src/nl.c

   1 /* nl -- number lines of files
   2    Copyright (C) 89, 92, 1995-2009 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 \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 "linebuffer.h"
  32 #include "quote.h"
  33 #include "xstrtol.h"
  34
  35 /* The official name of this program (e.g., no `g' prefix).  */
  36 #define PROGRAM_NAME "nl"
  37
  38 #define AUTHORS \
  39   proper_name ("Scott Bartram"), \
  40   proper_name ("David MacKenzie")
  41
  42 /* Line-number formats.  They are given an int width, an intmax_t
  43    value, and a string separator.  */
  44
  45 /* Right justified, no leading zeroes.  */
  46 static char const FORMAT_RIGHT_NOLZ[] = "%*" PRIdMAX "%s";
  47
  48 /* Right justified, leading zeroes.  */
  49 static char const FORMAT_RIGHT_LZ[] = "%0*" PRIdMAX "%s";
  50
  51 /* Left justified, no leading zeroes.  */
  52 static char const FORMAT_LEFT[] = "%-*" PRIdMAX "%s";
  53
  54 /* Default section delimiter characters.  */
  55 static char const DEFAULT_SECTION_DELIMITERS[] = "\\:";
  56
  57 /* Types of input lines: either one of the section delimiters,
  58    or text to output. */
  59 enum section
  60 {
  61   Header, Body, Footer, Text
  62 };
  63
  64 /* Format of body lines (-b).  */
  65 static char const *body_type = "t";
  66
  67 /* Format of header lines (-h).  */
  68 static char const *header_type = "n";
  69
  70 /* Format of footer lines (-f).  */
  71 static char const *footer_type = "n";
  72
  73 /* Format currently being used (body, header, or footer).  */
  74 static char const *current_type;
  75
  76 /* Regex for body lines to number (-bp).  */
  77 static struct re_pattern_buffer body_regex;
  78
  79 /* Regex for header lines to number (-hp).  */
  80 static struct re_pattern_buffer header_regex;
  81
  82 /* Regex for footer lines to number (-fp).  */
  83 static struct re_pattern_buffer footer_regex;
  84
  85 /* Fastmaps for the above.  */
  86 static char body_fastmap[UCHAR_MAX + 1];
  87 static char header_fastmap[UCHAR_MAX + 1];
  88 static char footer_fastmap[UCHAR_MAX + 1];
  89
  90 /* Pointer to current regex, if any.  */
  91 static struct re_pattern_buffer *current_regex = NULL;
  92
  93 /* Separator string to print after line number (-s).  */
  94 static char const *separator_str = "\t";
  95
  96 /* Input section delimiter string (-d).  */
  97 static char const *section_del = DEFAULT_SECTION_DELIMITERS;
  98
  99 /* Header delimiter string.  */
 100 static char *header_del = NULL;
 101
 102 /* Header section delimiter length.  */
 103 static size_t header_del_len;
 104
 105 /* Body delimiter string.  */
 106 static char *body_del = NULL;
 107
 108 /* Body section delimiter length.  */
 109 static size_t body_del_len;
 110
 111 /* Footer delimiter string.  */
 112 static char *footer_del = NULL;
 113
 114 /* Footer section delimiter length.  */
 115 static size_t footer_del_len;
 116
 117 /* Input buffer.  */
 118 static struct linebuffer line_buf;
 119
 120 /* printf format string for unnumbered lines.  */
 121 static char *print_no_line_fmt = NULL;
 122
 123 /* Starting line number on each page (-v).  */
 124 static intmax_t starting_line_number = 1;
 125
 126 /* Line number increment (-i).  */
 127 static intmax_t page_incr = 1;
 128
 129 /* If true, reset line number at start of each page (-p).  */
 130 static bool reset_numbers = true;
 131
 132 /* Number of blank lines to consider to be one line for numbering (-l).  */
 133 static intmax_t blank_join = 1;
 134
 135 /* Width of line numbers (-w).  */
 136 static int lineno_width = 6;
 137
 138 /* Line number format (-n).  */
 139 static char const *lineno_format = FORMAT_RIGHT_NOLZ;
 140
 141 /* Current print line number.  */
 142 static intmax_t line_no;
 143
 144 /* True if we have ever read standard input.  */
 145 static bool have_read_stdin;
 146
 147 enum
 148 {
 149   PAGE_INCREMENT_OPTION_DEPRECATED = CHAR_MAX + 1
 150 };
 151
 152 static struct option const longopts[] =
 153 {
 154   {"header-numbering", required_argument, NULL, 'h'},
 155   {"body-numbering", required_argument, NULL, 'b'},
 156   {"footer-numbering", required_argument, NULL, 'f'},
 157   {"starting-line-number", required_argument, NULL, 'v'},
 158   {"line-increment", required_argument, NULL, 'i'},
 159   /* FIXME: page-increment is deprecated, remove in dec-2011.  */
 160   {"page-increment", required_argument, NULL, PAGE_INCREMENT_OPTION_DEPRECATED},
 161   {"no-renumber", no_argument, NULL, 'p'},
 162   {"join-blank-lines", required_argument, NULL, 'l'},
 163   {"number-separator", required_argument, NULL, 's'},
 164   {"number-width", required_argument, NULL, 'w'},
 165   {"number-format", required_argument, NULL, 'n'},
 166   {"section-delimiter", required_argument, NULL, 'd'},
 167   {GETOPT_HELP_OPTION_DECL},
 168   {GETOPT_VERSION_OPTION_DECL},
 169   {NULL, 0, NULL, 0}
 170 };
 171
 172 /* Print a usage message and quit. */
 173
 174 void
 175 usage (int status)
 176 {
 177   if (status != EXIT_SUCCESS)
 178     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 179              program_name);
 180   else
 181     {
 182       printf (_("\
 183 Usage: %s [OPTION]... [FILE]...\n\
 184 "),
 185               program_name);
 186       fputs (_("\
 187 Write each FILE to standard output, with line numbers added.\n\
 188 With no FILE, or when FILE is -, read standard input.\n\
 189 \n\
 190 "), stdout);
 191       fputs (_("\
 192 Mandatory arguments to long options are mandatory for short options too.\n\
 193 "), stdout);
 194       fputs (_("\
 195   -b, --body-numbering=STYLE      use STYLE for numbering body lines\n\
 196   -d, --section-delimiter=CC      use CC for separating logical pages\n\
 197   -f, --footer-numbering=STYLE    use STYLE for numbering footer lines\n\
 198 "), stdout);
 199       fputs (_("\
 200   -h, --header-numbering=STYLE    use STYLE for numbering header lines\n\
 201   -i, --line-increment=NUMBER     line number increment at each line\n\
 202   -l, --join-blank-lines=NUMBER   group of NUMBER empty lines counted as one\n\
 203   -n, --number-format=FORMAT      insert line numbers according to FORMAT\n\
 204   -p, --no-renumber               do not reset line numbers at logical pages\n\
 205   -s, --number-separator=STRING   add STRING after (possible) line number\n\
 206 "), stdout);
 207       fputs (_("\
 208   -v, --starting-line-number=NUMBER  first line number on each logical page\n\
 209   -w, --number-width=NUMBER       use NUMBER columns for line numbers\n\
 210 "), stdout);
 211       fputs (HELP_OPTION_DESCRIPTION, stdout);
 212       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 213       fputs (_("\
 214 \n\
 215 By default, selects -v1 -i1 -l1 -sTAB -w6 -nrn -hn -bt -fn.  CC are\n\
 216 two delimiter characters for separating logical pages, a missing\n\
 217 second character implies :.  Type \\\\ for \\.  STYLE is one of:\n\
 218 "), stdout);
 219       fputs (_("\
 220 \n\
 221   a         number all lines\n\
 222   t         number only nonempty lines\n\
 223   n         number no lines\n\
 224   pBRE      number only lines that contain a match for the basic regular\n\
 225             expression, BRE\n\
 226 \n\
 227 FORMAT is one of:\n\
 228 \n\
 229   ln   left justified, no leading zeros\n\
 230   rn   right justified, no leading zeros\n\
 231   rz   right justified, leading zeros\n\
 232 \n\
 233 "), stdout);
 234       emit_ancillary_info ();
 235     }
 236   exit (status);
 237 }
 238
 239 /* Set the command line flag TYPEP and possibly the regex pointer REGEXP,
 240    according to `optarg'.  */
 241
 242 static bool
 243 build_type_arg (char const **typep,
 244                 struct re_pattern_buffer *regexp, char *fastmap)
 245 {
 246   char const *errmsg;
 247   bool rval = true;
 248
 249   switch (*optarg)
 250     {
 251     case 'a':
 252     case 't':
 253     case 'n':
 254       *typep = optarg;
 255       break;
 256     case 'p':
 257       *typep = optarg++;
 258       regexp->buffer = NULL;
 259       regexp->allocated = 0;
 260       regexp->fastmap = fastmap;
 261       regexp->translate = NULL;
 262       re_syntax_options =
 263         RE_SYNTAX_POSIX_BASIC & ~RE_CONTEXT_INVALID_DUP & ~RE_NO_EMPTY_RANGES;
 264       errmsg = re_compile_pattern (optarg, strlen (optarg), regexp);
 265       if (errmsg)
 266         error (EXIT_FAILURE, 0, "%s", errmsg);
 267       break;
 268     default:
 269       rval = false;
 270       break;
 271     }
 272   return rval;
 273 }
 274
 275 /* Print the line number and separator; increment the line number. */
 276
 277 static void
 278 print_lineno (void)
 279 {
 280   intmax_t next_line_no;
 281
 282   printf (lineno_format, lineno_width, line_no, separator_str);
 283
 284   next_line_no = line_no + page_incr;
 285   if (next_line_no < line_no)
 286     error (EXIT_FAILURE, 0, _("line number overflow"));
 287   line_no = next_line_no;
 288 }
 289
 290 /* Switch to a header section. */
 291
 292 static void
 293 proc_header (void)
 294 {
 295   current_type = header_type;
 296   current_regex = &header_regex;
 297   if (reset_numbers)
 298     line_no = starting_line_number;
 299   putchar ('\n');
 300 }
 301
 302 /* Switch to a body section. */
 303
 304 static void
 305 proc_body (void)
 306 {
 307   current_type = body_type;
 308   current_regex = &body_regex;
 309   putchar ('\n');
 310 }
 311
 312 /* Switch to a footer section. */
 313
 314 static void
 315 proc_footer (void)
 316 {
 317   current_type = footer_type;
 318   current_regex = &footer_regex;
 319   putchar ('\n');
 320 }
 321
 322 /* Process a regular text line in `line_buf'. */
 323
 324 static void
 325 proc_text (void)
 326 {
 327   static intmax_t blank_lines = 0;      /* Consecutive blank lines so far. */
 328
 329   switch (*current_type)
 330     {
 331     case 'a':
 332       if (blank_join > 1)
 333         {
 334           if (1 < line_buf.length || ++blank_lines == blank_join)
 335             {
 336               print_lineno ();
 337               blank_lines = 0;
 338             }
 339           else
 340             fputs (print_no_line_fmt, stdout);
 341         }
 342       else
 343         print_lineno ();
 344       break;
 345     case 't':
 346       if (1 < line_buf.length)
 347         print_lineno ();
 348       else
 349         fputs (print_no_line_fmt, stdout);
 350       break;
 351     case 'n':
 352       fputs (print_no_line_fmt, stdout);
 353       break;
 354     case 'p':
 355       switch (re_search (current_regex, line_buf.buffer, line_buf.length - 1,
 356                          0, line_buf.length - 1, NULL))
 357         {
 358         case -2:
 359           error (EXIT_FAILURE, errno, _("error in regular expression search"));
 360
 361         case -1:
 362           fputs (print_no_line_fmt, stdout);
 363           break;
 364
 365         default:
 366           print_lineno ();
 367           break;
 368         }
 369     }
 370   fwrite (line_buf.buffer, sizeof (char), line_buf.length, stdout);
 371 }
 372
 373 /* Return the type of line in `line_buf'. */
 374
 375 static enum section
 376 check_section (void)
 377 {
 378   size_t len = line_buf.length - 1;
 379
 380   if (len < 2 || memcmp (line_buf.buffer, section_del, 2))
 381     return Text;
 382   if (len == header_del_len
 383       && !memcmp (line_buf.buffer, header_del, header_del_len))
 384     return Header;
 385   if (len == body_del_len
 386       && !memcmp (line_buf.buffer, body_del, body_del_len))
 387     return Body;
 388   if (len == footer_del_len
 389       && !memcmp (line_buf.buffer, footer_del, footer_del_len))
 390     return Footer;
 391   return Text;
 392 }
 393
 394 /* Read and process the file pointed to by FP. */
 395
 396 static void
 397 process_file (FILE *fp)
 398 {
 399   while (readlinebuffer (&line_buf, fp))
 400     {
 401       switch (check_section ())
 402         {
 403         case Header:
 404           proc_header ();
 405           break;
 406         case Body:
 407           proc_body ();
 408           break;
 409         case Footer:
 410           proc_footer ();
 411           break;
 412         case Text:
 413           proc_text ();
 414           break;
 415         }
 416     }
 417 }
 418
 419 /* Process file FILE to standard output.
 420    Return true if successful.  */
 421
 422 static bool
 423 nl_file (char const *file)
 424 {
 425   FILE *stream;
 426
 427   if (STREQ (file, "-"))
 428     {
 429       have_read_stdin = true;
 430       stream = stdin;
 431     }
 432   else
 433     {
 434       stream = fopen (file, "r");
 435       if (stream == NULL)
 436         {
 437           error (0, errno, "%s", file);
 438           return false;
 439         }
 440     }
 441
 442   process_file (stream);
 443
 444   if (ferror (stream))
 445     {
 446       error (0, errno, "%s", file);
 447       return false;
 448     }
 449   if (STREQ (file, "-"))
 450     clearerr (stream);          /* Also clear EOF. */
 451   else if (fclose (stream) == EOF)
 452     {
 453       error (0, errno, "%s", file);
 454       return false;
 455     }
 456   return true;
 457 }
 458
 459 int
 460 main (int argc, char **argv)
 461 {
 462   int c;
 463   size_t len;
 464   bool ok = true;
 465
 466   initialize_main (&argc, &argv);
 467   set_program_name (argv[0]);
 468   setlocale (LC_ALL, "");
 469   bindtextdomain (PACKAGE, LOCALEDIR);
 470   textdomain (PACKAGE);
 471
 472   atexit (close_stdout);
 473
 474   have_read_stdin = false;
 475
 476   while ((c = getopt_long (argc, argv, "h:b:f:v:i:pl:s:w:n:d:", longopts,
 477                            NULL)) != -1)
 478     {
 479       switch (c)
 480         {
 481         case 'h':
 482           if (! build_type_arg (&header_type, &header_regex, header_fastmap))
 483             {
 484               error (0, 0, _("invalid header numbering style: %s"),
 485                      quote (optarg));
 486               ok = false;
 487             }
 488           break;
 489         case 'b':
 490           if (! build_type_arg (&body_type, &body_regex, body_fastmap))
 491             {
 492               error (0, 0, _("invalid body numbering style: %s"),
 493                      quote (optarg));
 494               ok = false;
 495             }
 496           break;
 497         case 'f':
 498           if (! build_type_arg (&footer_type, &footer_regex, footer_fastmap))
 499             {
 500               error (0, 0, _("invalid footer numbering style: %s"),
 501                      quote (optarg));
 502               ok = false;
 503             }
 504           break;
 505         case 'v':
 506           if (xstrtoimax (optarg, NULL, 10, &starting_line_number, "")
 507               != LONGINT_OK)
 508             {
 509               error (0, 0, _("invalid starting line number: %s"),
 510                      quote (optarg));
 511               ok = false;
 512             }
 513           break;
 514   case PAGE_INCREMENT_OPTION_DEPRECATED:
 515     error (0, 0, _("WARNING: --page-increment is deprecated; "
 516                    "use --line-increment instead"));
 517     /* fall through */
 518         case 'i':
 519           if (! (xstrtoimax (optarg, NULL, 10, &page_incr, "") == LONGINT_OK
 520                  && 0 < page_incr))
 521             {
 522               error (0, 0, _("invalid line number increment: %s"),
 523                      quote (optarg));
 524               ok = false;
 525             }
 526           break;
 527         case 'p':
 528           reset_numbers = false;
 529           break;
 530         case 'l':
 531           if (! (xstrtoimax (optarg, NULL, 10, &blank_join, "") == LONGINT_OK
 532                  && 0 < blank_join))
 533             {
 534               error (0, 0, _("invalid number of blank lines: %s"),
 535                      quote (optarg));
 536               ok = false;
 537             }
 538           break;
 539         case 's':
 540           separator_str = optarg;
 541           break;
 542         case 'w':
 543           {
 544             long int tmp_long;
 545             if (xstrtol (optarg, NULL, 10, &tmp_long, "") != LONGINT_OK
 546                 || tmp_long <= 0 || tmp_long > INT_MAX)
 547               {
 548                 error (0, 0, _("invalid line number field width: %s"),
 549                        quote (optarg));
 550                 ok = false;
 551               }
 552             else
 553               {
 554                 lineno_width = tmp_long;
 555               }
 556           }
 557           break;
 558         case 'n':
 559           if (STREQ (optarg, "ln"))
 560             lineno_format = FORMAT_LEFT;
 561           else if (STREQ (optarg, "rn"))
 562             lineno_format = FORMAT_RIGHT_NOLZ;
 563           else if (STREQ (optarg, "rz"))
 564             lineno_format = FORMAT_RIGHT_LZ;
 565           else
 566             {
 567               error (0, 0, _("invalid line numbering format: %s"),
 568                      quote (optarg));
 569               ok = false;
 570             }
 571           break;
 572         case 'd':
 573           section_del = optarg;
 574           break;
 575         case_GETOPT_HELP_CHAR;
 576         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 577         default:
 578           ok = false;
 579           break;
 580         }
 581     }
 582
 583   if (!ok)
 584     usage (EXIT_FAILURE);
 585
 586   /* Initialize the section delimiters.  */
 587   len = strlen (section_del);
 588
 589   header_del_len = len * 3;
 590   header_del = xmalloc (header_del_len + 1);
 591   stpcpy (stpcpy (stpcpy (header_del, section_del), section_del), section_del);
 592
 593   body_del_len = len * 2;
 594   body_del = xmalloc (body_del_len + 1);
 595   stpcpy (stpcpy (body_del, section_del), section_del);
 596
 597   footer_del_len = len;
 598   footer_del = xmalloc (footer_del_len + 1);
 599   stpcpy (footer_del, section_del);
 600
 601   /* Initialize the input buffer.  */
 602   initbuffer (&line_buf);
 603
 604   /* Initialize the printf format for unnumbered lines. */
 605   len = strlen (separator_str);
 606   print_no_line_fmt = xmalloc (lineno_width + len + 1);
 607   memset (print_no_line_fmt, ' ', lineno_width + len);
 608   print_no_line_fmt[lineno_width + len] = '\0';
 609
 610   line_no = starting_line_number;
 611   current_type = body_type;
 612   current_regex = &body_regex;
 613
 614   /* Main processing. */
 615
 616   if (optind == argc)
 617     ok = nl_file ("-");
 618   else
 619     for (; optind < argc; optind++)
 620       ok &= nl_file (argv[optind]);
 621
 622   if (have_read_stdin && fclose (stdin) == EOF)
 623     error (EXIT_FAILURE, errno, "-");
 624
 625   exit (ok ? EXIT_SUCCESS : EXIT_FAILURE);
 626 }