b56542e7234161518a131d3e3edbb8c20c50fbab
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.
5 *
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"
15 const char git_attr__true[] = "(builtin)true";
16 const char git_attr__false[] = "\0(builtin)false";
17 static const char git_attr__unknown[] = "(builtin)unknown";
18 #define ATTR__TRUE git_attr__true
19 #define ATTR__FALSE git_attr__false
20 #define ATTR__UNSET NULL
21 #define ATTR__UNKNOWN git_attr__unknown
23 static const char *attributes_file;
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 char name[FLEX_ARRAY];
37 };
38 static int attr_nr;
40 static struct git_attr_check *check_all_attr;
41 static struct git_attr *(git_attr_hash[HASHSIZE]);
43 static unsigned hash_name(const char *name, int namelen)
44 {
45 unsigned val = 0, c;
47 while (namelen--) {
48 c = *name++;
49 val = ((val << 7) | (val >> 22)) ^ c;
50 }
51 return val;
52 }
54 static int invalid_attr_name(const char *name, int namelen)
55 {
56 /*
57 * Attribute name cannot begin with '-' and from
58 * [-A-Za-z0-9_.]. We'd specifically exclude '=' for now,
59 * as we might later want to allow non-binary value for
60 * attributes, e.g. "*.svg merge=special-merge-program-for-svg"
61 */
62 if (*name == '-')
63 return -1;
64 while (namelen--) {
65 char ch = *name++;
66 if (! (ch == '-' || ch == '.' || ch == '_' ||
67 ('0' <= ch && ch <= '9') ||
68 ('a' <= ch && ch <= 'z') ||
69 ('A' <= ch && ch <= 'Z')) )
70 return -1;
71 }
72 return 0;
73 }
75 static struct git_attr *git_attr_internal(const char *name, int len)
76 {
77 unsigned hval = hash_name(name, len);
78 unsigned pos = hval % HASHSIZE;
79 struct git_attr *a;
81 for (a = git_attr_hash[pos]; a; a = a->next) {
82 if (a->h == hval &&
83 !memcmp(a->name, name, len) && !a->name[len])
84 return a;
85 }
87 if (invalid_attr_name(name, len))
88 return NULL;
90 a = xmalloc(sizeof(*a) + len + 1);
91 memcpy(a->name, name, len);
92 a->name[len] = 0;
93 a->h = hval;
94 a->next = git_attr_hash[pos];
95 a->attr_nr = attr_nr++;
96 git_attr_hash[pos] = a;
98 check_all_attr = xrealloc(check_all_attr,
99 sizeof(*check_all_attr) * attr_nr);
100 check_all_attr[a->attr_nr].attr = a;
101 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
102 return a;
103 }
105 struct git_attr *git_attr(const char *name)
106 {
107 return git_attr_internal(name, strlen(name));
108 }
110 /* What does a matched pattern decide? */
111 struct attr_state {
112 struct git_attr *attr;
113 const char *setto;
114 };
116 /*
117 * One rule, as from a .gitattributes file.
118 *
119 * If is_macro is true, then u.attr is a pointer to the git_attr being
120 * defined.
121 *
122 * If is_macro is false, then u.pattern points at the filename pattern
123 * to which the rule applies. (The memory pointed to is part of the
124 * memory block allocated for the match_attr instance.)
125 *
126 * In either case, num_attr is the number of attributes affected by
127 * this rule, and state is an array listing them. The attributes are
128 * listed as they appear in the file (macros unexpanded).
129 */
130 struct match_attr {
131 union {
132 char *pattern;
133 struct git_attr *attr;
134 } u;
135 char is_macro;
136 unsigned num_attr;
137 struct attr_state state[FLEX_ARRAY];
138 };
140 static const char blank[] = " \t\r\n";
142 /*
143 * Parse a whitespace-delimited attribute state (i.e., "attr",
144 * "-attr", "!attr", or "attr=value") from the string starting at src.
145 * If e is not NULL, write the results to *e. Return a pointer to the
146 * remainder of the string (with leading whitespace removed), or NULL
147 * if there was an error.
148 */
149 static const char *parse_attr(const char *src, int lineno, const char *cp,
150 struct attr_state *e)
151 {
152 const char *ep, *equals;
153 int len;
155 ep = cp + strcspn(cp, blank);
156 equals = strchr(cp, '=');
157 if (equals && ep < equals)
158 equals = NULL;
159 if (equals)
160 len = equals - cp;
161 else
162 len = ep - cp;
163 if (!e) {
164 if (*cp == '-' || *cp == '!') {
165 cp++;
166 len--;
167 }
168 if (invalid_attr_name(cp, len)) {
169 fprintf(stderr,
170 "%.*s is not a valid attribute name: %s:%d\n",
171 len, cp, src, lineno);
172 return NULL;
173 }
174 } else {
175 if (*cp == '-' || *cp == '!') {
176 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
177 cp++;
178 len--;
179 }
180 else if (!equals)
181 e->setto = ATTR__TRUE;
182 else {
183 e->setto = xmemdupz(equals + 1, ep - equals - 1);
184 }
185 e->attr = git_attr_internal(cp, len);
186 }
187 return ep + strspn(ep, blank);
188 }
190 static struct match_attr *parse_attr_line(const char *line, const char *src,
191 int lineno, int macro_ok)
192 {
193 int namelen;
194 int num_attr;
195 const char *cp, *name, *states;
196 struct match_attr *res = NULL;
197 int pass;
198 int is_macro;
200 cp = line + strspn(line, blank);
201 if (!*cp || *cp == '#')
202 return NULL;
203 name = cp;
204 namelen = strcspn(name, blank);
205 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
206 !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
207 if (!macro_ok) {
208 fprintf(stderr, "%s not allowed: %s:%d\n",
209 name, src, lineno);
210 return NULL;
211 }
212 is_macro = 1;
213 name += strlen(ATTRIBUTE_MACRO_PREFIX);
214 name += strspn(name, blank);
215 namelen = strcspn(name, blank);
216 if (invalid_attr_name(name, namelen)) {
217 fprintf(stderr,
218 "%.*s is not a valid attribute name: %s:%d\n",
219 namelen, name, src, lineno);
220 return NULL;
221 }
222 }
223 else
224 is_macro = 0;
226 states = name + namelen;
227 states += strspn(states, blank);
229 for (pass = 0; pass < 2; pass++) {
230 /* pass 0 counts and allocates, pass 1 fills */
231 for (cp = states, num_attr = 0; *cp; num_attr++) {
232 cp = parse_attr(src, lineno, cp,
233 pass ? &(res->state[num_attr]) : NULL);
234 if (!cp)
235 return NULL;
236 }
237 if (pass)
238 break;
239 res = xcalloc(1,
240 sizeof(*res) +
241 sizeof(struct attr_state) * num_attr +
242 (is_macro ? 0 : namelen + 1));
243 if (is_macro)
244 res->u.attr = git_attr_internal(name, namelen);
245 else {
246 res->u.pattern = (char *)&(res->state[num_attr]);
247 memcpy(res->u.pattern, name, namelen);
248 res->u.pattern[namelen] = 0;
249 }
250 res->is_macro = is_macro;
251 res->num_attr = num_attr;
252 }
253 return res;
254 }
256 /*
257 * Like info/exclude and .gitignore, the attribute information can
258 * come from many places.
259 *
260 * (1) .gitattribute file of the same directory;
261 * (2) .gitattribute file of the parent directory if (1) does not have
262 * any match; this goes recursively upwards, just like .gitignore.
263 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
264 *
265 * In the same file, later entries override the earlier match, so in the
266 * global list, we would have entries from info/attributes the earliest
267 * (reading the file from top to bottom), .gitattribute of the root
268 * directory (again, reading the file from top to bottom) down to the
269 * current directory, and then scan the list backwards to find the first match.
270 * This is exactly the same as what excluded() does in dir.c to deal with
271 * .gitignore
272 */
274 static struct attr_stack {
275 struct attr_stack *prev;
276 char *origin;
277 unsigned num_matches;
278 unsigned alloc;
279 struct match_attr **attrs;
280 } *attr_stack;
282 static void free_attr_elem(struct attr_stack *e)
283 {
284 int i;
285 free(e->origin);
286 for (i = 0; i < e->num_matches; i++) {
287 struct match_attr *a = e->attrs[i];
288 int j;
289 for (j = 0; j < a->num_attr; j++) {
290 const char *setto = a->state[j].setto;
291 if (setto == ATTR__TRUE ||
292 setto == ATTR__FALSE ||
293 setto == ATTR__UNSET ||
294 setto == ATTR__UNKNOWN)
295 ;
296 else
297 free((char *) setto);
298 }
299 free(a);
300 }
301 free(e);
302 }
304 static const char *builtin_attr[] = {
305 "[attr]binary -diff -text",
306 NULL,
307 };
309 static void handle_attr_line(struct attr_stack *res,
310 const char *line,
311 const char *src,
312 int lineno,
313 int macro_ok)
314 {
315 struct match_attr *a;
317 a = parse_attr_line(line, src, lineno, macro_ok);
318 if (!a)
319 return;
320 if (res->alloc <= res->num_matches) {
321 res->alloc = alloc_nr(res->num_matches);
322 res->attrs = xrealloc(res->attrs,
323 sizeof(struct match_attr *) *
324 res->alloc);
325 }
326 res->attrs[res->num_matches++] = a;
327 }
329 static struct attr_stack *read_attr_from_array(const char **list)
330 {
331 struct attr_stack *res;
332 const char *line;
333 int lineno = 0;
335 res = xcalloc(1, sizeof(*res));
336 while ((line = *(list++)) != NULL)
337 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
338 return res;
339 }
341 static enum git_attr_direction direction;
342 static struct index_state *use_index;
344 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
345 {
346 FILE *fp = fopen(path, "r");
347 struct attr_stack *res;
348 char buf[2048];
349 int lineno = 0;
351 if (!fp)
352 return NULL;
353 res = xcalloc(1, sizeof(*res));
354 while (fgets(buf, sizeof(buf), fp))
355 handle_attr_line(res, buf, path, ++lineno, macro_ok);
356 fclose(fp);
357 return res;
358 }
360 static void *read_index_data(const char *path)
361 {
362 int pos, len;
363 unsigned long sz;
364 enum object_type type;
365 void *data;
366 struct index_state *istate = use_index ? use_index : &the_index;
368 len = strlen(path);
369 pos = index_name_pos(istate, path, len);
370 if (pos < 0) {
371 /*
372 * We might be in the middle of a merge, in which
373 * case we would read stage #2 (ours).
374 */
375 int i;
376 for (i = -pos - 1;
377 (pos < 0 && i < istate->cache_nr &&
378 !strcmp(istate->cache[i]->name, path));
379 i++)
380 if (ce_stage(istate->cache[i]) == 2)
381 pos = i;
382 }
383 if (pos < 0)
384 return NULL;
385 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
386 if (!data || type != OBJ_BLOB) {
387 free(data);
388 return NULL;
389 }
390 return data;
391 }
393 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
394 {
395 struct attr_stack *res;
396 char *buf, *sp;
397 int lineno = 0;
399 buf = read_index_data(path);
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++)
408 ;
409 more = (*ep == '\n');
410 *ep = '\0';
411 handle_attr_line(res, sp, path, ++lineno, macro_ok);
412 sp = ep + more;
413 }
414 free(buf);
415 return res;
416 }
418 static struct attr_stack *read_attr(const char *path, int macro_ok)
419 {
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);
426 }
427 else if (direction == GIT_ATTR_CHECKIN) {
428 res = read_attr_from_file(path, macro_ok);
429 if (!res)
430 /*
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.
434 */
435 res = read_attr_from_index(path, macro_ok);
436 }
437 else
438 res = read_attr_from_index(path, macro_ok);
439 if (!res)
440 res = xcalloc(1, sizeof(*res));
441 return res;
442 }
444 #if DEBUG_ATTR
445 static void debug_info(const char *what, struct attr_stack *elem)
446 {
447 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
448 }
449 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
450 {
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);
462 }
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)
472 {
473 while (attr_stack) {
474 struct attr_stack *elem = attr_stack;
475 attr_stack = elem->prev;
476 free_attr_elem(elem);
477 }
478 }
480 static const char *git_etc_gitattributes(void)
481 {
482 static const char *system_wide;
483 if (!system_wide)
484 system_wide = system_path(ETC_GITATTRIBUTES);
485 return system_wide;
486 }
488 static int git_attr_system(void)
489 {
490 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
491 }
493 static int git_attr_config(const char *var, const char *value, void *dummy)
494 {
495 if (!strcmp(var, "core.attributesfile"))
496 return git_config_pathname(&attributes_file, var, value);
498 return 0;
499 }
501 static void bootstrap_attr_stack(void)
502 {
503 if (!attr_stack) {
504 struct attr_stack *elem;
506 elem = read_attr_from_array(builtin_attr);
507 elem->origin = NULL;
508 elem->prev = attr_stack;
509 attr_stack = elem;
511 if (git_attr_system()) {
512 elem = read_attr_from_file(git_etc_gitattributes(), 1);
513 if (elem) {
514 elem->origin = NULL;
515 elem->prev = attr_stack;
516 attr_stack = elem;
517 }
518 }
520 git_config(git_attr_config, NULL);
521 if (attributes_file) {
522 elem = read_attr_from_file(attributes_file, 1);
523 if (elem) {
524 elem->origin = NULL;
525 elem->prev = attr_stack;
526 attr_stack = elem;
527 }
528 }
530 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
531 elem = read_attr(GITATTRIBUTES_FILE, 1);
532 elem->origin = strdup("");
533 elem->prev = attr_stack;
534 attr_stack = elem;
535 debug_push(elem);
536 }
538 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
539 if (!elem)
540 elem = xcalloc(1, sizeof(*elem));
541 elem->origin = NULL;
542 elem->prev = attr_stack;
543 attr_stack = elem;
544 }
545 }
547 static void prepare_attr_stack(const char *path, int dirlen)
548 {
549 struct attr_stack *elem, *info;
550 int len;
551 struct strbuf pathbuf;
553 strbuf_init(&pathbuf, dirlen+2+strlen(GITATTRIBUTES_FILE));
555 /*
556 * At the bottom of the attribute stack is the built-in
557 * set of attribute definitions, followed by the contents
558 * of $(prefix)/etc/gitattributes and a file specified by
559 * core.attributesfile. Then, contents from
560 * .gitattribute files from directories closer to the
561 * root to the ones in deeper directories are pushed
562 * to the stack. Finally, at the very top of the stack
563 * we always keep the contents of $GIT_DIR/info/attributes.
564 *
565 * When checking, we use entries from near the top of the
566 * stack, preferring $GIT_DIR/info/attributes, then
567 * .gitattributes in deeper directories to shallower ones,
568 * and finally use the built-in set as the default.
569 */
570 if (!attr_stack)
571 bootstrap_attr_stack();
573 /*
574 * Pop the "info" one that is always at the top of the stack.
575 */
576 info = attr_stack;
577 attr_stack = info->prev;
579 /*
580 * Pop the ones from directories that are not the prefix of
581 * the path we are checking.
582 */
583 while (attr_stack && attr_stack->origin) {
584 int namelen = strlen(attr_stack->origin);
586 elem = attr_stack;
587 if (namelen <= dirlen &&
588 !strncmp(elem->origin, path, namelen))
589 break;
591 debug_pop(elem);
592 attr_stack = elem->prev;
593 free_attr_elem(elem);
594 }
596 /*
597 * Read from parent directories and push them down
598 */
599 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
600 while (1) {
601 char *cp;
603 len = strlen(attr_stack->origin);
604 if (dirlen <= len)
605 break;
606 strbuf_reset(&pathbuf);
607 strbuf_add(&pathbuf, path, dirlen);
608 strbuf_addch(&pathbuf, '/');
609 cp = strchr(pathbuf.buf + len + 1, '/');
610 strcpy(cp + 1, GITATTRIBUTES_FILE);
611 elem = read_attr(pathbuf.buf, 0);
612 *cp = '\0';
613 elem->origin = strdup(pathbuf.buf);
614 elem->prev = attr_stack;
615 attr_stack = elem;
616 debug_push(elem);
617 }
618 }
620 strbuf_release(&pathbuf);
622 /*
623 * Finally push the "info" one at the top of the stack.
624 */
625 info->prev = attr_stack;
626 attr_stack = info;
627 }
629 static int path_matches(const char *pathname, int pathlen,
630 const char *pattern,
631 const char *base, int baselen)
632 {
633 if (!strchr(pattern, '/')) {
634 /* match basename */
635 const char *basename = strrchr(pathname, '/');
636 basename = basename ? basename + 1 : pathname;
637 return (fnmatch(pattern, basename, 0) == 0);
638 }
639 /*
640 * match with FNM_PATHNAME; the pattern has base implicitly
641 * in front of it.
642 */
643 if (*pattern == '/')
644 pattern++;
645 if (pathlen < baselen ||
646 (baselen && pathname[baselen] != '/') ||
647 strncmp(pathname, base, baselen))
648 return 0;
649 if (baselen != 0)
650 baselen++;
651 return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
652 }
654 static int macroexpand_one(int attr_nr, int rem);
656 static int fill_one(const char *what, struct match_attr *a, int rem)
657 {
658 struct git_attr_check *check = check_all_attr;
659 int i;
661 for (i = a->num_attr - 1; 0 < rem && 0 <= i; i--) {
662 struct git_attr *attr = a->state[i].attr;
663 const char **n = &(check[attr->attr_nr].value);
664 const char *v = a->state[i].setto;
666 if (*n == ATTR__UNKNOWN) {
667 debug_set(what,
668 a->is_macro ? a->u.attr->name : a->u.pattern,
669 attr, v);
670 *n = v;
671 rem--;
672 rem = macroexpand_one(attr->attr_nr, rem);
673 }
674 }
675 return rem;
676 }
678 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
679 {
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,
688 a->u.pattern, base, strlen(base)))
689 rem = fill_one("fill", a, rem);
690 }
691 return rem;
692 }
694 static int macroexpand_one(int attr_nr, int rem)
695 {
696 struct attr_stack *stk;
697 struct match_attr *a = NULL;
698 int i;
700 if (check_all_attr[attr_nr].value != ATTR__TRUE)
701 return rem;
703 for (stk = attr_stack; !a && stk; stk = stk->prev)
704 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
705 struct match_attr *ma = stk->attrs[i];
706 if (!ma->is_macro)
707 continue;
708 if (ma->u.attr->attr_nr == attr_nr)
709 a = ma;
710 }
712 if (a)
713 rem = fill_one("expand", a, rem);
715 return rem;
716 }
718 int git_checkattr(const char *path, int num, struct git_attr_check *check)
719 {
720 struct attr_stack *stk;
721 const char *cp;
722 int dirlen, pathlen, i, rem;
724 bootstrap_attr_stack();
725 for (i = 0; i < attr_nr; i++)
726 check_all_attr[i].value = ATTR__UNKNOWN;
728 pathlen = strlen(path);
729 cp = strrchr(path, '/');
730 if (!cp)
731 dirlen = 0;
732 else
733 dirlen = cp - path;
734 prepare_attr_stack(path, dirlen);
735 rem = attr_nr;
736 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
737 rem = fill(path, pathlen, stk, rem);
739 for (i = 0; i < num; i++) {
740 const char *value = check_all_attr[check[i].attr->attr_nr].value;
741 if (value == ATTR__UNKNOWN)
742 value = ATTR__UNSET;
743 check[i].value = value;
744 }
746 return 0;
747 }
749 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
750 {
751 enum git_attr_direction old = direction;
753 if (is_bare_repository() && new != GIT_ATTR_INDEX)
754 die("BUG: non-INDEX attr direction in a bare repo");
756 direction = new;
757 if (new != old)
758 drop_attr_stack();
759 use_index = istate;
760 }