1 # Pod::Text -- Convert POD data to formatted ASCII text.
2 # $Id: Text.pm,v 2.6 2000/10/10 02:13:17 eagle Exp $
4 # Copyright 1999, 2000 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 Pod::Text, 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);
27 use vars qw(@ISA @EXPORT %ESCAPES $VERSION);
29 # We inherit from Pod::Select instead of Pod::Parser so that we can be used
31 @ISA = qw(Pod::Select Exporter);
33 # We have to export pod2text for backward compatibility.
34 @EXPORT = qw(pod2text);
36 # Don't use the CVS revision as the version, since this module is also in
37 # Perl core and too many things could munge CVS magic revision strings.
38 # This number should ideally be the same as the CVS revision in podlators,
43 ############################################################################
44 # Table of supported E<> escapes
45 ############################################################################
47 # This table is taken near verbatim from Pod::PlainText in Pod::Parser,
48 # which got it near verbatim from the original Pod::Text. It is therefore
49 # credited to Tom Christiansen, and I'm glad I didn't have to write it. :)
50 # "iexcl" to "divide" added by Tim Jenness.
52 'amp' => '&', # ampersand
53 'lt' => '<', # left chevron, less-than
54 'gt' => '>', # right chevron, greater-than
55 'quot' => '"', # double quote
56 'sol' => '/', # solidus (forward slash)
57 'verbar' => '|', # vertical bar
59 "Aacute" => "\xC1", # capital A, acute accent
60 "aacute" => "\xE1", # small a, acute accent
61 "Acirc" => "\xC2", # capital A, circumflex accent
62 "acirc" => "\xE2", # small a, circumflex accent
63 "AElig" => "\xC6", # capital AE diphthong (ligature)
64 "aelig" => "\xE6", # small ae diphthong (ligature)
65 "Agrave" => "\xC0", # capital A, grave accent
66 "agrave" => "\xE0", # small a, grave accent
67 "Aring" => "\xC5", # capital A, ring
68 "aring" => "\xE5", # small a, ring
69 "Atilde" => "\xC3", # capital A, tilde
70 "atilde" => "\xE3", # small a, tilde
71 "Auml" => "\xC4", # capital A, dieresis or umlaut mark
72 "auml" => "\xE4", # small a, dieresis or umlaut mark
73 "Ccedil" => "\xC7", # capital C, cedilla
74 "ccedil" => "\xE7", # small c, cedilla
75 "Eacute" => "\xC9", # capital E, acute accent
76 "eacute" => "\xE9", # small e, acute accent
77 "Ecirc" => "\xCA", # capital E, circumflex accent
78 "ecirc" => "\xEA", # small e, circumflex accent
79 "Egrave" => "\xC8", # capital E, grave accent
80 "egrave" => "\xE8", # small e, grave accent
81 "ETH" => "\xD0", # capital Eth, Icelandic
82 "eth" => "\xF0", # small eth, Icelandic
83 "Euml" => "\xCB", # capital E, dieresis or umlaut mark
84 "euml" => "\xEB", # small e, dieresis or umlaut mark
85 "Iacute" => "\xCC", # capital I, acute accent
86 "iacute" => "\xEC", # small i, acute accent
87 "Icirc" => "\xCE", # capital I, circumflex accent
88 "icirc" => "\xEE", # small i, circumflex accent
89 "Igrave" => "\xCD", # capital I, grave accent
90 "igrave" => "\xED", # small i, grave accent
91 "Iuml" => "\xCF", # capital I, dieresis or umlaut mark
92 "iuml" => "\xEF", # small i, dieresis or umlaut mark
93 "Ntilde" => "\xD1", # capital N, tilde
94 "ntilde" => "\xF1", # small n, tilde
95 "Oacute" => "\xD3", # capital O, acute accent
96 "oacute" => "\xF3", # small o, acute accent
97 "Ocirc" => "\xD4", # capital O, circumflex accent
98 "ocirc" => "\xF4", # small o, circumflex accent
99 "Ograve" => "\xD2", # capital O, grave accent
100 "ograve" => "\xF2", # small o, grave accent
101 "Oslash" => "\xD8", # capital O, slash
102 "oslash" => "\xF8", # small o, slash
103 "Otilde" => "\xD5", # capital O, tilde
104 "otilde" => "\xF5", # small o, tilde
105 "Ouml" => "\xD6", # capital O, dieresis or umlaut mark
106 "ouml" => "\xF6", # small o, dieresis or umlaut mark
107 "szlig" => "\xDF", # small sharp s, German (sz ligature)
108 "THORN" => "\xDE", # capital THORN, Icelandic
109 "thorn" => "\xFE", # small thorn, Icelandic
110 "Uacute" => "\xDA", # capital U, acute accent
111 "uacute" => "\xFA", # small u, acute accent
112 "Ucirc" => "\xDB", # capital U, circumflex accent
113 "ucirc" => "\xFB", # small u, circumflex accent
114 "Ugrave" => "\xD9", # capital U, grave accent
115 "ugrave" => "\xF9", # small u, grave accent
116 "Uuml" => "\xDC", # capital U, dieresis or umlaut mark
117 "uuml" => "\xFC", # small u, dieresis or umlaut mark
118 "Yacute" => "\xDD", # capital Y, acute accent
119 "yacute" => "\xFD", # small y, acute accent
120 "yuml" => "\xFF", # small y, dieresis or umlaut mark
122 "laquo" => "\xAB", # left pointing double angle quotation mark
123 "lchevron" => "\xAB", # synonym (backwards compatibility)
124 "raquo" => "\xBB", # right pointing double angle quotation mark
125 "rchevron" => "\xBB", # synonym (backwards compatibility)
127 "iexcl" => "\xA1", # inverted exclamation mark
128 "cent" => "\xA2", # cent sign
129 "pound" => "\xA3", # (UK) pound sign
130 "curren" => "\xA4", # currency sign
131 "yen" => "\xA5", # yen sign
132 "brvbar" => "\xA6", # broken vertical bar
133 "sect" => "\xA7", # section sign
134 "uml" => "\xA8", # diaresis
135 "copy" => "\xA9", # Copyright symbol
136 "ordf" => "\xAA", # feminine ordinal indicator
137 "not" => "\xAC", # not sign
138 "shy" => "\xAD", # soft hyphen
139 "reg" => "\xAE", # registered trademark
140 "macr" => "\xAF", # macron, overline
141 "deg" => "\xB0", # degree sign
142 "plusmn" => "\xB1", # plus-minus sign
143 "sup2" => "\xB2", # superscript 2
144 "sup3" => "\xB3", # superscript 3
145 "acute" => "\xB4", # acute accent
146 "micro" => "\xB5", # micro sign
147 "para" => "\xB6", # pilcrow sign = paragraph sign
148 "middot" => "\xB7", # middle dot = Georgian comma
149 "cedil" => "\xB8", # cedilla
150 "sup1" => "\xB9", # superscript 1
151 "ordm" => "\xBA", # masculine ordinal indicator
152 "frac14" => "\xBC", # vulgar fraction one quarter
153 "frac12" => "\xBD", # vulgar fraction one half
154 "frac34" => "\xBE", # vulgar fraction three quarters
155 "iquest" => "\xBF", # inverted question mark
156 "times" => "\xD7", # multiplication sign
157 "divide" => "\xF7", # division sign
161 ############################################################################
163 ############################################################################
165 # Initialize the object. Must be sure to call our parent initializer.
169 $$self{alt} = 0 unless defined $$self{alt};
170 $$self{indent} = 4 unless defined $$self{indent};
171 $$self{loose} = 0 unless defined $$self{loose};
172 $$self{sentence} = 0 unless defined $$self{sentence};
173 $$self{width} = 76 unless defined $$self{width};
175 # Figure out what quotes we'll be using for C<> text.
176 $$self{quotes} ||= "'";
177 if ($$self{quotes} eq 'none') {
178 $$self{LQUOTE} = $$self{RQUOTE} = '';
179 } elsif (length ($$self{quotes}) == 1) {
180 $$self{LQUOTE} = $$self{RQUOTE} = $$self{quotes};
181 } elsif ($$self{quotes} =~ /^(.)(.)$/
182 || $$self{quotes} =~ /^(..)(..)$/) {
186 croak qq(Invalid quote specification "$$self{quotes}");
189 $$self{INDENTS} = []; # Stack of indentations.
190 $$self{MARGIN} = $$self{indent}; # Current left margin in spaces.
192 $self->SUPER::initialize;
196 ############################################################################
198 ############################################################################
200 # Called for each command paragraph. Gets the command, the associated
201 # paragraph, the line number, and a Pod::Paragraph object. Just dispatches
202 # the command to a method named the same as the command. =cut is handled
203 # internally by Pod::Parser.
207 return if $command eq 'pod';
208 return if ($$self{EXCLUDE} && $command ne 'end');
209 $self->item ("\n") if defined $$self{ITEM};
210 if ($self->can ('cmd_' . $command)) {
211 $command = 'cmd_' . $command;
212 $self->$command (@_);
214 my ($text, $line, $paragraph) = @_;
216 ($file, $line) = $paragraph->file_line;
218 $text = " $text" if ($text =~ /^\S/);
219 warn qq($file:$line: Unknown command paragraph "=$command$text"\n);
224 # Called for a verbatim paragraph. Gets the paragraph, the line number, and
225 # a Pod::Paragraph object. Just output it verbatim, but with tabs converted
229 return if $$self{EXCLUDE};
230 $self->item if defined $$self{ITEM};
233 s/^(\s*\S+)/(' ' x $$self{MARGIN}) . $1/gme;
237 # Called for a regular text block. Gets the paragraph, the line number, and
238 # a Pod::Paragraph object. Perform interpolation and output the results.
241 return if $$self{EXCLUDE};
242 $self->output ($_[0]), return if $$self{VERBATIM};
246 # Perform a little magic to collapse multiple L<> references. This is
247 # here mostly for backwards-compatibility. We'll just rewrite the whole
248 # thing into actual text at this part, bypassing the whole internal
249 # sequence parsing thing.
252 L< # A link of the form L</something>.
255 [:\w]+ # The item has to be a simple word...
256 (\(\))? # ...or simple function.
260 ,?\s+(and\s+)? # Allow lots of them, conjuncted.
273 my @items = split /(?:,?\s+(?:and\s+)?)/;
276 for ($i = 0; $i < @items; $i++) {
277 $string .= $items[$i];
278 $string .= ", " if @items > 2 && $i != $#items;
279 $string .= " and " if ($i == $#items - 1);
281 $string .= " entries elsewhere in this document";
285 # Now actually interpolate and output the paragraph.
286 $_ = $self->interpolate ($_, $line);
288 if (defined $$self{ITEM}) {
289 $self->item ($_ . "\n");
291 $self->output ($self->reformat ($_ . "\n"));
295 # Called for an interior sequence. Gets the command, argument, and a
296 # Pod::InteriorSequence object and is expected to return the resulting text.
297 # Calls code, bold, italic, file, and link to handle those types of
298 # sequences, and handles S<>, E<>, X<>, and Z<> directly.
299 sub interior_sequence {
303 return '' if ($command eq 'X' || $command eq 'Z');
305 # Expand escapes into the actual character now, carping if invalid.
306 if ($command eq 'E') {
310 return $ESCAPES{$_} if defined $ESCAPES{$_};
311 carp "Unknown escape: E<$_>";
316 # For all the other sequences, empty content produces no output.
319 # For S<>, compress all internal whitespace and then map spaces to \01.
320 # When we output the text, we'll map this back.
321 if ($command eq 'S') {
327 # Anything else needs to get dispatched to another method.
328 if ($command eq 'B') { return $self->seq_b ($_) }
329 elsif ($command eq 'C') { return $self->seq_c ($_) }
330 elsif ($command eq 'F') { return $self->seq_f ($_) }
331 elsif ($command eq 'I') { return $self->seq_i ($_) }
332 elsif ($command eq 'L') { return $self->seq_l ($_) }
333 else { carp "Unknown sequence $command<$_>" }
336 # Called for each paragraph that's actually part of the POD. We take
337 # advantage of this opportunity to untabify the input.
338 sub preprocess_paragraph {
341 1 while s/^(.*?)(\t+)/$1 . ' ' x (length ($2) * 8 - length ($1) % 8)/me;
346 ############################################################################
348 ############################################################################
350 # All command paragraphs take the paragraph and the line number.
352 # First level heading.
357 $_ = $self->interpolate ($_, shift);
359 $self->output ("\n==== $_ ====\n\n");
361 $_ .= "\n" if $$self{loose};
362 $self->output ($_ . "\n");
366 # Second level heading.
371 $_ = $self->interpolate ($_, shift);
373 $self->output ("\n== $_ ==\n\n");
375 $self->output (' ' x ($$self{indent} / 2) . $_ . "\n\n");
383 unless (/^[-+]?\d+\s+$/) { $_ = $$self{indent} }
384 push (@{ $$self{INDENTS} }, $$self{MARGIN});
385 $$self{MARGIN} += ($_ + 0);
391 $$self{MARGIN} = pop @{ $$self{INDENTS} };
392 unless (defined $$self{MARGIN}) {
393 carp "Unmatched =back";
394 $$self{MARGIN} = $$self{indent};
398 # An individual list item.
401 if (defined $$self{ITEM}) { $self->item }
404 $$self{ITEM} = $self->interpolate ($_);
407 # Begin a block for a particular translator. Setting VERBATIM triggers
408 # special handling in textblock().
412 my ($kind) = /^(\S+)/ or return;
413 if ($kind eq 'text') {
414 $$self{VERBATIM} = 1;
420 # End a block for a particular translator. We assume that all =begin/=end
421 # pairs are properly closed.
425 $$self{VERBATIM} = 0;
428 # One paragraph for a particular translator. Ignore it unless it's intended
429 # for text, in which case we treat it as a verbatim text block.
434 return unless s/^text\b[ \t]*\n?//;
435 $self->verbatim ($_, $line);
439 ############################################################################
441 ############################################################################
443 # The simple formatting ones. These are here mostly so that subclasses can
444 # override them and do more complicated things.
445 sub seq_b { return $_[0]{alt} ? "``$_[1]''" : $_[1] }
446 sub seq_f { return $_[0]{alt} ? "\"$_[1]\"" : $_[1] }
447 sub seq_i { return '*' . $_[1] . '*' }
449 return $_[0]{alt} ? "``$_[1]''" : "$_[0]{LQUOTE}$_[1]$_[0]{RQUOTE}"
452 # The complicated one. Handle links. Since this is plain text, we can't
453 # actually make any real links, so this is all to figure out what text we
459 # Smash whitespace in case we were split across multiple lines.
462 # If we were given any explicit text, just output it.
463 if (/^([^|]+)\|/) { return $1 }
465 # Okay, leading and trailing whitespace isn't important; get rid of it.
469 # Default to using the whole content of the link entry as a section
470 # name. Note that L<manpage/> forces a manpage interpretation, as does
471 # something looking like L<manpage(section)>. The latter is an
472 # enhancement over the original Pod::Text.
473 my ($manpage, $section) = ('', $_);
474 if (/^"\s*(.*?)\s*"$/) {
475 $section = '"' . $1 . '"';
476 } elsif (m/^[-:.\w]+(?:\(\S+\))?$/) {
477 ($manpage, $section) = ($_, '');
479 ($manpage, $section) = split (/\s*\/\s*/, $_, 2);
482 # Now build the actual output text.
484 if (!length $section) {
485 $text = "the $manpage manpage" if length $manpage;
486 } elsif ($section =~ /^[:\w]+(?:\(\))?/) {
487 $text .= 'the ' . $section . ' entry';
488 $text .= (length $manpage) ? " in the $manpage manpage"
489 : " elsewhere in this document";
491 $section =~ s/^\"\s*//;
492 $section =~ s/\s*\"$//;
493 $text .= 'the section on "' . $section . '"';
494 $text .= " in the $manpage manpage" if length $manpage;
500 ############################################################################
502 ############################################################################
504 # This method is called whenever an =item command is complete (in other
505 # words, we've seen its associated paragraph or know for certain that it
506 # doesn't have one). It gets the paragraph associated with the item as an
507 # argument. If that argument is empty, just output the item tag; if it
508 # contains a newline, output the item tag followed by the newline.
509 # Otherwise, see if there's enough room for us to output the item tag in the
510 # margin of the text or if we have to put it on a separate line.
514 my $tag = $$self{ITEM};
515 unless (defined $tag) {
516 carp "item called without tag";
520 my $indent = $$self{INDENTS}[-1];
521 unless (defined $indent) { $indent = $$self{indent} }
522 my $space = ' ' x $indent;
523 $space =~ s/^ /:/ if $$self{alt};
524 if (!$_ || /^\s+$/ || ($$self{MARGIN} - $indent < length ($tag) + 1)) {
525 my $margin = $$self{MARGIN};
526 $$self{MARGIN} = $indent;
527 my $output = $self->reformat ($tag);
528 $output =~ s/\n*$/\n/;
529 $self->output ($output);
530 $$self{MARGIN} = $margin;
531 $self->output ($self->reformat ($_)) if /\S/;
533 $_ = $self->reformat ($_);
534 s/^ /:/ if ($$self{alt} && $indent > 0);
535 my $tagspace = ' ' x length $tag;
536 s/^($space)$tagspace/$1$tag/ or warn "Bizarre space in item";
542 ############################################################################
544 ############################################################################
546 # Wrap a line, indenting by the current left margin. We can't use
547 # Text::Wrap because it plays games with tabs. We can't use formline, even
548 # though we'd really like to, because it screws up non-printing characters.
549 # So we have to do the wrapping ourselves.
554 my $spaces = ' ' x $$self{MARGIN};
555 my $width = $$self{width} - $$self{MARGIN};
556 while (length > $width) {
557 if (s/^([^\n]{0,$width})\s+// || s/^([^\n]{$width})//) {
558 $output .= $spaces . $1 . "\n";
563 $output .= $spaces . $_;
564 $output =~ s/\s+$/\n\n/;
568 # Reformat a paragraph of text for the current margin. Takes the text to
569 # reformat and returns the formatted text.
574 # If we're trying to preserve two spaces after sentences, do some
575 # munging to support that. Otherwise, smash all repeated whitespace.
576 if ($$self{sentence}) {
587 # Output text to the output device.
588 sub output { $_[1] =~ tr/\01/ /; print { $_[0]->output_handle } $_[1] }
591 ############################################################################
592 # Backwards compatibility
593 ############################################################################
595 # The old Pod::Text module did everything in a pod2text() function. This
596 # tries to provide the same interface for legacy applications.
600 # This is really ugly; I hate doing option parsing in the middle of a
601 # module. But the old Pod::Text module supported passing flags to its
602 # entry function, so handle -a and -<number>.
603 while ($_[0] =~ /^-/) {
605 if ($flag eq '-a') { push (@args, alt => 1) }
606 elsif ($flag =~ /^-(\d+)$/) { push (@args, width => $1) }
613 # Now that we know what arguments we're using, create the parser.
614 my $parser = Pod::Text->new (@args);
616 # If two arguments were given, the second argument is going to be a file
617 # handle. That means we want to call parse_from_filehandle(), which
618 # means we need to turn the first argument into a file handle. Magic
619 # open will handle the <&STDIN case automagically.
623 unless (open (IN, $fhs[0])) {
624 croak ("Can't open $fhs[0] for reading: $!\n");
628 return $parser->parse_from_filehandle (@fhs);
630 return $parser->parse_from_file (@_);
635 ############################################################################
636 # Module return value and documentation
637 ############################################################################
644 Pod::Text - Convert POD data to formatted ASCII text
649 my $parser = Pod::Text->new (sentence => 0, width => 78);
651 # Read POD from STDIN and write to STDOUT.
652 $parser->parse_from_filehandle;
654 # Read POD from file.pod and write to file.txt.
655 $parser->parse_from_file ('file.pod', 'file.txt');
659 Pod::Text is a module that can convert documentation in the POD format (the
660 preferred language for documenting Perl) into formatted ASCII. It uses no
661 special formatting controls or codes whatsoever, and its output is therefore
662 suitable for nearly any device.
664 As a derived class from Pod::Parser, Pod::Text supports the same methods and
665 interfaces. See L<Pod::Parser> for all the details; briefly, one creates a
666 new parser with C<Pod::Text-E<gt>new()> and then calls either
667 parse_from_filehandle() or parse_from_file().
669 new() can take options, in the form of key/value pairs, that control the
670 behavior of the parser. The currently recognized options are:
676 If set to a true value, selects an alternate output format that, among other
677 things, uses a different heading style and marks C<=item> entries with a
678 colon in the left margin. Defaults to false.
682 The number of spaces to indent regular text, and the default indentation for
683 C<=over> blocks. Defaults to 4.
687 If set to a true value, a blank line is printed after a C<=head1> heading.
688 If set to false (the default), no blank line is printed after C<=head1>,
689 although one is still printed after C<=head2>. This is the default because
690 it's the expected formatting for manual pages; if you're formatting
691 arbitrary text documents, setting this to true may result in more pleasing
696 Sets the quote marks used to surround CE<lt>> text. If the value is a
697 single character, it is used as both the left and right quote; if it is two
698 characters, the first character is used as the left quote and the second as
699 the right quoted; and if it is four characters, the first two are used as
700 the left quote and the second two as the right quote.
702 This may also be set to the special value C<none>, in which case no quote
703 marks are added around CE<lt>> text.
707 If set to a true value, Pod::Text will assume that each sentence ends in two
708 spaces, and will try to preserve that spacing. If set to false, all
709 consecutive whitespace in non-verbatim paragraphs is compressed into a
710 single space. Defaults to true.
714 The column at which to wrap text on the right-hand side. Defaults to 76.
718 The standard Pod::Parser method parse_from_filehandle() takes up to two
719 arguments, the first being the file handle to read POD from and the second
720 being the file handle to write the formatted output to. The first defaults
721 to STDIN if not given, and the second defaults to STDOUT. The method
722 parse_from_file() is almost identical, except that its two arguments are the
723 input and output disk files instead. See L<Pod::Parser> for the specific
730 =item Bizarre space in item
732 (W) Something has gone wrong in internal C<=item> processing. This message
733 indicates a bug in Pod::Text; you should never see it.
735 =item Can't open %s for reading: %s
737 (F) Pod::Text was invoked via the compatibility mode pod2text() interface
738 and the input file it was given could not be opened.
740 =item Invalid quote specification "%s"
742 (F) The quote specification given (the quotes option to the constructor) was
743 invalid. A quote specification must be one, two, or four characters long.
745 =item %s:%d: Unknown command paragraph "%s".
747 (W) The POD source contained a non-standard command paragraph (something of
748 the form C<=command args>) that Pod::Man didn't know about. It was ignored.
750 =item Unknown escape: %s
752 (W) The POD source contained an C<EE<lt>E<gt>> escape that Pod::Text didn't
755 =item Unknown sequence: %s
757 (W) The POD source contained a non-standard internal sequence (something of
758 the form C<XE<lt>E<gt>>) that Pod::Text didn't know about.
760 =item Unmatched =back
762 (W) Pod::Text encountered a C<=back> command that didn't correspond to an
769 Embedded Ctrl-As (octal 001) in the input will be mapped to spaces on
770 output, due to an internal implementation detail.
774 This is a replacement for an earlier Pod::Text module written by Tom
775 Christiansen. It has a revamped interface, since it now uses Pod::Parser,
776 but an interface roughly compatible with the old Pod::Text::pod2text()
777 function is still available. Please change to the new calling convention,
780 The original Pod::Text contained code to do formatting via termcap
781 sequences, although it wasn't turned on by default and it was problematic to
782 get it to work at all. This rewrite doesn't even try to do that, but a
783 subclass of it does. Look for L<Pod::Text::Termcap|Pod::Text::Termcap>.
787 L<Pod::Parser|Pod::Parser>, L<Pod::Text::Termcap|Pod::Text::Termcap>,
792 Russ Allbery E<lt>rra@stanford.eduE<gt>, based I<very> heavily on the
793 original Pod::Text by Tom Christiansen E<lt>tchrist@mox.perl.comE<gt> and
794 its conversion to Pod::Parser by Brad Appleton
795 E<lt>bradapp@enteract.comE<gt>.