bash: rename __git_find_subcommand() to __git_find_on_cmdline()
[git/spearce.git] / attr.c
blob55bdb7cdebea7f7ea551231fe7801f272d128d69
1 #define NO_THE_INDEX_COMPATIBILITY_MACROS
2 #include "cache.h"
3 #include "attr.h"
5 const char git_attr__true[] = "(builtin)true";
6 const char git_attr__false[] = "\0(builtin)false";
7 static 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
14 * The basic design decision here is that we are not going to have
15 * insanely large number of attributes.
17 * This is a randomly chosen prime.
19 #define HASHSIZE 257
21 #ifndef DEBUG_ATTR
22 #define DEBUG_ATTR 0
23 #endif
25 struct git_attr {
26 struct git_attr *next;
27 unsigned h;
28 int attr_nr;
29 char name[FLEX_ARRAY];
31 static int attr_nr;
33 static struct git_attr_check *check_all_attr;
34 static struct git_attr *(git_attr_hash[HASHSIZE]);
36 static unsigned hash_name(const char *name, int namelen)
38 unsigned val = 0, c;
40 while (namelen--) {
41 c = *name++;
42 val = ((val << 7) | (val >> 22)) ^ c;
44 return val;
47 static int invalid_attr_name(const char *name, int namelen)
50 * Attribute name cannot begin with '-' and from
51 * [-A-Za-z0-9_.]. We'd specifically exclude '=' for now,
52 * as we might later want to allow non-binary value for
53 * attributes, e.g. "*.svg merge=special-merge-program-for-svg"
55 if (*name == '-')
56 return -1;
57 while (namelen--) {
58 char ch = *name++;
59 if (! (ch == '-' || ch == '.' || ch == '_' ||
60 ('0' <= ch && ch <= '9') ||
61 ('a' <= ch && ch <= 'z') ||
62 ('A' <= ch && ch <= 'Z')) )
63 return -1;
65 return 0;
68 struct git_attr *git_attr(const char *name, int len)
70 unsigned hval = hash_name(name, len);
71 unsigned pos = hval % HASHSIZE;
72 struct git_attr *a;
74 for (a = git_attr_hash[pos]; a; a = a->next) {
75 if (a->h == hval &&
76 !memcmp(a->name, name, len) && !a->name[len])
77 return a;
80 if (invalid_attr_name(name, len))
81 return NULL;
83 a = xmalloc(sizeof(*a) + len + 1);
84 memcpy(a->name, name, len);
85 a->name[len] = 0;
86 a->h = hval;
87 a->next = git_attr_hash[pos];
88 a->attr_nr = attr_nr++;
89 git_attr_hash[pos] = a;
91 check_all_attr = xrealloc(check_all_attr,
92 sizeof(*check_all_attr) * attr_nr);
93 check_all_attr[a->attr_nr].attr = a;
94 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
95 return a;
99 * .gitattributes file is one line per record, each of which is
101 * (1) glob pattern.
102 * (2) whitespace
103 * (3) whitespace separated list of attribute names, each of which
104 * could be prefixed with '-' to mean "set to false", '!' to mean
105 * "unset".
108 /* What does a matched pattern decide? */
109 struct attr_state {
110 struct git_attr *attr;
111 const char *setto;
114 struct match_attr {
115 union {
116 char *pattern;
117 struct git_attr *attr;
118 } u;
119 char is_macro;
120 unsigned num_attr;
121 struct attr_state state[FLEX_ARRAY];
124 static const char blank[] = " \t\r\n";
126 static const char *parse_attr(const char *src, int lineno, const char *cp,
127 int *num_attr, struct match_attr *res)
129 const char *ep, *equals;
130 int len;
132 ep = cp + strcspn(cp, blank);
133 equals = strchr(cp, '=');
134 if (equals && ep < equals)
135 equals = NULL;
136 if (equals)
137 len = equals - cp;
138 else
139 len = ep - cp;
140 if (!res) {
141 if (*cp == '-' || *cp == '!') {
142 cp++;
143 len--;
145 if (invalid_attr_name(cp, len)) {
146 fprintf(stderr,
147 "%.*s is not a valid attribute name: %s:%d\n",
148 len, cp, src, lineno);
149 return NULL;
151 } else {
152 struct attr_state *e;
154 e = &(res->state[*num_attr]);
155 if (*cp == '-' || *cp == '!') {
156 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
157 cp++;
158 len--;
160 else if (!equals)
161 e->setto = ATTR__TRUE;
162 else {
163 e->setto = xmemdupz(equals + 1, ep - equals - 1);
165 e->attr = git_attr(cp, len);
167 (*num_attr)++;
168 return ep + strspn(ep, blank);
171 static struct match_attr *parse_attr_line(const char *line, const char *src,
172 int lineno, int macro_ok)
174 int namelen;
175 int num_attr;
176 const char *cp, *name;
177 struct match_attr *res = NULL;
178 int pass;
179 int is_macro;
181 cp = line + strspn(line, blank);
182 if (!*cp || *cp == '#')
183 return NULL;
184 name = cp;
185 namelen = strcspn(name, blank);
186 if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
187 !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
188 if (!macro_ok) {
189 fprintf(stderr, "%s not allowed: %s:%d\n",
190 name, src, lineno);
191 return NULL;
193 is_macro = 1;
194 name += strlen(ATTRIBUTE_MACRO_PREFIX);
195 name += strspn(name, blank);
196 namelen = strcspn(name, blank);
197 if (invalid_attr_name(name, namelen)) {
198 fprintf(stderr,
199 "%.*s is not a valid attribute name: %s:%d\n",
200 namelen, name, src, lineno);
201 return NULL;
204 else
205 is_macro = 0;
207 for (pass = 0; pass < 2; pass++) {
208 /* pass 0 counts and allocates, pass 1 fills */
209 num_attr = 0;
210 cp = name + namelen;
211 cp = cp + strspn(cp, blank);
212 while (*cp) {
213 cp = parse_attr(src, lineno, cp, &num_attr, res);
214 if (!cp)
215 return NULL;
217 if (pass)
218 break;
219 res = xcalloc(1,
220 sizeof(*res) +
221 sizeof(struct attr_state) * num_attr +
222 (is_macro ? 0 : namelen + 1));
223 if (is_macro)
224 res->u.attr = git_attr(name, namelen);
225 else {
226 res->u.pattern = (char *)&(res->state[num_attr]);
227 memcpy(res->u.pattern, name, namelen);
228 res->u.pattern[namelen] = 0;
230 res->is_macro = is_macro;
231 res->num_attr = num_attr;
233 return res;
237 * Like info/exclude and .gitignore, the attribute information can
238 * come from many places.
240 * (1) .gitattribute file of the same directory;
241 * (2) .gitattribute file of the parent directory if (1) does not have
242 * any match; this goes recursively upwards, just like .gitignore.
243 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
245 * In the same file, later entries override the earlier match, so in the
246 * global list, we would have entries from info/attributes the earliest
247 * (reading the file from top to bottom), .gitattribute of the root
248 * directory (again, reading the file from top to bottom) down to the
249 * current directory, and then scan the list backwards to find the first match.
250 * This is exactly the same as what excluded() does in dir.c to deal with
251 * .gitignore
254 static struct attr_stack {
255 struct attr_stack *prev;
256 char *origin;
257 unsigned num_matches;
258 unsigned alloc;
259 struct match_attr **attrs;
260 } *attr_stack;
262 static void free_attr_elem(struct attr_stack *e)
264 int i;
265 free(e->origin);
266 for (i = 0; i < e->num_matches; i++) {
267 struct match_attr *a = e->attrs[i];
268 int j;
269 for (j = 0; j < a->num_attr; j++) {
270 const char *setto = a->state[j].setto;
271 if (setto == ATTR__TRUE ||
272 setto == ATTR__FALSE ||
273 setto == ATTR__UNSET ||
274 setto == ATTR__UNKNOWN)
276 else
277 free((char *) setto);
279 free(a);
281 free(e);
284 static const char *builtin_attr[] = {
285 "[attr]binary -diff -crlf",
286 NULL,
289 static void handle_attr_line(struct attr_stack *res,
290 const char *line,
291 const char *src,
292 int lineno,
293 int macro_ok)
295 struct match_attr *a;
297 a = parse_attr_line(line, src, lineno, macro_ok);
298 if (!a)
299 return;
300 if (res->alloc <= res->num_matches) {
301 res->alloc = alloc_nr(res->num_matches);
302 res->attrs = xrealloc(res->attrs,
303 sizeof(struct match_attr *) *
304 res->alloc);
306 res->attrs[res->num_matches++] = a;
309 static struct attr_stack *read_attr_from_array(const char **list)
311 struct attr_stack *res;
312 const char *line;
313 int lineno = 0;
315 res = xcalloc(1, sizeof(*res));
316 while ((line = *(list++)) != NULL)
317 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
318 return res;
321 static enum git_attr_direction direction;
322 static struct index_state *use_index;
324 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
326 FILE *fp = fopen(path, "r");
327 struct attr_stack *res;
328 char buf[2048];
329 int lineno = 0;
331 if (!fp)
332 return NULL;
333 res = xcalloc(1, sizeof(*res));
334 while (fgets(buf, sizeof(buf), fp))
335 handle_attr_line(res, buf, path, ++lineno, macro_ok);
336 fclose(fp);
337 return res;
340 static void *read_index_data(const char *path)
342 int pos, len;
343 unsigned long sz;
344 enum object_type type;
345 void *data;
346 struct index_state *istate = use_index ? use_index : &the_index;
348 len = strlen(path);
349 pos = index_name_pos(istate, path, len);
350 if (pos < 0) {
352 * We might be in the middle of a merge, in which
353 * case we would read stage #2 (ours).
355 int i;
356 for (i = -pos - 1;
357 (pos < 0 && i < istate->cache_nr &&
358 !strcmp(istate->cache[i]->name, path));
359 i++)
360 if (ce_stage(istate->cache[i]) == 2)
361 pos = i;
363 if (pos < 0)
364 return NULL;
365 data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
366 if (!data || type != OBJ_BLOB) {
367 free(data);
368 return NULL;
370 return data;
373 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
375 struct attr_stack *res;
376 char *buf, *sp;
377 int lineno = 0;
379 buf = read_index_data(path);
380 if (!buf)
381 return NULL;
383 res = xcalloc(1, sizeof(*res));
384 for (sp = buf; *sp; ) {
385 char *ep;
386 int more;
387 for (ep = sp; *ep && *ep != '\n'; ep++)
389 more = (*ep == '\n');
390 *ep = '\0';
391 handle_attr_line(res, sp, path, ++lineno, macro_ok);
392 sp = ep + more;
394 free(buf);
395 return res;
398 static struct attr_stack *read_attr(const char *path, int macro_ok)
400 struct attr_stack *res;
402 if (direction == GIT_ATTR_CHECKOUT) {
403 res = read_attr_from_index(path, macro_ok);
404 if (!res)
405 res = read_attr_from_file(path, macro_ok);
407 else if (direction == GIT_ATTR_CHECKIN) {
408 res = read_attr_from_file(path, macro_ok);
409 if (!res)
411 * There is no checked out .gitattributes file there, but
412 * we might have it in the index. We allow operation in a
413 * sparsely checked out work tree, so read from it.
415 res = read_attr_from_index(path, macro_ok);
417 else
418 res = read_attr_from_index(path, macro_ok);
419 if (!res)
420 res = xcalloc(1, sizeof(*res));
421 return res;
424 #if DEBUG_ATTR
425 static void debug_info(const char *what, struct attr_stack *elem)
427 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
429 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
431 const char *value = v;
433 if (ATTR_TRUE(value))
434 value = "set";
435 else if (ATTR_FALSE(value))
436 value = "unset";
437 else if (ATTR_UNSET(value))
438 value = "unspecified";
440 fprintf(stderr, "%s: %s => %s (%s)\n",
441 what, attr->name, (char *) value, match);
443 #define debug_push(a) debug_info("push", (a))
444 #define debug_pop(a) debug_info("pop", (a))
445 #else
446 #define debug_push(a) do { ; } while (0)
447 #define debug_pop(a) do { ; } while (0)
448 #define debug_set(a,b,c,d) do { ; } while (0)
449 #endif
451 static void drop_attr_stack(void)
453 while (attr_stack) {
454 struct attr_stack *elem = attr_stack;
455 attr_stack = elem->prev;
456 free_attr_elem(elem);
460 static void bootstrap_attr_stack(void)
462 if (!attr_stack) {
463 struct attr_stack *elem;
465 elem = read_attr_from_array(builtin_attr);
466 elem->origin = NULL;
467 elem->prev = attr_stack;
468 attr_stack = elem;
470 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
471 elem = read_attr(GITATTRIBUTES_FILE, 1);
472 elem->origin = strdup("");
473 elem->prev = attr_stack;
474 attr_stack = elem;
475 debug_push(elem);
478 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
479 if (!elem)
480 elem = xcalloc(1, sizeof(*elem));
481 elem->origin = NULL;
482 elem->prev = attr_stack;
483 attr_stack = elem;
487 static void prepare_attr_stack(const char *path, int dirlen)
489 struct attr_stack *elem, *info;
490 int len;
491 struct strbuf pathbuf;
493 strbuf_init(&pathbuf, dirlen+2+strlen(GITATTRIBUTES_FILE));
496 * At the bottom of the attribute stack is the built-in
497 * set of attribute definitions. Then, contents from
498 * .gitattribute files from directories closer to the
499 * root to the ones in deeper directories are pushed
500 * to the stack. Finally, at the very top of the stack
501 * we always keep the contents of $GIT_DIR/info/attributes.
503 * When checking, we use entries from near the top of the
504 * stack, preferring $GIT_DIR/info/attributes, then
505 * .gitattributes in deeper directories to shallower ones,
506 * and finally use the built-in set as the default.
508 if (!attr_stack)
509 bootstrap_attr_stack();
512 * Pop the "info" one that is always at the top of the stack.
514 info = attr_stack;
515 attr_stack = info->prev;
518 * Pop the ones from directories that are not the prefix of
519 * the path we are checking.
521 while (attr_stack && attr_stack->origin) {
522 int namelen = strlen(attr_stack->origin);
524 elem = attr_stack;
525 if (namelen <= dirlen &&
526 !strncmp(elem->origin, path, namelen))
527 break;
529 debug_pop(elem);
530 attr_stack = elem->prev;
531 free_attr_elem(elem);
535 * Read from parent directories and push them down
537 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
538 while (1) {
539 char *cp;
541 len = strlen(attr_stack->origin);
542 if (dirlen <= len)
543 break;
544 strbuf_reset(&pathbuf);
545 strbuf_add(&pathbuf, path, dirlen);
546 strbuf_addch(&pathbuf, '/');
547 cp = strchr(pathbuf.buf + len + 1, '/');
548 strcpy(cp + 1, GITATTRIBUTES_FILE);
549 elem = read_attr(pathbuf.buf, 0);
550 *cp = '\0';
551 elem->origin = strdup(pathbuf.buf);
552 elem->prev = attr_stack;
553 attr_stack = elem;
554 debug_push(elem);
558 strbuf_release(&pathbuf);
561 * Finally push the "info" one at the top of the stack.
563 info->prev = attr_stack;
564 attr_stack = info;
567 static int path_matches(const char *pathname, int pathlen,
568 const char *pattern,
569 const char *base, int baselen)
571 if (!strchr(pattern, '/')) {
572 /* match basename */
573 const char *basename = strrchr(pathname, '/');
574 basename = basename ? basename + 1 : pathname;
575 return (fnmatch(pattern, basename, 0) == 0);
578 * match with FNM_PATHNAME; the pattern has base implicitly
579 * in front of it.
581 if (*pattern == '/')
582 pattern++;
583 if (pathlen < baselen ||
584 (baselen && pathname[baselen] != '/') ||
585 strncmp(pathname, base, baselen))
586 return 0;
587 if (baselen != 0)
588 baselen++;
589 return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
592 static int fill_one(const char *what, struct match_attr *a, int rem)
594 struct git_attr_check *check = check_all_attr;
595 int i;
597 for (i = 0; 0 < rem && i < a->num_attr; i++) {
598 struct git_attr *attr = a->state[i].attr;
599 const char **n = &(check[attr->attr_nr].value);
600 const char *v = a->state[i].setto;
602 if (*n == ATTR__UNKNOWN) {
603 debug_set(what, a->u.pattern, attr, v);
604 *n = v;
605 rem--;
608 return rem;
611 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
613 int i;
614 const char *base = stk->origin ? stk->origin : "";
616 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
617 struct match_attr *a = stk->attrs[i];
618 if (a->is_macro)
619 continue;
620 if (path_matches(path, pathlen,
621 a->u.pattern, base, strlen(base)))
622 rem = fill_one("fill", a, rem);
624 return rem;
627 static int macroexpand(struct attr_stack *stk, int rem)
629 int i;
630 struct git_attr_check *check = check_all_attr;
632 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
633 struct match_attr *a = stk->attrs[i];
634 if (!a->is_macro)
635 continue;
636 if (check[a->u.attr->attr_nr].value != ATTR__TRUE)
637 continue;
638 rem = fill_one("expand", a, rem);
640 return rem;
643 int git_checkattr(const char *path, int num, struct git_attr_check *check)
645 struct attr_stack *stk;
646 const char *cp;
647 int dirlen, pathlen, i, rem;
649 bootstrap_attr_stack();
650 for (i = 0; i < attr_nr; i++)
651 check_all_attr[i].value = ATTR__UNKNOWN;
653 pathlen = strlen(path);
654 cp = strrchr(path, '/');
655 if (!cp)
656 dirlen = 0;
657 else
658 dirlen = cp - path;
659 prepare_attr_stack(path, dirlen);
660 rem = attr_nr;
661 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
662 rem = fill(path, pathlen, stk, rem);
664 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
665 rem = macroexpand(stk, rem);
667 for (i = 0; i < num; i++) {
668 const char *value = check_all_attr[check[i].attr->attr_nr].value;
669 if (value == ATTR__UNKNOWN)
670 value = ATTR__UNSET;
671 check[i].value = value;
674 return 0;
677 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
679 enum git_attr_direction old = direction;
681 if (is_bare_repository() && new != GIT_ATTR_INDEX)
682 die("BUG: non-INDEX attr direction in a bare repo");
684 direction = new;
685 if (new != old)
686 drop_attr_stack();
687 use_index = istate;