git/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 (pass)
 220                         break;
 221                 res = xcalloc(1,
 222                               sizeof(*res) +
 223                               sizeof(struct attr_state) * num_attr +
 224                               (is_macro ? 0 : namelen + 1));
 225                 if (is_macro)
 226                         res->u.attr = git_attr(name, namelen);
 227                 else {
 228                         res->u.pattern = (char*)&(res->state[num_attr]);
 229                         memcpy(res->u.pattern, name, namelen);
 230                         res->u.pattern[namelen] = 0;
 231                 }
 232                 res->is_macro = is_macro;
 233                 res->num_attr = num_attr;
 234         }
 235         return res;
 236 }
 237
 238 /*
 239  * Like info/exclude and .gitignore, the attribute information can
 240  * come from many places.
 241  *
 242  * (1) .gitattribute file of the same directory;
 243  * (2) .gitattribute file of the parent directory if (1) does not have
 244  *      any match; this goes recursively upwards, just like .gitignore.
 245  * (3) $GIT_DIR/info/attributes, which overrides both of the above.
 246  *
 247  * In the same file, later entries override the earlier match, so in the
 248  * global list, we would have entries from info/attributes the earliest
 249  * (reading the file from top to bottom), .gitattribute of the root
 250  * directory (again, reading the file from top to bottom) down to the
 251  * current directory, and then scan the list backwards to find the first match.
 252  * This is exactly the same as what excluded() does in dir.c to deal with
 253  * .gitignore
 254  */
 255
 256 static struct attr_stack {
 257         struct attr_stack *prev;
 258         char *origin;
 259         unsigned num_matches;
 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 struct attr_stack *read_attr_from_array(const char **list)
 291 {
 292         struct attr_stack *res;
 293         const char *line;
 294         int lineno = 0;
 295
 296         res = xcalloc(1, sizeof(*res));
 297         while ((line = *(list++)) != NULL) {
 298                 struct match_attr *a;
 299
 300                 a = parse_attr_line(line, "[builtin]", ++lineno, 1);
 301                 if (!a)
 302                         continue;
 303                 res->attrs = xrealloc(res->attrs,
 304                         sizeof(struct match_attr *) * (res->num_matches + 1));
 305                 res->attrs[res->num_matches++] = a;
 306         }
 307         return res;
 308 }
 309
 310 static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
 311 {
 312         FILE *fp;
 313         struct attr_stack *res;
 314         char buf[2048];
 315         int lineno = 0;
 316
 317         res = xcalloc(1, sizeof(*res));
 318         fp = fopen(path, "r");
 319         if (!fp)
 320                 return res;
 321
 322         while (fgets(buf, sizeof(buf), fp)) {
 323                 struct match_attr *a;
 324
 325                 a = parse_attr_line(buf, path, ++lineno, macro_ok);
 326                 if (!a)
 327                         continue;
 328                 res->attrs = xrealloc(res->attrs,
 329                         sizeof(struct match_attr *) * (res->num_matches + 1));
 330                 res->attrs[res->num_matches++] = a;
 331         }
 332         fclose(fp);
 333         return res;
 334 }
 335
 336 #if DEBUG_ATTR
 337 static void debug_info(const char *what, struct attr_stack *elem)
 338 {
 339         fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
 340 }
 341 static void debug_set(const char *what, const char *match, struct git_attr *attr, void *v)
 342 {
 343         const char *value = v;
 344
 345         if (ATTR_TRUE(value))
 346                 value = "set";
 347         else if (ATTR_FALSE(value))
 348                 value = "unset";
 349         else if (ATTR_UNSET(value))
 350                 value = "unspecified";
 351
 352         fprintf(stderr, "%s: %s => %s (%s)\n",
 353                 what, attr->name, (char *) value, match);
 354 }
 355 #define debug_push(a) debug_info("push", (a))
 356 #define debug_pop(a) debug_info("pop", (a))
 357 #else
 358 #define debug_push(a) do { ; } while (0)
 359 #define debug_pop(a) do { ; } while (0)
 360 #define debug_set(a,b,c,d) do { ; } while (0)
 361 #endif
 362
 363 static void bootstrap_attr_stack(void)
 364 {
 365         if (!attr_stack) {
 366                 struct attr_stack *elem;
 367
 368                 elem = read_attr_from_array(builtin_attr);
 369                 elem->origin = NULL;
 370                 elem->prev = attr_stack;
 371                 attr_stack = elem;
 372
 373                 elem = read_attr_from_file(GITATTRIBUTES_FILE, 1);
 374                 elem->origin = strdup("");
 375                 elem->prev = attr_stack;
 376                 attr_stack = elem;
 377                 debug_push(elem);
 378
 379                 elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
 380                 elem->origin = NULL;
 381                 elem->prev = attr_stack;
 382                 attr_stack = elem;
 383         }
 384 }
 385
 386 static void prepare_attr_stack(const char *path, int dirlen)
 387 {
 388         struct attr_stack *elem, *info;
 389         int len;
 390         char pathbuf[PATH_MAX];
 391
 392         /*
 393          * At the bottom of the attribute stack is the built-in
 394          * set of attribute definitions.  Then, contents from
 395          * .gitattribute files from directories closer to the
 396          * root to the ones in deeper directories are pushed
 397          * to the stack.  Finally, at the very top of the stack
 398          * we always keep the contents of $GIT_DIR/info/attributes.
 399          *
 400          * When checking, we use entries from near the top of the
 401          * stack, preferring $GIT_DIR/info/attributes, then
 402          * .gitattributes in deeper directories to shallower ones,
 403          * and finally use the built-in set as the default.
 404          */
 405         if (!attr_stack)
 406                 bootstrap_attr_stack();
 407
 408         /*
 409          * Pop the "info" one that is always at the top of the stack.
 410          */
 411         info = attr_stack;
 412         attr_stack = info->prev;
 413
 414         /*
 415          * Pop the ones from directories that are not the prefix of
 416          * the path we are checking.
 417          */
 418         while (attr_stack && attr_stack->origin) {
 419                 int namelen = strlen(attr_stack->origin);
 420
 421                 elem = attr_stack;
 422                 if (namelen <= dirlen &&
 423                     !strncmp(elem->origin, path, namelen))
 424                         break;
 425
 426                 debug_pop(elem);
 427                 attr_stack = elem->prev;
 428                 free_attr_elem(elem);
 429         }
 430
 431         /*
 432          * Read from parent directories and push them down
 433          */
 434         while (1) {
 435                 char *cp;
 436
 437                 len = strlen(attr_stack->origin);
 438                 if (dirlen <= len)
 439                         break;
 440                 memcpy(pathbuf, path, dirlen);
 441                 memcpy(pathbuf + dirlen, "/", 2);
 442                 cp = strchr(pathbuf + len + 1, '/');
 443                 strcpy(cp + 1, GITATTRIBUTES_FILE);
 444                 elem = read_attr_from_file(pathbuf, 0);
 445                 *cp = '\0';
 446                 elem->origin = strdup(pathbuf);
 447                 elem->prev = attr_stack;
 448                 attr_stack = elem;
 449                 debug_push(elem);
 450         }
 451
 452         /*
 453          * Finally push the "info" one at the top of the stack.
 454          */
 455         info->prev = attr_stack;
 456         attr_stack = info;
 457 }
 458
 459 static int path_matches(const char *pathname, int pathlen,
 460                         const char *pattern,
 461                         const char *base, int baselen)
 462 {
 463         if (!strchr(pattern, '/')) {
 464                 /* match basename */
 465                 const char *basename = strrchr(pathname, '/');
 466                 basename = basename ? basename + 1 : pathname;
 467                 return (fnmatch(pattern, basename, 0) == 0);
 468         }
 469         /*
 470          * match with FNM_PATHNAME; the pattern has base implicitly
 471          * in front of it.
 472          */
 473         if (*pattern == '/')
 474                 pattern++;
 475         if (pathlen < baselen ||
 476             (baselen && pathname[baselen - 1] != '/') ||
 477             strncmp(pathname, base, baselen))
 478                 return 0;
 479         return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
 480 }
 481
 482 static int fill_one(const char *what, struct match_attr *a, int rem)
 483 {
 484         struct git_attr_check *check = check_all_attr;
 485         int i;
 486
 487         for (i = 0; 0 < rem && i < a->num_attr; i++) {
 488                 struct git_attr *attr = a->state[i].attr;
 489                 const char **n = &(check[attr->attr_nr].value);
 490                 const char *v = a->state[i].setto;
 491
 492                 if (*n == ATTR__UNKNOWN) {
 493                         debug_set(what, a->u.pattern, attr, v);
 494                         *n = v;
 495                         rem--;
 496                 }
 497         }
 498         return rem;
 499 }
 500
 501 static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
 502 {
 503         int i;
 504         const char *base = stk->origin ? stk->origin : "";
 505
 506         for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 507                 struct match_attr *a = stk->attrs[i];
 508                 if (a->is_macro)
 509                         continue;
 510                 if (path_matches(path, pathlen,
 511                                  a->u.pattern, base, strlen(base)))
 512                         rem = fill_one("fill", a, rem);
 513         }
 514         return rem;
 515 }
 516
 517 static int macroexpand(struct attr_stack *stk, int rem)
 518 {
 519         int i;
 520         struct git_attr_check *check = check_all_attr;
 521
 522         for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 523                 struct match_attr *a = stk->attrs[i];
 524                 if (!a->is_macro)
 525                         continue;
 526                 if (check[a->u.attr->attr_nr].value != ATTR__TRUE)
 527                         continue;
 528                 rem = fill_one("expand", a, rem);
 529         }
 530         return rem;
 531 }
 532
 533 int git_checkattr(const char *path, int num, struct git_attr_check *check)
 534 {
 535         struct attr_stack *stk;
 536         const char *cp;
 537         int dirlen, pathlen, i, rem;
 538
 539         bootstrap_attr_stack();
 540         for (i = 0; i < attr_nr; i++)
 541                 check_all_attr[i].value = ATTR__UNKNOWN;
 542
 543         pathlen = strlen(path);
 544         cp = strrchr(path, '/');
 545         if (!cp)
 546                 dirlen = 0;
 547         else
 548                 dirlen = cp - path;
 549         prepare_attr_stack(path, dirlen);
 550         rem = attr_nr;
 551         for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 552                 rem = fill(path, pathlen, stk, rem);
 553
 554         for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 555                 rem = macroexpand(stk, rem);
 556
 557         for (i = 0; i < num; i++) {
 558                 const char *value = check_all_attr[check[i].attr->attr_nr].value;
 559                 if (value == ATTR__UNKNOWN)
 560                         value = ATTR__UNSET;
 561                 check[i].value = value;
 562         }
 563
 564         return 0;
 565 }