src/split.c

   1 /* split.c -- split a file into pieces.
   2    Copyright (C) 1988, 1991, 1995-2008 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 \f
  17 /* By tege@sics.se, with rms.
  18
  19    To do:
  20    * Implement -t CHAR or -t REGEX to specify break characters other
  21      than newline. */
  22
  23 #include <config.h>
  24
  25 #include <stdio.h>
  26 #include <getopt.h>
  27 #include <sys/types.h>
  28
  29 #include "system.h"
  30 #include "error.h"
  31 #include "fd-reopen.h"
  32 #include "fcntl--.h"
  33 #include "full-read.h"
  34 #include "full-write.h"
  35 #include "quote.h"
  36 #include "safe-read.h"
  37 #include "xstrtol.h"
  38
  39 /* The official name of this program (e.g., no `g' prefix).  */
  40 #define PROGRAM_NAME "split"
  41
  42 #define AUTHORS \
  43   proper_name_utf8 ("Torbjorn Granlund", "Torbj\303\266rn Granlund"), \
  44   proper_name ("Richard M. Stallman")
  45
  46 #define DEFAULT_SUFFIX_LENGTH 2
  47
  48 /* Base name of output files.  */
  49 static char const *outbase;
  50
  51 /* Name of output files.  */
  52 static char *outfile;
  53
  54 /* Pointer to the end of the prefix in OUTFILE.
  55    Suffixes are inserted here.  */
  56 static char *outfile_mid;
  57
  58 /* Length of OUTFILE's suffix.  */
  59 static size_t suffix_length = DEFAULT_SUFFIX_LENGTH;
  60
  61 /* Alphabet of characters to use in suffix.  */
  62 static char const *suffix_alphabet = "abcdefghijklmnopqrstuvwxyz";
  63
  64 /* Name of input file.  May be "-".  */
  65 static char *infile;
  66
  67 /* Descriptor on which output file is open.  */
  68 static int output_desc;
  69
  70 /* If true, print a diagnostic on standard error just before each
  71    output file is opened. */
  72 static bool verbose;
  73
  74 /* For long options that have no equivalent short option, use a
  75    non-character as a pseudo short option, starting with CHAR_MAX + 1.  */
  76 enum
  77 {
  78   VERBOSE_OPTION = CHAR_MAX + 1
  79 };
  80
  81 static struct option const longopts[] =
  82 {
  83   {"bytes", required_argument, NULL, 'b'},
  84   {"lines", required_argument, NULL, 'l'},
  85   {"line-bytes", required_argument, NULL, 'C'},
  86   {"suffix-length", required_argument, NULL, 'a'},
  87   {"numeric-suffixes", no_argument, NULL, 'd'},
  88   {"verbose", no_argument, NULL, VERBOSE_OPTION},
  89   {GETOPT_HELP_OPTION_DECL},
  90   {GETOPT_VERSION_OPTION_DECL},
  91   {NULL, 0, NULL, 0}
  92 };
  93
  94 void
  95 usage (int status)
  96 {
  97   if (status != EXIT_SUCCESS)
  98     fprintf (stderr, _("Try `%s --help' for more information.\n"),
  99              program_name);
 100   else
 101     {
 102       printf (_("\
 103 Usage: %s [OPTION]... [INPUT [PREFIX]]\n\
 104 "),
 105               program_name);
 106     fputs (_("\
 107 Output fixed-size pieces of INPUT to PREFIXaa, PREFIXab, ...; default\n\
 108 size is 1000 lines, and default PREFIX is `x'.  With no INPUT, or when INPUT\n\
 109 is -, read standard input.\n\
 110 \n\
 111 "), stdout);
 112       fputs (_("\
 113 Mandatory arguments to long options are mandatory for short options too.\n\
 114 "), stdout);
 115       fprintf (stdout, _("\
 116   -a, --suffix-length=N   use suffixes of length N (default %d)\n\
 117   -b, --bytes=SIZE        put SIZE bytes per output file\n\
 118   -C, --line-bytes=SIZE   put at most SIZE bytes of lines per output file\n\
 119   -d, --numeric-suffixes  use numeric suffixes instead of alphabetic\n\
 120   -l, --lines=NUMBER      put NUMBER lines per output file\n\
 121 "), DEFAULT_SUFFIX_LENGTH);
 122       fputs (_("\
 123       --verbose           print a diagnostic just before each\n\
 124                             output file is opened\n\
 125 "), stdout);
 126       fputs (HELP_OPTION_DESCRIPTION, stdout);
 127       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 128       fputs (_("\
 129 \n\
 130 SIZE may have a multiplier suffix:\n\
 131 b 512, kB 1000, K 1024, MB 1000*1000, M 1024*1024,\n\
 132 GB 1000*1000*1000, G 1024*1024*1024, and so on for T, P, E, Z, Y.\n\
 133 "), stdout);
 134       emit_bug_reporting_address ();
 135     }
 136   exit (status);
 137 }
 138
 139 /* Compute the next sequential output file name and store it into the
 140    string `outfile'.  */
 141
 142 static void
 143 next_file_name (void)
 144 {
 145   /* Index in suffix_alphabet of each character in the suffix.  */
 146   static size_t *sufindex;
 147
 148   if (! outfile)
 149     {
 150       /* Allocate and initialize the first file name.  */
 151
 152       size_t outbase_length = strlen (outbase);
 153       size_t outfile_length = outbase_length + suffix_length;
 154       if (outfile_length + 1 < outbase_length)
 155         xalloc_die ();
 156       outfile = xmalloc (outfile_length + 1);
 157       outfile_mid = outfile + outbase_length;
 158       memcpy (outfile, outbase, outbase_length);
 159       memset (outfile_mid, suffix_alphabet[0], suffix_length);
 160       outfile[outfile_length] = 0;
 161       sufindex = xcalloc (suffix_length, sizeof *sufindex);
 162
 163 #if ! _POSIX_NO_TRUNC && HAVE_PATHCONF && defined _PC_NAME_MAX
 164       /* POSIX requires that if the output file name is too long for
 165          its directory, `split' must fail without creating any files.
 166          This must be checked for explicitly on operating systems that
 167          silently truncate file names.  */
 168       {
 169         char *dir = dir_name (outfile);
 170         long name_max = pathconf (dir, _PC_NAME_MAX);
 171         if (0 <= name_max && name_max < base_len (last_component (outfile)))
 172           error (EXIT_FAILURE, ENAMETOOLONG, "%s", outfile);
 173         free (dir);
 174       }
 175 #endif
 176     }
 177   else
 178     {
 179       /* Increment the suffix in place, if possible.  */
 180
 181       size_t i = suffix_length;
 182       while (i-- != 0)
 183         {
 184           sufindex[i]++;
 185           outfile_mid[i] = suffix_alphabet[sufindex[i]];
 186           if (outfile_mid[i])
 187             return;
 188           sufindex[i] = 0;
 189           outfile_mid[i] = suffix_alphabet[sufindex[i]];
 190         }
 191       error (EXIT_FAILURE, 0, _("output file suffixes exhausted"));
 192     }
 193 }
 194
 195 /* Write BYTES bytes at BP to an output file.
 196    If NEW_FILE_FLAG is true, open the next output file.
 197    Otherwise add to the same output file already in use.  */
 198
 199 static void
 200 cwrite (bool new_file_flag, const char *bp, size_t bytes)
 201 {
 202   if (new_file_flag)
 203     {
 204       if (output_desc >= 0 && close (output_desc) < 0)
 205         error (EXIT_FAILURE, errno, "%s", outfile);
 206
 207       next_file_name ();
 208       if (verbose)
 209         fprintf (stdout, _("creating file %s\n"), quote (outfile));
 210       output_desc = open (outfile,
 211                           O_WRONLY | O_CREAT | O_TRUNC | O_BINARY,
 212                           (S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP
 213                            | S_IROTH | S_IWOTH));
 214       if (output_desc < 0)
 215         error (EXIT_FAILURE, errno, "%s", outfile);
 216     }
 217   if (full_write (output_desc, bp, bytes) != bytes)
 218     error (EXIT_FAILURE, errno, "%s", outfile);
 219 }
 220
 221 /* Split into pieces of exactly N_BYTES bytes.
 222    Use buffer BUF, whose size is BUFSIZE.  */
 223
 224 static void
 225 bytes_split (uintmax_t n_bytes, char *buf, size_t bufsize)
 226 {
 227   size_t n_read;
 228   bool new_file_flag = true;
 229   size_t to_read;
 230   uintmax_t to_write = n_bytes;
 231   char *bp_out;
 232
 233   do
 234     {
 235       n_read = full_read (STDIN_FILENO, buf, bufsize);
 236       if (n_read == SAFE_READ_ERROR)
 237         error (EXIT_FAILURE, errno, "%s", infile);
 238       bp_out = buf;
 239       to_read = n_read;
 240       for (;;)
 241         {
 242           if (to_read < to_write)
 243             {
 244               if (to_read)      /* do not write 0 bytes! */
 245                 {
 246                   cwrite (new_file_flag, bp_out, to_read);
 247                   to_write -= to_read;
 248                   new_file_flag = false;
 249                 }
 250               break;
 251             }
 252           else
 253             {
 254               size_t w = to_write;
 255               cwrite (new_file_flag, bp_out, w);
 256               bp_out += w;
 257               to_read -= w;
 258               new_file_flag = true;
 259               to_write = n_bytes;
 260             }
 261         }
 262     }
 263   while (n_read == bufsize);
 264 }
 265
 266 /* Split into pieces of exactly N_LINES lines.
 267    Use buffer BUF, whose size is BUFSIZE.  */
 268
 269 static void
 270 lines_split (uintmax_t n_lines, char *buf, size_t bufsize)
 271 {
 272   size_t n_read;
 273   char *bp, *bp_out, *eob;
 274   bool new_file_flag = true;
 275   uintmax_t n = 0;
 276
 277   do
 278     {
 279       n_read = full_read (STDIN_FILENO, buf, bufsize);
 280       if (n_read == SAFE_READ_ERROR)
 281         error (EXIT_FAILURE, errno, "%s", infile);
 282       bp = bp_out = buf;
 283       eob = bp + n_read;
 284       *eob = '\n';
 285       for (;;)
 286         {
 287           bp = memchr (bp, '\n', eob - bp + 1);
 288           if (bp == eob)
 289             {
 290               if (eob != bp_out) /* do not write 0 bytes! */
 291                 {
 292                   size_t len = eob - bp_out;
 293                   cwrite (new_file_flag, bp_out, len);
 294                   new_file_flag = false;
 295                 }
 296               break;
 297             }
 298
 299           ++bp;
 300           if (++n >= n_lines)
 301             {
 302               cwrite (new_file_flag, bp_out, bp - bp_out);
 303               bp_out = bp;
 304               new_file_flag = true;
 305               n = 0;
 306             }
 307         }
 308     }
 309   while (n_read == bufsize);
 310 }
 311 \f
 312 /* Split into pieces that are as large as possible while still not more
 313    than N_BYTES bytes, and are split on line boundaries except
 314    where lines longer than N_BYTES bytes occur.
 315    FIXME: Allow N_BYTES to be any uintmax_t value, and don't require a
 316    buffer of size N_BYTES, in case N_BYTES is very large.  */
 317
 318 static void
 319 line_bytes_split (size_t n_bytes)
 320 {
 321   size_t n_read;
 322   char *bp;
 323   bool eof = false;
 324   size_t n_buffered = 0;
 325   char *buf = xmalloc (n_bytes);
 326
 327   do
 328     {
 329       /* Fill up the full buffer size from the input file.  */
 330
 331       n_read = full_read (STDIN_FILENO, buf + n_buffered, n_bytes - n_buffered);
 332       if (n_read == SAFE_READ_ERROR)
 333         error (EXIT_FAILURE, errno, "%s", infile);
 334
 335       n_buffered += n_read;
 336       if (n_buffered != n_bytes)
 337         {
 338           if (n_buffered == 0)
 339             break;
 340           eof = true;
 341         }
 342
 343       /* Find where to end this chunk.  */
 344       bp = buf + n_buffered;
 345       if (n_buffered == n_bytes)
 346         {
 347           while (bp > buf && bp[-1] != '\n')
 348             bp--;
 349         }
 350
 351       /* If chunk has no newlines, use all the chunk.  */
 352       if (bp == buf)
 353         bp = buf + n_buffered;
 354
 355       /* Output the chars as one output file.  */
 356       cwrite (true, buf, bp - buf);
 357
 358       /* Discard the chars we just output; move rest of chunk
 359          down to be the start of the next chunk.  Source and
 360          destination probably overlap.  */
 361       n_buffered -= bp - buf;
 362       if (n_buffered > 0)
 363         memmove (buf, bp, n_buffered);
 364     }
 365   while (!eof);
 366   free (buf);
 367 }
 368
 369 #define FAIL_ONLY_ONE_WAY()                                     \
 370   do                                                            \
 371     {                                                           \
 372       error (0, 0, _("cannot split in more than one way"));     \
 373       usage (EXIT_FAILURE);                                     \
 374     }                                                           \
 375   while (0)
 376
 377 int
 378 main (int argc, char **argv)
 379 {
 380   struct stat stat_buf;
 381   enum
 382     {
 383       type_undef, type_bytes, type_byteslines, type_lines, type_digits
 384     } split_type = type_undef;
 385   size_t in_blk_size;           /* optimal block size of input file device */
 386   char *buf;                    /* file i/o buffer */
 387   size_t page_size = getpagesize ();
 388   uintmax_t n_units;
 389   static char const multipliers[] = "bEGKkMmPTYZ0";
 390   int c;
 391   int digits_optind = 0;
 392
 393   initialize_main (&argc, &argv);
 394   set_program_name (argv[0]);
 395   setlocale (LC_ALL, "");
 396   bindtextdomain (PACKAGE, LOCALEDIR);
 397   textdomain (PACKAGE);
 398
 399   atexit (close_stdout);
 400
 401   /* Parse command line options.  */
 402
 403   infile = "-";
 404   outbase = "x";
 405
 406   while (1)
 407     {
 408       /* This is the argv-index of the option we will read next.  */
 409       int this_optind = optind ? optind : 1;
 410
 411       c = getopt_long (argc, argv, "0123456789C:a:b:dl:", longopts, NULL);
 412       if (c == -1)
 413         break;
 414
 415       switch (c)
 416         {
 417         case 'a':
 418           {
 419             unsigned long tmp;
 420             if (xstrtoul (optarg, NULL, 10, &tmp, "") != LONGINT_OK
 421                 || SIZE_MAX / sizeof (size_t) < tmp)
 422               {
 423                 error (0, 0, _("%s: invalid suffix length"), optarg);
 424                 usage (EXIT_FAILURE);
 425               }
 426             suffix_length = tmp;
 427           }
 428           break;
 429
 430         case 'b':
 431           if (split_type != type_undef)
 432             FAIL_ONLY_ONE_WAY ();
 433           split_type = type_bytes;
 434           if (xstrtoumax (optarg, NULL, 10, &n_units, multipliers) != LONGINT_OK
 435               || n_units == 0)
 436             {
 437               error (0, 0, _("%s: invalid number of bytes"), optarg);
 438               usage (EXIT_FAILURE);
 439             }
 440           break;
 441
 442         case 'l':
 443           if (split_type != type_undef)
 444             FAIL_ONLY_ONE_WAY ();
 445           split_type = type_lines;
 446           if (xstrtoumax (optarg, NULL, 10, &n_units, "") != LONGINT_OK
 447               || n_units == 0)
 448             {
 449               error (0, 0, _("%s: invalid number of lines"), optarg);
 450               usage (EXIT_FAILURE);
 451             }
 452           break;
 453
 454         case 'C':
 455           if (split_type != type_undef)
 456             FAIL_ONLY_ONE_WAY ();
 457           split_type = type_byteslines;
 458           if (xstrtoumax (optarg, NULL, 10, &n_units, multipliers) != LONGINT_OK
 459               || n_units == 0 || SIZE_MAX < n_units)
 460             {
 461               error (0, 0, _("%s: invalid number of bytes"), optarg);
 462               usage (EXIT_FAILURE);
 463             }
 464           break;
 465
 466         case '0':
 467         case '1':
 468         case '2':
 469         case '3':
 470         case '4':
 471         case '5':
 472         case '6':
 473         case '7':
 474         case '8':
 475         case '9':
 476           if (split_type == type_undef)
 477             {
 478               split_type = type_digits;
 479               n_units = 0;
 480             }
 481           if (split_type != type_undef && split_type != type_digits)
 482             FAIL_ONLY_ONE_WAY ();
 483           if (digits_optind != 0 && digits_optind != this_optind)
 484             n_units = 0;        /* More than one number given; ignore other. */
 485           digits_optind = this_optind;
 486           if (!DECIMAL_DIGIT_ACCUMULATE (n_units, c - '0', uintmax_t))
 487             {
 488               char buffer[INT_BUFSIZE_BOUND (uintmax_t)];
 489               error (EXIT_FAILURE, 0,
 490                      _("line count option -%s%c... is too large"),
 491                      umaxtostr (n_units, buffer), c);
 492             }
 493           break;
 494
 495         case 'd':
 496           suffix_alphabet = "0123456789";
 497           break;
 498
 499         case VERBOSE_OPTION:
 500           verbose = true;
 501           break;
 502
 503         case_GETOPT_HELP_CHAR;
 504
 505         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 506
 507         default:
 508           usage (EXIT_FAILURE);
 509         }
 510     }
 511
 512   /* Handle default case.  */
 513   if (split_type == type_undef)
 514     {
 515       split_type = type_lines;
 516       n_units = 1000;
 517     }
 518
 519   if (n_units == 0)
 520     {
 521       error (0, 0, _("invalid number of lines: 0"));
 522       usage (EXIT_FAILURE);
 523     }
 524
 525   /* Get out the filename arguments.  */
 526
 527   if (optind < argc)
 528     infile = argv[optind++];
 529
 530   if (optind < argc)
 531     outbase = argv[optind++];
 532
 533   if (optind < argc)
 534     {
 535       error (0, 0, _("extra operand %s"), quote (argv[optind]));
 536       usage (EXIT_FAILURE);
 537     }
 538
 539   /* Open the input file.  */
 540   if (! STREQ (infile, "-")
 541       && fd_reopen (STDIN_FILENO, infile, O_RDONLY, 0) < 0)
 542     error (EXIT_FAILURE, errno, _("cannot open %s for reading"),
 543            quote (infile));
 544
 545   /* Binary I/O is safer when bytecounts are used.  */
 546   if (O_BINARY && ! isatty (STDIN_FILENO))
 547     freopen (NULL, "rb", stdin);
 548
 549   /* No output file is open now.  */
 550   output_desc = -1;
 551
 552   /* Get the optimal block size of input device and make a buffer.  */
 553
 554   if (fstat (STDIN_FILENO, &stat_buf) != 0)
 555     error (EXIT_FAILURE, errno, "%s", infile);
 556   in_blk_size = ST_BLKSIZE (stat_buf);
 557
 558   buf = ptr_align (xmalloc (in_blk_size + 1 + page_size - 1), page_size);
 559
 560   switch (split_type)
 561     {
 562     case type_digits:
 563     case type_lines:
 564       lines_split (n_units, buf, in_blk_size);
 565       break;
 566
 567     case type_bytes:
 568       bytes_split (n_units, buf, in_blk_size);
 569       break;
 570
 571     case type_byteslines:
 572       line_bytes_split (n_units);
 573       break;
 574
 575     default:
 576       abort ();
 577     }
 578
 579   if (close (STDIN_FILENO) != 0)
 580     error (EXIT_FAILURE, errno, "%s", infile);
 581   if (output_desc >= 0 && close (output_desc) < 0)
 582     error (EXIT_FAILURE, errno, "%s", outfile);
 583
 584   exit (EXIT_SUCCESS);
 585 }