src/nl.c

   1 /* nl -- number lines of files
   2    Copyright (C) 1989-2015 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 /* 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 "xdectoint.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 static struct option const longopts[] =
 149 {
 150   {"header-numbering", required_argument, NULL, 'h'},
 151   {"body-numbering", required_argument, NULL, 'b'},
 152   {"footer-numbering", required_argument, NULL, 'f'},
 153   {"starting-line-number", required_argument, NULL, 'v'},
 154   {"line-increment", required_argument, NULL, 'i'},
 155   {"no-renumber", no_argument, NULL, 'p'},
 156   {"join-blank-lines", required_argument, NULL, 'l'},
 157   {"number-separator", required_argument, NULL, 's'},
 158   {"number-width", required_argument, NULL, 'w'},
 159   {"number-format", required_argument, NULL, 'n'},
 160   {"section-delimiter", required_argument, NULL, 'd'},
 161   {GETOPT_HELP_OPTION_DECL},
 162   {GETOPT_VERSION_OPTION_DECL},
 163   {NULL, 0, NULL, 0}
 164 };
 165
 166 /* Print a usage message and quit. */
 167
 168 void
 169 usage (int status)
 170 {
 171   if (status != EXIT_SUCCESS)
 172     emit_try_help ();
 173   else
 174     {
 175       printf (_("\
 176 Usage: %s [OPTION]... [FILE]...\n\
 177 "),
 178               program_name);
 179       fputs (_("\
 180 Write each FILE to standard output, with line numbers added.\n\
 181 "), stdout);
 182
 183       emit_stdin_note ();
 184       emit_mandatory_arg_note ();
 185
 186       fputs (_("\
 187   -b, --body-numbering=STYLE      use STYLE for numbering body lines\n\
 188   -d, --section-delimiter=CC      use CC for separating logical pages\n\
 189   -f, --footer-numbering=STYLE    use STYLE for numbering footer lines\n\
 190 "), stdout);
 191       fputs (_("\
 192   -h, --header-numbering=STYLE    use STYLE for numbering header lines\n\
 193   -i, --line-increment=NUMBER     line number increment at each line\n\
 194   -l, --join-blank-lines=NUMBER   group of NUMBER empty lines counted as one\n\
 195   -n, --number-format=FORMAT      insert line numbers according to FORMAT\n\
 196   -p, --no-renumber               do not reset line numbers at logical pages\n\
 197   -s, --number-separator=STRING   add STRING after (possible) line number\n\
 198 "), stdout);
 199       fputs (_("\
 200   -v, --starting-line-number=NUMBER  first line number on each logical page\n\
 201   -w, --number-width=NUMBER       use NUMBER columns for line numbers\n\
 202 "), stdout);
 203       fputs (HELP_OPTION_DESCRIPTION, stdout);
 204       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 205       fputs (_("\
 206 \n\
 207 By default, selects -v1 -i1 -l1 -sTAB -w6 -nrn -hn -bt -fn.  CC are\n\
 208 two delimiter characters for separating logical pages, a missing\n\
 209 second character implies :.  Type \\\\ for \\.  STYLE is one of:\n\
 210 "), stdout);
 211       fputs (_("\
 212 \n\
 213   a         number all lines\n\
 214   t         number only nonempty lines\n\
 215   n         number no lines\n\
 216   pBRE      number only lines that contain a match for the basic regular\n\
 217             expression, BRE\n\
 218 \n\
 219 FORMAT is one of:\n\
 220 \n\
 221   ln   left justified, no leading zeros\n\
 222   rn   right justified, no leading zeros\n\
 223   rz   right justified, leading zeros\n\
 224 \n\
 225 "), stdout);
 226       emit_ancillary_info (PROGRAM_NAME);
 227     }
 228   exit (status);
 229 }
 230
 231 /* Set the command line flag TYPEP and possibly the regex pointer REGEXP,
 232    according to 'optarg'.  */
 233
 234 static bool
 235 build_type_arg (char const **typep,
 236                 struct re_pattern_buffer *regexp, char *fastmap)
 237 {
 238   char const *errmsg;
 239   bool rval = true;
 240
 241   switch (*optarg)
 242     {
 243     case 'a':
 244     case 't':
 245     case 'n':
 246       *typep = optarg;
 247       break;
 248     case 'p':
 249       *typep = optarg++;
 250       regexp->buffer = NULL;
 251       regexp->allocated = 0;
 252       regexp->fastmap = fastmap;
 253       regexp->translate = NULL;
 254       re_syntax_options =
 255         RE_SYNTAX_POSIX_BASIC & ~RE_CONTEXT_INVALID_DUP & ~RE_NO_EMPTY_RANGES;
 256       errmsg = re_compile_pattern (optarg, strlen (optarg), regexp);
 257       if (errmsg)
 258         error (EXIT_FAILURE, 0, "%s", (errmsg));
 259       break;
 260     default:
 261       rval = false;
 262       break;
 263     }
 264   return rval;
 265 }
 266
 267 /* Print the line number and separator; increment the line number. */
 268
 269 static void
 270 print_lineno (void)
 271 {
 272   intmax_t next_line_no;
 273
 274   printf (lineno_format, lineno_width, line_no, separator_str);
 275
 276   next_line_no = line_no + page_incr;
 277   if (next_line_no < line_no)
 278     error (EXIT_FAILURE, 0, _("line number overflow"));
 279   line_no = next_line_no;
 280 }
 281
 282 /* Switch to a header section. */
 283
 284 static void
 285 proc_header (void)
 286 {
 287   current_type = header_type;
 288   current_regex = &header_regex;
 289   if (reset_numbers)
 290     line_no = starting_line_number;
 291   putchar ('\n');
 292 }
 293
 294 /* Switch to a body section. */
 295
 296 static void
 297 proc_body (void)
 298 {
 299   current_type = body_type;
 300   current_regex = &body_regex;
 301   putchar ('\n');
 302 }
 303
 304 /* Switch to a footer section. */
 305
 306 static void
 307 proc_footer (void)
 308 {
 309   current_type = footer_type;
 310   current_regex = &footer_regex;
 311   putchar ('\n');
 312 }
 313
 314 /* Process a regular text line in 'line_buf'. */
 315
 316 static void
 317 proc_text (void)
 318 {
 319   static intmax_t blank_lines = 0;      /* Consecutive blank lines so far. */
 320
 321   switch (*current_type)
 322     {
 323     case 'a':
 324       if (blank_join > 1)
 325         {
 326           if (1 < line_buf.length || ++blank_lines == blank_join)
 327             {
 328               print_lineno ();
 329               blank_lines = 0;
 330             }
 331           else
 332             fputs (print_no_line_fmt, stdout);
 333         }
 334       else
 335         print_lineno ();
 336       break;
 337     case 't':
 338       if (1 < line_buf.length)
 339         print_lineno ();
 340       else
 341         fputs (print_no_line_fmt, stdout);
 342       break;
 343     case 'n':
 344       fputs (print_no_line_fmt, stdout);
 345       break;
 346     case 'p':
 347       switch (re_search (current_regex, line_buf.buffer, line_buf.length - 1,
 348                          0, line_buf.length - 1, NULL))
 349         {
 350         case -2:
 351           error (EXIT_FAILURE, errno, _("error in regular expression search"));
 352
 353         case -1:
 354           fputs (print_no_line_fmt, stdout);
 355           break;
 356
 357         default:
 358           print_lineno ();
 359           break;
 360         }
 361     }
 362   fwrite (line_buf.buffer, sizeof (char), line_buf.length, stdout);
 363 }
 364
 365 /* Return the type of line in 'line_buf'. */
 366
 367 static enum section
 368 check_section (void)
 369 {
 370   size_t len = line_buf.length - 1;
 371
 372   if (len < 2 || memcmp (line_buf.buffer, section_del, 2))
 373     return Text;
 374   if (len == header_del_len
 375       && !memcmp (line_buf.buffer, header_del, header_del_len))
 376     return Header;
 377   if (len == body_del_len
 378       && !memcmp (line_buf.buffer, body_del, body_del_len))
 379     return Body;
 380   if (len == footer_del_len
 381       && !memcmp (line_buf.buffer, footer_del, footer_del_len))
 382     return Footer;
 383   return Text;
 384 }
 385
 386 /* Read and process the file pointed to by FP. */
 387
 388 static void
 389 process_file (FILE *fp)
 390 {
 391   while (readlinebuffer (&line_buf, fp))
 392     {
 393       switch (check_section ())
 394         {
 395         case Header:
 396           proc_header ();
 397           break;
 398         case Body:
 399           proc_body ();
 400           break;
 401         case Footer:
 402           proc_footer ();
 403           break;
 404         case Text:
 405           proc_text ();
 406           break;
 407         }
 408     }
 409 }
 410
 411 /* Process file FILE to standard output.
 412    Return true if successful.  */
 413
 414 static bool
 415 nl_file (char const *file)
 416 {
 417   FILE *stream;
 418
 419   if (STREQ (file, "-"))
 420     {
 421       have_read_stdin = true;
 422       stream = stdin;
 423     }
 424   else
 425     {
 426       stream = fopen (file, "r");
 427       if (stream == NULL)
 428         {
 429           error (0, errno, "%s", quotef (file));
 430           return false;
 431         }
 432     }
 433
 434   fadvise (stream, FADVISE_SEQUENTIAL);
 435
 436   process_file (stream);
 437
 438   if (ferror (stream))
 439     {
 440       error (0, errno, "%s", quotef (file));
 441       return false;
 442     }
 443   if (STREQ (file, "-"))
 444     clearerr (stream);          /* Also clear EOF. */
 445   else if (fclose (stream) == EOF)
 446     {
 447       error (0, errno, "%s", quotef (file));
 448       return false;
 449     }
 450   return true;
 451 }
 452
 453 int
 454 main (int argc, char **argv)
 455 {
 456   int c;
 457   size_t len;
 458   bool ok = true;
 459
 460   initialize_main (&argc, &argv);
 461   set_program_name (argv[0]);
 462   setlocale (LC_ALL, "");
 463   bindtextdomain (PACKAGE, LOCALEDIR);
 464   textdomain (PACKAGE);
 465
 466   atexit (close_stdout);
 467
 468   have_read_stdin = false;
 469
 470   while ((c = getopt_long (argc, argv, "h:b:f:v:i:pl:s:w:n:d:", longopts,
 471                            NULL)) != -1)
 472     {
 473       switch (c)
 474         {
 475         case 'h':
 476           if (! build_type_arg (&header_type, &header_regex, header_fastmap))
 477             {
 478               error (0, 0, _("invalid header numbering style: %s"),
 479                      quote (optarg));
 480               ok = false;
 481             }
 482           break;
 483         case 'b':
 484           if (! build_type_arg (&body_type, &body_regex, body_fastmap))
 485             {
 486               error (0, 0, _("invalid body numbering style: %s"),
 487                      quote (optarg));
 488               ok = false;
 489             }
 490           break;
 491         case 'f':
 492           if (! build_type_arg (&footer_type, &footer_regex, footer_fastmap))
 493             {
 494               error (0, 0, _("invalid footer numbering style: %s"),
 495                      quote (optarg));
 496               ok = false;
 497             }
 498           break;
 499         case 'v':
 500           starting_line_number = xdectoimax (optarg, INTMAX_MIN, INTMAX_MAX, "",
 501                                              _("invalid starting line number"),
 502                                              0);
 503           break;
 504         case 'i':
 505           page_incr = xdectoimax (optarg, 1, INTMAX_MAX, "",
 506                                   _("invalid line number increment"), 0);
 507           break;
 508         case 'p':
 509           reset_numbers = false;
 510           break;
 511         case 'l':
 512           blank_join = xdectoimax (optarg, 1, INTMAX_MAX, "",
 513                                    _("invalid line number of blank lines"), 0);
 514           break;
 515         case 's':
 516           separator_str = optarg;
 517           break;
 518         case 'w':
 519           lineno_width = xdectoimax (optarg, 1, INT_MAX, "",
 520                                      _("invalid line number field width"), 0);
 521           break;
 522         case 'n':
 523           if (STREQ (optarg, "ln"))
 524             lineno_format = FORMAT_LEFT;
 525           else if (STREQ (optarg, "rn"))
 526             lineno_format = FORMAT_RIGHT_NOLZ;
 527           else if (STREQ (optarg, "rz"))
 528             lineno_format = FORMAT_RIGHT_LZ;
 529           else
 530             {
 531               error (0, 0, _("invalid line numbering format: %s"),
 532                      quote (optarg));
 533               ok = false;
 534             }
 535           break;
 536         case 'd':
 537           section_del = optarg;
 538           break;
 539         case_GETOPT_HELP_CHAR;
 540         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 541         default:
 542           ok = false;
 543           break;
 544         }
 545     }
 546
 547   if (!ok)
 548     usage (EXIT_FAILURE);
 549
 550   /* Initialize the section delimiters.  */
 551   len = strlen (section_del);
 552
 553   header_del_len = len * 3;
 554   header_del = xmalloc (header_del_len + 1);
 555   stpcpy (stpcpy (stpcpy (header_del, section_del), section_del), section_del);
 556
 557   body_del_len = len * 2;
 558   body_del = xmalloc (body_del_len + 1);
 559   stpcpy (stpcpy (body_del, section_del), section_del);
 560
 561   footer_del_len = len;
 562   footer_del = xmalloc (footer_del_len + 1);
 563   stpcpy (footer_del, section_del);
 564
 565   /* Initialize the input buffer.  */
 566   initbuffer (&line_buf);
 567
 568   /* Initialize the printf format for unnumbered lines. */
 569   len = strlen (separator_str);
 570   print_no_line_fmt = xmalloc (lineno_width + len + 1);
 571   memset (print_no_line_fmt, ' ', lineno_width + len);
 572   print_no_line_fmt[lineno_width + len] = '\0';
 573
 574   line_no = starting_line_number;
 575   current_type = body_type;
 576   current_regex = &body_regex;
 577
 578   /* Main processing. */
 579
 580   if (optind == argc)
 581     ok = nl_file ("-");
 582   else
 583     for (; optind < argc; optind++)
 584       ok &= nl_file (argv[optind]);
 585
 586   if (have_read_stdin && fclose (stdin) == EOF)
 587     error (EXIT_FAILURE, errno, "-");
 588
 589   return ok ? EXIT_SUCCESS : EXIT_FAILURE;
 590 }