pathspec.con commit pathspec: support :(glob) syntax (bd30c2e)
   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 */
  18void 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                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 */
  48char *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_ICASE, '\0', "icase" },
  61 *      { PATHSPEC_RECURSIVE, '*', "recursive" },
  62 *      { PATHSPEC_REGEXP, '\0', "regexp" },
  63 *
  64 */
  65
  66static struct pathspec_magic {
  67        unsigned bit;
  68        char mnemonic; /* this cannot be ':'! */
  69        const char *name;
  70} pathspec_magic[] = {
  71        { PATHSPEC_FROMTOP, '/', "top" },
  72        { PATHSPEC_LITERAL,   0, "literal" },
  73        { PATHSPEC_GLOB,   '\0', "glob" },
  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 */
  89static 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        unsigned magic = 0, short_magic = 0, global_magic = 0;
  99        const char *copyfrom = elt, *long_magic_end = NULL;
 100        char *match;
 101        int i, pathspec_prefix = -1;
 102
 103        if (literal_global < 0)
 104                literal_global = git_env_bool(GIT_LITERAL_PATHSPECS_ENVIRONMENT, 0);
 105        if (literal_global)
 106                global_magic |= PATHSPEC_LITERAL;
 107
 108        if (glob_global < 0)
 109                glob_global = git_env_bool(GIT_GLOB_PATHSPECS_ENVIRONMENT, 0);
 110        if (glob_global)
 111                global_magic |= PATHSPEC_GLOB;
 112
 113        if (noglob_global < 0)
 114                noglob_global = git_env_bool(GIT_NOGLOB_PATHSPECS_ENVIRONMENT, 0);
 115
 116        if (glob_global && noglob_global)
 117                die(_("global 'glob' and 'noglob' pathspec settings are incompatible"));
 118
 119        if ((global_magic & PATHSPEC_LITERAL) &&
 120            (global_magic & ~PATHSPEC_LITERAL))
 121                die(_("global 'literal' pathspec setting is incompatible "
 122                      "with all other global pathspec settings"));
 123
 124        if (elt[0] != ':' || literal_global) {
 125                ; /* nothing to do */
 126        } else if (elt[1] == '(') {
 127                /* longhand */
 128                const char *nextat;
 129                for (copyfrom = elt + 2;
 130                     *copyfrom && *copyfrom != ')';
 131                     copyfrom = nextat) {
 132                        size_t len = strcspn(copyfrom, ",)");
 133                        if (copyfrom[len] == ',')
 134                                nextat = copyfrom + len + 1;
 135                        else
 136                                /* handle ')' and '\0' */
 137                                nextat = copyfrom + len;
 138                        if (!len)
 139                                continue;
 140                        for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 141                                if (strlen(pathspec_magic[i].name) == len &&
 142                                    !strncmp(pathspec_magic[i].name, copyfrom, len)) {
 143                                        magic |= pathspec_magic[i].bit;
 144                                        break;
 145                                }
 146                                if (!prefixcmp(copyfrom, "prefix:")) {
 147                                        char *endptr;
 148                                        pathspec_prefix = strtol(copyfrom + 7,
 149                                                                 &endptr, 10);
 150                                        if (endptr - copyfrom != len)
 151                                                die(_("invalid parameter for pathspec magic 'prefix'"));
 152                                        /* "i" would be wrong, but it does not matter */
 153                                        break;
 154                                }
 155                        }
 156                        if (ARRAY_SIZE(pathspec_magic) <= i)
 157                                die(_("Invalid pathspec magic '%.*s' in '%s'"),
 158                                    (int) len, copyfrom, elt);
 159                }
 160                if (*copyfrom != ')')
 161                        die(_("Missing ')' at the end of pathspec magic in '%s'"), elt);
 162                long_magic_end = copyfrom;
 163                copyfrom++;
 164        } else {
 165                /* shorthand */
 166                for (copyfrom = elt + 1;
 167                     *copyfrom && *copyfrom != ':';
 168                     copyfrom++) {
 169                        char ch = *copyfrom;
 170
 171                        if (!is_pathspec_magic(ch))
 172                                break;
 173                        for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++)
 174                                if (pathspec_magic[i].mnemonic == ch) {
 175                                        short_magic |= pathspec_magic[i].bit;
 176                                        break;
 177                                }
 178                        if (ARRAY_SIZE(pathspec_magic) <= i)
 179                                die(_("Unimplemented pathspec magic '%c' in '%s'"),
 180                                    ch, elt);
 181                }
 182                if (*copyfrom == ':')
 183                        copyfrom++;
 184        }
 185
 186        magic |= short_magic;
 187        *p_short_magic = short_magic;
 188
 189        /* --noglob-pathspec adds :(literal) _unless_ :(glob) is specifed */
 190        if (noglob_global && !(magic & PATHSPEC_GLOB))
 191                global_magic |= PATHSPEC_LITERAL;
 192
 193        /* --glob-pathspec is overriden by :(literal) */
 194        if ((global_magic & PATHSPEC_GLOB) && (magic & PATHSPEC_LITERAL))
 195                global_magic &= ~PATHSPEC_GLOB;
 196
 197        magic |= global_magic;
 198
 199        if (pathspec_prefix >= 0 &&
 200            (prefixlen || (prefix && *prefix)))
 201                die("BUG: 'prefix' magic is supposed to be used at worktree's root");
 202
 203        if ((magic & PATHSPEC_LITERAL) && (magic & PATHSPEC_GLOB))
 204                die(_("%s: 'literal' and 'glob' are incompatible"), elt);
 205
 206        if (pathspec_prefix >= 0) {
 207                match = xstrdup(copyfrom);
 208                prefixlen = pathspec_prefix;
 209        } else if (magic & PATHSPEC_FROMTOP) {
 210                match = xstrdup(copyfrom);
 211                prefixlen = 0;
 212        } else {
 213                match = prefix_path_gently(prefix, prefixlen, &prefixlen, copyfrom);
 214                if (!match)
 215                        die(_("%s: '%s' is outside repository"), elt, copyfrom);
 216        }
 217        *raw = item->match = match;
 218        /*
 219         * Prefix the pathspec (keep all magic) and assign to
 220         * original. Useful for passing to another command.
 221         */
 222        if (flags & PATHSPEC_PREFIX_ORIGIN) {
 223                struct strbuf sb = STRBUF_INIT;
 224                const char *start = elt;
 225                if (prefixlen && !literal_global) {
 226                        /* Preserve the actual prefix length of each pattern */
 227                        if (long_magic_end) {
 228                                strbuf_add(&sb, start, long_magic_end - start);
 229                                strbuf_addf(&sb, ",prefix:%d", prefixlen);
 230                                start = long_magic_end;
 231                        } else {
 232                                if (*start == ':')
 233                                        start++;
 234                                strbuf_addf(&sb, ":(prefix:%d)", prefixlen);
 235                        }
 236                }
 237                strbuf_add(&sb, start, copyfrom - start);
 238                strbuf_addstr(&sb, match);
 239                item->original = strbuf_detach(&sb, NULL);
 240        } else
 241                item->original = elt;
 242        item->len = strlen(item->match);
 243        item->prefix = prefixlen;
 244
 245        if ((flags & PATHSPEC_STRIP_SUBMODULE_SLASH_CHEAP) &&
 246            (item->len >= 1 && item->match[item->len - 1] == '/') &&
 247            (i = cache_name_pos(item->match, item->len - 1)) >= 0 &&
 248            S_ISGITLINK(active_cache[i]->ce_mode)) {
 249                item->len--;
 250                match[item->len] = '\0';
 251        }
 252
 253        if (flags & PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE)
 254                for (i = 0; i < active_nr; i++) {
 255                        struct cache_entry *ce = active_cache[i];
 256                        int ce_len = ce_namelen(ce);
 257
 258                        if (!S_ISGITLINK(ce->ce_mode))
 259                                continue;
 260
 261                        if (item->len <= ce_len || match[ce_len] != '/' ||
 262                            memcmp(ce->name, match, ce_len))
 263                                continue;
 264                        if (item->len == ce_len + 1) {
 265                                /* strip trailing slash */
 266                                item->len--;
 267                                match[item->len] = '\0';
 268                        } else
 269                                die (_("Pathspec '%s' is in submodule '%.*s'"),
 270                                     elt, ce_len, ce->name);
 271                }
 272
 273        if (magic & PATHSPEC_LITERAL)
 274                item->nowildcard_len = item->len;
 275        else {
 276                item->nowildcard_len = simple_length(item->match);
 277                if (item->nowildcard_len < prefixlen)
 278                        item->nowildcard_len = prefixlen;
 279        }
 280        item->flags = 0;
 281        if (magic & PATHSPEC_GLOB) {
 282                /*
 283                 * FIXME: should we enable ONESTAR in _GLOB for
 284                 * pattern "* * / * . c"?
 285                 */
 286        } else {
 287                if (item->nowildcard_len < item->len &&
 288                    item->match[item->nowildcard_len] == '*' &&
 289                    no_wildcard(item->match + item->nowildcard_len + 1))
 290                        item->flags |= PATHSPEC_ONESTAR;
 291        }
 292
 293        /* sanity checks, pathspec matchers assume these are sane */
 294        assert(item->nowildcard_len <= item->len &&
 295               item->prefix         <= item->len);
 296        return magic;
 297}
 298
 299static int pathspec_item_cmp(const void *a_, const void *b_)
 300{
 301        struct pathspec_item *a, *b;
 302
 303        a = (struct pathspec_item *)a_;
 304        b = (struct pathspec_item *)b_;
 305        return strcmp(a->match, b->match);
 306}
 307
 308static void NORETURN unsupported_magic(const char *pattern,
 309                                       unsigned magic,
 310                                       unsigned short_magic)
 311{
 312        struct strbuf sb = STRBUF_INIT;
 313        int i, n;
 314        for (n = i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
 315                const struct pathspec_magic *m = pathspec_magic + i;
 316                if (!(magic & m->bit))
 317                        continue;
 318                if (sb.len)
 319                        strbuf_addstr(&sb, " ");
 320                if (short_magic & m->bit)
 321                        strbuf_addf(&sb, "'%c'", m->mnemonic);
 322                else
 323                        strbuf_addf(&sb, "'%s'", m->name);
 324                n++;
 325        }
 326        /*
 327         * We may want to substitute "this command" with a command
 328         * name. E.g. when add--interactive dies when running
 329         * "checkout -p"
 330         */
 331        die(_("%s: pathspec magic not supported by this command: %s"),
 332            pattern, sb.buf);
 333}
 334
 335/*
 336 * Given command line arguments and a prefix, convert the input to
 337 * pathspec. die() if any magic in magic_mask is used.
 338 */
 339void parse_pathspec(struct pathspec *pathspec,
 340                    unsigned magic_mask, unsigned flags,
 341                    const char *prefix, const char **argv)
 342{
 343        struct pathspec_item *item;
 344        const char *entry = argv ? *argv : NULL;
 345        int i, n, prefixlen;
 346
 347        memset(pathspec, 0, sizeof(*pathspec));
 348
 349        if (flags & PATHSPEC_MAXDEPTH_VALID)
 350                pathspec->magic |= PATHSPEC_MAXDEPTH;
 351
 352        /* No arguments, no prefix -> no pathspec */
 353        if (!entry && !prefix)
 354                return;
 355
 356        if ((flags & PATHSPEC_PREFER_CWD) &&
 357            (flags & PATHSPEC_PREFER_FULL))
 358                die("BUG: PATHSPEC_PREFER_CWD and PATHSPEC_PREFER_FULL are incompatible");
 359
 360        /* No arguments with prefix -> prefix pathspec */
 361        if (!entry) {
 362                static const char *raw[2];
 363
 364                if (flags & PATHSPEC_PREFER_FULL)
 365                        return;
 366
 367                if (!(flags & PATHSPEC_PREFER_CWD))
 368                        die("BUG: PATHSPEC_PREFER_CWD requires arguments");
 369
 370                pathspec->items = item = xmalloc(sizeof(*item));
 371                memset(item, 0, sizeof(*item));
 372                item->match = prefix;
 373                item->original = prefix;
 374                item->nowildcard_len = item->len = strlen(prefix);
 375                item->prefix = item->len;
 376                raw[0] = prefix;
 377                raw[1] = NULL;
 378                pathspec->nr = 1;
 379                pathspec->_raw = raw;
 380                return;
 381        }
 382
 383        n = 0;
 384        while (argv[n])
 385                n++;
 386
 387        pathspec->nr = n;
 388        pathspec->items = item = xmalloc(sizeof(*item) * n);
 389        pathspec->_raw = argv;
 390        prefixlen = prefix ? strlen(prefix) : 0;
 391
 392        for (i = 0; i < n; i++) {
 393                unsigned short_magic;
 394                entry = argv[i];
 395
 396                item[i].magic = prefix_pathspec(item + i, &short_magic,
 397                                                argv + i, flags,
 398                                                prefix, prefixlen, entry);
 399                if (item[i].magic & magic_mask)
 400                        unsupported_magic(entry,
 401                                          item[i].magic & magic_mask,
 402                                          short_magic);
 403
 404                if ((flags & PATHSPEC_SYMLINK_LEADING_PATH) &&
 405                    has_symlink_leading_path(item[i].match, item[i].len)) {
 406                        die(_("pathspec '%s' is beyond a symbolic link"), entry);
 407                }
 408
 409                if (item[i].nowildcard_len < item[i].len)
 410                        pathspec->has_wildcard = 1;
 411                pathspec->magic |= item[i].magic;
 412        }
 413
 414
 415        if (pathspec->magic & PATHSPEC_MAXDEPTH) {
 416                if (flags & PATHSPEC_KEEP_ORDER)
 417                        die("BUG: PATHSPEC_MAXDEPTH_VALID and PATHSPEC_KEEP_ORDER are incompatible");
 418                qsort(pathspec->items, pathspec->nr,
 419                      sizeof(struct pathspec_item), pathspec_item_cmp);
 420        }
 421}
 422
 423/*
 424 * N.B. get_pathspec() is deprecated in favor of the "struct pathspec"
 425 * based interface - see pathspec.c:parse_pathspec().
 426 *
 427 * Arguments:
 428 *  - prefix - a path relative to the root of the working tree
 429 *  - pathspec - a list of paths underneath the prefix path
 430 *
 431 * Iterates over pathspec, prepending each path with prefix,
 432 * and return the resulting list.
 433 *
 434 * If pathspec is empty, return a singleton list containing prefix.
 435 *
 436 * If pathspec and prefix are both empty, return an empty list.
 437 *
 438 * This is typically used by built-in commands such as add.c, in order
 439 * to normalize argv arguments provided to the built-in into a list of
 440 * paths to process, all relative to the root of the working tree.
 441 */
 442const char **get_pathspec(const char *prefix, const char **pathspec)
 443{
 444        struct pathspec ps;
 445        parse_pathspec(&ps,
 446                       PATHSPEC_ALL_MAGIC &
 447                       ~(PATHSPEC_FROMTOP | PATHSPEC_LITERAL),
 448                       PATHSPEC_PREFER_CWD,
 449                       prefix, pathspec);
 450        return ps._raw;
 451}
 452
 453void copy_pathspec(struct pathspec *dst, const struct pathspec *src)
 454{
 455        *dst = *src;
 456        dst->items = xmalloc(sizeof(struct pathspec_item) * dst->nr);
 457        memcpy(dst->items, src->items,
 458               sizeof(struct pathspec_item) * dst->nr);
 459}
 460
 461void free_pathspec(struct pathspec *pathspec)
 462{
 463        free(pathspec->items);
 464        pathspec->items = NULL;
 465}