Andrew's git - gitweb.git/blob

   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
  20static const char * const builtin_gc_usage[] = {
  21        "git gc [options]",
  22        NULL
  23};
  24
  25static int pack_refs = 1;
  26static int aggressive_window = -1;
  27static int gc_auto_threshold = 6700;
  28static int gc_auto_pack_limit = 50;
  29static char *prune_expire = "2.weeks.ago";
  30
  31#define MAX_ADD 10
  32static const char *argv_pack_refs[] = {"pack-refs", "--all", "--prune", NULL};
  33static const char *argv_reflog[] = {"reflog", "expire", "--all", NULL};
  34static const char *argv_repack[MAX_ADD] = {"repack", "-d", "-l", NULL};
  35static const char *argv_prune[] = {"prune", "--expire", NULL, NULL};
  36static const char *argv_rerere[] = {"rerere", "gc", NULL};
  37
  38static 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)
  61                        return config_error_nonbool(var);
  62                if (strcmp(value, "now")) {
  63                        unsigned long now = approxidate("now");
  64                        if (approxidate(value) >= now)
  65                                return error("Invalid %s: '%s'", var, value);
  66                }
  67                prune_expire = xstrdup(value);
  68                return 0;
  69        }
  70        return git_default_config(var, value, cb);
  71}
  72
  73static void append_option(const char **cmd, const char *opt, int max_length)
  74{
  75        int i;
  76
  77        for (i = 0; cmd[i]; i++)
  78                ;
  79
  80        if (i + 2 >= max_length)
  81                die("Too many options specified");
  82        cmd[i++] = opt;
  83        cmd[i] = NULL;
  84}
  85
  86static 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
 127static 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
 150static int run_hook(void)
 151{
 152        const char *argv[2];
 153        struct child_process hook;
 154        int ret;
 155
 156        argv[0] = git_path("hooks/pre-auto-gc");
 157        argv[1] = NULL;
 158
 159        if (access(argv[0], X_OK) < 0)
 160                return 0;
 161
 162        memset(&hook, 0, sizeof(hook));
 163        hook.argv = argv;
 164        hook.no_stdin = 1;
 165        hook.stdout_to_stderr = 1;
 166
 167        ret = start_command(&hook);
 168        if (ret) {
 169                warning("Could not spawn %s", argv[0]);
 170                return ret;
 171        }
 172        ret = finish_command(&hook);
 173        if (ret == -ERR_RUN_COMMAND_WAITPID_SIGNAL)
 174                warning("%s exited due to uncaught signal", argv[0]);
 175        return ret;
 176}
 177
 178static int need_to_gc(void)
 179{
 180        /*
 181         * Setting gc.auto to 0 or negative can disable the
 182         * automatic gc.
 183         */
 184        if (gc_auto_threshold <= 0)
 185                return 0;
 186
 187        /*
 188         * If there are too many loose objects, but not too many
 189         * packs, we run "repack -d -l".  If there are too many packs,
 190         * we run "repack -A -d -l".  Otherwise we tell the caller
 191         * there is no need.
 192         */
 193        if (too_many_packs())
 194                append_option(argv_repack, "-A", MAX_ADD);
 195        else if (!too_many_loose_objects())
 196                return 0;
 197
 198        if (run_hook())
 199                return 0;
 200        return 1;
 201}
 202
 203int cmd_gc(int argc, const char **argv, const char *prefix)
 204{
 205        int prune = 0;
 206        int aggressive = 0;
 207        int auto_gc = 0;
 208        int quiet = 0;
 209        char buf[80];
 210
 211        struct option builtin_gc_options[] = {
 212                OPT_BOOLEAN(0, "prune", &prune, "prune unreferenced objects (deprecated)"),
 213                OPT_BOOLEAN(0, "aggressive", &aggressive, "be more thorough (increased runtime)"),
 214                OPT_BOOLEAN(0, "auto", &auto_gc, "enable auto-gc mode"),
 215                OPT_BOOLEAN('q', "quiet", &quiet, "suppress progress reports"),
 216                OPT_END()
 217        };
 218
 219        git_config(gc_config, NULL);
 220
 221        if (pack_refs < 0)
 222                pack_refs = !is_bare_repository();
 223
 224        argc = parse_options(argc, argv, builtin_gc_options, builtin_gc_usage, 0);
 225        if (argc > 0)
 226                usage_with_options(builtin_gc_usage, builtin_gc_options);
 227
 228        if (aggressive) {
 229                append_option(argv_repack, "-f", MAX_ADD);
 230                if (aggressive_window > 0) {
 231                        sprintf(buf, "--window=%d", aggressive_window);
 232                        append_option(argv_repack, buf, MAX_ADD);
 233                }
 234        }
 235        if (quiet)
 236                append_option(argv_repack, "-q", MAX_ADD);
 237
 238        if (auto_gc) {
 239                /*
 240                 * Auto-gc should be least intrusive as possible.
 241                 */
 242                if (!need_to_gc())
 243                        return 0;
 244                fprintf(stderr, "Auto packing your repository for optimum "
 245                        "performance. You may also\n"
 246                        "run \"git gc\" manually. See "
 247                        "\"git help gc\" for more information.\n");
 248        } else
 249                append_option(argv_repack, "-A", MAX_ADD);
 250
 251        if (pack_refs && run_command_v_opt(argv_pack_refs, RUN_GIT_CMD))
 252                return error(FAILED_RUN, argv_pack_refs[0]);
 253
 254        if (run_command_v_opt(argv_reflog, RUN_GIT_CMD))
 255                return error(FAILED_RUN, argv_reflog[0]);
 256
 257        if (run_command_v_opt(argv_repack, RUN_GIT_CMD))
 258                return error(FAILED_RUN, argv_repack[0]);
 259
 260        argv_prune[2] = prune_expire;
 261        if (run_command_v_opt(argv_prune, RUN_GIT_CMD))
 262                return error(FAILED_RUN, argv_prune[0]);
 263
 264        if (run_command_v_opt(argv_rerere, RUN_GIT_CMD))
 265                return error(FAILED_RUN, argv_rerere[0]);
 266
 267        if (auto_gc && too_many_loose_objects())
 268                warning("There are too many unreachable loose objects; "
 269                        "run 'git prune' to remove them.");
 270
 271        return 0;
 272}