src/split.c

   1 /* split.c -- split a file into pieces.
   2    Copyright (C) 1988, 1991, 1995 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
  16    Foundation, 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 #if HAVE_LIMITS_H
  31 # include <limits.h>
  32 #endif
  33
  34 #ifndef UINT_MAX
  35 # define UINT_MAX ((unsigned int) ~(unsigned int) 0)
  36 #endif
  37
  38 #ifndef INT_MAX
  39 # define INT_MAX ((int) (UINT_MAX >> 1))
  40 #endif
  41
  42 #include "system.h"
  43 #include "error.h"
  44 #include "xstrtol.h"
  45
  46 char *xmalloc ();
  47 int full_write ();
  48 int safe_read ();
  49
  50 /* The name this program was run with. */
  51 char *program_name;
  52
  53 /* Base 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 /* Pointer to the end of OUTFILE. */
  61 static char *outfile_end;
  62
  63 /* Name of input file.  May be "-".  */
  64 static char *infile;
  65
  66 /* Descriptor on which input file is open.  */
  67 static int input_desc;
  68
  69 /* Descriptor on which output file is open.  */
  70 static int output_desc;
  71
  72 /* If nonzero, display usage information and exit.  */
  73 static int show_help;
  74
  75 /* If nonzero, print the version on standard output then exit.  */
  76 static int show_version;
  77
  78 /* If nonzero, print a diagnostic on standard error just before each
  79    output file is opened. */
  80 static int verbose;
  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   {"verbose", no_argument, NULL, 2},
  88   {"help", no_argument, &show_help, 1},
  89   {"version", no_argument, &show_version, 1},
  90   {NULL, 0, NULL, 0}
  91 };
  92
  93 static void
  94 usage (int status, const char *reason)
  95 {
  96   if (reason != NULL)
  97     fprintf (stderr, "%s: %s\n", program_name, reason);
  98
  99   if (status != 0)
 100     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 101              program_name);
 102   else
 103     {
 104       printf (_("\
 105 Usage: %s [OPTION] [INPUT [PREFIX]]\n\
 106 "),
 107               program_name);
 108     printf (_("\
 109 Output fixed-size pieces of INPUT to PREFIXaa, PREFIXab, ...; default\n\
 110 PREFIX is `x'.  With no INPUT, or when INPUT is -, read standard input.\n\
 111 \n\
 112   -C, --line-bytes=SIZE   put at most SIZE bytes of lines per output file\n\
 113   -b, --bytes=SIZE        put SIZE bytes per output file\n\
 114   -l, --lines=NUMBER      put NUMBER lines per output file\n\
 115       --verbose           print a diagnostic to standard error just\n\
 116                             before each output file is opened\n\
 117   -NUMBER                 same as -l NUMBER\n\
 118       --help              display this help and exit\n\
 119       --version           output version information and exit\n\
 120 \n\
 121 SIZE may have a multiplier suffix: b for 512, k for 1K, m for 1 Meg.\n\
 122 "));
 123     }
 124   exit (status);
 125 }
 126
 127 /* Compute the next sequential output file name suffix and store it
 128    into the string `outfile' at the position pointed to by `outfile_mid'.  */
 129
 130 static void
 131 next_file_name (void)
 132 {
 133   int x;
 134   char *ne;
 135   unsigned int i;
 136
 137   static int first_call = 1;
 138
 139   /* Status for outfile name generation.  */
 140   static unsigned outfile_count = 0;
 141   static unsigned outfile_name_limit = 25 * 26;
 142   static unsigned outfile_name_generation = 1;
 143
 144   if (!first_call)
 145     outfile_count++;
 146   first_call = 0;
 147   if (outfile_count < outfile_name_limit)
 148     {
 149       for (ne = outfile_end - 1; ; ne--)
 150         {
 151           x = *ne;
 152           if (x != 'z')
 153             break;
 154           *ne = 'a';
 155         }
 156       *ne = x + 1;
 157       return;
 158     }
 159
 160   outfile_count = 0;
 161   outfile_name_limit *= 26;
 162   outfile_name_generation++;
 163   *outfile_mid++ = 'z';
 164   for (i = 0; i <= outfile_name_generation; i++)
 165     outfile_mid[i] = 'a';
 166   outfile_end += 2;
 167 }
 168
 169 /* Write BYTES bytes at BP to an output file.
 170    If NEW_FILE_FLAG is nonzero, open the next output file.
 171    Otherwise add to the same output file already in use.  */
 172
 173 static void
 174 cwrite (int new_file_flag, const char *bp, int bytes)
 175 {
 176   if (new_file_flag)
 177     {
 178       if (output_desc >= 0 && close (output_desc) < 0)
 179         error (1, errno, "%s", outfile);
 180
 181       next_file_name ();
 182       if (verbose)
 183         fprintf (stderr, _("creating file `%s'\n"), outfile);
 184       output_desc = open (outfile, O_WRONLY | O_CREAT | O_TRUNC, 0666);
 185       if (output_desc < 0)
 186         error (1, errno, "%s", outfile);
 187     }
 188   if (full_write (output_desc, bp, bytes) < 0)
 189     error (1, errno, "%s", outfile);
 190 }
 191
 192 /* Read NCHARS bytes from the input file into BUF.
 193    Return the number of bytes successfully read.
 194    If this is less than NCHARS, do not call `stdread' again.  */
 195
 196 static int
 197 stdread (char *buf, int nchars)
 198 {
 199   int n_read;
 200   int to_be_read = nchars;
 201
 202   while (to_be_read)
 203     {
 204       n_read = safe_read (input_desc, buf, to_be_read);
 205       if (n_read < 0)
 206         return -1;
 207       if (n_read == 0)
 208         break;
 209       to_be_read -= n_read;
 210       buf += n_read;
 211     }
 212   return nchars - to_be_read;
 213 }
 214
 215 /* Split into pieces of exactly NCHARS bytes.
 216    Use buffer BUF, whose size is BUFSIZE.  */
 217
 218 static void
 219 bytes_split (int nchars, char *buf, int bufsize)
 220 {
 221   int n_read;
 222   int new_file_flag = 1;
 223   int to_read;
 224   int to_write = nchars;
 225   char *bp_out;
 226
 227   do
 228     {
 229       n_read = stdread (buf, bufsize);
 230       if (n_read < 0)
 231         error (1, errno, "%s", infile);
 232       bp_out = buf;
 233       to_read = n_read;
 234       for (;;)
 235         {
 236           if (to_read < to_write)
 237             {
 238               if (to_read)      /* do not write 0 bytes! */
 239                 {
 240                   cwrite (new_file_flag, bp_out, to_read);
 241                   to_write -= to_read;
 242                   new_file_flag = 0;
 243                 }
 244               break;
 245             }
 246           else
 247             {
 248               cwrite (new_file_flag, bp_out, to_write);
 249               bp_out += to_write;
 250               to_read -= to_write;
 251               new_file_flag = 1;
 252               to_write = nchars;
 253             }
 254         }
 255     }
 256   while (n_read == bufsize);
 257 }
 258
 259 /* Split into pieces of exactly NLINES lines.
 260    Use buffer BUF, whose size is BUFSIZE.  */
 261
 262 static void
 263 lines_split (int nlines, char *buf, int bufsize)
 264 {
 265   int n_read;
 266   char *bp, *bp_out, *eob;
 267   int new_file_flag = 1;
 268   int n = 0;
 269
 270   do
 271     {
 272       n_read = stdread (buf, bufsize);
 273       if (n_read < 0)
 274         error (1, errno, "%s", infile);
 275       bp = bp_out = buf;
 276       eob = bp + n_read;
 277       *eob = '\n';
 278       for (;;)
 279         {
 280           while (*bp++ != '\n')
 281             ;                   /* this semicolon takes most of the time */
 282           if (bp > eob)
 283             {
 284               if (eob != bp_out) /* do not write 0 bytes! */
 285                 {
 286                   cwrite (new_file_flag, bp_out, eob - bp_out);
 287                   new_file_flag = 0;
 288                 }
 289               break;
 290             }
 291           else
 292             if (++n >= nlines)
 293               {
 294                 cwrite (new_file_flag, bp_out, bp - bp_out);
 295                 bp_out = bp;
 296                 new_file_flag = 1;
 297                 n = 0;
 298               }
 299         }
 300     }
 301   while (n_read == bufsize);
 302 }
 303 \f
 304 /* Split into pieces that are as large as possible while still not more
 305    than NCHARS bytes, and are split on line boundaries except
 306    where lines longer than NCHARS bytes occur. */
 307
 308 static void
 309 line_bytes_split (int nchars)
 310 {
 311   int n_read;
 312   char *bp;
 313   int eof = 0;
 314   int n_buffered = 0;
 315   char *buf = (char *) xmalloc (nchars);
 316
 317   do
 318     {
 319       /* Fill up the full buffer size from the input file.  */
 320
 321       n_read = stdread (buf + n_buffered, nchars - n_buffered);
 322       if (n_read < 0)
 323         error (1, errno, "%s", infile);
 324
 325       n_buffered += n_read;
 326       if (n_buffered != nchars)
 327         eof = 1;
 328
 329       /* Find where to end this chunk.  */
 330       bp = buf + n_buffered;
 331       if (n_buffered == nchars)
 332         {
 333           while (bp > buf && bp[-1] != '\n')
 334             bp--;
 335         }
 336
 337       /* If chunk has no newlines, use all the chunk.  */
 338       if (bp == buf)
 339         bp = buf + n_buffered;
 340
 341       /* Output the chars as one output file.  */
 342       cwrite (1, buf, bp - buf);
 343
 344       /* Discard the chars we just output; move rest of chunk
 345          down to be the start of the next chunk.  Source and
 346          destination probably overlap.  */
 347       n_buffered -= bp - buf;
 348       if (n_buffered > 0)
 349         memmove (buf, bp, n_buffered);
 350     }
 351   while (!eof);
 352   free (buf);
 353 }
 354
 355 int
 356 main (int argc, char **argv)
 357 {
 358   struct stat stat_buf;
 359   int num;                      /* numeric argument from command line */
 360   enum
 361     {
 362       type_undef, type_bytes, type_byteslines, type_lines, type_digits
 363     } split_type = type_undef;
 364   int in_blk_size;              /* optimal block size of input file device */
 365   char *buf;                    /* file i/o buffer */
 366   int accum = 0;
 367   char *outbase;
 368   int c;
 369   int digits_optind = 0;
 370
 371   program_name = argv[0];
 372   setlocale (LC_ALL, "");
 373   bindtextdomain (PACKAGE, LOCALEDIR);
 374   textdomain (PACKAGE);
 375
 376   /* Parse command line options.  */
 377
 378   infile = "-";
 379   outbase = "x";
 380
 381   while (1)
 382     {
 383       /* This is the argv-index of the option we will read next.  */
 384       int this_optind = optind ? optind : 1;
 385       long int tmp_long;
 386
 387       c = getopt_long (argc, argv, "0123456789vb:l:C:", longopts, (int *) 0);
 388       if (c == EOF)
 389         break;
 390
 391       switch (c)
 392         {
 393         case 0:
 394           break;
 395
 396         case 'b':
 397           if (split_type != type_undef)
 398             usage (2, _("cannot split in more than one way"));
 399           split_type = type_bytes;
 400           if (xstrtol (optarg, NULL, 10, &tmp_long, "bkm") != LONGINT_OK
 401               || tmp_long < 0 || tmp_long > INT_MAX)
 402             usage (2, _("invalid number of bytes"));
 403           accum = (int) tmp_long;
 404           break;
 405
 406         case 'l':
 407           if (split_type != type_undef)
 408             usage (2, _("cannot split in more than one way"));
 409           split_type = type_lines;
 410           if (xstrtol (optarg, NULL, 10, &tmp_long, "") != LONGINT_OK
 411               || tmp_long < 0 || tmp_long > INT_MAX)
 412             usage (2, _("invalid number of lines"));
 413           accum = (int) tmp_long;
 414           break;
 415
 416         case 'C':
 417           if (split_type != type_undef)
 418             usage (2, _("cannot split in more than one way"));
 419           split_type = type_byteslines;
 420           if (xstrtol (optarg, NULL, 10, &tmp_long, "bkm") != LONGINT_OK
 421               || tmp_long < 0 ||  tmp_long > INT_MAX)
 422             usage (2, _("invalid number of bytes"));
 423           accum = (int) tmp_long;
 424           break;
 425
 426         case '0':
 427         case '1':
 428         case '2':
 429         case '3':
 430         case '4':
 431         case '5':
 432         case '6':
 433         case '7':
 434         case '8':
 435         case '9':
 436           if (split_type != type_undef && split_type != type_digits)
 437             usage (2, _("cannot split in more than one way"));
 438           if (digits_optind != 0 && digits_optind != this_optind)
 439             accum = 0;          /* More than one number given; ignore other. */
 440           digits_optind = this_optind;
 441           split_type = type_digits;
 442           accum = accum * 10 + c - '0';
 443           break;
 444
 445         case 2:
 446           verbose = 1;
 447           break;
 448
 449         default:
 450           usage (2, (char *)0);
 451         }
 452     }
 453
 454   if (show_version)
 455     {
 456       printf ("split - %s\n", PACKAGE_VERSION);
 457       exit (0);
 458     }
 459
 460   if (show_help)
 461     usage (0, (char *)0);
 462
 463   /* Handle default case.  */
 464   if (split_type == type_undef)
 465     {
 466       split_type = type_lines;
 467       accum = 1000;
 468     }
 469
 470   if (accum < 1)
 471     usage (2, _("invalid number"));
 472   num = accum;
 473
 474   /* Get out the filename arguments.  */
 475
 476   if (optind < argc)
 477     infile = argv[optind++];
 478
 479   if (optind < argc)
 480     outbase = argv[optind++];
 481
 482   if (optind < argc)
 483     usage (2, _("too many arguments"));
 484
 485   /* Open the input file.  */
 486   if (!strcmp (infile, "-"))
 487     input_desc = 0;
 488   else
 489     {
 490       input_desc = open (infile, O_RDONLY);
 491       if (input_desc < 0)
 492         error (1, errno, "%s", infile);
 493     }
 494
 495   /* No output file is open now.  */
 496   output_desc = -1;
 497
 498   /* Copy the output file prefix so we can add suffixes to it.
 499      26**29 is certainly enough output files!  */
 500
 501   outfile = xmalloc (strlen (outbase) + 30);
 502   strcpy (outfile, outbase);
 503   outfile_mid = outfile + strlen (outfile);
 504   outfile_end = outfile_mid + 2;
 505   memset (outfile_mid, 0, 30);
 506   outfile_mid[0] = 'a';
 507   outfile_mid[1] = 'a' - 1;  /* first call to next_file_name makes it an 'a' */
 508
 509   /* Get the optimal block size of input device and make a buffer.  */
 510
 511   if (fstat (input_desc, &stat_buf) < 0)
 512     error (1, errno, "%s", infile);
 513   in_blk_size = ST_BLKSIZE (stat_buf);
 514
 515   buf = xmalloc (in_blk_size + 1);
 516
 517   switch (split_type)
 518     {
 519     case type_digits:
 520     case type_lines:
 521       lines_split (num, buf, in_blk_size);
 522       break;
 523
 524     case type_bytes:
 525       bytes_split (num, buf, in_blk_size);
 526       break;
 527
 528     case type_byteslines:
 529       line_bytes_split (num);
 530       break;
 531
 532     default:
 533       abort ();
 534     }
 535
 536   if (close (input_desc) < 0)
 537     error (1, errno, "%s", infile);
 538   if (output_desc >= 0 && close (output_desc) < 0)
 539     error (1, errno, "%s", outfile);
 540
 541   exit (0);
 542 }