1#!/usr/bin/perl -w
2
3# Generate a short man page from --help and --version output.
4# Copyright (C) 1997-2004, 2008-2010 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 of the License, or
9# (at your option) 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
22use 5.005;
23use strict;
24use Getopt::Long;
25use Text::Tabs qw(expand);
26use POSIX qw(strftime setlocale LC_ALL);
27use locale;
28
29my $this_program = 'help2man';
30my $this_version = '1.35';
31
32my $have_gettext;
33BEGIN {
34    eval {
35	require Locale::gettext;
36	Locale::gettext->import (qw(gettext textdomain));
37	$have_gettext = 1;
38    };
39
40    unless ($have_gettext)
41    {
42	*gettext = sub { $_[0] };
43	*textdomain = sub {};
44    }
45}
46
47sub _ { gettext @_ }
48sub N_ { $_[0] }
49
50textdomain $this_program;
51{
52    my ($user_locale) = grep defined && length,
53	(map $ENV{$_}, qw(LANGUAGE LC_ALL LC_MESSAGES LANG)), 'C';
54
55    sub kark # die with message formatted in the invoking user's locale
56    {
57	setlocale LC_ALL, $user_locale;
58	my $fmt = gettext shift;
59	die +(sprintf $fmt, @_), "\n";
60    }
61}
62
63my $version_info = sprintf _(<<'EOT'), $this_program, $this_version;
64GNU %s %s
65
66Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004 Free Software
67Foundation, Inc.
68This is free software; see the source for copying conditions.  There is NO
69warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
70
71Written by Brendan O'Dea <bod@debian.org>
72EOT
73
74my $help_info = sprintf _(<<'EOT'), $this_program, $this_program;
75`%s' generates a man page out of `--help' and `--version' output.
76
77Usage: %s [OPTION]... EXECUTABLE
78
79 -n, --name=STRING       description for the NAME paragraph
80 -s, --section=SECTION   section number for manual page (1, 6, 8)
81 -m, --manual=TEXT       name of manual (User Commands, ...)
82 -S, --source=TEXT       source of program (FSF, Debian, ...)
83 -L, --locale=STRING     select locale (default "C")
84 -i, --include=FILE      include material from `FILE'
85 -I, --opt-include=FILE  include material from `FILE' if it exists
86 -o, --output=FILE       send output to `FILE'
87 -p, --info-page=TEXT    name of Texinfo manual
88 -N, --no-info           suppress pointer to Texinfo manual
89     --help              print this help, then exit
90     --version           print version number, then exit
91
92EXECUTABLE should accept `--help' and `--version' options although
93alternatives may be specified using:
94
95 -h, --help-option=STRING     help option string
96 -v, --version-option=STRING  version option string
97
98Report bugs to <bug-help2man@gnu.org>.
99EOT
100
101my $section = 1;
102my $manual = '';
103my $source = '';
104my $locale = 'C';
105my $help_option = '--help';
106my $version_option = '--version';
107my ($opt_name, @opt_include, $opt_output, $opt_info, $opt_no_info);
108
109my %opt_def = (
110    'n|name=s'		 => \$opt_name,
111    's|section=s'	 => \$section,
112    'm|manual=s'	 => \$manual,
113    'S|source=s'	 => \$source,
114    'L|locale=s'	 => \$locale,
115    'i|include=s'	 => sub { push @opt_include, [ pop, 1 ] },
116    'I|opt-include=s'	 => sub { push @opt_include, [ pop, 0 ] },
117    'o|output=s'	 => \$opt_output,
118    'p|info-page=s'	 => \$opt_info,
119    'N|no-info'		 => \$opt_no_info,
120    'h|help-option=s'	 => \$help_option,
121    'v|version-option=s' => \$version_option,
122);
123
124# Parse options.
125Getopt::Long::config('bundling');
126GetOptions (%opt_def,
127    help    => sub { print $help_info; exit },
128    version => sub { print $version_info; exit },
129) or die $help_info;
130
131die $help_info unless @ARGV == 1;
132
133die "$this_program: no locale support (Locale::gettext required)\n"
134    unless $locale eq 'C' or $have_gettext;
135
136# Set localization of date and executable's output.
137delete @ENV{qw(LANGUAGE LC_MESSAGES LANG)};
138setlocale LC_ALL, $ENV{LC_ALL} = $locale;
139
140my %include = ();
141my %append = ();
142my @include = (); # retain order given in include file
143
144# Process include file (if given).  Format is:
145#
146#   [section name]
147#   verbatim text
148#
149# or
150#
151#   /pattern/
152#   verbatim text
153#
154
155while (@opt_include)
156{
157    my ($inc, $required) = @{shift @opt_include};
158
159    next unless -f $inc or $required;
160    kark N_("%s: can't open `%s' (%s)"), $this_program, $inc, $!
161	unless open INC, $inc;
162
163    my $key;
164    my $hash = \%include;
165
166    while (<INC>)
167    {
168	# [section]
169	if (/^\[([^]]+)\]/)
170	{
171	    $key = uc $1;
172	    $key =~ s/^\s+//;
173	    $key =~ s/\s+$//;
174	    $hash = \%include;
175	    push @include, $key unless $include{$key};
176	    next;
177	}
178
179	# /pattern/
180	if (m!^/(.*)/([ims]*)!)
181	{
182	    my $pat = $2 ? "(?$2)$1" : $1;
183
184	    # Check pattern.
185	    eval { $key = qr($pat) };
186	    if ($@)
187	    {
188		$@ =~ s/ at .*? line \d.*//;
189		die "$inc:$.:$@";
190	    }
191
192	    $hash = \%append;
193	    next;
194	}
195
196	# Check for options before the first section--anything else is
197	# silently ignored, allowing the first for comments and
198	# revision info.
199	unless ($key)
200	{
201	    # handle options
202	    if (/^-/)
203	    {
204		local @ARGV = split;
205		GetOptions %opt_def;
206	    }
207
208	    next;
209	}
210
211	$hash->{$key} ||= '';
212	$hash->{$key} .= $_;
213    }
214
215    close INC;
216
217    kark N_("%s: no valid information found in `%s'"), $this_program, $inc
218	unless $key;
219}
220
221# Compress trailing blank lines.
222for my $hash (\(%include, %append))
223{
224    for (keys %$hash) { $hash->{$_} =~ s/\n+$/\n/ }
225}
226
227# Grab help and version info from executable.
228my ($help_text, $version_text) = map {
229    join '', map { s/ +$//; expand $_ } `$ARGV[0] $_ 2>/dev/null`
230	or kark N_("%s: can't get `%s' info from %s"), $this_program,
231	    $_, $ARGV[0]
232} $help_option, $version_option;
233
234my $date = strftime "%B %Y", localtime;
235(my $program = $ARGV[0]) =~ s!.*/!!;
236my $package = $program;
237my $version;
238
239if ($opt_output)
240{
241    unlink $opt_output or kark N_("%s: can't unlink %s (%s)"),
242	$this_program, $opt_output, $! if -e $opt_output;
243
244    open STDOUT, ">$opt_output"
245	or kark N_("%s: can't create %s (%s)"), $this_program, $opt_output, $!;
246}
247
248# The first line of the --version information is assumed to be in one
249# of the following formats:
250#
251#   <version>
252#   <program> <version>
253#   {GNU,Free} <program> <version>
254#   <program> ({GNU,Free} <package>) <version>
255#   <program> - {GNU,Free} <package> <version>
256#
257# and seperated from any copyright/author details by a blank line.
258
259($_, $version_text) = split /\n+/, $version_text, 2;
260
261if (/^(\S+) +\(((?:GNU|Free) +[^)]+)\) +(.*)/ or
262    /^(\S+) +- *((?:GNU|Free) +\S+) +(.*)/)
263{
264    $program = $1;
265    $package = $2;
266    $version = $3;
267}
268elsif (/^((?:GNU|Free) +)?(\S+) +(.*)/)
269{
270    $program = $2;
271    $package = $1 ? "$1$2" : $2;
272    $version = $3;
273}
274else
275{
276    $version = $_;
277}
278
279$program =~ s!.*/!!;
280
281# No info for `info' itself.
282$opt_no_info = 1 if $program eq 'info';
283
284for ($include{_('NAME')})
285{
286    if ($opt_name) # --name overrides --include contents.
287    {
288	$_ = "$program \\- $opt_name\n";
289    }
290    elsif ($_) # Use first name given as $program
291    {
292	$program = $1 if /^([^\s,]+)(?:,?\s*[^\s,\\-]+)*\s+\\?-/;
293    }
294    else # Set a default (useless) NAME paragraph.
295    {
296	$_ = sprintf _("%s \\- manual page for %s %s") . "\n", $program,
297	    $program, $version;
298    }
299}
300
301# Man pages traditionally have the page title in caps.
302my $PROGRAM = uc $program;
303
304# Set default page head/footers
305$source ||= "$program $version";
306unless ($manual)
307{
308    for ($section)
309    {
310	if (/^(1[Mm]|8)/) { $manual = _('System Administration Utilities') }
311	elsif (/^6/)	  { $manual = _('Games') }
312	else		  { $manual = _('User Commands') }
313    }
314}
315
316# Extract usage clause(s) [if any] for SYNOPSIS.
317my $PAT_USAGE = _('Usage');
318my $PAT_USAGE_CONT = _('or');
319if ($help_text =~ s/^($PAT_USAGE):( +(\S+))(.*)((?:\n(?: {6}\1| *($PAT_USAGE_CONT): +\S).*)*)//om)
320{
321    my @syn = $3 . $4;
322
323    if ($_ = $5)
324    {
325	s/^\n//;
326	for (split /\n/) { s/^ *(($PAT_USAGE_CONT): +)?//o; push @syn, $_ }
327    }
328
329    my $synopsis = '';
330    for (@syn)
331    {
332	$synopsis .= ".br\n" if $synopsis;
333	s!^\S*/!!;
334	s/^(\S+) *//;
335	$synopsis .= ".B $1\n";
336	s/\s+$//;
337	s/(([][]|\.\.+)+)/\\fR$1\\fI/g;
338	s/^/\\fI/ unless s/^\\fR//;
339	$_ .= '\fR';
340	s/(\\fI)( *)/$2$1/g;
341	s/\\fI\\fR//g;
342	s/^\\fR//;
343	s/\\fI$//;
344	s/^\./\\&./;
345
346	$synopsis .= "$_\n";
347    }
348
349    $include{_('SYNOPSIS')} ||= $synopsis;
350}
351
352# Process text, initial section is DESCRIPTION.
353my $sect = _('DESCRIPTION');
354$_ = "$help_text\n\n$version_text";
355
356# Normalise paragraph breaks.
357s/^\n+//;
358s/\n*$/\n/;
359s/\n\n+/\n\n/g;
360
361# Join hyphenated lines.
362s/([A-Za-z])-\n *([A-Za-z])/$1$2/g;
363
364# Temporarily exchange leading dots, apostrophes and backslashes for
365# tokens.
366s/^\./\x80/mg;
367s/^'/\x81/mg;
368s/\\/\x82/g;
369
370my $PAT_BUGS		= _('Report +(?:\w+ +)?bugs|Email +bug +reports +to');
371my $PAT_AUTHOR		= _('Written +by');
372my $PAT_OPTIONS		= _('Options');
373my $PAT_EXAMPLES	= _('Examples');
374my $PAT_FREE_SOFTWARE	= _('This +is +free +software');
375my $PAT_INFO		= _('For +complete +documentation');
376
377# Start a new paragraph (if required) for these.
378s/([^\n])\n($PAT_BUGS|$PAT_AUTHOR)/$1\n\n$2/og;
379
380sub convert_option;
381
382while (length)
383{
384    # Convert some standard paragraph names.
385    if (s/^($PAT_OPTIONS): *\n//o)
386    {
387	$sect = _('OPTIONS');
388	next;
389    }
390    elsif (s/^($PAT_EXAMPLES): *\n//o)
391    {
392	$sect = _('EXAMPLES');
393	next;
394    }
395    # Skip any texinfo reference as that's handled separately
396    if (s/($PAT_INFO).*\n//o)
397    {
398	next;
399    }
400
401    # Copyright section
402    if (/^Copyright +[(\xa9]/)
403    {
404	$sect = _('COPYRIGHT');
405	$include{$sect} ||= '';
406	$include{$sect} .= ".PP\n" if $include{$sect};
407
408	my $copy;
409	($copy, $_) = split /\n\n/, $_, 2;
410
411	for ($copy)
412	{
413	    # Add back newline
414	    s/\n*$/\n/;
415
416	    # Convert iso9959-1 copyright symbol or (c) to nroff
417	    # character.
418	    s/^Copyright +(?:\xa9|\([Cc]\))/Copyright \\(co/mg;
419
420	    # Insert line breaks before additional copyright messages
421	    # and the disclaimer.
422	    s/(.)\n(Copyright |$PAT_FREE_SOFTWARE)/$1\n.br\n$2/og;
423	}
424
425	$include{$sect} .= $copy;
426	$_ ||= '';
427	next;
428    }
429
430    # Catch bug report text.
431    if (/^($PAT_BUGS) /o)
432    {
433	$sect = _('REPORTING BUGS');
434    }
435
436    # Author section.
437    elsif (/^($PAT_AUTHOR)/o)
438    {
439	$sect = _('AUTHOR');
440    }
441
442    # Examples, indicated by an indented leading $, % or > are
443    # rendered in a constant width font.
444    if (/^( +)([\$\%>] )\S/)
445    {
446	my $indent = $1;
447	my $prefix = $2;
448	my $break = '.IP';
449	$include{$sect} ||= '';
450	while (s/^$indent\Q$prefix\E(\S.*)\n*//)
451	{
452	    $include{$sect} .= "$break\n\\f(CW$prefix$1\\fR\n";
453	    $break = '.br';
454	}
455
456	next;
457    }
458
459    my $matched = '';
460    $include{$sect} ||= '';
461
462    # Sub-sections have a trailing colon and the second line indented.
463    if (s/^(\S.*:) *\n / /)
464    {
465	$matched .= $& if %append;
466	$include{$sect} .= qq(.SS "$1"\n);
467    }
468
469    my $indent = 0;
470    my $content = '';
471
472    # Option with description.
473    if (s/^( {1,10}([+-]\S.*?))(?:(  +(?!-))|\n( {20,}))(\S.*)\n//)
474    {
475	$matched .= $& if %append;
476	$indent = length ($4 || "$1$3");
477	$content = ".TP\n\x84$2\n\x84$5\n";
478	unless ($4)
479	{
480	    # Indent may be different on second line.
481	    $indent = length $& if /^ {20,}/;
482	}
483    }
484
485    # Option without description.
486    elsif (s/^ {1,10}([+-]\S.*)\n//)
487    {
488	$matched .= $& if %append;
489	$content = ".HP\n\x84$1\n";
490	$indent = 80; # not continued
491    }
492
493    # Indented paragraph with tag.
494    elsif (s/^( +(\S.*?)  +)(\S.*)\n//)
495    {
496	$matched .= $& if %append;
497	$indent = length $1;
498	$content = ".TP\n\x84$2\n\x84$3\n";
499    }
500
501    # Indented paragraph.
502    elsif (s/^( +)(\S.*)\n//)
503    {
504	$matched .= $& if %append;
505	$indent = length $1;
506	$content = ".IP\n\x84$2\n";
507    }
508
509    # Left justified paragraph.
510    else
511    {
512	s/(.*)\n//;
513	$matched .= $& if %append;
514	$content = ".PP\n" if $include{$sect};
515	$content .= "$1\n";
516    }
517
518    # Append continuations.
519    while ($indent ? s/^ {$indent}(\S.*)\n// : s/^(\S.*)\n//)
520    {
521	$matched .= $& if %append;
522	$content .= "\x84$1\n"
523    }
524
525    # Move to next paragraph.
526    s/^\n+//;
527
528    for ($content)
529    {
530	# Leading dot and apostrophe protection.
531	s/\x84\./\x80/g;
532	s/\x84'/\x81/g;
533	s/\x84//g;
534
535	# Convert options.
536	s/(^| |\()(-[][\w=-]+)/$1 . convert_option $2/mge;
537
538	# Escape remaining hyphens
539	s/-/\x83/g;
540    }
541
542    # Check if matched paragraph contains /pat/.
543    if (%append)
544    {
545	for my $pat (keys %append)
546	{
547	    if ($matched =~ $pat)
548	    {
549		$content .= ".PP\n" unless $append{$pat} =~ /^\./;
550		$content .= $append{$pat};
551	    }
552	}
553    }
554
555    $include{$sect} .= $content;
556}
557
558# Refer to the real documentation.
559unless ($opt_no_info)
560{
561    my $info_page = $opt_info || $program;
562
563    $sect = _('SEE ALSO');
564    $include{$sect} ||= '';
565    $include{$sect} .= ".PP\n" if $include{$sect};
566    $include{$sect} .= sprintf _(<<'EOT'), $program, $program, $info_page;
567The full documentation for
568.B %s
569is maintained as a Texinfo manual.  If the
570.B info
571and
572.B %s
573programs are properly installed at your site, the command
574.IP
575.B info coreutils \(aq%s invocation\(aq
576.PP
577should give you access to the complete manual.
578EOT
579}
580
581# Output header.
582print <<EOT;
583.\\" DO NOT MODIFY THIS FILE!  It was generated by $this_program $this_version.
584.TH $PROGRAM "$section" "$date" "$source" "$manual"
585EOT
586
587# Section ordering.
588my @pre = (_('NAME'), _('SYNOPSIS'), _('DESCRIPTION'), _('OPTIONS'),
589    _('EXAMPLES'));
590
591my @post = (_('AUTHOR'), _('REPORTING BUGS'), _('COPYRIGHT'), _('SEE ALSO'));
592my $filter = join '|', @pre, @post;
593
594# Output content.
595for my $sect (@pre, (grep ! /^($filter)$/o, @include), @post)
596{
597    if ($include{$sect})
598    {
599	my $lsect = gettext $sect;
600	my $quote = $lsect =~ /\W/ ? '"' : '';
601	print ".SH $quote$lsect$quote\n";
602
603	for ($include{$sect})
604	{
605	    # Replace leading dot, apostrophe, backslash and hyphen
606	    # tokens.
607	    s/\x80/\\&./g;
608	    s/\x81/\\&'/g;
609	    s/\x82/\\e/g;
610	    s/\x83/\\-/g;
611
612	    # Convert some latin1 chars to troff equivalents
613	    s/\xa0/\\ /g; # non-breaking space
614
615	    $sect eq 'REPORTING BUGS'
616		and s/\n(.)/\n.br\n$1/g;
617
618	    print;
619	}
620    }
621}
622
623close STDOUT or kark N_("%s: error writing to %s (%s)"), $this_program,
624    $opt_output || 'stdout', $!;
625
626exit;
627
628# Convert option dashes to \- to stop nroff from hyphenating 'em, and
629# embolden.  Option arguments get italicised.
630sub convert_option
631{
632    local $_ = '\fB' . shift;
633
634    s/-/\x83/g;
635    unless (s/\[=(.*)\]$/\\fR[=\\fI$1\\fR]/)
636    {
637	s/=(.)/\\fR=\\fI$1/;
638	s/ (.)/ \\fI$1/;
639	$_ .= '\fR';
640    }
641
642    $_;
643}
644