attr.c

   1 #include "cache.h"
   2 #include "attr.h"
   3
   4 const char git_attr__true[] = "(builtin)true";
   5 const char git_attr__false[] = "\0(builtin)false";
   6 static 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
  11
  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
  24 struct git_attr {
  25         struct git_attr *next;
  26         unsigned h;
  27         int attr_nr;
  28         char name[FLEX_ARRAY];
  29 };
  30 static int attr_nr;
  31
  32 static struct git_attr_check *check_all_attr;
  33 static struct git_attr *(git_attr_hash[HASHSIZE]);
  34
  35 static 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
  47 static 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
  68 struct 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
  80         if (invalid_attr_name(name, len))
  81                 return NULL;
  82
  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;
  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;
  94         check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
  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
 104  *     could be prefixed with '-' to mean "set to false", '!' to mean
 105  *     "unset".
 106  */
 107
 108 /* What does a matched pattern decide? */
 109 struct attr_state {
 110         struct git_attr *attr;
 111         const char *setto;
 112 };
 113
 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];
 122 };
 123
 124 static const char blank[] = " \t\r\n";
 125
 126 static 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 {
 163                         e->setto = xmemdupz(equals + 1, ep - equals - 1);
 164                 }
 165                 e->attr = git_attr(cp, len);
 166         }
 167         (*num_attr)++;
 168         return ep + strspn(ep, blank);
 169 }
 170
 171 static struct match_attr *parse_attr_line(const char *line, const char *src,
 172                                           int lineno, int macro_ok)
 173 {
 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;
 180
 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;
 192                 }
 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;
 202                 }
 203         }
 204         else
 205                 is_macro = 0;
 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);
 212                 while (*cp) {
 213                         cp = parse_attr(src, lineno, cp, &num_attr, res);
 214                         if (!cp)
 215                                 return NULL;
 216                 }
 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;
 229                 }
 230                 res->is_macro = is_macro;
 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;
 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.
 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
 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;
 261
 262 static void free_attr_elem(struct attr_stack *e)
 263 {
 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)
 275                                 ;
 276                         else
 277                                 free((char*) setto);
 278                 }
 279                 free(a);
 280         }
 281         free(e);
 282 }
 283
 284 static const char *builtin_attr[] = {
 285         "[attr]binary -diff -crlf",
 286         NULL,
 287 };
 288
 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)
 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
 309 static struct attr_stack *read_attr_from_array(const char **list)
 310 {
 311         struct attr_stack *res;
 312         const char *line;
 313         int lineno = 0;
 314
 315         res = xcalloc(1, sizeof(*res));
 316         while ((line = *(list++)) != NULL)
 317                 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
 318         return res;
 319 }
 320
 321 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
 322 {
 323         FILE *fp = fopen(path, "r");
 324         struct attr_stack *res;
 325         char buf[2048];
 326         int lineno = 0;
 327
 328         if (!fp)
 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 }
 336
 337 static 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
 369 static struct attr_stack *read_attr(const char *path, int macro_ok)
 370 {
 371         struct attr_stack *res;
 372         char *buf, *sp;
 373         int lineno = 0;
 374
 375         res = read_attr_from_file(path, macro_ok);
 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);
 401         return res;
 402 }
 403
 404 #if DEBUG_ATTR
 405 static void debug_info(const char *what, struct attr_stack *elem)
 406 {
 407         fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
 408 }
 409 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
 410 {
 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);
 422 }
 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)
 428 #define debug_set(a,b,c,d) do { ; } while (0)
 429 #endif
 430
 431 static 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
 441                 if (!is_bare_repository()) {
 442                         elem = read_attr(GITATTRIBUTES_FILE, 1);
 443                         elem->origin = strdup("");
 444                         elem->prev = attr_stack;
 445                         attr_stack = elem;
 446                         debug_push(elem);
 447                 }
 448
 449                 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
 450                 if (!elem)
 451                         elem = xcalloc(1, sizeof(*elem));
 452                 elem->origin = NULL;
 453                 elem->prev = attr_stack;
 454                 attr_stack = elem;
 455         }
 456 }
 457
 458 static void prepare_attr_stack(const char *path, int dirlen)
 459 {
 460         struct attr_stack *elem, *info;
 461         int len;
 462         struct strbuf pathbuf;
 463
 464         strbuf_init(&pathbuf, dirlen+2+strlen(GITATTRIBUTES_FILE));
 465
 466         /*
 467          * At the bottom of the attribute stack is the built-in
 468          * set of attribute definitions.  Then, contents from
 469          * .gitattribute files from directories closer to the
 470          * root to the ones in deeper directories are pushed
 471          * to the stack.  Finally, at the very top of the stack
 472          * we always keep the contents of $GIT_DIR/info/attributes.
 473          *
 474          * When checking, we use entries from near the top of the
 475          * stack, preferring $GIT_DIR/info/attributes, then
 476          * .gitattributes in deeper directories to shallower ones,
 477          * and finally use the built-in set as the default.
 478          */
 479         if (!attr_stack)
 480                 bootstrap_attr_stack();
 481
 482         /*
 483          * Pop the "info" one that is always at the top of the stack.
 484          */
 485         info = attr_stack;
 486         attr_stack = info->prev;
 487
 488         /*
 489          * Pop the ones from directories that are not the prefix of
 490          * the path we are checking.
 491          */
 492         while (attr_stack && attr_stack->origin) {
 493                 int namelen = strlen(attr_stack->origin);
 494
 495                 elem = attr_stack;
 496                 if (namelen <= dirlen &&
 497                     !strncmp(elem->origin, path, namelen))
 498                         break;
 499
 500                 debug_pop(elem);
 501                 attr_stack = elem->prev;
 502                 free_attr_elem(elem);
 503         }
 504
 505         /*
 506          * Read from parent directories and push them down
 507          */
 508         if (!is_bare_repository()) {
 509                 while (1) {
 510                         char *cp;
 511
 512                         len = strlen(attr_stack->origin);
 513                         if (dirlen <= len)
 514                                 break;
 515                         strbuf_reset(&pathbuf);
 516                         strbuf_add(&pathbuf, path, dirlen);
 517                         strbuf_addch(&pathbuf, '/');
 518                         cp = strchr(pathbuf.buf + len + 1, '/');
 519                         strcpy(cp + 1, GITATTRIBUTES_FILE);
 520                         elem = read_attr(pathbuf.buf, 0);
 521                         *cp = '\0';
 522                         elem->origin = strdup(pathbuf.buf);
 523                         elem->prev = attr_stack;
 524                         attr_stack = elem;
 525                         debug_push(elem);
 526                 }
 527         }
 528
 529         /*
 530          * Finally push the "info" one at the top of the stack.
 531          */
 532         info->prev = attr_stack;
 533         attr_stack = info;
 534 }
 535
 536 static int path_matches(const char *pathname, int pathlen,
 537                         const char *pattern,
 538                         const char *base, int baselen)
 539 {
 540         if (!strchr(pattern, '/')) {
 541                 /* match basename */
 542                 const char *basename = strrchr(pathname, '/');
 543                 basename = basename ? basename + 1 : pathname;
 544                 return (fnmatch(pattern, basename, 0) == 0);
 545         }
 546         /*
 547          * match with FNM_PATHNAME; the pattern has base implicitly
 548          * in front of it.
 549          */
 550         if (*pattern == '/')
 551                 pattern++;
 552         if (pathlen < baselen ||
 553             (baselen && pathname[baselen] != '/') ||
 554             strncmp(pathname, base, baselen))
 555                 return 0;
 556         if (baselen != 0)
 557                 baselen++;
 558         return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
 559 }
 560
 561 static int fill_one(const char *what, struct match_attr *a, int rem)
 562 {
 563         struct git_attr_check *check = check_all_attr;
 564         int i;
 565
 566         for (i = 0; 0 < rem && i < a->num_attr; i++) {
 567                 struct git_attr *attr = a->state[i].attr;
 568                 const char **n = &(check[attr->attr_nr].value);
 569                 const char *v = a->state[i].setto;
 570
 571                 if (*n == ATTR__UNKNOWN) {
 572                         debug_set(what, a->u.pattern, attr, v);
 573                         *n = v;
 574                         rem--;
 575                 }
 576         }
 577         return rem;
 578 }
 579
 580 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
 581 {
 582         int i;
 583         const char *base = stk->origin ? stk->origin : "";
 584
 585         for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 586                 struct match_attr *a = stk->attrs[i];
 587                 if (a->is_macro)
 588                         continue;
 589                 if (path_matches(path, pathlen,
 590                                  a->u.pattern, base, strlen(base)))
 591                         rem = fill_one("fill", a, rem);
 592         }
 593         return rem;
 594 }
 595
 596 static int macroexpand(struct attr_stack *stk, int rem)
 597 {
 598         int i;
 599         struct git_attr_check *check = check_all_attr;
 600
 601         for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 602                 struct match_attr *a = stk->attrs[i];
 603                 if (!a->is_macro)
 604                         continue;
 605                 if (check[a->u.attr->attr_nr].value != ATTR__TRUE)
 606                         continue;
 607                 rem = fill_one("expand", a, rem);
 608         }
 609         return rem;
 610 }
 611
 612 int git_checkattr(const char *path, int num, struct git_attr_check *check)
 613 {
 614         struct attr_stack *stk;
 615         const char *cp;
 616         int dirlen, pathlen, i, rem;
 617
 618         bootstrap_attr_stack();
 619         for (i = 0; i < attr_nr; i++)
 620                 check_all_attr[i].value = ATTR__UNKNOWN;
 621
 622         pathlen = strlen(path);
 623         cp = strrchr(path, '/');
 624         if (!cp)
 625                 dirlen = 0;
 626         else
 627                 dirlen = cp - path;
 628         prepare_attr_stack(path, dirlen);
 629         rem = attr_nr;
 630         for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 631                 rem = fill(path, pathlen, stk, rem);
 632
 633         for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 634                 rem = macroexpand(stk, rem);
 635
 636         for (i = 0; i < num; i++) {
 637                 const char *value = check_all_attr[check[i].attr->attr_nr].value;
 638                 if (value == ATTR__UNKNOWN)
 639                         value = ATTR__UNSET;
 640                 check[i].value = value;
 641         }
 642
 643         return 0;
 644 }