]>
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 | ||
09595258 LT |
17 | static int read_directory_recursive(struct dir_struct *dir, |
18 | const char *path, const char *base, int baselen, | |
19 | int check_only, const struct path_simplify *simplify); | |
20 | ||
3c6a370b LT |
21 | int common_prefix(const char **pathspec) |
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 | ||
e813d50e JH |
54 | /* |
55 | * Does 'match' matches the given name? | |
56 | * A match is found if | |
57 | * | |
58 | * (1) the 'match' string is leading directory of 'name', or | |
59 | * (2) the 'match' string is a wildcard and matches 'name', or | |
60 | * (3) the 'match' string is exactly the same as 'name'. | |
61 | * | |
62 | * and the return value tells which case it was. | |
63 | * | |
64 | * It returns 0 when there is no match. | |
65 | */ | |
3c6a370b LT |
66 | static int match_one(const char *match, const char *name, int namelen) |
67 | { | |
68 | int matchlen; | |
69 | ||
70 | /* If the match was just the prefix, we matched */ | |
71 | matchlen = strlen(match); | |
72 | if (!matchlen) | |
e813d50e | 73 | return MATCHED_RECURSIVELY; |
3c6a370b LT |
74 | |
75 | /* | |
76 | * If we don't match the matchstring exactly, | |
77 | * we need to match by fnmatch | |
78 | */ | |
79 | if (strncmp(match, name, matchlen)) | |
e813d50e | 80 | return !fnmatch(match, name, 0) ? MATCHED_FNMATCH : 0; |
3c6a370b | 81 | |
e813d50e JH |
82 | if (!name[matchlen]) |
83 | return MATCHED_EXACTLY; | |
84 | if (match[matchlen-1] == '/' || name[matchlen] == '/') | |
85 | return MATCHED_RECURSIVELY; | |
86 | return 0; | |
3c6a370b LT |
87 | } |
88 | ||
e813d50e JH |
89 | /* |
90 | * Given a name and a list of pathspecs, see if the name matches | |
91 | * any of the pathspecs. The caller is also interested in seeing | |
92 | * all pathspec matches some names it calls this function with | |
93 | * (otherwise the user could have mistyped the unmatched pathspec), | |
94 | * and a mark is left in seen[] array for pathspec element that | |
95 | * actually matched anything. | |
96 | */ | |
3c6a370b LT |
97 | int match_pathspec(const char **pathspec, const char *name, int namelen, int prefix, char *seen) |
98 | { | |
99 | int retval; | |
100 | const char *match; | |
101 | ||
102 | name += prefix; | |
103 | namelen -= prefix; | |
104 | ||
105 | for (retval = 0; (match = *pathspec++) != NULL; seen++) { | |
e813d50e JH |
106 | int how; |
107 | if (retval && *seen == MATCHED_EXACTLY) | |
3c6a370b LT |
108 | continue; |
109 | match += prefix; | |
e813d50e JH |
110 | how = match_one(match, name, namelen); |
111 | if (how) { | |
112 | if (retval < how) | |
113 | retval = how; | |
114 | if (*seen < how) | |
115 | *seen = how; | |
3c6a370b LT |
116 | } |
117 | } | |
118 | return retval; | |
119 | } | |
120 | ||
453ec4bd LT |
121 | void add_exclude(const char *string, const char *base, |
122 | int baselen, struct exclude_list *which) | |
123 | { | |
124 | struct exclude *x = xmalloc(sizeof (*x)); | |
125 | ||
126 | x->pattern = string; | |
127 | x->base = base; | |
128 | x->baselen = baselen; | |
129 | if (which->nr == which->alloc) { | |
130 | which->alloc = alloc_nr(which->alloc); | |
83572c1a JF |
131 | which->excludes = xrealloc(which->excludes, |
132 | which->alloc * sizeof(x)); | |
453ec4bd LT |
133 | } |
134 | which->excludes[which->nr++] = x; | |
135 | } | |
136 | ||
137 | static int add_excludes_from_file_1(const char *fname, | |
138 | const char *base, | |
139 | int baselen, | |
140 | struct exclude_list *which) | |
141 | { | |
c470701a | 142 | struct stat st; |
453ec4bd | 143 | int fd, i; |
dc49cd76 | 144 | size_t size; |
453ec4bd LT |
145 | char *buf, *entry; |
146 | ||
147 | fd = open(fname, O_RDONLY); | |
c470701a | 148 | if (fd < 0 || fstat(fd, &st) < 0) |
453ec4bd | 149 | goto err; |
dc49cd76 | 150 | size = xsize_t(st.st_size); |
453ec4bd LT |
151 | if (size == 0) { |
152 | close(fd); | |
153 | return 0; | |
154 | } | |
155 | buf = xmalloc(size+1); | |
93d26e4c | 156 | if (read_in_full(fd, buf, size) != size) |
453ec4bd LT |
157 | goto err; |
158 | close(fd); | |
159 | ||
160 | buf[size++] = '\n'; | |
161 | entry = buf; | |
162 | for (i = 0; i < size; i++) { | |
163 | if (buf[i] == '\n') { | |
164 | if (entry != buf + i && entry[0] != '#') { | |
165 | buf[i - (i && buf[i-1] == '\r')] = 0; | |
166 | add_exclude(entry, base, baselen, which); | |
167 | } | |
168 | entry = buf + i + 1; | |
169 | } | |
170 | } | |
171 | return 0; | |
172 | ||
173 | err: | |
174 | if (0 <= fd) | |
175 | close(fd); | |
176 | return -1; | |
177 | } | |
178 | ||
179 | void add_excludes_from_file(struct dir_struct *dir, const char *fname) | |
180 | { | |
181 | if (add_excludes_from_file_1(fname, "", 0, | |
182 | &dir->exclude_list[EXC_FILE]) < 0) | |
183 | die("cannot use %s as an exclude file", fname); | |
184 | } | |
185 | ||
f8a9d428 | 186 | int push_exclude_per_directory(struct dir_struct *dir, const char *base, int baselen) |
453ec4bd LT |
187 | { |
188 | char exclude_file[PATH_MAX]; | |
189 | struct exclude_list *el = &dir->exclude_list[EXC_DIRS]; | |
190 | int current_nr = el->nr; | |
191 | ||
192 | if (dir->exclude_per_dir) { | |
193 | memcpy(exclude_file, base, baselen); | |
194 | strcpy(exclude_file + baselen, dir->exclude_per_dir); | |
195 | add_excludes_from_file_1(exclude_file, base, baselen, el); | |
196 | } | |
197 | return current_nr; | |
198 | } | |
199 | ||
f8a9d428 | 200 | void pop_exclude_per_directory(struct dir_struct *dir, int stk) |
453ec4bd LT |
201 | { |
202 | struct exclude_list *el = &dir->exclude_list[EXC_DIRS]; | |
203 | ||
204 | while (stk < el->nr) | |
205 | free(el->excludes[--el->nr]); | |
206 | } | |
207 | ||
208 | /* Scan the list and let the last match determines the fate. | |
209 | * Return 1 for exclude, 0 for include and -1 for undecided. | |
210 | */ | |
211 | static int excluded_1(const char *pathname, | |
212 | int pathlen, | |
213 | struct exclude_list *el) | |
214 | { | |
215 | int i; | |
216 | ||
217 | if (el->nr) { | |
218 | for (i = el->nr - 1; 0 <= i; i--) { | |
219 | struct exclude *x = el->excludes[i]; | |
220 | const char *exclude = x->pattern; | |
221 | int to_exclude = 1; | |
222 | ||
223 | if (*exclude == '!') { | |
224 | to_exclude = 0; | |
225 | exclude++; | |
226 | } | |
227 | ||
228 | if (!strchr(exclude, '/')) { | |
229 | /* match basename */ | |
230 | const char *basename = strrchr(pathname, '/'); | |
231 | basename = (basename) ? basename+1 : pathname; | |
232 | if (fnmatch(exclude, basename, 0) == 0) | |
233 | return to_exclude; | |
234 | } | |
235 | else { | |
236 | /* match with FNM_PATHNAME: | |
237 | * exclude has base (baselen long) implicitly | |
238 | * in front of it. | |
239 | */ | |
240 | int baselen = x->baselen; | |
241 | if (*exclude == '/') | |
242 | exclude++; | |
243 | ||
244 | if (pathlen < baselen || | |
245 | (baselen && pathname[baselen-1] != '/') || | |
246 | strncmp(pathname, x->base, baselen)) | |
247 | continue; | |
248 | ||
249 | if (fnmatch(exclude, pathname+baselen, | |
250 | FNM_PATHNAME) == 0) | |
251 | return to_exclude; | |
252 | } | |
253 | } | |
254 | } | |
255 | return -1; /* undecided */ | |
256 | } | |
257 | ||
258 | int excluded(struct dir_struct *dir, const char *pathname) | |
259 | { | |
260 | int pathlen = strlen(pathname); | |
261 | int st; | |
262 | ||
263 | for (st = EXC_CMDL; st <= EXC_FILE; st++) { | |
264 | switch (excluded_1(pathname, pathlen, &dir->exclude_list[st])) { | |
265 | case 0: | |
266 | return 0; | |
267 | case 1: | |
268 | return 1; | |
269 | } | |
270 | } | |
271 | return 0; | |
272 | } | |
273 | ||
4d06f8ac | 274 | struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len) |
453ec4bd LT |
275 | { |
276 | struct dir_entry *ent; | |
277 | ||
278 | if (cache_name_pos(pathname, len) >= 0) | |
4d06f8ac | 279 | return NULL; |
453ec4bd LT |
280 | |
281 | if (dir->nr == dir->alloc) { | |
282 | int alloc = alloc_nr(dir->alloc); | |
283 | dir->alloc = alloc; | |
284 | dir->entries = xrealloc(dir->entries, alloc*sizeof(ent)); | |
285 | } | |
286 | ent = xmalloc(sizeof(*ent) + len + 1); | |
4d06f8ac | 287 | ent->ignored = ent->ignored_dir = 0; |
453ec4bd LT |
288 | ent->len = len; |
289 | memcpy(ent->name, pathname, len); | |
290 | ent->name[len] = 0; | |
291 | dir->entries[dir->nr++] = ent; | |
4d06f8ac | 292 | return ent; |
453ec4bd LT |
293 | } |
294 | ||
09595258 LT |
295 | enum exist_status { |
296 | index_nonexistent = 0, | |
297 | index_directory, | |
298 | index_gitdir, | |
299 | }; | |
300 | ||
301 | /* | |
302 | * The index sorts alphabetically by entry name, which | |
303 | * means that a gitlink sorts as '\0' at the end, while | |
304 | * a directory (which is defined not as an entry, but as | |
305 | * the files it contains) will sort with the '/' at the | |
306 | * end. | |
307 | */ | |
308 | static enum exist_status directory_exists_in_index(const char *dirname, int len) | |
453ec4bd LT |
309 | { |
310 | int pos = cache_name_pos(dirname, len); | |
09595258 LT |
311 | if (pos < 0) |
312 | pos = -pos-1; | |
313 | while (pos < active_nr) { | |
314 | struct cache_entry *ce = active_cache[pos++]; | |
315 | unsigned char endchar; | |
316 | ||
317 | if (strncmp(ce->name, dirname, len)) | |
318 | break; | |
319 | endchar = ce->name[len]; | |
320 | if (endchar > '/') | |
321 | break; | |
322 | if (endchar == '/') | |
323 | return index_directory; | |
302b9282 | 324 | if (!endchar && S_ISGITLINK(ntohl(ce->ce_mode))) |
09595258 LT |
325 | return index_gitdir; |
326 | } | |
327 | return index_nonexistent; | |
328 | } | |
329 | ||
330 | /* | |
331 | * When we find a directory when traversing the filesystem, we | |
332 | * have three distinct cases: | |
333 | * | |
334 | * - ignore it | |
335 | * - see it as a directory | |
336 | * - recurse into it | |
337 | * | |
338 | * and which one we choose depends on a combination of existing | |
339 | * git index contents and the flags passed into the directory | |
340 | * traversal routine. | |
341 | * | |
342 | * Case 1: If we *already* have entries in the index under that | |
343 | * directory name, we always recurse into the directory to see | |
344 | * all the files. | |
345 | * | |
346 | * Case 2: If we *already* have that directory name as a gitlink, | |
347 | * we always continue to see it as a gitlink, regardless of whether | |
348 | * there is an actual git directory there or not (it might not | |
349 | * be checked out as a subproject!) | |
350 | * | |
351 | * Case 3: if we didn't have it in the index previously, we | |
352 | * have a few sub-cases: | |
353 | * | |
354 | * (a) if "show_other_directories" is true, we show it as | |
355 | * just a directory, unless "hide_empty_directories" is | |
356 | * also true and the directory is empty, in which case | |
357 | * we just ignore it entirely. | |
358 | * (b) if it looks like a git directory, and we don't have | |
302b9282 | 359 | * 'no_gitlinks' set we treat it as a gitlink, and show it |
09595258 LT |
360 | * as a directory. |
361 | * (c) otherwise, we recurse into it. | |
362 | */ | |
363 | enum directory_treatment { | |
364 | show_directory, | |
365 | ignore_directory, | |
366 | recurse_into_directory, | |
367 | }; | |
368 | ||
369 | static enum directory_treatment treat_directory(struct dir_struct *dir, | |
370 | const char *dirname, int len, | |
371 | const struct path_simplify *simplify) | |
372 | { | |
373 | /* The "len-1" is to strip the final '/' */ | |
374 | switch (directory_exists_in_index(dirname, len-1)) { | |
375 | case index_directory: | |
376 | return recurse_into_directory; | |
377 | ||
378 | case index_gitdir: | |
ab22aed3 LT |
379 | if (dir->show_other_directories) |
380 | return ignore_directory; | |
09595258 LT |
381 | return show_directory; |
382 | ||
383 | case index_nonexistent: | |
384 | if (dir->show_other_directories) | |
385 | break; | |
302b9282 | 386 | if (!dir->no_gitlinks) { |
09595258 LT |
387 | unsigned char sha1[20]; |
388 | if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0) | |
389 | return show_directory; | |
390 | } | |
391 | return recurse_into_directory; | |
392 | } | |
393 | ||
394 | /* This is the "show_other_directories" case */ | |
395 | if (!dir->hide_empty_directories) | |
396 | return show_directory; | |
397 | if (!read_directory_recursive(dir, dirname, dirname, len, 1, simplify)) | |
398 | return ignore_directory; | |
399 | return show_directory; | |
453ec4bd LT |
400 | } |
401 | ||
9fc42d60 LT |
402 | /* |
403 | * This is an inexact early pruning of any recursive directory | |
404 | * reading - if the path cannot possibly be in the pathspec, | |
405 | * return true, and we'll skip it early. | |
406 | */ | |
407 | static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify) | |
408 | { | |
409 | if (simplify) { | |
410 | for (;;) { | |
411 | const char *match = simplify->path; | |
412 | int len = simplify->len; | |
413 | ||
414 | if (!match) | |
415 | break; | |
416 | if (len > pathlen) | |
417 | len = pathlen; | |
418 | if (!memcmp(path, match, len)) | |
419 | return 0; | |
420 | simplify++; | |
421 | } | |
422 | return 1; | |
423 | } | |
424 | return 0; | |
425 | } | |
426 | ||
453ec4bd LT |
427 | /* |
428 | * Read a directory tree. We currently ignore anything but | |
429 | * directories, regular files and symlinks. That's because git | |
430 | * doesn't handle them at all yet. Maybe that will change some | |
431 | * day. | |
432 | * | |
433 | * Also, we ignore the name ".git" (even if it is not a directory). | |
434 | * That likely will not change. | |
435 | */ | |
9fc42d60 | 436 | static int read_directory_recursive(struct dir_struct *dir, const char *path, const char *base, int baselen, int check_only, const struct path_simplify *simplify) |
453ec4bd LT |
437 | { |
438 | DIR *fdir = opendir(path); | |
439 | int contents = 0; | |
440 | ||
441 | if (fdir) { | |
442 | int exclude_stk; | |
443 | struct dirent *de; | |
095c424d | 444 | char fullname[PATH_MAX + 1]; |
453ec4bd LT |
445 | memcpy(fullname, base, baselen); |
446 | ||
447 | exclude_stk = push_exclude_per_directory(dir, base, baselen); | |
448 | ||
449 | while ((de = readdir(fdir)) != NULL) { | |
450 | int len; | |
b9916256 | 451 | int exclude; |
453ec4bd LT |
452 | |
453 | if ((de->d_name[0] == '.') && | |
454 | (de->d_name[1] == 0 || | |
455 | !strcmp(de->d_name + 1, ".") || | |
456 | !strcmp(de->d_name + 1, "git"))) | |
457 | continue; | |
458 | len = strlen(de->d_name); | |
5d5cea67 LT |
459 | /* Ignore overly long pathnames! */ |
460 | if (len + baselen + 8 > sizeof(fullname)) | |
461 | continue; | |
453ec4bd | 462 | memcpy(fullname + baselen, de->d_name, len+1); |
9fc42d60 LT |
463 | if (simplify_away(fullname, baselen + len, simplify)) |
464 | continue; | |
b9916256 MS |
465 | |
466 | exclude = excluded(dir, fullname); | |
467 | if (exclude != dir->show_ignored) { | |
c889763b JH |
468 | if (!dir->show_ignored || DTYPE(de) != DT_DIR) { |
469 | continue; | |
470 | } | |
471 | } | |
453ec4bd LT |
472 | |
473 | switch (DTYPE(de)) { | |
474 | struct stat st; | |
453ec4bd LT |
475 | default: |
476 | continue; | |
477 | case DT_UNKNOWN: | |
478 | if (lstat(fullname, &st)) | |
479 | continue; | |
480 | if (S_ISREG(st.st_mode) || S_ISLNK(st.st_mode)) | |
481 | break; | |
482 | if (!S_ISDIR(st.st_mode)) | |
483 | continue; | |
484 | /* fallthrough */ | |
485 | case DT_DIR: | |
486 | memcpy(fullname + baselen + len, "/", 2); | |
487 | len++; | |
09595258 LT |
488 | switch (treat_directory(dir, fullname, baselen + len, simplify)) { |
489 | case show_directory: | |
b9916256 MS |
490 | if (exclude != dir->show_ignored) |
491 | continue; | |
453ec4bd | 492 | break; |
09595258 LT |
493 | case recurse_into_directory: |
494 | contents += read_directory_recursive(dir, | |
495 | fullname, fullname, baselen + len, 0, simplify); | |
496 | continue; | |
497 | case ignore_directory: | |
498 | continue; | |
453ec4bd | 499 | } |
09595258 | 500 | break; |
453ec4bd LT |
501 | case DT_REG: |
502 | case DT_LNK: | |
503 | break; | |
504 | } | |
453ec4bd | 505 | contents++; |
07ccbff8 JS |
506 | if (check_only) |
507 | goto exit_early; | |
508 | else | |
4d06f8ac | 509 | dir_add_name(dir, fullname, baselen + len); |
453ec4bd | 510 | } |
07ccbff8 | 511 | exit_early: |
453ec4bd LT |
512 | closedir(fdir); |
513 | ||
514 | pop_exclude_per_directory(dir, exclude_stk); | |
515 | } | |
516 | ||
517 | return contents; | |
518 | } | |
519 | ||
520 | static int cmp_name(const void *p1, const void *p2) | |
521 | { | |
522 | const struct dir_entry *e1 = *(const struct dir_entry **)p1; | |
523 | const struct dir_entry *e2 = *(const struct dir_entry **)p2; | |
524 | ||
525 | return cache_name_compare(e1->name, e1->len, | |
526 | e2->name, e2->len); | |
527 | } | |
528 | ||
9fc42d60 LT |
529 | /* |
530 | * Return the length of the "simple" part of a path match limiter. | |
531 | */ | |
532 | static int simple_length(const char *match) | |
453ec4bd | 533 | { |
9fc42d60 LT |
534 | const char special[256] = { |
535 | [0] = 1, ['?'] = 1, | |
536 | ['\\'] = 1, ['*'] = 1, | |
537 | ['['] = 1 | |
538 | }; | |
539 | int len = -1; | |
540 | ||
541 | for (;;) { | |
542 | unsigned char c = *match++; | |
543 | len++; | |
544 | if (special[c]) | |
545 | return len; | |
546 | } | |
547 | } | |
548 | ||
549 | static struct path_simplify *create_simplify(const char **pathspec) | |
550 | { | |
551 | int nr, alloc = 0; | |
552 | struct path_simplify *simplify = NULL; | |
553 | ||
554 | if (!pathspec) | |
555 | return NULL; | |
556 | ||
557 | for (nr = 0 ; ; nr++) { | |
558 | const char *match; | |
559 | if (nr >= alloc) { | |
560 | alloc = alloc_nr(alloc); | |
561 | simplify = xrealloc(simplify, alloc * sizeof(*simplify)); | |
562 | } | |
563 | match = *pathspec++; | |
564 | if (!match) | |
565 | break; | |
566 | simplify[nr].path = match; | |
567 | simplify[nr].len = simple_length(match); | |
568 | } | |
569 | simplify[nr].path = NULL; | |
570 | simplify[nr].len = 0; | |
571 | return simplify; | |
572 | } | |
573 | ||
574 | static void free_simplify(struct path_simplify *simplify) | |
575 | { | |
576 | if (simplify) | |
577 | free(simplify); | |
578 | } | |
579 | ||
580 | int read_directory(struct dir_struct *dir, const char *path, const char *base, int baselen, const char **pathspec) | |
581 | { | |
582 | struct path_simplify *simplify = create_simplify(pathspec); | |
583 | ||
b4189aa8 LT |
584 | /* |
585 | * Make sure to do the per-directory exclude for all the | |
586 | * directories leading up to our base. | |
587 | */ | |
588 | if (baselen) { | |
589 | if (dir->exclude_per_dir) { | |
590 | char *p, *pp = xmalloc(baselen+1); | |
591 | memcpy(pp, base, baselen+1); | |
592 | p = pp; | |
593 | while (1) { | |
594 | char save = *p; | |
595 | *p = 0; | |
596 | push_exclude_per_directory(dir, pp, p-pp); | |
597 | *p++ = save; | |
598 | if (!save) | |
599 | break; | |
600 | p = strchr(p, '/'); | |
601 | if (p) | |
602 | p++; | |
603 | else | |
604 | p = pp + baselen; | |
605 | } | |
606 | free(pp); | |
607 | } | |
608 | } | |
609 | ||
9fc42d60 LT |
610 | read_directory_recursive(dir, path, base, baselen, 0, simplify); |
611 | free_simplify(simplify); | |
453ec4bd LT |
612 | qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name); |
613 | return dir->nr; | |
614 | } | |
c91f0d92 JK |
615 | |
616 | int | |
617 | file_exists(const char *f) | |
618 | { | |
619 | struct stat sb; | |
620 | return stat(f, &sb) == 0; | |
621 | } |