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