pathspec.c

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