src/split.c

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