update-index/diff-index: use core.preloadindex to improve performance
[git/mingw/j6t.git] / attr.c
blob887a9ae46b7044489845abf2072339271caf2d4a
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.
6 * One basic design decision here is that we are not going to support
7 * an insanely large number of attributes.
8 */
10 #define NO_THE_INDEX_COMPATIBILITY_MACROS
11 #include "cache.h"
12 #include "exec_cmd.h"
13 #include "attr.h"
14 #include "dir.h"
16 const char git_attr__true[] = "(builtin)true";
17 const char git_attr__false[] = "\0(builtin)false";
18 static const char git_attr__unknown[] = "(builtin)unknown";
19 #define ATTR__TRUE git_attr__true
20 #define ATTR__FALSE git_attr__false
21 #define ATTR__UNSET NULL
22 #define ATTR__UNKNOWN git_attr__unknown
24 /* This is a randomly chosen prime. */
25 #define HASHSIZE 257
27 #ifndef DEBUG_ATTR
28 #define DEBUG_ATTR 0
29 #endif
31 struct git_attr {
32 struct git_attr *next;
33 unsigned h;
34 int attr_nr;
35 char name[FLEX_ARRAY];
37 static int attr_nr;
39 static struct git_attr_check *check_all_attr;
40 static struct git_attr *(git_attr_hash[HASHSIZE]);
42 char *git_attr_name(struct git_attr *attr)
44 return attr->name;
47 static unsigned hash_name(const char *name, int namelen)
49 unsigned val = 0, c;
51 while (namelen--) {
52 c = *name++;
53 val = ((val << 7) | (val >> 22)) ^ c;
55 return val;
58 static int invalid_attr_name(const char *name, int namelen)
61 * Attribute name cannot begin with '-' and must consist of
62 * characters from [-A-Za-z0-9_.].
64 if (namelen <= 0 || *name == '-')
65 return -1;
66 while (namelen--) {
67 char ch = *name++;
68 if (! (ch == '-' || ch == '.' || ch == '_' ||
69 ('0' <= ch && ch <= '9') ||
70 ('a' <= ch && ch <= 'z') ||
71 ('A' <= ch && ch <= 'Z')) )
72 return -1;
74 return 0;
77 static struct git_attr *git_attr_internal(const char *name, int len)
79 unsigned hval = hash_name(name, len);
80 unsigned pos = hval % HASHSIZE;
81 struct git_attr *a;
83 for (a = git_attr_hash[pos]; a; a = a->next) {
84 if (a->h == hval &&
85 !memcmp(a->name, name, len) && !a->name[len])
86 return a;
89 if (invalid_attr_name(name, len))
90 return NULL;
92 a = xmalloc(sizeof(*a) + len + 1);
93 memcpy(a->name, name, len);
94 a->name[len] = 0;
95 a->h = hval;
96 a->next = git_attr_hash[pos];
97 a->attr_nr = attr_nr++;
98 git_attr_hash[pos] = a;
100 check_all_attr = xrealloc(check_all_attr,
101 sizeof(*check_all_attr) * attr_nr);
102 check_all_attr[a->attr_nr].attr = a;
103 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
104 return a;
107 struct git_attr *git_attr(const char *name)
109 return git_attr_internal(name, strlen(name));
112 /* What does a matched pattern decide? */
113 struct attr_state {
114 struct git_attr *attr;
115 const char *setto;
119 * One rule, as from a .gitattributes file.
121 * If is_macro is true, then u.attr is a pointer to the git_attr being
122 * defined.
124 * If is_macro is false, then u.pattern points at the filename pattern
125 * to which the rule applies. (The memory pointed to is part of the
126 * memory block allocated for the match_attr instance.)
128 * In either case, num_attr is the number of attributes affected by
129 * this rule, and state is an array listing them. The attributes are
130 * listed as they appear in the file (macros unexpanded).
132 struct match_attr {
133 union {
134 char *pattern;
135 struct git_attr *attr;
136 } u;
137 char is_macro;
138 unsigned num_attr;
139 struct attr_state state[FLEX_ARRAY];
142 static const char blank[] = " \t\r\n";
145 * Parse a whitespace-delimited attribute state (i.e., "attr",
146 * "-attr", "!attr", or "attr=value") from the string starting at src.
147 * If e is not NULL, write the results to *e. Return a pointer to the
148 * remainder of the string (with leading whitespace removed), or NULL
149 * if there was an error.
151 static const char *parse_attr(const char *src, int lineno, const char *cp,
152 struct attr_state *e)
154 const char *ep, *equals;
155 int len;
157 ep = cp + strcspn(cp, blank);
158 equals = strchr(cp, '=');
159 if (equals && ep < equals)
160 equals = NULL;
161 if (equals)
162 len = equals - cp;
163 else
164 len = ep - cp;
165 if (!e) {
166 if (*cp == '-' || *cp == '!') {
167 cp++;
168 len--;
170 if (invalid_attr_name(cp, len)) {
171 fprintf(stderr,
172 "%.*s is not a valid attribute name: %s:%d\n",
173 len, cp, src, lineno);
174 return NULL;
176 } else {
177 if (*cp == '-' || *cp == '!') {
178 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
179 cp++;
180 len--;
182 else if (!equals)
183 e->setto = ATTR__TRUE;
184 else {
185 e->setto = xmemdupz(equals + 1, ep - equals - 1);
187 e->attr = git_attr_internal(cp, len);
189 return ep + strspn(ep, blank);
192 static struct match_attr *parse_attr_line(const char *line, const char *src,
193 int lineno, int macro_ok)
195 int namelen;
196 int num_attr, i;
197 const char *cp, *name, *states;
198 struct match_attr *res = NULL;
199 int is_macro;
201 cp = line + strspn(line, blank);
202 if (!*cp || *cp == '#')
203 return NULL;
204 name = cp;
205 namelen = strcspn(name, blank);
206 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
207 !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
208 if (!macro_ok) {
209 fprintf(stderr, "%s not allowed: %s:%d\n",
210 name, src, lineno);
211 return NULL;
213 is_macro = 1;
214 name += strlen(ATTRIBUTE_MACRO_PREFIX);
215 name += strspn(name, blank);
216 namelen = strcspn(name, blank);
217 if (invalid_attr_name(name, namelen)) {
218 fprintf(stderr,
219 "%.*s is not a valid attribute name: %s:%d\n",
220 namelen, name, src, lineno);
221 return NULL;
224 else
225 is_macro = 0;
227 states = name + namelen;
228 states += strspn(states, blank);
230 /* First pass to count the attr_states */
231 for (cp = states, num_attr = 0; *cp; num_attr++) {
232 cp = parse_attr(src, lineno, cp, NULL);
233 if (!cp)
234 return NULL;
237 res = xcalloc(1,
238 sizeof(*res) +
239 sizeof(struct attr_state) * num_attr +
240 (is_macro ? 0 : namelen + 1));
241 if (is_macro)
242 res->u.attr = git_attr_internal(name, namelen);
243 else {
244 res->u.pattern = (char *)&(res->state[num_attr]);
245 memcpy(res->u.pattern, name, namelen);
246 res->u.pattern[namelen] = 0;
248 res->is_macro = is_macro;
249 res->num_attr = num_attr;
251 /* Second pass to fill the attr_states */
252 for (cp = states, i = 0; *cp; i++) {
253 cp = parse_attr(src, lineno, cp, &(res->state[i]));
256 return res;
260 * Like info/exclude and .gitignore, the attribute information can
261 * come from many places.
263 * (1) .gitattribute file of the same directory;
264 * (2) .gitattribute file of the parent directory if (1) does not have
265 * any match; this goes recursively upwards, just like .gitignore.
266 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
268 * In the same file, later entries override the earlier match, so in the
269 * global list, we would have entries from info/attributes the earliest
270 * (reading the file from top to bottom), .gitattribute of the root
271 * directory (again, reading the file from top to bottom) down to the
272 * current directory, and then scan the list backwards to find the first match.
273 * This is exactly the same as what excluded() does in dir.c to deal with
274 * .gitignore
277 static struct attr_stack {
278 struct attr_stack *prev;
279 char *origin;
280 unsigned num_matches;
281 unsigned alloc;
282 struct match_attr **attrs;
283 } *attr_stack;
285 static void free_attr_elem(struct attr_stack *e)
287 int i;
288 free(e->origin);
289 for (i = 0; i < e->num_matches; i++) {
290 struct match_attr *a = e->attrs[i];
291 int j;
292 for (j = 0; j < a->num_attr; j++) {
293 const char *setto = a->state[j].setto;
294 if (setto == ATTR__TRUE ||
295 setto == ATTR__FALSE ||
296 setto == ATTR__UNSET ||
297 setto == ATTR__UNKNOWN)
299 else
300 free((char *) setto);
302 free(a);
304 free(e->attrs);
305 free(e);
308 static const char *builtin_attr[] = {
309 "[attr]binary -diff -merge -text",
310 NULL,
313 static void handle_attr_line(struct attr_stack *res,
314 const char *line,
315 const char *src,
316 int lineno,
317 int macro_ok)
319 struct match_attr *a;
321 a = parse_attr_line(line, src, lineno, macro_ok);
322 if (!a)
323 return;
324 if (res->alloc <= res->num_matches) {
325 res->alloc = alloc_nr(res->num_matches);
326 res->attrs = xrealloc(res->attrs,
327 sizeof(struct match_attr *) *
328 res->alloc);
330 res->attrs[res->num_matches++] = a;
333 static struct attr_stack *read_attr_from_array(const char **list)
335 struct attr_stack *res;
336 const char *line;
337 int lineno = 0;
339 res = xcalloc(1, sizeof(*res));
340 while ((line = *(list++)) != NULL)
341 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
342 return res;
345 static enum git_attr_direction direction;
346 static struct index_state *use_index;
348 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
350 FILE *fp = fopen(path, "r");
351 struct attr_stack *res;
352 char buf[2048];
353 int lineno = 0;
355 if (!fp) {
356 if (errno != ENOENT && errno != ENOTDIR)
357 warn_on_inaccessible(path);
358 return NULL;
360 res = xcalloc(1, sizeof(*res));
361 while (fgets(buf, sizeof(buf), fp))
362 handle_attr_line(res, buf, path, ++lineno, macro_ok);
363 fclose(fp);
364 return res;
367 static void *read_index_data(const char *path)
369 int pos, len;
370 unsigned long sz;
371 enum object_type type;
372 void *data;
373 struct index_state *istate = use_index ? use_index : &the_index;
375 len = strlen(path);
376 pos = index_name_pos(istate, path, len);
377 if (pos < 0) {
379 * We might be in the middle of a merge, in which
380 * case we would read stage #2 (ours).
382 int i;
383 for (i = -pos - 1;
384 (pos < 0 && i < istate->cache_nr &&
385 !strcmp(istate->cache[i]->name, path));
386 i++)
387 if (ce_stage(istate->cache[i]) == 2)
388 pos = i;
390 if (pos < 0)
391 return NULL;
392 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
393 if (!data || type != OBJ_BLOB) {
394 free(data);
395 return NULL;
397 return data;
400 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
402 struct attr_stack *res;
403 char *buf, *sp;
404 int lineno = 0;
406 buf = read_index_data(path);
407 if (!buf)
408 return NULL;
410 res = xcalloc(1, sizeof(*res));
411 for (sp = buf; *sp; ) {
412 char *ep;
413 int more;
414 for (ep = sp; *ep && *ep != '\n'; ep++)
416 more = (*ep == '\n');
417 *ep = '\0';
418 handle_attr_line(res, sp, path, ++lineno, macro_ok);
419 sp = ep + more;
421 free(buf);
422 return res;
425 static struct attr_stack *read_attr(const char *path, int macro_ok)
427 struct attr_stack *res;
429 if (direction == GIT_ATTR_CHECKOUT) {
430 res = read_attr_from_index(path, macro_ok);
431 if (!res)
432 res = read_attr_from_file(path, macro_ok);
434 else if (direction == GIT_ATTR_CHECKIN) {
435 res = read_attr_from_file(path, macro_ok);
436 if (!res)
438 * There is no checked out .gitattributes file there, but
439 * we might have it in the index. We allow operation in a
440 * sparsely checked out work tree, so read from it.
442 res = read_attr_from_index(path, macro_ok);
444 else
445 res = read_attr_from_index(path, macro_ok);
446 if (!res)
447 res = xcalloc(1, sizeof(*res));
448 return res;
451 #if DEBUG_ATTR
452 static void debug_info(const char *what, struct attr_stack *elem)
454 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
456 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
458 const char *value = v;
460 if (ATTR_TRUE(value))
461 value = "set";
462 else if (ATTR_FALSE(value))
463 value = "unset";
464 else if (ATTR_UNSET(value))
465 value = "unspecified";
467 fprintf(stderr, "%s: %s => %s (%s)\n",
468 what, attr->name, (char *) value, match);
470 #define debug_push(a) debug_info("push", (a))
471 #define debug_pop(a) debug_info("pop", (a))
472 #else
473 #define debug_push(a) do { ; } while (0)
474 #define debug_pop(a) do { ; } while (0)
475 #define debug_set(a,b,c,d) do { ; } while (0)
476 #endif
478 static void drop_attr_stack(void)
480 while (attr_stack) {
481 struct attr_stack *elem = attr_stack;
482 attr_stack = elem->prev;
483 free_attr_elem(elem);
487 static const char *git_etc_gitattributes(void)
489 static const char *system_wide;
490 if (!system_wide)
491 system_wide = system_path(ETC_GITATTRIBUTES);
492 return system_wide;
495 static int git_attr_system(void)
497 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
500 static void bootstrap_attr_stack(void)
502 struct attr_stack *elem;
503 char *xdg_attributes_file;
505 if (attr_stack)
506 return;
508 elem = read_attr_from_array(builtin_attr);
509 elem->origin = NULL;
510 elem->prev = attr_stack;
511 attr_stack = elem;
513 if (git_attr_system()) {
514 elem = read_attr_from_file(git_etc_gitattributes(), 1);
515 if (elem) {
516 elem->origin = NULL;
517 elem->prev = attr_stack;
518 attr_stack = elem;
522 if (!git_attributes_file) {
523 home_config_paths(NULL, &xdg_attributes_file, "attributes");
524 git_attributes_file = xdg_attributes_file;
526 if (git_attributes_file) {
527 elem = read_attr_from_file(git_attributes_file, 1);
528 if (elem) {
529 elem->origin = NULL;
530 elem->prev = attr_stack;
531 attr_stack = elem;
535 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
536 elem = read_attr(GITATTRIBUTES_FILE, 1);
537 elem->origin = xstrdup("");
538 elem->prev = attr_stack;
539 attr_stack = elem;
540 debug_push(elem);
543 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
544 if (!elem)
545 elem = xcalloc(1, sizeof(*elem));
546 elem->origin = NULL;
547 elem->prev = attr_stack;
548 attr_stack = elem;
551 static void prepare_attr_stack(const char *path)
553 struct attr_stack *elem, *info;
554 int dirlen, len;
555 const char *cp;
557 cp = strrchr(path, '/');
558 if (!cp)
559 dirlen = 0;
560 else
561 dirlen = cp - path;
564 * At the bottom of the attribute stack is the built-in
565 * set of attribute definitions, followed by the contents
566 * of $(prefix)/etc/gitattributes and a file specified by
567 * core.attributesfile. Then, contents from
568 * .gitattribute files from directories closer to the
569 * root to the ones in deeper directories are pushed
570 * to the stack. Finally, at the very top of the stack
571 * we always keep the contents of $GIT_DIR/info/attributes.
573 * When checking, we use entries from near the top of the
574 * stack, preferring $GIT_DIR/info/attributes, then
575 * .gitattributes in deeper directories to shallower ones,
576 * and finally use the built-in set as the default.
578 bootstrap_attr_stack();
581 * Pop the "info" one that is always at the top of the stack.
583 info = attr_stack;
584 attr_stack = info->prev;
587 * Pop the ones from directories that are not the prefix of
588 * the path we are checking. Break out of the loop when we see
589 * the root one (whose origin is an empty string "") or the builtin
590 * one (whose origin is NULL) without popping it.
592 while (attr_stack->origin) {
593 int namelen = strlen(attr_stack->origin);
595 elem = attr_stack;
596 if (namelen <= dirlen &&
597 !strncmp(elem->origin, path, namelen) &&
598 (!namelen || path[namelen] == '/'))
599 break;
601 debug_pop(elem);
602 attr_stack = elem->prev;
603 free_attr_elem(elem);
607 * Read from parent directories and push them down
609 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
611 * bootstrap_attr_stack() should have added, and the
612 * above loop should have stopped before popping, the
613 * root element whose attr_stack->origin is set to an
614 * empty string.
616 struct strbuf pathbuf = STRBUF_INIT;
618 assert(attr_stack->origin);
619 while (1) {
620 len = strlen(attr_stack->origin);
621 if (dirlen <= len)
622 break;
623 cp = memchr(path + len + 1, '/', dirlen - len - 1);
624 if (!cp)
625 cp = path + dirlen;
626 strbuf_add(&pathbuf, path, cp - path);
627 strbuf_addch(&pathbuf, '/');
628 strbuf_addstr(&pathbuf, GITATTRIBUTES_FILE);
629 elem = read_attr(pathbuf.buf, 0);
630 strbuf_setlen(&pathbuf, cp - path);
631 elem->origin = strbuf_detach(&pathbuf, NULL);
632 elem->prev = attr_stack;
633 attr_stack = elem;
634 debug_push(elem);
637 strbuf_release(&pathbuf);
641 * Finally push the "info" one at the top of the stack.
643 info->prev = attr_stack;
644 attr_stack = info;
647 static int path_matches(const char *pathname, int pathlen,
648 const char *pattern,
649 const char *base, int baselen)
651 if (!strchr(pattern, '/')) {
652 /* match basename */
653 const char *basename = strrchr(pathname, '/');
654 basename = basename ? basename + 1 : pathname;
655 return (fnmatch_icase(pattern, basename, 0) == 0);
658 * match with FNM_PATHNAME; the pattern has base implicitly
659 * in front of it.
661 if (*pattern == '/')
662 pattern++;
663 if (pathlen < baselen ||
664 (baselen && pathname[baselen] != '/') ||
665 strncmp(pathname, base, baselen))
666 return 0;
667 if (baselen != 0)
668 baselen++;
669 return fnmatch_icase(pattern, pathname + baselen, FNM_PATHNAME) == 0;
672 static int macroexpand_one(int attr_nr, int rem);
674 static int fill_one(const char *what, struct match_attr *a, int rem)
676 struct git_attr_check *check = check_all_attr;
677 int i;
679 for (i = a->num_attr - 1; 0 < rem && 0 <= i; i--) {
680 struct git_attr *attr = a->state[i].attr;
681 const char **n = &(check[attr->attr_nr].value);
682 const char *v = a->state[i].setto;
684 if (*n == ATTR__UNKNOWN) {
685 debug_set(what,
686 a->is_macro ? a->u.attr->name : a->u.pattern,
687 attr, v);
688 *n = v;
689 rem--;
690 rem = macroexpand_one(attr->attr_nr, rem);
693 return rem;
696 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
698 int i;
699 const char *base = stk->origin ? stk->origin : "";
701 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
702 struct match_attr *a = stk->attrs[i];
703 if (a->is_macro)
704 continue;
705 if (path_matches(path, pathlen,
706 a->u.pattern, base, strlen(base)))
707 rem = fill_one("fill", a, rem);
709 return rem;
712 static int macroexpand_one(int attr_nr, int rem)
714 struct attr_stack *stk;
715 struct match_attr *a = NULL;
716 int i;
718 if (check_all_attr[attr_nr].value != ATTR__TRUE)
719 return rem;
721 for (stk = attr_stack; !a && stk; stk = stk->prev)
722 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
723 struct match_attr *ma = stk->attrs[i];
724 if (!ma->is_macro)
725 continue;
726 if (ma->u.attr->attr_nr == attr_nr)
727 a = ma;
730 if (a)
731 rem = fill_one("expand", a, rem);
733 return rem;
737 * Collect all attributes for path into the array pointed to by
738 * check_all_attr.
740 static void collect_all_attrs(const char *path)
742 struct attr_stack *stk;
743 int i, pathlen, rem;
745 prepare_attr_stack(path);
746 for (i = 0; i < attr_nr; i++)
747 check_all_attr[i].value = ATTR__UNKNOWN;
749 pathlen = strlen(path);
750 rem = attr_nr;
751 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
752 rem = fill(path, pathlen, stk, rem);
755 int git_check_attr(const char *path, int num, struct git_attr_check *check)
757 int i;
759 collect_all_attrs(path);
761 for (i = 0; i < num; i++) {
762 const char *value = check_all_attr[check[i].attr->attr_nr].value;
763 if (value == ATTR__UNKNOWN)
764 value = ATTR__UNSET;
765 check[i].value = value;
768 return 0;
771 int git_all_attrs(const char *path, int *num, struct git_attr_check **check)
773 int i, count, j;
775 collect_all_attrs(path);
777 /* Count the number of attributes that are set. */
778 count = 0;
779 for (i = 0; i < attr_nr; i++) {
780 const char *value = check_all_attr[i].value;
781 if (value != ATTR__UNSET && value != ATTR__UNKNOWN)
782 ++count;
784 *num = count;
785 *check = xmalloc(sizeof(**check) * count);
786 j = 0;
787 for (i = 0; i < attr_nr; i++) {
788 const char *value = check_all_attr[i].value;
789 if (value != ATTR__UNSET && value != ATTR__UNKNOWN) {
790 (*check)[j].attr = check_all_attr[i].attr;
791 (*check)[j].value = value;
792 ++j;
796 return 0;
799 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
801 enum git_attr_direction old = direction;
803 if (is_bare_repository() && new != GIT_ATTR_INDEX)
804 die("BUG: non-INDEX attr direction in a bare repo");
806 direction = new;
807 if (new != old)
808 drop_attr_stack();
809 use_index = istate;