]> git.ipfire.org Git - thirdparty/git.git/blob - attr.c
refs: print error message in debug output
[thirdparty/git.git] / attr.c
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 #include "cache.h"
10 #include "config.h"
11 #include "exec-cmd.h"
12 #include "attr.h"
13 #include "dir.h"
14 #include "utf8.h"
15 #include "quote.h"
16 #include "thread-utils.h"
17 #include "dir.h"
18
19 const char git_attr__true[] = "(builtin)true";
20 const char git_attr__false[] = "\0(builtin)false";
21 static const char git_attr__unknown[] = "(builtin)unknown";
22 #define ATTR__TRUE git_attr__true
23 #define ATTR__FALSE git_attr__false
24 #define ATTR__UNSET NULL
25 #define ATTR__UNKNOWN git_attr__unknown
26
27 #ifndef DEBUG_ATTR
28 #define DEBUG_ATTR 0
29 #endif
30
31 struct git_attr {
32 int attr_nr; /* unique attribute number */
33 char name[FLEX_ARRAY]; /* attribute name */
34 };
35
36 const char *git_attr_name(const struct git_attr *attr)
37 {
38 return attr->name;
39 }
40
41 struct attr_hashmap {
42 struct hashmap map;
43 pthread_mutex_t mutex;
44 };
45
46 static inline void hashmap_lock(struct attr_hashmap *map)
47 {
48 pthread_mutex_lock(&map->mutex);
49 }
50
51 static inline void hashmap_unlock(struct attr_hashmap *map)
52 {
53 pthread_mutex_unlock(&map->mutex);
54 }
55
56 /* The container for objects stored in "struct attr_hashmap" */
57 struct attr_hash_entry {
58 struct hashmap_entry ent;
59 const char *key; /* the key; memory should be owned by value */
60 size_t keylen; /* length of the key */
61 void *value; /* the stored value */
62 };
63
64 /* attr_hashmap comparison function */
65 static int attr_hash_entry_cmp(const void *unused_cmp_data,
66 const struct hashmap_entry *eptr,
67 const struct hashmap_entry *entry_or_key,
68 const void *unused_keydata)
69 {
70 const struct attr_hash_entry *a, *b;
71
72 a = container_of(eptr, const struct attr_hash_entry, ent);
73 b = container_of(entry_or_key, const struct attr_hash_entry, ent);
74 return (a->keylen != b->keylen) || strncmp(a->key, b->key, a->keylen);
75 }
76
77 /*
78 * The global dictionary of all interned attributes. This
79 * is a singleton object which is shared between threads.
80 * Access to this dictionary must be surrounded with a mutex.
81 */
82 static struct attr_hashmap g_attr_hashmap = {
83 HASHMAP_INIT(attr_hash_entry_cmp, NULL)
84 };
85
86 /*
87 * Retrieve the 'value' stored in a hashmap given the provided 'key'.
88 * If there is no matching entry, return NULL.
89 */
90 static void *attr_hashmap_get(struct attr_hashmap *map,
91 const char *key, size_t keylen)
92 {
93 struct attr_hash_entry k;
94 struct attr_hash_entry *e;
95
96 hashmap_entry_init(&k.ent, memhash(key, keylen));
97 k.key = key;
98 k.keylen = keylen;
99 e = hashmap_get_entry(&map->map, &k, ent, NULL);
100
101 return e ? e->value : NULL;
102 }
103
104 /* Add 'value' to a hashmap based on the provided 'key'. */
105 static void attr_hashmap_add(struct attr_hashmap *map,
106 const char *key, size_t keylen,
107 void *value)
108 {
109 struct attr_hash_entry *e;
110
111 e = xmalloc(sizeof(struct attr_hash_entry));
112 hashmap_entry_init(&e->ent, memhash(key, keylen));
113 e->key = key;
114 e->keylen = keylen;
115 e->value = value;
116
117 hashmap_add(&map->map, &e->ent);
118 }
119
120 struct all_attrs_item {
121 const struct git_attr *attr;
122 const char *value;
123 /*
124 * If 'macro' is non-NULL, indicates that 'attr' is a macro based on
125 * the current attribute stack and contains a pointer to the match_attr
126 * definition of the macro
127 */
128 const struct match_attr *macro;
129 };
130
131 /*
132 * Reallocate and reinitialize the array of all attributes (which is used in
133 * the attribute collection process) in 'check' based on the global dictionary
134 * of attributes.
135 */
136 static void all_attrs_init(struct attr_hashmap *map, struct attr_check *check)
137 {
138 int i;
139 unsigned int size;
140
141 hashmap_lock(map);
142
143 size = hashmap_get_size(&map->map);
144 if (size < check->all_attrs_nr)
145 BUG("interned attributes shouldn't be deleted");
146
147 /*
148 * If the number of attributes in the global dictionary has increased
149 * (or this attr_check instance doesn't have an initialized all_attrs
150 * field), reallocate the provided attr_check instance's all_attrs
151 * field and fill each entry with its corresponding git_attr.
152 */
153 if (size != check->all_attrs_nr) {
154 struct attr_hash_entry *e;
155 struct hashmap_iter iter;
156
157 REALLOC_ARRAY(check->all_attrs, size);
158 check->all_attrs_nr = size;
159
160 hashmap_for_each_entry(&map->map, &iter, e,
161 ent /* member name */) {
162 const struct git_attr *a = e->value;
163 check->all_attrs[a->attr_nr].attr = a;
164 }
165 }
166
167 hashmap_unlock(map);
168
169 /*
170 * Re-initialize every entry in check->all_attrs.
171 * This re-initialization can live outside of the locked region since
172 * the attribute dictionary is no longer being accessed.
173 */
174 for (i = 0; i < check->all_attrs_nr; i++) {
175 check->all_attrs[i].value = ATTR__UNKNOWN;
176 check->all_attrs[i].macro = NULL;
177 }
178 }
179
180 static int attr_name_valid(const char *name, size_t namelen)
181 {
182 /*
183 * Attribute name cannot begin with '-' and must consist of
184 * characters from [-A-Za-z0-9_.].
185 */
186 if (namelen <= 0 || *name == '-')
187 return 0;
188 while (namelen--) {
189 char ch = *name++;
190 if (! (ch == '-' || ch == '.' || ch == '_' ||
191 ('0' <= ch && ch <= '9') ||
192 ('a' <= ch && ch <= 'z') ||
193 ('A' <= ch && ch <= 'Z')) )
194 return 0;
195 }
196 return 1;
197 }
198
199 static void report_invalid_attr(const char *name, size_t len,
200 const char *src, int lineno)
201 {
202 struct strbuf err = STRBUF_INIT;
203 strbuf_addf(&err, _("%.*s is not a valid attribute name"),
204 (int) len, name);
205 fprintf(stderr, "%s: %s:%d\n", err.buf, src, lineno);
206 strbuf_release(&err);
207 }
208
209 /*
210 * Given a 'name', lookup and return the corresponding attribute in the global
211 * dictionary. If no entry is found, create a new attribute and store it in
212 * the dictionary.
213 */
214 static const struct git_attr *git_attr_internal(const char *name, int namelen)
215 {
216 struct git_attr *a;
217
218 if (!attr_name_valid(name, namelen))
219 return NULL;
220
221 hashmap_lock(&g_attr_hashmap);
222
223 a = attr_hashmap_get(&g_attr_hashmap, name, namelen);
224
225 if (!a) {
226 FLEX_ALLOC_MEM(a, name, name, namelen);
227 a->attr_nr = hashmap_get_size(&g_attr_hashmap.map);
228
229 attr_hashmap_add(&g_attr_hashmap, a->name, namelen, a);
230 assert(a->attr_nr ==
231 (hashmap_get_size(&g_attr_hashmap.map) - 1));
232 }
233
234 hashmap_unlock(&g_attr_hashmap);
235
236 return a;
237 }
238
239 const struct git_attr *git_attr(const char *name)
240 {
241 return git_attr_internal(name, strlen(name));
242 }
243
244 /* What does a matched pattern decide? */
245 struct attr_state {
246 const struct git_attr *attr;
247 const char *setto;
248 };
249
250 struct pattern {
251 const char *pattern;
252 int patternlen;
253 int nowildcardlen;
254 unsigned flags; /* PATTERN_FLAG_* */
255 };
256
257 /*
258 * One rule, as from a .gitattributes file.
259 *
260 * If is_macro is true, then u.attr is a pointer to the git_attr being
261 * defined.
262 *
263 * If is_macro is false, then u.pat is the filename pattern to which the
264 * rule applies.
265 *
266 * In either case, num_attr is the number of attributes affected by
267 * this rule, and state is an array listing them. The attributes are
268 * listed as they appear in the file (macros unexpanded).
269 */
270 struct match_attr {
271 union {
272 struct pattern pat;
273 const struct git_attr *attr;
274 } u;
275 char is_macro;
276 unsigned num_attr;
277 struct attr_state state[FLEX_ARRAY];
278 };
279
280 static const char blank[] = " \t\r\n";
281
282 /* Flags usable in read_attr() and parse_attr_line() family of functions. */
283 #define READ_ATTR_MACRO_OK (1<<0)
284 #define READ_ATTR_NOFOLLOW (1<<1)
285
286 /*
287 * Parse a whitespace-delimited attribute state (i.e., "attr",
288 * "-attr", "!attr", or "attr=value") from the string starting at src.
289 * If e is not NULL, write the results to *e. Return a pointer to the
290 * remainder of the string (with leading whitespace removed), or NULL
291 * if there was an error.
292 */
293 static const char *parse_attr(const char *src, int lineno, const char *cp,
294 struct attr_state *e)
295 {
296 const char *ep, *equals;
297 int len;
298
299 ep = cp + strcspn(cp, blank);
300 equals = strchr(cp, '=');
301 if (equals && ep < equals)
302 equals = NULL;
303 if (equals)
304 len = equals - cp;
305 else
306 len = ep - cp;
307 if (!e) {
308 if (*cp == '-' || *cp == '!') {
309 cp++;
310 len--;
311 }
312 if (!attr_name_valid(cp, len)) {
313 report_invalid_attr(cp, len, src, lineno);
314 return NULL;
315 }
316 } else {
317 /*
318 * As this function is always called twice, once with
319 * e == NULL in the first pass and then e != NULL in
320 * the second pass, no need for attr_name_valid()
321 * check here.
322 */
323 if (*cp == '-' || *cp == '!') {
324 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
325 cp++;
326 len--;
327 }
328 else if (!equals)
329 e->setto = ATTR__TRUE;
330 else {
331 e->setto = xmemdupz(equals + 1, ep - equals - 1);
332 }
333 e->attr = git_attr_internal(cp, len);
334 }
335 return ep + strspn(ep, blank);
336 }
337
338 static struct match_attr *parse_attr_line(const char *line, const char *src,
339 int lineno, unsigned flags)
340 {
341 int namelen;
342 int num_attr, i;
343 const char *cp, *name, *states;
344 struct match_attr *res = NULL;
345 int is_macro;
346 struct strbuf pattern = STRBUF_INIT;
347
348 cp = line + strspn(line, blank);
349 if (!*cp || *cp == '#')
350 return NULL;
351 name = cp;
352
353 if (*cp == '"' && !unquote_c_style(&pattern, name, &states)) {
354 name = pattern.buf;
355 namelen = pattern.len;
356 } else {
357 namelen = strcspn(name, blank);
358 states = name + namelen;
359 }
360
361 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
362 starts_with(name, ATTRIBUTE_MACRO_PREFIX)) {
363 if (!(flags & READ_ATTR_MACRO_OK)) {
364 fprintf_ln(stderr, _("%s not allowed: %s:%d"),
365 name, src, lineno);
366 goto fail_return;
367 }
368 is_macro = 1;
369 name += strlen(ATTRIBUTE_MACRO_PREFIX);
370 name += strspn(name, blank);
371 namelen = strcspn(name, blank);
372 if (!attr_name_valid(name, namelen)) {
373 report_invalid_attr(name, namelen, src, lineno);
374 goto fail_return;
375 }
376 }
377 else
378 is_macro = 0;
379
380 states += strspn(states, blank);
381
382 /* First pass to count the attr_states */
383 for (cp = states, num_attr = 0; *cp; num_attr++) {
384 cp = parse_attr(src, lineno, cp, NULL);
385 if (!cp)
386 goto fail_return;
387 }
388
389 res = xcalloc(1,
390 sizeof(*res) +
391 sizeof(struct attr_state) * num_attr +
392 (is_macro ? 0 : namelen + 1));
393 if (is_macro) {
394 res->u.attr = git_attr_internal(name, namelen);
395 } else {
396 char *p = (char *)&(res->state[num_attr]);
397 memcpy(p, name, namelen);
398 res->u.pat.pattern = p;
399 parse_path_pattern(&res->u.pat.pattern,
400 &res->u.pat.patternlen,
401 &res->u.pat.flags,
402 &res->u.pat.nowildcardlen);
403 if (res->u.pat.flags & PATTERN_FLAG_NEGATIVE) {
404 warning(_("Negative patterns are ignored in git attributes\n"
405 "Use '\\!' for literal leading exclamation."));
406 goto fail_return;
407 }
408 }
409 res->is_macro = is_macro;
410 res->num_attr = num_attr;
411
412 /* Second pass to fill the attr_states */
413 for (cp = states, i = 0; *cp; i++) {
414 cp = parse_attr(src, lineno, cp, &(res->state[i]));
415 }
416
417 strbuf_release(&pattern);
418 return res;
419
420 fail_return:
421 strbuf_release(&pattern);
422 free(res);
423 return NULL;
424 }
425
426 /*
427 * Like info/exclude and .gitignore, the attribute information can
428 * come from many places.
429 *
430 * (1) .gitattributes file of the same directory;
431 * (2) .gitattributes file of the parent directory if (1) does not have
432 * any match; this goes recursively upwards, just like .gitignore.
433 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
434 *
435 * In the same file, later entries override the earlier match, so in the
436 * global list, we would have entries from info/attributes the earliest
437 * (reading the file from top to bottom), .gitattributes of the root
438 * directory (again, reading the file from top to bottom) down to the
439 * current directory, and then scan the list backwards to find the first match.
440 * This is exactly the same as what is_excluded() does in dir.c to deal with
441 * .gitignore file and info/excludes file as a fallback.
442 */
443
444 struct attr_stack {
445 struct attr_stack *prev;
446 char *origin;
447 size_t originlen;
448 unsigned num_matches;
449 unsigned alloc;
450 struct match_attr **attrs;
451 };
452
453 static void attr_stack_free(struct attr_stack *e)
454 {
455 int i;
456 free(e->origin);
457 for (i = 0; i < e->num_matches; i++) {
458 struct match_attr *a = e->attrs[i];
459 int j;
460 for (j = 0; j < a->num_attr; j++) {
461 const char *setto = a->state[j].setto;
462 if (setto == ATTR__TRUE ||
463 setto == ATTR__FALSE ||
464 setto == ATTR__UNSET ||
465 setto == ATTR__UNKNOWN)
466 ;
467 else
468 free((char *) setto);
469 }
470 free(a);
471 }
472 free(e->attrs);
473 free(e);
474 }
475
476 static void drop_attr_stack(struct attr_stack **stack)
477 {
478 while (*stack) {
479 struct attr_stack *elem = *stack;
480 *stack = elem->prev;
481 attr_stack_free(elem);
482 }
483 }
484
485 /* List of all attr_check structs; access should be surrounded by mutex */
486 static struct check_vector {
487 size_t nr;
488 size_t alloc;
489 struct attr_check **checks;
490 pthread_mutex_t mutex;
491 } check_vector;
492
493 static inline void vector_lock(void)
494 {
495 pthread_mutex_lock(&check_vector.mutex);
496 }
497
498 static inline void vector_unlock(void)
499 {
500 pthread_mutex_unlock(&check_vector.mutex);
501 }
502
503 static void check_vector_add(struct attr_check *c)
504 {
505 vector_lock();
506
507 ALLOC_GROW(check_vector.checks,
508 check_vector.nr + 1,
509 check_vector.alloc);
510 check_vector.checks[check_vector.nr++] = c;
511
512 vector_unlock();
513 }
514
515 static void check_vector_remove(struct attr_check *check)
516 {
517 int i;
518
519 vector_lock();
520
521 /* Find entry */
522 for (i = 0; i < check_vector.nr; i++)
523 if (check_vector.checks[i] == check)
524 break;
525
526 if (i >= check_vector.nr)
527 BUG("no entry found");
528
529 /* shift entries over */
530 for (; i < check_vector.nr - 1; i++)
531 check_vector.checks[i] = check_vector.checks[i + 1];
532
533 check_vector.nr--;
534
535 vector_unlock();
536 }
537
538 /* Iterate through all attr_check instances and drop their stacks */
539 static void drop_all_attr_stacks(void)
540 {
541 int i;
542
543 vector_lock();
544
545 for (i = 0; i < check_vector.nr; i++) {
546 drop_attr_stack(&check_vector.checks[i]->stack);
547 }
548
549 vector_unlock();
550 }
551
552 struct attr_check *attr_check_alloc(void)
553 {
554 struct attr_check *c = xcalloc(1, sizeof(struct attr_check));
555
556 /* save pointer to the check struct */
557 check_vector_add(c);
558
559 return c;
560 }
561
562 struct attr_check *attr_check_initl(const char *one, ...)
563 {
564 struct attr_check *check;
565 int cnt;
566 va_list params;
567 const char *param;
568
569 va_start(params, one);
570 for (cnt = 1; (param = va_arg(params, const char *)) != NULL; cnt++)
571 ;
572 va_end(params);
573
574 check = attr_check_alloc();
575 check->nr = cnt;
576 check->alloc = cnt;
577 CALLOC_ARRAY(check->items, cnt);
578
579 check->items[0].attr = git_attr(one);
580 va_start(params, one);
581 for (cnt = 1; cnt < check->nr; cnt++) {
582 const struct git_attr *attr;
583 param = va_arg(params, const char *);
584 if (!param)
585 BUG("counted %d != ended at %d",
586 check->nr, cnt);
587 attr = git_attr(param);
588 if (!attr)
589 BUG("%s: not a valid attribute name", param);
590 check->items[cnt].attr = attr;
591 }
592 va_end(params);
593 return check;
594 }
595
596 struct attr_check *attr_check_dup(const struct attr_check *check)
597 {
598 struct attr_check *ret;
599
600 if (!check)
601 return NULL;
602
603 ret = attr_check_alloc();
604
605 ret->nr = check->nr;
606 ret->alloc = check->alloc;
607 ALLOC_ARRAY(ret->items, ret->nr);
608 COPY_ARRAY(ret->items, check->items, ret->nr);
609
610 return ret;
611 }
612
613 struct attr_check_item *attr_check_append(struct attr_check *check,
614 const struct git_attr *attr)
615 {
616 struct attr_check_item *item;
617
618 ALLOC_GROW(check->items, check->nr + 1, check->alloc);
619 item = &check->items[check->nr++];
620 item->attr = attr;
621 return item;
622 }
623
624 void attr_check_reset(struct attr_check *check)
625 {
626 check->nr = 0;
627 }
628
629 void attr_check_clear(struct attr_check *check)
630 {
631 FREE_AND_NULL(check->items);
632 check->alloc = 0;
633 check->nr = 0;
634
635 FREE_AND_NULL(check->all_attrs);
636 check->all_attrs_nr = 0;
637
638 drop_attr_stack(&check->stack);
639 }
640
641 void attr_check_free(struct attr_check *check)
642 {
643 if (check) {
644 /* Remove check from the check vector */
645 check_vector_remove(check);
646
647 attr_check_clear(check);
648 free(check);
649 }
650 }
651
652 static const char *builtin_attr[] = {
653 "[attr]binary -diff -merge -text",
654 NULL,
655 };
656
657 static void handle_attr_line(struct attr_stack *res,
658 const char *line,
659 const char *src,
660 int lineno,
661 unsigned flags)
662 {
663 struct match_attr *a;
664
665 a = parse_attr_line(line, src, lineno, flags);
666 if (!a)
667 return;
668 ALLOC_GROW(res->attrs, res->num_matches + 1, res->alloc);
669 res->attrs[res->num_matches++] = a;
670 }
671
672 static struct attr_stack *read_attr_from_array(const char **list)
673 {
674 struct attr_stack *res;
675 const char *line;
676 int lineno = 0;
677
678 CALLOC_ARRAY(res, 1);
679 while ((line = *(list++)) != NULL)
680 handle_attr_line(res, line, "[builtin]", ++lineno,
681 READ_ATTR_MACRO_OK);
682 return res;
683 }
684
685 /*
686 * Callers into the attribute system assume there is a single, system-wide
687 * global state where attributes are read from and when the state is flipped by
688 * calling git_attr_set_direction(), the stack frames that have been
689 * constructed need to be discarded so that subsequent calls into the
690 * attribute system will lazily read from the right place. Since changing
691 * direction causes a global paradigm shift, it should not ever be called while
692 * another thread could potentially be calling into the attribute system.
693 */
694 static enum git_attr_direction direction;
695
696 void git_attr_set_direction(enum git_attr_direction new_direction)
697 {
698 if (is_bare_repository() && new_direction != GIT_ATTR_INDEX)
699 BUG("non-INDEX attr direction in a bare repo");
700
701 if (new_direction != direction)
702 drop_all_attr_stacks();
703
704 direction = new_direction;
705 }
706
707 static struct attr_stack *read_attr_from_file(const char *path, unsigned flags)
708 {
709 int fd;
710 FILE *fp;
711 struct attr_stack *res;
712 char buf[2048];
713 int lineno = 0;
714
715 if (flags & READ_ATTR_NOFOLLOW)
716 fd = open_nofollow(path, O_RDONLY);
717 else
718 fd = open(path, O_RDONLY);
719
720 if (fd < 0) {
721 warn_on_fopen_errors(path);
722 return NULL;
723 }
724 fp = xfdopen(fd, "r");
725
726 CALLOC_ARRAY(res, 1);
727 while (fgets(buf, sizeof(buf), fp)) {
728 char *bufp = buf;
729 if (!lineno)
730 skip_utf8_bom(&bufp, strlen(bufp));
731 handle_attr_line(res, bufp, path, ++lineno, flags);
732 }
733 fclose(fp);
734 return res;
735 }
736
737 static struct attr_stack *read_attr_from_index(struct index_state *istate,
738 const char *path,
739 unsigned flags)
740 {
741 struct attr_stack *res;
742 char *buf, *sp;
743 int lineno = 0;
744
745 if (!istate)
746 return NULL;
747
748 /*
749 * The .gitattributes file only applies to files within its
750 * parent directory. In the case of cone-mode sparse-checkout,
751 * the .gitattributes file is sparse if and only if all paths
752 * within that directory are also sparse. Thus, don't load the
753 * .gitattributes file since it will not matter.
754 *
755 * In the case of a sparse index, it is critical that we don't go
756 * looking for a .gitattributes file, as doing so would cause the
757 * index to expand.
758 */
759 if (!path_in_cone_mode_sparse_checkout(path, istate))
760 return NULL;
761
762 buf = read_blob_data_from_index(istate, path, NULL);
763 if (!buf)
764 return NULL;
765
766 CALLOC_ARRAY(res, 1);
767 for (sp = buf; *sp; ) {
768 char *ep;
769 int more;
770
771 ep = strchrnul(sp, '\n');
772 more = (*ep == '\n');
773 *ep = '\0';
774 handle_attr_line(res, sp, path, ++lineno, flags);
775 sp = ep + more;
776 }
777 free(buf);
778 return res;
779 }
780
781 static struct attr_stack *read_attr(struct index_state *istate,
782 const char *path, unsigned flags)
783 {
784 struct attr_stack *res = NULL;
785
786 if (direction == GIT_ATTR_INDEX) {
787 res = read_attr_from_index(istate, path, flags);
788 } else if (!is_bare_repository()) {
789 if (direction == GIT_ATTR_CHECKOUT) {
790 res = read_attr_from_index(istate, path, flags);
791 if (!res)
792 res = read_attr_from_file(path, flags);
793 } else if (direction == GIT_ATTR_CHECKIN) {
794 res = read_attr_from_file(path, flags);
795 if (!res)
796 /*
797 * There is no checked out .gitattributes file
798 * there, but we might have it in the index.
799 * We allow operation in a sparsely checked out
800 * work tree, so read from it.
801 */
802 res = read_attr_from_index(istate, path, flags);
803 }
804 }
805
806 if (!res)
807 CALLOC_ARRAY(res, 1);
808 return res;
809 }
810
811 #if DEBUG_ATTR
812 static void debug_info(const char *what, struct attr_stack *elem)
813 {
814 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
815 }
816 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
817 {
818 const char *value = v;
819
820 if (ATTR_TRUE(value))
821 value = "set";
822 else if (ATTR_FALSE(value))
823 value = "unset";
824 else if (ATTR_UNSET(value))
825 value = "unspecified";
826
827 fprintf(stderr, "%s: %s => %s (%s)\n",
828 what, attr->name, (char *) value, match);
829 }
830 #define debug_push(a) debug_info("push", (a))
831 #define debug_pop(a) debug_info("pop", (a))
832 #else
833 #define debug_push(a) do { ; } while (0)
834 #define debug_pop(a) do { ; } while (0)
835 #define debug_set(a,b,c,d) do { ; } while (0)
836 #endif /* DEBUG_ATTR */
837
838 static const char *git_etc_gitattributes(void)
839 {
840 static const char *system_wide;
841 if (!system_wide)
842 system_wide = system_path(ETC_GITATTRIBUTES);
843 return system_wide;
844 }
845
846 static const char *get_home_gitattributes(void)
847 {
848 if (!git_attributes_file)
849 git_attributes_file = xdg_config_home("attributes");
850
851 return git_attributes_file;
852 }
853
854 static int git_attr_system(void)
855 {
856 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
857 }
858
859 static GIT_PATH_FUNC(git_path_info_attributes, INFOATTRIBUTES_FILE)
860
861 static void push_stack(struct attr_stack **attr_stack_p,
862 struct attr_stack *elem, char *origin, size_t originlen)
863 {
864 if (elem) {
865 elem->origin = origin;
866 if (origin)
867 elem->originlen = originlen;
868 elem->prev = *attr_stack_p;
869 *attr_stack_p = elem;
870 }
871 }
872
873 static void bootstrap_attr_stack(struct index_state *istate,
874 struct attr_stack **stack)
875 {
876 struct attr_stack *e;
877 unsigned flags = READ_ATTR_MACRO_OK;
878
879 if (*stack)
880 return;
881
882 /* builtin frame */
883 e = read_attr_from_array(builtin_attr);
884 push_stack(stack, e, NULL, 0);
885
886 /* system-wide frame */
887 if (git_attr_system()) {
888 e = read_attr_from_file(git_etc_gitattributes(), flags);
889 push_stack(stack, e, NULL, 0);
890 }
891
892 /* home directory */
893 if (get_home_gitattributes()) {
894 e = read_attr_from_file(get_home_gitattributes(), flags);
895 push_stack(stack, e, NULL, 0);
896 }
897
898 /* root directory */
899 e = read_attr(istate, GITATTRIBUTES_FILE, flags | READ_ATTR_NOFOLLOW);
900 push_stack(stack, e, xstrdup(""), 0);
901
902 /* info frame */
903 if (startup_info->have_repository)
904 e = read_attr_from_file(git_path_info_attributes(), flags);
905 else
906 e = NULL;
907 if (!e)
908 CALLOC_ARRAY(e, 1);
909 push_stack(stack, e, NULL, 0);
910 }
911
912 static void prepare_attr_stack(struct index_state *istate,
913 const char *path, int dirlen,
914 struct attr_stack **stack)
915 {
916 struct attr_stack *info;
917 struct strbuf pathbuf = STRBUF_INIT;
918
919 /*
920 * At the bottom of the attribute stack is the built-in
921 * set of attribute definitions, followed by the contents
922 * of $(prefix)/etc/gitattributes and a file specified by
923 * core.attributesfile. Then, contents from
924 * .gitattributes files from directories closer to the
925 * root to the ones in deeper directories are pushed
926 * to the stack. Finally, at the very top of the stack
927 * we always keep the contents of $GIT_DIR/info/attributes.
928 *
929 * When checking, we use entries from near the top of the
930 * stack, preferring $GIT_DIR/info/attributes, then
931 * .gitattributes in deeper directories to shallower ones,
932 * and finally use the built-in set as the default.
933 */
934 bootstrap_attr_stack(istate, stack);
935
936 /*
937 * Pop the "info" one that is always at the top of the stack.
938 */
939 info = *stack;
940 *stack = info->prev;
941
942 /*
943 * Pop the ones from directories that are not the prefix of
944 * the path we are checking. Break out of the loop when we see
945 * the root one (whose origin is an empty string "") or the builtin
946 * one (whose origin is NULL) without popping it.
947 */
948 while ((*stack)->origin) {
949 int namelen = (*stack)->originlen;
950 struct attr_stack *elem;
951
952 elem = *stack;
953 if (namelen <= dirlen &&
954 !strncmp(elem->origin, path, namelen) &&
955 (!namelen || path[namelen] == '/'))
956 break;
957
958 debug_pop(elem);
959 *stack = elem->prev;
960 attr_stack_free(elem);
961 }
962
963 /*
964 * bootstrap_attr_stack() should have added, and the
965 * above loop should have stopped before popping, the
966 * root element whose attr_stack->origin is set to an
967 * empty string.
968 */
969 assert((*stack)->origin);
970
971 strbuf_addstr(&pathbuf, (*stack)->origin);
972 /* Build up to the directory 'path' is in */
973 while (pathbuf.len < dirlen) {
974 size_t len = pathbuf.len;
975 struct attr_stack *next;
976 char *origin;
977
978 /* Skip path-separator */
979 if (len < dirlen && is_dir_sep(path[len]))
980 len++;
981 /* Find the end of the next component */
982 while (len < dirlen && !is_dir_sep(path[len]))
983 len++;
984
985 if (pathbuf.len > 0)
986 strbuf_addch(&pathbuf, '/');
987 strbuf_add(&pathbuf, path + pathbuf.len, (len - pathbuf.len));
988 strbuf_addf(&pathbuf, "/%s", GITATTRIBUTES_FILE);
989
990 next = read_attr(istate, pathbuf.buf, READ_ATTR_NOFOLLOW);
991
992 /* reset the pathbuf to not include "/.gitattributes" */
993 strbuf_setlen(&pathbuf, len);
994
995 origin = xstrdup(pathbuf.buf);
996 push_stack(stack, next, origin, len);
997 }
998
999 /*
1000 * Finally push the "info" one at the top of the stack.
1001 */
1002 push_stack(stack, info, NULL, 0);
1003
1004 strbuf_release(&pathbuf);
1005 }
1006
1007 static int path_matches(const char *pathname, int pathlen,
1008 int basename_offset,
1009 const struct pattern *pat,
1010 const char *base, int baselen)
1011 {
1012 const char *pattern = pat->pattern;
1013 int prefix = pat->nowildcardlen;
1014 int isdir = (pathlen && pathname[pathlen - 1] == '/');
1015
1016 if ((pat->flags & PATTERN_FLAG_MUSTBEDIR) && !isdir)
1017 return 0;
1018
1019 if (pat->flags & PATTERN_FLAG_NODIR) {
1020 return match_basename(pathname + basename_offset,
1021 pathlen - basename_offset - isdir,
1022 pattern, prefix,
1023 pat->patternlen, pat->flags);
1024 }
1025 return match_pathname(pathname, pathlen - isdir,
1026 base, baselen,
1027 pattern, prefix, pat->patternlen, pat->flags);
1028 }
1029
1030 static int macroexpand_one(struct all_attrs_item *all_attrs, int nr, int rem);
1031
1032 static int fill_one(const char *what, struct all_attrs_item *all_attrs,
1033 const struct match_attr *a, int rem)
1034 {
1035 int i;
1036
1037 for (i = a->num_attr - 1; rem > 0 && i >= 0; i--) {
1038 const struct git_attr *attr = a->state[i].attr;
1039 const char **n = &(all_attrs[attr->attr_nr].value);
1040 const char *v = a->state[i].setto;
1041
1042 if (*n == ATTR__UNKNOWN) {
1043 debug_set(what,
1044 a->is_macro ? a->u.attr->name : a->u.pat.pattern,
1045 attr, v);
1046 *n = v;
1047 rem--;
1048 rem = macroexpand_one(all_attrs, attr->attr_nr, rem);
1049 }
1050 }
1051 return rem;
1052 }
1053
1054 static int fill(const char *path, int pathlen, int basename_offset,
1055 const struct attr_stack *stack,
1056 struct all_attrs_item *all_attrs, int rem)
1057 {
1058 for (; rem > 0 && stack; stack = stack->prev) {
1059 int i;
1060 const char *base = stack->origin ? stack->origin : "";
1061
1062 for (i = stack->num_matches - 1; 0 < rem && 0 <= i; i--) {
1063 const struct match_attr *a = stack->attrs[i];
1064 if (a->is_macro)
1065 continue;
1066 if (path_matches(path, pathlen, basename_offset,
1067 &a->u.pat, base, stack->originlen))
1068 rem = fill_one("fill", all_attrs, a, rem);
1069 }
1070 }
1071
1072 return rem;
1073 }
1074
1075 static int macroexpand_one(struct all_attrs_item *all_attrs, int nr, int rem)
1076 {
1077 const struct all_attrs_item *item = &all_attrs[nr];
1078
1079 if (item->macro && item->value == ATTR__TRUE)
1080 return fill_one("expand", all_attrs, item->macro, rem);
1081 else
1082 return rem;
1083 }
1084
1085 /*
1086 * Marks the attributes which are macros based on the attribute stack.
1087 * This prevents having to search through the attribute stack each time
1088 * a macro needs to be expanded during the fill stage.
1089 */
1090 static void determine_macros(struct all_attrs_item *all_attrs,
1091 const struct attr_stack *stack)
1092 {
1093 for (; stack; stack = stack->prev) {
1094 int i;
1095 for (i = stack->num_matches - 1; i >= 0; i--) {
1096 const struct match_attr *ma = stack->attrs[i];
1097 if (ma->is_macro) {
1098 int n = ma->u.attr->attr_nr;
1099 if (!all_attrs[n].macro) {
1100 all_attrs[n].macro = ma;
1101 }
1102 }
1103 }
1104 }
1105 }
1106
1107 /*
1108 * Collect attributes for path into the array pointed to by check->all_attrs.
1109 * If check->check_nr is non-zero, only attributes in check[] are collected.
1110 * Otherwise all attributes are collected.
1111 */
1112 static void collect_some_attrs(struct index_state *istate,
1113 const char *path,
1114 struct attr_check *check)
1115 {
1116 int pathlen, rem, dirlen;
1117 const char *cp, *last_slash = NULL;
1118 int basename_offset;
1119
1120 for (cp = path; *cp; cp++) {
1121 if (*cp == '/' && cp[1])
1122 last_slash = cp;
1123 }
1124 pathlen = cp - path;
1125 if (last_slash) {
1126 basename_offset = last_slash + 1 - path;
1127 dirlen = last_slash - path;
1128 } else {
1129 basename_offset = 0;
1130 dirlen = 0;
1131 }
1132
1133 prepare_attr_stack(istate, path, dirlen, &check->stack);
1134 all_attrs_init(&g_attr_hashmap, check);
1135 determine_macros(check->all_attrs, check->stack);
1136
1137 rem = check->all_attrs_nr;
1138 fill(path, pathlen, basename_offset, check->stack, check->all_attrs, rem);
1139 }
1140
1141 void git_check_attr(struct index_state *istate,
1142 const char *path,
1143 struct attr_check *check)
1144 {
1145 int i;
1146
1147 collect_some_attrs(istate, path, check);
1148
1149 for (i = 0; i < check->nr; i++) {
1150 size_t n = check->items[i].attr->attr_nr;
1151 const char *value = check->all_attrs[n].value;
1152 if (value == ATTR__UNKNOWN)
1153 value = ATTR__UNSET;
1154 check->items[i].value = value;
1155 }
1156 }
1157
1158 void git_all_attrs(struct index_state *istate,
1159 const char *path, struct attr_check *check)
1160 {
1161 int i;
1162
1163 attr_check_reset(check);
1164 collect_some_attrs(istate, path, check);
1165
1166 for (i = 0; i < check->all_attrs_nr; i++) {
1167 const char *name = check->all_attrs[i].attr->name;
1168 const char *value = check->all_attrs[i].value;
1169 struct attr_check_item *item;
1170 if (value == ATTR__UNSET || value == ATTR__UNKNOWN)
1171 continue;
1172 item = attr_check_append(check, git_attr(name));
1173 item->value = value;
1174 }
1175 }
1176
1177 void attr_start(void)
1178 {
1179 pthread_mutex_init(&g_attr_hashmap.mutex, NULL);
1180 pthread_mutex_init(&check_vector.mutex, NULL);
1181 }