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 ($rev, $ver, $sver) = ($] =~ /^(\d+)\.(\d{3})(\d{0,3})$/);
342 $sver ||= 0; $sver *= 10 ** (3-length($sver));
343 $rev += 0; $ver += 0; $sver += 0;
344 $$self{release} = "perl v$rev.$ver.$sver";
347 # Double quotes in things that will be quoted.
348 for (qw/center date release/) { $$self{$_} =~ s/\"/\"\"/g }
350 $$self{INDENT} = 0; # Current indentation level.
351 $$self{INDENTS} = []; # Stack of indentations.
352 $$self{INDEX} = []; # Index keys waiting to be printed.
354 $self->SUPER::initialize;
357 # For each document we process, output the preamble first. Note that the
358 # fixed width font is a global default; once we interpolate it into the
359 # PREAMBLE, it ain't ever changing. Maybe fix this later.
363 # Try to figure out the name and section from the file name.
364 my $section = $$self{section} || 1;
365 my $name = $$self{name};
366 if (!defined $name) {
367 $name = $self->input_file;
368 $section = 3 if (!$$self{section} && $name =~ /\.pm$/i);
369 $name =~ s/\.p(od|[lm])$//i;
370 if ($section =~ /^1/) {
371 require File::Basename;
372 $name = uc File::Basename::basename ($name);
374 # Lose everything up to the first of
375 # */lib/*perl* standard or site_perl module
376 # */*perl*/lib from -D prefix=/opt/perl
377 # */*perl*/ random module hierarchy
378 # which works. Should be fixed to use File::Spec.
381 if ( s%^.*?/lib/[^/]*perl[^/]*/%%i
382 or s%^.*?/[^/]*perl[^/]*/(?:lib/)?%%i) {
383 s%^site(_perl)?/%%; # site and site_perl
384 s%^(.*-$^O|$^O-.*)/%%o; # arch
385 s%^\d+\.\d+%%; # version
392 # Modification date header. Try to use the modification time of our
394 if (!defined $$self{date}) {
395 my $time = (stat $self->input_file)[9] || time;
396 my ($day, $month, $year) = (localtime $time)[3,4,5];
399 $$self{date} = join ('-', $year, $month, $day);
402 # Now, print out the preamble and the title.
403 $PREAMBLE =~ s/\@CFONT\@/$$self{fixed}/;
405 print { $self->output_handle } <<"----END OF HEADER----";
406 .\\" Automatically generated by Pod::Man version $VERSION
407 .\\" @{[ scalar localtime ]}
409 .\\" Standard preamble:
410 .\\" ======================================================================
412 .\\" ======================================================================
414 .IX Title "$name $section"
415 .TH $name $section "$$self{release}" "$$self{date}" "$$self{center}"
417 ----END OF HEADER----
420 # Initialize a few per-file variables.
422 $$self{NEEDSPACE} = 0;
426 ############################################################################
428 ############################################################################
430 # Called for each command paragraph. Gets the command, the associated
431 # paragraph, the line number, and a Pod::Paragraph object. Just dispatches
432 # the command to a method named the same as the command. =cut is handled
433 # internally by Pod::Parser.
437 return if $command eq 'pod';
438 return if ($$self{EXCLUDE} && $command ne 'end');
439 $command = 'cmd_' . $command;
440 $self->$command (@_);
443 # Called for a verbatim paragraph. Gets the paragraph, the line number, and
444 # a Pod::Paragraph object. Rofficate backslashes, untabify, put a
445 # zero-width character at the beginning of each line to protect against
446 # commands, and wrap in .Vb/.Ve.
449 return if $$self{EXCLUDE};
453 my $lines = tr/\n/\n/;
454 1 while s/^(.*?)(\t+)/$1 . ' ' x (length ($2) * 8 - length ($1) % 8)/me;
456 s/^(\s*\S)/'\&' . $1/gme;
457 $self->makespace if $$self{NEEDSPACE};
458 $self->output (".Vb $lines\n$_.Ve\n");
459 $$self{NEEDSPACE} = 0;
462 # Called for a regular text block. Gets the paragraph, the line number, and
463 # a Pod::Paragraph object. Perform interpolation and output the results.
466 return if $$self{EXCLUDE};
467 $self->output ($_[0]), return if $$self{VERBATIM};
469 # Perform a little magic to collapse multiple L<> references. We'll
470 # just rewrite the whole thing into actual text at this part, bypassing
471 # the whole internal sequence parsing thing.
473 (L< # A link of the form L</something>.
476 [:\w]+ # The item has to be a simple word...
477 (\(\))? # ...or simple function.
481 ,?\s+(and\s+)? # Allow lots of them, conjuncted.
490 s{ L< / ([^>]+ ) } {$1}g;
491 my @items = split /(?:,?\s+(?:and\s+)?)/;
494 for ($i = 0; $i < @items; $i++) {
495 $string .= $items[$i];
496 $string .= ", " if @items > 2 && $i != $#items;
497 $string .= " and " if ($i == $#items - 1);
499 $string .= " entries elsewhere in this document";
503 # Parse the tree and output it. collapse knows about references to
504 # scalars as well as scalars and does the right thing with them.
505 local $_ = $self->parse (@_);
507 $self->makespace if $$self{NEEDSPACE};
508 $self->output (protect $self->mapfonts ($_));
510 $$self{NEEDSPACE} = 1;
513 # Called for an interior sequence. Takes a Pod::InteriorSequence object and
514 # returns a reference to a scalar. This scalar is the final formatted text.
515 # It's returned as a reference so that other interior sequences above us
516 # know that the text has already been processed.
518 my ($self, $seq) = @_;
519 my $command = $seq->cmd_name;
521 # Zero-width characters.
522 if ($command eq 'Z') {
523 my $v = '\&'; return bless \ $v, 'Pod::Man::String';
526 # C<>, L<>, X<>, and E<> don't apply guesswork to their contents.
527 local $_ = $self->collapse ($seq->parse_tree, $command =~ /^[CELX]$/);
529 # Handle E<> escapes.
530 if ($command eq 'E') {
532 return bless \ chr ($_), 'Pod::Man::String';
533 } elsif (exists $ESCAPES{$_}) {
534 return bless \ "$ESCAPES{$_}", 'Pod::Man::String';
536 carp "Unknown escape E<$1>";
537 return bless \ "E<$_>", 'Pod::Man::String';
541 # For all the other sequences, empty content produces no output.
542 return '' if $_ eq '';
544 # Handle formatting sequences.
545 if ($command eq 'B') {
546 return bless \ ('\f(BS' . $_ . '\f(BE'), 'Pod::Man::String';
547 } elsif ($command eq 'F') {
548 return bless \ ('\f(IS' . $_ . '\f(IE'), 'Pod::Man::String';
549 } elsif ($command eq 'I') {
550 return bless \ ('\f(IS' . $_ . '\f(IE'), 'Pod::Man::String';
551 } elsif ($command eq 'C') {
554 return bless \ ('\f(FS\*(C`' . $_ . "\\*(C'\\f(FE"),
559 if ($command eq 'L') {
560 # XXX bug in lvalue subroutines prevents this from working
561 #return bless \ ($self->buildlink ($_)), 'Pod::Man::String';
562 my $v = $self->buildlink($_);
563 return bless \$v, 'Pod::Man::String';
566 # Whitespace protection replaces whitespace with "\ ".
567 if ($command eq 'S') {
569 return bless \ "$_", 'Pod::Man::String';
572 # Add an index entry to the list of ones waiting to be output.
573 if ($command eq 'X') { push (@{ $$self{INDEX} }, $_); return '' }
575 # Anything else is unknown.
576 carp "Unknown sequence $command<$_>";
580 ############################################################################
582 ############################################################################
584 # All command paragraphs take the paragraph and the line number.
586 # First level heading. We can't output .IX in the NAME section due to a bug
587 # in some versions of catman, so don't output a .IX for that section. .SH
588 # already uses small caps, so remove any E<> sequences that would cause
592 local $_ = $self->parse (@_);
595 $self->output (switchquotes ('.SH', $self->mapfonts ($_)));
596 $self->outindex (($_ eq 'NAME') ? () : ('Header', $_));
597 $$self{NEEDSPACE} = 0;
600 # Second level heading.
603 local $_ = $self->parse (@_);
605 $self->output (switchquotes ('.Sh', $self->mapfonts ($_)));
606 $self->outindex ('Subsection', $_);
607 $$self{NEEDSPACE} = 0;
610 # Start a list. For indents after the first, wrap the outside indent in .RS
611 # so that hanging paragraph tags will be correct.
615 unless (/^[-+]?\d+\s+$/) { $_ = $$self{indent} }
616 if (@{ $$self{INDENTS} } > 0) {
617 $self->output (".RS $$self{INDENT}\n");
619 push (@{ $$self{INDENTS} }, $$self{INDENT});
620 $$self{INDENT} = ($_ + 0);
623 # End a list. If we've closed an embedded indent, we've mangled the hanging
624 # paragraph indent, so temporarily replace it with .RS and set WEIRDINDENT.
625 # We'll close that .RS at the next =back or =item.
628 $$self{INDENT} = pop @{ $$self{INDENTS} };
629 unless (defined $$self{INDENT}) {
630 carp "Unmatched =back";
633 if ($$self{WEIRDINDENT}) {
634 $self->output (".RE\n");
635 $$self{WEIRDINDENT} = 0;
637 if (@{ $$self{INDENTS} } > 0) {
638 $self->output (".RE\n");
639 $self->output (".RS $$self{INDENT}\n");
640 $$self{WEIRDINDENT} = 1;
642 $$self{NEEDSPACE} = 1;
645 # An individual list item. Emit an index entry for anything that's
646 # interesting, but don't emit index entries for things like bullets and
647 # numbers. rofficate bullets too while we're at it (so for nice output, use
648 # * for your lists rather than o or . or - or some other thing).
651 local $_ = $self->parse (@_);
654 if (/\w/ && !/^\w[.\)]\s*$/) {
656 $index =~ s/^\s*[-*+o.]?\s*//;
658 s/^\*(\s|\Z)/\\\(bu$1/;
659 if ($$self{WEIRDINDENT}) {
660 $self->output (".RE\n");
661 $$self{WEIRDINDENT} = 0;
663 $_ = $self->mapfonts ($_);
664 $self->output (switchquotes ('.Ip', $_, $$self{INDENT}));
665 $self->outindex ($index ? ('Item', $index) : ());
666 $$self{NEEDSPACE} = 0;
669 # Begin a block for a particular translator. Setting VERBATIM triggers
670 # special handling in textblock().
674 my ($kind) = /^(\S+)/ or return;
675 if ($kind eq 'man' || $kind eq 'roff') {
676 $$self{VERBATIM} = 1;
682 # End a block for a particular translator. We assume that all =begin/=end
683 # pairs are properly closed.
687 $$self{VERBATIM} = 0;
690 # One paragraph for a particular translator. Ignore it unless it's intended
691 # for man or roff, in which case we output it verbatim.
696 return unless s/^(?:man|roff)\b[ \t]*\n?//;
701 ############################################################################
703 ############################################################################
705 # Handle links. We can't actually make real hyperlinks, so this is all to
706 # figure out what text and formatting we print out.
711 # Smash whitespace in case we were split across multiple lines.
714 # If we were given any explicit text, just output it.
715 if (m{ ^ ([^|]+) \| }x) { return $1 }
717 # Okay, leading and trailing whitespace isn't important.
721 # Default to using the whole content of the link entry as a section
722 # name. Note that L<manpage/> forces a manpage interpretation, as does
723 # something looking like L<manpage(section)>. Do the same thing to
724 # L<manpage(section)> as we would to manpage(section) without the L<>;
725 # see guesswork(). If we've added italics, don't add the "manpage"
726 # text; markup is sufficient.
727 my ($manpage, $section) = ('', $_);
728 if (/^"\s*(.*?)\s*"$/) {
729 $section = '"' . $1 . '"';
730 } elsif (m{ ^ [-:.\w]+ (?: \( \S+ \) )? $ }x) {
731 ($manpage, $section) = ($_, '');
732 $manpage =~ s/^([^\(]+)\(/'\f(IS' . $1 . '\f(IE\|('/e;
734 ($manpage, $section) = split (/\s*\/\s*/, $_, 2);
735 if ($manpage =~ /^[-:.\w]+(?:\(\S+\))?$/) {
736 $manpage =~ s/^([^\(]+)\(/'\f(IS' . $1 . '\f(IE\|'/e;
738 $section =~ s/^\"\s*//;
739 $section =~ s/\s*\"$//;
741 if ($manpage && $manpage !~ /\\f\(IS/) {
742 $manpage = "the $manpage manpage";
745 # Now build the actual output text.
747 if (!length ($section) && !length ($manpage)) {
748 carp "Invalid link $_";
749 } elsif (!length ($section)) {
751 } elsif ($section =~ /^[:\w]+(?:\(\))?/) {
752 $text .= 'the ' . $section . ' entry';
753 $text .= (length $manpage) ? " in $manpage"
754 : " elsewhere in this document";
756 if ($section !~ /^".*"$/) { $section = '"' . $section . '"' }
757 $text .= 'the section on ' . $section;
758 $text .= " in $manpage" if length $manpage;
764 ############################################################################
765 # Escaping and fontification
766 ############################################################################
768 # At this point, we'll have embedded font codes of the form \f(<font>[SE]
769 # where <font> is one of B, I, or F. Turn those into the right font start
770 # or end codes. B<someI<thing> else> should map to \fBsome\f(BIthing\fB
771 # else\fR. The old pod2man didn't get this right; the second \fB was \fR,
772 # so nested sequences didn't work right. We take care of this by using
773 # variables as a combined pointer to our current font sequence, and set each
774 # to the number of current nestings of start tags for that font. Use them
775 # as a vector to look up what font sequence to use.
780 my ($fixed, $bold, $italic) = (0, 0, 0);
781 my %magic = (F => \$fixed, B => \$bold, I => \$italic);
783 ${ $magic{$1} } += ($2 eq 'S') ? 1 : -1;
784 $$self{FONTS}{($fixed && 1) . ($bold && 1) . ($italic && 1)};
790 ############################################################################
791 # *roff-specific parsing
792 ############################################################################
794 # Called instead of parse_text, calls parse_text with the right flags.
797 $self->parse_text ({ -expand_seq => 'sequence',
798 -expand_ptree => 'collapse' }, @_);
801 # Takes a parse tree and a flag saying whether or not to treat it as literal
802 # text (not call guesswork on it), and returns the concatenation of all of
803 # the text strings in that parse tree. If the literal flag isn't true,
804 # guesswork() will be called on all plain scalars in the parse tree.
805 # Assumes that everything in the parse tree is either a scalar or a
806 # reference to a scalar.
808 my ($self, $ptree, $literal) = @_;
810 return join ('', map {
819 return join ('', map {
820 ref ($_) ? $$_ : $self->guesswork ($_)
825 # Takes a text block to perform guesswork on; this is guaranteed not to
826 # contain any interior sequences. Returns the text block with remapping
832 # rofficate backslashes.
835 # Ensure double underbars have a tiny space between them.
838 # Make all caps a little smaller. Be careful here, since we don't want
839 # to make @ARGV into small caps, nor do we want to fix the MIME in
840 # MIME-Version, since it looks weird with the full-height V.
842 ( ^ | [\s\(\"\'\`\[\{<>] )
843 ( [A-Z] [A-Z] [/A-Z+:\d_\$&-]* )
844 (?: (?= [\s>\}\]\)\'\".?!,;:] | -- ) | $ )
845 } { $1 . '\s-1' . $2 . '\s0' . $3 }egx;
847 # Turn PI into a pretty pi.
848 s{ (?: \\s-1 | \b ) PI (?: \\s0 | \b ) } {\\*\(PI}gx;
850 # Italize functions in the form func().
854 [:\w]+ (?:\\s-1)? \(\)
856 } { '\f(IS' . $1 . '\f(IE' }egx;
858 # func(n) is a reference to a manual page. Make it \fIfunc\fR\|(n).
861 (\w[-:.\w]+ (?:\\s-1)?)
865 } { '\f(IS' . $1 . '\f(IE\|' . $2 }egx;
867 # Convert simple Perl variable references to a fixed-width font.
872 } { $1 . '\f(FS' . $2 . '\f(FE'}egx;
874 # Translate -- into a real em dash if it's used like one and fix up
875 # dashes, but keep hyphens hyphens.
876 s{ (\G|^|.) (-+) (\b|.) } {
877 my ($pre, $dash, $post) = ($1, $2, $3);
878 if (length ($dash) == 1) {
879 ($pre =~ /[a-zA-Z]/) ? "$pre-$post" : "$pre\\-$post";
880 } elsif (length ($dash) == 2
881 && ((!$pre && !$post)
882 || ($pre =~ /\w/ && !$post)
883 || ($pre eq ' ' && $post eq ' ')
884 || ($pre eq '=' && $post ne '=')
885 || ($pre ne '=' && $post eq '='))) {
888 $pre . ('\-' x length $dash) . $post;
892 # Fix up double quotes.
893 s{ \" ([^\"]+) \" } { '\*(L"' . $1 . '\*(R"' }egx;
895 # Make C++ into \*(C+, which is a squinched version.
896 s{ \b C\+\+ } {\\*\(C+}gx;
903 ############################################################################
905 ############################################################################
907 # Make vertical whitespace.
910 $self->output ($$self{INDENT} > 0 ? ".Sp\n" : ".PP\n");
913 # Output any pending index entries, and optionally an index entry given as
914 # an argument. Support multiple index entries in X<> separated by slashes,
915 # and strip special escapes from index entries.
917 my ($self, $section, $index) = @_;
918 my @entries = map { split m%\s*/\s*% } @{ $$self{INDEX} };
919 return unless ($section || @entries);
923 my $output = '.IX Xref "'
924 . join (' ', map { s/\"/\"\"/; $_ } @entries)
928 $index =~ s/\"/\"\"/;
930 $index =~ s/\\(?:s-?\d|.\(..|.)//g;
931 $output .= ".IX $section " . '"' . $index . '"' . "\n";
933 $self->output ($output);
936 # Output text to the output device.
937 sub output { print { $_[0]->output_handle } $_[1] }
941 .\" These are some extra bits of roff that I don't want to lose track of
942 .\" but that have been removed from the preamble to make it a bit shorter
943 .\" since they're not currently being used. They're accents and special
944 .\" characters we don't currently have escapes for.
951 . ds ? \s-2c\h'-\w'c'u*7/10'\u\h'\*(#H'\zi\d\s+2\h'\w'c'u*8/10'
952 . ds ! \s-2\(or\s+2\h'-\w'\(or'u'\v'-.8m'.\v'.8m'
953 . 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'
955 .ds v \\k:\h'-(\\n(.wu*9/10-\*(#H)'\v'-\*(#V'\*(#[\s-4v\s0\v'\*(#V'\h'|\\n:u'\*(#]
956 .ds _ \\k:\h'-(\\n(.wu*9/10-\*(#H+(\*(#F*2/3))'\v'-.4m'\z\(hy\v'.4m'\h'|\\n:u'
957 .ds . \\k:\h'-(\\n(.wu*8/10)'\v'\*(#V*4/10'\z.\v'-\*(#V*4/10'\h'|\\n:u'
958 .ds 3 \*(#[\v'.2m'\s-2\&3\s0\v'-.2m'\*(#]
959 .ds oe o\h'-(\w'o'u*4/10)'e
960 .ds Oe O\h'-(\w'O'u*4/10)'E
961 .if \n(.H>23 .if \n(.V>19 \
963 . ds v \h'-1'\o'\(aa\(ga'
971 ############################################################################
973 ############################################################################
977 Pod::Man - Convert POD data to formatted *roff input
982 my $parser = Pod::Man->new (release => $VERSION, section => 8);
984 # Read POD from STDIN and write to STDOUT.
985 $parser->parse_from_filehandle;
987 # Read POD from file.pod and write to file.1.
988 $parser->parse_from_file ('file.pod', 'file.1');
992 Pod::Man is a module to convert documentation in the POD format (the
993 preferred language for documenting Perl) into *roff input using the man
994 macro set. The resulting *roff code is suitable for display on a terminal
995 using nroff(1), normally via man(1), or printing using troff(1). It is
996 conventionally invoked using the driver script B<pod2man>, but it can also
999 As a derived class from Pod::Parser, Pod::Man supports the same methods and
1000 interfaces. See L<Pod::Parser> for all the details; briefly, one creates a
1001 new parser with C<Pod::Man-E<gt>new()> and then calls either
1002 parse_from_filehandle() or parse_from_file().
1004 new() can take options, in the form of key/value pairs that control the
1005 behavior of the parser. See below for details.
1007 If no options are given, Pod::Man uses the name of the input file with any
1008 trailing C<.pod>, C<.pm>, or C<.pl> stripped as the man page title, to
1009 section 1 unless the file ended in C<.pm> in which case it defaults to
1010 section 3, to a centered title of "User Contributed Perl Documentation", to
1011 a centered footer of the Perl version it is run with, and to a left-hand
1012 footer of the modification date of its input (or the current date if given
1015 Pod::Man assumes that your *roff formatters have a fixed-width font named
1016 CW. If yours is called something else (like CR), use the C<fixed> option to
1017 specify it. This generally only matters for troff output for printing.
1018 Similarly, you can set the fonts used for bold, italic, and bold italic
1021 Besides the obvious pod conversions, Pod::Man also takes care of formatting
1022 func(), func(n), and simple variable references like $foo or @bar so you
1023 don't have to use code escapes for them; complex expressions like
1024 C<$fred{'stuff'}> will still need to be escaped, though. It also translates
1025 dashes that aren't used as hyphens into en dashes, makes long dashes--like
1026 this--into proper em dashes, fixes "paired quotes," makes C++ and PI look
1027 right, puts a little space between double underbars, makes ALLCAPS a teeny
1028 bit smaller in troff(1), and escapes stuff that *roff treats as special so
1029 that you don't have to.
1031 The recognized options to new() are as follows. All options take a single
1038 Sets the centered page header to use instead of "User Contributed Perl
1043 Sets the left-hand footer. By default, the modification date of the input
1044 file will be used, or the current date if stat() can't find that file (the
1045 case if the input is from STDIN), and the date will be formatted as
1050 The fixed-width font to use for vertabim text and code. Defaults to CW.
1051 Some systems may want CR instead. Only matters for troff(1) output.
1055 Bold version of the fixed-width font. Defaults to CB. Only matters for
1060 Italic version of the fixed-width font (actually, something of a misnomer,
1061 since most fixed-width fonts only have an oblique version, not an italic
1062 version). Defaults to CI. Only matters for troff(1) output.
1064 =item fixedbolditalic
1066 Bold italic (probably actually oblique) version of the fixed-width font.
1067 Pod::Man doesn't assume you have this, and defaults to CB. Some systems
1068 (such as Solaris) have this font available as CX. Only matters for troff(1)
1073 Set the centered footer. By default, this is the version of Perl you run
1074 Pod::Man under. Note that some system an macro sets assume that the
1075 centered footer will be a modification date and will prepend something like
1076 "Last modified: "; if this is the case, you may want to set C<release> to
1077 the last modified date and C<date> to the version number.
1081 Set the section for the C<.TH> macro. The standard section numbering
1082 convention is to use 1 for user commands, 2 for system calls, 3 for
1083 functions, 4 for devices, 5 for file formats, 6 for games, 7 for
1084 miscellaneous information, and 8 for administrator commands. There is a lot
1085 of variation here, however; some systems (like Solaris) use 4 for file
1086 formats, 5 for miscellaneous information, and 7 for devices. Still others
1087 use 1m instead of 8, or some mix of both. About the only section numbers
1088 that are reliably consistent are 1, 2, and 3.
1090 By default, section 1 will be used unless the file ends in .pm in which case
1091 section 3 will be selected.
1095 The standard Pod::Parser method parse_from_filehandle() takes up to two
1096 arguments, the first being the file handle to read POD from and the second
1097 being the file handle to write the formatted output to. The first defaults
1098 to STDIN if not given, and the second defaults to STDOUT. The method
1099 parse_from_file() is almost identical, except that its two arguments are the
1100 input and output disk files instead. See L<Pod::Parser> for the specific
1107 =item roff font should be 1 or 2 chars, not `%s'
1109 (F) You specified a *roff font (using C<fixed>, C<fixedbold>, etc.) that
1110 wasn't either one or two characters. Pod::Man doesn't support *roff fonts
1111 longer than two characters, although some *roff extensions do (the canonical
1112 versions of nroff(1) and troff(1) don't either).
1114 =item Invalid link %s
1116 (W) The POD source contained a C<LE<lt>E<gt>> sequence that Pod::Man was
1117 unable to parse. You should never see this error message; it probably
1118 indicates a bug in Pod::Man.
1120 =item Unknown escape EE<lt>%sE<gt>
1122 (W) The POD source contained an C<EE<lt>E<gt>> escape that Pod::Man didn't
1123 know about. C<EE<lt>%sE<gt>> was printed verbatim in the output.
1125 =item Unknown sequence %s
1127 (W) The POD source contained a non-standard interior sequence (something of
1128 the form C<XE<lt>E<gt>>) that Pod::Man didn't know about. It was ignored.
1130 =item Unmatched =back
1132 (W) Pod::Man encountered a C<=back> command that didn't correspond to an
1139 The lint-like features and strict POD format checking done by B<pod2man> are
1140 not yet implemented and should be, along with the corresponding C<lax>
1143 The NAME section should be recognized specially and index entries emitted
1144 for everything in that section. This would have to be deferred until the
1145 next section, since extraneous things in NAME tends to confuse various man
1148 The handling of hyphens, en dashes, and em dashes is somewhat fragile, and
1149 one may get the wrong one under some circumstances. This should only matter
1150 for troff(1) output.
1152 When and whether to use small caps is somewhat tricky, and Pod::Man doesn't
1153 necessarily get it right.
1155 Pod::Man doesn't handle font names longer than two characters. Neither do
1156 most troff(1) implementations, but GNU troff does as an extension. It would
1157 be nice to support as an option for those who want to use it.
1159 The preamble added to each output file is rather verbose, and most of it is
1160 only necessary in the presence of EE<lt>E<gt> escapes for non-ASCII
1161 characters. It would ideally be nice if all of those definitions were only
1162 output if needed, perhaps on the fly as the characters are used.
1164 Some of the automagic applied to file names assumes Unix directory
1167 Pod::Man is excessively slow.
1171 The intention is for this module and its driver script to eventually replace
1172 B<pod2man> in Perl core.
1176 L<Pod::Parser|Pod::Parser>, perlpod(1), pod2man(1), nroff(1), troff(1),
1179 Ossanna, Joseph F., and Brian W. Kernighan. "Troff User's Manual,"
1180 Computing Science Technical Report No. 54, AT&T Bell Laboratories. This is
1181 the best documentation of standard nroff(1) and troff(1). At the time of
1182 this writing, it's available at http://www.cs.bell-labs.com/cm/cs/cstr.html.
1184 The man page documenting the man macro set may be man(5) instead of man(7)
1185 on your system. Also, please see pod2man(1) for extensive documentation on
1186 writing manual pages if you've not done it before and aren't familiar with
1191 Russ Allbery E<lt>rra@stanford.eduE<gt>, based I<very> heavily on the
1192 original B<pod2man> by Tom Christiansen E<lt>tchrist@mox.perl.comE<gt>.