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                 match_pathspec_depth(pathspec, ce->name, ce_namelen(ce), 0, 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 /*
  78  * Take an element of a pathspec and check for magic signatures.
  79  * Append the result to the prefix. Return the magic bitmap.
  80  *
  81  * For now, we only parse the syntax and throw out anything other than
  82  * "top" magic.
  83  *
  84  * NEEDSWORK: This needs to be rewritten when we start migrating
  85  * get_pathspec() users to use the "struct pathspec" interface.  For
  86  * example, a pathspec element may be marked as case-insensitive, but
  87  * the prefix part must always match literally, and a single stupid
  88  * string cannot express such a case.
  89  */
  90 static unsigned prefix_pathspec(struct pathspec_item *item,
  91                                 unsigned *p_short_magic,
  92                                 const char **raw, unsigned flags,
  93                                 const char *prefix, int prefixlen,
  94                                 const char *elt)
  95 {
  96         static int literal_global = -1;
  97         static int glob_global = -1;
  98         static int noglob_global = -1;
  99         static int icase_global = -1;
 100         unsigned magic = 0, short_magic = 0, global_magic = 0;
 101         const char *copyfrom = elt, *long_magic_end = NULL;
 102         char *match;
 103         int i, pathspec_prefix = -1;
 104
 105         if (literal_global < 0)
 106                 literal_global = git_env_bool(GIT_LITERAL_PATHSPECS_ENVIRONMENT, 0);
 107         if (literal_global)
 108                 global_magic |= PATHSPEC_LITERAL;
 109
 110         if (glob_global < 0)
 111                 glob_global = git_env_bool(GIT_GLOB_PATHSPECS_ENVIRONMENT, 0);
 112         if (glob_global)
 113                 global_magic |= PATHSPEC_GLOB;
 114
 115         if (noglob_global < 0)
 116                 noglob_global = git_env_bool(GIT_NOGLOB_PATHSPECS_ENVIRONMENT, 0);
 117
 118         if (glob_global && noglob_global)
 119                 die(_("global 'glob' and 'noglob' pathspec settings are incompatible"));
 120
 121
 122         if (icase_global < 0)
 123                 icase_global = git_env_bool(GIT_ICASE_PATHSPECS_ENVIRONMENT, 0);
 124         if (icase_global)
 125                 global_magic |= PATHSPEC_ICASE;
 126
 127         if ((global_magic & PATHSPEC_LITERAL) &&
 128             (global_magic & ~PATHSPEC_LITERAL))
 129                 die(_("global 'literal' pathspec setting is incompatible "
 130                       "with all other global pathspec settings"));
 131
 132         if (flags & PATHSPEC_LITERAL_PATH)
 133                 global_magic = 0;
 134
 135         if (elt[0] != ':' || literal_global ||
 136             (flags & PATHSPEC_LITERAL_PATH)) {
 137                 ; /* nothing to do */
 138         } else if (elt[1] == '(') {
 139                 /* longhand */
 140                 const char *nextat;
 141                 for (copyfrom = elt + 2;
 142                      *copyfrom && *copyfrom != ')';
 143                      copyfrom = nextat) {
 144                         size_t len = strcspn(copyfrom, ",)");
 145                         if (copyfrom[len] == ',')
 146                                 nextat = copyfrom + len + 1;
 147                         else
 148                                 /* handle ')' and '\0' */
 149                                 nextat = copyfrom + len;
 150                         if (!len)
 151                                 continue;
 152                         for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 153                                 if (strlen(pathspec_magic[i].name) == len &&
 154                                     !strncmp(pathspec_magic[i].name, copyfrom, len)) {
 155                                         magic |= pathspec_magic[i].bit;
 156                                         break;
 157                                 }
 158                                 if (!prefixcmp(copyfrom, "prefix:")) {
 159                                         char *endptr;
 160                                         pathspec_prefix = strtol(copyfrom + 7,
 161                                                                  &endptr, 10);
 162                                         if (endptr - copyfrom != len)
 163                                                 die(_("invalid parameter for pathspec magic 'prefix'"));
 164                                         /* "i" would be wrong, but it does not matter */
 165                                         break;
 166                                 }
 167                         }
 168                         if (ARRAY_SIZE(pathspec_magic) <= i)
 169                                 die(_("Invalid pathspec magic '%.*s' in '%s'"),
 170                                     (int) len, copyfrom, elt);
 171                 }
 172                 if (*copyfrom != ')')
 173                         die(_("Missing ')' at the end of pathspec magic in '%s'"), elt);
 174                 long_magic_end = copyfrom;
 175                 copyfrom++;
 176         } else {
 177                 /* shorthand */
 178                 for (copyfrom = elt + 1;
 179                      *copyfrom && *copyfrom != ':';
 180                      copyfrom++) {
 181                         char ch = *copyfrom;
 182
 183                         if (!is_pathspec_magic(ch))
 184                                 break;
 185                         for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++)
 186                                 if (pathspec_magic[i].mnemonic == ch) {
 187                                         short_magic |= pathspec_magic[i].bit;
 188                                         break;
 189                                 }
 190                         if (ARRAY_SIZE(pathspec_magic) <= i)
 191                                 die(_("Unimplemented pathspec magic '%c' in '%s'"),
 192                                     ch, elt);
 193                 }
 194                 if (*copyfrom == ':')
 195                         copyfrom++;
 196         }
 197
 198         magic |= short_magic;
 199         *p_short_magic = short_magic;
 200
 201         /* --noglob-pathspec adds :(literal) _unless_ :(glob) is specified */
 202         if (noglob_global && !(magic & PATHSPEC_GLOB))
 203                 global_magic |= PATHSPEC_LITERAL;
 204
 205         /* --glob-pathspec is overridden by :(literal) */
 206         if ((global_magic & PATHSPEC_GLOB) && (magic & PATHSPEC_LITERAL))
 207                 global_magic &= ~PATHSPEC_GLOB;
 208
 209         magic |= global_magic;
 210
 211         if (pathspec_prefix >= 0 &&
 212             (prefixlen || (prefix && *prefix)))
 213                 die("BUG: 'prefix' magic is supposed to be used at worktree's root");
 214
 215         if ((magic & PATHSPEC_LITERAL) && (magic & PATHSPEC_GLOB))
 216                 die(_("%s: 'literal' and 'glob' are incompatible"), elt);
 217
 218         if (pathspec_prefix >= 0) {
 219                 match = xstrdup(copyfrom);
 220                 prefixlen = pathspec_prefix;
 221         } else if (magic & PATHSPEC_FROMTOP) {
 222                 match = xstrdup(copyfrom);
 223                 prefixlen = 0;
 224         } else {
 225                 match = prefix_path_gently(prefix, prefixlen, &prefixlen, copyfrom);
 226                 if (!match)
 227                         die(_("%s: '%s' is outside repository"), elt, copyfrom);
 228         }
 229         *raw = item->match = match;
 230         /*
 231          * Prefix the pathspec (keep all magic) and assign to
 232          * original. Useful for passing to another command.
 233          */
 234         if (flags & PATHSPEC_PREFIX_ORIGIN) {
 235                 struct strbuf sb = STRBUF_INIT;
 236                 const char *start = elt;
 237                 if (prefixlen && !literal_global) {
 238                         /* Preserve the actual prefix length of each pattern */
 239                         if (short_magic)
 240                                 die("BUG: prefixing on short magic is not supported");
 241                         else if (long_magic_end) {
 242                                 strbuf_add(&sb, start, long_magic_end - start);
 243                                 strbuf_addf(&sb, ",prefix:%d", prefixlen);
 244                                 start = long_magic_end;
 245                         } else {
 246                                 if (*start == ':')
 247                                         start++;
 248                                 strbuf_addf(&sb, ":(prefix:%d)", prefixlen);
 249                         }
 250                 }
 251                 strbuf_add(&sb, start, copyfrom - start);
 252                 strbuf_addstr(&sb, match);
 253                 item->original = strbuf_detach(&sb, NULL);
 254         } else
 255                 item->original = elt;
 256         item->len = strlen(item->match);
 257         item->prefix = prefixlen;
 258
 259         if ((flags & PATHSPEC_STRIP_SUBMODULE_SLASH_CHEAP) &&
 260             (item->len >= 1 && item->match[item->len - 1] == '/') &&
 261             (i = cache_name_pos(item->match, item->len - 1)) >= 0 &&
 262             S_ISGITLINK(active_cache[i]->ce_mode)) {
 263                 item->len--;
 264                 match[item->len] = '\0';
 265         }
 266
 267         if (flags & PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE)
 268                 for (i = 0; i < active_nr; i++) {
 269                         struct cache_entry *ce = active_cache[i];
 270                         int ce_len = ce_namelen(ce);
 271
 272                         if (!S_ISGITLINK(ce->ce_mode))
 273                                 continue;
 274
 275                         if (item->len <= ce_len || match[ce_len] != '/' ||
 276                             memcmp(ce->name, match, ce_len))
 277                                 continue;
 278                         if (item->len == ce_len + 1) {
 279                                 /* strip trailing slash */
 280                                 item->len--;
 281                                 match[item->len] = '\0';
 282                         } else
 283                                 die (_("Pathspec '%s' is in submodule '%.*s'"),
 284                                      elt, ce_len, ce->name);
 285                 }
 286
 287         if (magic & PATHSPEC_LITERAL)
 288                 item->nowildcard_len = item->len;
 289         else {
 290                 item->nowildcard_len = simple_length(item->match);
 291                 if (item->nowildcard_len < prefixlen)
 292                         item->nowildcard_len = prefixlen;
 293         }
 294         item->flags = 0;
 295         if (magic & PATHSPEC_GLOB) {
 296                 /*
 297                  * FIXME: should we enable ONESTAR in _GLOB for
 298                  * pattern "* * / * . c"?
 299                  */
 300         } else {
 301                 if (item->nowildcard_len < item->len &&
 302                     item->match[item->nowildcard_len] == '*' &&
 303                     no_wildcard(item->match + item->nowildcard_len + 1))
 304                         item->flags |= PATHSPEC_ONESTAR;
 305         }
 306
 307         /* sanity checks, pathspec matchers assume these are sane */
 308         assert(item->nowildcard_len <= item->len &&
 309                item->prefix         <= item->len);
 310         return magic;
 311 }
 312
 313 static int pathspec_item_cmp(const void *a_, const void *b_)
 314 {
 315         struct pathspec_item *a, *b;
 316
 317         a = (struct pathspec_item *)a_;
 318         b = (struct pathspec_item *)b_;
 319         return strcmp(a->match, b->match);
 320 }
 321
 322 static void NORETURN unsupported_magic(const char *pattern,
 323                                        unsigned magic,
 324                                        unsigned short_magic)
 325 {
 326         struct strbuf sb = STRBUF_INIT;
 327         int i, n;
 328         for (n = i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 329                 const struct pathspec_magic *m = pathspec_magic + i;
 330                 if (!(magic & m->bit))
 331                         continue;
 332                 if (sb.len)
 333                         strbuf_addstr(&sb, " ");
 334                 if (short_magic & m->bit)
 335                         strbuf_addf(&sb, "'%c'", m->mnemonic);
 336                 else
 337                         strbuf_addf(&sb, "'%s'", m->name);
 338                 n++;
 339         }
 340         /*
 341          * We may want to substitute "this command" with a command
 342          * name. E.g. when add--interactive dies when running
 343          * "checkout -p"
 344          */
 345         die(_("%s: pathspec magic not supported by this command: %s"),
 346             pattern, sb.buf);
 347 }
 348
 349 /*
 350  * Given command line arguments and a prefix, convert the input to
 351  * pathspec. die() if any magic in magic_mask is used.
 352  */
 353 void parse_pathspec(struct pathspec *pathspec,
 354                     unsigned magic_mask, unsigned flags,
 355                     const char *prefix, const char **argv)
 356 {
 357         struct pathspec_item *item;
 358         const char *entry = argv ? *argv : NULL;
 359         int i, n, prefixlen, nr_exclude = 0;
 360
 361         memset(pathspec, 0, sizeof(*pathspec));
 362
 363         if (flags & PATHSPEC_MAXDEPTH_VALID)
 364                 pathspec->magic |= PATHSPEC_MAXDEPTH;
 365
 366         /* No arguments, no prefix -> no pathspec */
 367         if (!entry && !prefix)
 368                 return;
 369
 370         if ((flags & PATHSPEC_PREFER_CWD) &&
 371             (flags & PATHSPEC_PREFER_FULL))
 372                 die("BUG: PATHSPEC_PREFER_CWD and PATHSPEC_PREFER_FULL are incompatible");
 373
 374         /* No arguments with prefix -> prefix pathspec */
 375         if (!entry) {
 376                 static const char *raw[2];
 377
 378                 if (flags & PATHSPEC_PREFER_FULL)
 379                         return;
 380
 381                 if (!(flags & PATHSPEC_PREFER_CWD))
 382                         die("BUG: PATHSPEC_PREFER_CWD requires arguments");
 383
 384                 pathspec->items = item = xmalloc(sizeof(*item));
 385                 memset(item, 0, sizeof(*item));
 386                 item->match = prefix;
 387                 item->original = prefix;
 388                 item->nowildcard_len = item->len = strlen(prefix);
 389                 item->prefix = item->len;
 390                 raw[0] = prefix;
 391                 raw[1] = NULL;
 392                 pathspec->nr = 1;
 393                 pathspec->_raw = raw;
 394                 return;
 395         }
 396
 397         n = 0;
 398         while (argv[n])
 399                 n++;
 400
 401         pathspec->nr = n;
 402         pathspec->items = item = xmalloc(sizeof(*item) * n);
 403         pathspec->_raw = argv;
 404         prefixlen = prefix ? strlen(prefix) : 0;
 405
 406         for (i = 0; i < n; i++) {
 407                 unsigned short_magic;
 408                 entry = argv[i];
 409
 410                 item[i].magic = prefix_pathspec(item + i, &short_magic,
 411                                                 argv + i, flags,
 412                                                 prefix, prefixlen, entry);
 413                 if ((flags & PATHSPEC_LITERAL_PATH) &&
 414                     !(magic_mask & PATHSPEC_LITERAL))
 415                         item[i].magic |= PATHSPEC_LITERAL;
 416                 if (item[i].magic & PATHSPEC_EXCLUDE)
 417                         nr_exclude++;
 418                 if (item[i].magic & magic_mask)
 419                         unsupported_magic(entry,
 420                                           item[i].magic & magic_mask,
 421                                           short_magic);
 422
 423                 if ((flags & PATHSPEC_SYMLINK_LEADING_PATH) &&
 424                     has_symlink_leading_path(item[i].match, item[i].len)) {
 425                         die(_("pathspec '%s' is beyond a symbolic link"), entry);
 426                 }
 427
 428                 if (item[i].nowildcard_len < item[i].len)
 429                         pathspec->has_wildcard = 1;
 430                 pathspec->magic |= item[i].magic;
 431         }
 432
 433         if (nr_exclude == n)
 434                 die(_("There is nothing to exclude from by :(exclude) patterns.\n"
 435                       "Perhaps you forgot to add either ':/' or '.' ?"));
 436
 437
 438         if (pathspec->magic & PATHSPEC_MAXDEPTH) {
 439                 if (flags & PATHSPEC_KEEP_ORDER)
 440                         die("BUG: PATHSPEC_MAXDEPTH_VALID and PATHSPEC_KEEP_ORDER are incompatible");
 441                 qsort(pathspec->items, pathspec->nr,
 442                       sizeof(struct pathspec_item), pathspec_item_cmp);
 443         }
 444 }
 445
 446 /*
 447  * N.B. get_pathspec() is deprecated in favor of the "struct pathspec"
 448  * based interface - see pathspec.c:parse_pathspec().
 449  *
 450  * Arguments:
 451  *  - prefix - a path relative to the root of the working tree
 452  *  - pathspec - a list of paths underneath the prefix path
 453  *
 454  * Iterates over pathspec, prepending each path with prefix,
 455  * and return the resulting list.
 456  *
 457  * If pathspec is empty, return a singleton list containing prefix.
 458  *
 459  * If pathspec and prefix are both empty, return an empty list.
 460  *
 461  * This is typically used by built-in commands such as add.c, in order
 462  * to normalize argv arguments provided to the built-in into a list of
 463  * paths to process, all relative to the root of the working tree.
 464  */
 465 const char **get_pathspec(const char *prefix, const char **pathspec)
 466 {
 467         struct pathspec ps;
 468         parse_pathspec(&ps,
 469                        PATHSPEC_ALL_MAGIC &
 470                        ~(PATHSPEC_FROMTOP | PATHSPEC_LITERAL),
 471                        PATHSPEC_PREFER_CWD,
 472                        prefix, pathspec);
 473         return ps._raw;
 474 }
 475
 476 void copy_pathspec(struct pathspec *dst, const struct pathspec *src)
 477 {
 478         *dst = *src;
 479         dst->items = xmalloc(sizeof(struct pathspec_item) * dst->nr);
 480         memcpy(dst->items, src->items,
 481                sizeof(struct pathspec_item) * dst->nr);
 482 }
 483
 484 void free_pathspec(struct pathspec *pathspec)
 485 {
 486         free(pathspec->items);
 487         pathspec->items = NULL;
 488 }