t7810-grep.sh: fix a whitespace inconsistency
[git/git-svn.git] / attr.c
blob1f9eebd2ddb7fe8a561500c41f33a52d1fb4fdd1
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 int maybe_macro;
36 int maybe_real;
37 char name[FLEX_ARRAY];
39 static int attr_nr;
40 static int cannot_trust_maybe_real;
42 static struct git_attr_check *check_all_attr;
43 static struct git_attr *(git_attr_hash[HASHSIZE]);
45 char *git_attr_name(struct git_attr *attr)
47 return attr->name;
50 static unsigned hash_name(const char *name, int namelen)
52 unsigned val = 0, c;
54 while (namelen--) {
55 c = *name++;
56 val = ((val << 7) | (val >> 22)) ^ c;
58 return val;
61 static int invalid_attr_name(const char *name, int namelen)
64 * Attribute name cannot begin with '-' and must consist of
65 * characters from [-A-Za-z0-9_.].
67 if (namelen <= 0 || *name == '-')
68 return -1;
69 while (namelen--) {
70 char ch = *name++;
71 if (! (ch == '-' || ch == '.' || ch == '_' ||
72 ('0' <= ch && ch <= '9') ||
73 ('a' <= ch && ch <= 'z') ||
74 ('A' <= ch && ch <= 'Z')) )
75 return -1;
77 return 0;
80 static struct git_attr *git_attr_internal(const char *name, int len)
82 unsigned hval = hash_name(name, len);
83 unsigned pos = hval % HASHSIZE;
84 struct git_attr *a;
86 for (a = git_attr_hash[pos]; a; a = a->next) {
87 if (a->h == hval &&
88 !memcmp(a->name, name, len) && !a->name[len])
89 return a;
92 if (invalid_attr_name(name, len))
93 return NULL;
95 a = xmalloc(sizeof(*a) + len + 1);
96 memcpy(a->name, name, len);
97 a->name[len] = 0;
98 a->h = hval;
99 a->next = git_attr_hash[pos];
100 a->attr_nr = attr_nr++;
101 a->maybe_macro = 0;
102 a->maybe_real = 0;
103 git_attr_hash[pos] = a;
105 REALLOC_ARRAY(check_all_attr, attr_nr);
106 check_all_attr[a->attr_nr].attr = a;
107 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
108 return a;
111 struct git_attr *git_attr(const char *name)
113 return git_attr_internal(name, strlen(name));
116 /* What does a matched pattern decide? */
117 struct attr_state {
118 struct git_attr *attr;
119 const char *setto;
122 struct pattern {
123 const char *pattern;
124 int patternlen;
125 int nowildcardlen;
126 int flags; /* EXC_FLAG_* */
130 * One rule, as from a .gitattributes file.
132 * If is_macro is true, then u.attr is a pointer to the git_attr being
133 * defined.
135 * If is_macro is false, then u.pattern points at the filename pattern
136 * to which the rule applies. (The memory pointed to is part of the
137 * memory block allocated for the match_attr instance.)
139 * In either case, num_attr is the number of attributes affected by
140 * this rule, and state is an array listing them. The attributes are
141 * listed as they appear in the file (macros unexpanded).
143 struct match_attr {
144 union {
145 struct pattern pat;
146 struct git_attr *attr;
147 } u;
148 char is_macro;
149 unsigned num_attr;
150 struct attr_state state[FLEX_ARRAY];
153 static const char blank[] = " \t\r\n";
156 * Parse a whitespace-delimited attribute state (i.e., "attr",
157 * "-attr", "!attr", or "attr=value") from the string starting at src.
158 * If e is not NULL, write the results to *e. Return a pointer to the
159 * remainder of the string (with leading whitespace removed), or NULL
160 * if there was an error.
162 static const char *parse_attr(const char *src, int lineno, const char *cp,
163 struct attr_state *e)
165 const char *ep, *equals;
166 int len;
168 ep = cp + strcspn(cp, blank);
169 equals = strchr(cp, '=');
170 if (equals && ep < equals)
171 equals = NULL;
172 if (equals)
173 len = equals - cp;
174 else
175 len = ep - cp;
176 if (!e) {
177 if (*cp == '-' || *cp == '!') {
178 cp++;
179 len--;
181 if (invalid_attr_name(cp, len)) {
182 fprintf(stderr,
183 "%.*s is not a valid attribute name: %s:%d\n",
184 len, cp, src, lineno);
185 return NULL;
187 } else {
188 if (*cp == '-' || *cp == '!') {
189 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
190 cp++;
191 len--;
193 else if (!equals)
194 e->setto = ATTR__TRUE;
195 else {
196 e->setto = xmemdupz(equals + 1, ep - equals - 1);
198 e->attr = git_attr_internal(cp, len);
200 return ep + strspn(ep, blank);
203 static struct match_attr *parse_attr_line(const char *line, const char *src,
204 int lineno, int macro_ok)
206 int namelen;
207 int num_attr, i;
208 const char *cp, *name, *states;
209 struct match_attr *res = NULL;
210 int is_macro;
212 cp = line + strspn(line, blank);
213 if (!*cp || *cp == '#')
214 return NULL;
215 name = cp;
216 namelen = strcspn(name, blank);
217 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
218 starts_with(name, ATTRIBUTE_MACRO_PREFIX)) {
219 if (!macro_ok) {
220 fprintf(stderr, "%s not allowed: %s:%d\n",
221 name, src, lineno);
222 return NULL;
224 is_macro = 1;
225 name += strlen(ATTRIBUTE_MACRO_PREFIX);
226 name += strspn(name, blank);
227 namelen = strcspn(name, blank);
228 if (invalid_attr_name(name, namelen)) {
229 fprintf(stderr,
230 "%.*s is not a valid attribute name: %s:%d\n",
231 namelen, name, src, lineno);
232 return NULL;
235 else
236 is_macro = 0;
238 states = name + namelen;
239 states += strspn(states, blank);
241 /* First pass to count the attr_states */
242 for (cp = states, num_attr = 0; *cp; num_attr++) {
243 cp = parse_attr(src, lineno, cp, NULL);
244 if (!cp)
245 return NULL;
248 res = xcalloc(1,
249 sizeof(*res) +
250 sizeof(struct attr_state) * num_attr +
251 (is_macro ? 0 : namelen + 1));
252 if (is_macro) {
253 res->u.attr = git_attr_internal(name, namelen);
254 res->u.attr->maybe_macro = 1;
255 } else {
256 char *p = (char *)&(res->state[num_attr]);
257 memcpy(p, name, namelen);
258 res->u.pat.pattern = p;
259 parse_exclude_pattern(&res->u.pat.pattern,
260 &res->u.pat.patternlen,
261 &res->u.pat.flags,
262 &res->u.pat.nowildcardlen);
263 if (res->u.pat.flags & EXC_FLAG_NEGATIVE) {
264 warning(_("Negative patterns are ignored in git attributes\n"
265 "Use '\\!' for literal leading exclamation."));
266 return NULL;
269 res->is_macro = is_macro;
270 res->num_attr = num_attr;
272 /* Second pass to fill the attr_states */
273 for (cp = states, i = 0; *cp; i++) {
274 cp = parse_attr(src, lineno, cp, &(res->state[i]));
275 if (!is_macro)
276 res->state[i].attr->maybe_real = 1;
277 if (res->state[i].attr->maybe_macro)
278 cannot_trust_maybe_real = 1;
281 return res;
285 * Like info/exclude and .gitignore, the attribute information can
286 * come from many places.
288 * (1) .gitattribute file of the same directory;
289 * (2) .gitattribute file of the parent directory if (1) does not have
290 * any match; this goes recursively upwards, just like .gitignore.
291 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
293 * In the same file, later entries override the earlier match, so in the
294 * global list, we would have entries from info/attributes the earliest
295 * (reading the file from top to bottom), .gitattribute of the root
296 * directory (again, reading the file from top to bottom) down to the
297 * current directory, and then scan the list backwards to find the first match.
298 * This is exactly the same as what is_excluded() does in dir.c to deal with
299 * .gitignore
302 static struct attr_stack {
303 struct attr_stack *prev;
304 char *origin;
305 size_t originlen;
306 unsigned num_matches;
307 unsigned alloc;
308 struct match_attr **attrs;
309 } *attr_stack;
311 static void free_attr_elem(struct attr_stack *e)
313 int i;
314 free(e->origin);
315 for (i = 0; i < e->num_matches; i++) {
316 struct match_attr *a = e->attrs[i];
317 int j;
318 for (j = 0; j < a->num_attr; j++) {
319 const char *setto = a->state[j].setto;
320 if (setto == ATTR__TRUE ||
321 setto == ATTR__FALSE ||
322 setto == ATTR__UNSET ||
323 setto == ATTR__UNKNOWN)
325 else
326 free((char *) setto);
328 free(a);
330 free(e->attrs);
331 free(e);
334 static const char *builtin_attr[] = {
335 "[attr]binary -diff -merge -text",
336 NULL,
339 static void handle_attr_line(struct attr_stack *res,
340 const char *line,
341 const char *src,
342 int lineno,
343 int macro_ok)
345 struct match_attr *a;
347 a = parse_attr_line(line, src, lineno, macro_ok);
348 if (!a)
349 return;
350 ALLOC_GROW(res->attrs, res->num_matches + 1, res->alloc);
351 res->attrs[res->num_matches++] = a;
354 static struct attr_stack *read_attr_from_array(const char **list)
356 struct attr_stack *res;
357 const char *line;
358 int lineno = 0;
360 res = xcalloc(1, sizeof(*res));
361 while ((line = *(list++)) != NULL)
362 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
363 return res;
366 static enum git_attr_direction direction;
367 static struct index_state *use_index;
369 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
371 FILE *fp = fopen(path, "r");
372 struct attr_stack *res;
373 char buf[2048];
374 int lineno = 0;
376 if (!fp) {
377 if (errno != ENOENT && errno != ENOTDIR)
378 warn_on_inaccessible(path);
379 return NULL;
381 res = xcalloc(1, sizeof(*res));
382 while (fgets(buf, sizeof(buf), fp))
383 handle_attr_line(res, buf, path, ++lineno, macro_ok);
384 fclose(fp);
385 return res;
388 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
390 struct attr_stack *res;
391 char *buf, *sp;
392 int lineno = 0;
394 buf = read_blob_data_from_index(use_index ? use_index : &the_index, path, NULL);
395 if (!buf)
396 return NULL;
398 res = xcalloc(1, sizeof(*res));
399 for (sp = buf; *sp; ) {
400 char *ep;
401 int more;
402 for (ep = sp; *ep && *ep != '\n'; ep++)
404 more = (*ep == '\n');
405 *ep = '\0';
406 handle_attr_line(res, sp, path, ++lineno, macro_ok);
407 sp = ep + more;
409 free(buf);
410 return res;
413 static struct attr_stack *read_attr(const char *path, int macro_ok)
415 struct attr_stack *res;
417 if (direction == GIT_ATTR_CHECKOUT) {
418 res = read_attr_from_index(path, macro_ok);
419 if (!res)
420 res = read_attr_from_file(path, macro_ok);
422 else if (direction == GIT_ATTR_CHECKIN) {
423 res = read_attr_from_file(path, macro_ok);
424 if (!res)
426 * There is no checked out .gitattributes file there, but
427 * we might have it in the index. We allow operation in a
428 * sparsely checked out work tree, so read from it.
430 res = read_attr_from_index(path, macro_ok);
432 else
433 res = read_attr_from_index(path, macro_ok);
434 if (!res)
435 res = xcalloc(1, sizeof(*res));
436 return res;
439 #if DEBUG_ATTR
440 static void debug_info(const char *what, struct attr_stack *elem)
442 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
444 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
446 const char *value = v;
448 if (ATTR_TRUE(value))
449 value = "set";
450 else if (ATTR_FALSE(value))
451 value = "unset";
452 else if (ATTR_UNSET(value))
453 value = "unspecified";
455 fprintf(stderr, "%s: %s => %s (%s)\n",
456 what, attr->name, (char *) value, match);
458 #define debug_push(a) debug_info("push", (a))
459 #define debug_pop(a) debug_info("pop", (a))
460 #else
461 #define debug_push(a) do { ; } while (0)
462 #define debug_pop(a) do { ; } while (0)
463 #define debug_set(a,b,c,d) do { ; } while (0)
464 #endif
466 static void drop_attr_stack(void)
468 while (attr_stack) {
469 struct attr_stack *elem = attr_stack;
470 attr_stack = elem->prev;
471 free_attr_elem(elem);
475 static const char *git_etc_gitattributes(void)
477 static const char *system_wide;
478 if (!system_wide)
479 system_wide = system_path(ETC_GITATTRIBUTES);
480 return system_wide;
483 static int git_attr_system(void)
485 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
488 static void bootstrap_attr_stack(void)
490 struct attr_stack *elem;
491 char *xdg_attributes_file;
493 if (attr_stack)
494 return;
496 elem = read_attr_from_array(builtin_attr);
497 elem->origin = NULL;
498 elem->prev = attr_stack;
499 attr_stack = elem;
501 if (git_attr_system()) {
502 elem = read_attr_from_file(git_etc_gitattributes(), 1);
503 if (elem) {
504 elem->origin = NULL;
505 elem->prev = attr_stack;
506 attr_stack = elem;
510 if (!git_attributes_file) {
511 home_config_paths(NULL, &xdg_attributes_file, "attributes");
512 git_attributes_file = xdg_attributes_file;
514 if (git_attributes_file) {
515 elem = read_attr_from_file(git_attributes_file, 1);
516 if (elem) {
517 elem->origin = NULL;
518 elem->prev = attr_stack;
519 attr_stack = elem;
523 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
524 elem = read_attr(GITATTRIBUTES_FILE, 1);
525 elem->origin = xstrdup("");
526 elem->originlen = 0;
527 elem->prev = attr_stack;
528 attr_stack = elem;
529 debug_push(elem);
532 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
533 if (!elem)
534 elem = xcalloc(1, sizeof(*elem));
535 elem->origin = NULL;
536 elem->prev = attr_stack;
537 attr_stack = elem;
540 static void prepare_attr_stack(const char *path, int dirlen)
542 struct attr_stack *elem, *info;
543 int len;
544 const char *cp;
547 * At the bottom of the attribute stack is the built-in
548 * set of attribute definitions, followed by the contents
549 * of $(prefix)/etc/gitattributes and a file specified by
550 * core.attributesfile. Then, contents from
551 * .gitattribute files from directories closer to the
552 * root to the ones in deeper directories are pushed
553 * to the stack. Finally, at the very top of the stack
554 * we always keep the contents of $GIT_DIR/info/attributes.
556 * When checking, we use entries from near the top of the
557 * stack, preferring $GIT_DIR/info/attributes, then
558 * .gitattributes in deeper directories to shallower ones,
559 * and finally use the built-in set as the default.
561 bootstrap_attr_stack();
564 * Pop the "info" one that is always at the top of the stack.
566 info = attr_stack;
567 attr_stack = info->prev;
570 * Pop the ones from directories that are not the prefix of
571 * the path we are checking. Break out of the loop when we see
572 * the root one (whose origin is an empty string "") or the builtin
573 * one (whose origin is NULL) without popping it.
575 while (attr_stack->origin) {
576 int namelen = strlen(attr_stack->origin);
578 elem = attr_stack;
579 if (namelen <= dirlen &&
580 !strncmp(elem->origin, path, namelen) &&
581 (!namelen || path[namelen] == '/'))
582 break;
584 debug_pop(elem);
585 attr_stack = elem->prev;
586 free_attr_elem(elem);
590 * Read from parent directories and push them down
592 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
594 * bootstrap_attr_stack() should have added, and the
595 * above loop should have stopped before popping, the
596 * root element whose attr_stack->origin is set to an
597 * empty string.
599 struct strbuf pathbuf = STRBUF_INIT;
601 assert(attr_stack->origin);
602 while (1) {
603 len = strlen(attr_stack->origin);
604 if (dirlen <= len)
605 break;
606 cp = memchr(path + len + 1, '/', dirlen - len - 1);
607 if (!cp)
608 cp = path + dirlen;
609 strbuf_add(&pathbuf, path, cp - path);
610 strbuf_addch(&pathbuf, '/');
611 strbuf_addstr(&pathbuf, GITATTRIBUTES_FILE);
612 elem = read_attr(pathbuf.buf, 0);
613 strbuf_setlen(&pathbuf, cp - path);
614 elem->origin = strbuf_detach(&pathbuf, &elem->originlen);
615 elem->prev = attr_stack;
616 attr_stack = elem;
617 debug_push(elem);
620 strbuf_release(&pathbuf);
624 * Finally push the "info" one at the top of the stack.
626 info->prev = attr_stack;
627 attr_stack = info;
630 static int path_matches(const char *pathname, int pathlen,
631 int basename_offset,
632 const struct pattern *pat,
633 const char *base, int baselen)
635 const char *pattern = pat->pattern;
636 int prefix = pat->nowildcardlen;
637 int isdir = (pathlen && pathname[pathlen - 1] == '/');
639 if ((pat->flags & EXC_FLAG_MUSTBEDIR) && !isdir)
640 return 0;
642 if (pat->flags & EXC_FLAG_NODIR) {
643 return match_basename(pathname + basename_offset,
644 pathlen - basename_offset - isdir,
645 pattern, prefix,
646 pat->patternlen, pat->flags);
648 return match_pathname(pathname, pathlen - isdir,
649 base, baselen,
650 pattern, prefix, pat->patternlen, pat->flags);
653 static int macroexpand_one(int attr_nr, int rem);
655 static int fill_one(const char *what, struct match_attr *a, int rem)
657 struct git_attr_check *check = check_all_attr;
658 int i;
660 for (i = a->num_attr - 1; 0 < rem && 0 <= i; i--) {
661 struct git_attr *attr = a->state[i].attr;
662 const char **n = &(check[attr->attr_nr].value);
663 const char *v = a->state[i].setto;
665 if (*n == ATTR__UNKNOWN) {
666 debug_set(what,
667 a->is_macro ? a->u.attr->name : a->u.pat.pattern,
668 attr, v);
669 *n = v;
670 rem--;
671 rem = macroexpand_one(attr->attr_nr, rem);
674 return rem;
677 static int fill(const char *path, int pathlen, int basename_offset,
678 struct attr_stack *stk, int rem)
680 int i;
681 const char *base = stk->origin ? stk->origin : "";
683 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
684 struct match_attr *a = stk->attrs[i];
685 if (a->is_macro)
686 continue;
687 if (path_matches(path, pathlen, basename_offset,
688 &a->u.pat, base, stk->originlen))
689 rem = fill_one("fill", a, rem);
691 return rem;
694 static int macroexpand_one(int nr, int rem)
696 struct attr_stack *stk;
697 struct match_attr *a = NULL;
698 int i;
700 if (check_all_attr[nr].value != ATTR__TRUE ||
701 !check_all_attr[nr].attr->maybe_macro)
702 return rem;
704 for (stk = attr_stack; !a && stk; stk = stk->prev)
705 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
706 struct match_attr *ma = stk->attrs[i];
707 if (!ma->is_macro)
708 continue;
709 if (ma->u.attr->attr_nr == nr)
710 a = ma;
713 if (a)
714 rem = fill_one("expand", a, rem);
716 return rem;
720 * Collect attributes for path into the array pointed to by
721 * check_all_attr. If num is non-zero, only attributes in check[] are
722 * collected. Otherwise all attributes are collected.
724 static void collect_some_attrs(const char *path, int num,
725 struct git_attr_check *check)
728 struct attr_stack *stk;
729 int i, pathlen, rem, dirlen;
730 const char *cp, *last_slash = NULL;
731 int basename_offset;
733 for (cp = path; *cp; cp++) {
734 if (*cp == '/' && cp[1])
735 last_slash = cp;
737 pathlen = cp - path;
738 if (last_slash) {
739 basename_offset = last_slash + 1 - path;
740 dirlen = last_slash - path;
741 } else {
742 basename_offset = 0;
743 dirlen = 0;
746 prepare_attr_stack(path, dirlen);
747 for (i = 0; i < attr_nr; i++)
748 check_all_attr[i].value = ATTR__UNKNOWN;
749 if (num && !cannot_trust_maybe_real) {
750 rem = 0;
751 for (i = 0; i < num; i++) {
752 if (!check[i].attr->maybe_real) {
753 struct git_attr_check *c;
754 c = check_all_attr + check[i].attr->attr_nr;
755 c->value = ATTR__UNSET;
756 rem++;
759 if (rem == num)
760 return;
763 rem = attr_nr;
764 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
765 rem = fill(path, pathlen, basename_offset, stk, rem);
768 int git_check_attr(const char *path, int num, struct git_attr_check *check)
770 int i;
772 collect_some_attrs(path, num, check);
774 for (i = 0; i < num; i++) {
775 const char *value = check_all_attr[check[i].attr->attr_nr].value;
776 if (value == ATTR__UNKNOWN)
777 value = ATTR__UNSET;
778 check[i].value = value;
781 return 0;
784 int git_all_attrs(const char *path, int *num, struct git_attr_check **check)
786 int i, count, j;
788 collect_some_attrs(path, 0, NULL);
790 /* Count the number of attributes that are set. */
791 count = 0;
792 for (i = 0; i < attr_nr; i++) {
793 const char *value = check_all_attr[i].value;
794 if (value != ATTR__UNSET && value != ATTR__UNKNOWN)
795 ++count;
797 *num = count;
798 *check = xmalloc(sizeof(**check) * count);
799 j = 0;
800 for (i = 0; i < attr_nr; i++) {
801 const char *value = check_all_attr[i].value;
802 if (value != ATTR__UNSET && value != ATTR__UNKNOWN) {
803 (*check)[j].attr = check_all_attr[i].attr;
804 (*check)[j].value = value;
805 ++j;
809 return 0;
812 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
814 enum git_attr_direction old = direction;
816 if (is_bare_repository() && new != GIT_ATTR_INDEX)
817 die("BUG: non-INDEX attr direction in a bare repo");
819 direction = new;
820 if (new != old)
821 drop_attr_stack();
822 use_index = istate;