gitignore: make pattern parsing code a separate function
[git/mjg.git] / attr.c
blobe7caee4ddd94f4c82f136b89d3f66afe328e7497
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 size_t originlen;
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->attrs);
306 free(e);
309 static const char *builtin_attr[] = {
310 "[attr]binary -diff -text",
311 NULL,
314 static void handle_attr_line(struct attr_stack *res,
315 const char *line,
316 const char *src,
317 int lineno,
318 int macro_ok)
320 struct match_attr *a;
322 a = parse_attr_line(line, src, lineno, macro_ok);
323 if (!a)
324 return;
325 if (res->alloc <= res->num_matches) {
326 res->alloc = alloc_nr(res->num_matches);
327 res->attrs = xrealloc(res->attrs,
328 sizeof(struct match_attr *) *
329 res->alloc);
331 res->attrs[res->num_matches++] = a;
334 static struct attr_stack *read_attr_from_array(const char **list)
336 struct attr_stack *res;
337 const char *line;
338 int lineno = 0;
340 res = xcalloc(1, sizeof(*res));
341 while ((line = *(list++)) != NULL)
342 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
343 return res;
346 static enum git_attr_direction direction;
347 static struct index_state *use_index;
349 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
351 FILE *fp = fopen(path, "r");
352 struct attr_stack *res;
353 char buf[2048];
354 int lineno = 0;
356 if (!fp)
357 return NULL;
358 res = xcalloc(1, sizeof(*res));
359 while (fgets(buf, sizeof(buf), fp))
360 handle_attr_line(res, buf, path, ++lineno, macro_ok);
361 fclose(fp);
362 return res;
365 static void *read_index_data(const char *path)
367 int pos, len;
368 unsigned long sz;
369 enum object_type type;
370 void *data;
371 struct index_state *istate = use_index ? use_index : &the_index;
373 len = strlen(path);
374 pos = index_name_pos(istate, path, len);
375 if (pos < 0) {
377 * We might be in the middle of a merge, in which
378 * case we would read stage #2 (ours).
380 int i;
381 for (i = -pos - 1;
382 (pos < 0 && i < istate->cache_nr &&
383 !strcmp(istate->cache[i]->name, path));
384 i++)
385 if (ce_stage(istate->cache[i]) == 2)
386 pos = i;
388 if (pos < 0)
389 return NULL;
390 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
391 if (!data || type != OBJ_BLOB) {
392 free(data);
393 return NULL;
395 return data;
398 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
400 struct attr_stack *res;
401 char *buf, *sp;
402 int lineno = 0;
404 buf = read_index_data(path);
405 if (!buf)
406 return NULL;
408 res = xcalloc(1, sizeof(*res));
409 for (sp = buf; *sp; ) {
410 char *ep;
411 int more;
412 for (ep = sp; *ep && *ep != '\n'; ep++)
414 more = (*ep == '\n');
415 *ep = '\0';
416 handle_attr_line(res, sp, path, ++lineno, macro_ok);
417 sp = ep + more;
419 free(buf);
420 return res;
423 static struct attr_stack *read_attr(const char *path, int macro_ok)
425 struct attr_stack *res;
427 if (direction == GIT_ATTR_CHECKOUT) {
428 res = read_attr_from_index(path, macro_ok);
429 if (!res)
430 res = read_attr_from_file(path, macro_ok);
432 else if (direction == GIT_ATTR_CHECKIN) {
433 res = read_attr_from_file(path, macro_ok);
434 if (!res)
436 * There is no checked out .gitattributes file there, but
437 * we might have it in the index. We allow operation in a
438 * sparsely checked out work tree, so read from it.
440 res = read_attr_from_index(path, macro_ok);
442 else
443 res = read_attr_from_index(path, macro_ok);
444 if (!res)
445 res = xcalloc(1, sizeof(*res));
446 return res;
449 #if DEBUG_ATTR
450 static void debug_info(const char *what, struct attr_stack *elem)
452 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
454 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
456 const char *value = v;
458 if (ATTR_TRUE(value))
459 value = "set";
460 else if (ATTR_FALSE(value))
461 value = "unset";
462 else if (ATTR_UNSET(value))
463 value = "unspecified";
465 fprintf(stderr, "%s: %s => %s (%s)\n",
466 what, attr->name, (char *) value, match);
468 #define debug_push(a) debug_info("push", (a))
469 #define debug_pop(a) debug_info("pop", (a))
470 #else
471 #define debug_push(a) do { ; } while (0)
472 #define debug_pop(a) do { ; } while (0)
473 #define debug_set(a,b,c,d) do { ; } while (0)
474 #endif
476 static void drop_attr_stack(void)
478 while (attr_stack) {
479 struct attr_stack *elem = attr_stack;
480 attr_stack = elem->prev;
481 free_attr_elem(elem);
485 static const char *git_etc_gitattributes(void)
487 static const char *system_wide;
488 if (!system_wide)
489 system_wide = system_path(ETC_GITATTRIBUTES);
490 return system_wide;
493 static int git_attr_system(void)
495 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
498 static void bootstrap_attr_stack(void)
500 struct attr_stack *elem;
502 if (attr_stack)
503 return;
505 elem = read_attr_from_array(builtin_attr);
506 elem->origin = NULL;
507 elem->prev = attr_stack;
508 attr_stack = elem;
510 if (git_attr_system()) {
511 elem = read_attr_from_file(git_etc_gitattributes(), 1);
512 if (elem) {
513 elem->origin = NULL;
514 elem->prev = attr_stack;
515 attr_stack = elem;
519 if (git_attributes_file) {
520 elem = read_attr_from_file(git_attributes_file, 1);
521 if (elem) {
522 elem->origin = NULL;
523 elem->prev = attr_stack;
524 attr_stack = elem;
528 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
529 elem = read_attr(GITATTRIBUTES_FILE, 1);
530 elem->origin = xstrdup("");
531 elem->originlen = 0;
532 elem->prev = attr_stack;
533 attr_stack = elem;
534 debug_push(elem);
537 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
538 if (!elem)
539 elem = xcalloc(1, sizeof(*elem));
540 elem->origin = NULL;
541 elem->prev = attr_stack;
542 attr_stack = elem;
545 static void prepare_attr_stack(const char *path)
547 struct attr_stack *elem, *info;
548 int dirlen, len;
549 const char *cp;
551 cp = strrchr(path, '/');
552 if (!cp)
553 dirlen = 0;
554 else
555 dirlen = cp - path;
558 * At the bottom of the attribute stack is the built-in
559 * set of attribute definitions, followed by the contents
560 * of $(prefix)/etc/gitattributes and a file specified by
561 * core.attributesfile. Then, contents from
562 * .gitattribute files from directories closer to the
563 * root to the ones in deeper directories are pushed
564 * to the stack. Finally, at the very top of the stack
565 * we always keep the contents of $GIT_DIR/info/attributes.
567 * When checking, we use entries from near the top of the
568 * stack, preferring $GIT_DIR/info/attributes, then
569 * .gitattributes in deeper directories to shallower ones,
570 * and finally use the built-in set as the default.
572 bootstrap_attr_stack();
575 * Pop the "info" one that is always at the top of the stack.
577 info = attr_stack;
578 attr_stack = info->prev;
581 * Pop the ones from directories that are not the prefix of
582 * the path we are checking. Break out of the loop when we see
583 * the root one (whose origin is an empty string "") or the builtin
584 * one (whose origin is NULL) without popping it.
586 while (attr_stack->origin) {
587 int namelen = strlen(attr_stack->origin);
589 elem = attr_stack;
590 if (namelen <= dirlen &&
591 !strncmp(elem->origin, path, namelen) &&
592 (!namelen || path[namelen] == '/'))
593 break;
595 debug_pop(elem);
596 attr_stack = elem->prev;
597 free_attr_elem(elem);
601 * Read from parent directories and push them down
603 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
605 * bootstrap_attr_stack() should have added, and the
606 * above loop should have stopped before popping, the
607 * root element whose attr_stack->origin is set to an
608 * empty string.
610 struct strbuf pathbuf = STRBUF_INIT;
612 assert(attr_stack->origin);
613 while (1) {
614 len = strlen(attr_stack->origin);
615 if (dirlen <= len)
616 break;
617 cp = memchr(path + len + 1, '/', dirlen - len - 1);
618 if (!cp)
619 cp = path + dirlen;
620 strbuf_add(&pathbuf, path, cp - path);
621 strbuf_addch(&pathbuf, '/');
622 strbuf_addstr(&pathbuf, GITATTRIBUTES_FILE);
623 elem = read_attr(pathbuf.buf, 0);
624 strbuf_setlen(&pathbuf, cp - path);
625 elem->origin = strbuf_detach(&pathbuf, &elem->originlen);
626 elem->prev = attr_stack;
627 attr_stack = elem;
628 debug_push(elem);
631 strbuf_release(&pathbuf);
635 * Finally push the "info" one at the top of the stack.
637 info->prev = attr_stack;
638 attr_stack = info;
641 static int path_matches(const char *pathname, int pathlen,
642 const char *basename,
643 const char *pattern,
644 const char *base, int baselen)
646 if (!strchr(pattern, '/')) {
647 return (fnmatch_icase(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_icase(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, const char *basename,
689 struct attr_stack *stk, int rem)
691 int i;
692 const char *base = stk->origin ? stk->origin : "";
694 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
695 struct match_attr *a = stk->attrs[i];
696 if (a->is_macro)
697 continue;
698 if (path_matches(path, pathlen, basename,
699 a->u.pattern, base, stk->originlen))
700 rem = fill_one("fill", a, rem);
702 return rem;
705 static int macroexpand_one(int attr_nr, int rem)
707 struct attr_stack *stk;
708 struct match_attr *a = NULL;
709 int i;
711 if (check_all_attr[attr_nr].value != ATTR__TRUE)
712 return rem;
714 for (stk = attr_stack; !a && stk; stk = stk->prev)
715 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
716 struct match_attr *ma = stk->attrs[i];
717 if (!ma->is_macro)
718 continue;
719 if (ma->u.attr->attr_nr == attr_nr)
720 a = ma;
723 if (a)
724 rem = fill_one("expand", a, rem);
726 return rem;
730 * Collect all attributes for path into the array pointed to by
731 * check_all_attr.
733 static void collect_all_attrs(const char *path)
735 struct attr_stack *stk;
736 int i, pathlen, rem;
737 const char *basename;
739 prepare_attr_stack(path);
740 for (i = 0; i < attr_nr; i++)
741 check_all_attr[i].value = ATTR__UNKNOWN;
743 basename = strrchr(path, '/');
744 basename = basename ? basename + 1 : path;
746 pathlen = strlen(path);
747 rem = attr_nr;
748 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
749 rem = fill(path, pathlen, basename, stk, rem);
752 int git_check_attr(const char *path, int num, struct git_attr_check *check)
754 int i;
756 collect_all_attrs(path);
758 for (i = 0; i < num; i++) {
759 const char *value = check_all_attr[check[i].attr->attr_nr].value;
760 if (value == ATTR__UNKNOWN)
761 value = ATTR__UNSET;
762 check[i].value = value;
765 return 0;
768 int git_all_attrs(const char *path, int *num, struct git_attr_check **check)
770 int i, count, j;
772 collect_all_attrs(path);
774 /* Count the number of attributes that are set. */
775 count = 0;
776 for (i = 0; i < attr_nr; i++) {
777 const char *value = check_all_attr[i].value;
778 if (value != ATTR__UNSET && value != ATTR__UNKNOWN)
779 ++count;
781 *num = count;
782 *check = xmalloc(sizeof(**check) * count);
783 j = 0;
784 for (i = 0; i < attr_nr; i++) {
785 const char *value = check_all_attr[i].value;
786 if (value != ATTR__UNSET && value != ATTR__UNKNOWN) {
787 (*check)[j].attr = check_all_attr[i].attr;
788 (*check)[j].value = value;
789 ++j;
793 return 0;
796 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
798 enum git_attr_direction old = direction;
800 if (is_bare_repository() && new != GIT_ATTR_INDEX)
801 die("BUG: non-INDEX attr direction in a bare repo");
803 direction = new;
804 if (new != old)
805 drop_attr_stack();
806 use_index = istate;