dir.h

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