entry.c

   1 #include "git-compat-util.h"
   2 #include "object-store-ll.h"
   3 #include "dir.h"
   4 #include "environment.h"
   5 #include "gettext.h"
   6 #include "hex.h"
   7 #include "name-hash.h"
   8 #include "sparse-index.h"
   9 #include "streaming.h"
  10 #include "submodule.h"
  11 #include "symlinks.h"
  12 #include "progress.h"
  13 #include "fsmonitor.h"
  14 #include "entry.h"
  15 #include "parallel-checkout.h"
  16
  17 static void create_directories(const char *path, int path_len,
  18                                const struct checkout *state)
  19 {
  20         char *buf = xmallocz(path_len);
  21         int len = 0;
  22
  23         while (len < path_len) {
  24                 do {
  25                         buf[len] = path[len];
  26                         len++;
  27                 } while (len < path_len && path[len] != '/');
  28                 if (len >= path_len)
  29                         break;
  30                 buf[len] = 0;
  31
  32                 /*
  33                  * For 'checkout-index --prefix=<dir>', <dir> is
  34                  * allowed to be a symlink to an existing directory,
  35                  * and we set 'state->base_dir_len' below, such that
  36                  * we test the path components of the prefix with the
  37                  * stat() function instead of the lstat() function.
  38                  */
  39                 if (has_dirs_only_path(buf, len, state->base_dir_len))
  40                         continue; /* ok, it is already a directory. */
  41
  42                 /*
  43                  * If this mkdir() would fail, it could be that there
  44                  * is already a symlink or something else exists
  45                  * there, therefore we then try to unlink it and try
  46                  * one more time to create the directory.
  47                  */
  48                 if (mkdir(buf, 0777)) {
  49                         if (errno == EEXIST && state->force &&
  50                             !unlink_or_warn(buf) && !mkdir(buf, 0777))
  51                                 continue;
  52                         die_errno("cannot create directory at '%s'", buf);
  53                 }
  54         }
  55         free(buf);
  56 }
  57
  58 static void remove_subtree(struct strbuf *path)
  59 {
  60         DIR *dir = opendir(path->buf);
  61         struct dirent *de;
  62         int origlen = path->len;
  63
  64         if (!dir)
  65                 die_errno("cannot opendir '%s'", path->buf);
  66         while ((de = readdir_skip_dot_and_dotdot(dir)) != NULL) {
  67                 struct stat st;
  68
  69                 strbuf_addch(path, '/');
  70                 strbuf_addstr(path, de->d_name);
  71                 if (lstat(path->buf, &st))
  72                         die_errno("cannot lstat '%s'", path->buf);
  73                 if (S_ISDIR(st.st_mode))
  74                         remove_subtree(path);
  75                 else if (unlink(path->buf))
  76                         die_errno("cannot unlink '%s'", path->buf);
  77                 strbuf_setlen(path, origlen);
  78         }
  79         closedir(dir);
  80         if (rmdir(path->buf))
  81                 die_errno("cannot rmdir '%s'", path->buf);
  82 }
  83
  84 static int create_file(const char *path, unsigned int mode)
  85 {
  86         mode = (mode & 0100) ? 0777 : 0666;
  87         return open(path, O_WRONLY | O_CREAT | O_EXCL, mode);
  88 }
  89
  90 void *read_blob_entry(const struct cache_entry *ce, size_t *size)
  91 {
  92         enum object_type type;
  93         unsigned long ul;
  94         void *blob_data = repo_read_object_file(the_repository, &ce->oid,
  95                                                 &type, &ul);
  96
  97         *size = ul;
  98         if (blob_data) {
  99                 if (type == OBJ_BLOB)
 100                         return blob_data;
 101                 free(blob_data);
 102         }
 103         return NULL;
 104 }
 105
 106 static int open_output_fd(char *path, const struct cache_entry *ce, int to_tempfile)
 107 {
 108         int symlink = (ce->ce_mode & S_IFMT) != S_IFREG;
 109         if (to_tempfile) {
 110                 xsnprintf(path, TEMPORARY_FILENAME_LENGTH, "%s",
 111                           symlink ? ".merge_link_XXXXXX" : ".merge_file_XXXXXX");
 112                 return mkstemp(path);
 113         } else {
 114                 return create_file(path, !symlink ? ce->ce_mode : 0666);
 115         }
 116 }
 117
 118 int fstat_checkout_output(int fd, const struct checkout *state, struct stat *st)
 119 {
 120         /* use fstat() only when path == ce->name */
 121         if (fstat_is_reliable() &&
 122             state->refresh_cache && !state->base_dir_len) {
 123                 return !fstat(fd, st);
 124         }
 125         return 0;
 126 }
 127
 128 static int streaming_write_entry(const struct cache_entry *ce, char *path,
 129                                  struct stream_filter *filter,
 130                                  const struct checkout *state, int to_tempfile,
 131                                  int *fstat_done, struct stat *statbuf)
 132 {
 133         int result = 0;
 134         int fd;
 135
 136         fd = open_output_fd(path, ce, to_tempfile);
 137         if (fd < 0)
 138                 return -1;
 139
 140         result |= stream_blob_to_fd(fd, &ce->oid, filter, 1);
 141         *fstat_done = fstat_checkout_output(fd, state, statbuf);
 142         result |= close(fd);
 143
 144         if (result)
 145                 unlink(path);
 146         return result;
 147 }
 148
 149 void enable_delayed_checkout(struct checkout *state)
 150 {
 151         if (!state->delayed_checkout) {
 152                 state->delayed_checkout = xmalloc(sizeof(*state->delayed_checkout));
 153                 state->delayed_checkout->state = CE_CAN_DELAY;
 154                 string_list_init_nodup(&state->delayed_checkout->filters);
 155                 string_list_init_nodup(&state->delayed_checkout->paths);
 156         }
 157 }
 158
 159 static int remove_available_paths(struct string_list_item *item, void *cb_data)
 160 {
 161         struct string_list *available_paths = cb_data;
 162         struct string_list_item *available;
 163
 164         available = string_list_lookup(available_paths, item->string);
 165         if (available)
 166                 available->util = item->util;
 167         return !available;
 168 }
 169
 170 int finish_delayed_checkout(struct checkout *state, int show_progress)
 171 {
 172         int errs = 0;
 173         unsigned processed_paths = 0;
 174         off_t filtered_bytes = 0;
 175         struct string_list_item *filter, *path;
 176         struct progress *progress = NULL;
 177         struct delayed_checkout *dco = state->delayed_checkout;
 178
 179         if (!state->delayed_checkout)
 180                 return errs;
 181
 182         dco->state = CE_RETRY;
 183         if (show_progress)
 184                 progress = start_delayed_progress(_("Filtering content"), dco->paths.nr);
 185         while (dco->filters.nr > 0) {
 186                 for_each_string_list_item(filter, &dco->filters) {
 187                         struct string_list available_paths = STRING_LIST_INIT_NODUP;
 188
 189                         if (!async_query_available_blobs(filter->string, &available_paths)) {
 190                                 /* Filter reported an error */
 191                                 errs = 1;
 192                                 filter->string = "";
 193                                 continue;
 194                         }
 195                         if (available_paths.nr <= 0) {
 196                                 /*
 197                                  * Filter responded with no entries. That means
 198                                  * the filter is done and we can remove the
 199                                  * filter from the list (see
 200                                  * "string_list_remove_empty_items" call below).
 201                                  */
 202                                 filter->string = "";
 203                                 continue;
 204                         }
 205
 206                         /*
 207                          * In dco->paths we store a list of all delayed paths.
 208                          * The filter just send us a list of available paths.
 209                          * Remove them from the list.
 210                          */
 211                         filter_string_list(&dco->paths, 0,
 212                                 &remove_available_paths, &available_paths);
 213
 214                         for_each_string_list_item(path, &available_paths) {
 215                                 struct cache_entry* ce;
 216
 217                                 if (!path->util) {
 218                                         error("external filter '%s' signaled that '%s' "
 219                                               "is now available although it has not been "
 220                                               "delayed earlier",
 221                                               filter->string, path->string);
 222                                         errs |= 1;
 223
 224                                         /*
 225                                          * Do not ask the filter for available blobs,
 226                                          * again, as the filter is likely buggy.
 227                                          */
 228                                         filter->string = "";
 229                                         continue;
 230                                 }
 231                                 ce = index_file_exists(state->istate, path->string,
 232                                                        strlen(path->string), 0);
 233                                 if (ce) {
 234                                         display_progress(progress, ++processed_paths);
 235                                         errs |= checkout_entry(ce, state, NULL, path->util);
 236                                         filtered_bytes += ce->ce_stat_data.sd_size;
 237                                         display_throughput(progress, filtered_bytes);
 238                                 } else
 239                                         errs = 1;
 240                         }
 241                 }
 242                 string_list_remove_empty_items(&dco->filters, 0);
 243         }
 244         stop_progress(&progress);
 245         string_list_clear(&dco->filters, 0);
 246
 247         /* At this point we should not have any delayed paths anymore. */
 248         errs |= dco->paths.nr;
 249         for_each_string_list_item(path, &dco->paths) {
 250                 error("'%s' was not filtered properly", path->string);
 251         }
 252         string_list_clear(&dco->paths, 0);
 253
 254         free(dco);
 255         state->delayed_checkout = NULL;
 256
 257         return errs;
 258 }
 259
 260 void update_ce_after_write(const struct checkout *state, struct cache_entry *ce,
 261                            struct stat *st)
 262 {
 263         if (state->refresh_cache) {
 264                 assert(state->istate);
 265                 fill_stat_cache_info(state->istate, ce, st);
 266                 ce->ce_flags |= CE_UPDATE_IN_BASE;
 267                 mark_fsmonitor_invalid(state->istate, ce);
 268                 state->istate->cache_changed |= CE_ENTRY_CHANGED;
 269         }
 270 }
 271
 272 /* Note: ca is used (and required) iff the entry refers to a regular file. */
 273 static int write_entry(struct cache_entry *ce, char *path, struct conv_attrs *ca,
 274                        const struct checkout *state, int to_tempfile,
 275                        int *nr_checkouts)
 276 {
 277         unsigned int ce_mode_s_ifmt = ce->ce_mode & S_IFMT;
 278         struct delayed_checkout *dco = state->delayed_checkout;
 279         int fd, ret, fstat_done = 0;
 280         char *new_blob;
 281         struct strbuf buf = STRBUF_INIT;
 282         size_t size;
 283         ssize_t wrote;
 284         size_t newsize = 0;
 285         struct stat st;
 286         const struct submodule *sub;
 287         struct checkout_metadata meta;
 288         static int scratch_nr_checkouts;
 289
 290         clone_checkout_metadata(&meta, &state->meta, &ce->oid);
 291
 292         if (ce_mode_s_ifmt == S_IFREG) {
 293                 struct stream_filter *filter = get_stream_filter_ca(ca, &ce->oid);
 294                 if (filter &&
 295                     !streaming_write_entry(ce, path, filter,
 296                                            state, to_tempfile,
 297                                            &fstat_done, &st))
 298                         goto finish;
 299         }
 300
 301         switch (ce_mode_s_ifmt) {
 302         case S_IFLNK:
 303                 new_blob = read_blob_entry(ce, &size);
 304                 if (!new_blob)
 305                         return error("unable to read sha1 file of %s (%s)",
 306                                      ce->name, oid_to_hex(&ce->oid));
 307
 308                 /*
 309                  * We can't make a real symlink; write out a regular file entry
 310                  * with the symlink destination as its contents.
 311                  */
 312                 if (!has_symlinks || to_tempfile)
 313                         goto write_file_entry;
 314
 315                 ret = symlink(new_blob, path);
 316                 free(new_blob);
 317                 if (ret)
 318                         return error_errno("unable to create symlink %s", path);
 319                 break;
 320
 321         case S_IFREG:
 322                 /*
 323                  * We do not send the blob in case of a retry, so do not
 324                  * bother reading it at all.
 325                  */
 326                 if (dco && dco->state == CE_RETRY) {
 327                         new_blob = NULL;
 328                         size = 0;
 329                 } else {
 330                         new_blob = read_blob_entry(ce, &size);
 331                         if (!new_blob)
 332                                 return error("unable to read sha1 file of %s (%s)",
 333                                              ce->name, oid_to_hex(&ce->oid));
 334                 }
 335
 336                 /*
 337                  * Convert from git internal format to working tree format
 338                  */
 339                 if (dco && dco->state != CE_NO_DELAY) {
 340                         ret = async_convert_to_working_tree_ca(ca, ce->name,
 341                                                                new_blob, size,
 342                                                                &buf, &meta, dco);
 343                         if (ret) {
 344                                 struct string_list_item *item =
 345                                         string_list_lookup(&dco->paths, ce->name);
 346                                 if (item) {
 347                                         item->util = nr_checkouts ? nr_checkouts
 348                                                         : &scratch_nr_checkouts;
 349                                         free(new_blob);
 350                                         goto delayed;
 351                                 }
 352                         }
 353                 } else {
 354                         ret = convert_to_working_tree_ca(ca, ce->name, new_blob,
 355                                                          size, &buf, &meta);
 356                 }
 357
 358                 if (ret) {
 359                         free(new_blob);
 360                         new_blob = strbuf_detach(&buf, &newsize);
 361                         size = newsize;
 362                 }
 363                 /*
 364                  * No "else" here as errors from convert are OK at this
 365                  * point. If the error would have been fatal (e.g.
 366                  * filter is required), then we would have died already.
 367                  */
 368
 369         write_file_entry:
 370                 fd = open_output_fd(path, ce, to_tempfile);
 371                 if (fd < 0) {
 372                         free(new_blob);
 373                         return error_errno("unable to create file %s", path);
 374                 }
 375
 376                 wrote = write_in_full(fd, new_blob, size);
 377                 if (!to_tempfile)
 378                         fstat_done = fstat_checkout_output(fd, state, &st);
 379                 close(fd);
 380                 free(new_blob);
 381                 if (wrote < 0)
 382                         return error("unable to write file %s", path);
 383                 break;
 384
 385         case S_IFGITLINK:
 386                 if (to_tempfile)
 387                         return error("cannot create temporary submodule %s", ce->name);
 388                 if (mkdir(path, 0777) < 0)
 389                         return error("cannot create submodule directory %s", path);
 390                 sub = submodule_from_ce(ce);
 391                 if (sub)
 392                         return submodule_move_head(ce->name, state->super_prefix,
 393                                 NULL, oid_to_hex(&ce->oid),
 394                                 state->force ? SUBMODULE_MOVE_HEAD_FORCE : 0);
 395                 break;
 396
 397         default:
 398                 return error("unknown file mode for %s in index", ce->name);
 399         }
 400
 401 finish:
 402         if (state->refresh_cache) {
 403                 if (!fstat_done && lstat(ce->name, &st) < 0)
 404                         return error_errno("unable to stat just-written file %s",
 405                                            ce->name);
 406                 update_ce_after_write(state, ce , &st);
 407         }
 408         if (nr_checkouts)
 409                 (*nr_checkouts)++;
 410 delayed:
 411         return 0;
 412 }
 413
 414 /*
 415  * This is like 'lstat()', except it refuses to follow symlinks
 416  * in the path, after skipping "skiplen".
 417  */
 418 static int check_path(const char *path, int len, struct stat *st, int skiplen)
 419 {
 420         const char *slash = path + len;
 421
 422         while (path < slash && *slash != '/')
 423                 slash--;
 424         if (!has_dirs_only_path(path, slash - path, skiplen)) {
 425                 errno = ENOENT;
 426                 return -1;
 427         }
 428         return lstat(path, st);
 429 }
 430
 431 static void mark_colliding_entries(const struct checkout *state,
 432                                    struct cache_entry *ce, struct stat *st)
 433 {
 434         int i, trust_ino = check_stat;
 435
 436 #if defined(GIT_WINDOWS_NATIVE) || defined(__CYGWIN__)
 437         trust_ino = 0;
 438 #endif
 439
 440         ce->ce_flags |= CE_MATCHED;
 441
 442         /* TODO: audit for interaction with sparse-index. */
 443         ensure_full_index(state->istate);
 444         for (i = 0; i < state->istate->cache_nr; i++) {
 445                 struct cache_entry *dup = state->istate->cache[i];
 446
 447                 if (dup == ce) {
 448                         /*
 449                          * Parallel checkout doesn't create the files in index
 450                          * order. So the other side of the collision may appear
 451                          * after the given cache_entry in the array.
 452                          */
 453                         if (parallel_checkout_status() == PC_RUNNING)
 454                                 continue;
 455                         else
 456                                 break;
 457                 }
 458
 459                 if (dup->ce_flags & (CE_MATCHED | CE_VALID | CE_SKIP_WORKTREE))
 460                         continue;
 461
 462                 if ((trust_ino && !match_stat_data(&dup->ce_stat_data, st)) ||
 463                     paths_collide(ce->name, dup->name)) {
 464                         dup->ce_flags |= CE_MATCHED;
 465                         break;
 466                 }
 467         }
 468 }
 469
 470 int checkout_entry_ca(struct cache_entry *ce, struct conv_attrs *ca,
 471                       const struct checkout *state, char *topath,
 472                       int *nr_checkouts)
 473 {
 474         static struct strbuf path = STRBUF_INIT;
 475         struct stat st;
 476         struct conv_attrs ca_buf;
 477
 478         if (ce->ce_flags & CE_WT_REMOVE) {
 479                 if (topath)
 480                         /*
 481                          * No content and thus no path to create, so we have
 482                          * no pathname to return.
 483                          */
 484                         BUG("Can't remove entry to a path");
 485                 unlink_entry(ce, state->super_prefix);
 486                 return 0;
 487         }
 488
 489         if (topath) {
 490                 if (S_ISREG(ce->ce_mode) && !ca) {
 491                         convert_attrs(state->istate, &ca_buf, ce->name);
 492                         ca = &ca_buf;
 493                 }
 494                 return write_entry(ce, topath, ca, state, 1, nr_checkouts);
 495         }
 496
 497         strbuf_reset(&path);
 498         strbuf_add(&path, state->base_dir, state->base_dir_len);
 499         strbuf_add(&path, ce->name, ce_namelen(ce));
 500
 501         if (!check_path(path.buf, path.len, &st, state->base_dir_len)) {
 502                 const struct submodule *sub;
 503                 unsigned changed = ie_match_stat(state->istate, ce, &st,
 504                                                  CE_MATCH_IGNORE_VALID | CE_MATCH_IGNORE_SKIP_WORKTREE);
 505                 /*
 506                  * Needs to be checked before !changed returns early,
 507                  * as the possibly empty directory was not changed
 508                  */
 509                 sub = submodule_from_ce(ce);
 510                 if (sub) {
 511                         int err;
 512                         if (!is_submodule_populated_gently(ce->name, &err)) {
 513                                 struct stat sb;
 514                                 if (lstat(ce->name, &sb))
 515                                         die(_("could not stat file '%s'"), ce->name);
 516                                 if (!(st.st_mode & S_IFDIR))
 517                                         unlink_or_warn(ce->name);
 518
 519                                 return submodule_move_head(ce->name, state->super_prefix,
 520                                         NULL, oid_to_hex(&ce->oid), 0);
 521                         } else
 522                                 return submodule_move_head(ce->name, state->super_prefix,
 523                                         "HEAD", oid_to_hex(&ce->oid),
 524                                         state->force ? SUBMODULE_MOVE_HEAD_FORCE : 0);
 525                 }
 526
 527                 if (!changed)
 528                         return 0;
 529                 if (!state->force) {
 530                         if (!state->quiet)
 531                                 fprintf(stderr,
 532                                         "%s already exists, no checkout\n",
 533                                         path.buf);
 534                         return -1;
 535                 }
 536
 537                 if (state->clone)
 538                         mark_colliding_entries(state, ce, &st);
 539
 540                 /*
 541                  * We unlink the old file, to get the new one with the
 542                  * right permissions (including umask, which is nasty
 543                  * to emulate by hand - much easier to let the system
 544                  * just do the right thing)
 545                  */
 546                 if (S_ISDIR(st.st_mode)) {
 547                         /* If it is a gitlink, leave it alone! */
 548                         if (S_ISGITLINK(ce->ce_mode))
 549                                 return 0;
 550                         /*
 551                          * We must avoid replacing submodules' leading
 552                          * directories with symbolic links, lest recursive
 553                          * clones can write into arbitrary locations.
 554                          *
 555                          * Technically, this logic is not limited
 556                          * to recursive clones, or for that matter to
 557                          * submodules' paths colliding with symbolic links'
 558                          * paths. Yet it strikes a balance in favor of
 559                          * simplicity, and if paths are colliding, we might
 560                          * just as well keep the directories during a clone.
 561                          */
 562                         if (state->clone && S_ISLNK(ce->ce_mode))
 563                                 return 0;
 564                         remove_subtree(&path);
 565                 } else if (unlink(path.buf))
 566                         return error_errno("unable to unlink old '%s'", path.buf);
 567         } else if (state->not_new)
 568                 return 0;
 569
 570         create_directories(path.buf, path.len, state);
 571
 572         if (S_ISREG(ce->ce_mode) && !ca) {
 573                 convert_attrs(state->istate, &ca_buf, ce->name);
 574                 ca = &ca_buf;
 575         }
 576
 577         if (!enqueue_checkout(ce, ca, nr_checkouts))
 578                 return 0;
 579
 580         return write_entry(ce, path.buf, ca, state, 0, nr_checkouts);
 581 }
 582
 583 void unlink_entry(const struct cache_entry *ce, const char *super_prefix)
 584 {
 585         const struct submodule *sub = submodule_from_ce(ce);
 586         if (sub) {
 587                 /* state.force is set at the caller. */
 588                 submodule_move_head(ce->name, super_prefix, "HEAD", NULL,
 589                                     SUBMODULE_MOVE_HEAD_FORCE);
 590         }
 591         if (check_leading_path(ce->name, ce_namelen(ce), 1) >= 0)
 592                 return;
 593         if (remove_or_warn(ce->ce_mode, ce->name))
 594                 return;
 595         schedule_dir_for_removal(ce->name, ce_namelen(ce));
 596 }
 597
 598 int remove_or_warn(unsigned int mode, const char *file)
 599 {
 600         return S_ISGITLINK(mode) ? rmdir_or_warn(file) : unlink_or_warn(file);
 601 }