dir.c: get rid of the wildcard symbol set in no_wildcard()
[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;
344 if (no_wildcard(string))
345 x->flags |= EXC_FLAG_NOWILDCARD;
346 if (*string == '*' && no_wildcard(string+1))
347 x->flags |= EXC_FLAG_ENDSWITH;
686a4a06 348 ALLOC_GROW(which->excludes, which->nr + 1, which->alloc);
453ec4bd
LT
349 which->excludes[which->nr++] = x;
350}
351
c28b3d6e
NTND
352static void *read_skip_worktree_file_from_index(const char *path, size_t *size)
353{
354 int pos, len;
355 unsigned long sz;
356 enum object_type type;
357 void *data;
358 struct index_state *istate = &the_index;
359
360 len = strlen(path);
361 pos = index_name_pos(istate, path, len);
362 if (pos < 0)
363 return NULL;
364 if (!ce_skip_worktree(istate->cache[pos]))
365 return NULL;
366 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
367 if (!data || type != OBJ_BLOB) {
368 free(data);
369 return NULL;
370 }
371 *size = xsize_t(sz);
372 return data;
373}
374
0fd0e241
NTND
375void free_excludes(struct exclude_list *el)
376{
377 int i;
378
379 for (i = 0; i < el->nr; i++)
380 free(el->excludes[i]);
381 free(el->excludes);
382
383 el->nr = 0;
384 el->excludes = NULL;
385}
386
cb097534
NTND
387int add_excludes_from_file_to_list(const char *fname,
388 const char *base,
389 int baselen,
390 char **buf_p,
391 struct exclude_list *which,
392 int check_index)
453ec4bd 393{
c470701a 394 struct stat st;
453ec4bd 395 int fd, i;
9d14017a 396 size_t size = 0;
453ec4bd
LT
397 char *buf, *entry;
398
399 fd = open(fname, O_RDONLY);
c28b3d6e
NTND
400 if (fd < 0 || fstat(fd, &st) < 0) {
401 if (0 <= fd)
402 close(fd);
403 if (!check_index ||
404 (buf = read_skip_worktree_file_from_index(fname, &size)) == NULL)
405 return -1;
45d76f17
NTND
406 if (size == 0) {
407 free(buf);
408 return 0;
409 }
410 if (buf[size-1] != '\n') {
411 buf = xrealloc(buf, size+1);
412 buf[size++] = '\n';
413 }
453ec4bd 414 }
c28b3d6e
NTND
415 else {
416 size = xsize_t(st.st_size);
417 if (size == 0) {
418 close(fd);
419 return 0;
420 }
45d76f17 421 buf = xmalloc(size+1);
c28b3d6e 422 if (read_in_full(fd, buf, size) != size) {
45d76f17 423 free(buf);
c28b3d6e
NTND
424 close(fd);
425 return -1;
426 }
45d76f17 427 buf[size++] = '\n';
c28b3d6e 428 close(fd);
6ba78238 429 }
453ec4bd 430
63d285c8
JH
431 if (buf_p)
432 *buf_p = buf;
453ec4bd 433 entry = buf;
45d76f17
NTND
434 for (i = 0; i < size; i++) {
435 if (buf[i] == '\n') {
453ec4bd
LT
436 if (entry != buf + i && entry[0] != '#') {
437 buf[i - (i && buf[i-1] == '\r')] = 0;
438 add_exclude(entry, base, baselen, which);
439 }
440 entry = buf + i + 1;
441 }
442 }
443 return 0;
453ec4bd
LT
444}
445
446void add_excludes_from_file(struct dir_struct *dir, const char *fname)
447{
cb097534
NTND
448 if (add_excludes_from_file_to_list(fname, "", 0, NULL,
449 &dir->exclude_list[EXC_FILE], 0) < 0)
453ec4bd
LT
450 die("cannot use %s as an exclude file", fname);
451}
452
63d285c8 453static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
453ec4bd 454{
63d285c8
JH
455 struct exclude_list *el;
456 struct exclude_stack *stk = NULL;
457 int current;
458
459 if ((!dir->exclude_per_dir) ||
460 (baselen + strlen(dir->exclude_per_dir) >= PATH_MAX))
461 return; /* too long a path -- ignore */
462
463 /* Pop the ones that are not the prefix of the path being checked. */
464 el = &dir->exclude_list[EXC_DIRS];
465 while ((stk = dir->exclude_stack) != NULL) {
466 if (stk->baselen <= baselen &&
467 !strncmp(dir->basebuf, base, stk->baselen))
468 break;
469 dir->exclude_stack = stk->prev;
470 while (stk->exclude_ix < el->nr)
471 free(el->excludes[--el->nr]);
472 free(stk->filebuf);
473 free(stk);
453ec4bd 474 }
453ec4bd 475
63d285c8
JH
476 /* Read from the parent directories and push them down. */
477 current = stk ? stk->baselen : -1;
478 while (current < baselen) {
479 struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
480 const char *cp;
453ec4bd 481
63d285c8
JH
482 if (current < 0) {
483 cp = base;
484 current = 0;
485 }
486 else {
487 cp = strchr(base + current + 1, '/');
488 if (!cp)
489 die("oops in prep_exclude");
490 cp++;
491 }
492 stk->prev = dir->exclude_stack;
493 stk->baselen = cp - base;
494 stk->exclude_ix = el->nr;
495 memcpy(dir->basebuf + current, base + current,
496 stk->baselen - current);
497 strcpy(dir->basebuf + stk->baselen, dir->exclude_per_dir);
cb097534
NTND
498 add_excludes_from_file_to_list(dir->basebuf,
499 dir->basebuf, stk->baselen,
500 &stk->filebuf, el, 1);
63d285c8
JH
501 dir->exclude_stack = stk;
502 current = stk->baselen;
503 }
504 dir->basebuf[baselen] = '\0';
453ec4bd
LT
505}
506
2c5b0115 507/* Scan the list and let the last match determine the fate.
453ec4bd
LT
508 * Return 1 for exclude, 0 for include and -1 for undecided.
509 */
cb097534
NTND
510int excluded_from_list(const char *pathname,
511 int pathlen, const char *basename, int *dtype,
512 struct exclude_list *el)
453ec4bd
LT
513{
514 int i;
515
35a94d44
NTND
516 if (!el->nr)
517 return -1; /* undefined */
518
519 for (i = el->nr - 1; 0 <= i; i--) {
520 struct exclude *x = el->excludes[i];
521 const char *exclude = x->pattern;
522 int to_exclude = x->to_exclude;
523
524 if (x->flags & EXC_FLAG_MUSTBEDIR) {
525 if (*dtype == DT_UNKNOWN)
526 *dtype = get_dtype(NULL, pathname, pathlen);
527 if (*dtype != DT_DIR)
528 continue;
529 }
d6b8fc30 530
35a94d44
NTND
531 if (x->flags & EXC_FLAG_NODIR) {
532 /* match basename */
533 if (x->flags & EXC_FLAG_NOWILDCARD) {
534 if (!strcmp_icase(exclude, basename))
535 return to_exclude;
536 } else if (x->flags & EXC_FLAG_ENDSWITH) {
537 if (x->patternlen - 1 <= pathlen &&
538 !strcmp_icase(exclude + 1, pathname + pathlen - x->patternlen + 1))
539 return to_exclude;
540 } else {
541 if (fnmatch_icase(exclude, basename, 0) == 0)
542 return to_exclude;
453ec4bd 543 }
35a94d44
NTND
544 continue;
545 }
546
547
548 /* match with FNM_PATHNAME:
549 * exclude has base (baselen long) implicitly in front of it.
550 */
551 if (*exclude == '/')
552 exclude++;
553
554 if (pathlen < x->baselen ||
555 (x->baselen && pathname[x->baselen-1] != '/') ||
556 strncmp_icase(pathname, x->base, x->baselen))
557 continue;
558
559 if (x->flags & EXC_FLAG_NOWILDCARD) {
560 if (!strcmp_icase(exclude, pathname + x->baselen))
561 return to_exclude;
562 } else {
563 if (fnmatch_icase(exclude, pathname+x->baselen,
564 FNM_PATHNAME) == 0)
565 return to_exclude;
453ec4bd
LT
566 }
567 }
568 return -1; /* undecided */
569}
570
6831a88a 571int excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
453ec4bd
LT
572{
573 int pathlen = strlen(pathname);
574 int st;
68492fc7
LK
575 const char *basename = strrchr(pathname, '/');
576 basename = (basename) ? basename+1 : pathname;
453ec4bd 577
63d285c8 578 prep_exclude(dir, pathname, basename-pathname);
453ec4bd 579 for (st = EXC_CMDL; st <= EXC_FILE; st++) {
cb097534
NTND
580 switch (excluded_from_list(pathname, pathlen, basename,
581 dtype_p, &dir->exclude_list[st])) {
453ec4bd
LT
582 case 0:
583 return 0;
584 case 1:
585 return 1;
586 }
587 }
588 return 0;
589}
590
f3fa1838
JH
591static struct dir_entry *dir_entry_new(const char *pathname, int len)
592{
453ec4bd
LT
593 struct dir_entry *ent;
594
453ec4bd
LT
595 ent = xmalloc(sizeof(*ent) + len + 1);
596 ent->len = len;
597 memcpy(ent->name, pathname, len);
598 ent->name[len] = 0;
4d06f8ac 599 return ent;
453ec4bd
LT
600}
601
159b3212 602static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
6815e569 603{
0a9b88b7 604 if (cache_name_exists(pathname, len, ignore_case))
6815e569
JK
605 return NULL;
606
25fd2f7a 607 ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
6815e569
JK
608 return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
609}
610
108da0db 611struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
2abd31b0 612{
6e4f981f 613 if (!cache_name_is_other(pathname, len))
2abd31b0
JK
614 return NULL;
615
25fd2f7a 616 ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
2abd31b0
JK
617 return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
618}
619
09595258
LT
620enum exist_status {
621 index_nonexistent = 0,
622 index_directory,
4b05548f 623 index_gitdir
09595258
LT
624};
625
5102c617
JJ
626/*
627 * Do not use the alphabetically stored index to look up
628 * the directory name; instead, use the case insensitive
629 * name hash.
630 */
631static enum exist_status directory_exists_in_index_icase(const char *dirname, int len)
632{
633 struct cache_entry *ce = index_name_exists(&the_index, dirname, len + 1, ignore_case);
634 unsigned char endchar;
635
636 if (!ce)
637 return index_nonexistent;
638 endchar = ce->name[len];
639
640 /*
641 * The cache_entry structure returned will contain this dirname
642 * and possibly additional path components.
643 */
644 if (endchar == '/')
645 return index_directory;
646
647 /*
648 * If there are no additional path components, then this cache_entry
649 * represents a submodule. Submodules, despite being directories,
650 * are stored in the cache without a closing slash.
651 */
652 if (!endchar && S_ISGITLINK(ce->ce_mode))
653 return index_gitdir;
654
655 /* This should never be hit, but it exists just in case. */
656 return index_nonexistent;
657}
658
09595258
LT
659/*
660 * The index sorts alphabetically by entry name, which
661 * means that a gitlink sorts as '\0' at the end, while
662 * a directory (which is defined not as an entry, but as
663 * the files it contains) will sort with the '/' at the
664 * end.
665 */
666static enum exist_status directory_exists_in_index(const char *dirname, int len)
453ec4bd 667{
5102c617
JJ
668 int pos;
669
670 if (ignore_case)
671 return directory_exists_in_index_icase(dirname, len);
672
673 pos = cache_name_pos(dirname, len);
09595258
LT
674 if (pos < 0)
675 pos = -pos-1;
676 while (pos < active_nr) {
677 struct cache_entry *ce = active_cache[pos++];
678 unsigned char endchar;
679
680 if (strncmp(ce->name, dirname, len))
681 break;
682 endchar = ce->name[len];
683 if (endchar > '/')
684 break;
685 if (endchar == '/')
686 return index_directory;
7a51ed66 687 if (!endchar && S_ISGITLINK(ce->ce_mode))
09595258
LT
688 return index_gitdir;
689 }
690 return index_nonexistent;
691}
692
693/*
694 * When we find a directory when traversing the filesystem, we
695 * have three distinct cases:
696 *
697 * - ignore it
698 * - see it as a directory
699 * - recurse into it
700 *
701 * and which one we choose depends on a combination of existing
702 * git index contents and the flags passed into the directory
703 * traversal routine.
704 *
705 * Case 1: If we *already* have entries in the index under that
706 * directory name, we always recurse into the directory to see
707 * all the files.
708 *
709 * Case 2: If we *already* have that directory name as a gitlink,
710 * we always continue to see it as a gitlink, regardless of whether
711 * there is an actual git directory there or not (it might not
712 * be checked out as a subproject!)
713 *
714 * Case 3: if we didn't have it in the index previously, we
715 * have a few sub-cases:
716 *
717 * (a) if "show_other_directories" is true, we show it as
718 * just a directory, unless "hide_empty_directories" is
719 * also true and the directory is empty, in which case
720 * we just ignore it entirely.
721 * (b) if it looks like a git directory, and we don't have
302b9282 722 * 'no_gitlinks' set we treat it as a gitlink, and show it
09595258
LT
723 * as a directory.
724 * (c) otherwise, we recurse into it.
725 */
726enum directory_treatment {
727 show_directory,
728 ignore_directory,
4b05548f 729 recurse_into_directory
09595258
LT
730};
731
732static enum directory_treatment treat_directory(struct dir_struct *dir,
733 const char *dirname, int len,
734 const struct path_simplify *simplify)
735{
736 /* The "len-1" is to strip the final '/' */
737 switch (directory_exists_in_index(dirname, len-1)) {
738 case index_directory:
739 return recurse_into_directory;
740
741 case index_gitdir:
7c4c97c0 742 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
ab22aed3 743 return ignore_directory;
09595258
LT
744 return show_directory;
745
746 case index_nonexistent:
7c4c97c0 747 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
09595258 748 break;
7c4c97c0 749 if (!(dir->flags & DIR_NO_GITLINKS)) {
09595258
LT
750 unsigned char sha1[20];
751 if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
752 return show_directory;
753 }
754 return recurse_into_directory;
755 }
756
757 /* This is the "show_other_directories" case */
7c4c97c0 758 if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
09595258 759 return show_directory;
dba2e203 760 if (!read_directory_recursive(dir, dirname, len, 1, simplify))
09595258
LT
761 return ignore_directory;
762 return show_directory;
453ec4bd
LT
763}
764
9fc42d60
LT
765/*
766 * This is an inexact early pruning of any recursive directory
767 * reading - if the path cannot possibly be in the pathspec,
768 * return true, and we'll skip it early.
769 */
770static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
771{
772 if (simplify) {
773 for (;;) {
774 const char *match = simplify->path;
775 int len = simplify->len;
776
777 if (!match)
778 break;
779 if (len > pathlen)
780 len = pathlen;
781 if (!memcmp(path, match, len))
782 return 0;
783 simplify++;
784 }
785 return 1;
786 }
787 return 0;
788}
789
29209cbe
JK
790/*
791 * This function tells us whether an excluded path matches a
792 * list of "interesting" pathspecs. That is, whether a path matched
793 * by any of the pathspecs could possibly be ignored by excluding
794 * the specified path. This can happen if:
795 *
796 * 1. the path is mentioned explicitly in the pathspec
797 *
798 * 2. the path is a directory prefix of some element in the
799 * pathspec
800 */
801static int exclude_matches_pathspec(const char *path, int len,
802 const struct path_simplify *simplify)
e96980ef
JK
803{
804 if (simplify) {
805 for (; simplify->path; simplify++) {
806 if (len == simplify->len
807 && !memcmp(path, simplify->path, len))
808 return 1;
29209cbe
JK
809 if (len < simplify->len
810 && simplify->path[len] == '/'
811 && !memcmp(path, simplify->path, len))
812 return 1;
e96980ef
JK
813 }
814 }
815 return 0;
816}
817
443e061a
LT
818static int get_index_dtype(const char *path, int len)
819{
820 int pos;
821 struct cache_entry *ce;
822
823 ce = cache_name_exists(path, len, 0);
824 if (ce) {
825 if (!ce_uptodate(ce))
826 return DT_UNKNOWN;
827 if (S_ISGITLINK(ce->ce_mode))
828 return DT_DIR;
829 /*
830 * Nobody actually cares about the
831 * difference between DT_LNK and DT_REG
832 */
833 return DT_REG;
834 }
835
836 /* Try to look it up as a directory */
837 pos = cache_name_pos(path, len);
838 if (pos >= 0)
839 return DT_UNKNOWN;
840 pos = -pos-1;
841 while (pos < active_nr) {
842 ce = active_cache[pos++];
843 if (strncmp(ce->name, path, len))
844 break;
845 if (ce->name[len] > '/')
846 break;
847 if (ce->name[len] < '/')
848 continue;
849 if (!ce_uptodate(ce))
850 break; /* continue? */
851 return DT_DIR;
852 }
853 return DT_UNKNOWN;
854}
855
caa6b782 856static int get_dtype(struct dirent *de, const char *path, int len)
07134421 857{
6831a88a 858 int dtype = de ? DTYPE(de) : DT_UNKNOWN;
07134421
LT
859 struct stat st;
860
861 if (dtype != DT_UNKNOWN)
862 return dtype;
443e061a
LT
863 dtype = get_index_dtype(path, len);
864 if (dtype != DT_UNKNOWN)
865 return dtype;
866 if (lstat(path, &st))
07134421
LT
867 return dtype;
868 if (S_ISREG(st.st_mode))
869 return DT_REG;
870 if (S_ISDIR(st.st_mode))
871 return DT_DIR;
872 if (S_ISLNK(st.st_mode))
873 return DT_LNK;
874 return dtype;
875}
876
53cc5356
JH
877enum path_treatment {
878 path_ignored,
879 path_handled,
4b05548f 880 path_recurse
53cc5356
JH
881};
882
16e2cfa9 883static enum path_treatment treat_one_path(struct dir_struct *dir,
49dc2cc2 884 struct strbuf *path,
16e2cfa9
JH
885 const struct path_simplify *simplify,
886 int dtype, struct dirent *de)
53cc5356 887{
49dc2cc2 888 int exclude = excluded(dir, path->buf, &dtype);
53cc5356 889 if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
49dc2cc2
RS
890 && exclude_matches_pathspec(path->buf, path->len, simplify))
891 dir_add_ignored(dir, path->buf, path->len);
53cc5356
JH
892
893 /*
894 * Excluded? If we don't explicitly want to show
895 * ignored files, ignore it
896 */
897 if (exclude && !(dir->flags & DIR_SHOW_IGNORED))
898 return path_ignored;
899
900 if (dtype == DT_UNKNOWN)
49dc2cc2 901 dtype = get_dtype(de, path->buf, path->len);
53cc5356
JH
902
903 /*
904 * Do we want to see just the ignored files?
905 * We still need to recurse into directories,
906 * even if we don't ignore them, since the
907 * directory may contain files that we do..
908 */
909 if (!exclude && (dir->flags & DIR_SHOW_IGNORED)) {
910 if (dtype != DT_DIR)
911 return path_ignored;
912 }
913
914 switch (dtype) {
915 default:
916 return path_ignored;
917 case DT_DIR:
49dc2cc2
RS
918 strbuf_addch(path, '/');
919 switch (treat_directory(dir, path->buf, path->len, simplify)) {
53cc5356
JH
920 case show_directory:
921 if (exclude != !!(dir->flags
922 & DIR_SHOW_IGNORED))
923 return path_ignored;
924 break;
925 case recurse_into_directory:
926 return path_recurse;
927 case ignore_directory:
928 return path_ignored;
929 }
930 break;
931 case DT_REG:
932 case DT_LNK:
933 break;
934 }
935 return path_handled;
936}
937
16e2cfa9
JH
938static enum path_treatment treat_path(struct dir_struct *dir,
939 struct dirent *de,
49dc2cc2 940 struct strbuf *path,
16e2cfa9 941 int baselen,
49dc2cc2 942 const struct path_simplify *simplify)
16e2cfa9
JH
943{
944 int dtype;
945
946 if (is_dot_or_dotdot(de->d_name) || !strcmp(de->d_name, ".git"))
947 return path_ignored;
49dc2cc2
RS
948 strbuf_setlen(path, baselen);
949 strbuf_addstr(path, de->d_name);
950 if (simplify_away(path->buf, path->len, simplify))
16e2cfa9
JH
951 return path_ignored;
952
953 dtype = DTYPE(de);
49dc2cc2 954 return treat_one_path(dir, path, simplify, dtype, de);
16e2cfa9
JH
955}
956
453ec4bd
LT
957/*
958 * Read a directory tree. We currently ignore anything but
959 * directories, regular files and symlinks. That's because git
960 * doesn't handle them at all yet. Maybe that will change some
961 * day.
962 *
963 * Also, we ignore the name ".git" (even if it is not a directory).
964 * That likely will not change.
965 */
53cc5356
JH
966static int read_directory_recursive(struct dir_struct *dir,
967 const char *base, int baselen,
968 int check_only,
969 const struct path_simplify *simplify)
453ec4bd 970{
1528d247 971 DIR *fdir;
453ec4bd 972 int contents = 0;
02cb6753 973 struct dirent *de;
bef36921 974 struct strbuf path = STRBUF_INIT;
453ec4bd 975
bef36921 976 strbuf_add(&path, base, baselen);
02cb6753 977
1528d247
RS
978 fdir = opendir(path.len ? path.buf : ".");
979 if (!fdir)
980 goto out;
981
02cb6753 982 while ((de = readdir(fdir)) != NULL) {
bef36921 983 switch (treat_path(dir, de, &path, baselen, simplify)) {
02cb6753 984 case path_recurse:
bef36921
JH
985 contents += read_directory_recursive(dir, path.buf,
986 path.len, 0,
987 simplify);
02cb6753
NTND
988 continue;
989 case path_ignored:
990 continue;
991 case path_handled:
992 break;
453ec4bd 993 }
02cb6753
NTND
994 contents++;
995 if (check_only)
1528d247
RS
996 break;
997 dir_add_name(dir, path.buf, path.len);
453ec4bd 998 }
02cb6753 999 closedir(fdir);
1528d247 1000 out:
bef36921 1001 strbuf_release(&path);
453ec4bd
LT
1002
1003 return contents;
1004}
1005
1006static int cmp_name(const void *p1, const void *p2)
1007{
1008 const struct dir_entry *e1 = *(const struct dir_entry **)p1;
1009 const struct dir_entry *e2 = *(const struct dir_entry **)p2;
1010
1011 return cache_name_compare(e1->name, e1->len,
1012 e2->name, e2->len);
1013}
1014
9fc42d60
LT
1015static struct path_simplify *create_simplify(const char **pathspec)
1016{
1017 int nr, alloc = 0;
1018 struct path_simplify *simplify = NULL;
1019
1020 if (!pathspec)
1021 return NULL;
1022
1023 for (nr = 0 ; ; nr++) {
1024 const char *match;
1025 if (nr >= alloc) {
1026 alloc = alloc_nr(alloc);
1027 simplify = xrealloc(simplify, alloc * sizeof(*simplify));
1028 }
1029 match = *pathspec++;
1030 if (!match)
1031 break;
1032 simplify[nr].path = match;
1033 simplify[nr].len = simple_length(match);
1034 }
1035 simplify[nr].path = NULL;
1036 simplify[nr].len = 0;
1037 return simplify;
1038}
1039
1040static void free_simplify(struct path_simplify *simplify)
1041{
8e0f7003 1042 free(simplify);
9fc42d60
LT
1043}
1044
48ffef96
JH
1045static int treat_leading_path(struct dir_struct *dir,
1046 const char *path, int len,
1047 const struct path_simplify *simplify)
1048{
49dc2cc2
RS
1049 struct strbuf sb = STRBUF_INIT;
1050 int baselen, rc = 0;
48ffef96
JH
1051 const char *cp;
1052
1053 while (len && path[len - 1] == '/')
1054 len--;
1055 if (!len)
1056 return 1;
1057 baselen = 0;
1058 while (1) {
1059 cp = path + baselen + !!baselen;
1060 cp = memchr(cp, '/', path + len - cp);
1061 if (!cp)
1062 baselen = len;
1063 else
1064 baselen = cp - path;
49dc2cc2
RS
1065 strbuf_setlen(&sb, 0);
1066 strbuf_add(&sb, path, baselen);
1067 if (!is_directory(sb.buf))
1068 break;
1069 if (simplify_away(sb.buf, sb.len, simplify))
1070 break;
1071 if (treat_one_path(dir, &sb, simplify,
48ffef96 1072 DT_DIR, NULL) == path_ignored)
49dc2cc2
RS
1073 break; /* do not recurse into it */
1074 if (len <= baselen) {
1075 rc = 1;
1076 break; /* finished checking */
1077 }
48ffef96 1078 }
49dc2cc2
RS
1079 strbuf_release(&sb);
1080 return rc;
48ffef96
JH
1081}
1082
dba2e203 1083int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
9fc42d60 1084{
725b0605 1085 struct path_simplify *simplify;
b4189aa8 1086
dba2e203 1087 if (has_symlink_leading_path(path, len))
725b0605
JH
1088 return dir->nr;
1089
1090 simplify = create_simplify(pathspec);
48ffef96
JH
1091 if (!len || treat_leading_path(dir, path, len, simplify))
1092 read_directory_recursive(dir, path, len, 0, simplify);
9fc42d60 1093 free_simplify(simplify);
453ec4bd 1094 qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
2abd31b0 1095 qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
453ec4bd
LT
1096 return dir->nr;
1097}
c91f0d92 1098
686a4a06 1099int file_exists(const char *f)
c91f0d92 1100{
686a4a06 1101 struct stat sb;
a50f9fc5 1102 return lstat(f, &sb) == 0;
c91f0d92 1103}
e6636747
JS
1104
1105/*
9b125da4
NTND
1106 * Given two normalized paths (a trailing slash is ok), if subdir is
1107 * outside dir, return -1. Otherwise return the offset in subdir that
1108 * can be used as relative path to dir.
e6636747 1109 */
9b125da4 1110int dir_inside_of(const char *subdir, const char *dir)
e6636747 1111{
9b125da4 1112 int offset = 0;
e6636747 1113
9b125da4 1114 assert(dir && subdir && *dir && *subdir);
e6636747 1115
9b125da4 1116 while (*dir && *subdir && *dir == *subdir) {
e6636747 1117 dir++;
9b125da4
NTND
1118 subdir++;
1119 offset++;
490544b1 1120 }
9b125da4
NTND
1121
1122 /* hel[p]/me vs hel[l]/yeah */
1123 if (*dir && *subdir)
1124 return -1;
1125
1126 if (!*subdir)
1127 return !*dir ? offset : -1; /* same dir */
1128
1129 /* foo/[b]ar vs foo/[] */
1130 if (is_dir_sep(dir[-1]))
1131 return is_dir_sep(subdir[-1]) ? offset : -1;
1132
1133 /* foo[/]bar vs foo[] */
1134 return is_dir_sep(*subdir) ? offset + 1 : -1;
e6636747
JS
1135}
1136
1137int is_inside_dir(const char *dir)
1138{
b892913d
NTND
1139 char cwd[PATH_MAX];
1140 if (!dir)
1141 return 0;
1142 if (!getcwd(cwd, sizeof(cwd)))
1143 die_errno("can't find the current directory");
1144 return dir_inside_of(cwd, dir) >= 0;
e6636747 1145}
7155b727 1146
55892d23
AP
1147int is_empty_dir(const char *path)
1148{
1149 DIR *dir = opendir(path);
1150 struct dirent *e;
1151 int ret = 1;
1152
1153 if (!dir)
1154 return 0;
1155
1156 while ((e = readdir(dir)) != NULL)
1157 if (!is_dot_or_dotdot(e->d_name)) {
1158 ret = 0;
1159 break;
1160 }
1161
1162 closedir(dir);
1163 return ret;
1164}
1165
ae2f203e 1166static int remove_dir_recurse(struct strbuf *path, int flag, int *kept_up)
7155b727 1167{
a0f4afbe 1168 DIR *dir;
7155b727 1169 struct dirent *e;
ae2f203e 1170 int ret = 0, original_len = path->len, len, kept_down = 0;
a0f4afbe 1171 int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
c844a803 1172 int keep_toplevel = (flag & REMOVE_DIR_KEEP_TOPLEVEL);
a0f4afbe 1173 unsigned char submodule_head[20];
7155b727 1174
a0f4afbe 1175 if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
ae2f203e 1176 !resolve_gitlink_ref(path->buf, "HEAD", submodule_head)) {
a0f4afbe 1177 /* Do not descend and nuke a nested git work tree. */
ae2f203e
JH
1178 if (kept_up)
1179 *kept_up = 1;
a0f4afbe 1180 return 0;
ae2f203e 1181 }
a0f4afbe 1182
ae2f203e 1183 flag &= ~REMOVE_DIR_KEEP_TOPLEVEL;
a0f4afbe 1184 dir = opendir(path->buf);
c844a803 1185 if (!dir) {
ae2f203e 1186 /* an empty dir could be removed even if it is unreadble */
c844a803
JH
1187 if (!keep_toplevel)
1188 return rmdir(path->buf);
1189 else
1190 return -1;
1191 }
7155b727
JS
1192 if (path->buf[original_len - 1] != '/')
1193 strbuf_addch(path, '/');
1194
1195 len = path->len;
1196 while ((e = readdir(dir)) != NULL) {
1197 struct stat st;
8ca12c0d
AP
1198 if (is_dot_or_dotdot(e->d_name))
1199 continue;
7155b727
JS
1200
1201 strbuf_setlen(path, len);
1202 strbuf_addstr(path, e->d_name);
1203 if (lstat(path->buf, &st))
1204 ; /* fall thru */
1205 else if (S_ISDIR(st.st_mode)) {
ae2f203e 1206 if (!remove_dir_recurse(path, flag, &kept_down))
7155b727
JS
1207 continue; /* happy */
1208 } else if (!only_empty && !unlink(path->buf))
1209 continue; /* happy, too */
1210
1211 /* path too long, stat fails, or non-directory still exists */
1212 ret = -1;
1213 break;
1214 }
1215 closedir(dir);
1216
1217 strbuf_setlen(path, original_len);
ae2f203e 1218 if (!ret && !keep_toplevel && !kept_down)
7155b727 1219 ret = rmdir(path->buf);
ae2f203e
JH
1220 else if (kept_up)
1221 /*
1222 * report the uplevel that it is not an error that we
1223 * did not rmdir() our directory.
1224 */
1225 *kept_up = !ret;
7155b727
JS
1226 return ret;
1227}
039bc64e 1228
ae2f203e
JH
1229int remove_dir_recursively(struct strbuf *path, int flag)
1230{
1231 return remove_dir_recurse(path, flag, NULL);
1232}
1233
039bc64e
JH
1234void setup_standard_excludes(struct dir_struct *dir)
1235{
1236 const char *path;
1237
1238 dir->exclude_per_dir = ".gitignore";
1239 path = git_path("info/exclude");
1240 if (!access(path, R_OK))
1241 add_excludes_from_file(dir, path);
1242 if (excludes_file && !access(excludes_file, R_OK))
1243 add_excludes_from_file(dir, excludes_file);
1244}
4a92d1bf
AR
1245
1246int remove_path(const char *name)
1247{
1248 char *slash;
1249
1250 if (unlink(name) && errno != ENOENT)
1251 return -1;
1252
1253 slash = strrchr(name, '/');
1254 if (slash) {
1255 char *dirs = xstrdup(name);
1256 slash = dirs + (slash - name);
1257 do {
1258 *slash = '\0';
3fc0d131 1259 } while (rmdir(dirs) == 0 && (slash = strrchr(dirs, '/')));
4a92d1bf
AR
1260 free(dirs);
1261 }
1262 return 0;
1263}
1264
86e4ca69
NTND
1265static int pathspec_item_cmp(const void *a_, const void *b_)
1266{
1267 struct pathspec_item *a, *b;
1268
1269 a = (struct pathspec_item *)a_;
1270 b = (struct pathspec_item *)b_;
1271 return strcmp(a->match, b->match);
1272}
1273
0602f3e9
NTND
1274int init_pathspec(struct pathspec *pathspec, const char **paths)
1275{
1276 const char **p = paths;
1277 int i;
1278
1279 memset(pathspec, 0, sizeof(*pathspec));
1280 if (!p)
1281 return 0;
1282 while (*p)
1283 p++;
1284 pathspec->raw = paths;
1285 pathspec->nr = p - paths;
1286 if (!pathspec->nr)
1287 return 0;
1288
1289 pathspec->items = xmalloc(sizeof(struct pathspec_item)*pathspec->nr);
1290 for (i = 0; i < pathspec->nr; i++) {
1291 struct pathspec_item *item = pathspec->items+i;
1292 const char *path = paths[i];
1293
1294 item->match = path;
1295 item->len = strlen(path);
33e0f62b
JH
1296 item->use_wildcard = !no_wildcard(path);
1297 if (item->use_wildcard)
d38f2809 1298 pathspec->has_wildcard = 1;
0602f3e9 1299 }
86e4ca69
NTND
1300
1301 qsort(pathspec->items, pathspec->nr,
1302 sizeof(struct pathspec_item), pathspec_item_cmp);
1303
0602f3e9
NTND
1304 return 0;
1305}
1306
1307void free_pathspec(struct pathspec *pathspec)
1308{
1309 free(pathspec->items);
1310 pathspec->items = NULL;
1311}