src/du.c

   1 /* du -- summarize disk usage
   2    Copyright (C) 1988-1991, 1995-2010 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 <getopt.h>
  28 #include <sys/types.h>
  29 #include <assert.h>
  30 #include "system.h"
  31 #include "argmatch.h"
  32 #include "argv-iter.h"
  33 #include "di-set.h"
  34 #include "error.h"
  35 #include "exclude.h"
  36 #include "fprintftime.h"
  37 #include "human.h"
  38 #include "quote.h"
  39 #include "quotearg.h"
  40 #include "same.h"
  41 #include "stat-time.h"
  42 #include "stdio--.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 #else
  60 # define FTS_CROSS_CHECK(Fts)
  61 #endif
  62
  63 /* A set of dev/ino pairs.  */
  64 static struct di_set *di_set;
  65
  66 /* Define a class for collecting directory information. */
  67
  68 struct duinfo
  69 {
  70   /* Size of files in directory.  */
  71   uintmax_t size;
  72
  73   /* Latest time stamp found.  If tmax.tv_sec == TYPE_MINIMUM (time_t)
  74      && tmax.tv_nsec < 0, no time stamp has been found.  */
  75   struct timespec tmax;
  76 };
  77
  78 /* Initialize directory data.  */
  79 static inline void
  80 duinfo_init (struct duinfo *a)
  81 {
  82   a->size = 0;
  83   a->tmax.tv_sec = TYPE_MINIMUM (time_t);
  84   a->tmax.tv_nsec = -1;
  85 }
  86
  87 /* Set directory data.  */
  88 static inline void
  89 duinfo_set (struct duinfo *a, uintmax_t size, struct timespec tmax)
  90 {
  91   a->size = size;
  92   a->tmax = tmax;
  93 }
  94
  95 /* Accumulate directory data.  */
  96 static inline void
  97 duinfo_add (struct duinfo *a, struct duinfo const *b)
  98 {
  99   a->size += b->size;
 100   if (timespec_cmp (a->tmax, b->tmax) < 0)
 101     a->tmax = b->tmax;
 102 }
 103
 104 /* A structure for per-directory level information.  */
 105 struct dulevel
 106 {
 107   /* Entries in this directory.  */
 108   struct duinfo ent;
 109
 110   /* Total for subdirectories.  */
 111   struct duinfo subdir;
 112 };
 113
 114 /* If true, display counts for all files, not just directories.  */
 115 static bool opt_all = false;
 116
 117 /* If true, rather than using the disk usage of each file,
 118    use the apparent size (a la stat.st_size).  */
 119 static bool apparent_size = false;
 120
 121 /* If true, count each hard link of files with multiple links.  */
 122 static bool opt_count_all = false;
 123
 124 /* If true, hash all files to look for hard links.  */
 125 static bool hash_all;
 126
 127 /* If true, output the NUL byte instead of a newline at the end of each line. */
 128 static bool opt_nul_terminate_output = false;
 129
 130 /* If true, print a grand total at the end.  */
 131 static bool print_grand_total = false;
 132
 133 /* If nonzero, do not add sizes of subdirectories.  */
 134 static bool opt_separate_dirs = false;
 135
 136 /* Show the total for each directory (and file if --all) that is at
 137    most MAX_DEPTH levels down from the root of the hierarchy.  The root
 138    is at level 0, so `du --max-depth=0' is equivalent to `du -s'.  */
 139 static size_t max_depth = SIZE_MAX;
 140
 141 /* Human-readable options for output.  */
 142 static int human_output_opts;
 143
 144 /* If true, print most recently modified date, using the specified format.  */
 145 static bool opt_time = false;
 146
 147 /* Type of time to display. controlled by --time.  */
 148
 149 enum time_type
 150   {
 151     time_mtime,                 /* default */
 152     time_ctime,
 153     time_atime
 154   };
 155
 156 static enum time_type time_type = time_mtime;
 157
 158 /* User specified date / time style */
 159 static char const *time_style = NULL;
 160
 161 /* Format used to display date / time. Controlled by --time-style */
 162 static char const *time_format = NULL;
 163
 164 /* The units to use when printing sizes.  */
 165 static uintmax_t output_block_size;
 166
 167 /* File name patterns to exclude.  */
 168 static struct exclude *exclude;
 169
 170 /* Grand total size of all args, in bytes. Also latest modified date. */
 171 static struct duinfo tot_dui;
 172
 173 #define IS_DIR_TYPE(Type)       \
 174   ((Type) == FTS_DP             \
 175    || (Type) == FTS_DNR)
 176
 177 /* For long options that have no equivalent short option, use a
 178    non-character as a pseudo short option, starting with CHAR_MAX + 1.  */
 179 enum
 180 {
 181   APPARENT_SIZE_OPTION = CHAR_MAX + 1,
 182   EXCLUDE_OPTION,
 183   FILES0_FROM_OPTION,
 184   HUMAN_SI_OPTION,
 185   FTS_DEBUG,
 186   TIME_OPTION,
 187   TIME_STYLE_OPTION
 188 };
 189
 190 static struct option const long_options[] =
 191 {
 192   {"all", no_argument, NULL, 'a'},
 193   {"apparent-size", no_argument, NULL, APPARENT_SIZE_OPTION},
 194   {"block-size", required_argument, NULL, 'B'},
 195   {"bytes", no_argument, NULL, 'b'},
 196   {"count-links", no_argument, NULL, 'l'},
 197   /* {"-debug", no_argument, NULL, FTS_DEBUG}, */
 198   {"dereference", no_argument, NULL, 'L'},
 199   {"dereference-args", no_argument, NULL, 'D'},
 200   {"exclude", required_argument, NULL, EXCLUDE_OPTION},
 201   {"exclude-from", required_argument, NULL, 'X'},
 202   {"files0-from", required_argument, NULL, FILES0_FROM_OPTION},
 203   {"human-readable", no_argument, NULL, 'h'},
 204   {"si", no_argument, NULL, HUMAN_SI_OPTION},
 205   {"max-depth", required_argument, NULL, 'd'},
 206   {"null", no_argument, NULL, '0'},
 207   {"no-dereference", no_argument, NULL, 'P'},
 208   {"one-file-system", no_argument, NULL, 'x'},
 209   {"separate-dirs", no_argument, NULL, 'S'},
 210   {"summarize", no_argument, NULL, 's'},
 211   {"total", no_argument, NULL, 'c'},
 212   {"time", optional_argument, NULL, TIME_OPTION},
 213   {"time-style", required_argument, NULL, TIME_STYLE_OPTION},
 214   {GETOPT_HELP_OPTION_DECL},
 215   {GETOPT_VERSION_OPTION_DECL},
 216   {NULL, 0, NULL, 0}
 217 };
 218
 219 static char const *const time_args[] =
 220 {
 221   "atime", "access", "use", "ctime", "status", NULL
 222 };
 223 static enum time_type const time_types[] =
 224 {
 225   time_atime, time_atime, time_atime, time_ctime, time_ctime
 226 };
 227 ARGMATCH_VERIFY (time_args, time_types);
 228
 229 /* `full-iso' uses full ISO-style dates and times.  `long-iso' uses longer
 230    ISO-style time stamps, though shorter than `full-iso'.  `iso' uses shorter
 231    ISO-style time stamps.  */
 232 enum time_style
 233   {
 234     full_iso_time_style,       /* --time-style=full-iso */
 235     long_iso_time_style,       /* --time-style=long-iso */
 236     iso_time_style             /* --time-style=iso */
 237   };
 238
 239 static char const *const time_style_args[] =
 240 {
 241   "full-iso", "long-iso", "iso", NULL
 242 };
 243 static enum time_style const time_style_types[] =
 244 {
 245   full_iso_time_style, long_iso_time_style, iso_time_style
 246 };
 247 ARGMATCH_VERIFY (time_style_args, time_style_types);
 248
 249 void
 250 usage (int status)
 251 {
 252   if (status != EXIT_SUCCESS)
 253     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 254              program_name);
 255   else
 256     {
 257       printf (_("\
 258 Usage: %s [OPTION]... [FILE]...\n\
 259   or:  %s [OPTION]... --files0-from=F\n\
 260 "), program_name, program_name);
 261       fputs (_("\
 262 Summarize disk usage of each FILE, recursively for directories.\n\
 263 \n\
 264 "), stdout);
 265       fputs (_("\
 266 Mandatory arguments to long options are mandatory for short options too.\n\
 267 "), stdout);
 268       fputs (_("\
 269   -a, --all             write counts for all files, not just directories\n\
 270       --apparent-size   print apparent sizes, rather than disk usage; although\n\
 271                           the apparent size is usually smaller, it may be\n\
 272                           larger due to holes in (`sparse') files, internal\n\
 273                           fragmentation, indirect blocks, and the like\n\
 274 "), stdout);
 275       fputs (_("\
 276   -B, --block-size=SIZE  scale sizes by SIZE before printing them.  E.g.,\n\
 277                            `-BM' prints sizes in units of 1,048,576 bytes.\n\
 278                            See SIZE format below.\n\
 279   -b, --bytes           equivalent to `--apparent-size --block-size=1'\n\
 280   -c, --total           produce a grand total\n\
 281   -D, --dereference-args  dereference only symlinks that are listed on the\n\
 282                           command line\n\
 283 "), stdout);
 284       fputs (_("\
 285       --files0-from=F   summarize disk usage of the NUL-terminated file\n\
 286                           names specified in file F;\n\
 287                           If F is - then read names from standard input\n\
 288   -H                    equivalent to --dereference-args (-D)\n\
 289   -h, --human-readable  print sizes in human readable format (e.g., 1K 234M 2G)\n\
 290       --si              like -h, but use powers of 1000 not 1024\n\
 291 "), stdout);
 292       fputs (_("\
 293   -k                    like --block-size=1K\n\
 294   -l, --count-links     count sizes many times if hard linked\n\
 295   -m                    like --block-size=1M\n\
 296 "), stdout);
 297       fputs (_("\
 298   -L, --dereference     dereference all symbolic links\n\
 299   -P, --no-dereference  don't follow any symbolic links (this is the default)\n\
 300   -0, --null            end each output line with 0 byte rather than newline\n\
 301   -S, --separate-dirs   do not include size of subdirectories\n\
 302   -s, --summarize       display only a total for each argument\n\
 303 "), stdout);
 304       fputs (_("\
 305   -x, --one-file-system    skip directories on different file systems\n\
 306   -X, --exclude-from=FILE  exclude files that match any pattern in FILE\n\
 307       --exclude=PATTERN    exclude files that match PATTERN\n\
 308   -d, --max-depth=N     print the total for a directory (or file, with --all)\n\
 309                           only if it is N or fewer levels below the command\n\
 310                           line argument;  --max-depth=0 is the same as\n\
 311                           --summarize\n\
 312 "), stdout);
 313       fputs (_("\
 314       --time            show time of the last modification of any file in the\n\
 315                           directory, or any of its subdirectories\n\
 316       --time=WORD       show time as WORD instead of modification time:\n\
 317                           atime, access, use, ctime or status\n\
 318       --time-style=STYLE  show times using style STYLE:\n\
 319                           full-iso, long-iso, iso, +FORMAT\n\
 320                           FORMAT is interpreted like `date'\n\
 321 "), stdout);
 322       fputs (HELP_OPTION_DESCRIPTION, stdout);
 323       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 324       emit_blocksize_note ("DU");
 325       emit_size_note ();
 326       emit_ancillary_info ();
 327     }
 328   exit (status);
 329 }
 330
 331 /* Try to insert the INO/DEV pair into the global table, HTAB.
 332    Return true if the pair is successfully inserted,
 333    false if the pair is already in the table.  */
 334 static bool
 335 hash_ins (ino_t ino, dev_t dev)
 336 {
 337   int inserted = di_set_insert (di_set, dev, ino);
 338   if (inserted < 0)
 339     xalloc_die ();
 340   return inserted;
 341 }
 342
 343 /* FIXME: this code is nearly identical to code in date.c  */
 344 /* Display the date and time in WHEN according to the format specified
 345    in FORMAT.  */
 346
 347 static void
 348 show_date (const char *format, struct timespec when)
 349 {
 350   struct tm *tm = localtime (&when.tv_sec);
 351   if (! tm)
 352     {
 353       char buf[INT_BUFSIZE_BOUND (intmax_t)];
 354       error (0, 0, _("time %s is out of range"), timetostr (when.tv_sec, buf));
 355       fputs (buf, stdout);
 356       return;
 357     }
 358
 359   fprintftime (stdout, format, tm, 0, when.tv_nsec);
 360 }
 361
 362 /* Print N_BYTES.  Convert it to a readable value before printing.  */
 363
 364 static void
 365 print_only_size (uintmax_t n_bytes)
 366 {
 367   char buf[LONGEST_HUMAN_READABLE + 1];
 368   fputs (human_readable (n_bytes, buf, human_output_opts,
 369                          1, output_block_size), stdout);
 370 }
 371
 372 /* Print size (and optionally time) indicated by *PDUI, followed by STRING.  */
 373
 374 static void
 375 print_size (const struct duinfo *pdui, const char *string)
 376 {
 377   print_only_size (pdui->size);
 378   if (opt_time)
 379     {
 380       putchar ('\t');
 381       show_date (time_format, pdui->tmax);
 382     }
 383   printf ("\t%s%c", string, opt_nul_terminate_output ? '\0' : '\n');
 384   fflush (stdout);
 385 }
 386
 387 /* This function is called once for every file system object that fts
 388    encounters.  fts does a depth-first traversal.  This function knows
 389    that and accumulates per-directory totals based on changes in
 390    the depth of the current entry.  It returns true on success.  */
 391
 392 static bool
 393 process_file (FTS *fts, FTSENT *ent)
 394 {
 395   bool ok = true;
 396   struct duinfo dui;
 397   struct duinfo dui_to_print;
 398   size_t level;
 399   static size_t prev_level;
 400   static size_t n_alloc;
 401   /* First element of the structure contains:
 402      The sum of the st_size values of all entries in the single directory
 403      at the corresponding level.  Although this does include the st_size
 404      corresponding to each subdirectory, it does not include the size of
 405      any file in a subdirectory. Also corresponding last modified date.
 406      Second element of the structure contains:
 407      The sum of the sizes of all entries in the hierarchy at or below the
 408      directory at the specified level.  */
 409   static struct dulevel *dulvl;
 410
 411   const char *file = ent->fts_path;
 412   const struct stat *sb = ent->fts_statp;
 413   int info = ent->fts_info;
 414
 415   if (info == FTS_DNR)
 416     {
 417       /* An error occurred, but the size is known, so count it.  */
 418       error (0, ent->fts_errno, _("cannot read directory %s"), quote (file));
 419       ok = false;
 420     }
 421   else if (info != FTS_DP)
 422     {
 423       bool excluded = excluded_file_name (exclude, file);
 424       if (! excluded)
 425         {
 426           /* Make the stat buffer *SB valid, or fail noisily.  */
 427
 428           if (info == FTS_NSOK)
 429             {
 430               fts_set (fts, ent, FTS_AGAIN);
 431               FTSENT const *e = fts_read (fts);
 432               assert (e == ent);
 433               info = ent->fts_info;
 434             }
 435
 436           if (info == FTS_NS || info == FTS_SLNONE)
 437             {
 438               error (0, ent->fts_errno, _("cannot access %s"), quote (file));
 439               return false;
 440             }
 441         }
 442
 443       if (excluded
 444           || (! opt_count_all
 445               && (hash_all || (! S_ISDIR (sb->st_mode) && 1 < sb->st_nlink))
 446               && ! hash_ins (sb->st_ino, sb->st_dev)))
 447         {
 448           /* If ignoring a directory in preorder, skip its children.
 449              Ignore the next fts_read output too, as it's a postorder
 450              visit to the same directory.  */
 451           if (info == FTS_D)
 452             {
 453               fts_set (fts, ent, FTS_SKIP);
 454               FTSENT const *e = fts_read (fts);
 455               assert (e == ent);
 456             }
 457
 458           return true;
 459         }
 460
 461       switch (info)
 462         {
 463         case FTS_D:
 464           return true;
 465
 466         case FTS_ERR:
 467           /* An error occurred, but the size is known, so count it.  */
 468           error (0, ent->fts_errno, "%s", quote (file));
 469           ok = false;
 470           break;
 471
 472         case FTS_DC:
 473           if (cycle_warning_required (fts, ent))
 474             {
 475               emit_cycle_warning (file);
 476               return false;
 477             }
 478           return true;
 479         }
 480     }
 481
 482   duinfo_set (&dui,
 483               (apparent_size
 484                ? sb->st_size
 485                : (uintmax_t) ST_NBLOCKS (*sb) * ST_NBLOCKSIZE),
 486               (time_type == time_mtime ? get_stat_mtime (sb)
 487                : time_type == time_atime ? get_stat_atime (sb)
 488                : get_stat_ctime (sb)));
 489
 490   level = ent->fts_level;
 491   dui_to_print = dui;
 492
 493   if (n_alloc == 0)
 494     {
 495       n_alloc = level + 10;
 496       dulvl = xcalloc (n_alloc, sizeof *dulvl);
 497     }
 498   else
 499     {
 500       if (level == prev_level)
 501         {
 502           /* This is usually the most common case.  Do nothing.  */
 503         }
 504       else if (level > prev_level)
 505         {
 506           /* Descending the hierarchy.
 507              Clear the accumulators for *all* levels between prev_level
 508              and the current one.  The depth may change dramatically,
 509              e.g., from 1 to 10.  */
 510           size_t i;
 511
 512           if (n_alloc <= level)
 513             {
 514               dulvl = xnrealloc (dulvl, level, 2 * sizeof *dulvl);
 515               n_alloc = level * 2;
 516             }
 517
 518           for (i = prev_level + 1; i <= level; i++)
 519             {
 520               duinfo_init (&dulvl[i].ent);
 521               duinfo_init (&dulvl[i].subdir);
 522             }
 523         }
 524       else /* level < prev_level */
 525         {
 526           /* Ascending the hierarchy.
 527              Process a directory only after all entries in that
 528              directory have been processed.  When the depth decreases,
 529              propagate sums from the children (prev_level) to the parent.
 530              Here, the current level is always one smaller than the
 531              previous one.  */
 532           assert (level == prev_level - 1);
 533           duinfo_add (&dui_to_print, &dulvl[prev_level].ent);
 534           if (!opt_separate_dirs)
 535             duinfo_add (&dui_to_print, &dulvl[prev_level].subdir);
 536           duinfo_add (&dulvl[level].subdir, &dulvl[prev_level].ent);
 537           duinfo_add (&dulvl[level].subdir, &dulvl[prev_level].subdir);
 538         }
 539     }
 540
 541   prev_level = level;
 542
 543   /* Let the size of a directory entry contribute to the total for the
 544      containing directory, unless --separate-dirs (-S) is specified.  */
 545   if (! (opt_separate_dirs && IS_DIR_TYPE (info)))
 546     duinfo_add (&dulvl[level].ent, &dui);
 547
 548   /* Even if this directory is unreadable or we can't chdir into it,
 549      do let its size contribute to the total. */
 550   duinfo_add (&tot_dui, &dui);
 551
 552   if ((IS_DIR_TYPE (info) && level <= max_depth)
 553       || ((opt_all && level <= max_depth) || level == 0))
 554     print_size (&dui_to_print, file);
 555
 556   return ok;
 557 }
 558
 559 /* Recursively print the sizes of the directories (and, if selected, files)
 560    named in FILES, the last entry of which is NULL.
 561    BIT_FLAGS controls how fts works.
 562    Return true if successful.  */
 563
 564 static bool
 565 du_files (char **files, int bit_flags)
 566 {
 567   bool ok = true;
 568
 569   if (*files)
 570     {
 571       FTS *fts = xfts_open (files, bit_flags, NULL);
 572
 573       while (1)
 574         {
 575           FTSENT *ent;
 576
 577           ent = fts_read (fts);
 578           if (ent == NULL)
 579             {
 580               if (errno != 0)
 581                 {
 582                   /* FIXME: try to give a better message  */
 583                   error (0, errno, _("fts_read failed"));
 584                   ok = false;
 585                 }
 586               break;
 587             }
 588           FTS_CROSS_CHECK (fts);
 589
 590           ok &= process_file (fts, ent);
 591         }
 592
 593       if (fts_close (fts) != 0)
 594         {
 595           error (0, errno, _("fts_close failed"));
 596           ok = false;
 597         }
 598     }
 599
 600   return ok;
 601 }
 602
 603 int
 604 main (int argc, char **argv)
 605 {
 606   char *cwd_only[2];
 607   bool max_depth_specified = false;
 608   bool ok = true;
 609   char *files_from = NULL;
 610
 611   /* Bit flags that control how fts works.  */
 612   int bit_flags = FTS_NOSTAT;
 613
 614   /* Select one of the three FTS_ options that control if/when
 615      to follow a symlink.  */
 616   int symlink_deref_bits = FTS_PHYSICAL;
 617
 618   /* If true, display only a total for each argument. */
 619   bool opt_summarize_only = false;
 620
 621   cwd_only[0] = bad_cast (".");
 622   cwd_only[1] = NULL;
 623
 624   initialize_main (&argc, &argv);
 625   set_program_name (argv[0]);
 626   setlocale (LC_ALL, "");
 627   bindtextdomain (PACKAGE, LOCALEDIR);
 628   textdomain (PACKAGE);
 629
 630   atexit (close_stdout);
 631
 632   exclude = new_exclude ();
 633
 634   human_options (getenv ("DU_BLOCK_SIZE"),
 635                  &human_output_opts, &output_block_size);
 636
 637   while (true)
 638     {
 639       int oi = -1;
 640       int c = getopt_long (argc, argv, "0abd:chHklmsxB:DLPSX:",
 641                            long_options, &oi);
 642       if (c == -1)
 643         break;
 644
 645       switch (c)
 646         {
 647 #if DU_DEBUG
 648         case FTS_DEBUG:
 649           fts_debug = true;
 650           break;
 651 #endif
 652
 653         case '0':
 654           opt_nul_terminate_output = true;
 655           break;
 656
 657         case 'a':
 658           opt_all = true;
 659           break;
 660
 661         case APPARENT_SIZE_OPTION:
 662           apparent_size = true;
 663           break;
 664
 665         case 'b':
 666           apparent_size = true;
 667           human_output_opts = 0;
 668           output_block_size = 1;
 669           break;
 670
 671         case 'c':
 672           print_grand_total = true;
 673           break;
 674
 675         case 'h':
 676           human_output_opts = human_autoscale | human_SI | human_base_1024;
 677           output_block_size = 1;
 678           break;
 679
 680         case HUMAN_SI_OPTION:
 681           human_output_opts = human_autoscale | human_SI;
 682           output_block_size = 1;
 683           break;
 684
 685         case 'k':
 686           human_output_opts = 0;
 687           output_block_size = 1024;
 688           break;
 689
 690         case 'd':               /* --max-depth=N */
 691           {
 692             unsigned long int tmp_ulong;
 693             if (xstrtoul (optarg, NULL, 0, &tmp_ulong, NULL) == LONGINT_OK
 694                 && tmp_ulong <= SIZE_MAX)
 695               {
 696                 max_depth_specified = true;
 697                 max_depth = tmp_ulong;
 698               }
 699             else
 700               {
 701                 error (0, 0, _("invalid maximum depth %s"),
 702                        quote (optarg));
 703                 ok = false;
 704               }
 705           }
 706           break;
 707
 708         case 'm':
 709           human_output_opts = 0;
 710           output_block_size = 1024 * 1024;
 711           break;
 712
 713         case 'l':
 714           opt_count_all = true;
 715           break;
 716
 717         case 's':
 718           opt_summarize_only = true;
 719           break;
 720
 721         case 'x':
 722           bit_flags |= FTS_XDEV;
 723           break;
 724
 725         case 'B':
 726           {
 727             enum strtol_error e = human_options (optarg, &human_output_opts,
 728                                                  &output_block_size);
 729             if (e != LONGINT_OK)
 730               xstrtol_fatal (e, oi, c, long_options, optarg);
 731           }
 732           break;
 733
 734         case 'H':  /* NOTE: before 2008-12, -H was equivalent to --si.  */
 735         case 'D':
 736           symlink_deref_bits = FTS_COMFOLLOW | FTS_PHYSICAL;
 737           break;
 738
 739         case 'L': /* --dereference */
 740           symlink_deref_bits = FTS_LOGICAL;
 741           break;
 742
 743         case 'P': /* --no-dereference */
 744           symlink_deref_bits = FTS_PHYSICAL;
 745           break;
 746
 747         case 'S':
 748           opt_separate_dirs = true;
 749           break;
 750
 751         case 'X':
 752           if (add_exclude_file (add_exclude, exclude, optarg,
 753                                 EXCLUDE_WILDCARDS, '\n'))
 754             {
 755               error (0, errno, "%s", quotearg_colon (optarg));
 756               ok = false;
 757             }
 758           break;
 759
 760         case FILES0_FROM_OPTION:
 761           files_from = optarg;
 762           break;
 763
 764         case EXCLUDE_OPTION:
 765           add_exclude (exclude, optarg, EXCLUDE_WILDCARDS);
 766           break;
 767
 768         case TIME_OPTION:
 769           opt_time = true;
 770           time_type =
 771             (optarg
 772              ? XARGMATCH ("--time", optarg, time_args, time_types)
 773              : time_mtime);
 774           break;
 775
 776         case TIME_STYLE_OPTION:
 777           time_style = optarg;
 778           break;
 779
 780         case_GETOPT_HELP_CHAR;
 781
 782         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 783
 784         default:
 785           ok = false;
 786         }
 787     }
 788
 789   if (!ok)
 790     usage (EXIT_FAILURE);
 791
 792   if (opt_all && opt_summarize_only)
 793     {
 794       error (0, 0, _("cannot both summarize and show all entries"));
 795       usage (EXIT_FAILURE);
 796     }
 797
 798   if (opt_summarize_only && max_depth_specified && max_depth == 0)
 799     {
 800       error (0, 0,
 801              _("warning: summarizing is the same as using --max-depth=0"));
 802     }
 803
 804   if (opt_summarize_only && max_depth_specified && max_depth != 0)
 805     {
 806       unsigned long int d = max_depth;
 807       error (0, 0, _("warning: summarizing conflicts with --max-depth=%lu"), d);
 808       usage (EXIT_FAILURE);
 809     }
 810
 811   if (opt_summarize_only)
 812     max_depth = 0;
 813
 814   /* Process time style if printing last times.  */
 815   if (opt_time)
 816     {
 817       if (! time_style)
 818         {
 819           time_style = getenv ("TIME_STYLE");
 820
 821           /* Ignore TIMESTYLE="locale", for compatibility with ls.  */
 822           if (! time_style || STREQ (time_style, "locale"))
 823             time_style = "long-iso";
 824           else if (*time_style == '+')
 825             {
 826               /* Ignore anything after a newline, for compatibility
 827                  with ls.  */
 828               char *p = strchr (time_style, '\n');
 829               if (p)
 830                 *p = '\0';
 831             }
 832           else
 833             {
 834               /* Ignore "posix-" prefix, for compatibility with ls.  */
 835               static char const posix_prefix[] = "posix-";
 836               while (strncmp (time_style, posix_prefix, sizeof posix_prefix - 1)
 837                      == 0)
 838                 time_style += sizeof posix_prefix - 1;
 839             }
 840         }
 841
 842       if (*time_style == '+')
 843         time_format = time_style + 1;
 844       else
 845         {
 846           switch (XARGMATCH ("time style", time_style,
 847                              time_style_args, time_style_types))
 848             {
 849             case full_iso_time_style:
 850               time_format = "%Y-%m-%d %H:%M:%S.%N %z";
 851               break;
 852
 853             case long_iso_time_style:
 854               time_format = "%Y-%m-%d %H:%M";
 855               break;
 856
 857             case iso_time_style:
 858               time_format = "%Y-%m-%d";
 859               break;
 860             }
 861         }
 862     }
 863
 864   struct argv_iterator *ai;
 865   if (files_from)
 866     {
 867       /* When using --files0-from=F, you may not specify any files
 868          on the command-line.  */
 869       if (optind < argc)
 870         {
 871           error (0, 0, _("extra operand %s"), quote (argv[optind]));
 872           fprintf (stderr, "%s\n",
 873                    _("file operands cannot be combined with --files0-from"));
 874           usage (EXIT_FAILURE);
 875         }
 876
 877       if (! (STREQ (files_from, "-") || freopen (files_from, "r", stdin)))
 878         error (EXIT_FAILURE, errno, _("cannot open %s for reading"),
 879                quote (files_from));
 880
 881       ai = argv_iter_init_stream (stdin);
 882
 883       /* It's not easy here to count the arguments, so assume the
 884          worst.  */
 885       hash_all = true;
 886     }
 887   else
 888     {
 889       char **files = (optind < argc ? argv + optind : cwd_only);
 890       ai = argv_iter_init_argv (files);
 891
 892       /* Hash all dev,ino pairs if there are multiple arguments, or if
 893          following non-command-line symlinks, because in either case a
 894          file with just one hard link might be seen more than once.  */
 895       hash_all = (optind + 1 < argc || symlink_deref_bits == FTS_LOGICAL);
 896     }
 897
 898   if (!ai)
 899     xalloc_die ();
 900
 901   /* Initialize the set of dev,inode pairs.  */
 902   di_set = di_set_alloc ();
 903   if (!di_set)
 904     xalloc_die ();
 905
 906   /* If not hashing everything, process_file won't find cycles on its
 907      own, so ask fts_read to check for them accurately.  */
 908   if (opt_count_all || ! hash_all)
 909     bit_flags |= FTS_TIGHT_CYCLE_CHECK;
 910
 911   bit_flags |= symlink_deref_bits;
 912   static char *temp_argv[] = { NULL, NULL };
 913
 914   while (true)
 915     {
 916       bool skip_file = false;
 917       enum argv_iter_err ai_err;
 918       char *file_name = argv_iter (ai, &ai_err);
 919       if (ai_err == AI_ERR_EOF)
 920         break;
 921       if (!file_name)
 922         {
 923           switch (ai_err)
 924             {
 925             case AI_ERR_READ:
 926               error (0, errno, _("%s: read error"), quote (files_from));
 927               continue;
 928
 929             case AI_ERR_MEM:
 930               xalloc_die ();
 931
 932             default:
 933               assert (!"unexpected error code from argv_iter");
 934             }
 935         }
 936       if (files_from && STREQ (files_from, "-") && STREQ (file_name, "-"))
 937         {
 938           /* Give a better diagnostic in an unusual case:
 939              printf - | du --files0-from=- */
 940           error (0, 0, _("when reading file names from stdin, "
 941                          "no file name of %s allowed"),
 942                  quote (file_name));
 943           skip_file = true;
 944         }
 945
 946       /* Report and skip any empty file names before invoking fts.
 947          This works around a glitch in fts, which fails immediately
 948          (without looking at the other file names) when given an empty
 949          file name.  */
 950       if (!file_name[0])
 951         {
 952           /* Diagnose a zero-length file name.  When it's one
 953              among many, knowing the record number may help.
 954              FIXME: currently print the record number only with
 955              --files0-from=FILE.  Maybe do it for argv, too?  */
 956           if (files_from == NULL)
 957             error (0, 0, "%s", _("invalid zero-length file name"));
 958           else
 959             {
 960               /* Using the standard `filename:line-number:' prefix here is
 961                  not totally appropriate, since NUL is the separator, not NL,
 962                  but it might be better than nothing.  */
 963               unsigned long int file_number = argv_iter_n_args (ai);
 964               error (0, 0, "%s:%lu: %s", quotearg_colon (files_from),
 965                      file_number, _("invalid zero-length file name"));
 966             }
 967           skip_file = true;
 968         }
 969
 970       if (skip_file)
 971         ok = false;
 972       else
 973         {
 974           temp_argv[0] = file_name;
 975           ok &= du_files (temp_argv, bit_flags);
 976         }
 977     }
 978
 979   argv_iter_free (ai);
 980   di_set_free (di_set);
 981
 982   if (files_from && (ferror (stdin) || fclose (stdin) != 0))
 983     error (EXIT_FAILURE, 0, _("error reading %s"), quote (files_from));
 984
 985   if (print_grand_total)
 986     print_size (&tot_dui, _("total"));
 987
 988   exit (ok ? EXIT_SUCCESS : EXIT_FAILURE);
 989 }