src/split.c

   1 /* split.c -- split a file into pieces.
   2    Copyright (C) 1988, 1991, 1995-2010 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 "xfreopen.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       emit_size_note ();
 129       emit_ancillary_info ();
 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 (last_component (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 true, open the next output file.
 192    Otherwise add to the same output file already in use.  */
 193
 194 static void
 195 cwrite (bool 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 (stdout, _("creating file %s\n"), quote (outfile));
 205       output_desc = open (outfile,
 206                           O_WRONLY | O_CREAT | O_TRUNC | O_BINARY,
 207                           (S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP
 208                            | S_IROTH | S_IWOTH));
 209       if (output_desc < 0)
 210         error (EXIT_FAILURE, errno, "%s", outfile);
 211     }
 212   if (full_write (output_desc, bp, bytes) != bytes)
 213     error (EXIT_FAILURE, errno, "%s", outfile);
 214 }
 215
 216 /* Split into pieces of exactly N_BYTES bytes.
 217    Use buffer BUF, whose size is BUFSIZE.  */
 218
 219 static void
 220 bytes_split (uintmax_t n_bytes, char *buf, size_t bufsize)
 221 {
 222   size_t n_read;
 223   bool new_file_flag = true;
 224   size_t to_read;
 225   uintmax_t to_write = n_bytes;
 226   char *bp_out;
 227
 228   do
 229     {
 230       n_read = full_read (STDIN_FILENO, buf, bufsize);
 231       if (n_read < bufsize && errno)
 232         error (EXIT_FAILURE, errno, "%s", infile);
 233       bp_out = buf;
 234       to_read = n_read;
 235       while (true)
 236         {
 237           if (to_read < to_write)
 238             {
 239               if (to_read)      /* do not write 0 bytes! */
 240                 {
 241                   cwrite (new_file_flag, bp_out, to_read);
 242                   to_write -= to_read;
 243                   new_file_flag = false;
 244                 }
 245               break;
 246             }
 247           else
 248             {
 249               size_t w = to_write;
 250               cwrite (new_file_flag, bp_out, w);
 251               bp_out += w;
 252               to_read -= w;
 253               new_file_flag = true;
 254               to_write = n_bytes;
 255             }
 256         }
 257     }
 258   while (n_read == bufsize);
 259 }
 260
 261 /* Split into pieces of exactly N_LINES lines.
 262    Use buffer BUF, whose size is BUFSIZE.  */
 263
 264 static void
 265 lines_split (uintmax_t n_lines, char *buf, size_t bufsize)
 266 {
 267   size_t n_read;
 268   char *bp, *bp_out, *eob;
 269   bool new_file_flag = true;
 270   uintmax_t n = 0;
 271
 272   do
 273     {
 274       n_read = full_read (STDIN_FILENO, buf, bufsize);
 275       if (n_read < bufsize && errno)
 276         error (EXIT_FAILURE, errno, "%s", infile);
 277       bp = bp_out = buf;
 278       eob = bp + n_read;
 279       *eob = '\n';
 280       while (true)
 281         {
 282           bp = memchr (bp, '\n', eob - bp + 1);
 283           if (bp == eob)
 284             {
 285               if (eob != bp_out) /* do not write 0 bytes! */
 286                 {
 287                   size_t len = eob - bp_out;
 288                   cwrite (new_file_flag, bp_out, len);
 289                   new_file_flag = false;
 290                 }
 291               break;
 292             }
 293
 294           ++bp;
 295           if (++n >= n_lines)
 296             {
 297               cwrite (new_file_flag, bp_out, bp - bp_out);
 298               bp_out = bp;
 299               new_file_flag = true;
 300               n = 0;
 301             }
 302         }
 303     }
 304   while (n_read == bufsize);
 305 }
 306 \f
 307 /* Split into pieces that are as large as possible while still not more
 308    than N_BYTES bytes, and are split on line boundaries except
 309    where lines longer than N_BYTES bytes occur.
 310    FIXME: Allow N_BYTES to be any uintmax_t value, and don't require a
 311    buffer of size N_BYTES, in case N_BYTES is very large.  */
 312
 313 static void
 314 line_bytes_split (size_t n_bytes)
 315 {
 316   char *bp;
 317   bool eof = false;
 318   size_t n_buffered = 0;
 319   char *buf = xmalloc (n_bytes);
 320
 321   do
 322     {
 323       /* Fill up the full buffer size from the input file.  */
 324
 325       size_t to_read = n_bytes - n_buffered;
 326       size_t n_read = full_read (STDIN_FILENO, buf + n_buffered, to_read);
 327       if (n_read < to_read && errno)
 328         error (EXIT_FAILURE, errno, "%s", infile);
 329
 330       n_buffered += n_read;
 331       if (n_buffered != n_bytes)
 332         {
 333           if (n_buffered == 0)
 334             break;
 335           eof = true;
 336         }
 337
 338       /* Find where to end this chunk.  */
 339       bp = buf + n_buffered;
 340       if (n_buffered == n_bytes)
 341         {
 342           while (bp > buf && bp[-1] != '\n')
 343             bp--;
 344         }
 345
 346       /* If chunk has no newlines, use all the chunk.  */
 347       if (bp == buf)
 348         bp = buf + n_buffered;
 349
 350       /* Output the chars as one output file.  */
 351       cwrite (true, buf, bp - buf);
 352
 353       /* Discard the chars we just output; move rest of chunk
 354          down to be the start of the next chunk.  Source and
 355          destination probably overlap.  */
 356       n_buffered -= bp - buf;
 357       if (n_buffered > 0)
 358         memmove (buf, bp, n_buffered);
 359     }
 360   while (!eof);
 361   free (buf);
 362 }
 363
 364 #define FAIL_ONLY_ONE_WAY()                                     \
 365   do                                                            \
 366     {                                                           \
 367       error (0, 0, _("cannot split in more than one way"));     \
 368       usage (EXIT_FAILURE);                                     \
 369     }                                                           \
 370   while (0)
 371
 372 int
 373 main (int argc, char **argv)
 374 {
 375   struct stat stat_buf;
 376   enum
 377     {
 378       type_undef, type_bytes, type_byteslines, type_lines, type_digits
 379     } split_type = type_undef;
 380   size_t in_blk_size;           /* optimal block size of input file device */
 381   char *buf;                    /* file i/o buffer */
 382   size_t page_size = getpagesize ();
 383   uintmax_t n_units;
 384   static char const multipliers[] = "bEGKkMmPTYZ0";
 385   int c;
 386   int digits_optind = 0;
 387
 388   initialize_main (&argc, &argv);
 389   set_program_name (argv[0]);
 390   setlocale (LC_ALL, "");
 391   bindtextdomain (PACKAGE, LOCALEDIR);
 392   textdomain (PACKAGE);
 393
 394   atexit (close_stdout);
 395
 396   /* Parse command line options.  */
 397
 398   infile = bad_cast ( "-");
 399   outbase = bad_cast ("x");
 400
 401   while (1)
 402     {
 403       /* This is the argv-index of the option we will read next.  */
 404       int this_optind = optind ? optind : 1;
 405
 406       c = getopt_long (argc, argv, "0123456789C:a:b:dl:", longopts, NULL);
 407       if (c == -1)
 408         break;
 409
 410       switch (c)
 411         {
 412         case 'a':
 413           {
 414             unsigned long tmp;
 415             if (xstrtoul (optarg, NULL, 10, &tmp, "") != LONGINT_OK
 416                 || SIZE_MAX / sizeof (size_t) < tmp)
 417               {
 418                 error (0, 0, _("%s: invalid suffix length"), optarg);
 419                 usage (EXIT_FAILURE);
 420               }
 421             suffix_length = tmp;
 422           }
 423           break;
 424
 425         case 'b':
 426           if (split_type != type_undef)
 427             FAIL_ONLY_ONE_WAY ();
 428           split_type = type_bytes;
 429           if (xstrtoumax (optarg, NULL, 10, &n_units, multipliers) != LONGINT_OK
 430               || n_units == 0)
 431             {
 432               error (0, 0, _("%s: invalid number of bytes"), optarg);
 433               usage (EXIT_FAILURE);
 434             }
 435           break;
 436
 437         case 'l':
 438           if (split_type != type_undef)
 439             FAIL_ONLY_ONE_WAY ();
 440           split_type = type_lines;
 441           if (xstrtoumax (optarg, NULL, 10, &n_units, "") != LONGINT_OK
 442               || n_units == 0)
 443             {
 444               error (0, 0, _("%s: invalid number of lines"), optarg);
 445               usage (EXIT_FAILURE);
 446             }
 447           break;
 448
 449         case 'C':
 450           if (split_type != type_undef)
 451             FAIL_ONLY_ONE_WAY ();
 452           split_type = type_byteslines;
 453           if (xstrtoumax (optarg, NULL, 10, &n_units, multipliers) != LONGINT_OK
 454               || n_units == 0 || SIZE_MAX < n_units)
 455             {
 456               error (0, 0, _("%s: invalid number of bytes"), optarg);
 457               usage (EXIT_FAILURE);
 458             }
 459           break;
 460
 461         case '0':
 462         case '1':
 463         case '2':
 464         case '3':
 465         case '4':
 466         case '5':
 467         case '6':
 468         case '7':
 469         case '8':
 470         case '9':
 471           if (split_type == type_undef)
 472             {
 473               split_type = type_digits;
 474               n_units = 0;
 475             }
 476           if (split_type != type_undef && split_type != type_digits)
 477             FAIL_ONLY_ONE_WAY ();
 478           if (digits_optind != 0 && digits_optind != this_optind)
 479             n_units = 0;        /* More than one number given; ignore other. */
 480           digits_optind = this_optind;
 481           if (!DECIMAL_DIGIT_ACCUMULATE (n_units, c - '0', uintmax_t))
 482             {
 483               char buffer[INT_BUFSIZE_BOUND (uintmax_t)];
 484               error (EXIT_FAILURE, 0,
 485                      _("line count option -%s%c... is too large"),
 486                      umaxtostr (n_units, buffer), c);
 487             }
 488           break;
 489
 490         case 'd':
 491           suffix_alphabet = "0123456789";
 492           break;
 493
 494         case VERBOSE_OPTION:
 495           verbose = true;
 496           break;
 497
 498         case_GETOPT_HELP_CHAR;
 499
 500         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 501
 502         default:
 503           usage (EXIT_FAILURE);
 504         }
 505     }
 506
 507   /* Handle default case.  */
 508   if (split_type == type_undef)
 509     {
 510       split_type = type_lines;
 511       n_units = 1000;
 512     }
 513
 514   if (n_units == 0)
 515     {
 516       error (0, 0, _("invalid number of lines: 0"));
 517       usage (EXIT_FAILURE);
 518     }
 519
 520   /* Get out the filename arguments.  */
 521
 522   if (optind < argc)
 523     infile = argv[optind++];
 524
 525   if (optind < argc)
 526     outbase = argv[optind++];
 527
 528   if (optind < argc)
 529     {
 530       error (0, 0, _("extra operand %s"), quote (argv[optind]));
 531       usage (EXIT_FAILURE);
 532     }
 533
 534   /* Open the input file.  */
 535   if (! STREQ (infile, "-")
 536       && fd_reopen (STDIN_FILENO, infile, O_RDONLY, 0) < 0)
 537     error (EXIT_FAILURE, errno, _("cannot open %s for reading"),
 538            quote (infile));
 539
 540   /* Binary I/O is safer when bytecounts are used.  */
 541   if (O_BINARY && ! isatty (STDIN_FILENO))
 542     xfreopen (NULL, "rb", stdin);
 543
 544   /* No output file is open now.  */
 545   output_desc = -1;
 546
 547   /* Get the optimal block size of input device and make a buffer.  */
 548
 549   if (fstat (STDIN_FILENO, &stat_buf) != 0)
 550     error (EXIT_FAILURE, errno, "%s", infile);
 551   in_blk_size = io_blksize (stat_buf);
 552
 553   buf = ptr_align (xmalloc (in_blk_size + 1 + page_size - 1), page_size);
 554
 555   switch (split_type)
 556     {
 557     case type_digits:
 558     case type_lines:
 559       lines_split (n_units, buf, in_blk_size);
 560       break;
 561
 562     case type_bytes:
 563       bytes_split (n_units, buf, in_blk_size);
 564       break;
 565
 566     case type_byteslines:
 567       line_bytes_split (n_units);
 568       break;
 569
 570     default:
 571       abort ();
 572     }
 573
 574   if (close (STDIN_FILENO) != 0)
 575     error (EXIT_FAILURE, errno, "%s", infile);
 576   if (output_desc >= 0 && close (output_desc) < 0)
 577     error (EXIT_FAILURE, errno, "%s", outfile);
 578
 579   exit (EXIT_SUCCESS);
 580 }