Commit | Line | Data |
---|---|---|
6055f9d4 | 1 | # Pod::Text -- Convert POD data to formatted ASCII text. |
6055f9d4 | 2 | # |
8f202758 SP |
3 | # Copyright 1999, 2000, 2001, 2002, 2004, 2006 |
4 | # by Russ Allbery <rra@stanford.edu> | |
6055f9d4 | 5 | # |
3c014959 | 6 | # This program is free software; you may redistribute it and/or modify it |
6055f9d4 GS |
7 | # under the same terms as Perl itself. |
8 | # | |
5ec554fb JH |
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 JH |
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 GS |
26 | require 5.004; |
27 | ||
6055f9d4 | 28 | use strict; |
2e20e14f | 29 | use vars qw(@ISA @EXPORT %ESCAPES $VERSION); |
6055f9d4 | 30 | |
b7ae008f SP |
31 | use Carp qw(carp croak); |
32 | use Exporter (); | |
33 | use Pod::Simple (); | |
34 | ||
35 | @ISA = qw(Pod::Simple Exporter); | |
6055f9d4 | 36 | |
2e20e14f GS |
37 | # We have to export pod2text for backward compatibility. |
38 | @EXPORT = qw(pod2text); | |
39 | ||
3c014959 JH |
40 | # Don't use the CVS revision as the version, since this module is also in Perl |
41 | # core and too many things could munge CVS magic revision strings. This | |
42 | # number should ideally be the same as the CVS revision in podlators, however. | |
42ae9e1d | 43 | $VERSION = 3.08; |
69e00e79 | 44 | |
3c014959 | 45 | ############################################################################## |
6055f9d4 | 46 | # Initialization |
3c014959 | 47 | ############################################################################## |
69e00e79 | 48 | |
b7ae008f SP |
49 | # This function handles code blocks. It's registered as a callback to |
50 | # Pod::Simple and therefore doesn't work as a regular method call, but all it | |
51 | # does is call output_code with the line. | |
52 | sub handle_code { | |
53 | my ($line, $number, $parser) = @_; | |
54 | $parser->output_code ($line . "\n"); | |
55 | } | |
56 | ||
57 | # Initialize the object and set various Pod::Simple options that we need. | |
58 | # Here, we also process any additional options passed to the constructor or | |
59 | # set up defaults if none were given. Note that all internal object keys are | |
60 | # in all-caps, reserving all lower-case object keys for Pod::Simple and user | |
61 | # arguments. | |
62 | sub new { | |
63 | my $class = shift; | |
64 | my $self = $class->SUPER::new; | |
65 | ||
66 | # Tell Pod::Simple to handle S<> by automatically inserting . | |
67 | $self->nbsp_for_S (1); | |
68 | ||
69 | # Tell Pod::Simple to keep whitespace whenever possible. | |
70 | if ($self->can ('preserve_whitespace')) { | |
71 | $self->preserve_whitespace (1); | |
72 | } else { | |
73 | $self->fullstop_space_harden (1); | |
74 | } | |
69e00e79 | 75 | |
b7ae008f SP |
76 | # The =for and =begin targets that we accept. |
77 | $self->accept_targets (qw/text TEXT/); | |
78 | ||
79 | # Ensure that contiguous blocks of code are merged together. Otherwise, | |
80 | # some of the guesswork heuristics don't work right. | |
81 | $self->merge_text (1); | |
82 | ||
83 | # Pod::Simple doesn't do anything useful with our arguments, but we want | |
84 | # to put them in our object as hash keys and values. This could cause | |
85 | # problems if we ever clash with Pod::Simple's own internal class | |
86 | # variables. | |
87 | my %opts = @_; | |
88 | my @opts = map { ("opt_$_", $opts{$_}) } keys %opts; | |
89 | %$self = (%$self, @opts); | |
90 | ||
91 | # Initialize various things from our parameters. | |
92 | $$self{opt_alt} = 0 unless defined $$self{opt_alt}; | |
93 | $$self{opt_indent} = 4 unless defined $$self{opt_indent}; | |
94 | $$self{opt_margin} = 0 unless defined $$self{opt_margin}; | |
95 | $$self{opt_loose} = 0 unless defined $$self{opt_loose}; | |
96 | $$self{opt_sentence} = 0 unless defined $$self{opt_sentence}; | |
97 | $$self{opt_width} = 76 unless defined $$self{opt_width}; | |
69e00e79 | 98 | |
ab1f1d91 | 99 | # Figure out what quotes we'll be using for C<> text. |
b7ae008f SP |
100 | $$self{opt_quotes} ||= '"'; |
101 | if ($$self{opt_quotes} eq 'none') { | |
ab1f1d91 | 102 | $$self{LQUOTE} = $$self{RQUOTE} = ''; |
b7ae008f SP |
103 | } elsif (length ($$self{opt_quotes}) == 1) { |
104 | $$self{LQUOTE} = $$self{RQUOTE} = $$self{opt_quotes}; | |
105 | } elsif ($$self{opt_quotes} =~ /^(.)(.)$/ | |
106 | || $$self{opt_quotes} =~ /^(..)(..)$/) { | |
ab1f1d91 JH |
107 | $$self{LQUOTE} = $1; |
108 | $$self{RQUOTE} = $2; | |
109 | } else { | |
b7ae008f | 110 | croak qq(Invalid quote specification "$$self{opt_quotes}"); |
ab1f1d91 JH |
111 | } |
112 | ||
b7ae008f SP |
113 | # If requested, do something with the non-POD text. |
114 | $self->code_handler (\&handle_code) if $$self{opt_code}; | |
11f72409 | 115 | |
b7ae008f SP |
116 | # Return the created object. |
117 | return $self; | |
118 | } | |
69e00e79 | 119 | |
b7ae008f SP |
120 | ############################################################################## |
121 | # Core parsing | |
122 | ############################################################################## | |
59548eca | 123 | |
b7ae008f SP |
124 | # This is the glue that connects the code below with Pod::Simple itself. The |
125 | # goal is to convert the event stream coming from the POD parser into method | |
126 | # calls to handlers once the complete content of a tag has been seen. Each | |
127 | # paragraph or POD command will have textual content associated with it, and | |
128 | # as soon as all of a paragraph or POD command has been seen, that content | |
129 | # will be passed in to the corresponding method for handling that type of | |
130 | # object. The exceptions are handlers for lists, which have opening tag | |
131 | # handlers and closing tag handlers that will be called right away. | |
132 | # | |
133 | # The internal hash key PENDING is used to store the contents of a tag until | |
134 | # all of it has been seen. It holds a stack of open tags, each one | |
135 | # represented by a tuple of the attributes hash for the tag and the contents | |
136 | # of the tag. | |
137 | ||
138 | # Add a block of text to the contents of the current node, formatting it | |
139 | # according to the current formatting instructions as we do. | |
140 | sub _handle_text { | |
141 | my ($self, $text) = @_; | |
142 | my $tag = $$self{PENDING}[-1]; | |
143 | $$tag[1] .= $text; | |
6055f9d4 | 144 | } |
69e00e79 | 145 | |
b7ae008f SP |
146 | # Given an element name, get the corresponding method name. |
147 | sub method_for_element { | |
148 | my ($self, $element) = @_; | |
149 | $element =~ tr/-/_/; | |
150 | $element =~ tr/A-Z/a-z/; | |
151 | $element =~ tr/_a-z0-9//cd; | |
152 | return $element; | |
153 | } | |
69e00e79 | 154 | |
b7ae008f SP |
155 | # Handle the start of a new element. If cmd_element is defined, assume that |
156 | # we need to collect the entire tree for this element before passing it to the | |
157 | # element method, and create a new tree into which we'll collect blocks of | |
158 | # text and nested elements. Otherwise, if start_element is defined, call it. | |
159 | sub _handle_element_start { | |
160 | my ($self, $element, $attrs) = @_; | |
161 | my $method = $self->method_for_element ($element); | |
162 | ||
163 | # If we have a command handler, we need to accumulate the contents of the | |
164 | # tag before calling it. | |
165 | if ($self->can ("cmd_$method")) { | |
166 | push (@{ $$self{PENDING} }, [ $attrs, '' ]); | |
167 | } elsif ($self->can ("start_$method")) { | |
168 | my $method = 'start_' . $method; | |
169 | $self->$method ($attrs, ''); | |
170 | } | |
171 | } | |
6055f9d4 | 172 | |
b7ae008f SP |
173 | # Handle the end of an element. If we had a cmd_ method for this element, |
174 | # this is where we pass along the text that we've accumulated. Otherwise, if | |
175 | # we have an end_ method for the element, call that. | |
176 | sub _handle_element_end { | |
177 | my ($self, $element) = @_; | |
178 | my $method = $self->method_for_element ($element); | |
179 | ||
180 | # If we have a command handler, pull off the pending text and pass it to | |
181 | # the handler along with the saved attribute hash. | |
182 | if ($self->can ("cmd_$method")) { | |
183 | my $tag = pop @{ $$self{PENDING} }; | |
184 | my $method = 'cmd_' . $method; | |
185 | my $text = $self->$method (@$tag); | |
186 | if (defined $text) { | |
187 | if (@{ $$self{PENDING} } > 1) { | |
188 | $$self{PENDING}[-1][1] .= $text; | |
189 | } else { | |
190 | $self->output ($text); | |
191 | } | |
192 | } | |
193 | } elsif ($self->can ("end_$method")) { | |
194 | my $method = 'end_' . $method; | |
8f202758 | 195 | $self->$method (); |
ab1f1d91 | 196 | } |
6055f9d4 | 197 | } |
69e00e79 | 198 | |
b7ae008f SP |
199 | ############################################################################## |
200 | # Output formatting | |
201 | ############################################################################## | |
202 | ||
203 | # Wrap a line, indenting by the current left margin. We can't use Text::Wrap | |
204 | # because it plays games with tabs. We can't use formline, even though we'd | |
205 | # really like to, because it screws up non-printing characters. So we have to | |
206 | # do the wrapping ourselves. | |
207 | sub wrap { | |
6055f9d4 | 208 | my $self = shift; |
6055f9d4 | 209 | local $_ = shift; |
b7ae008f SP |
210 | my $output = ''; |
211 | my $spaces = ' ' x $$self{MARGIN}; | |
212 | my $width = $$self{opt_width} - $$self{MARGIN}; | |
213 | while (length > $width) { | |
214 | if (s/^([^\n]{0,$width})\s+// || s/^([^\n]{$width})//) { | |
215 | $output .= $spaces . $1 . "\n"; | |
216 | } else { | |
217 | last; | |
218 | } | |
219 | } | |
220 | $output .= $spaces . $_; | |
221 | $output =~ s/\s+$/\n\n/; | |
222 | return $output; | |
6055f9d4 | 223 | } |
69e00e79 | 224 | |
b7ae008f SP |
225 | # Reformat a paragraph of text for the current margin. Takes the text to |
226 | # reformat and returns the formatted text. | |
227 | sub reformat { | |
27f805f4 | 228 | my $self = shift; |
27f805f4 | 229 | local $_ = shift; |
6055f9d4 | 230 | |
b7ae008f SP |
231 | # If we're trying to preserve two spaces after sentences, do some munging |
232 | # to support that. Otherwise, smash all repeated whitespace. | |
233 | if ($$self{opt_sentence}) { | |
234 | s/ +$//mg; | |
235 | s/\.\n/. \n/g; | |
236 | s/\n/ /g; | |
237 | s/ +/ /g; | |
6055f9d4 | 238 | } else { |
b7ae008f | 239 | s/\s+/ /g; |
6055f9d4 | 240 | } |
b7ae008f | 241 | return $self->wrap ($_); |
6055f9d4 | 242 | } |
69e00e79 | 243 | |
b7ae008f SP |
244 | # Output text to the output device. |
245 | sub output { | |
246 | my ($self, $text) = @_; | |
247 | $text =~ tr/\240\255/ /d; | |
248 | print { $$self{output_fh} } $text; | |
249 | } | |
bf202ccd | 250 | |
b7ae008f SP |
251 | # Output a block of code (something that isn't part of the POD text). Called |
252 | # by preprocess_paragraph only if we were given the code option. Exists here | |
253 | # only so that it can be overridden by subclasses. | |
254 | sub output_code { $_[0]->output ($_[1]) } | |
69e00e79 | 255 | |
b7ae008f SP |
256 | ############################################################################## |
257 | # Document initialization | |
258 | ############################################################################## | |
259 | ||
260 | # Set up various things that have to be initialized on a per-document basis. | |
261 | sub start_document { | |
262 | my $self = shift; | |
263 | my $margin = $$self{opt_indent} + $$self{opt_margin}; | |
264 | ||
265 | # Initialize a few per-document variables. | |
266 | $$self{INDENTS} = []; # Stack of indentations. | |
267 | $$self{MARGIN} = $margin; # Default left margin. | |
268 | $$self{PENDING} = [[]]; # Pending output. | |
269 | ||
270 | return ''; | |
271 | } | |
272 | ||
273 | ############################################################################## | |
274 | # Text blocks | |
275 | ############################################################################## | |
276 | ||
277 | # This method is called whenever an =item command is complete (in other words, | |
278 | # we've seen its associated paragraph or know for certain that it doesn't have | |
279 | # one). It gets the paragraph associated with the item as an argument. If | |
280 | # that argument is empty, just output the item tag; if it contains a newline, | |
281 | # output the item tag followed by the newline. Otherwise, see if there's | |
282 | # enough room for us to output the item tag in the margin of the text or if we | |
283 | # have to put it on a separate line. | |
284 | sub item { | |
285 | my ($self, $text) = @_; | |
286 | my $tag = $$self{ITEM}; | |
287 | unless (defined $tag) { | |
288 | carp "Item called without tag"; | |
289 | return; | |
6055f9d4 | 290 | } |
b7ae008f | 291 | undef $$self{ITEM}; |
69e00e79 | 292 | |
b7ae008f SP |
293 | # Calculate the indentation and margin. $fits is set to true if the tag |
294 | # will fit into the margin of the paragraph given our indentation level. | |
295 | my $indent = $$self{INDENTS}[-1]; | |
296 | $indent = $$self{opt_indent} unless defined $indent; | |
297 | my $margin = ' ' x $$self{opt_margin}; | |
298 | my $fits = ($$self{MARGIN} - $indent >= length ($tag) + 1); | |
69e00e79 | 299 | |
b7ae008f SP |
300 | # If the tag doesn't fit, or if we have no associated text, print out the |
301 | # tag separately. Otherwise, put the tag in the margin of the paragraph. | |
302 | if (!$text || $text =~ /^\s+$/ || !$fits) { | |
303 | my $realindent = $$self{MARGIN}; | |
304 | $$self{MARGIN} = $indent; | |
305 | my $output = $self->reformat ($tag); | |
306 | $output =~ s/^$margin /$margin:/ if ($$self{opt_alt} && $indent > 0); | |
307 | $output =~ s/\n*$/\n/; | |
308 | ||
309 | # If the text is just whitespace, we have an empty item paragraph; | |
310 | # this can result from =over/=item/=back without any intermixed | |
311 | # paragraphs. Insert some whitespace to keep the =item from merging | |
312 | # into the next paragraph. | |
313 | $output .= "\n" if $text && $text =~ /^\s*$/; | |
314 | ||
315 | $self->output ($output); | |
316 | $$self{MARGIN} = $realindent; | |
317 | $self->output ($self->reformat ($text)) if ($text && $text =~ /\S/); | |
318 | } else { | |
319 | my $space = ' ' x $indent; | |
320 | $space =~ s/^$margin /$margin:/ if $$self{opt_alt}; | |
321 | $text = $self->reformat ($text); | |
322 | $text =~ s/^$margin /$margin:/ if ($$self{opt_alt} && $indent > 0); | |
323 | my $tagspace = ' ' x length $tag; | |
324 | $text =~ s/^($space)$tagspace/$1$tag/ or warn "Bizarre space in item"; | |
325 | $self->output ($text); | |
6055f9d4 | 326 | } |
b7ae008f | 327 | } |
69e00e79 | 328 | |
b7ae008f SP |
329 | # Handle a basic block of text. The only tricky thing here is that if there |
330 | # is a pending item tag, we need to format this as an item paragraph. | |
331 | sub cmd_para { | |
332 | my ($self, $attrs, $text) = @_; | |
333 | $text =~ s/\s+$/\n/; | |
334 | if (defined $$self{ITEM}) { | |
335 | $self->item ($text . "\n"); | |
336 | } else { | |
337 | $self->output ($self->reformat ($text . "\n")); | |
59548eca | 338 | } |
b7ae008f | 339 | return ''; |
6055f9d4 | 340 | } |
f02a87df | 341 | |
b7ae008f SP |
342 | # Handle a verbatim paragraph. Just print it out, but indent it according to |
343 | # our margin. | |
344 | sub cmd_verbatim { | |
345 | my ($self, $attrs, $text) = @_; | |
346 | $self->item if defined $$self{ITEM}; | |
347 | return if $text =~ /^\s*$/; | |
348 | $text =~ s/^(\n*)(\s*\S+)/$1 . (' ' x $$self{MARGIN}) . $2/gme; | |
349 | $text =~ s/\s*$/\n\n/; | |
350 | $self->output ($text); | |
351 | return ''; | |
6055f9d4 | 352 | } |
3ec07288 | 353 | |
b7ae008f SP |
354 | # Handle literal text (produced by =for and similar constructs). Just output |
355 | # it with the minimum of changes. | |
356 | sub cmd_data { | |
357 | my ($self, $attrs, $text) = @_; | |
358 | $text =~ s/^\n+//; | |
359 | $text =~ s/\n{0,2}$/\n/; | |
360 | $self->output ($text); | |
361 | return ''; | |
362 | } | |
69e00e79 | 363 | |
3c014959 | 364 | ############################################################################## |
b7ae008f | 365 | # Headings |
3c014959 | 366 | ############################################################################## |
f2506fb2 | 367 | |
b7ae008f SP |
368 | # The common code for handling all headers. Takes the header text, the |
369 | # indentation, and the surrounding marker for the alt formatting method. | |
370 | sub heading { | |
371 | my ($self, $text, $indent, $marker) = @_; | |
372 | $self->item ("\n\n") if defined $$self{ITEM}; | |
373 | $text =~ s/\s+$//; | |
374 | if ($$self{opt_alt}) { | |
375 | my $closemark = reverse (split (//, $marker)); | |
376 | my $margin = ' ' x $$self{opt_margin}; | |
377 | $self->output ("\n" . "$margin$marker $text $closemark" . "\n\n"); | |
378 | } else { | |
379 | $text .= "\n" if $$self{opt_loose}; | |
380 | my $margin = ' ' x ($$self{opt_margin} + $indent); | |
381 | $self->output ($margin . $text . "\n"); | |
382 | } | |
383 | return ''; | |
384 | } | |
69e00e79 | 385 | |
6055f9d4 GS |
386 | # First level heading. |
387 | sub cmd_head1 { | |
b7ae008f SP |
388 | my ($self, $attrs, $text) = @_; |
389 | $self->heading ($text, 0, '===='); | |
6055f9d4 | 390 | } |
69e00e79 | 391 | |
6055f9d4 GS |
392 | # Second level heading. |
393 | sub cmd_head2 { | |
b7ae008f SP |
394 | my ($self, $attrs, $text) = @_; |
395 | $self->heading ($text, $$self{opt_indent} / 2, '== '); | |
6055f9d4 | 396 | } |
69e00e79 | 397 | |
50a3fd2a RA |
398 | # Third level heading. |
399 | sub cmd_head3 { | |
b7ae008f SP |
400 | my ($self, $attrs, $text) = @_; |
401 | $self->heading ($text, $$self{opt_indent} * 2 / 3 + 0.5, '= '); | |
50a3fd2a RA |
402 | } |
403 | ||
b7ae008f | 404 | # Fourth level heading. |
50a3fd2a | 405 | sub cmd_head4 { |
b7ae008f SP |
406 | my ($self, $attrs, $text) = @_; |
407 | $self->heading ($text, $$self{opt_indent} * 3 / 4 + 0.5, '- '); | |
50a3fd2a RA |
408 | } |
409 | ||
b7ae008f SP |
410 | ############################################################################## |
411 | # List handling | |
412 | ############################################################################## | |
413 | ||
414 | # Handle the beginning of an =over block. Takes the type of the block as the | |
415 | # first argument, and then the attr hash. This is called by the handlers for | |
416 | # the four different types of lists (bullet, number, text, and block). | |
417 | sub over_common_start { | |
418 | my ($self, $attrs) = @_; | |
b616daaf | 419 | $self->item ("\n\n") if defined $$self{ITEM}; |
b7ae008f SP |
420 | |
421 | # Find the indentation level. | |
422 | my $indent = $$attrs{indent}; | |
423 | unless (defined ($indent) && $indent =~ /^\s*[-+]?\d{1,4}\s*$/) { | |
424 | $indent = $$self{opt_indent}; | |
425 | } | |
426 | ||
427 | # Add this to our stack of indents and increase our current margin. | |
6055f9d4 | 428 | push (@{ $$self{INDENTS} }, $$self{MARGIN}); |
b7ae008f SP |
429 | $$self{MARGIN} += ($indent + 0); |
430 | return ''; | |
6055f9d4 | 431 | } |
69e00e79 | 432 | |
b7ae008f SP |
433 | # End an =over block. Takes no options other than the class pointer. Output |
434 | # any pending items and then pop one level of indentation. | |
435 | sub over_common_end { | |
436 | my ($self) = @_; | |
b616daaf | 437 | $self->item ("\n\n") if defined $$self{ITEM}; |
6055f9d4 | 438 | $$self{MARGIN} = pop @{ $$self{INDENTS} }; |
b7ae008f | 439 | return ''; |
69e00e79 | 440 | } |
441 | ||
b7ae008f SP |
442 | # Dispatch the start and end calls as appropriate. |
443 | sub start_over_bullet { $_[0]->over_common_start ($_[1]) } | |
444 | sub start_over_number { $_[0]->over_common_start ($_[1]) } | |
445 | sub start_over_text { $_[0]->over_common_start ($_[1]) } | |
446 | sub start_over_block { $_[0]->over_common_start ($_[1]) } | |
447 | sub end_over_bullet { $_[0]->over_common_end } | |
448 | sub end_over_number { $_[0]->over_common_end } | |
449 | sub end_over_text { $_[0]->over_common_end } | |
450 | sub end_over_block { $_[0]->over_common_end } | |
451 | ||
452 | # The common handler for all item commands. Takes the type of the item, the | |
453 | # attributes, and then the text of the item. | |
454 | sub item_common { | |
455 | my ($self, $type, $attrs, $text) = @_; | |
456 | $self->item if defined $$self{ITEM}; | |
69e00e79 | 457 | |
b7ae008f SP |
458 | # Clean up the text. We want to end up with two variables, one ($text) |
459 | # which contains any body text after taking out the item portion, and | |
460 | # another ($item) which contains the actual item text. Note the use of | |
461 | # the internal Pod::Simple attribute here; that's a potential land mine. | |
462 | $text =~ s/\s+$//; | |
463 | my ($item, $index); | |
464 | if ($type eq 'bullet') { | |
465 | $item = '*'; | |
466 | } elsif ($type eq 'number') { | |
467 | $item = $$attrs{'~orig_content'}; | |
27f805f4 | 468 | } else { |
b7ae008f SP |
469 | $item = $text; |
470 | $item =~ s/\s*\n\s*/ /g; | |
471 | $text = ''; | |
27f805f4 | 472 | } |
b7ae008f | 473 | $$self{ITEM} = $item; |
6055f9d4 | 474 | |
b7ae008f SP |
475 | # If body text for this item was included, go ahead and output that now. |
476 | if ($text) { | |
477 | $text =~ s/\s*$/\n/; | |
478 | $self->item ($text); | |
479 | } | |
480 | return ''; | |
6055f9d4 | 481 | } |
f2506fb2 | 482 | |
b7ae008f SP |
483 | # Dispatch the item commands to the appropriate place. |
484 | sub cmd_item_bullet { my $self = shift; $self->item_common ('bullet', @_) } | |
485 | sub cmd_item_number { my $self = shift; $self->item_common ('number', @_) } | |
486 | sub cmd_item_text { my $self = shift; $self->item_common ('text', @_) } | |
487 | sub cmd_item_block { my $self = shift; $self->item_common ('block', @_) } | |
69e00e79 | 488 | |
3c014959 | 489 | ############################################################################## |
5ec554fb | 490 | # Formatting codes |
3c014959 | 491 | ############################################################################## |
69e00e79 | 492 | |
b7ae008f SP |
493 | # The simple ones. |
494 | sub cmd_b { return $_[0]{alt} ? "``$_[2]''" : $_[2] } | |
495 | sub cmd_f { return $_[0]{alt} ? "\"$_[2]\"" : $_[2] } | |
496 | sub cmd_i { return '*' . $_[2] . '*' } | |
497 | sub cmd_x { return '' } | |
3c014959 JH |
498 | |
499 | # Apply a whole bunch of messy heuristics to not quote things that don't | |
500 | # benefit from being quoted. These originally come from Barrie Slaymaker and | |
501 | # largely duplicate code in Pod::Man. | |
b7ae008f SP |
502 | sub cmd_c { |
503 | my ($self, $attrs, $text) = @_; | |
3c014959 JH |
504 | |
505 | # A regex that matches the portion of a variable reference that's the | |
506 | # array or hash index, separated out just because we want to use it in | |
507 | # several places in the following regex. | |
508 | my $index = '(?: \[.*\] | \{.*\} )?'; | |
509 | ||
510 | # Check for things that we don't want to quote, and if we find any of | |
511 | # them, return the string with just a font change and no quoting. | |
b7ae008f | 512 | $text =~ m{ |
3c014959 JH |
513 | ^\s* |
514 | (?: | |
515 | ( [\'\`\"] ) .* \1 # already quoted | |
516 | | \` .* \' # `quoted' | |
517 | | \$+ [\#^]? \S $index # special ($^Foo, $") | |
518 | | [\$\@%&*]+ \#? [:\'\w]+ $index # plain var or func | |
519 | | [\$\@%&*]* [:\'\w]+ (?: -> )? \(\s*[^\s,]\s*\) # 0/1-arg func call | |
f011ec7d | 520 | | [+-]? ( \d[\d.]* | \.\d+ ) (?: [eE][+-]?\d+ )? # a number |
3c014959 JH |
521 | | 0x [a-fA-F\d]+ # a hex constant |
522 | ) | |
523 | \s*\z | |
b7ae008f | 524 | }xo && return $text; |
3c014959 JH |
525 | |
526 | # If we didn't return, go ahead and quote the text. | |
b7ae008f SP |
527 | return $$self{opt_alt} |
528 | ? "``$text''" | |
529 | : "$$self{LQUOTE}$text$$self{RQUOTE}"; | |
69e00e79 | 530 | } |
531 | ||
b7ae008f SP |
532 | # Links reduce to the text that we're given, wrapped in angle brackets if it's |
533 | # a URL. | |
534 | sub cmd_l { | |
535 | my ($self, $attrs, $text) = @_; | |
536 | return $$attrs{type} eq 'url' ? "<$text>" : $text; | |
b616daaf JH |
537 | } |
538 | ||
3c014959 | 539 | ############################################################################## |
27f805f4 | 540 | # Backwards compatibility |
3c014959 | 541 | ############################################################################## |
27f805f4 GS |
542 | |
543 | # The old Pod::Text module did everything in a pod2text() function. This | |
544 | # tries to provide the same interface for legacy applications. | |
545 | sub pod2text { | |
546 | my @args; | |
547 | ||
548 | # This is really ugly; I hate doing option parsing in the middle of a | |
549 | # module. But the old Pod::Text module supported passing flags to its | |
550 | # entry function, so handle -a and -<number>. | |
551 | while ($_[0] =~ /^-/) { | |
552 | my $flag = shift; | |
553 | if ($flag eq '-a') { push (@args, alt => 1) } | |
554 | elsif ($flag =~ /^-(\d+)$/) { push (@args, width => $1) } | |
555 | else { | |
556 | unshift (@_, $flag); | |
557 | last; | |
558 | } | |
559 | } | |
560 | ||
561 | # Now that we know what arguments we're using, create the parser. | |
562 | my $parser = Pod::Text->new (@args); | |
563 | ||
564 | # If two arguments were given, the second argument is going to be a file | |
3c014959 JH |
565 | # handle. That means we want to call parse_from_filehandle(), which means |
566 | # we need to turn the first argument into a file handle. Magic open will | |
567 | # handle the <&STDIN case automagically. | |
27f805f4 | 568 | if (defined $_[1]) { |
ab1f1d91 | 569 | my @fhs = @_; |
27f805f4 | 570 | local *IN; |
ab1f1d91 JH |
571 | unless (open (IN, $fhs[0])) { |
572 | croak ("Can't open $fhs[0] for reading: $!\n"); | |
27f805f4 GS |
573 | return; |
574 | } | |
ab1f1d91 | 575 | $fhs[0] = \*IN; |
8f202758 SP |
576 | $parser->output_fh ($fhs[1]); |
577 | my $retval = $parser->parse_file ($fhs[0]); | |
578 | my $fh = $parser->output_fh (); | |
579 | close $fh; | |
580 | return $retval; | |
27f805f4 | 581 | } else { |
b7ae008f | 582 | return $parser->parse_file (@_); |
27f805f4 GS |
583 | } |
584 | } | |
585 | ||
8f202758 SP |
586 | # Reset the underlying Pod::Simple object between calls to parse_from_file so |
587 | # that the same object can be reused to convert multiple pages. | |
588 | sub parse_from_file { | |
589 | my $self = shift; | |
590 | $self->reinit; | |
42ae9e1d RGS |
591 | |
592 | # Fake the old cutting option to Pod::Parser. This fiddings with internal | |
593 | # Pod::Simple state and is quite ugly; we need a better approach. | |
594 | if (ref ($_[0]) eq 'HASH') { | |
595 | my $opts = shift @_; | |
596 | if (defined ($$opts{-cutting}) && !$$opts{-cutting}) { | |
597 | $$self{in_pod} = 1; | |
598 | $$self{last_was_blank} = 1; | |
599 | } | |
600 | } | |
601 | ||
602 | # Do the work. | |
8f202758 | 603 | my $retval = $self->Pod::Simple::parse_from_file (@_); |
42ae9e1d RGS |
604 | |
605 | # Flush output, since Pod::Simple doesn't do this. Ideally we should also | |
606 | # close the file descriptor if we had to open one, but we can't easily | |
607 | # figure this out. | |
8f202758 SP |
608 | my $fh = $self->output_fh (); |
609 | my $oldfh = select $fh; | |
610 | my $oldflush = $|; | |
611 | $| = 1; | |
612 | print $fh ''; | |
613 | $| = $oldflush; | |
614 | select $oldfh; | |
615 | return $retval; | |
616 | } | |
617 | ||
fcf69717 SP |
618 | # Pod::Simple failed to provide this backward compatibility function, so |
619 | # implement it ourselves. File handles are one of the inputs that | |
620 | # parse_from_file supports. | |
621 | sub parse_from_filehandle { | |
622 | my $self = shift; | |
623 | $self->parse_from_file (@_); | |
624 | } | |
625 | ||
3c014959 | 626 | ############################################################################## |
6055f9d4 | 627 | # Module return value and documentation |
3c014959 | 628 | ############################################################################## |
69e00e79 | 629 | |
6055f9d4 GS |
630 | 1; |
631 | __END__ | |
69e00e79 | 632 | |
6055f9d4 | 633 | =head1 NAME |
69e00e79 | 634 | |
6055f9d4 | 635 | Pod::Text - Convert POD data to formatted ASCII text |
69e00e79 | 636 | |
6055f9d4 | 637 | =head1 SYNOPSIS |
69e00e79 | 638 | |
6055f9d4 GS |
639 | use Pod::Text; |
640 | my $parser = Pod::Text->new (sentence => 0, width => 78); | |
69e00e79 | 641 | |
6055f9d4 GS |
642 | # Read POD from STDIN and write to STDOUT. |
643 | $parser->parse_from_filehandle; | |
69e00e79 | 644 | |
6055f9d4 GS |
645 | # Read POD from file.pod and write to file.txt. |
646 | $parser->parse_from_file ('file.pod', 'file.txt'); | |
69e00e79 | 647 | |
6055f9d4 | 648 | =head1 DESCRIPTION |
5491a304 | 649 | |
27f805f4 GS |
650 | Pod::Text is a module that can convert documentation in the POD format (the |
651 | preferred language for documenting Perl) into formatted ASCII. It uses no | |
652 | special formatting controls or codes whatsoever, and its output is therefore | |
653 | suitable for nearly any device. | |
69e00e79 | 654 | |
b7ae008f SP |
655 | As a derived class from Pod::Simple, Pod::Text supports the same methods and |
656 | interfaces. See L<Pod::Simple> for all the details; briefly, one creates a | |
657 | new parser with C<< Pod::Text->new() >> and then normally calls parse_file(). | |
6055f9d4 | 658 | |
27f805f4 | 659 | new() can take options, in the form of key/value pairs, that control the |
6055f9d4 GS |
660 | behavior of the parser. The currently recognized options are: |
661 | ||
662 | =over 4 | |
663 | ||
664 | =item alt | |
665 | ||
666 | If set to a true value, selects an alternate output format that, among other | |
667 | things, uses a different heading style and marks C<=item> entries with a | |
668 | colon in the left margin. Defaults to false. | |
669 | ||
59548eca JH |
670 | =item code |
671 | ||
672 | If set to a true value, the non-POD parts of the input file will be included | |
673 | in the output. Useful for viewing code documented with POD blocks with the | |
674 | POD rendered and the code left intact. | |
675 | ||
6055f9d4 GS |
676 | =item indent |
677 | ||
678 | The number of spaces to indent regular text, and the default indentation for | |
679 | C<=over> blocks. Defaults to 4. | |
680 | ||
681 | =item loose | |
682 | ||
683 | If set to a true value, a blank line is printed after a C<=head1> heading. | |
684 | If set to false (the default), no blank line is printed after C<=head1>, | |
685 | although one is still printed after C<=head2>. This is the default because | |
686 | it's the expected formatting for manual pages; if you're formatting | |
687 | arbitrary text documents, setting this to true may result in more pleasing | |
688 | output. | |
689 | ||
11f72409 RA |
690 | =item margin |
691 | ||
692 | The width of the left margin in spaces. Defaults to 0. This is the margin | |
693 | for all text, including headings, not the amount by which regular text is | |
694 | indented; for the latter, see the I<indent> option. To set the right | |
695 | margin, see the I<width> option. | |
696 | ||
ab1f1d91 JH |
697 | =item quotes |
698 | ||
699 | Sets the quote marks used to surround CE<lt>> text. If the value is a | |
700 | single character, it is used as both the left and right quote; if it is two | |
701 | characters, the first character is used as the left quote and the second as | |
702 | the right quoted; and if it is four characters, the first two are used as | |
703 | the left quote and the second two as the right quote. | |
704 | ||
705 | This may also be set to the special value C<none>, in which case no quote | |
706 | marks are added around CE<lt>> text. | |
707 | ||
6055f9d4 GS |
708 | =item sentence |
709 | ||
27f805f4 GS |
710 | If set to a true value, Pod::Text will assume that each sentence ends in two |
711 | spaces, and will try to preserve that spacing. If set to false, all | |
6055f9d4 GS |
712 | consecutive whitespace in non-verbatim paragraphs is compressed into a |
713 | single space. Defaults to true. | |
714 | ||
715 | =item width | |
716 | ||
717 | The column at which to wrap text on the right-hand side. Defaults to 76. | |
718 | ||
719 | =back | |
720 | ||
b7ae008f SP |
721 | The standard Pod::Simple method parse_file() takes one argument, the file or |
722 | file handle to read from, and writes output to standard output unless that | |
723 | has been changed with the output_fh() method. See L<Pod::Simple> for the | |
724 | specific details and for other alternative interfaces. | |
6055f9d4 GS |
725 | |
726 | =head1 DIAGNOSTICS | |
727 | ||
728 | =over 4 | |
729 | ||
27f805f4 GS |
730 | =item Bizarre space in item |
731 | ||
59548eca JH |
732 | =item Item called without tag |
733 | ||
734 | (W) Something has gone wrong in internal C<=item> processing. These | |
735 | messages indicate a bug in Pod::Text; you should never see them. | |
27f805f4 GS |
736 | |
737 | =item Can't open %s for reading: %s | |
738 | ||
739 | (F) Pod::Text was invoked via the compatibility mode pod2text() interface | |
740 | and the input file it was given could not be opened. | |
741 | ||
ab1f1d91 JH |
742 | =item Invalid quote specification "%s" |
743 | ||
744 | (F) The quote specification given (the quotes option to the constructor) was | |
745 | invalid. A quote specification must be one, two, or four characters long. | |
746 | ||
6055f9d4 GS |
747 | =back |
748 | ||
749 | =head1 NOTES | |
750 | ||
27f805f4 | 751 | This is a replacement for an earlier Pod::Text module written by Tom |
b7ae008f | 752 | Christiansen. It has a revamped interface, since it now uses Pod::Simple, |
27f805f4 GS |
753 | but an interface roughly compatible with the old Pod::Text::pod2text() |
754 | function is still available. Please change to the new calling convention, | |
755 | though. | |
6055f9d4 GS |
756 | |
757 | The original Pod::Text contained code to do formatting via termcap | |
758 | sequences, although it wasn't turned on by default and it was problematic to | |
27f805f4 | 759 | get it to work at all. This rewrite doesn't even try to do that, but a |
bf202ccd | 760 | subclass of it does. Look for L<Pod::Text::Termcap>. |
6055f9d4 GS |
761 | |
762 | =head1 SEE ALSO | |
763 | ||
b7ae008f | 764 | L<Pod::Simple>, L<Pod::Text::Termcap>, L<pod2text(1)> |
6055f9d4 | 765 | |
fd20da51 JH |
766 | The current version of this module is always available from its web site at |
767 | L<http://www.eyrie.org/~eagle/software/podlators/>. It is also part of the | |
768 | Perl core distribution as of 5.6.0. | |
769 | ||
6055f9d4 GS |
770 | =head1 AUTHOR |
771 | ||
bf202ccd JH |
772 | Russ Allbery <rra@stanford.edu>, based I<very> heavily on the original |
773 | Pod::Text by Tom Christiansen <tchrist@mox.perl.com> and its conversion to | |
b7ae008f SP |
774 | Pod::Parser by Brad Appleton <bradapp@enteract.com>. Sean Burke's initial |
775 | conversion of Pod::Man to use Pod::Simple provided much-needed guidance on | |
776 | how to use Pod::Simple. | |
6055f9d4 | 777 | |
3c014959 JH |
778 | =head1 COPYRIGHT AND LICENSE |
779 | ||
8f202758 | 780 | Copyright 1999, 2000, 2001, 2002, 2004, 2006 Russ Allbery <rra@stanford.edu>. |
3c014959 JH |
781 | |
782 | This program is free software; you may redistribute it and/or modify it | |
783 | under the same terms as Perl itself. | |
784 | ||
6055f9d4 | 785 | =cut |