entry.c

   1 #include "git-compat-util.h"
   2 #include "blob.h"
   3 #include "object-store.h"
   4 #include "dir.h"
   5 #include "environment.h"
   6 #include "gettext.h"
   7 #include "hex.h"
   8 #include "name-hash.h"
   9 #include "sparse-index.h"
  10 #include "streaming.h"
  11 #include "submodule.h"
  12 #include "symlinks.h"
  13 #include "progress.h"
  14 #include "fsmonitor.h"
  15 #include "entry.h"
  16 #include "parallel-checkout.h"
  17 #include "wrapper.h"
  18
  19 static void create_directories(const char *path, int path_len,
  20                                const struct checkout *state)
  21 {
  22         char *buf = xmallocz(path_len);
  23         int len = 0;
  24
  25         while (len < path_len) {
  26                 do {
  27                         buf[len] = path[len];
  28                         len++;
  29                 } while (len < path_len && path[len] != '/');
  30                 if (len >= path_len)
  31                         break;
  32                 buf[len] = 0;
  33
  34                 /*
  35                  * For 'checkout-index --prefix=<dir>', <dir> is
  36                  * allowed to be a symlink to an existing directory,
  37                  * and we set 'state->base_dir_len' below, such that
  38                  * we test the path components of the prefix with the
  39                  * stat() function instead of the lstat() function.
  40                  */
  41                 if (has_dirs_only_path(buf, len, state->base_dir_len))
  42                         continue; /* ok, it is already a directory. */
  43
  44                 /*
  45                  * If this mkdir() would fail, it could be that there
  46                  * is already a symlink or something else exists
  47                  * there, therefore we then try to unlink it and try
  48                  * one more time to create the directory.
  49                  */
  50                 if (mkdir(buf, 0777)) {
  51                         if (errno == EEXIST && state->force &&
  52                             !unlink_or_warn(buf) && !mkdir(buf, 0777))
  53                                 continue;
  54                         die_errno("cannot create directory at '%s'", buf);
  55                 }
  56         }
  57         free(buf);
  58 }
  59
  60 static void remove_subtree(struct strbuf *path)
  61 {
  62         DIR *dir = opendir(path->buf);
  63         struct dirent *de;
  64         int origlen = path->len;
  65
  66         if (!dir)
  67                 die_errno("cannot opendir '%s'", path->buf);
  68         while ((de = readdir_skip_dot_and_dotdot(dir)) != NULL) {
  69                 struct stat st;
  70
  71                 strbuf_addch(path, '/');
  72                 strbuf_addstr(path, de->d_name);
  73                 if (lstat(path->buf, &st))
  74                         die_errno("cannot lstat '%s'", path->buf);
  75                 if (S_ISDIR(st.st_mode))
  76                         remove_subtree(path);
  77                 else if (unlink(path->buf))
  78                         die_errno("cannot unlink '%s'", path->buf);
  79                 strbuf_setlen(path, origlen);
  80         }
  81         closedir(dir);
  82         if (rmdir(path->buf))
  83                 die_errno("cannot rmdir '%s'", path->buf);
  84 }
  85
  86 static int create_file(const char *path, unsigned int mode)
  87 {
  88         mode = (mode & 0100) ? 0777 : 0666;
  89         return open(path, O_WRONLY | O_CREAT | O_EXCL, mode);
  90 }
  91
  92 void *read_blob_entry(const struct cache_entry *ce, size_t *size)
  93 {
  94         enum object_type type;
  95         unsigned long ul;
  96         void *blob_data = repo_read_object_file(the_repository, &ce->oid,
  97                                                 &type, &ul);
  98
  99         *size = ul;
 100         if (blob_data) {
 101                 if (type == OBJ_BLOB)
 102                         return blob_data;
 103                 free(blob_data);
 104         }
 105         return NULL;
 106 }
 107
 108 static int open_output_fd(char *path, const struct cache_entry *ce, int to_tempfile)
 109 {
 110         int symlink = (ce->ce_mode & S_IFMT) != S_IFREG;
 111         if (to_tempfile) {
 112                 xsnprintf(path, TEMPORARY_FILENAME_LENGTH, "%s",
 113                           symlink ? ".merge_link_XXXXXX" : ".merge_file_XXXXXX");
 114                 return mkstemp(path);
 115         } else {
 116                 return create_file(path, !symlink ? ce->ce_mode : 0666);
 117         }
 118 }
 119
 120 int fstat_checkout_output(int fd, const struct checkout *state, struct stat *st)
 121 {
 122         /* use fstat() only when path == ce->name */
 123         if (fstat_is_reliable() &&
 124             state->refresh_cache && !state->base_dir_len) {
 125                 return !fstat(fd, st);
 126         }
 127         return 0;
 128 }
 129
 130 static int streaming_write_entry(const struct cache_entry *ce, char *path,
 131                                  struct stream_filter *filter,
 132                                  const struct checkout *state, int to_tempfile,
 133                                  int *fstat_done, struct stat *statbuf)
 134 {
 135         int result = 0;
 136         int fd;
 137
 138         fd = open_output_fd(path, ce, to_tempfile);
 139         if (fd < 0)
 140                 return -1;
 141
 142         result |= stream_blob_to_fd(fd, &ce->oid, filter, 1);
 143         *fstat_done = fstat_checkout_output(fd, state, statbuf);
 144         result |= close(fd);
 145
 146         if (result)
 147                 unlink(path);
 148         return result;
 149 }
 150
 151 void enable_delayed_checkout(struct checkout *state)
 152 {
 153         if (!state->delayed_checkout) {
 154                 state->delayed_checkout = xmalloc(sizeof(*state->delayed_checkout));
 155                 state->delayed_checkout->state = CE_CAN_DELAY;
 156                 string_list_init_nodup(&state->delayed_checkout->filters);
 157                 string_list_init_nodup(&state->delayed_checkout->paths);
 158         }
 159 }
 160
 161 static int remove_available_paths(struct string_list_item *item, void *cb_data)
 162 {
 163         struct string_list *available_paths = cb_data;
 164         struct string_list_item *available;
 165
 166         available = string_list_lookup(available_paths, item->string);
 167         if (available)
 168                 available->util = item->util;
 169         return !available;
 170 }
 171
 172 int finish_delayed_checkout(struct checkout *state, int show_progress)
 173 {
 174         int errs = 0;
 175         unsigned processed_paths = 0;
 176         off_t filtered_bytes = 0;
 177         struct string_list_item *filter, *path;
 178         struct progress *progress = NULL;
 179         struct delayed_checkout *dco = state->delayed_checkout;
 180
 181         if (!state->delayed_checkout)
 182                 return errs;
 183
 184         dco->state = CE_RETRY;
 185         if (show_progress)
 186                 progress = start_delayed_progress(_("Filtering content"), dco->paths.nr);
 187         while (dco->filters.nr > 0) {
 188                 for_each_string_list_item(filter, &dco->filters) {
 189                         struct string_list available_paths = STRING_LIST_INIT_NODUP;
 190
 191                         if (!async_query_available_blobs(filter->string, &available_paths)) {
 192                                 /* Filter reported an error */
 193                                 errs = 1;
 194                                 filter->string = "";
 195                                 continue;
 196                         }
 197                         if (available_paths.nr <= 0) {
 198                                 /*
 199                                  * Filter responded with no entries. That means
 200                                  * the filter is done and we can remove the
 201                                  * filter from the list (see
 202                                  * "string_list_remove_empty_items" call below).
 203                                  */
 204                                 filter->string = "";
 205                                 continue;
 206                         }
 207
 208                         /*
 209                          * In dco->paths we store a list of all delayed paths.
 210                          * The filter just send us a list of available paths.
 211                          * Remove them from the list.
 212                          */
 213                         filter_string_list(&dco->paths, 0,
 214                                 &remove_available_paths, &available_paths);
 215
 216                         for_each_string_list_item(path, &available_paths) {
 217                                 struct cache_entry* ce;
 218
 219                                 if (!path->util) {
 220                                         error("external filter '%s' signaled that '%s' "
 221                                               "is now available although it has not been "
 222                                               "delayed earlier",
 223                                               filter->string, path->string);
 224                                         errs |= 1;
 225
 226                                         /*
 227                                          * Do not ask the filter for available blobs,
 228                                          * again, as the filter is likely buggy.
 229                                          */
 230                                         filter->string = "";
 231                                         continue;
 232                                 }
 233                                 ce = index_file_exists(state->istate, path->string,
 234                                                        strlen(path->string), 0);
 235                                 if (ce) {
 236                                         display_progress(progress, ++processed_paths);
 237                                         errs |= checkout_entry(ce, state, NULL, path->util);
 238                                         filtered_bytes += ce->ce_stat_data.sd_size;
 239                                         display_throughput(progress, filtered_bytes);
 240                                 } else
 241                                         errs = 1;
 242                         }
 243                 }
 244                 string_list_remove_empty_items(&dco->filters, 0);
 245         }
 246         stop_progress(&progress);
 247         string_list_clear(&dco->filters, 0);
 248
 249         /* At this point we should not have any delayed paths anymore. */
 250         errs |= dco->paths.nr;
 251         for_each_string_list_item(path, &dco->paths) {
 252                 error("'%s' was not filtered properly", path->string);
 253         }
 254         string_list_clear(&dco->paths, 0);
 255
 256         free(dco);
 257         state->delayed_checkout = NULL;
 258
 259         return errs;
 260 }
 261
 262 void update_ce_after_write(const struct checkout *state, struct cache_entry *ce,
 263                            struct stat *st)
 264 {
 265         if (state->refresh_cache) {
 266                 assert(state->istate);
 267                 fill_stat_cache_info(state->istate, ce, st);
 268                 ce->ce_flags |= CE_UPDATE_IN_BASE;
 269                 mark_fsmonitor_invalid(state->istate, ce);
 270                 state->istate->cache_changed |= CE_ENTRY_CHANGED;
 271         }
 272 }
 273
 274 /* Note: ca is used (and required) iff the entry refers to a regular file. */
 275 static int write_entry(struct cache_entry *ce, char *path, struct conv_attrs *ca,
 276                        const struct checkout *state, int to_tempfile,
 277                        int *nr_checkouts)
 278 {
 279         unsigned int ce_mode_s_ifmt = ce->ce_mode & S_IFMT;
 280         struct delayed_checkout *dco = state->delayed_checkout;
 281         int fd, ret, fstat_done = 0;
 282         char *new_blob;
 283         struct strbuf buf = STRBUF_INIT;
 284         size_t size;
 285         ssize_t wrote;
 286         size_t newsize = 0;
 287         struct stat st;
 288         const struct submodule *sub;
 289         struct checkout_metadata meta;
 290         static int scratch_nr_checkouts;
 291
 292         clone_checkout_metadata(&meta, &state->meta, &ce->oid);
 293
 294         if (ce_mode_s_ifmt == S_IFREG) {
 295                 struct stream_filter *filter = get_stream_filter_ca(ca, &ce->oid);
 296                 if (filter &&
 297                     !streaming_write_entry(ce, path, filter,
 298                                            state, to_tempfile,
 299                                            &fstat_done, &st))
 300                         goto finish;
 301         }
 302
 303         switch (ce_mode_s_ifmt) {
 304         case S_IFLNK:
 305                 new_blob = read_blob_entry(ce, &size);
 306                 if (!new_blob)
 307                         return error("unable to read sha1 file of %s (%s)",
 308                                      ce->name, oid_to_hex(&ce->oid));
 309
 310                 /*
 311                  * We can't make a real symlink; write out a regular file entry
 312                  * with the symlink destination as its contents.
 313                  */
 314                 if (!has_symlinks || to_tempfile)
 315                         goto write_file_entry;
 316
 317                 ret = symlink(new_blob, path);
 318                 free(new_blob);
 319                 if (ret)
 320                         return error_errno("unable to create symlink %s", path);
 321                 break;
 322
 323         case S_IFREG:
 324                 /*
 325                  * We do not send the blob in case of a retry, so do not
 326                  * bother reading it at all.
 327                  */
 328                 if (dco && dco->state == CE_RETRY) {
 329                         new_blob = NULL;
 330                         size = 0;
 331                 } else {
 332                         new_blob = read_blob_entry(ce, &size);
 333                         if (!new_blob)
 334                                 return error("unable to read sha1 file of %s (%s)",
 335                                              ce->name, oid_to_hex(&ce->oid));
 336                 }
 337
 338                 /*
 339                  * Convert from git internal format to working tree format
 340                  */
 341                 if (dco && dco->state != CE_NO_DELAY) {
 342                         ret = async_convert_to_working_tree_ca(ca, ce->name,
 343                                                                new_blob, size,
 344                                                                &buf, &meta, dco);
 345                         if (ret) {
 346                                 struct string_list_item *item =
 347                                         string_list_lookup(&dco->paths, ce->name);
 348                                 if (item) {
 349                                         item->util = nr_checkouts ? nr_checkouts
 350                                                         : &scratch_nr_checkouts;
 351                                         free(new_blob);
 352                                         goto delayed;
 353                                 }
 354                         }
 355                 } else {
 356                         ret = convert_to_working_tree_ca(ca, ce->name, new_blob,
 357                                                          size, &buf, &meta);
 358                 }
 359
 360                 if (ret) {
 361                         free(new_blob);
 362                         new_blob = strbuf_detach(&buf, &newsize);
 363                         size = newsize;
 364                 }
 365                 /*
 366                  * No "else" here as errors from convert are OK at this
 367                  * point. If the error would have been fatal (e.g.
 368                  * filter is required), then we would have died already.
 369                  */
 370
 371         write_file_entry:
 372                 fd = open_output_fd(path, ce, to_tempfile);
 373                 if (fd < 0) {
 374                         free(new_blob);
 375                         return error_errno("unable to create file %s", path);
 376                 }
 377
 378                 wrote = write_in_full(fd, new_blob, size);
 379                 if (!to_tempfile)
 380                         fstat_done = fstat_checkout_output(fd, state, &st);
 381                 close(fd);
 382                 free(new_blob);
 383                 if (wrote < 0)
 384                         return error("unable to write file %s", path);
 385                 break;
 386
 387         case S_IFGITLINK:
 388                 if (to_tempfile)
 389                         return error("cannot create temporary submodule %s", ce->name);
 390                 if (mkdir(path, 0777) < 0)
 391                         return error("cannot create submodule directory %s", path);
 392                 sub = submodule_from_ce(ce);
 393                 if (sub)
 394                         return submodule_move_head(ce->name, state->super_prefix,
 395                                 NULL, oid_to_hex(&ce->oid),
 396                                 state->force ? SUBMODULE_MOVE_HEAD_FORCE : 0);
 397                 break;
 398
 399         default:
 400                 return error("unknown file mode for %s in index", ce->name);
 401         }
 402
 403 finish:
 404         if (state->refresh_cache) {
 405                 if (!fstat_done && lstat(ce->name, &st) < 0)
 406                         return error_errno("unable to stat just-written file %s",
 407                                            ce->name);
 408                 update_ce_after_write(state, ce , &st);
 409         }
 410         if (nr_checkouts)
 411                 (*nr_checkouts)++;
 412 delayed:
 413         return 0;
 414 }
 415
 416 /*
 417  * This is like 'lstat()', except it refuses to follow symlinks
 418  * in the path, after skipping "skiplen".
 419  */
 420 static int check_path(const char *path, int len, struct stat *st, int skiplen)
 421 {
 422         const char *slash = path + len;
 423
 424         while (path < slash && *slash != '/')
 425                 slash--;
 426         if (!has_dirs_only_path(path, slash - path, skiplen)) {
 427                 errno = ENOENT;
 428                 return -1;
 429         }
 430         return lstat(path, st);
 431 }
 432
 433 static void mark_colliding_entries(const struct checkout *state,
 434                                    struct cache_entry *ce, struct stat *st)
 435 {
 436         int i, trust_ino = check_stat;
 437
 438 #if defined(GIT_WINDOWS_NATIVE) || defined(__CYGWIN__)
 439         trust_ino = 0;
 440 #endif
 441
 442         ce->ce_flags |= CE_MATCHED;
 443
 444         /* TODO: audit for interaction with sparse-index. */
 445         ensure_full_index(state->istate);
 446         for (i = 0; i < state->istate->cache_nr; i++) {
 447                 struct cache_entry *dup = state->istate->cache[i];
 448
 449                 if (dup == ce) {
 450                         /*
 451                          * Parallel checkout doesn't create the files in index
 452                          * order. So the other side of the collision may appear
 453                          * after the given cache_entry in the array.
 454                          */
 455                         if (parallel_checkout_status() == PC_RUNNING)
 456                                 continue;
 457                         else
 458                                 break;
 459                 }
 460
 461                 if (dup->ce_flags & (CE_MATCHED | CE_VALID | CE_SKIP_WORKTREE))
 462                         continue;
 463
 464                 if ((trust_ino && !match_stat_data(&dup->ce_stat_data, st)) ||
 465                     (!trust_ino && !fspathcmp(ce->name, dup->name))) {
 466                         dup->ce_flags |= CE_MATCHED;
 467                         break;
 468                 }
 469         }
 470 }
 471
 472 int checkout_entry_ca(struct cache_entry *ce, struct conv_attrs *ca,
 473                       const struct checkout *state, char *topath,
 474                       int *nr_checkouts)
 475 {
 476         static struct strbuf path = STRBUF_INIT;
 477         struct stat st;
 478         struct conv_attrs ca_buf;
 479
 480         if (ce->ce_flags & CE_WT_REMOVE) {
 481                 if (topath)
 482                         /*
 483                          * No content and thus no path to create, so we have
 484                          * no pathname to return.
 485                          */
 486                         BUG("Can't remove entry to a path");
 487                 unlink_entry(ce, state->super_prefix);
 488                 return 0;
 489         }
 490
 491         if (topath) {
 492                 if (S_ISREG(ce->ce_mode) && !ca) {
 493                         convert_attrs(state->istate, &ca_buf, ce->name);
 494                         ca = &ca_buf;
 495                 }
 496                 return write_entry(ce, topath, ca, state, 1, nr_checkouts);
 497         }
 498
 499         strbuf_reset(&path);
 500         strbuf_add(&path, state->base_dir, state->base_dir_len);
 501         strbuf_add(&path, ce->name, ce_namelen(ce));
 502
 503         if (!check_path(path.buf, path.len, &st, state->base_dir_len)) {
 504                 const struct submodule *sub;
 505                 unsigned changed = ie_match_stat(state->istate, ce, &st,
 506                                                  CE_MATCH_IGNORE_VALID | CE_MATCH_IGNORE_SKIP_WORKTREE);
 507                 /*
 508                  * Needs to be checked before !changed returns early,
 509                  * as the possibly empty directory was not changed
 510                  */
 511                 sub = submodule_from_ce(ce);
 512                 if (sub) {
 513                         int err;
 514                         if (!is_submodule_populated_gently(ce->name, &err)) {
 515                                 struct stat sb;
 516                                 if (lstat(ce->name, &sb))
 517                                         die(_("could not stat file '%s'"), ce->name);
 518                                 if (!(st.st_mode & S_IFDIR))
 519                                         unlink_or_warn(ce->name);
 520
 521                                 return submodule_move_head(ce->name, state->super_prefix,
 522                                         NULL, oid_to_hex(&ce->oid), 0);
 523                         } else
 524                                 return submodule_move_head(ce->name, state->super_prefix,
 525                                         "HEAD", oid_to_hex(&ce->oid),
 526                                         state->force ? SUBMODULE_MOVE_HEAD_FORCE : 0);
 527                 }
 528
 529                 if (!changed)
 530                         return 0;
 531                 if (!state->force) {
 532                         if (!state->quiet)
 533                                 fprintf(stderr,
 534                                         "%s already exists, no checkout\n",
 535                                         path.buf);
 536                         return -1;
 537                 }
 538
 539                 if (state->clone)
 540                         mark_colliding_entries(state, ce, &st);
 541
 542                 /*
 543                  * We unlink the old file, to get the new one with the
 544                  * right permissions (including umask, which is nasty
 545                  * to emulate by hand - much easier to let the system
 546                  * just do the right thing)
 547                  */
 548                 if (S_ISDIR(st.st_mode)) {
 549                         /* If it is a gitlink, leave it alone! */
 550                         if (S_ISGITLINK(ce->ce_mode))
 551                                 return 0;
 552                         remove_subtree(&path);
 553                 } else if (unlink(path.buf))
 554                         return error_errno("unable to unlink old '%s'", path.buf);
 555         } else if (state->not_new)
 556                 return 0;
 557
 558         create_directories(path.buf, path.len, state);
 559
 560         if (S_ISREG(ce->ce_mode) && !ca) {
 561                 convert_attrs(state->istate, &ca_buf, ce->name);
 562                 ca = &ca_buf;
 563         }
 564
 565         if (!enqueue_checkout(ce, ca, nr_checkouts))
 566                 return 0;
 567
 568         return write_entry(ce, path.buf, ca, state, 0, nr_checkouts);
 569 }
 570
 571 void unlink_entry(const struct cache_entry *ce, const char *super_prefix)
 572 {
 573         const struct submodule *sub = submodule_from_ce(ce);
 574         if (sub) {
 575                 /* state.force is set at the caller. */
 576                 submodule_move_head(ce->name, super_prefix, "HEAD", NULL,
 577                                     SUBMODULE_MOVE_HEAD_FORCE);
 578         }
 579         if (check_leading_path(ce->name, ce_namelen(ce), 1) >= 0)
 580                 return;
 581         if (remove_or_warn(ce->ce_mode, ce->name))
 582                 return;
 583         schedule_dir_for_removal(ce->name, ce_namelen(ce));
 584 }