hashmap: migrate documentation from Documentation/technical into header
[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 */
7663cdc8
SB
79static int attr_hash_entry_cmp(void *unused_cmp_data,
80 const struct attr_hash_entry *a,
1a600b75 81 const struct attr_hash_entry *b,
7663cdc8 82 void *unused_keydata)
1a600b75
BW
83{
84 return (a->keylen != b->keylen) || strncmp(a->key, b->key, a->keylen);
85}
86
87/* Initialize an 'attr_hashmap' object */
88static void attr_hashmap_init(struct attr_hashmap *map)
89{
7663cdc8 90 hashmap_init(&map->map, (hashmap_cmp_fn) attr_hash_entry_cmp, NULL, 0);
1a600b75
BW
91}
92
93/*
94 * Retrieve the 'value' stored in a hashmap given the provided 'key'.
95 * If there is no matching entry, return NULL.
96 */
97static void *attr_hashmap_get(struct attr_hashmap *map,
98 const char *key, size_t keylen)
99{
100 struct attr_hash_entry k;
101 struct attr_hash_entry *e;
102
103 if (!map->map.tablesize)
104 attr_hashmap_init(map);
105
106 hashmap_entry_init(&k, memhash(key, keylen));
107 k.key = key;
108 k.keylen = keylen;
109 e = hashmap_get(&map->map, &k, NULL);
110
111 return e ? e->value : NULL;
112}
113
114/* Add 'value' to a hashmap based on the provided 'key'. */
115static 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 if (!map->map.tablesize)
122 attr_hashmap_init(map);
123
124 e = xmalloc(sizeof(struct attr_hash_entry));
125 hashmap_entry_init(e, memhash(key, keylen));
126 e->key = key;
127 e->keylen = keylen;
128 e->value = value;
129
130 hashmap_add(&map->map, e);
d0bfd026
JH
131}
132
685b2925
BW
133struct all_attrs_item {
134 const struct git_attr *attr;
135 const char *value;
60a12722
BW
136 /*
137 * If 'macro' is non-NULL, indicates that 'attr' is a macro based on
138 * the current attribute stack and contains a pointer to the match_attr
139 * definition of the macro
140 */
141 const struct match_attr *macro;
685b2925
BW
142};
143
144/*
145 * Reallocate and reinitialize the array of all attributes (which is used in
146 * the attribute collection process) in 'check' based on the global dictionary
147 * of attributes.
148 */
149static void all_attrs_init(struct attr_hashmap *map, struct attr_check *check)
150{
151 int i;
152
153 hashmap_lock(map);
154
155 if (map->map.size < check->all_attrs_nr)
156 die("BUG: interned attributes shouldn't be deleted");
157
158 /*
159 * If the number of attributes in the global dictionary has increased
160 * (or this attr_check instance doesn't have an initialized all_attrs
161 * field), reallocate the provided attr_check instance's all_attrs
162 * field and fill each entry with its corresponding git_attr.
163 */
164 if (map->map.size != check->all_attrs_nr) {
165 struct attr_hash_entry *e;
166 struct hashmap_iter iter;
167 hashmap_iter_init(&map->map, &iter);
168
169 REALLOC_ARRAY(check->all_attrs, map->map.size);
170 check->all_attrs_nr = map->map.size;
171
172 while ((e = hashmap_iter_next(&iter))) {
173 const struct git_attr *a = e->value;
174 check->all_attrs[a->attr_nr].attr = a;
175 }
176 }
177
178 hashmap_unlock(map);
179
180 /*
181 * Re-initialize every entry in check->all_attrs.
182 * This re-initialization can live outside of the locked region since
183 * the attribute dictionary is no longer being accessed.
184 */
185 for (i = 0; i < check->all_attrs_nr; i++) {
186 check->all_attrs[i].value = ATTR__UNKNOWN;
60a12722 187 check->all_attrs[i].macro = NULL;
685b2925
BW
188 }
189}
190
428103c7 191static int attr_name_valid(const char *name, size_t namelen)
e4aee10a
JH
192{
193 /*
d42453ab
MH
194 * Attribute name cannot begin with '-' and must consist of
195 * characters from [-A-Za-z0-9_.].
e4aee10a 196 */
c0b13b21 197 if (namelen <= 0 || *name == '-')
428103c7 198 return 0;
e4aee10a
JH
199 while (namelen--) {
200 char ch = *name++;
201 if (! (ch == '-' || ch == '.' || ch == '_' ||
202 ('0' <= ch && ch <= '9') ||
203 ('a' <= ch && ch <= 'z') ||
204 ('A' <= ch && ch <= 'Z')) )
428103c7 205 return 0;
e4aee10a 206 }
428103c7
JH
207 return 1;
208}
209
210static void report_invalid_attr(const char *name, size_t len,
211 const char *src, int lineno)
212{
213 struct strbuf err = STRBUF_INIT;
214 strbuf_addf(&err, _("%.*s is not a valid attribute name"),
215 (int) len, name);
216 fprintf(stderr, "%s: %s:%d\n", err.buf, src, lineno);
217 strbuf_release(&err);
e4aee10a
JH
218}
219
1a600b75
BW
220/*
221 * Given a 'name', lookup and return the corresponding attribute in the global
222 * dictionary. If no entry is found, create a new attribute and store it in
223 * the dictionary.
224 */
e810e063 225static const struct git_attr *git_attr_internal(const char *name, int namelen)
d0bfd026 226{
d0bfd026
JH
227 struct git_attr *a;
228
1a600b75 229 if (!attr_name_valid(name, namelen))
e4aee10a
JH
230 return NULL;
231
1a600b75
BW
232 hashmap_lock(&g_attr_hashmap);
233
234 a = attr_hashmap_get(&g_attr_hashmap, name, namelen);
235
236 if (!a) {
237 FLEX_ALLOC_MEM(a, name, name, namelen);
238 a->attr_nr = g_attr_hashmap.map.size;
1a600b75
BW
239
240 attr_hashmap_add(&g_attr_hashmap, a->name, namelen, a);
241 assert(a->attr_nr == (g_attr_hashmap.map.size - 1));
1a600b75
BW
242 }
243
244 hashmap_unlock(&g_attr_hashmap);
f48fd688 245
d0bfd026
JH
246 return a;
247}
248
e810e063 249const struct git_attr *git_attr(const char *name)
7fb0eaa2
JH
250{
251 return git_attr_internal(name, strlen(name));
252}
253
515106fa 254/* What does a matched pattern decide? */
d0bfd026 255struct attr_state {
e810e063 256 const struct git_attr *attr;
a5e92abd 257 const char *setto;
d0bfd026
JH
258};
259
82dce998
NTND
260struct pattern {
261 const char *pattern;
262 int patternlen;
263 int nowildcardlen;
f8708998 264 unsigned flags; /* EXC_FLAG_* */
82dce998
NTND
265};
266
ba845b75
MH
267/*
268 * One rule, as from a .gitattributes file.
269 *
270 * If is_macro is true, then u.attr is a pointer to the git_attr being
271 * defined.
272 *
4894f4ff
JH
273 * If is_macro is false, then u.pat is the filename pattern to which the
274 * rule applies.
ba845b75
MH
275 *
276 * In either case, num_attr is the number of attributes affected by
277 * this rule, and state is an array listing them. The attributes are
278 * listed as they appear in the file (macros unexpanded).
279 */
d0bfd026 280struct match_attr {
f48fd688 281 union {
82dce998 282 struct pattern pat;
e810e063 283 const struct git_attr *attr;
f48fd688
JH
284 } u;
285 char is_macro;
d0bfd026
JH
286 unsigned num_attr;
287 struct attr_state state[FLEX_ARRAY];
288};
289
290static const char blank[] = " \t\r\n";
291
d1751298
MH
292/*
293 * Parse a whitespace-delimited attribute state (i.e., "attr",
294 * "-attr", "!attr", or "attr=value") from the string starting at src.
295 * If e is not NULL, write the results to *e. Return a pointer to the
296 * remainder of the string (with leading whitespace removed), or NULL
297 * if there was an error.
298 */
515106fa 299static const char *parse_attr(const char *src, int lineno, const char *cp,
d1751298 300 struct attr_state *e)
515106fa
JH
301{
302 const char *ep, *equals;
303 int len;
304
305 ep = cp + strcspn(cp, blank);
306 equals = strchr(cp, '=');
307 if (equals && ep < equals)
308 equals = NULL;
309 if (equals)
310 len = equals - cp;
311 else
312 len = ep - cp;
d1751298 313 if (!e) {
515106fa
JH
314 if (*cp == '-' || *cp == '!') {
315 cp++;
316 len--;
317 }
428103c7
JH
318 if (!attr_name_valid(cp, len)) {
319 report_invalid_attr(cp, len, src, lineno);
515106fa
JH
320 return NULL;
321 }
322 } else {
5a884019
JH
323 /*
324 * As this function is always called twice, once with
325 * e == NULL in the first pass and then e != NULL in
428103c7 326 * the second pass, no need for attr_name_valid()
5a884019
JH
327 * check here.
328 */
515106fa
JH
329 if (*cp == '-' || *cp == '!') {
330 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
331 cp++;
332 len--;
333 }
334 else if (!equals)
335 e->setto = ATTR__TRUE;
336 else {
182af834 337 e->setto = xmemdupz(equals + 1, ep - equals - 1);
515106fa 338 }
7fb0eaa2 339 e->attr = git_attr_internal(cp, len);
515106fa 340 }
515106fa
JH
341 return ep + strspn(ep, blank);
342}
343
f48fd688
JH
344static struct match_attr *parse_attr_line(const char *line, const char *src,
345 int lineno, int macro_ok)
d0bfd026
JH
346{
347 int namelen;
d68e1c18 348 int num_attr, i;
85c4a0d0 349 const char *cp, *name, *states;
515106fa 350 struct match_attr *res = NULL;
f48fd688 351 int is_macro;
860a74d9 352 struct strbuf pattern = STRBUF_INIT;
d0bfd026
JH
353
354 cp = line + strspn(line, blank);
355 if (!*cp || *cp == '#')
356 return NULL;
357 name = cp;
860a74d9
NTND
358
359 if (*cp == '"' && !unquote_c_style(&pattern, name, &states)) {
360 name = pattern.buf;
361 namelen = pattern.len;
362 } else {
363 namelen = strcspn(name, blank);
364 states = name + namelen;
365 }
366
f48fd688 367 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
59556548 368 starts_with(name, ATTRIBUTE_MACRO_PREFIX)) {
f48fd688
JH
369 if (!macro_ok) {
370 fprintf(stderr, "%s not allowed: %s:%d\n",
371 name, src, lineno);
62af8969 372 goto fail_return;
f48fd688
JH
373 }
374 is_macro = 1;
375 name += strlen(ATTRIBUTE_MACRO_PREFIX);
376 name += strspn(name, blank);
377 namelen = strcspn(name, blank);
428103c7
JH
378 if (!attr_name_valid(name, namelen)) {
379 report_invalid_attr(name, namelen, src, lineno);
62af8969 380 goto fail_return;
e4aee10a 381 }
f48fd688
JH
382 }
383 else
384 is_macro = 0;
d0bfd026 385
85c4a0d0
MH
386 states += strspn(states, blank);
387
d68e1c18
MH
388 /* First pass to count the attr_states */
389 for (cp = states, num_attr = 0; *cp; num_attr++) {
390 cp = parse_attr(src, lineno, cp, NULL);
391 if (!cp)
62af8969 392 goto fail_return;
d0bfd026 393 }
d68e1c18
MH
394
395 res = xcalloc(1,
396 sizeof(*res) +
397 sizeof(struct attr_state) * num_attr +
398 (is_macro ? 0 : namelen + 1));
fad32bcd 399 if (is_macro) {
d68e1c18 400 res->u.attr = git_attr_internal(name, namelen);
fad32bcd 401 } else {
82dce998
NTND
402 char *p = (char *)&(res->state[num_attr]);
403 memcpy(p, name, namelen);
404 res->u.pat.pattern = p;
405 parse_exclude_pattern(&res->u.pat.pattern,
406 &res->u.pat.patternlen,
407 &res->u.pat.flags,
408 &res->u.pat.nowildcardlen);
8b1bd024
TR
409 if (res->u.pat.flags & EXC_FLAG_NEGATIVE) {
410 warning(_("Negative patterns are ignored in git attributes\n"
411 "Use '\\!' for literal leading exclamation."));
62af8969 412 goto fail_return;
8b1bd024 413 }
d0bfd026 414 }
d68e1c18
MH
415 res->is_macro = is_macro;
416 res->num_attr = num_attr;
417
418 /* Second pass to fill the attr_states */
419 for (cp = states, i = 0; *cp; i++) {
420 cp = parse_attr(src, lineno, cp, &(res->state[i]));
421 }
422
860a74d9 423 strbuf_release(&pattern);
d0bfd026 424 return res;
62af8969
JH
425
426fail_return:
860a74d9 427 strbuf_release(&pattern);
62af8969
JH
428 free(res);
429 return NULL;
d0bfd026
JH
430}
431
432/*
433 * Like info/exclude and .gitignore, the attribute information can
434 * come from many places.
435 *
436 * (1) .gitattribute file of the same directory;
515106fa
JH
437 * (2) .gitattribute file of the parent directory if (1) does not have
438 * any match; this goes recursively upwards, just like .gitignore.
439 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
d0bfd026
JH
440 *
441 * In the same file, later entries override the earlier match, so in the
442 * global list, we would have entries from info/attributes the earliest
443 * (reading the file from top to bottom), .gitattribute of the root
444 * directory (again, reading the file from top to bottom) down to the
445 * current directory, and then scan the list backwards to find the first match.
6d24e7a8 446 * This is exactly the same as what is_excluded() does in dir.c to deal with
6f416cba 447 * .gitignore file and info/excludes file as a fallback.
d0bfd026
JH
448 */
449
dc81cf37 450struct attr_stack {
d0bfd026
JH
451 struct attr_stack *prev;
452 char *origin;
cd6a0b26 453 size_t originlen;
d0bfd026 454 unsigned num_matches;
a4413118 455 unsigned alloc;
d0bfd026 456 struct match_attr **attrs;
dc81cf37 457};
d0bfd026 458
dc81cf37 459static void attr_stack_free(struct attr_stack *e)
d0bfd026
JH
460{
461 int i;
462 free(e->origin);
515106fa
JH
463 for (i = 0; i < e->num_matches; i++) {
464 struct match_attr *a = e->attrs[i];
465 int j;
466 for (j = 0; j < a->num_attr; j++) {
a5e92abd 467 const char *setto = a->state[j].setto;
515106fa
JH
468 if (setto == ATTR__TRUE ||
469 setto == ATTR__FALSE ||
470 setto == ATTR__UNSET ||
471 setto == ATTR__UNKNOWN)
472 ;
473 else
4b25d091 474 free((char *) setto);
515106fa
JH
475 }
476 free(a);
477 }
37475f97 478 free(e->attrs);
d0bfd026
JH
479 free(e);
480}
481
dc81cf37
BW
482static void drop_attr_stack(struct attr_stack **stack)
483{
484 while (*stack) {
485 struct attr_stack *elem = *stack;
486 *stack = elem->prev;
487 attr_stack_free(elem);
488 }
489}
490
491/* List of all attr_check structs; access should be surrounded by mutex */
492static struct check_vector {
493 size_t nr;
494 size_t alloc;
495 struct attr_check **checks;
496#ifndef NO_PTHREADS
497 pthread_mutex_t mutex;
498#endif
499} check_vector;
500
501static inline void vector_lock(void)
502{
503#ifndef NO_PTHREADS
504 pthread_mutex_lock(&check_vector.mutex);
505#endif
506}
507
508static inline void vector_unlock(void)
509{
510#ifndef NO_PTHREADS
511 pthread_mutex_unlock(&check_vector.mutex);
512#endif
513}
514
515static void check_vector_add(struct attr_check *c)
516{
517 vector_lock();
518
519 ALLOC_GROW(check_vector.checks,
520 check_vector.nr + 1,
521 check_vector.alloc);
522 check_vector.checks[check_vector.nr++] = c;
523
524 vector_unlock();
525}
526
527static void check_vector_remove(struct attr_check *check)
528{
529 int i;
530
531 vector_lock();
532
533 /* Find entry */
534 for (i = 0; i < check_vector.nr; i++)
535 if (check_vector.checks[i] == check)
536 break;
537
538 if (i >= check_vector.nr)
539 die("BUG: no entry found");
540
541 /* shift entries over */
542 for (; i < check_vector.nr - 1; i++)
543 check_vector.checks[i] = check_vector.checks[i + 1];
544
545 check_vector.nr--;
546
547 vector_unlock();
548}
549
550/* Iterate through all attr_check instances and drop their stacks */
551static void drop_all_attr_stacks(void)
552{
553 int i;
554
555 vector_lock();
556
557 for (i = 0; i < check_vector.nr; i++) {
558 drop_attr_stack(&check_vector.checks[i]->stack);
559 }
560
561 vector_unlock();
562}
563
37293768
JH
564struct attr_check *attr_check_alloc(void)
565{
dc81cf37
BW
566 struct attr_check *c = xcalloc(1, sizeof(struct attr_check));
567
568 /* save pointer to the check struct */
569 check_vector_add(c);
570
571 return c;
37293768
JH
572}
573
574struct attr_check *attr_check_initl(const char *one, ...)
575{
576 struct attr_check *check;
577 int cnt;
578 va_list params;
579 const char *param;
580
581 va_start(params, one);
582 for (cnt = 1; (param = va_arg(params, const char *)) != NULL; cnt++)
583 ;
584 va_end(params);
585
586 check = attr_check_alloc();
587 check->nr = cnt;
588 check->alloc = cnt;
589 check->items = xcalloc(cnt, sizeof(struct attr_check_item));
590
591 check->items[0].attr = git_attr(one);
592 va_start(params, one);
593 for (cnt = 1; cnt < check->nr; cnt++) {
594 const struct git_attr *attr;
595 param = va_arg(params, const char *);
596 if (!param)
597 die("BUG: counted %d != ended at %d",
598 check->nr, cnt);
599 attr = git_attr(param);
600 if (!attr)
601 die("BUG: %s: not a valid attribute name", param);
602 check->items[cnt].attr = attr;
603 }
604 va_end(params);
605 return check;
606}
607
b0db7046
BW
608struct attr_check *attr_check_dup(const struct attr_check *check)
609{
610 struct attr_check *ret;
611
612 if (!check)
613 return NULL;
614
615 ret = attr_check_alloc();
616
617 ret->nr = check->nr;
618 ret->alloc = check->alloc;
619 ALLOC_ARRAY(ret->items, ret->nr);
620 COPY_ARRAY(ret->items, check->items, ret->nr);
621
622 return ret;
623}
624
37293768
JH
625struct attr_check_item *attr_check_append(struct attr_check *check,
626 const struct git_attr *attr)
627{
628 struct attr_check_item *item;
629
630 ALLOC_GROW(check->items, check->nr + 1, check->alloc);
631 item = &check->items[check->nr++];
632 item->attr = attr;
633 return item;
634}
635
636void attr_check_reset(struct attr_check *check)
637{
638 check->nr = 0;
639}
640
641void attr_check_clear(struct attr_check *check)
642{
6a83d902 643 FREE_AND_NULL(check->items);
37293768
JH
644 check->alloc = 0;
645 check->nr = 0;
685b2925 646
6a83d902 647 FREE_AND_NULL(check->all_attrs);
685b2925 648 check->all_attrs_nr = 0;
dc81cf37
BW
649
650 drop_attr_stack(&check->stack);
37293768
JH
651}
652
653void attr_check_free(struct attr_check *check)
654{
dc81cf37
BW
655 if (check) {
656 /* Remove check from the check vector */
657 check_vector_remove(check);
658
659 attr_check_clear(check);
660 free(check);
661 }
37293768
JH
662}
663
d0bfd026 664static const char *builtin_attr[] = {
155a4b71 665 "[attr]binary -diff -merge -text",
d0bfd026
JH
666 NULL,
667};
668
a4413118
JH
669static void handle_attr_line(struct attr_stack *res,
670 const char *line,
671 const char *src,
672 int lineno,
673 int macro_ok)
674{
675 struct match_attr *a;
676
677 a = parse_attr_line(line, src, lineno, macro_ok);
678 if (!a)
679 return;
3a7fa03d 680 ALLOC_GROW(res->attrs, res->num_matches + 1, res->alloc);
a4413118
JH
681 res->attrs[res->num_matches++] = a;
682}
683
d0bfd026
JH
684static struct attr_stack *read_attr_from_array(const char **list)
685{
686 struct attr_stack *res;
687 const char *line;
f48fd688 688 int lineno = 0;
d0bfd026
JH
689
690 res = xcalloc(1, sizeof(*res));
a4413118
JH
691 while ((line = *(list++)) != NULL)
692 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
d0bfd026
JH
693 return res;
694}
695
7d42ec54 696/*
f0dd0421
BW
697 * Callers into the attribute system assume there is a single, system-wide
698 * global state where attributes are read from and when the state is flipped by
699 * calling git_attr_set_direction(), the stack frames that have been
700 * constructed need to be discarded so so that subsequent calls into the
701 * attribute system will lazily read from the right place. Since changing
702 * direction causes a global paradigm shift, it should not ever be called while
703 * another thread could potentially be calling into the attribute system.
7d42ec54 704 */
06f33c17
JH
705static enum git_attr_direction direction;
706static struct index_state *use_index;
707
f0dd0421
BW
708void git_attr_set_direction(enum git_attr_direction new_direction,
709 struct index_state *istate)
710{
711 if (is_bare_repository() && new_direction != GIT_ATTR_INDEX)
712 die("BUG: non-INDEX attr direction in a bare repo");
713
714 if (new_direction != direction)
715 drop_all_attr_stacks();
716
717 direction = new_direction;
718 use_index = istate;
719}
720
f48fd688 721static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
d0bfd026 722{
e9d983f1 723 FILE *fp = fopen_or_warn(path, "r");
d0bfd026
JH
724 struct attr_stack *res;
725 char buf[2048];
f48fd688 726 int lineno = 0;
d0bfd026 727
e9d983f1 728 if (!fp)
a4413118
JH
729 return NULL;
730 res = xcalloc(1, sizeof(*res));
27547e5f
JH
731 while (fgets(buf, sizeof(buf), fp)) {
732 char *bufp = buf;
733 if (!lineno)
734 skip_utf8_bom(&bufp, strlen(bufp));
735 handle_attr_line(res, bufp, path, ++lineno, macro_ok);
736 }
a4413118
JH
737 fclose(fp);
738 return res;
739}
d0bfd026 740
06f33c17 741static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
a4413118
JH
742{
743 struct attr_stack *res;
1a9d7e9b
JH
744 char *buf, *sp;
745 int lineno = 0;
f48fd688 746
ff366825 747 buf = read_blob_data_from_index(use_index ? use_index : &the_index, path, NULL);
1a9d7e9b 748 if (!buf)
06f33c17 749 return NULL;
1a9d7e9b 750
06f33c17 751 res = xcalloc(1, sizeof(*res));
1a9d7e9b
JH
752 for (sp = buf; *sp; ) {
753 char *ep;
754 int more;
263434f2
JH
755
756 ep = strchrnul(sp, '\n');
1a9d7e9b
JH
757 more = (*ep == '\n');
758 *ep = '\0';
759 handle_attr_line(res, sp, path, ++lineno, macro_ok);
760 sp = ep + more;
761 }
762 free(buf);
d0bfd026
JH
763 return res;
764}
765
06f33c17
JH
766static struct attr_stack *read_attr(const char *path, int macro_ok)
767{
0787dafd 768 struct attr_stack *res = NULL;
06f33c17 769
0787dafd 770 if (direction == GIT_ATTR_INDEX) {
06f33c17 771 res = read_attr_from_index(path, macro_ok);
0787dafd
BW
772 } else if (!is_bare_repository()) {
773 if (direction == GIT_ATTR_CHECKOUT) {
06f33c17 774 res = read_attr_from_index(path, macro_ok);
0787dafd
BW
775 if (!res)
776 res = read_attr_from_file(path, macro_ok);
777 } else if (direction == GIT_ATTR_CHECKIN) {
778 res = read_attr_from_file(path, macro_ok);
779 if (!res)
780 /*
781 * There is no checked out .gitattributes file
782 * there, but we might have it in the index.
783 * We allow operation in a sparsely checked out
784 * work tree, so read from it.
785 */
786 res = read_attr_from_index(path, macro_ok);
787 }
06f33c17 788 }
0787dafd 789
06f33c17
JH
790 if (!res)
791 res = xcalloc(1, sizeof(*res));
792 return res;
793}
794
d0bfd026
JH
795#if DEBUG_ATTR
796static void debug_info(const char *what, struct attr_stack *elem)
797{
798 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
799}
cf94ccda 800static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
f48fd688 801{
515106fa
JH
802 const char *value = v;
803
804 if (ATTR_TRUE(value))
805 value = "set";
806 else if (ATTR_FALSE(value))
807 value = "unset";
808 else if (ATTR_UNSET(value))
809 value = "unspecified";
810
811 fprintf(stderr, "%s: %s => %s (%s)\n",
812 what, attr->name, (char *) value, match);
f48fd688 813}
d0bfd026
JH
814#define debug_push(a) debug_info("push", (a))
815#define debug_pop(a) debug_info("pop", (a))
816#else
817#define debug_push(a) do { ; } while (0)
818#define debug_pop(a) do { ; } while (0)
f48fd688 819#define debug_set(a,b,c,d) do { ; } while (0)
034d35c9 820#endif /* DEBUG_ATTR */
d0bfd026 821
c5147722 822static const char *git_etc_gitattributes(void)
6df42ab9
PO
823{
824 static const char *system_wide;
825 if (!system_wide)
826 system_wide = system_path(ETC_GITATTRIBUTES);
827 return system_wide;
828}
829
dc81cf37
BW
830static const char *get_home_gitattributes(void)
831{
832 if (!git_attributes_file)
833 git_attributes_file = xdg_config_home("attributes");
834
835 return git_attributes_file;
836}
837
c5147722 838static int git_attr_system(void)
6df42ab9
PO
839{
840 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
841}
842
f932729c
JK
843static GIT_PATH_FUNC(git_path_info_attributes, INFOATTRIBUTES_FILE)
844
4c0ce074
JH
845static void push_stack(struct attr_stack **attr_stack_p,
846 struct attr_stack *elem, char *origin, size_t originlen)
847{
848 if (elem) {
849 elem->origin = origin;
850 if (origin)
851 elem->originlen = originlen;
852 elem->prev = *attr_stack_p;
853 *attr_stack_p = elem;
854 }
855}
856
dc81cf37 857static void bootstrap_attr_stack(struct attr_stack **stack)
f48fd688 858{
dc81cf37 859 struct attr_stack *e;
f48fd688 860
dc81cf37 861 if (*stack)
909ca7b9 862 return;
f48fd688 863
dc81cf37
BW
864 /* builtin frame */
865 e = read_attr_from_array(builtin_attr);
866 push_stack(stack, e, NULL, 0);
6df42ab9 867
dc81cf37
BW
868 /* system-wide frame */
869 if (git_attr_system()) {
870 e = read_attr_from_file(git_etc_gitattributes(), 1);
871 push_stack(stack, e, NULL, 0);
872 }
f48fd688 873
dc81cf37
BW
874 /* home directory */
875 if (get_home_gitattributes()) {
876 e = read_attr_from_file(get_home_gitattributes(), 1);
877 push_stack(stack, e, NULL, 0);
f48fd688 878 }
909ca7b9 879
dc81cf37 880 /* root directory */
0787dafd 881 e = read_attr(GITATTRIBUTES_FILE, 1);
dc81cf37 882 push_stack(stack, e, xstrdup(""), 0);
f0056f64 883
dc81cf37
BW
884 /* info frame */
885 if (startup_info->have_repository)
886 e = read_attr_from_file(git_path_info_attributes(), 1);
887 else
888 e = NULL;
889 if (!e)
890 e = xcalloc(1, sizeof(struct attr_stack));
891 push_stack(stack, e, NULL, 0);
f48fd688
JH
892}
893
dc81cf37
BW
894static void prepare_attr_stack(const char *path, int dirlen,
895 struct attr_stack **stack)
d0bfd026 896{
dc81cf37 897 struct attr_stack *info;
0787dafd 898 struct strbuf pathbuf = STRBUF_INIT;
a8727017 899
d0bfd026
JH
900 /*
901 * At the bottom of the attribute stack is the built-in
6df42ab9
PO
902 * set of attribute definitions, followed by the contents
903 * of $(prefix)/etc/gitattributes and a file specified by
904 * core.attributesfile. Then, contents from
d0bfd026
JH
905 * .gitattribute files from directories closer to the
906 * root to the ones in deeper directories are pushed
907 * to the stack. Finally, at the very top of the stack
908 * we always keep the contents of $GIT_DIR/info/attributes.
909 *
910 * When checking, we use entries from near the top of the
911 * stack, preferring $GIT_DIR/info/attributes, then
912 * .gitattributes in deeper directories to shallower ones,
913 * and finally use the built-in set as the default.
914 */
dc81cf37 915 bootstrap_attr_stack(stack);
d0bfd026
JH
916
917 /*
918 * Pop the "info" one that is always at the top of the stack.
919 */
dc81cf37
BW
920 info = *stack;
921 *stack = info->prev;
d0bfd026
JH
922
923 /*
924 * Pop the ones from directories that are not the prefix of
c432ef99
JH
925 * the path we are checking. Break out of the loop when we see
926 * the root one (whose origin is an empty string "") or the builtin
927 * one (whose origin is NULL) without popping it.
d0bfd026 928 */
dc81cf37
BW
929 while ((*stack)->origin) {
930 int namelen = (*stack)->originlen;
931 struct attr_stack *elem;
d0bfd026 932
dc81cf37 933 elem = *stack;
d0bfd026 934 if (namelen <= dirlen &&
1afca444
JK
935 !strncmp(elem->origin, path, namelen) &&
936 (!namelen || path[namelen] == '/'))
d0bfd026
JH
937 break;
938
939 debug_pop(elem);
dc81cf37
BW
940 *stack = elem->prev;
941 attr_stack_free(elem);
d0bfd026
JH
942 }
943
944 /*
0787dafd
BW
945 * bootstrap_attr_stack() should have added, and the
946 * above loop should have stopped before popping, the
947 * root element whose attr_stack->origin is set to an
948 * empty string.
d0bfd026 949 */
0787dafd
BW
950 assert((*stack)->origin);
951
952 strbuf_addstr(&pathbuf, (*stack)->origin);
953 /* Build up to the directory 'path' is in */
954 while (pathbuf.len < dirlen) {
955 size_t len = pathbuf.len;
956 struct attr_stack *next;
957 char *origin;
958
959 /* Skip path-separator */
960 if (len < dirlen && is_dir_sep(path[len]))
961 len++;
962 /* Find the end of the next component */
963 while (len < dirlen && !is_dir_sep(path[len]))
964 len++;
965
966 if (pathbuf.len > 0)
967 strbuf_addch(&pathbuf, '/');
968 strbuf_add(&pathbuf, path + pathbuf.len, (len - pathbuf.len));
969 strbuf_addf(&pathbuf, "/%s", GITATTRIBUTES_FILE);
970
971 next = read_attr(pathbuf.buf, 0);
972
973 /* reset the pathbuf to not include "/.gitattributes" */
974 strbuf_setlen(&pathbuf, len);
975
976 origin = xstrdup(pathbuf.buf);
977 push_stack(stack, next, origin, len);
97410b27 978 }
d4c98565 979
d0bfd026
JH
980 /*
981 * Finally push the "info" one at the top of the stack.
982 */
dc81cf37 983 push_stack(stack, info, NULL, 0);
0787dafd
BW
984
985 strbuf_release(&pathbuf);
d0bfd026
JH
986}
987
988static int path_matches(const char *pathname, int pathlen,
bd2f371d 989 int basename_offset,
82dce998 990 const struct pattern *pat,
d0bfd026
JH
991 const char *base, int baselen)
992{
82dce998
NTND
993 const char *pattern = pat->pattern;
994 int prefix = pat->nowildcardlen;
dc09e9ec 995 int isdir = (pathlen && pathname[pathlen - 1] == '/');
82dce998 996
dc09e9ec 997 if ((pat->flags & EXC_FLAG_MUSTBEDIR) && !isdir)
94bc671a
JNA
998 return 0;
999
82dce998 1000 if (pat->flags & EXC_FLAG_NODIR) {
bd2f371d 1001 return match_basename(pathname + basename_offset,
dc09e9ec 1002 pathlen - basename_offset - isdir,
82dce998
NTND
1003 pattern, prefix,
1004 pat->patternlen, pat->flags);
d0bfd026 1005 }
dc09e9ec 1006 return match_pathname(pathname, pathlen - isdir,
82dce998
NTND
1007 base, baselen,
1008 pattern, prefix, pat->patternlen, pat->flags);
d0bfd026
JH
1009}
1010
685b2925 1011static int macroexpand_one(struct all_attrs_item *all_attrs, int nr, int rem);
ec775c41 1012
685b2925 1013static int fill_one(const char *what, struct all_attrs_item *all_attrs,
60a12722 1014 const struct match_attr *a, int rem)
515106fa 1015{
515106fa
JH
1016 int i;
1017
685b2925 1018 for (i = a->num_attr - 1; rem > 0 && i >= 0; i--) {
e810e063 1019 const struct git_attr *attr = a->state[i].attr;
685b2925 1020 const char **n = &(all_attrs[attr->attr_nr].value);
a5e92abd 1021 const char *v = a->state[i].setto;
515106fa
JH
1022
1023 if (*n == ATTR__UNKNOWN) {
426c27b7 1024 debug_set(what,
712efb1a 1025 a->is_macro ? a->u.attr->name : a->u.pat.pattern,
426c27b7 1026 attr, v);
515106fa
JH
1027 *n = v;
1028 rem--;
685b2925 1029 rem = macroexpand_one(all_attrs, attr->attr_nr, rem);
515106fa
JH
1030 }
1031 }
1032 return rem;
1033}
1034
bd2f371d 1035static int fill(const char *path, int pathlen, int basename_offset,
dc81cf37
BW
1036 const struct attr_stack *stack,
1037 struct all_attrs_item *all_attrs, int rem)
d0bfd026 1038{
dc81cf37
BW
1039 for (; rem > 0 && stack; stack = stack->prev) {
1040 int i;
1041 const char *base = stack->origin ? stack->origin : "";
1042
1043 for (i = stack->num_matches - 1; 0 < rem && 0 <= i; i--) {
1044 const struct match_attr *a = stack->attrs[i];
1045 if (a->is_macro)
1046 continue;
1047 if (path_matches(path, pathlen, basename_offset,
1048 &a->u.pat, base, stack->originlen))
1049 rem = fill_one("fill", all_attrs, a, rem);
1050 }
d0bfd026 1051 }
dc81cf37 1052
d0bfd026
JH
1053 return rem;
1054}
1055
685b2925 1056static int macroexpand_one(struct all_attrs_item *all_attrs, int nr, int rem)
f48fd688 1057{
60a12722 1058 const struct all_attrs_item *item = &all_attrs[nr];
f48fd688 1059
60a12722
BW
1060 if (item->macro && item->value == ATTR__TRUE)
1061 return fill_one("expand", all_attrs, item->macro, rem);
1062 else
ec775c41 1063 return rem;
60a12722 1064}
ec775c41 1065
60a12722
BW
1066/*
1067 * Marks the attributes which are macros based on the attribute stack.
1068 * This prevents having to search through the attribute stack each time
1069 * a macro needs to be expanded during the fill stage.
1070 */
1071static void determine_macros(struct all_attrs_item *all_attrs,
1072 const struct attr_stack *stack)
1073{
1074 for (; stack; stack = stack->prev) {
1075 int i;
1076 for (i = stack->num_matches - 1; i >= 0; i--) {
1077 const struct match_attr *ma = stack->attrs[i];
1078 if (ma->is_macro) {
1079 int n = ma->u.attr->attr_nr;
1080 if (!all_attrs[n].macro) {
1081 all_attrs[n].macro = ma;
1082 }
1083 }
ec775c41 1084 }
4b0c6961 1085 }
f48fd688
JH
1086}
1087
2d721744 1088/*
685b2925
BW
1089 * Collect attributes for path into the array pointed to by check->all_attrs.
1090 * If check->check_nr is non-zero, only attributes in check[] are collected.
1091 * Otherwise all attributes are collected.
2d721744 1092 */
6bc2e3f7 1093static void collect_some_attrs(const char *path, struct attr_check *check)
d0bfd026 1094{
9db9eecf 1095 int i, pathlen, rem, dirlen;
bd2f371d
JH
1096 const char *cp, *last_slash = NULL;
1097 int basename_offset;
9db9eecf
NTND
1098
1099 for (cp = path; *cp; cp++) {
1100 if (*cp == '/' && cp[1])
1101 last_slash = cp;
1102 }
1103 pathlen = cp - path;
1104 if (last_slash) {
bd2f371d 1105 basename_offset = last_slash + 1 - path;
9db9eecf
NTND
1106 dirlen = last_slash - path;
1107 } else {
bd2f371d 1108 basename_offset = 0;
9db9eecf
NTND
1109 dirlen = 0;
1110 }
d0bfd026 1111
dc81cf37 1112 prepare_attr_stack(path, dirlen, &check->stack);
685b2925 1113 all_attrs_init(&g_attr_hashmap, check);
dc81cf37 1114 determine_macros(check->all_attrs, check->stack);
685b2925 1115
60a12722 1116 if (check->nr) {
06a604e6 1117 rem = 0;
6bc2e3f7 1118 for (i = 0; i < check->nr; i++) {
60a12722
BW
1119 int n = check->items[i].attr->attr_nr;
1120 struct all_attrs_item *item = &check->all_attrs[n];
1121 if (item->macro) {
1122 item->value = ATTR__UNSET;
06a604e6
NTND
1123 rem++;
1124 }
1125 }
6bc2e3f7 1126 if (rem == check->nr)
06a604e6
NTND
1127 return;
1128 }
d0bfd026 1129
685b2925 1130 rem = check->all_attrs_nr;
dc81cf37 1131 fill(path, pathlen, basename_offset, check->stack, check->all_attrs, rem);
2d721744
MH
1132}
1133
6bc2e3f7 1134int git_check_attr(const char *path, struct attr_check *check)
2d721744
MH
1135{
1136 int i;
1137
6bc2e3f7 1138 collect_some_attrs(path, check);
f48fd688 1139
6bc2e3f7 1140 for (i = 0; i < check->nr; i++) {
685b2925
BW
1141 size_t n = check->items[i].attr->attr_nr;
1142 const char *value = check->all_attrs[n].value;
515106fa
JH
1143 if (value == ATTR__UNKNOWN)
1144 value = ATTR__UNSET;
6bc2e3f7 1145 check->items[i].value = value;
515106fa 1146 }
f48fd688 1147
d0bfd026
JH
1148 return 0;
1149}
06f33c17 1150
7f864111 1151void git_all_attrs(const char *path, struct attr_check *check)
ee548df3 1152{
7f864111 1153 int i;
ee548df3 1154
7f864111 1155 attr_check_reset(check);
6bc2e3f7 1156 collect_some_attrs(path, check);
ee548df3 1157
685b2925
BW
1158 for (i = 0; i < check->all_attrs_nr; i++) {
1159 const char *name = check->all_attrs[i].attr->name;
1160 const char *value = check->all_attrs[i].value;
7f864111
JH
1161 struct attr_check_item *item;
1162 if (value == ATTR__UNSET || value == ATTR__UNKNOWN)
1163 continue;
1164 item = attr_check_append(check, git_attr(name));
1165 item->value = value;
ee548df3 1166 }
ee548df3
MH
1167}
1168
1a600b75
BW
1169void attr_start(void)
1170{
1171#ifndef NO_PTHREADS
1172 pthread_mutex_init(&g_attr_hashmap.mutex, NULL);
dc81cf37 1173 pthread_mutex_init(&check_vector.mutex, NULL);
1a600b75
BW
1174#endif
1175}