Raw
1 /*
2 * Handle git attributes. See gitattributes(5) for a description of
3 * the file syntax, and attr.h for a description of the API.
4 *
5 * One basic design decision here is that we are not going to support
6 * an insanely large number of attributes.
7 */
8
9 #define USE_THE_REPOSITORY_VARIABLE
10 #define DISABLE_SIGN_COMPARE_WARNINGS
11
12 #include "git-compat-util.h"
13 #include "config.h"
14 #include "environment.h"
15 #include "exec-cmd.h"
16 #include "attr.h"
17 #include "dir.h"
18 #include "gettext.h"
19 #include "path.h"
20 #include "utf8.h"
21 #include "quote.h"
22 #include "read-cache-ll.h"
23 #include "refs.h"
24 #include "revision.h"
25 #include "odb.h"
26 #include "setup.h"
27 #include "thread-utils.h"
28 #include "tree-walk.h"
29 #include "object-name.h"
30
31 char *git_attr_tree;
32
33 const char git_attr__true[] = "(builtin)true";
34 const char git_attr__false[] = "\0(builtin)false";
35 static const char git_attr__unknown[] = "(builtin)unknown";
36 #define ATTR__TRUE git_attr__true
37 #define ATTR__FALSE git_attr__false
38 #define ATTR__UNSET NULL
39 #define ATTR__UNKNOWN git_attr__unknown
40
41 struct git_attr {
42 unsigned int attr_nr; /* unique attribute number */
43 char name[FLEX_ARRAY]; /* attribute name */
44 };
45
46 const char *git_attr_name(const struct git_attr *attr)
47 {
48 return attr->name;
49 }
50
51 struct attr_hashmap {
52 struct hashmap map;
53 pthread_mutex_t mutex;
54 };
55
56 static inline void hashmap_lock(struct attr_hashmap *map)
57 {
58 pthread_mutex_lock(&map->mutex);
59 }
60
61 static inline void hashmap_unlock(struct attr_hashmap *map)
62 {
63 pthread_mutex_unlock(&map->mutex);
64 }
65
66 /* The container for objects stored in "struct attr_hashmap" */
67 struct attr_hash_entry {
68 struct hashmap_entry ent;
69 const char *key; /* the key; memory should be owned by value */
70 size_t keylen; /* length of the key */
71 void *value; /* the stored value */
72 };
73
74 /* attr_hashmap comparison function */
75 static int attr_hash_entry_cmp(const void *cmp_data UNUSED,
76 const struct hashmap_entry *eptr,
77 const struct hashmap_entry *entry_or_key,
78 const void *keydata UNUSED)
79 {
80 const struct attr_hash_entry *a, *b;
81
82 a = container_of(eptr, const struct attr_hash_entry, ent);
83 b = container_of(entry_or_key, const struct attr_hash_entry, ent);
84 return (a->keylen != b->keylen) || strncmp(a->key, b->key, a->keylen);
85 }
86
87 /*
88 * The global dictionary of all interned attributes. This
89 * is a singleton object which is shared between threads.
90 * Access to this dictionary must be surrounded with a mutex.
91 */
92 static struct attr_hashmap g_attr_hashmap = {
93 .map = HASHMAP_INIT(attr_hash_entry_cmp, NULL),
94 };
95
96 /*
97 * Retrieve the 'value' stored in a hashmap given the provided 'key'.
98 * If there is no matching entry, return NULL.
99 */
100 static void *attr_hashmap_get(struct attr_hashmap *map,
101 const char *key, size_t keylen)
102 {
103 struct attr_hash_entry k;
104 struct attr_hash_entry *e;
105
106 hashmap_entry_init(&k.ent, memhash(key, keylen));
107 k.key = key;
108 k.keylen = keylen;
109 e = hashmap_get_entry(&map->map, &k, ent, NULL);
110
111 return e ? e->value : NULL;
112 }
113
114 /* Add 'value' to a hashmap based on the provided 'key'. */
115 static void attr_hashmap_add(struct attr_hashmap *map,
116 const char *key, size_t keylen,
117 void *value)
118 {
119 struct attr_hash_entry *e;
120
121 e = xmalloc(sizeof(struct attr_hash_entry));
122 hashmap_entry_init(&e->ent, memhash(key, keylen));
123 e->key = key;
124 e->keylen = keylen;
125 e->value = value;
126
127 hashmap_add(&map->map, &e->ent);
128 }
129
130 struct all_attrs_item {
131 const struct git_attr *attr;
132 const char *value;
133 /*
134 * If 'macro' is non-NULL, indicates that 'attr' is a macro based on
135 * the current attribute stack and contains a pointer to the match_attr
136 * definition of the macro
137 */
138 const struct match_attr *macro;
139 };
140
141 /*
142 * Reallocate and reinitialize the array of all attributes (which is used in
143 * the attribute collection process) in 'check' based on the global dictionary
144 * of attributes.
145 */
146 static void all_attrs_init(struct attr_hashmap *map, struct attr_check *check)
147 {
148 int i;
149 unsigned int size;
150
151 hashmap_lock(map);
152
153 size = hashmap_get_size(&map->map);
154 if (size < check->all_attrs_nr)
155 BUG("interned attributes shouldn't be deleted");
156
157 /*
158 * If the number of attributes in the global dictionary has increased
159 * (or this attr_check instance doesn't have an initialized all_attrs
160 * field), reallocate the provided attr_check instance's all_attrs
161 * field and fill each entry with its corresponding git_attr.
162 */
163 if (size != check->all_attrs_nr) {
164 struct attr_hash_entry *e;
165 struct hashmap_iter iter;
166
167 REALLOC_ARRAY(check->all_attrs, size);
168 check->all_attrs_nr = size;
169
170 hashmap_for_each_entry(&map->map, &iter, e,
171 ent /* member name */) {
172 const struct git_attr *a = e->value;
173 check->all_attrs[a->attr_nr].attr = a;
174 }
175 }
176
177 hashmap_unlock(map);
178
179 /*
180 * Re-initialize every entry in check->all_attrs.
181 * This re-initialization can live outside of the locked region since
182 * the attribute dictionary is no longer being accessed.
183 */
184 for (i = 0; i < check->all_attrs_nr; i++) {
185 check->all_attrs[i].value = ATTR__UNKNOWN;
186 check->all_attrs[i].macro = NULL;
187 }
188 }
189
190 /*
191 * Attribute name cannot begin with "builtin_" which
192 * is a reserved namespace for built in attributes values.
193 */
194 static int attr_name_reserved(const char *name)
195 {
196 return starts_with(name, "builtin_");
197 }
198
199 static int attr_name_valid(const char *name, size_t namelen)
200 {
201 /*
202 * Attribute name cannot begin with '-' and must consist of
203 * characters from [-A-Za-z0-9_.].
204 */
205 if (namelen <= 0 || *name == '-')
206 return 0;
207 while (namelen--) {
208 char ch = *name++;
209 if (! (ch == '-' || ch == '.' || ch == '_' ||
210 ('0' <= ch && ch <= '9') ||
211 ('a' <= ch && ch <= 'z') ||
212 ('A' <= ch && ch <= 'Z')) )
213 return 0;
214 }
215 return 1;
216 }
217
218 static void report_invalid_attr(const char *name, size_t len,
219 const char *src, int lineno)
220 {
221 struct strbuf err = STRBUF_INIT;
222 strbuf_addf(&err, _("%.*s is not a valid attribute name"),
223 (int) len, name);
224 fprintf(stderr, "%s: %s:%d\n", err.buf, src, lineno);
225 strbuf_release(&err);
226 }
227
228 /*
229 * Given a 'name', lookup and return the corresponding attribute in the global
230 * dictionary. If no entry is found, create a new attribute and store it in
231 * the dictionary.
232 */
233 static const struct git_attr *git_attr_internal(const char *name, size_t namelen)
234 {
235 struct git_attr *a;
236
237 if (!attr_name_valid(name, namelen))
238 return NULL;
239
240 hashmap_lock(&g_attr_hashmap);
241
242 a = attr_hashmap_get(&g_attr_hashmap, name, namelen);
243
244 if (!a) {
245 FLEX_ALLOC_MEM(a, name, name, namelen);
246 a->attr_nr = hashmap_get_size(&g_attr_hashmap.map);
247
248 attr_hashmap_add(&g_attr_hashmap, a->name, namelen, a);
249 if (a->attr_nr != hashmap_get_size(&g_attr_hashmap.map) - 1)
250 die(_("unable to add additional attribute"));
251 }
252
253 hashmap_unlock(&g_attr_hashmap);
254
255 return a;
256 }
257
258 const struct git_attr *git_attr(const char *name)
259 {
260 return git_attr_internal(name, strlen(name));
261 }
262
263 static const char blank[] = " \t\r\n";
264
265 /* Flags usable in read_attr() and parse_attr_line() family of functions. */
266 #define READ_ATTR_MACRO_OK (1<<0)
267 #define READ_ATTR_NOFOLLOW (1<<1)
268
269 /*
270 * Parse a whitespace-delimited attribute state (i.e., "attr",
271 * "-attr", "!attr", or "attr=value") from the string starting at src.
272 * If e is not NULL, write the results to *e. Return a pointer to the
273 * remainder of the string (with leading whitespace removed), or NULL
274 * if there was an error.
275 */
276 static const char *parse_attr(const char *src, int lineno, const char *cp,
277 struct attr_state *e)
278 {
279 const char *ep, *equals;
280 size_t len;
281
282 ep = cp + strcspn(cp, blank);
283 equals = strchr(cp, '=');
284 if (equals && ep < equals)
285 equals = NULL;
286 if (equals)
287 len = equals - cp;
288 else
289 len = ep - cp;
290 if (!e) {
291 if (*cp == '-' || *cp == '!') {
292 cp++;
293 len--;
294 }
295 if (!attr_name_valid(cp, len) || attr_name_reserved(cp)) {
296 report_invalid_attr(cp, len, src, lineno);
297 return NULL;
298 }
299 } else {
300 /*
301 * As this function is always called twice, once with
302 * e == NULL in the first pass and then e != NULL in
303 * the second pass, no need for attr_name_valid()
304 * check here.
305 */
306 if (*cp == '-' || *cp == '!') {
307 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
308 cp++;
309 len--;
310 }
311 else if (!equals)
312 e->setto = ATTR__TRUE;
313 else {
314 e->setto = xmemdupz(equals + 1, ep - equals - 1);
315 }
316 e->attr = git_attr_internal(cp, len);
317 }
318 return ep + strspn(ep, blank);
319 }
320
321 struct match_attr *parse_attr_line(const char *line, const char *src,
322 int lineno, unsigned flags)
323 {
324 size_t namelen, num_attr, i;
325 const char *cp, *name, *states;
326 struct match_attr *res = NULL;
327 int is_macro;
328 struct strbuf pattern = STRBUF_INIT;
329
330 cp = line + strspn(line, blank);
331 if (!*cp || *cp == '#')
332 return NULL;
333 name = cp;
334
335 if (strlen(line) >= ATTR_MAX_LINE_LENGTH) {
336 warning(_("ignoring overly long attributes line %d"), lineno);
337 return NULL;
338 }
339
340 if (*cp == '"' && !unquote_c_style(&pattern, name, &states)) {
341 name = pattern.buf;
342 namelen = pattern.len;
343 } else {
344 namelen = strcspn(name, blank);
345 states = name + namelen;
346 }
347
348 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
349 starts_with(name, ATTRIBUTE_MACRO_PREFIX)) {
350 if (!(flags & READ_ATTR_MACRO_OK)) {
351 fprintf_ln(stderr, _("%s not allowed: %s:%d"),
352 name, src, lineno);
353 goto fail_return;
354 }
355 is_macro = 1;
356 name += strlen(ATTRIBUTE_MACRO_PREFIX);
357 name += strspn(name, blank);
358 namelen = strcspn(name, blank);
359 if (!attr_name_valid(name, namelen) || attr_name_reserved(name)) {
360 report_invalid_attr(name, namelen, src, lineno);
361 goto fail_return;
362 }
363 }
364 else
365 is_macro = 0;
366
367 states += strspn(states, blank);
368
369 /* First pass to count the attr_states */
370 for (cp = states, num_attr = 0; *cp; num_attr++) {
371 cp = parse_attr(src, lineno, cp, NULL);
372 if (!cp)
373 goto fail_return;
374 }
375
376 res = xcalloc(1, st_add3(sizeof(*res),
377 st_mult(sizeof(struct attr_state), num_attr),
378 is_macro ? 0 : namelen + 1));
379 if (is_macro) {
380 res->u.attr = git_attr_internal(name, namelen);
381 } else {
382 char *p = (char *)&(res->state[num_attr]);
383 memcpy(p, name, namelen);
384 res->u.pat.pattern = p;
385 parse_path_pattern(&res->u.pat.pattern,
386 &res->u.pat.patternlen,
387 &res->u.pat.flags,
388 &res->u.pat.nowildcardlen);
389 if (res->u.pat.flags & PATTERN_FLAG_NEGATIVE) {
390 warning(_("Negative patterns are ignored in git attributes\n"
391 "Use '\\!' for literal leading exclamation."));
392 goto fail_return;
393 }
394 }
395 res->is_macro = is_macro;
396 res->num_attr = num_attr;
397
398 /* Second pass to fill the attr_states */
399 for (cp = states, i = 0; *cp; i++) {
400 cp = parse_attr(src, lineno, cp, &(res->state[i]));
401 }
402
403 strbuf_release(&pattern);
404 return res;
405
406 fail_return:
407 strbuf_release(&pattern);
408 free(res);
409 return NULL;
410 }
411
412 /*
413 * Like info/exclude and .gitignore, the attribute information can
414 * come from many places.
415 *
416 * (1) .gitattributes file of the same directory;
417 * (2) .gitattributes file of the parent directory if (1) does not have
418 * any match; this goes recursively upwards, just like .gitignore.
419 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
420 *
421 * In the same file, later entries override the earlier match, so in the
422 * global list, we would have entries from info/attributes the earliest
423 * (reading the file from top to bottom), .gitattributes of the root
424 * directory (again, reading the file from top to bottom) down to the
425 * current directory, and then scan the list backwards to find the first match.
426 * This is exactly the same as what is_excluded() does in dir.c to deal with
427 * .gitignore file and info/excludes file as a fallback.
428 */
429
430 struct attr_stack {
431 struct attr_stack *prev;
432 char *origin;
433 size_t originlen;
434 unsigned num_matches;
435 unsigned alloc;
436 struct match_attr **attrs;
437 };
438
439 static void attr_stack_free(struct attr_stack *e)
440 {
441 unsigned i;
442 free(e->origin);
443 for (i = 0; i < e->num_matches; i++) {
444 struct match_attr *a = e->attrs[i];
445 size_t j;
446
447 for (j = 0; j < a->num_attr; j++) {
448 const char *setto = a->state[j].setto;
449 if (setto == ATTR__TRUE ||
450 setto == ATTR__FALSE ||
451 setto == ATTR__UNSET ||
452 setto == ATTR__UNKNOWN)
453 ;
454 else
455 free((char *) setto);
456 }
457 free(a);
458 }
459 free(e->attrs);
460 free(e);
461 }
462
463 static void drop_attr_stack(struct attr_stack **stack)
464 {
465 while (*stack) {
466 struct attr_stack *elem = *stack;
467 *stack = elem->prev;
468 attr_stack_free(elem);
469 }
470 }
471
472 /* List of all attr_check structs; access should be surrounded by mutex */
473 static struct check_vector {
474 size_t nr;
475 size_t alloc;
476 struct attr_check **checks;
477 pthread_mutex_t mutex;
478 } check_vector;
479
480 static inline void vector_lock(void)
481 {
482 pthread_mutex_lock(&check_vector.mutex);
483 }
484
485 static inline void vector_unlock(void)
486 {
487 pthread_mutex_unlock(&check_vector.mutex);
488 }
489
490 static void check_vector_add(struct attr_check *c)
491 {
492 vector_lock();
493
494 ALLOC_GROW(check_vector.checks,
495 check_vector.nr + 1,
496 check_vector.alloc);
497 check_vector.checks[check_vector.nr++] = c;
498
499 vector_unlock();
500 }
501
502 static void check_vector_remove(struct attr_check *check)
503 {
504 int i;
505
506 vector_lock();
507
508 /* Find entry */
509 for (i = 0; i < check_vector.nr; i++)
510 if (check_vector.checks[i] == check)
511 break;
512
513 if (i >= check_vector.nr)
514 BUG("no entry found");
515
516 /* shift entries over */
517 for (; i < check_vector.nr - 1; i++)
518 check_vector.checks[i] = check_vector.checks[i + 1];
519
520 check_vector.nr--;
521
522 vector_unlock();
523 }
524
525 /* Iterate through all attr_check instances and drop their stacks */
526 static void drop_all_attr_stacks(void)
527 {
528 int i;
529
530 vector_lock();
531
532 for (i = 0; i < check_vector.nr; i++) {
533 drop_attr_stack(&check_vector.checks[i]->stack);
534 }
535
536 vector_unlock();
537 }
538
539 struct attr_check *attr_check_alloc(void)
540 {
541 struct attr_check *c = xcalloc(1, sizeof(struct attr_check));
542
543 /* save pointer to the check struct */
544 check_vector_add(c);
545
546 return c;
547 }
548
549 struct attr_check *attr_check_initl(const char *one, ...)
550 {
551 struct attr_check *check;
552 int cnt;
553 va_list params;
554 const char *param;
555
556 va_start(params, one);
557 for (cnt = 1; (param = va_arg(params, const char *)) != NULL; cnt++)
558 ;
559 va_end(params);
560
561 check = attr_check_alloc();
562 check->nr = cnt;
563 check->alloc = cnt;
564 CALLOC_ARRAY(check->items, cnt);
565
566 check->items[0].attr = git_attr(one);
567 va_start(params, one);
568 for (cnt = 1; cnt < check->nr; cnt++) {
569 const struct git_attr *attr;
570 param = va_arg(params, const char *);
571 if (!param)
572 BUG("counted %d != ended at %d",
573 check->nr, cnt);
574 attr = git_attr(param);
575 if (!attr)
576 BUG("%s: not a valid attribute name", param);
577 check->items[cnt].attr = attr;
578 }
579 va_end(params);
580 return check;
581 }
582
583 struct attr_check *attr_check_dup(const struct attr_check *check)
584 {
585 struct attr_check *ret;
586
587 if (!check)
588 return NULL;
589
590 ret = attr_check_alloc();
591
592 ret->nr = check->nr;
593 ret->alloc = check->alloc;
594 DUP_ARRAY(ret->items, check->items, ret->nr);
595
596 return ret;
597 }
598
599 struct attr_check_item *attr_check_append(struct attr_check *check,
600 const struct git_attr *attr)
601 {
602 struct attr_check_item *item;
603
604 ALLOC_GROW(check->items, check->nr + 1, check->alloc);
605 item = &check->items[check->nr++];
606 item->attr = attr;
607 return item;
608 }
609
610 void attr_check_reset(struct attr_check *check)
611 {
612 check->nr = 0;
613 }
614
615 void attr_check_clear(struct attr_check *check)
616 {
617 FREE_AND_NULL(check->items);
618 check->alloc = 0;
619 check->nr = 0;
620
621 FREE_AND_NULL(check->all_attrs);
622 check->all_attrs_nr = 0;
623
624 drop_attr_stack(&check->stack);
625 }
626
627 void attr_check_free(struct attr_check *check)
628 {
629 if (check) {
630 /* Remove check from the check vector */
631 check_vector_remove(check);
632
633 attr_check_clear(check);
634 free(check);
635 }
636 }
637
638 static const char *builtin_attr[] = {
639 "[attr]binary -diff -merge -text",
640 NULL,
641 };
642
643 static void handle_attr_line(struct attr_stack *res,
644 const char *line,
645 const char *src,
646 int lineno,
647 unsigned flags)
648 {
649 struct match_attr *a;
650
651 a = parse_attr_line(line, src, lineno, flags);
652 if (!a)
653 return;
654 ALLOC_GROW_BY(res->attrs, res->num_matches, 1, res->alloc);
655 res->attrs[res->num_matches - 1] = a;
656 }
657
658 static struct attr_stack *read_attr_from_array(const char **list)
659 {
660 struct attr_stack *res;
661 const char *line;
662 int lineno = 0;
663
664 CALLOC_ARRAY(res, 1);
665 while ((line = *(list++)) != NULL)
666 handle_attr_line(res, line, "[builtin]", ++lineno,
667 READ_ATTR_MACRO_OK);
668 return res;
669 }
670
671 /*
672 * Callers into the attribute system assume there is a single, system-wide
673 * global state where attributes are read from and when the state is flipped by
674 * calling git_attr_set_direction(), the stack frames that have been
675 * constructed need to be discarded so that subsequent calls into the
676 * attribute system will lazily read from the right place. Since changing
677 * direction causes a global paradigm shift, it should not ever be called while
678 * another thread could potentially be calling into the attribute system.
679 */
680 static enum git_attr_direction direction;
681
682 void git_attr_set_direction(enum git_attr_direction new_direction)
683 {
684 if (is_bare_repository() && new_direction != GIT_ATTR_INDEX)
685 BUG("non-INDEX attr direction in a bare repo");
686
687 if (new_direction != direction)
688 drop_all_attr_stacks();
689
690 direction = new_direction;
691 }
692
693 static struct attr_stack *read_attr_from_file(const char *path, unsigned flags)
694 {
695 struct strbuf buf = STRBUF_INIT;
696 int fd;
697 FILE *fp;
698 struct attr_stack *res;
699 int lineno = 0;
700 struct stat st;
701
702 if (flags & READ_ATTR_NOFOLLOW)
703 fd = open_nofollow(path, O_RDONLY);
704 else
705 fd = open(path, O_RDONLY);
706
707 if (fd < 0) {
708 warn_on_fopen_errors(path);
709 return NULL;
710 }
711 fp = xfdopen(fd, "r");
712 if (fstat(fd, &st)) {
713 warning_errno(_("cannot fstat gitattributes file '%s'"), path);
714 fclose(fp);
715 return NULL;
716 }
717 if (st.st_size >= ATTR_MAX_FILE_SIZE) {
718 warning(_("ignoring overly large gitattributes file '%s'"), path);
719 fclose(fp);
720 return NULL;
721 }
722
723 CALLOC_ARRAY(res, 1);
724 while (strbuf_getline(&buf, fp) != EOF) {
725 if (!lineno && starts_with(buf.buf, utf8_bom))
726 strbuf_remove(&buf, 0, strlen(utf8_bom));
727 handle_attr_line(res, buf.buf, path, ++lineno, flags);
728 }
729
730 fclose(fp);
731 strbuf_release(&buf);
732 return res;
733 }
734
735 static struct attr_stack *read_attr_from_buf(char *buf, size_t length,
736 const char *path, unsigned flags)
737 {
738 struct attr_stack *res;
739 char *sp;
740 int lineno = 0;
741
742 if (!buf)
743 return NULL;
744 if (length >= ATTR_MAX_FILE_SIZE) {
745 warning(_("ignoring overly large gitattributes blob '%s'"), path);
746 free(buf);
747 return NULL;
748 }
749
750 CALLOC_ARRAY(res, 1);
751 for (sp = buf; *sp;) {
752 char *ep;
753 int more;
754
755 ep = strchrnul(sp, '\n');
756 more = (*ep == '\n');
757 *ep = '\0';
758 handle_attr_line(res, sp, path, ++lineno, flags);
759 sp = ep + more;
760 }
761 free(buf);
762
763 return res;
764 }
765
766 static struct attr_stack *read_attr_from_blob(struct index_state *istate,
767 const struct object_id *tree_oid,
768 const char *path, unsigned flags)
769 {
770 struct object_id oid;
771 size_t sz;
772 enum object_type type;
773 void *buf;
774 unsigned short mode;
775
776 if (!tree_oid)
777 return NULL;
778
779 if (get_tree_entry(istate->repo, tree_oid, path, &oid, &mode))
780 return NULL;
781
782 buf = odb_read_object(istate->repo->objects, &oid, &type, &sz);
783 if (!buf || type != OBJ_BLOB) {
784 free(buf);
785 return NULL;
786 }
787
788 return read_attr_from_buf(buf, sz, path, flags);
789 }
790
791 static struct attr_stack *read_attr_from_index(struct index_state *istate,
792 const char *path, unsigned flags)
793 {
794 struct attr_stack *stack = NULL;
795 char *buf;
796 unsigned long size;
797 int sparse_dir_pos = -1;
798
799 if (!istate)
800 return NULL;
801
802 /*
803 * When handling sparse-checkouts, .gitattributes files
804 * may reside within a sparse directory. We distinguish
805 * whether a path exists directly in the index or not by
806 * evaluating if 'pos' is negative.
807 * If 'pos' is negative, the path is not directly present
808 * in the index and is likely within a sparse directory.
809 * For paths not in the index, The absolute value of 'pos'
810 * minus 1 gives us the position where the path would be
811 * inserted in lexicographic order within the index.
812 * We then subtract another 1 from this value
813 * (sparse_dir_pos = -pos - 2) to find the position of the
814 * last index entry which is lexicographically smaller than
815 * the path. This would be the sparse directory containing
816 * the path. By identifying the sparse directory containing
817 * the path, we can correctly read the attributes specified
818 * in the .gitattributes file from the tree object of the
819 * sparse directory.
820 */
821 if (!path_in_cone_mode_sparse_checkout(path, istate)) {
822 int pos = index_name_pos_sparse(istate, path, strlen(path));
823
824 if (pos < 0)
825 sparse_dir_pos = -pos - 2;
826 }
827
828 if (sparse_dir_pos >= 0 &&
829 S_ISSPARSEDIR(istate->cache[sparse_dir_pos]->ce_mode) &&
830 !strncmp(istate->cache[sparse_dir_pos]->name, path, ce_namelen(istate->cache[sparse_dir_pos]))) {
831 const char *relative_path = path + ce_namelen(istate->cache[sparse_dir_pos]);
832 stack = read_attr_from_blob(istate, &istate->cache[sparse_dir_pos]->oid, relative_path, flags);
833 } else {
834 buf = read_blob_data_from_index(istate, path, &size);
835 if (buf)
836 stack = read_attr_from_buf(buf, size, path, flags);
837 }
838 return stack;
839 }
840
841 static struct attr_stack *read_attr(struct index_state *istate,
842 const struct object_id *tree_oid,
843 const char *path, unsigned flags)
844 {
845 struct attr_stack *res = NULL;
846
847 if (direction == GIT_ATTR_INDEX) {
848 res = read_attr_from_index(istate, path, flags);
849 } else if (tree_oid) {
850 res = read_attr_from_blob(istate, tree_oid, path, flags);
851 } else if (!is_bare_repository()) {
852 if (direction == GIT_ATTR_CHECKOUT) {
853 res = read_attr_from_index(istate, path, flags);
854 if (!res)
855 res = read_attr_from_file(path, flags);
856 } else if (direction == GIT_ATTR_CHECKIN) {
857 res = read_attr_from_file(path, flags);
858 if (!res)
859 /*
860 * There is no checked out .gitattributes file
861 * there, but we might have it in the index.
862 * We allow operation in a sparsely checked out
863 * work tree, so read from it.
864 */
865 res = read_attr_from_index(istate, path, flags);
866 }
867 }
868
869 if (!res)
870 CALLOC_ARRAY(res, 1);
871 return res;
872 }
873
874 const char *git_attr_system_file(void)
875 {
876 static const char *system_wide;
877 if (!system_wide)
878 system_wide = system_path(ETC_GITATTRIBUTES);
879 return system_wide;
880 }
881
882 const char *git_attr_global_file(void)
883 {
884 struct repo_config_values *cfg = repo_config_values(the_repository);
885 if (!cfg->attributes_file)
886 cfg->attributes_file = xdg_config_home("attributes");
887
888 return cfg->attributes_file;
889 }
890
891 int git_attr_system_is_enabled(void)
892 {
893 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
894 }
895
896 static GIT_PATH_FUNC(git_path_info_attributes, INFOATTRIBUTES_FILE)
897
898 static void push_stack(struct attr_stack **attr_stack_p,
899 struct attr_stack *elem, char *origin, size_t originlen)
900 {
901 if (elem) {
902 elem->origin = origin;
903 if (origin)
904 elem->originlen = originlen;
905 elem->prev = *attr_stack_p;
906 *attr_stack_p = elem;
907 }
908 }
909
910 static void bootstrap_attr_stack(struct index_state *istate,
911 const struct object_id *tree_oid,
912 struct attr_stack **stack)
913 {
914 struct attr_stack *e;
915 unsigned flags = READ_ATTR_MACRO_OK;
916
917 if (*stack)
918 return;
919
920 /* builtin frame */
921 e = read_attr_from_array(builtin_attr);
922 push_stack(stack, e, NULL, 0);
923
924 /* system-wide frame */
925 if (git_attr_system_is_enabled()) {
926 e = read_attr_from_file(git_attr_system_file(), flags);
927 push_stack(stack, e, NULL, 0);
928 }
929
930 /* home directory */
931 if (git_attr_global_file()) {
932 e = read_attr_from_file(git_attr_global_file(), flags);
933 push_stack(stack, e, NULL, 0);
934 }
935
936 /* root directory */
937 e = read_attr(istate, tree_oid, GITATTRIBUTES_FILE, flags | READ_ATTR_NOFOLLOW);
938 push_stack(stack, e, xstrdup(""), 0);
939
940 /* info frame */
941 if (startup_info->have_repository)
942 e = read_attr_from_file(git_path_info_attributes(), flags);
943 else
944 e = NULL;
945 if (!e)
946 CALLOC_ARRAY(e, 1);
947 push_stack(stack, e, NULL, 0);
948 }
949
950 static void prepare_attr_stack(struct index_state *istate,
951 const struct object_id *tree_oid,
952 const char *path, int dirlen,
953 struct attr_stack **stack)
954 {
955 struct attr_stack *info;
956 struct strbuf pathbuf = STRBUF_INIT;
957
958 /*
959 * At the bottom of the attribute stack is the built-in
960 * set of attribute definitions, followed by the contents
961 * of $(prefix)/etc/gitattributes and a file specified by
962 * core.attributesfile. Then, contents from
963 * .gitattributes files from directories closer to the
964 * root to the ones in deeper directories are pushed
965 * to the stack. Finally, at the very top of the stack
966 * we always keep the contents of $GIT_DIR/info/attributes.
967 *
968 * When checking, we use entries from near the top of the
969 * stack, preferring $GIT_DIR/info/attributes, then
970 * .gitattributes in deeper directories to shallower ones,
971 * and finally use the built-in set as the default.
972 */
973 bootstrap_attr_stack(istate, tree_oid, stack);
974
975 /*
976 * Pop the "info" one that is always at the top of the stack.
977 */
978 info = *stack;
979 *stack = info->prev;
980
981 /*
982 * Pop the ones from directories that are not the prefix of
983 * the path we are checking. Break out of the loop when we see
984 * the root one (whose origin is an empty string "") or the builtin
985 * one (whose origin is NULL) without popping it.
986 */
987 while ((*stack)->origin) {
988 int namelen = (*stack)->originlen;
989 struct attr_stack *elem;
990
991 elem = *stack;
992 if (namelen <= dirlen &&
993 !strncmp(elem->origin, path, namelen) &&
994 (!namelen || path[namelen] == '/'))
995 break;
996
997 *stack = elem->prev;
998 attr_stack_free(elem);
999 }
1000
1001 /*
1002 * bootstrap_attr_stack() should have added, and the
1003 * above loop should have stopped before popping, the
1004 * root element whose attr_stack->origin is set to an
1005 * empty string.
1006 */
1007 assert((*stack)->origin);
1008
1009 strbuf_addstr(&pathbuf, (*stack)->origin);
1010 /* Build up to the directory 'path' is in */
1011 while (pathbuf.len < dirlen) {
1012 size_t len = pathbuf.len;
1013 struct attr_stack *next;
1014 char *origin;
1015
1016 /* Skip path-separator */
1017 if (len < dirlen && is_dir_sep(path[len]))
1018 len++;
1019 /* Find the end of the next component */
1020 while (len < dirlen && !is_dir_sep(path[len]))
1021 len++;
1022
1023 if (pathbuf.len > 0)
1024 strbuf_addch(&pathbuf, '/');
1025 strbuf_add(&pathbuf, path + pathbuf.len, (len - pathbuf.len));
1026 strbuf_addf(&pathbuf, "/%s", GITATTRIBUTES_FILE);
1027
1028 next = read_attr(istate, tree_oid, pathbuf.buf, READ_ATTR_NOFOLLOW);
1029
1030 /* reset the pathbuf to not include "/.gitattributes" */
1031 strbuf_setlen(&pathbuf, len);
1032
1033 origin = xstrdup(pathbuf.buf);
1034 push_stack(stack, next, origin, len);
1035 }
1036
1037 /*
1038 * Finally push the "info" one at the top of the stack.
1039 */
1040 push_stack(stack, info, NULL, 0);
1041
1042 strbuf_release(&pathbuf);
1043 }
1044
1045 static int path_matches(const char *pathname, int pathlen,
1046 int basename_offset,
1047 const struct pattern *pat,
1048 const char *base, int baselen)
1049 {
1050 const char *pattern = pat->pattern;
1051 int prefix = pat->nowildcardlen;
1052 int isdir = (pathlen && pathname[pathlen - 1] == '/');
1053
1054 if ((pat->flags & PATTERN_FLAG_MUSTBEDIR) && !isdir)
1055 return 0;
1056
1057 if (pat->flags & PATTERN_FLAG_NODIR) {
1058 return match_basename(pathname + basename_offset,
1059 pathlen - basename_offset - isdir,
1060 pattern, prefix,
1061 pat->patternlen, pat->flags);
1062 }
1063 return match_pathname(pathname, pathlen - isdir,
1064 base, baselen,
1065 pattern, prefix, pat->patternlen);
1066 }
1067
1068 struct attr_state_queue {
1069 const struct attr_state **items;
1070 size_t alloc, nr;
1071 };
1072
1073 static void attr_state_queue_push(struct attr_state_queue *t,
1074 const struct match_attr *a)
1075 {
1076 for (size_t i = 0; i < a->num_attr; i++) {
1077 ALLOC_GROW(t->items, t->nr + 1, t->alloc);
1078 t->items[t->nr++] = &a->state[i];
1079 }
1080 }
1081
1082 static const struct attr_state *attr_state_queue_pop(struct attr_state_queue *t)
1083 {
1084 return t->nr ? t->items[--t->nr] : NULL;
1085 }
1086
1087 static void attr_state_queue_release(struct attr_state_queue *t)
1088 {
1089 free(t->items);
1090 }
1091
1092 static int fill_one(struct all_attrs_item *all_attrs,
1093 const struct match_attr *a, int rem)
1094 {
1095 struct attr_state_queue todo = { 0 };
1096 const struct attr_state *state;
1097
1098 attr_state_queue_push(&todo, a);
1099 while (rem > 0 && (state = attr_state_queue_pop(&todo))) {
1100 const struct git_attr *attr = state->attr;
1101 const char **n = &(all_attrs[attr->attr_nr].value);
1102 const char *v = state->setto;
1103
1104 if (*n == ATTR__UNKNOWN) {
1105 const struct all_attrs_item *item =
1106 &all_attrs[attr->attr_nr];
1107 *n = v;
1108 rem--;
1109 if (item->macro && item->value == ATTR__TRUE)
1110 attr_state_queue_push(&todo, item->macro);
1111 }
1112 }
1113 attr_state_queue_release(&todo);
1114 return rem;
1115 }
1116
1117 static int fill(const char *path, int pathlen, int basename_offset,
1118 const struct attr_stack *stack,
1119 struct all_attrs_item *all_attrs, int rem)
1120 {
1121 for (; rem > 0 && stack; stack = stack->prev) {
1122 unsigned i;
1123 const char *base = stack->origin ? stack->origin : "";
1124
1125 for (i = stack->num_matches; 0 < rem && 0 < i; i--) {
1126 const struct match_attr *a = stack->attrs[i - 1];
1127 if (a->is_macro)
1128 continue;
1129 if (path_matches(path, pathlen, basename_offset,
1130 &a->u.pat, base, stack->originlen))
1131 rem = fill_one(all_attrs, a, rem);
1132 }
1133 }
1134
1135 return rem;
1136 }
1137
1138 /*
1139 * Marks the attributes which are macros based on the attribute stack.
1140 * This prevents having to search through the attribute stack each time
1141 * a macro needs to be expanded during the fill stage.
1142 */
1143 static void determine_macros(struct all_attrs_item *all_attrs,
1144 const struct attr_stack *stack)
1145 {
1146 for (; stack; stack = stack->prev) {
1147 unsigned i;
1148 for (i = stack->num_matches; i > 0; i--) {
1149 const struct match_attr *ma = stack->attrs[i - 1];
1150 if (ma->is_macro) {
1151 unsigned int n = ma->u.attr->attr_nr;
1152 if (!all_attrs[n].macro) {
1153 all_attrs[n].macro = ma;
1154 }
1155 }
1156 }
1157 }
1158 }
1159
1160 /*
1161 * Collect attributes for path into the array pointed to by check->all_attrs.
1162 * If check->check_nr is non-zero, only attributes in check[] are collected.
1163 * Otherwise all attributes are collected.
1164 */
1165 static void collect_some_attrs(struct index_state *istate,
1166 const struct object_id *tree_oid,
1167 const char *path, struct attr_check *check)
1168 {
1169 int pathlen, rem, dirlen;
1170 const char *cp, *last_slash = NULL;
1171 int basename_offset;
1172
1173 for (cp = path; *cp; cp++) {
1174 if (*cp == '/' && cp[1])
1175 last_slash = cp;
1176 }
1177 pathlen = cp - path;
1178 if (last_slash) {
1179 basename_offset = last_slash + 1 - path;
1180 dirlen = last_slash - path;
1181 } else {
1182 basename_offset = 0;
1183 dirlen = 0;
1184 }
1185
1186 prepare_attr_stack(istate, tree_oid, path, dirlen, &check->stack);
1187 all_attrs_init(&g_attr_hashmap, check);
1188 determine_macros(check->all_attrs, check->stack);
1189
1190 rem = check->all_attrs_nr;
1191 fill(path, pathlen, basename_offset, check->stack, check->all_attrs, rem);
1192 }
1193
1194 static const char *default_attr_source_tree_object_name;
1195
1196 void set_git_attr_source(const char *tree_object_name)
1197 {
1198 default_attr_source_tree_object_name = xstrdup(tree_object_name);
1199 }
1200
1201 static int compute_default_attr_source(struct object_id *attr_source)
1202 {
1203 int ignore_bad_attr_tree = 0;
1204
1205 if (!default_attr_source_tree_object_name)
1206 default_attr_source_tree_object_name = getenv(GIT_ATTR_SOURCE_ENVIRONMENT);
1207
1208 if (!default_attr_source_tree_object_name && git_attr_tree) {
1209 default_attr_source_tree_object_name = git_attr_tree;
1210 ignore_bad_attr_tree = 1;
1211 }
1212
1213 if (!default_attr_source_tree_object_name)
1214 return 0;
1215
1216 if (!startup_info->have_repository) {
1217 if (!ignore_bad_attr_tree)
1218 die(_("cannot use --attr-source or GIT_ATTR_SOURCE without repo"));
1219 return 0;
1220 }
1221
1222 if (repo_get_oid_treeish(the_repository,
1223 default_attr_source_tree_object_name,
1224 attr_source)) {
1225 if (!ignore_bad_attr_tree)
1226 die(_("bad --attr-source or GIT_ATTR_SOURCE"));
1227 return 0;
1228 }
1229
1230 return 1;
1231 }
1232
1233 static struct object_id *default_attr_source(void)
1234 {
1235 static struct object_id attr_source;
1236 static int has_attr_source = -1;
1237
1238 if (has_attr_source < 0)
1239 has_attr_source = compute_default_attr_source(&attr_source);
1240 if (!has_attr_source)
1241 return NULL;
1242 return &attr_source;
1243 }
1244
1245 static const char *interned_mode_string(unsigned int mode)
1246 {
1247 static struct {
1248 unsigned int val;
1249 char str[7];
1250 } mode_string[] = {
1251 { .val = 0040000 },
1252 { .val = 0100644 },
1253 { .val = 0100755 },
1254 { .val = 0120000 },
1255 { .val = 0160000 },
1256 };
1257 int i;
1258
1259 for (i = 0; i < ARRAY_SIZE(mode_string); i++) {
1260 if (mode_string[i].val != mode)
1261 continue;
1262 if (!*mode_string[i].str)
1263 snprintf(mode_string[i].str, sizeof(mode_string[i].str),
1264 "%06o", mode);
1265 return mode_string[i].str;
1266 }
1267 BUG("Unsupported mode 0%o", mode);
1268 }
1269
1270 static const char *builtin_object_mode_attr(struct index_state *istate, const char *path)
1271 {
1272 unsigned int mode;
1273
1274 if (direction == GIT_ATTR_CHECKIN) {
1275 struct object_id oid;
1276 struct stat st;
1277 if (lstat(path, &st))
1278 die_errno(_("unable to stat '%s'"), path);
1279 mode = canon_mode(st.st_mode);
1280 if (S_ISDIR(mode)) {
1281 /*
1282 *`path` is either a directory or it is a submodule,
1283 * in which case it is already indexed as submodule
1284 * or it does not exist in the index yet and we need to
1285 * check if we can resolve to a ref.
1286 */
1287 int pos = index_name_pos(istate, path, strlen(path));
1288 if (pos >= 0) {
1289 if (S_ISGITLINK(istate->cache[pos]->ce_mode))
1290 mode = istate->cache[pos]->ce_mode;
1291 } else if (repo_resolve_gitlink_ref(the_repository, path,
1292 "HEAD", &oid) == 0) {
1293 mode = S_IFGITLINK;
1294 }
1295 }
1296 } else {
1297 /*
1298 * For GIT_ATTR_CHECKOUT and GIT_ATTR_INDEX we only check
1299 * for mode in the index.
1300 */
1301 int pos = index_name_pos(istate, path, strlen(path));
1302 if (pos >= 0)
1303 mode = istate->cache[pos]->ce_mode;
1304 else
1305 return ATTR__UNSET;
1306 }
1307
1308 return interned_mode_string(mode);
1309 }
1310
1311
1312 static const char *compute_builtin_attr(struct index_state *istate,
1313 const char *path,
1314 const struct git_attr *attr) {
1315 static const struct git_attr *object_mode_attr;
1316
1317 if (!object_mode_attr)
1318 object_mode_attr = git_attr("builtin_objectmode");
1319
1320 if (attr == object_mode_attr)
1321 return builtin_object_mode_attr(istate, path);
1322 return ATTR__UNSET;
1323 }
1324
1325 void git_check_attr(struct index_state *istate,
1326 const char *path,
1327 struct attr_check *check)
1328 {
1329 int i;
1330 const struct object_id *tree_oid = default_attr_source();
1331
1332 collect_some_attrs(istate, tree_oid, path, check);
1333
1334 for (i = 0; i < check->nr; i++) {
1335 unsigned int n = check->items[i].attr->attr_nr;
1336 const char *value = check->all_attrs[n].value;
1337 if (value == ATTR__UNKNOWN)
1338 value = compute_builtin_attr(istate, path, check->all_attrs[n].attr);
1339 check->items[i].value = value;
1340 }
1341 }
1342
1343 void git_all_attrs(struct index_state *istate,
1344 const char *path, struct attr_check *check)
1345 {
1346 int i;
1347 const struct object_id *tree_oid = default_attr_source();
1348
1349 attr_check_reset(check);
1350 collect_some_attrs(istate, tree_oid, path, check);
1351
1352 for (i = 0; i < check->all_attrs_nr; i++) {
1353 const char *name = check->all_attrs[i].attr->name;
1354 const char *value = check->all_attrs[i].value;
1355 struct attr_check_item *item;
1356 if (value == ATTR__UNSET || value == ATTR__UNKNOWN)
1357 continue;
1358 item = attr_check_append(check, git_attr(name));
1359 item->value = value;
1360 }
1361 }
1362
1363 void attr_start(void)
1364 {
1365 pthread_mutex_init(&g_attr_hashmap.mutex, NULL);
1366 pthread_mutex_init(&check_vector.mutex, NULL);
1367 }