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 static void prefix_short_magic(struct strbuf *sb, int prefixlen,
  78                                unsigned short_magic)
  79 {
  80         int i;
  81         strbuf_addstr(sb, ":(");
  82         for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++)
  83                 if (short_magic & pathspec_magic[i].bit) {
  84                         if (sb->buf[sb->len - 1] != '(')
  85                                 strbuf_addch(sb, ',');
  86                         strbuf_addstr(sb, pathspec_magic[i].name);
  87                 }
  88         strbuf_addf(sb, ",prefix:%d)", prefixlen);
  89 }
  90
  91 /*
  92  * Take an element of a pathspec and check for magic signatures.
  93  * Append the result to the prefix. Return the magic bitmap.
  94  *
  95  * For now, we only parse the syntax and throw out anything other than
  96  * "top" magic.
  97  *
  98  * NEEDSWORK: This needs to be rewritten when we start migrating
  99  * get_pathspec() users to use the "struct pathspec" interface.  For
 100  * example, a pathspec element may be marked as case-insensitive, but
 101  * the prefix part must always match literally, and a single stupid
 102  * string cannot express such a case.
 103  */
 104 static unsigned prefix_pathspec(struct pathspec_item *item,
 105                                 unsigned *p_short_magic,
 106                                 const char **raw, unsigned flags,
 107                                 const char *prefix, int prefixlen,
 108                                 const char *elt)
 109 {
 110         static int literal_global = -1;
 111         static int glob_global = -1;
 112         static int noglob_global = -1;
 113         static int icase_global = -1;
 114         unsigned magic = 0, short_magic = 0, global_magic = 0;
 115         const char *copyfrom = elt, *long_magic_end = NULL;
 116         char *match;
 117         int i, pathspec_prefix = -1;
 118
 119         if (literal_global < 0)
 120                 literal_global = git_env_bool(GIT_LITERAL_PATHSPECS_ENVIRONMENT, 0);
 121         if (literal_global)
 122                 global_magic |= PATHSPEC_LITERAL;
 123
 124         if (glob_global < 0)
 125                 glob_global = git_env_bool(GIT_GLOB_PATHSPECS_ENVIRONMENT, 0);
 126         if (glob_global)
 127                 global_magic |= PATHSPEC_GLOB;
 128
 129         if (noglob_global < 0)
 130                 noglob_global = git_env_bool(GIT_NOGLOB_PATHSPECS_ENVIRONMENT, 0);
 131
 132         if (glob_global && noglob_global)
 133                 die(_("global 'glob' and 'noglob' pathspec settings are incompatible"));
 134
 135
 136         if (icase_global < 0)
 137                 icase_global = git_env_bool(GIT_ICASE_PATHSPECS_ENVIRONMENT, 0);
 138         if (icase_global)
 139                 global_magic |= PATHSPEC_ICASE;
 140
 141         if ((global_magic & PATHSPEC_LITERAL) &&
 142             (global_magic & ~PATHSPEC_LITERAL))
 143                 die(_("global 'literal' pathspec setting is incompatible "
 144                       "with all other global pathspec settings"));
 145
 146         if (flags & PATHSPEC_LITERAL_PATH)
 147                 global_magic = 0;
 148
 149         if (elt[0] != ':' || literal_global ||
 150             (flags & PATHSPEC_LITERAL_PATH)) {
 151                 ; /* nothing to do */
 152         } else if (elt[1] == '(') {
 153                 /* longhand */
 154                 const char *nextat;
 155                 for (copyfrom = elt + 2;
 156                      *copyfrom && *copyfrom != ')';
 157                      copyfrom = nextat) {
 158                         size_t len = strcspn(copyfrom, ",)");
 159                         if (copyfrom[len] == ',')
 160                                 nextat = copyfrom + len + 1;
 161                         else
 162                                 /* handle ')' and '\0' */
 163                                 nextat = copyfrom + len;
 164                         if (!len)
 165                                 continue;
 166                         for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 167                                 if (strlen(pathspec_magic[i].name) == len &&
 168                                     !strncmp(pathspec_magic[i].name, copyfrom, len)) {
 169                                         magic |= pathspec_magic[i].bit;
 170                                         break;
 171                                 }
 172                                 if (starts_with(copyfrom, "prefix:")) {
 173                                         char *endptr;
 174                                         pathspec_prefix = strtol(copyfrom + 7,
 175                                                                  &endptr, 10);
 176                                         if (endptr - copyfrom != len)
 177                                                 die(_("invalid parameter for pathspec magic 'prefix'"));
 178                                         /* "i" would be wrong, but it does not matter */
 179                                         break;
 180                                 }
 181                         }
 182                         if (ARRAY_SIZE(pathspec_magic) <= i)
 183                                 die(_("Invalid pathspec magic '%.*s' in '%s'"),
 184                                     (int) len, copyfrom, elt);
 185                 }
 186                 if (*copyfrom != ')')
 187                         die(_("Missing ')' at the end of pathspec magic in '%s'"), elt);
 188                 long_magic_end = copyfrom;
 189                 copyfrom++;
 190         } else {
 191                 /* shorthand */
 192                 for (copyfrom = elt + 1;
 193                      *copyfrom && *copyfrom != ':';
 194                      copyfrom++) {
 195                         char ch = *copyfrom;
 196
 197                         if (!is_pathspec_magic(ch))
 198                                 break;
 199                         for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++)
 200                                 if (pathspec_magic[i].mnemonic == ch) {
 201                                         short_magic |= pathspec_magic[i].bit;
 202                                         break;
 203                                 }
 204                         if (ARRAY_SIZE(pathspec_magic) <= i)
 205                                 die(_("Unimplemented pathspec magic '%c' in '%s'"),
 206                                     ch, elt);
 207                 }
 208                 if (*copyfrom == ':')
 209                         copyfrom++;
 210         }
 211
 212         magic |= short_magic;
 213         *p_short_magic = short_magic;
 214
 215         /* --noglob-pathspec adds :(literal) _unless_ :(glob) is specified */
 216         if (noglob_global && !(magic & PATHSPEC_GLOB))
 217                 global_magic |= PATHSPEC_LITERAL;
 218
 219         /* --glob-pathspec is overridden by :(literal) */
 220         if ((global_magic & PATHSPEC_GLOB) && (magic & PATHSPEC_LITERAL))
 221                 global_magic &= ~PATHSPEC_GLOB;
 222
 223         magic |= global_magic;
 224
 225         if (pathspec_prefix >= 0 &&
 226             (prefixlen || (prefix && *prefix)))
 227                 die("BUG: 'prefix' magic is supposed to be used at worktree's root");
 228
 229         if ((magic & PATHSPEC_LITERAL) && (magic & PATHSPEC_GLOB))
 230                 die(_("%s: 'literal' and 'glob' are incompatible"), elt);
 231
 232         if (pathspec_prefix >= 0) {
 233                 match = xstrdup(copyfrom);
 234                 prefixlen = pathspec_prefix;
 235         } else if (magic & PATHSPEC_FROMTOP) {
 236                 match = xstrdup(copyfrom);
 237                 prefixlen = 0;
 238         } else {
 239                 match = prefix_path_gently(prefix, prefixlen, &prefixlen, copyfrom);
 240                 if (!match)
 241                         die(_("%s: '%s' is outside repository"), elt, copyfrom);
 242         }
 243         *raw = item->match = match;
 244         /*
 245          * Prefix the pathspec (keep all magic) and assign to
 246          * original. Useful for passing to another command.
 247          */
 248         if (flags & PATHSPEC_PREFIX_ORIGIN) {
 249                 struct strbuf sb = STRBUF_INIT;
 250                 if (prefixlen && !literal_global) {
 251                         /* Preserve the actual prefix length of each pattern */
 252                         if (short_magic)
 253                                 prefix_short_magic(&sb, prefixlen, short_magic);
 254                         else if (long_magic_end) {
 255                                 strbuf_add(&sb, elt, long_magic_end - elt);
 256                                 strbuf_addf(&sb, ",prefix:%d)", prefixlen);
 257                         } else
 258                                 strbuf_addf(&sb, ":(prefix:%d)", prefixlen);
 259                 }
 260                 strbuf_addstr(&sb, match);
 261                 item->original = strbuf_detach(&sb, NULL);
 262         } else
 263                 item->original = elt;
 264         item->len = strlen(item->match);
 265         item->prefix = prefixlen;
 266
 267         if ((flags & PATHSPEC_STRIP_SUBMODULE_SLASH_CHEAP) &&
 268             (item->len >= 1 && item->match[item->len - 1] == '/') &&
 269             (i = cache_name_pos(item->match, item->len - 1)) >= 0 &&
 270             S_ISGITLINK(active_cache[i]->ce_mode)) {
 271                 item->len--;
 272                 match[item->len] = '\0';
 273         }
 274
 275         if (flags & PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE)
 276                 for (i = 0; i < active_nr; i++) {
 277                         struct cache_entry *ce = active_cache[i];
 278                         int ce_len = ce_namelen(ce);
 279
 280                         if (!S_ISGITLINK(ce->ce_mode))
 281                                 continue;
 282
 283                         if (item->len <= ce_len || match[ce_len] != '/' ||
 284                             memcmp(ce->name, match, ce_len))
 285                                 continue;
 286                         if (item->len == ce_len + 1) {
 287                                 /* strip trailing slash */
 288                                 item->len--;
 289                                 match[item->len] = '\0';
 290                         } else
 291                                 die (_("Pathspec '%s' is in submodule '%.*s'"),
 292                                      elt, ce_len, ce->name);
 293                 }
 294
 295         if (magic & PATHSPEC_LITERAL)
 296                 item->nowildcard_len = item->len;
 297         else {
 298                 item->nowildcard_len = simple_length(item->match);
 299                 if (item->nowildcard_len < prefixlen)
 300                         item->nowildcard_len = prefixlen;
 301         }
 302         item->flags = 0;
 303         if (magic & PATHSPEC_GLOB) {
 304                 /*
 305                  * FIXME: should we enable ONESTAR in _GLOB for
 306                  * pattern "* * / * . c"?
 307                  */
 308         } else {
 309                 if (item->nowildcard_len < item->len &&
 310                     item->match[item->nowildcard_len] == '*' &&
 311                     no_wildcard(item->match + item->nowildcard_len + 1))
 312                         item->flags |= PATHSPEC_ONESTAR;
 313         }
 314
 315         /* sanity checks, pathspec matchers assume these are sane */
 316         assert(item->nowildcard_len <= item->len &&
 317                item->prefix         <= item->len);
 318         return magic;
 319 }
 320
 321 static int pathspec_item_cmp(const void *a_, const void *b_)
 322 {
 323         struct pathspec_item *a, *b;
 324
 325         a = (struct pathspec_item *)a_;
 326         b = (struct pathspec_item *)b_;
 327         return strcmp(a->match, b->match);
 328 }
 329
 330 static void NORETURN unsupported_magic(const char *pattern,
 331                                        unsigned magic,
 332                                        unsigned short_magic)
 333 {
 334         struct strbuf sb = STRBUF_INIT;
 335         int i, n;
 336         for (n = i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 337                 const struct pathspec_magic *m = pathspec_magic + i;
 338                 if (!(magic & m->bit))
 339                         continue;
 340                 if (sb.len)
 341                         strbuf_addstr(&sb, " ");
 342                 if (short_magic & m->bit)
 343                         strbuf_addf(&sb, "'%c'", m->mnemonic);
 344                 else
 345                         strbuf_addf(&sb, "'%s'", m->name);
 346                 n++;
 347         }
 348         /*
 349          * We may want to substitute "this command" with a command
 350          * name. E.g. when add--interactive dies when running
 351          * "checkout -p"
 352          */
 353         die(_("%s: pathspec magic not supported by this command: %s"),
 354             pattern, sb.buf);
 355 }
 356
 357 /*
 358  * Given command line arguments and a prefix, convert the input to
 359  * pathspec. die() if any magic in magic_mask is used.
 360  */
 361 void parse_pathspec(struct pathspec *pathspec,
 362                     unsigned magic_mask, unsigned flags,
 363                     const char *prefix, const char **argv)
 364 {
 365         struct pathspec_item *item;
 366         const char *entry = argv ? *argv : NULL;
 367         int i, n, prefixlen, nr_exclude = 0;
 368
 369         memset(pathspec, 0, sizeof(*pathspec));
 370
 371         if (flags & PATHSPEC_MAXDEPTH_VALID)
 372                 pathspec->magic |= PATHSPEC_MAXDEPTH;
 373
 374         /* No arguments, no prefix -> no pathspec */
 375         if (!entry && !prefix)
 376                 return;
 377
 378         if ((flags & PATHSPEC_PREFER_CWD) &&
 379             (flags & PATHSPEC_PREFER_FULL))
 380                 die("BUG: PATHSPEC_PREFER_CWD and PATHSPEC_PREFER_FULL are incompatible");
 381
 382         /* No arguments with prefix -> prefix pathspec */
 383         if (!entry) {
 384                 static const char *raw[2];
 385
 386                 if (flags & PATHSPEC_PREFER_FULL)
 387                         return;
 388
 389                 if (!(flags & PATHSPEC_PREFER_CWD))
 390                         die("BUG: PATHSPEC_PREFER_CWD requires arguments");
 391
 392                 pathspec->items = item = xmalloc(sizeof(*item));
 393                 memset(item, 0, sizeof(*item));
 394                 item->match = prefix;
 395                 item->original = prefix;
 396                 item->nowildcard_len = item->len = strlen(prefix);
 397                 item->prefix = item->len;
 398                 raw[0] = prefix;
 399                 raw[1] = NULL;
 400                 pathspec->nr = 1;
 401                 pathspec->_raw = raw;
 402                 return;
 403         }
 404
 405         n = 0;
 406         while (argv[n])
 407                 n++;
 408
 409         pathspec->nr = n;
 410         pathspec->items = item = xmalloc(sizeof(*item) * n);
 411         pathspec->_raw = argv;
 412         prefixlen = prefix ? strlen(prefix) : 0;
 413
 414         for (i = 0; i < n; i++) {
 415                 unsigned short_magic;
 416                 entry = argv[i];
 417
 418                 item[i].magic = prefix_pathspec(item + i, &short_magic,
 419                                                 argv + i, flags,
 420                                                 prefix, prefixlen, entry);
 421                 if ((flags & PATHSPEC_LITERAL_PATH) &&
 422                     !(magic_mask & PATHSPEC_LITERAL))
 423                         item[i].magic |= PATHSPEC_LITERAL;
 424                 if (item[i].magic & PATHSPEC_EXCLUDE)
 425                         nr_exclude++;
 426                 if (item[i].magic & magic_mask)
 427                         unsupported_magic(entry,
 428                                           item[i].magic & magic_mask,
 429                                           short_magic);
 430
 431                 if ((flags & PATHSPEC_SYMLINK_LEADING_PATH) &&
 432                     has_symlink_leading_path(item[i].match, item[i].len)) {
 433                         die(_("pathspec '%s' is beyond a symbolic link"), entry);
 434                 }
 435
 436                 if (item[i].nowildcard_len < item[i].len)
 437                         pathspec->has_wildcard = 1;
 438                 pathspec->magic |= item[i].magic;
 439         }
 440
 441         if (nr_exclude == n)
 442                 die(_("There is nothing to exclude from by :(exclude) patterns.\n"
 443                       "Perhaps you forgot to add either ':/' or '.' ?"));
 444
 445
 446         if (pathspec->magic & PATHSPEC_MAXDEPTH) {
 447                 if (flags & PATHSPEC_KEEP_ORDER)
 448                         die("BUG: PATHSPEC_MAXDEPTH_VALID and PATHSPEC_KEEP_ORDER are incompatible");
 449                 qsort(pathspec->items, pathspec->nr,
 450                       sizeof(struct pathspec_item), pathspec_item_cmp);
 451         }
 452 }
 453
 454 /*
 455  * N.B. get_pathspec() is deprecated in favor of the "struct pathspec"
 456  * based interface - see pathspec.c:parse_pathspec().
 457  *
 458  * Arguments:
 459  *  - prefix - a path relative to the root of the working tree
 460  *  - pathspec - a list of paths underneath the prefix path
 461  *
 462  * Iterates over pathspec, prepending each path with prefix,
 463  * and return the resulting list.
 464  *
 465  * If pathspec is empty, return a singleton list containing prefix.
 466  *
 467  * If pathspec and prefix are both empty, return an empty list.
 468  *
 469  * This is typically used by built-in commands such as add.c, in order
 470  * to normalize argv arguments provided to the built-in into a list of
 471  * paths to process, all relative to the root of the working tree.
 472  */
 473 const char **get_pathspec(const char *prefix, const char **pathspec)
 474 {
 475         struct pathspec ps;
 476         parse_pathspec(&ps,
 477                        PATHSPEC_ALL_MAGIC &
 478                        ~(PATHSPEC_FROMTOP | PATHSPEC_LITERAL),
 479                        PATHSPEC_PREFER_CWD,
 480                        prefix, pathspec);
 481         return ps._raw;
 482 }
 483
 484 void copy_pathspec(struct pathspec *dst, const struct pathspec *src)
 485 {
 486         *dst = *src;
 487         dst->items = xmalloc(sizeof(struct pathspec_item) * dst->nr);
 488         memcpy(dst->items, src->items,
 489                sizeof(struct pathspec_item) * dst->nr);
 490 }
 491
 492 void free_pathspec(struct pathspec *pathspec)
 493 {
 494         free(pathspec->items);
 495         pathspec->items = NULL;
 496 }