]> git.ipfire.org Git - thirdparty/git.git/blob - dir.h
Merge branch 'jc/diff-exit-code-with-w-fixes' into maint-2.42
[thirdparty/git.git] / dir.h
1 #ifndef DIR_H
2 #define DIR_H
3
4 #include "hash-ll.h"
5 #include "hashmap.h"
6 #include "pathspec.h"
7 #include "statinfo.h"
8 #include "strbuf.h"
9
10 struct repository;
11
12 /**
13 * The directory listing API is used to enumerate paths in the work tree,
14 * optionally taking `.git/info/exclude` and `.gitignore` files per directory
15 * into account.
16 */
17
18 /**
19 * Calling sequence
20 * ----------------
21 *
22 * Note: The index may be checked for .gitignore files that are
23 * CE_SKIP_WORKTREE marked. If you want to exclude files, make sure you have
24 * loaded the index first.
25 *
26 * - Prepare `struct dir_struct dir` using `dir_init()` function.
27 *
28 * - To add single exclude pattern, call `add_pattern_list()` and then
29 * `add_pattern()`.
30 *
31 * - To add patterns from a file (e.g. `.git/info/exclude`), call
32 * `add_patterns_from_file()` , and/or set `dir.exclude_per_dir`.
33 *
34 * - A short-hand function `setup_standard_excludes()` can be used to set
35 * up the standard set of exclude settings, instead of manually calling
36 * the add_pattern*() family of functions.
37 *
38 * - Call `fill_directory()`.
39 *
40 * - Use `dir.entries[]` and `dir.ignored[]`.
41 *
42 * - Call `dir_clear()` when the contained elements are no longer in use.
43 *
44 */
45
46 struct repository;
47
48 struct dir_entry {
49 unsigned int len;
50 char name[FLEX_ARRAY]; /* more */
51 };
52
53 #define PATTERN_FLAG_NODIR 1
54 #define PATTERN_FLAG_ENDSWITH 4
55 #define PATTERN_FLAG_MUSTBEDIR 8
56 #define PATTERN_FLAG_NEGATIVE 16
57
58 struct path_pattern {
59 /*
60 * This allows callers of last_matching_pattern() etc.
61 * to determine the origin of the matching pattern.
62 */
63 struct pattern_list *pl;
64
65 const char *pattern;
66 int patternlen;
67 int nowildcardlen;
68 const char *base;
69 int baselen;
70 unsigned flags; /* PATTERN_FLAG_* */
71
72 /*
73 * Counting starts from 1 for line numbers in ignore files,
74 * and from -1 decrementing for patterns from CLI args.
75 */
76 int srcpos;
77 };
78
79 /* used for hashmaps for cone patterns */
80 struct pattern_entry {
81 struct hashmap_entry ent;
82 char *pattern;
83 size_t patternlen;
84 };
85
86 /*
87 * Each excludes file will be parsed into a fresh exclude_list which
88 * is appended to the relevant exclude_list_group (either EXC_DIRS or
89 * EXC_FILE). An exclude_list within the EXC_CMDL exclude_list_group
90 * can also be used to represent the list of --exclude values passed
91 * via CLI args.
92 */
93 struct pattern_list {
94 int nr;
95 int alloc;
96
97 /* remember pointer to exclude file contents so we can free() */
98 char *filebuf;
99
100 /* origin of list, e.g. path to filename, or descriptive string */
101 const char *src;
102
103 struct path_pattern **patterns;
104
105 /*
106 * While scanning the excludes, we attempt to match the patterns
107 * with a more restricted set that allows us to use hashsets for
108 * matching logic, which is faster than the linear lookup in the
109 * excludes array above. If non-zero, that check succeeded.
110 */
111 unsigned use_cone_patterns;
112 unsigned full_cone;
113
114 /*
115 * Stores paths where everything starting with those paths
116 * is included.
117 */
118 struct hashmap recursive_hashmap;
119
120 /*
121 * Used to check single-level parents of blobs.
122 */
123 struct hashmap parent_hashmap;
124 };
125
126 /*
127 * The contents of the per-directory exclude files are lazily read on
128 * demand and then cached in memory, one per exclude_stack struct, in
129 * order to avoid opening and parsing each one every time that
130 * directory is traversed.
131 */
132 struct exclude_stack {
133 struct exclude_stack *prev; /* the struct exclude_stack for the parent directory */
134 int baselen;
135 int exclude_ix; /* index of exclude_list within EXC_DIRS exclude_list_group */
136 struct untracked_cache_dir *ucd;
137 };
138
139 struct exclude_list_group {
140 int nr, alloc;
141 struct pattern_list *pl;
142 };
143
144 struct oid_stat {
145 struct stat_data stat;
146 struct object_id oid;
147 int valid;
148 };
149
150 /*
151 * Untracked cache
152 *
153 * The following inputs are sufficient to determine what files in a
154 * directory are excluded:
155 *
156 * - The list of files and directories of the directory in question
157 * - The $GIT_DIR/index
158 * - dir_struct flags
159 * - The content of $GIT_DIR/info/exclude
160 * - The content of core.excludesfile
161 * - The content (or the lack) of .gitignore of all parent directories
162 * from $GIT_WORK_TREE
163 * - The check_only flag in read_directory_recursive (for
164 * DIR_HIDE_EMPTY_DIRECTORIES)
165 *
166 * The first input can be checked using directory mtime. In many
167 * filesystems, directory mtime (stat_data field) is updated when its
168 * files or direct subdirs are added or removed.
169 *
170 * The second one can be hooked from cache_tree_invalidate_path().
171 * Whenever a file (or a submodule) is added or removed from a
172 * directory, we invalidate that directory.
173 *
174 * The remaining inputs are easy, their SHA-1 could be used to verify
175 * their contents (exclude_sha1[], info_exclude_sha1[] and
176 * excludes_file_sha1[])
177 */
178 struct untracked_cache_dir {
179 struct untracked_cache_dir **dirs;
180 char **untracked;
181 struct stat_data stat_data;
182 unsigned int untracked_alloc, dirs_nr, dirs_alloc;
183 unsigned int untracked_nr;
184 unsigned int check_only : 1;
185 /* all data except 'dirs' in this struct are good */
186 unsigned int valid : 1;
187 unsigned int recurse : 1;
188 /* null object ID means this directory does not have .gitignore */
189 struct object_id exclude_oid;
190 char name[FLEX_ARRAY];
191 };
192
193 struct untracked_cache {
194 struct oid_stat ss_info_exclude;
195 struct oid_stat ss_excludes_file;
196 const char *exclude_per_dir;
197 char *exclude_per_dir_to_free;
198 struct strbuf ident;
199 /*
200 * dir_struct#flags must match dir_flags or the untracked
201 * cache is ignored.
202 */
203 unsigned dir_flags;
204 struct untracked_cache_dir *root;
205 /* Statistics */
206 int dir_created;
207 int gitignore_invalidated;
208 int dir_invalidated;
209 int dir_opened;
210 /* fsmonitor invalidation data */
211 unsigned int use_fsmonitor : 1;
212 };
213
214 /**
215 * structure is used to pass directory traversal options to the library and to
216 * record the paths discovered. A single `struct dir_struct` is used regardless
217 * of whether or not the traversal recursively descends into subdirectories.
218 */
219 struct dir_struct {
220
221 /* bit-field of options */
222 enum {
223
224 /**
225 * Return just ignored files in `entries[]`, not untracked files.
226 * This flag is mutually exclusive with `DIR_SHOW_IGNORED_TOO`.
227 */
228 DIR_SHOW_IGNORED = 1<<0,
229
230 /* Include a directory that is not tracked. */
231 DIR_SHOW_OTHER_DIRECTORIES = 1<<1,
232
233 /* Do not include a directory that is not tracked and is empty. */
234 DIR_HIDE_EMPTY_DIRECTORIES = 1<<2,
235
236 /**
237 * If set, recurse into a directory that looks like a Git directory.
238 * Otherwise it is shown as a directory.
239 */
240 DIR_NO_GITLINKS = 1<<3,
241
242 /**
243 * Special mode for git-add. Return ignored files in `ignored[]` and
244 * untracked files in `entries[]`. Only returns ignored files that match
245 * pathspec exactly (no wildcards). Does not recurse into ignored
246 * directories.
247 */
248 DIR_COLLECT_IGNORED = 1<<4,
249
250 /**
251 * Similar to `DIR_SHOW_IGNORED`, but return ignored files in
252 * `ignored[]` in addition to untracked files in `entries[]`.
253 * This flag is mutually exclusive with `DIR_SHOW_IGNORED`.
254 */
255 DIR_SHOW_IGNORED_TOO = 1<<5,
256
257 DIR_COLLECT_KILLED_ONLY = 1<<6,
258
259 /**
260 * Only has meaning if `DIR_SHOW_IGNORED_TOO` is also set; if this is
261 * set, the untracked contents of untracked directories are also
262 * returned in `entries[]`.
263 */
264 DIR_KEEP_UNTRACKED_CONTENTS = 1<<7,
265
266 /**
267 * Only has meaning if `DIR_SHOW_IGNORED_TOO` is also set; if this is
268 * set, returns ignored files and directories that match an exclude
269 * pattern. If a directory matches an exclude pattern, then the
270 * directory is returned and the contained paths are not. A directory
271 * that does not match an exclude pattern will not be returned even if
272 * all of its contents are ignored. In this case, the contents are
273 * returned as individual entries.
274 *
275 * If this is set, files and directories that explicitly match an ignore
276 * pattern are reported. Implicitly ignored directories (directories that
277 * do not match an ignore pattern, but whose contents are all ignored)
278 * are not reported, instead all of the contents are reported.
279 */
280 DIR_SHOW_IGNORED_TOO_MODE_MATCHING = 1<<8,
281
282 DIR_SKIP_NESTED_GIT = 1<<9
283 } flags;
284
285 /* The number of members in `entries[]` array. */
286 int nr; /* output only */
287
288 /* The number of members in `ignored[]` array. */
289 int ignored_nr; /* output only */
290
291 /* An array of `struct dir_entry`, each element of which describes a path. */
292 struct dir_entry **entries; /* output only */
293
294 /**
295 * used for ignored paths with the `DIR_SHOW_IGNORED_TOO` and
296 * `DIR_COLLECT_IGNORED` flags.
297 */
298 struct dir_entry **ignored; /* output only */
299
300 /* Enable/update untracked file cache if set */
301 struct untracked_cache *untracked;
302
303 /**
304 * Deprecated: ls-files is the only allowed caller; all other callers
305 * should leave this as NULL; it pre-dated the
306 * setup_standard_excludes() mechanism that replaces this.
307 *
308 * This field tracks the name of the file to be read in each directory
309 * for excluded files (typically `.gitignore`).
310 */
311 const char *exclude_per_dir;
312
313 struct dir_struct_internal {
314 /* Keeps track of allocation of `entries[]` array.*/
315 int alloc;
316
317 /* Keeps track of allocation of `ignored[]` array. */
318 int ignored_alloc;
319
320 /*
321 * We maintain three groups of exclude pattern lists:
322 *
323 * EXC_CMDL lists patterns explicitly given on the command line.
324 * EXC_DIRS lists patterns obtained from per-directory ignore
325 * files.
326 * EXC_FILE lists patterns from fallback ignore files, e.g.
327 * - .git/info/exclude
328 * - core.excludesfile
329 *
330 * Each group contains multiple exclude lists, a single list
331 * per source.
332 */
333 #define EXC_CMDL 0
334 #define EXC_DIRS 1
335 #define EXC_FILE 2
336 struct exclude_list_group exclude_list_group[3];
337
338 /*
339 * Temporary variables which are used during loading of the
340 * per-directory exclude lists.
341 *
342 * exclude_stack points to the top of the exclude_stack, and
343 * basebuf contains the full path to the current
344 * (sub)directory in the traversal. Exclude points to the
345 * matching exclude struct if the directory is excluded.
346 */
347 struct exclude_stack *exclude_stack;
348 struct path_pattern *pattern;
349 struct strbuf basebuf;
350
351 /* Additional metadata related to 'untracked' */
352 struct oid_stat ss_info_exclude;
353 struct oid_stat ss_excludes_file;
354 unsigned unmanaged_exclude_files;
355
356 /* Stats about the traversal */
357 unsigned visited_paths;
358 unsigned visited_directories;
359 } internal;
360 };
361
362 #define DIR_INIT { 0 }
363
364 struct dirent *readdir_skip_dot_and_dotdot(DIR *dirp);
365
366 /*Count the number of slashes for string s*/
367 int count_slashes(const char *s);
368
369 /*
370 * The ordering of these constants is significant, with
371 * higher-numbered match types signifying "closer" (i.e. more
372 * specific) matches which will override lower-numbered match types
373 * when populating the seen[] array.
374 */
375 #define MATCHED_RECURSIVELY 1
376 #define MATCHED_RECURSIVELY_LEADING_PATHSPEC 2
377 #define MATCHED_FNMATCH 3
378 #define MATCHED_EXACTLY 4
379 int simple_length(const char *match);
380 int no_wildcard(const char *string);
381 char *common_prefix(const struct pathspec *pathspec);
382 int report_path_error(const char *ps_matched, const struct pathspec *pathspec);
383 int within_depth(const char *name, int namelen, int depth, int max_depth);
384
385 int fill_directory(struct dir_struct *dir,
386 struct index_state *istate,
387 const struct pathspec *pathspec);
388 int read_directory(struct dir_struct *, struct index_state *istate,
389 const char *path, int len,
390 const struct pathspec *pathspec);
391
392 enum pattern_match_result {
393 UNDECIDED = -1,
394 NOT_MATCHED = 0,
395 MATCHED = 1,
396 MATCHED_RECURSIVE = 2,
397 };
398
399 /*
400 * Scan the list of patterns to determine if the ordered list
401 * of patterns matches on 'pathname'.
402 *
403 * Return 1 for a match, 0 for not matched and -1 for undecided.
404 */
405 enum pattern_match_result path_matches_pattern_list(const char *pathname,
406 int pathlen,
407 const char *basename, int *dtype,
408 struct pattern_list *pl,
409 struct index_state *istate);
410
411 int init_sparse_checkout_patterns(struct index_state *state);
412
413 int path_in_sparse_checkout(const char *path,
414 struct index_state *istate);
415 int path_in_cone_mode_sparse_checkout(const char *path,
416 struct index_state *istate);
417
418 struct dir_entry *dir_add_ignored(struct dir_struct *dir,
419 struct index_state *istate,
420 const char *pathname, int len);
421
422 /*
423 * these implement the matching logic for dir.c:excluded_from_list and
424 * attr.c:path_matches()
425 */
426 int match_basename(const char *, int,
427 const char *, int, int, unsigned);
428 int match_pathname(const char *, int,
429 const char *, int,
430 const char *, int, int);
431
432 struct path_pattern *last_matching_pattern(struct dir_struct *dir,
433 struct index_state *istate,
434 const char *name, int *dtype);
435
436 int is_excluded(struct dir_struct *dir,
437 struct index_state *istate,
438 const char *name, int *dtype);
439
440 int pl_hashmap_cmp(const void *unused_cmp_data,
441 const struct hashmap_entry *a,
442 const struct hashmap_entry *b,
443 const void *key);
444 int hashmap_contains_parent(struct hashmap *map,
445 const char *path,
446 struct strbuf *buffer);
447 struct pattern_list *add_pattern_list(struct dir_struct *dir,
448 int group_type, const char *src);
449 int add_patterns_from_file_to_list(const char *fname, const char *base, int baselen,
450 struct pattern_list *pl, struct index_state *istate,
451 unsigned flags);
452 void add_patterns_from_file(struct dir_struct *, const char *fname);
453 int add_patterns_from_blob_to_list(struct object_id *oid,
454 const char *base, int baselen,
455 struct pattern_list *pl);
456 void parse_path_pattern(const char **string, int *patternlen, unsigned *flags, int *nowildcardlen);
457 void add_pattern(const char *string, const char *base,
458 int baselen, struct pattern_list *pl, int srcpos);
459 void clear_pattern_list(struct pattern_list *pl);
460 void dir_clear(struct dir_struct *dir);
461
462 int repo_file_exists(struct repository *repo, const char *path);
463 int file_exists(const char *);
464
465 int is_inside_dir(const char *dir);
466 int dir_inside_of(const char *subdir, const char *dir);
467
468 static inline int is_dot_or_dotdot(const char *name)
469 {
470 return (name[0] == '.' &&
471 (name[1] == '\0' ||
472 (name[1] == '.' && name[2] == '\0')));
473 }
474
475 int is_empty_dir(const char *dir);
476
477 /*
478 * Retrieve the "humanish" basename of the given Git URL.
479 *
480 * For example:
481 * /path/to/repo.git => "repo"
482 * host.xz:foo/.git => "foo"
483 * http://example.com/user/bar.baz => "bar.baz"
484 */
485 char *git_url_basename(const char *repo, int is_bundle, int is_bare);
486 void strip_dir_trailing_slashes(char *dir);
487
488 void setup_standard_excludes(struct dir_struct *dir);
489
490 char *get_sparse_checkout_filename(void);
491 int get_sparse_checkout_patterns(struct pattern_list *pl);
492
493 /* Constants for remove_dir_recursively: */
494
495 /*
496 * If a non-directory is found within path, stop and return an error.
497 * (In this case some empty directories might already have been
498 * removed.)
499 */
500 #define REMOVE_DIR_EMPTY_ONLY 01
501
502 /*
503 * If any Git work trees are found within path, skip them without
504 * considering it an error.
505 */
506 #define REMOVE_DIR_KEEP_NESTED_GIT 02
507
508 /* Remove the contents of path, but leave path itself. */
509 #define REMOVE_DIR_KEEP_TOPLEVEL 04
510
511 /* Remove the_original_cwd too */
512 #define REMOVE_DIR_PURGE_ORIGINAL_CWD 0x08
513
514 /*
515 * Remove path and its contents, recursively. flags is a combination
516 * of the above REMOVE_DIR_* constants. Return 0 on success.
517 *
518 * This function uses path as temporary scratch space, but restores it
519 * before returning.
520 */
521 int remove_dir_recursively(struct strbuf *path, int flag);
522
523 /*
524 * Tries to remove the path, along with leading empty directories so long as
525 * those empty directories are not startup_info->original_cwd. Ignores
526 * ENOENT.
527 */
528 int remove_path(const char *path);
529
530 int fspathcmp(const char *a, const char *b);
531 int fspatheq(const char *a, const char *b);
532 int fspathncmp(const char *a, const char *b, size_t count);
533 unsigned int fspathhash(const char *str);
534
535 /*
536 * The prefix part of pattern must not contains wildcards.
537 */
538 struct pathspec_item;
539 int git_fnmatch(const struct pathspec_item *item,
540 const char *pattern, const char *string,
541 int prefix);
542
543 int submodule_path_match(struct index_state *istate,
544 const struct pathspec *ps,
545 const char *submodule_name,
546 char *seen);
547
548 static inline int dir_path_match(struct index_state *istate,
549 const struct dir_entry *ent,
550 const struct pathspec *pathspec,
551 int prefix, char *seen)
552 {
553 int has_trailing_dir = ent->len && ent->name[ent->len - 1] == '/';
554 int len = has_trailing_dir ? ent->len - 1 : ent->len;
555 return match_pathspec(istate, pathspec, ent->name, len, prefix, seen,
556 has_trailing_dir);
557 }
558
559 int cmp_dir_entry(const void *p1, const void *p2);
560 int check_dir_entry_contains(const struct dir_entry *out, const struct dir_entry *in);
561
562 void untracked_cache_invalidate_path(struct index_state *, const char *, int safe_path);
563 void untracked_cache_remove_from_index(struct index_state *, const char *);
564 void untracked_cache_add_to_index(struct index_state *, const char *);
565
566 void free_untracked_cache(struct untracked_cache *);
567 struct untracked_cache *read_untracked_extension(const void *data, unsigned long sz);
568 void write_untracked_extension(struct strbuf *out, struct untracked_cache *untracked);
569 void add_untracked_cache(struct index_state *istate);
570 void remove_untracked_cache(struct index_state *istate);
571
572 /*
573 * Connect a worktree to a git directory by creating (or overwriting) a
574 * '.git' file containing the location of the git directory. In the git
575 * directory set the core.worktree setting to indicate where the worktree is.
576 * When `recurse_into_nested` is set, recurse into any nested submodules,
577 * connecting them as well.
578 */
579 void connect_work_tree_and_git_dir(const char *work_tree,
580 const char *git_dir,
581 int recurse_into_nested);
582 void relocate_gitdir(const char *path,
583 const char *old_git_dir,
584 const char *new_git_dir);
585
586 /**
587 * The "enum path_matches_kind" determines how path_match_flags() will
588 * behave. The flags come in sets, and one (and only one) must be
589 * provided out of each "set":
590 *
591 * PATH_MATCH_NATIVE:
592 * Path separator is is_dir_sep()
593 * PATH_MATCH_XPLATFORM:
594 * Path separator is is_xplatform_dir_sep()
595 *
596 * Do we use is_dir_sep() to check for a directory separator
597 * (*_NATIVE), or do we always check for '/' or '\' (*_XPLATFORM). The
598 * "*_NATIVE" version on Windows is the same as "*_XPLATFORM",
599 * everywhere else "*_NATIVE" means "only /".
600 *
601 * PATH_MATCH_STARTS_WITH_DOT_SLASH:
602 * Match a path starting with "./"
603 * PATH_MATCH_STARTS_WITH_DOT_DOT_SLASH:
604 * Match a path starting with "../"
605 *
606 * The "/" in the above is adjusted based on the "*_NATIVE" and
607 * "*_XPLATFORM" flags.
608 */
609 enum path_match_flags {
610 PATH_MATCH_NATIVE = 1 << 0,
611 PATH_MATCH_XPLATFORM = 1 << 1,
612 PATH_MATCH_STARTS_WITH_DOT_SLASH = 1 << 2,
613 PATH_MATCH_STARTS_WITH_DOT_DOT_SLASH = 1 << 3,
614 };
615 #define PATH_MATCH_KINDS_MASK (PATH_MATCH_STARTS_WITH_DOT_SLASH | \
616 PATH_MATCH_STARTS_WITH_DOT_DOT_SLASH)
617 #define PATH_MATCH_PLATFORM_MASK (PATH_MATCH_NATIVE | PATH_MATCH_XPLATFORM)
618
619 /**
620 * path_match_flags() checks if a given "path" matches a given "enum
621 * path_match_flags" criteria.
622 */
623 int path_match_flags(const char *const path, const enum path_match_flags f);
624
625 /**
626 * starts_with_dot_slash_native(): convenience wrapper for
627 * path_match_flags() with PATH_MATCH_STARTS_WITH_DOT_SLASH and
628 * PATH_MATCH_NATIVE.
629 */
630 static inline int starts_with_dot_slash_native(const char *const path)
631 {
632 const enum path_match_flags what = PATH_MATCH_STARTS_WITH_DOT_SLASH;
633
634 return path_match_flags(path, what | PATH_MATCH_NATIVE);
635 }
636
637 /**
638 * starts_with_dot_slash_native(): convenience wrapper for
639 * path_match_flags() with PATH_MATCH_STARTS_WITH_DOT_DOT_SLASH and
640 * PATH_MATCH_NATIVE.
641 */
642 static inline int starts_with_dot_dot_slash_native(const char *const path)
643 {
644 const enum path_match_flags what = PATH_MATCH_STARTS_WITH_DOT_DOT_SLASH;
645
646 return path_match_flags(path, what | PATH_MATCH_NATIVE);
647 }
648
649 #endif