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, 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 static struct git_attr *git_attr_internal(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 struct git_attr *git_attr(const char *name)
  99 {
 100         return git_attr_internal(name, strlen(name));
 101 }
 102
 103 /*
 104  * .gitattributes file is one line per record, each of which is
 105  *
 106  * (1) glob pattern.
 107  * (2) whitespace
 108  * (3) whitespace separated list of attribute names, each of which
 109  *     could be prefixed with '-' to mean "set to false", '!' to mean
 110  *     "unset".
 111  */
 112
 113 /* What does a matched pattern decide? */
 114 struct attr_state {
 115         struct git_attr *attr;
 116         const char *setto;
 117 };
 118
 119 struct match_attr {
 120         union {
 121                 char *pattern;
 122                 struct git_attr *attr;
 123         } u;
 124         char is_macro;
 125         unsigned num_attr;
 126         struct attr_state state[FLEX_ARRAY];
 127 };
 128
 129 static const char blank[] = " \t\r\n";
 130
 131 static const char *parse_attr(const char *src, int lineno, const char *cp,
 132                               int *num_attr, struct match_attr *res)
 133 {
 134         const char *ep, *equals;
 135         int len;
 136
 137         ep = cp + strcspn(cp, blank);
 138         equals = strchr(cp, '=');
 139         if (equals && ep < equals)
 140                 equals = NULL;
 141         if (equals)
 142                 len = equals - cp;
 143         else
 144                 len = ep - cp;
 145         if (!res) {
 146                 if (*cp == '-' || *cp == '!') {
 147                         cp++;
 148                         len--;
 149                 }
 150                 if (invalid_attr_name(cp, len)) {
 151                         fprintf(stderr,
 152                                 "%.*s is not a valid attribute name: %s:%d\n",
 153                                 len, cp, src, lineno);
 154                         return NULL;
 155                 }
 156         } else {
 157                 struct attr_state *e;
 158
 159                 e = &(res->state[*num_attr]);
 160                 if (*cp == '-' || *cp == '!') {
 161                         e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
 162                         cp++;
 163                         len--;
 164                 }
 165                 else if (!equals)
 166                         e->setto = ATTR__TRUE;
 167                 else {
 168                         e->setto = xmemdupz(equals + 1, ep - equals - 1);
 169                 }
 170                 e->attr = git_attr_internal(cp, len);
 171         }
 172         (*num_attr)++;
 173         return ep + strspn(ep, blank);
 174 }
 175
 176 static struct match_attr *parse_attr_line(const char *line, const char *src,
 177                                           int lineno, int macro_ok)
 178 {
 179         int namelen;
 180         int num_attr;
 181         const char *cp, *name;
 182         struct match_attr *res = NULL;
 183         int pass;
 184         int is_macro;
 185
 186         cp = line + strspn(line, blank);
 187         if (!*cp || *cp == '#')
 188                 return NULL;
 189         name = cp;
 190         namelen = strcspn(name, blank);
 191         if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
 192             !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
 193                 if (!macro_ok) {
 194                         fprintf(stderr, "%s not allowed: %s:%d\n",
 195                                 name, src, lineno);
 196                         return NULL;
 197                 }
 198                 is_macro = 1;
 199                 name += strlen(ATTRIBUTE_MACRO_PREFIX);
 200                 name += strspn(name, blank);
 201                 namelen = strcspn(name, blank);
 202                 if (invalid_attr_name(name, namelen)) {
 203                         fprintf(stderr,
 204                                 "%.*s is not a valid attribute name: %s:%d\n",
 205                                 namelen, name, src, lineno);
 206                         return NULL;
 207                 }
 208         }
 209         else
 210                 is_macro = 0;
 211
 212         for (pass = 0; pass < 2; pass++) {
 213                 /* pass 0 counts and allocates, pass 1 fills */
 214                 num_attr = 0;
 215                 cp = name + namelen;
 216                 cp = cp + strspn(cp, blank);
 217                 while (*cp) {
 218                         cp = parse_attr(src, lineno, cp, &num_attr, res);
 219                         if (!cp)
 220                                 return NULL;
 221                 }
 222                 if (pass)
 223                         break;
 224                 res = xcalloc(1,
 225                               sizeof(*res) +
 226                               sizeof(struct attr_state) * num_attr +
 227                               (is_macro ? 0 : namelen + 1));
 228                 if (is_macro)
 229                         res->u.attr = git_attr_internal(name, namelen);
 230                 else {
 231                         res->u.pattern = (char *)&(res->state[num_attr]);
 232                         memcpy(res->u.pattern, name, namelen);
 233                         res->u.pattern[namelen] = 0;
 234                 }
 235                 res->is_macro = is_macro;
 236                 res->num_attr = num_attr;
 237         }
 238         return res;
 239 }
 240
 241 /*
 242  * Like info/exclude and .gitignore, the attribute information can
 243  * come from many places.
 244  *
 245  * (1) .gitattribute file of the same directory;
 246  * (2) .gitattribute file of the parent directory if (1) does not have
 247  *      any match; this goes recursively upwards, just like .gitignore.
 248  * (3) $GIT_DIR/info/attributes, which overrides both of the above.
 249  *
 250  * In the same file, later entries override the earlier match, so in the
 251  * global list, we would have entries from info/attributes the earliest
 252  * (reading the file from top to bottom), .gitattribute of the root
 253  * directory (again, reading the file from top to bottom) down to the
 254  * current directory, and then scan the list backwards to find the first match.
 255  * This is exactly the same as what excluded() does in dir.c to deal with
 256  * .gitignore
 257  */
 258
 259 static struct attr_stack {
 260         struct attr_stack *prev;
 261         char *origin;
 262         unsigned num_matches;
 263         unsigned alloc;
 264         struct match_attr **attrs;
 265 } *attr_stack;
 266
 267 static void free_attr_elem(struct attr_stack *e)
 268 {
 269         int i;
 270         free(e->origin);
 271         for (i = 0; i < e->num_matches; i++) {
 272                 struct match_attr *a = e->attrs[i];
 273                 int j;
 274                 for (j = 0; j < a->num_attr; j++) {
 275                         const char *setto = a->state[j].setto;
 276                         if (setto == ATTR__TRUE ||
 277                             setto == ATTR__FALSE ||
 278                             setto == ATTR__UNSET ||
 279                             setto == ATTR__UNKNOWN)
 280                                 ;
 281                         else
 282                                 free((char *) setto);
 283                 }
 284                 free(a);
 285         }
 286         free(e);
 287 }
 288
 289 static const char *builtin_attr[] = {
 290         "[attr]binary -diff -text",
 291         NULL,
 292 };
 293
 294 static void handle_attr_line(struct attr_stack *res,
 295                              const char *line,
 296                              const char *src,
 297                              int lineno,
 298                              int macro_ok)
 299 {
 300         struct match_attr *a;
 301
 302         a = parse_attr_line(line, src, lineno, macro_ok);
 303         if (!a)
 304                 return;
 305         if (res->alloc <= res->num_matches) {
 306                 res->alloc = alloc_nr(res->num_matches);
 307                 res->attrs = xrealloc(res->attrs,
 308                                       sizeof(struct match_attr *) *
 309                                       res->alloc);
 310         }
 311         res->attrs[res->num_matches++] = a;
 312 }
 313
 314 static struct attr_stack *read_attr_from_array(const char **list)
 315 {
 316         struct attr_stack *res;
 317         const char *line;
 318         int lineno = 0;
 319
 320         res = xcalloc(1, sizeof(*res));
 321         while ((line = *(list++)) != NULL)
 322                 handle_attr_line(res, line, "[builtin]", ++lineno, 1);
 323         return res;
 324 }
 325
 326 static enum git_attr_direction direction;
 327 static struct index_state *use_index;
 328
 329 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
 330 {
 331         FILE *fp = fopen(path, "r");
 332         struct attr_stack *res;
 333         char buf[2048];
 334         int lineno = 0;
 335
 336         if (!fp)
 337                 return NULL;
 338         res = xcalloc(1, sizeof(*res));
 339         while (fgets(buf, sizeof(buf), fp))
 340                 handle_attr_line(res, buf, path, ++lineno, macro_ok);
 341         fclose(fp);
 342         return res;
 343 }
 344
 345 static void *read_index_data(const char *path)
 346 {
 347         int pos, len;
 348         unsigned long sz;
 349         enum object_type type;
 350         void *data;
 351         struct index_state *istate = use_index ? use_index : &the_index;
 352
 353         len = strlen(path);
 354         pos = index_name_pos(istate, path, len);
 355         if (pos < 0) {
 356                 /*
 357                  * We might be in the middle of a merge, in which
 358                  * case we would read stage #2 (ours).
 359                  */
 360                 int i;
 361                 for (i = -pos - 1;
 362                      (pos < 0 && i < istate->cache_nr &&
 363                       !strcmp(istate->cache[i]->name, path));
 364                      i++)
 365                         if (ce_stage(istate->cache[i]) == 2)
 366                                 pos = i;
 367         }
 368         if (pos < 0)
 369                 return NULL;
 370         data = read_sha1_file(istate->cache[pos]->sha1, &type, &sz);
 371         if (!data || type != OBJ_BLOB) {
 372                 free(data);
 373                 return NULL;
 374         }
 375         return data;
 376 }
 377
 378 static struct attr_stack *read_attr_from_index(const char *path, int macro_ok)
 379 {
 380         struct attr_stack *res;
 381         char *buf, *sp;
 382         int lineno = 0;
 383
 384         buf = read_index_data(path);
 385         if (!buf)
 386                 return NULL;
 387
 388         res = xcalloc(1, sizeof(*res));
 389         for (sp = buf; *sp; ) {
 390                 char *ep;
 391                 int more;
 392                 for (ep = sp; *ep && *ep != '\n'; ep++)
 393                         ;
 394                 more = (*ep == '\n');
 395                 *ep = '\0';
 396                 handle_attr_line(res, sp, path, ++lineno, macro_ok);
 397                 sp = ep + more;
 398         }
 399         free(buf);
 400         return res;
 401 }
 402
 403 static struct attr_stack *read_attr(const char *path, int macro_ok)
 404 {
 405         struct attr_stack *res;
 406
 407         if (direction == GIT_ATTR_CHECKOUT) {
 408                 res = read_attr_from_index(path, macro_ok);
 409                 if (!res)
 410                         res = read_attr_from_file(path, macro_ok);
 411         }
 412         else if (direction == GIT_ATTR_CHECKIN) {
 413                 res = read_attr_from_file(path, macro_ok);
 414                 if (!res)
 415                         /*
 416                          * There is no checked out .gitattributes file there, but
 417                          * we might have it in the index.  We allow operation in a
 418                          * sparsely checked out work tree, so read from it.
 419                          */
 420                         res = read_attr_from_index(path, macro_ok);
 421         }
 422         else
 423                 res = read_attr_from_index(path, macro_ok);
 424         if (!res)
 425                 res = xcalloc(1, sizeof(*res));
 426         return res;
 427 }
 428
 429 #if DEBUG_ATTR
 430 static void debug_info(const char *what, struct attr_stack *elem)
 431 {
 432         fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
 433 }
 434 static void debug_set(const char *what, const char *match, struct git_attr *attr, const void *v)
 435 {
 436         const char *value = v;
 437
 438         if (ATTR_TRUE(value))
 439                 value = "set";
 440         else if (ATTR_FALSE(value))
 441                 value = "unset";
 442         else if (ATTR_UNSET(value))
 443                 value = "unspecified";
 444
 445         fprintf(stderr, "%s: %s => %s (%s)\n",
 446                 what, attr->name, (char *) value, match);
 447 }
 448 #define debug_push(a) debug_info("push", (a))
 449 #define debug_pop(a) debug_info("pop", (a))
 450 #else
 451 #define debug_push(a) do { ; } while (0)
 452 #define debug_pop(a) do { ; } while (0)
 453 #define debug_set(a,b,c,d) do { ; } while (0)
 454 #endif
 455
 456 static void drop_attr_stack(void)
 457 {
 458         while (attr_stack) {
 459                 struct attr_stack *elem = attr_stack;
 460                 attr_stack = elem->prev;
 461                 free_attr_elem(elem);
 462         }
 463 }
 464
 465 static void bootstrap_attr_stack(void)
 466 {
 467         if (!attr_stack) {
 468                 struct attr_stack *elem;
 469
 470                 elem = read_attr_from_array(builtin_attr);
 471                 elem->origin = NULL;
 472                 elem->prev = attr_stack;
 473                 attr_stack = elem;
 474
 475                 if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
 476                         elem = read_attr(GITATTRIBUTES_FILE, 1);
 477                         elem->origin = strdup("");
 478                         elem->prev = attr_stack;
 479                         attr_stack = elem;
 480                         debug_push(elem);
 481                 }
 482
 483                 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
 484                 if (!elem)
 485                         elem = xcalloc(1, sizeof(*elem));
 486                 elem->origin = NULL;
 487                 elem->prev = attr_stack;
 488                 attr_stack = elem;
 489         }
 490 }
 491
 492 static void prepare_attr_stack(const char *path, int dirlen)
 493 {
 494         struct attr_stack *elem, *info;
 495         int len;
 496         struct strbuf pathbuf;
 497
 498         strbuf_init(&pathbuf, dirlen+2+strlen(GITATTRIBUTES_FILE));
 499
 500         /*
 501          * At the bottom of the attribute stack is the built-in
 502          * set of attribute definitions.  Then, contents from
 503          * .gitattribute files from directories closer to the
 504          * root to the ones in deeper directories are pushed
 505          * to the stack.  Finally, at the very top of the stack
 506          * we always keep the contents of $GIT_DIR/info/attributes.
 507          *
 508          * When checking, we use entries from near the top of the
 509          * stack, preferring $GIT_DIR/info/attributes, then
 510          * .gitattributes in deeper directories to shallower ones,
 511          * and finally use the built-in set as the default.
 512          */
 513         if (!attr_stack)
 514                 bootstrap_attr_stack();
 515
 516         /*
 517          * Pop the "info" one that is always at the top of the stack.
 518          */
 519         info = attr_stack;
 520         attr_stack = info->prev;
 521
 522         /*
 523          * Pop the ones from directories that are not the prefix of
 524          * the path we are checking.
 525          */
 526         while (attr_stack && attr_stack->origin) {
 527                 int namelen = strlen(attr_stack->origin);
 528
 529                 elem = attr_stack;
 530                 if (namelen <= dirlen &&
 531                     !strncmp(elem->origin, path, namelen))
 532                         break;
 533
 534                 debug_pop(elem);
 535                 attr_stack = elem->prev;
 536                 free_attr_elem(elem);
 537         }
 538
 539         /*
 540          * Read from parent directories and push them down
 541          */
 542         if (!is_bare_repository() || direction == GIT_ATTR_INDEX) {
 543                 while (1) {
 544                         char *cp;
 545
 546                         len = strlen(attr_stack->origin);
 547                         if (dirlen <= len)
 548                                 break;
 549                         strbuf_reset(&pathbuf);
 550                         strbuf_add(&pathbuf, path, dirlen);
 551                         strbuf_addch(&pathbuf, '/');
 552                         cp = strchr(pathbuf.buf + len + 1, '/');
 553                         strcpy(cp + 1, GITATTRIBUTES_FILE);
 554                         elem = read_attr(pathbuf.buf, 0);
 555                         *cp = '\0';
 556                         elem->origin = strdup(pathbuf.buf);
 557                         elem->prev = attr_stack;
 558                         attr_stack = elem;
 559                         debug_push(elem);
 560                 }
 561         }
 562
 563         strbuf_release(&pathbuf);
 564
 565         /*
 566          * Finally push the "info" one at the top of the stack.
 567          */
 568         info->prev = attr_stack;
 569         attr_stack = info;
 570 }
 571
 572 static int path_matches(const char *pathname, int pathlen,
 573                         const char *pattern,
 574                         const char *base, int baselen)
 575 {
 576         if (!strchr(pattern, '/')) {
 577                 /* match basename */
 578                 const char *basename = strrchr(pathname, '/');
 579                 basename = basename ? basename + 1 : pathname;
 580                 return (fnmatch(pattern, basename, 0) == 0);
 581         }
 582         /*
 583          * match with FNM_PATHNAME; the pattern has base implicitly
 584          * in front of it.
 585          */
 586         if (*pattern == '/')
 587                 pattern++;
 588         if (pathlen < baselen ||
 589             (baselen && pathname[baselen] != '/') ||
 590             strncmp(pathname, base, baselen))
 591                 return 0;
 592         if (baselen != 0)
 593                 baselen++;
 594         return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
 595 }
 596
 597 static int macroexpand_one(int attr_nr, int rem);
 598
 599 static int fill_one(const char *what, struct match_attr *a, int rem)
 600 {
 601         struct git_attr_check *check = check_all_attr;
 602         int i;
 603
 604         for (i = a->num_attr - 1; 0 < rem && 0 <= i; i--) {
 605                 struct git_attr *attr = a->state[i].attr;
 606                 const char **n = &(check[attr->attr_nr].value);
 607                 const char *v = a->state[i].setto;
 608
 609                 if (*n == ATTR__UNKNOWN) {
 610                         debug_set(what,
 611                                   a->is_macro ? a->u.attr->name : a->u.pattern,
 612                                   attr, v);
 613                         *n = v;
 614                         rem--;
 615                         rem = macroexpand_one(attr->attr_nr, rem);
 616                 }
 617         }
 618         return rem;
 619 }
 620
 621 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
 622 {
 623         int i;
 624         const char *base = stk->origin ? stk->origin : "";
 625
 626         for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 627                 struct match_attr *a = stk->attrs[i];
 628                 if (a->is_macro)
 629                         continue;
 630                 if (path_matches(path, pathlen,
 631                                  a->u.pattern, base, strlen(base)))
 632                         rem = fill_one("fill", a, rem);
 633         }
 634         return rem;
 635 }
 636
 637 static int macroexpand_one(int attr_nr, int rem)
 638 {
 639         struct attr_stack *stk;
 640         struct match_attr *a = NULL;
 641         int i;
 642
 643         if (check_all_attr[attr_nr].value != ATTR__TRUE)
 644                 return rem;
 645
 646         for (stk = attr_stack; !a && stk; stk = stk->prev)
 647                 for (i = stk->num_matches - 1; !a && 0 <= i; i--) {
 648                         struct match_attr *ma = stk->attrs[i];
 649                         if (!ma->is_macro)
 650                                 continue;
 651                         if (ma->u.attr->attr_nr == attr_nr)
 652                                 a = ma;
 653                 }
 654
 655         if (a)
 656                 rem = fill_one("expand", a, rem);
 657
 658         return rem;
 659 }
 660
 661 int git_checkattr(const char *path, int num, struct git_attr_check *check)
 662 {
 663         struct attr_stack *stk;
 664         const char *cp;
 665         int dirlen, pathlen, i, rem;
 666
 667         bootstrap_attr_stack();
 668         for (i = 0; i < attr_nr; i++)
 669                 check_all_attr[i].value = ATTR__UNKNOWN;
 670
 671         pathlen = strlen(path);
 672         cp = strrchr(path, '/');
 673         if (!cp)
 674                 dirlen = 0;
 675         else
 676                 dirlen = cp - path;
 677         prepare_attr_stack(path, dirlen);
 678         rem = attr_nr;
 679         for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 680                 rem = fill(path, pathlen, stk, rem);
 681
 682         for (i = 0; i < num; i++) {
 683                 const char *value = check_all_attr[check[i].attr->attr_nr].value;
 684                 if (value == ATTR__UNKNOWN)
 685                         value = ATTR__UNSET;
 686                 check[i].value = value;
 687         }
 688
 689         return 0;
 690 }
 691
 692 void git_attr_set_direction(enum git_attr_direction new, struct index_state *istate)
 693 {
 694         enum git_attr_direction old = direction;
 695
 696         if (is_bare_repository() && new != GIT_ATTR_INDEX)
 697                 die("BUG: non-INDEX attr direction in a bare repo");
 698
 699         direction = new;
 700         if (new != old)
 701                 drop_attr_stack();
 702         use_index = istate;
 703 }