src/nl.c

   1 /* nl -- number lines of files
   2    Copyright (C) 1989-2012 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     emit_try_help ();
 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   fadvise (stream, FADVISE_SEQUENTIAL);
 443
 444   process_file (stream);
 445
 446   if (ferror (stream))
 447     {
 448       error (0, errno, "%s", file);
 449       return false;
 450     }
 451   if (STREQ (file, "-"))
 452     clearerr (stream);          /* Also clear EOF. */
 453   else if (fclose (stream) == EOF)
 454     {
 455       error (0, errno, "%s", file);
 456       return false;
 457     }
 458   return true;
 459 }
 460
 461 int
 462 main (int argc, char **argv)
 463 {
 464   int c;
 465   size_t len;
 466   bool ok = true;
 467
 468   initialize_main (&argc, &argv);
 469   set_program_name (argv[0]);
 470   setlocale (LC_ALL, "");
 471   bindtextdomain (PACKAGE, LOCALEDIR);
 472   textdomain (PACKAGE);
 473
 474   atexit (close_stdout);
 475
 476   have_read_stdin = false;
 477
 478   while ((c = getopt_long (argc, argv, "h:b:f:v:i:pl:s:w:n:d:", longopts,
 479                            NULL)) != -1)
 480     {
 481       switch (c)
 482         {
 483         case 'h':
 484           if (! build_type_arg (&header_type, &header_regex, header_fastmap))
 485             {
 486               error (0, 0, _("invalid header numbering style: %s"),
 487                      quote (optarg));
 488               ok = false;
 489             }
 490           break;
 491         case 'b':
 492           if (! build_type_arg (&body_type, &body_regex, body_fastmap))
 493             {
 494               error (0, 0, _("invalid body numbering style: %s"),
 495                      quote (optarg));
 496               ok = false;
 497             }
 498           break;
 499         case 'f':
 500           if (! build_type_arg (&footer_type, &footer_regex, footer_fastmap))
 501             {
 502               error (0, 0, _("invalid footer numbering style: %s"),
 503                      quote (optarg));
 504               ok = false;
 505             }
 506           break;
 507         case 'v':
 508           if (xstrtoimax (optarg, NULL, 10, &starting_line_number, "")
 509               != LONGINT_OK)
 510             {
 511               error (0, 0, _("invalid starting line number: %s"),
 512                      quote (optarg));
 513               ok = false;
 514             }
 515           break;
 516   case PAGE_INCREMENT_OPTION_DEPRECATED:
 517     error (0, 0, _("WARNING: --page-increment is deprecated; "
 518                    "use --line-increment instead"));
 519     /* fall through */
 520         case 'i':
 521           if (! (xstrtoimax (optarg, NULL, 10, &page_incr, "") == LONGINT_OK
 522                  && 0 < page_incr))
 523             {
 524               error (0, 0, _("invalid line number increment: %s"),
 525                      quote (optarg));
 526               ok = false;
 527             }
 528           break;
 529         case 'p':
 530           reset_numbers = false;
 531           break;
 532         case 'l':
 533           if (! (xstrtoimax (optarg, NULL, 10, &blank_join, "") == LONGINT_OK
 534                  && 0 < blank_join))
 535             {
 536               error (0, 0, _("invalid number of blank lines: %s"),
 537                      quote (optarg));
 538               ok = false;
 539             }
 540           break;
 541         case 's':
 542           separator_str = optarg;
 543           break;
 544         case 'w':
 545           {
 546             long int tmp_long;
 547             if (xstrtol (optarg, NULL, 10, &tmp_long, "") != LONGINT_OK
 548                 || tmp_long <= 0 || tmp_long > INT_MAX)
 549               {
 550                 error (0, 0, _("invalid line number field width: %s"),
 551                        quote (optarg));
 552                 ok = false;
 553               }
 554             else
 555               {
 556                 lineno_width = tmp_long;
 557               }
 558           }
 559           break;
 560         case 'n':
 561           if (STREQ (optarg, "ln"))
 562             lineno_format = FORMAT_LEFT;
 563           else if (STREQ (optarg, "rn"))
 564             lineno_format = FORMAT_RIGHT_NOLZ;
 565           else if (STREQ (optarg, "rz"))
 566             lineno_format = FORMAT_RIGHT_LZ;
 567           else
 568             {
 569               error (0, 0, _("invalid line numbering format: %s"),
 570                      quote (optarg));
 571               ok = false;
 572             }
 573           break;
 574         case 'd':
 575           section_del = optarg;
 576           break;
 577         case_GETOPT_HELP_CHAR;
 578         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 579         default:
 580           ok = false;
 581           break;
 582         }
 583     }
 584
 585   if (!ok)
 586     usage (EXIT_FAILURE);
 587
 588   /* Initialize the section delimiters.  */
 589   len = strlen (section_del);
 590
 591   header_del_len = len * 3;
 592   header_del = xmalloc (header_del_len + 1);
 593   stpcpy (stpcpy (stpcpy (header_del, section_del), section_del), section_del);
 594
 595   body_del_len = len * 2;
 596   body_del = xmalloc (body_del_len + 1);
 597   stpcpy (stpcpy (body_del, section_del), section_del);
 598
 599   footer_del_len = len;
 600   footer_del = xmalloc (footer_del_len + 1);
 601   stpcpy (footer_del, section_del);
 602
 603   /* Initialize the input buffer.  */
 604   initbuffer (&line_buf);
 605
 606   /* Initialize the printf format for unnumbered lines. */
 607   len = strlen (separator_str);
 608   print_no_line_fmt = xmalloc (lineno_width + len + 1);
 609   memset (print_no_line_fmt, ' ', lineno_width + len);
 610   print_no_line_fmt[lineno_width + len] = '\0';
 611
 612   line_no = starting_line_number;
 613   current_type = body_type;
 614   current_regex = &body_regex;
 615
 616   /* Main processing. */
 617
 618   if (optind == argc)
 619     ok = nl_file ("-");
 620   else
 621     for (; optind < argc; optind++)
 622       ok &= nl_file (argv[optind]);
 623
 624   if (have_read_stdin && fclose (stdin) == EOF)
 625     error (EXIT_FAILURE, errno, "-");
 626
 627   exit (ok ? EXIT_SUCCESS : EXIT_FAILURE);
 628 }