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       emit_blocksize_note ("DU");
 334       emit_size_note ();
 335       emit_ancillary_info ();
 336     }
 337   exit (status);
 338 }
 339
 340 static size_t
 341 entry_hash (void const *x, size_t table_size)
 342 {
 343   struct entry const *p = x;
 344
 345   /* Ignoring the device number here should be fine.  */
 346   /* The cast to uintmax_t prevents negative remainders
 347      if st_ino is negative.  */
 348   return (uintmax_t) p->st_ino % table_size;
 349 }
 350
 351 /* Compare two dev/ino pairs.  Return true if they are the same.  */
 352 static bool
 353 entry_compare (void const *x, void const *y)
 354 {
 355   struct entry const *a = x;
 356   struct entry const *b = y;
 357   return SAME_INODE (*a, *b) ? true : false;
 358 }
 359
 360 /* Try to insert the INO/DEV pair into the global table, HTAB.
 361    Return true if the pair is successfully inserted,
 362    false if the pair is already in the table.  */
 363 static bool
 364 hash_ins (ino_t ino, dev_t dev)
 365 {
 366   struct entry *ent;
 367   struct entry *ent_from_table;
 368
 369   ent = xmalloc (sizeof *ent);
 370   ent->st_ino = ino;
 371   ent->st_dev = dev;
 372
 373   ent_from_table = hash_insert (htab, ent);
 374   if (ent_from_table == NULL)
 375     {
 376       /* Insertion failed due to lack of memory.  */
 377       xalloc_die ();
 378     }
 379
 380   if (ent_from_table == ent)
 381     {
 382       /* Insertion succeeded.  */
 383       return true;
 384     }
 385
 386   /* That pair is already in the table, so ENT was not inserted.  Free it.  */
 387   free (ent);
 388
 389   return false;
 390 }
 391
 392 /* Initialize the hash table.  */
 393 static void
 394 hash_init (void)
 395 {
 396   htab = hash_initialize (INITIAL_TABLE_SIZE, NULL,
 397                           entry_hash, entry_compare, free);
 398   if (htab == NULL)
 399     xalloc_die ();
 400 }
 401
 402 /* FIXME: this code is nearly identical to code in date.c  */
 403 /* Display the date and time in WHEN according to the format specified
 404    in FORMAT.  */
 405
 406 static void
 407 show_date (const char *format, struct timespec when)
 408 {
 409   struct tm *tm = localtime (&when.tv_sec);
 410   if (! tm)
 411     {
 412       char buf[INT_BUFSIZE_BOUND (intmax_t)];
 413       error (0, 0, _("time %s is out of range"), timetostr (when.tv_sec, buf));
 414       fputs (buf, stdout);
 415       return;
 416     }
 417
 418   fprintftime (stdout, format, tm, 0, when.tv_nsec);
 419 }
 420
 421 /* Print N_BYTES.  Convert it to a readable value before printing.  */
 422
 423 static void
 424 print_only_size (uintmax_t n_bytes)
 425 {
 426   char buf[LONGEST_HUMAN_READABLE + 1];
 427   fputs (human_readable (n_bytes, buf, human_output_opts,
 428                          1, output_block_size), stdout);
 429 }
 430
 431 /* Print size (and optionally time) indicated by *PDUI, followed by STRING.  */
 432
 433 static void
 434 print_size (const struct duinfo *pdui, const char *string)
 435 {
 436   print_only_size (pdui->size);
 437   if (opt_time)
 438     {
 439       putchar ('\t');
 440       show_date (time_format, pdui->tmax);
 441     }
 442   printf ("\t%s%c", string, opt_nul_terminate_output ? '\0' : '\n');
 443   fflush (stdout);
 444 }
 445
 446 /* This function is called once for every file system object that fts
 447    encounters.  fts does a depth-first traversal.  This function knows
 448    that and accumulates per-directory totals based on changes in
 449    the depth of the current entry.  It returns true on success.  */
 450
 451 static bool
 452 process_file (FTS *fts, FTSENT *ent)
 453 {
 454   bool ok;
 455   struct duinfo dui;
 456   struct duinfo dui_to_print;
 457   size_t level;
 458   static size_t prev_level;
 459   static size_t n_alloc;
 460   /* First element of the structure contains:
 461      The sum of the st_size values of all entries in the single directory
 462      at the corresponding level.  Although this does include the st_size
 463      corresponding to each subdirectory, it does not include the size of
 464      any file in a subdirectory. Also corresponding last modified date.
 465      Second element of the structure contains:
 466      The sum of the sizes of all entries in the hierarchy at or below the
 467      directory at the specified level.  */
 468   static struct dulevel *dulvl;
 469   bool print = true;
 470
 471   const char *file = ent->fts_path;
 472   const struct stat *sb = ent->fts_statp;
 473   bool skip;
 474
 475   /* If necessary, set FTS_SKIP before returning.  */
 476   skip = excluded_file_name (exclude, file);
 477   if (skip)
 478     fts_set (fts, ent, FTS_SKIP);
 479
 480   switch (ent->fts_info)
 481     {
 482     case FTS_NS:
 483       error (0, ent->fts_errno, _("cannot access %s"), quote (file));
 484       return false;
 485
 486     case FTS_ERR:
 487       /* if (S_ISDIR (ent->fts_statp->st_mode) && FIXME */
 488       error (0, ent->fts_errno, _("%s"), quote (file));
 489       return false;
 490
 491     case FTS_DNR:
 492       /* Don't return just yet, since although the directory is not readable,
 493          we were able to stat it, so we do have a size.  */
 494       error (0, ent->fts_errno, _("cannot read directory %s"), quote (file));
 495       ok = false;
 496       break;
 497
 498     default:
 499       ok = true;
 500       break;
 501     }
 502
 503   /* If this is the first (pre-order) encounter with a directory,
 504      or if it's the second encounter for a skipped directory, then
 505      return right away.  */
 506   if (ent->fts_info == FTS_D || skip)
 507     return ok;
 508
 509   /* If the file is being excluded or if it has already been counted
 510      via a hard link, then don't let it contribute to the sums.  */
 511   if (skip
 512       || (!opt_count_all
 513           && ! S_ISDIR (sb->st_mode)
 514           && 1 < sb->st_nlink
 515           && ! hash_ins (sb->st_ino, sb->st_dev)))
 516     {
 517       /* Note that we must not simply return here.
 518          We still have to update prev_level and maybe propagate
 519          some sums up the hierarchy.  */
 520       duinfo_init (&dui);
 521       print = false;
 522     }
 523   else
 524     {
 525       duinfo_set (&dui,
 526                   (apparent_size
 527                    ? sb->st_size
 528                    : (uintmax_t) ST_NBLOCKS (*sb) * ST_NBLOCKSIZE),
 529                   (time_type == time_mtime ? get_stat_mtime (sb)
 530                    : time_type == time_atime ? get_stat_atime (sb)
 531                    : get_stat_ctime (sb)));
 532     }
 533
 534   level = ent->fts_level;
 535   dui_to_print = dui;
 536
 537   if (n_alloc == 0)
 538     {
 539       n_alloc = level + 10;
 540       dulvl = xcalloc (n_alloc, sizeof *dulvl);
 541     }
 542   else
 543     {
 544       if (level == prev_level)
 545         {
 546           /* This is usually the most common case.  Do nothing.  */
 547         }
 548       else if (level > prev_level)
 549         {
 550           /* Descending the hierarchy.
 551              Clear the accumulators for *all* levels between prev_level
 552              and the current one.  The depth may change dramatically,
 553              e.g., from 1 to 10.  */
 554           size_t i;
 555
 556           if (n_alloc <= level)
 557             {
 558               dulvl = xnrealloc (dulvl, level, 2 * sizeof *dulvl);
 559               n_alloc = level * 2;
 560             }
 561
 562           for (i = prev_level + 1; i <= level; i++)
 563             {
 564               duinfo_init (&dulvl[i].ent);
 565               duinfo_init (&dulvl[i].subdir);
 566             }
 567         }
 568       else /* level < prev_level */
 569         {
 570           /* Ascending the hierarchy.
 571              Process a directory only after all entries in that
 572              directory have been processed.  When the depth decreases,
 573              propagate sums from the children (prev_level) to the parent.
 574              Here, the current level is always one smaller than the
 575              previous one.  */
 576           assert (level == prev_level - 1);
 577           duinfo_add (&dui_to_print, &dulvl[prev_level].ent);
 578           if (!opt_separate_dirs)
 579             duinfo_add (&dui_to_print, &dulvl[prev_level].subdir);
 580           duinfo_add (&dulvl[level].subdir, &dulvl[prev_level].ent);
 581           duinfo_add (&dulvl[level].subdir, &dulvl[prev_level].subdir);
 582         }
 583     }
 584
 585   prev_level = level;
 586
 587   /* Let the size of a directory entry contribute to the total for the
 588      containing directory, unless --separate-dirs (-S) is specified.  */
 589   if ( ! (opt_separate_dirs && IS_DIR_TYPE (ent->fts_info)))
 590     duinfo_add (&dulvl[level].ent, &dui);
 591
 592   /* Even if this directory is unreadable or we can't chdir into it,
 593      do let its size contribute to the total. */
 594   duinfo_add (&tot_dui, &dui);
 595
 596   /* If we're not counting an entry, e.g., because it's a hard link
 597      to a file we've already counted (and --count-links), then don't
 598      print a line for it.  */
 599   if (!print)
 600     return ok;
 601
 602   if ((IS_DIR_TYPE (ent->fts_info) && level <= max_depth)
 603       || ((opt_all && level <= max_depth) || level == 0))
 604     print_size (&dui_to_print, file);
 605
 606   return ok;
 607 }
 608
 609 /* Recursively print the sizes of the directories (and, if selected, files)
 610    named in FILES, the last entry of which is NULL.
 611    BIT_FLAGS controls how fts works.
 612    Return true if successful.  */
 613
 614 static bool
 615 du_files (char **files, int bit_flags)
 616 {
 617   bool ok = true;
 618
 619   if (*files)
 620     {
 621       FTS *fts = xfts_open (files, bit_flags, NULL);
 622
 623       while (1)
 624         {
 625           FTSENT *ent;
 626
 627           ent = fts_read (fts);
 628           if (ent == NULL)
 629             {
 630               if (errno != 0)
 631                 {
 632                   /* FIXME: try to give a better message  */
 633                   error (0, errno, _("fts_read failed"));
 634                   ok = false;
 635                 }
 636               break;
 637             }
 638           FTS_CROSS_CHECK (fts);
 639
 640           ok &= process_file (fts, ent);
 641         }
 642
 643       if (fts_close (fts) != 0)
 644         {
 645           error (0, errno, _("fts_close failed"));
 646           ok = false;
 647         }
 648     }
 649
 650   return ok;
 651 }
 652
 653 int
 654 main (int argc, char **argv)
 655 {
 656   char *cwd_only[2];
 657   bool max_depth_specified = false;
 658   bool ok = true;
 659   char *files_from = NULL;
 660
 661   /* Bit flags that control how fts works.  */
 662   int bit_flags = FTS_TIGHT_CYCLE_CHECK | FTS_DEFER_STAT;
 663
 664   /* Select one of the three FTS_ options that control if/when
 665      to follow a symlink.  */
 666   int symlink_deref_bits = FTS_PHYSICAL;
 667
 668   /* If true, display only a total for each argument. */
 669   bool opt_summarize_only = false;
 670
 671   cwd_only[0] = bad_cast (".");
 672   cwd_only[1] = NULL;
 673
 674   initialize_main (&argc, &argv);
 675   set_program_name (argv[0]);
 676   setlocale (LC_ALL, "");
 677   bindtextdomain (PACKAGE, LOCALEDIR);
 678   textdomain (PACKAGE);
 679
 680   atexit (close_stdout);
 681
 682   exclude = new_exclude ();
 683
 684   human_options (getenv ("DU_BLOCK_SIZE"),
 685                  &human_output_opts, &output_block_size);
 686
 687   for (;;)
 688     {
 689       int oi = -1;
 690       int c = getopt_long (argc, argv, DEBUG_OPT "0abchHklmsxB:DLPSX:",
 691                            long_options, &oi);
 692       if (c == -1)
 693         break;
 694
 695       switch (c)
 696         {
 697 #if DU_DEBUG
 698         case 'd':
 699           fts_debug = true;
 700           break;
 701 #endif
 702
 703         case '0':
 704           opt_nul_terminate_output = true;
 705           break;
 706
 707         case 'a':
 708           opt_all = true;
 709           break;
 710
 711         case APPARENT_SIZE_OPTION:
 712           apparent_size = true;
 713           break;
 714
 715         case 'b':
 716           apparent_size = true;
 717           human_output_opts = 0;
 718           output_block_size = 1;
 719           break;
 720
 721         case 'c':
 722           print_grand_total = true;
 723           break;
 724
 725         case 'h':
 726           human_output_opts = human_autoscale | human_SI | human_base_1024;
 727           output_block_size = 1;
 728           break;
 729
 730         case HUMAN_SI_OPTION:
 731           human_output_opts = human_autoscale | human_SI;
 732           output_block_size = 1;
 733           break;
 734
 735         case 'k':
 736           human_output_opts = 0;
 737           output_block_size = 1024;
 738           break;
 739
 740         case MAX_DEPTH_OPTION:          /* --max-depth=N */
 741           {
 742             unsigned long int tmp_ulong;
 743             if (xstrtoul (optarg, NULL, 0, &tmp_ulong, NULL) == LONGINT_OK
 744                 && tmp_ulong <= SIZE_MAX)
 745               {
 746                 max_depth_specified = true;
 747                 max_depth = tmp_ulong;
 748               }
 749             else
 750               {
 751                 error (0, 0, _("invalid maximum depth %s"),
 752                        quote (optarg));
 753                 ok = false;
 754               }
 755           }
 756           break;
 757
 758         case MEGABYTES_LONG_OPTION: /* FIXME: remove in 2009 */
 759           error (0, 0,
 760                  _("the --megabytes option is deprecated; use -m instead"));
 761           /* fall through */
 762         case 'm':
 763           human_output_opts = 0;
 764           output_block_size = 1024 * 1024;
 765           break;
 766
 767         case 'l':
 768           opt_count_all = true;
 769           break;
 770
 771         case 's':
 772           opt_summarize_only = true;
 773           break;
 774
 775         case 'x':
 776           bit_flags |= FTS_XDEV;
 777           break;
 778
 779         case 'B':
 780           {
 781             enum strtol_error e = human_options (optarg, &human_output_opts,
 782                                                  &output_block_size);
 783             if (e != LONGINT_OK)
 784               xstrtol_fatal (e, oi, c, long_options, optarg);
 785           }
 786           break;
 787
 788         case 'H':  /* NOTE: before 2008-12, -H was equivalent to --si.  */
 789         case 'D':
 790           symlink_deref_bits = FTS_COMFOLLOW | FTS_PHYSICAL;
 791           break;
 792
 793         case 'L': /* --dereference */
 794           symlink_deref_bits = FTS_LOGICAL;
 795           break;
 796
 797         case 'P': /* --no-dereference */
 798           symlink_deref_bits = FTS_PHYSICAL;
 799           break;
 800
 801         case 'S':
 802           opt_separate_dirs = true;
 803           break;
 804
 805         case 'X':
 806           if (add_exclude_file (add_exclude, exclude, optarg,
 807                                 EXCLUDE_WILDCARDS, '\n'))
 808             {
 809               error (0, errno, "%s", quotearg_colon (optarg));
 810               ok = false;
 811             }
 812           break;
 813
 814         case FILES0_FROM_OPTION:
 815           files_from = optarg;
 816           break;
 817
 818         case EXCLUDE_OPTION:
 819           add_exclude (exclude, optarg, EXCLUDE_WILDCARDS);
 820           break;
 821
 822         case TIME_OPTION:
 823           opt_time = true;
 824           time_type =
 825             (optarg
 826              ? XARGMATCH ("--time", optarg, time_args, time_types)
 827              : time_mtime);
 828           break;
 829
 830         case TIME_STYLE_OPTION:
 831           time_style = optarg;
 832           break;
 833
 834         case_GETOPT_HELP_CHAR;
 835
 836         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 837
 838         default:
 839           ok = false;
 840         }
 841     }
 842
 843   if (!ok)
 844     usage (EXIT_FAILURE);
 845
 846   if (opt_all && opt_summarize_only)
 847     {
 848       error (0, 0, _("cannot both summarize and show all entries"));
 849       usage (EXIT_FAILURE);
 850     }
 851
 852   if (opt_summarize_only && max_depth_specified && max_depth == 0)
 853     {
 854       error (0, 0,
 855              _("warning: summarizing is the same as using --max-depth=0"));
 856     }
 857
 858   if (opt_summarize_only && max_depth_specified && max_depth != 0)
 859     {
 860       unsigned long int d = max_depth;
 861       error (0, 0, _("warning: summarizing conflicts with --max-depth=%lu"), d);
 862       usage (EXIT_FAILURE);
 863     }
 864
 865   if (opt_summarize_only)
 866     max_depth = 0;
 867
 868   /* Process time style if printing last times.  */
 869   if (opt_time)
 870     {
 871       if (! time_style)
 872         {
 873           time_style = getenv ("TIME_STYLE");
 874
 875           /* Ignore TIMESTYLE="locale", for compatibility with ls.  */
 876           if (! time_style || STREQ (time_style, "locale"))
 877             time_style = "long-iso";
 878           else if (*time_style == '+')
 879             {
 880               /* Ignore anything after a newline, for compatibility
 881                  with ls.  */
 882               char *p = strchr (time_style, '\n');
 883               if (p)
 884                 *p = '\0';
 885             }
 886           else
 887             {
 888               /* Ignore "posix-" prefix, for compatibility with ls.  */
 889               static char const posix_prefix[] = "posix-";
 890               while (strncmp (time_style, posix_prefix, sizeof posix_prefix - 1)
 891                      == 0)
 892                 time_style += sizeof posix_prefix - 1;
 893             }
 894         }
 895
 896       if (*time_style == '+')
 897         time_format = time_style + 1;
 898       else
 899         {
 900           switch (XARGMATCH ("time style", time_style,
 901                              time_style_args, time_style_types))
 902             {
 903             case full_iso_time_style:
 904               time_format = "%Y-%m-%d %H:%M:%S.%N %z";
 905               break;
 906
 907             case long_iso_time_style:
 908               time_format = "%Y-%m-%d %H:%M";
 909               break;
 910
 911             case iso_time_style:
 912               time_format = "%Y-%m-%d";
 913               break;
 914             }
 915         }
 916     }
 917
 918   struct argv_iterator *ai;
 919   if (files_from)
 920     {
 921       /* When using --files0-from=F, you may not specify any files
 922          on the command-line.  */
 923       if (optind < argc)
 924         {
 925           error (0, 0, _("extra operand %s"), quote (argv[optind]));
 926           fprintf (stderr, "%s\n",
 927                    _("file operands cannot be combined with --files0-from"));
 928           usage (EXIT_FAILURE);
 929         }
 930
 931       if (! (STREQ (files_from, "-") || freopen (files_from, "r", stdin)))
 932         error (EXIT_FAILURE, errno, _("cannot open %s for reading"),
 933                quote (files_from));
 934
 935       ai = argv_iter_init_stream (stdin);
 936     }
 937   else
 938     {
 939       char **files = (optind < argc ? argv + optind : cwd_only);
 940       ai = argv_iter_init_argv (files);
 941     }
 942
 943   if (!ai)
 944     xalloc_die ();
 945
 946   /* Initialize the hash structure for inode numbers.  */
 947   hash_init ();
 948
 949   bit_flags |= symlink_deref_bits;
 950   static char *temp_argv[] = { NULL, NULL };
 951
 952   while (true)
 953     {
 954       bool skip_file = false;
 955       enum argv_iter_err ai_err;
 956       char *file_name = argv_iter (ai, &ai_err);
 957       if (ai_err == AI_ERR_EOF)
 958         break;
 959       if (!file_name)
 960         {
 961           switch (ai_err)
 962             {
 963             case AI_ERR_READ:
 964               error (0, errno, _("%s: read error"), quote (files_from));
 965               continue;
 966
 967             case AI_ERR_MEM:
 968               xalloc_die ();
 969
 970             default:
 971               assert (!"unexpected error code from argv_iter");
 972             }
 973         }
 974       if (files_from && STREQ (files_from, "-") && STREQ (file_name, "-"))
 975         {
 976           /* Give a better diagnostic in an unusual case:
 977              printf - | du --files0-from=- */
 978           error (0, 0, _("when reading file names from stdin, "
 979                          "no file name of %s allowed"),
 980                  quote (file_name));
 981           skip_file = true;
 982         }
 983
 984       /* Report and skip any empty file names before invoking fts.
 985          This works around a glitch in fts, which fails immediately
 986          (without looking at the other file names) when given an empty
 987          file name.  */
 988       if (!file_name[0])
 989         {
 990           /* Diagnose a zero-length file name.  When it's one
 991              among many, knowing the record number may help.
 992              FIXME: currently print the record number only with
 993              --files0-from=FILE.  Maybe do it for argv, too?  */
 994           if (files_from == NULL)
 995             error (0, 0, "%s", _("invalid zero-length file name"));
 996           else
 997             {
 998               /* Using the standard `filename:line-number:' prefix here is
 999                  not totally appropriate, since NUL is the separator, not NL,
1000                  but it might be better than nothing.  */
1001               unsigned long int file_number = argv_iter_n_args (ai);
1002               error (0, 0, "%s:%lu: %s", quotearg_colon (files_from),
1003                      file_number, _("invalid zero-length file name"));
1004             }
1005           skip_file = true;
1006         }
1007
1008       if (skip_file)
1009         ok = false;
1010       else
1011         {
1012           temp_argv[0] = file_name;
1013           ok &= du_files (temp_argv, bit_flags);
1014         }
1015     }
1016
1017   argv_iter_free (ai);
1018
1019   if (files_from && (ferror (stdin) || fclose (stdin) != 0))
1020     error (EXIT_FAILURE, 0, _("error reading %s"), quote (files_from));
1021
1022   if (print_grand_total)
1023     print_size (&tot_dui, _("total"));
1024
1025   hash_free (htab);
1026
1027   exit (ok ? EXIT_SUCCESS : EXIT_FAILURE);
1028 }