]>
Commit | Line | Data |
---|---|---|
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 |
12 | struct path_simplify { |
13 | int len; | |
14 | const char *path; | |
15 | }; | |
16 | ||
dba2e203 | 17 | static int read_directory_recursive(struct dir_struct *dir, const char *path, int len, |
09595258 | 18 | int check_only, const struct path_simplify *simplify); |
caa6b782 | 19 | static int get_dtype(struct dirent *de, const char *path, int len); |
09595258 | 20 | |
1d8842d9 | 21 | static int common_prefix(const char **pathspec) |
3c6a370b LT |
22 | { |
23 | const char *path, *slash, *next; | |
24 | int prefix; | |
25 | ||
26 | if (!pathspec) | |
27 | return 0; | |
28 | ||
29 | path = *pathspec; | |
30 | slash = strrchr(path, '/'); | |
31 | if (!slash) | |
32 | return 0; | |
33 | ||
34 | prefix = slash - path + 1; | |
35 | while ((next = *++pathspec) != NULL) { | |
36 | int len = strlen(next); | |
c7f34c18 | 37 | if (len >= prefix && !memcmp(path, next, prefix)) |
3c6a370b | 38 | continue; |
c7f34c18 | 39 | len = prefix - 1; |
3c6a370b LT |
40 | for (;;) { |
41 | if (!len) | |
42 | return 0; | |
43 | if (next[--len] != '/') | |
44 | continue; | |
45 | if (memcmp(path, next, len+1)) | |
46 | continue; | |
47 | prefix = len + 1; | |
48 | break; | |
49 | } | |
50 | } | |
51 | return prefix; | |
52 | } | |
53 | ||
1d8842d9 LT |
54 | int fill_directory(struct dir_struct *dir, const char **pathspec) |
55 | { | |
dba2e203 LT |
56 | const char *path; |
57 | int len; | |
1d8842d9 LT |
58 | |
59 | /* | |
60 | * Calculate common prefix for the pathspec, and | |
61 | * use that to optimize the directory walk | |
62 | */ | |
dba2e203 | 63 | len = common_prefix(pathspec); |
1d8842d9 | 64 | path = ""; |
1d8842d9 | 65 | |
dba2e203 LT |
66 | if (len) |
67 | path = xmemdupz(*pathspec, len); | |
1d8842d9 LT |
68 | |
69 | /* Read the directory and prune it */ | |
dba2e203 LT |
70 | read_directory(dir, path, len, pathspec); |
71 | return len; | |
1d8842d9 LT |
72 | } |
73 | ||
e813d50e | 74 | /* |
2c5b0115 | 75 | * Does 'match' match the given name? |
e813d50e JH |
76 | * A match is found if |
77 | * | |
78 | * (1) the 'match' string is leading directory of 'name', or | |
79 | * (2) the 'match' string is a wildcard and matches 'name', or | |
80 | * (3) the 'match' string is exactly the same as 'name'. | |
81 | * | |
82 | * and the return value tells which case it was. | |
83 | * | |
84 | * It returns 0 when there is no match. | |
85 | */ | |
3c6a370b LT |
86 | static int match_one(const char *match, const char *name, int namelen) |
87 | { | |
88 | int matchlen; | |
89 | ||
90 | /* If the match was just the prefix, we matched */ | |
88ea8112 | 91 | if (!*match) |
e813d50e | 92 | return MATCHED_RECURSIVELY; |
3c6a370b | 93 | |
88ea8112 LT |
94 | for (;;) { |
95 | unsigned char c1 = *match; | |
96 | unsigned char c2 = *name; | |
8cc32992 | 97 | if (c1 == '\0' || is_glob_special(c1)) |
88ea8112 LT |
98 | break; |
99 | if (c1 != c2) | |
100 | return 0; | |
101 | match++; | |
102 | name++; | |
103 | namelen--; | |
104 | } | |
105 | ||
106 | ||
3c6a370b LT |
107 | /* |
108 | * If we don't match the matchstring exactly, | |
109 | * we need to match by fnmatch | |
110 | */ | |
88ea8112 | 111 | matchlen = strlen(match); |
3c6a370b | 112 | if (strncmp(match, name, matchlen)) |
e813d50e | 113 | return !fnmatch(match, name, 0) ? MATCHED_FNMATCH : 0; |
3c6a370b | 114 | |
f2d0df71 | 115 | if (namelen == matchlen) |
e813d50e JH |
116 | return MATCHED_EXACTLY; |
117 | if (match[matchlen-1] == '/' || name[matchlen] == '/') | |
118 | return MATCHED_RECURSIVELY; | |
119 | return 0; | |
3c6a370b LT |
120 | } |
121 | ||
e813d50e JH |
122 | /* |
123 | * Given a name and a list of pathspecs, see if the name matches | |
124 | * any of the pathspecs. The caller is also interested in seeing | |
125 | * all pathspec matches some names it calls this function with | |
126 | * (otherwise the user could have mistyped the unmatched pathspec), | |
127 | * and a mark is left in seen[] array for pathspec element that | |
128 | * actually matched anything. | |
129 | */ | |
0b50922a CB |
130 | int match_pathspec(const char **pathspec, const char *name, int namelen, |
131 | int prefix, char *seen) | |
3c6a370b | 132 | { |
0b50922a CB |
133 | int i, retval = 0; |
134 | ||
135 | if (!pathspec) | |
136 | return 1; | |
3c6a370b LT |
137 | |
138 | name += prefix; | |
139 | namelen -= prefix; | |
140 | ||
0b50922a | 141 | for (i = 0; pathspec[i] != NULL; i++) { |
e813d50e | 142 | int how; |
0b50922a CB |
143 | const char *match = pathspec[i] + prefix; |
144 | if (seen && seen[i] == MATCHED_EXACTLY) | |
3c6a370b | 145 | continue; |
e813d50e JH |
146 | how = match_one(match, name, namelen); |
147 | if (how) { | |
148 | if (retval < how) | |
149 | retval = how; | |
0b50922a CB |
150 | if (seen && seen[i] < how) |
151 | seen[i] = how; | |
3c6a370b LT |
152 | } |
153 | } | |
154 | return retval; | |
155 | } | |
156 | ||
68492fc7 LK |
157 | static int no_wildcard(const char *string) |
158 | { | |
dd482eea | 159 | return string[strcspn(string, "*?[{\\")] == '\0'; |
68492fc7 LK |
160 | } |
161 | ||
453ec4bd LT |
162 | void add_exclude(const char *string, const char *base, |
163 | int baselen, struct exclude_list *which) | |
164 | { | |
d6b8fc30 JH |
165 | struct exclude *x; |
166 | size_t len; | |
167 | int to_exclude = 1; | |
168 | int flags = 0; | |
453ec4bd | 169 | |
68492fc7 | 170 | if (*string == '!') { |
d6b8fc30 | 171 | to_exclude = 0; |
68492fc7 LK |
172 | string++; |
173 | } | |
d6b8fc30 JH |
174 | len = strlen(string); |
175 | if (len && string[len - 1] == '/') { | |
176 | char *s; | |
177 | x = xmalloc(sizeof(*x) + len); | |
4b25d091 | 178 | s = (char *)(x+1); |
d6b8fc30 JH |
179 | memcpy(s, string, len - 1); |
180 | s[len - 1] = '\0'; | |
181 | string = s; | |
182 | x->pattern = s; | |
183 | flags = EXC_FLAG_MUSTBEDIR; | |
184 | } else { | |
185 | x = xmalloc(sizeof(*x)); | |
186 | x->pattern = string; | |
187 | } | |
188 | x->to_exclude = to_exclude; | |
68492fc7 | 189 | x->patternlen = strlen(string); |
453ec4bd LT |
190 | x->base = base; |
191 | x->baselen = baselen; | |
d6b8fc30 | 192 | x->flags = flags; |
68492fc7 LK |
193 | if (!strchr(string, '/')) |
194 | x->flags |= EXC_FLAG_NODIR; | |
195 | if (no_wildcard(string)) | |
196 | x->flags |= EXC_FLAG_NOWILDCARD; | |
197 | if (*string == '*' && no_wildcard(string+1)) | |
198 | x->flags |= EXC_FLAG_ENDSWITH; | |
686a4a06 | 199 | ALLOC_GROW(which->excludes, which->nr + 1, which->alloc); |
453ec4bd LT |
200 | which->excludes[which->nr++] = x; |
201 | } | |
202 | ||
203 | static int add_excludes_from_file_1(const char *fname, | |
204 | const char *base, | |
205 | int baselen, | |
63d285c8 | 206 | char **buf_p, |
453ec4bd LT |
207 | struct exclude_list *which) |
208 | { | |
c470701a | 209 | struct stat st; |
453ec4bd | 210 | int fd, i; |
dc49cd76 | 211 | size_t size; |
453ec4bd LT |
212 | char *buf, *entry; |
213 | ||
214 | fd = open(fname, O_RDONLY); | |
c470701a | 215 | if (fd < 0 || fstat(fd, &st) < 0) |
453ec4bd | 216 | goto err; |
dc49cd76 | 217 | size = xsize_t(st.st_size); |
453ec4bd LT |
218 | if (size == 0) { |
219 | close(fd); | |
220 | return 0; | |
221 | } | |
222 | buf = xmalloc(size+1); | |
93d26e4c | 223 | if (read_in_full(fd, buf, size) != size) |
6ba78238 LH |
224 | { |
225 | free(buf); | |
453ec4bd | 226 | goto err; |
6ba78238 | 227 | } |
453ec4bd LT |
228 | close(fd); |
229 | ||
63d285c8 JH |
230 | if (buf_p) |
231 | *buf_p = buf; | |
453ec4bd | 232 | entry = buf; |
b5041c5f NTND |
233 | for (i = 0; i <= size; i++) { |
234 | if (i == size || buf[i] == '\n') { | |
453ec4bd LT |
235 | if (entry != buf + i && entry[0] != '#') { |
236 | buf[i - (i && buf[i-1] == '\r')] = 0; | |
237 | add_exclude(entry, base, baselen, which); | |
238 | } | |
239 | entry = buf + i + 1; | |
240 | } | |
241 | } | |
242 | return 0; | |
243 | ||
244 | err: | |
245 | if (0 <= fd) | |
246 | close(fd); | |
247 | return -1; | |
248 | } | |
249 | ||
250 | void add_excludes_from_file(struct dir_struct *dir, const char *fname) | |
251 | { | |
63d285c8 | 252 | if (add_excludes_from_file_1(fname, "", 0, NULL, |
453ec4bd LT |
253 | &dir->exclude_list[EXC_FILE]) < 0) |
254 | die("cannot use %s as an exclude file", fname); | |
255 | } | |
256 | ||
63d285c8 | 257 | static void prep_exclude(struct dir_struct *dir, const char *base, int baselen) |
453ec4bd | 258 | { |
63d285c8 JH |
259 | struct exclude_list *el; |
260 | struct exclude_stack *stk = NULL; | |
261 | int current; | |
262 | ||
263 | if ((!dir->exclude_per_dir) || | |
264 | (baselen + strlen(dir->exclude_per_dir) >= PATH_MAX)) | |
265 | return; /* too long a path -- ignore */ | |
266 | ||
267 | /* Pop the ones that are not the prefix of the path being checked. */ | |
268 | el = &dir->exclude_list[EXC_DIRS]; | |
269 | while ((stk = dir->exclude_stack) != NULL) { | |
270 | if (stk->baselen <= baselen && | |
271 | !strncmp(dir->basebuf, base, stk->baselen)) | |
272 | break; | |
273 | dir->exclude_stack = stk->prev; | |
274 | while (stk->exclude_ix < el->nr) | |
275 | free(el->excludes[--el->nr]); | |
276 | free(stk->filebuf); | |
277 | free(stk); | |
453ec4bd | 278 | } |
453ec4bd | 279 | |
63d285c8 JH |
280 | /* Read from the parent directories and push them down. */ |
281 | current = stk ? stk->baselen : -1; | |
282 | while (current < baselen) { | |
283 | struct exclude_stack *stk = xcalloc(1, sizeof(*stk)); | |
284 | const char *cp; | |
453ec4bd | 285 | |
63d285c8 JH |
286 | if (current < 0) { |
287 | cp = base; | |
288 | current = 0; | |
289 | } | |
290 | else { | |
291 | cp = strchr(base + current + 1, '/'); | |
292 | if (!cp) | |
293 | die("oops in prep_exclude"); | |
294 | cp++; | |
295 | } | |
296 | stk->prev = dir->exclude_stack; | |
297 | stk->baselen = cp - base; | |
298 | stk->exclude_ix = el->nr; | |
299 | memcpy(dir->basebuf + current, base + current, | |
300 | stk->baselen - current); | |
301 | strcpy(dir->basebuf + stk->baselen, dir->exclude_per_dir); | |
302 | add_excludes_from_file_1(dir->basebuf, | |
303 | dir->basebuf, stk->baselen, | |
304 | &stk->filebuf, el); | |
305 | dir->exclude_stack = stk; | |
306 | current = stk->baselen; | |
307 | } | |
308 | dir->basebuf[baselen] = '\0'; | |
453ec4bd LT |
309 | } |
310 | ||
2c5b0115 | 311 | /* Scan the list and let the last match determine the fate. |
453ec4bd LT |
312 | * Return 1 for exclude, 0 for include and -1 for undecided. |
313 | */ | |
314 | static int excluded_1(const char *pathname, | |
6831a88a | 315 | int pathlen, const char *basename, int *dtype, |
453ec4bd LT |
316 | struct exclude_list *el) |
317 | { | |
318 | int i; | |
319 | ||
320 | if (el->nr) { | |
321 | for (i = el->nr - 1; 0 <= i; i--) { | |
322 | struct exclude *x = el->excludes[i]; | |
323 | const char *exclude = x->pattern; | |
68492fc7 | 324 | int to_exclude = x->to_exclude; |
453ec4bd | 325 | |
6831a88a JH |
326 | if (x->flags & EXC_FLAG_MUSTBEDIR) { |
327 | if (*dtype == DT_UNKNOWN) | |
caa6b782 | 328 | *dtype = get_dtype(NULL, pathname, pathlen); |
6831a88a JH |
329 | if (*dtype != DT_DIR) |
330 | continue; | |
331 | } | |
d6b8fc30 | 332 | |
68492fc7 | 333 | if (x->flags & EXC_FLAG_NODIR) { |
453ec4bd | 334 | /* match basename */ |
68492fc7 LK |
335 | if (x->flags & EXC_FLAG_NOWILDCARD) { |
336 | if (!strcmp(exclude, basename)) | |
337 | return to_exclude; | |
338 | } else if (x->flags & EXC_FLAG_ENDSWITH) { | |
339 | if (x->patternlen - 1 <= pathlen && | |
340 | !strcmp(exclude + 1, pathname + pathlen - x->patternlen + 1)) | |
341 | return to_exclude; | |
342 | } else { | |
343 | if (fnmatch(exclude, basename, 0) == 0) | |
344 | return to_exclude; | |
345 | } | |
453ec4bd LT |
346 | } |
347 | else { | |
348 | /* match with FNM_PATHNAME: | |
349 | * exclude has base (baselen long) implicitly | |
350 | * in front of it. | |
351 | */ | |
352 | int baselen = x->baselen; | |
353 | if (*exclude == '/') | |
354 | exclude++; | |
355 | ||
356 | if (pathlen < baselen || | |
357 | (baselen && pathname[baselen-1] != '/') || | |
358 | strncmp(pathname, x->base, baselen)) | |
359 | continue; | |
360 | ||
68492fc7 LK |
361 | if (x->flags & EXC_FLAG_NOWILDCARD) { |
362 | if (!strcmp(exclude, pathname + baselen)) | |
363 | return to_exclude; | |
364 | } else { | |
365 | if (fnmatch(exclude, pathname+baselen, | |
366 | FNM_PATHNAME) == 0) | |
367 | return to_exclude; | |
368 | } | |
453ec4bd LT |
369 | } |
370 | } | |
371 | } | |
372 | return -1; /* undecided */ | |
373 | } | |
374 | ||
6831a88a | 375 | int excluded(struct dir_struct *dir, const char *pathname, int *dtype_p) |
453ec4bd LT |
376 | { |
377 | int pathlen = strlen(pathname); | |
378 | int st; | |
68492fc7 LK |
379 | const char *basename = strrchr(pathname, '/'); |
380 | basename = (basename) ? basename+1 : pathname; | |
453ec4bd | 381 | |
63d285c8 | 382 | prep_exclude(dir, pathname, basename-pathname); |
453ec4bd | 383 | for (st = EXC_CMDL; st <= EXC_FILE; st++) { |
d6b8fc30 | 384 | switch (excluded_1(pathname, pathlen, basename, |
6831a88a | 385 | dtype_p, &dir->exclude_list[st])) { |
453ec4bd LT |
386 | case 0: |
387 | return 0; | |
388 | case 1: | |
389 | return 1; | |
390 | } | |
391 | } | |
392 | return 0; | |
393 | } | |
394 | ||
f3fa1838 JH |
395 | static struct dir_entry *dir_entry_new(const char *pathname, int len) |
396 | { | |
453ec4bd LT |
397 | struct dir_entry *ent; |
398 | ||
453ec4bd LT |
399 | ent = xmalloc(sizeof(*ent) + len + 1); |
400 | ent->len = len; | |
401 | memcpy(ent->name, pathname, len); | |
402 | ent->name[len] = 0; | |
4d06f8ac | 403 | return ent; |
453ec4bd LT |
404 | } |
405 | ||
159b3212 | 406 | static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len) |
6815e569 | 407 | { |
0a9b88b7 | 408 | if (cache_name_exists(pathname, len, ignore_case)) |
6815e569 JK |
409 | return NULL; |
410 | ||
25fd2f7a | 411 | ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc); |
6815e569 JK |
412 | return dir->entries[dir->nr++] = dir_entry_new(pathname, len); |
413 | } | |
414 | ||
159b3212 | 415 | static struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len) |
2abd31b0 | 416 | { |
6e4f981f | 417 | if (!cache_name_is_other(pathname, len)) |
2abd31b0 JK |
418 | return NULL; |
419 | ||
25fd2f7a | 420 | ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc); |
2abd31b0 JK |
421 | return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len); |
422 | } | |
423 | ||
09595258 LT |
424 | enum exist_status { |
425 | index_nonexistent = 0, | |
426 | index_directory, | |
427 | index_gitdir, | |
428 | }; | |
429 | ||
430 | /* | |
431 | * The index sorts alphabetically by entry name, which | |
432 | * means that a gitlink sorts as '\0' at the end, while | |
433 | * a directory (which is defined not as an entry, but as | |
434 | * the files it contains) will sort with the '/' at the | |
435 | * end. | |
436 | */ | |
437 | static enum exist_status directory_exists_in_index(const char *dirname, int len) | |
453ec4bd LT |
438 | { |
439 | int pos = cache_name_pos(dirname, len); | |
09595258 LT |
440 | if (pos < 0) |
441 | pos = -pos-1; | |
442 | while (pos < active_nr) { | |
443 | struct cache_entry *ce = active_cache[pos++]; | |
444 | unsigned char endchar; | |
445 | ||
446 | if (strncmp(ce->name, dirname, len)) | |
447 | break; | |
448 | endchar = ce->name[len]; | |
449 | if (endchar > '/') | |
450 | break; | |
451 | if (endchar == '/') | |
452 | return index_directory; | |
7a51ed66 | 453 | if (!endchar && S_ISGITLINK(ce->ce_mode)) |
09595258 LT |
454 | return index_gitdir; |
455 | } | |
456 | return index_nonexistent; | |
457 | } | |
458 | ||
459 | /* | |
460 | * When we find a directory when traversing the filesystem, we | |
461 | * have three distinct cases: | |
462 | * | |
463 | * - ignore it | |
464 | * - see it as a directory | |
465 | * - recurse into it | |
466 | * | |
467 | * and which one we choose depends on a combination of existing | |
468 | * git index contents and the flags passed into the directory | |
469 | * traversal routine. | |
470 | * | |
471 | * Case 1: If we *already* have entries in the index under that | |
472 | * directory name, we always recurse into the directory to see | |
473 | * all the files. | |
474 | * | |
475 | * Case 2: If we *already* have that directory name as a gitlink, | |
476 | * we always continue to see it as a gitlink, regardless of whether | |
477 | * there is an actual git directory there or not (it might not | |
478 | * be checked out as a subproject!) | |
479 | * | |
480 | * Case 3: if we didn't have it in the index previously, we | |
481 | * have a few sub-cases: | |
482 | * | |
483 | * (a) if "show_other_directories" is true, we show it as | |
484 | * just a directory, unless "hide_empty_directories" is | |
485 | * also true and the directory is empty, in which case | |
486 | * we just ignore it entirely. | |
487 | * (b) if it looks like a git directory, and we don't have | |
302b9282 | 488 | * 'no_gitlinks' set we treat it as a gitlink, and show it |
09595258 LT |
489 | * as a directory. |
490 | * (c) otherwise, we recurse into it. | |
491 | */ | |
492 | enum directory_treatment { | |
493 | show_directory, | |
494 | ignore_directory, | |
495 | recurse_into_directory, | |
496 | }; | |
497 | ||
498 | static enum directory_treatment treat_directory(struct dir_struct *dir, | |
499 | const char *dirname, int len, | |
500 | const struct path_simplify *simplify) | |
501 | { | |
502 | /* The "len-1" is to strip the final '/' */ | |
503 | switch (directory_exists_in_index(dirname, len-1)) { | |
504 | case index_directory: | |
505 | return recurse_into_directory; | |
506 | ||
507 | case index_gitdir: | |
7c4c97c0 | 508 | if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES) |
ab22aed3 | 509 | return ignore_directory; |
09595258 LT |
510 | return show_directory; |
511 | ||
512 | case index_nonexistent: | |
7c4c97c0 | 513 | if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES) |
09595258 | 514 | break; |
7c4c97c0 | 515 | if (!(dir->flags & DIR_NO_GITLINKS)) { |
09595258 LT |
516 | unsigned char sha1[20]; |
517 | if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0) | |
518 | return show_directory; | |
519 | } | |
520 | return recurse_into_directory; | |
521 | } | |
522 | ||
523 | /* This is the "show_other_directories" case */ | |
7c4c97c0 | 524 | if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES)) |
09595258 | 525 | return show_directory; |
dba2e203 | 526 | if (!read_directory_recursive(dir, dirname, len, 1, simplify)) |
09595258 LT |
527 | return ignore_directory; |
528 | return show_directory; | |
453ec4bd LT |
529 | } |
530 | ||
9fc42d60 LT |
531 | /* |
532 | * This is an inexact early pruning of any recursive directory | |
533 | * reading - if the path cannot possibly be in the pathspec, | |
534 | * return true, and we'll skip it early. | |
535 | */ | |
536 | static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify) | |
537 | { | |
538 | if (simplify) { | |
539 | for (;;) { | |
540 | const char *match = simplify->path; | |
541 | int len = simplify->len; | |
542 | ||
543 | if (!match) | |
544 | break; | |
545 | if (len > pathlen) | |
546 | len = pathlen; | |
547 | if (!memcmp(path, match, len)) | |
548 | return 0; | |
549 | simplify++; | |
550 | } | |
551 | return 1; | |
552 | } | |
553 | return 0; | |
554 | } | |
555 | ||
e96980ef JK |
556 | static int in_pathspec(const char *path, int len, const struct path_simplify *simplify) |
557 | { | |
558 | if (simplify) { | |
559 | for (; simplify->path; simplify++) { | |
560 | if (len == simplify->len | |
561 | && !memcmp(path, simplify->path, len)) | |
562 | return 1; | |
563 | } | |
564 | } | |
565 | return 0; | |
566 | } | |
567 | ||
443e061a LT |
568 | static int get_index_dtype(const char *path, int len) |
569 | { | |
570 | int pos; | |
571 | struct cache_entry *ce; | |
572 | ||
573 | ce = cache_name_exists(path, len, 0); | |
574 | if (ce) { | |
575 | if (!ce_uptodate(ce)) | |
576 | return DT_UNKNOWN; | |
577 | if (S_ISGITLINK(ce->ce_mode)) | |
578 | return DT_DIR; | |
579 | /* | |
580 | * Nobody actually cares about the | |
581 | * difference between DT_LNK and DT_REG | |
582 | */ | |
583 | return DT_REG; | |
584 | } | |
585 | ||
586 | /* Try to look it up as a directory */ | |
587 | pos = cache_name_pos(path, len); | |
588 | if (pos >= 0) | |
589 | return DT_UNKNOWN; | |
590 | pos = -pos-1; | |
591 | while (pos < active_nr) { | |
592 | ce = active_cache[pos++]; | |
593 | if (strncmp(ce->name, path, len)) | |
594 | break; | |
595 | if (ce->name[len] > '/') | |
596 | break; | |
597 | if (ce->name[len] < '/') | |
598 | continue; | |
599 | if (!ce_uptodate(ce)) | |
600 | break; /* continue? */ | |
601 | return DT_DIR; | |
602 | } | |
603 | return DT_UNKNOWN; | |
604 | } | |
605 | ||
caa6b782 | 606 | static int get_dtype(struct dirent *de, const char *path, int len) |
07134421 | 607 | { |
6831a88a | 608 | int dtype = de ? DTYPE(de) : DT_UNKNOWN; |
07134421 LT |
609 | struct stat st; |
610 | ||
611 | if (dtype != DT_UNKNOWN) | |
612 | return dtype; | |
443e061a LT |
613 | dtype = get_index_dtype(path, len); |
614 | if (dtype != DT_UNKNOWN) | |
615 | return dtype; | |
616 | if (lstat(path, &st)) | |
07134421 LT |
617 | return dtype; |
618 | if (S_ISREG(st.st_mode)) | |
619 | return DT_REG; | |
620 | if (S_ISDIR(st.st_mode)) | |
621 | return DT_DIR; | |
622 | if (S_ISLNK(st.st_mode)) | |
623 | return DT_LNK; | |
624 | return dtype; | |
625 | } | |
626 | ||
453ec4bd LT |
627 | /* |
628 | * Read a directory tree. We currently ignore anything but | |
629 | * directories, regular files and symlinks. That's because git | |
630 | * doesn't handle them at all yet. Maybe that will change some | |
631 | * day. | |
632 | * | |
633 | * Also, we ignore the name ".git" (even if it is not a directory). | |
634 | * That likely will not change. | |
635 | */ | |
dba2e203 | 636 | static int read_directory_recursive(struct dir_struct *dir, const char *base, int baselen, int check_only, const struct path_simplify *simplify) |
453ec4bd | 637 | { |
dba2e203 | 638 | DIR *fdir = opendir(*base ? base : "."); |
453ec4bd LT |
639 | int contents = 0; |
640 | ||
641 | if (fdir) { | |
453ec4bd | 642 | struct dirent *de; |
dba2e203 LT |
643 | char path[PATH_MAX + 1]; |
644 | memcpy(path, base, baselen); | |
453ec4bd | 645 | |
453ec4bd | 646 | while ((de = readdir(fdir)) != NULL) { |
07134421 | 647 | int len, dtype; |
b9916256 | 648 | int exclude; |
453ec4bd | 649 | |
8ca12c0d AP |
650 | if (is_dot_or_dotdot(de->d_name) || |
651 | !strcmp(de->d_name, ".git")) | |
453ec4bd LT |
652 | continue; |
653 | len = strlen(de->d_name); | |
5d5cea67 | 654 | /* Ignore overly long pathnames! */ |
dba2e203 | 655 | if (len + baselen + 8 > sizeof(path)) |
5d5cea67 | 656 | continue; |
dba2e203 LT |
657 | memcpy(path + baselen, de->d_name, len+1); |
658 | len = baselen + len; | |
659 | if (simplify_away(path, len, simplify)) | |
9fc42d60 | 660 | continue; |
b9916256 | 661 | |
6831a88a | 662 | dtype = DTYPE(de); |
dba2e203 | 663 | exclude = excluded(dir, path, &dtype); |
7c4c97c0 | 664 | if (exclude && (dir->flags & DIR_COLLECT_IGNORED) |
dba2e203 LT |
665 | && in_pathspec(path, len, simplify)) |
666 | dir_add_ignored(dir, path,len); | |
07134421 LT |
667 | |
668 | /* | |
669 | * Excluded? If we don't explicitly want to show | |
670 | * ignored files, ignore it | |
671 | */ | |
7c4c97c0 | 672 | if (exclude && !(dir->flags & DIR_SHOW_IGNORED)) |
07134421 LT |
673 | continue; |
674 | ||
6831a88a | 675 | if (dtype == DT_UNKNOWN) |
caa6b782 | 676 | dtype = get_dtype(de, path, len); |
07134421 LT |
677 | |
678 | /* | |
679 | * Do we want to see just the ignored files? | |
680 | * We still need to recurse into directories, | |
681 | * even if we don't ignore them, since the | |
682 | * directory may contain files that we do.. | |
683 | */ | |
7c4c97c0 | 684 | if (!exclude && (dir->flags & DIR_SHOW_IGNORED)) { |
07134421 | 685 | if (dtype != DT_DIR) |
c889763b | 686 | continue; |
c889763b | 687 | } |
453ec4bd | 688 | |
07134421 | 689 | switch (dtype) { |
453ec4bd LT |
690 | default: |
691 | continue; | |
453ec4bd | 692 | case DT_DIR: |
dba2e203 | 693 | memcpy(path + len, "/", 2); |
453ec4bd | 694 | len++; |
dba2e203 | 695 | switch (treat_directory(dir, path, len, simplify)) { |
09595258 | 696 | case show_directory: |
7c4c97c0 JS |
697 | if (exclude != !!(dir->flags |
698 | & DIR_SHOW_IGNORED)) | |
b9916256 | 699 | continue; |
453ec4bd | 700 | break; |
09595258 LT |
701 | case recurse_into_directory: |
702 | contents += read_directory_recursive(dir, | |
dba2e203 | 703 | path, len, 0, simplify); |
09595258 LT |
704 | continue; |
705 | case ignore_directory: | |
706 | continue; | |
453ec4bd | 707 | } |
09595258 | 708 | break; |
453ec4bd LT |
709 | case DT_REG: |
710 | case DT_LNK: | |
711 | break; | |
712 | } | |
453ec4bd | 713 | contents++; |
07ccbff8 JS |
714 | if (check_only) |
715 | goto exit_early; | |
716 | else | |
dba2e203 | 717 | dir_add_name(dir, path, len); |
453ec4bd | 718 | } |
07ccbff8 | 719 | exit_early: |
453ec4bd | 720 | closedir(fdir); |
453ec4bd LT |
721 | } |
722 | ||
723 | return contents; | |
724 | } | |
725 | ||
726 | static int cmp_name(const void *p1, const void *p2) | |
727 | { | |
728 | const struct dir_entry *e1 = *(const struct dir_entry **)p1; | |
729 | const struct dir_entry *e2 = *(const struct dir_entry **)p2; | |
730 | ||
731 | return cache_name_compare(e1->name, e1->len, | |
732 | e2->name, e2->len); | |
733 | } | |
734 | ||
9fc42d60 LT |
735 | /* |
736 | * Return the length of the "simple" part of a path match limiter. | |
737 | */ | |
738 | static int simple_length(const char *match) | |
453ec4bd | 739 | { |
9fc42d60 LT |
740 | int len = -1; |
741 | ||
742 | for (;;) { | |
743 | unsigned char c = *match++; | |
744 | len++; | |
8cc32992 | 745 | if (c == '\0' || is_glob_special(c)) |
9fc42d60 LT |
746 | return len; |
747 | } | |
748 | } | |
749 | ||
750 | static struct path_simplify *create_simplify(const char **pathspec) | |
751 | { | |
752 | int nr, alloc = 0; | |
753 | struct path_simplify *simplify = NULL; | |
754 | ||
755 | if (!pathspec) | |
756 | return NULL; | |
757 | ||
758 | for (nr = 0 ; ; nr++) { | |
759 | const char *match; | |
760 | if (nr >= alloc) { | |
761 | alloc = alloc_nr(alloc); | |
762 | simplify = xrealloc(simplify, alloc * sizeof(*simplify)); | |
763 | } | |
764 | match = *pathspec++; | |
765 | if (!match) | |
766 | break; | |
767 | simplify[nr].path = match; | |
768 | simplify[nr].len = simple_length(match); | |
769 | } | |
770 | simplify[nr].path = NULL; | |
771 | simplify[nr].len = 0; | |
772 | return simplify; | |
773 | } | |
774 | ||
775 | static void free_simplify(struct path_simplify *simplify) | |
776 | { | |
8e0f7003 | 777 | free(simplify); |
9fc42d60 LT |
778 | } |
779 | ||
dba2e203 | 780 | int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec) |
9fc42d60 | 781 | { |
725b0605 | 782 | struct path_simplify *simplify; |
b4189aa8 | 783 | |
dba2e203 | 784 | if (has_symlink_leading_path(path, len)) |
725b0605 JH |
785 | return dir->nr; |
786 | ||
787 | simplify = create_simplify(pathspec); | |
dba2e203 | 788 | read_directory_recursive(dir, path, len, 0, simplify); |
9fc42d60 | 789 | free_simplify(simplify); |
453ec4bd | 790 | qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name); |
2abd31b0 | 791 | qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name); |
453ec4bd LT |
792 | return dir->nr; |
793 | } | |
c91f0d92 | 794 | |
686a4a06 | 795 | int file_exists(const char *f) |
c91f0d92 | 796 | { |
686a4a06 | 797 | struct stat sb; |
a50f9fc5 | 798 | return lstat(f, &sb) == 0; |
c91f0d92 | 799 | } |
e6636747 JS |
800 | |
801 | /* | |
802 | * get_relative_cwd() gets the prefix of the current working directory | |
803 | * relative to 'dir'. If we are not inside 'dir', it returns NULL. | |
420acb31 JS |
804 | * |
805 | * As a convenience, it also returns NULL if 'dir' is already NULL. The | |
806 | * reason for this behaviour is that it is natural for functions returning | |
807 | * directory names to return NULL to say "this directory does not exist" | |
808 | * or "this directory is invalid". These cases are usually handled the | |
809 | * same as if the cwd is not inside 'dir' at all, so get_relative_cwd() | |
810 | * returns NULL for both of them. | |
811 | * | |
812 | * Most notably, get_relative_cwd(buffer, size, get_git_work_tree()) | |
813 | * unifies the handling of "outside work tree" with "no work tree at all". | |
e6636747 JS |
814 | */ |
815 | char *get_relative_cwd(char *buffer, int size, const char *dir) | |
816 | { | |
817 | char *cwd = buffer; | |
818 | ||
e6636747 JS |
819 | if (!dir) |
820 | return NULL; | |
821 | if (!getcwd(buffer, size)) | |
d824cbba | 822 | die_errno("can't find the current directory"); |
e6636747 JS |
823 | |
824 | if (!is_absolute_path(dir)) | |
825 | dir = make_absolute_path(dir); | |
826 | ||
827 | while (*dir && *dir == *cwd) { | |
828 | dir++; | |
829 | cwd++; | |
830 | } | |
831 | if (*dir) | |
832 | return NULL; | |
833 | if (*cwd == '/') | |
834 | return cwd + 1; | |
835 | return cwd; | |
836 | } | |
837 | ||
838 | int is_inside_dir(const char *dir) | |
839 | { | |
840 | char buffer[PATH_MAX]; | |
841 | return get_relative_cwd(buffer, sizeof(buffer), dir) != NULL; | |
842 | } | |
7155b727 | 843 | |
55892d23 AP |
844 | int is_empty_dir(const char *path) |
845 | { | |
846 | DIR *dir = opendir(path); | |
847 | struct dirent *e; | |
848 | int ret = 1; | |
849 | ||
850 | if (!dir) | |
851 | return 0; | |
852 | ||
853 | while ((e = readdir(dir)) != NULL) | |
854 | if (!is_dot_or_dotdot(e->d_name)) { | |
855 | ret = 0; | |
856 | break; | |
857 | } | |
858 | ||
859 | closedir(dir); | |
860 | return ret; | |
861 | } | |
862 | ||
7155b727 JS |
863 | int remove_dir_recursively(struct strbuf *path, int only_empty) |
864 | { | |
865 | DIR *dir = opendir(path->buf); | |
866 | struct dirent *e; | |
867 | int ret = 0, original_len = path->len, len; | |
868 | ||
869 | if (!dir) | |
870 | return -1; | |
871 | if (path->buf[original_len - 1] != '/') | |
872 | strbuf_addch(path, '/'); | |
873 | ||
874 | len = path->len; | |
875 | while ((e = readdir(dir)) != NULL) { | |
876 | struct stat st; | |
8ca12c0d AP |
877 | if (is_dot_or_dotdot(e->d_name)) |
878 | continue; | |
7155b727 JS |
879 | |
880 | strbuf_setlen(path, len); | |
881 | strbuf_addstr(path, e->d_name); | |
882 | if (lstat(path->buf, &st)) | |
883 | ; /* fall thru */ | |
884 | else if (S_ISDIR(st.st_mode)) { | |
885 | if (!remove_dir_recursively(path, only_empty)) | |
886 | continue; /* happy */ | |
887 | } else if (!only_empty && !unlink(path->buf)) | |
888 | continue; /* happy, too */ | |
889 | ||
890 | /* path too long, stat fails, or non-directory still exists */ | |
891 | ret = -1; | |
892 | break; | |
893 | } | |
894 | closedir(dir); | |
895 | ||
896 | strbuf_setlen(path, original_len); | |
897 | if (!ret) | |
898 | ret = rmdir(path->buf); | |
899 | return ret; | |
900 | } | |
039bc64e JH |
901 | |
902 | void setup_standard_excludes(struct dir_struct *dir) | |
903 | { | |
904 | const char *path; | |
905 | ||
906 | dir->exclude_per_dir = ".gitignore"; | |
907 | path = git_path("info/exclude"); | |
908 | if (!access(path, R_OK)) | |
909 | add_excludes_from_file(dir, path); | |
910 | if (excludes_file && !access(excludes_file, R_OK)) | |
911 | add_excludes_from_file(dir, excludes_file); | |
912 | } | |
4a92d1bf AR |
913 | |
914 | int remove_path(const char *name) | |
915 | { | |
916 | char *slash; | |
917 | ||
918 | if (unlink(name) && errno != ENOENT) | |
919 | return -1; | |
920 | ||
921 | slash = strrchr(name, '/'); | |
922 | if (slash) { | |
923 | char *dirs = xstrdup(name); | |
924 | slash = dirs + (slash - name); | |
925 | do { | |
926 | *slash = '\0'; | |
927 | } while (rmdir(dirs) && (slash = strrchr(dirs, '/'))); | |
928 | free(dirs); | |
929 | } | |
930 | return 0; | |
931 | } | |
932 |