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       fputs (_("\
 130 \n\
 131 SIZE may have a multiplier suffix:\n\
 132 b 512, kB 1000, K 1024, MB 1000*1000, M 1024*1024,\n\
 133 GB 1000*1000*1000, G 1024*1024*1024, and so on for T, P, E, Z, Y.\n\
 134 "), stdout);
 135       emit_bug_reporting_address ();
 136     }
 137   exit (status);
 138 }
 139
 140 /* Compute the next sequential output file name and store it into the
 141    string `outfile'.  */
 142
 143 static void
 144 next_file_name (void)
 145 {
 146   /* Index in suffix_alphabet of each character in the suffix.  */
 147   static size_t *sufindex;
 148
 149   if (! outfile)
 150     {
 151       /* Allocate and initialize the first file name.  */
 152
 153       size_t outbase_length = strlen (outbase);
 154       size_t outfile_length = outbase_length + suffix_length;
 155       if (outfile_length + 1 < outbase_length)
 156         xalloc_die ();
 157       outfile = xmalloc (outfile_length + 1);
 158       outfile_mid = outfile + outbase_length;
 159       memcpy (outfile, outbase, outbase_length);
 160       memset (outfile_mid, suffix_alphabet[0], suffix_length);
 161       outfile[outfile_length] = 0;
 162       sufindex = xcalloc (suffix_length, sizeof *sufindex);
 163
 164 #if ! _POSIX_NO_TRUNC && HAVE_PATHCONF && defined _PC_NAME_MAX
 165       /* POSIX requires that if the output file name is too long for
 166          its directory, `split' must fail without creating any files.
 167          This must be checked for explicitly on operating systems that
 168          silently truncate file names.  */
 169       {
 170         char *dir = dir_name (outfile);
 171         long name_max = pathconf (dir, _PC_NAME_MAX);
 172         if (0 <= name_max && name_max < base_len (last_component (outfile)))
 173           error (EXIT_FAILURE, ENAMETOOLONG, "%s", outfile);
 174         free (dir);
 175       }
 176 #endif
 177     }
 178   else
 179     {
 180       /* Increment the suffix in place, if possible.  */
 181
 182       size_t i = suffix_length;
 183       while (i-- != 0)
 184         {
 185           sufindex[i]++;
 186           outfile_mid[i] = suffix_alphabet[sufindex[i]];
 187           if (outfile_mid[i])
 188             return;
 189           sufindex[i] = 0;
 190           outfile_mid[i] = suffix_alphabet[sufindex[i]];
 191         }
 192       error (EXIT_FAILURE, 0, _("output file suffixes exhausted"));
 193     }
 194 }
 195
 196 /* Write BYTES bytes at BP to an output file.
 197    If NEW_FILE_FLAG is true, open the next output file.
 198    Otherwise add to the same output file already in use.  */
 199
 200 static void
 201 cwrite (bool new_file_flag, const char *bp, size_t bytes)
 202 {
 203   if (new_file_flag)
 204     {
 205       if (output_desc >= 0 && close (output_desc) < 0)
 206         error (EXIT_FAILURE, errno, "%s", outfile);
 207
 208       next_file_name ();
 209       if (verbose)
 210         fprintf (stdout, _("creating file %s\n"), quote (outfile));
 211       output_desc = open (outfile,
 212                           O_WRONLY | O_CREAT | O_TRUNC | O_BINARY,
 213                           (S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP
 214                            | S_IROTH | S_IWOTH));
 215       if (output_desc < 0)
 216         error (EXIT_FAILURE, errno, "%s", outfile);
 217     }
 218   if (full_write (output_desc, bp, bytes) != bytes)
 219     error (EXIT_FAILURE, errno, "%s", outfile);
 220 }
 221
 222 /* Split into pieces of exactly N_BYTES bytes.
 223    Use buffer BUF, whose size is BUFSIZE.  */
 224
 225 static void
 226 bytes_split (uintmax_t n_bytes, char *buf, size_t bufsize)
 227 {
 228   size_t n_read;
 229   bool new_file_flag = true;
 230   size_t to_read;
 231   uintmax_t to_write = n_bytes;
 232   char *bp_out;
 233
 234   do
 235     {
 236       n_read = full_read (STDIN_FILENO, buf, bufsize);
 237       if (n_read == SAFE_READ_ERROR)
 238         error (EXIT_FAILURE, errno, "%s", infile);
 239       bp_out = buf;
 240       to_read = n_read;
 241       for (;;)
 242         {
 243           if (to_read < to_write)
 244             {
 245               if (to_read)      /* do not write 0 bytes! */
 246                 {
 247                   cwrite (new_file_flag, bp_out, to_read);
 248                   to_write -= to_read;
 249                   new_file_flag = false;
 250                 }
 251               break;
 252             }
 253           else
 254             {
 255               size_t w = to_write;
 256               cwrite (new_file_flag, bp_out, w);
 257               bp_out += w;
 258               to_read -= w;
 259               new_file_flag = true;
 260               to_write = n_bytes;
 261             }
 262         }
 263     }
 264   while (n_read == bufsize);
 265 }
 266
 267 /* Split into pieces of exactly N_LINES lines.
 268    Use buffer BUF, whose size is BUFSIZE.  */
 269
 270 static void
 271 lines_split (uintmax_t n_lines, char *buf, size_t bufsize)
 272 {
 273   size_t n_read;
 274   char *bp, *bp_out, *eob;
 275   bool new_file_flag = true;
 276   uintmax_t n = 0;
 277
 278   do
 279     {
 280       n_read = full_read (STDIN_FILENO, buf, bufsize);
 281       if (n_read == SAFE_READ_ERROR)
 282         error (EXIT_FAILURE, errno, "%s", infile);
 283       bp = bp_out = buf;
 284       eob = bp + n_read;
 285       *eob = '\n';
 286       for (;;)
 287         {
 288           bp = memchr (bp, '\n', eob - bp + 1);
 289           if (bp == eob)
 290             {
 291               if (eob != bp_out) /* do not write 0 bytes! */
 292                 {
 293                   size_t len = eob - bp_out;
 294                   cwrite (new_file_flag, bp_out, len);
 295                   new_file_flag = false;
 296                 }
 297               break;
 298             }
 299
 300           ++bp;
 301           if (++n >= n_lines)
 302             {
 303               cwrite (new_file_flag, bp_out, bp - bp_out);
 304               bp_out = bp;
 305               new_file_flag = true;
 306               n = 0;
 307             }
 308         }
 309     }
 310   while (n_read == bufsize);
 311 }
 312 \f
 313 /* Split into pieces that are as large as possible while still not more
 314    than N_BYTES bytes, and are split on line boundaries except
 315    where lines longer than N_BYTES bytes occur.
 316    FIXME: Allow N_BYTES to be any uintmax_t value, and don't require a
 317    buffer of size N_BYTES, in case N_BYTES is very large.  */
 318
 319 static void
 320 line_bytes_split (size_t n_bytes)
 321 {
 322   size_t n_read;
 323   char *bp;
 324   bool eof = false;
 325   size_t n_buffered = 0;
 326   char *buf = xmalloc (n_bytes);
 327
 328   do
 329     {
 330       /* Fill up the full buffer size from the input file.  */
 331
 332       n_read = full_read (STDIN_FILENO, buf + n_buffered, n_bytes - n_buffered);
 333       if (n_read == SAFE_READ_ERROR)
 334         error (EXIT_FAILURE, errno, "%s", infile);
 335
 336       n_buffered += n_read;
 337       if (n_buffered != n_bytes)
 338         {
 339           if (n_buffered == 0)
 340             break;
 341           eof = true;
 342         }
 343
 344       /* Find where to end this chunk.  */
 345       bp = buf + n_buffered;
 346       if (n_buffered == n_bytes)
 347         {
 348           while (bp > buf && bp[-1] != '\n')
 349             bp--;
 350         }
 351
 352       /* If chunk has no newlines, use all the chunk.  */
 353       if (bp == buf)
 354         bp = buf + n_buffered;
 355
 356       /* Output the chars as one output file.  */
 357       cwrite (true, buf, bp - buf);
 358
 359       /* Discard the chars we just output; move rest of chunk
 360          down to be the start of the next chunk.  Source and
 361          destination probably overlap.  */
 362       n_buffered -= bp - buf;
 363       if (n_buffered > 0)
 364         memmove (buf, bp, n_buffered);
 365     }
 366   while (!eof);
 367   free (buf);
 368 }
 369
 370 #define FAIL_ONLY_ONE_WAY()                                     \
 371   do                                                            \
 372     {                                                           \
 373       error (0, 0, _("cannot split in more than one way"));     \
 374       usage (EXIT_FAILURE);                                     \
 375     }                                                           \
 376   while (0)
 377
 378 int
 379 main (int argc, char **argv)
 380 {
 381   struct stat stat_buf;
 382   enum
 383     {
 384       type_undef, type_bytes, type_byteslines, type_lines, type_digits
 385     } split_type = type_undef;
 386   size_t in_blk_size;           /* optimal block size of input file device */
 387   char *buf;                    /* file i/o buffer */
 388   size_t page_size = getpagesize ();
 389   uintmax_t n_units;
 390   static char const multipliers[] = "bEGKkMmPTYZ0";
 391   int c;
 392   int digits_optind = 0;
 393
 394   initialize_main (&argc, &argv);
 395   set_program_name (argv[0]);
 396   setlocale (LC_ALL, "");
 397   bindtextdomain (PACKAGE, LOCALEDIR);
 398   textdomain (PACKAGE);
 399
 400   atexit (close_stdout);
 401
 402   /* Parse command line options.  */
 403
 404   infile = bad_cast ( "-");
 405   outbase = bad_cast ("x");
 406
 407   while (1)
 408     {
 409       /* This is the argv-index of the option we will read next.  */
 410       int this_optind = optind ? optind : 1;
 411
 412       c = getopt_long (argc, argv, "0123456789C:a:b:dl:", longopts, NULL);
 413       if (c == -1)
 414         break;
 415
 416       switch (c)
 417         {
 418         case 'a':
 419           {
 420             unsigned long tmp;
 421             if (xstrtoul (optarg, NULL, 10, &tmp, "") != LONGINT_OK
 422                 || SIZE_MAX / sizeof (size_t) < tmp)
 423               {
 424                 error (0, 0, _("%s: invalid suffix length"), optarg);
 425                 usage (EXIT_FAILURE);
 426               }
 427             suffix_length = tmp;
 428           }
 429           break;
 430
 431         case 'b':
 432           if (split_type != type_undef)
 433             FAIL_ONLY_ONE_WAY ();
 434           split_type = type_bytes;
 435           if (xstrtoumax (optarg, NULL, 10, &n_units, multipliers) != LONGINT_OK
 436               || n_units == 0)
 437             {
 438               error (0, 0, _("%s: invalid number of bytes"), optarg);
 439               usage (EXIT_FAILURE);
 440             }
 441           break;
 442
 443         case 'l':
 444           if (split_type != type_undef)
 445             FAIL_ONLY_ONE_WAY ();
 446           split_type = type_lines;
 447           if (xstrtoumax (optarg, NULL, 10, &n_units, "") != LONGINT_OK
 448               || n_units == 0)
 449             {
 450               error (0, 0, _("%s: invalid number of lines"), optarg);
 451               usage (EXIT_FAILURE);
 452             }
 453           break;
 454
 455         case 'C':
 456           if (split_type != type_undef)
 457             FAIL_ONLY_ONE_WAY ();
 458           split_type = type_byteslines;
 459           if (xstrtoumax (optarg, NULL, 10, &n_units, multipliers) != LONGINT_OK
 460               || n_units == 0 || SIZE_MAX < n_units)
 461             {
 462               error (0, 0, _("%s: invalid number of bytes"), optarg);
 463               usage (EXIT_FAILURE);
 464             }
 465           break;
 466
 467         case '0':
 468         case '1':
 469         case '2':
 470         case '3':
 471         case '4':
 472         case '5':
 473         case '6':
 474         case '7':
 475         case '8':
 476         case '9':
 477           if (split_type == type_undef)
 478             {
 479               split_type = type_digits;
 480               n_units = 0;
 481             }
 482           if (split_type != type_undef && split_type != type_digits)
 483             FAIL_ONLY_ONE_WAY ();
 484           if (digits_optind != 0 && digits_optind != this_optind)
 485             n_units = 0;        /* More than one number given; ignore other. */
 486           digits_optind = this_optind;
 487           if (!DECIMAL_DIGIT_ACCUMULATE (n_units, c - '0', uintmax_t))
 488             {
 489               char buffer[INT_BUFSIZE_BOUND (uintmax_t)];
 490               error (EXIT_FAILURE, 0,
 491                      _("line count option -%s%c... is too large"),
 492                      umaxtostr (n_units, buffer), c);
 493             }
 494           break;
 495
 496         case 'd':
 497           suffix_alphabet = "0123456789";
 498           break;
 499
 500         case VERBOSE_OPTION:
 501           verbose = true;
 502           break;
 503
 504         case_GETOPT_HELP_CHAR;
 505
 506         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 507
 508         default:
 509           usage (EXIT_FAILURE);
 510         }
 511     }
 512
 513   /* Handle default case.  */
 514   if (split_type == type_undef)
 515     {
 516       split_type = type_lines;
 517       n_units = 1000;
 518     }
 519
 520   if (n_units == 0)
 521     {
 522       error (0, 0, _("invalid number of lines: 0"));
 523       usage (EXIT_FAILURE);
 524     }
 525
 526   /* Get out the filename arguments.  */
 527
 528   if (optind < argc)
 529     infile = argv[optind++];
 530
 531   if (optind < argc)
 532     outbase = argv[optind++];
 533
 534   if (optind < argc)
 535     {
 536       error (0, 0, _("extra operand %s"), quote (argv[optind]));
 537       usage (EXIT_FAILURE);
 538     }
 539
 540   /* Open the input file.  */
 541   if (! STREQ (infile, "-")
 542       && fd_reopen (STDIN_FILENO, infile, O_RDONLY, 0) < 0)
 543     error (EXIT_FAILURE, errno, _("cannot open %s for reading"),
 544            quote (infile));
 545
 546   /* Binary I/O is safer when bytecounts are used.  */
 547   if (O_BINARY && ! isatty (STDIN_FILENO))
 548     xfreopen (NULL, "rb", stdin);
 549
 550   /* No output file is open now.  */
 551   output_desc = -1;
 552
 553   /* Get the optimal block size of input device and make a buffer.  */
 554
 555   if (fstat (STDIN_FILENO, &stat_buf) != 0)
 556     error (EXIT_FAILURE, errno, "%s", infile);
 557   in_blk_size = io_blksize (stat_buf);
 558
 559   buf = ptr_align (xmalloc (in_blk_size + 1 + page_size - 1), page_size);
 560
 561   switch (split_type)
 562     {
 563     case type_digits:
 564     case type_lines:
 565       lines_split (n_units, buf, in_blk_size);
 566       break;
 567
 568     case type_bytes:
 569       bytes_split (n_units, buf, in_blk_size);
 570       break;
 571
 572     case type_byteslines:
 573       line_bytes_split (n_units);
 574       break;
 575
 576     default:
 577       abort ();
 578     }
 579
 580   if (close (STDIN_FILENO) != 0)
 581     error (EXIT_FAILURE, errno, "%s", infile);
 582   if (output_desc >= 0 && close (output_desc) < 0)
 583     error (EXIT_FAILURE, errno, "%s", outfile);
 584
 585   exit (EXIT_SUCCESS);
 586 }