attr: store attribute stack in attr_check structure
[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
606struct attr_check_item *attr_check_append(struct attr_check *check,
607 const struct git_attr *attr)
608{
609 struct attr_check_item *item;
610
611 ALLOC_GROW(check->items, check->nr + 1, check->alloc);
612 item = &check->items[check->nr++];
613 item->attr = attr;
614 return item;
615}
616
617void attr_check_reset(struct attr_check *check)
618{
619 check->nr = 0;
620}
621
622void attr_check_clear(struct attr_check *check)
623{
624 free(check->items);
625 check->items = NULL;
626 check->alloc = 0;
627 check->nr = 0;
685b2925
BW
628
629 free(check->all_attrs);
630 check->all_attrs = NULL;
631 check->all_attrs_nr = 0;
dc81cf37
BW
632
633 drop_attr_stack(&check->stack);
37293768
JH
634}
635
636void attr_check_free(struct attr_check *check)
637{
dc81cf37
BW
638 if (check) {
639 /* Remove check from the check vector */
640 check_vector_remove(check);
641
642 attr_check_clear(check);
643 free(check);
644 }
37293768
JH
645}
646
d0bfd026 647static const char *builtin_attr[] = {
155a4b71 648 "[attr]binary -diff -merge -text",
d0bfd026
JH
649 NULL,
650};
651
a4413118
JH
652static void handle_attr_line(struct attr_stack *res,
653 const char *line,
654 const char *src,
655 int lineno,
656 int macro_ok)
657{
658 struct match_attr *a;
659
660 a = parse_attr_line(line, src, lineno, macro_ok);
661 if (!a)
662 return;
3a7fa03d 663 ALLOC_GROW(res->attrs, res->num_matches + 1, res->alloc);
a4413118
JH
664 res->attrs[res->num_matches++] = a;
665}
666
d0bfd026
JH
667static struct attr_stack *read_attr_from_array(const char **list)
668{
669 struct attr_stack *res;
670 const char *line;
f48fd688 671 int lineno = 0;
d0bfd026
JH
672
673 res = xcalloc(1, sizeof(*res));
a4413118
JH
674 while ((line = *(list++)) != NULL)
675 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
d0bfd026
JH
676 return res;
677}
678
7d42ec54
JH
679/*
680 * NEEDSWORK: these two are tricky. The callers assume there is a
681 * single, system-wide global state "where we read attributes from?"
682 * and when the state is flipped by calling git_attr_set_direction(),
683 * attr_stack is discarded so that subsequent attr_check will lazily
684 * read from the right place. And they do not know or care who called
685 * by them uses the attribute subsystem, hence have no knowledge of
686 * existing git_attr_check instances or future ones that will be
687 * created).
688 *
689 * Probably we need a thread_local that holds these two variables,
690 * and a list of git_attr_check instances (which need to be maintained
691 * by hooking into git_attr_check_alloc(), git_attr_check_initl(), and
692 * git_attr_check_clear(). Then git_attr_set_direction() updates the
693 * fields in that thread_local for these two variables, iterate over
694 * all the active git_attr_check instances and discard the attr_stack
695 * they hold. Yuck, but it sounds doable.
696 */
06f33c17
JH
697static enum git_attr_direction direction;
698static struct index_state *use_index;
699
f48fd688 700static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
d0bfd026 701{
a4413118 702 FILE *fp = fopen(path, "r");
d0bfd026
JH
703 struct attr_stack *res;
704 char buf[2048];
f48fd688 705 int lineno = 0;
d0bfd026 706
11e50b27 707 if (!fp) {
8e950dab 708 if (errno != ENOENT && errno != ENOTDIR)
55b38a48 709 warn_on_inaccessible(path);
a4413118 710 return NULL;
11e50b27 711 }
a4413118 712 res = xcalloc(1, sizeof(*res));
27547e5f
JH
713 while (fgets(buf, sizeof(buf), fp)) {
714 char *bufp = buf;
715 if (!lineno)
716 skip_utf8_bom(&bufp, strlen(bufp));
717 handle_attr_line(res, bufp, path, ++lineno, macro_ok);
718 }
a4413118
JH
719 fclose(fp);
720 return res;
721}
d0bfd026 722
06f33c17 723static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
a4413118
JH
724{
725 struct attr_stack *res;
1a9d7e9b
JH
726 char *buf, *sp;
727 int lineno = 0;
f48fd688 728
ff366825 729 buf = read_blob_data_from_index(use_index ? use_index : &the_index, path, NULL);
1a9d7e9b 730 if (!buf)
06f33c17 731 return NULL;
1a9d7e9b 732
06f33c17 733 res = xcalloc(1, sizeof(*res));
1a9d7e9b
JH
734 for (sp = buf; *sp; ) {
735 char *ep;
736 int more;
263434f2
JH
737
738 ep = strchrnul(sp, '\n');
1a9d7e9b
JH
739 more = (*ep == '\n');
740 *ep = '\0';
741 handle_attr_line(res, sp, path, ++lineno, macro_ok);
742 sp = ep + more;
743 }
744 free(buf);
d0bfd026
JH
745 return res;
746}
747
06f33c17
JH
748static struct attr_stack *read_attr(const char *path, int macro_ok)
749{
750 struct attr_stack *res;
751
752 if (direction == GIT_ATTR_CHECKOUT) {
753 res = read_attr_from_index(path, macro_ok);
754 if (!res)
755 res = read_attr_from_file(path, macro_ok);
756 }
4191e80a 757 else if (direction == GIT_ATTR_CHECKIN) {
06f33c17
JH
758 res = read_attr_from_file(path, macro_ok);
759 if (!res)
760 /*
761 * There is no checked out .gitattributes file there, but
762 * we might have it in the index. We allow operation in a
763 * sparsely checked out work tree, so read from it.
764 */
765 res = read_attr_from_index(path, macro_ok);
766 }
4191e80a
NTND
767 else
768 res = read_attr_from_index(path, macro_ok);
06f33c17
JH
769 if (!res)
770 res = xcalloc(1, sizeof(*res));
771 return res;
772}
773
d0bfd026
JH
774#if DEBUG_ATTR
775static void debug_info(const char *what, struct attr_stack *elem)
776{
777 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
778}
cf94ccda 779static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
f48fd688 780{
515106fa
JH
781 const char *value = v;
782
783 if (ATTR_TRUE(value))
784 value = "set";
785 else if (ATTR_FALSE(value))
786 value = "unset";
787 else if (ATTR_UNSET(value))
788 value = "unspecified";
789
790 fprintf(stderr, "%s: %s => %s (%s)\n",
791 what, attr->name, (char *) value, match);
f48fd688 792}
d0bfd026
JH
793#define debug_push(a) debug_info("push", (a))
794#define debug_pop(a) debug_info("pop", (a))
795#else
796#define debug_push(a) do { ; } while (0)
797#define debug_pop(a) do { ; } while (0)
f48fd688 798#define debug_set(a,b,c,d) do { ; } while (0)
034d35c9 799#endif /* DEBUG_ATTR */
d0bfd026 800
c5147722 801static const char *git_etc_gitattributes(void)
6df42ab9
PO
802{
803 static const char *system_wide;
804 if (!system_wide)
805 system_wide = system_path(ETC_GITATTRIBUTES);
806 return system_wide;
807}
808
dc81cf37
BW
809static const char *get_home_gitattributes(void)
810{
811 if (!git_attributes_file)
812 git_attributes_file = xdg_config_home("attributes");
813
814 return git_attributes_file;
815}
816
c5147722 817static int git_attr_system(void)
6df42ab9
PO
818{
819 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
820}
821
f932729c
JK
822static GIT_PATH_FUNC(git_path_info_attributes, INFOATTRIBUTES_FILE)
823
4c0ce074
JH
824static void push_stack(struct attr_stack **attr_stack_p,
825 struct attr_stack *elem, char *origin, size_t originlen)
826{
827 if (elem) {
828 elem->origin = origin;
829 if (origin)
830 elem->originlen = originlen;
831 elem->prev = *attr_stack_p;
832 *attr_stack_p = elem;
833 }
834}
835
dc81cf37 836static void bootstrap_attr_stack(struct attr_stack **stack)
f48fd688 837{
dc81cf37 838 struct attr_stack *e;
f48fd688 839
dc81cf37 840 if (*stack)
909ca7b9 841 return;
f48fd688 842
dc81cf37
BW
843 /* builtin frame */
844 e = read_attr_from_array(builtin_attr);
845 push_stack(stack, e, NULL, 0);
6df42ab9 846
dc81cf37
BW
847 /* system-wide frame */
848 if (git_attr_system()) {
849 e = read_attr_from_file(git_etc_gitattributes(), 1);
850 push_stack(stack, e, NULL, 0);
851 }
f48fd688 852
dc81cf37
BW
853 /* home directory */
854 if (get_home_gitattributes()) {
855 e = read_attr_from_file(get_home_gitattributes(), 1);
856 push_stack(stack, e, NULL, 0);
f48fd688 857 }
909ca7b9 858
dc81cf37
BW
859 /* root directory */
860 if (!is_bare_repository() || direction == GIT_ATTR_INDEX)
861 e = read_attr(GITATTRIBUTES_FILE, 1);
f0056f64 862 else
dc81cf37
BW
863 e = xcalloc(1, sizeof(struct attr_stack));
864 push_stack(stack, e, xstrdup(""), 0);
f0056f64 865
dc81cf37
BW
866 /* info frame */
867 if (startup_info->have_repository)
868 e = read_attr_from_file(git_path_info_attributes(), 1);
869 else
870 e = NULL;
871 if (!e)
872 e = xcalloc(1, sizeof(struct attr_stack));
873 push_stack(stack, e, NULL, 0);
f48fd688
JH
874}
875
dc81cf37
BW
876static void prepare_attr_stack(const char *path, int dirlen,
877 struct attr_stack **stack)
d0bfd026 878{
dc81cf37 879 struct attr_stack *info;
a8727017 880
d0bfd026
JH
881 /*
882 * At the bottom of the attribute stack is the built-in
6df42ab9
PO
883 * set of attribute definitions, followed by the contents
884 * of $(prefix)/etc/gitattributes and a file specified by
885 * core.attributesfile. Then, contents from
d0bfd026
JH
886 * .gitattribute files from directories closer to the
887 * root to the ones in deeper directories are pushed
888 * to the stack. Finally, at the very top of the stack
889 * we always keep the contents of $GIT_DIR/info/attributes.
890 *
891 * When checking, we use entries from near the top of the
892 * stack, preferring $GIT_DIR/info/attributes, then
893 * .gitattributes in deeper directories to shallower ones,
894 * and finally use the built-in set as the default.
895 */
dc81cf37 896 bootstrap_attr_stack(stack);
d0bfd026
JH
897
898 /*
899 * Pop the "info" one that is always at the top of the stack.
900 */
dc81cf37
BW
901 info = *stack;
902 *stack = info->prev;
d0bfd026
JH
903
904 /*
905 * Pop the ones from directories that are not the prefix of
c432ef99
JH
906 * the path we are checking. Break out of the loop when we see
907 * the root one (whose origin is an empty string "") or the builtin
908 * one (whose origin is NULL) without popping it.
d0bfd026 909 */
dc81cf37
BW
910 while ((*stack)->origin) {
911 int namelen = (*stack)->originlen;
912 struct attr_stack *elem;
d0bfd026 913
dc81cf37 914 elem = *stack;
d0bfd026 915 if (namelen <= dirlen &&
1afca444
JK
916 !strncmp(elem->origin, path, namelen) &&
917 (!namelen || path[namelen] == '/'))
d0bfd026
JH
918 break;
919
920 debug_pop(elem);
dc81cf37
BW
921 *stack = elem->prev;
922 attr_stack_free(elem);
d0bfd026
JH
923 }
924
925 /*
926 * Read from parent directories and push them down
927 */
4191e80a 928 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
c432ef99
JH
929 /*
930 * bootstrap_attr_stack() should have added, and the
931 * above loop should have stopped before popping, the
932 * root element whose attr_stack->origin is set to an
933 * empty string.
934 */
97410b27 935 struct strbuf pathbuf = STRBUF_INIT;
2d35d556 936
dc81cf37
BW
937 assert((*stack)->origin);
938 strbuf_addstr(&pathbuf, (*stack)->origin);
939 /* Build up to the directory 'path' is in */
940 while (pathbuf.len < dirlen) {
941 size_t len = pathbuf.len;
942 struct attr_stack *next;
4c0ce074
JH
943 char *origin;
944
dc81cf37
BW
945 /* Skip path-separator */
946 if (len < dirlen && is_dir_sep(path[len]))
947 len++;
948 /* Find the end of the next component */
949 while (len < dirlen && !is_dir_sep(path[len]))
950 len++;
951
952 if (pathbuf.len > 0)
953 strbuf_addch(&pathbuf, '/');
954 strbuf_add(&pathbuf, path + pathbuf.len,
955 (len - pathbuf.len));
956 strbuf_addf(&pathbuf, "/%s", GITATTRIBUTES_FILE);
957
958 next = read_attr(pathbuf.buf, 0);
d0bfd026 959
dc81cf37
BW
960 /* reset the pathbuf to not include "/.gitattributes" */
961 strbuf_setlen(&pathbuf, len);
962
963 origin = xstrdup(pathbuf.buf);
964 push_stack(stack, next, origin, len);
965
966 }
97410b27
BC
967 strbuf_release(&pathbuf);
968 }
d4c98565 969
d0bfd026
JH
970 /*
971 * Finally push the "info" one at the top of the stack.
972 */
dc81cf37 973 push_stack(stack, info, NULL, 0);
d0bfd026
JH
974}
975
976static int path_matches(const char *pathname, int pathlen,
bd2f371d 977 int basename_offset,
82dce998 978 const struct pattern *pat,
d0bfd026
JH
979 const char *base, int baselen)
980{
82dce998
NTND
981 const char *pattern = pat->pattern;
982 int prefix = pat->nowildcardlen;
dc09e9ec 983 int isdir = (pathlen && pathname[pathlen - 1] == '/');
82dce998 984
dc09e9ec 985 if ((pat->flags & EXC_FLAG_MUSTBEDIR) && !isdir)
94bc671a
JNA
986 return 0;
987
82dce998 988 if (pat->flags & EXC_FLAG_NODIR) {
bd2f371d 989 return match_basename(pathname + basename_offset,
dc09e9ec 990 pathlen - basename_offset - isdir,
82dce998
NTND
991 pattern, prefix,
992 pat->patternlen, pat->flags);
d0bfd026 993 }
dc09e9ec 994 return match_pathname(pathname, pathlen - isdir,
82dce998
NTND
995 base, baselen,
996 pattern, prefix, pat->patternlen, pat->flags);
d0bfd026
JH
997}
998
685b2925 999static int macroexpand_one(struct all_attrs_item *all_attrs, int nr, int rem);
ec775c41 1000
685b2925 1001static int fill_one(const char *what, struct all_attrs_item *all_attrs,
60a12722 1002 const struct match_attr *a, int rem)
515106fa 1003{
515106fa
JH
1004 int i;
1005
685b2925 1006 for (i = a->num_attr - 1; rem > 0 && i >= 0; i--) {
e810e063 1007 const struct git_attr *attr = a->state[i].attr;
685b2925 1008 const char **n = &(all_attrs[attr->attr_nr].value);
a5e92abd 1009 const char *v = a->state[i].setto;
515106fa
JH
1010
1011 if (*n == ATTR__UNKNOWN) {
426c27b7 1012 debug_set(what,
712efb1a 1013 a->is_macro ? a->u.attr->name : a->u.pat.pattern,
426c27b7 1014 attr, v);
515106fa
JH
1015 *n = v;
1016 rem--;
685b2925 1017 rem = macroexpand_one(all_attrs, attr->attr_nr, rem);
515106fa
JH
1018 }
1019 }
1020 return rem;
1021}
1022
bd2f371d 1023static int fill(const char *path, int pathlen, int basename_offset,
dc81cf37
BW
1024 const struct attr_stack *stack,
1025 struct all_attrs_item *all_attrs, int rem)
d0bfd026 1026{
dc81cf37
BW
1027 for (; rem > 0 && stack; stack = stack->prev) {
1028 int i;
1029 const char *base = stack->origin ? stack->origin : "";
1030
1031 for (i = stack->num_matches - 1; 0 < rem && 0 <= i; i--) {
1032 const struct match_attr *a = stack->attrs[i];
1033 if (a->is_macro)
1034 continue;
1035 if (path_matches(path, pathlen, basename_offset,
1036 &a->u.pat, base, stack->originlen))
1037 rem = fill_one("fill", all_attrs, a, rem);
1038 }
d0bfd026 1039 }
dc81cf37 1040
d0bfd026
JH
1041 return rem;
1042}
1043
685b2925 1044static int macroexpand_one(struct all_attrs_item *all_attrs, int nr, int rem)
f48fd688 1045{
60a12722 1046 const struct all_attrs_item *item = &all_attrs[nr];
f48fd688 1047
60a12722
BW
1048 if (item->macro && item->value == ATTR__TRUE)
1049 return fill_one("expand", all_attrs, item->macro, rem);
1050 else
ec775c41 1051 return rem;
60a12722 1052}
ec775c41 1053
60a12722
BW
1054/*
1055 * Marks the attributes which are macros based on the attribute stack.
1056 * This prevents having to search through the attribute stack each time
1057 * a macro needs to be expanded during the fill stage.
1058 */
1059static void determine_macros(struct all_attrs_item *all_attrs,
1060 const struct attr_stack *stack)
1061{
1062 for (; stack; stack = stack->prev) {
1063 int i;
1064 for (i = stack->num_matches - 1; i >= 0; i--) {
1065 const struct match_attr *ma = stack->attrs[i];
1066 if (ma->is_macro) {
1067 int n = ma->u.attr->attr_nr;
1068 if (!all_attrs[n].macro) {
1069 all_attrs[n].macro = ma;
1070 }
1071 }
ec775c41 1072 }
4b0c6961 1073 }
f48fd688
JH
1074}
1075
2d721744 1076/*
685b2925
BW
1077 * Collect attributes for path into the array pointed to by check->all_attrs.
1078 * If check->check_nr is non-zero, only attributes in check[] are collected.
1079 * Otherwise all attributes are collected.
2d721744 1080 */
6bc2e3f7 1081static void collect_some_attrs(const char *path, struct attr_check *check)
d0bfd026 1082{
9db9eecf 1083 int i, pathlen, rem, dirlen;
bd2f371d
JH
1084 const char *cp, *last_slash = NULL;
1085 int basename_offset;
9db9eecf
NTND
1086
1087 for (cp = path; *cp; cp++) {
1088 if (*cp == '/' && cp[1])
1089 last_slash = cp;
1090 }
1091 pathlen = cp - path;
1092 if (last_slash) {
bd2f371d 1093 basename_offset = last_slash + 1 - path;
9db9eecf
NTND
1094 dirlen = last_slash - path;
1095 } else {
bd2f371d 1096 basename_offset = 0;
9db9eecf
NTND
1097 dirlen = 0;
1098 }
d0bfd026 1099
dc81cf37 1100 prepare_attr_stack(path, dirlen, &check->stack);
685b2925 1101 all_attrs_init(&g_attr_hashmap, check);
dc81cf37 1102 determine_macros(check->all_attrs, check->stack);
685b2925 1103
60a12722 1104 if (check->nr) {
06a604e6 1105 rem = 0;
6bc2e3f7 1106 for (i = 0; i < check->nr; i++) {
60a12722
BW
1107 int n = check->items[i].attr->attr_nr;
1108 struct all_attrs_item *item = &check->all_attrs[n];
1109 if (item->macro) {
1110 item->value = ATTR__UNSET;
06a604e6
NTND
1111 rem++;
1112 }
1113 }
6bc2e3f7 1114 if (rem == check->nr)
06a604e6
NTND
1115 return;
1116 }
d0bfd026 1117
685b2925 1118 rem = check->all_attrs_nr;
dc81cf37 1119 fill(path, pathlen, basename_offset, check->stack, check->all_attrs, rem);
2d721744
MH
1120}
1121
6bc2e3f7 1122int git_check_attr(const char *path, struct attr_check *check)
2d721744
MH
1123{
1124 int i;
1125
6bc2e3f7 1126 collect_some_attrs(path, check);
f48fd688 1127
6bc2e3f7 1128 for (i = 0; i < check->nr; i++) {
685b2925
BW
1129 size_t n = check->items[i].attr->attr_nr;
1130 const char *value = check->all_attrs[n].value;
515106fa
JH
1131 if (value == ATTR__UNKNOWN)
1132 value = ATTR__UNSET;
6bc2e3f7 1133 check->items[i].value = value;
515106fa 1134 }
f48fd688 1135
d0bfd026
JH
1136 return 0;
1137}
06f33c17 1138
7f864111 1139void git_all_attrs(const char *path, struct attr_check *check)
ee548df3 1140{
7f864111 1141 int i;
ee548df3 1142
7f864111 1143 attr_check_reset(check);
6bc2e3f7 1144 collect_some_attrs(path, check);
ee548df3 1145
685b2925
BW
1146 for (i = 0; i < check->all_attrs_nr; i++) {
1147 const char *name = check->all_attrs[i].attr->name;
1148 const char *value = check->all_attrs[i].value;
7f864111
JH
1149 struct attr_check_item *item;
1150 if (value == ATTR__UNSET || value == ATTR__UNKNOWN)
1151 continue;
1152 item = attr_check_append(check, git_attr(name));
1153 item->value = value;
ee548df3 1154 }
ee548df3
MH
1155}
1156
06f33c17
JH
1157void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
1158{
1159 enum git_attr_direction old = direction;
4191e80a
NTND
1160
1161 if (is_bare_repository() && new != GIT_ATTR_INDEX)
1162 die("BUG: non-INDEX attr direction in a bare repo");
1163
06f33c17
JH
1164 direction = new;
1165 if (new != old)
dc81cf37 1166 drop_all_attr_stacks();
06f33c17
JH
1167 use_index = istate;
1168}
1a600b75
BW
1169
1170void attr_start(void)
1171{
1172#ifndef NO_PTHREADS
1173 pthread_mutex_init(&g_attr_hashmap.mutex, NULL);
dc81cf37 1174 pthread_mutex_init(&check_vector.mutex, NULL);
1a600b75
BW
1175#endif
1176}