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 "run-command.h"
  16
  17 #define FAILED_RUN "failed to run %s"
  18
  19 static const char builtin_gc_usage[] = "git-gc [--prune] [--aggressive]";
  20
  21 static int pack_refs = 1;
  22 static int aggressive_window = -1;
  23 static int gc_auto_threshold = 6700;
  24 static int gc_auto_pack_limit = 20;
  25
  26 #define MAX_ADD 10
  27 static const char *argv_pack_refs[] = {"pack-refs", "--all", "--prune", NULL};
  28 static const char *argv_reflog[] = {"reflog", "expire", "--all", NULL};
  29 static const char *argv_repack[MAX_ADD] = {"repack", "-d", "-l", NULL};
  30 static const char *argv_prune[] = {"prune", NULL};
  31 static const char *argv_rerere[] = {"rerere", "gc", NULL};
  32
  33 static int gc_config(const char *var, const char *value)
  34 {
  35         if (!strcmp(var, "gc.packrefs")) {
  36                 if (!strcmp(value, "notbare"))
  37                         pack_refs = -1;
  38                 else
  39                         pack_refs = git_config_bool(var, value);
  40                 return 0;
  41         }
  42         if (!strcmp(var, "gc.aggressivewindow")) {
  43                 aggressive_window = git_config_int(var, value);
  44                 return 0;
  45         }
  46         if (!strcmp(var, "gc.auto")) {
  47                 gc_auto_threshold = git_config_int(var, value);
  48                 return 0;
  49         }
  50         if (!strcmp(var, "gc.autopacklimit")) {
  51                 gc_auto_pack_limit = git_config_int(var, value);
  52                 return 0;
  53         }
  54         return git_default_config(var, value);
  55 }
  56
  57 static void append_option(const char **cmd, const char *opt, int max_length)
  58 {
  59         int i;
  60
  61         for (i = 0; cmd[i]; i++)
  62                 ;
  63
  64         if (i + 2 >= max_length)
  65                 die("Too many options specified");
  66         cmd[i++] = opt;
  67         cmd[i] = NULL;
  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                 char path[PATH_MAX];
 122                 size_t len;
 123                 int keep;
 124
 125                 if (!p->pack_local)
 126                         continue;
 127                 len = strlen(p->pack_name);
 128                 if (PATH_MAX <= len + 1)
 129                         continue; /* oops, give up */
 130                 memcpy(path, p->pack_name, len-5);
 131                 memcpy(path + len - 5, ".keep", 6);
 132                 keep = access(p->pack_name, F_OK) && (errno == ENOENT);
 133                 if (keep)
 134                         continue;
 135                 /*
 136                  * Perhaps check the size of the pack and count only
 137                  * very small ones here?
 138                  */
 139                 cnt++;
 140         }
 141         return gc_auto_pack_limit <= cnt;
 142 }
 143
 144 static int need_to_gc(void)
 145 {
 146         /*
 147          * Setting gc.auto and gc.autopacklimit to 0 or negative can
 148          * disable the automatic gc.
 149          */
 150         if (gc_auto_threshold <= 0 && gc_auto_pack_limit <= 0)
 151                 return 0;
 152
 153         /*
 154          * If there are too many loose objects, but not too many
 155          * packs, we run "repack -d -l".  If there are too many packs,
 156          * we run "repack -A -d -l".  Otherwise we tell the caller
 157          * there is no need.
 158          */
 159         if (too_many_packs())
 160                 append_option(argv_repack, "-A", MAX_ADD);
 161         else if (!too_many_loose_objects())
 162                 return 0;
 163         return 1;
 164 }
 165
 166 int cmd_gc(int argc, const char **argv, const char *prefix)
 167 {
 168         int i;
 169         int prune = 0;
 170         int auto_gc = 0;
 171         char buf[80];
 172
 173         git_config(gc_config);
 174
 175         if (pack_refs < 0)
 176                 pack_refs = !is_bare_repository();
 177
 178         for (i = 1; i < argc; i++) {
 179                 const char *arg = argv[i];
 180                 if (!strcmp(arg, "--prune")) {
 181                         prune = 1;
 182                         continue;
 183                 }
 184                 if (!strcmp(arg, "--aggressive")) {
 185                         append_option(argv_repack, "-f", MAX_ADD);
 186                         if (aggressive_window > 0) {
 187                                 sprintf(buf, "--window=%d", aggressive_window);
 188                                 append_option(argv_repack, buf, MAX_ADD);
 189                         }
 190                         continue;
 191                 }
 192                 if (!strcmp(arg, "--auto")) {
 193                         auto_gc = 1;
 194                         continue;
 195                 }
 196                 break;
 197         }
 198         if (i != argc)
 199                 usage(builtin_gc_usage);
 200
 201         if (auto_gc) {
 202                 /*
 203                  * Auto-gc should be least intrusive as possible.
 204                  */
 205                 prune = 0;
 206                 if (!need_to_gc())
 207                         return 0;
 208         } else {
 209                 /*
 210                  * Use safer (for shared repos) "-A" option to
 211                  * repack when not pruning. Auto-gc makes its
 212                  * own decision.
 213                  */
 214                 if (prune)
 215                         append_option(argv_repack, "-a", MAX_ADD);
 216                 else
 217                         append_option(argv_repack, "-A", MAX_ADD);
 218         }
 219
 220         if (pack_refs && run_command_v_opt(argv_pack_refs, RUN_GIT_CMD))
 221                 return error(FAILED_RUN, argv_pack_refs[0]);
 222
 223         if (run_command_v_opt(argv_reflog, RUN_GIT_CMD))
 224                 return error(FAILED_RUN, argv_reflog[0]);
 225
 226         if (run_command_v_opt(argv_repack, RUN_GIT_CMD))
 227                 return error(FAILED_RUN, argv_repack[0]);
 228
 229         if (prune && run_command_v_opt(argv_prune, RUN_GIT_CMD))
 230                 return error(FAILED_RUN, argv_prune[0]);
 231
 232         if (run_command_v_opt(argv_rerere, RUN_GIT_CMD))
 233                 return error(FAILED_RUN, argv_rerere[0]);
 234
 235         if (auto_gc && too_many_loose_objects())
 236                 warning("There are too many unreachable loose objects; "
 237                         "run 'git prune' to remove them.");
 238
 239         return 0;
 240 }