a7d1aa95d215b2259afeecbbe20506921d33fd6d
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 */
9
10#define NO_THE_INDEX_COMPATIBILITY_MACROS
11#include "cache.h"
12#include "exec_cmd.h"
13#include "attr.h"
14
15const char git_attr__true[] = "(builtin)true";
16const char git_attr__false[] = "\0(builtin)false";
17static 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
22
23static const char *attributes_file;
24
25/* This is a randomly chosen prime. */
26#define HASHSIZE 257
27
28#ifndef DEBUG_ATTR
29#define DEBUG_ATTR 0
30#endif
31
32struct git_attr {
33 struct git_attr *next;
34 unsigned h;
35 int attr_nr;
36 char name[FLEX_ARRAY];
37};
38static int attr_nr;
39
40static struct git_attr_check *check_all_attr;
41static struct git_attr *(git_attr_hash[HASHSIZE]);
42
43static unsigned hash_name(const char *name, int namelen)
44{
45 unsigned val = 0, c;
46
47 while (namelen--) {
48 c = *name++;
49 val = ((val << 7) | (val >> 22)) ^ c;
50 }
51 return val;
52}
53
54static 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}
74
75static 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;
80
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 }
86
87 if (invalid_attr_name(name, len))
88 return NULL;
89
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;
97
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}
104
105struct git_attr *git_attr(const char *name)
106{
107 return git_attr_internal(name, strlen(name));
108}
109
110/* What does a matched pattern decide? */
111struct attr_state {
112 struct git_attr *attr;
113 const char *setto;
114};
115
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 */
130struct 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};
139
140static const char blank[] = " \t\r\n";
141
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 */
149static 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;
154
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}
189
190static 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;
199
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;
225
226 states = name + namelen;
227 states += strspn(states, blank);
228
229 for (pass = 0; pass < 2; pass++) {
230 /* pass 0 counts and allocates, pass 1 fills */
231 num_attr = 0;
232 cp = states;
233 while (*cp) {
234 cp = parse_attr(src, lineno, cp,
235 pass ? &(res->state[num_attr]) : NULL);
236 if (!cp)
237 return NULL;
238 num_attr++;
239 }
240 if (pass)
241 break;
242 res = xcalloc(1,
243 sizeof(*res) +
244 sizeof(struct attr_state) * num_attr +
245 (is_macro ? 0 : namelen + 1));
246 if (is_macro)
247 res->u.attr = git_attr_internal(name, namelen);
248 else {
249 res->u.pattern = (char *)&(res->state[num_attr]);
250 memcpy(res->u.pattern, name, namelen);
251 res->u.pattern[namelen] = 0;
252 }
253 res->is_macro = is_macro;
254 res->num_attr = num_attr;
255 }
256 return res;
257}
258
259/*
260 * Like info/exclude and .gitignore, the attribute information can
261 * come from many places.
262 *
263 * (1) .gitattribute file of the same directory;
264 * (2) .gitattribute file of the parent directory if (1) does not have
265 * any match; this goes recursively upwards, just like .gitignore.
266 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
267 *
268 * In the same file, later entries override the earlier match, so in the
269 * global list, we would have entries from info/attributes the earliest
270 * (reading the file from top to bottom), .gitattribute of the root
271 * directory (again, reading the file from top to bottom) down to the
272 * current directory, and then scan the list backwards to find the first match.
273 * This is exactly the same as what excluded() does in dir.c to deal with
274 * .gitignore
275 */
276
277static struct attr_stack {
278 struct attr_stack *prev;
279 char *origin;
280 unsigned num_matches;
281 unsigned alloc;
282 struct match_attr **attrs;
283} *attr_stack;
284
285static void free_attr_elem(struct attr_stack *e)
286{
287 int i;
288 free(e->origin);
289 for (i = 0; i < e->num_matches; i++) {
290 struct match_attr *a = e->attrs[i];
291 int j;
292 for (j = 0; j < a->num_attr; j++) {
293 const char *setto = a->state[j].setto;
294 if (setto == ATTR__TRUE ||
295 setto == ATTR__FALSE ||
296 setto == ATTR__UNSET ||
297 setto == ATTR__UNKNOWN)
298 ;
299 else
300 free((char *) setto);
301 }
302 free(a);
303 }
304 free(e);
305}
306
307static const char *builtin_attr[] = {
308 "[attr]binary -diff -text",
309 NULL,
310};
311
312static void handle_attr_line(struct attr_stack *res,
313 const char *line,
314 const char *src,
315 int lineno,
316 int macro_ok)
317{
318 struct match_attr *a;
319
320 a = parse_attr_line(line, src, lineno, macro_ok);
321 if (!a)
322 return;
323 if (res->alloc <= res->num_matches) {
324 res->alloc = alloc_nr(res->num_matches);
325 res->attrs = xrealloc(res->attrs,
326 sizeof(struct match_attr *) *
327 res->alloc);
328 }
329 res->attrs[res->num_matches++] = a;
330}
331
332static struct attr_stack *read_attr_from_array(const char **list)
333{
334 struct attr_stack *res;
335 const char *line;
336 int lineno = 0;
337
338 res = xcalloc(1, sizeof(*res));
339 while ((line = *(list++)) != NULL)
340 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
341 return res;
342}
343
344static enum git_attr_direction direction;
345static struct index_state *use_index;
346
347static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
348{
349 FILE *fp = fopen(path, "r");
350 struct attr_stack *res;
351 char buf[2048];
352 int lineno = 0;
353
354 if (!fp)
355 return NULL;
356 res = xcalloc(1, sizeof(*res));
357 while (fgets(buf, sizeof(buf), fp))
358 handle_attr_line(res, buf, path, ++lineno, macro_ok);
359 fclose(fp);
360 return res;
361}
362
363static void *read_index_data(const char *path)
364{
365 int pos, len;
366 unsigned long sz;
367 enum object_type type;
368 void *data;
369 struct index_state *istate = use_index ? use_index : &the_index;
370
371 len = strlen(path);
372 pos = index_name_pos(istate, path, len);
373 if (pos < 0) {
374 /*
375 * We might be in the middle of a merge, in which
376 * case we would read stage #2 (ours).
377 */
378 int i;
379 for (i = -pos - 1;
380 (pos < 0 && i < istate->cache_nr &&
381 !strcmp(istate->cache[i]->name, path));
382 i++)
383 if (ce_stage(istate->cache[i]) == 2)
384 pos = i;
385 }
386 if (pos < 0)
387 return NULL;
388 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
389 if (!data || type != OBJ_BLOB) {
390 free(data);
391 return NULL;
392 }
393 return data;
394}
395
396static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
397{
398 struct attr_stack *res;
399 char *buf, *sp;
400 int lineno = 0;
401
402 buf = read_index_data(path);
403 if (!buf)
404 return NULL;
405
406 res = xcalloc(1, sizeof(*res));
407 for (sp = buf; *sp; ) {
408 char *ep;
409 int more;
410 for (ep = sp; *ep && *ep != '\n'; ep++)
411 ;
412 more = (*ep == '\n');
413 *ep = '\0';
414 handle_attr_line(res, sp, path, ++lineno, macro_ok);
415 sp = ep + more;
416 }
417 free(buf);
418 return res;
419}
420
421static struct attr_stack *read_attr(const char *path, int macro_ok)
422{
423 struct attr_stack *res;
424
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);
429 }
430 else if (direction == GIT_ATTR_CHECKIN) {
431 res = read_attr_from_file(path, macro_ok);
432 if (!res)
433 /*
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.
437 */
438 res = read_attr_from_index(path, macro_ok);
439 }
440 else
441 res = read_attr_from_index(path, macro_ok);
442 if (!res)
443 res = xcalloc(1, sizeof(*res));
444 return res;
445}
446
447#if DEBUG_ATTR
448static void debug_info(const char *what, struct attr_stack *elem)
449{
450 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
451}
452static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
453{
454 const char *value = v;
455
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";
462
463 fprintf(stderr, "%s: %s => %s (%s)\n",
464 what, attr->name, (char *) value, match);
465}
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
473
474static void drop_attr_stack(void)
475{
476 while (attr_stack) {
477 struct attr_stack *elem = attr_stack;
478 attr_stack = elem->prev;
479 free_attr_elem(elem);
480 }
481}
482
483static const char *git_etc_gitattributes(void)
484{
485 static const char *system_wide;
486 if (!system_wide)
487 system_wide = system_path(ETC_GITATTRIBUTES);
488 return system_wide;
489}
490
491static int git_attr_system(void)
492{
493 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
494}
495
496static int git_attr_config(const char *var, const char *value, void *dummy)
497{
498 if (!strcmp(var, "core.attributesfile"))
499 return git_config_pathname(&attributes_file, var, value);
500
501 return 0;
502}
503
504static void bootstrap_attr_stack(void)
505{
506 if (!attr_stack) {
507 struct attr_stack *elem;
508
509 elem = read_attr_from_array(builtin_attr);
510 elem->origin = NULL;
511 elem->prev = attr_stack;
512 attr_stack = elem;
513
514 if (git_attr_system()) {
515 elem = read_attr_from_file(git_etc_gitattributes(), 1);
516 if (elem) {
517 elem->origin = NULL;
518 elem->prev = attr_stack;
519 attr_stack = elem;
520 }
521 }
522
523 git_config(git_attr_config, NULL);
524 if (attributes_file) {
525 elem = read_attr_from_file(attributes_file, 1);
526 if (elem) {
527 elem->origin = NULL;
528 elem->prev = attr_stack;
529 attr_stack = elem;
530 }
531 }
532
533 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
534 elem = read_attr(GITATTRIBUTES_FILE, 1);
535 elem->origin = strdup("");
536 elem->prev = attr_stack;
537 attr_stack = elem;
538 debug_push(elem);
539 }
540
541 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
542 if (!elem)
543 elem = xcalloc(1, sizeof(*elem));
544 elem->origin = NULL;
545 elem->prev = attr_stack;
546 attr_stack = elem;
547 }
548}
549
550static void prepare_attr_stack(const char *path, int dirlen)
551{
552 struct attr_stack *elem, *info;
553 int len;
554 struct strbuf pathbuf;
555
556 strbuf_init(&pathbuf, dirlen+2+strlen(GITATTRIBUTES_FILE));
557
558 /*
559 * At the bottom of the attribute stack is the built-in
560 * set of attribute definitions, followed by the contents
561 * of $(prefix)/etc/gitattributes and a file specified by
562 * core.attributesfile. Then, contents from
563 * .gitattribute files from directories closer to the
564 * root to the ones in deeper directories are pushed
565 * to the stack. Finally, at the very top of the stack
566 * we always keep the contents of $GIT_DIR/info/attributes.
567 *
568 * When checking, we use entries from near the top of the
569 * stack, preferring $GIT_DIR/info/attributes, then
570 * .gitattributes in deeper directories to shallower ones,
571 * and finally use the built-in set as the default.
572 */
573 if (!attr_stack)
574 bootstrap_attr_stack();
575
576 /*
577 * Pop the "info" one that is always at the top of the stack.
578 */
579 info = attr_stack;
580 attr_stack = info->prev;
581
582 /*
583 * Pop the ones from directories that are not the prefix of
584 * the path we are checking.
585 */
586 while (attr_stack && attr_stack->origin) {
587 int namelen = strlen(attr_stack->origin);
588
589 elem = attr_stack;
590 if (namelen <= dirlen &&
591 !strncmp(elem->origin, path, namelen))
592 break;
593
594 debug_pop(elem);
595 attr_stack = elem->prev;
596 free_attr_elem(elem);
597 }
598
599 /*
600 * Read from parent directories and push them down
601 */
602 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
603 while (1) {
604 char *cp;
605
606 len = strlen(attr_stack->origin);
607 if (dirlen <= len)
608 break;
609 strbuf_reset(&pathbuf);
610 strbuf_add(&pathbuf, path, dirlen);
611 strbuf_addch(&pathbuf, '/');
612 cp = strchr(pathbuf.buf + len + 1, '/');
613 strcpy(cp + 1, GITATTRIBUTES_FILE);
614 elem = read_attr(pathbuf.buf, 0);
615 *cp = '\0';
616 elem->origin = strdup(pathbuf.buf);
617 elem->prev = attr_stack;
618 attr_stack = elem;
619 debug_push(elem);
620 }
621 }
622
623 strbuf_release(&pathbuf);
624
625 /*
626 * Finally push the "info" one at the top of the stack.
627 */
628 info->prev = attr_stack;
629 attr_stack = info;
630}
631
632static int path_matches(const char *pathname, int pathlen,
633 const char *pattern,
634 const char *base, int baselen)
635{
636 if (!strchr(pattern, '/')) {
637 /* match basename */
638 const char *basename = strrchr(pathname, '/');
639 basename = basename ? basename + 1 : pathname;
640 return (fnmatch(pattern, basename, 0) == 0);
641 }
642 /*
643 * match with FNM_PATHNAME; the pattern has base implicitly
644 * in front of it.
645 */
646 if (*pattern == '/')
647 pattern++;
648 if (pathlen < baselen ||
649 (baselen && pathname[baselen] != '/') ||
650 strncmp(pathname, base, baselen))
651 return 0;
652 if (baselen != 0)
653 baselen++;
654 return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
655}
656
657static int macroexpand_one(int attr_nr, int rem);
658
659static int fill_one(const char *what, struct match_attr *a, int rem)
660{
661 struct git_attr_check *check = check_all_attr;
662 int i;
663
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;
668
669 if (*n == ATTR__UNKNOWN) {
670 debug_set(what,
671 a->is_macro ? a->u.attr->name : a->u.pattern,
672 attr, v);
673 *n = v;
674 rem--;
675 rem = macroexpand_one(attr->attr_nr, rem);
676 }
677 }
678 return rem;
679}
680
681static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
682{
683 int i;
684 const char *base = stk->origin ? stk->origin : "";
685
686 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
687 struct match_attr *a = stk->attrs[i];
688 if (a->is_macro)
689 continue;
690 if (path_matches(path, pathlen,
691 a->u.pattern, base, strlen(base)))
692 rem = fill_one("fill", a, rem);
693 }
694 return rem;
695}
696
697static int macroexpand_one(int attr_nr, int rem)
698{
699 struct attr_stack *stk;
700 struct match_attr *a = NULL;
701 int i;
702
703 if (check_all_attr[attr_nr].value != ATTR__TRUE)
704 return rem;
705
706 for (stk = attr_stack; !a && stk; stk = stk->prev)
707 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
708 struct match_attr *ma = stk->attrs[i];
709 if (!ma->is_macro)
710 continue;
711 if (ma->u.attr->attr_nr == attr_nr)
712 a = ma;
713 }
714
715 if (a)
716 rem = fill_one("expand", a, rem);
717
718 return rem;
719}
720
721int git_checkattr(const char *path, int num, struct git_attr_check *check)
722{
723 struct attr_stack *stk;
724 const char *cp;
725 int dirlen, pathlen, i, rem;
726
727 bootstrap_attr_stack();
728 for (i = 0; i < attr_nr; i++)
729 check_all_attr[i].value = ATTR__UNKNOWN;
730
731 pathlen = strlen(path);
732 cp = strrchr(path, '/');
733 if (!cp)
734 dirlen = 0;
735 else
736 dirlen = cp - path;
737 prepare_attr_stack(path, dirlen);
738 rem = attr_nr;
739 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
740 rem = fill(path, pathlen, stk, rem);
741
742 for (i = 0; i < num; i++) {
743 const char *value = check_all_attr[check[i].attr->attr_nr].value;
744 if (value == ATTR__UNKNOWN)
745 value = ATTR__UNSET;
746 check[i].value = value;
747 }
748
749 return 0;
750}
751
752void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
753{
754 enum git_attr_direction old = direction;
755
756 if (is_bare_repository() && new != GIT_ATTR_INDEX)
757 die("BUG: non-INDEX attr direction in a bare repo");
758
759 direction = new;
760 if (new != old)
761 drop_attr_stack();
762 use_index = istate;
763}