src/split.c

   1 /* split.c -- split a file into pieces.
   2    Copyright (C) 88, 91, 1995-2004 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 2, or (at your option)
   7    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, write to the Free Software Foundation,
  16    Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  17 \f
  18 /* By tege@sics.se, with rms.
  19
  20    To do:
  21    * Implement -t CHAR or -t REGEX to specify break characters other
  22      than newline. */
  23
  24 #include <config.h>
  25
  26 #include <stdio.h>
  27 #include <getopt.h>
  28 #include <sys/types.h>
  29
  30 #include "system.h"
  31 #include "dirname.h"
  32 #include "error.h"
  33 #include "getpagesize.h"
  34 #include "full-read.h"
  35 #include "full-write.h"
  36 #include "inttostr.h"
  37 #include "posixver.h"
  38 #include "quote.h"
  39 #include "safe-read.h"
  40 #include "xstrtol.h"
  41
  42 /* The official name of this program (e.g., no `g' prefix).  */
  43 #define PROGRAM_NAME "split"
  44
  45 #define AUTHORS "Torbjorn Granlund", "Richard M. Stallman"
  46
  47 #define DEFAULT_SUFFIX_LENGTH 2
  48
  49 /* The name this program was run with. */
  50 char *program_name;
  51
  52 /* Base name of output files.  */
  53 static char const *outbase;
  54
  55 /* Name of output files.  */
  56 static char *outfile;
  57
  58 /* Pointer to the end of the prefix in OUTFILE.
  59    Suffixes are inserted here.  */
  60 static char *outfile_mid;
  61
  62 /* Length of OUTFILE's suffix.  */
  63 static size_t suffix_length = DEFAULT_SUFFIX_LENGTH;
  64
  65 /* Alphabet of characters to use in suffix.  */
  66 static char const *suffix_alphabet = "abcdefghijklmnopqrstuvwxyz";
  67
  68 /* Name of input file.  May be "-".  */
  69 static char *infile;
  70
  71 /* Descriptor on which input file is open.  */
  72 static int input_desc;
  73
  74 /* Descriptor on which output file is open.  */
  75 static int output_desc;
  76
  77 /* If true, print a diagnostic on standard error just before each
  78    output file is opened. */
  79 static bool verbose;
  80
  81 /* For long options that have no equivalent short option, use a
  82    non-character as a pseudo short option, starting with CHAR_MAX + 1.  */
  83 enum
  84 {
  85   VERBOSE_OPTION = CHAR_MAX + 1
  86 };
  87
  88 static struct option const longopts[] =
  89 {
  90   {"bytes", required_argument, NULL, 'b'},
  91   {"lines", required_argument, NULL, 'l'},
  92   {"line-bytes", required_argument, NULL, 'C'},
  93   {"suffix-length", required_argument, NULL, 'a'},
  94   {"numeric-suffixes", no_argument, NULL, 'd'},
  95   {"verbose", no_argument, NULL, VERBOSE_OPTION},
  96   {GETOPT_HELP_OPTION_DECL},
  97   {GETOPT_VERSION_OPTION_DECL},
  98   {NULL, 0, NULL, 0}
  99 };
 100
 101 void
 102 usage (int status)
 103 {
 104   if (status != EXIT_SUCCESS)
 105     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 106              program_name);
 107   else
 108     {
 109       printf (_("\
 110 Usage: %s [OPTION] [INPUT [PREFIX]]\n\
 111 "),
 112               program_name);
 113     fputs (_("\
 114 Output fixed-size pieces of INPUT to PREFIXaa, PREFIXab, ...; default\n\
 115 PREFIX is `x'.  With no INPUT, or when INPUT is -, read standard input.\n\
 116 \n\
 117 "), stdout);
 118       fputs (_("\
 119 Mandatory arguments to long options are mandatory for short options too.\n\
 120 "), stdout);
 121       fprintf (stdout, _("\
 122   -a, --suffix-length=N   use suffixes of length N (default %d)\n\
 123   -b, --bytes=SIZE        put SIZE bytes per output file\n\
 124   -C, --line-bytes=SIZE   put at most SIZE bytes of lines per output file\n\
 125   -d, --numeric-suffixes  use numeric suffixes instead of alphabetic\n\
 126   -l, --lines=NUMBER      put NUMBER lines per output file\n\
 127 "), DEFAULT_SUFFIX_LENGTH);
 128       fputs (_("\
 129       --verbose           print a diagnostic to standard error just\n\
 130                             before each output file is opened\n\
 131 "), stdout);
 132       fputs (HELP_OPTION_DESCRIPTION, stdout);
 133       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 134       fputs (_("\
 135 \n\
 136 SIZE may have a multiplier suffix: b for 512, k for 1K, m for 1 Meg.\n\
 137 "), stdout);
 138       printf (_("\nReport bugs to <%s>.\n"), PACKAGE_BUGREPORT);
 139     }
 140   exit (status);
 141 }
 142
 143 /* Compute the next sequential output file name and store it into the
 144    string `outfile'.  */
 145
 146 static void
 147 next_file_name (void)
 148 {
 149   /* Index in suffix_alphabet of each character in the suffix.  */
 150   static size_t *sufindex;
 151
 152   if (! outfile)
 153     {
 154       /* Allocate and initialize the first file name.  */
 155
 156       size_t outbase_length = strlen (outbase);
 157       size_t outfile_length = outbase_length + suffix_length;
 158       if (outfile_length + 1 < outbase_length)
 159         xalloc_die ();
 160       outfile = xmalloc (outfile_length + 1);
 161       outfile_mid = outfile + outbase_length;
 162       memcpy (outfile, outbase, outbase_length);
 163       memset (outfile_mid, suffix_alphabet[0], suffix_length);
 164       outfile[outfile_length] = 0;
 165       sufindex = xcalloc (suffix_length, sizeof *sufindex);
 166
 167 #if ! _POSIX_NO_TRUNC && HAVE_PATHCONF && defined _PC_NAME_MAX
 168       /* POSIX requires that if the output file name is too long for
 169          its directory, `split' must fail without creating any files.
 170          This must be checked for explicitly on operating systems that
 171          silently truncate file names.  */
 172       {
 173         char *dir = dir_name (outfile);
 174         long name_max = pathconf (dir, _PC_NAME_MAX);
 175         if (0 <= name_max && name_max < base_len (base_name (outfile)))
 176           error (EXIT_FAILURE, ENAMETOOLONG, "%s", outfile);
 177         free (dir);
 178       }
 179 #endif
 180     }
 181   else
 182     {
 183       /* Increment the suffix in place, if possible.  */
 184
 185       size_t i = suffix_length;
 186       while (i-- != 0)
 187         {
 188           sufindex[i]++;
 189           outfile_mid[i] = suffix_alphabet[sufindex[i]];
 190           if (outfile_mid[i])
 191             return;
 192           sufindex[i] = 0;
 193           outfile_mid[i] = suffix_alphabet[sufindex[i]];
 194         }
 195       error (EXIT_FAILURE, 0, _("Output file suffixes exhausted"));
 196     }
 197 }
 198
 199 /* Write BYTES bytes at BP to an output file.
 200    If NEW_FILE_FLAG is true, open the next output file.
 201    Otherwise add to the same output file already in use.  */
 202
 203 static void
 204 cwrite (bool new_file_flag, const char *bp, size_t bytes)
 205 {
 206   if (new_file_flag)
 207     {
 208       if (output_desc >= 0 && close (output_desc) < 0)
 209         error (EXIT_FAILURE, errno, "%s", outfile);
 210
 211       next_file_name ();
 212       if (verbose)
 213         fprintf (stderr, _("creating file `%s'\n"), outfile);
 214       output_desc = open (outfile,
 215                           O_WRONLY | O_CREAT | O_TRUNC | O_BINARY, 0666);
 216       if (output_desc < 0)
 217         error (EXIT_FAILURE, errno, "%s", outfile);
 218     }
 219   if (full_write (output_desc, bp, bytes) != bytes)
 220     error (EXIT_FAILURE, errno, "%s", outfile);
 221 }
 222
 223 /* Split into pieces of exactly N_BYTES bytes.
 224    Use buffer BUF, whose size is BUFSIZE.  */
 225
 226 static void
 227 bytes_split (uintmax_t n_bytes, char *buf, size_t bufsize)
 228 {
 229   size_t n_read;
 230   bool new_file_flag = true;
 231   size_t to_read;
 232   uintmax_t to_write = n_bytes;
 233   char *bp_out;
 234
 235   do
 236     {
 237       n_read = full_read (input_desc, buf, bufsize);
 238       if (n_read == SAFE_READ_ERROR)
 239         error (EXIT_FAILURE, errno, "%s", infile);
 240       bp_out = buf;
 241       to_read = n_read;
 242       for (;;)
 243         {
 244           if (to_read < to_write)
 245             {
 246               if (to_read)      /* do not write 0 bytes! */
 247                 {
 248                   cwrite (new_file_flag, bp_out, to_read);
 249                   to_write -= to_read;
 250                   new_file_flag = false;
 251                 }
 252               break;
 253             }
 254           else
 255             {
 256               size_t w = to_write;
 257               cwrite (new_file_flag, bp_out, w);
 258               bp_out += w;
 259               to_read -= w;
 260               new_file_flag = true;
 261               to_write = n_bytes;
 262             }
 263         }
 264     }
 265   while (n_read == bufsize);
 266 }
 267
 268 /* Split into pieces of exactly N_LINES lines.
 269    Use buffer BUF, whose size is BUFSIZE.  */
 270
 271 static void
 272 lines_split (uintmax_t n_lines, char *buf, size_t bufsize)
 273 {
 274   size_t n_read;
 275   char *bp, *bp_out, *eob;
 276   bool new_file_flag = true;
 277   uintmax_t n = 0;
 278
 279   do
 280     {
 281       n_read = full_read (input_desc, buf, bufsize);
 282       if (n_read == SAFE_READ_ERROR)
 283         error (EXIT_FAILURE, errno, "%s", infile);
 284       bp = bp_out = buf;
 285       eob = bp + n_read;
 286       *eob = '\n';
 287       for (;;)
 288         {
 289           bp = memchr (bp, '\n', eob - bp + 1);
 290           if (bp == eob)
 291             {
 292               if (eob != bp_out) /* do not write 0 bytes! */
 293                 {
 294                   size_t len = eob - bp_out;
 295                   cwrite (new_file_flag, bp_out, len);
 296                   new_file_flag = false;
 297                 }
 298               break;
 299             }
 300
 301           ++bp;
 302           if (++n >= n_lines)
 303             {
 304               cwrite (new_file_flag, bp_out, bp - bp_out);
 305               bp_out = bp;
 306               new_file_flag = true;
 307               n = 0;
 308             }
 309         }
 310     }
 311   while (n_read == bufsize);
 312 }
 313 \f
 314 /* Split into pieces that are as large as possible while still not more
 315    than N_BYTES bytes, and are split on line boundaries except
 316    where lines longer than N_BYTES bytes occur.
 317    FIXME: Allow N_BYTES to be any uintmax_t value, and don't require a
 318    buffer of size N_BYTES, in case N_BYTES is very large.  */
 319
 320 static void
 321 line_bytes_split (size_t n_bytes)
 322 {
 323   size_t n_read;
 324   char *bp;
 325   bool eof = false;
 326   size_t n_buffered = 0;
 327   char *buf = xmalloc (n_bytes);
 328
 329   do
 330     {
 331       /* Fill up the full buffer size from the input file.  */
 332
 333       n_read = full_read (input_desc, buf + n_buffered, n_bytes - n_buffered);
 334       if (n_read == SAFE_READ_ERROR)
 335         error (EXIT_FAILURE, errno, "%s", infile);
 336
 337       n_buffered += n_read;
 338       if (n_buffered != n_bytes)
 339         eof = true;
 340
 341       /* Find where to end this chunk.  */
 342       bp = buf + n_buffered;
 343       if (n_buffered == n_bytes)
 344         {
 345           while (bp > buf && bp[-1] != '\n')
 346             bp--;
 347         }
 348
 349       /* If chunk has no newlines, use all the chunk.  */
 350       if (bp == buf)
 351         bp = buf + n_buffered;
 352
 353       /* Output the chars as one output file.  */
 354       cwrite (true, buf, bp - buf);
 355
 356       /* Discard the chars we just output; move rest of chunk
 357          down to be the start of the next chunk.  Source and
 358          destination probably overlap.  */
 359       n_buffered -= bp - buf;
 360       if (n_buffered > 0)
 361         memmove (buf, bp, n_buffered);
 362     }
 363   while (!eof);
 364   free (buf);
 365 }
 366
 367 #define FAIL_ONLY_ONE_WAY()                                     \
 368   do                                                            \
 369     {                                                           \
 370       error (0, 0, _("cannot split in more than one way"));     \
 371       usage (EXIT_FAILURE);                                     \
 372     }                                                           \
 373   while (0)
 374
 375 int
 376 main (int argc, char **argv)
 377 {
 378   struct stat stat_buf;
 379   enum
 380     {
 381       type_undef, type_bytes, type_byteslines, type_lines, type_digits
 382     } split_type = type_undef;
 383   size_t in_blk_size;           /* optimal block size of input file device */
 384   char *buf;                    /* file i/o buffer */
 385   size_t page_size = getpagesize ();
 386   uintmax_t n_units;
 387   int c;
 388   int digits_optind = 0;
 389
 390   initialize_main (&argc, &argv);
 391   program_name = argv[0];
 392   setlocale (LC_ALL, "");
 393   bindtextdomain (PACKAGE, LOCALEDIR);
 394   textdomain (PACKAGE);
 395
 396   atexit (close_stdout);
 397
 398   /* Parse command line options.  */
 399
 400   infile = "-";
 401   outbase = "x";
 402
 403   while (1)
 404     {
 405       /* This is the argv-index of the option we will read next.  */
 406       int this_optind = optind ? optind : 1;
 407
 408       c = getopt_long (argc, argv, "0123456789C:a:b:dl:", longopts, NULL);
 409       if (c == -1)
 410         break;
 411
 412       switch (c)
 413         {
 414         case 'a':
 415           {
 416             unsigned long tmp;
 417             if (xstrtoul (optarg, NULL, 10, &tmp, "") != LONGINT_OK
 418                 || SIZE_MAX / sizeof (size_t) < tmp)
 419               {
 420                 error (0, 0, _("%s: invalid suffix length"), optarg);
 421                 usage (EXIT_FAILURE);
 422               }
 423             suffix_length = tmp;
 424           }
 425           break;
 426
 427         case 'b':
 428           if (split_type != type_undef)
 429             FAIL_ONLY_ONE_WAY ();
 430           split_type = type_bytes;
 431           if (xstrtoumax (optarg, NULL, 10, &n_units, "bkm") != LONGINT_OK
 432               || n_units == 0)
 433             {
 434               error (0, 0, _("%s: invalid number of bytes"), optarg);
 435               usage (EXIT_FAILURE);
 436             }
 437           break;
 438
 439         case 'l':
 440           if (split_type != type_undef)
 441             FAIL_ONLY_ONE_WAY ();
 442           split_type = type_lines;
 443           if (xstrtoumax (optarg, NULL, 10, &n_units, "") != LONGINT_OK
 444               || n_units == 0)
 445             {
 446               error (0, 0, _("%s: invalid number of lines"), optarg);
 447               usage (EXIT_FAILURE);
 448             }
 449           break;
 450
 451         case 'C':
 452           if (split_type != type_undef)
 453             FAIL_ONLY_ONE_WAY ();
 454           split_type = type_byteslines;
 455           if (xstrtoumax (optarg, NULL, 10, &n_units, "bkm") != LONGINT_OK
 456               || n_units == 0 || SIZE_MAX < n_units)
 457             {
 458               error (0, 0, _("%s: invalid number of bytes"), optarg);
 459               usage (EXIT_FAILURE);
 460             }
 461           break;
 462
 463         case '0':
 464         case '1':
 465         case '2':
 466         case '3':
 467         case '4':
 468         case '5':
 469         case '6':
 470         case '7':
 471         case '8':
 472         case '9':
 473           if (split_type == type_undef)
 474             {
 475               split_type = type_digits;
 476               n_units = 0;
 477             }
 478           if (split_type != type_undef && split_type != type_digits)
 479             FAIL_ONLY_ONE_WAY ();
 480           if (digits_optind != 0 && digits_optind != this_optind)
 481             n_units = 0;        /* More than one number given; ignore other. */
 482           digits_optind = this_optind;
 483           if (UINTMAX_MAX / 10 < n_units
 484               || n_units * 10 + c - '0' < n_units * 10)
 485             {
 486               char buffer[INT_BUFSIZE_BOUND (uintmax_t)];
 487               error (EXIT_FAILURE, 0,
 488                      _("line count option -%s%c... is too large"),
 489                      umaxtostr (n_units, buffer), c);
 490             }
 491           n_units = n_units * 10 + c - '0';
 492           break;
 493
 494         case 'd':
 495           suffix_alphabet = "0123456789";
 496           break;
 497
 498         case VERBOSE_OPTION:
 499           verbose = true;
 500           break;
 501
 502         case_GETOPT_HELP_CHAR;
 503
 504         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 505
 506         default:
 507           usage (EXIT_FAILURE);
 508         }
 509     }
 510
 511   if (digits_optind && 200112 <= posix2_version ())
 512     {
 513       char buffer[INT_BUFSIZE_BOUND (uintmax_t)];
 514       char const *a = umaxtostr (n_units, buffer);
 515       error (0, 0, _("`-%s' option is obsolete; use `-l %s'"), a, a);
 516       usage (EXIT_FAILURE);
 517     }
 518
 519   /* Handle default case.  */
 520   if (split_type == type_undef)
 521     {
 522       split_type = type_lines;
 523       n_units = 1000;
 524     }
 525
 526   if (n_units == 0)
 527     {
 528       /* FIXME: be sure to remove this block when removing
 529          support for obsolete options like `-10'.  */
 530       error (0, 0, _("invalid number of lines: 0"));
 531       usage (EXIT_FAILURE);
 532     }
 533
 534   /* Get out the filename arguments.  */
 535
 536   if (optind < argc)
 537     infile = argv[optind++];
 538
 539   if (optind < argc)
 540     outbase = argv[optind++];
 541
 542   if (optind < argc)
 543     {
 544       error (0, 0, _("extra operand %s"), quote (argv[optind]));
 545       usage (EXIT_FAILURE);
 546     }
 547
 548   /* Open the input file.  */
 549   if (STREQ (infile, "-"))
 550     input_desc = STDIN_FILENO;
 551   else
 552     {
 553       input_desc = open (infile, O_RDONLY);
 554       if (input_desc < 0)
 555         error (EXIT_FAILURE, errno, "%s", infile);
 556     }
 557   /* Binary I/O is safer when bytecounts are used.  */
 558   SET_BINARY (input_desc);
 559
 560   /* No output file is open now.  */
 561   output_desc = -1;
 562
 563   /* Get the optimal block size of input device and make a buffer.  */
 564
 565   if (fstat (input_desc, &stat_buf) < 0)
 566     error (EXIT_FAILURE, errno, "%s", infile);
 567   in_blk_size = ST_BLKSIZE (stat_buf);
 568
 569   buf = ptr_align (xmalloc (in_blk_size + 1 + page_size - 1), page_size);
 570
 571   switch (split_type)
 572     {
 573     case type_digits:
 574     case type_lines:
 575       lines_split (n_units, buf, in_blk_size);
 576       break;
 577
 578     case type_bytes:
 579       bytes_split (n_units, buf, in_blk_size);
 580       break;
 581
 582     case type_byteslines:
 583       line_bytes_split (n_units);
 584       break;
 585
 586     default:
 587       abort ();
 588     }
 589
 590   if (close (input_desc) < 0)
 591     error (EXIT_FAILURE, errno, "%s", infile);
 592   if (output_desc >= 0 && close (output_desc) < 0)
 593     error (EXIT_FAILURE, errno, "%s", outfile);
 594
 595   exit (EXIT_SUCCESS);
 596 }