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 (pass)
 215                         break;
 216                 res = xcalloc(1,
 217                               sizeof(*res) +
 218                               sizeof(struct attr_state) * num_attr +
 219                               (is_macro ? 0 : namelen + 1));
 220                 if (is_macro)
 221                         res->u.attr = git_attr(name, namelen);
 222                 else {
 223                         res->u.pattern = (char*)&(res->state[num_attr]);
 224                         memcpy(res->u.pattern, name, namelen);
 225                         res->u.pattern[namelen] = 0;
 226                 }
 227                 res->is_macro = is_macro;
 228                 res->num_attr = num_attr;
 229         }
 230         return res;
 231 }
 232
 233 /*
 234  * Like info/exclude and .gitignore, the attribute information can
 235  * come from many places.
 236  *
 237  * (1) .gitattribute file of the same directory;
 238  * (2) .gitattribute file of the parent directory if (1) does not have
 239  *      any match; this goes recursively upwards, just like .gitignore.
 240  * (3) $GIT_DIR/info/attributes, which overrides both of the above.
 241  *
 242  * In the same file, later entries override the earlier match, so in the
 243  * global list, we would have entries from info/attributes the earliest
 244  * (reading the file from top to bottom), .gitattribute of the root
 245  * directory (again, reading the file from top to bottom) down to the
 246  * current directory, and then scan the list backwards to find the first match.
 247  * This is exactly the same as what excluded() does in dir.c to deal with
 248  * .gitignore
 249  */
 250
 251 static struct attr_stack {
 252         struct attr_stack *prev;
 253         char *origin;
 254         unsigned num_matches;
 255         unsigned alloc;
 256         struct match_attr **attrs;
 257 } *attr_stack;
 258
 259 static void free_attr_elem(struct attr_stack *e)
 260 {
 261         int i;
 262         free(e->origin);
 263         for (i = 0; i < e->num_matches; i++) {
 264                 struct match_attr *a = e->attrs[i];
 265                 int j;
 266                 for (j = 0; j < a->num_attr; j++) {
 267                         const char *setto = a->state[j].setto;
 268                         if (setto == ATTR__TRUE ||
 269                             setto == ATTR__FALSE ||
 270                             setto == ATTR__UNSET ||
 271                             setto == ATTR__UNKNOWN)
 272                                 ;
 273                         else
 274                                 free((char*) setto);
 275                 }
 276                 free(a);
 277         }
 278         free(e);
 279 }
 280
 281 static const char *builtin_attr[] = {
 282         "[attr]binary -diff -crlf",
 283         NULL,
 284 };
 285
 286 static void handle_attr_line(struct attr_stack *res,
 287                              const char *line,
 288                              const char *src,
 289                              int lineno,
 290                              int macro_ok)
 291 {
 292         struct match_attr *a;
 293
 294         a = parse_attr_line(line, src, lineno, macro_ok);
 295         if (!a)
 296                 return;
 297         if (res->alloc <= res->num_matches) {
 298                 res->alloc = alloc_nr(res->num_matches);
 299                 res->attrs = xrealloc(res->attrs,
 300                                       sizeof(struct match_attr *) *
 301                                       res->alloc);
 302         }
 303         res->attrs[res->num_matches++] = a;
 304 }
 305
 306 static struct attr_stack *read_attr_from_array(const char **list)
 307 {
 308         struct attr_stack *res;
 309         const char *line;
 310         int lineno = 0;
 311
 312         res = xcalloc(1, sizeof(*res));
 313         while ((line = *(list++)) != NULL)
 314                 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
 315         return res;
 316 }
 317
 318 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
 319 {
 320         FILE *fp = fopen(path, "r");
 321         struct attr_stack *res;
 322         char buf[2048];
 323         int lineno = 0;
 324
 325         if (!fp)
 326                 return NULL;
 327         res = xcalloc(1, sizeof(*res));
 328         while (fgets(buf, sizeof(buf), fp))
 329                 handle_attr_line(res, buf, path, ++lineno, macro_ok);
 330         fclose(fp);
 331         return res;
 332 }
 333
 334 static void *read_index_data(const char *path)
 335 {
 336         int pos, len;
 337         unsigned long sz;
 338         enum object_type type;
 339         void *data;
 340
 341         len = strlen(path);
 342         pos = cache_name_pos(path, len);
 343         if (pos < 0) {
 344                 /*
 345                  * We might be in the middle of a merge, in which
 346                  * case we would read stage #2 (ours).
 347                  */
 348                 int i;
 349                 for (i = -pos - 1;
 350                      (pos < 0 && i < active_nr &&
 351                       !strcmp(active_cache[i]->name, path));
 352                      i++)
 353                         if (ce_stage(active_cache[i]) == 2)
 354                                 pos = i;
 355         }
 356         if (pos < 0)
 357                 return NULL;
 358         data = read_sha1_file(active_cache[pos]->sha1, &type, &sz);
 359         if (!data || type != OBJ_BLOB) {
 360                 free(data);
 361                 return NULL;
 362         }
 363         return data;
 364 }
 365
 366 static struct attr_stack *read_attr(const char *path, int macro_ok)
 367 {
 368         struct attr_stack *res;
 369         char *buf, *sp;
 370         int lineno = 0;
 371
 372         res = read_attr_from_file(path, macro_ok);
 373         if (res)
 374                 return res;
 375
 376         res = xcalloc(1, sizeof(*res));
 377
 378         /*
 379          * There is no checked out .gitattributes file there, but
 380          * we might have it in the index.  We allow operation in a
 381          * sparsely checked out work tree, so read from it.
 382          */
 383         buf = read_index_data(path);
 384         if (!buf)
 385                 return res;
 386
 387         for (sp = buf; *sp; ) {
 388                 char *ep;
 389                 int more;
 390                 for (ep = sp; *ep && *ep != '\n'; ep++)
 391                         ;
 392                 more = (*ep == '\n');
 393                 *ep = '\0';
 394                 handle_attr_line(res, sp, path, ++lineno, macro_ok);
 395                 sp = ep + more;
 396         }
 397         free(buf);
 398         return res;
 399 }
 400
 401 #if DEBUG_ATTR
 402 static void debug_info(const char *what, struct attr_stack *elem)
 403 {
 404         fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
 405 }
 406 static void debug_set(const char *what, const char *match, struct git_attr *attr, void *v)
 407 {
 408         const char *value = v;
 409
 410         if (ATTR_TRUE(value))
 411                 value = "set";
 412         else if (ATTR_FALSE(value))
 413                 value = "unset";
 414         else if (ATTR_UNSET(value))
 415                 value = "unspecified";
 416
 417         fprintf(stderr, "%s: %s => %s (%s)\n",
 418                 what, attr->name, (char *) value, match);
 419 }
 420 #define debug_push(a) debug_info("push", (a))
 421 #define debug_pop(a) debug_info("pop", (a))
 422 #else
 423 #define debug_push(a) do { ; } while (0)
 424 #define debug_pop(a) do { ; } while (0)
 425 #define debug_set(a,b,c,d) do { ; } while (0)
 426 #endif
 427
 428 static void bootstrap_attr_stack(void)
 429 {
 430         if (!attr_stack) {
 431                 struct attr_stack *elem;
 432
 433                 elem = read_attr_from_array(builtin_attr);
 434                 elem->origin = NULL;
 435                 elem->prev = attr_stack;
 436                 attr_stack = elem;
 437
 438                 elem = read_attr(GITATTRIBUTES_FILE, 1);
 439                 elem->origin = strdup("");
 440                 elem->prev = attr_stack;
 441                 attr_stack = elem;
 442                 debug_push(elem);
 443
 444                 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
 445                 if (!elem)
 446                         elem = xcalloc(1, sizeof(*elem));
 447                 elem->origin = NULL;
 448                 elem->prev = attr_stack;
 449                 attr_stack = elem;
 450         }
 451 }
 452
 453 static void prepare_attr_stack(const char *path, int dirlen)
 454 {
 455         struct attr_stack *elem, *info;
 456         int len;
 457         char pathbuf[PATH_MAX];
 458
 459         /*
 460          * At the bottom of the attribute stack is the built-in
 461          * set of attribute definitions.  Then, contents from
 462          * .gitattribute files from directories closer to the
 463          * root to the ones in deeper directories are pushed
 464          * to the stack.  Finally, at the very top of the stack
 465          * we always keep the contents of $GIT_DIR/info/attributes.
 466          *
 467          * When checking, we use entries from near the top of the
 468          * stack, preferring $GIT_DIR/info/attributes, then
 469          * .gitattributes in deeper directories to shallower ones,
 470          * and finally use the built-in set as the default.
 471          */
 472         if (!attr_stack)
 473                 bootstrap_attr_stack();
 474
 475         /*
 476          * Pop the "info" one that is always at the top of the stack.
 477          */
 478         info = attr_stack;
 479         attr_stack = info->prev;
 480
 481         /*
 482          * Pop the ones from directories that are not the prefix of
 483          * the path we are checking.
 484          */
 485         while (attr_stack && attr_stack->origin) {
 486                 int namelen = strlen(attr_stack->origin);
 487
 488                 elem = attr_stack;
 489                 if (namelen <= dirlen &&
 490                     !strncmp(elem->origin, path, namelen))
 491                         break;
 492
 493                 debug_pop(elem);
 494                 attr_stack = elem->prev;
 495                 free_attr_elem(elem);
 496         }
 497
 498         /*
 499          * Read from parent directories and push them down
 500          */
 501         while (1) {
 502                 char *cp;
 503
 504                 len = strlen(attr_stack->origin);
 505                 if (dirlen <= len)
 506                         break;
 507                 memcpy(pathbuf, path, dirlen);
 508                 memcpy(pathbuf + dirlen, "/", 2);
 509                 cp = strchr(pathbuf + len + 1, '/');
 510                 strcpy(cp + 1, GITATTRIBUTES_FILE);
 511                 elem = read_attr(pathbuf, 0);
 512                 *cp = '\0';
 513                 elem->origin = strdup(pathbuf);
 514                 elem->prev = attr_stack;
 515                 attr_stack = elem;
 516                 debug_push(elem);
 517         }
 518
 519         /*
 520          * Finally push the "info" one at the top of the stack.
 521          */
 522         info->prev = attr_stack;
 523         attr_stack = info;
 524 }
 525
 526 static int path_matches(const char *pathname, int pathlen,
 527                         const char *pattern,
 528                         const char *base, int baselen)
 529 {
 530         if (!strchr(pattern, '/')) {
 531                 /* match basename */
 532                 const char *basename = strrchr(pathname, '/');
 533                 basename = basename ? basename + 1 : pathname;
 534                 return (fnmatch(pattern, basename, 0) == 0);
 535         }
 536         /*
 537          * match with FNM_PATHNAME; the pattern has base implicitly
 538          * in front of it.
 539          */
 540         if (*pattern == '/')
 541                 pattern++;
 542         if (pathlen < baselen ||
 543             (baselen && pathname[baselen - 1] != '/') ||
 544             strncmp(pathname, base, baselen))
 545                 return 0;
 546         return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
 547 }
 548
 549 static int fill_one(const char *what, struct match_attr *a, int rem)
 550 {
 551         struct git_attr_check *check = check_all_attr;
 552         int i;
 553
 554         for (i = 0; 0 < rem && i < a->num_attr; i++) {
 555                 struct git_attr *attr = a->state[i].attr;
 556                 const char **n = &(check[attr->attr_nr].value);
 557                 const char *v = a->state[i].setto;
 558
 559                 if (*n == ATTR__UNKNOWN) {
 560                         debug_set(what, a->u.pattern, attr, v);
 561                         *n = v;
 562                         rem--;
 563                 }
 564         }
 565         return rem;
 566 }
 567
 568 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
 569 {
 570         int i;
 571         const char *base = stk->origin ? stk->origin : "";
 572
 573         for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 574                 struct match_attr *a = stk->attrs[i];
 575                 if (a->is_macro)
 576                         continue;
 577                 if (path_matches(path, pathlen,
 578                                  a->u.pattern, base, strlen(base)))
 579                         rem = fill_one("fill", a, rem);
 580         }
 581         return rem;
 582 }
 583
 584 static int macroexpand(struct attr_stack *stk, int rem)
 585 {
 586         int i;
 587         struct git_attr_check *check = check_all_attr;
 588
 589         for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 590                 struct match_attr *a = stk->attrs[i];
 591                 if (!a->is_macro)
 592                         continue;
 593                 if (check[a->u.attr->attr_nr].value != ATTR__TRUE)
 594                         continue;
 595                 rem = fill_one("expand", a, rem);
 596         }
 597         return rem;
 598 }
 599
 600 int git_checkattr(const char *path, int num, struct git_attr_check *check)
 601 {
 602         struct attr_stack *stk;
 603         const char *cp;
 604         int dirlen, pathlen, i, rem;
 605
 606         bootstrap_attr_stack();
 607         for (i = 0; i < attr_nr; i++)
 608                 check_all_attr[i].value = ATTR__UNKNOWN;
 609
 610         pathlen = strlen(path);
 611         cp = strrchr(path, '/');
 612         if (!cp)
 613                 dirlen = 0;
 614         else
 615                 dirlen = cp - path;
 616         prepare_attr_stack(path, dirlen);
 617         rem = attr_nr;
 618         for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 619                 rem = fill(path, pathlen, stk, rem);
 620
 621         for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 622                 rem = macroexpand(stk, rem);
 623
 624         for (i = 0; i < num; i++) {
 625                 const char *value = check_all_attr[check[i].attr->attr_nr].value;
 626                 if (value == ATTR__UNKNOWN)
 627                         value = ATTR__UNSET;
 628                 check[i].value = value;
 629         }
 630
 631         return 0;
 632 }