dir.c: unify is_excluded and is_path_excluded APIs
[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"
453ec4bd 14
9fc42d60
LT
15struct path_simplify {
16 int len;
17 const char *path;
18};
19
dba2e203 20static int read_directory_recursive(struct dir_struct *dir, const char *path, int len,
09595258 21 int check_only, const struct path_simplify *simplify);
caa6b782 22static int get_dtype(struct dirent *de, const char *path, int len);
09595258 23
8cf2a84e
JJ
24/* helper string functions with support for the ignore_case flag */
25int strcmp_icase(const char *a, const char *b)
26{
27 return ignore_case ? strcasecmp(a, b) : strcmp(a, b);
28}
29
30int strncmp_icase(const char *a, const char *b, size_t count)
31{
32 return ignore_case ? strncasecmp(a, b, count) : strncmp(a, b, count);
33}
34
35int fnmatch_icase(const char *pattern, const char *string, int flags)
36{
37 return fnmatch(pattern, string, flags | (ignore_case ? FNM_CASEFOLD : 0));
38}
39
5d74762d
NTND
40inline int git_fnmatch(const char *pattern, const char *string,
41 int flags, int prefix)
42{
43 int fnm_flags = 0;
44 if (flags & GFNM_PATHNAME)
45 fnm_flags |= FNM_PATHNAME;
46 if (prefix > 0) {
47 if (strncmp(pattern, string, prefix))
48 return FNM_NOMATCH;
49 pattern += prefix;
50 string += prefix;
51 }
8c6abbcd
NTND
52 if (flags & GFNM_ONESTAR) {
53 int pattern_len = strlen(++pattern);
54 int string_len = strlen(string);
55 return string_len < pattern_len ||
56 strcmp(pattern,
57 string + string_len - pattern_len);
58 }
5d74762d
NTND
59 return fnmatch(pattern, string, fnm_flags);
60}
61
0b6e56df
JH
62static int fnmatch_icase_mem(const char *pattern, int patternlen,
63 const char *string, int stringlen,
64 int flags)
65{
66 int match_status;
67 struct strbuf pat_buf = STRBUF_INIT;
68 struct strbuf str_buf = STRBUF_INIT;
69 const char *use_pat = pattern;
70 const char *use_str = string;
71
72 if (pattern[patternlen]) {
73 strbuf_add(&pat_buf, pattern, patternlen);
74 use_pat = pat_buf.buf;
75 }
76 if (string[stringlen]) {
77 strbuf_add(&str_buf, string, stringlen);
78 use_str = str_buf.buf;
79 }
80
f30366b2
JH
81 if (ignore_case)
82 flags |= WM_CASEFOLD;
83 match_status = wildmatch(use_pat, use_str, flags, NULL);
0b6e56df
JH
84
85 strbuf_release(&pat_buf);
86 strbuf_release(&str_buf);
87
88 return match_status;
89}
90
f950eb95 91static size_t common_prefix_len(const char **pathspec)
3c6a370b 92{
4a085b16
JH
93 const char *n, *first;
94 size_t max = 0;
823ab40f 95 int literal = limit_pathspec_to_literal();
3c6a370b
LT
96
97 if (!pathspec)
4a085b16
JH
98 return max;
99
100 first = *pathspec;
101 while ((n = *pathspec++)) {
102 size_t i, len = 0;
103 for (i = 0; first == n || i < max; i++) {
104 char c = n[i];
823ab40f 105 if (!c || c != first[i] || (!literal && is_glob_special(c)))
4a085b16
JH
106 break;
107 if (c == '/')
108 len = i + 1;
109 }
110 if (first == n || len < max) {
111 max = len;
112 if (!max)
113 break;
114 }
3c6a370b 115 }
4a085b16 116 return max;
3c6a370b
LT
117}
118
f950eb95
CB
119/*
120 * Returns a copy of the longest leading path common among all
121 * pathspecs.
122 */
123char *common_prefix(const char **pathspec)
124{
125 unsigned long len = common_prefix_len(pathspec);
126
127 return len ? xmemdupz(*pathspec, len) : NULL;
128}
129
1d8842d9
LT
130int fill_directory(struct dir_struct *dir, const char **pathspec)
131{
4a085b16 132 size_t len;
1d8842d9
LT
133
134 /*
135 * Calculate common prefix for the pathspec, and
136 * use that to optimize the directory walk
137 */
4a085b16 138 len = common_prefix_len(pathspec);
1d8842d9
LT
139
140 /* Read the directory and prune it */
2b189435 141 read_directory(dir, pathspec ? *pathspec : "", len, pathspec);
dba2e203 142 return len;
1d8842d9
LT
143}
144
bc96cc87
NTND
145int within_depth(const char *name, int namelen,
146 int depth, int max_depth)
147{
148 const char *cp = name, *cpe = name + namelen;
149
150 while (cp < cpe) {
151 if (*cp++ != '/')
152 continue;
153 depth++;
154 if (depth > max_depth)
155 return 0;
156 }
157 return 1;
158}
159
e813d50e 160/*
2c5b0115 161 * Does 'match' match the given name?
e813d50e
JH
162 * A match is found if
163 *
164 * (1) the 'match' string is leading directory of 'name', or
165 * (2) the 'match' string is a wildcard and matches 'name', or
166 * (3) the 'match' string is exactly the same as 'name'.
167 *
168 * and the return value tells which case it was.
169 *
170 * It returns 0 when there is no match.
171 */
3c6a370b
LT
172static int match_one(const char *match, const char *name, int namelen)
173{
174 int matchlen;
823ab40f 175 int literal = limit_pathspec_to_literal();
3c6a370b
LT
176
177 /* If the match was just the prefix, we matched */
88ea8112 178 if (!*match)
e813d50e 179 return MATCHED_RECURSIVELY;
3c6a370b 180
21444f18
JJ
181 if (ignore_case) {
182 for (;;) {
183 unsigned char c1 = tolower(*match);
184 unsigned char c2 = tolower(*name);
823ab40f 185 if (c1 == '\0' || (!literal && is_glob_special(c1)))
21444f18
JJ
186 break;
187 if (c1 != c2)
188 return 0;
189 match++;
190 name++;
191 namelen--;
192 }
193 } else {
194 for (;;) {
195 unsigned char c1 = *match;
196 unsigned char c2 = *name;
823ab40f 197 if (c1 == '\0' || (!literal && is_glob_special(c1)))
21444f18
JJ
198 break;
199 if (c1 != c2)
200 return 0;
201 match++;
202 name++;
203 namelen--;
204 }
88ea8112
LT
205 }
206
3c6a370b
LT
207 /*
208 * If we don't match the matchstring exactly,
209 * we need to match by fnmatch
210 */
88ea8112 211 matchlen = strlen(match);
823ab40f
JK
212 if (strncmp_icase(match, name, matchlen)) {
213 if (literal)
214 return 0;
21444f18 215 return !fnmatch_icase(match, name, 0) ? MATCHED_FNMATCH : 0;
823ab40f 216 }
3c6a370b 217
f2d0df71 218 if (namelen == matchlen)
e813d50e
JH
219 return MATCHED_EXACTLY;
220 if (match[matchlen-1] == '/' || name[matchlen] == '/')
221 return MATCHED_RECURSIVELY;
222 return 0;
3c6a370b
LT
223}
224
e813d50e 225/*
52ed1894
AS
226 * Given a name and a list of pathspecs, returns the nature of the
227 * closest (i.e. most specific) match of the name to any of the
228 * pathspecs.
229 *
230 * The caller typically calls this multiple times with the same
231 * pathspec and seen[] array but with different name/namelen
232 * (e.g. entries from the index) and is interested in seeing if and
233 * how each pathspec matches all the names it calls this function
234 * with. A mark is left in the seen[] array for each pathspec element
235 * indicating the closest type of match that element achieved, so if
236 * seen[n] remains zero after multiple invocations, that means the nth
237 * pathspec did not match any names, which could indicate that the
238 * user mistyped the nth pathspec.
e813d50e 239 */
0b50922a
CB
240int match_pathspec(const char **pathspec, const char *name, int namelen,
241 int prefix, char *seen)
3c6a370b 242{
0b50922a
CB
243 int i, retval = 0;
244
245 if (!pathspec)
246 return 1;
3c6a370b
LT
247
248 name += prefix;
249 namelen -= prefix;
250
0b50922a 251 for (i = 0; pathspec[i] != NULL; i++) {
e813d50e 252 int how;
0b50922a
CB
253 const char *match = pathspec[i] + prefix;
254 if (seen && seen[i] == MATCHED_EXACTLY)
3c6a370b 255 continue;
e813d50e
JH
256 how = match_one(match, name, namelen);
257 if (how) {
258 if (retval < how)
259 retval = how;
0b50922a
CB
260 if (seen && seen[i] < how)
261 seen[i] = how;
3c6a370b
LT
262 }
263 }
264 return retval;
265}
266
61cf2820
NTND
267/*
268 * Does 'match' match the given name?
269 * A match is found if
270 *
271 * (1) the 'match' string is leading directory of 'name', or
272 * (2) the 'match' string is a wildcard and matches 'name', or
273 * (3) the 'match' string is exactly the same as 'name'.
274 *
275 * and the return value tells which case it was.
276 *
277 * It returns 0 when there is no match.
278 */
279static int match_pathspec_item(const struct pathspec_item *item, int prefix,
280 const char *name, int namelen)
281{
282 /* name/namelen has prefix cut off by caller */
283 const char *match = item->match + prefix;
284 int matchlen = item->len - prefix;
285
286 /* If the match was just the prefix, we matched */
287 if (!*match)
288 return MATCHED_RECURSIVELY;
289
290 if (matchlen <= namelen && !strncmp(match, name, matchlen)) {
291 if (matchlen == namelen)
292 return MATCHED_EXACTLY;
293
294 if (match[matchlen-1] == '/' || name[matchlen] == '/')
295 return MATCHED_RECURSIVELY;
296 }
297
5d74762d 298 if (item->nowildcard_len < item->len &&
8c6abbcd
NTND
299 !git_fnmatch(match, name,
300 item->flags & PATHSPEC_ONESTAR ? GFNM_ONESTAR : 0,
301 item->nowildcard_len - prefix))
61cf2820
NTND
302 return MATCHED_FNMATCH;
303
304 return 0;
305}
306
307/*
52ed1894
AS
308 * Given a name and a list of pathspecs, returns the nature of the
309 * closest (i.e. most specific) match of the name to any of the
310 * pathspecs.
311 *
312 * The caller typically calls this multiple times with the same
313 * pathspec and seen[] array but with different name/namelen
314 * (e.g. entries from the index) and is interested in seeing if and
315 * how each pathspec matches all the names it calls this function
316 * with. A mark is left in the seen[] array for each pathspec element
317 * indicating the closest type of match that element achieved, so if
318 * seen[n] remains zero after multiple invocations, that means the nth
319 * pathspec did not match any names, which could indicate that the
320 * user mistyped the nth pathspec.
61cf2820
NTND
321 */
322int match_pathspec_depth(const struct pathspec *ps,
323 const char *name, int namelen,
324 int prefix, char *seen)
325{
326 int i, retval = 0;
327
328 if (!ps->nr) {
329 if (!ps->recursive || ps->max_depth == -1)
330 return MATCHED_RECURSIVELY;
331
332 if (within_depth(name, namelen, 0, ps->max_depth))
333 return MATCHED_EXACTLY;
334 else
335 return 0;
336 }
337
338 name += prefix;
339 namelen -= prefix;
340
341 for (i = ps->nr - 1; i >= 0; i--) {
342 int how;
343 if (seen && seen[i] == MATCHED_EXACTLY)
344 continue;
345 how = match_pathspec_item(ps->items+i, prefix, name, namelen);
346 if (ps->recursive && ps->max_depth != -1 &&
347 how && how != MATCHED_FNMATCH) {
348 int len = ps->items[i].len;
349 if (name[len] == '/')
350 len++;
351 if (within_depth(name+len, namelen-len, 0, ps->max_depth))
352 how = MATCHED_EXACTLY;
353 else
354 how = 0;
355 }
356 if (how) {
357 if (retval < how)
358 retval = how;
359 if (seen && seen[i] < how)
360 seen[i] = how;
361 }
362 }
363 return retval;
364}
365
fcd631ed
NTND
366/*
367 * Return the length of the "simple" part of a path match limiter.
368 */
369static int simple_length(const char *match)
370{
371 int len = -1;
372
373 for (;;) {
374 unsigned char c = *match++;
375 len++;
376 if (c == '\0' || is_glob_special(c))
377 return len;
378 }
379}
380
68492fc7
LK
381static int no_wildcard(const char *string)
382{
fcd631ed 383 return string[simple_length(string)] == '\0';
68492fc7
LK
384}
385
82dce998
NTND
386void parse_exclude_pattern(const char **pattern,
387 int *patternlen,
388 int *flags,
389 int *nowildcardlen)
84460eec
NTND
390{
391 const char *p = *pattern;
392 size_t i, len;
393
394 *flags = 0;
395 if (*p == '!') {
396 *flags |= EXC_FLAG_NEGATIVE;
397 p++;
398 }
399 len = strlen(p);
400 if (len && p[len - 1] == '/') {
401 len--;
402 *flags |= EXC_FLAG_MUSTBEDIR;
403 }
404 for (i = 0; i < len; i++) {
405 if (p[i] == '/')
406 break;
407 }
408 if (i == len)
409 *flags |= EXC_FLAG_NODIR;
410 *nowildcardlen = simple_length(p);
411 /*
412 * we should have excluded the trailing slash from 'p' too,
413 * but that's one more allocation. Instead just make sure
414 * nowildcardlen does not exceed real patternlen
415 */
416 if (*nowildcardlen > len)
417 *nowildcardlen = len;
418 if (*p == '*' && no_wildcard(p + 1))
419 *flags |= EXC_FLAG_ENDSWITH;
420 *pattern = p;
421 *patternlen = len;
422}
423
453ec4bd 424void add_exclude(const char *string, const char *base,
c04318e4 425 int baselen, struct exclude_list *el, int srcpos)
453ec4bd 426{
d6b8fc30 427 struct exclude *x;
84460eec
NTND
428 int patternlen;
429 int flags;
430 int nowildcardlen;
453ec4bd 431
84460eec
NTND
432 parse_exclude_pattern(&string, &patternlen, &flags, &nowildcardlen);
433 if (flags & EXC_FLAG_MUSTBEDIR) {
d6b8fc30 434 char *s;
84460eec 435 x = xmalloc(sizeof(*x) + patternlen + 1);
4b25d091 436 s = (char *)(x+1);
84460eec
NTND
437 memcpy(s, string, patternlen);
438 s[patternlen] = '\0';
d6b8fc30 439 x->pattern = s;
d6b8fc30
JH
440 } else {
441 x = xmalloc(sizeof(*x));
442 x->pattern = string;
443 }
84460eec
NTND
444 x->patternlen = patternlen;
445 x->nowildcardlen = nowildcardlen;
453ec4bd
LT
446 x->base = base;
447 x->baselen = baselen;
d6b8fc30 448 x->flags = flags;
c04318e4 449 x->srcpos = srcpos;
840fc334
AS
450 ALLOC_GROW(el->excludes, el->nr + 1, el->alloc);
451 el->excludes[el->nr++] = x;
c04318e4 452 x->el = el;
453ec4bd
LT
453}
454
c28b3d6e
NTND
455static void *read_skip_worktree_file_from_index(const char *path, size_t *size)
456{
457 int pos, len;
458 unsigned long sz;
459 enum object_type type;
460 void *data;
461 struct index_state *istate = &the_index;
462
463 len = strlen(path);
464 pos = index_name_pos(istate, path, len);
465 if (pos < 0)
466 return NULL;
467 if (!ce_skip_worktree(istate->cache[pos]))
468 return NULL;
469 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
470 if (!data || type != OBJ_BLOB) {
471 free(data);
472 return NULL;
473 }
474 *size = xsize_t(sz);
475 return data;
476}
477
f6198812
AS
478/*
479 * Frees memory within el which was allocated for exclude patterns and
480 * the file buffer. Does not free el itself.
481 */
482void clear_exclude_list(struct exclude_list *el)
0fd0e241
NTND
483{
484 int i;
485
486 for (i = 0; i < el->nr; i++)
487 free(el->excludes[i]);
488 free(el->excludes);
c082df24 489 free(el->filebuf);
0fd0e241
NTND
490
491 el->nr = 0;
492 el->excludes = NULL;
c082df24 493 el->filebuf = NULL;
0fd0e241
NTND
494}
495
cb097534
NTND
496int add_excludes_from_file_to_list(const char *fname,
497 const char *base,
498 int baselen,
840fc334 499 struct exclude_list *el,
cb097534 500 int check_index)
453ec4bd 501{
c470701a 502 struct stat st;
c04318e4 503 int fd, i, lineno = 1;
9d14017a 504 size_t size = 0;
453ec4bd
LT
505 char *buf, *entry;
506
507 fd = open(fname, O_RDONLY);
c28b3d6e 508 if (fd < 0 || fstat(fd, &st) < 0) {
69660731 509 if (errno != ENOENT)
55b38a48 510 warn_on_inaccessible(fname);
c28b3d6e
NTND
511 if (0 <= fd)
512 close(fd);
513 if (!check_index ||
514 (buf = read_skip_worktree_file_from_index(fname, &size)) == NULL)
515 return -1;
45d76f17
NTND
516 if (size == 0) {
517 free(buf);
518 return 0;
519 }
520 if (buf[size-1] != '\n') {
521 buf = xrealloc(buf, size+1);
522 buf[size++] = '\n';
523 }
453ec4bd 524 }
c28b3d6e
NTND
525 else {
526 size = xsize_t(st.st_size);
527 if (size == 0) {
528 close(fd);
529 return 0;
530 }
45d76f17 531 buf = xmalloc(size+1);
c28b3d6e 532 if (read_in_full(fd, buf, size) != size) {
45d76f17 533 free(buf);
c28b3d6e
NTND
534 close(fd);
535 return -1;
536 }
45d76f17 537 buf[size++] = '\n';
c28b3d6e 538 close(fd);
6ba78238 539 }
453ec4bd 540
c082df24 541 el->filebuf = buf;
453ec4bd 542 entry = buf;
45d76f17
NTND
543 for (i = 0; i < size; i++) {
544 if (buf[i] == '\n') {
453ec4bd
LT
545 if (entry != buf + i && entry[0] != '#') {
546 buf[i - (i && buf[i-1] == '\r')] = 0;
c04318e4 547 add_exclude(entry, base, baselen, el, lineno);
453ec4bd 548 }
c04318e4 549 lineno++;
453ec4bd
LT
550 entry = buf + i + 1;
551 }
552 }
553 return 0;
453ec4bd
LT
554}
555
c04318e4
AS
556struct exclude_list *add_exclude_list(struct dir_struct *dir,
557 int group_type, const char *src)
c082df24
AS
558{
559 struct exclude_list *el;
560 struct exclude_list_group *group;
561
562 group = &dir->exclude_list_group[group_type];
563 ALLOC_GROW(group->el, group->nr + 1, group->alloc);
564 el = &group->el[group->nr++];
565 memset(el, 0, sizeof(*el));
c04318e4 566 el->src = src;
c082df24
AS
567 return el;
568}
569
570/*
571 * Used to set up core.excludesfile and .git/info/exclude lists.
572 */
453ec4bd
LT
573void add_excludes_from_file(struct dir_struct *dir, const char *fname)
574{
c082df24 575 struct exclude_list *el;
c04318e4 576 el = add_exclude_list(dir, EXC_FILE, fname);
c082df24 577 if (add_excludes_from_file_to_list(fname, "", 0, el, 0) < 0)
453ec4bd
LT
578 die("cannot use %s as an exclude file", fname);
579}
580
82dce998
NTND
581int match_basename(const char *basename, int basenamelen,
582 const char *pattern, int prefix, int patternlen,
583 int flags)
593cb880
NTND
584{
585 if (prefix == patternlen) {
0b6e56df
JH
586 if (patternlen == basenamelen &&
587 !strncmp_icase(pattern, basename, basenamelen))
593cb880
NTND
588 return 1;
589 } else if (flags & EXC_FLAG_ENDSWITH) {
0b6e56df 590 /* "*literal" matching against "fooliteral" */
593cb880 591 if (patternlen - 1 <= basenamelen &&
0b6e56df
JH
592 !strncmp_icase(pattern + 1,
593 basename + basenamelen - (patternlen - 1),
594 patternlen - 1))
593cb880
NTND
595 return 1;
596 } else {
0b6e56df
JH
597 if (fnmatch_icase_mem(pattern, patternlen,
598 basename, basenamelen,
599 0) == 0)
593cb880
NTND
600 return 1;
601 }
602 return 0;
603}
604
82dce998
NTND
605int match_pathname(const char *pathname, int pathlen,
606 const char *base, int baselen,
607 const char *pattern, int prefix, int patternlen,
608 int flags)
b5592632
NTND
609{
610 const char *name;
611 int namelen;
612
613 /*
614 * match with FNM_PATHNAME; the pattern has base implicitly
615 * in front of it.
616 */
617 if (*pattern == '/') {
618 pattern++;
982ac873 619 patternlen--;
b5592632
NTND
620 prefix--;
621 }
622
623 /*
624 * baselen does not count the trailing slash. base[] may or
625 * may not end with a trailing slash though.
626 */
627 if (pathlen < baselen + 1 ||
628 (baselen && pathname[baselen] != '/') ||
629 strncmp_icase(pathname, base, baselen))
630 return 0;
631
632 namelen = baselen ? pathlen - baselen - 1 : pathlen;
633 name = pathname + pathlen - namelen;
634
635 if (prefix) {
636 /*
637 * if the non-wildcard part is longer than the
638 * remaining pathname, surely it cannot match.
639 */
640 if (prefix > namelen)
641 return 0;
642
643 if (strncmp_icase(pattern, name, prefix))
644 return 0;
645 pattern += prefix;
ab3aebc1 646 patternlen -= prefix;
b5592632
NTND
647 name += prefix;
648 namelen -= prefix;
ab3aebc1
JK
649
650 /*
651 * If the whole pattern did not have a wildcard,
652 * then our prefix match is all we need; we
653 * do not need to call fnmatch at all.
654 */
655 if (!patternlen && !namelen)
656 return 1;
b5592632
NTND
657 }
658
ab3aebc1
JK
659 return fnmatch_icase_mem(pattern, patternlen,
660 name, namelen,
f30366b2 661 WM_PATHNAME) == 0;
b5592632
NTND
662}
663
578cd7c3
AS
664/*
665 * Scan the given exclude list in reverse to see whether pathname
666 * should be ignored. The first match (i.e. the last on the list), if
667 * any, determines the fate. Returns the exclude_list element which
668 * matched, or NULL for undecided.
453ec4bd 669 */
578cd7c3
AS
670static struct exclude *last_exclude_matching_from_list(const char *pathname,
671 int pathlen,
672 const char *basename,
673 int *dtype,
674 struct exclude_list *el)
453ec4bd
LT
675{
676 int i;
677
35a94d44 678 if (!el->nr)
578cd7c3 679 return NULL; /* undefined */
d6b8fc30 680
35a94d44
NTND
681 for (i = el->nr - 1; 0 <= i; i--) {
682 struct exclude *x = el->excludes[i];
b5592632 683 const char *exclude = x->pattern;
b5592632 684 int prefix = x->nowildcardlen;
35a94d44
NTND
685
686 if (x->flags & EXC_FLAG_MUSTBEDIR) {
687 if (*dtype == DT_UNKNOWN)
688 *dtype = get_dtype(NULL, pathname, pathlen);
689 if (*dtype != DT_DIR)
690 continue;
691 }
d6b8fc30 692
35a94d44 693 if (x->flags & EXC_FLAG_NODIR) {
593cb880
NTND
694 if (match_basename(basename,
695 pathlen - (basename - pathname),
696 exclude, prefix, x->patternlen,
697 x->flags))
578cd7c3 698 return x;
35a94d44
NTND
699 continue;
700 }
701
b5592632
NTND
702 assert(x->baselen == 0 || x->base[x->baselen - 1] == '/');
703 if (match_pathname(pathname, pathlen,
704 x->base, x->baselen ? x->baselen - 1 : 0,
705 exclude, prefix, x->patternlen, x->flags))
578cd7c3 706 return x;
453ec4bd 707 }
578cd7c3
AS
708 return NULL; /* undecided */
709}
710
711/*
712 * Scan the list and let the last match determine the fate.
713 * Return 1 for exclude, 0 for include and -1 for undecided.
714 */
715int is_excluded_from_list(const char *pathname,
716 int pathlen, const char *basename, int *dtype,
717 struct exclude_list *el)
718{
719 struct exclude *exclude;
720 exclude = last_exclude_matching_from_list(pathname, pathlen, basename, dtype, el);
721 if (exclude)
722 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
453ec4bd
LT
723 return -1; /* undecided */
724}
725
46aa2f95
KB
726static struct exclude *last_exclude_matching_from_lists(struct dir_struct *dir,
727 const char *pathname, int pathlen, const char *basename,
728 int *dtype_p)
729{
730 int i, j;
731 struct exclude_list_group *group;
732 struct exclude *exclude;
733 for (i = EXC_CMDL; i <= EXC_FILE; i++) {
734 group = &dir->exclude_list_group[i];
735 for (j = group->nr - 1; j >= 0; j--) {
736 exclude = last_exclude_matching_from_list(
737 pathname, pathlen, basename, dtype_p,
738 &group->el[j]);
739 if (exclude)
740 return exclude;
741 }
742 }
743 return NULL;
744}
745
6cd5c582
KB
746/*
747 * Loads the per-directory exclude list for the substring of base
748 * which has a char length of baselen.
749 */
750static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
751{
752 struct exclude_list_group *group;
753 struct exclude_list *el;
754 struct exclude_stack *stk = NULL;
755 int current;
756
6cd5c582
KB
757 group = &dir->exclude_list_group[EXC_DIRS];
758
759 /* Pop the exclude lists from the EXCL_DIRS exclude_list_group
760 * which originate from directories not in the prefix of the
761 * path being checked. */
762 while ((stk = dir->exclude_stack) != NULL) {
763 if (stk->baselen <= baselen &&
764 !strncmp(dir->basebuf, base, stk->baselen))
765 break;
766 el = &group->el[dir->exclude_stack->exclude_ix];
767 dir->exclude_stack = stk->prev;
95c6f271 768 dir->exclude = NULL;
6cd5c582
KB
769 free((char *)el->src); /* see strdup() below */
770 clear_exclude_list(el);
771 free(stk);
772 group->nr--;
773 }
774
95c6f271
KB
775 /* Skip traversing into sub directories if the parent is excluded */
776 if (dir->exclude)
777 return;
778
6cd5c582
KB
779 /* Read from the parent directories and push them down. */
780 current = stk ? stk->baselen : -1;
781 while (current < baselen) {
782 struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
783 const char *cp;
784
785 if (current < 0) {
786 cp = base;
787 current = 0;
788 }
789 else {
790 cp = strchr(base + current + 1, '/');
791 if (!cp)
792 die("oops in prep_exclude");
793 cp++;
794 }
795 stk->prev = dir->exclude_stack;
796 stk->baselen = cp - base;
95c6f271
KB
797 stk->exclude_ix = group->nr;
798 el = add_exclude_list(dir, EXC_DIRS, NULL);
6cd5c582
KB
799 memcpy(dir->basebuf + current, base + current,
800 stk->baselen - current);
95c6f271
KB
801
802 /* Abort if the directory is excluded */
803 if (stk->baselen) {
804 int dt = DT_DIR;
805 dir->basebuf[stk->baselen - 1] = 0;
806 dir->exclude = last_exclude_matching_from_lists(dir,
807 dir->basebuf, stk->baselen - 1,
808 dir->basebuf + current, &dt);
809 dir->basebuf[stk->baselen - 1] = '/';
810 if (dir->exclude) {
811 dir->basebuf[stk->baselen] = 0;
812 dir->exclude_stack = stk;
813 return;
814 }
815 }
816
817 /* Try to read per-directory file unless path is too long */
818 if (dir->exclude_per_dir &&
819 stk->baselen + strlen(dir->exclude_per_dir) < PATH_MAX) {
820 strcpy(dir->basebuf + stk->baselen,
821 dir->exclude_per_dir);
822 /*
823 * dir->basebuf gets reused by the traversal, but we
824 * need fname to remain unchanged to ensure the src
825 * member of each struct exclude correctly
826 * back-references its source file. Other invocations
827 * of add_exclude_list provide stable strings, so we
828 * strdup() and free() here in the caller.
829 */
830 el->src = strdup(dir->basebuf);
831 add_excludes_from_file_to_list(dir->basebuf,
832 dir->basebuf, stk->baselen, el, 1);
833 }
6cd5c582
KB
834 dir->exclude_stack = stk;
835 current = stk->baselen;
836 }
837 dir->basebuf[baselen] = '\0';
838}
839
f4cd69a6
AS
840/*
841 * Loads the exclude lists for the directory containing pathname, then
842 * scans all exclude lists to determine whether pathname is excluded.
843 * Returns the exclude_list element which matched, or NULL for
844 * undecided.
845 */
846static struct exclude *last_exclude_matching(struct dir_struct *dir,
847 const char *pathname,
848 int *dtype_p)
453ec4bd
LT
849{
850 int pathlen = strlen(pathname);
68492fc7
LK
851 const char *basename = strrchr(pathname, '/');
852 basename = (basename) ? basename+1 : pathname;
453ec4bd 853
63d285c8 854 prep_exclude(dir, pathname, basename-pathname);
c082df24 855
95c6f271
KB
856 if (dir->exclude)
857 return dir->exclude;
858
46aa2f95
KB
859 return last_exclude_matching_from_lists(dir, pathname, pathlen,
860 basename, dtype_p);
f4cd69a6
AS
861}
862
863/*
864 * Loads the exclude lists for the directory containing pathname, then
865 * scans all exclude lists to determine whether pathname is excluded.
866 * Returns 1 if true, otherwise 0.
867 */
868static int is_excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
869{
870 struct exclude *exclude =
871 last_exclude_matching(dir, pathname, dtype_p);
872 if (exclude)
873 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
453ec4bd
LT
874 return 0;
875}
876
eb41775e
JH
877void path_exclude_check_init(struct path_exclude_check *check,
878 struct dir_struct *dir)
879{
880 check->dir = dir;
eb41775e
JH
881}
882
883void path_exclude_check_clear(struct path_exclude_check *check)
884{
eb41775e
JH
885}
886
93921b07 887/*
a35341a8
AS
888 * For each subdirectory in name, starting with the top-most, checks
889 * to see if that subdirectory is excluded, and if so, returns the
890 * corresponding exclude structure. Otherwise, checks whether name
891 * itself (which is presumably a file) is excluded.
93921b07
JH
892 *
893 * A path to a directory known to be excluded is left in check->path to
894 * optimize for repeated checks for files in the same excluded directory.
895 */
a35341a8
AS
896struct exclude *last_exclude_matching_path(struct path_exclude_check *check,
897 const char *name, int namelen,
898 int *dtype)
eb41775e 899{
a35341a8
AS
900 return last_exclude_matching(check->dir, name, dtype);
901}
902
903/*
904 * Is this name excluded? This is for a caller like show_files() that
905 * do not honor directory hierarchy and iterate through paths that are
906 * possibly in an ignored directory.
907 */
908int is_path_excluded(struct path_exclude_check *check,
909 const char *name, int namelen, int *dtype)
910{
911 struct exclude *exclude =
912 last_exclude_matching_path(check, name, namelen, dtype);
913 if (exclude)
914 return exclude->flags & EXC_FLAG_NEGATIVE ? 0 : 1;
915 return 0;
eb41775e
JH
916}
917
f3fa1838
JH
918static struct dir_entry *dir_entry_new(const char *pathname, int len)
919{
453ec4bd
LT
920 struct dir_entry *ent;
921
453ec4bd
LT
922 ent = xmalloc(sizeof(*ent) + len + 1);
923 ent->len = len;
924 memcpy(ent->name, pathname, len);
925 ent->name[len] = 0;
4d06f8ac 926 return ent;
453ec4bd
LT
927}
928
159b3212 929static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
6815e569 930{
5bd8e2d8 931 if (cache_name_exists(pathname, len, ignore_case))
6815e569
JK
932 return NULL;
933
25fd2f7a 934 ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
6815e569
JK
935 return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
936}
937
108da0db 938struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
2abd31b0 939{
6e4f981f 940 if (!cache_name_is_other(pathname, len))
2abd31b0
JK
941 return NULL;
942
25fd2f7a 943 ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
2abd31b0
JK
944 return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
945}
946
09595258
LT
947enum exist_status {
948 index_nonexistent = 0,
949 index_directory,
4b05548f 950 index_gitdir
09595258
LT
951};
952
5102c617
JJ
953/*
954 * Do not use the alphabetically stored index to look up
955 * the directory name; instead, use the case insensitive
956 * name hash.
957 */
958static enum exist_status directory_exists_in_index_icase(const char *dirname, int len)
959{
960 struct cache_entry *ce = index_name_exists(&the_index, dirname, len + 1, ignore_case);
961 unsigned char endchar;
962
963 if (!ce)
964 return index_nonexistent;
965 endchar = ce->name[len];
966
967 /*
968 * The cache_entry structure returned will contain this dirname
969 * and possibly additional path components.
970 */
971 if (endchar == '/')
972 return index_directory;
973
974 /*
975 * If there are no additional path components, then this cache_entry
976 * represents a submodule. Submodules, despite being directories,
977 * are stored in the cache without a closing slash.
978 */
979 if (!endchar && S_ISGITLINK(ce->ce_mode))
980 return index_gitdir;
981
982 /* This should never be hit, but it exists just in case. */
983 return index_nonexistent;
984}
985
09595258
LT
986/*
987 * The index sorts alphabetically by entry name, which
988 * means that a gitlink sorts as '\0' at the end, while
989 * a directory (which is defined not as an entry, but as
990 * the files it contains) will sort with the '/' at the
991 * end.
992 */
993static enum exist_status directory_exists_in_index(const char *dirname, int len)
453ec4bd 994{
5102c617
JJ
995 int pos;
996
997 if (ignore_case)
998 return directory_exists_in_index_icase(dirname, len);
999
1000 pos = cache_name_pos(dirname, len);
09595258
LT
1001 if (pos < 0)
1002 pos = -pos-1;
1003 while (pos < active_nr) {
1004 struct cache_entry *ce = active_cache[pos++];
1005 unsigned char endchar;
1006
1007 if (strncmp(ce->name, dirname, len))
1008 break;
1009 endchar = ce->name[len];
1010 if (endchar > '/')
1011 break;
1012 if (endchar == '/')
1013 return index_directory;
7a51ed66 1014 if (!endchar && S_ISGITLINK(ce->ce_mode))
09595258
LT
1015 return index_gitdir;
1016 }
1017 return index_nonexistent;
1018}
1019
1020/*
1021 * When we find a directory when traversing the filesystem, we
1022 * have three distinct cases:
1023 *
1024 * - ignore it
1025 * - see it as a directory
1026 * - recurse into it
1027 *
1028 * and which one we choose depends on a combination of existing
1029 * git index contents and the flags passed into the directory
1030 * traversal routine.
1031 *
1032 * Case 1: If we *already* have entries in the index under that
5bd8e2d8
KB
1033 * directory name, we always recurse into the directory to see
1034 * all the files.
09595258
LT
1035 *
1036 * Case 2: If we *already* have that directory name as a gitlink,
1037 * we always continue to see it as a gitlink, regardless of whether
1038 * there is an actual git directory there or not (it might not
1039 * be checked out as a subproject!)
1040 *
1041 * Case 3: if we didn't have it in the index previously, we
1042 * have a few sub-cases:
1043 *
1044 * (a) if "show_other_directories" is true, we show it as
1045 * just a directory, unless "hide_empty_directories" is
1046 * also true and the directory is empty, in which case
1047 * we just ignore it entirely.
721ac4ed
AP
1048 * if we are looking for ignored directories, look if it
1049 * contains only ignored files to decide if it must be shown as
1050 * ignored or not.
09595258 1051 * (b) if it looks like a git directory, and we don't have
302b9282 1052 * 'no_gitlinks' set we treat it as a gitlink, and show it
09595258
LT
1053 * as a directory.
1054 * (c) otherwise, we recurse into it.
1055 */
1056enum directory_treatment {
1057 show_directory,
1058 ignore_directory,
4b05548f 1059 recurse_into_directory
09595258
LT
1060};
1061
1062static enum directory_treatment treat_directory(struct dir_struct *dir,
721ac4ed 1063 const char *dirname, int len, int exclude,
09595258
LT
1064 const struct path_simplify *simplify)
1065{
1066 /* The "len-1" is to strip the final '/' */
1067 switch (directory_exists_in_index(dirname, len-1)) {
1068 case index_directory:
1069 return recurse_into_directory;
1070
1071 case index_gitdir:
7c4c97c0 1072 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
ab22aed3 1073 return ignore_directory;
09595258
LT
1074 return show_directory;
1075
1076 case index_nonexistent:
7c4c97c0 1077 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
09595258 1078 break;
7c4c97c0 1079 if (!(dir->flags & DIR_NO_GITLINKS)) {
09595258
LT
1080 unsigned char sha1[20];
1081 if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
1082 return show_directory;
1083 }
1084 return recurse_into_directory;
1085 }
1086
1087 /* This is the "show_other_directories" case */
721ac4ed 1088
560bb7a7
KB
1089 /* might be a sub directory in an excluded directory */
1090 if (!exclude) {
1091 struct path_exclude_check check;
1092 int dt = DT_DIR;
1093 path_exclude_check_init(&check, dir);
1094 exclude = is_path_excluded(&check, dirname, len, &dt);
1095 path_exclude_check_clear(&check);
1096 }
1097
721ac4ed
AP
1098 /*
1099 * We are looking for ignored files and our directory is not ignored,
c94ab010 1100 * check if it contains untracked files (i.e. is listed as untracked)
721ac4ed
AP
1101 */
1102 if ((dir->flags & DIR_SHOW_IGNORED) && !exclude) {
1103 int ignored;
1104 dir->flags &= ~DIR_SHOW_IGNORED;
721ac4ed 1105 ignored = read_directory_recursive(dir, dirname, len, 1, simplify);
721ac4ed
AP
1106 dir->flags |= DIR_SHOW_IGNORED;
1107
c94ab010
KB
1108 if (ignored)
1109 return ignore_directory;
721ac4ed 1110 }
184d2a8e
KB
1111
1112 if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
09595258 1113 return show_directory;
dba2e203 1114 if (!read_directory_recursive(dir, dirname, len, 1, simplify))
09595258
LT
1115 return ignore_directory;
1116 return show_directory;
453ec4bd
LT
1117}
1118
721ac4ed
AP
1119/*
1120 * Decide what to do when we find a file while traversing the
1121 * filesystem. Mostly two cases:
1122 *
1123 * 1. We are looking for ignored files
1124 * (a) File is ignored, include it
1125 * (b) File is in ignored path, include it
1126 * (c) File is not ignored, exclude it
1127 *
1128 * 2. Other scenarios, include the file if not excluded
1129 *
1130 * Return 1 for exclude, 0 for include.
1131 */
1132static int treat_file(struct dir_struct *dir, struct strbuf *path, int exclude, int *dtype)
1133{
1134 struct path_exclude_check check;
1135 int exclude_file = 0;
1136
0104c9e7
KB
1137 /* Always exclude indexed files */
1138 if (index_name_exists(&the_index, path->buf, path->len, ignore_case))
1139 return 1;
1140
721ac4ed
AP
1141 if (exclude)
1142 exclude_file = !(dir->flags & DIR_SHOW_IGNORED);
1143 else if (dir->flags & DIR_SHOW_IGNORED) {
721ac4ed
AP
1144 path_exclude_check_init(&check, dir);
1145
0a9a787f 1146 if (!is_path_excluded(&check, path->buf, path->len, dtype))
721ac4ed
AP
1147 exclude_file = 1;
1148
1149 path_exclude_check_clear(&check);
1150 }
1151
1152 return exclude_file;
1153}
1154
9fc42d60
LT
1155/*
1156 * This is an inexact early pruning of any recursive directory
1157 * reading - if the path cannot possibly be in the pathspec,
1158 * return true, and we'll skip it early.
1159 */
1160static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
1161{
1162 if (simplify) {
1163 for (;;) {
1164 const char *match = simplify->path;
1165 int len = simplify->len;
1166
1167 if (!match)
1168 break;
1169 if (len > pathlen)
1170 len = pathlen;
1171 if (!memcmp(path, match, len))
1172 return 0;
1173 simplify++;
1174 }
1175 return 1;
1176 }
1177 return 0;
1178}
1179
29209cbe
JK
1180/*
1181 * This function tells us whether an excluded path matches a
1182 * list of "interesting" pathspecs. That is, whether a path matched
1183 * by any of the pathspecs could possibly be ignored by excluding
1184 * the specified path. This can happen if:
1185 *
1186 * 1. the path is mentioned explicitly in the pathspec
1187 *
1188 * 2. the path is a directory prefix of some element in the
1189 * pathspec
1190 */
1191static int exclude_matches_pathspec(const char *path, int len,
1192 const struct path_simplify *simplify)
e96980ef
JK
1193{
1194 if (simplify) {
1195 for (; simplify->path; simplify++) {
1196 if (len == simplify->len
1197 && !memcmp(path, simplify->path, len))
1198 return 1;
29209cbe
JK
1199 if (len < simplify->len
1200 && simplify->path[len] == '/'
1201 && !memcmp(path, simplify->path, len))
1202 return 1;
e96980ef
JK
1203 }
1204 }
1205 return 0;
1206}
1207
443e061a
LT
1208static int get_index_dtype(const char *path, int len)
1209{
1210 int pos;
1211 struct cache_entry *ce;
1212
1213 ce = cache_name_exists(path, len, 0);
1214 if (ce) {
1215 if (!ce_uptodate(ce))
1216 return DT_UNKNOWN;
1217 if (S_ISGITLINK(ce->ce_mode))
1218 return DT_DIR;
1219 /*
1220 * Nobody actually cares about the
1221 * difference between DT_LNK and DT_REG
1222 */
1223 return DT_REG;
1224 }
1225
1226 /* Try to look it up as a directory */
1227 pos = cache_name_pos(path, len);
1228 if (pos >= 0)
1229 return DT_UNKNOWN;
1230 pos = -pos-1;
1231 while (pos < active_nr) {
1232 ce = active_cache[pos++];
1233 if (strncmp(ce->name, path, len))
1234 break;
1235 if (ce->name[len] > '/')
1236 break;
1237 if (ce->name[len] < '/')
1238 continue;
1239 if (!ce_uptodate(ce))
1240 break; /* continue? */
1241 return DT_DIR;
1242 }
1243 return DT_UNKNOWN;
1244}
1245
caa6b782 1246static int get_dtype(struct dirent *de, const char *path, int len)
07134421 1247{
6831a88a 1248 int dtype = de ? DTYPE(de) : DT_UNKNOWN;
07134421
LT
1249 struct stat st;
1250
1251 if (dtype != DT_UNKNOWN)
1252 return dtype;
443e061a
LT
1253 dtype = get_index_dtype(path, len);
1254 if (dtype != DT_UNKNOWN)
1255 return dtype;
1256 if (lstat(path, &st))
07134421
LT
1257 return dtype;
1258 if (S_ISREG(st.st_mode))
1259 return DT_REG;
1260 if (S_ISDIR(st.st_mode))
1261 return DT_DIR;
1262 if (S_ISLNK(st.st_mode))
1263 return DT_LNK;
1264 return dtype;
1265}
1266
53cc5356
JH
1267enum path_treatment {
1268 path_ignored,
1269 path_handled,
4b05548f 1270 path_recurse
53cc5356
JH
1271};
1272
16e2cfa9 1273static enum path_treatment treat_one_path(struct dir_struct *dir,
49dc2cc2 1274 struct strbuf *path,
16e2cfa9
JH
1275 const struct path_simplify *simplify,
1276 int dtype, struct dirent *de)
53cc5356 1277{
6d24e7a8 1278 int exclude = is_excluded(dir, path->buf, &dtype);
53cc5356 1279 if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
49dc2cc2
RS
1280 && exclude_matches_pathspec(path->buf, path->len, simplify))
1281 dir_add_ignored(dir, path->buf, path->len);
53cc5356
JH
1282
1283 /*
1284 * Excluded? If we don't explicitly want to show
1285 * ignored files, ignore it
1286 */
1287 if (exclude && !(dir->flags & DIR_SHOW_IGNORED))
1288 return path_ignored;
1289
1290 if (dtype == DT_UNKNOWN)
49dc2cc2 1291 dtype = get_dtype(de, path->buf, path->len);
53cc5356 1292
53cc5356
JH
1293 switch (dtype) {
1294 default:
1295 return path_ignored;
1296 case DT_DIR:
49dc2cc2 1297 strbuf_addch(path, '/');
721ac4ed 1298 switch (treat_directory(dir, path->buf, path->len, exclude, simplify)) {
53cc5356 1299 case show_directory:
53cc5356
JH
1300 break;
1301 case recurse_into_directory:
1302 return path_recurse;
1303 case ignore_directory:
1304 return path_ignored;
1305 }
1306 break;
1307 case DT_REG:
1308 case DT_LNK:
721ac4ed
AP
1309 switch (treat_file(dir, path, exclude, &dtype)) {
1310 case 1:
1311 return path_ignored;
1312 default:
1313 break;
1314 }
53cc5356
JH
1315 }
1316 return path_handled;
1317}
1318
16e2cfa9
JH
1319static enum path_treatment treat_path(struct dir_struct *dir,
1320 struct dirent *de,
49dc2cc2 1321 struct strbuf *path,
16e2cfa9 1322 int baselen,
49dc2cc2 1323 const struct path_simplify *simplify)
16e2cfa9
JH
1324{
1325 int dtype;
1326
1327 if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
1328 return path_ignored;
49dc2cc2
RS
1329 strbuf_setlen(path, baselen);
1330 strbuf_addstr(path, de->d_name);
1331 if (simplify_away(path->buf, path->len, simplify))
16e2cfa9
JH
1332 return path_ignored;
1333
1334 dtype = DTYPE(de);
49dc2cc2 1335 return treat_one_path(dir, path, simplify, dtype, de);
16e2cfa9
JH
1336}
1337
453ec4bd
LT
1338/*
1339 * Read a directory tree. We currently ignore anything but
1340 * directories, regular files and symlinks. That's because git
1341 * doesn't handle them at all yet. Maybe that will change some
1342 * day.
1343 *
1344 * Also, we ignore the name ".git" (even if it is not a directory).
1345 * That likely will not change.
1346 */
53cc5356
JH
1347static int read_directory_recursive(struct dir_struct *dir,
1348 const char *base, int baselen,
1349 int check_only,
1350 const struct path_simplify *simplify)
453ec4bd 1351{
1528d247 1352 DIR *fdir;
453ec4bd 1353 int contents = 0;
02cb6753 1354 struct dirent *de;
bef36921 1355 struct strbuf path = STRBUF_INIT;
453ec4bd 1356
bef36921 1357 strbuf_add(&path, base, baselen);
02cb6753 1358
1528d247
RS
1359 fdir = opendir(path.len ? path.buf : ".");
1360 if (!fdir)
1361 goto out;
1362
02cb6753 1363 while ((de = readdir(fdir)) != NULL) {
bef36921 1364 switch (treat_path(dir, de, &path, baselen, simplify)) {
02cb6753 1365 case path_recurse:
bef36921 1366 contents += read_directory_recursive(dir, path.buf,
289ff559 1367 path.len, check_only, simplify);
02cb6753
NTND
1368 continue;
1369 case path_ignored:
1370 continue;
1371 case path_handled:
1372 break;
453ec4bd 1373 }
02cb6753
NTND
1374 contents++;
1375 if (check_only)
1528d247
RS
1376 break;
1377 dir_add_name(dir, path.buf, path.len);
453ec4bd 1378 }
02cb6753 1379 closedir(fdir);
1528d247 1380 out:
bef36921 1381 strbuf_release(&path);
453ec4bd
LT
1382
1383 return contents;
1384}
1385
1386static int cmp_name(const void *p1, const void *p2)
1387{
1388 const struct dir_entry *e1 = *(const struct dir_entry **)p1;
1389 const struct dir_entry *e2 = *(const struct dir_entry **)p2;
1390
1391 return cache_name_compare(e1->name, e1->len,
1392 e2->name, e2->len);
1393}
1394
9fc42d60
LT
1395static struct path_simplify *create_simplify(const char **pathspec)
1396{
1397 int nr, alloc = 0;
1398 struct path_simplify *simplify = NULL;
1399
1400 if (!pathspec)
1401 return NULL;
1402
1403 for (nr = 0 ; ; nr++) {
1404 const char *match;
1405 if (nr >= alloc) {
1406 alloc = alloc_nr(alloc);
1407 simplify = xrealloc(simplify, alloc * sizeof(*simplify));
1408 }
1409 match = *pathspec++;
1410 if (!match)
1411 break;
1412 simplify[nr].path = match;
1413 simplify[nr].len = simple_length(match);
1414 }
1415 simplify[nr].path = NULL;
1416 simplify[nr].len = 0;
1417 return simplify;
1418}
1419
1420static void free_simplify(struct path_simplify *simplify)
1421{
8e0f7003 1422 free(simplify);
9fc42d60
LT
1423}
1424
48ffef96
JH
1425static int treat_leading_path(struct dir_struct *dir,
1426 const char *path, int len,
1427 const struct path_simplify *simplify)
1428{
49dc2cc2
RS
1429 struct strbuf sb = STRBUF_INIT;
1430 int baselen, rc = 0;
48ffef96 1431 const char *cp;
be8a84c5 1432 int old_flags = dir->flags;
48ffef96
JH
1433
1434 while (len && path[len - 1] == '/')
1435 len--;
1436 if (!len)
1437 return 1;
1438 baselen = 0;
be8a84c5 1439 dir->flags &= ~DIR_SHOW_OTHER_DIRECTORIES;
48ffef96
JH
1440 while (1) {
1441 cp = path + baselen + !!baselen;
1442 cp = memchr(cp, '/', path + len - cp);
1443 if (!cp)
1444 baselen = len;
1445 else
1446 baselen = cp - path;
49dc2cc2
RS
1447 strbuf_setlen(&sb, 0);
1448 strbuf_add(&sb, path, baselen);
1449 if (!is_directory(sb.buf))
1450 break;
1451 if (simplify_away(sb.buf, sb.len, simplify))
1452 break;
1453 if (treat_one_path(dir, &sb, simplify,
48ffef96 1454 DT_DIR, NULL) == path_ignored)
49dc2cc2
RS
1455 break; /* do not recurse into it */
1456 if (len <= baselen) {
1457 rc = 1;
1458 break; /* finished checking */
1459 }
48ffef96 1460 }
49dc2cc2 1461 strbuf_release(&sb);
be8a84c5 1462 dir->flags = old_flags;
49dc2cc2 1463 return rc;
48ffef96
JH
1464}
1465
dba2e203 1466int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
9fc42d60 1467{
725b0605 1468 struct path_simplify *simplify;
b4189aa8 1469
dba2e203 1470 if (has_symlink_leading_path(path, len))
725b0605
JH
1471 return dir->nr;
1472
1473 simplify = create_simplify(pathspec);
48ffef96
JH
1474 if (!len || treat_leading_path(dir, path, len, simplify))
1475 read_directory_recursive(dir, path, len, 0, simplify);
9fc42d60 1476 free_simplify(simplify);
453ec4bd 1477 qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
2abd31b0 1478 qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
453ec4bd
LT
1479 return dir->nr;
1480}
c91f0d92 1481
686a4a06 1482int file_exists(const char *f)
c91f0d92 1483{
686a4a06 1484 struct stat sb;
a50f9fc5 1485 return lstat(f, &sb) == 0;
c91f0d92 1486}
e6636747
JS
1487
1488/*
9b125da4
NTND
1489 * Given two normalized paths (a trailing slash is ok), if subdir is
1490 * outside dir, return -1. Otherwise return the offset in subdir that
1491 * can be used as relative path to dir.
e6636747 1492 */
9b125da4 1493int dir_inside_of(const char *subdir, const char *dir)
e6636747 1494{
9b125da4 1495 int offset = 0;
e6636747 1496
9b125da4 1497 assert(dir && subdir && *dir && *subdir);
e6636747 1498
9b125da4 1499 while (*dir && *subdir && *dir == *subdir) {
e6636747 1500 dir++;
9b125da4
NTND
1501 subdir++;
1502 offset++;
490544b1 1503 }
9b125da4
NTND
1504
1505 /* hel[p]/me vs hel[l]/yeah */
1506 if (*dir && *subdir)
1507 return -1;
1508
1509 if (!*subdir)
1510 return !*dir ? offset : -1; /* same dir */
1511
1512 /* foo/[b]ar vs foo/[] */
1513 if (is_dir_sep(dir[-1]))
1514 return is_dir_sep(subdir[-1]) ? offset : -1;
1515
1516 /* foo[/]bar vs foo[] */
1517 return is_dir_sep(*subdir) ? offset + 1 : -1;
e6636747
JS
1518}
1519
1520int is_inside_dir(const char *dir)
1521{
b892913d
NTND
1522 char cwd[PATH_MAX];
1523 if (!dir)
1524 return 0;
1525 if (!getcwd(cwd, sizeof(cwd)))
1526 die_errno("can't find the current directory");
1527 return dir_inside_of(cwd, dir) >= 0;
e6636747 1528}
7155b727 1529
55892d23
AP
1530int is_empty_dir(const char *path)
1531{
1532 DIR *dir = opendir(path);
1533 struct dirent *e;
1534 int ret = 1;
1535
1536 if (!dir)
1537 return 0;
1538
1539 while ((e = readdir(dir)) != NULL)
1540 if (!is_dot_or_dotdot(e->d_name)) {
1541 ret = 0;
1542 break;
1543 }
1544
1545 closedir(dir);
1546 return ret;
1547}
1548
ae2f203e 1549static int remove_dir_recurse(struct strbuf *path, int flag, int *kept_up)
7155b727 1550{
a0f4afbe 1551 DIR *dir;
7155b727 1552 struct dirent *e;
ae2f203e 1553 int ret = 0, original_len = path->len, len, kept_down = 0;
a0f4afbe 1554 int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
c844a803 1555 int keep_toplevel = (flag & REMOVE_DIR_KEEP_TOPLEVEL);
a0f4afbe 1556 unsigned char submodule_head[20];
7155b727 1557
a0f4afbe 1558 if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
ae2f203e 1559 !resolve_gitlink_ref(path->buf, "HEAD", submodule_head)) {
a0f4afbe 1560 /* Do not descend and nuke a nested git work tree. */
ae2f203e
JH
1561 if (kept_up)
1562 *kept_up = 1;
a0f4afbe 1563 return 0;
ae2f203e 1564 }
a0f4afbe 1565
ae2f203e 1566 flag &= ~REMOVE_DIR_KEEP_TOPLEVEL;
a0f4afbe 1567 dir = opendir(path->buf);
c844a803 1568 if (!dir) {
ae2f203e 1569 /* an empty dir could be removed even if it is unreadble */
c844a803
JH
1570 if (!keep_toplevel)
1571 return rmdir(path->buf);
1572 else
1573 return -1;
1574 }
7155b727
JS
1575 if (path->buf[original_len - 1] != '/')
1576 strbuf_addch(path, '/');
1577
1578 len = path->len;
1579 while ((e = readdir(dir)) != NULL) {
1580 struct stat st;
8ca12c0d
AP
1581 if (is_dot_or_dotdot(e->d_name))
1582 continue;
7155b727
JS
1583
1584 strbuf_setlen(path, len);
1585 strbuf_addstr(path, e->d_name);
1586 if (lstat(path->buf, &st))
1587 ; /* fall thru */
1588 else if (S_ISDIR(st.st_mode)) {
ae2f203e 1589 if (!remove_dir_recurse(path, flag, &kept_down))
7155b727
JS
1590 continue; /* happy */
1591 } else if (!only_empty && !unlink(path->buf))
1592 continue; /* happy, too */
1593
1594 /* path too long, stat fails, or non-directory still exists */
1595 ret = -1;
1596 break;
1597 }
1598 closedir(dir);
1599
1600 strbuf_setlen(path, original_len);
ae2f203e 1601 if (!ret && !keep_toplevel && !kept_down)
7155b727 1602 ret = rmdir(path->buf);
ae2f203e
JH
1603 else if (kept_up)
1604 /*
1605 * report the uplevel that it is not an error that we
1606 * did not rmdir() our directory.
1607 */
1608 *kept_up = !ret;
7155b727
JS
1609 return ret;
1610}
039bc64e 1611
ae2f203e
JH
1612int remove_dir_recursively(struct strbuf *path, int flag)
1613{
1614 return remove_dir_recurse(path, flag, NULL);
1615}
1616
039bc64e
JH
1617void setup_standard_excludes(struct dir_struct *dir)
1618{
1619 const char *path;
dc79687e 1620 char *xdg_path;
039bc64e
JH
1621
1622 dir->exclude_per_dir = ".gitignore";
1623 path = git_path("info/exclude");
dc79687e
HKNN
1624 if (!excludes_file) {
1625 home_config_paths(NULL, &xdg_path, "ignore");
1626 excludes_file = xdg_path;
1627 }
69660731 1628 if (!access_or_warn(path, R_OK))
039bc64e 1629 add_excludes_from_file(dir, path);
69660731 1630 if (excludes_file && !access_or_warn(excludes_file, R_OK))
039bc64e
JH
1631 add_excludes_from_file(dir, excludes_file);
1632}
4a92d1bf
AR
1633
1634int remove_path(const char *name)
1635{
1636 char *slash;
1637
9a6728d4 1638 if (unlink(name) && errno != ENOENT && errno != ENOTDIR)
4a92d1bf
AR
1639 return -1;
1640
1641 slash = strrchr(name, '/');
1642 if (slash) {
1643 char *dirs = xstrdup(name);
1644 slash = dirs + (slash - name);
1645 do {
1646 *slash = '\0';
3fc0d131 1647 } while (rmdir(dirs) == 0 && (slash = strrchr(dirs, '/')));
4a92d1bf
AR
1648 free(dirs);
1649 }
1650 return 0;
1651}
1652
86e4ca69
NTND
1653static int pathspec_item_cmp(const void *a_, const void *b_)
1654{
1655 struct pathspec_item *a, *b;
1656
1657 a = (struct pathspec_item *)a_;
1658 b = (struct pathspec_item *)b_;
1659 return strcmp(a->match, b->match);
1660}
1661
0602f3e9
NTND
1662int init_pathspec(struct pathspec *pathspec, const char **paths)
1663{
1664 const char **p = paths;
1665 int i;
1666
1667 memset(pathspec, 0, sizeof(*pathspec));
1668 if (!p)
1669 return 0;
1670 while (*p)
1671 p++;
1672 pathspec->raw = paths;
1673 pathspec->nr = p - paths;
1674 if (!pathspec->nr)
1675 return 0;
1676
1677 pathspec->items = xmalloc(sizeof(struct pathspec_item)*pathspec->nr);
1678 for (i = 0; i < pathspec->nr; i++) {
1679 struct pathspec_item *item = pathspec->items+i;
1680 const char *path = paths[i];
1681
1682 item->match = path;
1683 item->len = strlen(path);
8c6abbcd 1684 item->flags = 0;
971e829c
JH
1685 if (limit_pathspec_to_literal()) {
1686 item->nowildcard_len = item->len;
1687 } else {
1688 item->nowildcard_len = simple_length(path);
1689 if (item->nowildcard_len < item->len) {
1690 pathspec->has_wildcard = 1;
1691 if (path[item->nowildcard_len] == '*' &&
1692 no_wildcard(path + item->nowildcard_len + 1))
1693 item->flags |= PATHSPEC_ONESTAR;
1694 }
8c6abbcd 1695 }
0602f3e9 1696 }
86e4ca69
NTND
1697
1698 qsort(pathspec->items, pathspec->nr,
1699 sizeof(struct pathspec_item), pathspec_item_cmp);
1700
0602f3e9
NTND
1701 return 0;
1702}
1703
1704void free_pathspec(struct pathspec *pathspec)
1705{
1706 free(pathspec->items);
1707 pathspec->items = NULL;
1708}
823ab40f
JK
1709
1710int limit_pathspec_to_literal(void)
1711{
1712 static int flag = -1;
1713 if (flag < 0)
1714 flag = git_env_bool(GIT_LITERAL_PATHSPECS_ENVIRONMENT, 0);
1715 return flag;
1716}
a39b15b4 1717
270be816
AS
1718/*
1719 * Frees memory within dir which was allocated for exclude lists and
1720 * the exclude_stack. Does not free dir itself.
1721 */
1722void clear_directory(struct dir_struct *dir)
1723{
1724 int i, j;
1725 struct exclude_list_group *group;
1726 struct exclude_list *el;
1727 struct exclude_stack *stk;
1728
1729 for (i = EXC_CMDL; i <= EXC_FILE; i++) {
1730 group = &dir->exclude_list_group[i];
1731 for (j = 0; j < group->nr; j++) {
1732 el = &group->el[j];
1733 if (i == EXC_DIRS)
1734 free((char *)el->src);
1735 clear_exclude_list(el);
1736 }
1737 free(group->el);
1738 }
1739
1740 stk = dir->exclude_stack;
1741 while (stk) {
1742 struct exclude_stack *prev = stk->prev;
1743 free(stk);
1744 stk = prev;
1745 }
1746}