src/nl.c

   1 /* nl -- number lines of files
   2    Copyright (C) 1989, 1992, 1995-2011 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 /* Written by Scott Bartram (nancy!scott@uunet.uu.net)
  18    Revised by David MacKenzie (djm@gnu.ai.mit.edu) */
  19
  20 #include <config.h>
  21
  22 #include <stdio.h>
  23 #include <sys/types.h>
  24 #include <getopt.h>
  25
  26 #include "system.h"
  27
  28 #include <regex.h>
  29
  30 #include "error.h"
  31 #include "fadvise.h"
  32 #include "linebuffer.h"
  33 #include "quote.h"
  34 #include "xstrtol.h"
  35
  36 /* The official name of this program (e.g., no `g' prefix).  */
  37 #define PROGRAM_NAME "nl"
  38
  39 #define AUTHORS \
  40   proper_name ("Scott Bartram"), \
  41   proper_name ("David MacKenzie")
  42
  43 /* Line-number formats.  They are given an int width, an intmax_t
  44    value, and a string separator.  */
  45
  46 /* Right justified, no leading zeroes.  */
  47 static char const FORMAT_RIGHT_NOLZ[] = "%*" PRIdMAX "%s";
  48
  49 /* Right justified, leading zeroes.  */
  50 static char const FORMAT_RIGHT_LZ[] = "%0*" PRIdMAX "%s";
  51
  52 /* Left justified, no leading zeroes.  */
  53 static char const FORMAT_LEFT[] = "%-*" PRIdMAX "%s";
  54
  55 /* Default section delimiter characters.  */
  56 static char const DEFAULT_SECTION_DELIMITERS[] = "\\:";
  57
  58 /* Types of input lines: either one of the section delimiters,
  59    or text to output. */
  60 enum section
  61 {
  62   Header, Body, Footer, Text
  63 };
  64
  65 /* Format of body lines (-b).  */
  66 static char const *body_type = "t";
  67
  68 /* Format of header lines (-h).  */
  69 static char const *header_type = "n";
  70
  71 /* Format of footer lines (-f).  */
  72 static char const *footer_type = "n";
  73
  74 /* Format currently being used (body, header, or footer).  */
  75 static char const *current_type;
  76
  77 /* Regex for body lines to number (-bp).  */
  78 static struct re_pattern_buffer body_regex;
  79
  80 /* Regex for header lines to number (-hp).  */
  81 static struct re_pattern_buffer header_regex;
  82
  83 /* Regex for footer lines to number (-fp).  */
  84 static struct re_pattern_buffer footer_regex;
  85
  86 /* Fastmaps for the above.  */
  87 static char body_fastmap[UCHAR_MAX + 1];
  88 static char header_fastmap[UCHAR_MAX + 1];
  89 static char footer_fastmap[UCHAR_MAX + 1];
  90
  91 /* Pointer to current regex, if any.  */
  92 static struct re_pattern_buffer *current_regex = NULL;
  93
  94 /* Separator string to print after line number (-s).  */
  95 static char const *separator_str = "\t";
  96
  97 /* Input section delimiter string (-d).  */
  98 static char const *section_del = DEFAULT_SECTION_DELIMITERS;
  99
 100 /* Header delimiter string.  */
 101 static char *header_del = NULL;
 102
 103 /* Header section delimiter length.  */
 104 static size_t header_del_len;
 105
 106 /* Body delimiter string.  */
 107 static char *body_del = NULL;
 108
 109 /* Body section delimiter length.  */
 110 static size_t body_del_len;
 111
 112 /* Footer delimiter string.  */
 113 static char *footer_del = NULL;
 114
 115 /* Footer section delimiter length.  */
 116 static size_t footer_del_len;
 117
 118 /* Input buffer.  */
 119 static struct linebuffer line_buf;
 120
 121 /* printf format string for unnumbered lines.  */
 122 static char *print_no_line_fmt = NULL;
 123
 124 /* Starting line number on each page (-v).  */
 125 static intmax_t starting_line_number = 1;
 126
 127 /* Line number increment (-i).  */
 128 static intmax_t page_incr = 1;
 129
 130 /* If true, reset line number at start of each page (-p).  */
 131 static bool reset_numbers = true;
 132
 133 /* Number of blank lines to consider to be one line for numbering (-l).  */
 134 static intmax_t blank_join = 1;
 135
 136 /* Width of line numbers (-w).  */
 137 static int lineno_width = 6;
 138
 139 /* Line number format (-n).  */
 140 static char const *lineno_format = FORMAT_RIGHT_NOLZ;
 141
 142 /* Current print line number.  */
 143 static intmax_t line_no;
 144
 145 /* True if we have ever read standard input.  */
 146 static bool have_read_stdin;
 147
 148 enum
 149 {
 150   PAGE_INCREMENT_OPTION_DEPRECATED = CHAR_MAX + 1
 151 };
 152
 153 static struct option const longopts[] =
 154 {
 155   {"header-numbering", required_argument, NULL, 'h'},
 156   {"body-numbering", required_argument, NULL, 'b'},
 157   {"footer-numbering", required_argument, NULL, 'f'},
 158   {"starting-line-number", required_argument, NULL, 'v'},
 159   {"line-increment", required_argument, NULL, 'i'},
 160   /* FIXME: page-increment is deprecated, remove in dec-2011.  */
 161   {"page-increment", required_argument, NULL, PAGE_INCREMENT_OPTION_DEPRECATED},
 162   {"no-renumber", no_argument, NULL, 'p'},
 163   {"join-blank-lines", required_argument, NULL, 'l'},
 164   {"number-separator", required_argument, NULL, 's'},
 165   {"number-width", required_argument, NULL, 'w'},
 166   {"number-format", required_argument, NULL, 'n'},
 167   {"section-delimiter", required_argument, NULL, 'd'},
 168   {GETOPT_HELP_OPTION_DECL},
 169   {GETOPT_VERSION_OPTION_DECL},
 170   {NULL, 0, NULL, 0}
 171 };
 172
 173 /* Print a usage message and quit. */
 174
 175 void
 176 usage (int status)
 177 {
 178   if (status != EXIT_SUCCESS)
 179     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 180              program_name);
 181   else
 182     {
 183       printf (_("\
 184 Usage: %s [OPTION]... [FILE]...\n\
 185 "),
 186               program_name);
 187       fputs (_("\
 188 Write each FILE to standard output, with line numbers added.\n\
 189 With no FILE, or when FILE is -, read standard input.\n\
 190 \n\
 191 "), stdout);
 192       fputs (_("\
 193 Mandatory arguments to long options are mandatory for short options too.\n\
 194 "), stdout);
 195       fputs (_("\
 196   -b, --body-numbering=STYLE      use STYLE for numbering body lines\n\
 197   -d, --section-delimiter=CC      use CC for separating logical pages\n\
 198   -f, --footer-numbering=STYLE    use STYLE for numbering footer lines\n\
 199 "), stdout);
 200       fputs (_("\
 201   -h, --header-numbering=STYLE    use STYLE for numbering header lines\n\
 202   -i, --line-increment=NUMBER     line number increment at each line\n\
 203   -l, --join-blank-lines=NUMBER   group of NUMBER empty lines counted as one\n\
 204   -n, --number-format=FORMAT      insert line numbers according to FORMAT\n\
 205   -p, --no-renumber               do not reset line numbers at logical pages\n\
 206   -s, --number-separator=STRING   add STRING after (possible) line number\n\
 207 "), stdout);
 208       fputs (_("\
 209   -v, --starting-line-number=NUMBER  first line number on each logical page\n\
 210   -w, --number-width=NUMBER       use NUMBER columns for line numbers\n\
 211 "), stdout);
 212       fputs (HELP_OPTION_DESCRIPTION, stdout);
 213       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 214       fputs (_("\
 215 \n\
 216 By default, selects -v1 -i1 -l1 -sTAB -w6 -nrn -hn -bt -fn.  CC are\n\
 217 two delimiter characters for separating logical pages, a missing\n\
 218 second character implies :.  Type \\\\ for \\.  STYLE is one of:\n\
 219 "), stdout);
 220       fputs (_("\
 221 \n\
 222   a         number all lines\n\
 223   t         number only nonempty lines\n\
 224   n         number no lines\n\
 225   pBRE      number only lines that contain a match for the basic regular\n\
 226             expression, BRE\n\
 227 \n\
 228 FORMAT is one of:\n\
 229 \n\
 230   ln   left justified, no leading zeros\n\
 231   rn   right justified, no leading zeros\n\
 232   rz   right justified, leading zeros\n\
 233 \n\
 234 "), stdout);
 235       emit_ancillary_info ();
 236     }
 237   exit (status);
 238 }
 239
 240 /* Set the command line flag TYPEP and possibly the regex pointer REGEXP,
 241    according to `optarg'.  */
 242
 243 static bool
 244 build_type_arg (char const **typep,
 245                 struct re_pattern_buffer *regexp, char *fastmap)
 246 {
 247   char const *errmsg;
 248   bool rval = true;
 249
 250   switch (*optarg)
 251     {
 252     case 'a':
 253     case 't':
 254     case 'n':
 255       *typep = optarg;
 256       break;
 257     case 'p':
 258       *typep = optarg++;
 259       regexp->buffer = NULL;
 260       regexp->allocated = 0;
 261       regexp->fastmap = fastmap;
 262       regexp->translate = NULL;
 263       re_syntax_options =
 264         RE_SYNTAX_POSIX_BASIC & ~RE_CONTEXT_INVALID_DUP & ~RE_NO_EMPTY_RANGES;
 265       errmsg = re_compile_pattern (optarg, strlen (optarg), regexp);
 266       if (errmsg)
 267         error (EXIT_FAILURE, 0, "%s", errmsg);
 268       break;
 269     default:
 270       rval = false;
 271       break;
 272     }
 273   return rval;
 274 }
 275
 276 /* Print the line number and separator; increment the line number. */
 277
 278 static void
 279 print_lineno (void)
 280 {
 281   intmax_t next_line_no;
 282
 283   printf (lineno_format, lineno_width, line_no, separator_str);
 284
 285   next_line_no = line_no + page_incr;
 286   if (next_line_no < line_no)
 287     error (EXIT_FAILURE, 0, _("line number overflow"));
 288   line_no = next_line_no;
 289 }
 290
 291 /* Switch to a header section. */
 292
 293 static void
 294 proc_header (void)
 295 {
 296   current_type = header_type;
 297   current_regex = &header_regex;
 298   if (reset_numbers)
 299     line_no = starting_line_number;
 300   putchar ('\n');
 301 }
 302
 303 /* Switch to a body section. */
 304
 305 static void
 306 proc_body (void)
 307 {
 308   current_type = body_type;
 309   current_regex = &body_regex;
 310   putchar ('\n');
 311 }
 312
 313 /* Switch to a footer section. */
 314
 315 static void
 316 proc_footer (void)
 317 {
 318   current_type = footer_type;
 319   current_regex = &footer_regex;
 320   putchar ('\n');
 321 }
 322
 323 /* Process a regular text line in `line_buf'. */
 324
 325 static void
 326 proc_text (void)
 327 {
 328   static intmax_t blank_lines = 0;      /* Consecutive blank lines so far. */
 329
 330   switch (*current_type)
 331     {
 332     case 'a':
 333       if (blank_join > 1)
 334         {
 335           if (1 < line_buf.length || ++blank_lines == blank_join)
 336             {
 337               print_lineno ();
 338               blank_lines = 0;
 339             }
 340           else
 341             fputs (print_no_line_fmt, stdout);
 342         }
 343       else
 344         print_lineno ();
 345       break;
 346     case 't':
 347       if (1 < line_buf.length)
 348         print_lineno ();
 349       else
 350         fputs (print_no_line_fmt, stdout);
 351       break;
 352     case 'n':
 353       fputs (print_no_line_fmt, stdout);
 354       break;
 355     case 'p':
 356       switch (re_search (current_regex, line_buf.buffer, line_buf.length - 1,
 357                          0, line_buf.length - 1, NULL))
 358         {
 359         case -2:
 360           error (EXIT_FAILURE, errno, _("error in regular expression search"));
 361
 362         case -1:
 363           fputs (print_no_line_fmt, stdout);
 364           break;
 365
 366         default:
 367           print_lineno ();
 368           break;
 369         }
 370     }
 371   fwrite (line_buf.buffer, sizeof (char), line_buf.length, stdout);
 372 }
 373
 374 /* Return the type of line in `line_buf'. */
 375
 376 static enum section
 377 check_section (void)
 378 {
 379   size_t len = line_buf.length - 1;
 380
 381   if (len < 2 || memcmp (line_buf.buffer, section_del, 2))
 382     return Text;
 383   if (len == header_del_len
 384       && !memcmp (line_buf.buffer, header_del, header_del_len))
 385     return Header;
 386   if (len == body_del_len
 387       && !memcmp (line_buf.buffer, body_del, body_del_len))
 388     return Body;
 389   if (len == footer_del_len
 390       && !memcmp (line_buf.buffer, footer_del, footer_del_len))
 391     return Footer;
 392   return Text;
 393 }
 394
 395 /* Read and process the file pointed to by FP. */
 396
 397 static void
 398 process_file (FILE *fp)
 399 {
 400   while (readlinebuffer (&line_buf, fp))
 401     {
 402       switch (check_section ())
 403         {
 404         case Header:
 405           proc_header ();
 406           break;
 407         case Body:
 408           proc_body ();
 409           break;
 410         case Footer:
 411           proc_footer ();
 412           break;
 413         case Text:
 414           proc_text ();
 415           break;
 416         }
 417     }
 418 }
 419
 420 /* Process file FILE to standard output.
 421    Return true if successful.  */
 422
 423 static bool
 424 nl_file (char const *file)
 425 {
 426   FILE *stream;
 427
 428   if (STREQ (file, "-"))
 429     {
 430       have_read_stdin = true;
 431       stream = stdin;
 432     }
 433   else
 434     {
 435       stream = fopen (file, "r");
 436       if (stream == NULL)
 437         {
 438           error (0, errno, "%s", file);
 439           return false;
 440         }
 441     }
 442
 443   fadvise (stream, FADVISE_SEQUENTIAL);
 444
 445   process_file (stream);
 446
 447   if (ferror (stream))
 448     {
 449       error (0, errno, "%s", file);
 450       return false;
 451     }
 452   if (STREQ (file, "-"))
 453     clearerr (stream);          /* Also clear EOF. */
 454   else if (fclose (stream) == EOF)
 455     {
 456       error (0, errno, "%s", file);
 457       return false;
 458     }
 459   return true;
 460 }
 461
 462 int
 463 main (int argc, char **argv)
 464 {
 465   int c;
 466   size_t len;
 467   bool ok = true;
 468
 469   initialize_main (&argc, &argv);
 470   set_program_name (argv[0]);
 471   setlocale (LC_ALL, "");
 472   bindtextdomain (PACKAGE, LOCALEDIR);
 473   textdomain (PACKAGE);
 474
 475   atexit (close_stdout);
 476
 477   have_read_stdin = false;
 478
 479   while ((c = getopt_long (argc, argv, "h:b:f:v:i:pl:s:w:n:d:", longopts,
 480                            NULL)) != -1)
 481     {
 482       switch (c)
 483         {
 484         case 'h':
 485           if (! build_type_arg (&header_type, &header_regex, header_fastmap))
 486             {
 487               error (0, 0, _("invalid header numbering style: %s"),
 488                      quote (optarg));
 489               ok = false;
 490             }
 491           break;
 492         case 'b':
 493           if (! build_type_arg (&body_type, &body_regex, body_fastmap))
 494             {
 495               error (0, 0, _("invalid body numbering style: %s"),
 496                      quote (optarg));
 497               ok = false;
 498             }
 499           break;
 500         case 'f':
 501           if (! build_type_arg (&footer_type, &footer_regex, footer_fastmap))
 502             {
 503               error (0, 0, _("invalid footer numbering style: %s"),
 504                      quote (optarg));
 505               ok = false;
 506             }
 507           break;
 508         case 'v':
 509           if (xstrtoimax (optarg, NULL, 10, &starting_line_number, "")
 510               != LONGINT_OK)
 511             {
 512               error (0, 0, _("invalid starting line number: %s"),
 513                      quote (optarg));
 514               ok = false;
 515             }
 516           break;
 517   case PAGE_INCREMENT_OPTION_DEPRECATED:
 518     error (0, 0, _("WARNING: --page-increment is deprecated; "
 519                    "use --line-increment instead"));
 520     /* fall through */
 521         case 'i':
 522           if (! (xstrtoimax (optarg, NULL, 10, &page_incr, "") == LONGINT_OK
 523                  && 0 < page_incr))
 524             {
 525               error (0, 0, _("invalid line number increment: %s"),
 526                      quote (optarg));
 527               ok = false;
 528             }
 529           break;
 530         case 'p':
 531           reset_numbers = false;
 532           break;
 533         case 'l':
 534           if (! (xstrtoimax (optarg, NULL, 10, &blank_join, "") == LONGINT_OK
 535                  && 0 < blank_join))
 536             {
 537               error (0, 0, _("invalid number of blank lines: %s"),
 538                      quote (optarg));
 539               ok = false;
 540             }
 541           break;
 542         case 's':
 543           separator_str = optarg;
 544           break;
 545         case 'w':
 546           {
 547             long int tmp_long;
 548             if (xstrtol (optarg, NULL, 10, &tmp_long, "") != LONGINT_OK
 549                 || tmp_long <= 0 || tmp_long > INT_MAX)
 550               {
 551                 error (0, 0, _("invalid line number field width: %s"),
 552                        quote (optarg));
 553                 ok = false;
 554               }
 555             else
 556               {
 557                 lineno_width = tmp_long;
 558               }
 559           }
 560           break;
 561         case 'n':
 562           if (STREQ (optarg, "ln"))
 563             lineno_format = FORMAT_LEFT;
 564           else if (STREQ (optarg, "rn"))
 565             lineno_format = FORMAT_RIGHT_NOLZ;
 566           else if (STREQ (optarg, "rz"))
 567             lineno_format = FORMAT_RIGHT_LZ;
 568           else
 569             {
 570               error (0, 0, _("invalid line numbering format: %s"),
 571                      quote (optarg));
 572               ok = false;
 573             }
 574           break;
 575         case 'd':
 576           section_del = optarg;
 577           break;
 578         case_GETOPT_HELP_CHAR;
 579         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 580         default:
 581           ok = false;
 582           break;
 583         }
 584     }
 585
 586   if (!ok)
 587     usage (EXIT_FAILURE);
 588
 589   /* Initialize the section delimiters.  */
 590   len = strlen (section_del);
 591
 592   header_del_len = len * 3;
 593   header_del = xmalloc (header_del_len + 1);
 594   stpcpy (stpcpy (stpcpy (header_del, section_del), section_del), section_del);
 595
 596   body_del_len = len * 2;
 597   body_del = xmalloc (body_del_len + 1);
 598   stpcpy (stpcpy (body_del, section_del), section_del);
 599
 600   footer_del_len = len;
 601   footer_del = xmalloc (footer_del_len + 1);
 602   stpcpy (footer_del, section_del);
 603
 604   /* Initialize the input buffer.  */
 605   initbuffer (&line_buf);
 606
 607   /* Initialize the printf format for unnumbered lines. */
 608   len = strlen (separator_str);
 609   print_no_line_fmt = xmalloc (lineno_width + len + 1);
 610   memset (print_no_line_fmt, ' ', lineno_width + len);
 611   print_no_line_fmt[lineno_width + len] = '\0';
 612
 613   line_no = starting_line_number;
 614   current_type = body_type;
 615   current_regex = &body_regex;
 616
 617   /* Main processing. */
 618
 619   if (optind == argc)
 620     ok = nl_file ("-");
 621   else
 622     for (; optind < argc; optind++)
 623       ok &= nl_file (argv[optind]);
 624
 625   if (have_read_stdin && fclose (stdin) == EOF)
 626     error (EXIT_FAILURE, errno, "-");
 627
 628   exit (ok ? EXIT_SUCCESS : EXIT_FAILURE);
 629 }