_RUN, argv_prune[0]);
 248         }
 249
 250         if (run_command_v_opt(argv_rerere, RUN_GIT_CMD))
 251                 return error(FAILED_RUN, argv_rerere[0]);
 252
 253         if (do_rev_cache && run_command_v_opt(argv_rev_cache, RUN_GIT_CMD))
 254                 return error(FAILED_RUN, argv_rev_cache[0]);
 255
 256         if (auto_gc && too_many_loose_objects())
 257                 warning("There are too many unreachable loose objects; "
 258                         "run 'git prune' to remove them.");
 259
 260         return 0;
 261 }