1 # Pod::Man -- Convert POD data to formatted *roff input.
2 # $Id: Man.pm,v 0.8 1999/10/07 09:39:37 eagle Exp $
4 # Copyright 1999 by Russ Allbery <rra@stanford.edu>
6 # This program is free software; you can redistribute it and/or modify it
7 # under the same terms as Perl itself.
9 # This module is intended to be a replacement for pod2man, and attempts to
10 # match its output except for some specific circumstances where other
11 # decisions seemed to produce better output. It uses Pod::Parser and is
12 # designed to be very easy to subclass.
14 ############################################################################
15 # Modules and declarations
16 ############################################################################
22 use Carp qw(carp croak);
26 use subs qw(makespace);
27 use vars qw(@ISA %ESCAPES $PREAMBLE $VERSION);
29 @ISA = qw(Pod::Parser);
31 ($VERSION = (split (' ', q$Revision: 0.8 $ ))[1]) =~ s/\.(\d)$/.0$1/;
34 ############################################################################
35 # Preamble and *roff output tables
36 ############################################################################
38 # The following is the static preamble which starts all *roff output we
39 # generate. It's completely static except for the font to use as a
40 # fixed-width font, which is designed by @CFONT@. $PREAMBLE should
41 # therefore be run through s/\@CFONT\@/<font>/g before output.
42 $PREAMBLE = <<'----END OF PREAMBLE----';
43 .de Sh \" Subsection heading
51 .de Sp \" Vertical space (when we can't use .PP)
57 .ie \\n(.$>=3 .ne \\$3
61 .de Vb \" Begin verbatim text
66 .de Ve \" End verbatim text
71 .\" Set up some character translations and predefined strings. \*(-- will
72 .\" give an unbreakable dash, \*(PI will give pi, \*(L" will give a left
73 .\" double quote, and \*(R" will give a right double quote. | will give a
74 .\" real vertical bar. \*(C+ will give a nicer C++. Capital omega is used
75 .\" to do unbreakable dashes and therefore won't be available. \*(C` and
76 .\" \*(C' expand to `' in nroff, nothing in troff, for use with C<>
78 .ds C+ C\v'-.1v'\h'-1p'\s-2+\h'-1p'+\s0\v'.1v'\h'-1p'
82 . if (\n(.H=4u)&(1m=24u) .ds -- \(*W\h'-12u'\(*W\h'-12u'-\" diablo 10 pitch
83 . if (\n(.H=4u)&(1m=20u) .ds -- \(*W\h'-12u'\(*W\h'-8u'-\" diablo 12 pitch
96 .\" If the F register is turned on, we'll generate index entries on stderr
97 .\" for titles (.TH), headers (.SH), subsections (.Sh), items (.Ip), and
98 .\" index entries marked with X<> in POD. Of course, you'll have to process
99 .\" the output yourself in some meaningful fashion.
102 . tm Index:\\$1\t\\n%\t"\\$2"
108 .\" For nroff, turn off justification. Always turn off hyphenation; it
109 .\" makes way too many mistakes in technical documents.
113 .\" Accent mark definitions (@(#)ms.acc 1.5 88/02/08 SMI; from UCB 4.2).
114 .\" Fear. Run. Save yourself. No user-serviceable parts.
116 . \" fudge factors for nroff and troff
125 . ds #H ((1u-(\\\\n(.fu%2u))*.13m)
131 . \" simple accents for nroff and troff
141 . ds ' \\k:\h'-(\\n(.wu*8/10-\*(#H)'\'\h"|\\n:u"
142 . ds ` \\k:\h'-(\\n(.wu*8/10-\*(#H)'\`\h'|\\n:u'
143 . ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'^\h'|\\n:u'
144 . ds , \\k:\h'-(\\n(.wu*8/10)',\h'|\\n:u'
145 . ds ~ \\k:\h'-(\\n(.wu-\*(#H-.1m)'~\h'|\\n:u'
146 . ds / \\k:\h'-(\\n(.wu*8/10-\*(#H)'\z\(sl\h'|\\n:u'
148 . \" troff and (daisy-wheel) nroff accents
149 .ds : \\k:\h'-(\\n(.wu*8/10-\*(#H+.1m+\*(#F)'\v'-\*(#V'\z.\h'.2m+\*(#F'.\h'|\\n:u'\v'\*(#V'
150 .ds 8 \h'\*(#H'\(*b\h'-\*(#H'
151 .ds o \\k:\h'-(\\n(.wu+\w'\(de'u-\*(#H)/2u'\v'-.3n'\*(#[\z\(de\v'.3n'\h'|\\n:u'\*(#]
152 .ds d- \h'\*(#H'\(pd\h'-\w'~'u'\v'-.25m'\f2\(hy\fP\v'.25m'\h'-\*(#H'
153 .ds D- D\\k:\h'-\w'D'u'\v'-.11m'\z\(hy\v'.11m'\h'|\\n:u'
154 .ds th \*(#[\v'.3m'\s+1I\s-1\v'-.3m'\h'-(\w'I'u*2/3)'\s-1o\s+1\*(#]
155 .ds Th \*(#[\s+2I\s-2\h'-\w'I'u*3/5'\v'-.3m'o\v'.3m'\*(#]
156 .ds ae a\h'-(\w'a'u*4/10)'e
157 .ds Ae A\h'-(\w'A'u*4/10)'E
158 . \" corrections for vroff
159 .if v .ds ~ \\k:\h'-(\\n(.wu*9/10-\*(#H)'\s-2\u~\d\s+2\h'|\\n:u'
160 .if v .ds ^ \\k:\h'-(\\n(.wu*10/11-\*(#H)'\v'-.4m'^\v'.4m'\h'|\\n:u'
161 . \" for low resolution devices (crt and lpr)
162 .if \n(.H>23 .if \n(.V>19 \
175 ----END OF PREAMBLE----
177 # This table is taken nearly verbatim from Tom Christiansen's pod2man. It
178 # assumes that the standard preamble has already been printed, since that's
179 # what defines all of the accent marks. Note that some of these are quoted
180 # with double quotes since they contain embedded single quotes, so use \\
181 # uniformly for backslash for readability.
183 'amp' => '&', # ampersand
184 'lt' => '<', # left chevron, less-than
185 'gt' => '>', # right chevron, greater-than
186 'quot' => '"', # double quote
188 'Aacute' => "A\\*'", # capital A, acute accent
189 'aacute' => "a\\*'", # small a, acute accent
190 'Acirc' => 'A\\*^', # capital A, circumflex accent
191 'acirc' => 'a\\*^', # small a, circumflex accent
192 'AElig' => '\*(AE', # capital AE diphthong (ligature)
193 'aelig' => '\*(ae', # small ae diphthong (ligature)
194 'Agrave' => "A\\*`", # capital A, grave accent
195 'agrave' => "A\\*`", # small a, grave accent
196 'Aring' => 'A\\*o', # capital A, ring
197 'aring' => 'a\\*o', # small a, ring
198 'Atilde' => 'A\\*~', # capital A, tilde
199 'atilde' => 'a\\*~', # small a, tilde
200 'Auml' => 'A\\*:', # capital A, dieresis or umlaut mark
201 'auml' => 'a\\*:', # small a, dieresis or umlaut mark
202 'Ccedil' => 'C\\*,', # capital C, cedilla
203 'ccedil' => 'c\\*,', # small c, cedilla
204 'Eacute' => "E\\*'", # capital E, acute accent
205 'eacute' => "e\\*'", # small e, acute accent
206 'Ecirc' => 'E\\*^', # capital E, circumflex accent
207 'ecirc' => 'e\\*^', # small e, circumflex accent
208 'Egrave' => 'E\\*`', # capital E, grave accent
209 'egrave' => 'e\\*`', # small e, grave accent
210 'ETH' => '\\*(D-', # capital Eth, Icelandic
211 'eth' => '\\*(d-', # small eth, Icelandic
212 'Euml' => 'E\\*:', # capital E, dieresis or umlaut mark
213 'euml' => 'e\\*:', # small e, dieresis or umlaut mark
214 'Iacute' => "I\\*'", # capital I, acute accent
215 'iacute' => "i\\*'", # small i, acute accent
216 'Icirc' => 'I\\*^', # capital I, circumflex accent
217 'icirc' => 'i\\*^', # small i, circumflex accent
218 'Igrave' => 'I\\*`', # capital I, grave accent
219 'igrave' => 'i\\*`', # small i, grave accent
220 'Iuml' => 'I\\*:', # capital I, dieresis or umlaut mark
221 'iuml' => 'i\\*:', # small i, dieresis or umlaut mark
222 'Ntilde' => 'N\*~', # capital N, tilde
223 'ntilde' => 'n\*~', # small n, tilde
224 'Oacute' => "O\\*'", # capital O, acute accent
225 'oacute' => "o\\*'", # small o, acute accent
226 'Ocirc' => 'O\\*^', # capital O, circumflex accent
227 'ocirc' => 'o\\*^', # small o, circumflex accent
228 'Ograve' => 'O\\*`', # capital O, grave accent
229 'ograve' => 'o\\*`', # small o, grave accent
230 'Oslash' => 'O\\*/', # capital O, slash
231 'oslash' => 'o\\*/', # small o, slash
232 'Otilde' => 'O\\*~', # capital O, tilde
233 'otilde' => 'o\\*~', # small o, tilde
234 'Ouml' => 'O\\*:', # capital O, dieresis or umlaut mark
235 'ouml' => 'o\\*:', # small o, dieresis or umlaut mark
236 'szlig' => '\*8', # small sharp s, German (sz ligature)
237 'THORN' => '\\*(Th', # capital THORN, Icelandic
238 'thorn' => '\\*(th', # small thorn, Icelandic
239 'Uacute' => "U\\*'", # capital U, acute accent
240 'uacute' => "u\\*'", # small u, acute accent
241 'Ucirc' => 'U\\*^', # capital U, circumflex accent
242 'ucirc' => 'u\\*^', # small u, circumflex accent
243 'Ugrave' => 'U\\*`', # capital U, grave accent
244 'ugrave' => 'u\\*`', # small u, grave accent
245 'Uuml' => 'U\\*:', # capital U, dieresis or umlaut mark
246 'uuml' => 'u\\*:', # small u, dieresis or umlaut mark
247 'Yacute' => "Y\\*'", # capital Y, acute accent
248 'yacute' => "y\\*'", # small y, acute accent
249 'yuml' => 'y\\*:', # small y, dieresis or umlaut mark
253 ############################################################################
254 # Static helper functions
255 ############################################################################
257 # Protect leading quotes and periods against interpretation as commands.
258 sub protect { local $_ = shift; s/^([.\'])/\\&$1/mg; $_ }
260 # Given a command and a single argument that may or may not contain double
261 # quotes, handle double-quote formatting for it. If there are no double
262 # quotes, just return the command followed by the argument in double quotes.
263 # If there are double quotes, use an if statement to test for nroff, and for
264 # nroff output the command followed by the argument in double quotes with
265 # embedded double quotes doubled. For other formatters, remap paired double
266 # quotes to `` and ''.
275 $troff =~ s/\"\"([^\"]*)\"\"/\`\`$1\'\'/g;
276 s/\"/\"\"/g if $extra;
277 $troff =~ s/\"/\"\"/g if $extra;
278 $_ = qq("$_") . ($extra ? " $extra" : '');
279 $troff = qq("$troff") . ($extra ? " $extra" : '');
280 return ".if n $command $_\n.el $command $troff\n";
282 $_ = qq("$_") . ($extra ? " $extra" : '');
283 return "$command $_\n";
287 # Translate a font string into an escape.
288 sub toescape { (length ($_[0]) > 1 ? '\f(' : '\f') . $_[0] }
291 ############################################################################
293 ############################################################################
295 # Initialize the object. Here, we also process any additional options
296 # passed to the constructor or set up defaults if none were given. center
297 # is the centered title, release is the version number, and date is the date
298 # for the documentation. Note that we can't know what file name we're
299 # processing due to the architecture of Pod::Parser, so that *has* to either
300 # be passed to the constructor or set separately with Pod::Man::name().
304 # Figure out the fixed-width font. If user-supplied, make sure that
305 # they are the right length.
306 for (qw/fixed fixedbold fixeditalic fixedbolditalic/) {
307 if (defined $$self{$_}) {
308 if (length ($$self{$_}) < 1 || length ($$self{$_}) > 2) {
309 croak "roff font should be 1 or 2 chars, not `$$self{$_}'";
316 # Set the default fonts. We can't be sure what fixed bold-italic is
317 # going to be called, so default to just bold.
318 $$self{fixed} ||= 'CW';
319 $$self{fixedbold} ||= 'CB';
320 $$self{fixeditalic} ||= 'CI';
321 $$self{fixedbolditalic} ||= 'CB';
323 # Set up a table of font escapes. First number is fixed-width, second
324 # is bold, third is italic.
325 $$self{FONTS} = { '000' => '\fR', '001' => '\fI',
326 '010' => '\fB', '011' => '\f(BI',
327 '100' => toescape ($$self{fixed}),
328 '101' => toescape ($$self{fixeditalic}),
329 '110' => toescape ($$self{fixedbold}),
330 '111' => toescape ($$self{fixedbolditalic})};
332 # Extra stuff for page titles.
333 $$self{center} = 'User Contributed Perl Documentation'
334 unless defined $$self{center};
335 $$self{indent} = 4 unless defined $$self{indent};
337 # We used to try first to get the version number from a local binary,
338 # but we shouldn't need that any more. Get the version from the running
340 if (!defined $$self{release}) {
341 my ($version, $patch) = ($] =~ /^(.{5})(\d{2})?/);
342 $$self{release} = "perl $version";
343 $$self{release} .= ", patch $patch" if $patch;
346 # Double quotes in things that will be quoted.
347 for (qw/center date release/) { $$self{$_} =~ s/\"/\"\"/g }
349 $$self{INDENT} = 0; # Current indentation level.
350 $$self{INDENTS} = []; # Stack of indentations.
351 $$self{INDEX} = []; # Index keys waiting to be printed.
353 $self->SUPER::initialize;
356 # For each document we process, output the preamble first. Note that the
357 # fixed width font is a global default; once we interpolate it into the
358 # PREAMBLE, it ain't ever changing. Maybe fix this later.
362 # Try to figure out the name and section from the file name.
363 my $section = $$self{section} || 1;
364 my $name = $$self{name};
365 if (!defined $name) {
366 $name = $self->input_file;
367 $section = 3 if (!$$self{section} && $name =~ /\.pm$/i);
368 $name =~ s/\.p(od|[lm])$//i;
369 if ($section =~ /^1/) {
370 require File::Basename;
371 $name = uc File::Basename::basename ($name);
373 # Lose everything up to the first of
374 # */lib/*perl* standard or site_perl module
375 # */*perl*/lib from -D prefix=/opt/perl
376 # */*perl*/ random module hierarchy
377 # which works. Should be fixed to use File::Spec.
380 if ( s%^.*?/lib/[^/]*perl[^/]*/%%i
381 or s%^.*?/[^/]*perl[^/]*/(?:lib/)?%%i) {
382 s%^site(_perl)?/%%; # site and site_perl
383 s%^(.*-$^O|$^O-.*)/%%o; # arch
384 s%^\d+\.\d+%%; # version
391 # Modification date header. Try to use the modification time of our
393 if (!defined $$self{date}) {
394 my $time = (stat $self->input_file)[9] || time;
395 my ($day, $month, $year) = (localtime $time)[3,4,5];
398 $$self{date} = join ('-', $year, $month, $day);
401 # Now, print out the preamble and the title.
402 $PREAMBLE =~ s/\@CFONT\@/$$self{fixed}/;
404 print { $self->output_handle } <<"----END OF HEADER----";
405 .\\" Automatically generated by Pod::Man version $VERSION
406 .\\" @{[ scalar localtime ]}
408 .\\" Standard preamble:
409 .\\" ======================================================================
411 .\\" ======================================================================
413 .IX Title "$name $section"
414 .TH $name $section "$$self{release}" "$$self{date}" "$$self{center}"
416 ----END OF HEADER----
419 # Initialize a few per-file variables.
421 $$self{NEEDSPACE} = 0;
425 ############################################################################
427 ############################################################################
429 # Called for each command paragraph. Gets the command, the associated
430 # paragraph, the line number, and a Pod::Paragraph object. Just dispatches
431 # the command to a method named the same as the command. =cut is handled
432 # internally by Pod::Parser.
436 return if $command eq 'pod';
437 return if ($$self{EXCLUDE} && $command ne 'end');
438 $command = 'cmd_' . $command;
439 $self->$command (@_);
442 # Called for a verbatim paragraph. Gets the paragraph, the line number, and
443 # a Pod::Paragraph object. Rofficate backslashes, untabify, put a
444 # zero-width character at the beginning of each line to protect against
445 # commands, and wrap in .Vb/.Ve.
448 return if $$self{EXCLUDE};
452 my $lines = tr/\n/\n/;
453 1 while s/^(.*?)(\t+)/$1 . ' ' x (length ($2) * 8 - length ($1) % 8)/me;
455 s/^(\s*\S)/'\&' . $1/gme;
456 $self->makespace if $$self{NEEDSPACE};
457 $self->output (".Vb $lines\n$_.Ve\n");
458 $$self{NEEDSPACE} = 0;
461 # Called for a regular text block. Gets the paragraph, the line number, and
462 # a Pod::Paragraph object. Perform interpolation and output the results.
465 return if $$self{EXCLUDE};
466 $self->output ($_[0]), return if $$self{VERBATIM};
468 # Perform a little magic to collapse multiple L<> references. We'll
469 # just rewrite the whole thing into actual text at this part, bypassing
470 # the whole internal sequence parsing thing.
472 (L< # A link of the form L</something>.
475 [:\w]+ # The item has to be a simple word...
476 (\(\))? # ...or simple function.
480 ,?\s+(and\s+)? # Allow lots of them, conjuncted.
489 s{ L< / ([^>]+ ) } {$1}g;
490 my @items = split /(?:,?\s+(?:and\s+)?)/;
493 for ($i = 0; $i < @items; $i++) {
494 $string .= $items[$i];
495 $string .= ", " if @items > 2 && $i != $#items;
496 $string .= " and " if ($i == $#items - 1);
498 $string .= " entries elsewhere in this document";
502 # Parse the tree and output it. collapse knows about references to
503 # scalars as well as scalars and does the right thing with them.
504 local $_ = $self->parse (@_);
506 $self->makespace if $$self{NEEDSPACE};
507 $self->output (protect $self->mapfonts ($_));
509 $$self{NEEDSPACE} = 1;
512 # Called for an interior sequence. Takes a Pod::InteriorSequence object and
513 # returns a reference to a scalar. This scalar is the final formatted text.
514 # It's returned as a reference so that other interior sequences above us
515 # know that the text has already been processed.
517 my ($self, $seq) = @_;
518 my $command = $seq->cmd_name;
520 # Zero-width characters.
521 if ($command eq 'Z') { return bless \ '\&', 'Pod::Man::String' }
523 # C<>, L<>, X<>, and E<> don't apply guesswork to their contents.
524 local $_ = $self->collapse ($seq->parse_tree, $command =~ /^[CELX]$/);
526 # Handle E<> escapes.
527 if ($command eq 'E') {
529 return bless \ chr ($_), 'Pod::Man::String';
530 } elsif (exists $ESCAPES{$_}) {
531 return bless \ "$ESCAPES{$_}", 'Pod::Man::String';
533 carp "Unknown escape E<$1>";
534 return bless \ "E<$_>", 'Pod::Man::String';
538 # For all the other sequences, empty content produces no output.
539 return '' if $_ eq '';
541 # Handle formatting sequences.
542 if ($command eq 'B') {
543 return bless \ ('\f(BS' . $_ . '\f(BE'), 'Pod::Man::String';
544 } elsif ($command eq 'F') {
545 return bless \ ('\f(IS' . $_ . '\f(IE'), 'Pod::Man::String';
546 } elsif ($command eq 'I') {
547 return bless \ ('\f(IS' . $_ . '\f(IE'), 'Pod::Man::String';
548 } elsif ($command eq 'C') {
551 return bless \ ('\f(FS\*(C`' . $_ . "\\*(C'\\f(FE"),
556 if ($command eq 'L') {
557 return bless \ ($self->buildlink ($_)), 'Pod::Man::String';
560 # Whitespace protection replaces whitespace with "\ ".
561 if ($command eq 'S') {
563 return bless \ "$_", 'Pod::Man::String';
566 # Add an index entry to the list of ones waiting to be output.
567 if ($command eq 'X') { push (@{ $$self{INDEX} }, $_); return '' }
569 # Anything else is unknown.
570 carp "Unknown sequence $command<$_>";
574 ############################################################################
576 ############################################################################
578 # All command paragraphs take the paragraph and the line number.
580 # First level heading. We can't output .IX in the NAME section due to a bug
581 # in some versions of catman, so don't output a .IX for that section. .SH
582 # already uses small caps, so remove any E<> sequences that would cause
586 local $_ = $self->parse (@_);
589 $self->output (switchquotes ('.SH', $self->mapfonts ($_)));
590 $self->outindex (($_ eq 'NAME') ? () : ('Header', $_));
591 $$self{NEEDSPACE} = 0;
594 # Second level heading.
597 local $_ = $self->parse (@_);
599 $self->output (switchquotes ('.Sh', $self->mapfonts ($_)));
600 $self->outindex ('Subsection', $_);
601 $$self{NEEDSPACE} = 0;
604 # Start a list. For indents after the first, wrap the outside indent in .RS
605 # so that hanging paragraph tags will be correct.
609 unless (/^[-+]?\d+\s+$/) { $_ = $$self{indent} }
610 if (@{ $$self{INDENTS} } > 0) {
611 $self->output (".RS $$self{INDENT}\n");
613 push (@{ $$self{INDENTS} }, $$self{INDENT});
614 $$self{INDENT} = ($_ + 0);
617 # End a list. If we've closed an embedded indent, we've mangled the hanging
618 # paragraph indent, so temporarily replace it with .RS and set WEIRDINDENT.
619 # We'll close that .RS at the next =back or =item.
622 $$self{INDENT} = pop @{ $$self{INDENTS} };
623 unless (defined $$self{INDENT}) {
624 carp "Unmatched =back";
627 if ($$self{WEIRDINDENT}) {
628 $self->output (".RE\n");
629 $$self{WEIRDINDENT} = 0;
631 if (@{ $$self{INDENTS} } > 0) {
632 $self->output (".RE\n");
633 $self->output (".RS $$self{INDENT}\n");
634 $$self{WEIRDINDENT} = 1;
636 $$self{NEEDSPACE} = 1;
639 # An individual list item. Emit an index entry for anything that's
640 # interesting, but don't emit index entries for things like bullets and
641 # numbers. rofficate bullets too while we're at it (so for nice output, use
642 # * for your lists rather than o or . or - or some other thing).
645 local $_ = $self->parse (@_);
648 if (/\w/ && !/^\w[.\)]\s*$/) {
650 $index =~ s/^\s*[-*+o.]?\s*//;
652 s/^\*(\s|\Z)/\\\(bu$1/;
653 if ($$self{WEIRDINDENT}) {
654 $self->output (".RE\n");
655 $$self{WEIRDINDENT} = 0;
657 $_ = $self->mapfonts ($_);
658 $self->output (switchquotes ('.Ip', $_, $$self{INDENT}));
659 $self->outindex ($index ? ('Item', $index) : ());
660 $$self{NEEDSPACE} = 0;
663 # Begin a block for a particular translator. Setting VERBATIM triggers
664 # special handling in textblock().
668 my ($kind) = /^(\S+)/ or return;
669 if ($kind eq 'man' || $kind eq 'roff') {
670 $$self{VERBATIM} = 1;
676 # End a block for a particular translator. We assume that all =begin/=end
677 # pairs are properly closed.
681 $$self{VERBATIM} = 0;
684 # One paragraph for a particular translator. Ignore it unless it's intended
685 # for man or roff, in which case we output it verbatim.
690 return unless s/^(?:man|roff)\b[ \t]*\n?//;
695 ############################################################################
697 ############################################################################
699 # Handle links. We can't actually make real hyperlinks, so this is all to
700 # figure out what text and formatting we print out.
705 # Smash whitespace in case we were split across multiple lines.
708 # If we were given any explicit text, just output it.
709 if (m{ ^ ([^|]+) \| }x) { return $1 }
711 # Okay, leading and trailing whitespace isn't important.
715 # Default to using the whole content of the link entry as a section
716 # name. Note that L<manpage/> forces a manpage interpretation, as does
717 # something looking like L<manpage(section)>. Do the same thing to
718 # L<manpage(section)> as we would to manpage(section) without the L<>;
719 # see guesswork(). If we've added italics, don't add the "manpage"
720 # text; markup is sufficient.
721 my ($manpage, $section) = ('', $_);
722 if (/^"\s*(.*?)\s*"$/) {
723 $section = '"' . $1 . '"';
724 } elsif (m{ ^ [-:.\w]+ (?: \( \S+ \) )? $ }x) {
725 ($manpage, $section) = ($_, '');
726 $manpage =~ s/^([^\(]+)\(/'\f(IS' . $1 . '\f(IE\|('/e;
728 ($manpage, $section) = split (/\s*\/\s*/, $_, 2);
729 if ($manpage =~ /^[-:.\w]+(?:\(\S+\))?$/) {
730 $manpage =~ s/^([^\(]+)\(/'\f(IS' . $1 . '\f(IE\|'/e;
732 $section =~ s/^\"\s*//;
733 $section =~ s/\s*\"$//;
735 if ($manpage && $manpage !~ /\\f\(IS/) {
736 $manpage = "the $manpage manpage";
739 # Now build the actual output text.
741 if (!length ($section) && !length ($manpage)) {
742 carp "Invalid link $_";
743 } elsif (!length ($section)) {
745 } elsif ($section =~ /^[:\w]+(?:\(\))?/) {
746 $text .= 'the ' . $section . ' entry';
747 $text .= (length $manpage) ? " in $manpage"
748 : " elsewhere in this document";
750 if ($section !~ /^".*"$/) { $section = '"' . $section . '"' }
751 $text .= 'the section on ' . $section;
752 $text .= " in $manpage" if length $manpage;
758 ############################################################################
759 # Escaping and fontification
760 ############################################################################
762 # At this point, we'll have embedded font codes of the form \f(<font>[SE]
763 # where <font> is one of B, I, or F. Turn those into the right font start
764 # or end codes. B<someI<thing> else> should map to \fBsome\f(BIthing\fB
765 # else\fR. The old pod2man didn't get this right; the second \fB was \fR,
766 # so nested sequences didn't work right. We take care of this by using
767 # variables as a combined pointer to our current font sequence, and set each
768 # to the number of current nestings of start tags for that font. Use them
769 # as a vector to look up what font sequence to use.
774 my ($fixed, $bold, $italic) = (0, 0, 0);
775 my %magic = (F => \$fixed, B => \$bold, I => \$italic);
777 ${ $magic{$1} } += ($2 eq 'S') ? 1 : -1;
778 $$self{FONTS}{($fixed && 1) . ($bold && 1) . ($italic && 1)};
784 ############################################################################
785 # *roff-specific parsing
786 ############################################################################
788 # Called instead of parse_text, calls parse_text with the right flags.
791 $self->parse_text ({ -expand_seq => 'sequence',
792 -expand_ptree => 'collapse' }, @_);
795 # Takes a parse tree and a flag saying whether or not to treat it as literal
796 # text (not call guesswork on it), and returns the concatenation of all of
797 # the text strings in that parse tree. If the literal flag isn't true,
798 # guesswork() will be called on all plain scalars in the parse tree.
799 # Assumes that everything in the parse tree is either a scalar or a
800 # reference to a scalar.
802 my ($self, $ptree, $literal) = @_;
804 return join ('', map {
813 return join ('', map {
814 ref ($_) ? $$_ : $self->guesswork ($_)
819 # Takes a text block to perform guesswork on; this is guaranteed not to
820 # contain any interior sequences. Returns the text block with remapping
826 # rofficate backslashes.
829 # Ensure double underbars have a tiny space between them.
832 # Make all caps a little smaller. Be careful here, since we don't want
833 # to make @ARGV into small caps, nor do we want to fix the MIME in
834 # MIME-Version, since it looks weird with the full-height V.
836 ( ^ | [\s\(\"\'\`\[\{<>] )
837 ( [A-Z] [A-Z] [/A-Z+:\d_\$&-]* )
838 (?: (?= [\s>\}\]\)\'\".?!,;:] | -- ) | $ )
839 } { $1 . '\s-1' . $2 . '\s0' . $3 }egx;
841 # Turn PI into a pretty pi.
842 s{ (?: \\s-1 | \b ) PI (?: \\s0 | \b ) } {\\*\(PI}gx;
844 # Italize functions in the form func().
848 [:\w]+ (?:\\s-1)? \(\)
850 } { '\f(IS' . $1 . '\f(IE' }egx;
852 # func(n) is a reference to a manual page. Make it \fIfunc\fR\|(n).
855 (\w[-:.\w]+ (?:\\s-1)?)
859 } { '\f(IS' . $1 . '\f(IE\|' . $2 }egx;
861 # Convert simple Perl variable references to a fixed-width font.
866 } { $1 . '\f(FS' . $2 . '\f(FE'}egx;
868 # Translate -- into a real em dash if it's used like one and fix up
869 # dashes, but keep hyphens hyphens.
870 s{ (\G|^|.) (-+) (\b|.) } {
871 my ($pre, $dash, $post) = ($1, $2, $3);
872 if (length ($dash) == 1) {
873 ($pre =~ /[a-zA-Z]/) ? "$pre-$post" : "$pre\\-$post";
874 } elsif (length ($dash) == 2
875 && ((!$pre && !$post)
876 || ($pre =~ /\w/ && !$post)
877 || ($pre eq ' ' && $post eq ' ')
878 || ($pre eq '=' && $post ne '=')
879 || ($pre ne '=' && $post eq '='))) {
882 $pre . ('\-' x length $dash) . $post;
886 # Fix up double quotes.
887 s{ \" ([^\"]+) \" } { '\*(L"' . $1 . '\*(R"' }egx;
889 # Make C++ into \*(C+, which is a squinched version.
890 s{ \b C\+\+ } {\\*\(C+}gx;
897 ############################################################################
899 ############################################################################
901 # Make vertical whitespace.
904 $self->output ($$self{INDENT} > 0 ? ".Sp\n" : ".PP\n");
907 # Output any pending index entries, and optionally an index entry given as
908 # an argument. Support multiple index entries in X<> separated by slashes,
909 # and strip special escapes from index entries.
911 my ($self, $section, $index) = @_;
912 my @entries = map { split m%\s*/\s*% } @{ $$self{INDEX} };
913 return unless ($section || @entries);
917 my $output = '.IX Xref "'
918 . join (' ', map { s/\"/\"\"/; $_ } @entries)
922 $index =~ s/\"/\"\"/;
924 $index =~ s/\\(?:s-?\d|.\(..|.)//g;
925 $output .= ".IX $section " . '"' . $index . '"' . "\n";
927 $self->output ($output);
930 # Output text to the output device.
931 sub output { print { $_[0]->output_handle } $_[1] }
935 .\" These are some extra bits of roff that I don't want to lose track of
936 .\" but that have been removed from the preamble to make it a bit shorter
937 .\" since they're not currently being used. They're accents and special
938 .\" characters we don't currently have escapes for.
945 . ds ? \s-2c\h'-\w'c'u*7/10'\u\h'\*(#H'\zi\d\s+2\h'\w'c'u*8/10'
946 . ds ! \s-2\(or\s+2\h'-\w'\(or'u'\v'-.8m'.\v'.8m'
947 . ds q o\h'-\w'o'u*8/10'\s-4\v'.4m'\z\(*i\v'-.4m'\s+4\h'\w'o'u*8/10'
949 .ds v \\k:\h'-(\\n(.wu*9/10-\*(#H)'\v'-\*(#V'\*(#[\s-4v\s0\v'\*(#V'\h'|\\n:u'\*(#]
950 .ds _ \\k:\h'-(\\n(.wu*9/10-\*(#H+(\*(#F*2/3))'\v'-.4m'\z\(hy\v'.4m'\h'|\\n:u'
951 .ds . \\k:\h'-(\\n(.wu*8/10)'\v'\*(#V*4/10'\z.\v'-\*(#V*4/10'\h'|\\n:u'
952 .ds 3 \*(#[\v'.2m'\s-2\&3\s0\v'-.2m'\*(#]
953 .ds oe o\h'-(\w'o'u*4/10)'e
954 .ds Oe O\h'-(\w'O'u*4/10)'E
955 .if \n(.H>23 .if \n(.V>19 \
957 . ds v \h'-1'\o'\(aa\(ga'
965 ############################################################################
967 ############################################################################
971 Pod::Man - Convert POD data to formatted *roff input
976 my $parser = Pod::Man->new (release => $VERSION, section => 8);
978 # Read POD from STDIN and write to STDOUT.
979 $parser->parse_from_filehandle;
981 # Read POD from file.pod and write to file.1.
982 $parser->parse_from_file ('file.pod', 'file.1');
986 Pod::Man is a module to convert documentation in the POD format (the
987 preferred language for documenting Perl) into *roff input using the man
988 macro set. The resulting *roff code is suitable for display on a terminal
989 using nroff(1), normally via man(1), or printing using troff(1). It is
990 conventionally invoked using the driver script B<pod2roff>, but it can also
993 As a derived class from Pod::Parser, Pod::Man supports the same methods and
994 interfaces. See L<Pod::Parser> for all the details; briefly, one creates a
995 new parser with C<Pod::Man-E<gt>new()> and then calls either
996 parse_from_filehandle() or parse_from_file().
998 new() can take options, in the form of key/value pairs that control the
999 behavior of the parser. See below for details.
1001 If no options are given, Pod::Man uses the name of the input file with any
1002 trailing C<.pod>, C<.pm>, or C<.pl> stripped as the man page title, to
1003 section 1 unless the file ended in C<.pm> in which case it defaults to
1004 section 3, to a centered title of "User Contributed Perl Documentation", to
1005 a centered footer of the Perl version it is run with, and to a left-hand
1006 footer of the modification date of its input (or the current date if given
1009 Pod::Man assumes that your *roff formatters have a fixed-width font named
1010 CW. If yours is called something else (like CR), use the C<fixed> option to
1011 specify it. This generally only matters for troff output for printing.
1012 Similarly, you can set the fonts used for bold, italic, and bold italic
1015 Besides the obvious pod conversions, Pod::Man also takes care of formatting
1016 func(), func(n), and simple variable references like $foo or @bar so you
1017 don't have to use code escapes for them; complex expressions like
1018 C<$fred{'stuff'}> will still need to be escaped, though. It also translates
1019 dashes that aren't used as hyphens into en dashes, makes long dashes--like
1020 this--into proper em dashes, fixes "paired quotes," makes C++ and PI look
1021 right, puts a little space between double underbars, makes ALLCAPS a teeny
1022 bit smaller in troff(1), and escapes stuff that *roff treats as special so
1023 that you don't have to.
1025 The recognized options to new() are as follows. All options take a single
1032 Sets the centered page header to use instead of "User Contributed Perl
1037 Sets the left-hand footer. By default, the modification date of the input
1038 file will be used, or the current date if stat() can't find that file (the
1039 case if the input is from STDIN), and the date will be formatted as
1044 The fixed-width font to use for vertabim text and code. Defaults to CW.
1045 Some systems may want CR instead. Only matters for troff(1) output.
1049 Bold version of the fixed-width font. Defaults to CB. Only matters for
1054 Italic version of the fixed-width font (actually, something of a misnomer,
1055 since most fixed-width fonts only have an oblique version, not an italic
1056 version). Defaults to CI. Only matters for troff(1) output.
1058 =item fixedbolditalic
1060 Bold italic (probably actually oblique) version of the fixed-width font.
1061 Pod::Man doesn't assume you have this, and defaults to CB. Some systems
1062 (such as Solaris) have this font available as CX. Only matters for troff(1)
1067 Set the centered footer. By default, this is the version of Perl you run
1068 Pod::Man under. Note that some system an macro sets assume that the
1069 centered footer will be a modification date and will prepend something like
1070 "Last modified: "; if this is the case, you may want to set C<release> to
1071 the last modified date and C<date> to the version number.
1075 Set the section for the C<.TH> macro. The standard section numbering
1076 convention is to use 1 for user commands, 2 for system calls, 3 for
1077 functions, 4 for devices, 5 for file formats, 6 for games, 7 for
1078 miscellaneous information, and 8 for administrator commands. There is a lot
1079 of variation here, however; some systems (like Solaris) use 4 for file
1080 formats, 5 for miscellaneous information, and 7 for devices. Still others
1081 use 1m instead of 8, or some mix of both. About the only section numbers
1082 that are reliably consistent are 1, 2, and 3.
1084 By default, section 1 will be used unless the file ends in .pm in which case
1085 section 3 will be selected.
1089 The standard Pod::Parser method parse_from_filehandle() takes up to two
1090 arguments, the first being the file handle to read POD from and the second
1091 being the file handle to write the formatted output to. The first defaults
1092 to STDIN if not given, and the second defaults to STDOUT. The method
1093 parse_from_file() is almost identical, except that its two arguments are the
1094 input and output disk files instead. See L<Pod::Parser> for the specific
1101 =item roff font should be 1 or 2 chars, not `%s'
1103 (F) You specified a *roff font (using C<fixed>, C<fixedbold>, etc.) that
1104 wasn't either one or two characters. Pod::Man doesn't support *roff fonts
1105 longer than two characters, although some *roff extensions do (the canonical
1106 versions of nroff(1) and troff(1) don't either).
1108 =item Invalid link %s
1110 (W) The POD source contained a C<LE<lt>E<gt>> sequence that Pod::Man was
1111 unable to parse. You should never see this error message; it probably
1112 indicates a bug in Pod::Man.
1114 =item Unknown escape EE<lt>%sE<gt>
1116 (W) The POD source contained an C<EE<lt>E<gt>> escape that Pod::Man didn't
1117 know about. C<EE<lt>%sE<gt>> was printed verbatim in the output.
1119 =item Unknown sequence %s
1121 (W) The POD source contained a non-standard interior sequence (something of
1122 the form C<XE<lt>E<gt>>) that Pod::Man didn't know about. It was ignored.
1124 =item Unmatched =back
1126 (W) Pod::Man encountered a C<=back> command that didn't correspond to an
1133 The lint-like features and strict POD format checking done by B<pod2man> are
1134 not yet implemented and should be, along with the corresponding C<lax>
1137 The NAME section should be recognized specially and index entries emitted
1138 for everything in that section. This would have to be deferred until the
1139 next section, since extraneous things in NAME tends to confuse various man
1142 The handling of hyphens, en dashes, and em dashes is somewhat fragile, and
1143 one may get the wrong one under some circumstances. This should only matter
1144 for troff(1) output.
1146 When and whether to use small caps is somewhat tricky, and Pod::Man doesn't
1147 necessarily get it right.
1149 Pod::Man doesn't handle font names longer than two characters. Neither do
1150 most troff(1) implementations, but GNU troff does as an extension. It would
1151 be nice to support as an option for those who want to use it.
1153 The preamble added to each output file is rather verbose, and most of it is
1154 only necessary in the presence of EE<lt>E<gt> escapes for non-ASCII
1155 characters. It would ideally be nice if all of those definitions were only
1156 output if needed, perhaps on the fly as the characters are used.
1158 Some of the automagic applied to file names assumes Unix directory
1161 Pod::Man is excessively slow.
1165 The intention is for this module and its driver script to eventually replace
1166 B<pod2man> in Perl core.
1170 L<Pod::Parser|Pod::Parser>, perlpod(1), pod2roff(1), nroff(1), troff(1),
1173 Ossanna, Joseph F., and Brian W. Kernighan. "Troff User's Manual,"
1174 Computing Science Technical Report No. 54, AT&T Bell Laboratories. This is
1175 the best documentation of standard nroff(1) and troff(1). At the time of
1176 this writing, it's available at http://www.cs.bell-labs.com/cm/cs/cstr.html.
1178 The man page documenting the man macro set may be man(5) instead of man(7)
1179 on your system. Also, please see pod2roff(1) for extensive documentation on
1180 writing manual pages if you've not done it before and aren't familiar with
1185 Russ Allbery E<lt>rra@stanford.eduE<gt>, based I<very> heavily on the
1186 original B<pod2man> by Tom Christiansen E<lt>tchrist@mox.perl.comE<gt>.