Commit | Line | Data |
6055f9d4 |
1 | # Pod::Text -- Convert POD data to formatted ASCII text. |
6055f9d4 |
2 | # |
0e4e3f6e |
3 | # Copyright 1999, 2000, 2001, 2002, 2004, 2006, 2008 |
4 | # Russ Allbery <rra@stanford.edu> |
6055f9d4 |
5 | # |
3c014959 |
6 | # This program is free software; you may redistribute it and/or modify it |
6055f9d4 |
7 | # under the same terms as Perl itself. |
8 | # |
5ec554fb |
9 | # This module converts POD to formatted text. It replaces the old Pod::Text |
10 | # module that came with versions of Perl prior to 5.6.0 and attempts to match |
11 | # its output except for some specific circumstances where other decisions |
12 | # seemed to produce better output. It uses Pod::Parser and is designed to be |
13 | # very easy to subclass. |
3c014959 |
14 | # |
15 | # Perl core hackers, please note that this module is also separately |
16 | # maintained outside of the Perl core as part of the podlators. Please send |
17 | # me any patches at the address above in addition to sending them to the |
18 | # standard Perl mailing lists. |
6055f9d4 |
19 | |
3c014959 |
20 | ############################################################################## |
6055f9d4 |
21 | # Modules and declarations |
3c014959 |
22 | ############################################################################## |
69e00e79 |
23 | |
6055f9d4 |
24 | package Pod::Text; |
69e00e79 |
25 | |
6055f9d4 |
26 | require 5.004; |
27 | |
6055f9d4 |
28 | use strict; |
2e20e14f |
29 | use vars qw(@ISA @EXPORT %ESCAPES $VERSION); |
6055f9d4 |
30 | |
b7ae008f |
31 | use Carp qw(carp croak); |
32 | use Exporter (); |
33 | use Pod::Simple (); |
34 | |
35 | @ISA = qw(Pod::Simple Exporter); |
6055f9d4 |
36 | |
2e20e14f |
37 | # We have to export pod2text for backward compatibility. |
38 | @EXPORT = qw(pod2text); |
39 | |
2dab2cc1 |
40 | $VERSION = '3.12_01'; |
69e00e79 |
41 | |
3c014959 |
42 | ############################################################################## |
6055f9d4 |
43 | # Initialization |
3c014959 |
44 | ############################################################################## |
69e00e79 |
45 | |
b7ae008f |
46 | # This function handles code blocks. It's registered as a callback to |
47 | # Pod::Simple and therefore doesn't work as a regular method call, but all it |
48 | # does is call output_code with the line. |
49 | sub handle_code { |
50 | my ($line, $number, $parser) = @_; |
51 | $parser->output_code ($line . "\n"); |
52 | } |
53 | |
54 | # Initialize the object and set various Pod::Simple options that we need. |
55 | # Here, we also process any additional options passed to the constructor or |
56 | # set up defaults if none were given. Note that all internal object keys are |
57 | # in all-caps, reserving all lower-case object keys for Pod::Simple and user |
58 | # arguments. |
59 | sub new { |
60 | my $class = shift; |
61 | my $self = $class->SUPER::new; |
62 | |
63 | # Tell Pod::Simple to handle S<> by automatically inserting . |
64 | $self->nbsp_for_S (1); |
65 | |
66 | # Tell Pod::Simple to keep whitespace whenever possible. |
67 | if ($self->can ('preserve_whitespace')) { |
68 | $self->preserve_whitespace (1); |
69 | } else { |
70 | $self->fullstop_space_harden (1); |
71 | } |
69e00e79 |
72 | |
b7ae008f |
73 | # The =for and =begin targets that we accept. |
74 | $self->accept_targets (qw/text TEXT/); |
75 | |
76 | # Ensure that contiguous blocks of code are merged together. Otherwise, |
77 | # some of the guesswork heuristics don't work right. |
78 | $self->merge_text (1); |
79 | |
80 | # Pod::Simple doesn't do anything useful with our arguments, but we want |
81 | # to put them in our object as hash keys and values. This could cause |
82 | # problems if we ever clash with Pod::Simple's own internal class |
83 | # variables. |
84 | my %opts = @_; |
85 | my @opts = map { ("opt_$_", $opts{$_}) } keys %opts; |
86 | %$self = (%$self, @opts); |
87 | |
bc9c7511 |
88 | # Send errors to stderr if requested. |
89 | if ($$self{opt_stderr}) { |
90 | $self->no_errata_section (1); |
91 | $self->complain_stderr (1); |
92 | delete $$self{opt_stderr}; |
93 | } |
94 | |
b7ae008f |
95 | # Initialize various things from our parameters. |
96 | $$self{opt_alt} = 0 unless defined $$self{opt_alt}; |
97 | $$self{opt_indent} = 4 unless defined $$self{opt_indent}; |
98 | $$self{opt_margin} = 0 unless defined $$self{opt_margin}; |
99 | $$self{opt_loose} = 0 unless defined $$self{opt_loose}; |
100 | $$self{opt_sentence} = 0 unless defined $$self{opt_sentence}; |
101 | $$self{opt_width} = 76 unless defined $$self{opt_width}; |
69e00e79 |
102 | |
ab1f1d91 |
103 | # Figure out what quotes we'll be using for C<> text. |
b7ae008f |
104 | $$self{opt_quotes} ||= '"'; |
105 | if ($$self{opt_quotes} eq 'none') { |
ab1f1d91 |
106 | $$self{LQUOTE} = $$self{RQUOTE} = ''; |
b7ae008f |
107 | } elsif (length ($$self{opt_quotes}) == 1) { |
108 | $$self{LQUOTE} = $$self{RQUOTE} = $$self{opt_quotes}; |
109 | } elsif ($$self{opt_quotes} =~ /^(.)(.)$/ |
110 | || $$self{opt_quotes} =~ /^(..)(..)$/) { |
ab1f1d91 |
111 | $$self{LQUOTE} = $1; |
112 | $$self{RQUOTE} = $2; |
113 | } else { |
b7ae008f |
114 | croak qq(Invalid quote specification "$$self{opt_quotes}"); |
ab1f1d91 |
115 | } |
116 | |
b7ae008f |
117 | # If requested, do something with the non-POD text. |
118 | $self->code_handler (\&handle_code) if $$self{opt_code}; |
11f72409 |
119 | |
b7ae008f |
120 | # Return the created object. |
121 | return $self; |
122 | } |
69e00e79 |
123 | |
b7ae008f |
124 | ############################################################################## |
125 | # Core parsing |
126 | ############################################################################## |
59548eca |
127 | |
b7ae008f |
128 | # This is the glue that connects the code below with Pod::Simple itself. The |
129 | # goal is to convert the event stream coming from the POD parser into method |
130 | # calls to handlers once the complete content of a tag has been seen. Each |
131 | # paragraph or POD command will have textual content associated with it, and |
132 | # as soon as all of a paragraph or POD command has been seen, that content |
133 | # will be passed in to the corresponding method for handling that type of |
134 | # object. The exceptions are handlers for lists, which have opening tag |
135 | # handlers and closing tag handlers that will be called right away. |
136 | # |
137 | # The internal hash key PENDING is used to store the contents of a tag until |
138 | # all of it has been seen. It holds a stack of open tags, each one |
139 | # represented by a tuple of the attributes hash for the tag and the contents |
140 | # of the tag. |
141 | |
142 | # Add a block of text to the contents of the current node, formatting it |
143 | # according to the current formatting instructions as we do. |
144 | sub _handle_text { |
145 | my ($self, $text) = @_; |
146 | my $tag = $$self{PENDING}[-1]; |
147 | $$tag[1] .= $text; |
6055f9d4 |
148 | } |
69e00e79 |
149 | |
b7ae008f |
150 | # Given an element name, get the corresponding method name. |
151 | sub method_for_element { |
152 | my ($self, $element) = @_; |
153 | $element =~ tr/-/_/; |
154 | $element =~ tr/A-Z/a-z/; |
155 | $element =~ tr/_a-z0-9//cd; |
156 | return $element; |
157 | } |
69e00e79 |
158 | |
b7ae008f |
159 | # Handle the start of a new element. If cmd_element is defined, assume that |
160 | # we need to collect the entire tree for this element before passing it to the |
161 | # element method, and create a new tree into which we'll collect blocks of |
162 | # text and nested elements. Otherwise, if start_element is defined, call it. |
163 | sub _handle_element_start { |
164 | my ($self, $element, $attrs) = @_; |
165 | my $method = $self->method_for_element ($element); |
166 | |
167 | # If we have a command handler, we need to accumulate the contents of the |
168 | # tag before calling it. |
169 | if ($self->can ("cmd_$method")) { |
170 | push (@{ $$self{PENDING} }, [ $attrs, '' ]); |
171 | } elsif ($self->can ("start_$method")) { |
172 | my $method = 'start_' . $method; |
173 | $self->$method ($attrs, ''); |
174 | } |
175 | } |
6055f9d4 |
176 | |
b7ae008f |
177 | # Handle the end of an element. If we had a cmd_ method for this element, |
178 | # this is where we pass along the text that we've accumulated. Otherwise, if |
179 | # we have an end_ method for the element, call that. |
180 | sub _handle_element_end { |
181 | my ($self, $element) = @_; |
182 | my $method = $self->method_for_element ($element); |
183 | |
184 | # If we have a command handler, pull off the pending text and pass it to |
185 | # the handler along with the saved attribute hash. |
186 | if ($self->can ("cmd_$method")) { |
187 | my $tag = pop @{ $$self{PENDING} }; |
188 | my $method = 'cmd_' . $method; |
189 | my $text = $self->$method (@$tag); |
190 | if (defined $text) { |
191 | if (@{ $$self{PENDING} } > 1) { |
192 | $$self{PENDING}[-1][1] .= $text; |
193 | } else { |
194 | $self->output ($text); |
195 | } |
196 | } |
197 | } elsif ($self->can ("end_$method")) { |
198 | my $method = 'end_' . $method; |
8f202758 |
199 | $self->$method (); |
ab1f1d91 |
200 | } |
6055f9d4 |
201 | } |
69e00e79 |
202 | |
b7ae008f |
203 | ############################################################################## |
204 | # Output formatting |
205 | ############################################################################## |
206 | |
207 | # Wrap a line, indenting by the current left margin. We can't use Text::Wrap |
208 | # because it plays games with tabs. We can't use formline, even though we'd |
209 | # really like to, because it screws up non-printing characters. So we have to |
210 | # do the wrapping ourselves. |
211 | sub wrap { |
6055f9d4 |
212 | my $self = shift; |
6055f9d4 |
213 | local $_ = shift; |
b7ae008f |
214 | my $output = ''; |
215 | my $spaces = ' ' x $$self{MARGIN}; |
216 | my $width = $$self{opt_width} - $$self{MARGIN}; |
217 | while (length > $width) { |
218 | if (s/^([^\n]{0,$width})\s+// || s/^([^\n]{$width})//) { |
219 | $output .= $spaces . $1 . "\n"; |
220 | } else { |
221 | last; |
222 | } |
223 | } |
224 | $output .= $spaces . $_; |
225 | $output =~ s/\s+$/\n\n/; |
226 | return $output; |
6055f9d4 |
227 | } |
69e00e79 |
228 | |
b7ae008f |
229 | # Reformat a paragraph of text for the current margin. Takes the text to |
230 | # reformat and returns the formatted text. |
231 | sub reformat { |
27f805f4 |
232 | my $self = shift; |
27f805f4 |
233 | local $_ = shift; |
6055f9d4 |
234 | |
b7ae008f |
235 | # If we're trying to preserve two spaces after sentences, do some munging |
236 | # to support that. Otherwise, smash all repeated whitespace. |
237 | if ($$self{opt_sentence}) { |
238 | s/ +$//mg; |
239 | s/\.\n/. \n/g; |
240 | s/\n/ /g; |
241 | s/ +/ /g; |
6055f9d4 |
242 | } else { |
b7ae008f |
243 | s/\s+/ /g; |
6055f9d4 |
244 | } |
b7ae008f |
245 | return $self->wrap ($_); |
6055f9d4 |
246 | } |
69e00e79 |
247 | |
bc9c7511 |
248 | # Output text to the output device. Replace non-breaking spaces with spaces |
9f2f055a |
249 | # and soft hyphens with nothing, and then try to fix the output encoding if |
250 | # necessary to match the input encoding unless UTF-8 output is forced. This |
251 | # preserves the traditional pass-through behavior of Pod::Text. |
b7ae008f |
252 | sub output { |
253 | my ($self, $text) = @_; |
254 | $text =~ tr/\240\255/ /d; |
9f2f055a |
255 | unless ($$self{opt_utf8} || $$self{CHECKED_ENCODING}) { |
256 | my $encoding = $$self{encoding} || ''; |
257 | if ($encoding) { |
258 | eval { binmode ($$self{output_fh}, ":encoding($encoding)") }; |
259 | } |
260 | $$self{CHECKED_ENCODING} = 1; |
261 | } |
b7ae008f |
262 | print { $$self{output_fh} } $text; |
263 | } |
bf202ccd |
264 | |
b7ae008f |
265 | # Output a block of code (something that isn't part of the POD text). Called |
266 | # by preprocess_paragraph only if we were given the code option. Exists here |
267 | # only so that it can be overridden by subclasses. |
268 | sub output_code { $_[0]->output ($_[1]) } |
69e00e79 |
269 | |
b7ae008f |
270 | ############################################################################## |
271 | # Document initialization |
272 | ############################################################################## |
273 | |
274 | # Set up various things that have to be initialized on a per-document basis. |
275 | sub start_document { |
276 | my $self = shift; |
277 | my $margin = $$self{opt_indent} + $$self{opt_margin}; |
278 | |
279 | # Initialize a few per-document variables. |
280 | $$self{INDENTS} = []; # Stack of indentations. |
281 | $$self{MARGIN} = $margin; # Default left margin. |
282 | $$self{PENDING} = [[]]; # Pending output. |
283 | |
9f2f055a |
284 | # We have to redo encoding handling for each document. |
285 | delete $$self{CHECKED_ENCODING}; |
286 | |
287 | # If we were given the utf8 option, set an output encoding on our file |
288 | # handle. Wrap in an eval in case we're using a version of Perl too old |
289 | # to understand this. |
290 | # |
291 | # This is evil because it changes the global state of a file handle that |
292 | # we may not own. However, we can't just blindly encode all output, since |
293 | # there may be a pre-applied output encoding (such as from PERL_UNICODE) |
294 | # and then we would double-encode. This seems to be the least bad |
295 | # approach. |
296 | if ($$self{opt_utf8}) { |
297 | eval { binmode ($$self{output_fh}, ':encoding(UTF-8)') }; |
298 | } |
299 | |
b7ae008f |
300 | return ''; |
301 | } |
302 | |
303 | ############################################################################## |
304 | # Text blocks |
305 | ############################################################################## |
306 | |
307 | # This method is called whenever an =item command is complete (in other words, |
308 | # we've seen its associated paragraph or know for certain that it doesn't have |
309 | # one). It gets the paragraph associated with the item as an argument. If |
310 | # that argument is empty, just output the item tag; if it contains a newline, |
311 | # output the item tag followed by the newline. Otherwise, see if there's |
312 | # enough room for us to output the item tag in the margin of the text or if we |
313 | # have to put it on a separate line. |
314 | sub item { |
315 | my ($self, $text) = @_; |
316 | my $tag = $$self{ITEM}; |
317 | unless (defined $tag) { |
318 | carp "Item called without tag"; |
319 | return; |
6055f9d4 |
320 | } |
b7ae008f |
321 | undef $$self{ITEM}; |
69e00e79 |
322 | |
b7ae008f |
323 | # Calculate the indentation and margin. $fits is set to true if the tag |
324 | # will fit into the margin of the paragraph given our indentation level. |
325 | my $indent = $$self{INDENTS}[-1]; |
326 | $indent = $$self{opt_indent} unless defined $indent; |
327 | my $margin = ' ' x $$self{opt_margin}; |
328 | my $fits = ($$self{MARGIN} - $indent >= length ($tag) + 1); |
69e00e79 |
329 | |
b7ae008f |
330 | # If the tag doesn't fit, or if we have no associated text, print out the |
331 | # tag separately. Otherwise, put the tag in the margin of the paragraph. |
332 | if (!$text || $text =~ /^\s+$/ || !$fits) { |
333 | my $realindent = $$self{MARGIN}; |
334 | $$self{MARGIN} = $indent; |
335 | my $output = $self->reformat ($tag); |
336 | $output =~ s/^$margin /$margin:/ if ($$self{opt_alt} && $indent > 0); |
337 | $output =~ s/\n*$/\n/; |
338 | |
339 | # If the text is just whitespace, we have an empty item paragraph; |
340 | # this can result from =over/=item/=back without any intermixed |
341 | # paragraphs. Insert some whitespace to keep the =item from merging |
342 | # into the next paragraph. |
343 | $output .= "\n" if $text && $text =~ /^\s*$/; |
344 | |
345 | $self->output ($output); |
346 | $$self{MARGIN} = $realindent; |
347 | $self->output ($self->reformat ($text)) if ($text && $text =~ /\S/); |
348 | } else { |
349 | my $space = ' ' x $indent; |
350 | $space =~ s/^$margin /$margin:/ if $$self{opt_alt}; |
351 | $text = $self->reformat ($text); |
352 | $text =~ s/^$margin /$margin:/ if ($$self{opt_alt} && $indent > 0); |
353 | my $tagspace = ' ' x length $tag; |
354 | $text =~ s/^($space)$tagspace/$1$tag/ or warn "Bizarre space in item"; |
355 | $self->output ($text); |
6055f9d4 |
356 | } |
b7ae008f |
357 | } |
69e00e79 |
358 | |
b7ae008f |
359 | # Handle a basic block of text. The only tricky thing here is that if there |
360 | # is a pending item tag, we need to format this as an item paragraph. |
361 | sub cmd_para { |
362 | my ($self, $attrs, $text) = @_; |
363 | $text =~ s/\s+$/\n/; |
364 | if (defined $$self{ITEM}) { |
365 | $self->item ($text . "\n"); |
366 | } else { |
367 | $self->output ($self->reformat ($text . "\n")); |
59548eca |
368 | } |
b7ae008f |
369 | return ''; |
6055f9d4 |
370 | } |
f02a87df |
371 | |
b7ae008f |
372 | # Handle a verbatim paragraph. Just print it out, but indent it according to |
373 | # our margin. |
374 | sub cmd_verbatim { |
375 | my ($self, $attrs, $text) = @_; |
376 | $self->item if defined $$self{ITEM}; |
377 | return if $text =~ /^\s*$/; |
378 | $text =~ s/^(\n*)(\s*\S+)/$1 . (' ' x $$self{MARGIN}) . $2/gme; |
379 | $text =~ s/\s*$/\n\n/; |
380 | $self->output ($text); |
381 | return ''; |
6055f9d4 |
382 | } |
3ec07288 |
383 | |
b7ae008f |
384 | # Handle literal text (produced by =for and similar constructs). Just output |
385 | # it with the minimum of changes. |
386 | sub cmd_data { |
387 | my ($self, $attrs, $text) = @_; |
388 | $text =~ s/^\n+//; |
389 | $text =~ s/\n{0,2}$/\n/; |
390 | $self->output ($text); |
391 | return ''; |
392 | } |
69e00e79 |
393 | |
3c014959 |
394 | ############################################################################## |
b7ae008f |
395 | # Headings |
3c014959 |
396 | ############################################################################## |
f2506fb2 |
397 | |
b7ae008f |
398 | # The common code for handling all headers. Takes the header text, the |
399 | # indentation, and the surrounding marker for the alt formatting method. |
400 | sub heading { |
401 | my ($self, $text, $indent, $marker) = @_; |
402 | $self->item ("\n\n") if defined $$self{ITEM}; |
403 | $text =~ s/\s+$//; |
404 | if ($$self{opt_alt}) { |
405 | my $closemark = reverse (split (//, $marker)); |
406 | my $margin = ' ' x $$self{opt_margin}; |
407 | $self->output ("\n" . "$margin$marker $text $closemark" . "\n\n"); |
408 | } else { |
409 | $text .= "\n" if $$self{opt_loose}; |
410 | my $margin = ' ' x ($$self{opt_margin} + $indent); |
411 | $self->output ($margin . $text . "\n"); |
412 | } |
413 | return ''; |
414 | } |
69e00e79 |
415 | |
6055f9d4 |
416 | # First level heading. |
417 | sub cmd_head1 { |
b7ae008f |
418 | my ($self, $attrs, $text) = @_; |
419 | $self->heading ($text, 0, '===='); |
6055f9d4 |
420 | } |
69e00e79 |
421 | |
6055f9d4 |
422 | # Second level heading. |
423 | sub cmd_head2 { |
b7ae008f |
424 | my ($self, $attrs, $text) = @_; |
425 | $self->heading ($text, $$self{opt_indent} / 2, '== '); |
6055f9d4 |
426 | } |
69e00e79 |
427 | |
50a3fd2a |
428 | # Third level heading. |
429 | sub cmd_head3 { |
b7ae008f |
430 | my ($self, $attrs, $text) = @_; |
431 | $self->heading ($text, $$self{opt_indent} * 2 / 3 + 0.5, '= '); |
50a3fd2a |
432 | } |
433 | |
b7ae008f |
434 | # Fourth level heading. |
50a3fd2a |
435 | sub cmd_head4 { |
b7ae008f |
436 | my ($self, $attrs, $text) = @_; |
437 | $self->heading ($text, $$self{opt_indent} * 3 / 4 + 0.5, '- '); |
50a3fd2a |
438 | } |
439 | |
b7ae008f |
440 | ############################################################################## |
441 | # List handling |
442 | ############################################################################## |
443 | |
444 | # Handle the beginning of an =over block. Takes the type of the block as the |
445 | # first argument, and then the attr hash. This is called by the handlers for |
446 | # the four different types of lists (bullet, number, text, and block). |
447 | sub over_common_start { |
448 | my ($self, $attrs) = @_; |
b616daaf |
449 | $self->item ("\n\n") if defined $$self{ITEM}; |
b7ae008f |
450 | |
451 | # Find the indentation level. |
452 | my $indent = $$attrs{indent}; |
453 | unless (defined ($indent) && $indent =~ /^\s*[-+]?\d{1,4}\s*$/) { |
454 | $indent = $$self{opt_indent}; |
455 | } |
456 | |
457 | # Add this to our stack of indents and increase our current margin. |
6055f9d4 |
458 | push (@{ $$self{INDENTS} }, $$self{MARGIN}); |
b7ae008f |
459 | $$self{MARGIN} += ($indent + 0); |
460 | return ''; |
6055f9d4 |
461 | } |
69e00e79 |
462 | |
b7ae008f |
463 | # End an =over block. Takes no options other than the class pointer. Output |
464 | # any pending items and then pop one level of indentation. |
465 | sub over_common_end { |
466 | my ($self) = @_; |
b616daaf |
467 | $self->item ("\n\n") if defined $$self{ITEM}; |
6055f9d4 |
468 | $$self{MARGIN} = pop @{ $$self{INDENTS} }; |
b7ae008f |
469 | return ''; |
69e00e79 |
470 | } |
471 | |
b7ae008f |
472 | # Dispatch the start and end calls as appropriate. |
473 | sub start_over_bullet { $_[0]->over_common_start ($_[1]) } |
474 | sub start_over_number { $_[0]->over_common_start ($_[1]) } |
475 | sub start_over_text { $_[0]->over_common_start ($_[1]) } |
476 | sub start_over_block { $_[0]->over_common_start ($_[1]) } |
477 | sub end_over_bullet { $_[0]->over_common_end } |
478 | sub end_over_number { $_[0]->over_common_end } |
479 | sub end_over_text { $_[0]->over_common_end } |
480 | sub end_over_block { $_[0]->over_common_end } |
481 | |
482 | # The common handler for all item commands. Takes the type of the item, the |
483 | # attributes, and then the text of the item. |
484 | sub item_common { |
485 | my ($self, $type, $attrs, $text) = @_; |
486 | $self->item if defined $$self{ITEM}; |
69e00e79 |
487 | |
b7ae008f |
488 | # Clean up the text. We want to end up with two variables, one ($text) |
489 | # which contains any body text after taking out the item portion, and |
490 | # another ($item) which contains the actual item text. Note the use of |
491 | # the internal Pod::Simple attribute here; that's a potential land mine. |
492 | $text =~ s/\s+$//; |
493 | my ($item, $index); |
494 | if ($type eq 'bullet') { |
495 | $item = '*'; |
496 | } elsif ($type eq 'number') { |
497 | $item = $$attrs{'~orig_content'}; |
27f805f4 |
498 | } else { |
b7ae008f |
499 | $item = $text; |
500 | $item =~ s/\s*\n\s*/ /g; |
501 | $text = ''; |
27f805f4 |
502 | } |
b7ae008f |
503 | $$self{ITEM} = $item; |
6055f9d4 |
504 | |
b7ae008f |
505 | # If body text for this item was included, go ahead and output that now. |
506 | if ($text) { |
507 | $text =~ s/\s*$/\n/; |
508 | $self->item ($text); |
509 | } |
510 | return ''; |
6055f9d4 |
511 | } |
f2506fb2 |
512 | |
b7ae008f |
513 | # Dispatch the item commands to the appropriate place. |
514 | sub cmd_item_bullet { my $self = shift; $self->item_common ('bullet', @_) } |
515 | sub cmd_item_number { my $self = shift; $self->item_common ('number', @_) } |
516 | sub cmd_item_text { my $self = shift; $self->item_common ('text', @_) } |
517 | sub cmd_item_block { my $self = shift; $self->item_common ('block', @_) } |
69e00e79 |
518 | |
3c014959 |
519 | ############################################################################## |
5ec554fb |
520 | # Formatting codes |
3c014959 |
521 | ############################################################################## |
69e00e79 |
522 | |
b7ae008f |
523 | # The simple ones. |
524 | sub cmd_b { return $_[0]{alt} ? "``$_[2]''" : $_[2] } |
525 | sub cmd_f { return $_[0]{alt} ? "\"$_[2]\"" : $_[2] } |
526 | sub cmd_i { return '*' . $_[2] . '*' } |
527 | sub cmd_x { return '' } |
3c014959 |
528 | |
529 | # Apply a whole bunch of messy heuristics to not quote things that don't |
530 | # benefit from being quoted. These originally come from Barrie Slaymaker and |
531 | # largely duplicate code in Pod::Man. |
b7ae008f |
532 | sub cmd_c { |
533 | my ($self, $attrs, $text) = @_; |
3c014959 |
534 | |
535 | # A regex that matches the portion of a variable reference that's the |
536 | # array or hash index, separated out just because we want to use it in |
537 | # several places in the following regex. |
538 | my $index = '(?: \[.*\] | \{.*\} )?'; |
539 | |
540 | # Check for things that we don't want to quote, and if we find any of |
541 | # them, return the string with just a font change and no quoting. |
b7ae008f |
542 | $text =~ m{ |
3c014959 |
543 | ^\s* |
544 | (?: |
545 | ( [\'\`\"] ) .* \1 # already quoted |
546 | | \` .* \' # `quoted' |
547 | | \$+ [\#^]? \S $index # special ($^Foo, $") |
548 | | [\$\@%&*]+ \#? [:\'\w]+ $index # plain var or func |
549 | | [\$\@%&*]* [:\'\w]+ (?: -> )? \(\s*[^\s,]\s*\) # 0/1-arg func call |
f011ec7d |
550 | | [+-]? ( \d[\d.]* | \.\d+ ) (?: [eE][+-]?\d+ )? # a number |
3c014959 |
551 | | 0x [a-fA-F\d]+ # a hex constant |
552 | ) |
553 | \s*\z |
b7ae008f |
554 | }xo && return $text; |
3c014959 |
555 | |
556 | # If we didn't return, go ahead and quote the text. |
b7ae008f |
557 | return $$self{opt_alt} |
558 | ? "``$text''" |
559 | : "$$self{LQUOTE}$text$$self{RQUOTE}"; |
69e00e79 |
560 | } |
561 | |
b7ae008f |
562 | # Links reduce to the text that we're given, wrapped in angle brackets if it's |
563 | # a URL. |
564 | sub cmd_l { |
565 | my ($self, $attrs, $text) = @_; |
566 | return $$attrs{type} eq 'url' ? "<$text>" : $text; |
b616daaf |
567 | } |
568 | |
3c014959 |
569 | ############################################################################## |
27f805f4 |
570 | # Backwards compatibility |
3c014959 |
571 | ############################################################################## |
27f805f4 |
572 | |
573 | # The old Pod::Text module did everything in a pod2text() function. This |
574 | # tries to provide the same interface for legacy applications. |
575 | sub pod2text { |
576 | my @args; |
577 | |
578 | # This is really ugly; I hate doing option parsing in the middle of a |
579 | # module. But the old Pod::Text module supported passing flags to its |
580 | # entry function, so handle -a and -<number>. |
581 | while ($_[0] =~ /^-/) { |
582 | my $flag = shift; |
583 | if ($flag eq '-a') { push (@args, alt => 1) } |
584 | elsif ($flag =~ /^-(\d+)$/) { push (@args, width => $1) } |
585 | else { |
586 | unshift (@_, $flag); |
587 | last; |
588 | } |
589 | } |
590 | |
591 | # Now that we know what arguments we're using, create the parser. |
592 | my $parser = Pod::Text->new (@args); |
593 | |
594 | # If two arguments were given, the second argument is going to be a file |
3c014959 |
595 | # handle. That means we want to call parse_from_filehandle(), which means |
596 | # we need to turn the first argument into a file handle. Magic open will |
597 | # handle the <&STDIN case automagically. |
27f805f4 |
598 | if (defined $_[1]) { |
ab1f1d91 |
599 | my @fhs = @_; |
27f805f4 |
600 | local *IN; |
ab1f1d91 |
601 | unless (open (IN, $fhs[0])) { |
602 | croak ("Can't open $fhs[0] for reading: $!\n"); |
27f805f4 |
603 | return; |
604 | } |
ab1f1d91 |
605 | $fhs[0] = \*IN; |
8f202758 |
606 | $parser->output_fh ($fhs[1]); |
607 | my $retval = $parser->parse_file ($fhs[0]); |
608 | my $fh = $parser->output_fh (); |
609 | close $fh; |
610 | return $retval; |
27f805f4 |
611 | } else { |
b7ae008f |
612 | return $parser->parse_file (@_); |
27f805f4 |
613 | } |
614 | } |
615 | |
8f202758 |
616 | # Reset the underlying Pod::Simple object between calls to parse_from_file so |
617 | # that the same object can be reused to convert multiple pages. |
618 | sub parse_from_file { |
619 | my $self = shift; |
620 | $self->reinit; |
42ae9e1d |
621 | |
622 | # Fake the old cutting option to Pod::Parser. This fiddings with internal |
623 | # Pod::Simple state and is quite ugly; we need a better approach. |
624 | if (ref ($_[0]) eq 'HASH') { |
625 | my $opts = shift @_; |
626 | if (defined ($$opts{-cutting}) && !$$opts{-cutting}) { |
627 | $$self{in_pod} = 1; |
628 | $$self{last_was_blank} = 1; |
629 | } |
630 | } |
631 | |
632 | # Do the work. |
8f202758 |
633 | my $retval = $self->Pod::Simple::parse_from_file (@_); |
42ae9e1d |
634 | |
635 | # Flush output, since Pod::Simple doesn't do this. Ideally we should also |
636 | # close the file descriptor if we had to open one, but we can't easily |
637 | # figure this out. |
8f202758 |
638 | my $fh = $self->output_fh (); |
639 | my $oldfh = select $fh; |
640 | my $oldflush = $|; |
641 | $| = 1; |
642 | print $fh ''; |
643 | $| = $oldflush; |
644 | select $oldfh; |
645 | return $retval; |
646 | } |
647 | |
fcf69717 |
648 | # Pod::Simple failed to provide this backward compatibility function, so |
649 | # implement it ourselves. File handles are one of the inputs that |
650 | # parse_from_file supports. |
651 | sub parse_from_filehandle { |
652 | my $self = shift; |
653 | $self->parse_from_file (@_); |
654 | } |
655 | |
3c014959 |
656 | ############################################################################## |
6055f9d4 |
657 | # Module return value and documentation |
3c014959 |
658 | ############################################################################## |
69e00e79 |
659 | |
6055f9d4 |
660 | 1; |
661 | __END__ |
69e00e79 |
662 | |
6055f9d4 |
663 | =head1 NAME |
69e00e79 |
664 | |
6055f9d4 |
665 | Pod::Text - Convert POD data to formatted ASCII text |
69e00e79 |
666 | |
0e4e3f6e |
667 | =for stopwords |
9f2f055a |
668 | alt stderr Allbery Sean Burke's Christiansen UTF-8 pre-Unicode utf8 |
0e4e3f6e |
669 | |
6055f9d4 |
670 | =head1 SYNOPSIS |
69e00e79 |
671 | |
6055f9d4 |
672 | use Pod::Text; |
673 | my $parser = Pod::Text->new (sentence => 0, width => 78); |
69e00e79 |
674 | |
6055f9d4 |
675 | # Read POD from STDIN and write to STDOUT. |
676 | $parser->parse_from_filehandle; |
69e00e79 |
677 | |
6055f9d4 |
678 | # Read POD from file.pod and write to file.txt. |
679 | $parser->parse_from_file ('file.pod', 'file.txt'); |
69e00e79 |
680 | |
6055f9d4 |
681 | =head1 DESCRIPTION |
5491a304 |
682 | |
27f805f4 |
683 | Pod::Text is a module that can convert documentation in the POD format (the |
684 | preferred language for documenting Perl) into formatted ASCII. It uses no |
685 | special formatting controls or codes whatsoever, and its output is therefore |
686 | suitable for nearly any device. |
69e00e79 |
687 | |
b7ae008f |
688 | As a derived class from Pod::Simple, Pod::Text supports the same methods and |
689 | interfaces. See L<Pod::Simple> for all the details; briefly, one creates a |
690 | new parser with C<< Pod::Text->new() >> and then normally calls parse_file(). |
6055f9d4 |
691 | |
27f805f4 |
692 | new() can take options, in the form of key/value pairs, that control the |
6055f9d4 |
693 | behavior of the parser. The currently recognized options are: |
694 | |
695 | =over 4 |
696 | |
697 | =item alt |
698 | |
699 | If set to a true value, selects an alternate output format that, among other |
700 | things, uses a different heading style and marks C<=item> entries with a |
701 | colon in the left margin. Defaults to false. |
702 | |
59548eca |
703 | =item code |
704 | |
705 | If set to a true value, the non-POD parts of the input file will be included |
706 | in the output. Useful for viewing code documented with POD blocks with the |
707 | POD rendered and the code left intact. |
708 | |
6055f9d4 |
709 | =item indent |
710 | |
711 | The number of spaces to indent regular text, and the default indentation for |
712 | C<=over> blocks. Defaults to 4. |
713 | |
714 | =item loose |
715 | |
716 | If set to a true value, a blank line is printed after a C<=head1> heading. |
717 | If set to false (the default), no blank line is printed after C<=head1>, |
718 | although one is still printed after C<=head2>. This is the default because |
719 | it's the expected formatting for manual pages; if you're formatting |
720 | arbitrary text documents, setting this to true may result in more pleasing |
721 | output. |
722 | |
11f72409 |
723 | =item margin |
724 | |
725 | The width of the left margin in spaces. Defaults to 0. This is the margin |
726 | for all text, including headings, not the amount by which regular text is |
727 | indented; for the latter, see the I<indent> option. To set the right |
728 | margin, see the I<width> option. |
729 | |
ab1f1d91 |
730 | =item quotes |
731 | |
732 | Sets the quote marks used to surround CE<lt>> text. If the value is a |
733 | single character, it is used as both the left and right quote; if it is two |
734 | characters, the first character is used as the left quote and the second as |
735 | the right quoted; and if it is four characters, the first two are used as |
736 | the left quote and the second two as the right quote. |
737 | |
738 | This may also be set to the special value C<none>, in which case no quote |
739 | marks are added around CE<lt>> text. |
740 | |
6055f9d4 |
741 | =item sentence |
742 | |
27f805f4 |
743 | If set to a true value, Pod::Text will assume that each sentence ends in two |
744 | spaces, and will try to preserve that spacing. If set to false, all |
6055f9d4 |
745 | consecutive whitespace in non-verbatim paragraphs is compressed into a |
746 | single space. Defaults to true. |
747 | |
bc9c7511 |
748 | =item stderr |
749 | |
750 | Send error messages about invalid POD to standard error instead of |
751 | appending a POD ERRORS section to the generated output. |
752 | |
9f2f055a |
753 | =item utf8 |
754 | |
755 | By default, Pod::Text uses the same output encoding as the input encoding |
756 | of the POD source (provided that Perl was built with PerlIO; otherwise, it |
757 | doesn't encode its output). If this option is given, the output encoding |
758 | is forced to UTF-8. |
759 | |
760 | Be aware that, when using this option, the input encoding of your POD |
761 | source must be properly declared unless it is US-ASCII or Latin-1. POD |
762 | input without an C<=encoding> command will be assumed to be in Latin-1, |
763 | and if it's actually in UTF-8, the output will be double-encoded. See |
764 | L<perlpod(1)> for more information on the C<=encoding> command. |
765 | |
6055f9d4 |
766 | =item width |
767 | |
768 | The column at which to wrap text on the right-hand side. Defaults to 76. |
769 | |
770 | =back |
771 | |
b7ae008f |
772 | The standard Pod::Simple method parse_file() takes one argument, the file or |
773 | file handle to read from, and writes output to standard output unless that |
774 | has been changed with the output_fh() method. See L<Pod::Simple> for the |
775 | specific details and for other alternative interfaces. |
6055f9d4 |
776 | |
777 | =head1 DIAGNOSTICS |
778 | |
779 | =over 4 |
780 | |
27f805f4 |
781 | =item Bizarre space in item |
782 | |
59548eca |
783 | =item Item called without tag |
784 | |
785 | (W) Something has gone wrong in internal C<=item> processing. These |
786 | messages indicate a bug in Pod::Text; you should never see them. |
27f805f4 |
787 | |
788 | =item Can't open %s for reading: %s |
789 | |
790 | (F) Pod::Text was invoked via the compatibility mode pod2text() interface |
791 | and the input file it was given could not be opened. |
792 | |
ab1f1d91 |
793 | =item Invalid quote specification "%s" |
794 | |
795 | (F) The quote specification given (the quotes option to the constructor) was |
796 | invalid. A quote specification must be one, two, or four characters long. |
797 | |
6055f9d4 |
798 | =back |
799 | |
9f2f055a |
800 | =head1 BUGS |
801 | |
802 | Encoding handling assumes that PerlIO is available and does not work |
803 | properly if it isn't. The C<utf8> option is therefore not supported |
804 | unless Perl is built with PerlIO support. |
805 | |
806 | =head1 CAVEATS |
807 | |
808 | If Pod::Text is given the C<utf8> option, the encoding of its output file |
809 | handle will be forced to UTF-8 if possible, overriding any existing |
810 | encoding. This will be done even if the file handle is not created by |
811 | Pod::Text and was passed in from outside. This maintains consistency |
812 | regardless of PERL_UNICODE and other settings. |
813 | |
814 | If the C<utf8> option is not given, the encoding of its output file handle |
815 | will be forced to the detected encoding of the input POD, which preserves |
816 | whatever the input text is. This ensures backward compatibility with |
817 | earlier, pre-Unicode versions of this module, without large numbers of |
818 | Perl warnings. |
819 | |
820 | This is not ideal, but it seems to be the best compromise. If it doesn't |
821 | work for you, please let me know the details of how it broke. |
822 | |
6055f9d4 |
823 | =head1 NOTES |
824 | |
27f805f4 |
825 | This is a replacement for an earlier Pod::Text module written by Tom |
b7ae008f |
826 | Christiansen. It has a revamped interface, since it now uses Pod::Simple, |
27f805f4 |
827 | but an interface roughly compatible with the old Pod::Text::pod2text() |
828 | function is still available. Please change to the new calling convention, |
829 | though. |
6055f9d4 |
830 | |
831 | The original Pod::Text contained code to do formatting via termcap |
832 | sequences, although it wasn't turned on by default and it was problematic to |
27f805f4 |
833 | get it to work at all. This rewrite doesn't even try to do that, but a |
bf202ccd |
834 | subclass of it does. Look for L<Pod::Text::Termcap>. |
6055f9d4 |
835 | |
836 | =head1 SEE ALSO |
837 | |
9f2f055a |
838 | L<Pod::Simple>, L<Pod::Text::Termcap>, L<perlpod(1)>, L<pod2text(1)> |
6055f9d4 |
839 | |
fd20da51 |
840 | The current version of this module is always available from its web site at |
841 | L<http://www.eyrie.org/~eagle/software/podlators/>. It is also part of the |
842 | Perl core distribution as of 5.6.0. |
843 | |
6055f9d4 |
844 | =head1 AUTHOR |
845 | |
bf202ccd |
846 | Russ Allbery <rra@stanford.edu>, based I<very> heavily on the original |
847 | Pod::Text by Tom Christiansen <tchrist@mox.perl.com> and its conversion to |
b7ae008f |
848 | Pod::Parser by Brad Appleton <bradapp@enteract.com>. Sean Burke's initial |
849 | conversion of Pod::Man to use Pod::Simple provided much-needed guidance on |
850 | how to use Pod::Simple. |
6055f9d4 |
851 | |
3c014959 |
852 | =head1 COPYRIGHT AND LICENSE |
853 | |
0e4e3f6e |
854 | Copyright 1999, 2000, 2001, 2002, 2004, 2006, 2008 Russ Allbery |
855 | <rra@stanford.edu>. |
3c014959 |
856 | |
857 | This program is free software; you may redistribute it and/or modify it |
858 | under the same terms as Perl itself. |
859 | |
6055f9d4 |
860 | =cut |