Sync with 1.9.4
[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
c28b3d6e
NTND
469static void *read_skip_worktree_file_from_index(const char *path, size_t *size)
470{
471 int pos, len;
472 unsigned long sz;
473 enum object_type type;
474 void *data;
c28b3d6e
NTND
475
476 len = strlen(path);
71261027 477 pos = cache_name_pos(path, len);
c28b3d6e
NTND
478 if (pos < 0)
479 return NULL;
71261027 480 if (!ce_skip_worktree(active_cache[pos]))
c28b3d6e 481 return NULL;
71261027 482 data = read_sha1_file(active_cache[pos]->sha1, &type, &sz);
c28b3d6e
NTND
483 if (!data || type != OBJ_BLOB) {
484 free(data);
485 return NULL;
486 }
487 *size = xsize_t(sz);
488 return data;
489}
490
f6198812
AS
491/*
492 * Frees memory within el which was allocated for exclude patterns and
493 * the file buffer. Does not free el itself.
494 */
495void clear_exclude_list(struct exclude_list *el)
0fd0e241
NTND
496{
497 int i;
498
499 for (i = 0; i < el->nr; i++)
500 free(el->excludes[i]);
501 free(el->excludes);
c082df24 502 free(el->filebuf);
0fd0e241
NTND
503
504 el->nr = 0;
505 el->excludes = NULL;
c082df24 506 el->filebuf = NULL;
0fd0e241
NTND
507}
508
7e2e4b37 509static void trim_trailing_spaces(char *buf)
16402b99 510{
7e2e4b37 511 int i, last_space = -1, nr_spaces, len = strlen(buf);
16402b99
NTND
512 for (i = 0; i < len; i++)
513 if (buf[i] == '\\')
514 i++;
7e2e4b37
NTND
515 else if (buf[i] == ' ') {
516 if (last_space == -1) {
517 last_space = i;
518 nr_spaces = 1;
519 } else
520 nr_spaces++;
521 } else
16402b99
NTND
522 last_space = -1;
523
7e2e4b37
NTND
524 if (last_space != -1 && last_space + nr_spaces == len)
525 buf[last_space] = '\0';
16402b99
NTND
526}
527
cb097534
NTND
528int add_excludes_from_file_to_list(const char *fname,
529 const char *base,
530 int baselen,
840fc334 531 struct exclude_list *el,
cb097534 532 int check_index)
453ec4bd 533{
c470701a 534 struct stat st;
c04318e4 535 int fd, i, lineno = 1;
9d14017a 536 size_t size = 0;
453ec4bd
LT
537 char *buf, *entry;
538
539 fd = open(fname, O_RDONLY);
c28b3d6e 540 if (fd < 0 || fstat(fd, &st) < 0) {
69660731 541 if (errno != ENOENT)
55b38a48 542 warn_on_inaccessible(fname);
c28b3d6e
NTND
543 if (0 <= fd)
544 close(fd);
545 if (!check_index ||
546 (buf = read_skip_worktree_file_from_index(fname, &size)) == NULL)
547 return -1;
45d76f17
NTND
548 if (size == 0) {
549 free(buf);
550 return 0;
551 }
552 if (buf[size-1] != '\n') {
553 buf = xrealloc(buf, size+1);
554 buf[size++] = '\n';
555 }
453ec4bd 556 }
c28b3d6e
NTND
557 else {
558 size = xsize_t(st.st_size);
559 if (size == 0) {
560 close(fd);
561 return 0;
562 }
45d76f17 563 buf = xmalloc(size+1);
c28b3d6e 564 if (read_in_full(fd, buf, size) != size) {
45d76f17 565 free(buf);
c28b3d6e
NTND
566 close(fd);
567 return -1;
568 }
45d76f17 569 buf[size++] = '\n';
c28b3d6e 570 close(fd);
6ba78238 571 }
453ec4bd 572
c082df24 573 el->filebuf = buf;
453ec4bd 574 entry = buf;
45d76f17
NTND
575 for (i = 0; i < size; i++) {
576 if (buf[i] == '\n') {
453ec4bd
LT
577 if (entry != buf + i && entry[0] != '#') {
578 buf[i - (i && buf[i-1] == '\r')] = 0;
7e2e4b37 579 trim_trailing_spaces(entry);
c04318e4 580 add_exclude(entry, base, baselen, el, lineno);
453ec4bd 581 }
c04318e4 582 lineno++;
453ec4bd
LT
583 entry = buf + i + 1;
584 }
585 }
586 return 0;
453ec4bd
LT
587}
588
c04318e4
AS
589struct exclude_list *add_exclude_list(struct dir_struct *dir,
590 int group_type, const char *src)
c082df24
AS
591{
592 struct exclude_list *el;
593 struct exclude_list_group *group;
594
595 group = &dir->exclude_list_group[group_type];
596 ALLOC_GROW(group->el, group->nr + 1, group->alloc);
597 el = &group->el[group->nr++];
598 memset(el, 0, sizeof(*el));
c04318e4 599 el->src = src;
c082df24
AS
600 return el;
601}
602
603/*
604 * Used to set up core.excludesfile and .git/info/exclude lists.
605 */
453ec4bd
LT
606void add_excludes_from_file(struct dir_struct *dir, const char *fname)
607{
c082df24 608 struct exclude_list *el;
c04318e4 609 el = add_exclude_list(dir, EXC_FILE, fname);
c082df24 610 if (add_excludes_from_file_to_list(fname, "", 0, el, 0) < 0)
453ec4bd
LT
611 die("cannot use %s as an exclude file", fname);
612}
613
82dce998
NTND
614int match_basename(const char *basename, int basenamelen,
615 const char *pattern, int prefix, int patternlen,
616 int flags)
593cb880
NTND
617{
618 if (prefix == patternlen) {
0b6e56df
JH
619 if (patternlen == basenamelen &&
620 !strncmp_icase(pattern, basename, basenamelen))
593cb880
NTND
621 return 1;
622 } else if (flags & EXC_FLAG_ENDSWITH) {
0b6e56df 623 /* "*literal" matching against "fooliteral" */
593cb880 624 if (patternlen - 1 <= basenamelen &&
0b6e56df
JH
625 !strncmp_icase(pattern + 1,
626 basename + basenamelen - (patternlen - 1),
627 patternlen - 1))
593cb880
NTND
628 return 1;
629 } else {
0b6e56df
JH
630 if (fnmatch_icase_mem(pattern, patternlen,
631 basename, basenamelen,
632 0) == 0)
593cb880
NTND
633 return 1;
634 }
635 return 0;
636}
637
82dce998
NTND
638int match_pathname(const char *pathname, int pathlen,
639 const char *base, int baselen,
640 const char *pattern, int prefix, int patternlen,
641 int flags)
b5592632
NTND
642{
643 const char *name;
644 int namelen;
645
646 /*
647 * match with FNM_PATHNAME; the pattern has base implicitly
648 * in front of it.
649 */
650 if (*pattern == '/') {
651 pattern++;
982ac873 652 patternlen--;
b5592632
NTND
653 prefix--;
654 }
655
656 /*
657 * baselen does not count the trailing slash. base[] may or
658 * may not end with a trailing slash though.
659 */
660 if (pathlen < baselen + 1 ||
661 (baselen && pathname[baselen] != '/') ||
662 strncmp_icase(pathname, base, baselen))
663 return 0;
664
665 namelen = baselen ? pathlen - baselen - 1 : pathlen;
666 name = pathname + pathlen - namelen;
667
668 if (prefix) {
669 /*
670 * if the non-wildcard part is longer than the
671 * remaining pathname, surely it cannot match.
672 */
673 if (prefix > namelen)
674 return 0;
675
676 if (strncmp_icase(pattern, name, prefix))
677 return 0;
678 pattern += prefix;
ab3aebc1 679 patternlen -= prefix;
b5592632
NTND
680 name += prefix;
681 namelen -= prefix;
ab3aebc1
JK
682
683 /*
684 * If the whole pattern did not have a wildcard,
685 * then our prefix match is all we need; we
686 * do not need to call fnmatch at all.
687 */
688 if (!patternlen && !namelen)
689 return 1;
b5592632
NTND
690 }
691
ab3aebc1
JK
692 return fnmatch_icase_mem(pattern, patternlen,
693 name, namelen,
f30366b2 694 WM_PATHNAME) == 0;
b5592632
NTND
695}
696
578cd7c3
AS
697/*
698 * Scan the given exclude list in reverse to see whether pathname
699 * should be ignored. The first match (i.e. the last on the list), if
700 * any, determines the fate. Returns the exclude_list element which
701 * matched, or NULL for undecided.
453ec4bd 702 */
578cd7c3
AS
703static struct exclude *last_exclude_matching_from_list(const char *pathname,
704 int pathlen,
705 const char *basename,
706 int *dtype,
707 struct exclude_list *el)
453ec4bd
LT
708{
709 int i;
710
35a94d44 711 if (!el->nr)
578cd7c3 712 return NULL; /* undefined */
d6b8fc30 713
35a94d44
NTND
714 for (i = el->nr - 1; 0 <= i; i--) {
715 struct exclude *x = el->excludes[i];
b5592632 716 const char *exclude = x->pattern;
b5592632 717 int prefix = x->nowildcardlen;
35a94d44
NTND
718
719 if (x->flags & EXC_FLAG_MUSTBEDIR) {
720 if (*dtype == DT_UNKNOWN)
721 *dtype = get_dtype(NULL, pathname, pathlen);
722 if (*dtype != DT_DIR)
723 continue;
724 }
d6b8fc30 725
35a94d44 726 if (x->flags & EXC_FLAG_NODIR) {
593cb880
NTND
727 if (match_basename(basename,
728 pathlen - (basename - pathname),
729 exclude, prefix, x->patternlen,
730 x->flags))
578cd7c3 731 return x;
35a94d44
NTND
732 continue;
733 }
734
b5592632
NTND
735 assert(x->baselen == 0 || x->base[x->baselen - 1] == '/');
736 if (match_pathname(pathname, pathlen,
737 x->base, x->baselen ? x->baselen - 1 : 0,
738 exclude, prefix, x->patternlen, x->flags))
578cd7c3 739 return x;
453ec4bd 740 }
578cd7c3
AS
741 return NULL; /* undecided */
742}
743
744/*
745 * Scan the list and let the last match determine the fate.
746 * Return 1 for exclude, 0 for include and -1 for undecided.
747 */
748int is_excluded_from_list(const char *pathname,
749 int pathlen, const char *basename, int *dtype,
750 struct exclude_list *el)
751{
752 struct exclude *exclude;
753 exclude = last_exclude_matching_from_list(pathname, pathlen, basename, dtype, el);
754 if (exclude)
755 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
453ec4bd
LT
756 return -1; /* undecided */
757}
758
46aa2f95
KB
759static struct exclude *last_exclude_matching_from_lists(struct dir_struct *dir,
760 const char *pathname, int pathlen, const char *basename,
761 int *dtype_p)
762{
763 int i, j;
764 struct exclude_list_group *group;
765 struct exclude *exclude;
766 for (i = EXC_CMDL; i <= EXC_FILE; i++) {
767 group = &dir->exclude_list_group[i];
768 for (j = group->nr - 1; j >= 0; j--) {
769 exclude = last_exclude_matching_from_list(
770 pathname, pathlen, basename, dtype_p,
771 &group->el[j]);
772 if (exclude)
773 return exclude;
774 }
775 }
776 return NULL;
777}
778
6cd5c582
KB
779/*
780 * Loads the per-directory exclude list for the substring of base
781 * which has a char length of baselen.
782 */
783static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
784{
785 struct exclude_list_group *group;
786 struct exclude_list *el;
787 struct exclude_stack *stk = NULL;
788 int current;
789
6cd5c582
KB
790 group = &dir->exclude_list_group[EXC_DIRS];
791
792 /* Pop the exclude lists from the EXCL_DIRS exclude_list_group
793 * which originate from directories not in the prefix of the
794 * path being checked. */
795 while ((stk = dir->exclude_stack) != NULL) {
796 if (stk->baselen <= baselen &&
797 !strncmp(dir->basebuf, base, stk->baselen))
798 break;
799 el = &group->el[dir->exclude_stack->exclude_ix];
800 dir->exclude_stack = stk->prev;
95c6f271 801 dir->exclude = NULL;
6cd5c582
KB
802 free((char *)el->src); /* see strdup() below */
803 clear_exclude_list(el);
804 free(stk);
805 group->nr--;
806 }
807
95c6f271
KB
808 /* Skip traversing into sub directories if the parent is excluded */
809 if (dir->exclude)
810 return;
811
6cd5c582
KB
812 /* Read from the parent directories and push them down. */
813 current = stk ? stk->baselen : -1;
814 while (current < baselen) {
815 struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
816 const char *cp;
817
818 if (current < 0) {
819 cp = base;
820 current = 0;
821 }
822 else {
823 cp = strchr(base + current + 1, '/');
824 if (!cp)
825 die("oops in prep_exclude");
826 cp++;
827 }
828 stk->prev = dir->exclude_stack;
829 stk->baselen = cp - base;
95c6f271
KB
830 stk->exclude_ix = group->nr;
831 el = add_exclude_list(dir, EXC_DIRS, NULL);
6cd5c582
KB
832 memcpy(dir->basebuf + current, base + current,
833 stk->baselen - current);
95c6f271
KB
834
835 /* Abort if the directory is excluded */
836 if (stk->baselen) {
837 int dt = DT_DIR;
838 dir->basebuf[stk->baselen - 1] = 0;
839 dir->exclude = last_exclude_matching_from_lists(dir,
840 dir->basebuf, stk->baselen - 1,
841 dir->basebuf + current, &dt);
842 dir->basebuf[stk->baselen - 1] = '/';
c3c327de
KB
843 if (dir->exclude &&
844 dir->exclude->flags & EXC_FLAG_NEGATIVE)
845 dir->exclude = NULL;
95c6f271
KB
846 if (dir->exclude) {
847 dir->basebuf[stk->baselen] = 0;
848 dir->exclude_stack = stk;
849 return;
850 }
851 }
852
853 /* Try to read per-directory file unless path is too long */
854 if (dir->exclude_per_dir &&
855 stk->baselen + strlen(dir->exclude_per_dir) < PATH_MAX) {
856 strcpy(dir->basebuf + stk->baselen,
857 dir->exclude_per_dir);
858 /*
859 * dir->basebuf gets reused by the traversal, but we
860 * need fname to remain unchanged to ensure the src
861 * member of each struct exclude correctly
862 * back-references its source file. Other invocations
863 * of add_exclude_list provide stable strings, so we
864 * strdup() and free() here in the caller.
865 */
866 el->src = strdup(dir->basebuf);
867 add_excludes_from_file_to_list(dir->basebuf,
868 dir->basebuf, stk->baselen, el, 1);
869 }
6cd5c582
KB
870 dir->exclude_stack = stk;
871 current = stk->baselen;
872 }
873 dir->basebuf[baselen] = '\0';
874}
875
f4cd69a6
AS
876/*
877 * Loads the exclude lists for the directory containing pathname, then
878 * scans all exclude lists to determine whether pathname is excluded.
879 * Returns the exclude_list element which matched, or NULL for
880 * undecided.
881 */
b07bc8c8 882struct exclude *last_exclude_matching(struct dir_struct *dir,
f4cd69a6
AS
883 const char *pathname,
884 int *dtype_p)
453ec4bd
LT
885{
886 int pathlen = strlen(pathname);
68492fc7
LK
887 const char *basename = strrchr(pathname, '/');
888 basename = (basename) ? basename+1 : pathname;
453ec4bd 889
63d285c8 890 prep_exclude(dir, pathname, basename-pathname);
c082df24 891
95c6f271
KB
892 if (dir->exclude)
893 return dir->exclude;
894
46aa2f95
KB
895 return last_exclude_matching_from_lists(dir, pathname, pathlen,
896 basename, dtype_p);
f4cd69a6
AS
897}
898
899/*
900 * Loads the exclude lists for the directory containing pathname, then
901 * scans all exclude lists to determine whether pathname is excluded.
902 * Returns 1 if true, otherwise 0.
903 */
b07bc8c8 904int is_excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
f4cd69a6
AS
905{
906 struct exclude *exclude =
907 last_exclude_matching(dir, pathname, dtype_p);
908 if (exclude)
909 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
453ec4bd
LT
910 return 0;
911}
912
f3fa1838
JH
913static struct dir_entry *dir_entry_new(const char *pathname, int len)
914{
453ec4bd
LT
915 struct dir_entry *ent;
916
453ec4bd
LT
917 ent = xmalloc(sizeof(*ent) + len + 1);
918 ent->len = len;
919 memcpy(ent->name, pathname, len);
920 ent->name[len] = 0;
4d06f8ac 921 return ent;
453ec4bd
LT
922}
923
159b3212 924static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
6815e569 925{
ebbd7439 926 if (cache_file_exists(pathname, len, ignore_case))
6815e569
JK
927 return NULL;
928
25fd2f7a 929 ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
6815e569
JK
930 return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
931}
932
108da0db 933struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
2abd31b0 934{
6e4f981f 935 if (!cache_name_is_other(pathname, len))
2abd31b0
JK
936 return NULL;
937
25fd2f7a 938 ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
2abd31b0
JK
939 return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
940}
941
09595258
LT
942enum exist_status {
943 index_nonexistent = 0,
944 index_directory,
4b05548f 945 index_gitdir
09595258
LT
946};
947
5102c617 948/*
71261027 949 * Do not use the alphabetically sorted index to look up
5102c617 950 * the directory name; instead, use the case insensitive
ebbd7439 951 * directory hash.
5102c617
JJ
952 */
953static enum exist_status directory_exists_in_index_icase(const char *dirname, int len)
954{
d28eec26 955 const struct cache_entry *ce = cache_dir_exists(dirname, len);
5102c617
JJ
956 unsigned char endchar;
957
958 if (!ce)
959 return index_nonexistent;
960 endchar = ce->name[len];
961
962 /*
963 * The cache_entry structure returned will contain this dirname
964 * and possibly additional path components.
965 */
966 if (endchar == '/')
967 return index_directory;
968
969 /*
970 * If there are no additional path components, then this cache_entry
971 * represents a submodule. Submodules, despite being directories,
972 * are stored in the cache without a closing slash.
973 */
974 if (!endchar && S_ISGITLINK(ce->ce_mode))
975 return index_gitdir;
976
977 /* This should never be hit, but it exists just in case. */
978 return index_nonexistent;
979}
980
09595258
LT
981/*
982 * The index sorts alphabetically by entry name, which
983 * means that a gitlink sorts as '\0' at the end, while
984 * a directory (which is defined not as an entry, but as
985 * the files it contains) will sort with the '/' at the
986 * end.
987 */
988static enum exist_status directory_exists_in_index(const char *dirname, int len)
453ec4bd 989{
5102c617
JJ
990 int pos;
991
992 if (ignore_case)
993 return directory_exists_in_index_icase(dirname, len);
994
995 pos = cache_name_pos(dirname, len);
09595258
LT
996 if (pos < 0)
997 pos = -pos-1;
998 while (pos < active_nr) {
9c5e6c80 999 const struct cache_entry *ce = active_cache[pos++];
09595258
LT
1000 unsigned char endchar;
1001
1002 if (strncmp(ce->name, dirname, len))
1003 break;
1004 endchar = ce->name[len];
1005 if (endchar > '/')
1006 break;
1007 if (endchar == '/')
1008 return index_directory;
7a51ed66 1009 if (!endchar && S_ISGITLINK(ce->ce_mode))
09595258
LT
1010 return index_gitdir;
1011 }
1012 return index_nonexistent;
1013}
1014
1015/*
1016 * When we find a directory when traversing the filesystem, we
1017 * have three distinct cases:
1018 *
1019 * - ignore it
1020 * - see it as a directory
1021 * - recurse into it
1022 *
1023 * and which one we choose depends on a combination of existing
1024 * git index contents and the flags passed into the directory
1025 * traversal routine.
1026 *
1027 * Case 1: If we *already* have entries in the index under that
5bd8e2d8
KB
1028 * directory name, we always recurse into the directory to see
1029 * all the files.
09595258
LT
1030 *
1031 * Case 2: If we *already* have that directory name as a gitlink,
1032 * we always continue to see it as a gitlink, regardless of whether
1033 * there is an actual git directory there or not (it might not
1034 * be checked out as a subproject!)
1035 *
1036 * Case 3: if we didn't have it in the index previously, we
1037 * have a few sub-cases:
1038 *
1039 * (a) if "show_other_directories" is true, we show it as
1040 * just a directory, unless "hide_empty_directories" is
defd7c7b
KB
1041 * also true, in which case we need to check if it contains any
1042 * untracked and / or ignored files.
09595258 1043 * (b) if it looks like a git directory, and we don't have
302b9282 1044 * 'no_gitlinks' set we treat it as a gitlink, and show it
09595258
LT
1045 * as a directory.
1046 * (c) otherwise, we recurse into it.
1047 */
defd7c7b 1048static enum path_treatment treat_directory(struct dir_struct *dir,
721ac4ed 1049 const char *dirname, int len, int exclude,
09595258
LT
1050 const struct path_simplify *simplify)
1051{
1052 /* The "len-1" is to strip the final '/' */
1053 switch (directory_exists_in_index(dirname, len-1)) {
1054 case index_directory:
defd7c7b 1055 return path_recurse;
09595258
LT
1056
1057 case index_gitdir:
26c986e1 1058 return path_none;
09595258
LT
1059
1060 case index_nonexistent:
7c4c97c0 1061 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
09595258 1062 break;
7c4c97c0 1063 if (!(dir->flags & DIR_NO_GITLINKS)) {
09595258
LT
1064 unsigned char sha1[20];
1065 if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
defd7c7b 1066 return path_untracked;
09595258 1067 }
defd7c7b 1068 return path_recurse;
09595258
LT
1069 }
1070
1071 /* This is the "show_other_directories" case */
721ac4ed 1072
184d2a8e 1073 if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
defd7c7b
KB
1074 return exclude ? path_excluded : path_untracked;
1075
1076 return read_directory_recursive(dir, dirname, len, 1, simplify);
453ec4bd
LT
1077}
1078
9fc42d60
LT
1079/*
1080 * This is an inexact early pruning of any recursive directory
1081 * reading - if the path cannot possibly be in the pathspec,
1082 * return true, and we'll skip it early.
1083 */
1084static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
1085{
1086 if (simplify) {
1087 for (;;) {
1088 const char *match = simplify->path;
1089 int len = simplify->len;
1090
1091 if (!match)
1092 break;
1093 if (len > pathlen)
1094 len = pathlen;
1095 if (!memcmp(path, match, len))
1096 return 0;
1097 simplify++;
1098 }
1099 return 1;
1100 }
1101 return 0;
1102}
1103
29209cbe
JK
1104/*
1105 * This function tells us whether an excluded path matches a
1106 * list of "interesting" pathspecs. That is, whether a path matched
1107 * by any of the pathspecs could possibly be ignored by excluding
1108 * the specified path. This can happen if:
1109 *
1110 * 1. the path is mentioned explicitly in the pathspec
1111 *
1112 * 2. the path is a directory prefix of some element in the
1113 * pathspec
1114 */
1115static int exclude_matches_pathspec(const char *path, int len,
1116 const struct path_simplify *simplify)
e96980ef
JK
1117{
1118 if (simplify) {
1119 for (; simplify->path; simplify++) {
1120 if (len == simplify->len
1121 && !memcmp(path, simplify->path, len))
1122 return 1;
29209cbe
JK
1123 if (len < simplify->len
1124 && simplify->path[len] == '/'
1125 && !memcmp(path, simplify->path, len))
1126 return 1;
e96980ef
JK
1127 }
1128 }
1129 return 0;
1130}
1131
443e061a
LT
1132static int get_index_dtype(const char *path, int len)
1133{
1134 int pos;
9c5e6c80 1135 const struct cache_entry *ce;
443e061a 1136
ebbd7439 1137 ce = cache_file_exists(path, len, 0);
443e061a
LT
1138 if (ce) {
1139 if (!ce_uptodate(ce))
1140 return DT_UNKNOWN;
1141 if (S_ISGITLINK(ce->ce_mode))
1142 return DT_DIR;
1143 /*
1144 * Nobody actually cares about the
1145 * difference between DT_LNK and DT_REG
1146 */
1147 return DT_REG;
1148 }
1149
1150 /* Try to look it up as a directory */
1151 pos = cache_name_pos(path, len);
1152 if (pos >= 0)
1153 return DT_UNKNOWN;
1154 pos = -pos-1;
1155 while (pos < active_nr) {
1156 ce = active_cache[pos++];
1157 if (strncmp(ce->name, path, len))
1158 break;
1159 if (ce->name[len] > '/')
1160 break;
1161 if (ce->name[len] < '/')
1162 continue;
1163 if (!ce_uptodate(ce))
1164 break; /* continue? */
1165 return DT_DIR;
1166 }
1167 return DT_UNKNOWN;
1168}
1169
caa6b782 1170static int get_dtype(struct dirent *de, const char *path, int len)
07134421 1171{
6831a88a 1172 int dtype = de ? DTYPE(de) : DT_UNKNOWN;
07134421
LT
1173 struct stat st;
1174
1175 if (dtype != DT_UNKNOWN)
1176 return dtype;
443e061a
LT
1177 dtype = get_index_dtype(path, len);
1178 if (dtype != DT_UNKNOWN)
1179 return dtype;
1180 if (lstat(path, &st))
07134421
LT
1181 return dtype;
1182 if (S_ISREG(st.st_mode))
1183 return DT_REG;
1184 if (S_ISDIR(st.st_mode))
1185 return DT_DIR;
1186 if (S_ISLNK(st.st_mode))
1187 return DT_LNK;
1188 return dtype;
1189}
1190
16e2cfa9 1191static enum path_treatment treat_one_path(struct dir_struct *dir,
49dc2cc2 1192 struct strbuf *path,
16e2cfa9
JH
1193 const struct path_simplify *simplify,
1194 int dtype, struct dirent *de)
53cc5356 1195{
8aaf8d77 1196 int exclude;
ebbd7439 1197 int has_path_in_index = !!cache_file_exists(path->buf, path->len, ignore_case);
2eac2a4c 1198
8aaf8d77
KB
1199 if (dtype == DT_UNKNOWN)
1200 dtype = get_dtype(de, path->buf, path->len);
1201
1202 /* Always exclude indexed files */
2eac2a4c 1203 if (dtype != DT_DIR && has_path_in_index)
defd7c7b 1204 return path_none;
8aaf8d77 1205
2eac2a4c
JH
1206 /*
1207 * When we are looking at a directory P in the working tree,
1208 * there are three cases:
1209 *
1210 * (1) P exists in the index. Everything inside the directory P in
1211 * the working tree needs to go when P is checked out from the
1212 * index.
1213 *
1214 * (2) P does not exist in the index, but there is P/Q in the index.
1215 * We know P will stay a directory when we check out the contents
1216 * of the index, but we do not know yet if there is a directory
1217 * P/Q in the working tree to be killed, so we need to recurse.
1218 *
1219 * (3) P does not exist in the index, and there is no P/Q in the index
1220 * to require P to be a directory, either. Only in this case, we
1221 * know that everything inside P will not be killed without
1222 * recursing.
1223 */
1224 if ((dir->flags & DIR_COLLECT_KILLED_ONLY) &&
1225 (dtype == DT_DIR) &&
de372b1b
ES
1226 !has_path_in_index &&
1227 (directory_exists_in_index(path->buf, path->len) == index_nonexistent))
1228 return path_none;
8aaf8d77
KB
1229
1230 exclude = is_excluded(dir, path->buf, &dtype);
53cc5356
JH
1231
1232 /*
1233 * Excluded? If we don't explicitly want to show
1234 * ignored files, ignore it
1235 */
0aaf62b6 1236 if (exclude && !(dir->flags & (DIR_SHOW_IGNORED|DIR_SHOW_IGNORED_TOO)))
defd7c7b 1237 return path_excluded;
53cc5356 1238
53cc5356
JH
1239 switch (dtype) {
1240 default:
defd7c7b 1241 return path_none;
53cc5356 1242 case DT_DIR:
49dc2cc2 1243 strbuf_addch(path, '/');
defd7c7b
KB
1244 return treat_directory(dir, path->buf, path->len, exclude,
1245 simplify);
53cc5356
JH
1246 case DT_REG:
1247 case DT_LNK:
defd7c7b 1248 return exclude ? path_excluded : path_untracked;
53cc5356 1249 }
53cc5356
JH
1250}
1251
16e2cfa9
JH
1252static enum path_treatment treat_path(struct dir_struct *dir,
1253 struct dirent *de,
49dc2cc2 1254 struct strbuf *path,
16e2cfa9 1255 int baselen,
49dc2cc2 1256 const struct path_simplify *simplify)
16e2cfa9
JH
1257{
1258 int dtype;
1259
1260 if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
defd7c7b 1261 return path_none;
49dc2cc2
RS
1262 strbuf_setlen(path, baselen);
1263 strbuf_addstr(path, de->d_name);
1264 if (simplify_away(path->buf, path->len, simplify))
defd7c7b 1265 return path_none;
16e2cfa9
JH
1266
1267 dtype = DTYPE(de);
49dc2cc2 1268 return treat_one_path(dir, path, simplify, dtype, de);
16e2cfa9
JH
1269}
1270
453ec4bd
LT
1271/*
1272 * Read a directory tree. We currently ignore anything but
1273 * directories, regular files and symlinks. That's because git
1274 * doesn't handle them at all yet. Maybe that will change some
1275 * day.
1276 *
1277 * Also, we ignore the name ".git" (even if it is not a directory).
1278 * That likely will not change.
defd7c7b
KB
1279 *
1280 * Returns the most significant path_treatment value encountered in the scan.
453ec4bd 1281 */
defd7c7b 1282static enum path_treatment read_directory_recursive(struct dir_struct *dir,
53cc5356
JH
1283 const char *base, int baselen,
1284 int check_only,
1285 const struct path_simplify *simplify)
453ec4bd 1286{
1528d247 1287 DIR *fdir;
defd7c7b 1288 enum path_treatment state, subdir_state, dir_state = path_none;
02cb6753 1289 struct dirent *de;
bef36921 1290 struct strbuf path = STRBUF_INIT;
453ec4bd 1291
bef36921 1292 strbuf_add(&path, base, baselen);
02cb6753 1293
1528d247
RS
1294 fdir = opendir(path.len ? path.buf : ".");
1295 if (!fdir)
1296 goto out;
1297
02cb6753 1298 while ((de = readdir(fdir)) != NULL) {
defd7c7b
KB
1299 /* check how the file or directory should be treated */
1300 state = treat_path(dir, de, &path, baselen, simplify);
1301 if (state > dir_state)
1302 dir_state = state;
1303
1304 /* recurse into subdir if instructed by treat_path */
1305 if (state == path_recurse) {
1306 subdir_state = read_directory_recursive(dir, path.buf,
289ff559 1307 path.len, check_only, simplify);
defd7c7b
KB
1308 if (subdir_state > dir_state)
1309 dir_state = subdir_state;
1310 }
1311
1312 if (check_only) {
1313 /* abort early if maximum state has been reached */
1314 if (dir_state == path_untracked)
1315 break;
1316 /* skip the dir_add_* part */
02cb6753 1317 continue;
453ec4bd 1318 }
defd7c7b
KB
1319
1320 /* add the path to the appropriate result list */
1321 switch (state) {
1322 case path_excluded:
1323 if (dir->flags & DIR_SHOW_IGNORED)
1324 dir_add_name(dir, path.buf, path.len);
0aaf62b6
KB
1325 else if ((dir->flags & DIR_SHOW_IGNORED_TOO) ||
1326 ((dir->flags & DIR_COLLECT_IGNORED) &&
1327 exclude_matches_pathspec(path.buf, path.len,
1328 simplify)))
1329 dir_add_ignored(dir, path.buf, path.len);
defd7c7b
KB
1330 break;
1331
1332 case path_untracked:
1333 if (!(dir->flags & DIR_SHOW_IGNORED))
1334 dir_add_name(dir, path.buf, path.len);
1528d247 1335 break;
defd7c7b
KB
1336
1337 default:
1338 break;
1339 }
453ec4bd 1340 }
02cb6753 1341 closedir(fdir);
1528d247 1342 out:
bef36921 1343 strbuf_release(&path);
453ec4bd 1344
defd7c7b 1345 return dir_state;
453ec4bd
LT
1346}
1347
1348static int cmp_name(const void *p1, const void *p2)
1349{
1350 const struct dir_entry *e1 = *(const struct dir_entry **)p1;
1351 const struct dir_entry *e2 = *(const struct dir_entry **)p2;
1352
1353 return cache_name_compare(e1->name, e1->len,
1354 e2->name, e2->len);
1355}
1356
9fc42d60
LT
1357static struct path_simplify *create_simplify(const char **pathspec)
1358{
1359 int nr, alloc = 0;
1360 struct path_simplify *simplify = NULL;
1361
1362 if (!pathspec)
1363 return NULL;
1364
1365 for (nr = 0 ; ; nr++) {
1366 const char *match;
9af49f82 1367 ALLOC_GROW(simplify, nr + 1, alloc);
9fc42d60
LT
1368 match = *pathspec++;
1369 if (!match)
1370 break;
1371 simplify[nr].path = match;
1372 simplify[nr].len = simple_length(match);
1373 }
1374 simplify[nr].path = NULL;
1375 simplify[nr].len = 0;
1376 return simplify;
1377}
1378
1379static void free_simplify(struct path_simplify *simplify)
1380{
8e0f7003 1381 free(simplify);
9fc42d60
LT
1382}
1383
48ffef96
JH
1384static int treat_leading_path(struct dir_struct *dir,
1385 const char *path, int len,
1386 const struct path_simplify *simplify)
1387{
49dc2cc2
RS
1388 struct strbuf sb = STRBUF_INIT;
1389 int baselen, rc = 0;
48ffef96 1390 const char *cp;
be8a84c5 1391 int old_flags = dir->flags;
48ffef96
JH
1392
1393 while (len && path[len - 1] == '/')
1394 len--;
1395 if (!len)
1396 return 1;
1397 baselen = 0;
be8a84c5 1398 dir->flags &= ~DIR_SHOW_OTHER_DIRECTORIES;
48ffef96
JH
1399 while (1) {
1400 cp = path + baselen + !!baselen;
1401 cp = memchr(cp, '/', path + len - cp);
1402 if (!cp)
1403 baselen = len;
1404 else
1405 baselen = cp - path;
49dc2cc2
RS
1406 strbuf_setlen(&sb, 0);
1407 strbuf_add(&sb, path, baselen);
1408 if (!is_directory(sb.buf))
1409 break;
1410 if (simplify_away(sb.buf, sb.len, simplify))
1411 break;
1412 if (treat_one_path(dir, &sb, simplify,
defd7c7b 1413 DT_DIR, NULL) == path_none)
49dc2cc2
RS
1414 break; /* do not recurse into it */
1415 if (len <= baselen) {
1416 rc = 1;
1417 break; /* finished checking */
1418 }
48ffef96 1419 }
49dc2cc2 1420 strbuf_release(&sb);
be8a84c5 1421 dir->flags = old_flags;
49dc2cc2 1422 return rc;
48ffef96
JH
1423}
1424
7327d3d1 1425int read_directory(struct dir_struct *dir, const char *path, int len, const struct pathspec *pathspec)
9fc42d60 1426{
725b0605 1427 struct path_simplify *simplify;
b4189aa8 1428
7327d3d1
NTND
1429 /*
1430 * Check out create_simplify()
1431 */
1432 if (pathspec)
5c6933d2
NTND
1433 GUARD_PATHSPEC(pathspec,
1434 PATHSPEC_FROMTOP |
1435 PATHSPEC_MAXDEPTH |
bd30c2e4 1436 PATHSPEC_LITERAL |
93d93537 1437 PATHSPEC_GLOB |
ef79b1f8
NTND
1438 PATHSPEC_ICASE |
1439 PATHSPEC_EXCLUDE);
7327d3d1 1440
dba2e203 1441 if (has_symlink_leading_path(path, len))
725b0605
JH
1442 return dir->nr;
1443
ef79b1f8
NTND
1444 /*
1445 * exclude patterns are treated like positive ones in
1446 * create_simplify. Usually exclude patterns should be a
1447 * subset of positive ones, which has no impacts on
1448 * create_simplify().
1449 */
b3920bbd 1450 simplify = create_simplify(pathspec ? pathspec->_raw : NULL);
48ffef96
JH
1451 if (!len || treat_leading_path(dir, path, len, simplify))
1452 read_directory_recursive(dir, path, len, 0, simplify);
9fc42d60 1453 free_simplify(simplify);
453ec4bd 1454 qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
2abd31b0 1455 qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
453ec4bd
LT
1456 return dir->nr;
1457}
c91f0d92 1458
686a4a06 1459int file_exists(const char *f)
c91f0d92 1460{
686a4a06 1461 struct stat sb;
a50f9fc5 1462 return lstat(f, &sb) == 0;
c91f0d92 1463}
e6636747
JS
1464
1465/*
9b125da4
NTND
1466 * Given two normalized paths (a trailing slash is ok), if subdir is
1467 * outside dir, return -1. Otherwise return the offset in subdir that
1468 * can be used as relative path to dir.
e6636747 1469 */
9b125da4 1470int dir_inside_of(const char *subdir, const char *dir)
e6636747 1471{
9b125da4 1472 int offset = 0;
e6636747 1473
9b125da4 1474 assert(dir && subdir && *dir && *subdir);
e6636747 1475
9b125da4 1476 while (*dir && *subdir && *dir == *subdir) {
e6636747 1477 dir++;
9b125da4
NTND
1478 subdir++;
1479 offset++;
490544b1 1480 }
9b125da4
NTND
1481
1482 /* hel[p]/me vs hel[l]/yeah */
1483 if (*dir && *subdir)
1484 return -1;
1485
1486 if (!*subdir)
1487 return !*dir ? offset : -1; /* same dir */
1488
1489 /* foo/[b]ar vs foo/[] */
1490 if (is_dir_sep(dir[-1]))
1491 return is_dir_sep(subdir[-1]) ? offset : -1;
1492
1493 /* foo[/]bar vs foo[] */
1494 return is_dir_sep(*subdir) ? offset + 1 : -1;
e6636747
JS
1495}
1496
1497int is_inside_dir(const char *dir)
1498{
b892913d
NTND
1499 char cwd[PATH_MAX];
1500 if (!dir)
1501 return 0;
1502 if (!getcwd(cwd, sizeof(cwd)))
1503 die_errno("can't find the current directory");
1504 return dir_inside_of(cwd, dir) >= 0;
e6636747 1505}
7155b727 1506
55892d23
AP
1507int is_empty_dir(const char *path)
1508{
1509 DIR *dir = opendir(path);
1510 struct dirent *e;
1511 int ret = 1;
1512
1513 if (!dir)
1514 return 0;
1515
1516 while ((e = readdir(dir)) != NULL)
1517 if (!is_dot_or_dotdot(e->d_name)) {
1518 ret = 0;
1519 break;
1520 }
1521
1522 closedir(dir);
1523 return ret;
1524}
1525
ae2f203e 1526static int remove_dir_recurse(struct strbuf *path, int flag, int *kept_up)
7155b727 1527{
a0f4afbe 1528 DIR *dir;
7155b727 1529 struct dirent *e;
ae2f203e 1530 int ret = 0, original_len = path->len, len, kept_down = 0;
a0f4afbe 1531 int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
c844a803 1532 int keep_toplevel = (flag & REMOVE_DIR_KEEP_TOPLEVEL);
a0f4afbe 1533 unsigned char submodule_head[20];
7155b727 1534
a0f4afbe 1535 if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
ae2f203e 1536 !resolve_gitlink_ref(path->buf, "HEAD", submodule_head)) {
a0f4afbe 1537 /* Do not descend and nuke a nested git work tree. */
ae2f203e
JH
1538 if (kept_up)
1539 *kept_up = 1;
a0f4afbe 1540 return 0;
ae2f203e 1541 }
a0f4afbe 1542
ae2f203e 1543 flag &= ~REMOVE_DIR_KEEP_TOPLEVEL;
a0f4afbe 1544 dir = opendir(path->buf);
c844a803 1545 if (!dir) {
863808cd
MH
1546 if (errno == ENOENT)
1547 return keep_toplevel ? -1 : 0;
1548 else if (errno == EACCES && !keep_toplevel)
ecb2c282
MH
1549 /*
1550 * An empty dir could be removable even if it
1551 * is unreadable:
1552 */
c844a803
JH
1553 return rmdir(path->buf);
1554 else
1555 return -1;
1556 }
7155b727
JS
1557 if (path->buf[original_len - 1] != '/')
1558 strbuf_addch(path, '/');
1559
1560 len = path->len;
1561 while ((e = readdir(dir)) != NULL) {
1562 struct stat st;
8ca12c0d
AP
1563 if (is_dot_or_dotdot(e->d_name))
1564 continue;
7155b727
JS
1565
1566 strbuf_setlen(path, len);
1567 strbuf_addstr(path, e->d_name);
863808cd
MH
1568 if (lstat(path->buf, &st)) {
1569 if (errno == ENOENT)
1570 /*
1571 * file disappeared, which is what we
1572 * wanted anyway
1573 */
1574 continue;
1575 /* fall thru */
1576 } else if (S_ISDIR(st.st_mode)) {
ae2f203e 1577 if (!remove_dir_recurse(path, flag, &kept_down))
7155b727 1578 continue; /* happy */
863808cd
MH
1579 } else if (!only_empty &&
1580 (!unlink(path->buf) || errno == ENOENT)) {
7155b727 1581 continue; /* happy, too */
863808cd 1582 }
7155b727
JS
1583
1584 /* path too long, stat fails, or non-directory still exists */
1585 ret = -1;
1586 break;
1587 }
1588 closedir(dir);
1589
1590 strbuf_setlen(path, original_len);
ae2f203e 1591 if (!ret && !keep_toplevel && !kept_down)
863808cd 1592 ret = (!rmdir(path->buf) || errno == ENOENT) ? 0 : -1;
ae2f203e
JH
1593 else if (kept_up)
1594 /*
1595 * report the uplevel that it is not an error that we
1596 * did not rmdir() our directory.
1597 */
1598 *kept_up = !ret;
7155b727
JS
1599 return ret;
1600}
039bc64e 1601
ae2f203e
JH
1602int remove_dir_recursively(struct strbuf *path, int flag)
1603{
1604 return remove_dir_recurse(path, flag, NULL);
1605}
1606
039bc64e
JH
1607void setup_standard_excludes(struct dir_struct *dir)
1608{
1609 const char *path;
dc79687e 1610 char *xdg_path;
039bc64e
JH
1611
1612 dir->exclude_per_dir = ".gitignore";
1613 path = git_path("info/exclude");
dc79687e
HKNN
1614 if (!excludes_file) {
1615 home_config_paths(NULL, &xdg_path, "ignore");
1616 excludes_file = xdg_path;
1617 }
4698c8fe 1618 if (!access_or_warn(path, R_OK, 0))
039bc64e 1619 add_excludes_from_file(dir, path);
4698c8fe 1620 if (excludes_file && !access_or_warn(excludes_file, R_OK, 0))
039bc64e
JH
1621 add_excludes_from_file(dir, excludes_file);
1622}
4a92d1bf
AR
1623
1624int remove_path(const char *name)
1625{
1626 char *slash;
1627
9a6728d4 1628 if (unlink(name) && errno != ENOENT && errno != ENOTDIR)
4a92d1bf
AR
1629 return -1;
1630
1631 slash = strrchr(name, '/');
1632 if (slash) {
1633 char *dirs = xstrdup(name);
1634 slash = dirs + (slash - name);
1635 do {
1636 *slash = '\0';
3fc0d131 1637 } while (rmdir(dirs) == 0 && (slash = strrchr(dirs, '/')));
4a92d1bf
AR
1638 free(dirs);
1639 }
1640 return 0;
1641}
1642
270be816
AS
1643/*
1644 * Frees memory within dir which was allocated for exclude lists and
1645 * the exclude_stack. Does not free dir itself.
1646 */
1647void clear_directory(struct dir_struct *dir)
1648{
1649 int i, j;
1650 struct exclude_list_group *group;
1651 struct exclude_list *el;
1652 struct exclude_stack *stk;
1653
1654 for (i = EXC_CMDL; i <= EXC_FILE; i++) {
1655 group = &dir->exclude_list_group[i];
1656 for (j = 0; j < group->nr; j++) {
1657 el = &group->el[j];
1658 if (i == EXC_DIRS)
1659 free((char *)el->src);
1660 clear_exclude_list(el);
1661 }
1662 free(group->el);
1663 }
1664
1665 stk = dir->exclude_stack;
1666 while (stk) {
1667 struct exclude_stack *prev = stk->prev;
1668 free(stk);
1669 stk = prev;
1670 }
1671}