doc/help2man

   1 #!/usr/bin/perl -w
   2
   3 # Generate a short man page from --help and --version output.
   4 # Copyright (C) 1997-2012 Free Software Foundation, Inc.
   5
   6 # This program is free software; you can redistribute it and/or modify
   7 # it under the terms of the GNU General Public License as published by
   8 # the Free Software Foundation; either version 3, or (at your option)
   9 # any later version.
  10
  11 # This program is distributed in the hope that it will be useful,
  12 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 # GNU General Public License for more details.
  15
  16 # You should have received a copy of the GNU General Public License
  17 # along with this program; if not, see <http://www.gnu.org/licenses/>.
  18
  19 # Written by Brendan O'Dea <bod@debian.org>
  20 # Available from ftp://ftp.gnu.org/gnu/help2man/
  21
  22 use 5.006;
  23 use strict;
  24 use Getopt::Long;
  25 use Text::Tabs qw(expand);
  26 use POSIX qw(strftime setlocale LC_ALL);
  27
  28 my $this_program = 'help2man';
  29 my $this_version = '1.40.8';
  30
  31 sub _ { $_[0] }
  32 sub configure_locale
  33 {
  34     my $locale = shift;
  35     die "$this_program: no locale support (Locale::gettext required)\n"
  36         unless $locale eq 'C';
  37 }
  38
  39 sub dec { $_[0] }
  40 sub enc { $_[0] }
  41 sub enc_user { $_[0] }
  42 sub kark { die +(sprintf shift, @_), "\n" }
  43 sub N_ { $_[0] }
  44
  45 my $version_info = enc_user sprintf _(<<'EOT'), $this_program, $this_version;
  46 GNU %s %s
  47
  48 Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2009, 2010,
  49 2011, 2012 Free Software Foundation, Inc.
  50 This is free software; see the source for copying conditions.  There is NO
  51 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
  52
  53 Written by Brendan O'Dea <bod@debian.org>
  54 EOT
  55
  56 my $help_info = enc_user sprintf _(<<'EOT'), $this_program, $this_program;
  57 `%s' generates a man page out of `--help' and `--version' output.
  58
  59 Usage: %s [OPTION]... EXECUTABLE
  60
  61  -n, --name=STRING       description for the NAME paragraph
  62  -s, --section=SECTION   section number for manual page (1, 6, 8)
  63  -m, --manual=TEXT       name of manual (User Commands, ...)
  64  -S, --source=TEXT       source of program (FSF, Debian, ...)
  65  -L, --locale=STRING     select locale (default "C")
  66  -i, --include=FILE      include material from `FILE'
  67  -I, --opt-include=FILE  include material from `FILE' if it exists
  68  -o, --output=FILE       send output to `FILE'
  69  -p, --info-page=TEXT    name of Texinfo manual
  70  -N, --no-info           suppress pointer to Texinfo manual
  71  -l, --libtool           exclude the `lt-' from the program name
  72      --help              print this help, then exit
  73      --version           print version number, then exit
  74
  75 EXECUTABLE should accept `--help' and `--version' options and produce output on
  76 stdout although alternatives may be specified using:
  77
  78  -h, --help-option=STRING     help option string
  79  -v, --version-option=STRING  version option string
  80  --version-string=STRING      version string
  81  --no-discard-stderr          include stderr when parsing option output
  82
  83 Report bugs to <bug-help2man@gnu.org>.
  84 EOT
  85
  86 my $section = 1;
  87 my $manual = '';
  88 my $source = '';
  89 my $help_option = '--help';
  90 my $version_option = '--version';
  91 my $discard_stderr = 1;
  92 my ($opt_name, @opt_include, $opt_output, $opt_info, $opt_no_info, $opt_libtool,
  93     $version_text);
  94
  95 my %opt_def = (
  96     'n|name=s'           => \$opt_name,
  97     's|section=s'        => \$section,
  98     'm|manual=s'         => \$manual,
  99     'S|source=s'         => \$source,
 100     'L|locale=s'         => sub { configure_locale pop },
 101     'i|include=s'        => sub { push @opt_include, [ pop, 1 ] },
 102     'I|opt-include=s'    => sub { push @opt_include, [ pop, 0 ] },
 103     'o|output=s'         => \$opt_output,
 104     'p|info-page=s'      => \$opt_info,
 105     'N|no-info'          => \$opt_no_info,
 106     'l|libtool'          => \$opt_libtool,
 107     'help'               => sub { print $help_info; exit },
 108     'version'            => sub { print $version_info; exit },
 109     'h|help-option=s'    => \$help_option,
 110     'v|version-option=s' => \$version_option,
 111     'version-string=s'   => \$version_text,
 112     'discard-stderr!'    => \$discard_stderr,
 113 );
 114
 115 # Parse options.
 116 Getopt::Long::config('bundling');
 117 die $help_info unless GetOptions %opt_def and @ARGV == 1;
 118
 119 my %include = ();
 120 my %append = ();
 121 my @include = (); # retain order given in include file
 122
 123 # Process include file (if given).  Format is:
 124 #
 125 #   [section name]
 126 #   verbatim text
 127 #
 128 # or
 129 #
 130 #   /pattern/
 131 #   verbatim text
 132 #
 133
 134 while (@opt_include)
 135 {
 136     my ($inc, $required) = @{shift @opt_include};
 137
 138     next unless -f $inc or $required;
 139     kark N_("%s: can't open `%s' (%s)"), $this_program, $inc, $!
 140         unless open INC, $inc;
 141
 142     my $key;
 143     my $hash = \%include;
 144
 145     while (<INC>)
 146     {
 147         # Convert input to internal Perl format, so that multibyte
 148         # sequences are treated as single characters.
 149         $_ = dec $_;
 150
 151         # [section]
 152         if (/^\[([^]]+)\]\s*$/)
 153         {
 154             $key = uc $1;
 155             $key =~ s/^\s+//;
 156             $key =~ s/\s+$//;
 157             $hash = \%include;
 158             push @include, $key unless $include{$key};
 159             next;
 160         }
 161
 162         # /pattern/
 163         if (m!^/(.*)/([ims]*)\s*$!)
 164         {
 165             my $pat = $2 ? "(?$2)$1" : $1;
 166
 167             # Check pattern.
 168             eval { $key = qr($pat) };
 169             if ($@)
 170             {
 171                 $@ =~ s/ at .*? line \d.*//;
 172                 die "$inc:$.:$@";
 173             }
 174
 175             $hash = \%append;
 176             next;
 177         }
 178
 179         # Check for options before the first section--anything else is
 180         # silently ignored, allowing the first for comments and
 181         # revision info.
 182         unless ($key)
 183         {
 184             # handle options
 185             if (/^-/)
 186             {
 187                 local @ARGV = split;
 188                 GetOptions %opt_def;
 189             }
 190
 191             next;
 192         }
 193
 194         $hash->{$key} ||= '';
 195         $hash->{$key} .= $_;
 196     }
 197
 198     close INC;
 199
 200     kark N_("%s: no valid information found in `%s'"), $this_program, $inc
 201         unless $key;
 202 }
 203
 204 # Compress trailing blank lines.
 205 for my $hash (\(%include, %append))
 206 {
 207     for (keys %$hash) { $hash->{$_} =~ s/\n+$/\n/ }
 208 }
 209
 210 sub get_option_value;
 211
 212 # Grab help and version info from executable.
 213 my $help_text   = get_option_value $ARGV[0], $help_option;
 214 $version_text ||= get_option_value $ARGV[0], $version_option;
 215
 216 # Translators: the following message is a strftime(3) format string, which in
 217 # the English version expands to the month as a word and the full year.  It
 218 # is used on the footer of the generated manual pages.  If in doubt, you may
 219 # just use %x as the value (which should be the full locale-specific date).
 220 my $date = enc strftime _("%B %Y"), localtime;
 221 (my $program = $ARGV[0]) =~ s!.*/!!;
 222 my $package = $program;
 223 my $version;
 224
 225 if ($opt_output)
 226 {
 227     unlink $opt_output or kark N_("%s: can't unlink %s (%s)"),
 228         $this_program, $opt_output, $! if -e $opt_output;
 229
 230     open STDOUT, ">$opt_output"
 231         or kark N_("%s: can't create %s (%s)"), $this_program, $opt_output, $!;
 232 }
 233
 234 # The first line of the --version information is assumed to be in one
 235 # of the following formats:
 236 #
 237 #   <version>
 238 #   <program> <version>
 239 #   {GNU,Free} <program> <version>
 240 #   <program> ({GNU,Free} <package>) <version>
 241 #   <program> - {GNU,Free} <package> <version>
 242 #
 243 # and separated from any copyright/author details by a blank line.
 244
 245 ($_, $version_text) = ((split /\n+/, $version_text, 2), '');
 246
 247 if (/^(\S+) +\(((?:GNU|Free) +[^)]+)\) +(.*)/ or
 248     /^(\S+) +- *((?:GNU|Free) +\S+) +(.*)/)
 249 {
 250     $program = $1;
 251     $package = $2;
 252     $version = $3;
 253 }
 254 elsif (/^((?:GNU|Free) +)?(\S+) +(.*)/)
 255 {
 256     $program = $2;
 257     $package = $1 ? "$1$2" : $2;
 258     $version = $3;
 259 }
 260 else
 261 {
 262     $version = $_;
 263 }
 264
 265 $program =~ s!.*/!!;
 266
 267 # No info for `info' itself.
 268 $opt_no_info = 1 if $program eq 'info';
 269
 270 # Translators: "NAME", "SYNOPSIS" and other one or two word strings in all
 271 # upper case are manual page section headings.  The man(1) manual page in your
 272 # language, if available should provide the conventional translations.
 273 for ($include{_('NAME')})
 274 {
 275     if ($opt_name) # --name overrides --include contents.
 276     {
 277         $_ = "$program \\- $opt_name\n";
 278     }
 279     elsif ($_) # Use first name given as $program
 280     {
 281         $program = $1 if /^([^\s,]+)(?:,?\s*[^\s,\\-]+)*\s+\\?-/;
 282     }
 283     else # Set a default (useless) NAME paragraph.
 284     {
 285         $_ = sprintf _("%s \\- manual page for %s %s") . "\n", $program,
 286             $program, $version;
 287     }
 288 }
 289
 290 # Man pages traditionally have the page title in caps.
 291 my $PROGRAM = uc $program;
 292
 293 # Set default page head/footers
 294 $source ||= "$program $version";
 295 unless ($manual)
 296 {
 297     for ($section)
 298     {
 299         if (/^(1[Mm]|8)/) { $manual = enc _('System Administration Utilities') }
 300         elsif (/^6/)      { $manual = enc _('Games') }
 301         else              { $manual = enc _('User Commands') }
 302     }
 303 }
 304
 305 # Extract usage clause(s) [if any] for SYNOPSIS.
 306 # Translators: "Usage" and "or" here are patterns (regular expressions) which
 307 # are used to match the usage synopsis in program output.  An example from cp
 308 # (GNU coreutils) which contains both strings:
 309 #  Usage: cp [OPTION]... [-T] SOURCE DEST
 310 #    or:  cp [OPTION]... SOURCE... DIRECTORY
 311 #    or:  cp [OPTION]... -t DIRECTORY SOURCE...
 312 my $PAT_USAGE = _('Usage');
 313 my $PAT_USAGE_CONT = _('or');
 314 if ($help_text =~ s/^($PAT_USAGE):( +(\S+))(.*)((?:\n(?: {6}\1| *($PAT_USAGE_CONT): +\S).*)*)//om)
 315 {
 316     my @syn = $3 . $4;
 317
 318     if ($_ = $5)
 319     {
 320         s/^\n//;
 321         for (split /\n/) { s/^ *(($PAT_USAGE_CONT): +)?//o; push @syn, $_ }
 322     }
 323
 324     my $synopsis = '';
 325     for (@syn)
 326     {
 327         $synopsis .= ".br\n" if $synopsis;
 328         s!^\S*/!!;
 329         s/^lt-// if $opt_libtool;
 330         s/^(\S+) *//;
 331         $synopsis .= ".B $1\n";
 332         s/\s+$//;
 333         s/(([][]|\.\.+)+)/\\fR$1\\fI/g;
 334         s/^/\\fI/ unless s/^\\fR//;
 335         $_ .= '\fR';
 336         s/(\\fI)( *)/$2$1/g;
 337         s/\\fI\\fR//g;
 338         s/^\\fR//;
 339         s/\\fI$//;
 340         s/^\./\\&./;
 341
 342         $synopsis .= "$_\n";
 343     }
 344
 345     $include{_('SYNOPSIS')} ||= $synopsis;
 346 }
 347
 348 # Process text, initial section is DESCRIPTION.
 349 my $sect = _('DESCRIPTION');
 350 $_ = "$help_text\n\n$version_text";
 351
 352 # Normalise paragraph breaks.
 353 s/^\n+//;
 354 s/\n*$/\n/;
 355 s/\n\n+/\n\n/g;
 356
 357 # Join hyphenated lines.
 358 s/([A-Za-z])-\n *([A-Za-z])/$1$2/g;
 359
 360 # Temporarily exchange leading dots, apostrophes and backslashes for
 361 # tokens.
 362 s/^\./\x80/mg;
 363 s/^'/\x81/mg;
 364 s/\\/\x82/g;
 365
 366 # Translators: patterns are used to match common program output. In the source
 367 # these strings are all of the form of "my $PAT_something = _('...');" and are
 368 # regular expressions.  If there is more than one commonly used string, you
 369 # may separate alternatives with "|".  Spaces in these expressions are written
 370 # as " +" to indicate that more than one space may be matched.  The string
 371 # "(?:[\\w-]+ +)?" in the bug reporting pattern is used to indicate an
 372 # optional word, so that either "Report bugs" or "Report _program_ bugs" will
 373 # be matched.
 374 my $PAT_BUGS            = _('Report +(?:[\w-]+ +)?bugs|Email +bug +reports +to');
 375 my $PAT_AUTHOR          = _('Written +by');
 376 my $PAT_OPTIONS         = _('Options');
 377 my $PAT_ENVIRONMENT     = _('Environment');
 378 my $PAT_FILES           = _('Files');
 379 my $PAT_EXAMPLES        = _('Examples');
 380 my $PAT_FREE_SOFTWARE   = _('This +is +free +software');
 381
 382 # Start a new paragraph (if required) for these.
 383 s/([^\n])\n($PAT_BUGS|$PAT_AUTHOR) /$1\n\n$2 /og;
 384
 385 # Convert iso-8859-1 copyright symbol or (c) to nroff
 386 # character.
 387 s/^Copyright +(?:\xa9|\([Cc]\))/Copyright \\(co/mg;
 388
 389 sub convert_option;
 390
 391 while (length)
 392 {
 393     # Convert some standard paragraph names.
 394     if (s/^($PAT_OPTIONS): *\n//o)
 395     {
 396         $sect = _('OPTIONS');
 397         next;
 398     }
 399     if (s/^($PAT_ENVIRONMENT): *\n//o)
 400     {
 401         $sect = _('ENVIRONMENT');
 402         next;
 403     }
 404     if (s/^($PAT_FILES): *\n//o)
 405     {
 406         $sect = _('FILES');
 407         next;
 408     }
 409     elsif (s/^($PAT_EXAMPLES): *\n//o)
 410     {
 411         $sect = _('EXAMPLES');
 412         next;
 413     }
 414
 415     # Copyright section
 416     if (/^Copyright /)
 417     {
 418         $sect = _('COPYRIGHT');
 419     }
 420
 421     # Bug reporting section.
 422     elsif (/^($PAT_BUGS) /o)
 423     {
 424         $sect = _('REPORTING BUGS');
 425     }
 426
 427     # Author section.
 428     elsif (/^($PAT_AUTHOR)/o)
 429     {
 430         $sect = _('AUTHOR');
 431     }
 432
 433     # Examples, indicated by an indented leading $, % or > are
 434     # rendered in a constant width font.
 435     if (/^( +)([\$\%>] )\S/)
 436     {
 437         my $indent = $1;
 438         my $prefix = $2;
 439         my $break = '.IP';
 440         $include{$sect} ||= '';
 441         while (s/^$indent\Q$prefix\E(\S.*)\n*//)
 442         {
 443             $include{$sect} .= "$break\n\\f(CW$prefix$1\\fR\n";
 444             $break = '.br';
 445         }
 446
 447         next;
 448     }
 449
 450     my $matched = '';
 451     $include{$sect} ||= '';
 452
 453     # Sub-sections have a trailing colon and the second line indented.
 454     if (s/^(\S.*:) *\n / /)
 455     {
 456         $matched .= $& if %append;
 457         $include{$sect} .= qq(.SS "$1"\n);
 458     }
 459
 460     my $indent = 0;
 461     my $content = '';
 462
 463     # Option with description.
 464     if (s/^( {1,10}([+-]\S.*?))(?:(  +(?!-))|\n( {20,}))(\S.*)\n//)
 465     {
 466         $matched .= $& if %append;
 467         $indent = length ($4 || "$1$3");
 468         $content = ".TP\n\x84$2\n\x84$5\n";
 469         unless ($4)
 470         {
 471             # Indent may be different on second line.
 472             $indent = length $& if /^ {20,}/;
 473         }
 474     }
 475
 476     # Option without description.
 477     elsif (s/^ {1,10}([+-]\S.*)\n//)
 478     {
 479         $matched .= $& if %append;
 480         $content = ".HP\n\x84$1\n";
 481         $indent = 80; # not continued
 482     }
 483
 484     # Indented paragraph with tag.
 485     elsif (s/^( +(\S.*?)  +)(\S.*)\n//)
 486     {
 487         $matched .= $& if %append;
 488         $indent = length $1;
 489         $content = ".TP\n\x84$2\n\x84$3\n";
 490     }
 491
 492     # Indented paragraph.
 493     elsif (s/^( +)(\S.*)\n//)
 494     {
 495         $matched .= $& if %append;
 496         $indent = length $1;
 497         $content = ".IP\n\x84$2\n";
 498     }
 499
 500     # Left justified paragraph.
 501     else
 502     {
 503         s/(.*)\n//;
 504         $matched .= $& if %append;
 505         $content = ".PP\n" if $include{$sect};
 506         $content .= "$1\n";
 507     }
 508
 509     # Append continuations.
 510     while ($indent ? s/^ {$indent}(\S.*)\n// : s/^(\S.*)\n//)
 511     {
 512         $matched .= $& if %append;
 513         $content .= "\x84$1\n";
 514     }
 515
 516     # Move to next paragraph.
 517     s/^\n+//;
 518
 519     for ($content)
 520     {
 521         # Leading dot and apostrophe protection.
 522         s/\x84\./\x80/g;
 523         s/\x84'/\x81/g;
 524         s/\x84//g;
 525
 526         # Convert options.
 527         s/(^| |\()(-[][\w=-]+)/$1 . convert_option $2/mge;
 528
 529         # Escape remaining hyphens
 530         s/-/\x83/g;
 531
 532         if ($sect eq 'COPYRIGHT')
 533         {
 534             # Insert line breaks before additional copyright messages
 535             # and the disclaimer.
 536             s/\n(Copyright |$PAT_FREE_SOFTWARE)/\n.br\n$1/og;
 537         }
 538         elsif ($sect eq 'REPORTING BUGS')
 539         {
 540             # Handle multi-line bug reporting sections of the form:
 541             #
 542             #   Report <program> bugs to <addr>
 543             #   GNU <package> home page: <url>
 544             #   ...
 545             s/\n([[:upper:]])/\n.br\n$1/g;
 546         }
 547     }
 548
 549     # Check if matched paragraph contains /pat/.
 550     if (%append)
 551     {
 552         for my $pat (keys %append)
 553         {
 554             if ($matched =~ $pat)
 555             {
 556                 $content .= ".PP\n" unless $append{$pat} =~ /^\./;
 557                 $content .= $append{$pat};
 558             }
 559         }
 560     }
 561
 562     $include{$sect} .= $content;
 563 }
 564
 565 # Refer to the real documentation.
 566 unless ($opt_no_info)
 567 {
 568     my $info_page = $opt_info || $program;
 569
 570     $sect = _('SEE ALSO');
 571     $include{$sect} ||= '';
 572     $include{$sect} .= ".PP\n" if $include{$sect};
 573     $include{$sect} .= sprintf _(<<'EOT'), $program, $program, $info_page;
 574 The full documentation for
 575 .B %s
 576 is maintained as a Texinfo manual.  If the
 577 .B info
 578 and
 579 .B %s
 580 programs are properly installed at your site, the command
 581 .IP
 582 .B info %s
 583 .PP
 584 should give you access to the complete manual.
 585 EOT
 586 }
 587
 588 # Output header.
 589 print <<EOT;
 590 .\\" DO NOT MODIFY THIS FILE!  It was generated by $this_program $this_version.
 591 .TH $PROGRAM "$section" "$date" "$source" "$manual"
 592 EOT
 593
 594 # Section ordering.
 595 my @pre = (_('NAME'), _('SYNOPSIS'), _('DESCRIPTION'), _('OPTIONS'),
 596     _('ENVIRONMENT'), _('FILES'), _('EXAMPLES'));
 597
 598 my @post = (_('AUTHOR'), _('REPORTING BUGS'), _('COPYRIGHT'), _('SEE ALSO'));
 599 my $filter = join '|', @pre, @post;
 600
 601 # Output content.
 602 for my $sect (@pre, (grep ! /^($filter)$/o, @include), @post)
 603 {
 604     if ($include{$sect})
 605     {
 606         my $quote = $sect =~ /\W/ ? '"' : '';
 607         print enc ".SH $quote$sect$quote\n";
 608
 609         for ($include{$sect})
 610         {
 611             # Replace leading dot, apostrophe, backslash and hyphen
 612             # tokens.
 613             s/\x80/\\&./g;
 614             s/\x81/\\&'/g;
 615             s/\x82/\\e/g;
 616             s/\x83/\\-/g;
 617
 618             # Convert some latin1 chars to troff equivalents
 619             s/\xa0/\\ /g; # non-breaking space
 620
 621             print enc $_;
 622         }
 623     }
 624 }
 625
 626 close STDOUT or kark N_("%s: error writing to %s (%s)"), $this_program,
 627     $opt_output || 'stdout', $!;
 628
 629 exit;
 630
 631 # Call program with given option and return results.
 632 sub get_option_value
 633 {
 634     my ($prog, $opt) = @_;
 635     my $stderr = $discard_stderr ? '/dev/null' : '&1';
 636     my $value = join '',
 637         map { s/ +$//; expand $_ }
 638         map { dec $_ }
 639         `$prog $opt 2>$stderr`;
 640
 641     unless ($value)
 642     {
 643         my $err = N_("%s: can't get `%s' info from %s%s");
 644         my $extra = $discard_stderr
 645             ? "\n" . N_("Try `--no-discard-stderr' if option outputs to stderr")
 646             : '';
 647
 648         kark $err, $this_program, $opt, $prog, $extra;
 649     }
 650
 651     return $value;
 652 }
 653
 654 # Convert option dashes to \- to stop nroff from hyphenating 'em, and
 655 # embolden.  Option arguments get italicised.
 656 sub convert_option
 657 {
 658     local $_ = '\fB' . shift;
 659
 660     s/-/\x83/g;
 661     unless (s/\[=(.*)\]$/\\fR[=\\fI$1\\fR]/)
 662     {
 663         s/=(.)/\\fR=\\fI$1/;
 664         s/ (.)/ \\fI$1/;
 665         $_ .= '\fR';
 666     }
 667
 668     $_;
 669 }