Disallow the empty string as an attribute name
[alt-git.git] / attr.c
blobb1d1d6d791687afc98e273db983b8f8727fcd424
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 static unsigned hash_name(const char *name, int namelen)
41 unsigned val = 0, c;
43 while (namelen--) {
44 c = *name++;
45 val = ((val << 7) | (val >> 22)) ^ c;
47 return val;
50 static int invalid_attr_name(const char *name, int namelen)
53 * Attribute name cannot begin with '-' and must consist of
54 * characters from [-A-Za-z0-9_.].
56 if (namelen <= 0 || *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;
66 return 0;
69 static struct git_attr *git_attr_internal(const char *name, int len)
71 unsigned hval = hash_name(name, len);
72 unsigned pos = hval % HASHSIZE;
73 struct git_attr *a;
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;
81 if (invalid_attr_name(name, len))
82 return NULL;
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;
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;
99 struct git_attr *git_attr(const char *name)
101 return git_attr_internal(name, strlen(name));
105 * .gitattributes file is one line per record, each of which is
107 * (1) glob pattern.
108 * (2) whitespace
109 * (3) whitespace separated list of attribute names, each of which
110 * could be prefixed with '-' to mean "set to false", '!' to mean
111 * "unset".
114 /* What does a matched pattern decide? */
115 struct attr_state {
116 struct git_attr *attr;
117 const char *setto;
120 struct match_attr {
121 union {
122 char *pattern;
123 struct git_attr *attr;
124 } u;
125 char is_macro;
126 unsigned num_attr;
127 struct attr_state state[FLEX_ARRAY];
130 static const char blank[] = " \t\r\n";
132 static const char *parse_attr(const char *src, int lineno, const char *cp,
133 int *num_attr, struct match_attr *res)
135 const char *ep, *equals;
136 int len;
138 ep = cp + strcspn(cp, blank);
139 equals = strchr(cp, '=');
140 if (equals && ep < equals)
141 equals = NULL;
142 if (equals)
143 len = equals - cp;
144 else
145 len = ep - cp;
146 if (!res) {
147 if (*cp == '-' || *cp == '!') {
148 cp++;
149 len--;
151 if (invalid_attr_name(cp, len)) {
152 fprintf(stderr,
153 "%.*s is not a valid attribute name: %s:%d\n",
154 len, cp, src, lineno);
155 return NULL;
157 } else {
158 struct attr_state *e;
160 e = &(res->state[*num_attr]);
161 if (*cp == '-' || *cp == '!') {
162 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
163 cp++;
164 len--;
166 else if (!equals)
167 e->setto = ATTR__TRUE;
168 else {
169 e->setto = xmemdupz(equals + 1, ep - equals - 1);
171 e->attr = git_attr_internal(cp, len);
173 (*num_attr)++;
174 return ep + strspn(ep, blank);
177 static struct match_attr *parse_attr_line(const char *line, const char *src,
178 int lineno, int macro_ok)
180 int namelen;
181 int num_attr;
182 const char *cp, *name;
183 struct match_attr *res = NULL;
184 int pass;
185 int is_macro;
187 cp = line + strspn(line, blank);
188 if (!*cp || *cp == '#')
189 return NULL;
190 name = cp;
191 namelen = strcspn(name, blank);
192 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
193 !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
194 if (!macro_ok) {
195 fprintf(stderr, "%s not allowed: %s:%d\n",
196 name, src, lineno);
197 return NULL;
199 is_macro = 1;
200 name += strlen(ATTRIBUTE_MACRO_PREFIX);
201 name += strspn(name, blank);
202 namelen = strcspn(name, blank);
203 if (invalid_attr_name(name, namelen)) {
204 fprintf(stderr,
205 "%.*s is not a valid attribute name: %s:%d\n",
206 namelen, name, src, lineno);
207 return NULL;
210 else
211 is_macro = 0;
213 for (pass = 0; pass < 2; pass++) {
214 /* pass 0 counts and allocates, pass 1 fills */
215 num_attr = 0;
216 cp = name + namelen;
217 cp = cp + strspn(cp, blank);
218 while (*cp) {
219 cp = parse_attr(src, lineno, cp, &num_attr, res);
220 if (!cp)
221 return NULL;
223 if (pass)
224 break;
225 res = xcalloc(1,
226 sizeof(*res) +
227 sizeof(struct attr_state) * num_attr +
228 (is_macro ? 0 : namelen + 1));
229 if (is_macro)
230 res->u.attr = git_attr_internal(name, namelen);
231 else {
232 res->u.pattern = (char *)&(res->state[num_attr]);
233 memcpy(res->u.pattern, name, namelen);
234 res->u.pattern[namelen] = 0;
236 res->is_macro = is_macro;
237 res->num_attr = num_attr;
239 return res;
243 * Like info/exclude and .gitignore, the attribute information can
244 * come from many places.
246 * (1) .gitattribute file of the same directory;
247 * (2) .gitattribute file of the parent directory if (1) does not have
248 * any match; this goes recursively upwards, just like .gitignore.
249 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
251 * In the same file, later entries override the earlier match, so in the
252 * global list, we would have entries from info/attributes the earliest
253 * (reading the file from top to bottom), .gitattribute of the root
254 * directory (again, reading the file from top to bottom) down to the
255 * current directory, and then scan the list backwards to find the first match.
256 * This is exactly the same as what excluded() does in dir.c to deal with
257 * .gitignore
260 static struct attr_stack {
261 struct attr_stack *prev;
262 char *origin;
263 unsigned num_matches;
264 unsigned alloc;
265 struct match_attr **attrs;
266 } *attr_stack;
268 static void free_attr_elem(struct attr_stack *e)
270 int i;
271 free(e->origin);
272 for (i = 0; i < e->num_matches; i++) {
273 struct match_attr *a = e->attrs[i];
274 int j;
275 for (j = 0; j < a->num_attr; j++) {
276 const char *setto = a->state[j].setto;
277 if (setto == ATTR__TRUE ||
278 setto == ATTR__FALSE ||
279 setto == ATTR__UNSET ||
280 setto == ATTR__UNKNOWN)
282 else
283 free((char *) setto);
285 free(a);
287 free(e);
290 static const char *builtin_attr[] = {
291 "[attr]binary -diff -text",
292 NULL,
295 static void handle_attr_line(struct attr_stack *res,
296 const char *line,
297 const char *src,
298 int lineno,
299 int macro_ok)
301 struct match_attr *a;
303 a = parse_attr_line(line, src, lineno, macro_ok);
304 if (!a)
305 return;
306 if (res->alloc <= res->num_matches) {
307 res->alloc = alloc_nr(res->num_matches);
308 res->attrs = xrealloc(res->attrs,
309 sizeof(struct match_attr *) *
310 res->alloc);
312 res->attrs[res->num_matches++] = a;
315 static struct attr_stack *read_attr_from_array(const char **list)
317 struct attr_stack *res;
318 const char *line;
319 int lineno = 0;
321 res = xcalloc(1, sizeof(*res));
322 while ((line = *(list++)) != NULL)
323 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
324 return res;
327 static enum git_attr_direction direction;
328 static struct index_state *use_index;
330 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
332 FILE *fp = fopen(path, "r");
333 struct attr_stack *res;
334 char buf[2048];
335 int lineno = 0;
337 if (!fp)
338 return NULL;
339 res = xcalloc(1, sizeof(*res));
340 while (fgets(buf, sizeof(buf), fp))
341 handle_attr_line(res, buf, path, ++lineno, macro_ok);
342 fclose(fp);
343 return res;
346 static void *read_index_data(const char *path)
348 int pos, len;
349 unsigned long sz;
350 enum object_type type;
351 void *data;
352 struct index_state *istate = use_index ? use_index : &the_index;
354 len = strlen(path);
355 pos = index_name_pos(istate, path, len);
356 if (pos < 0) {
358 * We might be in the middle of a merge, in which
359 * case we would read stage #2 (ours).
361 int i;
362 for (i = -pos - 1;
363 (pos < 0 && i < istate->cache_nr &&
364 !strcmp(istate->cache[i]->name, path));
365 i++)
366 if (ce_stage(istate->cache[i]) == 2)
367 pos = i;
369 if (pos < 0)
370 return NULL;
371 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
372 if (!data || type != OBJ_BLOB) {
373 free(data);
374 return NULL;
376 return data;
379 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
381 struct attr_stack *res;
382 char *buf, *sp;
383 int lineno = 0;
385 buf = read_index_data(path);
386 if (!buf)
387 return NULL;
389 res = xcalloc(1, sizeof(*res));
390 for (sp = buf; *sp; ) {
391 char *ep;
392 int more;
393 for (ep = sp; *ep && *ep != '\n'; ep++)
395 more = (*ep == '\n');
396 *ep = '\0';
397 handle_attr_line(res, sp, path, ++lineno, macro_ok);
398 sp = ep + more;
400 free(buf);
401 return res;
404 static struct attr_stack *read_attr(const char *path, int macro_ok)
406 struct attr_stack *res;
408 if (direction == GIT_ATTR_CHECKOUT) {
409 res = read_attr_from_index(path, macro_ok);
410 if (!res)
411 res = read_attr_from_file(path, macro_ok);
413 else if (direction == GIT_ATTR_CHECKIN) {
414 res = read_attr_from_file(path, macro_ok);
415 if (!res)
417 * There is no checked out .gitattributes file there, but
418 * we might have it in the index. We allow operation in a
419 * sparsely checked out work tree, so read from it.
421 res = read_attr_from_index(path, macro_ok);
423 else
424 res = read_attr_from_index(path, macro_ok);
425 if (!res)
426 res = xcalloc(1, sizeof(*res));
427 return res;
430 #if DEBUG_ATTR
431 static void debug_info(const char *what, struct attr_stack *elem)
433 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
435 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
437 const char *value = v;
439 if (ATTR_TRUE(value))
440 value = "set";
441 else if (ATTR_FALSE(value))
442 value = "unset";
443 else if (ATTR_UNSET(value))
444 value = "unspecified";
446 fprintf(stderr, "%s: %s => %s (%s)\n",
447 what, attr->name, (char *) value, match);
449 #define debug_push(a) debug_info("push", (a))
450 #define debug_pop(a) debug_info("pop", (a))
451 #else
452 #define debug_push(a) do { ; } while (0)
453 #define debug_pop(a) do { ; } while (0)
454 #define debug_set(a,b,c,d) do { ; } while (0)
455 #endif
457 static void drop_attr_stack(void)
459 while (attr_stack) {
460 struct attr_stack *elem = attr_stack;
461 attr_stack = elem->prev;
462 free_attr_elem(elem);
466 static const char *git_etc_gitattributes(void)
468 static const char *system_wide;
469 if (!system_wide)
470 system_wide = system_path(ETC_GITATTRIBUTES);
471 return system_wide;
474 static int git_attr_system(void)
476 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
479 static int git_attr_config(const char *var, const char *value, void *dummy)
481 if (!strcmp(var, "core.attributesfile"))
482 return git_config_pathname(&attributes_file, var, value);
484 return 0;
487 static void bootstrap_attr_stack(void)
489 if (!attr_stack) {
490 struct attr_stack *elem;
492 elem = read_attr_from_array(builtin_attr);
493 elem->origin = NULL;
494 elem->prev = attr_stack;
495 attr_stack = elem;
497 if (git_attr_system()) {
498 elem = read_attr_from_file(git_etc_gitattributes(), 1);
499 if (elem) {
500 elem->origin = NULL;
501 elem->prev = attr_stack;
502 attr_stack = elem;
506 git_config(git_attr_config, NULL);
507 if (attributes_file) {
508 elem = read_attr_from_file(attributes_file, 1);
509 if (elem) {
510 elem->origin = NULL;
511 elem->prev = attr_stack;
512 attr_stack = elem;
516 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
517 elem = read_attr(GITATTRIBUTES_FILE, 1);
518 elem->origin = strdup("");
519 elem->prev = attr_stack;
520 attr_stack = elem;
521 debug_push(elem);
524 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
525 if (!elem)
526 elem = xcalloc(1, sizeof(*elem));
527 elem->origin = NULL;
528 elem->prev = attr_stack;
529 attr_stack = elem;
533 static void prepare_attr_stack(const char *path, int dirlen)
535 struct attr_stack *elem, *info;
536 int len;
537 struct strbuf pathbuf;
539 strbuf_init(&pathbuf, dirlen+2+strlen(GITATTRIBUTES_FILE));
542 * At the bottom of the attribute stack is the built-in
543 * set of attribute definitions, followed by the contents
544 * of $(prefix)/etc/gitattributes and a file specified by
545 * core.attributesfile. Then, contents from
546 * .gitattribute files from directories closer to the
547 * root to the ones in deeper directories are pushed
548 * to the stack. Finally, at the very top of the stack
549 * we always keep the contents of $GIT_DIR/info/attributes.
551 * When checking, we use entries from near the top of the
552 * stack, preferring $GIT_DIR/info/attributes, then
553 * .gitattributes in deeper directories to shallower ones,
554 * and finally use the built-in set as the default.
556 if (!attr_stack)
557 bootstrap_attr_stack();
560 * Pop the "info" one that is always at the top of the stack.
562 info = attr_stack;
563 attr_stack = info->prev;
566 * Pop the ones from directories that are not the prefix of
567 * the path we are checking.
569 while (attr_stack && attr_stack->origin) {
570 int namelen = strlen(attr_stack->origin);
572 elem = attr_stack;
573 if (namelen <= dirlen &&
574 !strncmp(elem->origin, path, namelen))
575 break;
577 debug_pop(elem);
578 attr_stack = elem->prev;
579 free_attr_elem(elem);
583 * Read from parent directories and push them down
585 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
586 while (1) {
587 char *cp;
589 len = strlen(attr_stack->origin);
590 if (dirlen <= len)
591 break;
592 strbuf_reset(&pathbuf);
593 strbuf_add(&pathbuf, path, dirlen);
594 strbuf_addch(&pathbuf, '/');
595 cp = strchr(pathbuf.buf + len + 1, '/');
596 strcpy(cp + 1, GITATTRIBUTES_FILE);
597 elem = read_attr(pathbuf.buf, 0);
598 *cp = '\0';
599 elem->origin = strdup(pathbuf.buf);
600 elem->prev = attr_stack;
601 attr_stack = elem;
602 debug_push(elem);
606 strbuf_release(&pathbuf);
609 * Finally push the "info" one at the top of the stack.
611 info->prev = attr_stack;
612 attr_stack = info;
615 static int path_matches(const char *pathname, int pathlen,
616 const char *pattern,
617 const char *base, int baselen)
619 if (!strchr(pattern, '/')) {
620 /* match basename */
621 const char *basename = strrchr(pathname, '/');
622 basename = basename ? basename + 1 : pathname;
623 return (fnmatch(pattern, basename, 0) == 0);
626 * match with FNM_PATHNAME; the pattern has base implicitly
627 * in front of it.
629 if (*pattern == '/')
630 pattern++;
631 if (pathlen < baselen ||
632 (baselen && pathname[baselen] != '/') ||
633 strncmp(pathname, base, baselen))
634 return 0;
635 if (baselen != 0)
636 baselen++;
637 return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
640 static int macroexpand_one(int attr_nr, int rem);
642 static int fill_one(const char *what, struct match_attr *a, int rem)
644 struct git_attr_check *check = check_all_attr;
645 int i;
647 for (i = a->num_attr - 1; 0 < rem && 0 <= i; i--) {
648 struct git_attr *attr = a->state[i].attr;
649 const char **n = &(check[attr->attr_nr].value);
650 const char *v = a->state[i].setto;
652 if (*n == ATTR__UNKNOWN) {
653 debug_set(what,
654 a->is_macro ? a->u.attr->name : a->u.pattern,
655 attr, v);
656 *n = v;
657 rem--;
658 rem = macroexpand_one(attr->attr_nr, rem);
661 return rem;
664 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
666 int i;
667 const char *base = stk->origin ? stk->origin : "";
669 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
670 struct match_attr *a = stk->attrs[i];
671 if (a->is_macro)
672 continue;
673 if (path_matches(path, pathlen,
674 a->u.pattern, base, strlen(base)))
675 rem = fill_one("fill", a, rem);
677 return rem;
680 static int macroexpand_one(int attr_nr, int rem)
682 struct attr_stack *stk;
683 struct match_attr *a = NULL;
684 int i;
686 if (check_all_attr[attr_nr].value != ATTR__TRUE)
687 return rem;
689 for (stk = attr_stack; !a && stk; stk = stk->prev)
690 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
691 struct match_attr *ma = stk->attrs[i];
692 if (!ma->is_macro)
693 continue;
694 if (ma->u.attr->attr_nr == attr_nr)
695 a = ma;
698 if (a)
699 rem = fill_one("expand", a, rem);
701 return rem;
704 int git_checkattr(const char *path, int num, struct git_attr_check *check)
706 struct attr_stack *stk;
707 const char *cp;
708 int dirlen, pathlen, i, rem;
710 bootstrap_attr_stack();
711 for (i = 0; i < attr_nr; i++)
712 check_all_attr[i].value = ATTR__UNKNOWN;
714 pathlen = strlen(path);
715 cp = strrchr(path, '/');
716 if (!cp)
717 dirlen = 0;
718 else
719 dirlen = cp - path;
720 prepare_attr_stack(path, dirlen);
721 rem = attr_nr;
722 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
723 rem = fill(path, pathlen, stk, rem);
725 for (i = 0; i < num; i++) {
726 const char *value = check_all_attr[check[i].attr->attr_nr].value;
727 if (value == ATTR__UNKNOWN)
728 value = ATTR__UNSET;
729 check[i].value = value;
732 return 0;
735 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
737 enum git_attr_direction old = direction;
739 if (is_bare_repository() && new != GIT_ATTR_INDEX)
740 die("BUG: non-INDEX attr direction in a bare repo");
742 direction = new;
743 if (new != old)
744 drop_attr_stack();
745 use_index = istate;