src/nl.c

   1 /* nl -- number lines of files
   2    Copyright (C) 1989-2024 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 <https://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 "fadvise.h"
  31 #include "linebuffer.h"
  32 #include "quote.h"
  33 #include "xdectoint.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[] = "%*jd%s";
  47
  48 /* Right justified, leading zeroes.  */
  49 static char const FORMAT_RIGHT_LZ[] = "%0*jd%s";
  50
  51 /* Left justified, no leading zeroes.  */
  52 static char const FORMAT_LEFT[] = "%-*jd%s";
  53
  54 /* Default section delimiter characters.  */
  55 static char 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 = nullptr;
  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 *section_del = DEFAULT_SECTION_DELIMITERS;
  98
  99 /* Header delimiter string.  */
 100 static char *header_del = nullptr;
 101
 102 /* Header section delimiter length.  */
 103 static size_t header_del_len;
 104
 105 /* Body delimiter string.  */
 106 static char *body_del = nullptr;
 107
 108 /* Body section delimiter length.  */
 109 static size_t body_del_len;
 110
 111 /* Footer delimiter string.  */
 112 static char *footer_del = nullptr;
 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 = nullptr;
 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 /* Whether the current line number has incremented past limits.  */
 145 static bool line_no_overflow;
 146
 147 /* True if we have ever read standard input.  */
 148 static bool have_read_stdin;
 149
 150 static struct option const longopts[] =
 151 {
 152   {"header-numbering", required_argument, nullptr, 'h'},
 153   {"body-numbering", required_argument, nullptr, 'b'},
 154   {"footer-numbering", required_argument, nullptr, 'f'},
 155   {"starting-line-number", required_argument, nullptr, 'v'},
 156   {"line-increment", required_argument, nullptr, 'i'},
 157   {"no-renumber", no_argument, nullptr, 'p'},
 158   {"join-blank-lines", required_argument, nullptr, 'l'},
 159   {"number-separator", required_argument, nullptr, 's'},
 160   {"number-width", required_argument, nullptr, 'w'},
 161   {"number-format", required_argument, nullptr, 'n'},
 162   {"section-delimiter", required_argument, nullptr, 'd'},
 163   {GETOPT_HELP_OPTION_DECL},
 164   {GETOPT_VERSION_OPTION_DECL},
 165   {nullptr, 0, nullptr, 0}
 166 };
 167
 168 /* Print a usage message and quit. */
 169
 170 void
 171 usage (int status)
 172 {
 173   if (status != EXIT_SUCCESS)
 174     emit_try_help ();
 175   else
 176     {
 177       printf (_("\
 178 Usage: %s [OPTION]... [FILE]...\n\
 179 "),
 180               program_name);
 181       fputs (_("\
 182 Write each FILE to standard output, with line numbers added.\n\
 183 "), stdout);
 184
 185       emit_stdin_note ();
 186       emit_mandatory_arg_note ();
 187
 188       fputs (_("\
 189   -b, --body-numbering=STYLE      use STYLE for numbering body lines\n\
 190   -d, --section-delimiter=CC      use CC for logical page delimiters\n\
 191   -f, --footer-numbering=STYLE    use STYLE for numbering footer lines\n\
 192 "), stdout);
 193       fputs (_("\
 194   -h, --header-numbering=STYLE    use STYLE for numbering header lines\n\
 195   -i, --line-increment=NUMBER     line number increment at each line\n\
 196   -l, --join-blank-lines=NUMBER   group of NUMBER empty lines counted as one\n\
 197   -n, --number-format=FORMAT      insert line numbers according to FORMAT\n\
 198   -p, --no-renumber               do not reset line numbers for each section\n\
 199   -s, --number-separator=STRING   add STRING after (possible) line number\n\
 200 "), stdout);
 201       fputs (_("\
 202   -v, --starting-line-number=NUMBER  first line number for each section\n\
 203   -w, --number-width=NUMBER       use NUMBER columns for line numbers\n\
 204 "), stdout);
 205       fputs (HELP_OPTION_DESCRIPTION, stdout);
 206       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 207       fputs (_("\
 208 \n\
 209 Default options are: -bt -d'\\:' -fn -hn -i1 -l1 -n'rn' -s<TAB> -v1 -w6\n\
 210 \n\
 211 CC are two delimiter characters used to construct logical page delimiters;\n\
 212 a missing second character implies ':'.  As a GNU extension one can specify\n\
 213 more than two characters, and also specifying the empty string (-d '')\n\
 214 disables section matching.\n\
 215 "), stdout);
 216       fputs (_("\
 217 \n\
 218 STYLE is one of:\n\
 219 \n\
 220   a      number all lines\n\
 221   t      number only nonempty lines\n\
 222   n      number no lines\n\
 223   pBRE   number only lines that contain a match for the basic regular\n\
 224          expression, BRE\n\
 225 "), stdout);
 226       fputs (_("\
 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 (PROGRAM_NAME);
 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 = nullptr;
 260       regexp->allocated = 0;
 261       regexp->fastmap = fastmap;
 262       regexp->translate = nullptr;
 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   if (line_no_overflow)
 282     error (EXIT_FAILURE, 0, _("line number overflow"));
 283
 284   printf (lineno_format, lineno_width, line_no, separator_str);
 285
 286   if (ckd_add (&line_no, line_no, page_incr))
 287     line_no_overflow = true;
 288 }
 289
 290 static void
 291 reset_lineno (void)
 292 {
 293   if (reset_numbers)
 294     {
 295       line_no = starting_line_number;
 296       line_no_overflow = false;
 297     }
 298 }
 299
 300 /* Switch to a header section. */
 301
 302 static void
 303 proc_header (void)
 304 {
 305   current_type = header_type;
 306   current_regex = &header_regex;
 307   reset_lineno ();
 308   putchar ('\n');
 309 }
 310
 311 /* Switch to a body section. */
 312
 313 static void
 314 proc_body (void)
 315 {
 316   current_type = body_type;
 317   current_regex = &body_regex;
 318   reset_lineno ();
 319   putchar ('\n');
 320 }
 321
 322 /* Switch to a footer section. */
 323
 324 static void
 325 proc_footer (void)
 326 {
 327   current_type = footer_type;
 328   current_regex = &footer_regex;
 329   reset_lineno ();
 330   putchar ('\n');
 331 }
 332
 333 /* Process a regular text line in 'line_buf'. */
 334
 335 static void
 336 proc_text (void)
 337 {
 338   static intmax_t blank_lines = 0;      /* Consecutive blank lines so far. */
 339
 340   switch (*current_type)
 341     {
 342     case 'a':
 343       if (blank_join > 1)
 344         {
 345           if (1 < line_buf.length || ++blank_lines == blank_join)
 346             {
 347               print_lineno ();
 348               blank_lines = 0;
 349             }
 350           else
 351             fputs (print_no_line_fmt, stdout);
 352         }
 353       else
 354         print_lineno ();
 355       break;
 356     case 't':
 357       if (1 < line_buf.length)
 358         print_lineno ();
 359       else
 360         fputs (print_no_line_fmt, stdout);
 361       break;
 362     case 'n':
 363       fputs (print_no_line_fmt, stdout);
 364       break;
 365     case 'p':
 366       switch (re_search (current_regex, line_buf.buffer, line_buf.length - 1,
 367                          0, line_buf.length - 1, nullptr))
 368         {
 369         case -2:
 370           error (EXIT_FAILURE, errno, _("error in regular expression search"));
 371
 372         case -1:
 373           fputs (print_no_line_fmt, stdout);
 374           break;
 375
 376         default:
 377           print_lineno ();
 378           break;
 379         }
 380     }
 381   fwrite (line_buf.buffer, sizeof (char), line_buf.length, stdout);
 382 }
 383
 384 /* Return the type of line in 'line_buf'. */
 385
 386 static enum section
 387 check_section (void)
 388 {
 389   size_t len = line_buf.length - 1;
 390
 391   if (len < 2 || footer_del_len < 2
 392       || memcmp (line_buf.buffer, section_del, 2))
 393     return Text;
 394   if (len == header_del_len
 395       && !memcmp (line_buf.buffer, header_del, header_del_len))
 396     return Header;
 397   if (len == body_del_len
 398       && !memcmp (line_buf.buffer, body_del, body_del_len))
 399     return Body;
 400   if (len == footer_del_len
 401       && !memcmp (line_buf.buffer, footer_del, footer_del_len))
 402     return Footer;
 403   return Text;
 404 }
 405
 406 /* Read and process the file pointed to by FP. */
 407
 408 static void
 409 process_file (FILE *fp)
 410 {
 411   while (readlinebuffer (&line_buf, fp))
 412     {
 413       switch (check_section ())
 414         {
 415         case Header:
 416           proc_header ();
 417           break;
 418         case Body:
 419           proc_body ();
 420           break;
 421         case Footer:
 422           proc_footer ();
 423           break;
 424         case Text:
 425           proc_text ();
 426           break;
 427         }
 428     }
 429 }
 430
 431 /* Process file FILE to standard output.
 432    Return true if successful.  */
 433
 434 static bool
 435 nl_file (char const *file)
 436 {
 437   FILE *stream;
 438
 439   if (STREQ (file, "-"))
 440     {
 441       have_read_stdin = true;
 442       stream = stdin;
 443       assume (stream);  /* Pacify GCC bug#109613.  */
 444     }
 445   else
 446     {
 447       stream = fopen (file, "r");
 448       if (stream == nullptr)
 449         {
 450           error (0, errno, "%s", quotef (file));
 451           return false;
 452         }
 453     }
 454
 455   fadvise (stream, FADVISE_SEQUENTIAL);
 456
 457   process_file (stream);
 458
 459   int err = errno;
 460   if (!ferror (stream))
 461     err = 0;
 462   if (STREQ (file, "-"))
 463     clearerr (stream);          /* Also clear EOF. */
 464   else if (fclose (stream) != 0 && !err)
 465     err = errno;
 466   if (err)
 467     {
 468       error (0, err, "%s", quotef (file));
 469       return false;
 470     }
 471   return true;
 472 }
 473
 474 int
 475 main (int argc, char **argv)
 476 {
 477   int c;
 478   size_t len;
 479   bool ok = true;
 480
 481   initialize_main (&argc, &argv);
 482   set_program_name (argv[0]);
 483   setlocale (LC_ALL, "");
 484   bindtextdomain (PACKAGE, LOCALEDIR);
 485   textdomain (PACKAGE);
 486
 487   atexit (close_stdout);
 488
 489   have_read_stdin = false;
 490
 491   while ((c = getopt_long (argc, argv, "h:b:f:v:i:pl:s:w:n:d:", longopts,
 492                            nullptr))
 493          != -1)
 494     {
 495       switch (c)
 496         {
 497         case 'h':
 498           if (! build_type_arg (&header_type, &header_regex, header_fastmap))
 499             {
 500               error (0, 0, _("invalid header numbering style: %s"),
 501                      quote (optarg));
 502               ok = false;
 503             }
 504           break;
 505         case 'b':
 506           if (! build_type_arg (&body_type, &body_regex, body_fastmap))
 507             {
 508               error (0, 0, _("invalid body numbering style: %s"),
 509                      quote (optarg));
 510               ok = false;
 511             }
 512           break;
 513         case 'f':
 514           if (! build_type_arg (&footer_type, &footer_regex, footer_fastmap))
 515             {
 516               error (0, 0, _("invalid footer numbering style: %s"),
 517                      quote (optarg));
 518               ok = false;
 519             }
 520           break;
 521         case 'v':
 522           starting_line_number = xdectoimax (optarg, INTMAX_MIN, INTMAX_MAX, "",
 523                                              _("invalid starting line number"),
 524                                              0);
 525           break;
 526         case 'i':
 527           page_incr = xdectoimax (optarg, INTMAX_MIN, INTMAX_MAX, "",
 528                                   _("invalid line number increment"), 0);
 529           break;
 530         case 'p':
 531           reset_numbers = false;
 532           break;
 533         case 'l':
 534           blank_join = xdectoimax (optarg, 1, INTMAX_MAX, "",
 535                                    _("invalid line number of blank lines"), 0);
 536           break;
 537         case 's':
 538           separator_str = optarg;
 539           break;
 540         case 'w':
 541           lineno_width = xdectoimax (optarg, 1, INT_MAX, "",
 542                                      _("invalid line number field width"), 0);
 543           break;
 544         case 'n':
 545           if (STREQ (optarg, "ln"))
 546             lineno_format = FORMAT_LEFT;
 547           else if (STREQ (optarg, "rn"))
 548             lineno_format = FORMAT_RIGHT_NOLZ;
 549           else if (STREQ (optarg, "rz"))
 550             lineno_format = FORMAT_RIGHT_LZ;
 551           else
 552             {
 553               error (0, 0, _("invalid line numbering format: %s"),
 554                      quote (optarg));
 555               ok = false;
 556             }
 557           break;
 558         case 'd':
 559           len = strlen (optarg);
 560           if (len == 1 || len == 2)  /* POSIX.  */
 561             {
 562               char *p = section_del;
 563               while (*optarg)
 564                 *p++ = *optarg++;
 565             }
 566           else
 567             section_del = optarg;  /* GNU extension.  */
 568           break;
 569         case_GETOPT_HELP_CHAR;
 570         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 571         default:
 572           ok = false;
 573           break;
 574         }
 575     }
 576
 577   if (!ok)
 578     usage (EXIT_FAILURE);
 579
 580   /* Initialize the section delimiters.  */
 581   len = strlen (section_del);
 582
 583   header_del_len = len * 3;
 584   header_del = xmalloc (header_del_len + 1);
 585   stpcpy (stpcpy (stpcpy (header_del, section_del), section_del), section_del);
 586
 587   body_del_len = len * 2;
 588   body_del = header_del + len;
 589
 590   footer_del_len = len;
 591   footer_del = body_del + len;
 592
 593   /* Initialize the input buffer.  */
 594   initbuffer (&line_buf);
 595
 596   /* Initialize the printf format for unnumbered lines. */
 597   len = strlen (separator_str);
 598   print_no_line_fmt = xmalloc (lineno_width + len + 1);
 599   memset (print_no_line_fmt, ' ', lineno_width + len);
 600   print_no_line_fmt[lineno_width + len] = '\0';
 601
 602   line_no = starting_line_number;
 603   current_type = body_type;
 604   current_regex = &body_regex;
 605
 606   /* Main processing. */
 607
 608   if (optind == argc)
 609     ok = nl_file ("-");
 610   else
 611     for (; optind < argc; optind++)
 612       ok &= nl_file (argv[optind]);
 613
 614   if (have_read_stdin && fclose (stdin) == EOF)
 615     error (EXIT_FAILURE, errno, "-");
 616
 617   return ok ? EXIT_SUCCESS : EXIT_FAILURE;
 618 }