hashmap: add API to disable item counting when threaded
[git/git.git] / attr.c
CommitLineData
86ab7f0c
MH
1/*
2 * Handle git attributes. See gitattributes(5) for a description of
3 * the file syntax, and Documentation/technical/api-gitattributes.txt
4 * for a description of the API.
5 *
6 * One basic design decision here is that we are not going to support
7 * an insanely large number of attributes.
8 */
9
06f33c17 10#define NO_THE_INDEX_COMPATIBILITY_MACROS
d0bfd026 11#include "cache.h"
b2141fc1 12#include "config.h"
6df42ab9 13#include "exec_cmd.h"
d0bfd026 14#include "attr.h"
6eba6210 15#include "dir.h"
27547e5f 16#include "utf8.h"
860a74d9 17#include "quote.h"
1a600b75 18#include "thread-utils.h"
d0bfd026 19
a5e92abd
JH
20const char git_attr__true[] = "(builtin)true";
21const char git_attr__false[] = "\0(builtin)false";
22static const char git_attr__unknown[] = "(builtin)unknown";
23#define ATTR__TRUE git_attr__true
24#define ATTR__FALSE git_attr__false
25#define ATTR__UNSET NULL
26#define ATTR__UNKNOWN git_attr__unknown
515106fa 27
d0bfd026
JH
28#ifndef DEBUG_ATTR
29#define DEBUG_ATTR 0
30#endif
31
32struct git_attr {
1a600b75 33 int attr_nr; /* unique attribute number */
1a600b75 34 char name[FLEX_ARRAY]; /* attribute name */
d0bfd026 35};
7d42ec54 36
ec4d77aa 37const char *git_attr_name(const struct git_attr *attr)
352404ac
MH
38{
39 return attr->name;
40}
41
1a600b75
BW
42struct attr_hashmap {
43 struct hashmap map;
44#ifndef NO_PTHREADS
45 pthread_mutex_t mutex;
46#endif
47};
48
49static inline void hashmap_lock(struct attr_hashmap *map)
50{
51#ifndef NO_PTHREADS
52 pthread_mutex_lock(&map->mutex);
53#endif
54}
55
56static inline void hashmap_unlock(struct attr_hashmap *map)
d0bfd026 57{
1a600b75
BW
58#ifndef NO_PTHREADS
59 pthread_mutex_unlock(&map->mutex);
60#endif
61}
d0bfd026 62
1a600b75
BW
63/*
64 * The global dictionary of all interned attributes. This
65 * is a singleton object which is shared between threads.
66 * Access to this dictionary must be surrounded with a mutex.
67 */
68static struct attr_hashmap g_attr_hashmap;
69
70/* The container for objects stored in "struct attr_hashmap" */
71struct attr_hash_entry {
72 struct hashmap_entry ent; /* must be the first member! */
73 const char *key; /* the key; memory should be owned by value */
74 size_t keylen; /* length of the key */
75 void *value; /* the stored value */
76};
77
78/* attr_hashmap comparison function */
201c14e3
SB
79static int attr_hash_entry_cmp(const void *unused_cmp_data,
80 const void *entry,
81 const void *entry_or_key,
82 const void *unused_keydata)
1a600b75 83{
201c14e3
SB
84 const struct attr_hash_entry *a = entry;
85 const struct attr_hash_entry *b = entry_or_key;
1a600b75
BW
86 return (a->keylen != b->keylen) || strncmp(a->key, b->key, a->keylen);
87}
88
89/* Initialize an 'attr_hashmap' object */
90static void attr_hashmap_init(struct attr_hashmap *map)
91{
201c14e3 92 hashmap_init(&map->map, attr_hash_entry_cmp, NULL, 0);
1a600b75
BW
93}
94
95/*
96 * Retrieve the 'value' stored in a hashmap given the provided 'key'.
97 * If there is no matching entry, return NULL.
98 */
99static void *attr_hashmap_get(struct attr_hashmap *map,
100 const char *key, size_t keylen)
101{
102 struct attr_hash_entry k;
103 struct attr_hash_entry *e;
104
105 if (!map->map.tablesize)
106 attr_hashmap_init(map);
107
108 hashmap_entry_init(&k, memhash(key, keylen));
109 k.key = key;
110 k.keylen = keylen;
111 e = hashmap_get(&map->map, &k, NULL);
112
113 return e ? e->value : NULL;
114}
115
116/* Add 'value' to a hashmap based on the provided 'key'. */
117static void attr_hashmap_add(struct attr_hashmap *map,
118 const char *key, size_t keylen,
119 void *value)
120{
121 struct attr_hash_entry *e;
122
123 if (!map->map.tablesize)
124 attr_hashmap_init(map);
125
126 e = xmalloc(sizeof(struct attr_hash_entry));
127 hashmap_entry_init(e, memhash(key, keylen));
128 e->key = key;
129 e->keylen = keylen;
130 e->value = value;
131
132 hashmap_add(&map->map, e);
d0bfd026
JH
133}
134
685b2925
BW
135struct all_attrs_item {
136 const struct git_attr *attr;
137 const char *value;
60a12722
BW
138 /*
139 * If 'macro' is non-NULL, indicates that 'attr' is a macro based on
140 * the current attribute stack and contains a pointer to the match_attr
141 * definition of the macro
142 */
143 const struct match_attr *macro;
685b2925
BW
144};
145
146/*
147 * Reallocate and reinitialize the array of all attributes (which is used in
148 * the attribute collection process) in 'check' based on the global dictionary
149 * of attributes.
150 */
151static void all_attrs_init(struct attr_hashmap *map, struct attr_check *check)
152{
153 int i;
8b604d19 154 unsigned int size;
685b2925
BW
155
156 hashmap_lock(map);
157
8b604d19
JH
158 size = hashmap_get_size(&map->map);
159 if (size < check->all_attrs_nr)
685b2925
BW
160 die("BUG: interned attributes shouldn't be deleted");
161
162 /*
163 * If the number of attributes in the global dictionary has increased
164 * (or this attr_check instance doesn't have an initialized all_attrs
165 * field), reallocate the provided attr_check instance's all_attrs
166 * field and fill each entry with its corresponding git_attr.
167 */
8b604d19 168 if (size != check->all_attrs_nr) {
685b2925
BW
169 struct attr_hash_entry *e;
170 struct hashmap_iter iter;
171 hashmap_iter_init(&map->map, &iter);
172
8b604d19
JH
173 REALLOC_ARRAY(check->all_attrs, size);
174 check->all_attrs_nr = size;
685b2925
BW
175
176 while ((e = hashmap_iter_next(&iter))) {
177 const struct git_attr *a = e->value;
178 check->all_attrs[a->attr_nr].attr = a;
179 }
180 }
181
182 hashmap_unlock(map);
183
184 /*
185 * Re-initialize every entry in check->all_attrs.
186 * This re-initialization can live outside of the locked region since
187 * the attribute dictionary is no longer being accessed.
188 */
189 for (i = 0; i < check->all_attrs_nr; i++) {
190 check->all_attrs[i].value = ATTR__UNKNOWN;
60a12722 191 check->all_attrs[i].macro = NULL;
685b2925
BW
192 }
193}
194
428103c7 195static int attr_name_valid(const char *name, size_t namelen)
e4aee10a
JH
196{
197 /*
d42453ab
MH
198 * Attribute name cannot begin with '-' and must consist of
199 * characters from [-A-Za-z0-9_.].
e4aee10a 200 */
c0b13b21 201 if (namelen <= 0 || *name == '-')
428103c7 202 return 0;
e4aee10a
JH
203 while (namelen--) {
204 char ch = *name++;
205 if (! (ch == '-' || ch == '.' || ch == '_' ||
206 ('0' <= ch && ch <= '9') ||
207 ('a' <= ch && ch <= 'z') ||
208 ('A' <= ch && ch <= 'Z')) )
428103c7 209 return 0;
e4aee10a 210 }
428103c7
JH
211 return 1;
212}
213
214static void report_invalid_attr(const char *name, size_t len,
215 const char *src, int lineno)
216{
217 struct strbuf err = STRBUF_INIT;
218 strbuf_addf(&err, _("%.*s is not a valid attribute name"),
219 (int) len, name);
220 fprintf(stderr, "%s: %s:%d\n", err.buf, src, lineno);
221 strbuf_release(&err);
e4aee10a
JH
222}
223
1a600b75
BW
224/*
225 * Given a 'name', lookup and return the corresponding attribute in the global
226 * dictionary. If no entry is found, create a new attribute and store it in
227 * the dictionary.
228 */
e810e063 229static const struct git_attr *git_attr_internal(const char *name, int namelen)
d0bfd026 230{
d0bfd026
JH
231 struct git_attr *a;
232
1a600b75 233 if (!attr_name_valid(name, namelen))
e4aee10a
JH
234 return NULL;
235
1a600b75
BW
236 hashmap_lock(&g_attr_hashmap);
237
238 a = attr_hashmap_get(&g_attr_hashmap, name, namelen);
239
240 if (!a) {
241 FLEX_ALLOC_MEM(a, name, name, namelen);
8b604d19 242 a->attr_nr = hashmap_get_size(&g_attr_hashmap.map);
1a600b75
BW
243
244 attr_hashmap_add(&g_attr_hashmap, a->name, namelen, a);
8b604d19
JH
245 assert(a->attr_nr ==
246 (hashmap_get_size(&g_attr_hashmap.map) - 1));
1a600b75
BW
247 }
248
249 hashmap_unlock(&g_attr_hashmap);
f48fd688 250
d0bfd026
JH
251 return a;
252}
253
e810e063 254const struct git_attr *git_attr(const char *name)
7fb0eaa2
JH
255{
256 return git_attr_internal(name, strlen(name));
257}
258
515106fa 259/* What does a matched pattern decide? */
d0bfd026 260struct attr_state {
e810e063 261 const struct git_attr *attr;
a5e92abd 262 const char *setto;
d0bfd026
JH
263};
264
82dce998
NTND
265struct pattern {
266 const char *pattern;
267 int patternlen;
268 int nowildcardlen;
f8708998 269 unsigned flags; /* EXC_FLAG_* */
82dce998
NTND
270};
271
ba845b75
MH
272/*
273 * One rule, as from a .gitattributes file.
274 *
275 * If is_macro is true, then u.attr is a pointer to the git_attr being
276 * defined.
277 *
4894f4ff
JH
278 * If is_macro is false, then u.pat is the filename pattern to which the
279 * rule applies.
ba845b75
MH
280 *
281 * In either case, num_attr is the number of attributes affected by
282 * this rule, and state is an array listing them. The attributes are
283 * listed as they appear in the file (macros unexpanded).
284 */
d0bfd026 285struct match_attr {
f48fd688 286 union {
82dce998 287 struct pattern pat;
e810e063 288 const struct git_attr *attr;
f48fd688
JH
289 } u;
290 char is_macro;
d0bfd026
JH
291 unsigned num_attr;
292 struct attr_state state[FLEX_ARRAY];
293};
294
295static const char blank[] = " \t\r\n";
296
d1751298
MH
297/*
298 * Parse a whitespace-delimited attribute state (i.e., "attr",
299 * "-attr", "!attr", or "attr=value") from the string starting at src.
300 * If e is not NULL, write the results to *e. Return a pointer to the
301 * remainder of the string (with leading whitespace removed), or NULL
302 * if there was an error.
303 */
515106fa 304static const char *parse_attr(const char *src, int lineno, const char *cp,
d1751298 305 struct attr_state *e)
515106fa
JH
306{
307 const char *ep, *equals;
308 int len;
309
310 ep = cp + strcspn(cp, blank);
311 equals = strchr(cp, '=');
312 if (equals && ep < equals)
313 equals = NULL;
314 if (equals)
315 len = equals - cp;
316 else
317 len = ep - cp;
d1751298 318 if (!e) {
515106fa
JH
319 if (*cp == '-' || *cp == '!') {
320 cp++;
321 len--;
322 }
428103c7
JH
323 if (!attr_name_valid(cp, len)) {
324 report_invalid_attr(cp, len, src, lineno);
515106fa
JH
325 return NULL;
326 }
327 } else {
5a884019
JH
328 /*
329 * As this function is always called twice, once with
330 * e == NULL in the first pass and then e != NULL in
428103c7 331 * the second pass, no need for attr_name_valid()
5a884019
JH
332 * check here.
333 */
515106fa
JH
334 if (*cp == '-' || *cp == '!') {
335 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
336 cp++;
337 len--;
338 }
339 else if (!equals)
340 e->setto = ATTR__TRUE;
341 else {
182af834 342 e->setto = xmemdupz(equals + 1, ep - equals - 1);
515106fa 343 }
7fb0eaa2 344 e->attr = git_attr_internal(cp, len);
515106fa 345 }
515106fa
JH
346 return ep + strspn(ep, blank);
347}
348
f48fd688
JH
349static struct match_attr *parse_attr_line(const char *line, const char *src,
350 int lineno, int macro_ok)
d0bfd026
JH
351{
352 int namelen;
d68e1c18 353 int num_attr, i;
85c4a0d0 354 const char *cp, *name, *states;
515106fa 355 struct match_attr *res = NULL;
f48fd688 356 int is_macro;
860a74d9 357 struct strbuf pattern = STRBUF_INIT;
d0bfd026
JH
358
359 cp = line + strspn(line, blank);
360 if (!*cp || *cp == '#')
361 return NULL;
362 name = cp;
860a74d9
NTND
363
364 if (*cp == '"' && !unquote_c_style(&pattern, name, &states)) {
365 name = pattern.buf;
366 namelen = pattern.len;
367 } else {
368 namelen = strcspn(name, blank);
369 states = name + namelen;
370 }
371
f48fd688 372 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
59556548 373 starts_with(name, ATTRIBUTE_MACRO_PREFIX)) {
f48fd688
JH
374 if (!macro_ok) {
375 fprintf(stderr, "%s not allowed: %s:%d\n",
376 name, src, lineno);
62af8969 377 goto fail_return;
f48fd688
JH
378 }
379 is_macro = 1;
380 name += strlen(ATTRIBUTE_MACRO_PREFIX);
381 name += strspn(name, blank);
382 namelen = strcspn(name, blank);
428103c7
JH
383 if (!attr_name_valid(name, namelen)) {
384 report_invalid_attr(name, namelen, src, lineno);
62af8969 385 goto fail_return;
e4aee10a 386 }
f48fd688
JH
387 }
388 else
389 is_macro = 0;
d0bfd026 390
85c4a0d0
MH
391 states += strspn(states, blank);
392
d68e1c18
MH
393 /* First pass to count the attr_states */
394 for (cp = states, num_attr = 0; *cp; num_attr++) {
395 cp = parse_attr(src, lineno, cp, NULL);
396 if (!cp)
62af8969 397 goto fail_return;
d0bfd026 398 }
d68e1c18
MH
399
400 res = xcalloc(1,
401 sizeof(*res) +
402 sizeof(struct attr_state) * num_attr +
403 (is_macro ? 0 : namelen + 1));
fad32bcd 404 if (is_macro) {
d68e1c18 405 res->u.attr = git_attr_internal(name, namelen);
fad32bcd 406 } else {
82dce998
NTND
407 char *p = (char *)&(res->state[num_attr]);
408 memcpy(p, name, namelen);
409 res->u.pat.pattern = p;
410 parse_exclude_pattern(&res->u.pat.pattern,
411 &res->u.pat.patternlen,
412 &res->u.pat.flags,
413 &res->u.pat.nowildcardlen);
8b1bd024
TR
414 if (res->u.pat.flags & EXC_FLAG_NEGATIVE) {
415 warning(_("Negative patterns are ignored in git attributes\n"
416 "Use '\\!' for literal leading exclamation."));
62af8969 417 goto fail_return;
8b1bd024 418 }
d0bfd026 419 }
d68e1c18
MH
420 res->is_macro = is_macro;
421 res->num_attr = num_attr;
422
423 /* Second pass to fill the attr_states */
424 for (cp = states, i = 0; *cp; i++) {
425 cp = parse_attr(src, lineno, cp, &(res->state[i]));
426 }
427
860a74d9 428 strbuf_release(&pattern);
d0bfd026 429 return res;
62af8969
JH
430
431fail_return:
860a74d9 432 strbuf_release(&pattern);
62af8969
JH
433 free(res);
434 return NULL;
d0bfd026
JH
435}
436
437/*
438 * Like info/exclude and .gitignore, the attribute information can
439 * come from many places.
440 *
441 * (1) .gitattribute file of the same directory;
515106fa
JH
442 * (2) .gitattribute file of the parent directory if (1) does not have
443 * any match; this goes recursively upwards, just like .gitignore.
444 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
d0bfd026
JH
445 *
446 * In the same file, later entries override the earlier match, so in the
447 * global list, we would have entries from info/attributes the earliest
448 * (reading the file from top to bottom), .gitattribute of the root
449 * directory (again, reading the file from top to bottom) down to the
450 * current directory, and then scan the list backwards to find the first match.
6d24e7a8 451 * This is exactly the same as what is_excluded() does in dir.c to deal with
6f416cba 452 * .gitignore file and info/excludes file as a fallback.
d0bfd026
JH
453 */
454
dc81cf37 455struct attr_stack {
d0bfd026
JH
456 struct attr_stack *prev;
457 char *origin;
cd6a0b26 458 size_t originlen;
d0bfd026 459 unsigned num_matches;
a4413118 460 unsigned alloc;
d0bfd026 461 struct match_attr **attrs;
dc81cf37 462};
d0bfd026 463
dc81cf37 464static void attr_stack_free(struct attr_stack *e)
d0bfd026
JH
465{
466 int i;
467 free(e->origin);
515106fa
JH
468 for (i = 0; i < e->num_matches; i++) {
469 struct match_attr *a = e->attrs[i];
470 int j;
471 for (j = 0; j < a->num_attr; j++) {
a5e92abd 472 const char *setto = a->state[j].setto;
515106fa
JH
473 if (setto == ATTR__TRUE ||
474 setto == ATTR__FALSE ||
475 setto == ATTR__UNSET ||
476 setto == ATTR__UNKNOWN)
477 ;
478 else
4b25d091 479 free((char *) setto);
515106fa
JH
480 }
481 free(a);
482 }
37475f97 483 free(e->attrs);
d0bfd026
JH
484 free(e);
485}
486
dc81cf37
BW
487static void drop_attr_stack(struct attr_stack **stack)
488{
489 while (*stack) {
490 struct attr_stack *elem = *stack;
491 *stack = elem->prev;
492 attr_stack_free(elem);
493 }
494}
495
496/* List of all attr_check structs; access should be surrounded by mutex */
497static struct check_vector {
498 size_t nr;
499 size_t alloc;
500 struct attr_check **checks;
501#ifndef NO_PTHREADS
502 pthread_mutex_t mutex;
503#endif
504} check_vector;
505
506static inline void vector_lock(void)
507{
508#ifndef NO_PTHREADS
509 pthread_mutex_lock(&check_vector.mutex);
510#endif
511}
512
513static inline void vector_unlock(void)
514{
515#ifndef NO_PTHREADS
516 pthread_mutex_unlock(&check_vector.mutex);
517#endif
518}
519
520static void check_vector_add(struct attr_check *c)
521{
522 vector_lock();
523
524 ALLOC_GROW(check_vector.checks,
525 check_vector.nr + 1,
526 check_vector.alloc);
527 check_vector.checks[check_vector.nr++] = c;
528
529 vector_unlock();
530}
531
532static void check_vector_remove(struct attr_check *check)
533{
534 int i;
535
536 vector_lock();
537
538 /* Find entry */
539 for (i = 0; i < check_vector.nr; i++)
540 if (check_vector.checks[i] == check)
541 break;
542
543 if (i >= check_vector.nr)
544 die("BUG: no entry found");
545
546 /* shift entries over */
547 for (; i < check_vector.nr - 1; i++)
548 check_vector.checks[i] = check_vector.checks[i + 1];
549
550 check_vector.nr--;
551
552 vector_unlock();
553}
554
555/* Iterate through all attr_check instances and drop their stacks */
556static void drop_all_attr_stacks(void)
557{
558 int i;
559
560 vector_lock();
561
562 for (i = 0; i < check_vector.nr; i++) {
563 drop_attr_stack(&check_vector.checks[i]->stack);
564 }
565
566 vector_unlock();
567}
568
37293768
JH
569struct attr_check *attr_check_alloc(void)
570{
dc81cf37
BW
571 struct attr_check *c = xcalloc(1, sizeof(struct attr_check));
572
573 /* save pointer to the check struct */
574 check_vector_add(c);
575
576 return c;
37293768
JH
577}
578
579struct attr_check *attr_check_initl(const char *one, ...)
580{
581 struct attr_check *check;
582 int cnt;
583 va_list params;
584 const char *param;
585
586 va_start(params, one);
587 for (cnt = 1; (param = va_arg(params, const char *)) != NULL; cnt++)
588 ;
589 va_end(params);
590
591 check = attr_check_alloc();
592 check->nr = cnt;
593 check->alloc = cnt;
594 check->items = xcalloc(cnt, sizeof(struct attr_check_item));
595
596 check->items[0].attr = git_attr(one);
597 va_start(params, one);
598 for (cnt = 1; cnt < check->nr; cnt++) {
599 const struct git_attr *attr;
600 param = va_arg(params, const char *);
601 if (!param)
602 die("BUG: counted %d != ended at %d",
603 check->nr, cnt);
604 attr = git_attr(param);
605 if (!attr)
606 die("BUG: %s: not a valid attribute name", param);
607 check->items[cnt].attr = attr;
608 }
609 va_end(params);
610 return check;
611}
612
b0db7046
BW
613struct attr_check *attr_check_dup(const struct attr_check *check)
614{
615 struct attr_check *ret;
616
617 if (!check)
618 return NULL;
619
620 ret = attr_check_alloc();
621
622 ret->nr = check->nr;
623 ret->alloc = check->alloc;
624 ALLOC_ARRAY(ret->items, ret->nr);
625 COPY_ARRAY(ret->items, check->items, ret->nr);
626
627 return ret;
628}
629
37293768
JH
630struct attr_check_item *attr_check_append(struct attr_check *check,
631 const struct git_attr *attr)
632{
633 struct attr_check_item *item;
634
635 ALLOC_GROW(check->items, check->nr + 1, check->alloc);
636 item = &check->items[check->nr++];
637 item->attr = attr;
638 return item;
639}
640
641void attr_check_reset(struct attr_check *check)
642{
643 check->nr = 0;
644}
645
646void attr_check_clear(struct attr_check *check)
647{
6a83d902 648 FREE_AND_NULL(check->items);
37293768
JH
649 check->alloc = 0;
650 check->nr = 0;
685b2925 651
6a83d902 652 FREE_AND_NULL(check->all_attrs);
685b2925 653 check->all_attrs_nr = 0;
dc81cf37
BW
654
655 drop_attr_stack(&check->stack);
37293768
JH
656}
657
658void attr_check_free(struct attr_check *check)
659{
dc81cf37
BW
660 if (check) {
661 /* Remove check from the check vector */
662 check_vector_remove(check);
663
664 attr_check_clear(check);
665 free(check);
666 }
37293768
JH
667}
668
d0bfd026 669static const char *builtin_attr[] = {
155a4b71 670 "[attr]binary -diff -merge -text",
d0bfd026
JH
671 NULL,
672};
673
a4413118
JH
674static void handle_attr_line(struct attr_stack *res,
675 const char *line,
676 const char *src,
677 int lineno,
678 int macro_ok)
679{
680 struct match_attr *a;
681
682 a = parse_attr_line(line, src, lineno, macro_ok);
683 if (!a)
684 return;
3a7fa03d 685 ALLOC_GROW(res->attrs, res->num_matches + 1, res->alloc);
a4413118
JH
686 res->attrs[res->num_matches++] = a;
687}
688
d0bfd026
JH
689static struct attr_stack *read_attr_from_array(const char **list)
690{
691 struct attr_stack *res;
692 const char *line;
f48fd688 693 int lineno = 0;
d0bfd026
JH
694
695 res = xcalloc(1, sizeof(*res));
a4413118
JH
696 while ((line = *(list++)) != NULL)
697 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
d0bfd026
JH
698 return res;
699}
700
7d42ec54 701/*
f0dd0421
BW
702 * Callers into the attribute system assume there is a single, system-wide
703 * global state where attributes are read from and when the state is flipped by
704 * calling git_attr_set_direction(), the stack frames that have been
705 * constructed need to be discarded so so that subsequent calls into the
706 * attribute system will lazily read from the right place. Since changing
707 * direction causes a global paradigm shift, it should not ever be called while
708 * another thread could potentially be calling into the attribute system.
7d42ec54 709 */
06f33c17
JH
710static enum git_attr_direction direction;
711static struct index_state *use_index;
712
f0dd0421
BW
713void git_attr_set_direction(enum git_attr_direction new_direction,
714 struct index_state *istate)
715{
716 if (is_bare_repository() && new_direction != GIT_ATTR_INDEX)
717 die("BUG: non-INDEX attr direction in a bare repo");
718
719 if (new_direction != direction)
720 drop_all_attr_stacks();
721
722 direction = new_direction;
723 use_index = istate;
724}
725
f48fd688 726static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
d0bfd026 727{
e9d983f1 728 FILE *fp = fopen_or_warn(path, "r");
d0bfd026
JH
729 struct attr_stack *res;
730 char buf[2048];
f48fd688 731 int lineno = 0;
d0bfd026 732
e9d983f1 733 if (!fp)
a4413118
JH
734 return NULL;
735 res = xcalloc(1, sizeof(*res));
27547e5f
JH
736 while (fgets(buf, sizeof(buf), fp)) {
737 char *bufp = buf;
738 if (!lineno)
739 skip_utf8_bom(&bufp, strlen(bufp));
740 handle_attr_line(res, bufp, path, ++lineno, macro_ok);
741 }
a4413118
JH
742 fclose(fp);
743 return res;
744}
d0bfd026 745
06f33c17 746static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
a4413118
JH
747{
748 struct attr_stack *res;
1a9d7e9b
JH
749 char *buf, *sp;
750 int lineno = 0;
f48fd688 751
ff366825 752 buf = read_blob_data_from_index(use_index ? use_index : &the_index, path, NULL);
1a9d7e9b 753 if (!buf)
06f33c17 754 return NULL;
1a9d7e9b 755
06f33c17 756 res = xcalloc(1, sizeof(*res));
1a9d7e9b
JH
757 for (sp = buf; *sp; ) {
758 char *ep;
759 int more;
263434f2
JH
760
761 ep = strchrnul(sp, '\n');
1a9d7e9b
JH
762 more = (*ep == '\n');
763 *ep = '\0';
764 handle_attr_line(res, sp, path, ++lineno, macro_ok);
765 sp = ep + more;
766 }
767 free(buf);
d0bfd026
JH
768 return res;
769}
770
06f33c17
JH
771static struct attr_stack *read_attr(const char *path, int macro_ok)
772{
0787dafd 773 struct attr_stack *res = NULL;
06f33c17 774
0787dafd 775 if (direction == GIT_ATTR_INDEX) {
06f33c17 776 res = read_attr_from_index(path, macro_ok);
0787dafd
BW
777 } else if (!is_bare_repository()) {
778 if (direction == GIT_ATTR_CHECKOUT) {
06f33c17 779 res = read_attr_from_index(path, macro_ok);
0787dafd
BW
780 if (!res)
781 res = read_attr_from_file(path, macro_ok);
782 } else if (direction == GIT_ATTR_CHECKIN) {
783 res = read_attr_from_file(path, macro_ok);
784 if (!res)
785 /*
786 * There is no checked out .gitattributes file
787 * there, but we might have it in the index.
788 * We allow operation in a sparsely checked out
789 * work tree, so read from it.
790 */
791 res = read_attr_from_index(path, macro_ok);
792 }
06f33c17 793 }
0787dafd 794
06f33c17
JH
795 if (!res)
796 res = xcalloc(1, sizeof(*res));
797 return res;
798}
799
d0bfd026
JH
800#if DEBUG_ATTR
801static void debug_info(const char *what, struct attr_stack *elem)
802{
803 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
804}
cf94ccda 805static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
f48fd688 806{
515106fa
JH
807 const char *value = v;
808
809 if (ATTR_TRUE(value))
810 value = "set";
811 else if (ATTR_FALSE(value))
812 value = "unset";
813 else if (ATTR_UNSET(value))
814 value = "unspecified";
815
816 fprintf(stderr, "%s: %s => %s (%s)\n",
817 what, attr->name, (char *) value, match);
f48fd688 818}
d0bfd026
JH
819#define debug_push(a) debug_info("push", (a))
820#define debug_pop(a) debug_info("pop", (a))
821#else
822#define debug_push(a) do { ; } while (0)
823#define debug_pop(a) do { ; } while (0)
f48fd688 824#define debug_set(a,b,c,d) do { ; } while (0)
034d35c9 825#endif /* DEBUG_ATTR */
d0bfd026 826
c5147722 827static const char *git_etc_gitattributes(void)
6df42ab9
PO
828{
829 static const char *system_wide;
830 if (!system_wide)
831 system_wide = system_path(ETC_GITATTRIBUTES);
832 return system_wide;
833}
834
dc81cf37
BW
835static const char *get_home_gitattributes(void)
836{
837 if (!git_attributes_file)
838 git_attributes_file = xdg_config_home("attributes");
839
840 return git_attributes_file;
841}
842
c5147722 843static int git_attr_system(void)
6df42ab9
PO
844{
845 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
846}
847
f932729c
JK
848static GIT_PATH_FUNC(git_path_info_attributes, INFOATTRIBUTES_FILE)
849
4c0ce074
JH
850static void push_stack(struct attr_stack **attr_stack_p,
851 struct attr_stack *elem, char *origin, size_t originlen)
852{
853 if (elem) {
854 elem->origin = origin;
855 if (origin)
856 elem->originlen = originlen;
857 elem->prev = *attr_stack_p;
858 *attr_stack_p = elem;
859 }
860}
861
dc81cf37 862static void bootstrap_attr_stack(struct attr_stack **stack)
f48fd688 863{
dc81cf37 864 struct attr_stack *e;
f48fd688 865
dc81cf37 866 if (*stack)
909ca7b9 867 return;
f48fd688 868
dc81cf37
BW
869 /* builtin frame */
870 e = read_attr_from_array(builtin_attr);
871 push_stack(stack, e, NULL, 0);
6df42ab9 872
dc81cf37
BW
873 /* system-wide frame */
874 if (git_attr_system()) {
875 e = read_attr_from_file(git_etc_gitattributes(), 1);
876 push_stack(stack, e, NULL, 0);
877 }
f48fd688 878
dc81cf37
BW
879 /* home directory */
880 if (get_home_gitattributes()) {
881 e = read_attr_from_file(get_home_gitattributes(), 1);
882 push_stack(stack, e, NULL, 0);
f48fd688 883 }
909ca7b9 884
dc81cf37 885 /* root directory */
0787dafd 886 e = read_attr(GITATTRIBUTES_FILE, 1);
dc81cf37 887 push_stack(stack, e, xstrdup(""), 0);
f0056f64 888
dc81cf37
BW
889 /* info frame */
890 if (startup_info->have_repository)
891 e = read_attr_from_file(git_path_info_attributes(), 1);
892 else
893 e = NULL;
894 if (!e)
895 e = xcalloc(1, sizeof(struct attr_stack));
896 push_stack(stack, e, NULL, 0);
f48fd688
JH
897}
898
dc81cf37
BW
899static void prepare_attr_stack(const char *path, int dirlen,
900 struct attr_stack **stack)
d0bfd026 901{
dc81cf37 902 struct attr_stack *info;
0787dafd 903 struct strbuf pathbuf = STRBUF_INIT;
a8727017 904
d0bfd026
JH
905 /*
906 * At the bottom of the attribute stack is the built-in
6df42ab9
PO
907 * set of attribute definitions, followed by the contents
908 * of $(prefix)/etc/gitattributes and a file specified by
909 * core.attributesfile. Then, contents from
d0bfd026
JH
910 * .gitattribute files from directories closer to the
911 * root to the ones in deeper directories are pushed
912 * to the stack. Finally, at the very top of the stack
913 * we always keep the contents of $GIT_DIR/info/attributes.
914 *
915 * When checking, we use entries from near the top of the
916 * stack, preferring $GIT_DIR/info/attributes, then
917 * .gitattributes in deeper directories to shallower ones,
918 * and finally use the built-in set as the default.
919 */
dc81cf37 920 bootstrap_attr_stack(stack);
d0bfd026
JH
921
922 /*
923 * Pop the "info" one that is always at the top of the stack.
924 */
dc81cf37
BW
925 info = *stack;
926 *stack = info->prev;
d0bfd026
JH
927
928 /*
929 * Pop the ones from directories that are not the prefix of
c432ef99
JH
930 * the path we are checking. Break out of the loop when we see
931 * the root one (whose origin is an empty string "") or the builtin
932 * one (whose origin is NULL) without popping it.
d0bfd026 933 */
dc81cf37
BW
934 while ((*stack)->origin) {
935 int namelen = (*stack)->originlen;
936 struct attr_stack *elem;
d0bfd026 937
dc81cf37 938 elem = *stack;
d0bfd026 939 if (namelen <= dirlen &&
1afca444
JK
940 !strncmp(elem->origin, path, namelen) &&
941 (!namelen || path[namelen] == '/'))
d0bfd026
JH
942 break;
943
944 debug_pop(elem);
dc81cf37
BW
945 *stack = elem->prev;
946 attr_stack_free(elem);
d0bfd026
JH
947 }
948
949 /*
0787dafd
BW
950 * bootstrap_attr_stack() should have added, and the
951 * above loop should have stopped before popping, the
952 * root element whose attr_stack->origin is set to an
953 * empty string.
d0bfd026 954 */
0787dafd
BW
955 assert((*stack)->origin);
956
957 strbuf_addstr(&pathbuf, (*stack)->origin);
958 /* Build up to the directory 'path' is in */
959 while (pathbuf.len < dirlen) {
960 size_t len = pathbuf.len;
961 struct attr_stack *next;
962 char *origin;
963
964 /* Skip path-separator */
965 if (len < dirlen && is_dir_sep(path[len]))
966 len++;
967 /* Find the end of the next component */
968 while (len < dirlen && !is_dir_sep(path[len]))
969 len++;
970
971 if (pathbuf.len > 0)
972 strbuf_addch(&pathbuf, '/');
973 strbuf_add(&pathbuf, path + pathbuf.len, (len - pathbuf.len));
974 strbuf_addf(&pathbuf, "/%s", GITATTRIBUTES_FILE);
975
976 next = read_attr(pathbuf.buf, 0);
977
978 /* reset the pathbuf to not include "/.gitattributes" */
979 strbuf_setlen(&pathbuf, len);
980
981 origin = xstrdup(pathbuf.buf);
982 push_stack(stack, next, origin, len);
97410b27 983 }
d4c98565 984
d0bfd026
JH
985 /*
986 * Finally push the "info" one at the top of the stack.
987 */
dc81cf37 988 push_stack(stack, info, NULL, 0);
0787dafd
BW
989
990 strbuf_release(&pathbuf);
d0bfd026
JH
991}
992
993static int path_matches(const char *pathname, int pathlen,
bd2f371d 994 int basename_offset,
82dce998 995 const struct pattern *pat,
d0bfd026
JH
996 const char *base, int baselen)
997{
82dce998
NTND
998 const char *pattern = pat->pattern;
999 int prefix = pat->nowildcardlen;
dc09e9ec 1000 int isdir = (pathlen && pathname[pathlen - 1] == '/');
82dce998 1001
dc09e9ec 1002 if ((pat->flags & EXC_FLAG_MUSTBEDIR) && !isdir)
94bc671a
JNA
1003 return 0;
1004
82dce998 1005 if (pat->flags & EXC_FLAG_NODIR) {
bd2f371d 1006 return match_basename(pathname + basename_offset,
dc09e9ec 1007 pathlen - basename_offset - isdir,
82dce998
NTND
1008 pattern, prefix,
1009 pat->patternlen, pat->flags);
d0bfd026 1010 }
dc09e9ec 1011 return match_pathname(pathname, pathlen - isdir,
82dce998
NTND
1012 base, baselen,
1013 pattern, prefix, pat->patternlen, pat->flags);
d0bfd026
JH
1014}
1015
685b2925 1016static int macroexpand_one(struct all_attrs_item *all_attrs, int nr, int rem);
ec775c41 1017
685b2925 1018static int fill_one(const char *what, struct all_attrs_item *all_attrs,
60a12722 1019 const struct match_attr *a, int rem)
515106fa 1020{
515106fa
JH
1021 int i;
1022
685b2925 1023 for (i = a->num_attr - 1; rem > 0 && i >= 0; i--) {
e810e063 1024 const struct git_attr *attr = a->state[i].attr;
685b2925 1025 const char **n = &(all_attrs[attr->attr_nr].value);
a5e92abd 1026 const char *v = a->state[i].setto;
515106fa
JH
1027
1028 if (*n == ATTR__UNKNOWN) {
426c27b7 1029 debug_set(what,
712efb1a 1030 a->is_macro ? a->u.attr->name : a->u.pat.pattern,
426c27b7 1031 attr, v);
515106fa
JH
1032 *n = v;
1033 rem--;
685b2925 1034 rem = macroexpand_one(all_attrs, attr->attr_nr, rem);
515106fa
JH
1035 }
1036 }
1037 return rem;
1038}
1039
bd2f371d 1040static int fill(const char *path, int pathlen, int basename_offset,
dc81cf37
BW
1041 const struct attr_stack *stack,
1042 struct all_attrs_item *all_attrs, int rem)
d0bfd026 1043{
dc81cf37
BW
1044 for (; rem > 0 && stack; stack = stack->prev) {
1045 int i;
1046 const char *base = stack->origin ? stack->origin : "";
1047
1048 for (i = stack->num_matches - 1; 0 < rem && 0 <= i; i--) {
1049 const struct match_attr *a = stack->attrs[i];
1050 if (a->is_macro)
1051 continue;
1052 if (path_matches(path, pathlen, basename_offset,
1053 &a->u.pat, base, stack->originlen))
1054 rem = fill_one("fill", all_attrs, a, rem);
1055 }
d0bfd026 1056 }
dc81cf37 1057
d0bfd026
JH
1058 return rem;
1059}
1060
685b2925 1061static int macroexpand_one(struct all_attrs_item *all_attrs, int nr, int rem)
f48fd688 1062{
60a12722 1063 const struct all_attrs_item *item = &all_attrs[nr];
f48fd688 1064
60a12722
BW
1065 if (item->macro && item->value == ATTR__TRUE)
1066 return fill_one("expand", all_attrs, item->macro, rem);
1067 else
ec775c41 1068 return rem;
60a12722 1069}
ec775c41 1070
60a12722
BW
1071/*
1072 * Marks the attributes which are macros based on the attribute stack.
1073 * This prevents having to search through the attribute stack each time
1074 * a macro needs to be expanded during the fill stage.
1075 */
1076static void determine_macros(struct all_attrs_item *all_attrs,
1077 const struct attr_stack *stack)
1078{
1079 for (; stack; stack = stack->prev) {
1080 int i;
1081 for (i = stack->num_matches - 1; i >= 0; i--) {
1082 const struct match_attr *ma = stack->attrs[i];
1083 if (ma->is_macro) {
1084 int n = ma->u.attr->attr_nr;
1085 if (!all_attrs[n].macro) {
1086 all_attrs[n].macro = ma;
1087 }
1088 }
ec775c41 1089 }
4b0c6961 1090 }
f48fd688
JH
1091}
1092
2d721744 1093/*
685b2925
BW
1094 * Collect attributes for path into the array pointed to by check->all_attrs.
1095 * If check->check_nr is non-zero, only attributes in check[] are collected.
1096 * Otherwise all attributes are collected.
2d721744 1097 */
6bc2e3f7 1098static void collect_some_attrs(const char *path, struct attr_check *check)
d0bfd026 1099{
9db9eecf 1100 int i, pathlen, rem, dirlen;
bd2f371d
JH
1101 const char *cp, *last_slash = NULL;
1102 int basename_offset;
9db9eecf
NTND
1103
1104 for (cp = path; *cp; cp++) {
1105 if (*cp == '/' && cp[1])
1106 last_slash = cp;
1107 }
1108 pathlen = cp - path;
1109 if (last_slash) {
bd2f371d 1110 basename_offset = last_slash + 1 - path;
9db9eecf
NTND
1111 dirlen = last_slash - path;
1112 } else {
bd2f371d 1113 basename_offset = 0;
9db9eecf
NTND
1114 dirlen = 0;
1115 }
d0bfd026 1116
dc81cf37 1117 prepare_attr_stack(path, dirlen, &check->stack);
685b2925 1118 all_attrs_init(&g_attr_hashmap, check);
dc81cf37 1119 determine_macros(check->all_attrs, check->stack);
685b2925 1120
60a12722 1121 if (check->nr) {
06a604e6 1122 rem = 0;
6bc2e3f7 1123 for (i = 0; i < check->nr; i++) {
60a12722
BW
1124 int n = check->items[i].attr->attr_nr;
1125 struct all_attrs_item *item = &check->all_attrs[n];
1126 if (item->macro) {
1127 item->value = ATTR__UNSET;
06a604e6
NTND
1128 rem++;
1129 }
1130 }
6bc2e3f7 1131 if (rem == check->nr)
06a604e6
NTND
1132 return;
1133 }
d0bfd026 1134
685b2925 1135 rem = check->all_attrs_nr;
dc81cf37 1136 fill(path, pathlen, basename_offset, check->stack, check->all_attrs, rem);
2d721744
MH
1137}
1138
6bc2e3f7 1139int git_check_attr(const char *path, struct attr_check *check)
2d721744
MH
1140{
1141 int i;
1142
6bc2e3f7 1143 collect_some_attrs(path, check);
f48fd688 1144
6bc2e3f7 1145 for (i = 0; i < check->nr; i++) {
685b2925
BW
1146 size_t n = check->items[i].attr->attr_nr;
1147 const char *value = check->all_attrs[n].value;
515106fa
JH
1148 if (value == ATTR__UNKNOWN)
1149 value = ATTR__UNSET;
6bc2e3f7 1150 check->items[i].value = value;
515106fa 1151 }
f48fd688 1152
d0bfd026
JH
1153 return 0;
1154}
06f33c17 1155
7f864111 1156void git_all_attrs(const char *path, struct attr_check *check)
ee548df3 1157{
7f864111 1158 int i;
ee548df3 1159
7f864111 1160 attr_check_reset(check);
6bc2e3f7 1161 collect_some_attrs(path, check);
ee548df3 1162
685b2925
BW
1163 for (i = 0; i < check->all_attrs_nr; i++) {
1164 const char *name = check->all_attrs[i].attr->name;
1165 const char *value = check->all_attrs[i].value;
7f864111
JH
1166 struct attr_check_item *item;
1167 if (value == ATTR__UNSET || value == ATTR__UNKNOWN)
1168 continue;
1169 item = attr_check_append(check, git_attr(name));
1170 item->value = value;
ee548df3 1171 }
ee548df3
MH
1172}
1173
1a600b75
BW
1174void attr_start(void)
1175{
1176#ifndef NO_PTHREADS
1177 pthread_mutex_init(&g_attr_hashmap.mutex, NULL);
dc81cf37 1178 pthread_mutex_init(&check_vector.mutex, NULL);
1a600b75
BW
1179#endif
1180}