attr.c: tighten constness around "git_attr" structure
[git.git] / attr.c
blobf7cf7ae3060fc401fe3a2c4c44c24c32915d4601
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"
15 #include "utf8.h"
17 const char git_attr__true[] = "(builtin)true";
18 const char git_attr__false[] = "\0(builtin)false";
19 static const char git_attr__unknown[] = "(builtin)unknown";
20 #define ATTR__TRUE git_attr__true
21 #define ATTR__FALSE git_attr__false
22 #define ATTR__UNSET NULL
23 #define ATTR__UNKNOWN git_attr__unknown
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 int maybe_macro;
37 int maybe_real;
38 char name[FLEX_ARRAY];
40 static int attr_nr;
41 static int cannot_trust_maybe_real;
43 static struct git_attr_check *check_all_attr;
44 static struct git_attr *(git_attr_hash[HASHSIZE]);
46 const char *git_attr_name(const struct git_attr *attr)
48 return attr->name;
51 static unsigned hash_name(const char *name, int namelen)
53 unsigned val = 0, c;
55 while (namelen--) {
56 c = *name++;
57 val = ((val << 7) | (val >> 22)) ^ c;
59 return val;
62 static int invalid_attr_name(const char *name, int namelen)
65 * Attribute name cannot begin with '-' and must consist of
66 * characters from [-A-Za-z0-9_.].
68 if (namelen <= 0 || *name == '-')
69 return -1;
70 while (namelen--) {
71 char ch = *name++;
72 if (! (ch == '-' || ch == '.' || ch == '_' ||
73 ('0' <= ch && ch <= '9') ||
74 ('a' <= ch && ch <= 'z') ||
75 ('A' <= ch && ch <= 'Z')) )
76 return -1;
78 return 0;
81 static struct git_attr *git_attr_internal(const char *name, int len)
83 unsigned hval = hash_name(name, len);
84 unsigned pos = hval % HASHSIZE;
85 struct git_attr *a;
87 for (a = git_attr_hash[pos]; a; a = a->next) {
88 if (a->h == hval &&
89 !memcmp(a->name, name, len) && !a->name[len])
90 return a;
93 if (invalid_attr_name(name, len))
94 return NULL;
96 FLEX_ALLOC_MEM(a, name, name, len);
97 a->h = hval;
98 a->next = git_attr_hash[pos];
99 a->attr_nr = attr_nr++;
100 a->maybe_macro = 0;
101 a->maybe_real = 0;
102 git_attr_hash[pos] = a;
104 REALLOC_ARRAY(check_all_attr, attr_nr);
105 check_all_attr[a->attr_nr].attr = a;
106 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
107 return a;
110 struct git_attr *git_attr(const char *name)
112 return git_attr_internal(name, strlen(name));
115 /* What does a matched pattern decide? */
116 struct attr_state {
117 struct git_attr *attr;
118 const char *setto;
121 struct pattern {
122 const char *pattern;
123 int patternlen;
124 int nowildcardlen;
125 unsigned flags; /* EXC_FLAG_* */
129 * One rule, as from a .gitattributes file.
131 * If is_macro is true, then u.attr is a pointer to the git_attr being
132 * defined.
134 * If is_macro is false, then u.pat is the filename pattern to which the
135 * rule applies.
137 * In either case, num_attr is the number of attributes affected by
138 * this rule, and state is an array listing them. The attributes are
139 * listed as they appear in the file (macros unexpanded).
141 struct match_attr {
142 union {
143 struct pattern pat;
144 struct git_attr *attr;
145 } u;
146 char is_macro;
147 unsigned num_attr;
148 struct attr_state state[FLEX_ARRAY];
151 static const char blank[] = " \t\r\n";
154 * Parse a whitespace-delimited attribute state (i.e., "attr",
155 * "-attr", "!attr", or "attr=value") from the string starting at src.
156 * If e is not NULL, write the results to *e. Return a pointer to the
157 * remainder of the string (with leading whitespace removed), or NULL
158 * if there was an error.
160 static const char *parse_attr(const char *src, int lineno, const char *cp,
161 struct attr_state *e)
163 const char *ep, *equals;
164 int len;
166 ep = cp + strcspn(cp, blank);
167 equals = strchr(cp, '=');
168 if (equals && ep < equals)
169 equals = NULL;
170 if (equals)
171 len = equals - cp;
172 else
173 len = ep - cp;
174 if (!e) {
175 if (*cp == '-' || *cp == '!') {
176 cp++;
177 len--;
179 if (invalid_attr_name(cp, len)) {
180 fprintf(stderr,
181 "%.*s is not a valid attribute name: %s:%d\n",
182 len, cp, src, lineno);
183 return NULL;
185 } else {
187 * As this function is always called twice, once with
188 * e == NULL in the first pass and then e != NULL in
189 * the second pass, no need for invalid_attr_name()
190 * check here.
192 if (*cp == '-' || *cp == '!') {
193 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
194 cp++;
195 len--;
197 else if (!equals)
198 e->setto = ATTR__TRUE;
199 else {
200 e->setto = xmemdupz(equals + 1, ep - equals - 1);
202 e->attr = git_attr_internal(cp, len);
204 return ep + strspn(ep, blank);
207 static struct match_attr *parse_attr_line(const char *line, const char *src,
208 int lineno, int macro_ok)
210 int namelen;
211 int num_attr, i;
212 const char *cp, *name, *states;
213 struct match_attr *res = NULL;
214 int is_macro;
216 cp = line + strspn(line, blank);
217 if (!*cp || *cp == '#')
218 return NULL;
219 name = cp;
220 namelen = strcspn(name, blank);
221 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
222 starts_with(name, ATTRIBUTE_MACRO_PREFIX)) {
223 if (!macro_ok) {
224 fprintf(stderr, "%s not allowed: %s:%d\n",
225 name, src, lineno);
226 return NULL;
228 is_macro = 1;
229 name += strlen(ATTRIBUTE_MACRO_PREFIX);
230 name += strspn(name, blank);
231 namelen = strcspn(name, blank);
232 if (invalid_attr_name(name, namelen)) {
233 fprintf(stderr,
234 "%.*s is not a valid attribute name: %s:%d\n",
235 namelen, name, src, lineno);
236 return NULL;
239 else
240 is_macro = 0;
242 states = name + namelen;
243 states += strspn(states, blank);
245 /* First pass to count the attr_states */
246 for (cp = states, num_attr = 0; *cp; num_attr++) {
247 cp = parse_attr(src, lineno, cp, NULL);
248 if (!cp)
249 return NULL;
252 res = xcalloc(1,
253 sizeof(*res) +
254 sizeof(struct attr_state) * num_attr +
255 (is_macro ? 0 : namelen + 1));
256 if (is_macro) {
257 res->u.attr = git_attr_internal(name, namelen);
258 res->u.attr->maybe_macro = 1;
259 } else {
260 char *p = (char *)&(res->state[num_attr]);
261 memcpy(p, name, namelen);
262 res->u.pat.pattern = p;
263 parse_exclude_pattern(&res->u.pat.pattern,
264 &res->u.pat.patternlen,
265 &res->u.pat.flags,
266 &res->u.pat.nowildcardlen);
267 if (res->u.pat.flags & EXC_FLAG_NEGATIVE) {
268 warning(_("Negative patterns are ignored in git attributes\n"
269 "Use '\\!' for literal leading exclamation."));
270 return NULL;
273 res->is_macro = is_macro;
274 res->num_attr = num_attr;
276 /* Second pass to fill the attr_states */
277 for (cp = states, i = 0; *cp; i++) {
278 cp = parse_attr(src, lineno, cp, &(res->state[i]));
279 if (!is_macro)
280 res->state[i].attr->maybe_real = 1;
281 if (res->state[i].attr->maybe_macro)
282 cannot_trust_maybe_real = 1;
285 return res;
289 * Like info/exclude and .gitignore, the attribute information can
290 * come from many places.
292 * (1) .gitattribute file of the same directory;
293 * (2) .gitattribute file of the parent directory if (1) does not have
294 * any match; this goes recursively upwards, just like .gitignore.
295 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
297 * In the same file, later entries override the earlier match, so in the
298 * global list, we would have entries from info/attributes the earliest
299 * (reading the file from top to bottom), .gitattribute of the root
300 * directory (again, reading the file from top to bottom) down to the
301 * current directory, and then scan the list backwards to find the first match.
302 * This is exactly the same as what is_excluded() does in dir.c to deal with
303 * .gitignore file and info/excludes file as a fallback.
306 static struct attr_stack {
307 struct attr_stack *prev;
308 char *origin;
309 size_t originlen;
310 unsigned num_matches;
311 unsigned alloc;
312 struct match_attr **attrs;
313 } *attr_stack;
315 static void free_attr_elem(struct attr_stack *e)
317 int i;
318 free(e->origin);
319 for (i = 0; i < e->num_matches; i++) {
320 struct match_attr *a = e->attrs[i];
321 int j;
322 for (j = 0; j < a->num_attr; j++) {
323 const char *setto = a->state[j].setto;
324 if (setto == ATTR__TRUE ||
325 setto == ATTR__FALSE ||
326 setto == ATTR__UNSET ||
327 setto == ATTR__UNKNOWN)
329 else
330 free((char *) setto);
332 free(a);
334 free(e->attrs);
335 free(e);
338 static const char *builtin_attr[] = {
339 "[attr]binary -diff -merge -text",
340 NULL,
343 static void handle_attr_line(struct attr_stack *res,
344 const char *line,
345 const char *src,
346 int lineno,
347 int macro_ok)
349 struct match_attr *a;
351 a = parse_attr_line(line, src, lineno, macro_ok);
352 if (!a)
353 return;
354 ALLOC_GROW(res->attrs, res->num_matches + 1, res->alloc);
355 res->attrs[res->num_matches++] = a;
358 static struct attr_stack *read_attr_from_array(const char **list)
360 struct attr_stack *res;
361 const char *line;
362 int lineno = 0;
364 res = xcalloc(1, sizeof(*res));
365 while ((line = *(list++)) != NULL)
366 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
367 return res;
370 static enum git_attr_direction direction;
371 static struct index_state *use_index;
373 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
375 FILE *fp = fopen(path, "r");
376 struct attr_stack *res;
377 char buf[2048];
378 int lineno = 0;
380 if (!fp) {
381 if (errno != ENOENT && errno != ENOTDIR)
382 warn_on_inaccessible(path);
383 return NULL;
385 res = xcalloc(1, sizeof(*res));
386 while (fgets(buf, sizeof(buf), fp)) {
387 char *bufp = buf;
388 if (!lineno)
389 skip_utf8_bom(&bufp, strlen(bufp));
390 handle_attr_line(res, bufp, path, ++lineno, macro_ok);
392 fclose(fp);
393 return res;
396 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
398 struct attr_stack *res;
399 char *buf, *sp;
400 int lineno = 0;
402 buf = read_blob_data_from_index(use_index ? use_index : &the_index, path, NULL);
403 if (!buf)
404 return NULL;
406 res = xcalloc(1, sizeof(*res));
407 for (sp = buf; *sp; ) {
408 char *ep;
409 int more;
411 ep = strchrnul(sp, '\n');
412 more = (*ep == '\n');
413 *ep = '\0';
414 handle_attr_line(res, sp, path, ++lineno, macro_ok);
415 sp = ep + more;
417 free(buf);
418 return res;
421 static struct attr_stack *read_attr(const char *path, int macro_ok)
423 struct attr_stack *res;
425 if (direction == GIT_ATTR_CHECKOUT) {
426 res = read_attr_from_index(path, macro_ok);
427 if (!res)
428 res = read_attr_from_file(path, macro_ok);
430 else if (direction == GIT_ATTR_CHECKIN) {
431 res = read_attr_from_file(path, macro_ok);
432 if (!res)
434 * There is no checked out .gitattributes file there, but
435 * we might have it in the index. We allow operation in a
436 * sparsely checked out work tree, so read from it.
438 res = read_attr_from_index(path, macro_ok);
440 else
441 res = read_attr_from_index(path, macro_ok);
442 if (!res)
443 res = xcalloc(1, sizeof(*res));
444 return res;
447 #if DEBUG_ATTR
448 static void debug_info(const char *what, struct attr_stack *elem)
450 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
452 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
454 const char *value = v;
456 if (ATTR_TRUE(value))
457 value = "set";
458 else if (ATTR_FALSE(value))
459 value = "unset";
460 else if (ATTR_UNSET(value))
461 value = "unspecified";
463 fprintf(stderr, "%s: %s => %s (%s)\n",
464 what, attr->name, (char *) value, match);
466 #define debug_push(a) debug_info("push", (a))
467 #define debug_pop(a) debug_info("pop", (a))
468 #else
469 #define debug_push(a) do { ; } while (0)
470 #define debug_pop(a) do { ; } while (0)
471 #define debug_set(a,b,c,d) do { ; } while (0)
472 #endif /* DEBUG_ATTR */
474 static void drop_attr_stack(void)
476 while (attr_stack) {
477 struct attr_stack *elem = attr_stack;
478 attr_stack = elem->prev;
479 free_attr_elem(elem);
483 static const char *git_etc_gitattributes(void)
485 static const char *system_wide;
486 if (!system_wide)
487 system_wide = system_path(ETC_GITATTRIBUTES);
488 return system_wide;
491 static int git_attr_system(void)
493 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
496 static GIT_PATH_FUNC(git_path_info_attributes, INFOATTRIBUTES_FILE)
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 git_attributes_file = xdg_config_home("attributes");
521 if (git_attributes_file) {
522 elem = read_attr_from_file(git_attributes_file, 1);
523 if (elem) {
524 elem->origin = NULL;
525 elem->prev = attr_stack;
526 attr_stack = elem;
530 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
531 elem = read_attr(GITATTRIBUTES_FILE, 1);
532 elem->origin = xstrdup("");
533 elem->originlen = 0;
534 elem->prev = attr_stack;
535 attr_stack = elem;
536 debug_push(elem);
539 if (startup_info->have_repository)
540 elem = read_attr_from_file(git_path_info_attributes(), 1);
541 else
542 elem = NULL;
544 if (!elem)
545 elem = xcalloc(1, sizeof(*elem));
546 elem->origin = NULL;
547 elem->prev = attr_stack;
548 attr_stack = elem;
551 static void prepare_attr_stack(const char *path, int dirlen)
553 struct attr_stack *elem, *info;
554 int len;
555 const char *cp;
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 int basename_offset,
643 const struct pattern *pat,
644 const char *base, int baselen)
646 const char *pattern = pat->pattern;
647 int prefix = pat->nowildcardlen;
648 int isdir = (pathlen && pathname[pathlen - 1] == '/');
650 if ((pat->flags & EXC_FLAG_MUSTBEDIR) && !isdir)
651 return 0;
653 if (pat->flags & EXC_FLAG_NODIR) {
654 return match_basename(pathname + basename_offset,
655 pathlen - basename_offset - isdir,
656 pattern, prefix,
657 pat->patternlen, pat->flags);
659 return match_pathname(pathname, pathlen - isdir,
660 base, baselen,
661 pattern, prefix, pat->patternlen, pat->flags);
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.pat.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, int basename_offset,
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_offset,
699 &a->u.pat, base, stk->originlen))
700 rem = fill_one("fill", a, rem);
702 return rem;
705 static int macroexpand_one(int nr, int rem)
707 struct attr_stack *stk;
708 int i;
710 if (check_all_attr[nr].value != ATTR__TRUE ||
711 !check_all_attr[nr].attr->maybe_macro)
712 return rem;
714 for (stk = attr_stack; stk; stk = stk->prev) {
715 for (i = stk->num_matches - 1; 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 == nr)
720 return fill_one("expand", ma, rem);
724 return rem;
728 * Collect attributes for path into the array pointed to by
729 * check_all_attr. If num is non-zero, only attributes in check[] are
730 * collected. Otherwise all attributes are collected.
732 static void collect_some_attrs(const char *path, int num,
733 struct git_attr_check *check)
736 struct attr_stack *stk;
737 int i, pathlen, rem, dirlen;
738 const char *cp, *last_slash = NULL;
739 int basename_offset;
741 for (cp = path; *cp; cp++) {
742 if (*cp == '/' && cp[1])
743 last_slash = cp;
745 pathlen = cp - path;
746 if (last_slash) {
747 basename_offset = last_slash + 1 - path;
748 dirlen = last_slash - path;
749 } else {
750 basename_offset = 0;
751 dirlen = 0;
754 prepare_attr_stack(path, dirlen);
755 for (i = 0; i < attr_nr; i++)
756 check_all_attr[i].value = ATTR__UNKNOWN;
757 if (num && !cannot_trust_maybe_real) {
758 rem = 0;
759 for (i = 0; i < num; i++) {
760 if (!check[i].attr->maybe_real) {
761 struct git_attr_check *c;
762 c = check_all_attr + check[i].attr->attr_nr;
763 c->value = ATTR__UNSET;
764 rem++;
767 if (rem == num)
768 return;
771 rem = attr_nr;
772 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
773 rem = fill(path, pathlen, basename_offset, stk, rem);
776 int git_check_attr(const char *path, int num, struct git_attr_check *check)
778 int i;
780 collect_some_attrs(path, num, check);
782 for (i = 0; i < num; i++) {
783 const char *value = check_all_attr[check[i].attr->attr_nr].value;
784 if (value == ATTR__UNKNOWN)
785 value = ATTR__UNSET;
786 check[i].value = value;
789 return 0;
792 int git_all_attrs(const char *path, int *num, struct git_attr_check **check)
794 int i, count, j;
796 collect_some_attrs(path, 0, NULL);
798 /* Count the number of attributes that are set. */
799 count = 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 ++count;
805 *num = count;
806 ALLOC_ARRAY(*check, count);
807 j = 0;
808 for (i = 0; i < attr_nr; i++) {
809 const char *value = check_all_attr[i].value;
810 if (value != ATTR__UNSET && value != ATTR__UNKNOWN) {
811 (*check)[j].attr = check_all_attr[i].attr;
812 (*check)[j].value = value;
813 ++j;
817 return 0;
820 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
822 enum git_attr_direction old = direction;
824 if (is_bare_repository() && new != GIT_ATTR_INDEX)
825 die("BUG: non-INDEX attr direction in a bare repo");
827 direction = new;
828 if (new != old)
829 drop_attr_stack();
830 use_index = istate;