src/unexpand.c

   1 /* unexpand - convert blanks to tabs
   2    Copyright (C) 89, 91, 1995-2006, 2008-2009 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
  17 /* By default, convert only maximal strings of initial blanks and tabs
  18    into tabs.
  19    Preserves backspace characters in the output; they decrement the
  20    column count for tab calculations.
  21    The default action is equivalent to -8.
  22
  23    Options:
  24    --tabs=tab1[,tab2[,...]]
  25    -t tab1[,tab2[,...]]
  26    -tab1[,tab2[,...]]   If only one tab stop is given, set the tabs tab1
  27                         columns apart instead of the default 8.  Otherwise,
  28                         set the tabs at columns tab1, tab2, etc. (numbered from
  29                         0); preserve any blanks beyond the tab stops given.
  30    --all
  31    -a                   Use tabs wherever they would replace 2 or more blanks,
  32                         not just at the beginnings of lines.
  33
  34    David MacKenzie <djm@gnu.ai.mit.edu> */
  35
  36 #include <config.h>
  37
  38 #include <stdio.h>
  39 #include <getopt.h>
  40 #include <sys/types.h>
  41 #include "system.h"
  42 #include "error.h"
  43 #include "quote.h"
  44 #include "xstrndup.h"
  45
  46 /* The official name of this program (e.g., no `g' prefix).  */
  47 #define PROGRAM_NAME "unexpand"
  48
  49 #define AUTHORS proper_name ("David MacKenzie")
  50
  51 /* If true, convert blanks even after nonblank characters have been
  52    read on the line.  */
  53 static bool convert_entire_line;
  54
  55 /* If nonzero, the size of all tab stops.  If zero, use `tab_list' instead.  */
  56 static size_t tab_size;
  57
  58 /* The maximum distance between tab stops.  */
  59 static size_t max_column_width;
  60
  61 /* Array of the explicit column numbers of the tab stops;
  62    after `tab_list' is exhausted, the rest of the line is printed
  63    unchanged.  The first column is column 0.  */
  64 static uintmax_t *tab_list;
  65
  66 /* The number of allocated entries in `tab_list'.  */
  67 static size_t n_tabs_allocated;
  68
  69 /* The index of the first invalid element of `tab_list',
  70    where the next element can be added.  */
  71 static size_t first_free_tab;
  72
  73 /* Null-terminated array of input filenames.  */
  74 static char **file_list;
  75
  76 /* Default for `file_list' if no files are given on the command line.  */
  77 static char *stdin_argv[] =
  78 {
  79   (char *) "-", NULL
  80 };
  81
  82 /* True if we have ever read standard input.  */
  83 static bool have_read_stdin;
  84
  85 /* The desired exit status.  */
  86 static int exit_status;
  87
  88 /* For long options that have no equivalent short option, use a
  89    non-character as a pseudo short option, starting with CHAR_MAX + 1.  */
  90 enum
  91 {
  92   CONVERT_FIRST_ONLY_OPTION = CHAR_MAX + 1
  93 };
  94
  95 static struct option const longopts[] =
  96 {
  97   {"tabs", required_argument, NULL, 't'},
  98   {"all", no_argument, NULL, 'a'},
  99   {"first-only", no_argument, NULL, CONVERT_FIRST_ONLY_OPTION},
 100   {GETOPT_HELP_OPTION_DECL},
 101   {GETOPT_VERSION_OPTION_DECL},
 102   {NULL, 0, NULL, 0}
 103 };
 104
 105 void
 106 usage (int status)
 107 {
 108   if (status != EXIT_SUCCESS)
 109     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 110              program_name);
 111   else
 112     {
 113       printf (_("\
 114 Usage: %s [OPTION]... [FILE]...\n\
 115 "),
 116               program_name);
 117       fputs (_("\
 118 Convert blanks in each FILE to tabs, writing to standard output.\n\
 119 With no FILE, or when FILE is -, read standard input.\n\
 120 \n\
 121 "), stdout);
 122       fputs (_("\
 123 Mandatory arguments to long options are mandatory for short options too.\n\
 124 "), stdout);
 125       fputs (_("\
 126   -a, --all        convert all blanks, instead of just initial blanks\n\
 127       --first-only  convert only leading sequences of blanks (overrides -a)\n\
 128   -t, --tabs=N     have tabs N characters apart instead of 8 (enables -a)\n\
 129   -t, --tabs=LIST  use comma separated LIST of tab positions (enables -a)\n\
 130 "), stdout);
 131       fputs (HELP_OPTION_DESCRIPTION, stdout);
 132       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 133       emit_ancillary_info ();
 134     }
 135   exit (status);
 136 }
 137
 138 /* Add tab stop TABVAL to the end of `tab_list'.  */
 139
 140 static void
 141 add_tab_stop (uintmax_t tabval)
 142 {
 143   uintmax_t prev_column = first_free_tab ? tab_list[first_free_tab - 1] : 0;
 144   uintmax_t column_width = prev_column <= tabval ? tabval - prev_column : 0;
 145
 146   if (first_free_tab == n_tabs_allocated)
 147     tab_list = X2NREALLOC (tab_list, &n_tabs_allocated);
 148   tab_list[first_free_tab++] = tabval;
 149
 150   if (max_column_width < column_width)
 151     {
 152       if (SIZE_MAX < column_width)
 153         error (EXIT_FAILURE, 0, _("tabs are too far apart"));
 154       max_column_width = column_width;
 155     }
 156 }
 157
 158 /* Add the comma or blank separated list of tab stops STOPS
 159    to the list of tab stops.  */
 160
 161 static void
 162 parse_tab_stops (char const *stops)
 163 {
 164   bool have_tabval = false;
 165   uintmax_t tabval IF_LINT (= 0);
 166   char const *num_start IF_LINT (= NULL);
 167   bool ok = true;
 168
 169   for (; *stops; stops++)
 170     {
 171       if (*stops == ',' || isblank (to_uchar (*stops)))
 172         {
 173           if (have_tabval)
 174             add_tab_stop (tabval);
 175           have_tabval = false;
 176         }
 177       else if (ISDIGIT (*stops))
 178         {
 179           if (!have_tabval)
 180             {
 181               tabval = 0;
 182               have_tabval = true;
 183               num_start = stops;
 184             }
 185
 186           /* Detect overflow.  */
 187           if (!DECIMAL_DIGIT_ACCUMULATE (tabval, *stops - '0', uintmax_t))
 188             {
 189               size_t len = strspn (num_start, "0123456789");
 190               char *bad_num = xstrndup (num_start, len);
 191               error (0, 0, _("tab stop is too large %s"), quote (bad_num));
 192               free (bad_num);
 193               ok = false;
 194               stops = num_start + len - 1;
 195             }
 196         }
 197       else
 198         {
 199           error (0, 0, _("tab size contains invalid character(s): %s"),
 200                  quote (stops));
 201           ok = false;
 202           break;
 203         }
 204     }
 205
 206   if (!ok)
 207     exit (EXIT_FAILURE);
 208
 209   if (have_tabval)
 210     add_tab_stop (tabval);
 211 }
 212
 213 /* Check that the list of tab stops TABS, with ENTRIES entries,
 214    contains only nonzero, ascending values.  */
 215
 216 static void
 217 validate_tab_stops (uintmax_t const *tabs, size_t entries)
 218 {
 219   uintmax_t prev_tab = 0;
 220   size_t i;
 221
 222   for (i = 0; i < entries; i++)
 223     {
 224       if (tabs[i] == 0)
 225         error (EXIT_FAILURE, 0, _("tab size cannot be 0"));
 226       if (tabs[i] <= prev_tab)
 227         error (EXIT_FAILURE, 0, _("tab sizes must be ascending"));
 228       prev_tab = tabs[i];
 229     }
 230 }
 231
 232 /* Close the old stream pointer FP if it is non-NULL,
 233    and return a new one opened to read the next input file.
 234    Open a filename of `-' as the standard input.
 235    Return NULL if there are no more input files.  */
 236
 237 static FILE *
 238 next_file (FILE *fp)
 239 {
 240   static char *prev_file;
 241   char *file;
 242
 243   if (fp)
 244     {
 245       if (ferror (fp))
 246         {
 247           error (0, errno, "%s", prev_file);
 248           exit_status = EXIT_FAILURE;
 249         }
 250       if (STREQ (prev_file, "-"))
 251         clearerr (fp);          /* Also clear EOF.  */
 252       else if (fclose (fp) != 0)
 253         {
 254           error (0, errno, "%s", prev_file);
 255           exit_status = EXIT_FAILURE;
 256         }
 257     }
 258
 259   while ((file = *file_list++) != NULL)
 260     {
 261       if (STREQ (file, "-"))
 262         {
 263           have_read_stdin = true;
 264           prev_file = file;
 265           return stdin;
 266         }
 267       fp = fopen (file, "r");
 268       if (fp)
 269         {
 270           prev_file = file;
 271           return fp;
 272         }
 273       error (0, errno, "%s", file);
 274       exit_status = EXIT_FAILURE;
 275     }
 276   return NULL;
 277 }
 278
 279 /* Change blanks to tabs, writing to stdout.
 280    Read each file in `file_list', in order.  */
 281
 282 static void
 283 unexpand (void)
 284 {
 285   /* Input stream.  */
 286   FILE *fp = next_file (NULL);
 287
 288   /* The array of pending blanks.  In non-POSIX locales, blanks can
 289      include characters other than spaces, so the blanks must be
 290      stored, not merely counted.  */
 291   char *pending_blank;
 292
 293   if (!fp)
 294     return;
 295
 296   /* The worst case is a non-blank character, then one blank, then a
 297      tab stop, then MAX_COLUMN_WIDTH - 1 blanks, then a non-blank; so
 298      allocate MAX_COLUMN_WIDTH bytes to store the blanks.  */
 299   pending_blank = xmalloc (max_column_width);
 300
 301   for (;;)
 302     {
 303       /* Input character, or EOF.  */
 304       int c;
 305
 306       /* If true, perform translations.  */
 307       bool convert = true;
 308
 309
 310       /* The following variables have valid values only when CONVERT
 311          is true:  */
 312
 313       /* Column of next input character.  */
 314       uintmax_t column = 0;
 315
 316       /* Column the next input tab stop is on.  */
 317       uintmax_t next_tab_column = 0;
 318
 319       /* Index in TAB_LIST of next tab stop to examine.  */
 320       size_t tab_index = 0;
 321
 322       /* If true, the first pending blank came just before a tab stop.  */
 323       bool one_blank_before_tab_stop = false;
 324
 325       /* If true, the previous input character was a blank.  This is
 326          initially true, since initial strings of blanks are treated
 327          as if the line was preceded by a blank.  */
 328       bool prev_blank = true;
 329
 330       /* Number of pending columns of blanks.  */
 331       size_t pending = 0;
 332
 333
 334       /* Convert a line of text.  */
 335
 336       do
 337         {
 338           while ((c = getc (fp)) < 0 && (fp = next_file (fp)))
 339             continue;
 340
 341           if (convert)
 342             {
 343               bool blank = !! isblank (c);
 344
 345               if (blank)
 346                 {
 347                   if (next_tab_column <= column)
 348                     {
 349                       if (tab_size)
 350                         next_tab_column =
 351                           column + (tab_size - column % tab_size);
 352                       else
 353                         for (;;)
 354                           if (tab_index == first_free_tab)
 355                             {
 356                               convert = false;
 357                               break;
 358                             }
 359                           else
 360                             {
 361                               uintmax_t tab = tab_list[tab_index++];
 362                               if (column < tab)
 363                                 {
 364                                   next_tab_column = tab;
 365                                   break;
 366                                 }
 367                             }
 368                     }
 369
 370                   if (convert)
 371                     {
 372                       if (next_tab_column < column)
 373                         error (EXIT_FAILURE, 0, _("input line is too long"));
 374
 375                       if (c == '\t')
 376                         {
 377                           column = next_tab_column;
 378
 379                           /* Discard pending blanks, unless it was a single
 380                              blank just before the previous tab stop.  */
 381                           if (! (pending == 1 && one_blank_before_tab_stop))
 382                             {
 383                               pending = 0;
 384                               one_blank_before_tab_stop = false;
 385                             }
 386                         }
 387                       else
 388                         {
 389                           column++;
 390
 391                           if (! (prev_blank && column == next_tab_column))
 392                             {
 393                               /* It is not yet known whether the pending blanks
 394                                  will be replaced by tabs.  */
 395                               if (column == next_tab_column)
 396                                 one_blank_before_tab_stop = true;
 397                               pending_blank[pending++] = c;
 398                               prev_blank = true;
 399                               continue;
 400                             }
 401
 402                           /* Replace the pending blanks by a tab or two.  */
 403                           pending_blank[0] = c = '\t';
 404                           pending = one_blank_before_tab_stop;
 405                         }
 406                     }
 407                 }
 408               else if (c == '\b')
 409                 {
 410                   /* Go back one column, and force recalculation of the
 411                      next tab stop.  */
 412                   column -= !!column;
 413                   next_tab_column = column;
 414                   tab_index -= !!tab_index;
 415                 }
 416               else
 417                 {
 418                   column++;
 419                   if (!column)
 420                     error (EXIT_FAILURE, 0, _("input line is too long"));
 421                 }
 422
 423               if (pending)
 424                 {
 425                   if (fwrite (pending_blank, 1, pending, stdout) != pending)
 426                     error (EXIT_FAILURE, errno, _("write error"));
 427                   pending = 0;
 428                   one_blank_before_tab_stop = false;
 429                 }
 430
 431               prev_blank = blank;
 432               convert &= convert_entire_line || blank;
 433             }
 434
 435           if (c < 0)
 436             {
 437               free (pending_blank);
 438               return;
 439             }
 440
 441           if (putchar (c) < 0)
 442             error (EXIT_FAILURE, errno, _("write error"));
 443         }
 444       while (c != '\n');
 445     }
 446 }
 447
 448 int
 449 main (int argc, char **argv)
 450 {
 451   bool have_tabval = false;
 452   uintmax_t tabval IF_LINT (= 0);
 453   int c;
 454
 455   /* If true, cancel the effect of any -a (explicit or implicit in -t),
 456      so that only leading blanks will be considered.  */
 457   bool convert_first_only = false;
 458
 459   initialize_main (&argc, &argv);
 460   set_program_name (argv[0]);
 461   setlocale (LC_ALL, "");
 462   bindtextdomain (PACKAGE, LOCALEDIR);
 463   textdomain (PACKAGE);
 464
 465   atexit (close_stdout);
 466
 467   have_read_stdin = false;
 468   exit_status = EXIT_SUCCESS;
 469   convert_entire_line = false;
 470   tab_list = NULL;
 471   first_free_tab = 0;
 472
 473   while ((c = getopt_long (argc, argv, ",0123456789at:", longopts, NULL))
 474          != -1)
 475     {
 476       switch (c)
 477         {
 478         case '?':
 479           usage (EXIT_FAILURE);
 480         case 'a':
 481           convert_entire_line = true;
 482           break;
 483         case 't':
 484           convert_entire_line = true;
 485           parse_tab_stops (optarg);
 486           break;
 487         case CONVERT_FIRST_ONLY_OPTION:
 488           convert_first_only = true;
 489           break;
 490         case ',':
 491           if (have_tabval)
 492             add_tab_stop (tabval);
 493           have_tabval = false;
 494           break;
 495         case_GETOPT_HELP_CHAR;
 496         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 497         default:
 498           if (!have_tabval)
 499             {
 500               tabval = 0;
 501               have_tabval = true;
 502             }
 503           if (!DECIMAL_DIGIT_ACCUMULATE (tabval, c - '0', uintmax_t))
 504             error (EXIT_FAILURE, 0, _("tab stop value is too large"));
 505           break;
 506         }
 507     }
 508
 509   if (convert_first_only)
 510     convert_entire_line = false;
 511
 512   if (have_tabval)
 513     add_tab_stop (tabval);
 514
 515   validate_tab_stops (tab_list, first_free_tab);
 516
 517   if (first_free_tab == 0)
 518     tab_size = max_column_width = 8;
 519   else if (first_free_tab == 1)
 520     tab_size = tab_list[0];
 521   else
 522     tab_size = 0;
 523
 524   file_list = (optind < argc ? &argv[optind] : stdin_argv);
 525
 526   unexpand ();
 527
 528   if (have_read_stdin && fclose (stdin) != 0)
 529     error (EXIT_FAILURE, errno, "-");
 530
 531   exit (exit_status);
 532 }