1#define NO_THE_INDEX_COMPATIBILITY_MACROS
2#include "cache.h"
3#include "attr.h"
4
5const char git_attr__true[] = "(builtin)true";
6const char git_attr__false[] = "\0(builtin)false";
7static const char git_attr__unknown[] = "(builtin)unknown";
8#define ATTR__TRUE git_attr__true
9#define ATTR__FALSE git_attr__false
10#define ATTR__UNSET NULL
11#define ATTR__UNKNOWN git_attr__unknown
12
13/*
14 * The basic design decision here is that we are not going to have
15 * insanely large number of attributes.
16 *
17 * This is a randomly chosen prime.
18 */
19#define HASHSIZE 257
20
21#ifndef DEBUG_ATTR
22#define DEBUG_ATTR 0
23#endif
24
25struct git_attr {
26 struct git_attr *next;
27 unsigned h;
28 int attr_nr;
29 char name[FLEX_ARRAY];
30};
31static int attr_nr;
32
33static struct git_attr_check *check_all_attr;
34static struct git_attr *(git_attr_hash[HASHSIZE]);
35
36static unsigned hash_name(const char *name, int namelen)
37{
38 unsigned val = 0;
39 unsigned char c;
40
41 while (namelen--) {
42 c = *name++;
43 val = ((val << 7) | (val >> 22)) ^ c;
44 }
45 return val;
46}
47
48static int invalid_attr_name(const char *name, int namelen)
49{
50 /*
51 * Attribute name cannot begin with '-' and from
52 * [-A-Za-z0-9_.]. We'd specifically exclude '=' for now,
53 * as we might later want to allow non-binary value for
54 * attributes, e.g. "*.svg merge=special-merge-program-for-svg"
55 */
56 if (*name == '-')
57 return -1;
58 while (namelen--) {
59 char ch = *name++;
60 if (! (ch == '-' || ch == '.' || ch == '_' ||
61 ('0' <= ch && ch <= '9') ||
62 ('a' <= ch && ch <= 'z') ||
63 ('A' <= ch && ch <= 'Z')) )
64 return -1;
65 }
66 return 0;
67}
68
69struct git_attr *git_attr(const char *name, int len)
70{
71 unsigned hval = hash_name(name, len);
72 unsigned pos = hval % HASHSIZE;
73 struct git_attr *a;
74
75 for (a = git_attr_hash[pos]; a; a = a->next) {
76 if (a->h == hval &&
77 !memcmp(a->name, name, len) && !a->name[len])
78 return a;
79 }
80
81 if (invalid_attr_name(name, len))
82 return NULL;
83
84 a = xmalloc(sizeof(*a) + len + 1);
85 memcpy(a->name, name, len);
86 a->name[len] = 0;
87 a->h = hval;
88 a->next = git_attr_hash[pos];
89 a->attr_nr = attr_nr++;
90 git_attr_hash[pos] = a;
91
92 check_all_attr = xrealloc(check_all_attr,
93 sizeof(*check_all_attr) * attr_nr);
94 check_all_attr[a->attr_nr].attr = a;
95 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
96 return a;
97}
98
99/*
100 * .gitattributes file is one line per record, each of which is
101 *
102 * (1) glob pattern.
103 * (2) whitespace
104 * (3) whitespace separated list of attribute names, each of which
105 * could be prefixed with '-' to mean "set to false", '!' to mean
106 * "unset".
107 */
108
109/* What does a matched pattern decide? */
110struct attr_state {
111 struct git_attr *attr;
112 const char *setto;
113};
114
115struct match_attr {
116 union {
117 char *pattern;
118 struct git_attr *attr;
119 } u;
120 char is_macro;
121 unsigned num_attr;
122 struct attr_state state[FLEX_ARRAY];
123};
124
125static const char blank[] = " \t\r\n";
126
127static const char *parse_attr(const char *src, int lineno, const char *cp,
128 int *num_attr, struct match_attr *res)
129{
130 const char *ep, *equals;
131 int len;
132
133 ep = cp + strcspn(cp, blank);
134 equals = strchr(cp, '=');
135 if (equals && ep < equals)
136 equals = NULL;
137 if (equals)
138 len = equals - cp;
139 else
140 len = ep - cp;
141 if (!res) {
142 if (*cp == '-' || *cp == '!') {
143 cp++;
144 len--;
145 }
146 if (invalid_attr_name(cp, len)) {
147 fprintf(stderr,
148 "%.*s is not a valid attribute name: %s:%d\n",
149 len, cp, src, lineno);
150 return NULL;
151 }
152 } else {
153 struct attr_state *e;
154
155 e = &(res->state[*num_attr]);
156 if (*cp == '-' || *cp == '!') {
157 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
158 cp++;
159 len--;
160 }
161 else if (!equals)
162 e->setto = ATTR__TRUE;
163 else {
164 e->setto = xmemdupz(equals + 1, ep - equals - 1);
165 }
166 e->attr = git_attr(cp, len);
167 }
168 (*num_attr)++;
169 return ep + strspn(ep, blank);
170}
171
172static struct match_attr *parse_attr_line(const char *line, const char *src,
173 int lineno, int macro_ok)
174{
175 int namelen;
176 int num_attr;
177 const char *cp, *name;
178 struct match_attr *res = NULL;
179 int pass;
180 int is_macro;
181
182 cp = line + strspn(line, blank);
183 if (!*cp || *cp == '#')
184 return NULL;
185 name = cp;
186 namelen = strcspn(name, blank);
187 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
188 !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
189 if (!macro_ok) {
190 fprintf(stderr, "%s not allowed: %s:%d\n",
191 name, src, lineno);
192 return NULL;
193 }
194 is_macro = 1;
195 name += strlen(ATTRIBUTE_MACRO_PREFIX);
196 name += strspn(name, blank);
197 namelen = strcspn(name, blank);
198 if (invalid_attr_name(name, namelen)) {
199 fprintf(stderr,
200 "%.*s is not a valid attribute name: %s:%d\n",
201 namelen, name, src, lineno);
202 return NULL;
203 }
204 }
205 else
206 is_macro = 0;
207
208 for (pass = 0; pass < 2; pass++) {
209 /* pass 0 counts and allocates, pass 1 fills */
210 num_attr = 0;
211 cp = name + namelen;
212 cp = cp + strspn(cp, blank);
213 while (*cp) {
214 cp = parse_attr(src, lineno, cp, &num_attr, res);
215 if (!cp)
216 return NULL;
217 }
218 if (pass)
219 break;
220 res = xcalloc(1,
221 sizeof(*res) +
222 sizeof(struct attr_state) * num_attr +
223 (is_macro ? 0 : namelen + 1));
224 if (is_macro)
225 res->u.attr = git_attr(name, namelen);
226 else {
227 res->u.pattern = (char*)&(res->state[num_attr]);
228 memcpy(res->u.pattern, name, namelen);
229 res->u.pattern[namelen] = 0;
230 }
231 res->is_macro = is_macro;
232 res->num_attr = num_attr;
233 }
234 return res;
235}
236
237/*
238 * Like info/exclude and .gitignore, the attribute information can
239 * come from many places.
240 *
241 * (1) .gitattribute file of the same directory;
242 * (2) .gitattribute file of the parent directory if (1) does not have
243 * any match; this goes recursively upwards, just like .gitignore.
244 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
245 *
246 * In the same file, later entries override the earlier match, so in the
247 * global list, we would have entries from info/attributes the earliest
248 * (reading the file from top to bottom), .gitattribute of the root
249 * directory (again, reading the file from top to bottom) down to the
250 * current directory, and then scan the list backwards to find the first match.
251 * This is exactly the same as what excluded() does in dir.c to deal with
252 * .gitignore
253 */
254
255static struct attr_stack {
256 struct attr_stack *prev;
257 char *origin;
258 unsigned num_matches;
259 unsigned alloc;
260 struct match_attr **attrs;
261} *attr_stack;
262
263static void free_attr_elem(struct attr_stack *e)
264{
265 int i;
266 free(e->origin);
267 for (i = 0; i < e->num_matches; i++) {
268 struct match_attr *a = e->attrs[i];
269 int j;
270 for (j = 0; j < a->num_attr; j++) {
271 const char *setto = a->state[j].setto;
272 if (setto == ATTR__TRUE ||
273 setto == ATTR__FALSE ||
274 setto == ATTR__UNSET ||
275 setto == ATTR__UNKNOWN)
276 ;
277 else
278 free((char*) setto);
279 }
280 free(a);
281 }
282 free(e);
283}
284
285static const char *builtin_attr[] = {
286 "[attr]binary -diff -crlf",
287 NULL,
288};
289
290static void handle_attr_line(struct attr_stack *res,
291 const char *line,
292 const char *src,
293 int lineno,
294 int macro_ok)
295{
296 struct match_attr *a;
297
298 a = parse_attr_line(line, src, lineno, macro_ok);
299 if (!a)
300 return;
301 if (res->alloc <= res->num_matches) {
302 res->alloc = alloc_nr(res->num_matches);
303 res->attrs = xrealloc(res->attrs,
304 sizeof(struct match_attr *) *
305 res->alloc);
306 }
307 res->attrs[res->num_matches++] = a;
308}
309
310static struct attr_stack *read_attr_from_array(const char **list)
311{
312 struct attr_stack *res;
313 const char *line;
314 int lineno = 0;
315
316 res = xcalloc(1, sizeof(*res));
317 while ((line = *(list++)) != NULL)
318 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
319 return res;
320}
321
322static enum git_attr_direction direction;
323static struct index_state *use_index;
324
325static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
326{
327 FILE *fp = fopen(path, "r");
328 struct attr_stack *res;
329 char buf[2048];
330 int lineno = 0;
331
332 if (!fp)
333 return NULL;
334 res = xcalloc(1, sizeof(*res));
335 while (fgets(buf, sizeof(buf), fp))
336 handle_attr_line(res, buf, path, ++lineno, macro_ok);
337 fclose(fp);
338 return res;
339}
340
341static void *read_index_data(const char *path)
342{
343 int pos, len;
344 unsigned long sz;
345 enum object_type type;
346 void *data;
347 struct index_state *istate = use_index ? use_index : &the_index;
348
349 len = strlen(path);
350 pos = index_name_pos(istate, path, len);
351 if (pos < 0) {
352 /*
353 * We might be in the middle of a merge, in which
354 * case we would read stage #2 (ours).
355 */
356 int i;
357 for (i = -pos - 1;
358 (pos < 0 && i < istate->cache_nr &&
359 !strcmp(istate->cache[i]->name, path));
360 i++)
361 if (ce_stage(istate->cache[i]) == 2)
362 pos = i;
363 }
364 if (pos < 0)
365 return NULL;
366 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
367 if (!data || type != OBJ_BLOB) {
368 free(data);
369 return NULL;
370 }
371 return data;
372}
373
374static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
375{
376 struct attr_stack *res;
377 char *buf, *sp;
378 int lineno = 0;
379
380 buf = read_index_data(path);
381 if (!buf)
382 return NULL;
383
384 res = xcalloc(1, sizeof(*res));
385 for (sp = buf; *sp; ) {
386 char *ep;
387 int more;
388 for (ep = sp; *ep && *ep != '\n'; ep++)
389 ;
390 more = (*ep == '\n');
391 *ep = '\0';
392 handle_attr_line(res, sp, path, ++lineno, macro_ok);
393 sp = ep + more;
394 }
395 free(buf);
396 return res;
397}
398
399static struct attr_stack *read_attr(const char *path, int macro_ok)
400{
401 struct attr_stack *res;
402
403 if (direction == GIT_ATTR_CHECKOUT) {
404 res = read_attr_from_index(path, macro_ok);
405 if (!res)
406 res = read_attr_from_file(path, macro_ok);
407 }
408 else {
409 res = read_attr_from_file(path, macro_ok);
410 if (!res)
411 /*
412 * There is no checked out .gitattributes file there, but
413 * we might have it in the index. We allow operation in a
414 * sparsely checked out work tree, so read from it.
415 */
416 res = read_attr_from_index(path, macro_ok);
417 }
418 if (!res)
419 res = xcalloc(1, sizeof(*res));
420 return res;
421}
422
423#if DEBUG_ATTR
424static void debug_info(const char *what, struct attr_stack *elem)
425{
426 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
427}
428static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
429{
430 const char *value = v;
431
432 if (ATTR_TRUE(value))
433 value = "set";
434 else if (ATTR_FALSE(value))
435 value = "unset";
436 else if (ATTR_UNSET(value))
437 value = "unspecified";
438
439 fprintf(stderr, "%s: %s => %s (%s)\n",
440 what, attr->name, (char *) value, match);
441}
442#define debug_push(a) debug_info("push", (a))
443#define debug_pop(a) debug_info("pop", (a))
444#else
445#define debug_push(a) do { ; } while (0)
446#define debug_pop(a) do { ; } while (0)
447#define debug_set(a,b,c,d) do { ; } while (0)
448#endif
449
450static void drop_attr_stack(void)
451{
452 while (attr_stack) {
453 struct attr_stack *elem = attr_stack;
454 attr_stack = elem->prev;
455 free_attr_elem(elem);
456 }
457}
458
459static void bootstrap_attr_stack(void)
460{
461 if (!attr_stack) {
462 struct attr_stack *elem;
463
464 elem = read_attr_from_array(builtin_attr);
465 elem->origin = NULL;
466 elem->prev = attr_stack;
467 attr_stack = elem;
468
469 if (!is_bare_repository()) {
470 elem = read_attr(GITATTRIBUTES_FILE, 1);
471 elem->origin = strdup("");
472 elem->prev = attr_stack;
473 attr_stack = elem;
474 debug_push(elem);
475 }
476
477 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
478 if (!elem)
479 elem = xcalloc(1, sizeof(*elem));
480 elem->origin = NULL;
481 elem->prev = attr_stack;
482 attr_stack = elem;
483 }
484}
485
486static void prepare_attr_stack(const char *path, int dirlen)
487{
488 struct attr_stack *elem, *info;
489 int len;
490 struct strbuf pathbuf;
491
492 strbuf_init(&pathbuf, dirlen+2+strlen(GITATTRIBUTES_FILE));
493
494 /*
495 * At the bottom of the attribute stack is the built-in
496 * set of attribute definitions. Then, contents from
497 * .gitattribute files from directories closer to the
498 * root to the ones in deeper directories are pushed
499 * to the stack. Finally, at the very top of the stack
500 * we always keep the contents of $GIT_DIR/info/attributes.
501 *
502 * When checking, we use entries from near the top of the
503 * stack, preferring $GIT_DIR/info/attributes, then
504 * .gitattributes in deeper directories to shallower ones,
505 * and finally use the built-in set as the default.
506 */
507 if (!attr_stack)
508 bootstrap_attr_stack();
509
510 /*
511 * Pop the "info" one that is always at the top of the stack.
512 */
513 info = attr_stack;
514 attr_stack = info->prev;
515
516 /*
517 * Pop the ones from directories that are not the prefix of
518 * the path we are checking.
519 */
520 while (attr_stack && attr_stack->origin) {
521 int namelen = strlen(attr_stack->origin);
522
523 elem = attr_stack;
524 if (namelen <= dirlen &&
525 !strncmp(elem->origin, path, namelen))
526 break;
527
528 debug_pop(elem);
529 attr_stack = elem->prev;
530 free_attr_elem(elem);
531 }
532
533 /*
534 * Read from parent directories and push them down
535 */
536 if (!is_bare_repository()) {
537 while (1) {
538 char *cp;
539
540 len = strlen(attr_stack->origin);
541 if (dirlen <= len)
542 break;
543 strbuf_reset(&pathbuf);
544 strbuf_add(&pathbuf, path, dirlen);
545 strbuf_addch(&pathbuf, '/');
546 cp = strchr(pathbuf.buf + len + 1, '/');
547 strcpy(cp + 1, GITATTRIBUTES_FILE);
548 elem = read_attr(pathbuf.buf, 0);
549 *cp = '\0';
550 elem->origin = strdup(pathbuf.buf);
551 elem->prev = attr_stack;
552 attr_stack = elem;
553 debug_push(elem);
554 }
555 }
556
557 /*
558 * Finally push the "info" one at the top of the stack.
559 */
560 info->prev = attr_stack;
561 attr_stack = info;
562}
563
564static int path_matches(const char *pathname, int pathlen,
565 const char *pattern,
566 const char *base, int baselen)
567{
568 if (!strchr(pattern, '/')) {
569 /* match basename */
570 const char *basename = strrchr(pathname, '/');
571 basename = basename ? basename + 1 : pathname;
572 return (fnmatch(pattern, basename, 0) == 0);
573 }
574 /*
575 * match with FNM_PATHNAME; the pattern has base implicitly
576 * in front of it.
577 */
578 if (*pattern == '/')
579 pattern++;
580 if (pathlen < baselen ||
581 (baselen && pathname[baselen] != '/') ||
582 strncmp(pathname, base, baselen))
583 return 0;
584 if (baselen != 0)
585 baselen++;
586 return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
587}
588
589static int fill_one(const char *what, struct match_attr *a, int rem)
590{
591 struct git_attr_check *check = check_all_attr;
592 int i;
593
594 for (i = 0; 0 < rem && i < a->num_attr; i++) {
595 struct git_attr *attr = a->state[i].attr;
596 const char **n = &(check[attr->attr_nr].value);
597 const char *v = a->state[i].setto;
598
599 if (*n == ATTR__UNKNOWN) {
600 debug_set(what, a->u.pattern, attr, v);
601 *n = v;
602 rem--;
603 }
604 }
605 return rem;
606}
607
608static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
609{
610 int i;
611 const char *base = stk->origin ? stk->origin : "";
612
613 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
614 struct match_attr *a = stk->attrs[i];
615 if (a->is_macro)
616 continue;
617 if (path_matches(path, pathlen,
618 a->u.pattern, base, strlen(base)))
619 rem = fill_one("fill", a, rem);
620 }
621 return rem;
622}
623
624static int macroexpand(struct attr_stack *stk, int rem)
625{
626 int i;
627 struct git_attr_check *check = check_all_attr;
628
629 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
630 struct match_attr *a = stk->attrs[i];
631 if (!a->is_macro)
632 continue;
633 if (check[a->u.attr->attr_nr].value != ATTR__TRUE)
634 continue;
635 rem = fill_one("expand", a, rem);
636 }
637 return rem;
638}
639
640int git_checkattr(const char *path, int num, struct git_attr_check *check)
641{
642 struct attr_stack *stk;
643 const char *cp;
644 int dirlen, pathlen, i, rem;
645
646 bootstrap_attr_stack();
647 for (i = 0; i < attr_nr; i++)
648 check_all_attr[i].value = ATTR__UNKNOWN;
649
650 pathlen = strlen(path);
651 cp = strrchr(path, '/');
652 if (!cp)
653 dirlen = 0;
654 else
655 dirlen = cp - path;
656 prepare_attr_stack(path, dirlen);
657 rem = attr_nr;
658 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
659 rem = fill(path, pathlen, stk, rem);
660
661 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
662 rem = macroexpand(stk, rem);
663
664 for (i = 0; i < num; i++) {
665 const char *value = check_all_attr[check[i].attr->attr_nr].value;
666 if (value == ATTR__UNKNOWN)
667 value = ATTR__UNSET;
668 check[i].value = value;
669 }
670
671 return 0;
672}
673
674void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
675{
676 enum git_attr_direction old = direction;
677 direction = new;
678 if (new != old)
679 drop_attr_stack();
680 use_index = istate;
681}