attr: more matching optimizations from .gitignore
[git/jnareb-git.git] / attr.c
blob2fc6353628a2cd7ffe363a7b600dc8f4283d0e1f
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;
118 struct pattern {
119 const char *pattern;
120 int patternlen;
121 int nowildcardlen;
122 int flags; /* EXC_FLAG_* */
126 * One rule, as from a .gitattributes file.
128 * If is_macro is true, then u.attr is a pointer to the git_attr being
129 * defined.
131 * If is_macro is false, then u.pattern points at the filename pattern
132 * to which the rule applies. (The memory pointed to is part of the
133 * memory block allocated for the match_attr instance.)
135 * In either case, num_attr is the number of attributes affected by
136 * this rule, and state is an array listing them. The attributes are
137 * listed as they appear in the file (macros unexpanded).
139 struct match_attr {
140 union {
141 struct pattern pat;
142 struct git_attr *attr;
143 } u;
144 char is_macro;
145 unsigned num_attr;
146 struct attr_state state[FLEX_ARRAY];
149 static const char blank[] = " \t\r\n";
152 * Parse a whitespace-delimited attribute state (i.e., "attr",
153 * "-attr", "!attr", or "attr=value") from the string starting at src.
154 * If e is not NULL, write the results to *e. Return a pointer to the
155 * remainder of the string (with leading whitespace removed), or NULL
156 * if there was an error.
158 static const char *parse_attr(const char *src, int lineno, const char *cp,
159 struct attr_state *e)
161 const char *ep, *equals;
162 int len;
164 ep = cp + strcspn(cp, blank);
165 equals = strchr(cp, '=');
166 if (equals && ep < equals)
167 equals = NULL;
168 if (equals)
169 len = equals - cp;
170 else
171 len = ep - cp;
172 if (!e) {
173 if (*cp == '-' || *cp == '!') {
174 cp++;
175 len--;
177 if (invalid_attr_name(cp, len)) {
178 fprintf(stderr,
179 "%.*s is not a valid attribute name: %s:%d\n",
180 len, cp, src, lineno);
181 return NULL;
183 } else {
184 if (*cp == '-' || *cp == '!') {
185 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
186 cp++;
187 len--;
189 else if (!equals)
190 e->setto = ATTR__TRUE;
191 else {
192 e->setto = xmemdupz(equals + 1, ep - equals - 1);
194 e->attr = git_attr_internal(cp, len);
196 return ep + strspn(ep, blank);
199 static struct match_attr *parse_attr_line(const char *line, const char *src,
200 int lineno, int macro_ok)
202 int namelen;
203 int num_attr, i;
204 const char *cp, *name, *states;
205 struct match_attr *res = NULL;
206 int is_macro;
208 cp = line + strspn(line, blank);
209 if (!*cp || *cp == '#')
210 return NULL;
211 name = cp;
212 namelen = strcspn(name, blank);
213 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
214 !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
215 if (!macro_ok) {
216 fprintf(stderr, "%s not allowed: %s:%d\n",
217 name, src, lineno);
218 return NULL;
220 is_macro = 1;
221 name += strlen(ATTRIBUTE_MACRO_PREFIX);
222 name += strspn(name, blank);
223 namelen = strcspn(name, blank);
224 if (invalid_attr_name(name, namelen)) {
225 fprintf(stderr,
226 "%.*s is not a valid attribute name: %s:%d\n",
227 namelen, name, src, lineno);
228 return NULL;
231 else
232 is_macro = 0;
234 states = name + namelen;
235 states += strspn(states, blank);
237 /* First pass to count the attr_states */
238 for (cp = states, num_attr = 0; *cp; num_attr++) {
239 cp = parse_attr(src, lineno, cp, NULL);
240 if (!cp)
241 return NULL;
244 res = xcalloc(1,
245 sizeof(*res) +
246 sizeof(struct attr_state) * num_attr +
247 (is_macro ? 0 : namelen + 1));
248 if (is_macro)
249 res->u.attr = git_attr_internal(name, namelen);
250 else {
251 char *p = (char *)&(res->state[num_attr]);
252 memcpy(p, name, namelen);
253 res->u.pat.pattern = p;
254 parse_exclude_pattern(&res->u.pat.pattern,
255 &res->u.pat.patternlen,
256 &res->u.pat.flags,
257 &res->u.pat.nowildcardlen);
258 if (res->u.pat.flags & EXC_FLAG_NEGATIVE)
259 die(_("Negative patterns are forbidden in git attributes\n"
260 "Use '\\!' for literal leading exclamation."));
262 res->is_macro = is_macro;
263 res->num_attr = num_attr;
265 /* Second pass to fill the attr_states */
266 for (cp = states, i = 0; *cp; i++) {
267 cp = parse_attr(src, lineno, cp, &(res->state[i]));
270 return res;
274 * Like info/exclude and .gitignore, the attribute information can
275 * come from many places.
277 * (1) .gitattribute file of the same directory;
278 * (2) .gitattribute file of the parent directory if (1) does not have
279 * any match; this goes recursively upwards, just like .gitignore.
280 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
282 * In the same file, later entries override the earlier match, so in the
283 * global list, we would have entries from info/attributes the earliest
284 * (reading the file from top to bottom), .gitattribute of the root
285 * directory (again, reading the file from top to bottom) down to the
286 * current directory, and then scan the list backwards to find the first match.
287 * This is exactly the same as what excluded() does in dir.c to deal with
288 * .gitignore
291 static struct attr_stack {
292 struct attr_stack *prev;
293 char *origin;
294 size_t originlen;
295 unsigned num_matches;
296 unsigned alloc;
297 struct match_attr **attrs;
298 } *attr_stack;
300 static void free_attr_elem(struct attr_stack *e)
302 int i;
303 free(e->origin);
304 for (i = 0; i < e->num_matches; i++) {
305 struct match_attr *a = e->attrs[i];
306 int j;
307 for (j = 0; j < a->num_attr; j++) {
308 const char *setto = a->state[j].setto;
309 if (setto == ATTR__TRUE ||
310 setto == ATTR__FALSE ||
311 setto == ATTR__UNSET ||
312 setto == ATTR__UNKNOWN)
314 else
315 free((char *) setto);
317 free(a);
319 free(e->attrs);
320 free(e);
323 static const char *builtin_attr[] = {
324 "[attr]binary -diff -text",
325 NULL,
328 static void handle_attr_line(struct attr_stack *res,
329 const char *line,
330 const char *src,
331 int lineno,
332 int macro_ok)
334 struct match_attr *a;
336 a = parse_attr_line(line, src, lineno, macro_ok);
337 if (!a)
338 return;
339 if (res->alloc <= res->num_matches) {
340 res->alloc = alloc_nr(res->num_matches);
341 res->attrs = xrealloc(res->attrs,
342 sizeof(struct match_attr *) *
343 res->alloc);
345 res->attrs[res->num_matches++] = a;
348 static struct attr_stack *read_attr_from_array(const char **list)
350 struct attr_stack *res;
351 const char *line;
352 int lineno = 0;
354 res = xcalloc(1, sizeof(*res));
355 while ((line = *(list++)) != NULL)
356 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
357 return res;
360 static enum git_attr_direction direction;
361 static struct index_state *use_index;
363 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
365 FILE *fp = fopen(path, "r");
366 struct attr_stack *res;
367 char buf[2048];
368 int lineno = 0;
370 if (!fp)
371 return NULL;
372 res = xcalloc(1, sizeof(*res));
373 while (fgets(buf, sizeof(buf), fp))
374 handle_attr_line(res, buf, path, ++lineno, macro_ok);
375 fclose(fp);
376 return res;
379 static void *read_index_data(const char *path)
381 int pos, len;
382 unsigned long sz;
383 enum object_type type;
384 void *data;
385 struct index_state *istate = use_index ? use_index : &the_index;
387 len = strlen(path);
388 pos = index_name_pos(istate, path, len);
389 if (pos < 0) {
391 * We might be in the middle of a merge, in which
392 * case we would read stage #2 (ours).
394 int i;
395 for (i = -pos - 1;
396 (pos < 0 && i < istate->cache_nr &&
397 !strcmp(istate->cache[i]->name, path));
398 i++)
399 if (ce_stage(istate->cache[i]) == 2)
400 pos = i;
402 if (pos < 0)
403 return NULL;
404 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
405 if (!data || type != OBJ_BLOB) {
406 free(data);
407 return NULL;
409 return data;
412 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
414 struct attr_stack *res;
415 char *buf, *sp;
416 int lineno = 0;
418 buf = read_index_data(path);
419 if (!buf)
420 return NULL;
422 res = xcalloc(1, sizeof(*res));
423 for (sp = buf; *sp; ) {
424 char *ep;
425 int more;
426 for (ep = sp; *ep && *ep != '\n'; ep++)
428 more = (*ep == '\n');
429 *ep = '\0';
430 handle_attr_line(res, sp, path, ++lineno, macro_ok);
431 sp = ep + more;
433 free(buf);
434 return res;
437 static struct attr_stack *read_attr(const char *path, int macro_ok)
439 struct attr_stack *res;
441 if (direction == GIT_ATTR_CHECKOUT) {
442 res = read_attr_from_index(path, macro_ok);
443 if (!res)
444 res = read_attr_from_file(path, macro_ok);
446 else if (direction == GIT_ATTR_CHECKIN) {
447 res = read_attr_from_file(path, macro_ok);
448 if (!res)
450 * There is no checked out .gitattributes file there, but
451 * we might have it in the index. We allow operation in a
452 * sparsely checked out work tree, so read from it.
454 res = read_attr_from_index(path, macro_ok);
456 else
457 res = read_attr_from_index(path, macro_ok);
458 if (!res)
459 res = xcalloc(1, sizeof(*res));
460 return res;
463 #if DEBUG_ATTR
464 static void debug_info(const char *what, struct attr_stack *elem)
466 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
468 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
470 const char *value = v;
472 if (ATTR_TRUE(value))
473 value = "set";
474 else if (ATTR_FALSE(value))
475 value = "unset";
476 else if (ATTR_UNSET(value))
477 value = "unspecified";
479 fprintf(stderr, "%s: %s => %s (%s)\n",
480 what, attr->name, (char *) value, match);
482 #define debug_push(a) debug_info("push", (a))
483 #define debug_pop(a) debug_info("pop", (a))
484 #else
485 #define debug_push(a) do { ; } while (0)
486 #define debug_pop(a) do { ; } while (0)
487 #define debug_set(a,b,c,d) do { ; } while (0)
488 #endif
490 static void drop_attr_stack(void)
492 while (attr_stack) {
493 struct attr_stack *elem = attr_stack;
494 attr_stack = elem->prev;
495 free_attr_elem(elem);
499 static const char *git_etc_gitattributes(void)
501 static const char *system_wide;
502 if (!system_wide)
503 system_wide = system_path(ETC_GITATTRIBUTES);
504 return system_wide;
507 static int git_attr_system(void)
509 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
512 static void bootstrap_attr_stack(void)
514 struct attr_stack *elem;
516 if (attr_stack)
517 return;
519 elem = read_attr_from_array(builtin_attr);
520 elem->origin = NULL;
521 elem->prev = attr_stack;
522 attr_stack = elem;
524 if (git_attr_system()) {
525 elem = read_attr_from_file(git_etc_gitattributes(), 1);
526 if (elem) {
527 elem->origin = NULL;
528 elem->prev = attr_stack;
529 attr_stack = elem;
533 if (git_attributes_file) {
534 elem = read_attr_from_file(git_attributes_file, 1);
535 if (elem) {
536 elem->origin = NULL;
537 elem->prev = attr_stack;
538 attr_stack = elem;
542 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
543 elem = read_attr(GITATTRIBUTES_FILE, 1);
544 elem->origin = xstrdup("");
545 elem->originlen = 0;
546 elem->prev = attr_stack;
547 attr_stack = elem;
548 debug_push(elem);
551 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
552 if (!elem)
553 elem = xcalloc(1, sizeof(*elem));
554 elem->origin = NULL;
555 elem->prev = attr_stack;
556 attr_stack = elem;
559 static void prepare_attr_stack(const char *path)
561 struct attr_stack *elem, *info;
562 int dirlen, len;
563 const char *cp;
565 cp = strrchr(path, '/');
566 if (!cp)
567 dirlen = 0;
568 else
569 dirlen = cp - path;
572 * At the bottom of the attribute stack is the built-in
573 * set of attribute definitions, followed by the contents
574 * of $(prefix)/etc/gitattributes and a file specified by
575 * core.attributesfile. Then, contents from
576 * .gitattribute files from directories closer to the
577 * root to the ones in deeper directories are pushed
578 * to the stack. Finally, at the very top of the stack
579 * we always keep the contents of $GIT_DIR/info/attributes.
581 * When checking, we use entries from near the top of the
582 * stack, preferring $GIT_DIR/info/attributes, then
583 * .gitattributes in deeper directories to shallower ones,
584 * and finally use the built-in set as the default.
586 bootstrap_attr_stack();
589 * Pop the "info" one that is always at the top of the stack.
591 info = attr_stack;
592 attr_stack = info->prev;
595 * Pop the ones from directories that are not the prefix of
596 * the path we are checking. Break out of the loop when we see
597 * the root one (whose origin is an empty string "") or the builtin
598 * one (whose origin is NULL) without popping it.
600 while (attr_stack->origin) {
601 int namelen = strlen(attr_stack->origin);
603 elem = attr_stack;
604 if (namelen <= dirlen &&
605 !strncmp(elem->origin, path, namelen) &&
606 (!namelen || path[namelen] == '/'))
607 break;
609 debug_pop(elem);
610 attr_stack = elem->prev;
611 free_attr_elem(elem);
615 * Read from parent directories and push them down
617 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
619 * bootstrap_attr_stack() should have added, and the
620 * above loop should have stopped before popping, the
621 * root element whose attr_stack->origin is set to an
622 * empty string.
624 struct strbuf pathbuf = STRBUF_INIT;
626 assert(attr_stack->origin);
627 while (1) {
628 len = strlen(attr_stack->origin);
629 if (dirlen <= len)
630 break;
631 cp = memchr(path + len + 1, '/', dirlen - len - 1);
632 if (!cp)
633 cp = path + dirlen;
634 strbuf_add(&pathbuf, path, cp - path);
635 strbuf_addch(&pathbuf, '/');
636 strbuf_addstr(&pathbuf, GITATTRIBUTES_FILE);
637 elem = read_attr(pathbuf.buf, 0);
638 strbuf_setlen(&pathbuf, cp - path);
639 elem->origin = strbuf_detach(&pathbuf, &elem->originlen);
640 elem->prev = attr_stack;
641 attr_stack = elem;
642 debug_push(elem);
645 strbuf_release(&pathbuf);
649 * Finally push the "info" one at the top of the stack.
651 info->prev = attr_stack;
652 attr_stack = info;
655 static int path_matches(const char *pathname, int pathlen,
656 const char *basename,
657 const struct pattern *pat,
658 const char *base, int baselen)
660 const char *pattern = pat->pattern;
661 int prefix = pat->nowildcardlen;
663 if (pat->flags & EXC_FLAG_NODIR) {
664 return match_basename(basename,
665 pathlen - (basename - pathname),
666 pattern, prefix,
667 pat->patternlen, pat->flags);
669 return match_pathname(pathname, pathlen,
670 base, baselen,
671 pattern, prefix, pat->patternlen, pat->flags);
674 static int macroexpand_one(int attr_nr, int rem);
676 static int fill_one(const char *what, struct match_attr *a, int rem)
678 struct git_attr_check *check = check_all_attr;
679 int i;
681 for (i = a->num_attr - 1; 0 < rem && 0 <= i; i--) {
682 struct git_attr *attr = a->state[i].attr;
683 const char **n = &(check[attr->attr_nr].value);
684 const char *v = a->state[i].setto;
686 if (*n == ATTR__UNKNOWN) {
687 debug_set(what,
688 a->is_macro ? a->u.attr->name : a->u.pattern,
689 attr, v);
690 *n = v;
691 rem--;
692 rem = macroexpand_one(attr->attr_nr, rem);
695 return rem;
698 static int fill(const char *path, int pathlen, const char *basename,
699 struct attr_stack *stk, int rem)
701 int i;
702 const char *base = stk->origin ? stk->origin : "";
704 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
705 struct match_attr *a = stk->attrs[i];
706 if (a->is_macro)
707 continue;
708 if (path_matches(path, pathlen, basename,
709 &a->u.pat, base, stk->originlen))
710 rem = fill_one("fill", a, rem);
712 return rem;
715 static int macroexpand_one(int attr_nr, int rem)
717 struct attr_stack *stk;
718 struct match_attr *a = NULL;
719 int i;
721 if (check_all_attr[attr_nr].value != ATTR__TRUE)
722 return rem;
724 for (stk = attr_stack; !a && stk; stk = stk->prev)
725 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
726 struct match_attr *ma = stk->attrs[i];
727 if (!ma->is_macro)
728 continue;
729 if (ma->u.attr->attr_nr == attr_nr)
730 a = ma;
733 if (a)
734 rem = fill_one("expand", a, rem);
736 return rem;
740 * Collect all attributes for path into the array pointed to by
741 * check_all_attr.
743 static void collect_all_attrs(const char *path)
745 struct attr_stack *stk;
746 int i, pathlen, rem;
747 const char *basename;
749 prepare_attr_stack(path);
750 for (i = 0; i < attr_nr; i++)
751 check_all_attr[i].value = ATTR__UNKNOWN;
753 basename = strrchr(path, '/');
754 basename = basename ? basename + 1 : path;
756 pathlen = strlen(path);
757 rem = attr_nr;
758 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
759 rem = fill(path, pathlen, basename, stk, rem);
762 int git_check_attr(const char *path, int num, struct git_attr_check *check)
764 int i;
766 collect_all_attrs(path);
768 for (i = 0; i < num; i++) {
769 const char *value = check_all_attr[check[i].attr->attr_nr].value;
770 if (value == ATTR__UNKNOWN)
771 value = ATTR__UNSET;
772 check[i].value = value;
775 return 0;
778 int git_all_attrs(const char *path, int *num, struct git_attr_check **check)
780 int i, count, j;
782 collect_all_attrs(path);
784 /* Count the number of attributes that are set. */
785 count = 0;
786 for (i = 0; i < attr_nr; i++) {
787 const char *value = check_all_attr[i].value;
788 if (value != ATTR__UNSET && value != ATTR__UNKNOWN)
789 ++count;
791 *num = count;
792 *check = xmalloc(sizeof(**check) * count);
793 j = 0;
794 for (i = 0; i < attr_nr; i++) {
795 const char *value = check_all_attr[i].value;
796 if (value != ATTR__UNSET && value != ATTR__UNKNOWN) {
797 (*check)[j].attr = check_all_attr[i].attr;
798 (*check)[j].value = value;
799 ++j;
803 return 0;
806 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
808 enum git_attr_direction old = direction;
810 if (is_bare_repository() && new != GIT_ATTR_INDEX)
811 die("BUG: non-INDEX attr direction in a bare repo");
813 direction = new;
814 if (new != old)
815 drop_attr_stack();
816 use_index = istate;