attr.c

   1 #define NO_THE_INDEX_COMPATIBILITY_MACROS
   2 #include "cache.h"
   3 #include "attr.h"
   4
   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
  12
  13 /*
  14  * The basic design decision here is that we are not going to have
  15  * insanely large number of attributes.
  16  *
  17  * This is a randomly chosen prime.
  18  */
  19 #define HASHSIZE 257
  20
  21 #ifndef DEBUG_ATTR
  22 #define DEBUG_ATTR 0
  23 #endif
  24
  25 struct git_attr {
  26         struct git_attr *next;
  27         unsigned h;
  28         int attr_nr;
  29         char name[FLEX_ARRAY];
  30 };
  31 static int attr_nr;
  32
  33 static struct git_attr_check *check_all_attr;
  34 static struct git_attr *(git_attr_hash[HASHSIZE]);
  35
  36 static unsigned hash_name(const char *name, int namelen)
  37 {
  38         unsigned val = 0;
  39         unsigned char c;
  40
  41         while (namelen--) {
  42                 c = *name++;
  43                 val = ((val << 7) | (val >> 22)) ^ c;
  44         }
  45         return val;
  46 }
  47
  48 static int invalid_attr_name(const char *name, int namelen)
  49 {
  50         /*
  51          * Attribute name cannot begin with '-' and from
  52          * [-A-Za-z0-9_.].  We'd specifically exclude '=' for now,
  53          * as we might later want to allow non-binary value for
  54          * attributes, e.g. "*.svg      merge=special-merge-program-for-svg"
  55          */
  56         if (*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;
  65         }
  66         return 0;
  67 }
  68
  69 struct git_attr *git_attr(const char *name, int len)
  70 {
  71         unsigned hval = hash_name(name, len);
  72         unsigned pos = hval % HASHSIZE;
  73         struct git_attr *a;
  74
  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;
  79         }
  80
  81         if (invalid_attr_name(name, len))
  82                 return NULL;
  83
  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;
  91
  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;
  97 }
  98
  99 /*
 100  * .gitattributes file is one line per record, each of which is
 101  *
 102  * (1) glob pattern.
 103  * (2) whitespace
 104  * (3) whitespace separated list of attribute names, each of which
 105  *     could be prefixed with '-' to mean "set to false", '!' to mean
 106  *     "unset".
 107  */
 108
 109 /* What does a matched pattern decide? */
 110 struct attr_state {
 111         struct git_attr *attr;
 112         const char *setto;
 113 };
 114
 115 struct match_attr {
 116         union {
 117                 char *pattern;
 118                 struct git_attr *attr;
 119         } u;
 120         char is_macro;
 121         unsigned num_attr;
 122         struct attr_state state[FLEX_ARRAY];
 123 };
 124
 125 static const char blank[] = " \t\r\n";
 126
 127 static const char *parse_attr(const char *src, int lineno, const char *cp,
 128                               int *num_attr, struct match_attr *res)
 129 {
 130         const char *ep, *equals;
 131         int len;
 132
 133         ep = cp + strcspn(cp, blank);
 134         equals = strchr(cp, '=');
 135         if (equals && ep < equals)
 136                 equals = NULL;
 137         if (equals)
 138                 len = equals - cp;
 139         else
 140                 len = ep - cp;
 141         if (!res) {
 142                 if (*cp == '-' || *cp == '!') {
 143                         cp++;
 144                         len--;
 145                 }
 146                 if (invalid_attr_name(cp, len)) {
 147                         fprintf(stderr,
 148                                 "%.*s is not a valid attribute name: %s:%d\n",
 149                                 len, cp, src, lineno);
 150                         return NULL;
 151                 }
 152         } else {
 153                 struct attr_state *e;
 154
 155                 e = &(res->state[*num_attr]);
 156                 if (*cp == '-' || *cp == '!') {
 157                         e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
 158                         cp++;
 159                         len--;
 160                 }
 161                 else if (!equals)
 162                         e->setto = ATTR__TRUE;
 163                 else {
 164                         e->setto = xmemdupz(equals + 1, ep - equals - 1);
 165                 }
 166                 e->attr = git_attr(cp, len);
 167         }
 168         (*num_attr)++;
 169         return ep + strspn(ep, blank);
 170 }
 171
 172 static struct match_attr *parse_attr_line(const char *line, const char *src,
 173                                           int lineno, int macro_ok)
 174 {
 175         int namelen;
 176         int num_attr;
 177         const char *cp, *name;
 178         struct match_attr *res = NULL;
 179         int pass;
 180         int is_macro;
 181
 182         cp = line + strspn(line, blank);
 183         if (!*cp || *cp == '#')
 184                 return NULL;
 185         name = cp;
 186         namelen = strcspn(name, blank);
 187         if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
 188             !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
 189                 if (!macro_ok) {
 190                         fprintf(stderr, "%s not allowed: %s:%d\n",
 191                                 name, src, lineno);
 192                         return NULL;
 193                 }
 194                 is_macro = 1;
 195                 name += strlen(ATTRIBUTE_MACRO_PREFIX);
 196                 name += strspn(name, blank);
 197                 namelen = strcspn(name, blank);
 198                 if (invalid_attr_name(name, namelen)) {
 199                         fprintf(stderr,
 200                                 "%.*s is not a valid attribute name: %s:%d\n",
 201                                 namelen, name, src, lineno);
 202                         return NULL;
 203                 }
 204         }
 205         else
 206                 is_macro = 0;
 207
 208         for (pass = 0; pass < 2; pass++) {
 209                 /* pass 0 counts and allocates, pass 1 fills */
 210                 num_attr = 0;
 211                 cp = name + namelen;
 212                 cp = cp + strspn(cp, blank);
 213                 while (*cp) {
 214                         cp = parse_attr(src, lineno, cp, &num_attr, res);
 215                         if (!cp)
 216                                 return NULL;
 217                 }
 218                 if (pass)
 219                         break;
 220                 res = xcalloc(1,
 221                               sizeof(*res) +
 222                               sizeof(struct attr_state) * num_attr +
 223                               (is_macro ? 0 : namelen + 1));
 224                 if (is_macro)
 225                         res->u.attr = git_attr(name, namelen);
 226                 else {
 227                         res->u.pattern = (char *)&(res->state[num_attr]);
 228                         memcpy(res->u.pattern, name, namelen);
 229                         res->u.pattern[namelen] = 0;
 230                 }
 231                 res->is_macro = is_macro;
 232                 res->num_attr = num_attr;
 233         }
 234         return res;
 235 }
 236
 237 /*
 238  * Like info/exclude and .gitignore, the attribute information can
 239  * come from many places.
 240  *
 241  * (1) .gitattribute file of the same directory;
 242  * (2) .gitattribute file of the parent directory if (1) does not have
 243  *      any match; this goes recursively upwards, just like .gitignore.
 244  * (3) $GIT_DIR/info/attributes, which overrides both of the above.
 245  *
 246  * In the same file, later entries override the earlier match, so in the
 247  * global list, we would have entries from info/attributes the earliest
 248  * (reading the file from top to bottom), .gitattribute of the root
 249  * directory (again, reading the file from top to bottom) down to the
 250  * current directory, and then scan the list backwards to find the first match.
 251  * This is exactly the same as what excluded() does in dir.c to deal with
 252  * .gitignore
 253  */
 254
 255 static struct attr_stack {
 256         struct attr_stack *prev;
 257         char *origin;
 258         unsigned num_matches;
 259         unsigned alloc;
 260         struct match_attr **attrs;
 261 } *attr_stack;
 262
 263 static void free_attr_elem(struct attr_stack *e)
 264 {
 265         int i;
 266         free(e->origin);
 267         for (i = 0; i < e->num_matches; i++) {
 268                 struct match_attr *a = e->attrs[i];
 269                 int j;
 270                 for (j = 0; j < a->num_attr; j++) {
 271                         const char *setto = a->state[j].setto;
 272                         if (setto == ATTR__TRUE ||
 273                             setto == ATTR__FALSE ||
 274                             setto == ATTR__UNSET ||
 275                             setto == ATTR__UNKNOWN)
 276                                 ;
 277                         else
 278                                 free((char *) setto);
 279                 }
 280                 free(a);
 281         }
 282         free(e);
 283 }
 284
 285 static const char *builtin_attr[] = {
 286         "[attr]binary -diff -crlf",
 287         NULL,
 288 };
 289
 290 static void handle_attr_line(struct attr_stack *res,
 291                              const char *line,
 292                              const char *src,
 293                              int lineno,
 294                              int macro_ok)
 295 {
 296         struct match_attr *a;
 297
 298         a = parse_attr_line(line, src, lineno, macro_ok);
 299         if (!a)
 300                 return;
 301         if (res->alloc <= res->num_matches) {
 302                 res->alloc = alloc_nr(res->num_matches);
 303                 res->attrs = xrealloc(res->attrs,
 304                                       sizeof(struct match_attr *) *
 305                                       res->alloc);
 306         }
 307         res->attrs[res->num_matches++] = a;
 308 }
 309
 310 static struct attr_stack *read_attr_from_array(const char **list)
 311 {
 312         struct attr_stack *res;
 313         const char *line;
 314         int lineno = 0;
 315
 316         res = xcalloc(1, sizeof(*res));
 317         while ((line = *(list++)) != NULL)
 318                 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
 319         return res;
 320 }
 321
 322 static enum git_attr_direction direction;
 323 static struct index_state *use_index;
 324
 325 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
 326 {
 327         FILE *fp = fopen(path, "r");
 328         struct attr_stack *res;
 329         char buf[2048];
 330         int lineno = 0;
 331
 332         if (!fp)
 333                 return NULL;
 334         res = xcalloc(1, sizeof(*res));
 335         while (fgets(buf, sizeof(buf), fp))
 336                 handle_attr_line(res, buf, path, ++lineno, macro_ok);
 337         fclose(fp);
 338         return res;
 339 }
 340
 341 static void *read_index_data(const char *path)
 342 {
 343         int pos, len;
 344         unsigned long sz;
 345         enum object_type type;
 346         void *data;
 347         struct index_state *istate = use_index ? use_index : &the_index;
 348
 349         len = strlen(path);
 350         pos = index_name_pos(istate, path, len);
 351         if (pos < 0) {
 352                 /*
 353                  * We might be in the middle of a merge, in which
 354                  * case we would read stage #2 (ours).
 355                  */
 356                 int i;
 357                 for (i = -pos - 1;
 358                      (pos < 0 && i < istate->cache_nr &&
 359                       !strcmp(istate->cache[i]->name, path));
 360                      i++)
 361                         if (ce_stage(istate->cache[i]) == 2)
 362                                 pos = i;
 363         }
 364         if (pos < 0)
 365                 return NULL;
 366         data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
 367         if (!data || type != OBJ_BLOB) {
 368                 free(data);
 369                 return NULL;
 370         }
 371         return data;
 372 }
 373
 374 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
 375 {
 376         struct attr_stack *res;
 377         char *buf, *sp;
 378         int lineno = 0;
 379
 380         buf = read_index_data(path);
 381         if (!buf)
 382                 return NULL;
 383
 384         res = xcalloc(1, sizeof(*res));
 385         for (sp = buf; *sp; ) {
 386                 char *ep;
 387                 int more;
 388                 for (ep = sp; *ep && *ep != '\n'; ep++)
 389                         ;
 390                 more = (*ep == '\n');
 391                 *ep = '\0';
 392                 handle_attr_line(res, sp, path, ++lineno, macro_ok);
 393                 sp = ep + more;
 394         }
 395         free(buf);
 396         return res;
 397 }
 398
 399 static struct attr_stack *read_attr(const char *path, int macro_ok)
 400 {
 401         struct attr_stack *res;
 402
 403         if (direction == GIT_ATTR_CHECKOUT) {
 404                 res = read_attr_from_index(path, macro_ok);
 405                 if (!res)
 406                         res = read_attr_from_file(path, macro_ok);
 407         }
 408         else if (direction == GIT_ATTR_CHECKIN) {
 409                 res = read_attr_from_file(path, macro_ok);
 410                 if (!res)
 411                         /*
 412                          * There is no checked out .gitattributes file there, but
 413                          * we might have it in the index.  We allow operation in a
 414                          * sparsely checked out work tree, so read from it.
 415                          */
 416                         res = read_attr_from_index(path, macro_ok);
 417         }
 418         else
 419                 res = read_attr_from_index(path, macro_ok);
 420         if (!res)
 421                 res = xcalloc(1, sizeof(*res));
 422         return res;
 423 }
 424
 425 #if DEBUG_ATTR
 426 static void debug_info(const char *what, struct attr_stack *elem)
 427 {
 428         fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
 429 }
 430 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
 431 {
 432         const char *value = v;
 433
 434         if (ATTR_TRUE(value))
 435                 value = "set";
 436         else if (ATTR_FALSE(value))
 437                 value = "unset";
 438         else if (ATTR_UNSET(value))
 439                 value = "unspecified";
 440
 441         fprintf(stderr, "%s: %s => %s (%s)\n",
 442                 what, attr->name, (char *) value, match);
 443 }
 444 #define debug_push(a) debug_info("push", (a))
 445 #define debug_pop(a) debug_info("pop", (a))
 446 #else
 447 #define debug_push(a) do { ; } while (0)
 448 #define debug_pop(a) do { ; } while (0)
 449 #define debug_set(a,b,c,d) do { ; } while (0)
 450 #endif
 451
 452 static void drop_attr_stack(void)
 453 {
 454         while (attr_stack) {
 455                 struct attr_stack *elem = attr_stack;
 456                 attr_stack = elem->prev;
 457                 free_attr_elem(elem);
 458         }
 459 }
 460
 461 static void bootstrap_attr_stack(void)
 462 {
 463         if (!attr_stack) {
 464                 struct attr_stack *elem;
 465
 466                 elem = read_attr_from_array(builtin_attr);
 467                 elem->origin = NULL;
 468                 elem->prev = attr_stack;
 469                 attr_stack = elem;
 470
 471                 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
 472                         elem = read_attr(GITATTRIBUTES_FILE, 1);
 473                         elem->origin = strdup("");
 474                         elem->prev = attr_stack;
 475                         attr_stack = elem;
 476                         debug_push(elem);
 477                 }
 478
 479                 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
 480                 if (!elem)
 481                         elem = xcalloc(1, sizeof(*elem));
 482                 elem->origin = NULL;
 483                 elem->prev = attr_stack;
 484                 attr_stack = elem;
 485         }
 486 }
 487
 488 static void prepare_attr_stack(const char *path, int dirlen)
 489 {
 490         struct attr_stack *elem, *info;
 491         int len;
 492         struct strbuf pathbuf;
 493
 494         strbuf_init(&pathbuf, dirlen+2+strlen(GITATTRIBUTES_FILE));
 495
 496         /*
 497          * At the bottom of the attribute stack is the built-in
 498          * set of attribute definitions.  Then, contents from
 499          * .gitattribute files from directories closer to the
 500          * root to the ones in deeper directories are pushed
 501          * to the stack.  Finally, at the very top of the stack
 502          * we always keep the contents of $GIT_DIR/info/attributes.
 503          *
 504          * When checking, we use entries from near the top of the
 505          * stack, preferring $GIT_DIR/info/attributes, then
 506          * .gitattributes in deeper directories to shallower ones,
 507          * and finally use the built-in set as the default.
 508          */
 509         if (!attr_stack)
 510                 bootstrap_attr_stack();
 511
 512         /*
 513          * Pop the "info" one that is always at the top of the stack.
 514          */
 515         info = attr_stack;
 516         attr_stack = info->prev;
 517
 518         /*
 519          * Pop the ones from directories that are not the prefix of
 520          * the path we are checking.
 521          */
 522         while (attr_stack && attr_stack->origin) {
 523                 int namelen = strlen(attr_stack->origin);
 524
 525                 elem = attr_stack;
 526                 if (namelen <= dirlen &&
 527                     !strncmp(elem->origin, path, namelen))
 528                         break;
 529
 530                 debug_pop(elem);
 531                 attr_stack = elem->prev;
 532                 free_attr_elem(elem);
 533         }
 534
 535         /*
 536          * Read from parent directories and push them down
 537          */
 538         if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
 539                 while (1) {
 540                         char *cp;
 541
 542                         len = strlen(attr_stack->origin);
 543                         if (dirlen <= len)
 544                                 break;
 545                         strbuf_reset(&pathbuf);
 546                         strbuf_add(&pathbuf, path, dirlen);
 547                         strbuf_addch(&pathbuf, '/');
 548                         cp = strchr(pathbuf.buf + len + 1, '/');
 549                         strcpy(cp + 1, GITATTRIBUTES_FILE);
 550                         elem = read_attr(pathbuf.buf, 0);
 551                         *cp = '\0';
 552                         elem->origin = strdup(pathbuf.buf);
 553                         elem->prev = attr_stack;
 554                         attr_stack = elem;
 555                         debug_push(elem);
 556                 }
 557         }
 558
 559         /*
 560          * Finally push the "info" one at the top of the stack.
 561          */
 562         info->prev = attr_stack;
 563         attr_stack = info;
 564 }
 565
 566 static int path_matches(const char *pathname, int pathlen,
 567                         const char *pattern,
 568                         const char *base, int baselen)
 569 {
 570         if (!strchr(pattern, '/')) {
 571                 /* match basename */
 572                 const char *basename = strrchr(pathname, '/');
 573                 basename = basename ? basename + 1 : pathname;
 574                 return (fnmatch(pattern, basename, 0) == 0);
 575         }
 576         /*
 577          * match with FNM_PATHNAME; the pattern has base implicitly
 578          * in front of it.
 579          */
 580         if (*pattern == '/')
 581                 pattern++;
 582         if (pathlen < baselen ||
 583             (baselen && pathname[baselen] != '/') ||
 584             strncmp(pathname, base, baselen))
 585                 return 0;
 586         if (baselen != 0)
 587                 baselen++;
 588         return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
 589 }
 590
 591 static int fill_one(const char *what, struct match_attr *a, int rem)
 592 {
 593         struct git_attr_check *check = check_all_attr;
 594         int i;
 595
 596         for (i = 0; 0 < rem && i < a->num_attr; i++) {
 597                 struct git_attr *attr = a->state[i].attr;
 598                 const char **n = &(check[attr->attr_nr].value);
 599                 const char *v = a->state[i].setto;
 600
 601                 if (*n == ATTR__UNKNOWN) {
 602                         debug_set(what, a->u.pattern, attr, v);
 603                         *n = v;
 604                         rem--;
 605                 }
 606         }
 607         return rem;
 608 }
 609
 610 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
 611 {
 612         int i;
 613         const char *base = stk->origin ? stk->origin : "";
 614
 615         for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 616                 struct match_attr *a = stk->attrs[i];
 617                 if (a->is_macro)
 618                         continue;
 619                 if (path_matches(path, pathlen,
 620                                  a->u.pattern, base, strlen(base)))
 621                         rem = fill_one("fill", a, rem);
 622         }
 623         return rem;
 624 }
 625
 626 static int macroexpand(struct attr_stack *stk, int rem)
 627 {
 628         int i;
 629         struct git_attr_check *check = check_all_attr;
 630
 631         for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 632                 struct match_attr *a = stk->attrs[i];
 633                 if (!a->is_macro)
 634                         continue;
 635                 if (check[a->u.attr->attr_nr].value != ATTR__TRUE)
 636                         continue;
 637                 rem = fill_one("expand", a, rem);
 638         }
 639         return rem;
 640 }
 641
 642 int git_checkattr(const char *path, int num, struct git_attr_check *check)
 643 {
 644         struct attr_stack *stk;
 645         const char *cp;
 646         int dirlen, pathlen, i, rem;
 647
 648         bootstrap_attr_stack();
 649         for (i = 0; i < attr_nr; i++)
 650                 check_all_attr[i].value = ATTR__UNKNOWN;
 651
 652         pathlen = strlen(path);
 653         cp = strrchr(path, '/');
 654         if (!cp)
 655                 dirlen = 0;
 656         else
 657                 dirlen = cp - path;
 658         prepare_attr_stack(path, dirlen);
 659         rem = attr_nr;
 660         for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 661                 rem = fill(path, pathlen, stk, rem);
 662
 663         for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 664                 rem = macroexpand(stk, rem);
 665
 666         for (i = 0; i < num; i++) {
 667                 const char *value = check_all_attr[check[i].attr->attr_nr].value;
 668                 if (value == ATTR__UNKNOWN)
 669                         value = ATTR__UNSET;
 670                 check[i].value = value;
 671         }
 672
 673         return 0;
 674 }
 675
 676 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
 677 {
 678         enum git_attr_direction old = direction;
 679
 680         if (is_bare_repository() && new != GIT_ATTR_INDEX)
 681                 die("BUG: non-INDEX attr direction in a bare repo");
 682
 683         direction = new;
 684         if (new != old)
 685                 drop_attr_stack();
 686         use_index = istate;
 687 }