dir.c: optionally compute sha-1 of a .gitignore file
[git/git.git] / dir.c
CommitLineData
453ec4bd
LT
1/*
2 * This handles recursive filename detection with exclude
3 * files, index knowledge etc..
4 *
95a68344
AS
5 * See Documentation/technical/api-directory-listing.txt
6 *
453ec4bd
LT
7 * Copyright (C) Linus Torvalds, 2005-2006
8 * Junio Hamano, 2005-2006
9 */
453ec4bd
LT
10#include "cache.h"
11#include "dir.h"
09595258 12#include "refs.h"
237ec6e4 13#include "wildmatch.h"
64acde94 14#include "pathspec.h"
453ec4bd 15
9fc42d60
LT
16struct path_simplify {
17 int len;
18 const char *path;
19};
20
defd7c7b
KB
21/*
22 * Tells read_directory_recursive how a file or directory should be treated.
23 * Values are ordered by significance, e.g. if a directory contains both
24 * excluded and untracked files, it is listed as untracked because
25 * path_untracked > path_excluded.
26 */
27enum path_treatment {
28 path_none = 0,
29 path_recurse,
30 path_excluded,
31 path_untracked
32};
33
34static enum path_treatment read_directory_recursive(struct dir_struct *dir,
35 const char *path, int len,
09595258 36 int check_only, const struct path_simplify *simplify);
caa6b782 37static int get_dtype(struct dirent *de, const char *path, int len);
09595258 38
8cf2a84e
JJ
39/* helper string functions with support for the ignore_case flag */
40int strcmp_icase(const char *a, const char *b)
41{
42 return ignore_case ? strcasecmp(a, b) : strcmp(a, b);
43}
44
45int strncmp_icase(const char *a, const char *b, size_t count)
46{
47 return ignore_case ? strncasecmp(a, b, count) : strncmp(a, b, count);
48}
49
50int fnmatch_icase(const char *pattern, const char *string, int flags)
51{
eb07894f
NTND
52 return wildmatch(pattern, string,
53 flags | (ignore_case ? WM_CASEFOLD : 0),
54 NULL);
8cf2a84e
JJ
55}
56
1f26ce61
CB
57int git_fnmatch(const struct pathspec_item *item,
58 const char *pattern, const char *string,
59 int prefix)
5d74762d 60{
5d74762d 61 if (prefix > 0) {
93d93537 62 if (ps_strncmp(item, pattern, string, prefix))
eb07894f 63 return WM_NOMATCH;
5d74762d
NTND
64 pattern += prefix;
65 string += prefix;
66 }
bd30c2e4 67 if (item->flags & PATHSPEC_ONESTAR) {
8c6abbcd
NTND
68 int pattern_len = strlen(++pattern);
69 int string_len = strlen(string);
70 return string_len < pattern_len ||
93d93537
NTND
71 ps_strcmp(item, pattern,
72 string + string_len - pattern_len);
8c6abbcd 73 }
bd30c2e4 74 if (item->magic & PATHSPEC_GLOB)
93d93537
NTND
75 return wildmatch(pattern, string,
76 WM_PATHNAME |
77 (item->magic & PATHSPEC_ICASE ? WM_CASEFOLD : 0),
78 NULL);
bd30c2e4
NTND
79 else
80 /* wildmatch has not learned no FNM_PATHNAME mode yet */
eb07894f
NTND
81 return wildmatch(pattern, string,
82 item->magic & PATHSPEC_ICASE ? WM_CASEFOLD : 0,
83 NULL);
5d74762d
NTND
84}
85
0b6e56df
JH
86static int fnmatch_icase_mem(const char *pattern, int patternlen,
87 const char *string, int stringlen,
88 int flags)
89{
90 int match_status;
91 struct strbuf pat_buf = STRBUF_INIT;
92 struct strbuf str_buf = STRBUF_INIT;
93 const char *use_pat = pattern;
94 const char *use_str = string;
95
96 if (pattern[patternlen]) {
97 strbuf_add(&pat_buf, pattern, patternlen);
98 use_pat = pat_buf.buf;
99 }
100 if (string[stringlen]) {
101 strbuf_add(&str_buf, string, stringlen);
102 use_str = str_buf.buf;
103 }
104
f30366b2
JH
105 if (ignore_case)
106 flags |= WM_CASEFOLD;
107 match_status = wildmatch(use_pat, use_str, flags, NULL);
0b6e56df
JH
108
109 strbuf_release(&pat_buf);
110 strbuf_release(&str_buf);
111
112 return match_status;
113}
114
827f4d6c 115static size_t common_prefix_len(const struct pathspec *pathspec)
3c6a370b 116{
827f4d6c 117 int n;
4a085b16 118 size_t max = 0;
3c6a370b 119
93d93537
NTND
120 /*
121 * ":(icase)path" is treated as a pathspec full of
122 * wildcard. In other words, only prefix is considered common
123 * prefix. If the pathspec is abc/foo abc/bar, running in
124 * subdir xyz, the common prefix is still xyz, not xuz/abc as
125 * in non-:(icase).
126 */
5c6933d2
NTND
127 GUARD_PATHSPEC(pathspec,
128 PATHSPEC_FROMTOP |
129 PATHSPEC_MAXDEPTH |
bd30c2e4 130 PATHSPEC_LITERAL |
93d93537 131 PATHSPEC_GLOB |
ef79b1f8
NTND
132 PATHSPEC_ICASE |
133 PATHSPEC_EXCLUDE);
4a085b16 134
827f4d6c 135 for (n = 0; n < pathspec->nr; n++) {
93d93537 136 size_t i = 0, len = 0, item_len;
ef79b1f8
NTND
137 if (pathspec->items[n].magic & PATHSPEC_EXCLUDE)
138 continue;
93d93537
NTND
139 if (pathspec->items[n].magic & PATHSPEC_ICASE)
140 item_len = pathspec->items[n].prefix;
141 else
142 item_len = pathspec->items[n].nowildcard_len;
143 while (i < item_len && (n == 0 || i < max)) {
827f4d6c
NTND
144 char c = pathspec->items[n].match[i];
145 if (c != pathspec->items[0].match[i])
4a085b16
JH
146 break;
147 if (c == '/')
148 len = i + 1;
827f4d6c 149 i++;
4a085b16 150 }
827f4d6c 151 if (n == 0 || len < max) {
4a085b16
JH
152 max = len;
153 if (!max)
154 break;
155 }
3c6a370b 156 }
4a085b16 157 return max;
3c6a370b
LT
158}
159
f950eb95
CB
160/*
161 * Returns a copy of the longest leading path common among all
162 * pathspecs.
163 */
827f4d6c 164char *common_prefix(const struct pathspec *pathspec)
f950eb95
CB
165{
166 unsigned long len = common_prefix_len(pathspec);
167
827f4d6c 168 return len ? xmemdupz(pathspec->items[0].match, len) : NULL;
f950eb95
CB
169}
170
7327d3d1 171int fill_directory(struct dir_struct *dir, const struct pathspec *pathspec)
1d8842d9 172{
4a085b16 173 size_t len;
1d8842d9
LT
174
175 /*
176 * Calculate common prefix for the pathspec, and
177 * use that to optimize the directory walk
178 */
4a085b16 179 len = common_prefix_len(pathspec);
1d8842d9
LT
180
181 /* Read the directory and prune it */
b3920bbd 182 read_directory(dir, pathspec->nr ? pathspec->_raw[0] : "", len, pathspec);
dba2e203 183 return len;
1d8842d9
LT
184}
185
bc96cc87
NTND
186int within_depth(const char *name, int namelen,
187 int depth, int max_depth)
188{
189 const char *cp = name, *cpe = name + namelen;
190
191 while (cp < cpe) {
192 if (*cp++ != '/')
193 continue;
194 depth++;
195 if (depth > max_depth)
196 return 0;
197 }
198 return 1;
199}
200
42b0874a 201#define DO_MATCH_EXCLUDE 1
68690fdd 202#define DO_MATCH_DIRECTORY 2
42b0874a 203
61cf2820
NTND
204/*
205 * Does 'match' match the given name?
206 * A match is found if
207 *
208 * (1) the 'match' string is leading directory of 'name', or
209 * (2) the 'match' string is a wildcard and matches 'name', or
210 * (3) the 'match' string is exactly the same as 'name'.
211 *
212 * and the return value tells which case it was.
213 *
214 * It returns 0 when there is no match.
215 */
216static int match_pathspec_item(const struct pathspec_item *item, int prefix,
42b0874a 217 const char *name, int namelen, unsigned flags)
61cf2820
NTND
218{
219 /* name/namelen has prefix cut off by caller */
220 const char *match = item->match + prefix;
221 int matchlen = item->len - prefix;
222
93d93537
NTND
223 /*
224 * The normal call pattern is:
225 * 1. prefix = common_prefix_len(ps);
226 * 2. prune something, or fill_directory
854b0959 227 * 3. match_pathspec()
93d93537
NTND
228 *
229 * 'prefix' at #1 may be shorter than the command's prefix and
230 * it's ok for #2 to match extra files. Those extras will be
231 * trimmed at #3.
232 *
233 * Suppose the pathspec is 'foo' and '../bar' running from
234 * subdir 'xyz'. The common prefix at #1 will be empty, thanks
235 * to "../". We may have xyz/foo _and_ XYZ/foo after #2. The
236 * user does not want XYZ/foo, only the "foo" part should be
237 * case-insensitive. We need to filter out XYZ/foo here. In
238 * other words, we do not trust the caller on comparing the
239 * prefix part when :(icase) is involved. We do exact
240 * comparison ourselves.
241 *
242 * Normally the caller (common_prefix_len() in fact) does
243 * _exact_ matching on name[-prefix+1..-1] and we do not need
244 * to check that part. Be defensive and check it anyway, in
245 * case common_prefix_len is changed, or a new caller is
246 * introduced that does not use common_prefix_len.
247 *
248 * If the penalty turns out too high when prefix is really
249 * long, maybe change it to
250 * strncmp(match, name, item->prefix - prefix)
251 */
252 if (item->prefix && (item->magic & PATHSPEC_ICASE) &&
253 strncmp(item->match, name - prefix, item->prefix))
254 return 0;
255
61cf2820
NTND
256 /* If the match was just the prefix, we matched */
257 if (!*match)
258 return MATCHED_RECURSIVELY;
259
93d93537 260 if (matchlen <= namelen && !ps_strncmp(item, match, name, matchlen)) {
61cf2820
NTND
261 if (matchlen == namelen)
262 return MATCHED_EXACTLY;
263
264 if (match[matchlen-1] == '/' || name[matchlen] == '/')
265 return MATCHED_RECURSIVELY;
68690fdd
NTND
266 } else if ((flags & DO_MATCH_DIRECTORY) &&
267 match[matchlen - 1] == '/' &&
268 namelen == matchlen - 1 &&
269 !ps_strncmp(item, match, name, namelen))
270 return MATCHED_EXACTLY;
61cf2820 271
5d74762d 272 if (item->nowildcard_len < item->len &&
bd30c2e4 273 !git_fnmatch(item, match, name,
8c6abbcd 274 item->nowildcard_len - prefix))
61cf2820
NTND
275 return MATCHED_FNMATCH;
276
277 return 0;
278}
279
280/*
52ed1894
AS
281 * Given a name and a list of pathspecs, returns the nature of the
282 * closest (i.e. most specific) match of the name to any of the
283 * pathspecs.
284 *
285 * The caller typically calls this multiple times with the same
286 * pathspec and seen[] array but with different name/namelen
287 * (e.g. entries from the index) and is interested in seeing if and
288 * how each pathspec matches all the names it calls this function
289 * with. A mark is left in the seen[] array for each pathspec element
290 * indicating the closest type of match that element achieved, so if
291 * seen[n] remains zero after multiple invocations, that means the nth
292 * pathspec did not match any names, which could indicate that the
293 * user mistyped the nth pathspec.
61cf2820 294 */
854b0959
NTND
295static int do_match_pathspec(const struct pathspec *ps,
296 const char *name, int namelen,
297 int prefix, char *seen,
42b0874a 298 unsigned flags)
61cf2820 299{
42b0874a 300 int i, retval = 0, exclude = flags & DO_MATCH_EXCLUDE;
61cf2820 301
5c6933d2
NTND
302 GUARD_PATHSPEC(ps,
303 PATHSPEC_FROMTOP |
304 PATHSPEC_MAXDEPTH |
bd30c2e4 305 PATHSPEC_LITERAL |
93d93537 306 PATHSPEC_GLOB |
ef79b1f8
NTND
307 PATHSPEC_ICASE |
308 PATHSPEC_EXCLUDE);
8f4f8f45 309
61cf2820 310 if (!ps->nr) {
6330a171
NTND
311 if (!ps->recursive ||
312 !(ps->magic & PATHSPEC_MAXDEPTH) ||
313 ps->max_depth == -1)
61cf2820
NTND
314 return MATCHED_RECURSIVELY;
315
316 if (within_depth(name, namelen, 0, ps->max_depth))
317 return MATCHED_EXACTLY;
318 else
319 return 0;
320 }
321
322 name += prefix;
323 namelen -= prefix;
324
325 for (i = ps->nr - 1; i >= 0; i--) {
326 int how;
ef79b1f8
NTND
327
328 if ((!exclude && ps->items[i].magic & PATHSPEC_EXCLUDE) ||
329 ( exclude && !(ps->items[i].magic & PATHSPEC_EXCLUDE)))
330 continue;
331
61cf2820
NTND
332 if (seen && seen[i] == MATCHED_EXACTLY)
333 continue;
ef79b1f8
NTND
334 /*
335 * Make exclude patterns optional and never report
336 * "pathspec ':(exclude)foo' matches no files"
337 */
338 if (seen && ps->items[i].magic & PATHSPEC_EXCLUDE)
339 seen[i] = MATCHED_FNMATCH;
42b0874a
NTND
340 how = match_pathspec_item(ps->items+i, prefix, name,
341 namelen, flags);
6330a171
NTND
342 if (ps->recursive &&
343 (ps->magic & PATHSPEC_MAXDEPTH) &&
344 ps->max_depth != -1 &&
61cf2820
NTND
345 how && how != MATCHED_FNMATCH) {
346 int len = ps->items[i].len;
347 if (name[len] == '/')
348 len++;
349 if (within_depth(name+len, namelen-len, 0, ps->max_depth))
350 how = MATCHED_EXACTLY;
351 else
352 how = 0;
353 }
354 if (how) {
355 if (retval < how)
356 retval = how;
357 if (seen && seen[i] < how)
358 seen[i] = how;
359 }
360 }
361 return retval;
362}
363
854b0959
NTND
364int match_pathspec(const struct pathspec *ps,
365 const char *name, int namelen,
ae8d0824 366 int prefix, char *seen, int is_dir)
ef79b1f8
NTND
367{
368 int positive, negative;
ae8d0824 369 unsigned flags = is_dir ? DO_MATCH_DIRECTORY : 0;
42b0874a
NTND
370 positive = do_match_pathspec(ps, name, namelen,
371 prefix, seen, flags);
ef79b1f8
NTND
372 if (!(ps->magic & PATHSPEC_EXCLUDE) || !positive)
373 return positive;
42b0874a
NTND
374 negative = do_match_pathspec(ps, name, namelen,
375 prefix, seen,
376 flags | DO_MATCH_EXCLUDE);
ef79b1f8
NTND
377 return negative ? 0 : positive;
378}
379
fcd631ed
NTND
380/*
381 * Return the length of the "simple" part of a path match limiter.
382 */
87323bda 383int simple_length(const char *match)
fcd631ed
NTND
384{
385 int len = -1;
386
387 for (;;) {
388 unsigned char c = *match++;
389 len++;
390 if (c == '\0' || is_glob_special(c))
391 return len;
392 }
393}
394
87323bda 395int no_wildcard(const char *string)
68492fc7 396{
fcd631ed 397 return string[simple_length(string)] == '\0';
68492fc7
LK
398}
399
82dce998
NTND
400void parse_exclude_pattern(const char **pattern,
401 int *patternlen,
402 int *flags,
403 int *nowildcardlen)
84460eec
NTND
404{
405 const char *p = *pattern;
406 size_t i, len;
407
408 *flags = 0;
409 if (*p == '!') {
410 *flags |= EXC_FLAG_NEGATIVE;
411 p++;
412 }
413 len = strlen(p);
414 if (len && p[len - 1] == '/') {
415 len--;
416 *flags |= EXC_FLAG_MUSTBEDIR;
417 }
418 for (i = 0; i < len; i++) {
419 if (p[i] == '/')
420 break;
421 }
422 if (i == len)
423 *flags |= EXC_FLAG_NODIR;
424 *nowildcardlen = simple_length(p);
425 /*
426 * we should have excluded the trailing slash from 'p' too,
427 * but that's one more allocation. Instead just make sure
428 * nowildcardlen does not exceed real patternlen
429 */
430 if (*nowildcardlen > len)
431 *nowildcardlen = len;
432 if (*p == '*' && no_wildcard(p + 1))
433 *flags |= EXC_FLAG_ENDSWITH;
434 *pattern = p;
435 *patternlen = len;
436}
437
453ec4bd 438void add_exclude(const char *string, const char *base,
c04318e4 439 int baselen, struct exclude_list *el, int srcpos)
453ec4bd 440{
d6b8fc30 441 struct exclude *x;
84460eec
NTND
442 int patternlen;
443 int flags;
444 int nowildcardlen;
453ec4bd 445
84460eec
NTND
446 parse_exclude_pattern(&string, &patternlen, &flags, &nowildcardlen);
447 if (flags & EXC_FLAG_MUSTBEDIR) {
d6b8fc30 448 char *s;
84460eec 449 x = xmalloc(sizeof(*x) + patternlen + 1);
4b25d091 450 s = (char *)(x+1);
84460eec
NTND
451 memcpy(s, string, patternlen);
452 s[patternlen] = '\0';
d6b8fc30 453 x->pattern = s;
d6b8fc30
JH
454 } else {
455 x = xmalloc(sizeof(*x));
456 x->pattern = string;
457 }
84460eec
NTND
458 x->patternlen = patternlen;
459 x->nowildcardlen = nowildcardlen;
453ec4bd
LT
460 x->base = base;
461 x->baselen = baselen;
d6b8fc30 462 x->flags = flags;
c04318e4 463 x->srcpos = srcpos;
840fc334
AS
464 ALLOC_GROW(el->excludes, el->nr + 1, el->alloc);
465 el->excludes[el->nr++] = x;
c04318e4 466 x->el = el;
453ec4bd
LT
467}
468
55fe6f51
NTND
469static void *read_skip_worktree_file_from_index(const char *path, size_t *size,
470 struct sha1_stat *sha1_stat)
c28b3d6e
NTND
471{
472 int pos, len;
473 unsigned long sz;
474 enum object_type type;
475 void *data;
c28b3d6e
NTND
476
477 len = strlen(path);
71261027 478 pos = cache_name_pos(path, len);
c28b3d6e
NTND
479 if (pos < 0)
480 return NULL;
71261027 481 if (!ce_skip_worktree(active_cache[pos]))
c28b3d6e 482 return NULL;
71261027 483 data = read_sha1_file(active_cache[pos]->sha1, &type, &sz);
c28b3d6e
NTND
484 if (!data || type != OBJ_BLOB) {
485 free(data);
486 return NULL;
487 }
488 *size = xsize_t(sz);
55fe6f51
NTND
489 if (sha1_stat) {
490 memset(&sha1_stat->stat, 0, sizeof(sha1_stat->stat));
491 hashcpy(sha1_stat->sha1, active_cache[pos]->sha1);
492 }
c28b3d6e
NTND
493 return data;
494}
495
f6198812
AS
496/*
497 * Frees memory within el which was allocated for exclude patterns and
498 * the file buffer. Does not free el itself.
499 */
500void clear_exclude_list(struct exclude_list *el)
0fd0e241
NTND
501{
502 int i;
503
504 for (i = 0; i < el->nr; i++)
505 free(el->excludes[i]);
506 free(el->excludes);
c082df24 507 free(el->filebuf);
0fd0e241
NTND
508
509 el->nr = 0;
510 el->excludes = NULL;
c082df24 511 el->filebuf = NULL;
0fd0e241
NTND
512}
513
7e2e4b37 514static void trim_trailing_spaces(char *buf)
16402b99 515{
e61a6c1d
PB
516 char *p, *last_space = NULL;
517
518 for (p = buf; *p; p++)
519 switch (*p) {
520 case ' ':
521 if (!last_space)
522 last_space = p;
523 break;
524 case '\\':
525 p++;
526 if (!*p)
527 return;
528 /* fallthrough */
529 default:
530 last_space = NULL;
531 }
532
533 if (last_space)
534 *last_space = '\0';
16402b99
NTND
535}
536
55fe6f51
NTND
537/*
538 * Given a file with name "fname", read it (either from disk, or from
539 * the index if "check_index" is non-zero), parse it and store the
540 * exclude rules in "el".
541 *
542 * If "ss" is not NULL, compute SHA-1 of the exclude file and fill
543 * stat data from disk (only valid if add_excludes returns zero). If
544 * ss_valid is non-zero, "ss" must contain good value as input.
545 */
546static int add_excludes(const char *fname, const char *base, int baselen,
547 struct exclude_list *el, int check_index,
548 struct sha1_stat *sha1_stat)
453ec4bd 549{
c470701a 550 struct stat st;
c04318e4 551 int fd, i, lineno = 1;
9d14017a 552 size_t size = 0;
453ec4bd
LT
553 char *buf, *entry;
554
555 fd = open(fname, O_RDONLY);
c28b3d6e 556 if (fd < 0 || fstat(fd, &st) < 0) {
69660731 557 if (errno != ENOENT)
55b38a48 558 warn_on_inaccessible(fname);
c28b3d6e
NTND
559 if (0 <= fd)
560 close(fd);
561 if (!check_index ||
55fe6f51 562 (buf = read_skip_worktree_file_from_index(fname, &size, sha1_stat)) == NULL)
c28b3d6e 563 return -1;
45d76f17
NTND
564 if (size == 0) {
565 free(buf);
566 return 0;
567 }
568 if (buf[size-1] != '\n') {
569 buf = xrealloc(buf, size+1);
570 buf[size++] = '\n';
571 }
d961baa8 572 } else {
c28b3d6e
NTND
573 size = xsize_t(st.st_size);
574 if (size == 0) {
55fe6f51
NTND
575 if (sha1_stat) {
576 fill_stat_data(&sha1_stat->stat, &st);
577 hashcpy(sha1_stat->sha1, EMPTY_BLOB_SHA1_BIN);
578 sha1_stat->valid = 1;
579 }
c28b3d6e
NTND
580 close(fd);
581 return 0;
582 }
45d76f17 583 buf = xmalloc(size+1);
c28b3d6e 584 if (read_in_full(fd, buf, size) != size) {
45d76f17 585 free(buf);
c28b3d6e
NTND
586 close(fd);
587 return -1;
588 }
45d76f17 589 buf[size++] = '\n';
c28b3d6e 590 close(fd);
55fe6f51
NTND
591 if (sha1_stat) {
592 int pos;
593 if (sha1_stat->valid &&
594 !match_stat_data(&sha1_stat->stat, &st))
595 ; /* no content change, ss->sha1 still good */
596 else if (check_index &&
597 (pos = cache_name_pos(fname, strlen(fname))) >= 0 &&
598 !ce_stage(active_cache[pos]) &&
599 ce_uptodate(active_cache[pos]) &&
600 !would_convert_to_git(fname))
601 hashcpy(sha1_stat->sha1, active_cache[pos]->sha1);
602 else
603 hash_sha1_file(buf, size, "blob", sha1_stat->sha1);
604 fill_stat_data(&sha1_stat->stat, &st);
605 sha1_stat->valid = 1;
606 }
6ba78238 607 }
453ec4bd 608
c082df24 609 el->filebuf = buf;
453ec4bd 610 entry = buf;
45d76f17
NTND
611 for (i = 0; i < size; i++) {
612 if (buf[i] == '\n') {
453ec4bd
LT
613 if (entry != buf + i && entry[0] != '#') {
614 buf[i - (i && buf[i-1] == '\r')] = 0;
7e2e4b37 615 trim_trailing_spaces(entry);
c04318e4 616 add_exclude(entry, base, baselen, el, lineno);
453ec4bd 617 }
c04318e4 618 lineno++;
453ec4bd
LT
619 entry = buf + i + 1;
620 }
621 }
622 return 0;
453ec4bd
LT
623}
624
55fe6f51
NTND
625int add_excludes_from_file_to_list(const char *fname, const char *base,
626 int baselen, struct exclude_list *el,
627 int check_index)
628{
629 return add_excludes(fname, base, baselen, el, check_index, NULL);
630}
631
c04318e4
AS
632struct exclude_list *add_exclude_list(struct dir_struct *dir,
633 int group_type, const char *src)
c082df24
AS
634{
635 struct exclude_list *el;
636 struct exclude_list_group *group;
637
638 group = &dir->exclude_list_group[group_type];
639 ALLOC_GROW(group->el, group->nr + 1, group->alloc);
640 el = &group->el[group->nr++];
641 memset(el, 0, sizeof(*el));
c04318e4 642 el->src = src;
c082df24
AS
643 return el;
644}
645
646/*
647 * Used to set up core.excludesfile and .git/info/exclude lists.
648 */
453ec4bd
LT
649void add_excludes_from_file(struct dir_struct *dir, const char *fname)
650{
c082df24 651 struct exclude_list *el;
c04318e4 652 el = add_exclude_list(dir, EXC_FILE, fname);
c082df24 653 if (add_excludes_from_file_to_list(fname, "", 0, el, 0) < 0)
453ec4bd
LT
654 die("cannot use %s as an exclude file", fname);
655}
656
82dce998
NTND
657int match_basename(const char *basename, int basenamelen,
658 const char *pattern, int prefix, int patternlen,
659 int flags)
593cb880
NTND
660{
661 if (prefix == patternlen) {
0b6e56df
JH
662 if (patternlen == basenamelen &&
663 !strncmp_icase(pattern, basename, basenamelen))
593cb880
NTND
664 return 1;
665 } else if (flags & EXC_FLAG_ENDSWITH) {
0b6e56df 666 /* "*literal" matching against "fooliteral" */
593cb880 667 if (patternlen - 1 <= basenamelen &&
0b6e56df
JH
668 !strncmp_icase(pattern + 1,
669 basename + basenamelen - (patternlen - 1),
670 patternlen - 1))
593cb880
NTND
671 return 1;
672 } else {
0b6e56df
JH
673 if (fnmatch_icase_mem(pattern, patternlen,
674 basename, basenamelen,
675 0) == 0)
593cb880
NTND
676 return 1;
677 }
678 return 0;
679}
680
82dce998
NTND
681int match_pathname(const char *pathname, int pathlen,
682 const char *base, int baselen,
683 const char *pattern, int prefix, int patternlen,
684 int flags)
b5592632
NTND
685{
686 const char *name;
687 int namelen;
688
689 /*
690 * match with FNM_PATHNAME; the pattern has base implicitly
691 * in front of it.
692 */
693 if (*pattern == '/') {
694 pattern++;
982ac873 695 patternlen--;
b5592632
NTND
696 prefix--;
697 }
698
699 /*
700 * baselen does not count the trailing slash. base[] may or
701 * may not end with a trailing slash though.
702 */
703 if (pathlen < baselen + 1 ||
704 (baselen && pathname[baselen] != '/') ||
705 strncmp_icase(pathname, base, baselen))
706 return 0;
707
708 namelen = baselen ? pathlen - baselen - 1 : pathlen;
709 name = pathname + pathlen - namelen;
710
711 if (prefix) {
712 /*
713 * if the non-wildcard part is longer than the
714 * remaining pathname, surely it cannot match.
715 */
716 if (prefix > namelen)
717 return 0;
718
719 if (strncmp_icase(pattern, name, prefix))
720 return 0;
721 pattern += prefix;
ab3aebc1 722 patternlen -= prefix;
b5592632
NTND
723 name += prefix;
724 namelen -= prefix;
ab3aebc1
JK
725
726 /*
727 * If the whole pattern did not have a wildcard,
728 * then our prefix match is all we need; we
729 * do not need to call fnmatch at all.
730 */
731 if (!patternlen && !namelen)
732 return 1;
b5592632
NTND
733 }
734
ab3aebc1
JK
735 return fnmatch_icase_mem(pattern, patternlen,
736 name, namelen,
f30366b2 737 WM_PATHNAME) == 0;
b5592632
NTND
738}
739
578cd7c3
AS
740/*
741 * Scan the given exclude list in reverse to see whether pathname
742 * should be ignored. The first match (i.e. the last on the list), if
743 * any, determines the fate. Returns the exclude_list element which
744 * matched, or NULL for undecided.
453ec4bd 745 */
578cd7c3
AS
746static struct exclude *last_exclude_matching_from_list(const char *pathname,
747 int pathlen,
748 const char *basename,
749 int *dtype,
750 struct exclude_list *el)
453ec4bd
LT
751{
752 int i;
753
35a94d44 754 if (!el->nr)
578cd7c3 755 return NULL; /* undefined */
d6b8fc30 756
35a94d44
NTND
757 for (i = el->nr - 1; 0 <= i; i--) {
758 struct exclude *x = el->excludes[i];
b5592632 759 const char *exclude = x->pattern;
b5592632 760 int prefix = x->nowildcardlen;
35a94d44
NTND
761
762 if (x->flags & EXC_FLAG_MUSTBEDIR) {
763 if (*dtype == DT_UNKNOWN)
764 *dtype = get_dtype(NULL, pathname, pathlen);
765 if (*dtype != DT_DIR)
766 continue;
767 }
d6b8fc30 768
35a94d44 769 if (x->flags & EXC_FLAG_NODIR) {
593cb880
NTND
770 if (match_basename(basename,
771 pathlen - (basename - pathname),
772 exclude, prefix, x->patternlen,
773 x->flags))
578cd7c3 774 return x;
35a94d44
NTND
775 continue;
776 }
777
b5592632
NTND
778 assert(x->baselen == 0 || x->base[x->baselen - 1] == '/');
779 if (match_pathname(pathname, pathlen,
780 x->base, x->baselen ? x->baselen - 1 : 0,
781 exclude, prefix, x->patternlen, x->flags))
578cd7c3 782 return x;
453ec4bd 783 }
578cd7c3
AS
784 return NULL; /* undecided */
785}
786
787/*
788 * Scan the list and let the last match determine the fate.
789 * Return 1 for exclude, 0 for include and -1 for undecided.
790 */
791int is_excluded_from_list(const char *pathname,
792 int pathlen, const char *basename, int *dtype,
793 struct exclude_list *el)
794{
795 struct exclude *exclude;
796 exclude = last_exclude_matching_from_list(pathname, pathlen, basename, dtype, el);
797 if (exclude)
798 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
453ec4bd
LT
799 return -1; /* undecided */
800}
801
46aa2f95
KB
802static struct exclude *last_exclude_matching_from_lists(struct dir_struct *dir,
803 const char *pathname, int pathlen, const char *basename,
804 int *dtype_p)
805{
806 int i, j;
807 struct exclude_list_group *group;
808 struct exclude *exclude;
809 for (i = EXC_CMDL; i <= EXC_FILE; i++) {
810 group = &dir->exclude_list_group[i];
811 for (j = group->nr - 1; j >= 0; j--) {
812 exclude = last_exclude_matching_from_list(
813 pathname, pathlen, basename, dtype_p,
814 &group->el[j]);
815 if (exclude)
816 return exclude;
817 }
818 }
819 return NULL;
820}
821
6cd5c582
KB
822/*
823 * Loads the per-directory exclude list for the substring of base
824 * which has a char length of baselen.
825 */
826static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
827{
828 struct exclude_list_group *group;
829 struct exclude_list *el;
830 struct exclude_stack *stk = NULL;
831 int current;
832
6cd5c582
KB
833 group = &dir->exclude_list_group[EXC_DIRS];
834
d961baa8
NTND
835 /*
836 * Pop the exclude lists from the EXCL_DIRS exclude_list_group
6cd5c582 837 * which originate from directories not in the prefix of the
d961baa8
NTND
838 * path being checked.
839 */
6cd5c582
KB
840 while ((stk = dir->exclude_stack) != NULL) {
841 if (stk->baselen <= baselen &&
aceb9429 842 !strncmp(dir->basebuf.buf, base, stk->baselen))
6cd5c582
KB
843 break;
844 el = &group->el[dir->exclude_stack->exclude_ix];
845 dir->exclude_stack = stk->prev;
95c6f271 846 dir->exclude = NULL;
aceb9429 847 free((char *)el->src); /* see strbuf_detach() below */
6cd5c582
KB
848 clear_exclude_list(el);
849 free(stk);
850 group->nr--;
851 }
852
95c6f271
KB
853 /* Skip traversing into sub directories if the parent is excluded */
854 if (dir->exclude)
855 return;
856
aceb9429
NTND
857 /*
858 * Lazy initialization. All call sites currently just
859 * memset(dir, 0, sizeof(*dir)) before use. Changing all of
860 * them seems lots of work for little benefit.
861 */
862 if (!dir->basebuf.buf)
863 strbuf_init(&dir->basebuf, PATH_MAX);
864
6cd5c582
KB
865 /* Read from the parent directories and push them down. */
866 current = stk ? stk->baselen : -1;
aceb9429 867 strbuf_setlen(&dir->basebuf, current < 0 ? 0 : current);
6cd5c582 868 while (current < baselen) {
6cd5c582
KB
869 const char *cp;
870
03e11a71 871 stk = xcalloc(1, sizeof(*stk));
6cd5c582
KB
872 if (current < 0) {
873 cp = base;
874 current = 0;
d961baa8 875 } else {
6cd5c582
KB
876 cp = strchr(base + current + 1, '/');
877 if (!cp)
878 die("oops in prep_exclude");
879 cp++;
880 }
881 stk->prev = dir->exclude_stack;
882 stk->baselen = cp - base;
95c6f271
KB
883 stk->exclude_ix = group->nr;
884 el = add_exclude_list(dir, EXC_DIRS, NULL);
aceb9429
NTND
885 strbuf_add(&dir->basebuf, base + current, stk->baselen - current);
886 assert(stk->baselen == dir->basebuf.len);
95c6f271
KB
887
888 /* Abort if the directory is excluded */
889 if (stk->baselen) {
890 int dt = DT_DIR;
aceb9429 891 dir->basebuf.buf[stk->baselen - 1] = 0;
95c6f271 892 dir->exclude = last_exclude_matching_from_lists(dir,
aceb9429
NTND
893 dir->basebuf.buf, stk->baselen - 1,
894 dir->basebuf.buf + current, &dt);
895 dir->basebuf.buf[stk->baselen - 1] = '/';
c3c327de
KB
896 if (dir->exclude &&
897 dir->exclude->flags & EXC_FLAG_NEGATIVE)
898 dir->exclude = NULL;
95c6f271 899 if (dir->exclude) {
95c6f271
KB
900 dir->exclude_stack = stk;
901 return;
902 }
903 }
904
aceb9429
NTND
905 /* Try to read per-directory file */
906 if (dir->exclude_per_dir) {
95c6f271
KB
907 /*
908 * dir->basebuf gets reused by the traversal, but we
909 * need fname to remain unchanged to ensure the src
910 * member of each struct exclude correctly
911 * back-references its source file. Other invocations
912 * of add_exclude_list provide stable strings, so we
aceb9429 913 * strbuf_detach() and free() here in the caller.
95c6f271 914 */
aceb9429
NTND
915 struct strbuf sb = STRBUF_INIT;
916 strbuf_addbuf(&sb, &dir->basebuf);
917 strbuf_addstr(&sb, dir->exclude_per_dir);
918 el->src = strbuf_detach(&sb, NULL);
919 add_excludes_from_file_to_list(el->src, el->src,
920 stk->baselen, el, 1);
95c6f271 921 }
6cd5c582
KB
922 dir->exclude_stack = stk;
923 current = stk->baselen;
924 }
aceb9429 925 strbuf_setlen(&dir->basebuf, baselen);
6cd5c582
KB
926}
927
f4cd69a6
AS
928/*
929 * Loads the exclude lists for the directory containing pathname, then
930 * scans all exclude lists to determine whether pathname is excluded.
931 * Returns the exclude_list element which matched, or NULL for
932 * undecided.
933 */
b07bc8c8 934struct exclude *last_exclude_matching(struct dir_struct *dir,
f4cd69a6
AS
935 const char *pathname,
936 int *dtype_p)
453ec4bd
LT
937{
938 int pathlen = strlen(pathname);
68492fc7
LK
939 const char *basename = strrchr(pathname, '/');
940 basename = (basename) ? basename+1 : pathname;
453ec4bd 941
63d285c8 942 prep_exclude(dir, pathname, basename-pathname);
c082df24 943
95c6f271
KB
944 if (dir->exclude)
945 return dir->exclude;
946
46aa2f95
KB
947 return last_exclude_matching_from_lists(dir, pathname, pathlen,
948 basename, dtype_p);
f4cd69a6
AS
949}
950
951/*
952 * Loads the exclude lists for the directory containing pathname, then
953 * scans all exclude lists to determine whether pathname is excluded.
954 * Returns 1 if true, otherwise 0.
955 */
b07bc8c8 956int is_excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
f4cd69a6
AS
957{
958 struct exclude *exclude =
959 last_exclude_matching(dir, pathname, dtype_p);
960 if (exclude)
961 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
453ec4bd
LT
962 return 0;
963}
964
f3fa1838
JH
965static struct dir_entry *dir_entry_new(const char *pathname, int len)
966{
453ec4bd
LT
967 struct dir_entry *ent;
968
453ec4bd
LT
969 ent = xmalloc(sizeof(*ent) + len + 1);
970 ent->len = len;
971 memcpy(ent->name, pathname, len);
972 ent->name[len] = 0;
4d06f8ac 973 return ent;
453ec4bd
LT
974}
975
159b3212 976static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
6815e569 977{
ebbd7439 978 if (cache_file_exists(pathname, len, ignore_case))
6815e569
JK
979 return NULL;
980
25fd2f7a 981 ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
6815e569
JK
982 return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
983}
984
108da0db 985struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
2abd31b0 986{
6e4f981f 987 if (!cache_name_is_other(pathname, len))
2abd31b0
JK
988 return NULL;
989
25fd2f7a 990 ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
2abd31b0
JK
991 return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
992}
993
09595258
LT
994enum exist_status {
995 index_nonexistent = 0,
996 index_directory,
4b05548f 997 index_gitdir
09595258
LT
998};
999
5102c617 1000/*
71261027 1001 * Do not use the alphabetically sorted index to look up
5102c617 1002 * the directory name; instead, use the case insensitive
ebbd7439 1003 * directory hash.
5102c617
JJ
1004 */
1005static enum exist_status directory_exists_in_index_icase(const char *dirname, int len)
1006{
d28eec26 1007 const struct cache_entry *ce = cache_dir_exists(dirname, len);
5102c617
JJ
1008 unsigned char endchar;
1009
1010 if (!ce)
1011 return index_nonexistent;
1012 endchar = ce->name[len];
1013
1014 /*
1015 * The cache_entry structure returned will contain this dirname
1016 * and possibly additional path components.
1017 */
1018 if (endchar == '/')
1019 return index_directory;
1020
1021 /*
1022 * If there are no additional path components, then this cache_entry
1023 * represents a submodule. Submodules, despite being directories,
1024 * are stored in the cache without a closing slash.
1025 */
1026 if (!endchar && S_ISGITLINK(ce->ce_mode))
1027 return index_gitdir;
1028
1029 /* This should never be hit, but it exists just in case. */
1030 return index_nonexistent;
1031}
1032
09595258
LT
1033/*
1034 * The index sorts alphabetically by entry name, which
1035 * means that a gitlink sorts as '\0' at the end, while
1036 * a directory (which is defined not as an entry, but as
1037 * the files it contains) will sort with the '/' at the
1038 * end.
1039 */
1040static enum exist_status directory_exists_in_index(const char *dirname, int len)
453ec4bd 1041{
5102c617
JJ
1042 int pos;
1043
1044 if (ignore_case)
1045 return directory_exists_in_index_icase(dirname, len);
1046
1047 pos = cache_name_pos(dirname, len);
09595258
LT
1048 if (pos < 0)
1049 pos = -pos-1;
1050 while (pos < active_nr) {
9c5e6c80 1051 const struct cache_entry *ce = active_cache[pos++];
09595258
LT
1052 unsigned char endchar;
1053
1054 if (strncmp(ce->name, dirname, len))
1055 break;
1056 endchar = ce->name[len];
1057 if (endchar > '/')
1058 break;
1059 if (endchar == '/')
1060 return index_directory;
7a51ed66 1061 if (!endchar && S_ISGITLINK(ce->ce_mode))
09595258
LT
1062 return index_gitdir;
1063 }
1064 return index_nonexistent;
1065}
1066
1067/*
1068 * When we find a directory when traversing the filesystem, we
1069 * have three distinct cases:
1070 *
1071 * - ignore it
1072 * - see it as a directory
1073 * - recurse into it
1074 *
1075 * and which one we choose depends on a combination of existing
1076 * git index contents and the flags passed into the directory
1077 * traversal routine.
1078 *
1079 * Case 1: If we *already* have entries in the index under that
5bd8e2d8
KB
1080 * directory name, we always recurse into the directory to see
1081 * all the files.
09595258
LT
1082 *
1083 * Case 2: If we *already* have that directory name as a gitlink,
1084 * we always continue to see it as a gitlink, regardless of whether
1085 * there is an actual git directory there or not (it might not
1086 * be checked out as a subproject!)
1087 *
1088 * Case 3: if we didn't have it in the index previously, we
1089 * have a few sub-cases:
1090 *
1091 * (a) if "show_other_directories" is true, we show it as
1092 * just a directory, unless "hide_empty_directories" is
defd7c7b
KB
1093 * also true, in which case we need to check if it contains any
1094 * untracked and / or ignored files.
09595258 1095 * (b) if it looks like a git directory, and we don't have
302b9282 1096 * 'no_gitlinks' set we treat it as a gitlink, and show it
09595258
LT
1097 * as a directory.
1098 * (c) otherwise, we recurse into it.
1099 */
defd7c7b 1100static enum path_treatment treat_directory(struct dir_struct *dir,
721ac4ed 1101 const char *dirname, int len, int exclude,
09595258
LT
1102 const struct path_simplify *simplify)
1103{
1104 /* The "len-1" is to strip the final '/' */
1105 switch (directory_exists_in_index(dirname, len-1)) {
1106 case index_directory:
defd7c7b 1107 return path_recurse;
09595258
LT
1108
1109 case index_gitdir:
26c986e1 1110 return path_none;
09595258
LT
1111
1112 case index_nonexistent:
7c4c97c0 1113 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
09595258 1114 break;
7c4c97c0 1115 if (!(dir->flags & DIR_NO_GITLINKS)) {
09595258
LT
1116 unsigned char sha1[20];
1117 if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
defd7c7b 1118 return path_untracked;
09595258 1119 }
defd7c7b 1120 return path_recurse;
09595258
LT
1121 }
1122
1123 /* This is the "show_other_directories" case */
721ac4ed 1124
184d2a8e 1125 if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
defd7c7b
KB
1126 return exclude ? path_excluded : path_untracked;
1127
1128 return read_directory_recursive(dir, dirname, len, 1, simplify);
453ec4bd
LT
1129}
1130
9fc42d60
LT
1131/*
1132 * This is an inexact early pruning of any recursive directory
1133 * reading - if the path cannot possibly be in the pathspec,
1134 * return true, and we'll skip it early.
1135 */
1136static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
1137{
1138 if (simplify) {
1139 for (;;) {
1140 const char *match = simplify->path;
1141 int len = simplify->len;
1142
1143 if (!match)
1144 break;
1145 if (len > pathlen)
1146 len = pathlen;
1147 if (!memcmp(path, match, len))
1148 return 0;
1149 simplify++;
1150 }
1151 return 1;
1152 }
1153 return 0;
1154}
1155
29209cbe
JK
1156/*
1157 * This function tells us whether an excluded path matches a
1158 * list of "interesting" pathspecs. That is, whether a path matched
1159 * by any of the pathspecs could possibly be ignored by excluding
1160 * the specified path. This can happen if:
1161 *
1162 * 1. the path is mentioned explicitly in the pathspec
1163 *
1164 * 2. the path is a directory prefix of some element in the
1165 * pathspec
1166 */
1167static int exclude_matches_pathspec(const char *path, int len,
1168 const struct path_simplify *simplify)
e96980ef
JK
1169{
1170 if (simplify) {
1171 for (; simplify->path; simplify++) {
1172 if (len == simplify->len
1173 && !memcmp(path, simplify->path, len))
1174 return 1;
29209cbe
JK
1175 if (len < simplify->len
1176 && simplify->path[len] == '/'
1177 && !memcmp(path, simplify->path, len))
1178 return 1;
e96980ef
JK
1179 }
1180 }
1181 return 0;
1182}
1183
443e061a
LT
1184static int get_index_dtype(const char *path, int len)
1185{
1186 int pos;
9c5e6c80 1187 const struct cache_entry *ce;
443e061a 1188
ebbd7439 1189 ce = cache_file_exists(path, len, 0);
443e061a
LT
1190 if (ce) {
1191 if (!ce_uptodate(ce))
1192 return DT_UNKNOWN;
1193 if (S_ISGITLINK(ce->ce_mode))
1194 return DT_DIR;
1195 /*
1196 * Nobody actually cares about the
1197 * difference between DT_LNK and DT_REG
1198 */
1199 return DT_REG;
1200 }
1201
1202 /* Try to look it up as a directory */
1203 pos = cache_name_pos(path, len);
1204 if (pos >= 0)
1205 return DT_UNKNOWN;
1206 pos = -pos-1;
1207 while (pos < active_nr) {
1208 ce = active_cache[pos++];
1209 if (strncmp(ce->name, path, len))
1210 break;
1211 if (ce->name[len] > '/')
1212 break;
1213 if (ce->name[len] < '/')
1214 continue;
1215 if (!ce_uptodate(ce))
1216 break; /* continue? */
1217 return DT_DIR;
1218 }
1219 return DT_UNKNOWN;
1220}
1221
caa6b782 1222static int get_dtype(struct dirent *de, const char *path, int len)
07134421 1223{
6831a88a 1224 int dtype = de ? DTYPE(de) : DT_UNKNOWN;
07134421
LT
1225 struct stat st;
1226
1227 if (dtype != DT_UNKNOWN)
1228 return dtype;
443e061a
LT
1229 dtype = get_index_dtype(path, len);
1230 if (dtype != DT_UNKNOWN)
1231 return dtype;
1232 if (lstat(path, &st))
07134421
LT
1233 return dtype;
1234 if (S_ISREG(st.st_mode))
1235 return DT_REG;
1236 if (S_ISDIR(st.st_mode))
1237 return DT_DIR;
1238 if (S_ISLNK(st.st_mode))
1239 return DT_LNK;
1240 return dtype;
1241}
1242
16e2cfa9 1243static enum path_treatment treat_one_path(struct dir_struct *dir,
49dc2cc2 1244 struct strbuf *path,
16e2cfa9
JH
1245 const struct path_simplify *simplify,
1246 int dtype, struct dirent *de)
53cc5356 1247{
8aaf8d77 1248 int exclude;
ebbd7439 1249 int has_path_in_index = !!cache_file_exists(path->buf, path->len, ignore_case);
2eac2a4c 1250
8aaf8d77
KB
1251 if (dtype == DT_UNKNOWN)
1252 dtype = get_dtype(de, path->buf, path->len);
1253
1254 /* Always exclude indexed files */
2eac2a4c 1255 if (dtype != DT_DIR && has_path_in_index)
defd7c7b 1256 return path_none;
8aaf8d77 1257
2eac2a4c
JH
1258 /*
1259 * When we are looking at a directory P in the working tree,
1260 * there are three cases:
1261 *
1262 * (1) P exists in the index. Everything inside the directory P in
1263 * the working tree needs to go when P is checked out from the
1264 * index.
1265 *
1266 * (2) P does not exist in the index, but there is P/Q in the index.
1267 * We know P will stay a directory when we check out the contents
1268 * of the index, but we do not know yet if there is a directory
1269 * P/Q in the working tree to be killed, so we need to recurse.
1270 *
1271 * (3) P does not exist in the index, and there is no P/Q in the index
1272 * to require P to be a directory, either. Only in this case, we
1273 * know that everything inside P will not be killed without
1274 * recursing.
1275 */
1276 if ((dir->flags & DIR_COLLECT_KILLED_ONLY) &&
1277 (dtype == DT_DIR) &&
de372b1b
ES
1278 !has_path_in_index &&
1279 (directory_exists_in_index(path->buf, path->len) == index_nonexistent))
1280 return path_none;
8aaf8d77
KB
1281
1282 exclude = is_excluded(dir, path->buf, &dtype);
53cc5356
JH
1283
1284 /*
1285 * Excluded? If we don't explicitly want to show
1286 * ignored files, ignore it
1287 */
0aaf62b6 1288 if (exclude && !(dir->flags & (DIR_SHOW_IGNORED|DIR_SHOW_IGNORED_TOO)))
defd7c7b 1289 return path_excluded;
53cc5356 1290
53cc5356
JH
1291 switch (dtype) {
1292 default:
defd7c7b 1293 return path_none;
53cc5356 1294 case DT_DIR:
49dc2cc2 1295 strbuf_addch(path, '/');
defd7c7b
KB
1296 return treat_directory(dir, path->buf, path->len, exclude,
1297 simplify);
53cc5356
JH
1298 case DT_REG:
1299 case DT_LNK:
defd7c7b 1300 return exclude ? path_excluded : path_untracked;
53cc5356 1301 }
53cc5356
JH
1302}
1303
16e2cfa9
JH
1304static enum path_treatment treat_path(struct dir_struct *dir,
1305 struct dirent *de,
49dc2cc2 1306 struct strbuf *path,
16e2cfa9 1307 int baselen,
49dc2cc2 1308 const struct path_simplify *simplify)
16e2cfa9
JH
1309{
1310 int dtype;
1311
1312 if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
defd7c7b 1313 return path_none;
49dc2cc2
RS
1314 strbuf_setlen(path, baselen);
1315 strbuf_addstr(path, de->d_name);
1316 if (simplify_away(path->buf, path->len, simplify))
defd7c7b 1317 return path_none;
16e2cfa9
JH
1318
1319 dtype = DTYPE(de);
49dc2cc2 1320 return treat_one_path(dir, path, simplify, dtype, de);
16e2cfa9
JH
1321}
1322
453ec4bd
LT
1323/*
1324 * Read a directory tree. We currently ignore anything but
1325 * directories, regular files and symlinks. That's because git
1326 * doesn't handle them at all yet. Maybe that will change some
1327 * day.
1328 *
1329 * Also, we ignore the name ".git" (even if it is not a directory).
1330 * That likely will not change.
defd7c7b
KB
1331 *
1332 * Returns the most significant path_treatment value encountered in the scan.
453ec4bd 1333 */
defd7c7b 1334static enum path_treatment read_directory_recursive(struct dir_struct *dir,
53cc5356
JH
1335 const char *base, int baselen,
1336 int check_only,
1337 const struct path_simplify *simplify)
453ec4bd 1338{
1528d247 1339 DIR *fdir;
defd7c7b 1340 enum path_treatment state, subdir_state, dir_state = path_none;
02cb6753 1341 struct dirent *de;
bef36921 1342 struct strbuf path = STRBUF_INIT;
453ec4bd 1343
bef36921 1344 strbuf_add(&path, base, baselen);
02cb6753 1345
1528d247
RS
1346 fdir = opendir(path.len ? path.buf : ".");
1347 if (!fdir)
1348 goto out;
1349
02cb6753 1350 while ((de = readdir(fdir)) != NULL) {
defd7c7b
KB
1351 /* check how the file or directory should be treated */
1352 state = treat_path(dir, de, &path, baselen, simplify);
1353 if (state > dir_state)
1354 dir_state = state;
1355
1356 /* recurse into subdir if instructed by treat_path */
1357 if (state == path_recurse) {
1358 subdir_state = read_directory_recursive(dir, path.buf,
289ff559 1359 path.len, check_only, simplify);
defd7c7b
KB
1360 if (subdir_state > dir_state)
1361 dir_state = subdir_state;
1362 }
1363
1364 if (check_only) {
1365 /* abort early if maximum state has been reached */
1366 if (dir_state == path_untracked)
1367 break;
1368 /* skip the dir_add_* part */
02cb6753 1369 continue;
453ec4bd 1370 }
defd7c7b
KB
1371
1372 /* add the path to the appropriate result list */
1373 switch (state) {
1374 case path_excluded:
1375 if (dir->flags & DIR_SHOW_IGNORED)
1376 dir_add_name(dir, path.buf, path.len);
0aaf62b6
KB
1377 else if ((dir->flags & DIR_SHOW_IGNORED_TOO) ||
1378 ((dir->flags & DIR_COLLECT_IGNORED) &&
1379 exclude_matches_pathspec(path.buf, path.len,
1380 simplify)))
1381 dir_add_ignored(dir, path.buf, path.len);
defd7c7b
KB
1382 break;
1383
1384 case path_untracked:
1385 if (!(dir->flags & DIR_SHOW_IGNORED))
1386 dir_add_name(dir, path.buf, path.len);
1528d247 1387 break;
defd7c7b
KB
1388
1389 default:
1390 break;
1391 }
453ec4bd 1392 }
02cb6753 1393 closedir(fdir);
1528d247 1394 out:
bef36921 1395 strbuf_release(&path);
453ec4bd 1396
defd7c7b 1397 return dir_state;
453ec4bd
LT
1398}
1399
1400static int cmp_name(const void *p1, const void *p2)
1401{
1402 const struct dir_entry *e1 = *(const struct dir_entry **)p1;
1403 const struct dir_entry *e2 = *(const struct dir_entry **)p2;
1404
ccdd4a0f 1405 return name_compare(e1->name, e1->len, e2->name, e2->len);
453ec4bd
LT
1406}
1407
9fc42d60
LT
1408static struct path_simplify *create_simplify(const char **pathspec)
1409{
1410 int nr, alloc = 0;
1411 struct path_simplify *simplify = NULL;
1412
1413 if (!pathspec)
1414 return NULL;
1415
1416 for (nr = 0 ; ; nr++) {
1417 const char *match;
9af49f82 1418 ALLOC_GROW(simplify, nr + 1, alloc);
9fc42d60
LT
1419 match = *pathspec++;
1420 if (!match)
1421 break;
1422 simplify[nr].path = match;
1423 simplify[nr].len = simple_length(match);
1424 }
1425 simplify[nr].path = NULL;
1426 simplify[nr].len = 0;
1427 return simplify;
1428}
1429
1430static void free_simplify(struct path_simplify *simplify)
1431{
8e0f7003 1432 free(simplify);
9fc42d60
LT
1433}
1434
48ffef96
JH
1435static int treat_leading_path(struct dir_struct *dir,
1436 const char *path, int len,
1437 const struct path_simplify *simplify)
1438{
49dc2cc2
RS
1439 struct strbuf sb = STRBUF_INIT;
1440 int baselen, rc = 0;
48ffef96 1441 const char *cp;
be8a84c5 1442 int old_flags = dir->flags;
48ffef96
JH
1443
1444 while (len && path[len - 1] == '/')
1445 len--;
1446 if (!len)
1447 return 1;
1448 baselen = 0;
be8a84c5 1449 dir->flags &= ~DIR_SHOW_OTHER_DIRECTORIES;
48ffef96
JH
1450 while (1) {
1451 cp = path + baselen + !!baselen;
1452 cp = memchr(cp, '/', path + len - cp);
1453 if (!cp)
1454 baselen = len;
1455 else
1456 baselen = cp - path;
49dc2cc2
RS
1457 strbuf_setlen(&sb, 0);
1458 strbuf_add(&sb, path, baselen);
1459 if (!is_directory(sb.buf))
1460 break;
1461 if (simplify_away(sb.buf, sb.len, simplify))
1462 break;
1463 if (treat_one_path(dir, &sb, simplify,
defd7c7b 1464 DT_DIR, NULL) == path_none)
49dc2cc2
RS
1465 break; /* do not recurse into it */
1466 if (len <= baselen) {
1467 rc = 1;
1468 break; /* finished checking */
1469 }
48ffef96 1470 }
49dc2cc2 1471 strbuf_release(&sb);
be8a84c5 1472 dir->flags = old_flags;
49dc2cc2 1473 return rc;
48ffef96
JH
1474}
1475
7327d3d1 1476int read_directory(struct dir_struct *dir, const char *path, int len, const struct pathspec *pathspec)
9fc42d60 1477{
725b0605 1478 struct path_simplify *simplify;
b4189aa8 1479
7327d3d1
NTND
1480 /*
1481 * Check out create_simplify()
1482 */
1483 if (pathspec)
5c6933d2
NTND
1484 GUARD_PATHSPEC(pathspec,
1485 PATHSPEC_FROMTOP |
1486 PATHSPEC_MAXDEPTH |
bd30c2e4 1487 PATHSPEC_LITERAL |
93d93537 1488 PATHSPEC_GLOB |
ef79b1f8
NTND
1489 PATHSPEC_ICASE |
1490 PATHSPEC_EXCLUDE);
7327d3d1 1491
dba2e203 1492 if (has_symlink_leading_path(path, len))
725b0605
JH
1493 return dir->nr;
1494
ef79b1f8
NTND
1495 /*
1496 * exclude patterns are treated like positive ones in
1497 * create_simplify. Usually exclude patterns should be a
1498 * subset of positive ones, which has no impacts on
1499 * create_simplify().
1500 */
b3920bbd 1501 simplify = create_simplify(pathspec ? pathspec->_raw : NULL);
48ffef96
JH
1502 if (!len || treat_leading_path(dir, path, len, simplify))
1503 read_directory_recursive(dir, path, len, 0, simplify);
9fc42d60 1504 free_simplify(simplify);
453ec4bd 1505 qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
2abd31b0 1506 qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
453ec4bd
LT
1507 return dir->nr;
1508}
c91f0d92 1509
686a4a06 1510int file_exists(const char *f)
c91f0d92 1511{
686a4a06 1512 struct stat sb;
a50f9fc5 1513 return lstat(f, &sb) == 0;
c91f0d92 1514}
e6636747
JS
1515
1516/*
9b125da4
NTND
1517 * Given two normalized paths (a trailing slash is ok), if subdir is
1518 * outside dir, return -1. Otherwise return the offset in subdir that
1519 * can be used as relative path to dir.
e6636747 1520 */
9b125da4 1521int dir_inside_of(const char *subdir, const char *dir)
e6636747 1522{
9b125da4 1523 int offset = 0;
e6636747 1524
9b125da4 1525 assert(dir && subdir && *dir && *subdir);
e6636747 1526
9b125da4 1527 while (*dir && *subdir && *dir == *subdir) {
e6636747 1528 dir++;
9b125da4
NTND
1529 subdir++;
1530 offset++;
490544b1 1531 }
9b125da4
NTND
1532
1533 /* hel[p]/me vs hel[l]/yeah */
1534 if (*dir && *subdir)
1535 return -1;
1536
1537 if (!*subdir)
1538 return !*dir ? offset : -1; /* same dir */
1539
1540 /* foo/[b]ar vs foo/[] */
1541 if (is_dir_sep(dir[-1]))
1542 return is_dir_sep(subdir[-1]) ? offset : -1;
1543
1544 /* foo[/]bar vs foo[] */
1545 return is_dir_sep(*subdir) ? offset + 1 : -1;
e6636747
JS
1546}
1547
1548int is_inside_dir(const char *dir)
1549{
56b9f6e7
RS
1550 char *cwd;
1551 int rc;
1552
b892913d
NTND
1553 if (!dir)
1554 return 0;
56b9f6e7
RS
1555
1556 cwd = xgetcwd();
1557 rc = (dir_inside_of(cwd, dir) >= 0);
1558 free(cwd);
1559 return rc;
e6636747 1560}
7155b727 1561
55892d23
AP
1562int is_empty_dir(const char *path)
1563{
1564 DIR *dir = opendir(path);
1565 struct dirent *e;
1566 int ret = 1;
1567
1568 if (!dir)
1569 return 0;
1570
1571 while ((e = readdir(dir)) != NULL)
1572 if (!is_dot_or_dotdot(e->d_name)) {
1573 ret = 0;
1574 break;
1575 }
1576
1577 closedir(dir);
1578 return ret;
1579}
1580
ae2f203e 1581static int remove_dir_recurse(struct strbuf *path, int flag, int *kept_up)
7155b727 1582{
a0f4afbe 1583 DIR *dir;
7155b727 1584 struct dirent *e;
ae2f203e 1585 int ret = 0, original_len = path->len, len, kept_down = 0;
a0f4afbe 1586 int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
c844a803 1587 int keep_toplevel = (flag & REMOVE_DIR_KEEP_TOPLEVEL);
a0f4afbe 1588 unsigned char submodule_head[20];
7155b727 1589
a0f4afbe 1590 if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
ae2f203e 1591 !resolve_gitlink_ref(path->buf, "HEAD", submodule_head)) {
a0f4afbe 1592 /* Do not descend and nuke a nested git work tree. */
ae2f203e
JH
1593 if (kept_up)
1594 *kept_up = 1;
a0f4afbe 1595 return 0;
ae2f203e 1596 }
a0f4afbe 1597
ae2f203e 1598 flag &= ~REMOVE_DIR_KEEP_TOPLEVEL;
a0f4afbe 1599 dir = opendir(path->buf);
c844a803 1600 if (!dir) {
863808cd
MH
1601 if (errno == ENOENT)
1602 return keep_toplevel ? -1 : 0;
1603 else if (errno == EACCES && !keep_toplevel)
ecb2c282
MH
1604 /*
1605 * An empty dir could be removable even if it
1606 * is unreadable:
1607 */
c844a803
JH
1608 return rmdir(path->buf);
1609 else
1610 return -1;
1611 }
7155b727
JS
1612 if (path->buf[original_len - 1] != '/')
1613 strbuf_addch(path, '/');
1614
1615 len = path->len;
1616 while ((e = readdir(dir)) != NULL) {
1617 struct stat st;
8ca12c0d
AP
1618 if (is_dot_or_dotdot(e->d_name))
1619 continue;
7155b727
JS
1620
1621 strbuf_setlen(path, len);
1622 strbuf_addstr(path, e->d_name);
863808cd
MH
1623 if (lstat(path->buf, &st)) {
1624 if (errno == ENOENT)
1625 /*
1626 * file disappeared, which is what we
1627 * wanted anyway
1628 */
1629 continue;
1630 /* fall thru */
1631 } else if (S_ISDIR(st.st_mode)) {
ae2f203e 1632 if (!remove_dir_recurse(path, flag, &kept_down))
7155b727 1633 continue; /* happy */
863808cd
MH
1634 } else if (!only_empty &&
1635 (!unlink(path->buf) || errno == ENOENT)) {
7155b727 1636 continue; /* happy, too */
863808cd 1637 }
7155b727
JS
1638
1639 /* path too long, stat fails, or non-directory still exists */
1640 ret = -1;
1641 break;
1642 }
1643 closedir(dir);
1644
1645 strbuf_setlen(path, original_len);
ae2f203e 1646 if (!ret && !keep_toplevel && !kept_down)
863808cd 1647 ret = (!rmdir(path->buf) || errno == ENOENT) ? 0 : -1;
ae2f203e
JH
1648 else if (kept_up)
1649 /*
1650 * report the uplevel that it is not an error that we
1651 * did not rmdir() our directory.
1652 */
1653 *kept_up = !ret;
7155b727
JS
1654 return ret;
1655}
039bc64e 1656
ae2f203e
JH
1657int remove_dir_recursively(struct strbuf *path, int flag)
1658{
1659 return remove_dir_recurse(path, flag, NULL);
1660}
1661
039bc64e
JH
1662void setup_standard_excludes(struct dir_struct *dir)
1663{
1664 const char *path;
dc79687e 1665 char *xdg_path;
039bc64e
JH
1666
1667 dir->exclude_per_dir = ".gitignore";
1668 path = git_path("info/exclude");
dc79687e
HKNN
1669 if (!excludes_file) {
1670 home_config_paths(NULL, &xdg_path, "ignore");
1671 excludes_file = xdg_path;
1672 }
4698c8fe 1673 if (!access_or_warn(path, R_OK, 0))
039bc64e 1674 add_excludes_from_file(dir, path);
4698c8fe 1675 if (excludes_file && !access_or_warn(excludes_file, R_OK, 0))
039bc64e
JH
1676 add_excludes_from_file(dir, excludes_file);
1677}
4a92d1bf
AR
1678
1679int remove_path(const char *name)
1680{
1681 char *slash;
1682
9a6728d4 1683 if (unlink(name) && errno != ENOENT && errno != ENOTDIR)
4a92d1bf
AR
1684 return -1;
1685
1686 slash = strrchr(name, '/');
1687 if (slash) {
1688 char *dirs = xstrdup(name);
1689 slash = dirs + (slash - name);
1690 do {
1691 *slash = '\0';
3fc0d131 1692 } while (rmdir(dirs) == 0 && (slash = strrchr(dirs, '/')));
4a92d1bf
AR
1693 free(dirs);
1694 }
1695 return 0;
1696}
1697
270be816
AS
1698/*
1699 * Frees memory within dir which was allocated for exclude lists and
1700 * the exclude_stack. Does not free dir itself.
1701 */
1702void clear_directory(struct dir_struct *dir)
1703{
1704 int i, j;
1705 struct exclude_list_group *group;
1706 struct exclude_list *el;
1707 struct exclude_stack *stk;
1708
1709 for (i = EXC_CMDL; i <= EXC_FILE; i++) {
1710 group = &dir->exclude_list_group[i];
1711 for (j = 0; j < group->nr; j++) {
1712 el = &group->el[j];
1713 if (i == EXC_DIRS)
1714 free((char *)el->src);
1715 clear_exclude_list(el);
1716 }
1717 free(group->el);
1718 }
1719
1720 stk = dir->exclude_stack;
1721 while (stk) {
1722 struct exclude_stack *prev = stk->prev;
1723 free(stk);
1724 stk = prev;
1725 }
aceb9429 1726 strbuf_release(&dir->basebuf);
270be816 1727}