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                 /* Special case alias for '!' */
 228                 if (ch == '^') {
 229                         *magic |= PATHSPEC_EXCLUDE;
 230                         continue;
 231                 }
 232
 233                 if (!is_pathspec_magic(ch))
 234                         break;
 235
 236                 for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 237                         if (pathspec_magic[i].mnemonic == ch) {
 238                                 *magic |= pathspec_magic[i].bit;
 239                                 break;
 240                         }
 241                 }
 242
 243                 if (ARRAY_SIZE(pathspec_magic) <= i)
 244                         die(_("Unimplemented pathspec magic '%c' in '%s'"),
 245                             ch, elem);
 246         }
 247
 248         if (*pos == ':')
 249                 pos++;
 250
 251         return pos;
 252 }
 253
 254 static const char *parse_element_magic(unsigned *magic, int *prefix_len,
 255                                        const char *elem)
 256 {
 257         if (elem[0] != ':' || get_literal_global())
 258                 return elem; /* nothing to do */
 259         else if (elem[1] == '(')
 260                 /* longhand */
 261                 return parse_long_magic(magic, prefix_len, elem);
 262         else
 263                 /* shorthand */
 264                 return parse_short_magic(magic, elem);
 265 }
 266
 267 static void strip_submodule_slash_cheap(struct pathspec_item *item)
 268 {
 269         if (item->len >= 1 && item->match[item->len - 1] == '/') {
 270                 int i = cache_name_pos(item->match, item->len - 1);
 271
 272                 if (i >= 0 && S_ISGITLINK(active_cache[i]->ce_mode)) {
 273                         item->len--;
 274                         item->match[item->len] = '\0';
 275                 }
 276         }
 277 }
 278
 279 static void strip_submodule_slash_expensive(struct pathspec_item *item)
 280 {
 281         int i;
 282
 283         for (i = 0; i < active_nr; i++) {
 284                 struct cache_entry *ce = active_cache[i];
 285                 int ce_len = ce_namelen(ce);
 286
 287                 if (!S_ISGITLINK(ce->ce_mode))
 288                         continue;
 289
 290                 if (item->len <= ce_len || item->match[ce_len] != '/' ||
 291                     memcmp(ce->name, item->match, ce_len))
 292                         continue;
 293
 294                 if (item->len == ce_len + 1) {
 295                         /* strip trailing slash */
 296                         item->len--;
 297                         item->match[item->len] = '\0';
 298                 } else {
 299                         die(_("Pathspec '%s' is in submodule '%.*s'"),
 300                             item->original, ce_len, ce->name);
 301                 }
 302         }
 303 }
 304
 305 static void die_inside_submodule_path(struct pathspec_item *item)
 306 {
 307         int i;
 308
 309         for (i = 0; i < active_nr; i++) {
 310                 struct cache_entry *ce = active_cache[i];
 311                 int ce_len = ce_namelen(ce);
 312
 313                 if (!S_ISGITLINK(ce->ce_mode))
 314                         continue;
 315
 316                 if (item->len < ce_len ||
 317                     !(item->match[ce_len] == '/' || item->match[ce_len] == '\0') ||
 318                     memcmp(ce->name, item->match, ce_len))
 319                         continue;
 320
 321                 die(_("Pathspec '%s' is in submodule '%.*s'"),
 322                     item->original, ce_len, ce->name);
 323         }
 324 }
 325
 326 /*
 327  * Perform the initialization of a pathspec_item based on a pathspec element.
 328  */
 329 static void init_pathspec_item(struct pathspec_item *item, unsigned flags,
 330                                const char *prefix, int prefixlen,
 331                                const char *elt)
 332 {
 333         unsigned magic = 0, element_magic = 0;
 334         const char *copyfrom = elt;
 335         char *match;
 336         int pathspec_prefix = -1;
 337
 338         /* PATHSPEC_LITERAL_PATH ignores magic */
 339         if (flags & PATHSPEC_LITERAL_PATH) {
 340                 magic = PATHSPEC_LITERAL;
 341         } else {
 342                 copyfrom = parse_element_magic(&element_magic,
 343                                                &pathspec_prefix,
 344                                                elt);
 345                 magic |= element_magic;
 346                 magic |= get_global_magic(element_magic);
 347         }
 348
 349         item->magic = magic;
 350
 351         if (pathspec_prefix >= 0 &&
 352             (prefixlen || (prefix && *prefix)))
 353                 die("BUG: 'prefix' magic is supposed to be used at worktree's root");
 354
 355         if ((magic & PATHSPEC_LITERAL) && (magic & PATHSPEC_GLOB))
 356                 die(_("%s: 'literal' and 'glob' are incompatible"), elt);
 357
 358         /* Create match string which will be used for pathspec matching */
 359         if (pathspec_prefix >= 0) {
 360                 match = xstrdup(copyfrom);
 361                 prefixlen = pathspec_prefix;
 362         } else if (magic & PATHSPEC_FROMTOP) {
 363                 match = xstrdup(copyfrom);
 364                 prefixlen = 0;
 365         } else {
 366                 match = prefix_path_gently(prefix, prefixlen,
 367                                            &prefixlen, copyfrom);
 368                 if (!match)
 369                         die(_("%s: '%s' is outside repository"), elt, copyfrom);
 370         }
 371
 372         item->match = match;
 373         item->len = strlen(item->match);
 374         item->prefix = prefixlen;
 375
 376         /*
 377          * Prefix the pathspec (keep all magic) and assign to
 378          * original. Useful for passing to another command.
 379          */
 380         if ((flags & PATHSPEC_PREFIX_ORIGIN) &&
 381             prefixlen && !get_literal_global()) {
 382                 struct strbuf sb = STRBUF_INIT;
 383
 384                 /* Preserve the actual prefix length of each pattern */
 385                 prefix_magic(&sb, prefixlen, element_magic);
 386
 387                 strbuf_addstr(&sb, match);
 388                 item->original = strbuf_detach(&sb, NULL);
 389         } else {
 390                 item->original = xstrdup(elt);
 391         }
 392
 393         if (flags & PATHSPEC_STRIP_SUBMODULE_SLASH_CHEAP)
 394                 strip_submodule_slash_cheap(item);
 395
 396         if (flags & PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE)
 397                 strip_submodule_slash_expensive(item);
 398
 399         if (magic & PATHSPEC_LITERAL) {
 400                 item->nowildcard_len = item->len;
 401         } else {
 402                 item->nowildcard_len = simple_length(item->match);
 403                 if (item->nowildcard_len < prefixlen)
 404                         item->nowildcard_len = prefixlen;
 405         }
 406
 407         item->flags = 0;
 408         if (magic & PATHSPEC_GLOB) {
 409                 /*
 410                  * FIXME: should we enable ONESTAR in _GLOB for
 411                  * pattern "* * / * . c"?
 412                  */
 413         } else {
 414                 if (item->nowildcard_len < item->len &&
 415                     item->match[item->nowildcard_len] == '*' &&
 416                     no_wildcard(item->match + item->nowildcard_len + 1))
 417                         item->flags |= PATHSPEC_ONESTAR;
 418         }
 419
 420         /* sanity checks, pathspec matchers assume these are sane */
 421         if (item->nowildcard_len > item->len ||
 422             item->prefix         > item->len) {
 423                 /*
 424                  * This case can be triggered by the user pointing us to a
 425                  * pathspec inside a submodule, which is an input error.
 426                  * Detect that here and complain, but fallback in the
 427                  * non-submodule case to a BUG, as we have no idea what
 428                  * would trigger that.
 429                  */
 430                 die_inside_submodule_path(item);
 431                 die ("BUG: item->nowildcard_len > item->len || item->prefix > item->len)");
 432         }
 433 }
 434
 435 static int pathspec_item_cmp(const void *a_, const void *b_)
 436 {
 437         struct pathspec_item *a, *b;
 438
 439         a = (struct pathspec_item *)a_;
 440         b = (struct pathspec_item *)b_;
 441         return strcmp(a->match, b->match);
 442 }
 443
 444 static void NORETURN unsupported_magic(const char *pattern,
 445                                        unsigned magic)
 446 {
 447         struct strbuf sb = STRBUF_INIT;
 448         int i;
 449         for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 450                 const struct pathspec_magic *m = pathspec_magic + i;
 451                 if (!(magic & m->bit))
 452                         continue;
 453                 if (sb.len)
 454                         strbuf_addstr(&sb, ", ");
 455
 456                 if (m->mnemonic)
 457                         strbuf_addf(&sb, _("'%s' (mnemonic: '%c')"),
 458                                     m->name, m->mnemonic);
 459                 else
 460                         strbuf_addf(&sb, "'%s'", m->name);
 461         }
 462         /*
 463          * We may want to substitute "this command" with a command
 464          * name. E.g. when add--interactive dies when running
 465          * "checkout -p"
 466          */
 467         die(_("%s: pathspec magic not supported by this command: %s"),
 468             pattern, sb.buf);
 469 }
 470
 471 /*
 472  * Given command line arguments and a prefix, convert the input to
 473  * pathspec. die() if any magic in magic_mask is used.
 474  */
 475 void parse_pathspec(struct pathspec *pathspec,
 476                     unsigned magic_mask, unsigned flags,
 477                     const char *prefix, const char **argv)
 478 {
 479         struct pathspec_item *item;
 480         const char *entry = argv ? *argv : NULL;
 481         int i, n, prefixlen, warn_empty_string, nr_exclude = 0;
 482
 483         memset(pathspec, 0, sizeof(*pathspec));
 484
 485         if (flags & PATHSPEC_MAXDEPTH_VALID)
 486                 pathspec->magic |= PATHSPEC_MAXDEPTH;
 487
 488         /* No arguments, no prefix -> no pathspec */
 489         if (!entry && !prefix)
 490                 return;
 491
 492         if ((flags & PATHSPEC_PREFER_CWD) &&
 493             (flags & PATHSPEC_PREFER_FULL))
 494                 die("BUG: PATHSPEC_PREFER_CWD and PATHSPEC_PREFER_FULL are incompatible");
 495
 496         /* No arguments with prefix -> prefix pathspec */
 497         if (!entry) {
 498                 if (flags & PATHSPEC_PREFER_FULL)
 499                         return;
 500
 501                 if (!(flags & PATHSPEC_PREFER_CWD))
 502                         die("BUG: PATHSPEC_PREFER_CWD requires arguments");
 503
 504                 pathspec->items = item = xcalloc(1, sizeof(*item));
 505                 item->match = xstrdup(prefix);
 506                 item->original = xstrdup(prefix);
 507                 item->nowildcard_len = item->len = strlen(prefix);
 508                 item->prefix = item->len;
 509                 pathspec->nr = 1;
 510                 return;
 511         }
 512
 513         n = 0;
 514         warn_empty_string = 1;
 515         while (argv[n]) {
 516                 if (*argv[n] == '\0' && warn_empty_string) {
 517                         warning(_("empty strings as pathspecs will be made invalid in upcoming releases. "
 518                                   "please use . instead if you meant to match all paths"));
 519                         warn_empty_string = 0;
 520                 }
 521                 n++;
 522         }
 523
 524         pathspec->nr = n;
 525         ALLOC_ARRAY(pathspec->items, n + 1);
 526         item = pathspec->items;
 527         prefixlen = prefix ? strlen(prefix) : 0;
 528
 529         for (i = 0; i < n; i++) {
 530                 entry = argv[i];
 531
 532                 init_pathspec_item(item + i, flags, prefix, prefixlen, entry);
 533
 534                 if (item[i].magic & PATHSPEC_EXCLUDE)
 535                         nr_exclude++;
 536                 if (item[i].magic & magic_mask)
 537                         unsupported_magic(entry, item[i].magic & magic_mask);
 538
 539                 if ((flags & PATHSPEC_SYMLINK_LEADING_PATH) &&
 540                     has_symlink_leading_path(item[i].match, item[i].len)) {
 541                         die(_("pathspec '%s' is beyond a symbolic link"), entry);
 542                 }
 543
 544                 if (item[i].nowildcard_len < item[i].len)
 545                         pathspec->has_wildcard = 1;
 546                 pathspec->magic |= item[i].magic;
 547         }
 548
 549         /*
 550          * If everything is an exclude pattern, add one positive pattern
 551          * that matches everyting. We allocated an extra one for this.
 552          */
 553         if (nr_exclude == n) {
 554                 int plen = (!(flags & PATHSPEC_PREFER_CWD)) ? 0 : prefixlen;
 555                 init_pathspec_item(item + n, 0, prefix, plen, "");
 556                 pathspec->nr++;
 557         }
 558
 559         if (pathspec->magic & PATHSPEC_MAXDEPTH) {
 560                 if (flags & PATHSPEC_KEEP_ORDER)
 561                         die("BUG: PATHSPEC_MAXDEPTH_VALID and PATHSPEC_KEEP_ORDER are incompatible");
 562                 QSORT(pathspec->items, pathspec->nr, pathspec_item_cmp);
 563         }
 564 }
 565
 566 void copy_pathspec(struct pathspec *dst, const struct pathspec *src)
 567 {
 568         int i;
 569
 570         *dst = *src;
 571         ALLOC_ARRAY(dst->items, dst->nr);
 572         COPY_ARRAY(dst->items, src->items, dst->nr);
 573
 574         for (i = 0; i < dst->nr; i++) {
 575                 dst->items[i].match = xstrdup(src->items[i].match);
 576                 dst->items[i].original = xstrdup(src->items[i].original);
 577         }
 578 }
 579
 580 void clear_pathspec(struct pathspec *pathspec)
 581 {
 582         int i;
 583
 584         for (i = 0; i < pathspec->nr; i++) {
 585                 free(pathspec->items[i].match);
 586                 free(pathspec->items[i].original);
 587         }
 588         free(pathspec->items);
 589         pathspec->items = NULL;
 590         pathspec->nr = 0;
 591 }