submodule: Use cat instead of echo to avoid DOS line-endings
[git/dscho.git] / attr.c
blob33cb4e4d113cbb3816ba824cb06bf494a4bd9bc3
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"
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. */
26 #define HASHSIZE 257
28 #ifndef DEBUG_ATTR
29 #define DEBUG_ATTR 0
30 #endif
32 struct git_attr {
33 struct git_attr *next;
34 unsigned h;
35 int attr_nr;
36 char name[FLEX_ARRAY];
38 static int attr_nr;
40 static struct git_attr_check *check_all_attr;
41 static struct git_attr *(git_attr_hash[HASHSIZE]);
43 char *git_attr_name(struct git_attr *attr)
45 return attr->name;
48 static unsigned hash_name(const char *name, int namelen)
50 unsigned val = 0, c;
52 while (namelen--) {
53 c = *name++;
54 val = ((val << 7) | (val >> 22)) ^ c;
56 return val;
59 static int invalid_attr_name(const char *name, int namelen)
62 * Attribute name cannot begin with '-' and must consist of
63 * characters from [-A-Za-z0-9_.].
65 if (namelen <= 0 || *name == '-')
66 return -1;
67 while (namelen--) {
68 char ch = *name++;
69 if (! (ch == '-' || ch == '.' || ch == '_' ||
70 ('0' <= ch && ch <= '9') ||
71 ('a' <= ch && ch <= 'z') ||
72 ('A' <= ch && ch <= 'Z')) )
73 return -1;
75 return 0;
78 static struct git_attr *git_attr_internal(const char *name, int len)
80 unsigned hval = hash_name(name, len);
81 unsigned pos = hval % HASHSIZE;
82 struct git_attr *a;
84 for (a = git_attr_hash[pos]; a; a = a->next) {
85 if (a->h == hval &&
86 !memcmp(a->name, name, len) && !a->name[len])
87 return a;
90 if (invalid_attr_name(name, len))
91 return NULL;
93 a = xmalloc(sizeof(*a) + len + 1);
94 memcpy(a->name, name, len);
95 a->name[len] = 0;
96 a->h = hval;
97 a->next = git_attr_hash[pos];
98 a->attr_nr = attr_nr++;
99 git_attr_hash[pos] = a;
101 check_all_attr = xrealloc(check_all_attr,
102 sizeof(*check_all_attr) * attr_nr);
103 check_all_attr[a->attr_nr].attr = a;
104 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
105 return a;
108 struct git_attr *git_attr(const char *name)
110 return git_attr_internal(name, strlen(name));
113 /* What does a matched pattern decide? */
114 struct attr_state {
115 struct git_attr *attr;
116 const char *setto;
120 * One rule, as from a .gitattributes file.
122 * If is_macro is true, then u.attr is a pointer to the git_attr being
123 * defined.
125 * If is_macro is false, then u.pattern points at the filename pattern
126 * to which the rule applies. (The memory pointed to is part of the
127 * memory block allocated for the match_attr instance.)
129 * In either case, num_attr is the number of attributes affected by
130 * this rule, and state is an array listing them. The attributes are
131 * listed as they appear in the file (macros unexpanded).
133 struct match_attr {
134 union {
135 char *pattern;
136 struct git_attr *attr;
137 } u;
138 char is_macro;
139 unsigned num_attr;
140 struct attr_state state[FLEX_ARRAY];
143 static const char blank[] = " \t\r\n";
146 * Parse a whitespace-delimited attribute state (i.e., "attr",
147 * "-attr", "!attr", or "attr=value") from the string starting at src.
148 * If e is not NULL, write the results to *e. Return a pointer to the
149 * remainder of the string (with leading whitespace removed), or NULL
150 * if there was an error.
152 static const char *parse_attr(const char *src, int lineno, const char *cp,
153 struct attr_state *e)
155 const char *ep, *equals;
156 int len;
158 ep = cp + strcspn(cp, blank);
159 equals = strchr(cp, '=');
160 if (equals && ep < equals)
161 equals = NULL;
162 if (equals)
163 len = equals - cp;
164 else
165 len = ep - cp;
166 if (!e) {
167 if (*cp == '-' || *cp == '!') {
168 cp++;
169 len--;
171 if (invalid_attr_name(cp, len)) {
172 fprintf(stderr,
173 "%.*s is not a valid attribute name: %s:%d\n",
174 len, cp, src, lineno);
175 return NULL;
177 } else {
178 if (*cp == '-' || *cp == '!') {
179 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
180 cp++;
181 len--;
183 else if (!equals)
184 e->setto = ATTR__TRUE;
185 else {
186 e->setto = xmemdupz(equals + 1, ep - equals - 1);
188 e->attr = git_attr_internal(cp, len);
190 return ep + strspn(ep, blank);
193 static struct match_attr *parse_attr_line(const char *line, const char *src,
194 int lineno, int macro_ok)
196 int namelen;
197 int num_attr, i;
198 const char *cp, *name, *states;
199 struct match_attr *res = NULL;
200 int is_macro;
202 cp = line + strspn(line, blank);
203 if (!*cp || *cp == '#')
204 return NULL;
205 name = cp;
206 namelen = strcspn(name, blank);
207 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
208 !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
209 if (!macro_ok) {
210 fprintf(stderr, "%s not allowed: %s:%d\n",
211 name, src, lineno);
212 return NULL;
214 is_macro = 1;
215 name += strlen(ATTRIBUTE_MACRO_PREFIX);
216 name += strspn(name, blank);
217 namelen = strcspn(name, blank);
218 if (invalid_attr_name(name, namelen)) {
219 fprintf(stderr,
220 "%.*s is not a valid attribute name: %s:%d\n",
221 namelen, name, src, lineno);
222 return NULL;
225 else
226 is_macro = 0;
228 states = name + namelen;
229 states += strspn(states, blank);
231 /* First pass to count the attr_states */
232 for (cp = states, num_attr = 0; *cp; num_attr++) {
233 cp = parse_attr(src, lineno, cp, NULL);
234 if (!cp)
235 return NULL;
238 res = xcalloc(1,
239 sizeof(*res) +
240 sizeof(struct attr_state) * num_attr +
241 (is_macro ? 0 : namelen + 1));
242 if (is_macro)
243 res->u.attr = git_attr_internal(name, namelen);
244 else {
245 res->u.pattern = (char *)&(res->state[num_attr]);
246 memcpy(res->u.pattern, name, namelen);
247 res->u.pattern[namelen] = 0;
249 res->is_macro = is_macro;
250 res->num_attr = num_attr;
252 /* Second pass to fill the attr_states */
253 for (cp = states, i = 0; *cp; i++) {
254 cp = parse_attr(src, lineno, cp, &(res->state[i]));
257 return res;
261 * Like info/exclude and .gitignore, the attribute information can
262 * come from many places.
264 * (1) .gitattribute file of the same directory;
265 * (2) .gitattribute file of the parent directory if (1) does not have
266 * any match; this goes recursively upwards, just like .gitignore.
267 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
269 * In the same file, later entries override the earlier match, so in the
270 * global list, we would have entries from info/attributes the earliest
271 * (reading the file from top to bottom), .gitattribute of the root
272 * directory (again, reading the file from top to bottom) down to the
273 * current directory, and then scan the list backwards to find the first match.
274 * This is exactly the same as what excluded() does in dir.c to deal with
275 * .gitignore
278 static struct attr_stack {
279 struct attr_stack *prev;
280 char *origin;
281 unsigned num_matches;
282 unsigned alloc;
283 struct match_attr **attrs;
284 } *attr_stack;
286 static void free_attr_elem(struct attr_stack *e)
288 int i;
289 free(e->origin);
290 for (i = 0; i < e->num_matches; i++) {
291 struct match_attr *a = e->attrs[i];
292 int j;
293 for (j = 0; j < a->num_attr; j++) {
294 const char *setto = a->state[j].setto;
295 if (setto == ATTR__TRUE ||
296 setto == ATTR__FALSE ||
297 setto == ATTR__UNSET ||
298 setto == ATTR__UNKNOWN)
300 else
301 free((char *) setto);
303 free(a);
305 free(e);
308 static const char *builtin_attr[] = {
309 "[attr]binary -diff -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 return NULL;
357 res = xcalloc(1, sizeof(*res));
358 while (fgets(buf, sizeof(buf), fp))
359 handle_attr_line(res, buf, path, ++lineno, macro_ok);
360 fclose(fp);
361 return res;
364 static void *read_index_data(const char *path)
366 int pos, len;
367 unsigned long sz;
368 enum object_type type;
369 void *data;
370 struct index_state *istate = use_index ? use_index : &the_index;
372 len = strlen(path);
373 pos = index_name_pos(istate, path, len);
374 if (pos < 0) {
376 * We might be in the middle of a merge, in which
377 * case we would read stage #2 (ours).
379 int i;
380 for (i = -pos - 1;
381 (pos < 0 && i < istate->cache_nr &&
382 !strcmp(istate->cache[i]->name, path));
383 i++)
384 if (ce_stage(istate->cache[i]) == 2)
385 pos = i;
387 if (pos < 0)
388 return NULL;
389 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
390 if (!data || type != OBJ_BLOB) {
391 free(data);
392 return NULL;
394 return data;
397 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
399 struct attr_stack *res;
400 char *buf, *sp;
401 int lineno = 0;
403 buf = read_index_data(path);
404 if (!buf)
405 return NULL;
407 res = xcalloc(1, sizeof(*res));
408 for (sp = buf; *sp; ) {
409 char *ep;
410 int more;
411 for (ep = sp; *ep && *ep != '\n'; ep++)
413 more = (*ep == '\n');
414 *ep = '\0';
415 handle_attr_line(res, sp, path, ++lineno, macro_ok);
416 sp = ep + more;
418 free(buf);
419 return res;
422 static struct attr_stack *read_attr(const char *path, int macro_ok)
424 struct attr_stack *res;
426 if (direction == GIT_ATTR_CHECKOUT) {
427 res = read_attr_from_index(path, macro_ok);
428 if (!res)
429 res = read_attr_from_file(path, macro_ok);
431 else if (direction == GIT_ATTR_CHECKIN) {
432 res = read_attr_from_file(path, macro_ok);
433 if (!res)
435 * There is no checked out .gitattributes file there, but
436 * we might have it in the index. We allow operation in a
437 * sparsely checked out work tree, so read from it.
439 res = read_attr_from_index(path, macro_ok);
441 else
442 res = read_attr_from_index(path, macro_ok);
443 if (!res)
444 res = xcalloc(1, sizeof(*res));
445 return res;
448 #if DEBUG_ATTR
449 static void debug_info(const char *what, struct attr_stack *elem)
451 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
453 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
455 const char *value = v;
457 if (ATTR_TRUE(value))
458 value = "set";
459 else if (ATTR_FALSE(value))
460 value = "unset";
461 else if (ATTR_UNSET(value))
462 value = "unspecified";
464 fprintf(stderr, "%s: %s => %s (%s)\n",
465 what, attr->name, (char *) value, match);
467 #define debug_push(a) debug_info("push", (a))
468 #define debug_pop(a) debug_info("pop", (a))
469 #else
470 #define debug_push(a) do { ; } while (0)
471 #define debug_pop(a) do { ; } while (0)
472 #define debug_set(a,b,c,d) do { ; } while (0)
473 #endif
475 static void drop_attr_stack(void)
477 while (attr_stack) {
478 struct attr_stack *elem = attr_stack;
479 attr_stack = elem->prev;
480 free_attr_elem(elem);
484 static const char *git_etc_gitattributes(void)
486 static const char *system_wide;
487 if (!system_wide)
488 system_wide = system_path(ETC_GITATTRIBUTES);
489 return system_wide;
492 static int git_attr_system(void)
494 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
497 static int git_attr_config(const char *var, const char *value, void *dummy)
499 if (!strcmp(var, "core.attributesfile"))
500 return git_config_pathname(&attributes_file, var, value);
502 return 0;
505 static void bootstrap_attr_stack(void)
507 if (!attr_stack) {
508 struct attr_stack *elem;
510 elem = read_attr_from_array(builtin_attr);
511 elem->origin = NULL;
512 elem->prev = attr_stack;
513 attr_stack = elem;
515 if (git_attr_system()) {
516 elem = read_attr_from_file(git_etc_gitattributes(), 1);
517 if (elem) {
518 elem->origin = NULL;
519 elem->prev = attr_stack;
520 attr_stack = elem;
524 git_config(git_attr_config, NULL);
525 if (attributes_file) {
526 elem = read_attr_from_file(attributes_file, 1);
527 if (elem) {
528 elem->origin = NULL;
529 elem->prev = attr_stack;
530 attr_stack = elem;
534 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
535 elem = read_attr(GITATTRIBUTES_FILE, 1);
536 elem->origin = strdup("");
537 elem->prev = attr_stack;
538 attr_stack = elem;
539 debug_push(elem);
542 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
543 if (!elem)
544 elem = xcalloc(1, sizeof(*elem));
545 elem->origin = NULL;
546 elem->prev = attr_stack;
547 attr_stack = elem;
551 static void prepare_attr_stack(const char *path)
553 struct attr_stack *elem, *info;
554 int dirlen, len;
555 struct strbuf pathbuf;
556 const char *cp;
558 cp = strrchr(path, '/');
559 if (!cp)
560 dirlen = 0;
561 else
562 dirlen = cp - path;
564 strbuf_init(&pathbuf, dirlen+2+strlen(GITATTRIBUTES_FILE));
567 * At the bottom of the attribute stack is the built-in
568 * set of attribute definitions, followed by the contents
569 * of $(prefix)/etc/gitattributes and a file specified by
570 * core.attributesfile. Then, contents from
571 * .gitattribute files from directories closer to the
572 * root to the ones in deeper directories are pushed
573 * to the stack. Finally, at the very top of the stack
574 * we always keep the contents of $GIT_DIR/info/attributes.
576 * When checking, we use entries from near the top of the
577 * stack, preferring $GIT_DIR/info/attributes, then
578 * .gitattributes in deeper directories to shallower ones,
579 * and finally use the built-in set as the default.
581 bootstrap_attr_stack();
584 * Pop the "info" one that is always at the top of the stack.
586 info = attr_stack;
587 attr_stack = info->prev;
590 * Pop the ones from directories that are not the prefix of
591 * the path we are checking.
593 while (attr_stack && attr_stack->origin) {
594 int namelen = strlen(attr_stack->origin);
596 elem = attr_stack;
597 if (namelen <= dirlen &&
598 !strncmp(elem->origin, 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) {
610 while (1) {
611 char *cp;
613 len = strlen(attr_stack->origin);
614 if (dirlen <= len)
615 break;
616 strbuf_reset(&pathbuf);
617 strbuf_add(&pathbuf, path, dirlen);
618 strbuf_addch(&pathbuf, '/');
619 cp = strchr(pathbuf.buf + len + 1, '/');
620 strcpy(cp + 1, GITATTRIBUTES_FILE);
621 elem = read_attr(pathbuf.buf, 0);
622 *cp = '\0';
623 elem->origin = strdup(pathbuf.buf);
624 elem->prev = attr_stack;
625 attr_stack = elem;
626 debug_push(elem);
630 strbuf_release(&pathbuf);
633 * Finally push the "info" one at the top of the stack.
635 info->prev = attr_stack;
636 attr_stack = info;
639 static int path_matches(const char *pathname, int pathlen,
640 const char *pattern,
641 const char *base, int baselen)
643 if (!strchr(pattern, '/')) {
644 /* match basename */
645 const char *basename = strrchr(pathname, '/');
646 basename = basename ? basename + 1 : pathname;
647 return (fnmatch(pattern, basename, 0) == 0);
650 * match with FNM_PATHNAME; the pattern has base implicitly
651 * in front of it.
653 if (*pattern == '/')
654 pattern++;
655 if (pathlen < baselen ||
656 (baselen && pathname[baselen] != '/') ||
657 strncmp(pathname, base, baselen))
658 return 0;
659 if (baselen != 0)
660 baselen++;
661 return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
664 static int macroexpand_one(int attr_nr, int rem);
666 static int fill_one(const char *what, struct match_attr *a, int rem)
668 struct git_attr_check *check = check_all_attr;
669 int i;
671 for (i = a->num_attr - 1; 0 < rem && 0 <= i; i--) {
672 struct git_attr *attr = a->state[i].attr;
673 const char **n = &(check[attr->attr_nr].value);
674 const char *v = a->state[i].setto;
676 if (*n == ATTR__UNKNOWN) {
677 debug_set(what,
678 a->is_macro ? a->u.attr->name : a->u.pattern,
679 attr, v);
680 *n = v;
681 rem--;
682 rem = macroexpand_one(attr->attr_nr, rem);
685 return rem;
688 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
690 int i;
691 const char *base = stk->origin ? stk->origin : "";
693 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
694 struct match_attr *a = stk->attrs[i];
695 if (a->is_macro)
696 continue;
697 if (path_matches(path, pathlen,
698 a->u.pattern, base, strlen(base)))
699 rem = fill_one("fill", a, rem);
701 return rem;
704 static int macroexpand_one(int attr_nr, int rem)
706 struct attr_stack *stk;
707 struct match_attr *a = NULL;
708 int i;
710 if (check_all_attr[attr_nr].value != ATTR__TRUE)
711 return rem;
713 for (stk = attr_stack; !a && stk; stk = stk->prev)
714 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
715 struct match_attr *ma = stk->attrs[i];
716 if (!ma->is_macro)
717 continue;
718 if (ma->u.attr->attr_nr == attr_nr)
719 a = ma;
722 if (a)
723 rem = fill_one("expand", a, rem);
725 return rem;
729 * Collect all attributes for path into the array pointed to by
730 * check_all_attr.
732 static void collect_all_attrs(const char *path)
734 struct attr_stack *stk;
735 int i, pathlen, rem;
737 prepare_attr_stack(path);
738 for (i = 0; i < attr_nr; i++)
739 check_all_attr[i].value = ATTR__UNKNOWN;
741 pathlen = strlen(path);
742 rem = attr_nr;
743 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
744 rem = fill(path, pathlen, stk, rem);
747 int git_check_attr(const char *path, int num, struct git_attr_check *check)
749 int i;
751 collect_all_attrs(path);
753 for (i = 0; i < num; i++) {
754 const char *value = check_all_attr[check[i].attr->attr_nr].value;
755 if (value == ATTR__UNKNOWN)
756 value = ATTR__UNSET;
757 check[i].value = value;
760 return 0;
763 int git_all_attrs(const char *path, int *num, struct git_attr_check **check)
765 int i, count, j;
767 collect_all_attrs(path);
769 /* Count the number of attributes that are set. */
770 count = 0;
771 for (i = 0; i < attr_nr; i++) {
772 const char *value = check_all_attr[i].value;
773 if (value != ATTR__UNSET && value != ATTR__UNKNOWN)
774 ++count;
776 *num = count;
777 *check = xmalloc(sizeof(**check) * count);
778 j = 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 (*check)[j].attr = check_all_attr[i].attr;
783 (*check)[j].value = value;
784 ++j;
788 return 0;
791 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
793 enum git_attr_direction old = direction;
795 if (is_bare_repository() && new != GIT_ATTR_INDEX)
796 die("BUG: non-INDEX attr direction in a bare repo");
798 direction = new;
799 if (new != old)
800 drop_attr_stack();
801 use_index = istate;