src/shuf.c

   1 /* Shuffle lines of text.
   2
   3    Copyright (C) 2006-2013 Free Software Foundation, Inc.
   4
   5    This program is free software: you can redistribute it and/or modify
   6    it under the terms of the GNU General Public License as published by
   7    the Free Software Foundation, either version 3 of the License, or
   8    (at your option) any later version.
   9
  10    This program is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13    GNU General Public License for more details.
  14
  15    You should have received a copy of the GNU General Public License
  16    along with this program.  If not, see <http://www.gnu.org/licenses/>.
  17
  18    Written by Paul Eggert.  */
  19
  20 #include <config.h>
  21
  22 #include <sys/types.h>
  23 #include "system.h"
  24
  25 #include "error.h"
  26 #include "fadvise.h"
  27 #include "getopt.h"
  28 #include "quote.h"
  29 #include "quotearg.h"
  30 #include "randint.h"
  31 #include "randperm.h"
  32 #include "read-file.h"
  33 #include "stdio--.h"
  34 #include "xstrtol.h"
  35
  36 /* The official name of this program (e.g., no 'g' prefix).  */
  37 #define PROGRAM_NAME "shuf"
  38
  39 #define AUTHORS proper_name ("Paul Eggert")
  40
  41 void
  42 usage (int status)
  43 {
  44   if (status != EXIT_SUCCESS)
  45     emit_try_help ();
  46   else
  47     {
  48       printf (_("\
  49 Usage: %s [OPTION]... [FILE]\n\
  50   or:  %s -e [OPTION]... [ARG]...\n\
  51   or:  %s -i LO-HI [OPTION]...\n\
  52 "),
  53               program_name, program_name, program_name);
  54       fputs (_("\
  55 Write a random permutation of the input lines to standard output.\n\
  56 "), stdout);
  57
  58       emit_mandatory_arg_note ();
  59
  60       fputs (_("\
  61   -e, --echo                treat each ARG as an input line\n\
  62   -i, --input-range=LO-HI   treat each number LO through HI as an input line\n\
  63   -n, --head-count=COUNT    output at most COUNT lines\n\
  64   -o, --output=FILE         write result to FILE instead of standard output\n\
  65       --random-source=FILE  get random bytes from FILE\n\
  66   -z, --zero-terminated     end lines with 0 byte, not newline\n\
  67 "), stdout);
  68       fputs (HELP_OPTION_DESCRIPTION, stdout);
  69       fputs (VERSION_OPTION_DESCRIPTION, stdout);
  70       fputs (_("\
  71 \n\
  72 With no FILE, or when FILE is -, read standard input.\n\
  73 "), stdout);
  74       emit_ancillary_info ();
  75     }
  76
  77   exit (status);
  78 }
  79
  80 /* For long options that have no equivalent short option, use a
  81    non-character as a pseudo short option, starting with CHAR_MAX + 1.  */
  82 enum
  83 {
  84   RANDOM_SOURCE_OPTION = CHAR_MAX + 1
  85 };
  86
  87 static struct option const long_opts[] =
  88 {
  89   {"echo", no_argument, NULL, 'e'},
  90   {"input-range", required_argument, NULL, 'i'},
  91   {"head-count", required_argument, NULL, 'n'},
  92   {"output", required_argument, NULL, 'o'},
  93   {"random-source", required_argument, NULL, RANDOM_SOURCE_OPTION},
  94   {"zero-terminated", no_argument, NULL, 'z'},
  95   {GETOPT_HELP_OPTION_DECL},
  96   {GETOPT_VERSION_OPTION_DECL},
  97   {0, 0, 0, 0},
  98 };
  99
 100 static bool
 101 input_numbers_option_used (size_t lo_input, size_t hi_input)
 102 {
 103   return ! (lo_input == SIZE_MAX && hi_input == 0);
 104 }
 105
 106 static void
 107 input_from_argv (char **operand, int n_operands, char eolbyte)
 108 {
 109   char *p;
 110   size_t size = n_operands;
 111   int i;
 112
 113   for (i = 0; i < n_operands; i++)
 114     size += strlen (operand[i]);
 115   p = xmalloc (size);
 116
 117   for (i = 0; i < n_operands; i++)
 118     {
 119       char *p1 = stpcpy (p, operand[i]);
 120       operand[i] = p;
 121       p = p1;
 122       *p++ = eolbyte;
 123     }
 124
 125   operand[n_operands] = p;
 126 }
 127
 128 /* Return the start of the next line after LINE.  The current line
 129    ends in EOLBYTE, and is guaranteed to end before LINE + N.  */
 130
 131 static char *
 132 next_line (char *line, char eolbyte, size_t n)
 133 {
 134   char *p = memchr (line, eolbyte, n);
 135   return p + 1;
 136 }
 137
 138 /* Read data from file IN.  Input lines are delimited by EOLBYTE;
 139    silently append a trailing EOLBYTE if the file ends in some other
 140    byte.  Store a pointer to the resulting array of lines into *PLINE.
 141    Return the number of lines read.  Report an error and exit on
 142    failure.  */
 143
 144 static size_t
 145 read_input (FILE *in, char eolbyte, char ***pline)
 146 {
 147   char *p;
 148   char *buf = NULL;
 149   size_t used;
 150   char *lim;
 151   char **line;
 152   size_t i;
 153   size_t n_lines;
 154
 155   if (!(buf = fread_file (in, &used)))
 156     error (EXIT_FAILURE, errno, _("read error"));
 157
 158   if (used && buf[used - 1] != eolbyte)
 159     buf[used++] = eolbyte;
 160
 161   lim = buf + used;
 162
 163   n_lines = 0;
 164   for (p = buf; p < lim; p = next_line (p, eolbyte, lim - p))
 165     n_lines++;
 166
 167   *pline = line = xnmalloc (n_lines + 1, sizeof *line);
 168
 169   line[0] = p = buf;
 170   for (i = 1; i <= n_lines; i++)
 171     line[i] = p = next_line (p, eolbyte, lim - p);
 172
 173   return n_lines;
 174 }
 175
 176 static int
 177 write_permuted_output (size_t n_lines, char * const *line, size_t lo_input,
 178                        size_t const *permutation, char eolbyte)
 179 {
 180   size_t i;
 181
 182   if (line)
 183     for (i = 0; i < n_lines; i++)
 184       {
 185         char * const *p = line + permutation[i];
 186         size_t len = p[1] - p[0];
 187         if (fwrite (p[0], sizeof *p[0], len, stdout) != len)
 188           return -1;
 189       }
 190   else
 191     for (i = 0; i < n_lines; i++)
 192       {
 193         unsigned long int n = lo_input + permutation[i];
 194         if (printf ("%lu%c", n, eolbyte) < 0)
 195           return -1;
 196       }
 197
 198   return 0;
 199 }
 200
 201 int
 202 main (int argc, char **argv)
 203 {
 204   bool echo = false;
 205   size_t lo_input = SIZE_MAX;
 206   size_t hi_input = 0;
 207   size_t head_lines = SIZE_MAX;
 208   char const *outfile = NULL;
 209   char *random_source = NULL;
 210   char eolbyte = '\n';
 211   char **input_lines = NULL;
 212
 213   int optc;
 214   int n_operands;
 215   char **operand;
 216   size_t n_lines;
 217   char **line;
 218   struct randint_source *randint_source;
 219   size_t *permutation;
 220
 221   initialize_main (&argc, &argv);
 222   set_program_name (argv[0]);
 223   setlocale (LC_ALL, "");
 224   bindtextdomain (PACKAGE, LOCALEDIR);
 225   textdomain (PACKAGE);
 226
 227   atexit (close_stdout);
 228
 229   while ((optc = getopt_long (argc, argv, "ei:n:o:z", long_opts, NULL)) != -1)
 230     switch (optc)
 231       {
 232       case 'e':
 233         echo = true;
 234         break;
 235
 236       case 'i':
 237         {
 238           unsigned long int argval = 0;
 239           char *p = strchr (optarg, '-');
 240           char const *hi_optarg = optarg;
 241           bool invalid = !p;
 242
 243           if (input_numbers_option_used (lo_input, hi_input))
 244             error (EXIT_FAILURE, 0, _("multiple -i options specified"));
 245
 246           if (p)
 247             {
 248               *p = '\0';
 249               invalid = ((xstrtoul (optarg, NULL, 10, &argval, NULL)
 250                           != LONGINT_OK)
 251                          || SIZE_MAX < argval);
 252               *p = '-';
 253               lo_input = argval;
 254               hi_optarg = p + 1;
 255             }
 256
 257           invalid |= ((xstrtoul (hi_optarg, NULL, 10, &argval, NULL)
 258                        != LONGINT_OK)
 259                       || SIZE_MAX < argval);
 260           hi_input = argval;
 261           n_lines = hi_input - lo_input + 1;
 262           invalid |= ((lo_input <= hi_input) == (n_lines == 0));
 263           if (invalid)
 264             error (EXIT_FAILURE, 0, _("invalid input range %s"),
 265                    quote (optarg));
 266         }
 267         break;
 268
 269       case 'n':
 270         {
 271           unsigned long int argval;
 272           strtol_error e = xstrtoul (optarg, NULL, 10, &argval, NULL);
 273
 274           if (e == LONGINT_OK)
 275             head_lines = MIN (head_lines, argval);
 276           else if (e != LONGINT_OVERFLOW)
 277             error (EXIT_FAILURE, 0, _("invalid line count %s"),
 278                    quote (optarg));
 279         }
 280         break;
 281
 282       case 'o':
 283         if (outfile && !STREQ (outfile, optarg))
 284           error (EXIT_FAILURE, 0, _("multiple output files specified"));
 285         outfile = optarg;
 286         break;
 287
 288       case RANDOM_SOURCE_OPTION:
 289         if (random_source && !STREQ (random_source, optarg))
 290           error (EXIT_FAILURE, 0, _("multiple random sources specified"));
 291         random_source = optarg;
 292         break;
 293
 294       case 'z':
 295         eolbyte = '\0';
 296         break;
 297
 298       case_GETOPT_HELP_CHAR;
 299       case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 300       default:
 301         usage (EXIT_FAILURE);
 302       }
 303
 304   n_operands = argc - optind;
 305   operand = argv + optind;
 306
 307   if (echo)
 308     {
 309       if (input_numbers_option_used (lo_input, hi_input))
 310         error (EXIT_FAILURE, 0, _("cannot combine -e and -i options"));
 311       input_from_argv (operand, n_operands, eolbyte);
 312       n_lines = n_operands;
 313       line = operand;
 314     }
 315   else if (input_numbers_option_used (lo_input, hi_input))
 316     {
 317       if (n_operands)
 318         {
 319           error (0, 0, _("extra operand %s"), quote (operand[0]));
 320           usage (EXIT_FAILURE);
 321         }
 322       n_lines = hi_input - lo_input + 1;
 323       line = NULL;
 324     }
 325   else
 326     {
 327       switch (n_operands)
 328         {
 329         case 0:
 330           break;
 331
 332         case 1:
 333           if (! (STREQ (operand[0], "-") || freopen (operand[0], "r", stdin)))
 334             error (EXIT_FAILURE, errno, "%s", operand[0]);
 335           break;
 336
 337         default:
 338           error (0, 0, _("extra operand %s"), quote (operand[1]));
 339           usage (EXIT_FAILURE);
 340         }
 341
 342       fadvise (stdin, FADVISE_SEQUENTIAL);
 343
 344       n_lines = read_input (stdin, eolbyte, &input_lines);
 345       line = input_lines;
 346     }
 347
 348   head_lines = MIN (head_lines, n_lines);
 349
 350   randint_source = randint_all_new (random_source,
 351                                     randperm_bound (head_lines, n_lines));
 352   if (! randint_source)
 353     error (EXIT_FAILURE, errno, "%s", quotearg_colon (random_source));
 354
 355   /* Close stdin now, rather than earlier, so that randint_all_new
 356      doesn't have to worry about opening something other than
 357      stdin.  */
 358   if (! (echo || input_numbers_option_used (lo_input, hi_input))
 359       && (fclose (stdin) != 0))
 360     error (EXIT_FAILURE, errno, _("read error"));
 361
 362   permutation = randperm_new (randint_source, head_lines, n_lines);
 363
 364   if (outfile && ! freopen (outfile, "w", stdout))
 365     error (EXIT_FAILURE, errno, "%s", quotearg_colon (outfile));
 366   if (write_permuted_output (head_lines, line, lo_input, permutation, eolbyte)
 367       != 0)
 368     error (EXIT_FAILURE, errno, _("write error"));
 369
 370 #ifdef lint
 371   free (permutation);
 372   randint_all_free (randint_source);
 373   if (input_lines)
 374     {
 375       free (input_lines[0]);
 376       free (input_lines);
 377     }
 378 #endif
 379
 380   return EXIT_SUCCESS;
 381 }