Extract a function collect_all_attrs()
[alt-git.git] / attr.c
blobbc589f0b78994b6d5dd890f3fafb1c29e61394d8
1 #define NO_THE_INDEX_COMPATIBILITY_MACROS
2 #include "cache.h"
3 #include "exec_cmd.h"
4 #include "attr.h"
6 const char git_attr__true[] = "(builtin)true";
7 const char git_attr__false[] = "\0(builtin)false";
8 static const char git_attr__unknown[] = "(builtin)unknown";
9 #define ATTR__TRUE git_attr__true
10 #define ATTR__FALSE git_attr__false
11 #define ATTR__UNSET NULL
12 #define ATTR__UNKNOWN git_attr__unknown
14 static const char *attributes_file;
17 * The basic design decision here is that we are not going to have
18 * insanely large number of attributes.
20 * This is a randomly chosen prime.
22 #define HASHSIZE 257
24 #ifndef DEBUG_ATTR
25 #define DEBUG_ATTR 0
26 #endif
28 struct git_attr {
29 struct git_attr *next;
30 unsigned h;
31 int attr_nr;
32 char name[FLEX_ARRAY];
34 static int attr_nr;
36 static struct git_attr_check *check_all_attr;
37 static struct git_attr *(git_attr_hash[HASHSIZE]);
39 char *git_attr_name(struct git_attr *attr)
41 return attr->name;
44 static unsigned hash_name(const char *name, int namelen)
46 unsigned val = 0, c;
48 while (namelen--) {
49 c = *name++;
50 val = ((val << 7) | (val >> 22)) ^ c;
52 return val;
55 static int invalid_attr_name(const char *name, int namelen)
58 * Attribute name cannot begin with '-' and must consist of
59 * characters from [-A-Za-z0-9_.].
61 if (namelen <= 0 || *name == '-')
62 return -1;
63 while (namelen--) {
64 char ch = *name++;
65 if (! (ch == '-' || ch == '.' || ch == '_' ||
66 ('0' <= ch && ch <= '9') ||
67 ('a' <= ch && ch <= 'z') ||
68 ('A' <= ch && ch <= 'Z')) )
69 return -1;
71 return 0;
74 static struct git_attr *git_attr_internal(const char *name, int len)
76 unsigned hval = hash_name(name, len);
77 unsigned pos = hval % HASHSIZE;
78 struct git_attr *a;
80 for (a = git_attr_hash[pos]; a; a = a->next) {
81 if (a->h == hval &&
82 !memcmp(a->name, name, len) && !a->name[len])
83 return a;
86 if (invalid_attr_name(name, len))
87 return NULL;
89 a = xmalloc(sizeof(*a) + len + 1);
90 memcpy(a->name, name, len);
91 a->name[len] = 0;
92 a->h = hval;
93 a->next = git_attr_hash[pos];
94 a->attr_nr = attr_nr++;
95 git_attr_hash[pos] = a;
97 check_all_attr = xrealloc(check_all_attr,
98 sizeof(*check_all_attr) * attr_nr);
99 check_all_attr[a->attr_nr].attr = a;
100 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
101 return a;
104 struct git_attr *git_attr(const char *name)
106 return git_attr_internal(name, strlen(name));
110 * .gitattributes file is one line per record, each of which is
112 * (1) glob pattern.
113 * (2) whitespace
114 * (3) whitespace separated list of attribute names, each of which
115 * could be prefixed with '-' to mean "set to false", '!' to mean
116 * "unset".
119 /* What does a matched pattern decide? */
120 struct attr_state {
121 struct git_attr *attr;
122 const char *setto;
125 struct match_attr {
126 union {
127 char *pattern;
128 struct git_attr *attr;
129 } u;
130 char is_macro;
131 unsigned num_attr;
132 struct attr_state state[FLEX_ARRAY];
135 static const char blank[] = " \t\r\n";
137 static const char *parse_attr(const char *src, int lineno, const char *cp,
138 int *num_attr, struct match_attr *res)
140 const char *ep, *equals;
141 int len;
143 ep = cp + strcspn(cp, blank);
144 equals = strchr(cp, '=');
145 if (equals && ep < equals)
146 equals = NULL;
147 if (equals)
148 len = equals - cp;
149 else
150 len = ep - cp;
151 if (!res) {
152 if (*cp == '-' || *cp == '!') {
153 cp++;
154 len--;
156 if (invalid_attr_name(cp, len)) {
157 fprintf(stderr,
158 "%.*s is not a valid attribute name: %s:%d\n",
159 len, cp, src, lineno);
160 return NULL;
162 } else {
163 struct attr_state *e;
165 e = &(res->state[*num_attr]);
166 if (*cp == '-' || *cp == '!') {
167 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
168 cp++;
169 len--;
171 else if (!equals)
172 e->setto = ATTR__TRUE;
173 else {
174 e->setto = xmemdupz(equals + 1, ep - equals - 1);
176 e->attr = git_attr_internal(cp, len);
178 (*num_attr)++;
179 return ep + strspn(ep, blank);
182 static struct match_attr *parse_attr_line(const char *line, const char *src,
183 int lineno, int macro_ok)
185 int namelen;
186 int num_attr;
187 const char *cp, *name;
188 struct match_attr *res = NULL;
189 int pass;
190 int is_macro;
192 cp = line + strspn(line, blank);
193 if (!*cp || *cp == '#')
194 return NULL;
195 name = cp;
196 namelen = strcspn(name, blank);
197 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
198 !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
199 if (!macro_ok) {
200 fprintf(stderr, "%s not allowed: %s:%d\n",
201 name, src, lineno);
202 return NULL;
204 is_macro = 1;
205 name += strlen(ATTRIBUTE_MACRO_PREFIX);
206 name += strspn(name, blank);
207 namelen = strcspn(name, blank);
208 if (invalid_attr_name(name, namelen)) {
209 fprintf(stderr,
210 "%.*s is not a valid attribute name: %s:%d\n",
211 namelen, name, src, lineno);
212 return NULL;
215 else
216 is_macro = 0;
218 for (pass = 0; pass < 2; pass++) {
219 /* pass 0 counts and allocates, pass 1 fills */
220 num_attr = 0;
221 cp = name + namelen;
222 cp = cp + strspn(cp, blank);
223 while (*cp) {
224 cp = parse_attr(src, lineno, cp, &num_attr, res);
225 if (!cp)
226 return NULL;
228 if (pass)
229 break;
230 res = xcalloc(1,
231 sizeof(*res) +
232 sizeof(struct attr_state) * num_attr +
233 (is_macro ? 0 : namelen + 1));
234 if (is_macro)
235 res->u.attr = git_attr_internal(name, namelen);
236 else {
237 res->u.pattern = (char *)&(res->state[num_attr]);
238 memcpy(res->u.pattern, name, namelen);
239 res->u.pattern[namelen] = 0;
241 res->is_macro = is_macro;
242 res->num_attr = num_attr;
244 return res;
248 * Like info/exclude and .gitignore, the attribute information can
249 * come from many places.
251 * (1) .gitattribute file of the same directory;
252 * (2) .gitattribute file of the parent directory if (1) does not have
253 * any match; this goes recursively upwards, just like .gitignore.
254 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
256 * In the same file, later entries override the earlier match, so in the
257 * global list, we would have entries from info/attributes the earliest
258 * (reading the file from top to bottom), .gitattribute of the root
259 * directory (again, reading the file from top to bottom) down to the
260 * current directory, and then scan the list backwards to find the first match.
261 * This is exactly the same as what excluded() does in dir.c to deal with
262 * .gitignore
265 static struct attr_stack {
266 struct attr_stack *prev;
267 char *origin;
268 unsigned num_matches;
269 unsigned alloc;
270 struct match_attr **attrs;
271 } *attr_stack;
273 static void free_attr_elem(struct attr_stack *e)
275 int i;
276 free(e->origin);
277 for (i = 0; i < e->num_matches; i++) {
278 struct match_attr *a = e->attrs[i];
279 int j;
280 for (j = 0; j < a->num_attr; j++) {
281 const char *setto = a->state[j].setto;
282 if (setto == ATTR__TRUE ||
283 setto == ATTR__FALSE ||
284 setto == ATTR__UNSET ||
285 setto == ATTR__UNKNOWN)
287 else
288 free((char *) setto);
290 free(a);
292 free(e);
295 static const char *builtin_attr[] = {
296 "[attr]binary -diff -text",
297 NULL,
300 static void handle_attr_line(struct attr_stack *res,
301 const char *line,
302 const char *src,
303 int lineno,
304 int macro_ok)
306 struct match_attr *a;
308 a = parse_attr_line(line, src, lineno, macro_ok);
309 if (!a)
310 return;
311 if (res->alloc <= res->num_matches) {
312 res->alloc = alloc_nr(res->num_matches);
313 res->attrs = xrealloc(res->attrs,
314 sizeof(struct match_attr *) *
315 res->alloc);
317 res->attrs[res->num_matches++] = a;
320 static struct attr_stack *read_attr_from_array(const char **list)
322 struct attr_stack *res;
323 const char *line;
324 int lineno = 0;
326 res = xcalloc(1, sizeof(*res));
327 while ((line = *(list++)) != NULL)
328 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
329 return res;
332 static enum git_attr_direction direction;
333 static struct index_state *use_index;
335 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
337 FILE *fp = fopen(path, "r");
338 struct attr_stack *res;
339 char buf[2048];
340 int lineno = 0;
342 if (!fp)
343 return NULL;
344 res = xcalloc(1, sizeof(*res));
345 while (fgets(buf, sizeof(buf), fp))
346 handle_attr_line(res, buf, path, ++lineno, macro_ok);
347 fclose(fp);
348 return res;
351 static void *read_index_data(const char *path)
353 int pos, len;
354 unsigned long sz;
355 enum object_type type;
356 void *data;
357 struct index_state *istate = use_index ? use_index : &the_index;
359 len = strlen(path);
360 pos = index_name_pos(istate, path, len);
361 if (pos < 0) {
363 * We might be in the middle of a merge, in which
364 * case we would read stage #2 (ours).
366 int i;
367 for (i = -pos - 1;
368 (pos < 0 && i < istate->cache_nr &&
369 !strcmp(istate->cache[i]->name, path));
370 i++)
371 if (ce_stage(istate->cache[i]) == 2)
372 pos = i;
374 if (pos < 0)
375 return NULL;
376 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
377 if (!data || type != OBJ_BLOB) {
378 free(data);
379 return NULL;
381 return data;
384 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
386 struct attr_stack *res;
387 char *buf, *sp;
388 int lineno = 0;
390 buf = read_index_data(path);
391 if (!buf)
392 return NULL;
394 res = xcalloc(1, sizeof(*res));
395 for (sp = buf; *sp; ) {
396 char *ep;
397 int more;
398 for (ep = sp; *ep && *ep != '\n'; ep++)
400 more = (*ep == '\n');
401 *ep = '\0';
402 handle_attr_line(res, sp, path, ++lineno, macro_ok);
403 sp = ep + more;
405 free(buf);
406 return res;
409 static struct attr_stack *read_attr(const char *path, int macro_ok)
411 struct attr_stack *res;
413 if (direction == GIT_ATTR_CHECKOUT) {
414 res = read_attr_from_index(path, macro_ok);
415 if (!res)
416 res = read_attr_from_file(path, macro_ok);
418 else if (direction == GIT_ATTR_CHECKIN) {
419 res = read_attr_from_file(path, macro_ok);
420 if (!res)
422 * There is no checked out .gitattributes file there, but
423 * we might have it in the index. We allow operation in a
424 * sparsely checked out work tree, so read from it.
426 res = read_attr_from_index(path, macro_ok);
428 else
429 res = read_attr_from_index(path, macro_ok);
430 if (!res)
431 res = xcalloc(1, sizeof(*res));
432 return res;
435 #if DEBUG_ATTR
436 static void debug_info(const char *what, struct attr_stack *elem)
438 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
440 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
442 const char *value = v;
444 if (ATTR_TRUE(value))
445 value = "set";
446 else if (ATTR_FALSE(value))
447 value = "unset";
448 else if (ATTR_UNSET(value))
449 value = "unspecified";
451 fprintf(stderr, "%s: %s => %s (%s)\n",
452 what, attr->name, (char *) value, match);
454 #define debug_push(a) debug_info("push", (a))
455 #define debug_pop(a) debug_info("pop", (a))
456 #else
457 #define debug_push(a) do { ; } while (0)
458 #define debug_pop(a) do { ; } while (0)
459 #define debug_set(a,b,c,d) do { ; } while (0)
460 #endif
462 static void drop_attr_stack(void)
464 while (attr_stack) {
465 struct attr_stack *elem = attr_stack;
466 attr_stack = elem->prev;
467 free_attr_elem(elem);
471 static const char *git_etc_gitattributes(void)
473 static const char *system_wide;
474 if (!system_wide)
475 system_wide = system_path(ETC_GITATTRIBUTES);
476 return system_wide;
479 static int git_attr_system(void)
481 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
484 static int git_attr_config(const char *var, const char *value, void *dummy)
486 if (!strcmp(var, "core.attributesfile"))
487 return git_config_pathname(&attributes_file, var, value);
489 return 0;
492 static void bootstrap_attr_stack(void)
494 if (!attr_stack) {
495 struct attr_stack *elem;
497 elem = read_attr_from_array(builtin_attr);
498 elem->origin = NULL;
499 elem->prev = attr_stack;
500 attr_stack = elem;
502 if (git_attr_system()) {
503 elem = read_attr_from_file(git_etc_gitattributes(), 1);
504 if (elem) {
505 elem->origin = NULL;
506 elem->prev = attr_stack;
507 attr_stack = elem;
511 git_config(git_attr_config, NULL);
512 if (attributes_file) {
513 elem = read_attr_from_file(attributes_file, 1);
514 if (elem) {
515 elem->origin = NULL;
516 elem->prev = attr_stack;
517 attr_stack = elem;
521 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
522 elem = read_attr(GITATTRIBUTES_FILE, 1);
523 elem->origin = strdup("");
524 elem->prev = attr_stack;
525 attr_stack = elem;
526 debug_push(elem);
529 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
530 if (!elem)
531 elem = xcalloc(1, sizeof(*elem));
532 elem->origin = NULL;
533 elem->prev = attr_stack;
534 attr_stack = elem;
538 static void prepare_attr_stack(const char *path)
540 struct attr_stack *elem, *info;
541 int dirlen, len;
542 struct strbuf pathbuf;
543 const char *cp;
545 cp = strrchr(path, '/');
546 if (!cp)
547 dirlen = 0;
548 else
549 dirlen = cp - path;
551 strbuf_init(&pathbuf, dirlen+2+strlen(GITATTRIBUTES_FILE));
554 * At the bottom of the attribute stack is the built-in
555 * set of attribute definitions, followed by the contents
556 * of $(prefix)/etc/gitattributes and a file specified by
557 * core.attributesfile. Then, contents from
558 * .gitattribute files from directories closer to the
559 * root to the ones in deeper directories are pushed
560 * to the stack. Finally, at the very top of the stack
561 * we always keep the contents of $GIT_DIR/info/attributes.
563 * When checking, we use entries from near the top of the
564 * stack, preferring $GIT_DIR/info/attributes, then
565 * .gitattributes in deeper directories to shallower ones,
566 * and finally use the built-in set as the default.
568 if (!attr_stack)
569 bootstrap_attr_stack();
572 * Pop the "info" one that is always at the top of the stack.
574 info = attr_stack;
575 attr_stack = info->prev;
578 * Pop the ones from directories that are not the prefix of
579 * the path we are checking.
581 while (attr_stack && attr_stack->origin) {
582 int namelen = strlen(attr_stack->origin);
584 elem = attr_stack;
585 if (namelen <= dirlen &&
586 !strncmp(elem->origin, path, namelen))
587 break;
589 debug_pop(elem);
590 attr_stack = elem->prev;
591 free_attr_elem(elem);
595 * Read from parent directories and push them down
597 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
598 while (1) {
599 char *cp;
601 len = strlen(attr_stack->origin);
602 if (dirlen <= len)
603 break;
604 strbuf_reset(&pathbuf);
605 strbuf_add(&pathbuf, path, dirlen);
606 strbuf_addch(&pathbuf, '/');
607 cp = strchr(pathbuf.buf + len + 1, '/');
608 strcpy(cp + 1, GITATTRIBUTES_FILE);
609 elem = read_attr(pathbuf.buf, 0);
610 *cp = '\0';
611 elem->origin = strdup(pathbuf.buf);
612 elem->prev = attr_stack;
613 attr_stack = elem;
614 debug_push(elem);
618 strbuf_release(&pathbuf);
621 * Finally push the "info" one at the top of the stack.
623 info->prev = attr_stack;
624 attr_stack = info;
627 static int path_matches(const char *pathname, int pathlen,
628 const char *pattern,
629 const char *base, int baselen)
631 if (!strchr(pattern, '/')) {
632 /* match basename */
633 const char *basename = strrchr(pathname, '/');
634 basename = basename ? basename + 1 : pathname;
635 return (fnmatch(pattern, basename, 0) == 0);
638 * match with FNM_PATHNAME; the pattern has base implicitly
639 * in front of it.
641 if (*pattern == '/')
642 pattern++;
643 if (pathlen < baselen ||
644 (baselen && pathname[baselen] != '/') ||
645 strncmp(pathname, base, baselen))
646 return 0;
647 if (baselen != 0)
648 baselen++;
649 return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
652 static int macroexpand_one(int attr_nr, int rem);
654 static int fill_one(const char *what, struct match_attr *a, int rem)
656 struct git_attr_check *check = check_all_attr;
657 int i;
659 for (i = a->num_attr - 1; 0 < rem && 0 <= i; i--) {
660 struct git_attr *attr = a->state[i].attr;
661 const char **n = &(check[attr->attr_nr].value);
662 const char *v = a->state[i].setto;
664 if (*n == ATTR__UNKNOWN) {
665 debug_set(what,
666 a->is_macro ? a->u.attr->name : a->u.pattern,
667 attr, v);
668 *n = v;
669 rem--;
670 rem = macroexpand_one(attr->attr_nr, rem);
673 return rem;
676 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
678 int i;
679 const char *base = stk->origin ? stk->origin : "";
681 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
682 struct match_attr *a = stk->attrs[i];
683 if (a->is_macro)
684 continue;
685 if (path_matches(path, pathlen,
686 a->u.pattern, base, strlen(base)))
687 rem = fill_one("fill", a, rem);
689 return rem;
692 static int macroexpand_one(int attr_nr, int rem)
694 struct attr_stack *stk;
695 struct match_attr *a = NULL;
696 int i;
698 if (check_all_attr[attr_nr].value != ATTR__TRUE)
699 return rem;
701 for (stk = attr_stack; !a && stk; stk = stk->prev)
702 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
703 struct match_attr *ma = stk->attrs[i];
704 if (!ma->is_macro)
705 continue;
706 if (ma->u.attr->attr_nr == attr_nr)
707 a = ma;
710 if (a)
711 rem = fill_one("expand", a, rem);
713 return rem;
717 * Collect all attributes for path into the array pointed to by
718 * check_all_attr.
720 static void collect_all_attrs(const char *path)
722 struct attr_stack *stk;
723 int i, pathlen, rem;
725 bootstrap_attr_stack();
726 prepare_attr_stack(path);
727 for (i = 0; i < attr_nr; i++)
728 check_all_attr[i].value = ATTR__UNKNOWN;
730 pathlen = strlen(path);
731 rem = attr_nr;
732 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
733 rem = fill(path, pathlen, stk, rem);
736 int git_checkattr(const char *path, int num, struct git_attr_check *check)
738 int i;
740 collect_all_attrs(path);
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;
749 return 0;
752 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
754 enum git_attr_direction old = direction;
756 if (is_bare_repository() && new != GIT_ATTR_INDEX)
757 die("BUG: non-INDEX attr direction in a bare repo");
759 direction = new;
760 if (new != old)
761 drop_attr_stack();
762 use_index = istate;