1 # Pod::Text -- Convert POD data to formatted ASCII text.
2 # $Id: Text.pm,v 2.4 2000/03/17 00:17:08 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
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 $$self{INDENTS} = []; # Stack of indentations.
176 $$self{MARGIN} = $$self{indent}; # Current left margin in spaces.
178 $self->SUPER::initialize;
182 ############################################################################
184 ############################################################################
186 # Called for each command paragraph. Gets the command, the associated
187 # paragraph, the line number, and a Pod::Paragraph object. Just dispatches
188 # the command to a method named the same as the command. =cut is handled
189 # internally by Pod::Parser.
193 return if $command eq 'pod';
194 return if ($$self{EXCLUDE} && $command ne 'end');
195 $self->item ("\n") if defined $$self{ITEM};
196 $command = 'cmd_' . $command;
197 $self->$command (@_);
200 # Called for a verbatim paragraph. Gets the paragraph, the line number, and
201 # a Pod::Paragraph object. Just output it verbatim, but with tabs converted
205 return if $$self{EXCLUDE};
206 $self->item if defined $$self{ITEM};
209 s/^(\s*\S+)/(' ' x $$self{MARGIN}) . $1/gme;
213 # Called for a regular text block. Gets the paragraph, the line number, and
214 # a Pod::Paragraph object. Perform interpolation and output the results.
217 return if $$self{EXCLUDE};
218 $self->output ($_[0]), return if $$self{VERBATIM};
222 # Perform a little magic to collapse multiple L<> references. This is
223 # here mostly for backwards-compatibility. We'll just rewrite the whole
224 # thing into actual text at this part, bypassing the whole internal
225 # sequence parsing thing.
228 L< # A link of the form L</something>.
231 [:\w]+ # The item has to be a simple word...
232 (\(\))? # ...or simple function.
236 ,?\s+(and\s+)? # Allow lots of them, conjuncted.
249 my @items = split /(?:,?\s+(?:and\s+)?)/;
252 for ($i = 0; $i < @items; $i++) {
253 $string .= $items[$i];
254 $string .= ", " if @items > 2 && $i != $#items;
255 $string .= " and " if ($i == $#items - 1);
257 $string .= " entries elsewhere in this document";
261 # Now actually interpolate and output the paragraph.
262 $_ = $self->interpolate ($_, $line);
264 if (defined $$self{ITEM}) {
265 $self->item ($_ . "\n");
267 $self->output ($self->reformat ($_ . "\n"));
271 # Called for an interior sequence. Gets the command, argument, and a
272 # Pod::InteriorSequence object and is expected to return the resulting text.
273 # Calls code, bold, italic, file, and link to handle those types of
274 # sequences, and handles S<>, E<>, X<>, and Z<> directly.
275 sub interior_sequence {
279 return '' if ($command eq 'X' || $command eq 'Z');
281 # Expand escapes into the actual character now, carping if invalid.
282 if ($command eq 'E') {
286 return $ESCAPES{$_} if defined $ESCAPES{$_};
287 carp "Unknown escape: E<$_>";
292 # For all the other sequences, empty content produces no output.
295 # For S<>, compress all internal whitespace and then map spaces to \01.
296 # When we output the text, we'll map this back.
297 if ($command eq 'S') {
303 # Anything else needs to get dispatched to another method.
304 if ($command eq 'B') { return $self->seq_b ($_) }
305 elsif ($command eq 'C') { return $self->seq_c ($_) }
306 elsif ($command eq 'F') { return $self->seq_f ($_) }
307 elsif ($command eq 'I') { return $self->seq_i ($_) }
308 elsif ($command eq 'L') { return $self->seq_l ($_) }
309 else { carp "Unknown sequence $command<$_>" }
312 # Called for each paragraph that's actually part of the POD. We take
313 # advantage of this opportunity to untabify the input.
314 sub preprocess_paragraph {
317 1 while s/^(.*?)(\t+)/$1 . ' ' x (length ($2) * 8 - length ($1) % 8)/me;
322 ############################################################################
324 ############################################################################
326 # All command paragraphs take the paragraph and the line number.
328 # First level heading.
333 $_ = $self->interpolate ($_, shift);
335 $self->output ("\n==== $_ ====\n\n");
337 $_ .= "\n" if $$self{loose};
338 $self->output ($_ . "\n");
342 # Second level heading.
347 $_ = $self->interpolate ($_, shift);
349 $self->output ("\n== $_ ==\n\n");
351 $self->output (' ' x ($$self{indent} / 2) . $_ . "\n\n");
359 unless (/^[-+]?\d+\s+$/) { $_ = $$self{indent} }
360 push (@{ $$self{INDENTS} }, $$self{MARGIN});
361 $$self{MARGIN} += ($_ + 0);
367 $$self{MARGIN} = pop @{ $$self{INDENTS} };
368 unless (defined $$self{MARGIN}) {
369 carp "Unmatched =back";
370 $$self{MARGIN} = $$self{indent};
374 # An individual list item.
377 if (defined $$self{ITEM}) { $self->item }
380 $$self{ITEM} = $self->interpolate ($_);
383 # Begin a block for a particular translator. Setting VERBATIM triggers
384 # special handling in textblock().
388 my ($kind) = /^(\S+)/ or return;
389 if ($kind eq 'text') {
390 $$self{VERBATIM} = 1;
396 # End a block for a particular translator. We assume that all =begin/=end
397 # pairs are properly closed.
401 $$self{VERBATIM} = 0;
404 # One paragraph for a particular translator. Ignore it unless it's intended
405 # for text, in which case we treat it as a verbatim text block.
410 return unless s/^text\b[ \t]*\n?//;
411 $self->verbatim ($_, $line);
415 ############################################################################
417 ############################################################################
419 # The simple formatting ones. These are here mostly so that subclasses can
420 # override them and do more complicated things.
421 sub seq_b { return $_[0]{alt} ? "``$_[1]''" : $_[1] }
422 sub seq_c { return $_[0]{alt} ? "``$_[1]''" : "`$_[1]'" }
423 sub seq_f { return $_[0]{alt} ? "\"$_[1]\"" : $_[1] }
424 sub seq_i { return '*' . $_[1] . '*' }
426 # The complicated one. Handle links. Since this is plain text, we can't
427 # actually make any real links, so this is all to figure out what text we
433 # Smash whitespace in case we were split across multiple lines.
436 # If we were given any explicit text, just output it.
437 if (/^([^|]+)\|/) { return $1 }
439 # Okay, leading and trailing whitespace isn't important; get rid of it.
443 # Default to using the whole content of the link entry as a section
444 # name. Note that L<manpage/> forces a manpage interpretation, as does
445 # something looking like L<manpage(section)>. The latter is an
446 # enhancement over the original Pod::Text.
447 my ($manpage, $section) = ('', $_);
448 if (/^"\s*(.*?)\s*"$/) {
449 $section = '"' . $1 . '"';
450 } elsif (m/^[-:.\w]+(?:\(\S+\))?$/) {
451 ($manpage, $section) = ($_, '');
453 ($manpage, $section) = split (/\s*\/\s*/, $_, 2);
456 # Now build the actual output text.
458 if (!length $section) {
459 $text = "the $manpage manpage" if length $manpage;
460 } elsif ($section =~ /^[:\w]+(?:\(\))?/) {
461 $text .= 'the ' . $section . ' entry';
462 $text .= (length $manpage) ? " in the $manpage manpage"
463 : " elsewhere in this document";
465 $section =~ s/^\"\s*//;
466 $section =~ s/\s*\"$//;
467 $text .= 'the section on "' . $section . '"';
468 $text .= " in the $manpage manpage" if length $manpage;
474 ############################################################################
476 ############################################################################
478 # This method is called whenever an =item command is complete (in other
479 # words, we've seen its associated paragraph or know for certain that it
480 # doesn't have one). It gets the paragraph associated with the item as an
481 # argument. If that argument is empty, just output the item tag; if it
482 # contains a newline, output the item tag followed by the newline.
483 # Otherwise, see if there's enough room for us to output the item tag in the
484 # margin of the text or if we have to put it on a separate line.
488 my $tag = $$self{ITEM};
489 unless (defined $tag) {
490 carp "item called without tag";
494 my $indent = $$self{INDENTS}[-1];
495 unless (defined $indent) { $indent = $$self{indent} }
496 my $space = ' ' x $indent;
497 $space =~ s/^ /:/ if $$self{alt};
498 if (!$_ || /^\s+$/ || ($$self{MARGIN} - $indent < length ($tag) + 1)) {
499 my $margin = $$self{MARGIN};
500 $$self{MARGIN} = $indent;
501 my $output = $self->reformat ($tag);
502 $output =~ s/\n*$/\n/;
503 $self->output ($output);
504 $$self{MARGIN} = $margin;
505 $self->output ($self->reformat ($_)) if /\S/;
507 $_ = $self->reformat ($_);
508 s/^ /:/ if ($$self{alt} && $indent > 0);
509 my $tagspace = ' ' x length $tag;
510 s/^($space)$tagspace/$1$tag/ or warn "Bizarre space in item";
516 ############################################################################
518 ############################################################################
520 # Wrap a line, indenting by the current left margin. We can't use
521 # Text::Wrap because it plays games with tabs. We can't use formline, even
522 # though we'd really like to, because it screws up non-printing characters.
523 # So we have to do the wrapping ourselves.
528 my $spaces = ' ' x $$self{MARGIN};
529 my $width = $$self{width} - $$self{MARGIN};
530 while (length > $width) {
531 if (s/^([^\n]{0,$width})\s+// || s/^([^\n]{$width})//) {
532 $output .= $spaces . $1 . "\n";
537 $output .= $spaces . $_;
538 $output =~ s/\s+$/\n\n/;
542 # Reformat a paragraph of text for the current margin. Takes the text to
543 # reformat and returns the formatted text.
548 # If we're trying to preserve two spaces after sentences, do some
549 # munging to support that. Otherwise, smash all repeated whitespace.
550 if ($$self{sentence}) {
561 # Output text to the output device.
562 sub output { $_[1] =~ tr/\01/ /; print { $_[0]->output_handle } $_[1] }
565 ############################################################################
566 # Backwards compatibility
567 ############################################################################
569 # The old Pod::Text module did everything in a pod2text() function. This
570 # tries to provide the same interface for legacy applications.
574 # This is really ugly; I hate doing option parsing in the middle of a
575 # module. But the old Pod::Text module supported passing flags to its
576 # entry function, so handle -a and -<number>.
577 while ($_[0] =~ /^-/) {
579 if ($flag eq '-a') { push (@args, alt => 1) }
580 elsif ($flag =~ /^-(\d+)$/) { push (@args, width => $1) }
587 # Now that we know what arguments we're using, create the parser.
588 my $parser = Pod::Text->new (@args);
590 # If two arguments were given, the second argument is going to be a file
591 # handle. That means we want to call parse_from_filehandle(), which
592 # means we need to turn the first argument into a file handle. Magic
593 # open will handle the <&STDIN case automagically.
596 unless (open (IN, $_[0])) {
597 croak ("Can't open $_[0] for reading: $!\n");
601 return $parser->parse_from_filehandle (@_);
603 return $parser->parse_from_file (@_);
608 ############################################################################
609 # Module return value and documentation
610 ############################################################################
617 Pod::Text - Convert POD data to formatted ASCII text
622 my $parser = Pod::Text->new (sentence => 0, width => 78);
624 # Read POD from STDIN and write to STDOUT.
625 $parser->parse_from_filehandle;
627 # Read POD from file.pod and write to file.txt.
628 $parser->parse_from_file ('file.pod', 'file.txt');
632 Pod::Text is a module that can convert documentation in the POD format (the
633 preferred language for documenting Perl) into formatted ASCII. It uses no
634 special formatting controls or codes whatsoever, and its output is therefore
635 suitable for nearly any device.
637 As a derived class from Pod::Parser, Pod::Text supports the same methods and
638 interfaces. See L<Pod::Parser> for all the details; briefly, one creates a
639 new parser with C<Pod::Text-E<gt>new()> and then calls either
640 parse_from_filehandle() or parse_from_file().
642 new() can take options, in the form of key/value pairs, that control the
643 behavior of the parser. The currently recognized options are:
649 If set to a true value, selects an alternate output format that, among other
650 things, uses a different heading style and marks C<=item> entries with a
651 colon in the left margin. Defaults to false.
655 The number of spaces to indent regular text, and the default indentation for
656 C<=over> blocks. Defaults to 4.
660 If set to a true value, a blank line is printed after a C<=head1> heading.
661 If set to false (the default), no blank line is printed after C<=head1>,
662 although one is still printed after C<=head2>. This is the default because
663 it's the expected formatting for manual pages; if you're formatting
664 arbitrary text documents, setting this to true may result in more pleasing
669 If set to a true value, Pod::Text will assume that each sentence ends in two
670 spaces, and will try to preserve that spacing. If set to false, all
671 consecutive whitespace in non-verbatim paragraphs is compressed into a
672 single space. Defaults to true.
676 The column at which to wrap text on the right-hand side. Defaults to 76.
680 The standard Pod::Parser method parse_from_filehandle() takes up to two
681 arguments, the first being the file handle to read POD from and the second
682 being the file handle to write the formatted output to. The first defaults
683 to STDIN if not given, and the second defaults to STDOUT. The method
684 parse_from_file() is almost identical, except that its two arguments are the
685 input and output disk files instead. See L<Pod::Parser> for the specific
692 =item Bizarre space in item
694 (W) Something has gone wrong in internal C<=item> processing. This message
695 indicates a bug in Pod::Text; you should never see it.
697 =item Can't open %s for reading: %s
699 (F) Pod::Text was invoked via the compatibility mode pod2text() interface
700 and the input file it was given could not be opened.
702 =item Unknown escape: %s
704 (W) The POD source contained an C<EE<lt>E<gt>> escape that Pod::Text didn't
707 =item Unknown sequence: %s
709 (W) The POD source contained a non-standard internal sequence (something of
710 the form C<XE<lt>E<gt>>) that Pod::Text didn't know about.
712 =item Unmatched =back
714 (W) Pod::Text encountered a C<=back> command that didn't correspond to an
721 Embedded Ctrl-As (octal 001) in the input will be mapped to spaces on
722 output, due to an internal implementation detail.
726 This is a replacement for an earlier Pod::Text module written by Tom
727 Christiansen. It has a revamped interface, since it now uses Pod::Parser,
728 but an interface roughly compatible with the old Pod::Text::pod2text()
729 function is still available. Please change to the new calling convention,
732 The original Pod::Text contained code to do formatting via termcap
733 sequences, although it wasn't turned on by default and it was problematic to
734 get it to work at all. This rewrite doesn't even try to do that, but a
735 subclass of it does. Look for L<Pod::Text::Termcap|Pod::Text::Termcap>.
739 L<Pod::Parser|Pod::Parser>, L<Pod::Text::Termcap|Pod::Text::Termcap>,
744 Russ Allbery E<lt>rra@stanford.eduE<gt>, based I<very> heavily on the
745 original Pod::Text by Tom Christiansen E<lt>tchrist@mox.perl.comE<gt> and
746 its conversion to Pod::Parser by Brad Appleton
747 E<lt>bradapp@enteract.comE<gt>.