shortlog: use strbufs to read from stdin
[git/gitweb.git] / attr.c
blob086c08dcfab613ef2927ff2769797fa40cd03308
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 char *git_attr_name(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 a = xmalloc(sizeof(*a) + len + 1);
97 memcpy(a->name, name, len);
98 a->name[len] = 0;
99 a->h = hval;
100 a->next = git_attr_hash[pos];
101 a->attr_nr = attr_nr++;
102 a->maybe_macro = 0;
103 a->maybe_real = 0;
104 git_attr_hash[pos] = a;
106 REALLOC_ARRAY(check_all_attr, attr_nr);
107 check_all_attr[a->attr_nr].attr = a;
108 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
109 return a;
112 struct git_attr *git_attr(const char *name)
114 return git_attr_internal(name, strlen(name));
117 /* What does a matched pattern decide? */
118 struct attr_state {
119 struct git_attr *attr;
120 const char *setto;
123 struct pattern {
124 const char *pattern;
125 int patternlen;
126 int nowildcardlen;
127 int flags; /* EXC_FLAG_* */
131 * One rule, as from a .gitattributes file.
133 * If is_macro is true, then u.attr is a pointer to the git_attr being
134 * defined.
136 * If is_macro is false, then u.pattern points at the filename pattern
137 * to which the rule applies. (The memory pointed to is part of the
138 * memory block allocated for the match_attr instance.)
140 * In either case, num_attr is the number of attributes affected by
141 * this rule, and state is an array listing them. The attributes are
142 * listed as they appear in the file (macros unexpanded).
144 struct match_attr {
145 union {
146 struct pattern pat;
147 struct git_attr *attr;
148 } u;
149 char is_macro;
150 unsigned num_attr;
151 struct attr_state state[FLEX_ARRAY];
154 static const char blank[] = " \t\r\n";
157 * Parse a whitespace-delimited attribute state (i.e., "attr",
158 * "-attr", "!attr", or "attr=value") from the string starting at src.
159 * If e is not NULL, write the results to *e. Return a pointer to the
160 * remainder of the string (with leading whitespace removed), or NULL
161 * if there was an error.
163 static const char *parse_attr(const char *src, int lineno, const char *cp,
164 struct attr_state *e)
166 const char *ep, *equals;
167 int len;
169 ep = cp + strcspn(cp, blank);
170 equals = strchr(cp, '=');
171 if (equals && ep < equals)
172 equals = NULL;
173 if (equals)
174 len = equals - cp;
175 else
176 len = ep - cp;
177 if (!e) {
178 if (*cp == '-' || *cp == '!') {
179 cp++;
180 len--;
182 if (invalid_attr_name(cp, len)) {
183 fprintf(stderr,
184 "%.*s is not a valid attribute name: %s:%d\n",
185 len, cp, src, lineno);
186 return NULL;
188 } else {
189 if (*cp == '-' || *cp == '!') {
190 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
191 cp++;
192 len--;
194 else if (!equals)
195 e->setto = ATTR__TRUE;
196 else {
197 e->setto = xmemdupz(equals + 1, ep - equals - 1);
199 e->attr = git_attr_internal(cp, len);
201 return ep + strspn(ep, blank);
204 static struct match_attr *parse_attr_line(const char *line, const char *src,
205 int lineno, int macro_ok)
207 int namelen;
208 int num_attr, i;
209 const char *cp, *name, *states;
210 struct match_attr *res = NULL;
211 int is_macro;
213 cp = line + strspn(line, blank);
214 if (!*cp || *cp == '#')
215 return NULL;
216 name = cp;
217 namelen = strcspn(name, blank);
218 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
219 starts_with(name, ATTRIBUTE_MACRO_PREFIX)) {
220 if (!macro_ok) {
221 fprintf(stderr, "%s not allowed: %s:%d\n",
222 name, src, lineno);
223 return NULL;
225 is_macro = 1;
226 name += strlen(ATTRIBUTE_MACRO_PREFIX);
227 name += strspn(name, blank);
228 namelen = strcspn(name, blank);
229 if (invalid_attr_name(name, namelen)) {
230 fprintf(stderr,
231 "%.*s is not a valid attribute name: %s:%d\n",
232 namelen, name, src, lineno);
233 return NULL;
236 else
237 is_macro = 0;
239 states = name + namelen;
240 states += strspn(states, blank);
242 /* First pass to count the attr_states */
243 for (cp = states, num_attr = 0; *cp; num_attr++) {
244 cp = parse_attr(src, lineno, cp, NULL);
245 if (!cp)
246 return NULL;
249 res = xcalloc(1,
250 sizeof(*res) +
251 sizeof(struct attr_state) * num_attr +
252 (is_macro ? 0 : namelen + 1));
253 if (is_macro) {
254 res->u.attr = git_attr_internal(name, namelen);
255 res->u.attr->maybe_macro = 1;
256 } else {
257 char *p = (char *)&(res->state[num_attr]);
258 memcpy(p, name, namelen);
259 res->u.pat.pattern = p;
260 parse_exclude_pattern(&res->u.pat.pattern,
261 &res->u.pat.patternlen,
262 &res->u.pat.flags,
263 &res->u.pat.nowildcardlen);
264 if (res->u.pat.flags & EXC_FLAG_NEGATIVE) {
265 warning(_("Negative patterns are ignored in git attributes\n"
266 "Use '\\!' for literal leading exclamation."));
267 return NULL;
270 res->is_macro = is_macro;
271 res->num_attr = num_attr;
273 /* Second pass to fill the attr_states */
274 for (cp = states, i = 0; *cp; i++) {
275 cp = parse_attr(src, lineno, cp, &(res->state[i]));
276 if (!is_macro)
277 res->state[i].attr->maybe_real = 1;
278 if (res->state[i].attr->maybe_macro)
279 cannot_trust_maybe_real = 1;
282 return res;
286 * Like info/exclude and .gitignore, the attribute information can
287 * come from many places.
289 * (1) .gitattribute file of the same directory;
290 * (2) .gitattribute file of the parent directory if (1) does not have
291 * any match; this goes recursively upwards, just like .gitignore.
292 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
294 * In the same file, later entries override the earlier match, so in the
295 * global list, we would have entries from info/attributes the earliest
296 * (reading the file from top to bottom), .gitattribute of the root
297 * directory (again, reading the file from top to bottom) down to the
298 * current directory, and then scan the list backwards to find the first match.
299 * This is exactly the same as what is_excluded() does in dir.c to deal with
300 * .gitignore
303 static struct attr_stack {
304 struct attr_stack *prev;
305 char *origin;
306 size_t originlen;
307 unsigned num_matches;
308 unsigned alloc;
309 struct match_attr **attrs;
310 } *attr_stack;
312 static void free_attr_elem(struct attr_stack *e)
314 int i;
315 free(e->origin);
316 for (i = 0; i < e->num_matches; i++) {
317 struct match_attr *a = e->attrs[i];
318 int j;
319 for (j = 0; j < a->num_attr; j++) {
320 const char *setto = a->state[j].setto;
321 if (setto == ATTR__TRUE ||
322 setto == ATTR__FALSE ||
323 setto == ATTR__UNSET ||
324 setto == ATTR__UNKNOWN)
326 else
327 free((char *) setto);
329 free(a);
331 free(e->attrs);
332 free(e);
335 static const char *builtin_attr[] = {
336 "[attr]binary -diff -merge -text",
337 NULL,
340 static void handle_attr_line(struct attr_stack *res,
341 const char *line,
342 const char *src,
343 int lineno,
344 int macro_ok)
346 struct match_attr *a;
348 a = parse_attr_line(line, src, lineno, macro_ok);
349 if (!a)
350 return;
351 ALLOC_GROW(res->attrs, res->num_matches + 1, res->alloc);
352 res->attrs[res->num_matches++] = a;
355 static struct attr_stack *read_attr_from_array(const char **list)
357 struct attr_stack *res;
358 const char *line;
359 int lineno = 0;
361 res = xcalloc(1, sizeof(*res));
362 while ((line = *(list++)) != NULL)
363 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
364 return res;
367 static enum git_attr_direction direction;
368 static struct index_state *use_index;
370 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
372 FILE *fp = fopen(path, "r");
373 struct attr_stack *res;
374 char buf[2048];
375 int lineno = 0;
377 if (!fp) {
378 if (errno != ENOENT && errno != ENOTDIR)
379 warn_on_inaccessible(path);
380 return NULL;
382 res = xcalloc(1, sizeof(*res));
383 while (fgets(buf, sizeof(buf), fp)) {
384 char *bufp = buf;
385 if (!lineno)
386 skip_utf8_bom(&bufp, strlen(bufp));
387 handle_attr_line(res, bufp, path, ++lineno, macro_ok);
389 fclose(fp);
390 return res;
393 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
395 struct attr_stack *res;
396 char *buf, *sp;
397 int lineno = 0;
399 buf = read_blob_data_from_index(use_index ? use_index : &the_index, path, NULL);
400 if (!buf)
401 return NULL;
403 res = xcalloc(1, sizeof(*res));
404 for (sp = buf; *sp; ) {
405 char *ep;
406 int more;
407 for (ep = sp; *ep && *ep != '\n'; ep++)
409 more = (*ep == '\n');
410 *ep = '\0';
411 handle_attr_line(res, sp, path, ++lineno, macro_ok);
412 sp = ep + more;
414 free(buf);
415 return res;
418 static struct attr_stack *read_attr(const char *path, int macro_ok)
420 struct attr_stack *res;
422 if (direction == GIT_ATTR_CHECKOUT) {
423 res = read_attr_from_index(path, macro_ok);
424 if (!res)
425 res = read_attr_from_file(path, macro_ok);
427 else if (direction == GIT_ATTR_CHECKIN) {
428 res = read_attr_from_file(path, macro_ok);
429 if (!res)
431 * There is no checked out .gitattributes file there, but
432 * we might have it in the index. We allow operation in a
433 * sparsely checked out work tree, so read from it.
435 res = read_attr_from_index(path, macro_ok);
437 else
438 res = read_attr_from_index(path, macro_ok);
439 if (!res)
440 res = xcalloc(1, sizeof(*res));
441 return res;
444 #if DEBUG_ATTR
445 static void debug_info(const char *what, struct attr_stack *elem)
447 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
449 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
451 const char *value = v;
453 if (ATTR_TRUE(value))
454 value = "set";
455 else if (ATTR_FALSE(value))
456 value = "unset";
457 else if (ATTR_UNSET(value))
458 value = "unspecified";
460 fprintf(stderr, "%s: %s => %s (%s)\n",
461 what, attr->name, (char *) value, match);
463 #define debug_push(a) debug_info("push", (a))
464 #define debug_pop(a) debug_info("pop", (a))
465 #else
466 #define debug_push(a) do { ; } while (0)
467 #define debug_pop(a) do { ; } while (0)
468 #define debug_set(a,b,c,d) do { ; } while (0)
469 #endif
471 static void drop_attr_stack(void)
473 while (attr_stack) {
474 struct attr_stack *elem = attr_stack;
475 attr_stack = elem->prev;
476 free_attr_elem(elem);
480 static const char *git_etc_gitattributes(void)
482 static const char *system_wide;
483 if (!system_wide)
484 system_wide = system_path(ETC_GITATTRIBUTES);
485 return system_wide;
488 static int git_attr_system(void)
490 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
493 static GIT_PATH_FUNC(git_path_info_attributes, INFOATTRIBUTES_FILE)
495 static void bootstrap_attr_stack(void)
497 struct attr_stack *elem;
499 if (attr_stack)
500 return;
502 elem = read_attr_from_array(builtin_attr);
503 elem->origin = NULL;
504 elem->prev = attr_stack;
505 attr_stack = elem;
507 if (git_attr_system()) {
508 elem = read_attr_from_file(git_etc_gitattributes(), 1);
509 if (elem) {
510 elem->origin = NULL;
511 elem->prev = attr_stack;
512 attr_stack = elem;
516 if (!git_attributes_file)
517 git_attributes_file = xdg_config_home("attributes");
518 if (git_attributes_file) {
519 elem = read_attr_from_file(git_attributes_file, 1);
520 if (elem) {
521 elem->origin = NULL;
522 elem->prev = attr_stack;
523 attr_stack = elem;
527 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
528 elem = read_attr(GITATTRIBUTES_FILE, 1);
529 elem->origin = xstrdup("");
530 elem->originlen = 0;
531 elem->prev = attr_stack;
532 attr_stack = elem;
533 debug_push(elem);
536 elem = read_attr_from_file(git_path_info_attributes(), 1);
537 if (!elem)
538 elem = xcalloc(1, sizeof(*elem));
539 elem->origin = NULL;
540 elem->prev = attr_stack;
541 attr_stack = elem;
544 static void prepare_attr_stack(const char *path, int dirlen)
546 struct attr_stack *elem, *info;
547 int len;
548 const char *cp;
551 * At the bottom of the attribute stack is the built-in
552 * set of attribute definitions, followed by the contents
553 * of $(prefix)/etc/gitattributes and a file specified by
554 * core.attributesfile. Then, contents from
555 * .gitattribute files from directories closer to the
556 * root to the ones in deeper directories are pushed
557 * to the stack. Finally, at the very top of the stack
558 * we always keep the contents of $GIT_DIR/info/attributes.
560 * When checking, we use entries from near the top of the
561 * stack, preferring $GIT_DIR/info/attributes, then
562 * .gitattributes in deeper directories to shallower ones,
563 * and finally use the built-in set as the default.
565 bootstrap_attr_stack();
568 * Pop the "info" one that is always at the top of the stack.
570 info = attr_stack;
571 attr_stack = info->prev;
574 * Pop the ones from directories that are not the prefix of
575 * the path we are checking. Break out of the loop when we see
576 * the root one (whose origin is an empty string "") or the builtin
577 * one (whose origin is NULL) without popping it.
579 while (attr_stack->origin) {
580 int namelen = strlen(attr_stack->origin);
582 elem = attr_stack;
583 if (namelen <= dirlen &&
584 !strncmp(elem->origin, path, namelen) &&
585 (!namelen || path[namelen] == '/'))
586 break;
588 debug_pop(elem);
589 attr_stack = elem->prev;
590 free_attr_elem(elem);
594 * Read from parent directories and push them down
596 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
598 * bootstrap_attr_stack() should have added, and the
599 * above loop should have stopped before popping, the
600 * root element whose attr_stack->origin is set to an
601 * empty string.
603 struct strbuf pathbuf = STRBUF_INIT;
605 assert(attr_stack->origin);
606 while (1) {
607 len = strlen(attr_stack->origin);
608 if (dirlen <= len)
609 break;
610 cp = memchr(path + len + 1, '/', dirlen - len - 1);
611 if (!cp)
612 cp = path + dirlen;
613 strbuf_add(&pathbuf, path, cp - path);
614 strbuf_addch(&pathbuf, '/');
615 strbuf_addstr(&pathbuf, GITATTRIBUTES_FILE);
616 elem = read_attr(pathbuf.buf, 0);
617 strbuf_setlen(&pathbuf, cp - path);
618 elem->origin = strbuf_detach(&pathbuf, &elem->originlen);
619 elem->prev = attr_stack;
620 attr_stack = elem;
621 debug_push(elem);
624 strbuf_release(&pathbuf);
628 * Finally push the "info" one at the top of the stack.
630 info->prev = attr_stack;
631 attr_stack = info;
634 static int path_matches(const char *pathname, int pathlen,
635 int basename_offset,
636 const struct pattern *pat,
637 const char *base, int baselen)
639 const char *pattern = pat->pattern;
640 int prefix = pat->nowildcardlen;
641 int isdir = (pathlen && pathname[pathlen - 1] == '/');
643 if ((pat->flags & EXC_FLAG_MUSTBEDIR) && !isdir)
644 return 0;
646 if (pat->flags & EXC_FLAG_NODIR) {
647 return match_basename(pathname + basename_offset,
648 pathlen - basename_offset - isdir,
649 pattern, prefix,
650 pat->patternlen, pat->flags);
652 return match_pathname(pathname, pathlen - isdir,
653 base, baselen,
654 pattern, prefix, pat->patternlen, pat->flags);
657 static int macroexpand_one(int attr_nr, int rem);
659 static int fill_one(const char *what, struct match_attr *a, int rem)
661 struct git_attr_check *check = check_all_attr;
662 int i;
664 for (i = a->num_attr - 1; 0 < rem && 0 <= i; i--) {
665 struct git_attr *attr = a->state[i].attr;
666 const char **n = &(check[attr->attr_nr].value);
667 const char *v = a->state[i].setto;
669 if (*n == ATTR__UNKNOWN) {
670 debug_set(what,
671 a->is_macro ? a->u.attr->name : a->u.pat.pattern,
672 attr, v);
673 *n = v;
674 rem--;
675 rem = macroexpand_one(attr->attr_nr, rem);
678 return rem;
681 static int fill(const char *path, int pathlen, int basename_offset,
682 struct attr_stack *stk, int rem)
684 int i;
685 const char *base = stk->origin ? stk->origin : "";
687 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
688 struct match_attr *a = stk->attrs[i];
689 if (a->is_macro)
690 continue;
691 if (path_matches(path, pathlen, basename_offset,
692 &a->u.pat, base, stk->originlen))
693 rem = fill_one("fill", a, rem);
695 return rem;
698 static int macroexpand_one(int nr, int rem)
700 struct attr_stack *stk;
701 struct match_attr *a = NULL;
702 int i;
704 if (check_all_attr[nr].value != ATTR__TRUE ||
705 !check_all_attr[nr].attr->maybe_macro)
706 return rem;
708 for (stk = attr_stack; !a && stk; stk = stk->prev)
709 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
710 struct match_attr *ma = stk->attrs[i];
711 if (!ma->is_macro)
712 continue;
713 if (ma->u.attr->attr_nr == nr)
714 a = ma;
717 if (a)
718 rem = fill_one("expand", a, rem);
720 return rem;
724 * Collect attributes for path into the array pointed to by
725 * check_all_attr. If num is non-zero, only attributes in check[] are
726 * collected. Otherwise all attributes are collected.
728 static void collect_some_attrs(const char *path, int num,
729 struct git_attr_check *check)
732 struct attr_stack *stk;
733 int i, pathlen, rem, dirlen;
734 const char *cp, *last_slash = NULL;
735 int basename_offset;
737 for (cp = path; *cp; cp++) {
738 if (*cp == '/' && cp[1])
739 last_slash = cp;
741 pathlen = cp - path;
742 if (last_slash) {
743 basename_offset = last_slash + 1 - path;
744 dirlen = last_slash - path;
745 } else {
746 basename_offset = 0;
747 dirlen = 0;
750 prepare_attr_stack(path, dirlen);
751 for (i = 0; i < attr_nr; i++)
752 check_all_attr[i].value = ATTR__UNKNOWN;
753 if (num && !cannot_trust_maybe_real) {
754 rem = 0;
755 for (i = 0; i < num; i++) {
756 if (!check[i].attr->maybe_real) {
757 struct git_attr_check *c;
758 c = check_all_attr + check[i].attr->attr_nr;
759 c->value = ATTR__UNSET;
760 rem++;
763 if (rem == num)
764 return;
767 rem = attr_nr;
768 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
769 rem = fill(path, pathlen, basename_offset, stk, rem);
772 int git_check_attr(const char *path, int num, struct git_attr_check *check)
774 int i;
776 collect_some_attrs(path, num, check);
778 for (i = 0; i < num; i++) {
779 const char *value = check_all_attr[check[i].attr->attr_nr].value;
780 if (value == ATTR__UNKNOWN)
781 value = ATTR__UNSET;
782 check[i].value = value;
785 return 0;
788 int git_all_attrs(const char *path, int *num, struct git_attr_check **check)
790 int i, count, j;
792 collect_some_attrs(path, 0, NULL);
794 /* Count the number of attributes that are set. */
795 count = 0;
796 for (i = 0; i < attr_nr; i++) {
797 const char *value = check_all_attr[i].value;
798 if (value != ATTR__UNSET && value != ATTR__UNKNOWN)
799 ++count;
801 *num = count;
802 *check = xmalloc(sizeof(**check) * count);
803 j = 0;
804 for (i = 0; i < attr_nr; i++) {
805 const char *value = check_all_attr[i].value;
806 if (value != ATTR__UNSET && value != ATTR__UNKNOWN) {
807 (*check)[j].attr = check_all_attr[i].attr;
808 (*check)[j].value = value;
809 ++j;
813 return 0;
816 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
818 enum git_attr_direction old = direction;
820 if (is_bare_repository() && new != GIT_ATTR_INDEX)
821 die("BUG: non-INDEX attr direction in a bare repo");
823 direction = new;
824 if (new != old)
825 drop_attr_stack();
826 use_index = istate;