src/du.c

   1 /* du -- summarize disk usage
   2    Copyright (C) 1988-1991, 1995-2007 Free Software Foundation, Inc.
   3
   4    This program is free software: you can redistribute it and/or modify
   5    it under the terms of the GNU General Public License as published by
   6    the Free Software Foundation, either version 3 of the License, or
   7    (at your option) any later version.
   8
   9    This program is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12    GNU General Public License for more details.
  13
  14    You should have received a copy of the GNU General Public License
  15    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  16
  17 /* Differences from the Unix du:
  18    * Doesn't simply ignore the names of regular files given as arguments
  19      when -a is given.
  20
  21    By tege@sics.se, Torbjorn Granlund,
  22    and djm@ai.mit.edu, David MacKenzie.
  23    Variable blocks added by lm@sgi.com and eggert@twinsun.com.
  24    Rewritten to use nftw, then to use fts by Jim Meyering.  */
  25
  26 #include <config.h>
  27 #include <stdio.h>
  28 #include <getopt.h>
  29 #include <sys/types.h>
  30 #include <assert.h>
  31 #include "system.h"
  32 #include "argmatch.h"
  33 #include "error.h"
  34 #include "exclude.h"
  35 #include "fprintftime.h"
  36 #include "hash.h"
  37 #include "human.h"
  38 #include "inttostr.h"
  39 #include "quote.h"
  40 #include "quotearg.h"
  41 #include "readtokens0.h"
  42 #include "same.h"
  43 #include "stat-time.h"
  44 #include "xfts.h"
  45 #include "xstrtol.h"
  46
  47 extern bool fts_debug;
  48
  49 /* The official name of this program (e.g., no `g' prefix).  */
  50 #define PROGRAM_NAME "du"
  51
  52 #define AUTHORS \
  53   "Torbjorn Granlund", "David MacKenzie, Paul Eggert", "Jim Meyering"
  54
  55 #if DU_DEBUG
  56 # define FTS_CROSS_CHECK(Fts) fts_cross_check (Fts)
  57 # define DEBUG_OPT "d"
  58 #else
  59 # define FTS_CROSS_CHECK(Fts)
  60 # define DEBUG_OPT
  61 #endif
  62
  63 /* Initial size of the hash table.  */
  64 #define INITIAL_TABLE_SIZE 103
  65
  66 /* Hash structure for inode and device numbers.  The separate entry
  67    structure makes it easier to rehash "in place".  */
  68
  69 struct entry
  70 {
  71   ino_t st_ino;
  72   dev_t st_dev;
  73 };
  74
  75 /* A set of dev/ino pairs.  */
  76 static Hash_table *htab;
  77
  78 /* Define a class for collecting directory information. */
  79
  80 struct duinfo
  81 {
  82   /* Size of files in directory.  */
  83   uintmax_t size;
  84
  85   /* Latest time stamp found.  If tmax.tv_sec == TYPE_MINIMUM (time_t)
  86      && tmax.tv_nsec < 0, no time stamp has been found.  */
  87   struct timespec tmax;
  88 };
  89
  90 /* Initialize directory data.  */
  91 static inline void
  92 duinfo_init (struct duinfo *a)
  93 {
  94   a->size = 0;
  95   a->tmax.tv_sec = TYPE_MINIMUM (time_t);
  96   a->tmax.tv_nsec = -1;
  97 }
  98
  99 /* Set directory data.  */
 100 static inline void
 101 duinfo_set (struct duinfo *a, uintmax_t size, struct timespec tmax)
 102 {
 103   a->size = size;
 104   a->tmax = tmax;
 105 }
 106
 107 /* Accumulate directory data.  */
 108 static inline void
 109 duinfo_add (struct duinfo *a, struct duinfo const *b)
 110 {
 111   a->size += b->size;
 112   if (timespec_cmp (a->tmax, b->tmax) < 0)
 113     a->tmax = b->tmax;
 114 }
 115
 116 /* A structure for per-directory level information.  */
 117 struct dulevel
 118 {
 119   /* Entries in this directory.  */
 120   struct duinfo ent;
 121
 122   /* Total for subdirectories.  */
 123   struct duinfo subdir;
 124 };
 125
 126 /* Name under which this program was invoked.  */
 127 char *program_name;
 128
 129 /* If true, display counts for all files, not just directories.  */
 130 static bool opt_all = false;
 131
 132 /* If true, rather than using the disk usage of each file,
 133    use the apparent size (a la stat.st_size).  */
 134 static bool apparent_size = false;
 135
 136 /* If true, count each hard link of files with multiple links.  */
 137 static bool opt_count_all = false;
 138
 139 /* If true, output the NUL byte instead of a newline at the end of each line. */
 140 static bool opt_nul_terminate_output = false;
 141
 142 /* If true, print a grand total at the end.  */
 143 static bool print_grand_total = false;
 144
 145 /* If nonzero, do not add sizes of subdirectories.  */
 146 static bool opt_separate_dirs = false;
 147
 148 /* Show the total for each directory (and file if --all) that is at
 149    most MAX_DEPTH levels down from the root of the hierarchy.  The root
 150    is at level 0, so `du --max-depth=0' is equivalent to `du -s'.  */
 151 static size_t max_depth = SIZE_MAX;
 152
 153 /* Human-readable options for output.  */
 154 static int human_output_opts;
 155
 156 /* If true, print most recently modified date, using the specified format.  */
 157 static bool opt_time = false;
 158
 159 /* Type of time to display. controlled by --time.  */
 160
 161 enum time_type
 162   {
 163     time_mtime,                 /* default */
 164     time_ctime,
 165     time_atime
 166   };
 167
 168 static enum time_type time_type = time_mtime;
 169
 170 /* User specified date / time style */
 171 static char const *time_style = NULL;
 172
 173 /* Format used to display date / time. Controlled by --time-style */
 174 static char const *time_format = NULL;
 175
 176 /* The units to use when printing sizes.  */
 177 static uintmax_t output_block_size;
 178
 179 /* File name patterns to exclude.  */
 180 static struct exclude *exclude;
 181
 182 /* Grand total size of all args, in bytes. Also latest modified date. */
 183 static struct duinfo tot_dui;
 184
 185 #define IS_DIR_TYPE(Type)       \
 186   ((Type) == FTS_DP             \
 187    || (Type) == FTS_DNR)
 188
 189 /* For long options that have no equivalent short option, use a
 190    non-character as a pseudo short option, starting with CHAR_MAX + 1.  */
 191 enum
 192 {
 193   APPARENT_SIZE_OPTION = CHAR_MAX + 1,
 194   EXCLUDE_OPTION,
 195   FILES0_FROM_OPTION,
 196   HUMAN_SI_OPTION,
 197   MAX_DEPTH_OPTION,
 198   MEGABYTES_LONG_OPTION,
 199   TIME_OPTION,
 200   TIME_STYLE_OPTION
 201 };
 202
 203 static struct option const long_options[] =
 204 {
 205   {"all", no_argument, NULL, 'a'},
 206   {"apparent-size", no_argument, NULL, APPARENT_SIZE_OPTION},
 207   {"block-size", required_argument, NULL, 'B'},
 208   {"bytes", no_argument, NULL, 'b'},
 209   {"count-links", no_argument, NULL, 'l'},
 210   {"dereference", no_argument, NULL, 'L'},
 211   {"dereference-args", no_argument, NULL, 'D'},
 212   {"exclude", required_argument, NULL, EXCLUDE_OPTION},
 213   {"exclude-from", required_argument, NULL, 'X'},
 214   {"files0-from", required_argument, NULL, FILES0_FROM_OPTION},
 215   {"human-readable", no_argument, NULL, 'h'},
 216   {"si", no_argument, NULL, HUMAN_SI_OPTION},
 217   {"max-depth", required_argument, NULL, MAX_DEPTH_OPTION},
 218   {"null", no_argument, NULL, '0'},
 219   {"no-dereference", no_argument, NULL, 'P'},
 220   {"one-file-system", no_argument, NULL, 'x'},
 221   {"separate-dirs", no_argument, NULL, 'S'},
 222   {"summarize", no_argument, NULL, 's'},
 223   {"total", no_argument, NULL, 'c'},
 224   {"time", optional_argument, NULL, TIME_OPTION},
 225   {"time-style", required_argument, NULL, TIME_STYLE_OPTION},
 226   {GETOPT_HELP_OPTION_DECL},
 227   {GETOPT_VERSION_OPTION_DECL},
 228   {NULL, 0, NULL, 0}
 229 };
 230
 231 static char const *const time_args[] =
 232 {
 233   "atime", "access", "use", "ctime", "status", NULL
 234 };
 235 static enum time_type const time_types[] =
 236 {
 237   time_atime, time_atime, time_atime, time_ctime, time_ctime
 238 };
 239 ARGMATCH_VERIFY (time_args, time_types);
 240
 241 /* `full-iso' uses full ISO-style dates and times.  `long-iso' uses longer
 242    ISO-style time stamps, though shorter than `full-iso'.  `iso' uses shorter
 243    ISO-style time stamps.  */
 244 enum time_style
 245   {
 246     full_iso_time_style,       /* --time-style=full-iso */
 247     long_iso_time_style,       /* --time-style=long-iso */
 248     iso_time_style             /* --time-style=iso */
 249   };
 250
 251 static char const *const time_style_args[] =
 252 {
 253   "full-iso", "long-iso", "iso", NULL
 254 };
 255 static enum time_style const time_style_types[] =
 256 {
 257   full_iso_time_style, long_iso_time_style, iso_time_style
 258 };
 259 ARGMATCH_VERIFY (time_style_args, time_style_types);
 260
 261 void
 262 usage (int status)
 263 {
 264   if (status != EXIT_SUCCESS)
 265     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 266              program_name);
 267   else
 268     {
 269       printf (_("\
 270 Usage: %s [OPTION]... [FILE]...\n\
 271   or:  %s [OPTION]... --files0-from=F\n\
 272 "), program_name, program_name);
 273       fputs (_("\
 274 Summarize disk usage of each FILE, recursively for directories.\n\
 275 \n\
 276 "), stdout);
 277       fputs (_("\
 278 Mandatory arguments to long options are mandatory for short options too.\n\
 279 "), stdout);
 280       fputs (_("\
 281   -a, --all             write counts for all files, not just directories\n\
 282       --apparent-size   print apparent sizes, rather than disk usage; although\n\
 283                           the apparent size is usually smaller, it may be\n\
 284                           larger due to holes in (`sparse') files, internal\n\
 285                           fragmentation, indirect blocks, and the like\n\
 286 "), stdout);
 287       fputs (_("\
 288   -B, --block-size=SIZE  use SIZE-byte blocks\n\
 289   -b, --bytes           equivalent to `--apparent-size --block-size=1'\n\
 290   -c, --total           produce a grand total\n\
 291   -D, --dereference-args  dereference only symlinks that are listed on the\n\
 292                           command line\n\
 293 "), stdout);
 294       fputs (_("\
 295       --files0-from=F   summarize disk usage of the NUL-terminated file\n\
 296                           names specified in file F\n\
 297   -H                    like --si, but also evokes a warning; will soon\n\
 298                           change to be equivalent to --dereference-args (-D)\n\
 299   -h, --human-readable  print sizes in human readable format (e.g., 1K 234M 2G)\n\
 300       --si              like -h, but use powers of 1000 not 1024\n\
 301 "), stdout);
 302       fputs (_("\
 303   -k                    like --block-size=1K\n\
 304   -l, --count-links     count sizes many times if hard linked\n\
 305   -m                    like --block-size=1M\n\
 306 "), stdout);
 307       fputs (_("\
 308   -L, --dereference     dereference all symbolic links\n\
 309   -P, --no-dereference  don't follow any symbolic links (this is the default)\n\
 310   -0, --null            end each output line with 0 byte rather than newline\n\
 311   -S, --separate-dirs   do not include size of subdirectories\n\
 312   -s, --summarize       display only a total for each argument\n\
 313 "), stdout);
 314       fputs (_("\
 315   -x, --one-file-system  skip directories on different file systems\n\
 316   -X FILE, --exclude-from=FILE  Exclude files that match any pattern in FILE.\n\
 317       --exclude=PATTERN  Exclude files that match PATTERN.\n\
 318       --max-depth=N     print the total for a directory (or file, with --all)\n\
 319                           only if it is N or fewer levels below the command\n\
 320                           line argument;  --max-depth=0 is the same as\n\
 321                           --summarize\n\
 322 "), stdout);
 323       fputs (_("\
 324       --time            show time of the last modification of any file in the\n\
 325                           directory, or any of its subdirectories\n\
 326       --time=WORD       show time as WORD instead of modification time:\n\
 327                           atime, access, use, ctime or status\n\
 328       --time-style=STYLE  show times using style STYLE:\n\
 329                           full-iso, long-iso, iso, +FORMAT\n\
 330                           FORMAT is interpreted like `date'\n\
 331 "), stdout);
 332       fputs (HELP_OPTION_DESCRIPTION, stdout);
 333       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 334       fputs (_("\n\
 335 SIZE may be (or may be an integer optionally followed by) one of following:\n\
 336 kB 1000, K 1024, MB 1000*1000, M 1024*1024, and so on for G, T, P, E, Z, Y.\n\
 337 "), stdout);
 338       emit_bug_reporting_address ();
 339     }
 340   exit (status);
 341 }
 342
 343 static size_t
 344 entry_hash (void const *x, size_t table_size)
 345 {
 346   struct entry const *p = x;
 347
 348   /* Ignoring the device number here should be fine.  */
 349   /* The cast to uintmax_t prevents negative remainders
 350      if st_ino is negative.  */
 351   return (uintmax_t) p->st_ino % table_size;
 352 }
 353
 354 /* Compare two dev/ino pairs.  Return true if they are the same.  */
 355 static bool
 356 entry_compare (void const *x, void const *y)
 357 {
 358   struct entry const *a = x;
 359   struct entry const *b = y;
 360   return SAME_INODE (*a, *b) ? true : false;
 361 }
 362
 363 /* Try to insert the INO/DEV pair into the global table, HTAB.
 364    Return true if the pair is successfully inserted,
 365    false if the pair is already in the table.  */
 366 static bool
 367 hash_ins (ino_t ino, dev_t dev)
 368 {
 369   struct entry *ent;
 370   struct entry *ent_from_table;
 371
 372   ent = xmalloc (sizeof *ent);
 373   ent->st_ino = ino;
 374   ent->st_dev = dev;
 375
 376   ent_from_table = hash_insert (htab, ent);
 377   if (ent_from_table == NULL)
 378     {
 379       /* Insertion failed due to lack of memory.  */
 380       xalloc_die ();
 381     }
 382
 383   if (ent_from_table == ent)
 384     {
 385       /* Insertion succeeded.  */
 386       return true;
 387     }
 388
 389   /* That pair is already in the table, so ENT was not inserted.  Free it.  */
 390   free (ent);
 391
 392   return false;
 393 }
 394
 395 /* Initialize the hash table.  */
 396 static void
 397 hash_init (void)
 398 {
 399   htab = hash_initialize (INITIAL_TABLE_SIZE, NULL,
 400                           entry_hash, entry_compare, free);
 401   if (htab == NULL)
 402     xalloc_die ();
 403 }
 404
 405 /* FIXME: this code is nearly identical to code in date.c  */
 406 /* Display the date and time in WHEN according to the format specified
 407    in FORMAT.  */
 408
 409 static void
 410 show_date (const char *format, struct timespec when)
 411 {
 412   struct tm *tm = localtime (&when.tv_sec);
 413   if (! tm)
 414     {
 415       char buf[INT_BUFSIZE_BOUND (intmax_t)];
 416       error (0, 0, _("time %s is out of range"),
 417              (TYPE_SIGNED (time_t)
 418               ? imaxtostr (when.tv_sec, buf)
 419               : umaxtostr (when.tv_sec, buf)));
 420       fputs (buf, stdout);
 421       return;
 422     }
 423
 424   fprintftime (stdout, format, tm, 0, when.tv_nsec);
 425 }
 426
 427 /* Print N_BYTES.  Convert it to a readable value before printing.  */
 428
 429 static void
 430 print_only_size (uintmax_t n_bytes)
 431 {
 432   char buf[LONGEST_HUMAN_READABLE + 1];
 433   fputs (human_readable (n_bytes, buf, human_output_opts,
 434                          1, output_block_size), stdout);
 435 }
 436
 437 /* Print size (and optionally time) indicated by *PDUI, followed by STRING.  */
 438
 439 static void
 440 print_size (const struct duinfo *pdui, const char *string)
 441 {
 442   print_only_size (pdui->size);
 443   if (opt_time)
 444     {
 445       putchar ('\t');
 446       show_date (time_format, pdui->tmax);
 447     }
 448   printf ("\t%s%c", string, opt_nul_terminate_output ? '\0' : '\n');
 449   fflush (stdout);
 450 }
 451
 452 /* This function is called once for every file system object that fts
 453    encounters.  fts does a depth-first traversal.  This function knows
 454    that and accumulates per-directory totals based on changes in
 455    the depth of the current entry.  It returns true on success.  */
 456
 457 static bool
 458 process_file (FTS *fts, FTSENT *ent)
 459 {
 460   bool ok;
 461   struct duinfo dui;
 462   struct duinfo dui_to_print;
 463   size_t level;
 464   static size_t prev_level;
 465   static size_t n_alloc;
 466   /* First element of the structure contains:
 467      The sum of the st_size values of all entries in the single directory
 468      at the corresponding level.  Although this does include the st_size
 469      corresponding to each subdirectory, it does not include the size of
 470      any file in a subdirectory. Also corresponding last modified date.
 471      Second element of the structure contains:
 472      The sum of the sizes of all entries in the hierarchy at or below the
 473      directory at the specified level.  */
 474   static struct dulevel *dulvl;
 475   bool print = true;
 476
 477   const char *file = ent->fts_path;
 478   const struct stat *sb = ent->fts_statp;
 479   bool skip;
 480
 481   /* If necessary, set FTS_SKIP before returning.  */
 482   skip = excluded_file_name (exclude, ent->fts_path);
 483   if (skip)
 484     fts_set (fts, ent, FTS_SKIP);
 485
 486   switch (ent->fts_info)
 487     {
 488     case FTS_NS:
 489       error (0, ent->fts_errno, _("cannot access %s"), quote (file));
 490       return false;
 491
 492     case FTS_ERR:
 493       /* if (S_ISDIR (ent->fts_statp->st_mode) && FIXME */
 494       error (0, ent->fts_errno, _("%s"), quote (file));
 495       return false;
 496
 497     case FTS_DNR:
 498       /* Don't return just yet, since although the directory is not readable,
 499          we were able to stat it, so we do have a size.  */
 500       error (0, ent->fts_errno, _("cannot read directory %s"), quote (file));
 501       ok = false;
 502       break;
 503
 504     default:
 505       ok = true;
 506       break;
 507     }
 508
 509   /* If this is the first (pre-order) encounter with a directory,
 510      or if it's the second encounter for a skipped directory, then
 511      return right away.  */
 512   if (ent->fts_info == FTS_D || skip)
 513     return ok;
 514
 515   /* If the file is being excluded or if it has already been counted
 516      via a hard link, then don't let it contribute to the sums.  */
 517   if (skip
 518       || (!opt_count_all
 519           && ! S_ISDIR (sb->st_mode)
 520           && 1 < sb->st_nlink
 521           && ! hash_ins (sb->st_ino, sb->st_dev)))
 522     {
 523       /* Note that we must not simply return here.
 524          We still have to update prev_level and maybe propagate
 525          some sums up the hierarchy.  */
 526       duinfo_init (&dui);
 527       print = false;
 528     }
 529   else
 530     {
 531       duinfo_set (&dui,
 532                   (apparent_size
 533                    ? sb->st_size
 534                    : (uintmax_t) ST_NBLOCKS (*sb) * ST_NBLOCKSIZE),
 535                   (time_type == time_mtime ? get_stat_mtime (sb)
 536                    : time_type == time_atime ? get_stat_atime (sb)
 537                    : get_stat_ctime (sb)));
 538     }
 539
 540   level = ent->fts_level;
 541   dui_to_print = dui;
 542
 543   if (n_alloc == 0)
 544     {
 545       n_alloc = level + 10;
 546       dulvl = xcalloc (n_alloc, sizeof *dulvl);
 547     }
 548   else
 549     {
 550       if (level == prev_level)
 551         {
 552           /* This is usually the most common case.  Do nothing.  */
 553         }
 554       else if (level > prev_level)
 555         {
 556           /* Descending the hierarchy.
 557              Clear the accumulators for *all* levels between prev_level
 558              and the current one.  The depth may change dramatically,
 559              e.g., from 1 to 10.  */
 560           size_t i;
 561
 562           if (n_alloc <= level)
 563             {
 564               dulvl = xnrealloc (dulvl, level, 2 * sizeof *dulvl);
 565               n_alloc = level * 2;
 566             }
 567
 568           for (i = prev_level + 1; i <= level; i++)
 569             {
 570               duinfo_init (&dulvl[i].ent);
 571               duinfo_init (&dulvl[i].subdir);
 572             }
 573         }
 574       else /* level < prev_level */
 575         {
 576           /* Ascending the hierarchy.
 577              Process a directory only after all entries in that
 578              directory have been processed.  When the depth decreases,
 579              propagate sums from the children (prev_level) to the parent.
 580              Here, the current level is always one smaller than the
 581              previous one.  */
 582           assert (level == prev_level - 1);
 583           duinfo_add (&dui_to_print, &dulvl[prev_level].ent);
 584           if (!opt_separate_dirs)
 585             duinfo_add (&dui_to_print, &dulvl[prev_level].subdir);
 586           duinfo_add (&dulvl[level].subdir, &dulvl[prev_level].ent);
 587           duinfo_add (&dulvl[level].subdir, &dulvl[prev_level].subdir);
 588         }
 589     }
 590
 591   prev_level = level;
 592
 593   /* Let the size of a directory entry contribute to the total for the
 594      containing directory, unless --separate-dirs (-S) is specified.  */
 595   if ( ! (opt_separate_dirs && IS_DIR_TYPE (ent->fts_info)))
 596     duinfo_add (&dulvl[level].ent, &dui);
 597
 598   /* Even if this directory is unreadable or we can't chdir into it,
 599      do let its size contribute to the total. */
 600   duinfo_add (&tot_dui, &dui);
 601
 602   /* If we're not counting an entry, e.g., because it's a hard link
 603      to a file we've already counted (and --count-links), then don't
 604      print a line for it.  */
 605   if (!print)
 606     return ok;
 607
 608   if ((IS_DIR_TYPE (ent->fts_info) && level <= max_depth)
 609       || ((opt_all && level <= max_depth) || level == 0))
 610     print_size (&dui_to_print, file);
 611
 612   return ok;
 613 }
 614
 615 /* Recursively print the sizes of the directories (and, if selected, files)
 616    named in FILES, the last entry of which is NULL.
 617    BIT_FLAGS controls how fts works.
 618    Return true if successful.  */
 619
 620 static bool
 621 du_files (char **files, int bit_flags)
 622 {
 623   bool ok = true;
 624
 625   if (*files)
 626     {
 627       FTS *fts = xfts_open (files, bit_flags, NULL);
 628
 629       while (1)
 630         {
 631           FTSENT *ent;
 632
 633           ent = fts_read (fts);
 634           if (ent == NULL)
 635             {
 636               if (errno != 0)
 637                 {
 638                   /* FIXME: try to give a better message  */
 639                   error (0, errno, _("fts_read failed"));
 640                   ok = false;
 641                 }
 642               break;
 643             }
 644           FTS_CROSS_CHECK (fts);
 645
 646           ok &= process_file (fts, ent);
 647         }
 648
 649       /* Ignore failure, since the only way it can do so is in failing to
 650          return to the original directory, and since we're about to exit,
 651          that doesn't matter.  */
 652       fts_close (fts);
 653     }
 654
 655   if (print_grand_total)
 656     print_size (&tot_dui, _("total"));
 657
 658   return ok;
 659 }
 660
 661 int
 662 main (int argc, char **argv)
 663 {
 664   char *cwd_only[2];
 665   bool max_depth_specified = false;
 666   char **files;
 667   bool ok = true;
 668   char *files_from = NULL;
 669   struct Tokens tok;
 670
 671   /* Bit flags that control how fts works.  */
 672   int bit_flags = FTS_TIGHT_CYCLE_CHECK;
 673
 674   /* Select one of the three FTS_ options that control if/when
 675      to follow a symlink.  */
 676   int symlink_deref_bits = FTS_PHYSICAL;
 677
 678   /* If true, display only a total for each argument. */
 679   bool opt_summarize_only = false;
 680
 681   cwd_only[0] = ".";
 682   cwd_only[1] = NULL;
 683
 684   initialize_main (&argc, &argv);
 685   program_name = argv[0];
 686   setlocale (LC_ALL, "");
 687   bindtextdomain (PACKAGE, LOCALEDIR);
 688   textdomain (PACKAGE);
 689
 690   atexit (close_stdout);
 691
 692   exclude = new_exclude ();
 693
 694   human_options (getenv ("DU_BLOCK_SIZE"),
 695                  &human_output_opts, &output_block_size);
 696
 697   for (;;)
 698     {
 699       int oi = -1;
 700       int c = getopt_long (argc, argv, DEBUG_OPT "0abchHklmsxB:DLPSX:",
 701                            long_options, &oi);
 702       if (c == -1)
 703         break;
 704
 705       switch (c)
 706         {
 707 #if DU_DEBUG
 708         case 'd':
 709           fts_debug = true;
 710           break;
 711 #endif
 712
 713         case '0':
 714           opt_nul_terminate_output = true;
 715           break;
 716
 717         case 'a':
 718           opt_all = true;
 719           break;
 720
 721         case APPARENT_SIZE_OPTION:
 722           apparent_size = true;
 723           break;
 724
 725         case 'b':
 726           apparent_size = true;
 727           human_output_opts = 0;
 728           output_block_size = 1;
 729           break;
 730
 731         case 'c':
 732           print_grand_total = true;
 733           break;
 734
 735         case 'h':
 736           human_output_opts = human_autoscale | human_SI | human_base_1024;
 737           output_block_size = 1;
 738           break;
 739
 740         case 'H':  /* FIXME: remove warning and move this "case 'H'" to
 741                       precede --dereference-args in late 2006.  */
 742           error (0, 0, _("WARNING: use --si, not -H; the meaning of the -H\
 743  option will soon\nchange to be the same as that of --dereference-args (-D)"));
 744           /* fall through */
 745         case HUMAN_SI_OPTION:
 746           human_output_opts = human_autoscale | human_SI;
 747           output_block_size = 1;
 748           break;
 749
 750         case 'k':
 751           human_output_opts = 0;
 752           output_block_size = 1024;
 753           break;
 754
 755         case MAX_DEPTH_OPTION:          /* --max-depth=N */
 756           {
 757             unsigned long int tmp_ulong;
 758             if (xstrtoul (optarg, NULL, 0, &tmp_ulong, NULL) == LONGINT_OK
 759                 && tmp_ulong <= SIZE_MAX)
 760               {
 761                 max_depth_specified = true;
 762                 max_depth = tmp_ulong;
 763               }
 764             else
 765               {
 766                 error (0, 0, _("invalid maximum depth %s"),
 767                        quote (optarg));
 768                 ok = false;
 769               }
 770           }
 771           break;
 772
 773         case MEGABYTES_LONG_OPTION:
 774           error (0, 0,
 775                  _("the --megabytes option is deprecated; use -m instead"));
 776           /* fall through */
 777         case 'm':
 778           human_output_opts = 0;
 779           output_block_size = 1024 * 1024;
 780           break;
 781
 782         case 'l':
 783           opt_count_all = true;
 784           break;
 785
 786         case 's':
 787           opt_summarize_only = true;
 788           break;
 789
 790         case 'x':
 791           bit_flags |= FTS_XDEV;
 792           break;
 793
 794         case 'B':
 795           {
 796             enum strtol_error e = human_options (optarg, &human_output_opts,
 797                                                  &output_block_size);
 798             if (e != LONGINT_OK)
 799               xstrtol_fatal (e, oi, c, long_options, optarg);
 800           }
 801           break;
 802
 803         case 'D': /* This will eventually be 'H' (-H), too.  */
 804           symlink_deref_bits = FTS_COMFOLLOW | FTS_PHYSICAL;
 805           break;
 806
 807         case 'L': /* --dereference */
 808           symlink_deref_bits = FTS_LOGICAL;
 809           break;
 810
 811         case 'P': /* --no-dereference */
 812           symlink_deref_bits = FTS_PHYSICAL;
 813           break;
 814
 815         case 'S':
 816           opt_separate_dirs = true;
 817           break;
 818
 819         case 'X':
 820           if (add_exclude_file (add_exclude, exclude, optarg,
 821                                 EXCLUDE_WILDCARDS, '\n'))
 822             {
 823               error (0, errno, "%s", quotearg_colon (optarg));
 824               ok = false;
 825             }
 826           break;
 827
 828         case FILES0_FROM_OPTION:
 829           files_from = optarg;
 830           break;
 831
 832         case EXCLUDE_OPTION:
 833           add_exclude (exclude, optarg, EXCLUDE_WILDCARDS);
 834           break;
 835
 836         case TIME_OPTION:
 837           opt_time = true;
 838           time_type =
 839             (optarg
 840              ? XARGMATCH ("--time", optarg, time_args, time_types)
 841              : time_mtime);
 842           break;
 843
 844         case TIME_STYLE_OPTION:
 845           time_style = optarg;
 846           break;
 847
 848         case_GETOPT_HELP_CHAR;
 849
 850         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 851
 852         default:
 853           ok = false;
 854         }
 855     }
 856
 857   if (!ok)
 858     usage (EXIT_FAILURE);
 859
 860   if (opt_all & opt_summarize_only)
 861     {
 862       error (0, 0, _("cannot both summarize and show all entries"));
 863       usage (EXIT_FAILURE);
 864     }
 865
 866   if (opt_summarize_only && max_depth_specified && max_depth == 0)
 867     {
 868       error (0, 0,
 869              _("warning: summarizing is the same as using --max-depth=0"));
 870     }
 871
 872   if (opt_summarize_only && max_depth_specified && max_depth != 0)
 873     {
 874       unsigned long int d = max_depth;
 875       error (0, 0, _("warning: summarizing conflicts with --max-depth=%lu"), d);
 876       usage (EXIT_FAILURE);
 877     }
 878
 879   if (opt_summarize_only)
 880     max_depth = 0;
 881
 882   /* Process time style if printing last times.  */
 883   if (opt_time)
 884     {
 885       if (! time_style)
 886         {
 887           time_style = getenv ("TIME_STYLE");
 888
 889           /* Ignore TIMESTYLE="locale", for compatibility with ls.  */
 890           if (! time_style || STREQ (time_style, "locale"))
 891             time_style = "long-iso";
 892           else if (*time_style == '+')
 893             {
 894               /* Ignore anything after a newline, for compatibility
 895                  with ls.  */
 896               char *p = strchr (time_style, '\n');
 897               if (p)
 898                 *p = '\0';
 899             }
 900           else
 901             {
 902               /* Ignore "posix-" prefix, for compatibility with ls.  */
 903               static char const posix_prefix[] = "posix-";
 904               while (strncmp (time_style, posix_prefix, sizeof posix_prefix - 1)
 905                      == 0)
 906                 time_style += sizeof posix_prefix - 1;
 907             }
 908         }
 909
 910       if (*time_style == '+')
 911         time_format = time_style + 1;
 912       else
 913         {
 914           switch (XARGMATCH ("time style", time_style,
 915                              time_style_args, time_style_types))
 916             {
 917             case full_iso_time_style:
 918               time_format = "%Y-%m-%d %H:%M:%S.%N %z";
 919               break;
 920
 921             case long_iso_time_style:
 922               time_format = "%Y-%m-%d %H:%M";
 923               break;
 924
 925             case iso_time_style:
 926               time_format = "%Y-%m-%d";
 927               break;
 928             }
 929         }
 930     }
 931
 932   if (files_from)
 933     {
 934       /* When using --files0-from=F, you may not specify any files
 935          on the command-line.  */
 936       if (optind < argc)
 937         {
 938           error (0, 0, _("extra operand %s"), quote (argv[optind]));
 939           fprintf (stderr, "%s\n",
 940                    _("File operands cannot be combined with --files0-from."));
 941           usage (EXIT_FAILURE);
 942         }
 943
 944       if (! (STREQ (files_from, "-") || freopen (files_from, "r", stdin)))
 945         error (EXIT_FAILURE, errno, _("cannot open %s for reading"),
 946                quote (files_from));
 947
 948       readtokens0_init (&tok);
 949
 950       if (! readtokens0 (stdin, &tok) || fclose (stdin) != 0)
 951         error (EXIT_FAILURE, 0, _("cannot read file names from %s"),
 952                quote (files_from));
 953
 954       files = tok.tok;
 955     }
 956   else
 957     {
 958       files = (optind < argc ? argv + optind : cwd_only);
 959     }
 960
 961   /* Initialize the hash structure for inode numbers.  */
 962   hash_init ();
 963
 964   /* Report and filter out any empty file names before invoking fts.
 965      This works around a glitch in fts, which fails immediately
 966      (without looking at the other file names) when given an empty
 967      file name.  */
 968   {
 969     size_t i = 0;
 970     size_t j;
 971
 972     for (j = 0; ; j++)
 973       {
 974         if (i != j)
 975           files[i] = files[j];
 976
 977         if ( ! files[i])
 978           break;
 979
 980         if (files[i][0])
 981           i++;
 982         else
 983           {
 984             if (files_from)
 985               {
 986                 /* Using the standard `filename:line-number:' prefix here is
 987                    not totally appropriate, since NUL is the separator, not NL,
 988                    but it might be better than nothing.  */
 989                 unsigned long int file_number = j + 1;
 990                 error (0, 0, "%s:%lu: %s", quotearg_colon (files_from),
 991                        file_number, _("invalid zero-length file name"));
 992               }
 993             else
 994               error (0, 0, "%s", _("invalid zero-length file name"));
 995           }
 996       }
 997
 998     ok = (i == j);
 999   }
1000
1001   bit_flags |= symlink_deref_bits;
1002   ok &= du_files (files, bit_flags);
1003
1004   /* This isn't really necessary, but it does ensure we
1005      exercise this function.  */
1006   if (files_from)
1007     readtokens0_free (&tok);
1008
1009   hash_free (htab);
1010
1011   exit (ok ? EXIT_SUCCESS : EXIT_FAILURE);
1012 }