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