entry.c

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