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 };
  75
  76 /*
  77  * Take an element of a pathspec and check for magic signatures.
  78  * Append the result to the prefix. Return the magic bitmap.
  79  *
  80  * For now, we only parse the syntax and throw out anything other than
  81  * "top" magic.
  82  *
  83  * NEEDSWORK: This needs to be rewritten when we start migrating
  84  * get_pathspec() users to use the "struct pathspec" interface.  For
  85  * example, a pathspec element may be marked as case-insensitive, but
  86  * the prefix part must always match literally, and a single stupid
  87  * string cannot express such a case.
  88  */
  89 static unsigned prefix_pathspec(struct pathspec_item *item,
  90                                 unsigned *p_short_magic,
  91                                 const char **raw, unsigned flags,
  92                                 const char *prefix, int prefixlen,
  93                                 const char *elt)
  94 {
  95         static int literal_global = -1;
  96         static int glob_global = -1;
  97         static int noglob_global = -1;
  98         static int icase_global = -1;
  99         unsigned magic = 0, short_magic = 0, global_magic = 0;
 100         const char *copyfrom = elt, *long_magic_end = NULL;
 101         char *match;
 102         int i, pathspec_prefix = -1;
 103
 104         if (literal_global < 0)
 105                 literal_global = git_env_bool(GIT_LITERAL_PATHSPECS_ENVIRONMENT, 0);
 106         if (literal_global)
 107                 global_magic |= PATHSPEC_LITERAL;
 108
 109         if (glob_global < 0)
 110                 glob_global = git_env_bool(GIT_GLOB_PATHSPECS_ENVIRONMENT, 0);
 111         if (glob_global)
 112                 global_magic |= PATHSPEC_GLOB;
 113
 114         if (noglob_global < 0)
 115                 noglob_global = git_env_bool(GIT_NOGLOB_PATHSPECS_ENVIRONMENT, 0);
 116
 117         if (glob_global && noglob_global)
 118                 die(_("global 'glob' and 'noglob' pathspec settings are incompatible"));
 119
 120
 121         if (icase_global < 0)
 122                 icase_global = git_env_bool(GIT_ICASE_PATHSPECS_ENVIRONMENT, 0);
 123         if (icase_global)
 124                 global_magic |= PATHSPEC_ICASE;
 125
 126         if ((global_magic & PATHSPEC_LITERAL) &&
 127             (global_magic & ~PATHSPEC_LITERAL))
 128                 die(_("global 'literal' pathspec setting is incompatible "
 129                       "with all other global pathspec settings"));
 130
 131         if (elt[0] != ':' || literal_global) {
 132                 ; /* nothing to do */
 133         } else if (elt[1] == '(') {
 134                 /* longhand */
 135                 const char *nextat;
 136                 for (copyfrom = elt + 2;
 137                      *copyfrom && *copyfrom != ')';
 138                      copyfrom = nextat) {
 139                         size_t len = strcspn(copyfrom, ",)");
 140                         if (copyfrom[len] == ',')
 141                                 nextat = copyfrom + len + 1;
 142                         else
 143                                 /* handle ')' and '\0' */
 144                                 nextat = copyfrom + len;
 145                         if (!len)
 146                                 continue;
 147                         for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 148                                 if (strlen(pathspec_magic[i].name) == len &&
 149                                     !strncmp(pathspec_magic[i].name, copyfrom, len)) {
 150                                         magic |= pathspec_magic[i].bit;
 151                                         break;
 152                                 }
 153                                 if (!prefixcmp(copyfrom, "prefix:")) {
 154                                         char *endptr;
 155                                         pathspec_prefix = strtol(copyfrom + 7,
 156                                                                  &endptr, 10);
 157                                         if (endptr - copyfrom != len)
 158                                                 die(_("invalid parameter for pathspec magic 'prefix'"));
 159                                         /* "i" would be wrong, but it does not matter */
 160                                         break;
 161                                 }
 162                         }
 163                         if (ARRAY_SIZE(pathspec_magic) <= i)
 164                                 die(_("Invalid pathspec magic '%.*s' in '%s'"),
 165                                     (int) len, copyfrom, elt);
 166                 }
 167                 if (*copyfrom != ')')
 168                         die(_("Missing ')' at the end of pathspec magic in '%s'"), elt);
 169                 long_magic_end = copyfrom;
 170                 copyfrom++;
 171         } else {
 172                 /* shorthand */
 173                 for (copyfrom = elt + 1;
 174                      *copyfrom && *copyfrom != ':';
 175                      copyfrom++) {
 176                         char ch = *copyfrom;
 177
 178                         if (!is_pathspec_magic(ch))
 179                                 break;
 180                         for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++)
 181                                 if (pathspec_magic[i].mnemonic == ch) {
 182                                         short_magic |= pathspec_magic[i].bit;
 183                                         break;
 184                                 }
 185                         if (ARRAY_SIZE(pathspec_magic) <= i)
 186                                 die(_("Unimplemented pathspec magic '%c' in '%s'"),
 187                                     ch, elt);
 188                 }
 189                 if (*copyfrom == ':')
 190                         copyfrom++;
 191         }
 192
 193         magic |= short_magic;
 194         *p_short_magic = short_magic;
 195
 196         /* --noglob-pathspec adds :(literal) _unless_ :(glob) is specifed */
 197         if (noglob_global && !(magic & PATHSPEC_GLOB))
 198                 global_magic |= PATHSPEC_LITERAL;
 199
 200         /* --glob-pathspec is overriden by :(literal) */
 201         if ((global_magic & PATHSPEC_GLOB) && (magic & PATHSPEC_LITERAL))
 202                 global_magic &= ~PATHSPEC_GLOB;
 203
 204         magic |= global_magic;
 205
 206         if (pathspec_prefix >= 0 &&
 207             (prefixlen || (prefix && *prefix)))
 208                 die("BUG: 'prefix' magic is supposed to be used at worktree's root");
 209
 210         if ((magic & PATHSPEC_LITERAL) && (magic & PATHSPEC_GLOB))
 211                 die(_("%s: 'literal' and 'glob' are incompatible"), elt);
 212
 213         if (pathspec_prefix >= 0) {
 214                 match = xstrdup(copyfrom);
 215                 prefixlen = pathspec_prefix;
 216         } else if (magic & PATHSPEC_FROMTOP) {
 217                 match = xstrdup(copyfrom);
 218                 prefixlen = 0;
 219         } else {
 220                 match = prefix_path_gently(prefix, prefixlen, &prefixlen, copyfrom);
 221                 if (!match)
 222                         die(_("%s: '%s' is outside repository"), elt, copyfrom);
 223         }
 224         *raw = item->match = match;
 225         /*
 226          * Prefix the pathspec (keep all magic) and assign to
 227          * original. Useful for passing to another command.
 228          */
 229         if (flags & PATHSPEC_PREFIX_ORIGIN) {
 230                 struct strbuf sb = STRBUF_INIT;
 231                 const char *start = elt;
 232                 if (prefixlen && !literal_global) {
 233                         /* Preserve the actual prefix length of each pattern */
 234                         if (long_magic_end) {
 235                                 strbuf_add(&sb, start, long_magic_end - start);
 236                                 strbuf_addf(&sb, ",prefix:%d", prefixlen);
 237                                 start = long_magic_end;
 238                         } else {
 239                                 if (*start == ':')
 240                                         start++;
 241                                 strbuf_addf(&sb, ":(prefix:%d)", prefixlen);
 242                         }
 243                 }
 244                 strbuf_add(&sb, start, copyfrom - start);
 245                 strbuf_addstr(&sb, match);
 246                 item->original = strbuf_detach(&sb, NULL);
 247         } else
 248                 item->original = elt;
 249         item->len = strlen(item->match);
 250         item->prefix = prefixlen;
 251
 252         if ((flags & PATHSPEC_STRIP_SUBMODULE_SLASH_CHEAP) &&
 253             (item->len >= 1 && item->match[item->len - 1] == '/') &&
 254             (i = cache_name_pos(item->match, item->len - 1)) >= 0 &&
 255             S_ISGITLINK(active_cache[i]->ce_mode)) {
 256                 item->len--;
 257                 match[item->len] = '\0';
 258         }
 259
 260         if (flags & PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE)
 261                 for (i = 0; i < active_nr; i++) {
 262                         struct cache_entry *ce = active_cache[i];
 263                         int ce_len = ce_namelen(ce);
 264
 265                         if (!S_ISGITLINK(ce->ce_mode))
 266                                 continue;
 267
 268                         if (item->len <= ce_len || match[ce_len] != '/' ||
 269                             memcmp(ce->name, match, ce_len))
 270                                 continue;
 271                         if (item->len == ce_len + 1) {
 272                                 /* strip trailing slash */
 273                                 item->len--;
 274                                 match[item->len] = '\0';
 275                         } else
 276                                 die (_("Pathspec '%s' is in submodule '%.*s'"),
 277                                      elt, ce_len, ce->name);
 278                 }
 279
 280         if (magic & PATHSPEC_LITERAL)
 281                 item->nowildcard_len = item->len;
 282         else {
 283                 item->nowildcard_len = simple_length(item->match);
 284                 if (item->nowildcard_len < prefixlen)
 285                         item->nowildcard_len = prefixlen;
 286         }
 287         item->flags = 0;
 288         if (magic & PATHSPEC_GLOB) {
 289                 /*
 290                  * FIXME: should we enable ONESTAR in _GLOB for
 291                  * pattern "* * / * . c"?
 292                  */
 293         } else {
 294                 if (item->nowildcard_len < item->len &&
 295                     item->match[item->nowildcard_len] == '*' &&
 296                     no_wildcard(item->match + item->nowildcard_len + 1))
 297                         item->flags |= PATHSPEC_ONESTAR;
 298         }
 299
 300         /* sanity checks, pathspec matchers assume these are sane */
 301         assert(item->nowildcard_len <= item->len &&
 302                item->prefix         <= item->len);
 303         return magic;
 304 }
 305
 306 static int pathspec_item_cmp(const void *a_, const void *b_)
 307 {
 308         struct pathspec_item *a, *b;
 309
 310         a = (struct pathspec_item *)a_;
 311         b = (struct pathspec_item *)b_;
 312         return strcmp(a->match, b->match);
 313 }
 314
 315 static void NORETURN unsupported_magic(const char *pattern,
 316                                        unsigned magic,
 317                                        unsigned short_magic)
 318 {
 319         struct strbuf sb = STRBUF_INIT;
 320         int i, n;
 321         for (n = i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 322                 const struct pathspec_magic *m = pathspec_magic + i;
 323                 if (!(magic & m->bit))
 324                         continue;
 325                 if (sb.len)
 326                         strbuf_addstr(&sb, " ");
 327                 if (short_magic & m->bit)
 328                         strbuf_addf(&sb, "'%c'", m->mnemonic);
 329                 else
 330                         strbuf_addf(&sb, "'%s'", m->name);
 331                 n++;
 332         }
 333         /*
 334          * We may want to substitute "this command" with a command
 335          * name. E.g. when add--interactive dies when running
 336          * "checkout -p"
 337          */
 338         die(_("%s: pathspec magic not supported by this command: %s"),
 339             pattern, sb.buf);
 340 }
 341
 342 /*
 343  * Given command line arguments and a prefix, convert the input to
 344  * pathspec. die() if any magic in magic_mask is used.
 345  */
 346 void parse_pathspec(struct pathspec *pathspec,
 347                     unsigned magic_mask, unsigned flags,
 348                     const char *prefix, const char **argv)
 349 {
 350         struct pathspec_item *item;
 351         const char *entry = argv ? *argv : NULL;
 352         int i, n, prefixlen;
 353
 354         memset(pathspec, 0, sizeof(*pathspec));
 355
 356         if (flags & PATHSPEC_MAXDEPTH_VALID)
 357                 pathspec->magic |= PATHSPEC_MAXDEPTH;
 358
 359         /* No arguments, no prefix -> no pathspec */
 360         if (!entry && !prefix)
 361                 return;
 362
 363         if ((flags & PATHSPEC_PREFER_CWD) &&
 364             (flags & PATHSPEC_PREFER_FULL))
 365                 die("BUG: PATHSPEC_PREFER_CWD and PATHSPEC_PREFER_FULL are incompatible");
 366
 367         /* No arguments with prefix -> prefix pathspec */
 368         if (!entry) {
 369                 static const char *raw[2];
 370
 371                 if (flags & PATHSPEC_PREFER_FULL)
 372                         return;
 373
 374                 if (!(flags & PATHSPEC_PREFER_CWD))
 375                         die("BUG: PATHSPEC_PREFER_CWD requires arguments");
 376
 377                 pathspec->items = item = xmalloc(sizeof(*item));
 378                 memset(item, 0, sizeof(*item));
 379                 item->match = prefix;
 380                 item->original = prefix;
 381                 item->nowildcard_len = item->len = strlen(prefix);
 382                 item->prefix = item->len;
 383                 raw[0] = prefix;
 384                 raw[1] = NULL;
 385                 pathspec->nr = 1;
 386                 pathspec->_raw = raw;
 387                 return;
 388         }
 389
 390         n = 0;
 391         while (argv[n])
 392                 n++;
 393
 394         pathspec->nr = n;
 395         pathspec->items = item = xmalloc(sizeof(*item) * n);
 396         pathspec->_raw = argv;
 397         prefixlen = prefix ? strlen(prefix) : 0;
 398
 399         for (i = 0; i < n; i++) {
 400                 unsigned short_magic;
 401                 entry = argv[i];
 402
 403                 item[i].magic = prefix_pathspec(item + i, &short_magic,
 404                                                 argv + i, flags,
 405                                                 prefix, prefixlen, entry);
 406                 if (item[i].magic & magic_mask)
 407                         unsupported_magic(entry,
 408                                           item[i].magic & magic_mask,
 409                                           short_magic);
 410
 411                 if ((flags & PATHSPEC_SYMLINK_LEADING_PATH) &&
 412                     has_symlink_leading_path(item[i].match, item[i].len)) {
 413                         die(_("pathspec '%s' is beyond a symbolic link"), entry);
 414                 }
 415
 416                 if (item[i].nowildcard_len < item[i].len)
 417                         pathspec->has_wildcard = 1;
 418                 pathspec->magic |= item[i].magic;
 419         }
 420
 421
 422         if (pathspec->magic & PATHSPEC_MAXDEPTH) {
 423                 if (flags & PATHSPEC_KEEP_ORDER)
 424                         die("BUG: PATHSPEC_MAXDEPTH_VALID and PATHSPEC_KEEP_ORDER are incompatible");
 425                 qsort(pathspec->items, pathspec->nr,
 426                       sizeof(struct pathspec_item), pathspec_item_cmp);
 427         }
 428 }
 429
 430 /*
 431  * N.B. get_pathspec() is deprecated in favor of the "struct pathspec"
 432  * based interface - see pathspec.c:parse_pathspec().
 433  *
 434  * Arguments:
 435  *  - prefix - a path relative to the root of the working tree
 436  *  - pathspec - a list of paths underneath the prefix path
 437  *
 438  * Iterates over pathspec, prepending each path with prefix,
 439  * and return the resulting list.
 440  *
 441  * If pathspec is empty, return a singleton list containing prefix.
 442  *
 443  * If pathspec and prefix are both empty, return an empty list.
 444  *
 445  * This is typically used by built-in commands such as add.c, in order
 446  * to normalize argv arguments provided to the built-in into a list of
 447  * paths to process, all relative to the root of the working tree.
 448  */
 449 const char **get_pathspec(const char *prefix, const char **pathspec)
 450 {
 451         struct pathspec ps;
 452         parse_pathspec(&ps,
 453                        PATHSPEC_ALL_MAGIC &
 454                        ~(PATHSPEC_FROMTOP | PATHSPEC_LITERAL),
 455                        PATHSPEC_PREFER_CWD,
 456                        prefix, pathspec);
 457         return ps._raw;
 458 }
 459
 460 void copy_pathspec(struct pathspec *dst, const struct pathspec *src)
 461 {
 462         *dst = *src;
 463         dst->items = xmalloc(sizeof(struct pathspec_item) * dst->nr);
 464         memcpy(dst->items, src->items,
 465                sizeof(struct pathspec_item) * dst->nr);
 466 }
 467
 468 void free_pathspec(struct pathspec *pathspec)
 469 {
 470         free(pathspec->items);
 471         pathspec->items = NULL;
 472 }