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                         char *value;
 164                         int vallen = ep - equals;
 165                         value = xmalloc(vallen);
 166                         memcpy(value, equals+1, vallen-1);
 167                         value[vallen-1] = 0;
 168                         e->setto = value;
 169                 }
 170                 e->attr = git_attr(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(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 -crlf",
 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 struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
 327 {
 328         FILE *fp = fopen(path, "r");
 329         struct attr_stack *res;
 330         char buf[2048];
 331         int lineno = 0;
 332
 333         if (!fp)
 334                 return NULL;
 335         res = xcalloc(1, sizeof(*res));
 336         while (fgets(buf, sizeof(buf), fp))
 337                 handle_attr_line(res, buf, path, ++lineno, macro_ok);
 338         fclose(fp);
 339         return res;
 340 }
 341
 342 static void *read_index_data(const char *path)
 343 {
 344         int pos, len;
 345         unsigned long sz;
 346         enum object_type type;
 347         void *data;
 348
 349         len = strlen(path);
 350         pos = cache_name_pos(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 < active_nr &&
 359                       !strcmp(active_cache[i]->name, path));
 360                      i++)
 361                         if (ce_stage(active_cache[i]) == 2)
 362                                 pos = i;
 363         }
 364         if (pos < 0)
 365                 return NULL;
 366         data = read_sha1_file(active_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(const char *path, int macro_ok)
 375 {
 376         struct attr_stack *res;
 377         char *buf, *sp;
 378         int lineno = 0;
 379
 380         res = read_attr_from_file(path, macro_ok);
 381         if (res)
 382                 return res;
 383
 384         res = xcalloc(1, sizeof(*res));
 385
 386         /*
 387          * There is no checked out .gitattributes file there, but
 388          * we might have it in the index.  We allow operation in a
 389          * sparsely checked out work tree, so read from it.
 390          */
 391         buf = read_index_data(path);
 392         if (!buf)
 393                 return res;
 394
 395         for (sp = buf; *sp; ) {
 396                 char *ep;
 397                 int more;
 398                 for (ep = sp; *ep && *ep != '\n'; ep++)
 399                         ;
 400                 more = (*ep == '\n');
 401                 *ep = '\0';
 402                 handle_attr_line(res, sp, path, ++lineno, macro_ok);
 403                 sp = ep + more;
 404         }
 405         free(buf);
 406         return res;
 407 }
 408
 409 #if DEBUG_ATTR
 410 static void debug_info(const char *what, struct attr_stack *elem)
 411 {
 412         fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
 413 }
 414 static void debug_set(const char *what, const char *match, struct git_attr *attr, void *v)
 415 {
 416         const char *value = v;
 417
 418         if (ATTR_TRUE(value))
 419                 value = "set";
 420         else if (ATTR_FALSE(value))
 421                 value = "unset";
 422         else if (ATTR_UNSET(value))
 423                 value = "unspecified";
 424
 425         fprintf(stderr, "%s: %s => %s (%s)\n",
 426                 what, attr->name, (char *) value, match);
 427 }
 428 #define debug_push(a) debug_info("push", (a))
 429 #define debug_pop(a) debug_info("pop", (a))
 430 #else
 431 #define debug_push(a) do { ; } while (0)
 432 #define debug_pop(a) do { ; } while (0)
 433 #define debug_set(a,b,c,d) do { ; } while (0)
 434 #endif
 435
 436 static void bootstrap_attr_stack(void)
 437 {
 438         if (!attr_stack) {
 439                 struct attr_stack *elem;
 440
 441                 elem = read_attr_from_array(builtin_attr);
 442                 elem->origin = NULL;
 443                 elem->prev = attr_stack;
 444                 attr_stack = elem;
 445
 446                 elem = read_attr(GITATTRIBUTES_FILE, 1);
 447                 elem->origin = strdup("");
 448                 elem->prev = attr_stack;
 449                 attr_stack = elem;
 450                 debug_push(elem);
 451
 452                 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
 453                 if (!elem)
 454                         elem = xcalloc(1, sizeof(*elem));
 455                 elem->origin = NULL;
 456                 elem->prev = attr_stack;
 457                 attr_stack = elem;
 458         }
 459 }
 460
 461 static void prepare_attr_stack(const char *path, int dirlen)
 462 {
 463         struct attr_stack *elem, *info;
 464         int len;
 465         char pathbuf[PATH_MAX];
 466
 467         /*
 468          * At the bottom of the attribute stack is the built-in
 469          * set of attribute definitions.  Then, contents from
 470          * .gitattribute files from directories closer to the
 471          * root to the ones in deeper directories are pushed
 472          * to the stack.  Finally, at the very top of the stack
 473          * we always keep the contents of $GIT_DIR/info/attributes.
 474          *
 475          * When checking, we use entries from near the top of the
 476          * stack, preferring $GIT_DIR/info/attributes, then
 477          * .gitattributes in deeper directories to shallower ones,
 478          * and finally use the built-in set as the default.
 479          */
 480         if (!attr_stack)
 481                 bootstrap_attr_stack();
 482
 483         /*
 484          * Pop the "info" one that is always at the top of the stack.
 485          */
 486         info = attr_stack;
 487         attr_stack = info->prev;
 488
 489         /*
 490          * Pop the ones from directories that are not the prefix of
 491          * the path we are checking.
 492          */
 493         while (attr_stack && attr_stack->origin) {
 494                 int namelen = strlen(attr_stack->origin);
 495
 496                 elem = attr_stack;
 497                 if (namelen <= dirlen &&
 498                     !strncmp(elem->origin, path, namelen))
 499                         break;
 500
 501                 debug_pop(elem);
 502                 attr_stack = elem->prev;
 503                 free_attr_elem(elem);
 504         }
 505
 506         /*
 507          * Read from parent directories and push them down
 508          */
 509         while (1) {
 510                 char *cp;
 511
 512                 len = strlen(attr_stack->origin);
 513                 if (dirlen <= len)
 514                         break;
 515                 memcpy(pathbuf, path, dirlen);
 516                 memcpy(pathbuf + dirlen, "/", 2);
 517                 cp = strchr(pathbuf + len + 1, '/');
 518                 strcpy(cp + 1, GITATTRIBUTES_FILE);
 519                 elem = read_attr(pathbuf, 0);
 520                 *cp = '\0';
 521                 elem->origin = strdup(pathbuf);
 522                 elem->prev = attr_stack;
 523                 attr_stack = elem;
 524                 debug_push(elem);
 525         }
 526
 527         /*
 528          * Finally push the "info" one at the top of the stack.
 529          */
 530         info->prev = attr_stack;
 531         attr_stack = info;
 532 }
 533
 534 static int path_matches(const char *pathname, int pathlen,
 535                         const char *pattern,
 536                         const char *base, int baselen)
 537 {
 538         if (!strchr(pattern, '/')) {
 539                 /* match basename */
 540                 const char *basename = strrchr(pathname, '/');
 541                 basename = basename ? basename + 1 : pathname;
 542                 return (fnmatch(pattern, basename, 0) == 0);
 543         }
 544         /*
 545          * match with FNM_PATHNAME; the pattern has base implicitly
 546          * in front of it.
 547          */
 548         if (*pattern == '/')
 549                 pattern++;
 550         if (pathlen < baselen ||
 551             (baselen && pathname[baselen - 1] != '/') ||
 552             strncmp(pathname, base, baselen))
 553                 return 0;
 554         return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
 555 }
 556
 557 static int fill_one(const char *what, struct match_attr *a, int rem)
 558 {
 559         struct git_attr_check *check = check_all_attr;
 560         int i;
 561
 562         for (i = 0; 0 < rem && i < a->num_attr; i++) {
 563                 struct git_attr *attr = a->state[i].attr;
 564                 const char **n = &(check[attr->attr_nr].value);
 565                 const char *v = a->state[i].setto;
 566
 567                 if (*n == ATTR__UNKNOWN) {
 568                         debug_set(what, a->u.pattern, attr, v);
 569                         *n = v;
 570                         rem--;
 571                 }
 572         }
 573         return rem;
 574 }
 575
 576 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
 577 {
 578         int i;
 579         const char *base = stk->origin ? stk->origin : "";
 580
 581         for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 582                 struct match_attr *a = stk->attrs[i];
 583                 if (a->is_macro)
 584                         continue;
 585                 if (path_matches(path, pathlen,
 586                                  a->u.pattern, base, strlen(base)))
 587                         rem = fill_one("fill", a, rem);
 588         }
 589         return rem;
 590 }
 591
 592 static int macroexpand(struct attr_stack *stk, int rem)
 593 {
 594         int i;
 595         struct git_attr_check *check = check_all_attr;
 596
 597         for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 598                 struct match_attr *a = stk->attrs[i];
 599                 if (!a->is_macro)
 600                         continue;
 601                 if (check[a->u.attr->attr_nr].value != ATTR__TRUE)
 602                         continue;
 603                 rem = fill_one("expand", a, rem);
 604         }
 605         return rem;
 606 }
 607
 608 int git_checkattr(const char *path, int num, struct git_attr_check *check)
 609 {
 610         struct attr_stack *stk;
 611         const char *cp;
 612         int dirlen, pathlen, i, rem;
 613
 614         bootstrap_attr_stack();
 615         for (i = 0; i < attr_nr; i++)
 616                 check_all_attr[i].value = ATTR__UNKNOWN;
 617
 618         pathlen = strlen(path);
 619         cp = strrchr(path, '/');
 620         if (!cp)
 621                 dirlen = 0;
 622         else
 623                 dirlen = cp - path;
 624         prepare_attr_stack(path, dirlen);
 625         rem = attr_nr;
 626         for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 627                 rem = fill(path, pathlen, stk, rem);
 628
 629         for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 630                 rem = macroexpand(stk, rem);
 631
 632         for (i = 0; i < num; i++) {
 633                 const char *value = check_all_attr[check[i].attr->attr_nr].value;
 634                 if (value == ATTR__UNKNOWN)
 635                         value = ATTR__UNSET;
 636                 check[i].value = value;
 637         }
 638
 639         return 0;
 640 }