attr.con commit Allow low-level driver to specify different behaviour during internal merge. (3086486)
   1#include "cache.h"
   2#include "attr.h"
   3
   4#define ATTR__UNKNOWN   ((void *) -2)
   5
   6/*
   7 * The basic design decision here is that we are not going to have
   8 * insanely large number of attributes.
   9 *
  10 * This is a randomly chosen prime.
  11 */
  12#define HASHSIZE 257
  13
  14#ifndef DEBUG_ATTR
  15#define DEBUG_ATTR 0
  16#endif
  17
  18struct git_attr {
  19        struct git_attr *next;
  20        unsigned h;
  21        int attr_nr;
  22        char name[FLEX_ARRAY];
  23};
  24static int attr_nr;
  25
  26static struct git_attr_check *check_all_attr;
  27static struct git_attr *(git_attr_hash[HASHSIZE]);
  28
  29static unsigned hash_name(const char *name, int namelen)
  30{
  31        unsigned val = 0;
  32        unsigned char c;
  33
  34        while (namelen--) {
  35                c = *name++;
  36                val = ((val << 7) | (val >> 22)) ^ c;
  37        }
  38        return val;
  39}
  40
  41static int invalid_attr_name(const char *name, int namelen)
  42{
  43        /*
  44         * Attribute name cannot begin with '-' and from
  45         * [-A-Za-z0-9_.].  We'd specifically exclude '=' for now,
  46         * as we might later want to allow non-binary value for
  47         * attributes, e.g. "*.svg      merge=special-merge-program-for-svg"
  48         */
  49        if (*name == '-')
  50                return -1;
  51        while (namelen--) {
  52                char ch = *name++;
  53                if (! (ch == '-' || ch == '.' || ch == '_' ||
  54                       ('0' <= ch && ch <= '9') ||
  55                       ('a' <= ch && ch <= 'z') ||
  56                       ('A' <= ch && ch <= 'Z')) )
  57                        return -1;
  58        }
  59        return 0;
  60}
  61
  62struct git_attr *git_attr(const char *name, int len)
  63{
  64        unsigned hval = hash_name(name, len);
  65        unsigned pos = hval % HASHSIZE;
  66        struct git_attr *a;
  67
  68        for (a = git_attr_hash[pos]; a; a = a->next) {
  69                if (a->h == hval &&
  70                    !memcmp(a->name, name, len) && !a->name[len])
  71                        return a;
  72        }
  73
  74        if (invalid_attr_name(name, len))
  75                return NULL;
  76
  77        a = xmalloc(sizeof(*a) + len + 1);
  78        memcpy(a->name, name, len);
  79        a->name[len] = 0;
  80        a->h = hval;
  81        a->next = git_attr_hash[pos];
  82        a->attr_nr = attr_nr++;
  83        git_attr_hash[pos] = a;
  84
  85        check_all_attr = xrealloc(check_all_attr,
  86                                  sizeof(*check_all_attr) * attr_nr);
  87        check_all_attr[a->attr_nr].attr = a;
  88        check_all_attr[a->attr_nr].value = ATTR__UNKNOWN;
  89        return a;
  90}
  91
  92/*
  93 * .gitattributes file is one line per record, each of which is
  94 *
  95 * (1) glob pattern.
  96 * (2) whitespace
  97 * (3) whitespace separated list of attribute names, each of which
  98 *     could be prefixed with '-' to mean "set to false", '!' to mean
  99 *     "unset".
 100 */
 101
 102/* What does a matched pattern decide? */
 103struct attr_state {
 104        struct git_attr *attr;
 105        void *setto;
 106};
 107
 108struct match_attr {
 109        union {
 110                char *pattern;
 111                struct git_attr *attr;
 112        } u;
 113        char is_macro;
 114        unsigned num_attr;
 115        struct attr_state state[FLEX_ARRAY];
 116};
 117
 118static const char blank[] = " \t\r\n";
 119
 120static const char *parse_attr(const char *src, int lineno, const char *cp,
 121                              int *num_attr, struct match_attr *res)
 122{
 123        const char *ep, *equals;
 124        int len;
 125
 126        ep = cp + strcspn(cp, blank);
 127        equals = strchr(cp, '=');
 128        if (equals && ep < equals)
 129                equals = NULL;
 130        if (equals)
 131                len = equals - cp;
 132        else
 133                len = ep - cp;
 134        if (!res) {
 135                if (*cp == '-' || *cp == '!') {
 136                        cp++;
 137                        len--;
 138                }
 139                if (invalid_attr_name(cp, len)) {
 140                        fprintf(stderr,
 141                                "%.*s is not a valid attribute name: %s:%d\n",
 142                                len, cp, src, lineno);
 143                        return NULL;
 144                }
 145        } else {
 146                struct attr_state *e;
 147
 148                e = &(res->state[*num_attr]);
 149                if (*cp == '-' || *cp == '!') {
 150                        e->setto = (*cp == '-') ? ATTR__FALSE : ATTR__UNSET;
 151                        cp++;
 152                        len--;
 153                }
 154                else if (!equals)
 155                        e->setto = ATTR__TRUE;
 156                else {
 157                        char *value;
 158                        int vallen = ep - equals;
 159                        value = xmalloc(vallen);
 160                        memcpy(value, equals+1, vallen-1);
 161                        value[vallen-1] = 0;
 162                        e->setto = value;
 163                }
 164                e->attr = git_attr(cp, len);
 165        }
 166        (*num_attr)++;
 167        return ep + strspn(ep, blank);
 168}
 169
 170static struct match_attr *parse_attr_line(const char *line, const char *src,
 171                                          int lineno, int macro_ok)
 172{
 173        int namelen;
 174        int num_attr;
 175        const char *cp, *name;
 176        struct match_attr *res = NULL;
 177        int pass;
 178        int is_macro;
 179
 180        cp = line + strspn(line, blank);
 181        if (!*cp || *cp == '#')
 182                return NULL;
 183        name = cp;
 184        namelen = strcspn(name, blank);
 185        if (strlen(ATTRIBUTE_MACRO_PREFIX) < namelen &&
 186            !prefixcmp(name, ATTRIBUTE_MACRO_PREFIX)) {
 187                if (!macro_ok) {
 188                        fprintf(stderr, "%s not allowed: %s:%d\n",
 189                                name, src, lineno);
 190                        return NULL;
 191                }
 192                is_macro = 1;
 193                name += strlen(ATTRIBUTE_MACRO_PREFIX);
 194                name += strspn(name, blank);
 195                namelen = strcspn(name, blank);
 196                if (invalid_attr_name(name, namelen)) {
 197                        fprintf(stderr,
 198                                "%.*s is not a valid attribute name: %s:%d\n",
 199                                namelen, name, src, lineno);
 200                        return NULL;
 201                }
 202        }
 203        else
 204                is_macro = 0;
 205
 206        for (pass = 0; pass < 2; pass++) {
 207                /* pass 0 counts and allocates, pass 1 fills */
 208                num_attr = 0;
 209                cp = name + namelen;
 210                cp = cp + strspn(cp, blank);
 211                while (*cp)
 212                        cp = parse_attr(src, lineno, cp, &num_attr, res);
 213                if (pass)
 214                        break;
 215                res = xcalloc(1,
 216                              sizeof(*res) +
 217                              sizeof(struct attr_state) * num_attr +
 218                              (is_macro ? 0 : namelen + 1));
 219                if (is_macro)
 220                        res->u.attr = git_attr(name, namelen);
 221                else {
 222                        res->u.pattern = (char*)&(res->state[num_attr]);
 223                        memcpy(res->u.pattern, name, namelen);
 224                        res->u.pattern[namelen] = 0;
 225                }
 226                res->is_macro = is_macro;
 227                res->num_attr = num_attr;
 228        }
 229        return res;
 230}
 231
 232/*
 233 * Like info/exclude and .gitignore, the attribute information can
 234 * come from many places.
 235 *
 236 * (1) .gitattribute file of the same directory;
 237 * (2) .gitattribute file of the parent directory if (1) does not have
 238 *      any match; this goes recursively upwards, just like .gitignore.
 239 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
 240 *
 241 * In the same file, later entries override the earlier match, so in the
 242 * global list, we would have entries from info/attributes the earliest
 243 * (reading the file from top to bottom), .gitattribute of the root
 244 * directory (again, reading the file from top to bottom) down to the
 245 * current directory, and then scan the list backwards to find the first match.
 246 * This is exactly the same as what excluded() does in dir.c to deal with
 247 * .gitignore
 248 */
 249
 250static struct attr_stack {
 251        struct attr_stack *prev;
 252        char *origin;
 253        unsigned num_matches;
 254        struct match_attr **attrs;
 255} *attr_stack;
 256
 257static void free_attr_elem(struct attr_stack *e)
 258{
 259        int i;
 260        free(e->origin);
 261        for (i = 0; i < e->num_matches; i++) {
 262                struct match_attr *a = e->attrs[i];
 263                int j;
 264                for (j = 0; j < a->num_attr; j++) {
 265                        void *setto = a->state[j].setto;
 266                        if (setto == ATTR__TRUE ||
 267                            setto == ATTR__FALSE ||
 268                            setto == ATTR__UNSET ||
 269                            setto == ATTR__UNKNOWN)
 270                                ;
 271                        else
 272                                free(setto);
 273                }
 274                free(a);
 275        }
 276        free(e);
 277}
 278
 279static const char *builtin_attr[] = {
 280        "[attr]binary -diff -crlf",
 281        NULL,
 282};
 283
 284static struct attr_stack *read_attr_from_array(const char **list)
 285{
 286        struct attr_stack *res;
 287        const char *line;
 288        int lineno = 0;
 289
 290        res = xcalloc(1, sizeof(*res));
 291        while ((line = *(list++)) != NULL) {
 292                struct match_attr *a;
 293
 294                a = parse_attr_line(line, "[builtin]", ++lineno, 1);
 295                if (!a)
 296                        continue;
 297                res->attrs = xrealloc(res->attrs, res->num_matches + 1);
 298                res->attrs[res->num_matches++] = a;
 299        }
 300        return res;
 301}
 302
 303static struct attr_stack *read_attr_from_file(const char *path, int macro_ok)
 304{
 305        FILE *fp;
 306        struct attr_stack *res;
 307        char buf[2048];
 308        int lineno = 0;
 309
 310        res = xcalloc(1, sizeof(*res));
 311        fp = fopen(path, "r");
 312        if (!fp)
 313                return res;
 314
 315        while (fgets(buf, sizeof(buf), fp)) {
 316                struct match_attr *a;
 317
 318                a = parse_attr_line(buf, path, ++lineno, macro_ok);
 319                if (!a)
 320                        continue;
 321                res->attrs = xrealloc(res->attrs, res->num_matches + 1);
 322                res->attrs[res->num_matches++] = a;
 323        }
 324        fclose(fp);
 325        return res;
 326}
 327
 328#if DEBUG_ATTR
 329static void debug_info(const char *what, struct attr_stack *elem)
 330{
 331        fprintf(stderr, "%s: %s\n", what, elem->origin ? elem->origin : "()");
 332}
 333static void debug_set(const char *what, const char *match, struct git_attr *attr, void *v)
 334{
 335        const char *value = v;
 336
 337        if (ATTR_TRUE(value))
 338                value = "set";
 339        else if (ATTR_FALSE(value))
 340                value = "unset";
 341        else if (ATTR_UNSET(value))
 342                value = "unspecified";
 343
 344        fprintf(stderr, "%s: %s => %s (%s)\n",
 345                what, attr->name, (char *) value, match);
 346}
 347#define debug_push(a) debug_info("push", (a))
 348#define debug_pop(a) debug_info("pop", (a))
 349#else
 350#define debug_push(a) do { ; } while (0)
 351#define debug_pop(a) do { ; } while (0)
 352#define debug_set(a,b,c,d) do { ; } while (0)
 353#endif
 354
 355static void bootstrap_attr_stack(void)
 356{
 357        if (!attr_stack) {
 358                struct attr_stack *elem;
 359
 360                elem = read_attr_from_array(builtin_attr);
 361                elem->origin = NULL;
 362                elem->prev = attr_stack;
 363                attr_stack = elem;
 364
 365                elem = read_attr_from_file(GITATTRIBUTES_FILE, 1);
 366                elem->origin = strdup("");
 367                elem->prev = attr_stack;
 368                attr_stack = elem;
 369                debug_push(elem);
 370
 371                elem = read_attr_from_file(git_path(INFOATTRIBUTES_FILE), 1);
 372                elem->origin = NULL;
 373                elem->prev = attr_stack;
 374                attr_stack = elem;
 375        }
 376}
 377
 378static void prepare_attr_stack(const char *path, int dirlen)
 379{
 380        struct attr_stack *elem, *info;
 381        int len;
 382        char pathbuf[PATH_MAX];
 383
 384        /*
 385         * At the bottom of the attribute stack is the built-in
 386         * set of attribute definitions.  Then, contents from
 387         * .gitattribute files from directories closer to the
 388         * root to the ones in deeper directories are pushed
 389         * to the stack.  Finally, at the very top of the stack
 390         * we always keep the contents of $GIT_DIR/info/attributes.
 391         *
 392         * When checking, we use entries from near the top of the
 393         * stack, preferring $GIT_DIR/info/attributes, then
 394         * .gitattributes in deeper directories to shallower ones,
 395         * and finally use the built-in set as the default.
 396         */
 397        if (!attr_stack)
 398                bootstrap_attr_stack();
 399
 400        /*
 401         * Pop the "info" one that is always at the top of the stack.
 402         */
 403        info = attr_stack;
 404        attr_stack = info->prev;
 405
 406        /*
 407         * Pop the ones from directories that are not the prefix of
 408         * the path we are checking.
 409         */
 410        while (attr_stack && attr_stack->origin) {
 411                int namelen = strlen(attr_stack->origin);
 412
 413                elem = attr_stack;
 414                if (namelen <= dirlen &&
 415                    !strncmp(elem->origin, path, namelen))
 416                        break;
 417
 418                debug_pop(elem);
 419                attr_stack = elem->prev;
 420                free_attr_elem(elem);
 421        }
 422
 423        /*
 424         * Read from parent directories and push them down
 425         */
 426        while (1) {
 427                char *cp;
 428
 429                len = strlen(attr_stack->origin);
 430                if (dirlen <= len)
 431                        break;
 432                memcpy(pathbuf, path, dirlen);
 433                memcpy(pathbuf + dirlen, "/", 2);
 434                cp = strchr(pathbuf + len + 1, '/');
 435                strcpy(cp + 1, GITATTRIBUTES_FILE);
 436                elem = read_attr_from_file(pathbuf, 0);
 437                *cp = '\0';
 438                elem->origin = strdup(pathbuf);
 439                elem->prev = attr_stack;
 440                attr_stack = elem;
 441                debug_push(elem);
 442        }
 443
 444        /*
 445         * Finally push the "info" one at the top of the stack.
 446         */
 447        info->prev = attr_stack;
 448        attr_stack = info;
 449}
 450
 451static int path_matches(const char *pathname, int pathlen,
 452                        const char *pattern,
 453                        const char *base, int baselen)
 454{
 455        if (!strchr(pattern, '/')) {
 456                /* match basename */
 457                const char *basename = strrchr(pathname, '/');
 458                basename = basename ? basename + 1 : pathname;
 459                return (fnmatch(pattern, basename, 0) == 0);
 460        }
 461        /*
 462         * match with FNM_PATHNAME; the pattern has base implicitly
 463         * in front of it.
 464         */
 465        if (*pattern == '/')
 466                pattern++;
 467        if (pathlen < baselen ||
 468            (baselen && pathname[baselen - 1] != '/') ||
 469            strncmp(pathname, base, baselen))
 470                return 0;
 471        return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0;
 472}
 473
 474static int fill_one(const char *what, struct match_attr *a, int rem)
 475{
 476        struct git_attr_check *check = check_all_attr;
 477        int i;
 478
 479        for (i = 0; 0 < rem && i < a->num_attr; i++) {
 480                struct git_attr *attr = a->state[i].attr;
 481                void **n = &(check[attr->attr_nr].value);
 482                void *v = a->state[i].setto;
 483
 484                if (*n == ATTR__UNKNOWN) {
 485                        debug_set(what, a->u.pattern, attr, v);
 486                        *n = v;
 487                        rem--;
 488                }
 489        }
 490        return rem;
 491}
 492
 493static int fill(const char *path, int pathlen, struct attr_stack *stk, int rem)
 494{
 495        int i;
 496        const char *base = stk->origin ? stk->origin : "";
 497
 498        for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 499                struct match_attr *a = stk->attrs[i];
 500                if (a->is_macro)
 501                        continue;
 502                if (path_matches(path, pathlen,
 503                                 a->u.pattern, base, strlen(base)))
 504                        rem = fill_one("fill", a, rem);
 505        }
 506        return rem;
 507}
 508
 509static int macroexpand(struct attr_stack *stk, int rem)
 510{
 511        int i;
 512        struct git_attr_check *check = check_all_attr;
 513
 514        for (i = stk->num_matches - 1; 0 < rem && 0 <= i; i--) {
 515                struct match_attr *a = stk->attrs[i];
 516                if (!a->is_macro)
 517                        continue;
 518                if (check[a->u.attr->attr_nr].value != ATTR__TRUE)
 519                        continue;
 520                rem = fill_one("expand", a, rem);
 521        }
 522        return rem;
 523}
 524
 525int git_checkattr(const char *path, int num, struct git_attr_check *check)
 526{
 527        struct attr_stack *stk;
 528        const char *cp;
 529        int dirlen, pathlen, i, rem;
 530
 531        bootstrap_attr_stack();
 532        for (i = 0; i < attr_nr; i++)
 533                check_all_attr[i].value = ATTR__UNKNOWN;
 534
 535        pathlen = strlen(path);
 536        cp = strrchr(path, '/');
 537        if (!cp)
 538                dirlen = 0;
 539        else
 540                dirlen = cp - path;
 541        prepare_attr_stack(path, dirlen);
 542        rem = attr_nr;
 543        for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 544                rem = fill(path, pathlen, stk, rem);
 545
 546        for (stk = attr_stack; 0 < rem && stk; stk = stk->prev)
 547                rem = macroexpand(stk, rem);
 548
 549        for (i = 0; i < num; i++) {
 550                void *value = check_all_attr[check[i].attr->attr_nr].value;
 551                if (value == ATTR__UNKNOWN)
 552                        value = ATTR__UNSET;
 553                check[i].value = value;
 554        }
 555
 556        return 0;
 557}