pathspec: allow querying for attributes
[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"
6df42ab9 12#include "exec_cmd.h"
d0bfd026 13#include "attr.h"
6eba6210 14#include "dir.h"
27547e5f 15#include "utf8.h"
860a74d9 16#include "quote.h"
1a600b75 17#include "thread-utils.h"
d0bfd026 18
a5e92abd
JH
19const char git_attr__true[] = "(builtin)true";
20const char git_attr__false[] = "\0(builtin)false";
21static 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
515106fa 26
d0bfd026
JH
27#ifndef DEBUG_ATTR
28#define DEBUG_ATTR 0
29#endif
30
31struct git_attr {
1a600b75 32 int attr_nr; /* unique attribute number */
1a600b75 33 char name[FLEX_ARRAY]; /* attribute name */
d0bfd026 34};
7d42ec54 35
ec4d77aa 36const char *git_attr_name(const struct git_attr *attr)
352404ac
MH
37{
38 return attr->name;
39}
40
1a600b75
BW
41struct attr_hashmap {
42 struct hashmap map;
43#ifndef NO_PTHREADS
44 pthread_mutex_t mutex;
45#endif
46};
47
48static inline void hashmap_lock(struct attr_hashmap *map)
49{
50#ifndef NO_PTHREADS
51 pthread_mutex_lock(&map->mutex);
52#endif
53}
54
55static inline void hashmap_unlock(struct attr_hashmap *map)
d0bfd026 56{
1a600b75
BW
57#ifndef NO_PTHREADS
58 pthread_mutex_unlock(&map->mutex);
59#endif
60}
d0bfd026 61
1a600b75
BW
62/*
63 * The global dictionary of all interned attributes. This
64 * is a singleton object which is shared between threads.
65 * Access to this dictionary must be surrounded with a mutex.
66 */
67static struct attr_hashmap g_attr_hashmap;
68
69/* The container for objects stored in "struct attr_hashmap" */
70struct attr_hash_entry {
71 struct hashmap_entry ent; /* must be the first member! */
72 const char *key; /* the key; memory should be owned by value */
73 size_t keylen; /* length of the key */
74 void *value; /* the stored value */
75};
76
77/* attr_hashmap comparison function */
78static int attr_hash_entry_cmp(const struct attr_hash_entry *a,
79 const struct attr_hash_entry *b,
80 void *unused)
81{
82 return (a->keylen != b->keylen) || strncmp(a->key, b->key, a->keylen);
83}
84
85/* Initialize an 'attr_hashmap' object */
86static void attr_hashmap_init(struct attr_hashmap *map)
87{
88 hashmap_init(&map->map, (hashmap_cmp_fn) attr_hash_entry_cmp, 0);
89}
90
91/*
92 * Retrieve the 'value' stored in a hashmap given the provided 'key'.
93 * If there is no matching entry, return NULL.
94 */
95static void *attr_hashmap_get(struct attr_hashmap *map,
96 const char *key, size_t keylen)
97{
98 struct attr_hash_entry k;
99 struct attr_hash_entry *e;
100
101 if (!map->map.tablesize)
102 attr_hashmap_init(map);
103
104 hashmap_entry_init(&k, memhash(key, keylen));
105 k.key = key;
106 k.keylen = keylen;
107 e = hashmap_get(&map->map, &k, NULL);
108
109 return e ? e->value : NULL;
110}
111
112/* Add 'value' to a hashmap based on the provided 'key'. */
113static void attr_hashmap_add(struct attr_hashmap *map,
114 const char *key, size_t keylen,
115 void *value)
116{
117 struct attr_hash_entry *e;
118
119 if (!map->map.tablesize)
120 attr_hashmap_init(map);
121
122 e = xmalloc(sizeof(struct attr_hash_entry));
123 hashmap_entry_init(e, memhash(key, keylen));
124 e->key = key;
125 e->keylen = keylen;
126 e->value = value;
127
128 hashmap_add(&map->map, e);
d0bfd026
JH
129}
130
685b2925
BW
131struct all_attrs_item {
132 const struct git_attr *attr;
133 const char *value;
60a12722
BW
134 /*
135 * If 'macro' is non-NULL, indicates that 'attr' is a macro based on
136 * the current attribute stack and contains a pointer to the match_attr
137 * definition of the macro
138 */
139 const struct match_attr *macro;
685b2925
BW
140};
141
142/*
143 * Reallocate and reinitialize the array of all attributes (which is used in
144 * the attribute collection process) in 'check' based on the global dictionary
145 * of attributes.
146 */
147static void all_attrs_init(struct attr_hashmap *map, struct attr_check *check)
148{
149 int i;
150
151 hashmap_lock(map);
152
153 if (map->map.size < check->all_attrs_nr)
154 die("BUG: interned attributes shouldn't be deleted");
155
156 /*
157 * If the number of attributes in the global dictionary has increased
158 * (or this attr_check instance doesn't have an initialized all_attrs
159 * field), reallocate the provided attr_check instance's all_attrs
160 * field and fill each entry with its corresponding git_attr.
161 */
162 if (map->map.size != check->all_attrs_nr) {
163 struct attr_hash_entry *e;
164 struct hashmap_iter iter;
165 hashmap_iter_init(&map->map, &iter);
166
167 REALLOC_ARRAY(check->all_attrs, map->map.size);
168 check->all_attrs_nr = map->map.size;
169
170 while ((e = hashmap_iter_next(&iter))) {
171 const struct git_attr *a = e->value;
172 check->all_attrs[a->attr_nr].attr = a;
173 }
174 }
175
176 hashmap_unlock(map);
177
178 /*
179 * Re-initialize every entry in check->all_attrs.
180 * This re-initialization can live outside of the locked region since
181 * the attribute dictionary is no longer being accessed.
182 */
183 for (i = 0; i < check->all_attrs_nr; i++) {
184 check->all_attrs[i].value = ATTR__UNKNOWN;
60a12722 185 check->all_attrs[i].macro = NULL;
685b2925
BW
186 }
187}
188
428103c7 189static int attr_name_valid(const char *name, size_t namelen)
e4aee10a
JH
190{
191 /*
d42453ab
MH
192 * Attribute name cannot begin with '-' and must consist of
193 * characters from [-A-Za-z0-9_.].
e4aee10a 194 */
c0b13b21 195 if (namelen <= 0 || *name == '-')
428103c7 196 return 0;
e4aee10a
JH
197 while (namelen--) {
198 char ch = *name++;
199 if (! (ch == '-' || ch == '.' || ch == '_' ||
200 ('0' <= ch && ch <= '9') ||
201 ('a' <= ch && ch <= 'z') ||
202 ('A' <= ch && ch <= 'Z')) )
428103c7 203 return 0;
e4aee10a 204 }
428103c7
JH
205 return 1;
206}
207
208static void report_invalid_attr(const char *name, size_t len,
209 const char *src, int lineno)
210{
211 struct strbuf err = STRBUF_INIT;
212 strbuf_addf(&err, _("%.*s is not a valid attribute name"),
213 (int) len, name);
214 fprintf(stderr, "%s: %s:%d\n", err.buf, src, lineno);
215 strbuf_release(&err);
e4aee10a
JH
216}
217
1a600b75
BW
218/*
219 * Given a 'name', lookup and return the corresponding attribute in the global
220 * dictionary. If no entry is found, create a new attribute and store it in
221 * the dictionary.
222 */
e810e063 223static const struct git_attr *git_attr_internal(const char *name, int namelen)
d0bfd026 224{
d0bfd026
JH
225 struct git_attr *a;
226
1a600b75 227 if (!attr_name_valid(name, namelen))
e4aee10a
JH
228 return NULL;
229
1a600b75
BW
230 hashmap_lock(&g_attr_hashmap);
231
232 a = attr_hashmap_get(&g_attr_hashmap, name, namelen);
233
234 if (!a) {
235 FLEX_ALLOC_MEM(a, name, name, namelen);
236 a->attr_nr = g_attr_hashmap.map.size;
1a600b75
BW
237
238 attr_hashmap_add(&g_attr_hashmap, a->name, namelen, a);
239 assert(a->attr_nr == (g_attr_hashmap.map.size - 1));
1a600b75
BW
240 }
241
242 hashmap_unlock(&g_attr_hashmap);
f48fd688 243
d0bfd026
JH
244 return a;
245}
246
e810e063 247const struct git_attr *git_attr(const char *name)
7fb0eaa2
JH
248{
249 return git_attr_internal(name, strlen(name));
250}
251
515106fa 252/* What does a matched pattern decide? */
d0bfd026 253struct attr_state {
e810e063 254 const struct git_attr *attr;
a5e92abd 255 const char *setto;
d0bfd026
JH
256};
257
82dce998
NTND
258struct pattern {
259 const char *pattern;
260 int patternlen;
261 int nowildcardlen;
f8708998 262 unsigned flags; /* EXC_FLAG_* */
82dce998
NTND
263};
264
ba845b75
MH
265/*
266 * One rule, as from a .gitattributes file.
267 *
268 * If is_macro is true, then u.attr is a pointer to the git_attr being
269 * defined.
270 *
4894f4ff
JH
271 * If is_macro is false, then u.pat is the filename pattern to which the
272 * rule applies.
ba845b75
MH
273 *
274 * In either case, num_attr is the number of attributes affected by
275 * this rule, and state is an array listing them. The attributes are
276 * listed as they appear in the file (macros unexpanded).
277 */
d0bfd026 278struct match_attr {
f48fd688 279 union {
82dce998 280 struct pattern pat;
e810e063 281 const struct git_attr *attr;
f48fd688
JH
282 } u;
283 char is_macro;
d0bfd026
JH
284 unsigned num_attr;
285 struct attr_state state[FLEX_ARRAY];
286};
287
288static const char blank[] = " \t\r\n";
289
d1751298
MH
290/*
291 * Parse a whitespace-delimited attribute state (i.e., "attr",
292 * "-attr", "!attr", or "attr=value") from the string starting at src.
293 * If e is not NULL, write the results to *e. Return a pointer to the
294 * remainder of the string (with leading whitespace removed), or NULL
295 * if there was an error.
296 */
515106fa 297static const char *parse_attr(const char *src, int lineno, const char *cp,
d1751298 298 struct attr_state *e)
515106fa
JH
299{
300 const char *ep, *equals;
301 int len;
302
303 ep = cp + strcspn(cp, blank);
304 equals = strchr(cp, '=');
305 if (equals && ep < equals)
306 equals = NULL;
307 if (equals)
308 len = equals - cp;
309 else
310 len = ep - cp;
d1751298 311 if (!e) {
515106fa
JH
312 if (*cp == '-' || *cp == '!') {
313 cp++;
314 len--;
315 }
428103c7
JH
316 if (!attr_name_valid(cp, len)) {
317 report_invalid_attr(cp, len, src, lineno);
515106fa
JH
318 return NULL;
319 }
320 } else {
5a884019
JH
321 /*
322 * As this function is always called twice, once with
323 * e == NULL in the first pass and then e != NULL in
428103c7 324 * the second pass, no need for attr_name_valid()
5a884019
JH
325 * check here.
326 */
515106fa
JH
327 if (*cp == '-' || *cp == '!') {
328 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
329 cp++;
330 len--;
331 }
332 else if (!equals)
333 e->setto = ATTR__TRUE;
334 else {
182af834 335 e->setto = xmemdupz(equals + 1, ep - equals - 1);
515106fa 336 }
7fb0eaa2 337 e->attr = git_attr_internal(cp, len);
515106fa 338 }
515106fa
JH
339 return ep + strspn(ep, blank);
340}
341
f48fd688
JH
342static struct match_attr *parse_attr_line(const char *line, const char *src,
343 int lineno, int macro_ok)
d0bfd026
JH
344{
345 int namelen;
d68e1c18 346 int num_attr, i;
85c4a0d0 347 const char *cp, *name, *states;
515106fa 348 struct match_attr *res = NULL;
f48fd688 349 int is_macro;
860a74d9 350 struct strbuf pattern = STRBUF_INIT;
d0bfd026
JH
351
352 cp = line + strspn(line, blank);
353 if (!*cp || *cp == '#')
354 return NULL;
355 name = cp;
860a74d9
NTND
356
357 if (*cp == '"' && !unquote_c_style(&pattern, name, &states)) {
358 name = pattern.buf;
359 namelen = pattern.len;
360 } else {
361 namelen = strcspn(name, blank);
362 states = name + namelen;
363 }
364
f48fd688 365 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
59556548 366 starts_with(name, ATTRIBUTE_MACRO_PREFIX)) {
f48fd688
JH
367 if (!macro_ok) {
368 fprintf(stderr, "%s not allowed: %s:%d\n",
369 name, src, lineno);
62af8969 370 goto fail_return;
f48fd688
JH
371 }
372 is_macro = 1;
373 name += strlen(ATTRIBUTE_MACRO_PREFIX);
374 name += strspn(name, blank);
375 namelen = strcspn(name, blank);
428103c7
JH
376 if (!attr_name_valid(name, namelen)) {
377 report_invalid_attr(name, namelen, src, lineno);
62af8969 378 goto fail_return;
e4aee10a 379 }
f48fd688
JH
380 }
381 else
382 is_macro = 0;
d0bfd026 383
85c4a0d0
MH
384 states += strspn(states, blank);
385
d68e1c18
MH
386 /* First pass to count the attr_states */
387 for (cp = states, num_attr = 0; *cp; num_attr++) {
388 cp = parse_attr(src, lineno, cp, NULL);
389 if (!cp)
62af8969 390 goto fail_return;
d0bfd026 391 }
d68e1c18
MH
392
393 res = xcalloc(1,
394 sizeof(*res) +
395 sizeof(struct attr_state) * num_attr +
396 (is_macro ? 0 : namelen + 1));
fad32bcd 397 if (is_macro) {
d68e1c18 398 res->u.attr = git_attr_internal(name, namelen);
fad32bcd 399 } else {
82dce998
NTND
400 char *p = (char *)&(res->state[num_attr]);
401 memcpy(p, name, namelen);
402 res->u.pat.pattern = p;
403 parse_exclude_pattern(&res->u.pat.pattern,
404 &res->u.pat.patternlen,
405 &res->u.pat.flags,
406 &res->u.pat.nowildcardlen);
8b1bd024
TR
407 if (res->u.pat.flags & EXC_FLAG_NEGATIVE) {
408 warning(_("Negative patterns are ignored in git attributes\n"
409 "Use '\\!' for literal leading exclamation."));
62af8969 410 goto fail_return;
8b1bd024 411 }
d0bfd026 412 }
d68e1c18
MH
413 res->is_macro = is_macro;
414 res->num_attr = num_attr;
415
416 /* Second pass to fill the attr_states */
417 for (cp = states, i = 0; *cp; i++) {
418 cp = parse_attr(src, lineno, cp, &(res->state[i]));
419 }
420
860a74d9 421 strbuf_release(&pattern);
d0bfd026 422 return res;
62af8969
JH
423
424fail_return:
860a74d9 425 strbuf_release(&pattern);
62af8969
JH
426 free(res);
427 return NULL;
d0bfd026
JH
428}
429
430/*
431 * Like info/exclude and .gitignore, the attribute information can
432 * come from many places.
433 *
434 * (1) .gitattribute file of the same directory;
515106fa
JH
435 * (2) .gitattribute file of the parent directory if (1) does not have
436 * any match; this goes recursively upwards, just like .gitignore.
437 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
d0bfd026
JH
438 *
439 * In the same file, later entries override the earlier match, so in the
440 * global list, we would have entries from info/attributes the earliest
441 * (reading the file from top to bottom), .gitattribute of the root
442 * directory (again, reading the file from top to bottom) down to the
443 * current directory, and then scan the list backwards to find the first match.
6d24e7a8 444 * This is exactly the same as what is_excluded() does in dir.c to deal with
6f416cba 445 * .gitignore file and info/excludes file as a fallback.
d0bfd026
JH
446 */
447
dc81cf37 448struct attr_stack {
d0bfd026
JH
449 struct attr_stack *prev;
450 char *origin;
cd6a0b26 451 size_t originlen;
d0bfd026 452 unsigned num_matches;
a4413118 453 unsigned alloc;
d0bfd026 454 struct match_attr **attrs;
dc81cf37 455};
d0bfd026 456
dc81cf37 457static void attr_stack_free(struct attr_stack *e)
d0bfd026
JH
458{
459 int i;
460 free(e->origin);
515106fa
JH
461 for (i = 0; i < e->num_matches; i++) {
462 struct match_attr *a = e->attrs[i];
463 int j;
464 for (j = 0; j < a->num_attr; j++) {
a5e92abd 465 const char *setto = a->state[j].setto;
515106fa
JH
466 if (setto == ATTR__TRUE ||
467 setto == ATTR__FALSE ||
468 setto == ATTR__UNSET ||
469 setto == ATTR__UNKNOWN)
470 ;
471 else
4b25d091 472 free((char *) setto);
515106fa
JH
473 }
474 free(a);
475 }
37475f97 476 free(e->attrs);
d0bfd026
JH
477 free(e);
478}
479
dc81cf37
BW
480static void drop_attr_stack(struct attr_stack **stack)
481{
482 while (*stack) {
483 struct attr_stack *elem = *stack;
484 *stack = elem->prev;
485 attr_stack_free(elem);
486 }
487}
488
489/* List of all attr_check structs; access should be surrounded by mutex */
490static struct check_vector {
491 size_t nr;
492 size_t alloc;
493 struct attr_check **checks;
494#ifndef NO_PTHREADS
495 pthread_mutex_t mutex;
496#endif
497} check_vector;
498
499static inline void vector_lock(void)
500{
501#ifndef NO_PTHREADS
502 pthread_mutex_lock(&check_vector.mutex);
503#endif
504}
505
506static inline void vector_unlock(void)
507{
508#ifndef NO_PTHREADS
509 pthread_mutex_unlock(&check_vector.mutex);
510#endif
511}
512
513static void check_vector_add(struct attr_check *c)
514{
515 vector_lock();
516
517 ALLOC_GROW(check_vector.checks,
518 check_vector.nr + 1,
519 check_vector.alloc);
520 check_vector.checks[check_vector.nr++] = c;
521
522 vector_unlock();
523}
524
525static void check_vector_remove(struct attr_check *check)
526{
527 int i;
528
529 vector_lock();
530
531 /* Find entry */
532 for (i = 0; i < check_vector.nr; i++)
533 if (check_vector.checks[i] == check)
534 break;
535
536 if (i >= check_vector.nr)
537 die("BUG: no entry found");
538
539 /* shift entries over */
540 for (; i < check_vector.nr - 1; i++)
541 check_vector.checks[i] = check_vector.checks[i + 1];
542
543 check_vector.nr--;
544
545 vector_unlock();
546}
547
548/* Iterate through all attr_check instances and drop their stacks */
549static void drop_all_attr_stacks(void)
550{
551 int i;
552
553 vector_lock();
554
555 for (i = 0; i < check_vector.nr; i++) {
556 drop_attr_stack(&check_vector.checks[i]->stack);
557 }
558
559 vector_unlock();
560}
561
37293768
JH
562struct attr_check *attr_check_alloc(void)
563{
dc81cf37
BW
564 struct attr_check *c = xcalloc(1, sizeof(struct attr_check));
565
566 /* save pointer to the check struct */
567 check_vector_add(c);
568
569 return c;
37293768
JH
570}
571
572struct attr_check *attr_check_initl(const char *one, ...)
573{
574 struct attr_check *check;
575 int cnt;
576 va_list params;
577 const char *param;
578
579 va_start(params, one);
580 for (cnt = 1; (param = va_arg(params, const char *)) != NULL; cnt++)
581 ;
582 va_end(params);
583
584 check = attr_check_alloc();
585 check->nr = cnt;
586 check->alloc = cnt;
587 check->items = xcalloc(cnt, sizeof(struct attr_check_item));
588
589 check->items[0].attr = git_attr(one);
590 va_start(params, one);
591 for (cnt = 1; cnt < check->nr; cnt++) {
592 const struct git_attr *attr;
593 param = va_arg(params, const char *);
594 if (!param)
595 die("BUG: counted %d != ended at %d",
596 check->nr, cnt);
597 attr = git_attr(param);
598 if (!attr)
599 die("BUG: %s: not a valid attribute name", param);
600 check->items[cnt].attr = attr;
601 }
602 va_end(params);
603 return check;
604}
605
b0db7046
BW
606struct attr_check *attr_check_dup(const struct attr_check *check)
607{
608 struct attr_check *ret;
609
610 if (!check)
611 return NULL;
612
613 ret = attr_check_alloc();
614
615 ret->nr = check->nr;
616 ret->alloc = check->alloc;
617 ALLOC_ARRAY(ret->items, ret->nr);
618 COPY_ARRAY(ret->items, check->items, ret->nr);
619
620 return ret;
621}
622
37293768
JH
623struct attr_check_item *attr_check_append(struct attr_check *check,
624 const struct git_attr *attr)
625{
626 struct attr_check_item *item;
627
628 ALLOC_GROW(check->items, check->nr + 1, check->alloc);
629 item = &check->items[check->nr++];
630 item->attr = attr;
631 return item;
632}
633
634void attr_check_reset(struct attr_check *check)
635{
636 check->nr = 0;
637}
638
639void attr_check_clear(struct attr_check *check)
640{
641 free(check->items);
642 check->items = NULL;
643 check->alloc = 0;
644 check->nr = 0;
685b2925
BW
645
646 free(check->all_attrs);
647 check->all_attrs = NULL;
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{
a4413118 723 FILE *fp = fopen(path, "r");
d0bfd026
JH
724 struct attr_stack *res;
725 char buf[2048];
f48fd688 726 int lineno = 0;
d0bfd026 727
11e50b27 728 if (!fp) {
8e950dab 729 if (errno != ENOENT && errno != ENOTDIR)
55b38a48 730 warn_on_inaccessible(path);
a4413118 731 return NULL;
11e50b27 732 }
a4413118 733 res = xcalloc(1, sizeof(*res));
27547e5f
JH
734 while (fgets(buf, sizeof(buf), fp)) {
735 char *bufp = buf;
736 if (!lineno)
737 skip_utf8_bom(&bufp, strlen(bufp));
738 handle_attr_line(res, bufp, path, ++lineno, macro_ok);
739 }
a4413118
JH
740 fclose(fp);
741 return res;
742}
d0bfd026 743
06f33c17 744static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
a4413118
JH
745{
746 struct attr_stack *res;
1a9d7e9b
JH
747 char *buf, *sp;
748 int lineno = 0;
f48fd688 749
ff366825 750 buf = read_blob_data_from_index(use_index ? use_index : &the_index, path, NULL);
1a9d7e9b 751 if (!buf)
06f33c17 752 return NULL;
1a9d7e9b 753
06f33c17 754 res = xcalloc(1, sizeof(*res));
1a9d7e9b
JH
755 for (sp = buf; *sp; ) {
756 char *ep;
757 int more;
263434f2
JH
758
759 ep = strchrnul(sp, '\n');
1a9d7e9b
JH
760 more = (*ep == '\n');
761 *ep = '\0';
762 handle_attr_line(res, sp, path, ++lineno, macro_ok);
763 sp = ep + more;
764 }
765 free(buf);
d0bfd026
JH
766 return res;
767}
768
06f33c17
JH
769static struct attr_stack *read_attr(const char *path, int macro_ok)
770{
0787dafd 771 struct attr_stack *res = NULL;
06f33c17 772
0787dafd 773 if (direction == GIT_ATTR_INDEX) {
06f33c17 774 res = read_attr_from_index(path, macro_ok);
0787dafd
BW
775 } else if (!is_bare_repository()) {
776 if (direction == GIT_ATTR_CHECKOUT) {
06f33c17 777 res = read_attr_from_index(path, macro_ok);
0787dafd
BW
778 if (!res)
779 res = read_attr_from_file(path, macro_ok);
780 } else if (direction == GIT_ATTR_CHECKIN) {
781 res = read_attr_from_file(path, macro_ok);
782 if (!res)
783 /*
784 * There is no checked out .gitattributes file
785 * there, but we might have it in the index.
786 * We allow operation in a sparsely checked out
787 * work tree, so read from it.
788 */
789 res = read_attr_from_index(path, macro_ok);
790 }
06f33c17 791 }
0787dafd 792
06f33c17
JH
793 if (!res)
794 res = xcalloc(1, sizeof(*res));
795 return res;
796}
797
d0bfd026
JH
798#if DEBUG_ATTR
799static void debug_info(const char *what, struct attr_stack *elem)
800{
801 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
802}
cf94ccda 803static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
f48fd688 804{
515106fa
JH
805 const char *value = v;
806
807 if (ATTR_TRUE(value))
808 value = "set";
809 else if (ATTR_FALSE(value))
810 value = "unset";
811 else if (ATTR_UNSET(value))
812 value = "unspecified";
813
814 fprintf(stderr, "%s: %s => %s (%s)\n",
815 what, attr->name, (char *) value, match);
f48fd688 816}
d0bfd026
JH
817#define debug_push(a) debug_info("push", (a))
818#define debug_pop(a) debug_info("pop", (a))
819#else
820#define debug_push(a) do { ; } while (0)
821#define debug_pop(a) do { ; } while (0)
f48fd688 822#define debug_set(a,b,c,d) do { ; } while (0)
034d35c9 823#endif /* DEBUG_ATTR */
d0bfd026 824
c5147722 825static const char *git_etc_gitattributes(void)
6df42ab9
PO
826{
827 static const char *system_wide;
828 if (!system_wide)
829 system_wide = system_path(ETC_GITATTRIBUTES);
830 return system_wide;
831}
832
dc81cf37
BW
833static const char *get_home_gitattributes(void)
834{
835 if (!git_attributes_file)
836 git_attributes_file = xdg_config_home("attributes");
837
838 return git_attributes_file;
839}
840
c5147722 841static int git_attr_system(void)
6df42ab9
PO
842{
843 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
844}
845
f932729c
JK
846static GIT_PATH_FUNC(git_path_info_attributes, INFOATTRIBUTES_FILE)
847
4c0ce074
JH
848static void push_stack(struct attr_stack **attr_stack_p,
849 struct attr_stack *elem, char *origin, size_t originlen)
850{
851 if (elem) {
852 elem->origin = origin;
853 if (origin)
854 elem->originlen = originlen;
855 elem->prev = *attr_stack_p;
856 *attr_stack_p = elem;
857 }
858}
859
dc81cf37 860static void bootstrap_attr_stack(struct attr_stack **stack)
f48fd688 861{
dc81cf37 862 struct attr_stack *e;
f48fd688 863
dc81cf37 864 if (*stack)
909ca7b9 865 return;
f48fd688 866
dc81cf37
BW
867 /* builtin frame */
868 e = read_attr_from_array(builtin_attr);
869 push_stack(stack, e, NULL, 0);
6df42ab9 870
dc81cf37
BW
871 /* system-wide frame */
872 if (git_attr_system()) {
873 e = read_attr_from_file(git_etc_gitattributes(), 1);
874 push_stack(stack, e, NULL, 0);
875 }
f48fd688 876
dc81cf37
BW
877 /* home directory */
878 if (get_home_gitattributes()) {
879 e = read_attr_from_file(get_home_gitattributes(), 1);
880 push_stack(stack, e, NULL, 0);
f48fd688 881 }
909ca7b9 882
dc81cf37 883 /* root directory */
0787dafd 884 e = read_attr(GITATTRIBUTES_FILE, 1);
dc81cf37 885 push_stack(stack, e, xstrdup(""), 0);
f0056f64 886
dc81cf37
BW
887 /* info frame */
888 if (startup_info->have_repository)
889 e = read_attr_from_file(git_path_info_attributes(), 1);
890 else
891 e = NULL;
892 if (!e)
893 e = xcalloc(1, sizeof(struct attr_stack));
894 push_stack(stack, e, NULL, 0);
f48fd688
JH
895}
896
dc81cf37
BW
897static void prepare_attr_stack(const char *path, int dirlen,
898 struct attr_stack **stack)
d0bfd026 899{
dc81cf37 900 struct attr_stack *info;
0787dafd 901 struct strbuf pathbuf = STRBUF_INIT;
a8727017 902
d0bfd026
JH
903 /*
904 * At the bottom of the attribute stack is the built-in
6df42ab9
PO
905 * set of attribute definitions, followed by the contents
906 * of $(prefix)/etc/gitattributes and a file specified by
907 * core.attributesfile. Then, contents from
d0bfd026
JH
908 * .gitattribute files from directories closer to the
909 * root to the ones in deeper directories are pushed
910 * to the stack. Finally, at the very top of the stack
911 * we always keep the contents of $GIT_DIR/info/attributes.
912 *
913 * When checking, we use entries from near the top of the
914 * stack, preferring $GIT_DIR/info/attributes, then
915 * .gitattributes in deeper directories to shallower ones,
916 * and finally use the built-in set as the default.
917 */
dc81cf37 918 bootstrap_attr_stack(stack);
d0bfd026
JH
919
920 /*
921 * Pop the "info" one that is always at the top of the stack.
922 */
dc81cf37
BW
923 info = *stack;
924 *stack = info->prev;
d0bfd026
JH
925
926 /*
927 * Pop the ones from directories that are not the prefix of
c432ef99
JH
928 * the path we are checking. Break out of the loop when we see
929 * the root one (whose origin is an empty string "") or the builtin
930 * one (whose origin is NULL) without popping it.
d0bfd026 931 */
dc81cf37
BW
932 while ((*stack)->origin) {
933 int namelen = (*stack)->originlen;
934 struct attr_stack *elem;
d0bfd026 935
dc81cf37 936 elem = *stack;
d0bfd026 937 if (namelen <= dirlen &&
1afca444
JK
938 !strncmp(elem->origin, path, namelen) &&
939 (!namelen || path[namelen] == '/'))
d0bfd026
JH
940 break;
941
942 debug_pop(elem);
dc81cf37
BW
943 *stack = elem->prev;
944 attr_stack_free(elem);
d0bfd026
JH
945 }
946
947 /*
0787dafd
BW
948 * bootstrap_attr_stack() should have added, and the
949 * above loop should have stopped before popping, the
950 * root element whose attr_stack->origin is set to an
951 * empty string.
d0bfd026 952 */
0787dafd
BW
953 assert((*stack)->origin);
954
955 strbuf_addstr(&pathbuf, (*stack)->origin);
956 /* Build up to the directory 'path' is in */
957 while (pathbuf.len < dirlen) {
958 size_t len = pathbuf.len;
959 struct attr_stack *next;
960 char *origin;
961
962 /* Skip path-separator */
963 if (len < dirlen && is_dir_sep(path[len]))
964 len++;
965 /* Find the end of the next component */
966 while (len < dirlen && !is_dir_sep(path[len]))
967 len++;
968
969 if (pathbuf.len > 0)
970 strbuf_addch(&pathbuf, '/');
971 strbuf_add(&pathbuf, path + pathbuf.len, (len - pathbuf.len));
972 strbuf_addf(&pathbuf, "/%s", GITATTRIBUTES_FILE);
973
974 next = read_attr(pathbuf.buf, 0);
975
976 /* reset the pathbuf to not include "/.gitattributes" */
977 strbuf_setlen(&pathbuf, len);
978
979 origin = xstrdup(pathbuf.buf);
980 push_stack(stack, next, origin, len);
97410b27 981 }
d4c98565 982
d0bfd026
JH
983 /*
984 * Finally push the "info" one at the top of the stack.
985 */
dc81cf37 986 push_stack(stack, info, NULL, 0);
0787dafd
BW
987
988 strbuf_release(&pathbuf);
d0bfd026
JH
989}
990
991static int path_matches(const char *pathname, int pathlen,
bd2f371d 992 int basename_offset,
82dce998 993 const struct pattern *pat,
d0bfd026
JH
994 const char *base, int baselen)
995{
82dce998
NTND
996 const char *pattern = pat->pattern;
997 int prefix = pat->nowildcardlen;
dc09e9ec 998 int isdir = (pathlen && pathname[pathlen - 1] == '/');
82dce998 999
dc09e9ec 1000 if ((pat->flags & EXC_FLAG_MUSTBEDIR) && !isdir)
94bc671a
JNA
1001 return 0;
1002
82dce998 1003 if (pat->flags & EXC_FLAG_NODIR) {
bd2f371d 1004 return match_basename(pathname + basename_offset,
dc09e9ec 1005 pathlen - basename_offset - isdir,
82dce998
NTND
1006 pattern, prefix,
1007 pat->patternlen, pat->flags);
d0bfd026 1008 }
dc09e9ec 1009 return match_pathname(pathname, pathlen - isdir,
82dce998
NTND
1010 base, baselen,
1011 pattern, prefix, pat->patternlen, pat->flags);
d0bfd026
JH
1012}
1013
685b2925 1014static int macroexpand_one(struct all_attrs_item *all_attrs, int nr, int rem);
ec775c41 1015
685b2925 1016static int fill_one(const char *what, struct all_attrs_item *all_attrs,
60a12722 1017 const struct match_attr *a, int rem)
515106fa 1018{
515106fa
JH
1019 int i;
1020
685b2925 1021 for (i = a->num_attr - 1; rem > 0 && i >= 0; i--) {
e810e063 1022 const struct git_attr *attr = a->state[i].attr;
685b2925 1023 const char **n = &(all_attrs[attr->attr_nr].value);
a5e92abd 1024 const char *v = a->state[i].setto;
515106fa
JH
1025
1026 if (*n == ATTR__UNKNOWN) {
426c27b7 1027 debug_set(what,
712efb1a 1028 a->is_macro ? a->u.attr->name : a->u.pat.pattern,
426c27b7 1029 attr, v);
515106fa
JH
1030 *n = v;
1031 rem--;
685b2925 1032 rem = macroexpand_one(all_attrs, attr->attr_nr, rem);
515106fa
JH
1033 }
1034 }
1035 return rem;
1036}
1037
bd2f371d 1038static int fill(const char *path, int pathlen, int basename_offset,
dc81cf37
BW
1039 const struct attr_stack *stack,
1040 struct all_attrs_item *all_attrs, int rem)
d0bfd026 1041{
dc81cf37
BW
1042 for (; rem > 0 && stack; stack = stack->prev) {
1043 int i;
1044 const char *base = stack->origin ? stack->origin : "";
1045
1046 for (i = stack->num_matches - 1; 0 < rem && 0 <= i; i--) {
1047 const struct match_attr *a = stack->attrs[i];
1048 if (a->is_macro)
1049 continue;
1050 if (path_matches(path, pathlen, basename_offset,
1051 &a->u.pat, base, stack->originlen))
1052 rem = fill_one("fill", all_attrs, a, rem);
1053 }
d0bfd026 1054 }
dc81cf37 1055
d0bfd026
JH
1056 return rem;
1057}
1058
685b2925 1059static int macroexpand_one(struct all_attrs_item *all_attrs, int nr, int rem)
f48fd688 1060{
60a12722 1061 const struct all_attrs_item *item = &all_attrs[nr];
f48fd688 1062
60a12722
BW
1063 if (item->macro && item->value == ATTR__TRUE)
1064 return fill_one("expand", all_attrs, item->macro, rem);
1065 else
ec775c41 1066 return rem;
60a12722 1067}
ec775c41 1068
60a12722
BW
1069/*
1070 * Marks the attributes which are macros based on the attribute stack.
1071 * This prevents having to search through the attribute stack each time
1072 * a macro needs to be expanded during the fill stage.
1073 */
1074static void determine_macros(struct all_attrs_item *all_attrs,
1075 const struct attr_stack *stack)
1076{
1077 for (; stack; stack = stack->prev) {
1078 int i;
1079 for (i = stack->num_matches - 1; i >= 0; i--) {
1080 const struct match_attr *ma = stack->attrs[i];
1081 if (ma->is_macro) {
1082 int n = ma->u.attr->attr_nr;
1083 if (!all_attrs[n].macro) {
1084 all_attrs[n].macro = ma;
1085 }
1086 }
ec775c41 1087 }
4b0c6961 1088 }
f48fd688
JH
1089}
1090
2d721744 1091/*
685b2925
BW
1092 * Collect attributes for path into the array pointed to by check->all_attrs.
1093 * If check->check_nr is non-zero, only attributes in check[] are collected.
1094 * Otherwise all attributes are collected.
2d721744 1095 */
6bc2e3f7 1096static void collect_some_attrs(const char *path, struct attr_check *check)
d0bfd026 1097{
9db9eecf 1098 int i, pathlen, rem, dirlen;
bd2f371d
JH
1099 const char *cp, *last_slash = NULL;
1100 int basename_offset;
9db9eecf
NTND
1101
1102 for (cp = path; *cp; cp++) {
1103 if (*cp == '/' && cp[1])
1104 last_slash = cp;
1105 }
1106 pathlen = cp - path;
1107 if (last_slash) {
bd2f371d 1108 basename_offset = last_slash + 1 - path;
9db9eecf
NTND
1109 dirlen = last_slash - path;
1110 } else {
bd2f371d 1111 basename_offset = 0;
9db9eecf
NTND
1112 dirlen = 0;
1113 }
d0bfd026 1114
dc81cf37 1115 prepare_attr_stack(path, dirlen, &check->stack);
685b2925 1116 all_attrs_init(&g_attr_hashmap, check);
dc81cf37 1117 determine_macros(check->all_attrs, check->stack);
685b2925 1118
60a12722 1119 if (check->nr) {
06a604e6 1120 rem = 0;
6bc2e3f7 1121 for (i = 0; i < check->nr; i++) {
60a12722
BW
1122 int n = check->items[i].attr->attr_nr;
1123 struct all_attrs_item *item = &check->all_attrs[n];
1124 if (item->macro) {
1125 item->value = ATTR__UNSET;
06a604e6
NTND
1126 rem++;
1127 }
1128 }
6bc2e3f7 1129 if (rem == check->nr)
06a604e6
NTND
1130 return;
1131 }
d0bfd026 1132
685b2925 1133 rem = check->all_attrs_nr;
dc81cf37 1134 fill(path, pathlen, basename_offset, check->stack, check->all_attrs, rem);
2d721744
MH
1135}
1136
6bc2e3f7 1137int git_check_attr(const char *path, struct attr_check *check)
2d721744
MH
1138{
1139 int i;
1140
6bc2e3f7 1141 collect_some_attrs(path, check);
f48fd688 1142
6bc2e3f7 1143 for (i = 0; i < check->nr; i++) {
685b2925
BW
1144 size_t n = check->items[i].attr->attr_nr;
1145 const char *value = check->all_attrs[n].value;
515106fa
JH
1146 if (value == ATTR__UNKNOWN)
1147 value = ATTR__UNSET;
6bc2e3f7 1148 check->items[i].value = value;
515106fa 1149 }
f48fd688 1150
d0bfd026
JH
1151 return 0;
1152}
06f33c17 1153
7f864111 1154void git_all_attrs(const char *path, struct attr_check *check)
ee548df3 1155{
7f864111 1156 int i;
ee548df3 1157
7f864111 1158 attr_check_reset(check);
6bc2e3f7 1159 collect_some_attrs(path, check);
ee548df3 1160
685b2925
BW
1161 for (i = 0; i < check->all_attrs_nr; i++) {
1162 const char *name = check->all_attrs[i].attr->name;
1163 const char *value = check->all_attrs[i].value;
7f864111
JH
1164 struct attr_check_item *item;
1165 if (value == ATTR__UNSET || value == ATTR__UNKNOWN)
1166 continue;
1167 item = attr_check_append(check, git_attr(name));
1168 item->value = value;
ee548df3 1169 }
ee548df3
MH
1170}
1171
1a600b75
BW
1172void attr_start(void)
1173{
1174#ifndef NO_PTHREADS
1175 pthread_mutex_init(&g_attr_hashmap.mutex, NULL);
dc81cf37 1176 pthread_mutex_init(&check_vector.mutex, NULL);
1a600b75
BW
1177#endif
1178}