entry.c

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