dir.h

   1 #ifndef DIR_H
   2 #define DIR_H
   3
   4 #include "cache.h"
   5 #include "hashmap.h"
   6 #include "strbuf.h"
   7
   8 /**
   9  * The directory listing API is used to enumerate paths in the work tree,
  10  * optionally taking `.git/info/exclude` and `.gitignore` files per directory
  11  * into account.
  12  */
  13
  14 /**
  15  * Calling sequence
  16  * ----------------
  17  *
  18  * Note: The index may be checked for .gitignore files that are
  19  * CE_SKIP_WORKTREE marked. If you want to exclude files, make sure you have
  20  * loaded the index first.
  21  *
  22  * - Prepare `struct dir_struct dir` and clear it with `memset(&dir, 0,
  23  * sizeof(dir))`.
  24  *
  25  * - To add single exclude pattern, call `add_pattern_list()` and then
  26  *   `add_pattern()`.
  27  *
  28  * - To add patterns from a file (e.g. `.git/info/exclude`), call
  29  *   `add_patterns_from_file()` , and/or set `dir.exclude_per_dir`.  A
  30  *   short-hand function `setup_standard_excludes()` can be used to set
  31  *   up the standard set of exclude settings.
  32  *
  33  * - Set options described in the Data Structure section above.
  34  *
  35  * - Call `read_directory()`.
  36  *
  37  * - Use `dir.entries[]`.
  38  *
  39  * - Call `clear_directory()` when none of the contained elements are no longer in use.
  40  *
  41  */
  42
  43 struct dir_entry {
  44         unsigned int len;
  45         char name[FLEX_ARRAY]; /* more */
  46 };
  47
  48 #define PATTERN_FLAG_NODIR 1
  49 #define PATTERN_FLAG_ENDSWITH 4
  50 #define PATTERN_FLAG_MUSTBEDIR 8
  51 #define PATTERN_FLAG_NEGATIVE 16
  52
  53 struct path_pattern {
  54         /*
  55          * This allows callers of last_matching_pattern() etc.
  56          * to determine the origin of the matching pattern.
  57          */
  58         struct pattern_list *pl;
  59
  60         const char *pattern;
  61         int patternlen;
  62         int nowildcardlen;
  63         const char *base;
  64         int baselen;
  65         unsigned flags;         /* PATTERN_FLAG_* */
  66
  67         /*
  68          * Counting starts from 1 for line numbers in ignore files,
  69          * and from -1 decrementing for patterns from CLI args.
  70          */
  71         int srcpos;
  72 };
  73
  74 /* used for hashmaps for cone patterns */
  75 struct pattern_entry {
  76         struct hashmap_entry ent;
  77         char *pattern;
  78         size_t patternlen;
  79 };
  80
  81 /*
  82  * Each excludes file will be parsed into a fresh exclude_list which
  83  * is appended to the relevant exclude_list_group (either EXC_DIRS or
  84  * EXC_FILE).  An exclude_list within the EXC_CMDL exclude_list_group
  85  * can also be used to represent the list of --exclude values passed
  86  * via CLI args.
  87  */
  88 struct pattern_list {
  89         int nr;
  90         int alloc;
  91
  92         /* remember pointer to exclude file contents so we can free() */
  93         char *filebuf;
  94
  95         /* origin of list, e.g. path to filename, or descriptive string */
  96         const char *src;
  97
  98         struct path_pattern **patterns;
  99
 100         /*
 101          * While scanning the excludes, we attempt to match the patterns
 102          * with a more restricted set that allows us to use hashsets for
 103          * matching logic, which is faster than the linear lookup in the
 104          * excludes array above. If non-zero, that check succeeded.
 105          */
 106         unsigned use_cone_patterns;
 107         unsigned full_cone;
 108
 109         /*
 110          * Stores paths where everything starting with those paths
 111          * is included.
 112          */
 113         struct hashmap recursive_hashmap;
 114
 115         /*
 116          * Used to check single-level parents of blobs.
 117          */
 118         struct hashmap parent_hashmap;
 119 };
 120
 121 /*
 122  * The contents of the per-directory exclude files are lazily read on
 123  * demand and then cached in memory, one per exclude_stack struct, in
 124  * order to avoid opening and parsing each one every time that
 125  * directory is traversed.
 126  */
 127 struct exclude_stack {
 128         struct exclude_stack *prev; /* the struct exclude_stack for the parent directory */
 129         int baselen;
 130         int exclude_ix; /* index of exclude_list within EXC_DIRS exclude_list_group */
 131         struct untracked_cache_dir *ucd;
 132 };
 133
 134 struct exclude_list_group {
 135         int nr, alloc;
 136         struct pattern_list *pl;
 137 };
 138
 139 struct oid_stat {
 140         struct stat_data stat;
 141         struct object_id oid;
 142         int valid;
 143 };
 144
 145 /*
 146  *  Untracked cache
 147  *
 148  *  The following inputs are sufficient to determine what files in a
 149  *  directory are excluded:
 150  *
 151  *   - The list of files and directories of the directory in question
 152  *   - The $GIT_DIR/index
 153  *   - dir_struct flags
 154  *   - The content of $GIT_DIR/info/exclude
 155  *   - The content of core.excludesfile
 156  *   - The content (or the lack) of .gitignore of all parent directories
 157  *     from $GIT_WORK_TREE
 158  *   - The check_only flag in read_directory_recursive (for
 159  *     DIR_HIDE_EMPTY_DIRECTORIES)
 160  *
 161  *  The first input can be checked using directory mtime. In many
 162  *  filesystems, directory mtime (stat_data field) is updated when its
 163  *  files or direct subdirs are added or removed.
 164  *
 165  *  The second one can be hooked from cache_tree_invalidate_path().
 166  *  Whenever a file (or a submodule) is added or removed from a
 167  *  directory, we invalidate that directory.
 168  *
 169  *  The remaining inputs are easy, their SHA-1 could be used to verify
 170  *  their contents (exclude_sha1[], info_exclude_sha1[] and
 171  *  excludes_file_sha1[])
 172  */
 173 struct untracked_cache_dir {
 174         struct untracked_cache_dir **dirs;
 175         char **untracked;
 176         struct stat_data stat_data;
 177         unsigned int untracked_alloc, dirs_nr, dirs_alloc;
 178         unsigned int untracked_nr;
 179         unsigned int check_only : 1;
 180         /* all data except 'dirs' in this struct are good */
 181         unsigned int valid : 1;
 182         unsigned int recurse : 1;
 183         /* null object ID means this directory does not have .gitignore */
 184         struct object_id exclude_oid;
 185         char name[FLEX_ARRAY];
 186 };
 187
 188 struct untracked_cache {
 189         struct oid_stat ss_info_exclude;
 190         struct oid_stat ss_excludes_file;
 191         const char *exclude_per_dir;
 192         struct strbuf ident;
 193         /*
 194          * dir_struct#flags must match dir_flags or the untracked
 195          * cache is ignored.
 196          */
 197         unsigned dir_flags;
 198         struct untracked_cache_dir *root;
 199         /* Statistics */
 200         int dir_created;
 201         int gitignore_invalidated;
 202         int dir_invalidated;
 203         int dir_opened;
 204         /* fsmonitor invalidation data */
 205         unsigned int use_fsmonitor : 1;
 206 };
 207
 208 /**
 209  * structure is used to pass directory traversal options to the library and to
 210  * record the paths discovered. A single `struct dir_struct` is used regardless
 211  * of whether or not the traversal recursively descends into subdirectories.
 212  */
 213 struct dir_struct {
 214
 215         /* The number of members in `entries[]` array. */
 216         int nr;
 217
 218         /* Internal use; keeps track of allocation of `entries[]` array.*/
 219         int alloc;
 220
 221         /* The number of members in `ignored[]` array. */
 222         int ignored_nr;
 223
 224         int ignored_alloc;
 225
 226         /* bit-field of options */
 227         enum {
 228
 229                 /**
 230                  * Return just ignored files in `entries[]`, not untracked files.
 231                  * This flag is mutually exclusive with `DIR_SHOW_IGNORED_TOO`.
 232                  */
 233                 DIR_SHOW_IGNORED = 1<<0,
 234
 235                 /* Include a directory that is not tracked. */
 236                 DIR_SHOW_OTHER_DIRECTORIES = 1<<1,
 237
 238                 /* Do not include a directory that is not tracked and is empty. */
 239                 DIR_HIDE_EMPTY_DIRECTORIES = 1<<2,
 240
 241                 /**
 242                  * If set, recurse into a directory that looks like a Git directory.
 243                  * Otherwise it is shown as a directory.
 244                  */
 245                 DIR_NO_GITLINKS = 1<<3,
 246
 247                 /**
 248                  * Special mode for git-add. Return ignored files in `ignored[]` and
 249                  * untracked files in `entries[]`. Only returns ignored files that match
 250                  * pathspec exactly (no wildcards). Does not recurse into ignored
 251                  * directories.
 252                  */
 253                 DIR_COLLECT_IGNORED = 1<<4,
 254
 255                 /**
 256                  * Similar to `DIR_SHOW_IGNORED`, but return ignored files in
 257                  * `ignored[]` in addition to untracked files in `entries[]`.
 258                  * This flag is mutually exclusive with `DIR_SHOW_IGNORED`.
 259                  */
 260                 DIR_SHOW_IGNORED_TOO = 1<<5,
 261
 262                 DIR_COLLECT_KILLED_ONLY = 1<<6,
 263
 264                 /**
 265                  * Only has meaning if `DIR_SHOW_IGNORED_TOO` is also set; if this is
 266                  * set, the untracked contents of untracked directories are also
 267                  * returned in `entries[]`.
 268                  */
 269                 DIR_KEEP_UNTRACKED_CONTENTS = 1<<7,
 270
 271                 /**
 272                  * Only has meaning if `DIR_SHOW_IGNORED_TOO` is also set; if this is
 273                  * set, returns ignored files and directories that match an exclude
 274                  * pattern. If a directory matches an exclude pattern, then the
 275                  * directory is returned and the contained paths are not. A directory
 276                  * that does not match an exclude pattern will not be returned even if
 277                  * all of its contents are ignored. In this case, the contents are
 278                  * returned as individual entries.
 279                  *
 280                  * If this is set, files and directories that explicitly match an ignore
 281                  * pattern are reported. Implicitly ignored directories (directories that
 282                  * do not match an ignore pattern, but whose contents are all ignored)
 283                  * are not reported, instead all of the contents are reported.
 284                  */
 285                 DIR_SHOW_IGNORED_TOO_MODE_MATCHING = 1<<8,
 286
 287                 DIR_SKIP_NESTED_GIT = 1<<9
 288         } flags;
 289
 290         /* An array of `struct dir_entry`, each element of which describes a path. */
 291         struct dir_entry **entries;
 292
 293         /**
 294          * used for ignored paths with the `DIR_SHOW_IGNORED_TOO` and
 295          * `DIR_COLLECT_IGNORED` flags.
 296          */
 297         struct dir_entry **ignored;
 298
 299         /**
 300          * The name of the file to be read in each directory for excluded files
 301          * (typically `.gitignore`).
 302          */
 303         const char *exclude_per_dir;
 304
 305         /*
 306          * We maintain three groups of exclude pattern lists:
 307          *
 308          * EXC_CMDL lists patterns explicitly given on the command line.
 309          * EXC_DIRS lists patterns obtained from per-directory ignore files.
 310          * EXC_FILE lists patterns from fallback ignore files, e.g.
 311          *   - .git/info/exclude
 312          *   - core.excludesfile
 313          *
 314          * Each group contains multiple exclude lists, a single list
 315          * per source.
 316          */
 317 #define EXC_CMDL 0
 318 #define EXC_DIRS 1
 319 #define EXC_FILE 2
 320         struct exclude_list_group exclude_list_group[3];
 321
 322         /*
 323          * Temporary variables which are used during loading of the
 324          * per-directory exclude lists.
 325          *
 326          * exclude_stack points to the top of the exclude_stack, and
 327          * basebuf contains the full path to the current
 328          * (sub)directory in the traversal. Exclude points to the
 329          * matching exclude struct if the directory is excluded.
 330          */
 331         struct exclude_stack *exclude_stack;
 332         struct path_pattern *pattern;
 333         struct strbuf basebuf;
 334
 335         /* Enable untracked file cache if set */
 336         struct untracked_cache *untracked;
 337         struct oid_stat ss_info_exclude;
 338         struct oid_stat ss_excludes_file;
 339         unsigned unmanaged_exclude_files;
 340 };
 341
 342 /*Count the number of slashes for string s*/
 343 int count_slashes(const char *s);
 344
 345 /*
 346  * The ordering of these constants is significant, with
 347  * higher-numbered match types signifying "closer" (i.e. more
 348  * specific) matches which will override lower-numbered match types
 349  * when populating the seen[] array.
 350  */
 351 #define MATCHED_RECURSIVELY 1
 352 #define MATCHED_RECURSIVELY_LEADING_PATHSPEC 2
 353 #define MATCHED_FNMATCH 3
 354 #define MATCHED_EXACTLY 4
 355 int simple_length(const char *match);
 356 int no_wildcard(const char *string);
 357 char *common_prefix(const struct pathspec *pathspec);
 358 int match_pathspec(const struct index_state *istate,
 359                    const struct pathspec *pathspec,
 360                    const char *name, int namelen,
 361                    int prefix, char *seen, int is_dir);
 362 int report_path_error(const char *ps_matched, const struct pathspec *pathspec);
 363 int within_depth(const char *name, int namelen, int depth, int max_depth);
 364
 365 int fill_directory(struct dir_struct *dir,
 366                    struct index_state *istate,
 367                    const struct pathspec *pathspec);
 368 int read_directory(struct dir_struct *, struct index_state *istate,
 369                    const char *path, int len,
 370                    const struct pathspec *pathspec);
 371
 372 enum pattern_match_result {
 373         UNDECIDED = -1,
 374         NOT_MATCHED = 0,
 375         MATCHED = 1,
 376         MATCHED_RECURSIVE = 2,
 377 };
 378
 379 /*
 380  * Scan the list of patterns to determine if the ordered list
 381  * of patterns matches on 'pathname'.
 382  *
 383  * Return 1 for a match, 0 for not matched and -1 for undecided.
 384  */
 385 enum pattern_match_result path_matches_pattern_list(const char *pathname,
 386                                 int pathlen,
 387                                 const char *basename, int *dtype,
 388                                 struct pattern_list *pl,
 389                                 struct index_state *istate);
 390 struct dir_entry *dir_add_ignored(struct dir_struct *dir,
 391                                   struct index_state *istate,
 392                                   const char *pathname, int len);
 393
 394 /*
 395  * these implement the matching logic for dir.c:excluded_from_list and
 396  * attr.c:path_matches()
 397  */
 398 int match_basename(const char *, int,
 399                    const char *, int, int, unsigned);
 400 int match_pathname(const char *, int,
 401                    const char *, int,
 402                    const char *, int, int, unsigned);
 403
 404 struct path_pattern *last_matching_pattern(struct dir_struct *dir,
 405                                            struct index_state *istate,
 406                                            const char *name, int *dtype);
 407
 408 int is_excluded(struct dir_struct *dir,
 409                 struct index_state *istate,
 410                 const char *name, int *dtype);
 411
 412 int pl_hashmap_cmp(const void *unused_cmp_data,
 413                    const struct hashmap_entry *a,
 414                    const struct hashmap_entry *b,
 415                    const void *key);
 416 int hashmap_contains_parent(struct hashmap *map,
 417                             const char *path,
 418                             struct strbuf *buffer);
 419 struct pattern_list *add_pattern_list(struct dir_struct *dir,
 420                                       int group_type, const char *src);
 421 int add_patterns_from_file_to_list(const char *fname, const char *base, int baselen,
 422                                    struct pattern_list *pl, struct  index_state *istate);
 423 void add_patterns_from_file(struct dir_struct *, const char *fname);
 424 int add_patterns_from_blob_to_list(struct object_id *oid,
 425                                    const char *base, int baselen,
 426                                    struct pattern_list *pl);
 427 void parse_path_pattern(const char **string, int *patternlen, unsigned *flags, int *nowildcardlen);
 428 void add_pattern(const char *string, const char *base,
 429                  int baselen, struct pattern_list *pl, int srcpos);
 430 void clear_pattern_list(struct pattern_list *pl);
 431 void clear_directory(struct dir_struct *dir);
 432
 433 int repo_file_exists(struct repository *repo, const char *path);
 434 int file_exists(const char *);
 435
 436 int is_inside_dir(const char *dir);
 437 int dir_inside_of(const char *subdir, const char *dir);
 438
 439 static inline int is_dot_or_dotdot(const char *name)
 440 {
 441         return (name[0] == '.' &&
 442                 (name[1] == '\0' ||
 443                  (name[1] == '.' && name[2] == '\0')));
 444 }
 445
 446 int is_empty_dir(const char *dir);
 447
 448 void setup_standard_excludes(struct dir_struct *dir);
 449
 450
 451 /* Constants for remove_dir_recursively: */
 452
 453 /*
 454  * If a non-directory is found within path, stop and return an error.
 455  * (In this case some empty directories might already have been
 456  * removed.)
 457  */
 458 #define REMOVE_DIR_EMPTY_ONLY 01
 459
 460 /*
 461  * If any Git work trees are found within path, skip them without
 462  * considering it an error.
 463  */
 464 #define REMOVE_DIR_KEEP_NESTED_GIT 02
 465
 466 /* Remove the contents of path, but leave path itself. */
 467 #define REMOVE_DIR_KEEP_TOPLEVEL 04
 468
 469 /*
 470  * Remove path and its contents, recursively. flags is a combination
 471  * of the above REMOVE_DIR_* constants. Return 0 on success.
 472  *
 473  * This function uses path as temporary scratch space, but restores it
 474  * before returning.
 475  */
 476 int remove_dir_recursively(struct strbuf *path, int flag);
 477
 478 /* tries to remove the path with empty directories along it, ignores ENOENT */
 479 int remove_path(const char *path);
 480
 481 int fspathcmp(const char *a, const char *b);
 482 int fspathncmp(const char *a, const char *b, size_t count);
 483
 484 /*
 485  * The prefix part of pattern must not contains wildcards.
 486  */
 487 struct pathspec_item;
 488 int git_fnmatch(const struct pathspec_item *item,
 489                 const char *pattern, const char *string,
 490                 int prefix);
 491
 492 int submodule_path_match(const struct index_state *istate,
 493                          const struct pathspec *ps,
 494                          const char *submodule_name,
 495                          char *seen);
 496
 497 static inline int ce_path_match(const struct index_state *istate,
 498                                 const struct cache_entry *ce,
 499                                 const struct pathspec *pathspec,
 500                                 char *seen)
 501 {
 502         return match_pathspec(istate, pathspec, ce->name, ce_namelen(ce), 0, seen,
 503                               S_ISDIR(ce->ce_mode) || S_ISGITLINK(ce->ce_mode));
 504 }
 505
 506 static inline int dir_path_match(const struct index_state *istate,
 507                                  const struct dir_entry *ent,
 508                                  const struct pathspec *pathspec,
 509                                  int prefix, char *seen)
 510 {
 511         int has_trailing_dir = ent->len && ent->name[ent->len - 1] == '/';
 512         int len = has_trailing_dir ? ent->len - 1 : ent->len;
 513         return match_pathspec(istate, pathspec, ent->name, len, prefix, seen,
 514                               has_trailing_dir);
 515 }
 516
 517 int cmp_dir_entry(const void *p1, const void *p2);
 518 int check_dir_entry_contains(const struct dir_entry *out, const struct dir_entry *in);
 519
 520 void untracked_cache_invalidate_path(struct index_state *, const char *, int safe_path);
 521 void untracked_cache_remove_from_index(struct index_state *, const char *);
 522 void untracked_cache_add_to_index(struct index_state *, const char *);
 523
 524 void free_untracked_cache(struct untracked_cache *);
 525 struct untracked_cache *read_untracked_extension(const void *data, unsigned long sz);
 526 void write_untracked_extension(struct strbuf *out, struct untracked_cache *untracked);
 527 void add_untracked_cache(struct index_state *istate);
 528 void remove_untracked_cache(struct index_state *istate);
 529
 530 /*
 531  * Connect a worktree to a git directory by creating (or overwriting) a
 532  * '.git' file containing the location of the git directory. In the git
 533  * directory set the core.worktree setting to indicate where the worktree is.
 534  * When `recurse_into_nested` is set, recurse into any nested submodules,
 535  * connecting them as well.
 536  */
 537 void connect_work_tree_and_git_dir(const char *work_tree,
 538                                    const char *git_dir,
 539                                    int recurse_into_nested);
 540 void relocate_gitdir(const char *path,
 541                      const char *old_git_dir,
 542                      const char *new_git_dir);
 543 #endif