builtin/rm.c

   1 /*
   2  * "git rm" builtin command
   3  *
   4  * Copyright (C) Linus Torvalds 2006
   5  */
   6 #include "cache.h"
   7 #include "builtin.h"
   8 #include "dir.h"
   9 #include "cache-tree.h"
  10 #include "tree-walk.h"
  11 #include "parse-options.h"
  12 #include "submodule.h"
  13
  14 static const char * const builtin_rm_usage[] = {
  15         N_("git rm [options] [--] <file>..."),
  16         NULL
  17 };
  18
  19 static struct {
  20         int nr, alloc;
  21         struct {
  22                 const char *name;
  23                 char is_submodule;
  24         } *entry;
  25 } list;
  26
  27 static int get_ours_cache_pos(const char *path, int pos)
  28 {
  29         int i = -pos - 1;
  30
  31         while ((i < active_nr) && !strcmp(active_cache[i]->name, path)) {
  32                 if (ce_stage(active_cache[i]) == 2)
  33                         return i;
  34                 i++;
  35         }
  36         return -1;
  37 }
  38
  39 static int check_submodules_use_gitfiles(void)
  40 {
  41         int i;
  42         int errs = 0;
  43
  44         for (i = 0; i < list.nr; i++) {
  45                 const char *name = list.entry[i].name;
  46                 int pos;
  47                 struct cache_entry *ce;
  48                 struct stat st;
  49
  50                 pos = cache_name_pos(name, strlen(name));
  51                 if (pos < 0) {
  52                         pos = get_ours_cache_pos(name, pos);
  53                         if (pos < 0)
  54                                 continue;
  55                 }
  56                 ce = active_cache[pos];
  57
  58                 if (!S_ISGITLINK(ce->ce_mode) ||
  59                     (lstat(ce->name, &st) < 0) ||
  60                     is_empty_dir(name))
  61                         continue;
  62
  63                 if (!submodule_uses_gitfile(name))
  64                         errs = error(_("submodule '%s' (or one of its nested "
  65                                      "submodules) uses a .git directory\n"
  66                                      "(use 'rm -rf' if you really want to remove "
  67                                      "it including all of its history)"), name);
  68         }
  69
  70         return errs;
  71 }
  72
  73 static int check_local_mod(unsigned char *head, int index_only)
  74 {
  75         /*
  76          * Items in list are already sorted in the cache order,
  77          * so we could do this a lot more efficiently by using
  78          * tree_desc based traversal if we wanted to, but I am
  79          * lazy, and who cares if removal of files is a tad
  80          * slower than the theoretical maximum speed?
  81          */
  82         int i, no_head;
  83         int errs = 0;
  84
  85         no_head = is_null_sha1(head);
  86         for (i = 0; i < list.nr; i++) {
  87                 struct stat st;
  88                 int pos;
  89                 struct cache_entry *ce;
  90                 const char *name = list.entry[i].name;
  91                 unsigned char sha1[20];
  92                 unsigned mode;
  93                 int local_changes = 0;
  94                 int staged_changes = 0;
  95
  96                 pos = cache_name_pos(name, strlen(name));
  97                 if (pos < 0) {
  98                         /*
  99                          * Skip unmerged entries except for populated submodules
 100                          * that could lose history when removed.
 101                          */
 102                         pos = get_ours_cache_pos(name, pos);
 103                         if (pos < 0)
 104                                 continue;
 105
 106                         if (!S_ISGITLINK(active_cache[pos]->ce_mode) ||
 107                             is_empty_dir(name))
 108                                 continue;
 109                 }
 110                 ce = active_cache[pos];
 111
 112                 if (lstat(ce->name, &st) < 0) {
 113                         if (errno != ENOENT && errno != ENOTDIR)
 114                                 warning("'%s': %s", ce->name, strerror(errno));
 115                         /* It already vanished from the working tree */
 116                         continue;
 117                 }
 118                 else if (S_ISDIR(st.st_mode)) {
 119                         /* if a file was removed and it is now a
 120                          * directory, that is the same as ENOENT as
 121                          * far as git is concerned; we do not track
 122                          * directories unless they are submodules.
 123                          */
 124                         if (!S_ISGITLINK(ce->ce_mode))
 125                                 continue;
 126                 }
 127
 128                 /*
 129                  * "rm" of a path that has changes need to be treated
 130                  * carefully not to allow losing local changes
 131                  * accidentally.  A local change could be (1) file in
 132                  * work tree is different since the index; and/or (2)
 133                  * the user staged a content that is different from
 134                  * the current commit in the index.
 135                  *
 136                  * In such a case, you would need to --force the
 137                  * removal.  However, "rm --cached" (remove only from
 138                  * the index) is safe if the index matches the file in
 139                  * the work tree or the HEAD commit, as it means that
 140                  * the content being removed is available elsewhere.
 141                  */
 142
 143                 /*
 144                  * Is the index different from the file in the work tree?
 145                  * If it's a submodule, is its work tree modified?
 146                  */
 147                 if (ce_match_stat(ce, &st, 0) ||
 148                     (S_ISGITLINK(ce->ce_mode) &&
 149                      !ok_to_remove_submodule(ce->name)))
 150                         local_changes = 1;
 151
 152                 /*
 153                  * Is the index different from the HEAD commit?  By
 154                  * definition, before the very initial commit,
 155                  * anything staged in the index is treated by the same
 156                  * way as changed from the HEAD.
 157                  */
 158                 if (no_head
 159                      || get_tree_entry(head, name, sha1, &mode)
 160                      || ce->ce_mode != create_ce_mode(mode)
 161                      || hashcmp(ce->sha1, sha1))
 162                         staged_changes = 1;
 163
 164                 /*
 165                  * If the index does not match the file in the work
 166                  * tree and if it does not match the HEAD commit
 167                  * either, (1) "git rm" without --cached definitely
 168                  * will lose information; (2) "git rm --cached" will
 169                  * lose information unless it is about removing an
 170                  * "intent to add" entry.
 171                  */
 172                 if (local_changes && staged_changes) {
 173                         if (!index_only || !(ce->ce_flags & CE_INTENT_TO_ADD))
 174                                 errs = error(_("'%s' has staged content different "
 175                                              "from both the file and the HEAD\n"
 176                                              "(use -f to force removal)"), name);
 177                 }
 178                 else if (!index_only) {
 179                         if (staged_changes)
 180                                 errs = error(_("'%s' has changes staged in the index\n"
 181                                              "(use --cached to keep the file, "
 182                                              "or -f to force removal)"), name);
 183                         if (local_changes) {
 184                                 if (S_ISGITLINK(ce->ce_mode) &&
 185                                     !submodule_uses_gitfile(name)) {
 186                                         errs = error(_("submodule '%s' (or one of its nested "
 187                                                      "submodules) uses a .git directory\n"
 188                                                      "(use 'rm -rf' if you really want to remove "
 189                                                      "it including all of its history)"), name);
 190                                 } else
 191                                         errs = error(_("'%s' has local modifications\n"
 192                                                      "(use --cached to keep the file, "
 193                                                      "or -f to force removal)"), name);
 194                         }
 195                 }
 196         }
 197         return errs;
 198 }
 199
 200 static struct lock_file lock_file;
 201
 202 static int show_only = 0, force = 0, index_only = 0, recursive = 0, quiet = 0;
 203 static int ignore_unmatch = 0;
 204
 205 static struct option builtin_rm_options[] = {
 206         OPT__DRY_RUN(&show_only, N_("dry run")),
 207         OPT__QUIET(&quiet, N_("do not list removed files")),
 208         OPT_BOOLEAN( 0 , "cached",         &index_only, N_("only remove from the index")),
 209         OPT__FORCE(&force, N_("override the up-to-date check")),
 210         OPT_BOOLEAN('r', NULL,             &recursive,  N_("allow recursive removal")),
 211         OPT_BOOLEAN( 0 , "ignore-unmatch", &ignore_unmatch,
 212                                 N_("exit with a zero status even if nothing matched")),
 213         OPT_END(),
 214 };
 215
 216 int cmd_rm(int argc, const char **argv, const char *prefix)
 217 {
 218         int i, newfd;
 219         const char **pathspec;
 220         char *seen;
 221
 222         git_config(git_default_config, NULL);
 223
 224         argc = parse_options(argc, argv, prefix, builtin_rm_options,
 225                              builtin_rm_usage, 0);
 226         if (!argc)
 227                 usage_with_options(builtin_rm_usage, builtin_rm_options);
 228
 229         if (!index_only)
 230                 setup_work_tree();
 231
 232         newfd = hold_locked_index(&lock_file, 1);
 233
 234         if (read_cache() < 0)
 235                 die(_("index file corrupt"));
 236
 237         /*
 238          * Drop trailing directory separators from directories so we'll find
 239          * submodules in the index.
 240          */
 241         for (i = 0; i < argc; i++) {
 242                 size_t pathlen = strlen(argv[i]);
 243                 if (pathlen && is_dir_sep(argv[i][pathlen - 1]) &&
 244                     is_directory(argv[i])) {
 245                         do {
 246                                 pathlen--;
 247                         } while (pathlen && is_dir_sep(argv[i][pathlen - 1]));
 248                         argv[i] = xmemdupz(argv[i], pathlen);
 249                 }
 250         }
 251
 252         pathspec = get_pathspec(prefix, argv);
 253         refresh_index(&the_index, REFRESH_QUIET, pathspec, NULL, NULL);
 254
 255         seen = NULL;
 256         for (i = 0; pathspec[i] ; i++)
 257                 /* nothing */;
 258         seen = xcalloc(i, 1);
 259
 260         for (i = 0; i < active_nr; i++) {
 261                 struct cache_entry *ce = active_cache[i];
 262                 if (!match_pathspec(pathspec, ce->name, ce_namelen(ce), 0, seen))
 263                         continue;
 264                 ALLOC_GROW(list.entry, list.nr + 1, list.alloc);
 265                 list.entry[list.nr].name = ce->name;
 266                 list.entry[list.nr++].is_submodule = S_ISGITLINK(ce->ce_mode);
 267         }
 268
 269         if (pathspec) {
 270                 const char *match;
 271                 int seen_any = 0;
 272                 for (i = 0; (match = pathspec[i]) != NULL ; i++) {
 273                         if (!seen[i]) {
 274                                 if (!ignore_unmatch) {
 275                                         die(_("pathspec '%s' did not match any files"),
 276                                             match);
 277                                 }
 278                         }
 279                         else {
 280                                 seen_any = 1;
 281                         }
 282                         if (!recursive && seen[i] == MATCHED_RECURSIVELY)
 283                                 die(_("not removing '%s' recursively without -r"),
 284                                     *match ? match : ".");
 285                 }
 286
 287                 if (! seen_any)
 288                         exit(0);
 289         }
 290
 291         /*
 292          * If not forced, the file, the index and the HEAD (if exists)
 293          * must match; but the file can already been removed, since
 294          * this sequence is a natural "novice" way:
 295          *
 296          *      rm F; git rm F
 297          *
 298          * Further, if HEAD commit exists, "diff-index --cached" must
 299          * report no changes unless forced.
 300          */
 301         if (!force) {
 302                 unsigned char sha1[20];
 303                 if (get_sha1("HEAD", sha1))
 304                         hashclr(sha1);
 305                 if (check_local_mod(sha1, index_only))
 306                         exit(1);
 307         } else if (!index_only) {
 308                 if (check_submodules_use_gitfiles())
 309                         exit(1);
 310         }
 311
 312         /*
 313          * First remove the names from the index: we won't commit
 314          * the index unless all of them succeed.
 315          */
 316         for (i = 0; i < list.nr; i++) {
 317                 const char *path = list.entry[i].name;
 318                 if (!quiet)
 319                         printf("rm '%s'\n", path);
 320
 321                 if (remove_file_from_cache(path))
 322                         die(_("git rm: unable to remove %s"), path);
 323         }
 324
 325         if (show_only)
 326                 return 0;
 327
 328         /*
 329          * Then, unless we used "--cached", remove the filenames from
 330          * the workspace. If we fail to remove the first one, we
 331          * abort the "git rm" (but once we've successfully removed
 332          * any file at all, we'll go ahead and commit to it all:
 333          * by then we've already committed ourselves and can't fail
 334          * in the middle)
 335          */
 336         if (!index_only) {
 337                 int removed = 0;
 338                 for (i = 0; i < list.nr; i++) {
 339                         const char *path = list.entry[i].name;
 340                         if (list.entry[i].is_submodule) {
 341                                 if (is_empty_dir(path)) {
 342                                         if (!rmdir(path)) {
 343                                                 removed = 1;
 344                                                 continue;
 345                                         }
 346                                 } else {
 347                                         struct strbuf buf = STRBUF_INIT;
 348                                         strbuf_addstr(&buf, path);
 349                                         if (!remove_dir_recursively(&buf, 0)) {
 350                                                 removed = 1;
 351                                                 strbuf_release(&buf);
 352                                                 continue;
 353                                         }
 354                                         strbuf_release(&buf);
 355                                         /* Fallthrough and let remove_path() fail. */
 356                                 }
 357                         }
 358                         if (!remove_path(path)) {
 359                                 removed = 1;
 360                                 continue;
 361                         }
 362                         if (!removed)
 363                                 die_errno("git rm: '%s'", path);
 364                 }
 365         }
 366
 367         if (active_cache_changed) {
 368                 if (write_cache(newfd, active_cache, active_nr) ||
 369                     commit_locked_index(&lock_file))
 370                         die(_("Unable to write new index file"));
 371         }
 372
 373         return 0;
 374 }