src/du.c

   1 /* du -- summarize disk usage
   2    Copyright (C) 1988-1991, 1995-2009 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 "argv-iter.h"
  34 #include "error.h"
  35 #include "exclude.h"
  36 #include "fprintftime.h"
  37 #include "hash.h"
  38 #include "human.h"
  39 #include "quote.h"
  40 #include "quotearg.h"
  41 #include "same.h"
  42 #include "stat-time.h"
  43 #include "xfts.h"
  44 #include "xstrtol.h"
  45
  46 extern bool fts_debug;
  47
  48 /* The official name of this program (e.g., no `g' prefix).  */
  49 #define PROGRAM_NAME "du"
  50
  51 #define AUTHORS \
  52   proper_name_utf8 ("Torbjorn Granlund", "Torbj\303\266rn Granlund"), \
  53   proper_name ("David MacKenzie"), \
  54   proper_name ("Paul Eggert"), \
  55   proper_name ("Jim Meyering")
  56
  57 #if DU_DEBUG
  58 # define FTS_CROSS_CHECK(Fts) fts_cross_check (Fts)
  59 # define DEBUG_OPT "d"
  60 #else
  61 # define FTS_CROSS_CHECK(Fts)
  62 # define DEBUG_OPT
  63 #endif
  64
  65 /* Initial size of the hash table.  */
  66 #define INITIAL_TABLE_SIZE 103
  67
  68 /* Hash structure for inode and device numbers.  The separate entry
  69    structure makes it easier to rehash "in place".  */
  70
  71 struct entry
  72 {
  73   ino_t st_ino;
  74   dev_t st_dev;
  75 };
  76
  77 /* A set of dev/ino pairs.  */
  78 static Hash_table *htab;
  79
  80 /* Define a class for collecting directory information. */
  81
  82 struct duinfo
  83 {
  84   /* Size of files in directory.  */
  85   uintmax_t size;
  86
  87   /* Latest time stamp found.  If tmax.tv_sec == TYPE_MINIMUM (time_t)
  88      && tmax.tv_nsec < 0, no time stamp has been found.  */
  89   struct timespec tmax;
  90 };
  91
  92 /* Initialize directory data.  */
  93 static inline void
  94 duinfo_init (struct duinfo *a)
  95 {
  96   a->size = 0;
  97   a->tmax.tv_sec = TYPE_MINIMUM (time_t);
  98   a->tmax.tv_nsec = -1;
  99 }
 100
 101 /* Set directory data.  */
 102 static inline void
 103 duinfo_set (struct duinfo *a, uintmax_t size, struct timespec tmax)
 104 {
 105   a->size = size;
 106   a->tmax = tmax;
 107 }
 108
 109 /* Accumulate directory data.  */
 110 static inline void
 111 duinfo_add (struct duinfo *a, struct duinfo const *b)
 112 {
 113   a->size += b->size;
 114   if (timespec_cmp (a->tmax, b->tmax) < 0)
 115     a->tmax = b->tmax;
 116 }
 117
 118 /* A structure for per-directory level information.  */
 119 struct dulevel
 120 {
 121   /* Entries in this directory.  */
 122   struct duinfo ent;
 123
 124   /* Total for subdirectories.  */
 125   struct duinfo subdir;
 126 };
 127
 128 /* If true, display counts for all files, not just directories.  */
 129 static bool opt_all = false;
 130
 131 /* If true, rather than using the disk usage of each file,
 132    use the apparent size (a la stat.st_size).  */
 133 static bool apparent_size = false;
 134
 135 /* If true, count each hard link of files with multiple links.  */
 136 static bool opt_count_all = false;
 137
 138 /* If true, output the NUL byte instead of a newline at the end of each line. */
 139 static bool opt_nul_terminate_output = false;
 140
 141 /* If true, print a grand total at the end.  */
 142 static bool print_grand_total = false;
 143
 144 /* If nonzero, do not add sizes of subdirectories.  */
 145 static bool opt_separate_dirs = false;
 146
 147 /* Show the total for each directory (and file if --all) that is at
 148    most MAX_DEPTH levels down from the root of the hierarchy.  The root
 149    is at level 0, so `du --max-depth=0' is equivalent to `du -s'.  */
 150 static size_t max_depth = SIZE_MAX;
 151
 152 /* Human-readable options for output.  */
 153 static int human_output_opts;
 154
 155 /* If true, print most recently modified date, using the specified format.  */
 156 static bool opt_time = false;
 157
 158 /* Type of time to display. controlled by --time.  */
 159
 160 enum time_type
 161   {
 162     time_mtime,                 /* default */
 163     time_ctime,
 164     time_atime
 165   };
 166
 167 static enum time_type time_type = time_mtime;
 168
 169 /* User specified date / time style */
 170 static char const *time_style = NULL;
 171
 172 /* Format used to display date / time. Controlled by --time-style */
 173 static char const *time_format = NULL;
 174
 175 /* The units to use when printing sizes.  */
 176 static uintmax_t output_block_size;
 177
 178 /* File name patterns to exclude.  */
 179 static struct exclude *exclude;
 180
 181 /* Grand total size of all args, in bytes. Also latest modified date. */
 182 static struct duinfo tot_dui;
 183
 184 #define IS_DIR_TYPE(Type)       \
 185   ((Type) == FTS_DP             \
 186    || (Type) == FTS_DNR)
 187
 188 /* For long options that have no equivalent short option, use a
 189    non-character as a pseudo short option, starting with CHAR_MAX + 1.  */
 190 enum
 191 {
 192   APPARENT_SIZE_OPTION = CHAR_MAX + 1,
 193   EXCLUDE_OPTION,
 194   FILES0_FROM_OPTION,
 195   HUMAN_SI_OPTION,
 196   MAX_DEPTH_OPTION,
 197   MEGABYTES_LONG_OPTION,
 198   TIME_OPTION,
 199   TIME_STYLE_OPTION
 200 };
 201
 202 static struct option const long_options[] =
 203 {
 204   {"all", no_argument, NULL, 'a'},
 205   {"apparent-size", no_argument, NULL, APPARENT_SIZE_OPTION},
 206   {"block-size", required_argument, NULL, 'B'},
 207   {"bytes", no_argument, NULL, 'b'},
 208   {"count-links", no_argument, NULL, 'l'},
 209   {"dereference", no_argument, NULL, 'L'},
 210   {"dereference-args", no_argument, NULL, 'D'},
 211   {"exclude", required_argument, NULL, EXCLUDE_OPTION},
 212   {"exclude-from", required_argument, NULL, 'X'},
 213   {"files0-from", required_argument, NULL, FILES0_FROM_OPTION},
 214   {"human-readable", no_argument, NULL, 'h'},
 215   {"si", no_argument, NULL, HUMAN_SI_OPTION},
 216   {"max-depth", required_argument, NULL, MAX_DEPTH_OPTION},
 217   {"null", no_argument, NULL, '0'},
 218   {"no-dereference", no_argument, NULL, 'P'},
 219   {"one-file-system", no_argument, NULL, 'x'},
 220   {"separate-dirs", no_argument, NULL, 'S'},
 221   {"summarize", no_argument, NULL, 's'},
 222   {"total", no_argument, NULL, 'c'},
 223   {"time", optional_argument, NULL, TIME_OPTION},
 224   {"time-style", required_argument, NULL, TIME_STYLE_OPTION},
 225   {GETOPT_HELP_OPTION_DECL},
 226   {GETOPT_VERSION_OPTION_DECL},
 227   {NULL, 0, NULL, 0}
 228 };
 229
 230 static char const *const time_args[] =
 231 {
 232   "atime", "access", "use", "ctime", "status", NULL
 233 };
 234 static enum time_type const time_types[] =
 235 {
 236   time_atime, time_atime, time_atime, time_ctime, time_ctime
 237 };
 238 ARGMATCH_VERIFY (time_args, time_types);
 239
 240 /* `full-iso' uses full ISO-style dates and times.  `long-iso' uses longer
 241    ISO-style time stamps, though shorter than `full-iso'.  `iso' uses shorter
 242    ISO-style time stamps.  */
 243 enum time_style
 244   {
 245     full_iso_time_style,       /* --time-style=full-iso */
 246     long_iso_time_style,       /* --time-style=long-iso */
 247     iso_time_style             /* --time-style=iso */
 248   };
 249
 250 static char const *const time_style_args[] =
 251 {
 252   "full-iso", "long-iso", "iso", NULL
 253 };
 254 static enum time_style const time_style_types[] =
 255 {
 256   full_iso_time_style, long_iso_time_style, iso_time_style
 257 };
 258 ARGMATCH_VERIFY (time_style_args, time_style_types);
 259
 260 void
 261 usage (int status)
 262 {
 263   if (status != EXIT_SUCCESS)
 264     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 265              program_name);
 266   else
 267     {
 268       printf (_("\
 269 Usage: %s [OPTION]... [FILE]...\n\
 270   or:  %s [OPTION]... --files0-from=F\n\
 271 "), program_name, program_name);
 272       fputs (_("\
 273 Summarize disk usage of each FILE, recursively for directories.\n\
 274 \n\
 275 "), stdout);
 276       fputs (_("\
 277 Mandatory arguments to long options are mandatory for short options too.\n\
 278 "), stdout);
 279       fputs (_("\
 280   -a, --all             write counts for all files, not just directories\n\
 281       --apparent-size   print apparent sizes, rather than disk usage; although\n\
 282                           the apparent size is usually smaller, it may be\n\
 283                           larger due to holes in (`sparse') files, internal\n\
 284                           fragmentation, indirect blocks, and the like\n\
 285 "), stdout);
 286       fputs (_("\
 287   -B, --block-size=SIZE  use SIZE-byte blocks\n\
 288   -b, --bytes           equivalent to `--apparent-size --block-size=1'\n\
 289   -c, --total           produce a grand total\n\
 290   -D, --dereference-args  dereference only symlinks that are listed on the\n\
 291                           command line\n\
 292 "), stdout);
 293       fputs (_("\
 294       --files0-from=F   summarize disk usage of the NUL-terminated file\n\
 295                           names specified in file F;\n\
 296                           If F is - then read names from standard input\n\
 297   -H                    equivalent to --dereference-args (-D)\n\
 298   -h, --human-readable  print sizes in human readable format (e.g., 1K 234M 2G)\n\
 299       --si              like -h, but use powers of 1000 not 1024\n\
 300 "), stdout);
 301       fputs (_("\
 302   -k                    like --block-size=1K\n\
 303   -l, --count-links     count sizes many times if hard linked\n\
 304   -m                    like --block-size=1M\n\
 305 "), stdout);
 306       fputs (_("\
 307   -L, --dereference     dereference all symbolic links\n\
 308   -P, --no-dereference  don't follow any symbolic links (this is the default)\n\
 309   -0, --null            end each output line with 0 byte rather than newline\n\
 310   -S, --separate-dirs   do not include size of subdirectories\n\
 311   -s, --summarize       display only a total for each argument\n\
 312 "), stdout);
 313       fputs (_("\
 314   -x, --one-file-system    skip directories on different file systems\n\
 315   -X, --exclude-from=FILE  exclude files that match any pattern in FILE\n\
 316       --exclude=PATTERN    exclude files that match PATTERN\n\
 317       --max-depth=N     print the total for a directory (or file, with --all)\n\
 318                           only if it is N or fewer levels below the command\n\
 319                           line argument;  --max-depth=0 is the same as\n\
 320                           --summarize\n\
 321 "), stdout);
 322       fputs (_("\
 323       --time            show time of the last modification of any file in the\n\
 324                           directory, or any of its subdirectories\n\
 325       --time=WORD       show time as WORD instead of modification time:\n\
 326                           atime, access, use, ctime or status\n\
 327       --time-style=STYLE  show times using style STYLE:\n\
 328                           full-iso, long-iso, iso, +FORMAT\n\
 329                           FORMAT is interpreted like `date'\n\
 330 "), stdout);
 331       fputs (HELP_OPTION_DESCRIPTION, stdout);
 332       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 333       fputs (_("\n\
 334 SIZE may be (or may be an integer optionally followed by) one of following:\n\
 335 kB 1000, K 1024, MB 1000*1000, M 1024*1024, and so on for G, T, P, E, Z, Y.\n\
 336 "), stdout);
 337       emit_bug_reporting_address ();
 338     }
 339   exit (status);
 340 }
 341
 342 static size_t
 343 entry_hash (void const *x, size_t table_size)
 344 {
 345   struct entry const *p = x;
 346
 347   /* Ignoring the device number here should be fine.  */
 348   /* The cast to uintmax_t prevents negative remainders
 349      if st_ino is negative.  */
 350   return (uintmax_t) p->st_ino % table_size;
 351 }
 352
 353 /* Compare two dev/ino pairs.  Return true if they are the same.  */
 354 static bool
 355 entry_compare (void const *x, void const *y)
 356 {
 357   struct entry const *a = x;
 358   struct entry const *b = y;
 359   return SAME_INODE (*a, *b) ? true : false;
 360 }
 361
 362 /* Try to insert the INO/DEV pair into the global table, HTAB.
 363    Return true if the pair is successfully inserted,
 364    false if the pair is already in the table.  */
 365 static bool
 366 hash_ins (ino_t ino, dev_t dev)
 367 {
 368   struct entry *ent;
 369   struct entry *ent_from_table;
 370
 371   ent = xmalloc (sizeof *ent);
 372   ent->st_ino = ino;
 373   ent->st_dev = dev;
 374
 375   ent_from_table = hash_insert (htab, ent);
 376   if (ent_from_table == NULL)
 377     {
 378       /* Insertion failed due to lack of memory.  */
 379       xalloc_die ();
 380     }
 381
 382   if (ent_from_table == ent)
 383     {
 384       /* Insertion succeeded.  */
 385       return true;
 386     }
 387
 388   /* That pair is already in the table, so ENT was not inserted.  Free it.  */
 389   free (ent);
 390
 391   return false;
 392 }
 393
 394 /* Initialize the hash table.  */
 395 static void
 396 hash_init (void)
 397 {
 398   htab = hash_initialize (INITIAL_TABLE_SIZE, NULL,
 399                           entry_hash, entry_compare, free);
 400   if (htab == NULL)
 401     xalloc_die ();
 402 }
 403
 404 /* FIXME: this code is nearly identical to code in date.c  */
 405 /* Display the date and time in WHEN according to the format specified
 406    in FORMAT.  */
 407
 408 static void
 409 show_date (const char *format, struct timespec when)
 410 {
 411   struct tm *tm = localtime (&when.tv_sec);
 412   if (! tm)
 413     {
 414       char buf[INT_BUFSIZE_BOUND (intmax_t)];
 415       error (0, 0, _("time %s is out of range"), timetostr (when.tv_sec, buf));
 416       fputs (buf, stdout);
 417       return;
 418     }
 419
 420   fprintftime (stdout, format, tm, 0, when.tv_nsec);
 421 }
 422
 423 /* Print N_BYTES.  Convert it to a readable value before printing.  */
 424
 425 static void
 426 print_only_size (uintmax_t n_bytes)
 427 {
 428   char buf[LONGEST_HUMAN_READABLE + 1];
 429   fputs (human_readable (n_bytes, buf, human_output_opts,
 430                          1, output_block_size), stdout);
 431 }
 432
 433 /* Print size (and optionally time) indicated by *PDUI, followed by STRING.  */
 434
 435 static void
 436 print_size (const struct duinfo *pdui, const char *string)
 437 {
 438   print_only_size (pdui->size);
 439   if (opt_time)
 440     {
 441       putchar ('\t');
 442       show_date (time_format, pdui->tmax);
 443     }
 444   printf ("\t%s%c", string, opt_nul_terminate_output ? '\0' : '\n');
 445   fflush (stdout);
 446 }
 447
 448 /* This function is called once for every file system object that fts
 449    encounters.  fts does a depth-first traversal.  This function knows
 450    that and accumulates per-directory totals based on changes in
 451    the depth of the current entry.  It returns true on success.  */
 452
 453 static bool
 454 process_file (FTS *fts, FTSENT *ent)
 455 {
 456   bool ok;
 457   struct duinfo dui;
 458   struct duinfo dui_to_print;
 459   size_t level;
 460   static size_t prev_level;
 461   static size_t n_alloc;
 462   /* First element of the structure contains:
 463      The sum of the st_size values of all entries in the single directory
 464      at the corresponding level.  Although this does include the st_size
 465      corresponding to each subdirectory, it does not include the size of
 466      any file in a subdirectory. Also corresponding last modified date.
 467      Second element of the structure contains:
 468      The sum of the sizes of all entries in the hierarchy at or below the
 469      directory at the specified level.  */
 470   static struct dulevel *dulvl;
 471   bool print = true;
 472
 473   const char *file = ent->fts_path;
 474   const struct stat *sb = ent->fts_statp;
 475   bool skip;
 476
 477   /* If necessary, set FTS_SKIP before returning.  */
 478   skip = excluded_file_name (exclude, file);
 479   if (skip)
 480     fts_set (fts, ent, FTS_SKIP);
 481
 482   switch (ent->fts_info)
 483     {
 484     case FTS_NS:
 485       error (0, ent->fts_errno, _("cannot access %s"), quote (file));
 486       return false;
 487
 488     case FTS_ERR:
 489       /* if (S_ISDIR (ent->fts_statp->st_mode) && FIXME */
 490       error (0, ent->fts_errno, _("%s"), quote (file));
 491       return false;
 492
 493     case FTS_DNR:
 494       /* Don't return just yet, since although the directory is not readable,
 495          we were able to stat it, so we do have a size.  */
 496       error (0, ent->fts_errno, _("cannot read directory %s"), quote (file));
 497       ok = false;
 498       break;
 499
 500     default:
 501       ok = true;
 502       break;
 503     }
 504
 505   /* If this is the first (pre-order) encounter with a directory,
 506      or if it's the second encounter for a skipped directory, then
 507      return right away.  */
 508   if (ent->fts_info == FTS_D || skip)
 509     return ok;
 510
 511   /* If the file is being excluded or if it has already been counted
 512      via a hard link, then don't let it contribute to the sums.  */
 513   if (skip
 514       || (!opt_count_all
 515           && ! S_ISDIR (sb->st_mode)
 516           && 1 < sb->st_nlink
 517           && ! hash_ins (sb->st_ino, sb->st_dev)))
 518     {
 519       /* Note that we must not simply return here.
 520          We still have to update prev_level and maybe propagate
 521          some sums up the hierarchy.  */
 522       duinfo_init (&dui);
 523       print = false;
 524     }
 525   else
 526     {
 527       duinfo_set (&dui,
 528                   (apparent_size
 529                    ? sb->st_size
 530                    : (uintmax_t) ST_NBLOCKS (*sb) * ST_NBLOCKSIZE),
 531                   (time_type == time_mtime ? get_stat_mtime (sb)
 532                    : time_type == time_atime ? get_stat_atime (sb)
 533                    : get_stat_ctime (sb)));
 534     }
 535
 536   level = ent->fts_level;
 537   dui_to_print = dui;
 538
 539   if (n_alloc == 0)
 540     {
 541       n_alloc = level + 10;
 542       dulvl = xcalloc (n_alloc, sizeof *dulvl);
 543     }
 544   else
 545     {
 546       if (level == prev_level)
 547         {
 548           /* This is usually the most common case.  Do nothing.  */
 549         }
 550       else if (level > prev_level)
 551         {
 552           /* Descending the hierarchy.
 553              Clear the accumulators for *all* levels between prev_level
 554              and the current one.  The depth may change dramatically,
 555              e.g., from 1 to 10.  */
 556           size_t i;
 557
 558           if (n_alloc <= level)
 559             {
 560               dulvl = xnrealloc (dulvl, level, 2 * sizeof *dulvl);
 561               n_alloc = level * 2;
 562             }
 563
 564           for (i = prev_level + 1; i <= level; i++)
 565             {
 566               duinfo_init (&dulvl[i].ent);
 567               duinfo_init (&dulvl[i].subdir);
 568             }
 569         }
 570       else /* level < prev_level */
 571         {
 572           /* Ascending the hierarchy.
 573              Process a directory only after all entries in that
 574              directory have been processed.  When the depth decreases,
 575              propagate sums from the children (prev_level) to the parent.
 576              Here, the current level is always one smaller than the
 577              previous one.  */
 578           assert (level == prev_level - 1);
 579           duinfo_add (&dui_to_print, &dulvl[prev_level].ent);
 580           if (!opt_separate_dirs)
 581             duinfo_add (&dui_to_print, &dulvl[prev_level].subdir);
 582           duinfo_add (&dulvl[level].subdir, &dulvl[prev_level].ent);
 583           duinfo_add (&dulvl[level].subdir, &dulvl[prev_level].subdir);
 584         }
 585     }
 586
 587   prev_level = level;
 588
 589   /* Let the size of a directory entry contribute to the total for the
 590      containing directory, unless --separate-dirs (-S) is specified.  */
 591   if ( ! (opt_separate_dirs && IS_DIR_TYPE (ent->fts_info)))
 592     duinfo_add (&dulvl[level].ent, &dui);
 593
 594   /* Even if this directory is unreadable or we can't chdir into it,
 595      do let its size contribute to the total. */
 596   duinfo_add (&tot_dui, &dui);
 597
 598   /* If we're not counting an entry, e.g., because it's a hard link
 599      to a file we've already counted (and --count-links), then don't
 600      print a line for it.  */
 601   if (!print)
 602     return ok;
 603
 604   if ((IS_DIR_TYPE (ent->fts_info) && level <= max_depth)
 605       || ((opt_all && level <= max_depth) || level == 0))
 606     print_size (&dui_to_print, file);
 607
 608   return ok;
 609 }
 610
 611 /* Recursively print the sizes of the directories (and, if selected, files)
 612    named in FILES, the last entry of which is NULL.
 613    BIT_FLAGS controls how fts works.
 614    Return true if successful.  */
 615
 616 static bool
 617 du_files (char **files, int bit_flags)
 618 {
 619   bool ok = true;
 620
 621   if (*files)
 622     {
 623       FTS *fts = xfts_open (files, bit_flags, NULL);
 624
 625       while (1)
 626         {
 627           FTSENT *ent;
 628
 629           ent = fts_read (fts);
 630           if (ent == NULL)
 631             {
 632               if (errno != 0)
 633                 {
 634                   /* FIXME: try to give a better message  */
 635                   error (0, errno, _("fts_read failed"));
 636                   ok = false;
 637                 }
 638               break;
 639             }
 640           FTS_CROSS_CHECK (fts);
 641
 642           ok &= process_file (fts, ent);
 643         }
 644
 645       /* Ignore failure, since the only way it can do so is in failing to
 646          return to the original directory, and since we're about to exit,
 647          that doesn't matter.  */
 648       fts_close (fts);
 649     }
 650
 651   return ok;
 652 }
 653
 654 int
 655 main (int argc, char **argv)
 656 {
 657   char *cwd_only[2];
 658   bool max_depth_specified = false;
 659   bool ok = true;
 660   char *files_from = NULL;
 661
 662   /* Bit flags that control how fts works.  */
 663   int bit_flags = FTS_TIGHT_CYCLE_CHECK | FTS_DEFER_STAT;
 664
 665   /* Select one of the three FTS_ options that control if/when
 666      to follow a symlink.  */
 667   int symlink_deref_bits = FTS_PHYSICAL;
 668
 669   /* If true, display only a total for each argument. */
 670   bool opt_summarize_only = false;
 671
 672   cwd_only[0] = bad_cast (".");
 673   cwd_only[1] = NULL;
 674
 675   initialize_main (&argc, &argv);
 676   set_program_name (argv[0]);
 677   setlocale (LC_ALL, "");
 678   bindtextdomain (PACKAGE, LOCALEDIR);
 679   textdomain (PACKAGE);
 680
 681   atexit (close_stdout);
 682
 683   exclude = new_exclude ();
 684
 685   human_options (getenv ("DU_BLOCK_SIZE"),
 686                  &human_output_opts, &output_block_size);
 687
 688   for (;;)
 689     {
 690       int oi = -1;
 691       int c = getopt_long (argc, argv, DEBUG_OPT "0abchHklmsxB:DLPSX:",
 692                            long_options, &oi);
 693       if (c == -1)
 694         break;
 695
 696       switch (c)
 697         {
 698 #if DU_DEBUG
 699         case 'd':
 700           fts_debug = true;
 701           break;
 702 #endif
 703
 704         case '0':
 705           opt_nul_terminate_output = true;
 706           break;
 707
 708         case 'a':
 709           opt_all = true;
 710           break;
 711
 712         case APPARENT_SIZE_OPTION:
 713           apparent_size = true;
 714           break;
 715
 716         case 'b':
 717           apparent_size = true;
 718           human_output_opts = 0;
 719           output_block_size = 1;
 720           break;
 721
 722         case 'c':
 723           print_grand_total = true;
 724           break;
 725
 726         case 'h':
 727           human_output_opts = human_autoscale | human_SI | human_base_1024;
 728           output_block_size = 1;
 729           break;
 730
 731         case HUMAN_SI_OPTION:
 732           human_output_opts = human_autoscale | human_SI;
 733           output_block_size = 1;
 734           break;
 735
 736         case 'k':
 737           human_output_opts = 0;
 738           output_block_size = 1024;
 739           break;
 740
 741         case MAX_DEPTH_OPTION:          /* --max-depth=N */
 742           {
 743             unsigned long int tmp_ulong;
 744             if (xstrtoul (optarg, NULL, 0, &tmp_ulong, NULL) == LONGINT_OK
 745                 && tmp_ulong <= SIZE_MAX)
 746               {
 747                 max_depth_specified = true;
 748                 max_depth = tmp_ulong;
 749               }
 750             else
 751               {
 752                 error (0, 0, _("invalid maximum depth %s"),
 753                        quote (optarg));
 754                 ok = false;
 755               }
 756           }
 757           break;
 758
 759         case MEGABYTES_LONG_OPTION: /* FIXME: remove in 2009 */
 760           error (0, 0,
 761                  _("the --megabytes option is deprecated; use -m instead"));
 762           /* fall through */
 763         case 'm':
 764           human_output_opts = 0;
 765           output_block_size = 1024 * 1024;
 766           break;
 767
 768         case 'l':
 769           opt_count_all = true;
 770           break;
 771
 772         case 's':
 773           opt_summarize_only = true;
 774           break;
 775
 776         case 'x':
 777           bit_flags |= FTS_XDEV;
 778           break;
 779
 780         case 'B':
 781           {
 782             enum strtol_error e = human_options (optarg, &human_output_opts,
 783                                                  &output_block_size);
 784             if (e != LONGINT_OK)
 785               xstrtol_fatal (e, oi, c, long_options, optarg);
 786           }
 787           break;
 788
 789         case 'H':  /* NOTE: before 2008-12, -H was equivalent to --si.  */
 790         case 'D':
 791           symlink_deref_bits = FTS_COMFOLLOW | FTS_PHYSICAL;
 792           break;
 793
 794         case 'L': /* --dereference */
 795           symlink_deref_bits = FTS_LOGICAL;
 796           break;
 797
 798         case 'P': /* --no-dereference */
 799           symlink_deref_bits = FTS_PHYSICAL;
 800           break;
 801
 802         case 'S':
 803           opt_separate_dirs = true;
 804           break;
 805
 806         case 'X':
 807           if (add_exclude_file (add_exclude, exclude, optarg,
 808                                 EXCLUDE_WILDCARDS, '\n'))
 809             {
 810               error (0, errno, "%s", quotearg_colon (optarg));
 811               ok = false;
 812             }
 813           break;
 814
 815         case FILES0_FROM_OPTION:
 816           files_from = optarg;
 817           break;
 818
 819         case EXCLUDE_OPTION:
 820           add_exclude (exclude, optarg, EXCLUDE_WILDCARDS);
 821           break;
 822
 823         case TIME_OPTION:
 824           opt_time = true;
 825           time_type =
 826             (optarg
 827              ? XARGMATCH ("--time", optarg, time_args, time_types)
 828              : time_mtime);
 829           break;
 830
 831         case TIME_STYLE_OPTION:
 832           time_style = optarg;
 833           break;
 834
 835         case_GETOPT_HELP_CHAR;
 836
 837         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 838
 839         default:
 840           ok = false;
 841         }
 842     }
 843
 844   if (!ok)
 845     usage (EXIT_FAILURE);
 846
 847   if (opt_all & opt_summarize_only)
 848     {
 849       error (0, 0, _("cannot both summarize and show all entries"));
 850       usage (EXIT_FAILURE);
 851     }
 852
 853   if (opt_summarize_only && max_depth_specified && max_depth == 0)
 854     {
 855       error (0, 0,
 856              _("warning: summarizing is the same as using --max-depth=0"));
 857     }
 858
 859   if (opt_summarize_only && max_depth_specified && max_depth != 0)
 860     {
 861       unsigned long int d = max_depth;
 862       error (0, 0, _("warning: summarizing conflicts with --max-depth=%lu"), d);
 863       usage (EXIT_FAILURE);
 864     }
 865
 866   if (opt_summarize_only)
 867     max_depth = 0;
 868
 869   /* Process time style if printing last times.  */
 870   if (opt_time)
 871     {
 872       if (! time_style)
 873         {
 874           time_style = getenv ("TIME_STYLE");
 875
 876           /* Ignore TIMESTYLE="locale", for compatibility with ls.  */
 877           if (! time_style || STREQ (time_style, "locale"))
 878             time_style = "long-iso";
 879           else if (*time_style == '+')
 880             {
 881               /* Ignore anything after a newline, for compatibility
 882                  with ls.  */
 883               char *p = strchr (time_style, '\n');
 884               if (p)
 885                 *p = '\0';
 886             }
 887           else
 888             {
 889               /* Ignore "posix-" prefix, for compatibility with ls.  */
 890               static char const posix_prefix[] = "posix-";
 891               while (strncmp (time_style, posix_prefix, sizeof posix_prefix - 1)
 892                      == 0)
 893                 time_style += sizeof posix_prefix - 1;
 894             }
 895         }
 896
 897       if (*time_style == '+')
 898         time_format = time_style + 1;
 899       else
 900         {
 901           switch (XARGMATCH ("time style", time_style,
 902                              time_style_args, time_style_types))
 903             {
 904             case full_iso_time_style:
 905               time_format = "%Y-%m-%d %H:%M:%S.%N %z";
 906               break;
 907
 908             case long_iso_time_style:
 909               time_format = "%Y-%m-%d %H:%M";
 910               break;
 911
 912             case iso_time_style:
 913               time_format = "%Y-%m-%d";
 914               break;
 915             }
 916         }
 917     }
 918
 919   struct argv_iterator *ai;
 920   if (files_from)
 921     {
 922       /* When using --files0-from=F, you may not specify any files
 923          on the command-line.  */
 924       if (optind < argc)
 925         {
 926           error (0, 0, _("extra operand %s"), quote (argv[optind]));
 927           fprintf (stderr, "%s\n",
 928                    _("file operands cannot be combined with --files0-from"));
 929           usage (EXIT_FAILURE);
 930         }
 931
 932       if (! (STREQ (files_from, "-") || freopen (files_from, "r", stdin)))
 933         error (EXIT_FAILURE, errno, _("cannot open %s for reading"),
 934                quote (files_from));
 935
 936       ai = argv_iter_init_stream (stdin);
 937     }
 938   else
 939     {
 940       char **files = (optind < argc ? argv + optind : cwd_only);
 941       ai = argv_iter_init_argv (files);
 942     }
 943
 944   if (!ai)
 945     xalloc_die ();
 946
 947   /* Initialize the hash structure for inode numbers.  */
 948   hash_init ();
 949
 950   bit_flags |= symlink_deref_bits;
 951   static char *temp_argv[] = { NULL, NULL };
 952
 953   while (true)
 954     {
 955       bool skip_file = false;
 956       enum argv_iter_err ai_err;
 957       char *file_name = argv_iter (ai, &ai_err);
 958       if (ai_err == AI_ERR_EOF)
 959         break;
 960       if (!file_name)
 961         {
 962           switch (ai_err)
 963             {
 964             case AI_ERR_READ:
 965               error (0, errno, _("%s: read error"), quote (files_from));
 966               skip_file = true;
 967               continue;
 968
 969             case AI_ERR_MEM:
 970               xalloc_die ();
 971
 972             default:
 973               assert (!"unexpected error code from argv_iter");
 974             }
 975         }
 976       if (files_from && STREQ (files_from, "-") && STREQ (file_name, "-"))
 977         {
 978           /* Give a better diagnostic in an unusual case:
 979              printf - | du --files0-from=- */
 980           error (0, 0, _("when reading file names from stdin, "
 981                          "no file name of %s allowed"),
 982                  quote (file_name));
 983           skip_file = true;
 984         }
 985
 986       /* Report and skip any empty file names before invoking fts.
 987          This works around a glitch in fts, which fails immediately
 988          (without looking at the other file names) when given an empty
 989          file name.  */
 990       if (!file_name[0])
 991         {
 992           /* Diagnose a zero-length file name.  When it's one
 993              among many, knowing the record number may help.
 994              FIXME: currently print the record number only with
 995              --files0-from=FILE.  Maybe do it for argv, too?  */
 996           if (files_from == NULL)
 997             error (0, 0, "%s", _("invalid zero-length file name"));
 998           else
 999             {
1000               /* Using the standard `filename:line-number:' prefix here is
1001                  not totally appropriate, since NUL is the separator, not NL,
1002                  but it might be better than nothing.  */
1003               unsigned long int file_number = argv_iter_n_args (ai);
1004               error (0, 0, "%s:%lu: %s", quotearg_colon (files_from),
1005                      file_number, _("invalid zero-length file name"));
1006             }
1007           skip_file = true;
1008         }
1009
1010       if (skip_file)
1011         ok = false;
1012       else
1013         {
1014           temp_argv[0] = file_name;
1015           ok &= du_files (temp_argv, bit_flags);
1016         }
1017     }
1018
1019   argv_iter_free (ai);
1020
1021   if (files_from && (ferror (stdin) || fclose (stdin) != 0))
1022     error (EXIT_FAILURE, 0, _("error reading %s"), quote (files_from));
1023
1024   if (print_grand_total)
1025     print_size (&tot_dui, _("total"));
1026
1027   hash_free (htab);
1028
1029   exit (ok ? EXIT_SUCCESS : EXIT_FAILURE);
1030 }