exclude: split pathname matching code into a separate function
[git/git.git] / dir.c
CommitLineData
453ec4bd
LT
1/*
2 * This handles recursive filename detection with exclude
3 * files, index knowledge etc..
4 *
5 * Copyright (C) Linus Torvalds, 2005-2006
6 * Junio Hamano, 2005-2006
7 */
453ec4bd
LT
8#include "cache.h"
9#include "dir.h"
09595258 10#include "refs.h"
453ec4bd 11
9fc42d60
LT
12struct path_simplify {
13 int len;
14 const char *path;
15};
16
dba2e203 17static int read_directory_recursive(struct dir_struct *dir, const char *path, int len,
09595258 18 int check_only, const struct path_simplify *simplify);
caa6b782 19static int get_dtype(struct dirent *de, const char *path, int len);
09595258 20
8cf2a84e
JJ
21/* helper string functions with support for the ignore_case flag */
22int strcmp_icase(const char *a, const char *b)
23{
24 return ignore_case ? strcasecmp(a, b) : strcmp(a, b);
25}
26
27int strncmp_icase(const char *a, const char *b, size_t count)
28{
29 return ignore_case ? strncasecmp(a, b, count) : strncmp(a, b, count);
30}
31
32int fnmatch_icase(const char *pattern, const char *string, int flags)
33{
34 return fnmatch(pattern, string, flags | (ignore_case ? FNM_CASEFOLD : 0));
35}
36
f950eb95 37static size_t common_prefix_len(const char **pathspec)
3c6a370b 38{
4a085b16
JH
39 const char *n, *first;
40 size_t max = 0;
3c6a370b
LT
41
42 if (!pathspec)
4a085b16
JH
43 return max;
44
45 first = *pathspec;
46 while ((n = *pathspec++)) {
47 size_t i, len = 0;
48 for (i = 0; first == n || i < max; i++) {
49 char c = n[i];
50 if (!c || c != first[i] || is_glob_special(c))
51 break;
52 if (c == '/')
53 len = i + 1;
54 }
55 if (first == n || len < max) {
56 max = len;
57 if (!max)
58 break;
59 }
3c6a370b 60 }
4a085b16 61 return max;
3c6a370b
LT
62}
63
f950eb95
CB
64/*
65 * Returns a copy of the longest leading path common among all
66 * pathspecs.
67 */
68char *common_prefix(const char **pathspec)
69{
70 unsigned long len = common_prefix_len(pathspec);
71
72 return len ? xmemdupz(*pathspec, len) : NULL;
73}
74
1d8842d9
LT
75int fill_directory(struct dir_struct *dir, const char **pathspec)
76{
4a085b16 77 size_t len;
1d8842d9
LT
78
79 /*
80 * Calculate common prefix for the pathspec, and
81 * use that to optimize the directory walk
82 */
4a085b16 83 len = common_prefix_len(pathspec);
1d8842d9
LT
84
85 /* Read the directory and prune it */
2b189435 86 read_directory(dir, pathspec ? *pathspec : "", len, pathspec);
dba2e203 87 return len;
1d8842d9
LT
88}
89
bc96cc87
NTND
90int within_depth(const char *name, int namelen,
91 int depth, int max_depth)
92{
93 const char *cp = name, *cpe = name + namelen;
94
95 while (cp < cpe) {
96 if (*cp++ != '/')
97 continue;
98 depth++;
99 if (depth > max_depth)
100 return 0;
101 }
102 return 1;
103}
104
e813d50e 105/*
2c5b0115 106 * Does 'match' match the given name?
e813d50e
JH
107 * A match is found if
108 *
109 * (1) the 'match' string is leading directory of 'name', or
110 * (2) the 'match' string is a wildcard and matches 'name', or
111 * (3) the 'match' string is exactly the same as 'name'.
112 *
113 * and the return value tells which case it was.
114 *
115 * It returns 0 when there is no match.
116 */
3c6a370b
LT
117static int match_one(const char *match, const char *name, int namelen)
118{
119 int matchlen;
120
121 /* If the match was just the prefix, we matched */
88ea8112 122 if (!*match)
e813d50e 123 return MATCHED_RECURSIVELY;
3c6a370b 124
21444f18
JJ
125 if (ignore_case) {
126 for (;;) {
127 unsigned char c1 = tolower(*match);
128 unsigned char c2 = tolower(*name);
129 if (c1 == '\0' || is_glob_special(c1))
130 break;
131 if (c1 != c2)
132 return 0;
133 match++;
134 name++;
135 namelen--;
136 }
137 } else {
138 for (;;) {
139 unsigned char c1 = *match;
140 unsigned char c2 = *name;
141 if (c1 == '\0' || is_glob_special(c1))
142 break;
143 if (c1 != c2)
144 return 0;
145 match++;
146 name++;
147 namelen--;
148 }
88ea8112
LT
149 }
150
151
3c6a370b
LT
152 /*
153 * If we don't match the matchstring exactly,
154 * we need to match by fnmatch
155 */
88ea8112 156 matchlen = strlen(match);
21444f18
JJ
157 if (strncmp_icase(match, name, matchlen))
158 return !fnmatch_icase(match, name, 0) ? MATCHED_FNMATCH : 0;
3c6a370b 159
f2d0df71 160 if (namelen == matchlen)
e813d50e
JH
161 return MATCHED_EXACTLY;
162 if (match[matchlen-1] == '/' || name[matchlen] == '/')
163 return MATCHED_RECURSIVELY;
164 return 0;
3c6a370b
LT
165}
166
e813d50e
JH
167/*
168 * Given a name and a list of pathspecs, see if the name matches
169 * any of the pathspecs. The caller is also interested in seeing
170 * all pathspec matches some names it calls this function with
171 * (otherwise the user could have mistyped the unmatched pathspec),
172 * and a mark is left in seen[] array for pathspec element that
173 * actually matched anything.
174 */
0b50922a
CB
175int match_pathspec(const char **pathspec, const char *name, int namelen,
176 int prefix, char *seen)
3c6a370b 177{
0b50922a
CB
178 int i, retval = 0;
179
180 if (!pathspec)
181 return 1;
3c6a370b
LT
182
183 name += prefix;
184 namelen -= prefix;
185
0b50922a 186 for (i = 0; pathspec[i] != NULL; i++) {
e813d50e 187 int how;
0b50922a
CB
188 const char *match = pathspec[i] + prefix;
189 if (seen && seen[i] == MATCHED_EXACTLY)
3c6a370b 190 continue;
e813d50e
JH
191 how = match_one(match, name, namelen);
192 if (how) {
193 if (retval < how)
194 retval = how;
0b50922a
CB
195 if (seen && seen[i] < how)
196 seen[i] = how;
3c6a370b
LT
197 }
198 }
199 return retval;
200}
201
61cf2820
NTND
202/*
203 * Does 'match' match the given name?
204 * A match is found if
205 *
206 * (1) the 'match' string is leading directory of 'name', or
207 * (2) the 'match' string is a wildcard and matches 'name', or
208 * (3) the 'match' string is exactly the same as 'name'.
209 *
210 * and the return value tells which case it was.
211 *
212 * It returns 0 when there is no match.
213 */
214static int match_pathspec_item(const struct pathspec_item *item, int prefix,
215 const char *name, int namelen)
216{
217 /* name/namelen has prefix cut off by caller */
218 const char *match = item->match + prefix;
219 int matchlen = item->len - prefix;
220
221 /* If the match was just the prefix, we matched */
222 if (!*match)
223 return MATCHED_RECURSIVELY;
224
225 if (matchlen <= namelen && !strncmp(match, name, matchlen)) {
226 if (matchlen == namelen)
227 return MATCHED_EXACTLY;
228
229 if (match[matchlen-1] == '/' || name[matchlen] == '/')
230 return MATCHED_RECURSIVELY;
231 }
232
33e0f62b 233 if (item->use_wildcard && !fnmatch(match, name, 0))
61cf2820
NTND
234 return MATCHED_FNMATCH;
235
236 return 0;
237}
238
239/*
240 * Given a name and a list of pathspecs, see if the name matches
241 * any of the pathspecs. The caller is also interested in seeing
242 * all pathspec matches some names it calls this function with
243 * (otherwise the user could have mistyped the unmatched pathspec),
244 * and a mark is left in seen[] array for pathspec element that
245 * actually matched anything.
246 */
247int match_pathspec_depth(const struct pathspec *ps,
248 const char *name, int namelen,
249 int prefix, char *seen)
250{
251 int i, retval = 0;
252
253 if (!ps->nr) {
254 if (!ps->recursive || ps->max_depth == -1)
255 return MATCHED_RECURSIVELY;
256
257 if (within_depth(name, namelen, 0, ps->max_depth))
258 return MATCHED_EXACTLY;
259 else
260 return 0;
261 }
262
263 name += prefix;
264 namelen -= prefix;
265
266 for (i = ps->nr - 1; i >= 0; i--) {
267 int how;
268 if (seen && seen[i] == MATCHED_EXACTLY)
269 continue;
270 how = match_pathspec_item(ps->items+i, prefix, name, namelen);
271 if (ps->recursive && ps->max_depth != -1 &&
272 how && how != MATCHED_FNMATCH) {
273 int len = ps->items[i].len;
274 if (name[len] == '/')
275 len++;
276 if (within_depth(name+len, namelen-len, 0, ps->max_depth))
277 how = MATCHED_EXACTLY;
278 else
279 how = 0;
280 }
281 if (how) {
282 if (retval < how)
283 retval = how;
284 if (seen && seen[i] < how)
285 seen[i] = how;
286 }
287 }
288 return retval;
289}
290
fcd631ed
NTND
291/*
292 * Return the length of the "simple" part of a path match limiter.
293 */
294static int simple_length(const char *match)
295{
296 int len = -1;
297
298 for (;;) {
299 unsigned char c = *match++;
300 len++;
301 if (c == '\0' || is_glob_special(c))
302 return len;
303 }
304}
305
68492fc7
LK
306static int no_wildcard(const char *string)
307{
fcd631ed 308 return string[simple_length(string)] == '\0';
68492fc7
LK
309}
310
453ec4bd
LT
311void add_exclude(const char *string, const char *base,
312 int baselen, struct exclude_list *which)
313{
d6b8fc30
JH
314 struct exclude *x;
315 size_t len;
316 int to_exclude = 1;
317 int flags = 0;
453ec4bd 318
68492fc7 319 if (*string == '!') {
d6b8fc30 320 to_exclude = 0;
68492fc7
LK
321 string++;
322 }
d6b8fc30
JH
323 len = strlen(string);
324 if (len && string[len - 1] == '/') {
325 char *s;
326 x = xmalloc(sizeof(*x) + len);
4b25d091 327 s = (char *)(x+1);
d6b8fc30
JH
328 memcpy(s, string, len - 1);
329 s[len - 1] = '\0';
330 string = s;
331 x->pattern = s;
332 flags = EXC_FLAG_MUSTBEDIR;
333 } else {
334 x = xmalloc(sizeof(*x));
335 x->pattern = string;
336 }
337 x->to_exclude = to_exclude;
68492fc7 338 x->patternlen = strlen(string);
453ec4bd
LT
339 x->base = base;
340 x->baselen = baselen;
d6b8fc30 341 x->flags = flags;
68492fc7
LK
342 if (!strchr(string, '/'))
343 x->flags |= EXC_FLAG_NODIR;
f9f6e2ce 344 x->nowildcardlen = simple_length(string);
68492fc7
LK
345 if (*string == '*' && no_wildcard(string+1))
346 x->flags |= EXC_FLAG_ENDSWITH;
686a4a06 347 ALLOC_GROW(which->excludes, which->nr + 1, which->alloc);
453ec4bd
LT
348 which->excludes[which->nr++] = x;
349}
350
c28b3d6e
NTND
351static void *read_skip_worktree_file_from_index(const char *path, size_t *size)
352{
353 int pos, len;
354 unsigned long sz;
355 enum object_type type;
356 void *data;
357 struct index_state *istate = &the_index;
358
359 len = strlen(path);
360 pos = index_name_pos(istate, path, len);
361 if (pos < 0)
362 return NULL;
363 if (!ce_skip_worktree(istate->cache[pos]))
364 return NULL;
365 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
366 if (!data || type != OBJ_BLOB) {
367 free(data);
368 return NULL;
369 }
370 *size = xsize_t(sz);
371 return data;
372}
373
0fd0e241
NTND
374void free_excludes(struct exclude_list *el)
375{
376 int i;
377
378 for (i = 0; i < el->nr; i++)
379 free(el->excludes[i]);
380 free(el->excludes);
381
382 el->nr = 0;
383 el->excludes = NULL;
384}
385
cb097534
NTND
386int add_excludes_from_file_to_list(const char *fname,
387 const char *base,
388 int baselen,
389 char **buf_p,
390 struct exclude_list *which,
391 int check_index)
453ec4bd 392{
c470701a 393 struct stat st;
453ec4bd 394 int fd, i;
9d14017a 395 size_t size = 0;
453ec4bd
LT
396 char *buf, *entry;
397
398 fd = open(fname, O_RDONLY);
c28b3d6e
NTND
399 if (fd < 0 || fstat(fd, &st) < 0) {
400 if (0 <= fd)
401 close(fd);
402 if (!check_index ||
403 (buf = read_skip_worktree_file_from_index(fname, &size)) == NULL)
404 return -1;
45d76f17
NTND
405 if (size == 0) {
406 free(buf);
407 return 0;
408 }
409 if (buf[size-1] != '\n') {
410 buf = xrealloc(buf, size+1);
411 buf[size++] = '\n';
412 }
453ec4bd 413 }
c28b3d6e
NTND
414 else {
415 size = xsize_t(st.st_size);
416 if (size == 0) {
417 close(fd);
418 return 0;
419 }
45d76f17 420 buf = xmalloc(size+1);
c28b3d6e 421 if (read_in_full(fd, buf, size) != size) {
45d76f17 422 free(buf);
c28b3d6e
NTND
423 close(fd);
424 return -1;
425 }
45d76f17 426 buf[size++] = '\n';
c28b3d6e 427 close(fd);
6ba78238 428 }
453ec4bd 429
63d285c8
JH
430 if (buf_p)
431 *buf_p = buf;
453ec4bd 432 entry = buf;
45d76f17
NTND
433 for (i = 0; i < size; i++) {
434 if (buf[i] == '\n') {
453ec4bd
LT
435 if (entry != buf + i && entry[0] != '#') {
436 buf[i - (i && buf[i-1] == '\r')] = 0;
437 add_exclude(entry, base, baselen, which);
438 }
439 entry = buf + i + 1;
440 }
441 }
442 return 0;
453ec4bd
LT
443}
444
445void add_excludes_from_file(struct dir_struct *dir, const char *fname)
446{
cb097534
NTND
447 if (add_excludes_from_file_to_list(fname, "", 0, NULL,
448 &dir->exclude_list[EXC_FILE], 0) < 0)
453ec4bd
LT
449 die("cannot use %s as an exclude file", fname);
450}
451
63d285c8 452static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
453ec4bd 453{
63d285c8
JH
454 struct exclude_list *el;
455 struct exclude_stack *stk = NULL;
456 int current;
457
458 if ((!dir->exclude_per_dir) ||
459 (baselen + strlen(dir->exclude_per_dir) >= PATH_MAX))
460 return; /* too long a path -- ignore */
461
462 /* Pop the ones that are not the prefix of the path being checked. */
463 el = &dir->exclude_list[EXC_DIRS];
464 while ((stk = dir->exclude_stack) != NULL) {
465 if (stk->baselen <= baselen &&
466 !strncmp(dir->basebuf, base, stk->baselen))
467 break;
468 dir->exclude_stack = stk->prev;
469 while (stk->exclude_ix < el->nr)
470 free(el->excludes[--el->nr]);
471 free(stk->filebuf);
472 free(stk);
453ec4bd 473 }
453ec4bd 474
63d285c8
JH
475 /* Read from the parent directories and push them down. */
476 current = stk ? stk->baselen : -1;
477 while (current < baselen) {
478 struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
479 const char *cp;
453ec4bd 480
63d285c8
JH
481 if (current < 0) {
482 cp = base;
483 current = 0;
484 }
485 else {
486 cp = strchr(base + current + 1, '/');
487 if (!cp)
488 die("oops in prep_exclude");
489 cp++;
490 }
491 stk->prev = dir->exclude_stack;
492 stk->baselen = cp - base;
493 stk->exclude_ix = el->nr;
494 memcpy(dir->basebuf + current, base + current,
495 stk->baselen - current);
496 strcpy(dir->basebuf + stk->baselen, dir->exclude_per_dir);
cb097534
NTND
497 add_excludes_from_file_to_list(dir->basebuf,
498 dir->basebuf, stk->baselen,
499 &stk->filebuf, el, 1);
63d285c8
JH
500 dir->exclude_stack = stk;
501 current = stk->baselen;
502 }
503 dir->basebuf[baselen] = '\0';
453ec4bd
LT
504}
505
593cb880
NTND
506static int match_basename(const char *basename, int basenamelen,
507 const char *pattern, int prefix, int patternlen,
508 int flags)
509{
510 if (prefix == patternlen) {
511 if (!strcmp_icase(pattern, basename))
512 return 1;
513 } else if (flags & EXC_FLAG_ENDSWITH) {
514 if (patternlen - 1 <= basenamelen &&
515 !strcmp_icase(pattern + 1,
516 basename + basenamelen - patternlen + 1))
517 return 1;
518 } else {
519 if (fnmatch_icase(pattern, basename, 0) == 0)
520 return 1;
521 }
522 return 0;
523}
524
b5592632
NTND
525static int match_pathname(const char *pathname, int pathlen,
526 const char *base, int baselen,
527 const char *pattern, int prefix, int patternlen,
528 int flags)
529{
530 const char *name;
531 int namelen;
532
533 /*
534 * match with FNM_PATHNAME; the pattern has base implicitly
535 * in front of it.
536 */
537 if (*pattern == '/') {
538 pattern++;
539 prefix--;
540 }
541
542 /*
543 * baselen does not count the trailing slash. base[] may or
544 * may not end with a trailing slash though.
545 */
546 if (pathlen < baselen + 1 ||
547 (baselen && pathname[baselen] != '/') ||
548 strncmp_icase(pathname, base, baselen))
549 return 0;
550
551 namelen = baselen ? pathlen - baselen - 1 : pathlen;
552 name = pathname + pathlen - namelen;
553
554 if (prefix) {
555 /*
556 * if the non-wildcard part is longer than the
557 * remaining pathname, surely it cannot match.
558 */
559 if (prefix > namelen)
560 return 0;
561
562 if (strncmp_icase(pattern, name, prefix))
563 return 0;
564 pattern += prefix;
565 name += prefix;
566 namelen -= prefix;
567 }
568
569 return fnmatch_icase(pattern, name, FNM_PATHNAME) == 0;
570}
571
2c5b0115 572/* Scan the list and let the last match determine the fate.
453ec4bd
LT
573 * Return 1 for exclude, 0 for include and -1 for undecided.
574 */
cb097534
NTND
575int excluded_from_list(const char *pathname,
576 int pathlen, const char *basename, int *dtype,
577 struct exclude_list *el)
453ec4bd
LT
578{
579 int i;
580
35a94d44
NTND
581 if (!el->nr)
582 return -1; /* undefined */
d6b8fc30 583
35a94d44
NTND
584 for (i = el->nr - 1; 0 <= i; i--) {
585 struct exclude *x = el->excludes[i];
b5592632 586 const char *exclude = x->pattern;
35a94d44 587 int to_exclude = x->to_exclude;
b5592632 588 int prefix = x->nowildcardlen;
35a94d44
NTND
589
590 if (x->flags & EXC_FLAG_MUSTBEDIR) {
591 if (*dtype == DT_UNKNOWN)
592 *dtype = get_dtype(NULL, pathname, pathlen);
593 if (*dtype != DT_DIR)
594 continue;
595 }
d6b8fc30 596
35a94d44 597 if (x->flags & EXC_FLAG_NODIR) {
593cb880
NTND
598 if (match_basename(basename,
599 pathlen - (basename - pathname),
600 exclude, prefix, x->patternlen,
601 x->flags))
602 return to_exclude;
35a94d44
NTND
603 continue;
604 }
605
b5592632
NTND
606 assert(x->baselen == 0 || x->base[x->baselen - 1] == '/');
607 if (match_pathname(pathname, pathlen,
608 x->base, x->baselen ? x->baselen - 1 : 0,
609 exclude, prefix, x->patternlen, x->flags))
f9f6e2ce 610 return to_exclude;
453ec4bd
LT
611 }
612 return -1; /* undecided */
613}
614
0d316f0c 615static int excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
453ec4bd
LT
616{
617 int pathlen = strlen(pathname);
618 int st;
68492fc7
LK
619 const char *basename = strrchr(pathname, '/');
620 basename = (basename) ? basename+1 : pathname;
453ec4bd 621
63d285c8 622 prep_exclude(dir, pathname, basename-pathname);
453ec4bd 623 for (st = EXC_CMDL; st <= EXC_FILE; st++) {
cb097534
NTND
624 switch (excluded_from_list(pathname, pathlen, basename,
625 dtype_p, &dir->exclude_list[st])) {
453ec4bd
LT
626 case 0:
627 return 0;
628 case 1:
629 return 1;
630 }
631 }
632 return 0;
633}
634
eb41775e
JH
635void path_exclude_check_init(struct path_exclude_check *check,
636 struct dir_struct *dir)
637{
638 check->dir = dir;
639 strbuf_init(&check->path, 256);
640}
641
642void path_exclude_check_clear(struct path_exclude_check *check)
643{
644 strbuf_release(&check->path);
645}
646
93921b07 647/*
782cd4c0 648 * Is this name excluded? This is for a caller like show_files() that
93921b07
JH
649 * do not honor directory hierarchy and iterate through paths that are
650 * possibly in an ignored directory.
651 *
652 * A path to a directory known to be excluded is left in check->path to
653 * optimize for repeated checks for files in the same excluded directory.
654 */
782cd4c0
JH
655int path_excluded(struct path_exclude_check *check,
656 const char *name, int namelen, int *dtype)
eb41775e 657{
782cd4c0 658 int i;
eb41775e
JH
659 struct strbuf *path = &check->path;
660
782cd4c0
JH
661 /*
662 * we allow the caller to pass namelen as an optimization; it
663 * must match the length of the name, as we eventually call
664 * excluded() on the whole name string.
665 */
666 if (namelen < 0)
667 namelen = strlen(name);
668
93921b07 669 if (path->len &&
782cd4c0
JH
670 path->len <= namelen &&
671 !memcmp(name, path->buf, path->len) &&
672 (!name[path->len] || name[path->len] == '/'))
93921b07
JH
673 return 1;
674
eb41775e 675 strbuf_setlen(path, 0);
782cd4c0
JH
676 for (i = 0; name[i]; i++) {
677 int ch = name[i];
eb41775e
JH
678
679 if (ch == '/') {
782cd4c0
JH
680 int dt = DT_DIR;
681 if (excluded(check->dir, path->buf, &dt))
eb41775e
JH
682 return 1;
683 }
684 strbuf_addch(path, ch);
685 }
93921b07
JH
686
687 /* An entry in the index; cannot be a directory with subentries */
688 strbuf_setlen(path, 0);
689
782cd4c0 690 return excluded(check->dir, name, dtype);
eb41775e
JH
691}
692
f3fa1838
JH
693static struct dir_entry *dir_entry_new(const char *pathname, int len)
694{
453ec4bd
LT
695 struct dir_entry *ent;
696
453ec4bd
LT
697 ent = xmalloc(sizeof(*ent) + len + 1);
698 ent->len = len;
699 memcpy(ent->name, pathname, len);
700 ent->name[len] = 0;
4d06f8ac 701 return ent;
453ec4bd
LT
702}
703
159b3212 704static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
6815e569 705{
0a9b88b7 706 if (cache_name_exists(pathname, len, ignore_case))
6815e569
JK
707 return NULL;
708
25fd2f7a 709 ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
6815e569
JK
710 return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
711}
712
108da0db 713struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
2abd31b0 714{
6e4f981f 715 if (!cache_name_is_other(pathname, len))
2abd31b0
JK
716 return NULL;
717
25fd2f7a 718 ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
2abd31b0
JK
719 return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
720}
721
09595258
LT
722enum exist_status {
723 index_nonexistent = 0,
724 index_directory,
4b05548f 725 index_gitdir
09595258
LT
726};
727
5102c617
JJ
728/*
729 * Do not use the alphabetically stored index to look up
730 * the directory name; instead, use the case insensitive
731 * name hash.
732 */
733static enum exist_status directory_exists_in_index_icase(const char *dirname, int len)
734{
735 struct cache_entry *ce = index_name_exists(&the_index, dirname, len + 1, ignore_case);
736 unsigned char endchar;
737
738 if (!ce)
739 return index_nonexistent;
740 endchar = ce->name[len];
741
742 /*
743 * The cache_entry structure returned will contain this dirname
744 * and possibly additional path components.
745 */
746 if (endchar == '/')
747 return index_directory;
748
749 /*
750 * If there are no additional path components, then this cache_entry
751 * represents a submodule. Submodules, despite being directories,
752 * are stored in the cache without a closing slash.
753 */
754 if (!endchar && S_ISGITLINK(ce->ce_mode))
755 return index_gitdir;
756
757 /* This should never be hit, but it exists just in case. */
758 return index_nonexistent;
759}
760
09595258
LT
761/*
762 * The index sorts alphabetically by entry name, which
763 * means that a gitlink sorts as '\0' at the end, while
764 * a directory (which is defined not as an entry, but as
765 * the files it contains) will sort with the '/' at the
766 * end.
767 */
768static enum exist_status directory_exists_in_index(const char *dirname, int len)
453ec4bd 769{
5102c617
JJ
770 int pos;
771
772 if (ignore_case)
773 return directory_exists_in_index_icase(dirname, len);
774
775 pos = cache_name_pos(dirname, len);
09595258
LT
776 if (pos < 0)
777 pos = -pos-1;
778 while (pos < active_nr) {
779 struct cache_entry *ce = active_cache[pos++];
780 unsigned char endchar;
781
782 if (strncmp(ce->name, dirname, len))
783 break;
784 endchar = ce->name[len];
785 if (endchar > '/')
786 break;
787 if (endchar == '/')
788 return index_directory;
7a51ed66 789 if (!endchar && S_ISGITLINK(ce->ce_mode))
09595258
LT
790 return index_gitdir;
791 }
792 return index_nonexistent;
793}
794
795/*
796 * When we find a directory when traversing the filesystem, we
797 * have three distinct cases:
798 *
799 * - ignore it
800 * - see it as a directory
801 * - recurse into it
802 *
803 * and which one we choose depends on a combination of existing
804 * git index contents and the flags passed into the directory
805 * traversal routine.
806 *
807 * Case 1: If we *already* have entries in the index under that
808 * directory name, we always recurse into the directory to see
809 * all the files.
810 *
811 * Case 2: If we *already* have that directory name as a gitlink,
812 * we always continue to see it as a gitlink, regardless of whether
813 * there is an actual git directory there or not (it might not
814 * be checked out as a subproject!)
815 *
816 * Case 3: if we didn't have it in the index previously, we
817 * have a few sub-cases:
818 *
819 * (a) if "show_other_directories" is true, we show it as
820 * just a directory, unless "hide_empty_directories" is
821 * also true and the directory is empty, in which case
822 * we just ignore it entirely.
823 * (b) if it looks like a git directory, and we don't have
302b9282 824 * 'no_gitlinks' set we treat it as a gitlink, and show it
09595258
LT
825 * as a directory.
826 * (c) otherwise, we recurse into it.
827 */
828enum directory_treatment {
829 show_directory,
830 ignore_directory,
4b05548f 831 recurse_into_directory
09595258
LT
832};
833
834static enum directory_treatment treat_directory(struct dir_struct *dir,
835 const char *dirname, int len,
836 const struct path_simplify *simplify)
837{
838 /* The "len-1" is to strip the final '/' */
839 switch (directory_exists_in_index(dirname, len-1)) {
840 case index_directory:
841 return recurse_into_directory;
842
843 case index_gitdir:
7c4c97c0 844 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
ab22aed3 845 return ignore_directory;
09595258
LT
846 return show_directory;
847
848 case index_nonexistent:
7c4c97c0 849 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
09595258 850 break;
7c4c97c0 851 if (!(dir->flags & DIR_NO_GITLINKS)) {
09595258
LT
852 unsigned char sha1[20];
853 if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
854 return show_directory;
855 }
856 return recurse_into_directory;
857 }
858
859 /* This is the "show_other_directories" case */
7c4c97c0 860 if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
09595258 861 return show_directory;
dba2e203 862 if (!read_directory_recursive(dir, dirname, len, 1, simplify))
09595258
LT
863 return ignore_directory;
864 return show_directory;
453ec4bd
LT
865}
866
9fc42d60
LT
867/*
868 * This is an inexact early pruning of any recursive directory
869 * reading - if the path cannot possibly be in the pathspec,
870 * return true, and we'll skip it early.
871 */
872static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
873{
874 if (simplify) {
875 for (;;) {
876 const char *match = simplify->path;
877 int len = simplify->len;
878
879 if (!match)
880 break;
881 if (len > pathlen)
882 len = pathlen;
883 if (!memcmp(path, match, len))
884 return 0;
885 simplify++;
886 }
887 return 1;
888 }
889 return 0;
890}
891
29209cbe
JK
892/*
893 * This function tells us whether an excluded path matches a
894 * list of "interesting" pathspecs. That is, whether a path matched
895 * by any of the pathspecs could possibly be ignored by excluding
896 * the specified path. This can happen if:
897 *
898 * 1. the path is mentioned explicitly in the pathspec
899 *
900 * 2. the path is a directory prefix of some element in the
901 * pathspec
902 */
903static int exclude_matches_pathspec(const char *path, int len,
904 const struct path_simplify *simplify)
e96980ef
JK
905{
906 if (simplify) {
907 for (; simplify->path; simplify++) {
908 if (len == simplify->len
909 && !memcmp(path, simplify->path, len))
910 return 1;
29209cbe
JK
911 if (len < simplify->len
912 && simplify->path[len] == '/'
913 && !memcmp(path, simplify->path, len))
914 return 1;
e96980ef
JK
915 }
916 }
917 return 0;
918}
919
443e061a
LT
920static int get_index_dtype(const char *path, int len)
921{
922 int pos;
923 struct cache_entry *ce;
924
925 ce = cache_name_exists(path, len, 0);
926 if (ce) {
927 if (!ce_uptodate(ce))
928 return DT_UNKNOWN;
929 if (S_ISGITLINK(ce->ce_mode))
930 return DT_DIR;
931 /*
932 * Nobody actually cares about the
933 * difference between DT_LNK and DT_REG
934 */
935 return DT_REG;
936 }
937
938 /* Try to look it up as a directory */
939 pos = cache_name_pos(path, len);
940 if (pos >= 0)
941 return DT_UNKNOWN;
942 pos = -pos-1;
943 while (pos < active_nr) {
944 ce = active_cache[pos++];
945 if (strncmp(ce->name, path, len))
946 break;
947 if (ce->name[len] > '/')
948 break;
949 if (ce->name[len] < '/')
950 continue;
951 if (!ce_uptodate(ce))
952 break; /* continue? */
953 return DT_DIR;
954 }
955 return DT_UNKNOWN;
956}
957
caa6b782 958static int get_dtype(struct dirent *de, const char *path, int len)
07134421 959{
6831a88a 960 int dtype = de ? DTYPE(de) : DT_UNKNOWN;
07134421
LT
961 struct stat st;
962
963 if (dtype != DT_UNKNOWN)
964 return dtype;
443e061a
LT
965 dtype = get_index_dtype(path, len);
966 if (dtype != DT_UNKNOWN)
967 return dtype;
968 if (lstat(path, &st))
07134421
LT
969 return dtype;
970 if (S_ISREG(st.st_mode))
971 return DT_REG;
972 if (S_ISDIR(st.st_mode))
973 return DT_DIR;
974 if (S_ISLNK(st.st_mode))
975 return DT_LNK;
976 return dtype;
977}
978
53cc5356
JH
979enum path_treatment {
980 path_ignored,
981 path_handled,
4b05548f 982 path_recurse
53cc5356
JH
983};
984
16e2cfa9 985static enum path_treatment treat_one_path(struct dir_struct *dir,
49dc2cc2 986 struct strbuf *path,
16e2cfa9
JH
987 const struct path_simplify *simplify,
988 int dtype, struct dirent *de)
53cc5356 989{
49dc2cc2 990 int exclude = excluded(dir, path->buf, &dtype);
53cc5356 991 if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
49dc2cc2
RS
992 && exclude_matches_pathspec(path->buf, path->len, simplify))
993 dir_add_ignored(dir, path->buf, path->len);
53cc5356
JH
994
995 /*
996 * Excluded? If we don't explicitly want to show
997 * ignored files, ignore it
998 */
999 if (exclude && !(dir->flags & DIR_SHOW_IGNORED))
1000 return path_ignored;
1001
1002 if (dtype == DT_UNKNOWN)
49dc2cc2 1003 dtype = get_dtype(de, path->buf, path->len);
53cc5356
JH
1004
1005 /*
1006 * Do we want to see just the ignored files?
1007 * We still need to recurse into directories,
1008 * even if we don't ignore them, since the
1009 * directory may contain files that we do..
1010 */
1011 if (!exclude && (dir->flags & DIR_SHOW_IGNORED)) {
1012 if (dtype != DT_DIR)
1013 return path_ignored;
1014 }
1015
1016 switch (dtype) {
1017 default:
1018 return path_ignored;
1019 case DT_DIR:
49dc2cc2
RS
1020 strbuf_addch(path, '/');
1021 switch (treat_directory(dir, path->buf, path->len, simplify)) {
53cc5356
JH
1022 case show_directory:
1023 if (exclude != !!(dir->flags
1024 & DIR_SHOW_IGNORED))
1025 return path_ignored;
1026 break;
1027 case recurse_into_directory:
1028 return path_recurse;
1029 case ignore_directory:
1030 return path_ignored;
1031 }
1032 break;
1033 case DT_REG:
1034 case DT_LNK:
1035 break;
1036 }
1037 return path_handled;
1038}
1039
16e2cfa9
JH
1040static enum path_treatment treat_path(struct dir_struct *dir,
1041 struct dirent *de,
49dc2cc2 1042 struct strbuf *path,
16e2cfa9 1043 int baselen,
49dc2cc2 1044 const struct path_simplify *simplify)
16e2cfa9
JH
1045{
1046 int dtype;
1047
1048 if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
1049 return path_ignored;
49dc2cc2
RS
1050 strbuf_setlen(path, baselen);
1051 strbuf_addstr(path, de->d_name);
1052 if (simplify_away(path->buf, path->len, simplify))
16e2cfa9
JH
1053 return path_ignored;
1054
1055 dtype = DTYPE(de);
49dc2cc2 1056 return treat_one_path(dir, path, simplify, dtype, de);
16e2cfa9
JH
1057}
1058
453ec4bd
LT
1059/*
1060 * Read a directory tree. We currently ignore anything but
1061 * directories, regular files and symlinks. That's because git
1062 * doesn't handle them at all yet. Maybe that will change some
1063 * day.
1064 *
1065 * Also, we ignore the name ".git" (even if it is not a directory).
1066 * That likely will not change.
1067 */
53cc5356
JH
1068static int read_directory_recursive(struct dir_struct *dir,
1069 const char *base, int baselen,
1070 int check_only,
1071 const struct path_simplify *simplify)
453ec4bd 1072{
1528d247 1073 DIR *fdir;
453ec4bd 1074 int contents = 0;
02cb6753 1075 struct dirent *de;
bef36921 1076 struct strbuf path = STRBUF_INIT;
453ec4bd 1077
bef36921 1078 strbuf_add(&path, base, baselen);
02cb6753 1079
1528d247
RS
1080 fdir = opendir(path.len ? path.buf : ".");
1081 if (!fdir)
1082 goto out;
1083
02cb6753 1084 while ((de = readdir(fdir)) != NULL) {
bef36921 1085 switch (treat_path(dir, de, &path, baselen, simplify)) {
02cb6753 1086 case path_recurse:
bef36921
JH
1087 contents += read_directory_recursive(dir, path.buf,
1088 path.len, 0,
1089 simplify);
02cb6753
NTND
1090 continue;
1091 case path_ignored:
1092 continue;
1093 case path_handled:
1094 break;
453ec4bd 1095 }
02cb6753
NTND
1096 contents++;
1097 if (check_only)
1528d247
RS
1098 break;
1099 dir_add_name(dir, path.buf, path.len);
453ec4bd 1100 }
02cb6753 1101 closedir(fdir);
1528d247 1102 out:
bef36921 1103 strbuf_release(&path);
453ec4bd
LT
1104
1105 return contents;
1106}
1107
1108static int cmp_name(const void *p1, const void *p2)
1109{
1110 const struct dir_entry *e1 = *(const struct dir_entry **)p1;
1111 const struct dir_entry *e2 = *(const struct dir_entry **)p2;
1112
1113 return cache_name_compare(e1->name, e1->len,
1114 e2->name, e2->len);
1115}
1116
9fc42d60
LT
1117static struct path_simplify *create_simplify(const char **pathspec)
1118{
1119 int nr, alloc = 0;
1120 struct path_simplify *simplify = NULL;
1121
1122 if (!pathspec)
1123 return NULL;
1124
1125 for (nr = 0 ; ; nr++) {
1126 const char *match;
1127 if (nr >= alloc) {
1128 alloc = alloc_nr(alloc);
1129 simplify = xrealloc(simplify, alloc * sizeof(*simplify));
1130 }
1131 match = *pathspec++;
1132 if (!match)
1133 break;
1134 simplify[nr].path = match;
1135 simplify[nr].len = simple_length(match);
1136 }
1137 simplify[nr].path = NULL;
1138 simplify[nr].len = 0;
1139 return simplify;
1140}
1141
1142static void free_simplify(struct path_simplify *simplify)
1143{
8e0f7003 1144 free(simplify);
9fc42d60
LT
1145}
1146
48ffef96
JH
1147static int treat_leading_path(struct dir_struct *dir,
1148 const char *path, int len,
1149 const struct path_simplify *simplify)
1150{
49dc2cc2
RS
1151 struct strbuf sb = STRBUF_INIT;
1152 int baselen, rc = 0;
48ffef96
JH
1153 const char *cp;
1154
1155 while (len && path[len - 1] == '/')
1156 len--;
1157 if (!len)
1158 return 1;
1159 baselen = 0;
1160 while (1) {
1161 cp = path + baselen + !!baselen;
1162 cp = memchr(cp, '/', path + len - cp);
1163 if (!cp)
1164 baselen = len;
1165 else
1166 baselen = cp - path;
49dc2cc2
RS
1167 strbuf_setlen(&sb, 0);
1168 strbuf_add(&sb, path, baselen);
1169 if (!is_directory(sb.buf))
1170 break;
1171 if (simplify_away(sb.buf, sb.len, simplify))
1172 break;
1173 if (treat_one_path(dir, &sb, simplify,
48ffef96 1174 DT_DIR, NULL) == path_ignored)
49dc2cc2
RS
1175 break; /* do not recurse into it */
1176 if (len <= baselen) {
1177 rc = 1;
1178 break; /* finished checking */
1179 }
48ffef96 1180 }
49dc2cc2
RS
1181 strbuf_release(&sb);
1182 return rc;
48ffef96
JH
1183}
1184
dba2e203 1185int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
9fc42d60 1186{
725b0605 1187 struct path_simplify *simplify;
b4189aa8 1188
dba2e203 1189 if (has_symlink_leading_path(path, len))
725b0605
JH
1190 return dir->nr;
1191
1192 simplify = create_simplify(pathspec);
48ffef96
JH
1193 if (!len || treat_leading_path(dir, path, len, simplify))
1194 read_directory_recursive(dir, path, len, 0, simplify);
9fc42d60 1195 free_simplify(simplify);
453ec4bd 1196 qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
2abd31b0 1197 qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
453ec4bd
LT
1198 return dir->nr;
1199}
c91f0d92 1200
686a4a06 1201int file_exists(const char *f)
c91f0d92 1202{
686a4a06 1203 struct stat sb;
a50f9fc5 1204 return lstat(f, &sb) == 0;
c91f0d92 1205}
e6636747
JS
1206
1207/*
9b125da4
NTND
1208 * Given two normalized paths (a trailing slash is ok), if subdir is
1209 * outside dir, return -1. Otherwise return the offset in subdir that
1210 * can be used as relative path to dir.
e6636747 1211 */
9b125da4 1212int dir_inside_of(const char *subdir, const char *dir)
e6636747 1213{
9b125da4 1214 int offset = 0;
e6636747 1215
9b125da4 1216 assert(dir && subdir && *dir && *subdir);
e6636747 1217
9b125da4 1218 while (*dir && *subdir && *dir == *subdir) {
e6636747 1219 dir++;
9b125da4
NTND
1220 subdir++;
1221 offset++;
490544b1 1222 }
9b125da4
NTND
1223
1224 /* hel[p]/me vs hel[l]/yeah */
1225 if (*dir && *subdir)
1226 return -1;
1227
1228 if (!*subdir)
1229 return !*dir ? offset : -1; /* same dir */
1230
1231 /* foo/[b]ar vs foo/[] */
1232 if (is_dir_sep(dir[-1]))
1233 return is_dir_sep(subdir[-1]) ? offset : -1;
1234
1235 /* foo[/]bar vs foo[] */
1236 return is_dir_sep(*subdir) ? offset + 1 : -1;
e6636747
JS
1237}
1238
1239int is_inside_dir(const char *dir)
1240{
b892913d
NTND
1241 char cwd[PATH_MAX];
1242 if (!dir)
1243 return 0;
1244 if (!getcwd(cwd, sizeof(cwd)))
1245 die_errno("can't find the current directory");
1246 return dir_inside_of(cwd, dir) >= 0;
e6636747 1247}
7155b727 1248
55892d23
AP
1249int is_empty_dir(const char *path)
1250{
1251 DIR *dir = opendir(path);
1252 struct dirent *e;
1253 int ret = 1;
1254
1255 if (!dir)
1256 return 0;
1257
1258 while ((e = readdir(dir)) != NULL)
1259 if (!is_dot_or_dotdot(e->d_name)) {
1260 ret = 0;
1261 break;
1262 }
1263
1264 closedir(dir);
1265 return ret;
1266}
1267
ae2f203e 1268static int remove_dir_recurse(struct strbuf *path, int flag, int *kept_up)
7155b727 1269{
a0f4afbe 1270 DIR *dir;
7155b727 1271 struct dirent *e;
ae2f203e 1272 int ret = 0, original_len = path->len, len, kept_down = 0;
a0f4afbe 1273 int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
c844a803 1274 int keep_toplevel = (flag & REMOVE_DIR_KEEP_TOPLEVEL);
a0f4afbe 1275 unsigned char submodule_head[20];
7155b727 1276
a0f4afbe 1277 if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
ae2f203e 1278 !resolve_gitlink_ref(path->buf, "HEAD", submodule_head)) {
a0f4afbe 1279 /* Do not descend and nuke a nested git work tree. */
ae2f203e
JH
1280 if (kept_up)
1281 *kept_up = 1;
a0f4afbe 1282 return 0;
ae2f203e 1283 }
a0f4afbe 1284
ae2f203e 1285 flag &= ~REMOVE_DIR_KEEP_TOPLEVEL;
a0f4afbe 1286 dir = opendir(path->buf);
c844a803 1287 if (!dir) {
ae2f203e 1288 /* an empty dir could be removed even if it is unreadble */
c844a803
JH
1289 if (!keep_toplevel)
1290 return rmdir(path->buf);
1291 else
1292 return -1;
1293 }
7155b727
JS
1294 if (path->buf[original_len - 1] != '/')
1295 strbuf_addch(path, '/');
1296
1297 len = path->len;
1298 while ((e = readdir(dir)) != NULL) {
1299 struct stat st;
8ca12c0d
AP
1300 if (is_dot_or_dotdot(e->d_name))
1301 continue;
7155b727
JS
1302
1303 strbuf_setlen(path, len);
1304 strbuf_addstr(path, e->d_name);
1305 if (lstat(path->buf, &st))
1306 ; /* fall thru */
1307 else if (S_ISDIR(st.st_mode)) {
ae2f203e 1308 if (!remove_dir_recurse(path, flag, &kept_down))
7155b727
JS
1309 continue; /* happy */
1310 } else if (!only_empty && !unlink(path->buf))
1311 continue; /* happy, too */
1312
1313 /* path too long, stat fails, or non-directory still exists */
1314 ret = -1;
1315 break;
1316 }
1317 closedir(dir);
1318
1319 strbuf_setlen(path, original_len);
ae2f203e 1320 if (!ret && !keep_toplevel && !kept_down)
7155b727 1321 ret = rmdir(path->buf);
ae2f203e
JH
1322 else if (kept_up)
1323 /*
1324 * report the uplevel that it is not an error that we
1325 * did not rmdir() our directory.
1326 */
1327 *kept_up = !ret;
7155b727
JS
1328 return ret;
1329}
039bc64e 1330
ae2f203e
JH
1331int remove_dir_recursively(struct strbuf *path, int flag)
1332{
1333 return remove_dir_recurse(path, flag, NULL);
1334}
1335
039bc64e
JH
1336void setup_standard_excludes(struct dir_struct *dir)
1337{
1338 const char *path;
1339
1340 dir->exclude_per_dir = ".gitignore";
1341 path = git_path("info/exclude");
1342 if (!access(path, R_OK))
1343 add_excludes_from_file(dir, path);
1344 if (excludes_file && !access(excludes_file, R_OK))
1345 add_excludes_from_file(dir, excludes_file);
1346}
4a92d1bf
AR
1347
1348int remove_path(const char *name)
1349{
1350 char *slash;
1351
1352 if (unlink(name) && errno != ENOENT)
1353 return -1;
1354
1355 slash = strrchr(name, '/');
1356 if (slash) {
1357 char *dirs = xstrdup(name);
1358 slash = dirs + (slash - name);
1359 do {
1360 *slash = '\0';
3fc0d131 1361 } while (rmdir(dirs) == 0 && (slash = strrchr(dirs, '/')));
4a92d1bf
AR
1362 free(dirs);
1363 }
1364 return 0;
1365}
1366
86e4ca69
NTND
1367static int pathspec_item_cmp(const void *a_, const void *b_)
1368{
1369 struct pathspec_item *a, *b;
1370
1371 a = (struct pathspec_item *)a_;
1372 b = (struct pathspec_item *)b_;
1373 return strcmp(a->match, b->match);
1374}
1375
0602f3e9
NTND
1376int init_pathspec(struct pathspec *pathspec, const char **paths)
1377{
1378 const char **p = paths;
1379 int i;
1380
1381 memset(pathspec, 0, sizeof(*pathspec));
1382 if (!p)
1383 return 0;
1384 while (*p)
1385 p++;
1386 pathspec->raw = paths;
1387 pathspec->nr = p - paths;
1388 if (!pathspec->nr)
1389 return 0;
1390
1391 pathspec->items = xmalloc(sizeof(struct pathspec_item)*pathspec->nr);
1392 for (i = 0; i < pathspec->nr; i++) {
1393 struct pathspec_item *item = pathspec->items+i;
1394 const char *path = paths[i];
1395
1396 item->match = path;
1397 item->len = strlen(path);
33e0f62b
JH
1398 item->use_wildcard = !no_wildcard(path);
1399 if (item->use_wildcard)
d38f2809 1400 pathspec->has_wildcard = 1;
0602f3e9 1401 }
86e4ca69
NTND
1402
1403 qsort(pathspec->items, pathspec->nr,
1404 sizeof(struct pathspec_item), pathspec_item_cmp);
1405
0602f3e9
NTND
1406 return 0;
1407}
1408
1409void free_pathspec(struct pathspec *pathspec)
1410{
1411 free(pathspec->items);
1412 pathspec->items = NULL;
1413}