convert refresh_index to take struct pathspec
[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{
52 return fnmatch(pattern, string, flags | (ignore_case ? FNM_CASEFOLD : 0));
53}
54
5d74762d
NTND
55inline int git_fnmatch(const char *pattern, const char *string,
56 int flags, int prefix)
57{
58 int fnm_flags = 0;
59 if (flags & GFNM_PATHNAME)
60 fnm_flags |= FNM_PATHNAME;
61 if (prefix > 0) {
62 if (strncmp(pattern, string, prefix))
63 return FNM_NOMATCH;
64 pattern += prefix;
65 string += prefix;
66 }
8c6abbcd
NTND
67 if (flags & GFNM_ONESTAR) {
68 int pattern_len = strlen(++pattern);
69 int string_len = strlen(string);
70 return string_len < pattern_len ||
71 strcmp(pattern,
72 string + string_len - pattern_len);
73 }
5d74762d
NTND
74 return fnmatch(pattern, string, fnm_flags);
75}
76
0b6e56df
JH
77static int fnmatch_icase_mem(const char *pattern, int patternlen,
78 const char *string, int stringlen,
79 int flags)
80{
81 int match_status;
82 struct strbuf pat_buf = STRBUF_INIT;
83 struct strbuf str_buf = STRBUF_INIT;
84 const char *use_pat = pattern;
85 const char *use_str = string;
86
87 if (pattern[patternlen]) {
88 strbuf_add(&pat_buf, pattern, patternlen);
89 use_pat = pat_buf.buf;
90 }
91 if (string[stringlen]) {
92 strbuf_add(&str_buf, string, stringlen);
93 use_str = str_buf.buf;
94 }
95
f30366b2
JH
96 if (ignore_case)
97 flags |= WM_CASEFOLD;
98 match_status = wildmatch(use_pat, use_str, flags, NULL);
0b6e56df
JH
99
100 strbuf_release(&pat_buf);
101 strbuf_release(&str_buf);
102
103 return match_status;
104}
105
f950eb95 106static size_t common_prefix_len(const char **pathspec)
3c6a370b 107{
4a085b16
JH
108 const char *n, *first;
109 size_t max = 0;
823ab40f 110 int literal = limit_pathspec_to_literal();
3c6a370b
LT
111
112 if (!pathspec)
4a085b16
JH
113 return max;
114
115 first = *pathspec;
116 while ((n = *pathspec++)) {
117 size_t i, len = 0;
118 for (i = 0; first == n || i < max; i++) {
119 char c = n[i];
823ab40f 120 if (!c || c != first[i] || (!literal && is_glob_special(c)))
4a085b16
JH
121 break;
122 if (c == '/')
123 len = i + 1;
124 }
125 if (first == n || len < max) {
126 max = len;
127 if (!max)
128 break;
129 }
3c6a370b 130 }
4a085b16 131 return max;
3c6a370b
LT
132}
133
f950eb95
CB
134/*
135 * Returns a copy of the longest leading path common among all
136 * pathspecs.
137 */
138char *common_prefix(const char **pathspec)
139{
140 unsigned long len = common_prefix_len(pathspec);
141
142 return len ? xmemdupz(*pathspec, len) : NULL;
143}
144
1d8842d9
LT
145int fill_directory(struct dir_struct *dir, const char **pathspec)
146{
4a085b16 147 size_t len;
1d8842d9
LT
148
149 /*
150 * Calculate common prefix for the pathspec, and
151 * use that to optimize the directory walk
152 */
4a085b16 153 len = common_prefix_len(pathspec);
1d8842d9
LT
154
155 /* Read the directory and prune it */
2b189435 156 read_directory(dir, pathspec ? *pathspec : "", len, pathspec);
dba2e203 157 return len;
1d8842d9
LT
158}
159
bc96cc87
NTND
160int within_depth(const char *name, int namelen,
161 int depth, int max_depth)
162{
163 const char *cp = name, *cpe = name + namelen;
164
165 while (cp < cpe) {
166 if (*cp++ != '/')
167 continue;
168 depth++;
169 if (depth > max_depth)
170 return 0;
171 }
172 return 1;
173}
174
e813d50e 175/*
2c5b0115 176 * Does 'match' match the given name?
e813d50e
JH
177 * A match is found if
178 *
179 * (1) the 'match' string is leading directory of 'name', or
180 * (2) the 'match' string is a wildcard and matches 'name', or
181 * (3) the 'match' string is exactly the same as 'name'.
182 *
183 * and the return value tells which case it was.
184 *
185 * It returns 0 when there is no match.
186 */
3c6a370b
LT
187static int match_one(const char *match, const char *name, int namelen)
188{
189 int matchlen;
823ab40f 190 int literal = limit_pathspec_to_literal();
3c6a370b
LT
191
192 /* If the match was just the prefix, we matched */
88ea8112 193 if (!*match)
e813d50e 194 return MATCHED_RECURSIVELY;
3c6a370b 195
21444f18
JJ
196 if (ignore_case) {
197 for (;;) {
198 unsigned char c1 = tolower(*match);
199 unsigned char c2 = tolower(*name);
823ab40f 200 if (c1 == '\0' || (!literal && is_glob_special(c1)))
21444f18
JJ
201 break;
202 if (c1 != c2)
203 return 0;
204 match++;
205 name++;
206 namelen--;
207 }
208 } else {
209 for (;;) {
210 unsigned char c1 = *match;
211 unsigned char c2 = *name;
823ab40f 212 if (c1 == '\0' || (!literal && is_glob_special(c1)))
21444f18
JJ
213 break;
214 if (c1 != c2)
215 return 0;
216 match++;
217 name++;
218 namelen--;
219 }
88ea8112
LT
220 }
221
3c6a370b
LT
222 /*
223 * If we don't match the matchstring exactly,
224 * we need to match by fnmatch
225 */
88ea8112 226 matchlen = strlen(match);
823ab40f
JK
227 if (strncmp_icase(match, name, matchlen)) {
228 if (literal)
229 return 0;
21444f18 230 return !fnmatch_icase(match, name, 0) ? MATCHED_FNMATCH : 0;
823ab40f 231 }
3c6a370b 232
f2d0df71 233 if (namelen == matchlen)
e813d50e
JH
234 return MATCHED_EXACTLY;
235 if (match[matchlen-1] == '/' || name[matchlen] == '/')
236 return MATCHED_RECURSIVELY;
237 return 0;
3c6a370b
LT
238}
239
e813d50e 240/*
52ed1894
AS
241 * Given a name and a list of pathspecs, returns the nature of the
242 * closest (i.e. most specific) match of the name to any of the
243 * pathspecs.
244 *
245 * The caller typically calls this multiple times with the same
246 * pathspec and seen[] array but with different name/namelen
247 * (e.g. entries from the index) and is interested in seeing if and
248 * how each pathspec matches all the names it calls this function
249 * with. A mark is left in the seen[] array for each pathspec element
250 * indicating the closest type of match that element achieved, so if
251 * seen[n] remains zero after multiple invocations, that means the nth
252 * pathspec did not match any names, which could indicate that the
253 * user mistyped the nth pathspec.
e813d50e 254 */
0b50922a
CB
255int match_pathspec(const char **pathspec, const char *name, int namelen,
256 int prefix, char *seen)
3c6a370b 257{
0b50922a
CB
258 int i, retval = 0;
259
260 if (!pathspec)
261 return 1;
3c6a370b
LT
262
263 name += prefix;
264 namelen -= prefix;
265
0b50922a 266 for (i = 0; pathspec[i] != NULL; i++) {
e813d50e 267 int how;
0b50922a
CB
268 const char *match = pathspec[i] + prefix;
269 if (seen && seen[i] == MATCHED_EXACTLY)
3c6a370b 270 continue;
e813d50e
JH
271 how = match_one(match, name, namelen);
272 if (how) {
273 if (retval < how)
274 retval = how;
0b50922a
CB
275 if (seen && seen[i] < how)
276 seen[i] = how;
3c6a370b
LT
277 }
278 }
279 return retval;
280}
281
61cf2820
NTND
282/*
283 * Does 'match' match the given name?
284 * A match is found if
285 *
286 * (1) the 'match' string is leading directory of 'name', or
287 * (2) the 'match' string is a wildcard and matches 'name', or
288 * (3) the 'match' string is exactly the same as 'name'.
289 *
290 * and the return value tells which case it was.
291 *
292 * It returns 0 when there is no match.
293 */
294static int match_pathspec_item(const struct pathspec_item *item, int prefix,
295 const char *name, int namelen)
296{
297 /* name/namelen has prefix cut off by caller */
298 const char *match = item->match + prefix;
299 int matchlen = item->len - prefix;
300
301 /* If the match was just the prefix, we matched */
302 if (!*match)
303 return MATCHED_RECURSIVELY;
304
305 if (matchlen <= namelen && !strncmp(match, name, matchlen)) {
306 if (matchlen == namelen)
307 return MATCHED_EXACTLY;
308
309 if (match[matchlen-1] == '/' || name[matchlen] == '/')
310 return MATCHED_RECURSIVELY;
311 }
312
5d74762d 313 if (item->nowildcard_len < item->len &&
8c6abbcd
NTND
314 !git_fnmatch(match, name,
315 item->flags & PATHSPEC_ONESTAR ? GFNM_ONESTAR : 0,
316 item->nowildcard_len - prefix))
61cf2820
NTND
317 return MATCHED_FNMATCH;
318
319 return 0;
320}
321
322/*
52ed1894
AS
323 * Given a name and a list of pathspecs, returns the nature of the
324 * closest (i.e. most specific) match of the name to any of the
325 * pathspecs.
326 *
327 * The caller typically calls this multiple times with the same
328 * pathspec and seen[] array but with different name/namelen
329 * (e.g. entries from the index) and is interested in seeing if and
330 * how each pathspec matches all the names it calls this function
331 * with. A mark is left in the seen[] array for each pathspec element
332 * indicating the closest type of match that element achieved, so if
333 * seen[n] remains zero after multiple invocations, that means the nth
334 * pathspec did not match any names, which could indicate that the
335 * user mistyped the nth pathspec.
61cf2820
NTND
336 */
337int match_pathspec_depth(const struct pathspec *ps,
338 const char *name, int namelen,
339 int prefix, char *seen)
340{
341 int i, retval = 0;
342
8f4f8f45
NTND
343 GUARD_PATHSPEC(ps, PATHSPEC_FROMTOP | PATHSPEC_MAXDEPTH);
344
61cf2820 345 if (!ps->nr) {
6330a171
NTND
346 if (!ps->recursive ||
347 !(ps->magic & PATHSPEC_MAXDEPTH) ||
348 ps->max_depth == -1)
61cf2820
NTND
349 return MATCHED_RECURSIVELY;
350
351 if (within_depth(name, namelen, 0, ps->max_depth))
352 return MATCHED_EXACTLY;
353 else
354 return 0;
355 }
356
357 name += prefix;
358 namelen -= prefix;
359
360 for (i = ps->nr - 1; i >= 0; i--) {
361 int how;
362 if (seen && seen[i] == MATCHED_EXACTLY)
363 continue;
364 how = match_pathspec_item(ps->items+i, prefix, name, namelen);
6330a171
NTND
365 if (ps->recursive &&
366 (ps->magic & PATHSPEC_MAXDEPTH) &&
367 ps->max_depth != -1 &&
61cf2820
NTND
368 how && how != MATCHED_FNMATCH) {
369 int len = ps->items[i].len;
370 if (name[len] == '/')
371 len++;
372 if (within_depth(name+len, namelen-len, 0, ps->max_depth))
373 how = MATCHED_EXACTLY;
374 else
375 how = 0;
376 }
377 if (how) {
378 if (retval < how)
379 retval = how;
380 if (seen && seen[i] < how)
381 seen[i] = how;
382 }
383 }
384 return retval;
385}
386
fcd631ed
NTND
387/*
388 * Return the length of the "simple" part of a path match limiter.
389 */
87323bda 390int simple_length(const char *match)
fcd631ed
NTND
391{
392 int len = -1;
393
394 for (;;) {
395 unsigned char c = *match++;
396 len++;
397 if (c == '\0' || is_glob_special(c))
398 return len;
399 }
400}
401
87323bda 402int no_wildcard(const char *string)
68492fc7 403{
fcd631ed 404 return string[simple_length(string)] == '\0';
68492fc7
LK
405}
406
82dce998
NTND
407void parse_exclude_pattern(const char **pattern,
408 int *patternlen,
409 int *flags,
410 int *nowildcardlen)
84460eec
NTND
411{
412 const char *p = *pattern;
413 size_t i, len;
414
415 *flags = 0;
416 if (*p == '!') {
417 *flags |= EXC_FLAG_NEGATIVE;
418 p++;
419 }
420 len = strlen(p);
421 if (len && p[len - 1] == '/') {
422 len--;
423 *flags |= EXC_FLAG_MUSTBEDIR;
424 }
425 for (i = 0; i < len; i++) {
426 if (p[i] == '/')
427 break;
428 }
429 if (i == len)
430 *flags |= EXC_FLAG_NODIR;
431 *nowildcardlen = simple_length(p);
432 /*
433 * we should have excluded the trailing slash from 'p' too,
434 * but that's one more allocation. Instead just make sure
435 * nowildcardlen does not exceed real patternlen
436 */
437 if (*nowildcardlen > len)
438 *nowildcardlen = len;
439 if (*p == '*' && no_wildcard(p + 1))
440 *flags |= EXC_FLAG_ENDSWITH;
441 *pattern = p;
442 *patternlen = len;
443}
444
453ec4bd 445void add_exclude(const char *string, const char *base,
c04318e4 446 int baselen, struct exclude_list *el, int srcpos)
453ec4bd 447{
d6b8fc30 448 struct exclude *x;
84460eec
NTND
449 int patternlen;
450 int flags;
451 int nowildcardlen;
453ec4bd 452
84460eec
NTND
453 parse_exclude_pattern(&string, &patternlen, &flags, &nowildcardlen);
454 if (flags & EXC_FLAG_MUSTBEDIR) {
d6b8fc30 455 char *s;
84460eec 456 x = xmalloc(sizeof(*x) + patternlen + 1);
4b25d091 457 s = (char *)(x+1);
84460eec
NTND
458 memcpy(s, string, patternlen);
459 s[patternlen] = '\0';
d6b8fc30 460 x->pattern = s;
d6b8fc30
JH
461 } else {
462 x = xmalloc(sizeof(*x));
463 x->pattern = string;
464 }
84460eec
NTND
465 x->patternlen = patternlen;
466 x->nowildcardlen = nowildcardlen;
453ec4bd
LT
467 x->base = base;
468 x->baselen = baselen;
d6b8fc30 469 x->flags = flags;
c04318e4 470 x->srcpos = srcpos;
840fc334
AS
471 ALLOC_GROW(el->excludes, el->nr + 1, el->alloc);
472 el->excludes[el->nr++] = x;
c04318e4 473 x->el = el;
453ec4bd
LT
474}
475
c28b3d6e
NTND
476static void *read_skip_worktree_file_from_index(const char *path, size_t *size)
477{
478 int pos, len;
479 unsigned long sz;
480 enum object_type type;
481 void *data;
482 struct index_state *istate = &the_index;
483
484 len = strlen(path);
485 pos = index_name_pos(istate, path, len);
486 if (pos < 0)
487 return NULL;
488 if (!ce_skip_worktree(istate->cache[pos]))
489 return NULL;
490 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
491 if (!data || type != OBJ_BLOB) {
492 free(data);
493 return NULL;
494 }
495 *size = xsize_t(sz);
496 return data;
497}
498
f6198812
AS
499/*
500 * Frees memory within el which was allocated for exclude patterns and
501 * the file buffer. Does not free el itself.
502 */
503void clear_exclude_list(struct exclude_list *el)
0fd0e241
NTND
504{
505 int i;
506
507 for (i = 0; i < el->nr; i++)
508 free(el->excludes[i]);
509 free(el->excludes);
c082df24 510 free(el->filebuf);
0fd0e241
NTND
511
512 el->nr = 0;
513 el->excludes = NULL;
c082df24 514 el->filebuf = NULL;
0fd0e241
NTND
515}
516
cb097534
NTND
517int add_excludes_from_file_to_list(const char *fname,
518 const char *base,
519 int baselen,
840fc334 520 struct exclude_list *el,
cb097534 521 int check_index)
453ec4bd 522{
c470701a 523 struct stat st;
c04318e4 524 int fd, i, lineno = 1;
9d14017a 525 size_t size = 0;
453ec4bd
LT
526 char *buf, *entry;
527
528 fd = open(fname, O_RDONLY);
c28b3d6e 529 if (fd < 0 || fstat(fd, &st) < 0) {
69660731 530 if (errno != ENOENT)
55b38a48 531 warn_on_inaccessible(fname);
c28b3d6e
NTND
532 if (0 <= fd)
533 close(fd);
534 if (!check_index ||
535 (buf = read_skip_worktree_file_from_index(fname, &size)) == NULL)
536 return -1;
45d76f17
NTND
537 if (size == 0) {
538 free(buf);
539 return 0;
540 }
541 if (buf[size-1] != '\n') {
542 buf = xrealloc(buf, size+1);
543 buf[size++] = '\n';
544 }
453ec4bd 545 }
c28b3d6e
NTND
546 else {
547 size = xsize_t(st.st_size);
548 if (size == 0) {
549 close(fd);
550 return 0;
551 }
45d76f17 552 buf = xmalloc(size+1);
c28b3d6e 553 if (read_in_full(fd, buf, size) != size) {
45d76f17 554 free(buf);
c28b3d6e
NTND
555 close(fd);
556 return -1;
557 }
45d76f17 558 buf[size++] = '\n';
c28b3d6e 559 close(fd);
6ba78238 560 }
453ec4bd 561
c082df24 562 el->filebuf = buf;
453ec4bd 563 entry = buf;
45d76f17
NTND
564 for (i = 0; i < size; i++) {
565 if (buf[i] == '\n') {
453ec4bd
LT
566 if (entry != buf + i && entry[0] != '#') {
567 buf[i - (i && buf[i-1] == '\r')] = 0;
c04318e4 568 add_exclude(entry, base, baselen, el, lineno);
453ec4bd 569 }
c04318e4 570 lineno++;
453ec4bd
LT
571 entry = buf + i + 1;
572 }
573 }
574 return 0;
453ec4bd
LT
575}
576
c04318e4
AS
577struct exclude_list *add_exclude_list(struct dir_struct *dir,
578 int group_type, const char *src)
c082df24
AS
579{
580 struct exclude_list *el;
581 struct exclude_list_group *group;
582
583 group = &dir->exclude_list_group[group_type];
584 ALLOC_GROW(group->el, group->nr + 1, group->alloc);
585 el = &group->el[group->nr++];
586 memset(el, 0, sizeof(*el));
c04318e4 587 el->src = src;
c082df24
AS
588 return el;
589}
590
591/*
592 * Used to set up core.excludesfile and .git/info/exclude lists.
593 */
453ec4bd
LT
594void add_excludes_from_file(struct dir_struct *dir, const char *fname)
595{
c082df24 596 struct exclude_list *el;
c04318e4 597 el = add_exclude_list(dir, EXC_FILE, fname);
c082df24 598 if (add_excludes_from_file_to_list(fname, "", 0, el, 0) < 0)
453ec4bd
LT
599 die("cannot use %s as an exclude file", fname);
600}
601
82dce998
NTND
602int match_basename(const char *basename, int basenamelen,
603 const char *pattern, int prefix, int patternlen,
604 int flags)
593cb880
NTND
605{
606 if (prefix == patternlen) {
0b6e56df
JH
607 if (patternlen == basenamelen &&
608 !strncmp_icase(pattern, basename, basenamelen))
593cb880
NTND
609 return 1;
610 } else if (flags & EXC_FLAG_ENDSWITH) {
0b6e56df 611 /* "*literal" matching against "fooliteral" */
593cb880 612 if (patternlen - 1 <= basenamelen &&
0b6e56df
JH
613 !strncmp_icase(pattern + 1,
614 basename + basenamelen - (patternlen - 1),
615 patternlen - 1))
593cb880
NTND
616 return 1;
617 } else {
0b6e56df
JH
618 if (fnmatch_icase_mem(pattern, patternlen,
619 basename, basenamelen,
620 0) == 0)
593cb880
NTND
621 return 1;
622 }
623 return 0;
624}
625
82dce998
NTND
626int match_pathname(const char *pathname, int pathlen,
627 const char *base, int baselen,
628 const char *pattern, int prefix, int patternlen,
629 int flags)
b5592632
NTND
630{
631 const char *name;
632 int namelen;
633
634 /*
635 * match with FNM_PATHNAME; the pattern has base implicitly
636 * in front of it.
637 */
638 if (*pattern == '/') {
639 pattern++;
982ac873 640 patternlen--;
b5592632
NTND
641 prefix--;
642 }
643
644 /*
645 * baselen does not count the trailing slash. base[] may or
646 * may not end with a trailing slash though.
647 */
648 if (pathlen < baselen + 1 ||
649 (baselen && pathname[baselen] != '/') ||
650 strncmp_icase(pathname, base, baselen))
651 return 0;
652
653 namelen = baselen ? pathlen - baselen - 1 : pathlen;
654 name = pathname + pathlen - namelen;
655
656 if (prefix) {
657 /*
658 * if the non-wildcard part is longer than the
659 * remaining pathname, surely it cannot match.
660 */
661 if (prefix > namelen)
662 return 0;
663
664 if (strncmp_icase(pattern, name, prefix))
665 return 0;
666 pattern += prefix;
ab3aebc1 667 patternlen -= prefix;
b5592632
NTND
668 name += prefix;
669 namelen -= prefix;
ab3aebc1
JK
670
671 /*
672 * If the whole pattern did not have a wildcard,
673 * then our prefix match is all we need; we
674 * do not need to call fnmatch at all.
675 */
676 if (!patternlen && !namelen)
677 return 1;
b5592632
NTND
678 }
679
ab3aebc1
JK
680 return fnmatch_icase_mem(pattern, patternlen,
681 name, namelen,
f30366b2 682 WM_PATHNAME) == 0;
b5592632
NTND
683}
684
578cd7c3
AS
685/*
686 * Scan the given exclude list in reverse to see whether pathname
687 * should be ignored. The first match (i.e. the last on the list), if
688 * any, determines the fate. Returns the exclude_list element which
689 * matched, or NULL for undecided.
453ec4bd 690 */
578cd7c3
AS
691static struct exclude *last_exclude_matching_from_list(const char *pathname,
692 int pathlen,
693 const char *basename,
694 int *dtype,
695 struct exclude_list *el)
453ec4bd
LT
696{
697 int i;
698
35a94d44 699 if (!el->nr)
578cd7c3 700 return NULL; /* undefined */
d6b8fc30 701
35a94d44
NTND
702 for (i = el->nr - 1; 0 <= i; i--) {
703 struct exclude *x = el->excludes[i];
b5592632 704 const char *exclude = x->pattern;
b5592632 705 int prefix = x->nowildcardlen;
35a94d44
NTND
706
707 if (x->flags & EXC_FLAG_MUSTBEDIR) {
708 if (*dtype == DT_UNKNOWN)
709 *dtype = get_dtype(NULL, pathname, pathlen);
710 if (*dtype != DT_DIR)
711 continue;
712 }
d6b8fc30 713
35a94d44 714 if (x->flags & EXC_FLAG_NODIR) {
593cb880
NTND
715 if (match_basename(basename,
716 pathlen - (basename - pathname),
717 exclude, prefix, x->patternlen,
718 x->flags))
578cd7c3 719 return x;
35a94d44
NTND
720 continue;
721 }
722
b5592632
NTND
723 assert(x->baselen == 0 || x->base[x->baselen - 1] == '/');
724 if (match_pathname(pathname, pathlen,
725 x->base, x->baselen ? x->baselen - 1 : 0,
726 exclude, prefix, x->patternlen, x->flags))
578cd7c3 727 return x;
453ec4bd 728 }
578cd7c3
AS
729 return NULL; /* undecided */
730}
731
732/*
733 * Scan the list and let the last match determine the fate.
734 * Return 1 for exclude, 0 for include and -1 for undecided.
735 */
736int is_excluded_from_list(const char *pathname,
737 int pathlen, const char *basename, int *dtype,
738 struct exclude_list *el)
739{
740 struct exclude *exclude;
741 exclude = last_exclude_matching_from_list(pathname, pathlen, basename, dtype, el);
742 if (exclude)
743 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
453ec4bd
LT
744 return -1; /* undecided */
745}
746
46aa2f95
KB
747static struct exclude *last_exclude_matching_from_lists(struct dir_struct *dir,
748 const char *pathname, int pathlen, const char *basename,
749 int *dtype_p)
750{
751 int i, j;
752 struct exclude_list_group *group;
753 struct exclude *exclude;
754 for (i = EXC_CMDL; i <= EXC_FILE; i++) {
755 group = &dir->exclude_list_group[i];
756 for (j = group->nr - 1; j >= 0; j--) {
757 exclude = last_exclude_matching_from_list(
758 pathname, pathlen, basename, dtype_p,
759 &group->el[j]);
760 if (exclude)
761 return exclude;
762 }
763 }
764 return NULL;
765}
766
6cd5c582
KB
767/*
768 * Loads the per-directory exclude list for the substring of base
769 * which has a char length of baselen.
770 */
771static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
772{
773 struct exclude_list_group *group;
774 struct exclude_list *el;
775 struct exclude_stack *stk = NULL;
776 int current;
777
6cd5c582
KB
778 group = &dir->exclude_list_group[EXC_DIRS];
779
780 /* Pop the exclude lists from the EXCL_DIRS exclude_list_group
781 * which originate from directories not in the prefix of the
782 * path being checked. */
783 while ((stk = dir->exclude_stack) != NULL) {
784 if (stk->baselen <= baselen &&
785 !strncmp(dir->basebuf, base, stk->baselen))
786 break;
787 el = &group->el[dir->exclude_stack->exclude_ix];
788 dir->exclude_stack = stk->prev;
95c6f271 789 dir->exclude = NULL;
6cd5c582
KB
790 free((char *)el->src); /* see strdup() below */
791 clear_exclude_list(el);
792 free(stk);
793 group->nr--;
794 }
795
95c6f271
KB
796 /* Skip traversing into sub directories if the parent is excluded */
797 if (dir->exclude)
798 return;
799
6cd5c582
KB
800 /* Read from the parent directories and push them down. */
801 current = stk ? stk->baselen : -1;
802 while (current < baselen) {
803 struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
804 const char *cp;
805
806 if (current < 0) {
807 cp = base;
808 current = 0;
809 }
810 else {
811 cp = strchr(base + current + 1, '/');
812 if (!cp)
813 die("oops in prep_exclude");
814 cp++;
815 }
816 stk->prev = dir->exclude_stack;
817 stk->baselen = cp - base;
95c6f271
KB
818 stk->exclude_ix = group->nr;
819 el = add_exclude_list(dir, EXC_DIRS, NULL);
6cd5c582
KB
820 memcpy(dir->basebuf + current, base + current,
821 stk->baselen - current);
95c6f271
KB
822
823 /* Abort if the directory is excluded */
824 if (stk->baselen) {
825 int dt = DT_DIR;
826 dir->basebuf[stk->baselen - 1] = 0;
827 dir->exclude = last_exclude_matching_from_lists(dir,
828 dir->basebuf, stk->baselen - 1,
829 dir->basebuf + current, &dt);
830 dir->basebuf[stk->baselen - 1] = '/';
c3c327de
KB
831 if (dir->exclude &&
832 dir->exclude->flags & EXC_FLAG_NEGATIVE)
833 dir->exclude = NULL;
95c6f271
KB
834 if (dir->exclude) {
835 dir->basebuf[stk->baselen] = 0;
836 dir->exclude_stack = stk;
837 return;
838 }
839 }
840
841 /* Try to read per-directory file unless path is too long */
842 if (dir->exclude_per_dir &&
843 stk->baselen + strlen(dir->exclude_per_dir) < PATH_MAX) {
844 strcpy(dir->basebuf + stk->baselen,
845 dir->exclude_per_dir);
846 /*
847 * dir->basebuf gets reused by the traversal, but we
848 * need fname to remain unchanged to ensure the src
849 * member of each struct exclude correctly
850 * back-references its source file. Other invocations
851 * of add_exclude_list provide stable strings, so we
852 * strdup() and free() here in the caller.
853 */
854 el->src = strdup(dir->basebuf);
855 add_excludes_from_file_to_list(dir->basebuf,
856 dir->basebuf, stk->baselen, el, 1);
857 }
6cd5c582
KB
858 dir->exclude_stack = stk;
859 current = stk->baselen;
860 }
861 dir->basebuf[baselen] = '\0';
862}
863
f4cd69a6
AS
864/*
865 * Loads the exclude lists for the directory containing pathname, then
866 * scans all exclude lists to determine whether pathname is excluded.
867 * Returns the exclude_list element which matched, or NULL for
868 * undecided.
869 */
b07bc8c8 870struct exclude *last_exclude_matching(struct dir_struct *dir,
f4cd69a6
AS
871 const char *pathname,
872 int *dtype_p)
453ec4bd
LT
873{
874 int pathlen = strlen(pathname);
68492fc7
LK
875 const char *basename = strrchr(pathname, '/');
876 basename = (basename) ? basename+1 : pathname;
453ec4bd 877
63d285c8 878 prep_exclude(dir, pathname, basename-pathname);
c082df24 879
95c6f271
KB
880 if (dir->exclude)
881 return dir->exclude;
882
46aa2f95
KB
883 return last_exclude_matching_from_lists(dir, pathname, pathlen,
884 basename, dtype_p);
f4cd69a6
AS
885}
886
887/*
888 * Loads the exclude lists for the directory containing pathname, then
889 * scans all exclude lists to determine whether pathname is excluded.
890 * Returns 1 if true, otherwise 0.
891 */
b07bc8c8 892int is_excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
f4cd69a6
AS
893{
894 struct exclude *exclude =
895 last_exclude_matching(dir, pathname, dtype_p);
896 if (exclude)
897 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
453ec4bd
LT
898 return 0;
899}
900
f3fa1838
JH
901static struct dir_entry *dir_entry_new(const char *pathname, int len)
902{
453ec4bd
LT
903 struct dir_entry *ent;
904
453ec4bd
LT
905 ent = xmalloc(sizeof(*ent) + len + 1);
906 ent->len = len;
907 memcpy(ent->name, pathname, len);
908 ent->name[len] = 0;
4d06f8ac 909 return ent;
453ec4bd
LT
910}
911
159b3212 912static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
6815e569 913{
5bd8e2d8 914 if (cache_name_exists(pathname, len, ignore_case))
6815e569
JK
915 return NULL;
916
25fd2f7a 917 ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
6815e569
JK
918 return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
919}
920
108da0db 921struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
2abd31b0 922{
6e4f981f 923 if (!cache_name_is_other(pathname, len))
2abd31b0
JK
924 return NULL;
925
25fd2f7a 926 ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
2abd31b0
JK
927 return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
928}
929
09595258
LT
930enum exist_status {
931 index_nonexistent = 0,
932 index_directory,
4b05548f 933 index_gitdir
09595258
LT
934};
935
5102c617
JJ
936/*
937 * Do not use the alphabetically stored index to look up
938 * the directory name; instead, use the case insensitive
939 * name hash.
940 */
941static enum exist_status directory_exists_in_index_icase(const char *dirname, int len)
942{
943 struct cache_entry *ce = index_name_exists(&the_index, dirname, len + 1, ignore_case);
944 unsigned char endchar;
945
946 if (!ce)
947 return index_nonexistent;
948 endchar = ce->name[len];
949
950 /*
951 * The cache_entry structure returned will contain this dirname
952 * and possibly additional path components.
953 */
954 if (endchar == '/')
955 return index_directory;
956
957 /*
958 * If there are no additional path components, then this cache_entry
959 * represents a submodule. Submodules, despite being directories,
960 * are stored in the cache without a closing slash.
961 */
962 if (!endchar && S_ISGITLINK(ce->ce_mode))
963 return index_gitdir;
964
965 /* This should never be hit, but it exists just in case. */
966 return index_nonexistent;
967}
968
09595258
LT
969/*
970 * The index sorts alphabetically by entry name, which
971 * means that a gitlink sorts as '\0' at the end, while
972 * a directory (which is defined not as an entry, but as
973 * the files it contains) will sort with the '/' at the
974 * end.
975 */
976static enum exist_status directory_exists_in_index(const char *dirname, int len)
453ec4bd 977{
5102c617
JJ
978 int pos;
979
980 if (ignore_case)
981 return directory_exists_in_index_icase(dirname, len);
982
983 pos = cache_name_pos(dirname, len);
09595258
LT
984 if (pos < 0)
985 pos = -pos-1;
986 while (pos < active_nr) {
987 struct cache_entry *ce = active_cache[pos++];
988 unsigned char endchar;
989
990 if (strncmp(ce->name, dirname, len))
991 break;
992 endchar = ce->name[len];
993 if (endchar > '/')
994 break;
995 if (endchar == '/')
996 return index_directory;
7a51ed66 997 if (!endchar && S_ISGITLINK(ce->ce_mode))
09595258
LT
998 return index_gitdir;
999 }
1000 return index_nonexistent;
1001}
1002
1003/*
1004 * When we find a directory when traversing the filesystem, we
1005 * have three distinct cases:
1006 *
1007 * - ignore it
1008 * - see it as a directory
1009 * - recurse into it
1010 *
1011 * and which one we choose depends on a combination of existing
1012 * git index contents and the flags passed into the directory
1013 * traversal routine.
1014 *
1015 * Case 1: If we *already* have entries in the index under that
5bd8e2d8
KB
1016 * directory name, we always recurse into the directory to see
1017 * all the files.
09595258
LT
1018 *
1019 * Case 2: If we *already* have that directory name as a gitlink,
1020 * we always continue to see it as a gitlink, regardless of whether
1021 * there is an actual git directory there or not (it might not
1022 * be checked out as a subproject!)
1023 *
1024 * Case 3: if we didn't have it in the index previously, we
1025 * have a few sub-cases:
1026 *
1027 * (a) if "show_other_directories" is true, we show it as
1028 * just a directory, unless "hide_empty_directories" is
defd7c7b
KB
1029 * also true, in which case we need to check if it contains any
1030 * untracked and / or ignored files.
09595258 1031 * (b) if it looks like a git directory, and we don't have
302b9282 1032 * 'no_gitlinks' set we treat it as a gitlink, and show it
09595258
LT
1033 * as a directory.
1034 * (c) otherwise, we recurse into it.
1035 */
defd7c7b 1036static enum path_treatment treat_directory(struct dir_struct *dir,
721ac4ed 1037 const char *dirname, int len, int exclude,
09595258
LT
1038 const struct path_simplify *simplify)
1039{
1040 /* The "len-1" is to strip the final '/' */
1041 switch (directory_exists_in_index(dirname, len-1)) {
1042 case index_directory:
defd7c7b 1043 return path_recurse;
09595258
LT
1044
1045 case index_gitdir:
26c986e1 1046 return path_none;
09595258
LT
1047
1048 case index_nonexistent:
7c4c97c0 1049 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
09595258 1050 break;
7c4c97c0 1051 if (!(dir->flags & DIR_NO_GITLINKS)) {
09595258
LT
1052 unsigned char sha1[20];
1053 if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
defd7c7b 1054 return path_untracked;
09595258 1055 }
defd7c7b 1056 return path_recurse;
09595258
LT
1057 }
1058
1059 /* This is the "show_other_directories" case */
721ac4ed 1060
184d2a8e 1061 if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
defd7c7b
KB
1062 return exclude ? path_excluded : path_untracked;
1063
1064 return read_directory_recursive(dir, dirname, len, 1, simplify);
453ec4bd
LT
1065}
1066
9fc42d60
LT
1067/*
1068 * This is an inexact early pruning of any recursive directory
1069 * reading - if the path cannot possibly be in the pathspec,
1070 * return true, and we'll skip it early.
1071 */
1072static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
1073{
1074 if (simplify) {
1075 for (;;) {
1076 const char *match = simplify->path;
1077 int len = simplify->len;
1078
1079 if (!match)
1080 break;
1081 if (len > pathlen)
1082 len = pathlen;
1083 if (!memcmp(path, match, len))
1084 return 0;
1085 simplify++;
1086 }
1087 return 1;
1088 }
1089 return 0;
1090}
1091
29209cbe
JK
1092/*
1093 * This function tells us whether an excluded path matches a
1094 * list of "interesting" pathspecs. That is, whether a path matched
1095 * by any of the pathspecs could possibly be ignored by excluding
1096 * the specified path. This can happen if:
1097 *
1098 * 1. the path is mentioned explicitly in the pathspec
1099 *
1100 * 2. the path is a directory prefix of some element in the
1101 * pathspec
1102 */
1103static int exclude_matches_pathspec(const char *path, int len,
1104 const struct path_simplify *simplify)
e96980ef
JK
1105{
1106 if (simplify) {
1107 for (; simplify->path; simplify++) {
1108 if (len == simplify->len
1109 && !memcmp(path, simplify->path, len))
1110 return 1;
29209cbe
JK
1111 if (len < simplify->len
1112 && simplify->path[len] == '/'
1113 && !memcmp(path, simplify->path, len))
1114 return 1;
e96980ef
JK
1115 }
1116 }
1117 return 0;
1118}
1119
443e061a
LT
1120static int get_index_dtype(const char *path, int len)
1121{
1122 int pos;
1123 struct cache_entry *ce;
1124
1125 ce = cache_name_exists(path, len, 0);
1126 if (ce) {
1127 if (!ce_uptodate(ce))
1128 return DT_UNKNOWN;
1129 if (S_ISGITLINK(ce->ce_mode))
1130 return DT_DIR;
1131 /*
1132 * Nobody actually cares about the
1133 * difference between DT_LNK and DT_REG
1134 */
1135 return DT_REG;
1136 }
1137
1138 /* Try to look it up as a directory */
1139 pos = cache_name_pos(path, len);
1140 if (pos >= 0)
1141 return DT_UNKNOWN;
1142 pos = -pos-1;
1143 while (pos < active_nr) {
1144 ce = active_cache[pos++];
1145 if (strncmp(ce->name, path, len))
1146 break;
1147 if (ce->name[len] > '/')
1148 break;
1149 if (ce->name[len] < '/')
1150 continue;
1151 if (!ce_uptodate(ce))
1152 break; /* continue? */
1153 return DT_DIR;
1154 }
1155 return DT_UNKNOWN;
1156}
1157
caa6b782 1158static int get_dtype(struct dirent *de, const char *path, int len)
07134421 1159{
6831a88a 1160 int dtype = de ? DTYPE(de) : DT_UNKNOWN;
07134421
LT
1161 struct stat st;
1162
1163 if (dtype != DT_UNKNOWN)
1164 return dtype;
443e061a
LT
1165 dtype = get_index_dtype(path, len);
1166 if (dtype != DT_UNKNOWN)
1167 return dtype;
1168 if (lstat(path, &st))
07134421
LT
1169 return dtype;
1170 if (S_ISREG(st.st_mode))
1171 return DT_REG;
1172 if (S_ISDIR(st.st_mode))
1173 return DT_DIR;
1174 if (S_ISLNK(st.st_mode))
1175 return DT_LNK;
1176 return dtype;
1177}
1178
16e2cfa9 1179static enum path_treatment treat_one_path(struct dir_struct *dir,
49dc2cc2 1180 struct strbuf *path,
16e2cfa9
JH
1181 const struct path_simplify *simplify,
1182 int dtype, struct dirent *de)
53cc5356 1183{
8aaf8d77
KB
1184 int exclude;
1185 if (dtype == DT_UNKNOWN)
1186 dtype = get_dtype(de, path->buf, path->len);
1187
1188 /* Always exclude indexed files */
1189 if (dtype != DT_DIR &&
1190 cache_name_exists(path->buf, path->len, ignore_case))
defd7c7b 1191 return path_none;
8aaf8d77
KB
1192
1193 exclude = is_excluded(dir, path->buf, &dtype);
53cc5356
JH
1194
1195 /*
1196 * Excluded? If we don't explicitly want to show
1197 * ignored files, ignore it
1198 */
0aaf62b6 1199 if (exclude && !(dir->flags & (DIR_SHOW_IGNORED|DIR_SHOW_IGNORED_TOO)))
defd7c7b 1200 return path_excluded;
53cc5356 1201
53cc5356
JH
1202 switch (dtype) {
1203 default:
defd7c7b 1204 return path_none;
53cc5356 1205 case DT_DIR:
49dc2cc2 1206 strbuf_addch(path, '/');
defd7c7b
KB
1207 return treat_directory(dir, path->buf, path->len, exclude,
1208 simplify);
53cc5356
JH
1209 case DT_REG:
1210 case DT_LNK:
defd7c7b 1211 return exclude ? path_excluded : path_untracked;
53cc5356 1212 }
53cc5356
JH
1213}
1214
16e2cfa9
JH
1215static enum path_treatment treat_path(struct dir_struct *dir,
1216 struct dirent *de,
49dc2cc2 1217 struct strbuf *path,
16e2cfa9 1218 int baselen,
49dc2cc2 1219 const struct path_simplify *simplify)
16e2cfa9
JH
1220{
1221 int dtype;
1222
1223 if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
defd7c7b 1224 return path_none;
49dc2cc2
RS
1225 strbuf_setlen(path, baselen);
1226 strbuf_addstr(path, de->d_name);
1227 if (simplify_away(path->buf, path->len, simplify))
defd7c7b 1228 return path_none;
16e2cfa9
JH
1229
1230 dtype = DTYPE(de);
49dc2cc2 1231 return treat_one_path(dir, path, simplify, dtype, de);
16e2cfa9
JH
1232}
1233
453ec4bd
LT
1234/*
1235 * Read a directory tree. We currently ignore anything but
1236 * directories, regular files and symlinks. That's because git
1237 * doesn't handle them at all yet. Maybe that will change some
1238 * day.
1239 *
1240 * Also, we ignore the name ".git" (even if it is not a directory).
1241 * That likely will not change.
defd7c7b
KB
1242 *
1243 * Returns the most significant path_treatment value encountered in the scan.
453ec4bd 1244 */
defd7c7b 1245static enum path_treatment read_directory_recursive(struct dir_struct *dir,
53cc5356
JH
1246 const char *base, int baselen,
1247 int check_only,
1248 const struct path_simplify *simplify)
453ec4bd 1249{
1528d247 1250 DIR *fdir;
defd7c7b 1251 enum path_treatment state, subdir_state, dir_state = path_none;
02cb6753 1252 struct dirent *de;
bef36921 1253 struct strbuf path = STRBUF_INIT;
453ec4bd 1254
bef36921 1255 strbuf_add(&path, base, baselen);
02cb6753 1256
1528d247
RS
1257 fdir = opendir(path.len ? path.buf : ".");
1258 if (!fdir)
1259 goto out;
1260
02cb6753 1261 while ((de = readdir(fdir)) != NULL) {
defd7c7b
KB
1262 /* check how the file or directory should be treated */
1263 state = treat_path(dir, de, &path, baselen, simplify);
1264 if (state > dir_state)
1265 dir_state = state;
1266
1267 /* recurse into subdir if instructed by treat_path */
1268 if (state == path_recurse) {
1269 subdir_state = read_directory_recursive(dir, path.buf,
289ff559 1270 path.len, check_only, simplify);
defd7c7b
KB
1271 if (subdir_state > dir_state)
1272 dir_state = subdir_state;
1273 }
1274
1275 if (check_only) {
1276 /* abort early if maximum state has been reached */
1277 if (dir_state == path_untracked)
1278 break;
1279 /* skip the dir_add_* part */
02cb6753 1280 continue;
453ec4bd 1281 }
defd7c7b
KB
1282
1283 /* add the path to the appropriate result list */
1284 switch (state) {
1285 case path_excluded:
1286 if (dir->flags & DIR_SHOW_IGNORED)
1287 dir_add_name(dir, path.buf, path.len);
0aaf62b6
KB
1288 else if ((dir->flags & DIR_SHOW_IGNORED_TOO) ||
1289 ((dir->flags & DIR_COLLECT_IGNORED) &&
1290 exclude_matches_pathspec(path.buf, path.len,
1291 simplify)))
1292 dir_add_ignored(dir, path.buf, path.len);
defd7c7b
KB
1293 break;
1294
1295 case path_untracked:
1296 if (!(dir->flags & DIR_SHOW_IGNORED))
1297 dir_add_name(dir, path.buf, path.len);
1528d247 1298 break;
defd7c7b
KB
1299
1300 default:
1301 break;
1302 }
453ec4bd 1303 }
02cb6753 1304 closedir(fdir);
1528d247 1305 out:
bef36921 1306 strbuf_release(&path);
453ec4bd 1307
defd7c7b 1308 return dir_state;
453ec4bd
LT
1309}
1310
1311static int cmp_name(const void *p1, const void *p2)
1312{
1313 const struct dir_entry *e1 = *(const struct dir_entry **)p1;
1314 const struct dir_entry *e2 = *(const struct dir_entry **)p2;
1315
1316 return cache_name_compare(e1->name, e1->len,
1317 e2->name, e2->len);
1318}
1319
9fc42d60
LT
1320static struct path_simplify *create_simplify(const char **pathspec)
1321{
1322 int nr, alloc = 0;
1323 struct path_simplify *simplify = NULL;
1324
1325 if (!pathspec)
1326 return NULL;
1327
1328 for (nr = 0 ; ; nr++) {
1329 const char *match;
1330 if (nr >= alloc) {
1331 alloc = alloc_nr(alloc);
1332 simplify = xrealloc(simplify, alloc * sizeof(*simplify));
1333 }
1334 match = *pathspec++;
1335 if (!match)
1336 break;
1337 simplify[nr].path = match;
1338 simplify[nr].len = simple_length(match);
1339 }
1340 simplify[nr].path = NULL;
1341 simplify[nr].len = 0;
1342 return simplify;
1343}
1344
1345static void free_simplify(struct path_simplify *simplify)
1346{
8e0f7003 1347 free(simplify);
9fc42d60
LT
1348}
1349
48ffef96
JH
1350static int treat_leading_path(struct dir_struct *dir,
1351 const char *path, int len,
1352 const struct path_simplify *simplify)
1353{
49dc2cc2
RS
1354 struct strbuf sb = STRBUF_INIT;
1355 int baselen, rc = 0;
48ffef96 1356 const char *cp;
be8a84c5 1357 int old_flags = dir->flags;
48ffef96
JH
1358
1359 while (len && path[len - 1] == '/')
1360 len--;
1361 if (!len)
1362 return 1;
1363 baselen = 0;
be8a84c5 1364 dir->flags &= ~DIR_SHOW_OTHER_DIRECTORIES;
48ffef96
JH
1365 while (1) {
1366 cp = path + baselen + !!baselen;
1367 cp = memchr(cp, '/', path + len - cp);
1368 if (!cp)
1369 baselen = len;
1370 else
1371 baselen = cp - path;
49dc2cc2
RS
1372 strbuf_setlen(&sb, 0);
1373 strbuf_add(&sb, path, baselen);
1374 if (!is_directory(sb.buf))
1375 break;
1376 if (simplify_away(sb.buf, sb.len, simplify))
1377 break;
1378 if (treat_one_path(dir, &sb, simplify,
defd7c7b 1379 DT_DIR, NULL) == path_none)
49dc2cc2
RS
1380 break; /* do not recurse into it */
1381 if (len <= baselen) {
1382 rc = 1;
1383 break; /* finished checking */
1384 }
48ffef96 1385 }
49dc2cc2 1386 strbuf_release(&sb);
be8a84c5 1387 dir->flags = old_flags;
49dc2cc2 1388 return rc;
48ffef96
JH
1389}
1390
dba2e203 1391int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
9fc42d60 1392{
725b0605 1393 struct path_simplify *simplify;
b4189aa8 1394
dba2e203 1395 if (has_symlink_leading_path(path, len))
725b0605
JH
1396 return dir->nr;
1397
1398 simplify = create_simplify(pathspec);
48ffef96
JH
1399 if (!len || treat_leading_path(dir, path, len, simplify))
1400 read_directory_recursive(dir, path, len, 0, simplify);
9fc42d60 1401 free_simplify(simplify);
453ec4bd 1402 qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
2abd31b0 1403 qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
453ec4bd
LT
1404 return dir->nr;
1405}
c91f0d92 1406
686a4a06 1407int file_exists(const char *f)
c91f0d92 1408{
686a4a06 1409 struct stat sb;
a50f9fc5 1410 return lstat(f, &sb) == 0;
c91f0d92 1411}
e6636747
JS
1412
1413/*
9b125da4
NTND
1414 * Given two normalized paths (a trailing slash is ok), if subdir is
1415 * outside dir, return -1. Otherwise return the offset in subdir that
1416 * can be used as relative path to dir.
e6636747 1417 */
9b125da4 1418int dir_inside_of(const char *subdir, const char *dir)
e6636747 1419{
9b125da4 1420 int offset = 0;
e6636747 1421
9b125da4 1422 assert(dir && subdir && *dir && *subdir);
e6636747 1423
9b125da4 1424 while (*dir && *subdir && *dir == *subdir) {
e6636747 1425 dir++;
9b125da4
NTND
1426 subdir++;
1427 offset++;
490544b1 1428 }
9b125da4
NTND
1429
1430 /* hel[p]/me vs hel[l]/yeah */
1431 if (*dir && *subdir)
1432 return -1;
1433
1434 if (!*subdir)
1435 return !*dir ? offset : -1; /* same dir */
1436
1437 /* foo/[b]ar vs foo/[] */
1438 if (is_dir_sep(dir[-1]))
1439 return is_dir_sep(subdir[-1]) ? offset : -1;
1440
1441 /* foo[/]bar vs foo[] */
1442 return is_dir_sep(*subdir) ? offset + 1 : -1;
e6636747
JS
1443}
1444
1445int is_inside_dir(const char *dir)
1446{
b892913d
NTND
1447 char cwd[PATH_MAX];
1448 if (!dir)
1449 return 0;
1450 if (!getcwd(cwd, sizeof(cwd)))
1451 die_errno("can't find the current directory");
1452 return dir_inside_of(cwd, dir) >= 0;
e6636747 1453}
7155b727 1454
55892d23
AP
1455int is_empty_dir(const char *path)
1456{
1457 DIR *dir = opendir(path);
1458 struct dirent *e;
1459 int ret = 1;
1460
1461 if (!dir)
1462 return 0;
1463
1464 while ((e = readdir(dir)) != NULL)
1465 if (!is_dot_or_dotdot(e->d_name)) {
1466 ret = 0;
1467 break;
1468 }
1469
1470 closedir(dir);
1471 return ret;
1472}
1473
ae2f203e 1474static int remove_dir_recurse(struct strbuf *path, int flag, int *kept_up)
7155b727 1475{
a0f4afbe 1476 DIR *dir;
7155b727 1477 struct dirent *e;
ae2f203e 1478 int ret = 0, original_len = path->len, len, kept_down = 0;
a0f4afbe 1479 int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
c844a803 1480 int keep_toplevel = (flag & REMOVE_DIR_KEEP_TOPLEVEL);
a0f4afbe 1481 unsigned char submodule_head[20];
7155b727 1482
a0f4afbe 1483 if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
ae2f203e 1484 !resolve_gitlink_ref(path->buf, "HEAD", submodule_head)) {
a0f4afbe 1485 /* Do not descend and nuke a nested git work tree. */
ae2f203e
JH
1486 if (kept_up)
1487 *kept_up = 1;
a0f4afbe 1488 return 0;
ae2f203e 1489 }
a0f4afbe 1490
ae2f203e 1491 flag &= ~REMOVE_DIR_KEEP_TOPLEVEL;
a0f4afbe 1492 dir = opendir(path->buf);
c844a803 1493 if (!dir) {
ae2f203e 1494 /* an empty dir could be removed even if it is unreadble */
c844a803
JH
1495 if (!keep_toplevel)
1496 return rmdir(path->buf);
1497 else
1498 return -1;
1499 }
7155b727
JS
1500 if (path->buf[original_len - 1] != '/')
1501 strbuf_addch(path, '/');
1502
1503 len = path->len;
1504 while ((e = readdir(dir)) != NULL) {
1505 struct stat st;
8ca12c0d
AP
1506 if (is_dot_or_dotdot(e->d_name))
1507 continue;
7155b727
JS
1508
1509 strbuf_setlen(path, len);
1510 strbuf_addstr(path, e->d_name);
1511 if (lstat(path->buf, &st))
1512 ; /* fall thru */
1513 else if (S_ISDIR(st.st_mode)) {
ae2f203e 1514 if (!remove_dir_recurse(path, flag, &kept_down))
7155b727
JS
1515 continue; /* happy */
1516 } else if (!only_empty && !unlink(path->buf))
1517 continue; /* happy, too */
1518
1519 /* path too long, stat fails, or non-directory still exists */
1520 ret = -1;
1521 break;
1522 }
1523 closedir(dir);
1524
1525 strbuf_setlen(path, original_len);
ae2f203e 1526 if (!ret && !keep_toplevel && !kept_down)
7155b727 1527 ret = rmdir(path->buf);
ae2f203e
JH
1528 else if (kept_up)
1529 /*
1530 * report the uplevel that it is not an error that we
1531 * did not rmdir() our directory.
1532 */
1533 *kept_up = !ret;
7155b727
JS
1534 return ret;
1535}
039bc64e 1536
ae2f203e
JH
1537int remove_dir_recursively(struct strbuf *path, int flag)
1538{
1539 return remove_dir_recurse(path, flag, NULL);
1540}
1541
039bc64e
JH
1542void setup_standard_excludes(struct dir_struct *dir)
1543{
1544 const char *path;
dc79687e 1545 char *xdg_path;
039bc64e
JH
1546
1547 dir->exclude_per_dir = ".gitignore";
1548 path = git_path("info/exclude");
dc79687e
HKNN
1549 if (!excludes_file) {
1550 home_config_paths(NULL, &xdg_path, "ignore");
1551 excludes_file = xdg_path;
1552 }
4698c8fe 1553 if (!access_or_warn(path, R_OK, 0))
039bc64e 1554 add_excludes_from_file(dir, path);
4698c8fe 1555 if (excludes_file && !access_or_warn(excludes_file, R_OK, 0))
039bc64e
JH
1556 add_excludes_from_file(dir, excludes_file);
1557}
4a92d1bf
AR
1558
1559int remove_path(const char *name)
1560{
1561 char *slash;
1562
9a6728d4 1563 if (unlink(name) && errno != ENOENT && errno != ENOTDIR)
4a92d1bf
AR
1564 return -1;
1565
1566 slash = strrchr(name, '/');
1567 if (slash) {
1568 char *dirs = xstrdup(name);
1569 slash = dirs + (slash - name);
1570 do {
1571 *slash = '\0';
3fc0d131 1572 } while (rmdir(dirs) == 0 && (slash = strrchr(dirs, '/')));
4a92d1bf
AR
1573 free(dirs);
1574 }
1575 return 0;
1576}
1577
86e4ca69
NTND
1578static int pathspec_item_cmp(const void *a_, const void *b_)
1579{
1580 struct pathspec_item *a, *b;
1581
1582 a = (struct pathspec_item *)a_;
1583 b = (struct pathspec_item *)b_;
1584 return strcmp(a->match, b->match);
1585}
1586
0602f3e9
NTND
1587int init_pathspec(struct pathspec *pathspec, const char **paths)
1588{
1589 const char **p = paths;
1590 int i;
1591
1592 memset(pathspec, 0, sizeof(*pathspec));
1593 if (!p)
1594 return 0;
1595 while (*p)
1596 p++;
1597 pathspec->raw = paths;
1598 pathspec->nr = p - paths;
1599 if (!pathspec->nr)
1600 return 0;
1601
1602 pathspec->items = xmalloc(sizeof(struct pathspec_item)*pathspec->nr);
1603 for (i = 0; i < pathspec->nr; i++) {
1604 struct pathspec_item *item = pathspec->items+i;
1605 const char *path = paths[i];
1606
1607 item->match = path;
d2ce1331 1608 item->original = path;
0602f3e9 1609 item->len = strlen(path);
8c6abbcd 1610 item->flags = 0;
971e829c
JH
1611 if (limit_pathspec_to_literal()) {
1612 item->nowildcard_len = item->len;
1613 } else {
1614 item->nowildcard_len = simple_length(path);
1615 if (item->nowildcard_len < item->len) {
1616 pathspec->has_wildcard = 1;
1617 if (path[item->nowildcard_len] == '*' &&
1618 no_wildcard(path + item->nowildcard_len + 1))
1619 item->flags |= PATHSPEC_ONESTAR;
1620 }
8c6abbcd 1621 }
0602f3e9 1622 }
86e4ca69
NTND
1623
1624 qsort(pathspec->items, pathspec->nr,
1625 sizeof(struct pathspec_item), pathspec_item_cmp);
1626
0602f3e9
NTND
1627 return 0;
1628}
1629
1630void free_pathspec(struct pathspec *pathspec)
1631{
1632 free(pathspec->items);
1633 pathspec->items = NULL;
1634}
823ab40f
JK
1635
1636int limit_pathspec_to_literal(void)
1637{
1638 static int flag = -1;
1639 if (flag < 0)
1640 flag = git_env_bool(GIT_LITERAL_PATHSPECS_ENVIRONMENT, 0);
1641 return flag;
1642}
a39b15b4 1643
270be816
AS
1644/*
1645 * Frees memory within dir which was allocated for exclude lists and
1646 * the exclude_stack. Does not free dir itself.
1647 */
1648void clear_directory(struct dir_struct *dir)
1649{
1650 int i, j;
1651 struct exclude_list_group *group;
1652 struct exclude_list *el;
1653 struct exclude_stack *stk;
1654
1655 for (i = EXC_CMDL; i <= EXC_FILE; i++) {
1656 group = &dir->exclude_list_group[i];
1657 for (j = 0; j < group->nr; j++) {
1658 el = &group->el[j];
1659 if (i == EXC_DIRS)
1660 free((char *)el->src);
1661 clear_exclude_list(el);
1662 }
1663 free(group->el);
1664 }
1665
1666 stk = dir->exclude_stack;
1667 while (stk) {
1668 struct exclude_stack *prev = stk->prev;
1669 free(stk);
1670 stk = prev;
1671 }
1672}