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.
10 #define NO_THE_INDEX_COMPATIBILITY_MACROS
15 const char git_attr__true[] = "(builtin)true";
16 const char git_attr__false[] = "\0(builtin)false";
17 static const char git_attr__unknown[] = "(builtin)unknown";
18 #define ATTR__TRUE git_attr__true
19 #define ATTR__FALSE git_attr__false
20 #define ATTR__UNSET NULL
21 #define ATTR__UNKNOWN git_attr__unknown
23 static const char *attributes_file;
25 /* This is a randomly chosen prime. */
33 struct git_attr *next;
36 char name[FLEX_ARRAY];
40 static struct git_attr_check *check_all_attr;
41 static struct git_attr *(git_attr_hash[HASHSIZE]);
43 static unsigned hash_name(const char *name, int namelen)
49 val = ((val << 7) | (val >> 22)) ^ c;
54 static int invalid_attr_name(const char *name, int namelen)
57 * Attribute name cannot begin with '-' and from
58 * [-A-Za-z0-9_.]. We'd specifically exclude '=' for now,
59 * as we might later want to allow non-binary value for
60 * attributes, e.g. "*.svg merge=special-merge-program-for-svg"
66 if (! (ch == '-' || ch == '.' || ch == '_' ||
67 ('0' <= ch && ch <= '9') ||
68 ('a' <= ch && ch <= 'z') ||
69 ('A' <= ch && ch <= 'Z')) )
75 static struct git_attr *git_attr_internal(const char *name, int len)
77 unsigned hval = hash_name(name, len);
78 unsigned pos = hval % HASHSIZE;
81 for (a = git_attr_hash[pos]; a; a = a->next) {
83 !memcmp(a->name, name, len) && !a->name[len])
87 if (invalid_attr_name(name, len))
90 a = xmalloc(sizeof(*a) + len + 1);
91 memcpy(a->name, name, len);
94 a->next = git_attr_hash[pos];
95 a->attr_nr = attr_nr++;
96 git_attr_hash[pos] = a;
98 check_all_attr = xrealloc(check_all_attr,
99 sizeof(*check_all_attr) * attr_nr);
100 check_all_attr[a->attr_nr].attr = a;
101 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
105 struct git_attr *git_attr(const char *name)
107 return git_attr_internal(name, strlen(name));
110 /* What does a matched pattern decide? */
112 struct git_attr *attr;
117 * One rule, as from a .gitattributes file.
119 * If is_macro is true, then u.attr is a pointer to the git_attr being
122 * If is_macro is false, then u.pattern points at the filename pattern
123 * to which the rule applies. (The memory pointed to is part of the
124 * memory block allocated for the match_attr instance.)
126 * In either case, num_attr is the number of attributes affected by
127 * this rule, and state is an array listing them. The attributes are
128 * listed as they appear in the file (macros unexpanded).
133 struct git_attr *attr;
137 struct attr_state state[FLEX_ARRAY];
140 static const char blank[] = " \t\r\n";
142 static const char *parse_attr(const char *src, int lineno, const char *cp,
143 int *num_attr, struct match_attr *res)
145 const char *ep, *equals;
148 ep = cp + strcspn(cp, blank);
149 equals = strchr(cp, '=');
150 if (equals && ep < equals)
157 if (*cp == '-' || *cp == '!') {
161 if (invalid_attr_name(cp, len)) {
163 "%.*s is not a valid attribute name: %s:%d\n",
164 len, cp, src, lineno);
168 struct attr_state *e;
170 e = &(res->state[*num_attr]);
171 if (*cp == '-' || *cp == '!') {
172 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
177 e->setto = ATTR__TRUE;
179 e->setto = xmemdupz(equals + 1, ep - equals - 1);
181 e->attr = git_attr_internal(cp, len);
184 return ep + strspn(ep, blank);
187 static struct match_attr *parse_attr_line(const char *line, const char *src,
188 int lineno, int macro_ok)
192 const char *cp, *name;
193 struct match_attr *res = NULL;
197 cp = line + strspn(line, blank);
198 if (!*cp || *cp == '#')
201 namelen = strcspn(name, blank);
202 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
203 !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
205 fprintf(stderr, "%s not allowed: %s:%d\n",
210 name += strlen(ATTRIBUTE_MACRO_PREFIX);
211 name += strspn(name, blank);
212 namelen = strcspn(name, blank);
213 if (invalid_attr_name(name, namelen)) {
215 "%.*s is not a valid attribute name: %s:%d\n",
216 namelen, name, src, lineno);
223 for (pass = 0; pass < 2; pass++) {
224 /* pass 0 counts and allocates, pass 1 fills */
227 cp = cp + strspn(cp, blank);
229 cp = parse_attr(src, lineno, cp, &num_attr, res);
237 sizeof(struct attr_state) * num_attr +
238 (is_macro ? 0 : namelen + 1));
240 res->u.attr = git_attr_internal(name, namelen);
242 res->u.pattern = (char *)&(res->state[num_attr]);
243 memcpy(res->u.pattern, name, namelen);
244 res->u.pattern[namelen] = 0;
246 res->is_macro = is_macro;
247 res->num_attr = num_attr;
253 * Like info/exclude and .gitignore, the attribute information can
254 * come from many places.
256 * (1) .gitattribute file of the same directory;
257 * (2) .gitattribute file of the parent directory if (1) does not have
258 * any match; this goes recursively upwards, just like .gitignore.
259 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
261 * In the same file, later entries override the earlier match, so in the
262 * global list, we would have entries from info/attributes the earliest
263 * (reading the file from top to bottom), .gitattribute of the root
264 * directory (again, reading the file from top to bottom) down to the
265 * current directory, and then scan the list backwards to find the first match.
266 * This is exactly the same as what excluded() does in dir.c to deal with
270 static struct attr_stack {
271 struct attr_stack *prev;
273 unsigned num_matches;
275 struct match_attr **attrs;
278 static void free_attr_elem(struct attr_stack *e)
282 for (i = 0; i < e->num_matches; i++) {
283 struct match_attr *a = e->attrs[i];
285 for (j = 0; j < a->num_attr; j++) {
286 const char *setto = a->state[j].setto;
287 if (setto == ATTR__TRUE ||
288 setto == ATTR__FALSE ||
289 setto == ATTR__UNSET ||
290 setto == ATTR__UNKNOWN)
293 free((char *) setto);
300 static const char *builtin_attr[] = {
301 "[attr]binary -diff -text",
305 static void handle_attr_line(struct attr_stack *res,
311 struct match_attr *a;
313 a = parse_attr_line(line, src, lineno, macro_ok);
316 if (res->alloc <= res->num_matches) {
317 res->alloc = alloc_nr(res->num_matches);
318 res->attrs = xrealloc(res->attrs,
319 sizeof(struct match_attr *) *
322 res->attrs[res->num_matches++] = a;
325 static struct attr_stack *read_attr_from_array(const char **list)
327 struct attr_stack *res;
331 res = xcalloc(1, sizeof(*res));
332 while ((line = *(list++)) != NULL)
333 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
337 static enum git_attr_direction direction;
338 static struct index_state *use_index;
340 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
342 FILE *fp = fopen(path, "r");
343 struct attr_stack *res;
349 res = xcalloc(1, sizeof(*res));
350 while (fgets(buf, sizeof(buf), fp))
351 handle_attr_line(res, buf, path, ++lineno, macro_ok);
356 static void *read_index_data(const char *path)
360 enum object_type type;
362 struct index_state *istate = use_index ? use_index : &the_index;
365 pos = index_name_pos(istate, path, len);
368 * We might be in the middle of a merge, in which
369 * case we would read stage #2 (ours).
373 (pos < 0 && i < istate->cache_nr &&
374 !strcmp(istate->cache[i]->name, path));
376 if (ce_stage(istate->cache[i]) == 2)
381 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
382 if (!data || type != OBJ_BLOB) {
389 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
391 struct attr_stack *res;
395 buf = read_index_data(path);
399 res = xcalloc(1, sizeof(*res));
400 for (sp = buf; *sp; ) {
403 for (ep = sp; *ep && *ep != '\n'; ep++)
405 more = (*ep == '\n');
407 handle_attr_line(res, sp, path, ++lineno, macro_ok);
414 static struct attr_stack *read_attr(const char *path, int macro_ok)
416 struct attr_stack *res;
418 if (direction == GIT_ATTR_CHECKOUT) {
419 res = read_attr_from_index(path, macro_ok);
421 res = read_attr_from_file(path, macro_ok);
423 else if (direction == GIT_ATTR_CHECKIN) {
424 res = read_attr_from_file(path, macro_ok);
427 * There is no checked out .gitattributes file there, but
428 * we might have it in the index. We allow operation in a
429 * sparsely checked out work tree, so read from it.
431 res = read_attr_from_index(path, macro_ok);
434 res = read_attr_from_index(path, macro_ok);
436 res = xcalloc(1, sizeof(*res));
441 static void debug_info(const char *what, struct attr_stack *elem)
443 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
445 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
447 const char *value = v;
449 if (ATTR_TRUE(value))
451 else if (ATTR_FALSE(value))
453 else if (ATTR_UNSET(value))
454 value = "unspecified";
456 fprintf(stderr, "%s: %s => %s (%s)\n",
457 what, attr->name, (char *) value, match);
459 #define debug_push(a) debug_info("push", (a))
460 #define debug_pop(a) debug_info("pop", (a))
462 #define debug_push(a) do { ; } while (0)
463 #define debug_pop(a) do { ; } while (0)
464 #define debug_set(a,b,c,d) do { ; } while (0)
467 static void drop_attr_stack(void)
470 struct attr_stack *elem = attr_stack;
471 attr_stack = elem->prev;
472 free_attr_elem(elem);
476 static const char *git_etc_gitattributes(void)
478 static const char *system_wide;
480 system_wide = system_path(ETC_GITATTRIBUTES);
484 static int git_attr_system(void)
486 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
489 static int git_attr_config(const char *var, const char *value, void *dummy)
491 if (!strcmp(var, "core.attributesfile"))
492 return git_config_pathname(&attributes_file, var, value);
497 static void bootstrap_attr_stack(void)
500 struct attr_stack *elem;
502 elem = read_attr_from_array(builtin_attr);
504 elem->prev = attr_stack;
507 if (git_attr_system()) {
508 elem = read_attr_from_file(git_etc_gitattributes(), 1);
511 elem->prev = attr_stack;
516 git_config(git_attr_config, NULL);
517 if (attributes_file) {
518 elem = read_attr_from_file(attributes_file, 1);
521 elem->prev = attr_stack;
526 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
527 elem = read_attr(GITATTRIBUTES_FILE, 1);
528 elem->origin = strdup("");
529 elem->prev = attr_stack;
534 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
536 elem = xcalloc(1, sizeof(*elem));
538 elem->prev = attr_stack;
543 static void prepare_attr_stack(const char *path, int dirlen)
545 struct attr_stack *elem, *info;
547 struct strbuf pathbuf;
549 strbuf_init(&pathbuf, dirlen+2+strlen(GITATTRIBUTES_FILE));
552 * At the bottom of the attribute stack is the built-in
553 * set of attribute definitions, followed by the contents
554 * of $(prefix)/etc/gitattributes and a file specified by
555 * core.attributesfile. Then, contents from
556 * .gitattribute files from directories closer to the
557 * root to the ones in deeper directories are pushed
558 * to the stack. Finally, at the very top of the stack
559 * we always keep the contents of $GIT_DIR/info/attributes.
561 * When checking, we use entries from near the top of the
562 * stack, preferring $GIT_DIR/info/attributes, then
563 * .gitattributes in deeper directories to shallower ones,
564 * and finally use the built-in set as the default.
567 bootstrap_attr_stack();
570 * Pop the "info" one that is always at the top of the stack.
573 attr_stack = info->prev;
576 * Pop the ones from directories that are not the prefix of
577 * the path we are checking.
579 while (attr_stack && attr_stack->origin) {
580 int namelen = strlen(attr_stack->origin);
583 if (namelen <= dirlen &&
584 !strncmp(elem->origin, path, namelen))
588 attr_stack = elem->prev;
589 free_attr_elem(elem);
593 * Read from parent directories and push them down
595 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
599 len = strlen(attr_stack->origin);
602 strbuf_reset(&pathbuf);
603 strbuf_add(&pathbuf, path, dirlen);
604 strbuf_addch(&pathbuf, '/');
605 cp = strchr(pathbuf.buf + len + 1, '/');
606 strcpy(cp + 1, GITATTRIBUTES_FILE);
607 elem = read_attr(pathbuf.buf, 0);
609 elem->origin = strdup(pathbuf.buf);
610 elem->prev = attr_stack;
616 strbuf_release(&pathbuf);
619 * Finally push the "info" one at the top of the stack.
621 info->prev = attr_stack;
625 static int path_matches(const char *pathname, int pathlen,
627 const char *base, int baselen)
629 if (!strchr(pattern, '/')) {
631 const char *basename = strrchr(pathname, '/');
632 basename = basename ? basename + 1 : pathname;
633 return (fnmatch(pattern, basename, 0) == 0);
636 * match with FNM_PATHNAME; the pattern has base implicitly
641 if (pathlen < baselen ||
642 (baselen && pathname[baselen] != '/') ||
643 strncmp(pathname, base, baselen))
647 return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
650 static int macroexpand_one(int attr_nr, int rem);
652 static int fill_one(const char *what, struct match_attr *a, int rem)
654 struct git_attr_check *check = check_all_attr;
657 for (i = a->num_attr - 1; 0 < rem && 0 <= i; i--) {
658 struct git_attr *attr = a->state[i].attr;
659 const char **n = &(check[attr->attr_nr].value);
660 const char *v = a->state[i].setto;
662 if (*n == ATTR__UNKNOWN) {
664 a->is_macro ? a->u.attr->name : a->u.pattern,
668 rem = macroexpand_one(attr->attr_nr, rem);
674 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
677 const char *base = stk->origin ? stk->origin : "";
679 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
680 struct match_attr *a = stk->attrs[i];
683 if (path_matches(path, pathlen,
684 a->u.pattern, base, strlen(base)))
685 rem = fill_one("fill", a, rem);
690 static int macroexpand_one(int attr_nr, int rem)
692 struct attr_stack *stk;
693 struct match_attr *a = NULL;
696 if (check_all_attr[attr_nr].value != ATTR__TRUE)
699 for (stk = attr_stack; !a && stk; stk = stk->prev)
700 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
701 struct match_attr *ma = stk->attrs[i];
704 if (ma->u.attr->attr_nr == attr_nr)
709 rem = fill_one("expand", a, rem);
714 int git_checkattr(const char *path, int num, struct git_attr_check *check)
716 struct attr_stack *stk;
718 int dirlen, pathlen, i, rem;
720 bootstrap_attr_stack();
721 for (i = 0; i < attr_nr; i++)
722 check_all_attr[i].value = ATTR__UNKNOWN;
724 pathlen = strlen(path);
725 cp = strrchr(path, '/');
730 prepare_attr_stack(path, dirlen);
732 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
733 rem = fill(path, pathlen, stk, rem);
735 for (i = 0; i < num; i++) {
736 const char *value = check_all_attr[check[i].attr->attr_nr].value;
737 if (value == ATTR__UNKNOWN)
739 check[i].value = value;
745 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
747 enum git_attr_direction old = direction;
749 if (is_bare_repository() && new != GIT_ATTR_INDEX)
750 die("BUG: non-INDEX attr direction in a bare repo");