Interested in helping open source friends on HP-UX?
[git/mjg.git] / attr.c
blobcd5469770a6f7841ae327e6ba7141dd99d9cacc1
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 REALLOC_ARRAY(check_all_attr, attr_nr);
101 check_all_attr[a->attr_nr].attr = a;
102 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
103 return a;
106 struct git_attr *git_attr(const char *name)
108 return git_attr_internal(name, strlen(name));
111 /* What does a matched pattern decide? */
112 struct attr_state {
113 struct git_attr *attr;
114 const char *setto;
117 struct pattern {
118 const char *pattern;
119 int patternlen;
120 int nowildcardlen;
121 int flags; /* EXC_FLAG_* */
125 * One rule, as from a .gitattributes file.
127 * If is_macro is true, then u.attr is a pointer to the git_attr being
128 * defined.
130 * If is_macro is false, then u.pattern points at the filename pattern
131 * to which the rule applies. (The memory pointed to is part of the
132 * memory block allocated for the match_attr instance.)
134 * In either case, num_attr is the number of attributes affected by
135 * this rule, and state is an array listing them. The attributes are
136 * listed as they appear in the file (macros unexpanded).
138 struct match_attr {
139 union {
140 struct pattern pat;
141 struct git_attr *attr;
142 } u;
143 char is_macro;
144 unsigned num_attr;
145 struct attr_state state[FLEX_ARRAY];
148 static const char blank[] = " \t\r\n";
151 * Parse a whitespace-delimited attribute state (i.e., "attr",
152 * "-attr", "!attr", or "attr=value") from the string starting at src.
153 * If e is not NULL, write the results to *e. Return a pointer to the
154 * remainder of the string (with leading whitespace removed), or NULL
155 * if there was an error.
157 static const char *parse_attr(const char *src, int lineno, const char *cp,
158 struct attr_state *e)
160 const char *ep, *equals;
161 int len;
163 ep = cp + strcspn(cp, blank);
164 equals = strchr(cp, '=');
165 if (equals && ep < equals)
166 equals = NULL;
167 if (equals)
168 len = equals - cp;
169 else
170 len = ep - cp;
171 if (!e) {
172 if (*cp == '-' || *cp == '!') {
173 cp++;
174 len--;
176 if (invalid_attr_name(cp, len)) {
177 fprintf(stderr,
178 "%.*s is not a valid attribute name: %s:%d\n",
179 len, cp, src, lineno);
180 return NULL;
182 } else {
183 if (*cp == '-' || *cp == '!') {
184 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
185 cp++;
186 len--;
188 else if (!equals)
189 e->setto = ATTR__TRUE;
190 else {
191 e->setto = xmemdupz(equals + 1, ep - equals - 1);
193 e->attr = git_attr_internal(cp, len);
195 return ep + strspn(ep, blank);
198 static struct match_attr *parse_attr_line(const char *line, const char *src,
199 int lineno, int macro_ok)
201 int namelen;
202 int num_attr, i;
203 const char *cp, *name, *states;
204 struct match_attr *res = NULL;
205 int is_macro;
207 cp = line + strspn(line, blank);
208 if (!*cp || *cp == '#')
209 return NULL;
210 name = cp;
211 namelen = strcspn(name, blank);
212 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
213 starts_with(name, ATTRIBUTE_MACRO_PREFIX)) {
214 if (!macro_ok) {
215 fprintf(stderr, "%s not allowed: %s:%d\n",
216 name, src, lineno);
217 return NULL;
219 is_macro = 1;
220 name += strlen(ATTRIBUTE_MACRO_PREFIX);
221 name += strspn(name, blank);
222 namelen = strcspn(name, blank);
223 if (invalid_attr_name(name, namelen)) {
224 fprintf(stderr,
225 "%.*s is not a valid attribute name: %s:%d\n",
226 namelen, name, src, lineno);
227 return NULL;
230 else
231 is_macro = 0;
233 states = name + namelen;
234 states += strspn(states, blank);
236 /* First pass to count the attr_states */
237 for (cp = states, num_attr = 0; *cp; num_attr++) {
238 cp = parse_attr(src, lineno, cp, NULL);
239 if (!cp)
240 return NULL;
243 res = xcalloc(1,
244 sizeof(*res) +
245 sizeof(struct attr_state) * num_attr +
246 (is_macro ? 0 : namelen + 1));
247 if (is_macro)
248 res->u.attr = git_attr_internal(name, namelen);
249 else {
250 char *p = (char *)&(res->state[num_attr]);
251 memcpy(p, name, namelen);
252 res->u.pat.pattern = p;
253 parse_exclude_pattern(&res->u.pat.pattern,
254 &res->u.pat.patternlen,
255 &res->u.pat.flags,
256 &res->u.pat.nowildcardlen);
257 if (res->u.pat.flags & EXC_FLAG_NEGATIVE) {
258 warning(_("Negative patterns are ignored in git attributes\n"
259 "Use '\\!' for literal leading exclamation."));
260 return NULL;
263 res->is_macro = is_macro;
264 res->num_attr = num_attr;
266 /* Second pass to fill the attr_states */
267 for (cp = states, i = 0; *cp; i++) {
268 cp = parse_attr(src, lineno, cp, &(res->state[i]));
271 return res;
275 * Like info/exclude and .gitignore, the attribute information can
276 * come from many places.
278 * (1) .gitattribute file of the same directory;
279 * (2) .gitattribute file of the parent directory if (1) does not have
280 * any match; this goes recursively upwards, just like .gitignore.
281 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
283 * In the same file, later entries override the earlier match, so in the
284 * global list, we would have entries from info/attributes the earliest
285 * (reading the file from top to bottom), .gitattribute of the root
286 * directory (again, reading the file from top to bottom) down to the
287 * current directory, and then scan the list backwards to find the first match.
288 * This is exactly the same as what is_excluded() does in dir.c to deal with
289 * .gitignore
292 static struct attr_stack {
293 struct attr_stack *prev;
294 char *origin;
295 size_t originlen;
296 unsigned num_matches;
297 unsigned alloc;
298 struct match_attr **attrs;
299 } *attr_stack;
301 static void free_attr_elem(struct attr_stack *e)
303 int i;
304 free(e->origin);
305 for (i = 0; i < e->num_matches; i++) {
306 struct match_attr *a = e->attrs[i];
307 int j;
308 for (j = 0; j < a->num_attr; j++) {
309 const char *setto = a->state[j].setto;
310 if (setto == ATTR__TRUE ||
311 setto == ATTR__FALSE ||
312 setto == ATTR__UNSET ||
313 setto == ATTR__UNKNOWN)
315 else
316 free((char *) setto);
318 free(a);
320 free(e->attrs);
321 free(e);
324 static const char *builtin_attr[] = {
325 "[attr]binary -diff -merge -text",
326 NULL,
329 static void handle_attr_line(struct attr_stack *res,
330 const char *line,
331 const char *src,
332 int lineno,
333 int macro_ok)
335 struct match_attr *a;
337 a = parse_attr_line(line, src, lineno, macro_ok);
338 if (!a)
339 return;
340 ALLOC_GROW(res->attrs, res->num_matches + 1, res->alloc);
341 res->attrs[res->num_matches++] = a;
344 static struct attr_stack *read_attr_from_array(const char **list)
346 struct attr_stack *res;
347 const char *line;
348 int lineno = 0;
350 res = xcalloc(1, sizeof(*res));
351 while ((line = *(list++)) != NULL)
352 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
353 return res;
356 static enum git_attr_direction direction;
357 static struct index_state *use_index;
359 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
361 FILE *fp = fopen(path, "r");
362 struct attr_stack *res;
363 char buf[2048];
364 int lineno = 0;
366 if (!fp) {
367 if (errno != ENOENT && errno != ENOTDIR)
368 warn_on_inaccessible(path);
369 return NULL;
371 res = xcalloc(1, sizeof(*res));
372 while (fgets(buf, sizeof(buf), fp))
373 handle_attr_line(res, buf, path, ++lineno, macro_ok);
374 fclose(fp);
375 return res;
378 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
380 struct attr_stack *res;
381 char *buf, *sp;
382 int lineno = 0;
384 buf = read_blob_data_from_index(use_index ? use_index : &the_index, path, NULL);
385 if (!buf)
386 return NULL;
388 res = xcalloc(1, sizeof(*res));
389 for (sp = buf; *sp; ) {
390 char *ep;
391 int more;
392 for (ep = sp; *ep && *ep != '\n'; ep++)
394 more = (*ep == '\n');
395 *ep = '\0';
396 handle_attr_line(res, sp, path, ++lineno, macro_ok);
397 sp = ep + more;
399 free(buf);
400 return res;
403 static struct attr_stack *read_attr(const char *path, int macro_ok)
405 struct attr_stack *res;
407 if (direction == GIT_ATTR_CHECKOUT) {
408 res = read_attr_from_index(path, macro_ok);
409 if (!res)
410 res = read_attr_from_file(path, macro_ok);
412 else if (direction == GIT_ATTR_CHECKIN) {
413 res = read_attr_from_file(path, macro_ok);
414 if (!res)
416 * There is no checked out .gitattributes file there, but
417 * we might have it in the index. We allow operation in a
418 * sparsely checked out work tree, so read from it.
420 res = read_attr_from_index(path, macro_ok);
422 else
423 res = read_attr_from_index(path, macro_ok);
424 if (!res)
425 res = xcalloc(1, sizeof(*res));
426 return res;
429 #if DEBUG_ATTR
430 static void debug_info(const char *what, struct attr_stack *elem)
432 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
434 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
436 const char *value = v;
438 if (ATTR_TRUE(value))
439 value = "set";
440 else if (ATTR_FALSE(value))
441 value = "unset";
442 else if (ATTR_UNSET(value))
443 value = "unspecified";
445 fprintf(stderr, "%s: %s => %s (%s)\n",
446 what, attr->name, (char *) value, match);
448 #define debug_push(a) debug_info("push", (a))
449 #define debug_pop(a) debug_info("pop", (a))
450 #else
451 #define debug_push(a) do { ; } while (0)
452 #define debug_pop(a) do { ; } while (0)
453 #define debug_set(a,b,c,d) do { ; } while (0)
454 #endif
456 static void drop_attr_stack(void)
458 while (attr_stack) {
459 struct attr_stack *elem = attr_stack;
460 attr_stack = elem->prev;
461 free_attr_elem(elem);
465 static const char *git_etc_gitattributes(void)
467 static const char *system_wide;
468 if (!system_wide)
469 system_wide = system_path(ETC_GITATTRIBUTES);
470 return system_wide;
473 static int git_attr_system(void)
475 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
478 static void bootstrap_attr_stack(void)
480 struct attr_stack *elem;
481 char *xdg_attributes_file;
483 if (attr_stack)
484 return;
486 elem = read_attr_from_array(builtin_attr);
487 elem->origin = NULL;
488 elem->prev = attr_stack;
489 attr_stack = elem;
491 if (git_attr_system()) {
492 elem = read_attr_from_file(git_etc_gitattributes(), 1);
493 if (elem) {
494 elem->origin = NULL;
495 elem->prev = attr_stack;
496 attr_stack = elem;
500 if (!git_attributes_file) {
501 home_config_paths(NULL, &xdg_attributes_file, "attributes");
502 git_attributes_file = xdg_attributes_file;
504 if (git_attributes_file) {
505 elem = read_attr_from_file(git_attributes_file, 1);
506 if (elem) {
507 elem->origin = NULL;
508 elem->prev = attr_stack;
509 attr_stack = elem;
513 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
514 elem = read_attr(GITATTRIBUTES_FILE, 1);
515 elem->origin = xstrdup("");
516 elem->originlen = 0;
517 elem->prev = attr_stack;
518 attr_stack = elem;
519 debug_push(elem);
522 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
523 if (!elem)
524 elem = xcalloc(1, sizeof(*elem));
525 elem->origin = NULL;
526 elem->prev = attr_stack;
527 attr_stack = elem;
530 static void prepare_attr_stack(const char *path, int dirlen)
532 struct attr_stack *elem, *info;
533 int len;
534 const char *cp;
537 * At the bottom of the attribute stack is the built-in
538 * set of attribute definitions, followed by the contents
539 * of $(prefix)/etc/gitattributes and a file specified by
540 * core.attributesfile. Then, contents from
541 * .gitattribute files from directories closer to the
542 * root to the ones in deeper directories are pushed
543 * to the stack. Finally, at the very top of the stack
544 * we always keep the contents of $GIT_DIR/info/attributes.
546 * When checking, we use entries from near the top of the
547 * stack, preferring $GIT_DIR/info/attributes, then
548 * .gitattributes in deeper directories to shallower ones,
549 * and finally use the built-in set as the default.
551 bootstrap_attr_stack();
554 * Pop the "info" one that is always at the top of the stack.
556 info = attr_stack;
557 attr_stack = info->prev;
560 * Pop the ones from directories that are not the prefix of
561 * the path we are checking. Break out of the loop when we see
562 * the root one (whose origin is an empty string "") or the builtin
563 * one (whose origin is NULL) without popping it.
565 while (attr_stack->origin) {
566 int namelen = strlen(attr_stack->origin);
568 elem = attr_stack;
569 if (namelen <= dirlen &&
570 !strncmp(elem->origin, path, namelen) &&
571 (!namelen || path[namelen] == '/'))
572 break;
574 debug_pop(elem);
575 attr_stack = elem->prev;
576 free_attr_elem(elem);
580 * Read from parent directories and push them down
582 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
584 * bootstrap_attr_stack() should have added, and the
585 * above loop should have stopped before popping, the
586 * root element whose attr_stack->origin is set to an
587 * empty string.
589 struct strbuf pathbuf = STRBUF_INIT;
591 assert(attr_stack->origin);
592 while (1) {
593 len = strlen(attr_stack->origin);
594 if (dirlen <= len)
595 break;
596 cp = memchr(path + len + 1, '/', dirlen - len - 1);
597 if (!cp)
598 cp = path + dirlen;
599 strbuf_add(&pathbuf, path, cp - path);
600 strbuf_addch(&pathbuf, '/');
601 strbuf_addstr(&pathbuf, GITATTRIBUTES_FILE);
602 elem = read_attr(pathbuf.buf, 0);
603 strbuf_setlen(&pathbuf, cp - path);
604 elem->origin = strbuf_detach(&pathbuf, &elem->originlen);
605 elem->prev = attr_stack;
606 attr_stack = elem;
607 debug_push(elem);
610 strbuf_release(&pathbuf);
614 * Finally push the "info" one at the top of the stack.
616 info->prev = attr_stack;
617 attr_stack = info;
620 static int path_matches(const char *pathname, int pathlen,
621 int basename_offset,
622 const struct pattern *pat,
623 const char *base, int baselen)
625 const char *pattern = pat->pattern;
626 int prefix = pat->nowildcardlen;
627 int isdir = (pathlen && pathname[pathlen - 1] == '/');
629 if ((pat->flags & EXC_FLAG_MUSTBEDIR) && !isdir)
630 return 0;
632 if (pat->flags & EXC_FLAG_NODIR) {
633 return match_basename(pathname + basename_offset,
634 pathlen - basename_offset - isdir,
635 pattern, prefix,
636 pat->patternlen, pat->flags);
638 return match_pathname(pathname, pathlen - isdir,
639 base, baselen,
640 pattern, prefix, pat->patternlen, pat->flags);
643 static int macroexpand_one(int attr_nr, int rem);
645 static int fill_one(const char *what, struct match_attr *a, int rem)
647 struct git_attr_check *check = check_all_attr;
648 int i;
650 for (i = a->num_attr - 1; 0 < rem && 0 <= i; i--) {
651 struct git_attr *attr = a->state[i].attr;
652 const char **n = &(check[attr->attr_nr].value);
653 const char *v = a->state[i].setto;
655 if (*n == ATTR__UNKNOWN) {
656 debug_set(what,
657 a->is_macro ? a->u.attr->name : a->u.pat.pattern,
658 attr, v);
659 *n = v;
660 rem--;
661 rem = macroexpand_one(attr->attr_nr, rem);
664 return rem;
667 static int fill(const char *path, int pathlen, int basename_offset,
668 struct attr_stack *stk, int rem)
670 int i;
671 const char *base = stk->origin ? stk->origin : "";
673 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
674 struct match_attr *a = stk->attrs[i];
675 if (a->is_macro)
676 continue;
677 if (path_matches(path, pathlen, basename_offset,
678 &a->u.pat, base, stk->originlen))
679 rem = fill_one("fill", a, rem);
681 return rem;
684 static int macroexpand_one(int attr_nr, int rem)
686 struct attr_stack *stk;
687 struct match_attr *a = NULL;
688 int i;
690 if (check_all_attr[attr_nr].value != ATTR__TRUE)
691 return rem;
693 for (stk = attr_stack; !a && stk; stk = stk->prev)
694 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
695 struct match_attr *ma = stk->attrs[i];
696 if (!ma->is_macro)
697 continue;
698 if (ma->u.attr->attr_nr == attr_nr)
699 a = ma;
702 if (a)
703 rem = fill_one("expand", a, rem);
705 return rem;
709 * Collect all attributes for path into the array pointed to by
710 * check_all_attr.
712 static void collect_all_attrs(const char *path)
714 struct attr_stack *stk;
715 int i, pathlen, rem, dirlen;
716 const char *cp, *last_slash = NULL;
717 int basename_offset;
719 for (cp = path; *cp; cp++) {
720 if (*cp == '/' && cp[1])
721 last_slash = cp;
723 pathlen = cp - path;
724 if (last_slash) {
725 basename_offset = last_slash + 1 - path;
726 dirlen = last_slash - path;
727 } else {
728 basename_offset = 0;
729 dirlen = 0;
732 prepare_attr_stack(path, dirlen);
733 for (i = 0; i < attr_nr; i++)
734 check_all_attr[i].value = ATTR__UNKNOWN;
736 rem = attr_nr;
737 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
738 rem = fill(path, pathlen, basename_offset, stk, rem);
741 int git_check_attr(const char *path, int num, struct git_attr_check *check)
743 int i;
745 collect_all_attrs(path);
747 for (i = 0; i < num; i++) {
748 const char *value = check_all_attr[check[i].attr->attr_nr].value;
749 if (value == ATTR__UNKNOWN)
750 value = ATTR__UNSET;
751 check[i].value = value;
754 return 0;
757 int git_all_attrs(const char *path, int *num, struct git_attr_check **check)
759 int i, count, j;
761 collect_all_attrs(path);
763 /* Count the number of attributes that are set. */
764 count = 0;
765 for (i = 0; i < attr_nr; i++) {
766 const char *value = check_all_attr[i].value;
767 if (value != ATTR__UNSET && value != ATTR__UNKNOWN)
768 ++count;
770 *num = count;
771 *check = xmalloc(sizeof(**check) * count);
772 j = 0;
773 for (i = 0; i < attr_nr; i++) {
774 const char *value = check_all_attr[i].value;
775 if (value != ATTR__UNSET && value != ATTR__UNKNOWN) {
776 (*check)[j].attr = check_all_attr[i].attr;
777 (*check)[j].value = value;
778 ++j;
782 return 0;
785 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
787 enum git_attr_direction old = direction;
789 if (is_bare_repository() && new != GIT_ATTR_INDEX)
790 die("BUG: non-INDEX attr direction in a bare repo");
792 direction = new;
793 if (new != old)
794 drop_attr_stack();
795 use_index = istate;