1 # Pod::Text -- Convert POD data to formatted ASCII text.
2 # $Id: Text.pm,v 2.5 2000/09/03 09:23:29 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" => "\xCD", # capital I, acute accent
86 "iacute" => "\xED", # 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, $paragraph) = @_;
215 my ($file, $line) = $paragraph->file_line;
217 $text = " $text" if ($text =~ /^\S/);
218 warn qq($file:$line: Unknown command paragraph "=$command$text"\n);
223 # Called for a verbatim paragraph. Gets the paragraph, the line number, and
224 # a Pod::Paragraph object. Just output it verbatim, but with tabs converted
228 return if $$self{EXCLUDE};
229 $self->item if defined $$self{ITEM};
232 s/^(\s*\S+)/(' ' x $$self{MARGIN}) . $1/gme;
236 # Called for a regular text block. Gets the paragraph, the line number, and
237 # a Pod::Paragraph object. Perform interpolation and output the results.
240 return if $$self{EXCLUDE};
241 $self->output ($_[0]), return if $$self{VERBATIM};
245 # Perform a little magic to collapse multiple L<> references. This is
246 # here mostly for backwards-compatibility. We'll just rewrite the whole
247 # thing into actual text at this part, bypassing the whole internal
248 # sequence parsing thing.
251 L< # A link of the form L</something>.
254 [:\w]+ # The item has to be a simple word...
255 (\(\))? # ...or simple function.
259 ,?\s+(and\s+)? # Allow lots of them, conjuncted.
272 my @items = split /(?:,?\s+(?:and\s+)?)/;
275 for ($i = 0; $i < @items; $i++) {
276 $string .= $items[$i];
277 $string .= ", " if @items > 2 && $i != $#items;
278 $string .= " and " if ($i == $#items - 1);
280 $string .= " entries elsewhere in this document";
284 # Now actually interpolate and output the paragraph.
285 $_ = $self->interpolate ($_, $line);
287 if (defined $$self{ITEM}) {
288 $self->item ($_ . "\n");
290 $self->output ($self->reformat ($_ . "\n"));
294 # Called for an interior sequence. Gets the command, argument, and a
295 # Pod::InteriorSequence object and is expected to return the resulting text.
296 # Calls code, bold, italic, file, and link to handle those types of
297 # sequences, and handles S<>, E<>, X<>, and Z<> directly.
298 sub interior_sequence {
302 return '' if ($command eq 'X' || $command eq 'Z');
304 # Expand escapes into the actual character now, carping if invalid.
305 if ($command eq 'E') {
309 return $ESCAPES{$_} if defined $ESCAPES{$_};
310 carp "Unknown escape: E<$_>";
315 # For all the other sequences, empty content produces no output.
318 # For S<>, compress all internal whitespace and then map spaces to \01.
319 # When we output the text, we'll map this back.
320 if ($command eq 'S') {
326 # Anything else needs to get dispatched to another method.
327 if ($command eq 'B') { return $self->seq_b ($_) }
328 elsif ($command eq 'C') { return $self->seq_c ($_) }
329 elsif ($command eq 'F') { return $self->seq_f ($_) }
330 elsif ($command eq 'I') { return $self->seq_i ($_) }
331 elsif ($command eq 'L') { return $self->seq_l ($_) }
332 else { carp "Unknown sequence $command<$_>" }
335 # Called for each paragraph that's actually part of the POD. We take
336 # advantage of this opportunity to untabify the input.
337 sub preprocess_paragraph {
340 1 while s/^(.*?)(\t+)/$1 . ' ' x (length ($2) * 8 - length ($1) % 8)/me;
345 ############################################################################
347 ############################################################################
349 # All command paragraphs take the paragraph and the line number.
351 # First level heading.
356 $_ = $self->interpolate ($_, shift);
358 $self->output ("\n==== $_ ====\n\n");
360 $_ .= "\n" if $$self{loose};
361 $self->output ($_ . "\n");
365 # Second level heading.
370 $_ = $self->interpolate ($_, shift);
372 $self->output ("\n== $_ ==\n\n");
374 $self->output (' ' x ($$self{indent} / 2) . $_ . "\n\n");
382 unless (/^[-+]?\d+\s+$/) { $_ = $$self{indent} }
383 push (@{ $$self{INDENTS} }, $$self{MARGIN});
384 $$self{MARGIN} += ($_ + 0);
390 $$self{MARGIN} = pop @{ $$self{INDENTS} };
391 unless (defined $$self{MARGIN}) {
392 carp "Unmatched =back";
393 $$self{MARGIN} = $$self{indent};
397 # An individual list item.
400 if (defined $$self{ITEM}) { $self->item }
403 $$self{ITEM} = $self->interpolate ($_);
406 # Begin a block for a particular translator. Setting VERBATIM triggers
407 # special handling in textblock().
411 my ($kind) = /^(\S+)/ or return;
412 if ($kind eq 'text') {
413 $$self{VERBATIM} = 1;
419 # End a block for a particular translator. We assume that all =begin/=end
420 # pairs are properly closed.
424 $$self{VERBATIM} = 0;
427 # One paragraph for a particular translator. Ignore it unless it's intended
428 # for text, in which case we treat it as a verbatim text block.
433 return unless s/^text\b[ \t]*\n?//;
434 $self->verbatim ($_, $line);
438 ############################################################################
440 ############################################################################
442 # The simple formatting ones. These are here mostly so that subclasses can
443 # override them and do more complicated things.
444 sub seq_b { return $_[0]{alt} ? "``$_[1]''" : $_[1] }
445 sub seq_f { return $_[0]{alt} ? "\"$_[1]\"" : $_[1] }
446 sub seq_i { return '*' . $_[1] . '*' }
448 return $_[0]{alt} ? "``$_[1]''" : "$_[0]{LQUOTE}$_[1]$_[0]{RQUOTE}"
451 # The complicated one. Handle links. Since this is plain text, we can't
452 # actually make any real links, so this is all to figure out what text we
458 # Smash whitespace in case we were split across multiple lines.
461 # If we were given any explicit text, just output it.
462 if (/^([^|]+)\|/) { return $1 }
464 # Okay, leading and trailing whitespace isn't important; get rid of it.
468 # Default to using the whole content of the link entry as a section
469 # name. Note that L<manpage/> forces a manpage interpretation, as does
470 # something looking like L<manpage(section)>. The latter is an
471 # enhancement over the original Pod::Text.
472 my ($manpage, $section) = ('', $_);
473 if (/^"\s*(.*?)\s*"$/) {
474 $section = '"' . $1 . '"';
475 } elsif (m/^[-:.\w]+(?:\(\S+\))?$/) {
476 ($manpage, $section) = ($_, '');
478 ($manpage, $section) = split (/\s*\/\s*/, $_, 2);
481 # Now build the actual output text.
483 if (!length $section) {
484 $text = "the $manpage manpage" if length $manpage;
485 } elsif ($section =~ /^[:\w]+(?:\(\))?/) {
486 $text .= 'the ' . $section . ' entry';
487 $text .= (length $manpage) ? " in the $manpage manpage"
488 : " elsewhere in this document";
490 $section =~ s/^\"\s*//;
491 $section =~ s/\s*\"$//;
492 $text .= 'the section on "' . $section . '"';
493 $text .= " in the $manpage manpage" if length $manpage;
499 ############################################################################
501 ############################################################################
503 # This method is called whenever an =item command is complete (in other
504 # words, we've seen its associated paragraph or know for certain that it
505 # doesn't have one). It gets the paragraph associated with the item as an
506 # argument. If that argument is empty, just output the item tag; if it
507 # contains a newline, output the item tag followed by the newline.
508 # Otherwise, see if there's enough room for us to output the item tag in the
509 # margin of the text or if we have to put it on a separate line.
513 my $tag = $$self{ITEM};
514 unless (defined $tag) {
515 carp "item called without tag";
519 my $indent = $$self{INDENTS}[-1];
520 unless (defined $indent) { $indent = $$self{indent} }
521 my $space = ' ' x $indent;
522 $space =~ s/^ /:/ if $$self{alt};
523 if (!$_ || /^\s+$/ || ($$self{MARGIN} - $indent < length ($tag) + 1)) {
524 my $margin = $$self{MARGIN};
525 $$self{MARGIN} = $indent;
526 my $output = $self->reformat ($tag);
527 $output =~ s/\n*$/\n/;
528 $self->output ($output);
529 $$self{MARGIN} = $margin;
530 $self->output ($self->reformat ($_)) if /\S/;
532 $_ = $self->reformat ($_);
533 s/^ /:/ if ($$self{alt} && $indent > 0);
534 my $tagspace = ' ' x length $tag;
535 s/^($space)$tagspace/$1$tag/ or warn "Bizarre space in item";
541 ############################################################################
543 ############################################################################
545 # Wrap a line, indenting by the current left margin. We can't use
546 # Text::Wrap because it plays games with tabs. We can't use formline, even
547 # though we'd really like to, because it screws up non-printing characters.
548 # So we have to do the wrapping ourselves.
553 my $spaces = ' ' x $$self{MARGIN};
554 my $width = $$self{width} - $$self{MARGIN};
555 while (length > $width) {
556 if (s/^([^\n]{0,$width})\s+// || s/^([^\n]{$width})//) {
557 $output .= $spaces . $1 . "\n";
562 $output .= $spaces . $_;
563 $output =~ s/\s+$/\n\n/;
567 # Reformat a paragraph of text for the current margin. Takes the text to
568 # reformat and returns the formatted text.
573 # If we're trying to preserve two spaces after sentences, do some
574 # munging to support that. Otherwise, smash all repeated whitespace.
575 if ($$self{sentence}) {
586 # Output text to the output device.
587 sub output { $_[1] =~ tr/\01/ /; print { $_[0]->output_handle } $_[1] }
590 ############################################################################
591 # Backwards compatibility
592 ############################################################################
594 # The old Pod::Text module did everything in a pod2text() function. This
595 # tries to provide the same interface for legacy applications.
599 # This is really ugly; I hate doing option parsing in the middle of a
600 # module. But the old Pod::Text module supported passing flags to its
601 # entry function, so handle -a and -<number>.
602 while ($_[0] =~ /^-/) {
604 if ($flag eq '-a') { push (@args, alt => 1) }
605 elsif ($flag =~ /^-(\d+)$/) { push (@args, width => $1) }
612 # Now that we know what arguments we're using, create the parser.
613 my $parser = Pod::Text->new (@args);
615 # If two arguments were given, the second argument is going to be a file
616 # handle. That means we want to call parse_from_filehandle(), which
617 # means we need to turn the first argument into a file handle. Magic
618 # open will handle the <&STDIN case automagically.
622 unless (open (IN, $fhs[0])) {
623 croak ("Can't open $fhs[0] for reading: $!\n");
627 return $parser->parse_from_filehandle (@fhs);
629 return $parser->parse_from_file (@_);
634 ############################################################################
635 # Module return value and documentation
636 ############################################################################
643 Pod::Text - Convert POD data to formatted ASCII text
648 my $parser = Pod::Text->new (sentence => 0, width => 78);
650 # Read POD from STDIN and write to STDOUT.
651 $parser->parse_from_filehandle;
653 # Read POD from file.pod and write to file.txt.
654 $parser->parse_from_file ('file.pod', 'file.txt');
658 Pod::Text is a module that can convert documentation in the POD format (the
659 preferred language for documenting Perl) into formatted ASCII. It uses no
660 special formatting controls or codes whatsoever, and its output is therefore
661 suitable for nearly any device.
663 As a derived class from Pod::Parser, Pod::Text supports the same methods and
664 interfaces. See L<Pod::Parser> for all the details; briefly, one creates a
665 new parser with C<Pod::Text-E<gt>new()> and then calls either
666 parse_from_filehandle() or parse_from_file().
668 new() can take options, in the form of key/value pairs, that control the
669 behavior of the parser. The currently recognized options are:
675 If set to a true value, selects an alternate output format that, among other
676 things, uses a different heading style and marks C<=item> entries with a
677 colon in the left margin. Defaults to false.
681 The number of spaces to indent regular text, and the default indentation for
682 C<=over> blocks. Defaults to 4.
686 If set to a true value, a blank line is printed after a C<=head1> heading.
687 If set to false (the default), no blank line is printed after C<=head1>,
688 although one is still printed after C<=head2>. This is the default because
689 it's the expected formatting for manual pages; if you're formatting
690 arbitrary text documents, setting this to true may result in more pleasing
695 Sets the quote marks used to surround CE<lt>> text. If the value is a
696 single character, it is used as both the left and right quote; if it is two
697 characters, the first character is used as the left quote and the second as
698 the right quoted; and if it is four characters, the first two are used as
699 the left quote and the second two as the right quote.
701 This may also be set to the special value C<none>, in which case no quote
702 marks are added around CE<lt>> text.
706 If set to a true value, Pod::Text will assume that each sentence ends in two
707 spaces, and will try to preserve that spacing. If set to false, all
708 consecutive whitespace in non-verbatim paragraphs is compressed into a
709 single space. Defaults to true.
713 The column at which to wrap text on the right-hand side. Defaults to 76.
717 The standard Pod::Parser method parse_from_filehandle() takes up to two
718 arguments, the first being the file handle to read POD from and the second
719 being the file handle to write the formatted output to. The first defaults
720 to STDIN if not given, and the second defaults to STDOUT. The method
721 parse_from_file() is almost identical, except that its two arguments are the
722 input and output disk files instead. See L<Pod::Parser> for the specific
729 =item Bizarre space in item
731 (W) Something has gone wrong in internal C<=item> processing. This message
732 indicates a bug in Pod::Text; you should never see it.
734 =item Can't open %s for reading: %s
736 (F) Pod::Text was invoked via the compatibility mode pod2text() interface
737 and the input file it was given could not be opened.
739 =item Invalid quote specification "%s"
741 (F) The quote specification given (the quotes option to the constructor) was
742 invalid. A quote specification must be one, two, or four characters long.
744 =item %s:%d: Unknown command paragraph "%s".
746 (W) The POD source contained a non-standard command paragraph (something of
747 the form C<=command args>) that Pod::Man didn't know about. It was ignored.
749 =item Unknown escape: %s
751 (W) The POD source contained an C<EE<lt>E<gt>> escape that Pod::Text didn't
754 =item Unknown sequence: %s
756 (W) The POD source contained a non-standard internal sequence (something of
757 the form C<XE<lt>E<gt>>) that Pod::Text didn't know about.
759 =item Unmatched =back
761 (W) Pod::Text encountered a C<=back> command that didn't correspond to an
768 Embedded Ctrl-As (octal 001) in the input will be mapped to spaces on
769 output, due to an internal implementation detail.
773 This is a replacement for an earlier Pod::Text module written by Tom
774 Christiansen. It has a revamped interface, since it now uses Pod::Parser,
775 but an interface roughly compatible with the old Pod::Text::pod2text()
776 function is still available. Please change to the new calling convention,
779 The original Pod::Text contained code to do formatting via termcap
780 sequences, although it wasn't turned on by default and it was problematic to
781 get it to work at all. This rewrite doesn't even try to do that, but a
782 subclass of it does. Look for L<Pod::Text::Termcap|Pod::Text::Termcap>.
786 L<Pod::Parser|Pod::Parser>, L<Pod::Text::Termcap|Pod::Text::Termcap>,
791 Russ Allbery E<lt>rra@stanford.eduE<gt>, based I<very> heavily on the
792 original Pod::Text by Tom Christiansen E<lt>tchrist@mox.perl.comE<gt> and
793 its conversion to Pod::Parser by Brad Appleton
794 E<lt>bradapp@enteract.comE<gt>.