ls-files.con commit git-upload-pack: More efficient usage of the has_sha1 array (794f9fe)
   1/*
   2 * This merges the file listing in the directory cache index
   3 * with the actual working directory list, and shows different
   4 * combinations of the two.
   5 *
   6 * Copyright (C) Linus Torvalds, 2005
   7 */
   8#include <dirent.h>
   9#include <fnmatch.h>
  10
  11#include "cache.h"
  12#include "quote.h"
  13
  14static int show_deleted = 0;
  15static int show_cached = 0;
  16static int show_others = 0;
  17static int show_ignored = 0;
  18static int show_stage = 0;
  19static int show_unmerged = 0;
  20static int show_modified = 0;
  21static int show_killed = 0;
  22static int line_terminator = '\n';
  23
  24static int prefix_len = 0, prefix_offset = 0;
  25static const char *prefix = NULL;
  26static const char **pathspec = NULL;
  27
  28static const char *tag_cached = "";
  29static const char *tag_unmerged = "";
  30static const char *tag_removed = "";
  31static const char *tag_other = "";
  32static const char *tag_killed = "";
  33static const char *tag_modified = "";
  34
  35static const char *exclude_per_dir = NULL;
  36
  37/* We maintain three exclude pattern lists:
  38 * EXC_CMDL lists patterns explicitly given on the command line.
  39 * EXC_DIRS lists patterns obtained from per-directory ignore files.
  40 * EXC_FILE lists patterns from fallback ignore files.
  41 */
  42#define EXC_CMDL 0
  43#define EXC_DIRS 1
  44#define EXC_FILE 2
  45static struct exclude_list {
  46        int nr;
  47        int alloc;
  48        struct exclude {
  49                const char *pattern;
  50                const char *base;
  51                int baselen;
  52        } **excludes;
  53} exclude_list[3];
  54
  55static void add_exclude(const char *string, const char *base,
  56                        int baselen, struct exclude_list *which)
  57{
  58        struct exclude *x = xmalloc(sizeof (*x));
  59
  60        x->pattern = string;
  61        x->base = base;
  62        x->baselen = baselen;
  63        if (which->nr == which->alloc) {
  64                which->alloc = alloc_nr(which->alloc);
  65                which->excludes = realloc(which->excludes,
  66                                          which->alloc * sizeof(x));
  67        }
  68        which->excludes[which->nr++] = x;
  69}
  70
  71static int add_excludes_from_file_1(const char *fname,
  72                                    const char *base,
  73                                    int baselen,
  74                                    struct exclude_list *which)
  75{
  76        int fd, i;
  77        long size;
  78        char *buf, *entry;
  79
  80        fd = open(fname, O_RDONLY);
  81        if (fd < 0)
  82                goto err;
  83        size = lseek(fd, 0, SEEK_END);
  84        if (size < 0)
  85                goto err;
  86        lseek(fd, 0, SEEK_SET);
  87        if (size == 0) {
  88                close(fd);
  89                return 0;
  90        }
  91        buf = xmalloc(size);
  92        if (read(fd, buf, size) != size)
  93                goto err;
  94        close(fd);
  95
  96        entry = buf;
  97        for (i = 0; i < size; i++) {
  98                if (buf[i] == '\n') {
  99                        if (entry != buf + i && entry[0] != '#') {
 100                                buf[i] = 0;
 101                                add_exclude(entry, base, baselen, which);
 102                        }
 103                        entry = buf + i + 1;
 104                }
 105        }
 106        return 0;
 107
 108 err:
 109        if (0 <= fd)
 110                close(fd);
 111        return -1;
 112}
 113
 114static void add_excludes_from_file(const char *fname)
 115{
 116        if (add_excludes_from_file_1(fname, "", 0,
 117                                     &exclude_list[EXC_FILE]) < 0)
 118                die("cannot use %s as an exclude file", fname);
 119}
 120
 121static int push_exclude_per_directory(const char *base, int baselen)
 122{
 123        char exclude_file[PATH_MAX];
 124        struct exclude_list *el = &exclude_list[EXC_DIRS];
 125        int current_nr = el->nr;
 126
 127        if (exclude_per_dir) {
 128                memcpy(exclude_file, base, baselen);
 129                strcpy(exclude_file + baselen, exclude_per_dir);
 130                add_excludes_from_file_1(exclude_file, base, baselen, el);
 131        }
 132        return current_nr;
 133}
 134
 135static void pop_exclude_per_directory(int stk)
 136{
 137        struct exclude_list *el = &exclude_list[EXC_DIRS];
 138
 139        while (stk < el->nr)
 140                free(el->excludes[--el->nr]);
 141}
 142
 143/* Scan the list and let the last match determines the fate.
 144 * Return 1 for exclude, 0 for include and -1 for undecided.
 145 */
 146static int excluded_1(const char *pathname,
 147                      int pathlen,
 148                      struct exclude_list *el)
 149{
 150        int i;
 151
 152        if (el->nr) {
 153                for (i = el->nr - 1; 0 <= i; i--) {
 154                        struct exclude *x = el->excludes[i];
 155                        const char *exclude = x->pattern;
 156                        int to_exclude = 1;
 157
 158                        if (*exclude == '!') {
 159                                to_exclude = 0;
 160                                exclude++;
 161                        }
 162
 163                        if (!strchr(exclude, '/')) {
 164                                /* match basename */
 165                                const char *basename = strrchr(pathname, '/');
 166                                basename = (basename) ? basename+1 : pathname;
 167                                if (fnmatch(exclude, basename, 0) == 0)
 168                                        return to_exclude;
 169                        }
 170                        else {
 171                                /* match with FNM_PATHNAME:
 172                                 * exclude has base (baselen long) inplicitly
 173                                 * in front of it.
 174                                 */
 175                                int baselen = x->baselen;
 176                                if (*exclude == '/')
 177                                        exclude++;
 178
 179                                if (pathlen < baselen ||
 180                                    (baselen && pathname[baselen-1] != '/') ||
 181                                    strncmp(pathname, x->base, baselen))
 182                                    continue;
 183
 184                                if (fnmatch(exclude, pathname+baselen,
 185                                            FNM_PATHNAME) == 0)
 186                                        return to_exclude;
 187                        }
 188                }
 189        }
 190        return -1; /* undecided */
 191}
 192
 193static int excluded(const char *pathname)
 194{
 195        int pathlen = strlen(pathname);
 196        int st;
 197
 198        for (st = EXC_CMDL; st <= EXC_FILE; st++) {
 199                switch (excluded_1(pathname, pathlen, &exclude_list[st])) {
 200                case 0:
 201                        return 0;
 202                case 1:
 203                        return 1;
 204                }
 205        }
 206        return 0;
 207}
 208
 209struct nond_on_fs {
 210        int len;
 211        char name[0];
 212};
 213
 214static struct nond_on_fs **dir;
 215static int nr_dir;
 216static int dir_alloc;
 217
 218static void add_name(const char *pathname, int len)
 219{
 220        struct nond_on_fs *ent;
 221
 222        if (cache_name_pos(pathname, len) >= 0)
 223                return;
 224
 225        if (nr_dir == dir_alloc) {
 226                dir_alloc = alloc_nr(dir_alloc);
 227                dir = xrealloc(dir, dir_alloc*sizeof(ent));
 228        }
 229        ent = xmalloc(sizeof(*ent) + len + 1);
 230        ent->len = len;
 231        memcpy(ent->name, pathname, len);
 232        ent->name[len] = 0;
 233        dir[nr_dir++] = ent;
 234}
 235
 236/*
 237 * Read a directory tree. We currently ignore anything but
 238 * directories, regular files and symlinks. That's because git
 239 * doesn't handle them at all yet. Maybe that will change some
 240 * day.
 241 *
 242 * Also, we ignore the name ".git" (even if it is not a directory).
 243 * That likely will not change.
 244 */
 245static void read_directory(const char *path, const char *base, int baselen)
 246{
 247        DIR *dir = opendir(path);
 248
 249        if (dir) {
 250                int exclude_stk;
 251                struct dirent *de;
 252                char fullname[MAXPATHLEN + 1];
 253                memcpy(fullname, base, baselen);
 254
 255                exclude_stk = push_exclude_per_directory(base, baselen);
 256
 257                while ((de = readdir(dir)) != NULL) {
 258                        int len;
 259
 260                        if ((de->d_name[0] == '.') &&
 261                            (de->d_name[1] == 0 ||
 262                             !strcmp(de->d_name + 1, ".") ||
 263                             !strcmp(de->d_name + 1, "git")))
 264                                continue;
 265                        len = strlen(de->d_name);
 266                        memcpy(fullname + baselen, de->d_name, len+1);
 267                        if (excluded(fullname) != show_ignored)
 268                                continue;
 269
 270                        switch (DTYPE(de)) {
 271                        struct stat st;
 272                        default:
 273                                continue;
 274                        case DT_UNKNOWN:
 275                                if (lstat(fullname, &st))
 276                                        continue;
 277                                if (S_ISREG(st.st_mode) || S_ISLNK(st.st_mode))
 278                                        break;
 279                                if (!S_ISDIR(st.st_mode))
 280                                        continue;
 281                                /* fallthrough */
 282                        case DT_DIR:
 283                                memcpy(fullname + baselen + len, "/", 2);
 284                                read_directory(fullname, fullname,
 285                                               baselen + len + 1);
 286                                continue;
 287                        case DT_REG:
 288                        case DT_LNK:
 289                                break;
 290                        }
 291                        add_name(fullname, baselen + len);
 292                }
 293                closedir(dir);
 294
 295                pop_exclude_per_directory(exclude_stk);
 296        }
 297}
 298
 299static int cmp_name(const void *p1, const void *p2)
 300{
 301        const struct nond_on_fs *e1 = *(const struct nond_on_fs **)p1;
 302        const struct nond_on_fs *e2 = *(const struct nond_on_fs **)p2;
 303
 304        return cache_name_compare(e1->name, e1->len,
 305                                  e2->name, e2->len);
 306}
 307
 308/*
 309 * Match a pathspec against a filename. The first "len" characters
 310 * are the common prefix
 311 */
 312static int match(const char **spec, const char *filename, int len)
 313{
 314        const char *m;
 315
 316        while ((m = *spec++) != NULL) {
 317                int matchlen = strlen(m + len);
 318
 319                if (!matchlen)
 320                        return 1;
 321                if (!strncmp(m + len, filename + len, matchlen)) {
 322                        if (m[len + matchlen - 1] == '/')
 323                                return 1;
 324                        switch (filename[len + matchlen]) {
 325                        case '/': case '\0':
 326                                return 1;
 327                        }
 328                }
 329                if (!fnmatch(m + len, filename + len, 0))
 330                        return 1;
 331        }
 332        return 0;
 333}
 334
 335static void show_dir_entry(const char *tag, struct nond_on_fs *ent)
 336{
 337        int len = prefix_len;
 338        int offset = prefix_offset;
 339
 340        if (len >= ent->len)
 341                die("git-ls-files: internal error - directory entry not superset of prefix");
 342
 343        if (pathspec && !match(pathspec, ent->name, len))
 344                return;
 345
 346        fputs(tag, stdout);
 347        write_name_quoted("", ent->name + offset, line_terminator, stdout);
 348        putchar(line_terminator);
 349}
 350
 351static void show_killed_files(void)
 352{
 353        int i;
 354        for (i = 0; i < nr_dir; i++) {
 355                struct nond_on_fs *ent = dir[i];
 356                char *cp, *sp;
 357                int pos, len, killed = 0;
 358
 359                for (cp = ent->name; cp - ent->name < ent->len; cp = sp + 1) {
 360                        sp = strchr(cp, '/');
 361                        if (!sp) {
 362                                /* If ent->name is prefix of an entry in the
 363                                 * cache, it will be killed.
 364                                 */
 365                                pos = cache_name_pos(ent->name, ent->len);
 366                                if (0 <= pos)
 367                                        die("bug in show-killed-files");
 368                                pos = -pos - 1;
 369                                while (pos < active_nr &&
 370                                       ce_stage(active_cache[pos]))
 371                                        pos++; /* skip unmerged */
 372                                if (active_nr <= pos)
 373                                        break;
 374                                /* pos points at a name immediately after
 375                                 * ent->name in the cache.  Does it expect
 376                                 * ent->name to be a directory?
 377                                 */
 378                                len = ce_namelen(active_cache[pos]);
 379                                if ((ent->len < len) &&
 380                                    !strncmp(active_cache[pos]->name,
 381                                             ent->name, ent->len) &&
 382                                    active_cache[pos]->name[ent->len] == '/')
 383                                        killed = 1;
 384                                break;
 385                        }
 386                        if (0 <= cache_name_pos(ent->name, sp - ent->name)) {
 387                                /* If any of the leading directories in
 388                                 * ent->name is registered in the cache,
 389                                 * ent->name will be killed.
 390                                 */
 391                                killed = 1;
 392                                break;
 393                        }
 394                }
 395                if (killed)
 396                        show_dir_entry(tag_killed, dir[i]);
 397        }
 398}
 399
 400static void show_ce_entry(const char *tag, struct cache_entry *ce)
 401{
 402        int len = prefix_len;
 403        int offset = prefix_offset;
 404
 405        if (len >= ce_namelen(ce))
 406                die("git-ls-files: internal error - cache entry not superset of prefix");
 407
 408        if (pathspec && !match(pathspec, ce->name, len))
 409                return;
 410
 411        if (!show_stage) {
 412                fputs(tag, stdout);
 413                write_name_quoted("", ce->name + offset, line_terminator, stdout);
 414                putchar(line_terminator);
 415        }
 416        else {
 417                printf("%s%06o %s %d\t",
 418                       tag,
 419                       ntohl(ce->ce_mode),
 420                       sha1_to_hex(ce->sha1),
 421                       ce_stage(ce));
 422                write_name_quoted("", ce->name + offset, line_terminator, stdout);
 423                putchar(line_terminator);
 424        }
 425}
 426
 427static void show_files(void)
 428{
 429        int i;
 430
 431        /* For cached/deleted files we don't need to even do the readdir */
 432        if (show_others || show_killed) {
 433                const char *path = ".", *base = "";
 434                int baselen = prefix_len;
 435
 436                if (baselen)
 437                        path = base = prefix;
 438                read_directory(path, base, baselen);
 439                qsort(dir, nr_dir, sizeof(struct nond_on_fs *), cmp_name);
 440                if (show_others)
 441                        for (i = 0; i < nr_dir; i++)
 442                                show_dir_entry(tag_other, dir[i]);
 443                if (show_killed)
 444                        show_killed_files();
 445        }
 446        if (show_cached | show_stage) {
 447                for (i = 0; i < active_nr; i++) {
 448                        struct cache_entry *ce = active_cache[i];
 449                        if (excluded(ce->name) != show_ignored)
 450                                continue;
 451                        if (show_unmerged && !ce_stage(ce))
 452                                continue;
 453                        show_ce_entry(ce_stage(ce) ? tag_unmerged : tag_cached, ce);
 454                }
 455        }
 456        if (show_deleted | show_modified) {
 457                for (i = 0; i < active_nr; i++) {
 458                        struct cache_entry *ce = active_cache[i];
 459                        struct stat st;
 460                        int err;
 461                        if (excluded(ce->name) != show_ignored)
 462                                continue;
 463                        err = lstat(ce->name, &st);
 464                        if (show_deleted && err)
 465                                show_ce_entry(tag_removed, ce);
 466                        if (show_modified && ce_modified(ce, &st))
 467                                show_ce_entry(tag_modified, ce);
 468                }
 469        }
 470}
 471
 472/*
 473 * Prune the index to only contain stuff starting with "prefix"
 474 */
 475static void prune_cache(void)
 476{
 477        int pos = cache_name_pos(prefix, prefix_len);
 478        unsigned int first, last;
 479
 480        if (pos < 0)
 481                pos = -pos-1;
 482        active_cache += pos;
 483        active_nr -= pos;
 484        first = 0;
 485        last = active_nr;
 486        while (last > first) {
 487                int next = (last + first) >> 1;
 488                struct cache_entry *ce = active_cache[next];
 489                if (!strncmp(ce->name, prefix, prefix_len)) {
 490                        first = next+1;
 491                        continue;
 492                }
 493                last = next;
 494        }
 495        active_nr = last;
 496}
 497
 498static void verify_pathspec(void)
 499{
 500        const char **p, *n, *prev;
 501        char *real_prefix;
 502        unsigned long max;
 503
 504        prev = NULL;
 505        max = PATH_MAX;
 506        for (p = pathspec; (n = *p) != NULL; p++) {
 507                int i, len = 0;
 508                for (i = 0; i < max; i++) {
 509                        char c = n[i];
 510                        if (prev && prev[i] != c)
 511                                break;
 512                        if (!c || c == '*' || c == '?')
 513                                break;
 514                        if (c == '/')
 515                                len = i+1;
 516                }
 517                prev = n;
 518                if (len < max) {
 519                        max = len;
 520                        if (!max)
 521                                break;
 522                }
 523        }
 524
 525        if (prefix_offset > max || memcmp(prev, prefix, prefix_offset))
 526                die("git-ls-files: cannot generate relative filenames containing '..'");
 527
 528        real_prefix = NULL;
 529        prefix_len = max;
 530        if (max) {
 531                real_prefix = xmalloc(max + 1);
 532                memcpy(real_prefix, prev, max);
 533                real_prefix[max] = 0;
 534        }
 535        prefix = real_prefix;
 536}
 537
 538static const char ls_files_usage[] =
 539        "git-ls-files [-z] [-t] (--[cached|deleted|others|stage|unmerged|killed|modified])* "
 540        "[ --ignored ] [--exclude=<pattern>] [--exclude-from=<file>] "
 541        "[ --exclude-per-directory=<filename> ] [--] [<file>]*";
 542
 543int main(int argc, const char **argv)
 544{
 545        int i;
 546        int exc_given = 0;
 547
 548        prefix = setup_git_directory();
 549        if (prefix)
 550                prefix_offset = strlen(prefix);
 551
 552        for (i = 1; i < argc; i++) {
 553                const char *arg = argv[i];
 554
 555                if (!strcmp(arg, "--")) {
 556                        i++;
 557                        break;
 558                }
 559                if (!strcmp(arg, "-z")) {
 560                        line_terminator = 0;
 561                        continue;
 562                }
 563                if (!strcmp(arg, "-t")) {
 564                        tag_cached = "H ";
 565                        tag_unmerged = "M ";
 566                        tag_removed = "R ";
 567                        tag_modified = "C ";
 568                        tag_other = "? ";
 569                        tag_killed = "K ";
 570                        continue;
 571                }
 572                if (!strcmp(arg, "-c") || !strcmp(arg, "--cached")) {
 573                        show_cached = 1;
 574                        continue;
 575                }
 576                if (!strcmp(arg, "-d") || !strcmp(arg, "--deleted")) {
 577                        show_deleted = 1;
 578                        continue;
 579                }
 580                if (!strcmp(arg, "-m") || !strcmp(arg, "--modified")) {
 581                        show_modified = 1;
 582                        continue;
 583                }
 584                if (!strcmp(arg, "-o") || !strcmp(arg, "--others")) {
 585                        show_others = 1;
 586                        continue;
 587                }
 588                if (!strcmp(arg, "-i") || !strcmp(arg, "--ignored")) {
 589                        show_ignored = 1;
 590                        continue;
 591                }
 592                if (!strcmp(arg, "-s") || !strcmp(arg, "--stage")) {
 593                        show_stage = 1;
 594                        continue;
 595                }
 596                if (!strcmp(arg, "-k") || !strcmp(arg, "--killed")) {
 597                        show_killed = 1;
 598                        continue;
 599                }
 600                if (!strcmp(arg, "-u") || !strcmp(arg, "--unmerged")) {
 601                        /* There's no point in showing unmerged unless
 602                         * you also show the stage information.
 603                         */
 604                        show_stage = 1;
 605                        show_unmerged = 1;
 606                        continue;
 607                }
 608                if (!strcmp(arg, "-x") && i+1 < argc) {
 609                        exc_given = 1;
 610                        add_exclude(argv[++i], "", 0, &exclude_list[EXC_CMDL]);
 611                        continue;
 612                }
 613                if (!strncmp(arg, "--exclude=", 10)) {
 614                        exc_given = 1;
 615                        add_exclude(arg+10, "", 0, &exclude_list[EXC_CMDL]);
 616                        continue;
 617                }
 618                if (!strcmp(arg, "-X") && i+1 < argc) {
 619                        exc_given = 1;
 620                        add_excludes_from_file(argv[++i]);
 621                        continue;
 622                }
 623                if (!strncmp(arg, "--exclude-from=", 15)) {
 624                        exc_given = 1;
 625                        add_excludes_from_file(arg+15);
 626                        continue;
 627                }
 628                if (!strncmp(arg, "--exclude-per-directory=", 24)) {
 629                        exc_given = 1;
 630                        exclude_per_dir = arg + 24;
 631                        continue;
 632                }
 633                if (!strcmp(arg, "--full-name")) {
 634                        prefix_offset = 0;
 635                        continue;
 636                }
 637                if (*arg == '-')
 638                        usage(ls_files_usage);
 639                break;
 640        }
 641
 642        pathspec = get_pathspec(prefix, argv + i);
 643
 644        /* Verify that the pathspec matches the prefix */
 645        if (pathspec)
 646                verify_pathspec();
 647
 648        if (show_ignored && !exc_given) {
 649                fprintf(stderr, "%s: --ignored needs some exclude pattern\n",
 650                        argv[0]);
 651                exit(1);
 652        }
 653
 654        /* With no flags, we default to showing the cached files */
 655        if (!(show_stage | show_deleted | show_others | show_unmerged |
 656              show_killed | show_modified))
 657                show_cached = 1;
 658
 659        read_cache();
 660        if (prefix)
 661                prune_cache();
 662        show_files();
 663        return 0;
 664}