pathspec.c

   1 #define NO_THE_INDEX_COMPATIBILITY_MACROS
   2 #include "cache.h"
   3 #include "config.h"
   4 #include "dir.h"
   5 #include "pathspec.h"
   6 #include "attr.h"
   7
   8 /*
   9  * Finds which of the given pathspecs match items in the index.
  10  *
  11  * For each pathspec, sets the corresponding entry in the seen[] array
  12  * (which should be specs items long, i.e. the same size as pathspec)
  13  * to the nature of the "closest" (i.e. most specific) match found for
  14  * that pathspec in the index, if it was a closer type of match than
  15  * the existing entry.  As an optimization, matching is skipped
  16  * altogether if seen[] already only contains non-zero entries.
  17  *
  18  * If seen[] has not already been written to, it may make sense
  19  * to use find_pathspecs_matching_against_index() instead.
  20  */
  21 void add_pathspec_matches_against_index(const struct pathspec *pathspec,
  22                                         const struct index_state *istate,
  23                                         char *seen)
  24 {
  25         int num_unmatched = 0, i;
  26
  27         /*
  28          * Since we are walking the index as if we were walking the directory,
  29          * we have to mark the matched pathspec as seen; otherwise we will
  30          * mistakenly think that the user gave a pathspec that did not match
  31          * anything.
  32          */
  33         for (i = 0; i < pathspec->nr; i++)
  34                 if (!seen[i])
  35                         num_unmatched++;
  36         if (!num_unmatched)
  37                 return;
  38         for (i = 0; i < istate->cache_nr; i++) {
  39                 const struct cache_entry *ce = istate->cache[i];
  40                 ce_path_match(ce, pathspec, seen);
  41         }
  42 }
  43
  44 /*
  45  * Finds which of the given pathspecs match items in the index.
  46  *
  47  * This is a one-shot wrapper around add_pathspec_matches_against_index()
  48  * which allocates, populates, and returns a seen[] array indicating the
  49  * nature of the "closest" (i.e. most specific) matches which each of the
  50  * given pathspecs achieves against all items in the index.
  51  */
  52 char *find_pathspecs_matching_against_index(const struct pathspec *pathspec,
  53                                             const struct index_state *istate)
  54 {
  55         char *seen = xcalloc(pathspec->nr, 1);
  56         add_pathspec_matches_against_index(pathspec, istate, seen);
  57         return seen;
  58 }
  59
  60 /*
  61  * Magic pathspec
  62  *
  63  * Possible future magic semantics include stuff like:
  64  *
  65  *      { PATHSPEC_RECURSIVE, '*', "recursive" },
  66  *      { PATHSPEC_REGEXP, '\0', "regexp" },
  67  *
  68  */
  69
  70 static struct pathspec_magic {
  71         unsigned bit;
  72         char mnemonic; /* this cannot be ':'! */
  73         const char *name;
  74 } pathspec_magic[] = {
  75         { PATHSPEC_FROMTOP,  '/', "top" },
  76         { PATHSPEC_LITERAL, '\0', "literal" },
  77         { PATHSPEC_GLOB,    '\0', "glob" },
  78         { PATHSPEC_ICASE,   '\0', "icase" },
  79         { PATHSPEC_EXCLUDE,  '!', "exclude" },
  80         { PATHSPEC_ATTR,    '\0', "attr" },
  81 };
  82
  83 static void prefix_magic(struct strbuf *sb, int prefixlen, unsigned magic)
  84 {
  85         int i;
  86         strbuf_addstr(sb, ":(");
  87         for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++)
  88                 if (magic & pathspec_magic[i].bit) {
  89                         if (sb->buf[sb->len - 1] != '(')
  90                                 strbuf_addch(sb, ',');
  91                         strbuf_addstr(sb, pathspec_magic[i].name);
  92                 }
  93         strbuf_addf(sb, ",prefix:%d)", prefixlen);
  94 }
  95
  96 static size_t strcspn_escaped(const char *s, const char *stop)
  97 {
  98         const char *i;
  99
 100         for (i = s; *i; i++) {
 101                 /* skip the escaped character */
 102                 if (i[0] == '\\' && i[1]) {
 103                         i++;
 104                         continue;
 105                 }
 106
 107                 if (strchr(stop, *i))
 108                         break;
 109         }
 110         return i - s;
 111 }
 112
 113 static inline int invalid_value_char(const char ch)
 114 {
 115         if (isalnum(ch) || strchr(",-_", ch))
 116                 return 0;
 117         return -1;
 118 }
 119
 120 static char *attr_value_unescape(const char *value)
 121 {
 122         const char *src;
 123         char *dst, *ret;
 124
 125         ret = xmallocz(strlen(value));
 126         for (src = value, dst = ret; *src; src++, dst++) {
 127                 if (*src == '\\') {
 128                         if (!src[1])
 129                                 die(_("Escape character '\\' not allowed as "
 130                                       "last character in attr value"));
 131                         src++;
 132                 }
 133                 if (invalid_value_char(*src))
 134                         die("cannot use '%c' for value matching", *src);
 135                 *dst = *src;
 136         }
 137         *dst = '\0';
 138         return ret;
 139 }
 140
 141 static void parse_pathspec_attr_match(struct pathspec_item *item, const char *value)
 142 {
 143         struct string_list_item *si;
 144         struct string_list list = STRING_LIST_INIT_DUP;
 145
 146         if (item->attr_check || item->attr_match)
 147                 die(_("Only one 'attr:' specification is allowed."));
 148
 149         if (!value || !*value)
 150                 die(_("attr spec must not be empty"));
 151
 152         string_list_split(&list, value, ' ', -1);
 153         string_list_remove_empty_items(&list, 0);
 154
 155         item->attr_check = attr_check_alloc();
 156         item->attr_match = xcalloc(list.nr, sizeof(struct attr_match));
 157
 158         for_each_string_list_item(si, &list) {
 159                 size_t attr_len;
 160                 char *attr_name;
 161                 const struct git_attr *a;
 162
 163                 int j = item->attr_match_nr++;
 164                 const char *attr = si->string;
 165                 struct attr_match *am = &item->attr_match[j];
 166
 167                 switch (*attr) {
 168                 case '!':
 169                         am->match_mode = MATCH_UNSPECIFIED;
 170                         attr++;
 171                         attr_len = strlen(attr);
 172                         break;
 173                 case '-':
 174                         am->match_mode = MATCH_UNSET;
 175                         attr++;
 176                         attr_len = strlen(attr);
 177                         break;
 178                 default:
 179                         attr_len = strcspn(attr, "=");
 180                         if (attr[attr_len] != '=')
 181                                 am->match_mode = MATCH_SET;
 182                         else {
 183                                 const char *v = &attr[attr_len + 1];
 184                                 am->match_mode = MATCH_VALUE;
 185                                 am->value = attr_value_unescape(v);
 186                         }
 187                         break;
 188                 }
 189
 190                 attr_name = xmemdupz(attr, attr_len);
 191                 a = git_attr(attr_name);
 192                 if (!a)
 193                         die(_("invalid attribute name %s"), attr_name);
 194
 195                 attr_check_append(item->attr_check, a);
 196
 197                 free(attr_name);
 198         }
 199
 200         if (item->attr_check->nr != item->attr_match_nr)
 201                 die("BUG: should have same number of entries");
 202
 203         string_list_clear(&list, 0);
 204 }
 205
 206 static inline int get_literal_global(void)
 207 {
 208         static int literal = -1;
 209
 210         if (literal < 0)
 211                 literal = git_env_bool(GIT_LITERAL_PATHSPECS_ENVIRONMENT, 0);
 212
 213         return literal;
 214 }
 215
 216 static inline int get_glob_global(void)
 217 {
 218         static int glob = -1;
 219
 220         if (glob < 0)
 221                 glob = git_env_bool(GIT_GLOB_PATHSPECS_ENVIRONMENT, 0);
 222
 223         return glob;
 224 }
 225
 226 static inline int get_noglob_global(void)
 227 {
 228         static int noglob = -1;
 229
 230         if (noglob < 0)
 231                 noglob = git_env_bool(GIT_NOGLOB_PATHSPECS_ENVIRONMENT, 0);
 232
 233         return noglob;
 234 }
 235
 236 static inline int get_icase_global(void)
 237 {
 238         static int icase = -1;
 239
 240         if (icase < 0)
 241                 icase = git_env_bool(GIT_ICASE_PATHSPECS_ENVIRONMENT, 0);
 242
 243         return icase;
 244 }
 245
 246 static int get_global_magic(int element_magic)
 247 {
 248         int global_magic = 0;
 249
 250         if (get_literal_global())
 251                 global_magic |= PATHSPEC_LITERAL;
 252
 253         /* --glob-pathspec is overridden by :(literal) */
 254         if (get_glob_global() && !(element_magic & PATHSPEC_LITERAL))
 255                 global_magic |= PATHSPEC_GLOB;
 256
 257         if (get_glob_global() && get_noglob_global())
 258                 die(_("global 'glob' and 'noglob' pathspec settings are incompatible"));
 259
 260         if (get_icase_global())
 261                 global_magic |= PATHSPEC_ICASE;
 262
 263         if ((global_magic & PATHSPEC_LITERAL) &&
 264             (global_magic & ~PATHSPEC_LITERAL))
 265                 die(_("global 'literal' pathspec setting is incompatible "
 266                       "with all other global pathspec settings"));
 267
 268         /* --noglob-pathspec adds :(literal) _unless_ :(glob) is specified */
 269         if (get_noglob_global() && !(element_magic & PATHSPEC_GLOB))
 270                 global_magic |= PATHSPEC_LITERAL;
 271
 272         return global_magic;
 273 }
 274
 275 /*
 276  * Parse the pathspec element looking for long magic
 277  *
 278  * saves all magic in 'magic'
 279  * if prefix magic is used, save the prefix length in 'prefix_len'
 280  * returns the position in 'elem' after all magic has been parsed
 281  */
 282 static const char *parse_long_magic(unsigned *magic, int *prefix_len,
 283                                     struct pathspec_item *item,
 284                                     const char *elem)
 285 {
 286         const char *pos;
 287         const char *nextat;
 288
 289         for (pos = elem + 2; *pos && *pos != ')'; pos = nextat) {
 290                 size_t len = strcspn_escaped(pos, ",)");
 291                 int i;
 292
 293                 if (pos[len] == ',')
 294                         nextat = pos + len + 1; /* handle ',' */
 295                 else
 296                         nextat = pos + len; /* handle ')' and '\0' */
 297
 298                 if (!len)
 299                         continue;
 300
 301                 if (starts_with(pos, "prefix:")) {
 302                         char *endptr;
 303                         *prefix_len = strtol(pos + 7, &endptr, 10);
 304                         if (endptr - pos != len)
 305                                 die(_("invalid parameter for pathspec magic 'prefix'"));
 306                         continue;
 307                 }
 308
 309                 if (starts_with(pos, "attr:")) {
 310                         char *attr_body = xmemdupz(pos + 5, len - 5);
 311                         parse_pathspec_attr_match(item, attr_body);
 312                         *magic |= PATHSPEC_ATTR;
 313                         free(attr_body);
 314                         continue;
 315                 }
 316
 317                 for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 318                         if (strlen(pathspec_magic[i].name) == len &&
 319                             !strncmp(pathspec_magic[i].name, pos, len)) {
 320                                 *magic |= pathspec_magic[i].bit;
 321                                 break;
 322                         }
 323                 }
 324
 325                 if (ARRAY_SIZE(pathspec_magic) <= i)
 326                         die(_("Invalid pathspec magic '%.*s' in '%s'"),
 327                             (int) len, pos, elem);
 328         }
 329
 330         if (*pos != ')')
 331                 die(_("Missing ')' at the end of pathspec magic in '%s'"),
 332                     elem);
 333         pos++;
 334
 335         return pos;
 336 }
 337
 338 /*
 339  * Parse the pathspec element looking for short magic
 340  *
 341  * saves all magic in 'magic'
 342  * returns the position in 'elem' after all magic has been parsed
 343  */
 344 static const char *parse_short_magic(unsigned *magic, const char *elem)
 345 {
 346         const char *pos;
 347
 348         for (pos = elem + 1; *pos && *pos != ':'; pos++) {
 349                 char ch = *pos;
 350                 int i;
 351
 352                 /* Special case alias for '!' */
 353                 if (ch == '^') {
 354                         *magic |= PATHSPEC_EXCLUDE;
 355                         continue;
 356                 }
 357
 358                 if (!is_pathspec_magic(ch))
 359                         break;
 360
 361                 for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 362                         if (pathspec_magic[i].mnemonic == ch) {
 363                                 *magic |= pathspec_magic[i].bit;
 364                                 break;
 365                         }
 366                 }
 367
 368                 if (ARRAY_SIZE(pathspec_magic) <= i)
 369                         die(_("Unimplemented pathspec magic '%c' in '%s'"),
 370                             ch, elem);
 371         }
 372
 373         if (*pos == ':')
 374                 pos++;
 375
 376         return pos;
 377 }
 378
 379 static const char *parse_element_magic(unsigned *magic, int *prefix_len,
 380                                        struct pathspec_item *item,
 381                                        const char *elem)
 382 {
 383         if (elem[0] != ':' || get_literal_global())
 384                 return elem; /* nothing to do */
 385         else if (elem[1] == '(')
 386                 /* longhand */
 387                 return parse_long_magic(magic, prefix_len, item, elem);
 388         else
 389                 /* shorthand */
 390                 return parse_short_magic(magic, elem);
 391 }
 392
 393 /*
 394  * Perform the initialization of a pathspec_item based on a pathspec element.
 395  */
 396 static void init_pathspec_item(struct pathspec_item *item, unsigned flags,
 397                                const char *prefix, int prefixlen,
 398                                const char *elt)
 399 {
 400         unsigned magic = 0, element_magic = 0;
 401         const char *copyfrom = elt;
 402         char *match;
 403         int pathspec_prefix = -1;
 404
 405         item->attr_check = NULL;
 406         item->attr_match = NULL;
 407         item->attr_match_nr = 0;
 408
 409         /* PATHSPEC_LITERAL_PATH ignores magic */
 410         if (flags & PATHSPEC_LITERAL_PATH) {
 411                 magic = PATHSPEC_LITERAL;
 412         } else {
 413                 copyfrom = parse_element_magic(&element_magic,
 414                                                &pathspec_prefix,
 415                                                item,
 416                                                elt);
 417                 magic |= element_magic;
 418                 magic |= get_global_magic(element_magic);
 419         }
 420
 421         item->magic = magic;
 422
 423         if (pathspec_prefix >= 0 &&
 424             (prefixlen || (prefix && *prefix)))
 425                 die("BUG: 'prefix' magic is supposed to be used at worktree's root");
 426
 427         if ((magic & PATHSPEC_LITERAL) && (magic & PATHSPEC_GLOB))
 428                 die(_("%s: 'literal' and 'glob' are incompatible"), elt);
 429
 430         /* Create match string which will be used for pathspec matching */
 431         if (pathspec_prefix >= 0) {
 432                 match = xstrdup(copyfrom);
 433                 prefixlen = pathspec_prefix;
 434         } else if (magic & PATHSPEC_FROMTOP) {
 435                 match = xstrdup(copyfrom);
 436                 prefixlen = 0;
 437         } else {
 438                 match = prefix_path_gently(prefix, prefixlen,
 439                                            &prefixlen, copyfrom);
 440                 if (!match)
 441                         die(_("%s: '%s' is outside repository"), elt, copyfrom);
 442         }
 443
 444         item->match = match;
 445         item->len = strlen(item->match);
 446         item->prefix = prefixlen;
 447
 448         /*
 449          * Prefix the pathspec (keep all magic) and assign to
 450          * original. Useful for passing to another command.
 451          */
 452         if ((flags & PATHSPEC_PREFIX_ORIGIN) &&
 453             !get_literal_global()) {
 454                 struct strbuf sb = STRBUF_INIT;
 455
 456                 /* Preserve the actual prefix length of each pattern */
 457                 prefix_magic(&sb, prefixlen, element_magic);
 458
 459                 strbuf_addstr(&sb, match);
 460                 item->original = strbuf_detach(&sb, NULL);
 461         } else {
 462                 item->original = xstrdup(elt);
 463         }
 464
 465         if (magic & PATHSPEC_LITERAL) {
 466                 item->nowildcard_len = item->len;
 467         } else {
 468                 item->nowildcard_len = simple_length(item->match);
 469                 if (item->nowildcard_len < prefixlen)
 470                         item->nowildcard_len = prefixlen;
 471         }
 472
 473         item->flags = 0;
 474         if (magic & PATHSPEC_GLOB) {
 475                 /*
 476                  * FIXME: should we enable ONESTAR in _GLOB for
 477                  * pattern "* * / * . c"?
 478                  */
 479         } else {
 480                 if (item->nowildcard_len < item->len &&
 481                     item->match[item->nowildcard_len] == '*' &&
 482                     no_wildcard(item->match + item->nowildcard_len + 1))
 483                         item->flags |= PATHSPEC_ONESTAR;
 484         }
 485
 486         /* sanity checks, pathspec matchers assume these are sane */
 487         if (item->nowildcard_len > item->len ||
 488             item->prefix         > item->len) {
 489                 die ("BUG: error initializing pathspec_item");
 490         }
 491 }
 492
 493 static int pathspec_item_cmp(const void *a_, const void *b_)
 494 {
 495         struct pathspec_item *a, *b;
 496
 497         a = (struct pathspec_item *)a_;
 498         b = (struct pathspec_item *)b_;
 499         return strcmp(a->match, b->match);
 500 }
 501
 502 static void NORETURN unsupported_magic(const char *pattern,
 503                                        unsigned magic)
 504 {
 505         struct strbuf sb = STRBUF_INIT;
 506         int i;
 507         for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 508                 const struct pathspec_magic *m = pathspec_magic + i;
 509                 if (!(magic & m->bit))
 510                         continue;
 511                 if (sb.len)
 512                         strbuf_addstr(&sb, ", ");
 513
 514                 if (m->mnemonic)
 515                         strbuf_addf(&sb, _("'%s' (mnemonic: '%c')"),
 516                                     m->name, m->mnemonic);
 517                 else
 518                         strbuf_addf(&sb, "'%s'", m->name);
 519         }
 520         /*
 521          * We may want to substitute "this command" with a command
 522          * name. E.g. when add--interactive dies when running
 523          * "checkout -p"
 524          */
 525         die(_("%s: pathspec magic not supported by this command: %s"),
 526             pattern, sb.buf);
 527 }
 528
 529 /*
 530  * Given command line arguments and a prefix, convert the input to
 531  * pathspec. die() if any magic in magic_mask is used.
 532  */
 533 void parse_pathspec(struct pathspec *pathspec,
 534                     unsigned magic_mask, unsigned flags,
 535                     const char *prefix, const char **argv)
 536 {
 537         struct pathspec_item *item;
 538         const char *entry = argv ? *argv : NULL;
 539         int i, n, prefixlen, warn_empty_string, nr_exclude = 0;
 540
 541         memset(pathspec, 0, sizeof(*pathspec));
 542
 543         if (flags & PATHSPEC_MAXDEPTH_VALID)
 544                 pathspec->magic |= PATHSPEC_MAXDEPTH;
 545
 546         /* No arguments, no prefix -> no pathspec */
 547         if (!entry && !prefix)
 548                 return;
 549
 550         if ((flags & PATHSPEC_PREFER_CWD) &&
 551             (flags & PATHSPEC_PREFER_FULL))
 552                 die("BUG: PATHSPEC_PREFER_CWD and PATHSPEC_PREFER_FULL are incompatible");
 553
 554         /* No arguments with prefix -> prefix pathspec */
 555         if (!entry) {
 556                 if (flags & PATHSPEC_PREFER_FULL)
 557                         return;
 558
 559                 if (!(flags & PATHSPEC_PREFER_CWD))
 560                         die("BUG: PATHSPEC_PREFER_CWD requires arguments");
 561
 562                 pathspec->items = item = xcalloc(1, sizeof(*item));
 563                 item->match = xstrdup(prefix);
 564                 item->original = xstrdup(prefix);
 565                 item->nowildcard_len = item->len = strlen(prefix);
 566                 item->prefix = item->len;
 567                 pathspec->nr = 1;
 568                 return;
 569         }
 570
 571         n = 0;
 572         warn_empty_string = 1;
 573         while (argv[n]) {
 574                 if (*argv[n] == '\0' && warn_empty_string) {
 575                         warning(_("empty strings as pathspecs will be made invalid in upcoming releases. "
 576                                   "please use . instead if you meant to match all paths"));
 577                         warn_empty_string = 0;
 578                 }
 579                 n++;
 580         }
 581
 582         pathspec->nr = n;
 583         ALLOC_ARRAY(pathspec->items, n + 1);
 584         item = pathspec->items;
 585         prefixlen = prefix ? strlen(prefix) : 0;
 586
 587         for (i = 0; i < n; i++) {
 588                 entry = argv[i];
 589
 590                 init_pathspec_item(item + i, flags, prefix, prefixlen, entry);
 591
 592                 if (item[i].magic & PATHSPEC_EXCLUDE)
 593                         nr_exclude++;
 594                 if (item[i].magic & magic_mask)
 595                         unsupported_magic(entry, item[i].magic & magic_mask);
 596
 597                 if ((flags & PATHSPEC_SYMLINK_LEADING_PATH) &&
 598                     has_symlink_leading_path(item[i].match, item[i].len)) {
 599                         die(_("pathspec '%s' is beyond a symbolic link"), entry);
 600                 }
 601
 602                 if (item[i].nowildcard_len < item[i].len)
 603                         pathspec->has_wildcard = 1;
 604                 pathspec->magic |= item[i].magic;
 605         }
 606
 607         /*
 608          * If everything is an exclude pattern, add one positive pattern
 609          * that matches everything. We allocated an extra one for this.
 610          */
 611         if (nr_exclude == n) {
 612                 int plen = (!(flags & PATHSPEC_PREFER_CWD)) ? 0 : prefixlen;
 613                 init_pathspec_item(item + n, 0, prefix, plen, "");
 614                 pathspec->nr++;
 615         }
 616
 617         if (pathspec->magic & PATHSPEC_MAXDEPTH) {
 618                 if (flags & PATHSPEC_KEEP_ORDER)
 619                         die("BUG: PATHSPEC_MAXDEPTH_VALID and PATHSPEC_KEEP_ORDER are incompatible");
 620                 QSORT(pathspec->items, pathspec->nr, pathspec_item_cmp);
 621         }
 622 }
 623
 624 void copy_pathspec(struct pathspec *dst, const struct pathspec *src)
 625 {
 626         int i, j;
 627
 628         *dst = *src;
 629         ALLOC_ARRAY(dst->items, dst->nr);
 630         COPY_ARRAY(dst->items, src->items, dst->nr);
 631
 632         for (i = 0; i < dst->nr; i++) {
 633                 struct pathspec_item *d = &dst->items[i];
 634                 struct pathspec_item *s = &src->items[i];
 635
 636                 d->match = xstrdup(s->match);
 637                 d->original = xstrdup(s->original);
 638
 639                 ALLOC_ARRAY(d->attr_match, d->attr_match_nr);
 640                 COPY_ARRAY(d->attr_match, s->attr_match, d->attr_match_nr);
 641                 for (j = 0; j < d->attr_match_nr; j++) {
 642                         const char *value = s->attr_match[j].value;
 643                         d->attr_match[j].value = xstrdup_or_null(value);
 644                 }
 645
 646                 d->attr_check = attr_check_dup(s->attr_check);
 647         }
 648 }
 649
 650 void clear_pathspec(struct pathspec *pathspec)
 651 {
 652         int i, j;
 653
 654         for (i = 0; i < pathspec->nr; i++) {
 655                 free(pathspec->items[i].match);
 656                 free(pathspec->items[i].original);
 657
 658                 for (j = 0; j < pathspec->items[i].attr_match_nr; j++)
 659                         free(pathspec->items[i].attr_match[j].value);
 660                 free(pathspec->items[i].attr_match);
 661
 662                 if (pathspec->items[i].attr_check)
 663                         attr_check_free(pathspec->items[i].attr_check);
 664         }
 665
 666         FREE_AND_NULL(pathspec->items);
 667         pathspec->nr = 0;
 668 }