Port to 12 other Platforms.
[git/git.git] / attr.c
CommitLineData
d0bfd026
JH
1#include "cache.h"
2#include "attr.h"
3
a5e92abd
JH
4const char git_attr__true[] = "(builtin)true";
5const char git_attr__false[] = "\0(builtin)false";
6static const char git_attr__unknown[] = "(builtin)unknown";
7#define ATTR__TRUE git_attr__true
8#define ATTR__FALSE git_attr__false
9#define ATTR__UNSET NULL
10#define ATTR__UNKNOWN git_attr__unknown
515106fa 11
d0bfd026
JH
12/*
13 * The basic design decision here is that we are not going to have
14 * insanely large number of attributes.
15 *
16 * This is a randomly chosen prime.
17 */
18#define HASHSIZE 257
19
20#ifndef DEBUG_ATTR
21#define DEBUG_ATTR 0
22#endif
23
24struct git_attr {
25 struct git_attr *next;
26 unsigned h;
f48fd688 27 int attr_nr;
d0bfd026
JH
28 char name[FLEX_ARRAY];
29};
f48fd688 30static int attr_nr;
d0bfd026 31
f48fd688 32static struct git_attr_check *check_all_attr;
d0bfd026
JH
33static struct git_attr *(git_attr_hash[HASHSIZE]);
34
35static unsigned hash_name(const char *name, int namelen)
36{
37 unsigned val = 0;
38 unsigned char c;
39
40 while (namelen--) {
41 c = *name++;
42 val = ((val << 7) | (val >> 22)) ^ c;
43 }
44 return val;
45}
46
e4aee10a
JH
47static int invalid_attr_name(const char *name, int namelen)
48{
49 /*
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"
54 */
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;
64 }
65 return 0;
66}
67
d0bfd026
JH
68struct git_attr *git_attr(const char *name, int len)
69{
70 unsigned hval = hash_name(name, len);
71 unsigned pos = hval % HASHSIZE;
72 struct git_attr *a;
73
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;
78 }
79
e4aee10a
JH
80 if (invalid_attr_name(name, len))
81 return NULL;
82
d0bfd026
JH
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];
f48fd688 88 a->attr_nr = attr_nr++;
d0bfd026 89 git_attr_hash[pos] = a;
f48fd688
JH
90
91 check_all_attr = xrealloc(check_all_attr,
92 sizeof(*check_all_attr) * attr_nr);
93 check_all_attr[a->attr_nr].attr = a;
515106fa 94 check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
d0bfd026
JH
95 return a;
96}
97
98/*
99 * .gitattributes file is one line per record, each of which is
100 *
101 * (1) glob pattern.
102 * (2) whitespace
103 * (3) whitespace separated list of attribute names, each of which
515106fa
JH
104 * could be prefixed with '-' to mean "set to false", '!' to mean
105 * "unset".
d0bfd026
JH
106 */
107
515106fa 108/* What does a matched pattern decide? */
d0bfd026 109struct attr_state {
d0bfd026 110 struct git_attr *attr;
a5e92abd 111 const char *setto;
d0bfd026
JH
112};
113
114struct match_attr {
f48fd688
JH
115 union {
116 char *pattern;
117 struct git_attr *attr;
118 } u;
119 char is_macro;
d0bfd026
JH
120 unsigned num_attr;
121 struct attr_state state[FLEX_ARRAY];
122};
123
124static const char blank[] = " \t\r\n";
125
515106fa
JH
126static const char *parse_attr(const char *src, int lineno, const char *cp,
127 int *num_attr, struct match_attr *res)
128{
129 const char *ep, *equals;
130 int len;
131
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--;
144 }
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;
150 }
151 } else {
152 struct attr_state *e;
153
154 e = &(res->state[*num_attr]);
155 if (*cp == '-' || *cp == '!') {
156 e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
157 cp++;
158 len--;
159 }
160 else if (!equals)
161 e->setto = ATTR__TRUE;
162 else {
182af834 163 e->setto = xmemdupz(equals + 1, ep - equals - 1);
515106fa
JH
164 }
165 e->attr = git_attr(cp, len);
166 }
167 (*num_attr)++;
168 return ep + strspn(ep, blank);
169}
170
f48fd688
JH
171static struct match_attr *parse_attr_line(const char *line, const char *src,
172 int lineno, int macro_ok)
d0bfd026
JH
173{
174 int namelen;
175 int num_attr;
176 const char *cp, *name;
515106fa 177 struct match_attr *res = NULL;
d0bfd026 178 int pass;
f48fd688 179 int is_macro;
d0bfd026
JH
180
181 cp = line + strspn(line, blank);
182 if (!*cp || *cp == '#')
183 return NULL;
184 name = cp;
185 namelen = strcspn(name, blank);
f48fd688
JH
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;
192 }
193 is_macro = 1;
194 name += strlen(ATTRIBUTE_MACRO_PREFIX);
195 name += strspn(name, blank);
196 namelen = strcspn(name, blank);
e4aee10a
JH
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;
202 }
f48fd688
JH
203 }
204 else
205 is_macro = 0;
d0bfd026
JH
206
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);
d7b0a093 212 while (*cp) {
515106fa 213 cp = parse_attr(src, lineno, cp, &num_attr, res);
d7b0a093
SP
214 if (!cp)
215 return NULL;
216 }
d0bfd026
JH
217 if (pass)
218 break;
219 res = xcalloc(1,
220 sizeof(*res) +
221 sizeof(struct attr_state) * num_attr +
f48fd688 222 (is_macro ? 0 : namelen + 1));
515106fa 223 if (is_macro)
f48fd688
JH
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;
229 }
230 res->is_macro = is_macro;
d0bfd026
JH
231 res->num_attr = num_attr;
232 }
233 return res;
234}
235
236/*
237 * Like info/exclude and .gitignore, the attribute information can
238 * come from many places.
239 *
240 * (1) .gitattribute file of the same directory;
515106fa
JH
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.
d0bfd026
JH
244 *
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
252 */
253
254static struct attr_stack {
255 struct attr_stack *prev;
256 char *origin;
257 unsigned num_matches;
a4413118 258 unsigned alloc;
d0bfd026
JH
259 struct match_attr **attrs;
260} *attr_stack;
261
262static void free_attr_elem(struct attr_stack *e)
263{
264 int i;
265 free(e->origin);
515106fa
JH
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++) {
a5e92abd 270 const char *setto = a->state[j].setto;
515106fa
JH
271 if (setto == ATTR__TRUE ||
272 setto == ATTR__FALSE ||
273 setto == ATTR__UNSET ||
274 setto == ATTR__UNKNOWN)
275 ;
276 else
a5e92abd 277 free((char*) setto);
515106fa
JH
278 }
279 free(a);
280 }
d0bfd026
JH
281 free(e);
282}
283
284static const char *builtin_attr[] = {
e4aee10a 285 "[attr]binary -diff -crlf",
d0bfd026
JH
286 NULL,
287};
288
a4413118
JH
289static void handle_attr_line(struct attr_stack *res,
290 const char *line,
291 const char *src,
292 int lineno,
293 int macro_ok)
294{
295 struct match_attr *a;
296
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);
305 }
306 res->attrs[res->num_matches++] = a;
307}
308
d0bfd026
JH
309static struct attr_stack *read_attr_from_array(const char **list)
310{
311 struct attr_stack *res;
312 const char *line;
f48fd688 313 int lineno = 0;
d0bfd026
JH
314
315 res = xcalloc(1, sizeof(*res));
a4413118
JH
316 while ((line = *(list++)) != NULL)
317 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
d0bfd026
JH
318 return res;
319}
320
f48fd688 321static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
d0bfd026 322{
a4413118 323 FILE *fp = fopen(path, "r");
d0bfd026
JH
324 struct attr_stack *res;
325 char buf[2048];
f48fd688 326 int lineno = 0;
d0bfd026 327
d0bfd026 328 if (!fp)
a4413118
JH
329 return NULL;
330 res = xcalloc(1, sizeof(*res));
331 while (fgets(buf, sizeof(buf), fp))
332 handle_attr_line(res, buf, path, ++lineno, macro_ok);
333 fclose(fp);
334 return res;
335}
d0bfd026 336
1a9d7e9b
JH
337static void *read_index_data(const char *path)
338{
339 int pos, len;
340 unsigned long sz;
341 enum object_type type;
342 void *data;
343
344 len = strlen(path);
345 pos = cache_name_pos(path, len);
346 if (pos < 0) {
347 /*
348 * We might be in the middle of a merge, in which
349 * case we would read stage #2 (ours).
350 */
351 int i;
352 for (i = -pos - 1;
353 (pos < 0 && i < active_nr &&
354 !strcmp(active_cache[i]->name, path));
355 i++)
356 if (ce_stage(active_cache[i]) == 2)
357 pos = i;
358 }
359 if (pos < 0)
360 return NULL;
361 data = read_sha1_file(active_cache[pos]->sha1, &type, &sz);
362 if (!data || type != OBJ_BLOB) {
363 free(data);
364 return NULL;
365 }
366 return data;
367}
368
a4413118
JH
369static struct attr_stack *read_attr(const char *path, int macro_ok)
370{
371 struct attr_stack *res;
1a9d7e9b
JH
372 char *buf, *sp;
373 int lineno = 0;
f48fd688 374
a4413118 375 res = read_attr_from_file(path, macro_ok);
1a9d7e9b
JH
376 if (res)
377 return res;
378
379 res = xcalloc(1, sizeof(*res));
380
381 /*
382 * There is no checked out .gitattributes file there, but
383 * we might have it in the index. We allow operation in a
384 * sparsely checked out work tree, so read from it.
385 */
386 buf = read_index_data(path);
387 if (!buf)
388 return res;
389
390 for (sp = buf; *sp; ) {
391 char *ep;
392 int more;
393 for (ep = sp; *ep && *ep != '\n'; ep++)
394 ;
395 more = (*ep == '\n');
396 *ep = '\0';
397 handle_attr_line(res, sp, path, ++lineno, macro_ok);
398 sp = ep + more;
399 }
400 free(buf);
d0bfd026
JH
401 return res;
402}
403
404#if DEBUG_ATTR
405static void debug_info(const char *what, struct attr_stack *elem)
406{
407 fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
408}
cf94ccda 409static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
f48fd688 410{
515106fa
JH
411 const char *value = v;
412
413 if (ATTR_TRUE(value))
414 value = "set";
415 else if (ATTR_FALSE(value))
416 value = "unset";
417 else if (ATTR_UNSET(value))
418 value = "unspecified";
419
420 fprintf(stderr, "%s: %s => %s (%s)\n",
421 what, attr->name, (char *) value, match);
f48fd688 422}
d0bfd026
JH
423#define debug_push(a) debug_info("push", (a))
424#define debug_pop(a) debug_info("pop", (a))
425#else
426#define debug_push(a) do { ; } while (0)
427#define debug_pop(a) do { ; } while (0)
f48fd688 428#define debug_set(a,b,c,d) do { ; } while (0)
d0bfd026
JH
429#endif
430
f48fd688
JH
431static void bootstrap_attr_stack(void)
432{
433 if (!attr_stack) {
434 struct attr_stack *elem;
435
436 elem = read_attr_from_array(builtin_attr);
437 elem->origin = NULL;
438 elem->prev = attr_stack;
439 attr_stack = elem;
440
a4413118 441 elem = read_attr(GITATTRIBUTES_FILE, 1);
f48fd688
JH
442 elem->origin = strdup("");
443 elem->prev = attr_stack;
444 attr_stack = elem;
445 debug_push(elem);
446
447 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
a4413118
JH
448 if (!elem)
449 elem = xcalloc(1, sizeof(*elem));
f48fd688
JH
450 elem->origin = NULL;
451 elem->prev = attr_stack;
452 attr_stack = elem;
453 }
454}
455
d0bfd026
JH
456static void prepare_attr_stack(const char *path, int dirlen)
457{
458 struct attr_stack *elem, *info;
459 int len;
460 char pathbuf[PATH_MAX];
461
462 /*
463 * At the bottom of the attribute stack is the built-in
464 * set of attribute definitions. Then, contents from
465 * .gitattribute files from directories closer to the
466 * root to the ones in deeper directories are pushed
467 * to the stack. Finally, at the very top of the stack
468 * we always keep the contents of $GIT_DIR/info/attributes.
469 *
470 * When checking, we use entries from near the top of the
471 * stack, preferring $GIT_DIR/info/attributes, then
472 * .gitattributes in deeper directories to shallower ones,
473 * and finally use the built-in set as the default.
474 */
f48fd688
JH
475 if (!attr_stack)
476 bootstrap_attr_stack();
d0bfd026
JH
477
478 /*
479 * Pop the "info" one that is always at the top of the stack.
480 */
481 info = attr_stack;
482 attr_stack = info->prev;
483
484 /*
485 * Pop the ones from directories that are not the prefix of
486 * the path we are checking.
487 */
488 while (attr_stack && attr_stack->origin) {
489 int namelen = strlen(attr_stack->origin);
490
491 elem = attr_stack;
492 if (namelen <= dirlen &&
493 !strncmp(elem->origin, path, namelen))
494 break;
495
496 debug_pop(elem);
497 attr_stack = elem->prev;
498 free_attr_elem(elem);
499 }
500
501 /*
502 * Read from parent directories and push them down
503 */
504 while (1) {
505 char *cp;
506
507 len = strlen(attr_stack->origin);
508 if (dirlen <= len)
509 break;
510 memcpy(pathbuf, path, dirlen);
511 memcpy(pathbuf + dirlen, "/", 2);
512 cp = strchr(pathbuf + len + 1, '/');
513 strcpy(cp + 1, GITATTRIBUTES_FILE);
a4413118 514 elem = read_attr(pathbuf, 0);
d0bfd026
JH
515 *cp = '\0';
516 elem->origin = strdup(pathbuf);
517 elem->prev = attr_stack;
518 attr_stack = elem;
519 debug_push(elem);
520 }
521
522 /*
523 * Finally push the "info" one at the top of the stack.
524 */
525 info->prev = attr_stack;
526 attr_stack = info;
527}
528
529static int path_matches(const char *pathname, int pathlen,
530 const char *pattern,
531 const char *base, int baselen)
532{
533 if (!strchr(pattern, '/')) {
534 /* match basename */
535 const char *basename = strrchr(pathname, '/');
536 basename = basename ? basename + 1 : pathname;
537 return (fnmatch(pattern, basename, 0) == 0);
538 }
539 /*
540 * match with FNM_PATHNAME; the pattern has base implicitly
541 * in front of it.
542 */
543 if (*pattern == '/')
544 pattern++;
545 if (pathlen < baselen ||
cf94ccda 546 (baselen && pathname[baselen] != '/') ||
d0bfd026
JH
547 strncmp(pathname, base, baselen))
548 return 0;
82881b38
MO
549 if (baselen != 0)
550 baselen++;
551 return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
d0bfd026
JH
552}
553
515106fa
JH
554static int fill_one(const char *what, struct match_attr *a, int rem)
555{
556 struct git_attr_check *check = check_all_attr;
557 int i;
558
559 for (i = 0; 0 < rem && i < a->num_attr; i++) {
560 struct git_attr *attr = a->state[i].attr;
a5e92abd
JH
561 const char **n = &(check[attr->attr_nr].value);
562 const char *v = a->state[i].setto;
515106fa
JH
563
564 if (*n == ATTR__UNKNOWN) {
565 debug_set(what, a->u.pattern, attr, v);
566 *n = v;
567 rem--;
568 }
569 }
570 return rem;
571}
572
f48fd688 573static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
d0bfd026 574{
515106fa 575 int i;
d0bfd026
JH
576 const char *base = stk->origin ? stk->origin : "";
577
578 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
579 struct match_attr *a = stk->attrs[i];
f48fd688
JH
580 if (a->is_macro)
581 continue;
d0bfd026 582 if (path_matches(path, pathlen,
515106fa
JH
583 a->u.pattern, base, strlen(base)))
584 rem = fill_one("fill", a, rem);
d0bfd026
JH
585 }
586 return rem;
587}
588
f48fd688
JH
589static int macroexpand(struct attr_stack *stk, int rem)
590{
515106fa 591 int i;
f48fd688
JH
592 struct git_attr_check *check = check_all_attr;
593
594 for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
595 struct match_attr *a = stk->attrs[i];
596 if (!a->is_macro)
597 continue;
515106fa 598 if (check[a->u.attr->attr_nr].value != ATTR__TRUE)
f48fd688 599 continue;
515106fa 600 rem = fill_one("expand", a, rem);
f48fd688
JH
601 }
602 return rem;
603}
604
d0bfd026
JH
605int git_checkattr(const char *path, int num, struct git_attr_check *check)
606{
607 struct attr_stack *stk;
608 const char *cp;
609 int dirlen, pathlen, i, rem;
610
f48fd688
JH
611 bootstrap_attr_stack();
612 for (i = 0; i < attr_nr; i++)
515106fa 613 check_all_attr[i].value = ATTR__UNKNOWN;
d0bfd026
JH
614
615 pathlen = strlen(path);
616 cp = strrchr(path, '/');
617 if (!cp)
618 dirlen = 0;
619 else
620 dirlen = cp - path;
621 prepare_attr_stack(path, dirlen);
f48fd688
JH
622 rem = attr_nr;
623 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
624 rem = fill(path, pathlen, stk, rem);
625
d0bfd026 626 for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
f48fd688
JH
627 rem = macroexpand(stk, rem);
628
515106fa 629 for (i = 0; i < num; i++) {
a5e92abd 630 const char *value = check_all_attr[check[i].attr->attr_nr].value;
515106fa
JH
631 if (value == ATTR__UNKNOWN)
632 value = ATTR__UNSET;
633 check[i].value = value;
634 }
f48fd688 635
d0bfd026
JH
636 return 0;
637}