| 1 | # Pod::Text -- Convert POD data to formatted ASCII text. |
| 2 | # $Id: Text.pm,v 2.20 2002/07/15 05:46:00 eagle Exp $ |
| 3 | # |
| 4 | # Copyright 1999, 2000, 2001, 2002 by Russ Allbery <rra@stanford.edu> |
| 5 | # |
| 6 | # This program is free software; you may redistribute it and/or modify it |
| 7 | # under the same terms as Perl itself. |
| 8 | # |
| 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. |
| 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. |
| 19 | |
| 20 | ############################################################################## |
| 21 | # Modules and declarations |
| 22 | ############################################################################## |
| 23 | |
| 24 | package Pod::Text; |
| 25 | |
| 26 | require 5.004; |
| 27 | |
| 28 | use Carp qw(carp croak); |
| 29 | use Exporter (); |
| 30 | use Pod::ParseLink qw(parselink); |
| 31 | use Pod::Select (); |
| 32 | |
| 33 | use strict; |
| 34 | use vars qw(@ISA @EXPORT %ESCAPES $VERSION); |
| 35 | |
| 36 | # We inherit from Pod::Select instead of Pod::Parser so that we can be used by |
| 37 | # Pod::Usage. |
| 38 | @ISA = qw(Pod::Select Exporter); |
| 39 | |
| 40 | # We have to export pod2text for backward compatibility. |
| 41 | @EXPORT = qw(pod2text); |
| 42 | |
| 43 | # Don't use the CVS revision as the version, since this module is also in Perl |
| 44 | # core and too many things could munge CVS magic revision strings. This |
| 45 | # number should ideally be the same as the CVS revision in podlators, however. |
| 46 | $VERSION = 2.20; |
| 47 | |
| 48 | |
| 49 | ############################################################################## |
| 50 | # Table of supported E<> escapes |
| 51 | ############################################################################## |
| 52 | |
| 53 | # This table is taken near verbatim from Pod::PlainText in Pod::Parser, which |
| 54 | # got it near verbatim from the original Pod::Text. It is therefore credited |
| 55 | # to Tom Christiansen, and I'm glad I didn't have to write it. :) "iexcl" to |
| 56 | # "divide" added by Tim Jenness. |
| 57 | %ESCAPES = ( |
| 58 | 'amp' => '&', # ampersand |
| 59 | 'apos' => "'", # apostrophe |
| 60 | 'lt' => '<', # left chevron, less-than |
| 61 | 'gt' => '>', # right chevron, greater-than |
| 62 | 'quot' => '"', # double quote |
| 63 | 'sol' => '/', # solidus (forward slash) |
| 64 | 'verbar' => '|', # vertical bar |
| 65 | |
| 66 | "Aacute" => "\xC1", # capital A, acute accent |
| 67 | "aacute" => "\xE1", # small a, acute accent |
| 68 | "Acirc" => "\xC2", # capital A, circumflex accent |
| 69 | "acirc" => "\xE2", # small a, circumflex accent |
| 70 | "AElig" => "\xC6", # capital AE diphthong (ligature) |
| 71 | "aelig" => "\xE6", # small ae diphthong (ligature) |
| 72 | "Agrave" => "\xC0", # capital A, grave accent |
| 73 | "agrave" => "\xE0", # small a, grave accent |
| 74 | "Aring" => "\xC5", # capital A, ring |
| 75 | "aring" => "\xE5", # small a, ring |
| 76 | "Atilde" => "\xC3", # capital A, tilde |
| 77 | "atilde" => "\xE3", # small a, tilde |
| 78 | "Auml" => "\xC4", # capital A, dieresis or umlaut mark |
| 79 | "auml" => "\xE4", # small a, dieresis or umlaut mark |
| 80 | "Ccedil" => "\xC7", # capital C, cedilla |
| 81 | "ccedil" => "\xE7", # small c, cedilla |
| 82 | "Eacute" => "\xC9", # capital E, acute accent |
| 83 | "eacute" => "\xE9", # small e, acute accent |
| 84 | "Ecirc" => "\xCA", # capital E, circumflex accent |
| 85 | "ecirc" => "\xEA", # small e, circumflex accent |
| 86 | "Egrave" => "\xC8", # capital E, grave accent |
| 87 | "egrave" => "\xE8", # small e, grave accent |
| 88 | "ETH" => "\xD0", # capital Eth, Icelandic |
| 89 | "eth" => "\xF0", # small eth, Icelandic |
| 90 | "Euml" => "\xCB", # capital E, dieresis or umlaut mark |
| 91 | "euml" => "\xEB", # small e, dieresis or umlaut mark |
| 92 | "Iacute" => "\xCD", # capital I, acute accent |
| 93 | "iacute" => "\xED", # small i, acute accent |
| 94 | "Icirc" => "\xCE", # capital I, circumflex accent |
| 95 | "icirc" => "\xEE", # small i, circumflex accent |
| 96 | "Igrave" => "\xCC", # capital I, grave accent |
| 97 | "igrave" => "\xEC", # small i, grave accent |
| 98 | "Iuml" => "\xCF", # capital I, dieresis or umlaut mark |
| 99 | "iuml" => "\xEF", # small i, dieresis or umlaut mark |
| 100 | "Ntilde" => "\xD1", # capital N, tilde |
| 101 | "ntilde" => "\xF1", # small n, tilde |
| 102 | "Oacute" => "\xD3", # capital O, acute accent |
| 103 | "oacute" => "\xF3", # small o, acute accent |
| 104 | "Ocirc" => "\xD4", # capital O, circumflex accent |
| 105 | "ocirc" => "\xF4", # small o, circumflex accent |
| 106 | "Ograve" => "\xD2", # capital O, grave accent |
| 107 | "ograve" => "\xF2", # small o, grave accent |
| 108 | "Oslash" => "\xD8", # capital O, slash |
| 109 | "oslash" => "\xF8", # small o, slash |
| 110 | "Otilde" => "\xD5", # capital O, tilde |
| 111 | "otilde" => "\xF5", # small o, tilde |
| 112 | "Ouml" => "\xD6", # capital O, dieresis or umlaut mark |
| 113 | "ouml" => "\xF6", # small o, dieresis or umlaut mark |
| 114 | "szlig" => "\xDF", # small sharp s, German (sz ligature) |
| 115 | "THORN" => "\xDE", # capital THORN, Icelandic |
| 116 | "thorn" => "\xFE", # small thorn, Icelandic |
| 117 | "Uacute" => "\xDA", # capital U, acute accent |
| 118 | "uacute" => "\xFA", # small u, acute accent |
| 119 | "Ucirc" => "\xDB", # capital U, circumflex accent |
| 120 | "ucirc" => "\xFB", # small u, circumflex accent |
| 121 | "Ugrave" => "\xD9", # capital U, grave accent |
| 122 | "ugrave" => "\xF9", # small u, grave accent |
| 123 | "Uuml" => "\xDC", # capital U, dieresis or umlaut mark |
| 124 | "uuml" => "\xFC", # small u, dieresis or umlaut mark |
| 125 | "Yacute" => "\xDD", # capital Y, acute accent |
| 126 | "yacute" => "\xFD", # small y, acute accent |
| 127 | "yuml" => "\xFF", # small y, dieresis or umlaut mark |
| 128 | |
| 129 | "laquo" => "\xAB", # left pointing double angle quotation mark |
| 130 | "lchevron" => "\xAB", # synonym (backwards compatibility) |
| 131 | "raquo" => "\xBB", # right pointing double angle quotation mark |
| 132 | "rchevron" => "\xBB", # synonym (backwards compatibility) |
| 133 | |
| 134 | "iexcl" => "\xA1", # inverted exclamation mark |
| 135 | "cent" => "\xA2", # cent sign |
| 136 | "pound" => "\xA3", # (UK) pound sign |
| 137 | "curren" => "\xA4", # currency sign |
| 138 | "yen" => "\xA5", # yen sign |
| 139 | "brvbar" => "\xA6", # broken vertical bar |
| 140 | "sect" => "\xA7", # section sign |
| 141 | "uml" => "\xA8", # diaresis |
| 142 | "copy" => "\xA9", # Copyright symbol |
| 143 | "ordf" => "\xAA", # feminine ordinal indicator |
| 144 | "not" => "\xAC", # not sign |
| 145 | "shy" => '', # soft (discretionary) hyphen |
| 146 | "reg" => "\xAE", # registered trademark |
| 147 | "macr" => "\xAF", # macron, overline |
| 148 | "deg" => "\xB0", # degree sign |
| 149 | "plusmn" => "\xB1", # plus-minus sign |
| 150 | "sup2" => "\xB2", # superscript 2 |
| 151 | "sup3" => "\xB3", # superscript 3 |
| 152 | "acute" => "\xB4", # acute accent |
| 153 | "micro" => "\xB5", # micro sign |
| 154 | "para" => "\xB6", # pilcrow sign = paragraph sign |
| 155 | "middot" => "\xB7", # middle dot = Georgian comma |
| 156 | "cedil" => "\xB8", # cedilla |
| 157 | "sup1" => "\xB9", # superscript 1 |
| 158 | "ordm" => "\xBA", # masculine ordinal indicator |
| 159 | "frac14" => "\xBC", # vulgar fraction one quarter |
| 160 | "frac12" => "\xBD", # vulgar fraction one half |
| 161 | "frac34" => "\xBE", # vulgar fraction three quarters |
| 162 | "iquest" => "\xBF", # inverted question mark |
| 163 | "times" => "\xD7", # multiplication sign |
| 164 | "divide" => "\xF7", # division sign |
| 165 | |
| 166 | "nbsp" => "\x01", # non-breaking space |
| 167 | ); |
| 168 | |
| 169 | |
| 170 | ############################################################################## |
| 171 | # Initialization |
| 172 | ############################################################################## |
| 173 | |
| 174 | # Initialize the object. Must be sure to call our parent initializer. |
| 175 | sub initialize { |
| 176 | my $self = shift; |
| 177 | |
| 178 | $$self{alt} = 0 unless defined $$self{alt}; |
| 179 | $$self{indent} = 4 unless defined $$self{indent}; |
| 180 | $$self{loose} = 0 unless defined $$self{loose}; |
| 181 | $$self{sentence} = 0 unless defined $$self{sentence}; |
| 182 | $$self{width} = 76 unless defined $$self{width}; |
| 183 | |
| 184 | # Figure out what quotes we'll be using for C<> text. |
| 185 | $$self{quotes} ||= '"'; |
| 186 | if ($$self{quotes} eq 'none') { |
| 187 | $$self{LQUOTE} = $$self{RQUOTE} = ''; |
| 188 | } elsif (length ($$self{quotes}) == 1) { |
| 189 | $$self{LQUOTE} = $$self{RQUOTE} = $$self{quotes}; |
| 190 | } elsif ($$self{quotes} =~ /^(.)(.)$/ |
| 191 | || $$self{quotes} =~ /^(..)(..)$/) { |
| 192 | $$self{LQUOTE} = $1; |
| 193 | $$self{RQUOTE} = $2; |
| 194 | } else { |
| 195 | croak qq(Invalid quote specification "$$self{quotes}"); |
| 196 | } |
| 197 | |
| 198 | $$self{INDENTS} = []; # Stack of indentations. |
| 199 | $$self{MARGIN} = $$self{indent}; # Current left margin in spaces. |
| 200 | |
| 201 | $self->SUPER::initialize; |
| 202 | |
| 203 | # Tell Pod::Parser that we want the non-POD stuff too if code was set. |
| 204 | $self->parseopts ('-want_nonPODs' => 1) if $$self{code}; |
| 205 | } |
| 206 | |
| 207 | |
| 208 | ############################################################################## |
| 209 | # Core overrides |
| 210 | ############################################################################## |
| 211 | |
| 212 | # Called for each command paragraph. Gets the command, the associated |
| 213 | # paragraph, the line number, and a Pod::Paragraph object. Just dispatches |
| 214 | # the command to a method named the same as the command. =cut is handled |
| 215 | # internally by Pod::Parser. |
| 216 | sub command { |
| 217 | my $self = shift; |
| 218 | my $command = shift; |
| 219 | return if $command eq 'pod'; |
| 220 | return if ($$self{EXCLUDE} && $command ne 'end'); |
| 221 | if ($self->can ('cmd_' . $command)) { |
| 222 | $command = 'cmd_' . $command; |
| 223 | $self->$command (@_); |
| 224 | } else { |
| 225 | my ($text, $line, $paragraph) = @_; |
| 226 | my $file; |
| 227 | ($file, $line) = $paragraph->file_line; |
| 228 | $text =~ s/\n+\z//; |
| 229 | $text = " $text" if ($text =~ /^\S/); |
| 230 | warn qq($file:$line: Unknown command paragraph: =$command$text\n); |
| 231 | return; |
| 232 | } |
| 233 | } |
| 234 | |
| 235 | # Called for a verbatim paragraph. Gets the paragraph, the line number, and a |
| 236 | # Pod::Paragraph object. Just output it verbatim, but with tabs converted to |
| 237 | # spaces. |
| 238 | sub verbatim { |
| 239 | my $self = shift; |
| 240 | return if $$self{EXCLUDE}; |
| 241 | $self->item if defined $$self{ITEM}; |
| 242 | local $_ = shift; |
| 243 | return if /^\s*$/; |
| 244 | s/^(\s*\S+)/(' ' x $$self{MARGIN}) . $1/gme; |
| 245 | $self->output ($_); |
| 246 | } |
| 247 | |
| 248 | # Called for a regular text block. Gets the paragraph, the line number, and a |
| 249 | # Pod::Paragraph object. Perform interpolation and output the results. |
| 250 | sub textblock { |
| 251 | my $self = shift; |
| 252 | return if $$self{EXCLUDE}; |
| 253 | $self->output ($_[0]), return if $$self{VERBATIM}; |
| 254 | local $_ = shift; |
| 255 | my $line = shift; |
| 256 | |
| 257 | # Interpolate and output the paragraph. |
| 258 | $_ = $self->interpolate ($_, $line); |
| 259 | s/\s+$/\n/; |
| 260 | if (defined $$self{ITEM}) { |
| 261 | $self->item ($_ . "\n"); |
| 262 | } else { |
| 263 | $self->output ($self->reformat ($_ . "\n")); |
| 264 | } |
| 265 | } |
| 266 | |
| 267 | # Called for a formatting code. Gets the command, argument, and a |
| 268 | # Pod::InteriorSequence object and is expected to return the resulting text. |
| 269 | # Calls methods for code, bold, italic, file, and link to handle those types |
| 270 | # of codes, and handles S<>, E<>, X<>, and Z<> directly. |
| 271 | sub interior_sequence { |
| 272 | local $_; |
| 273 | my ($self, $command, $seq); |
| 274 | ($self, $command, $_, $seq) = @_; |
| 275 | |
| 276 | # We have to defer processing of the inside of an L<> formatting code. If |
| 277 | # this code is nested inside an L<> code, return the literal raw text of |
| 278 | # it. |
| 279 | my $parent = $seq->nested; |
| 280 | while (defined $parent) { |
| 281 | return $seq->raw_text if ($parent->cmd_name eq 'L'); |
| 282 | $parent = $parent->nested; |
| 283 | } |
| 284 | |
| 285 | # Index entries are ignored in plain text. |
| 286 | return '' if ($command eq 'X' || $command eq 'Z'); |
| 287 | |
| 288 | # Expand escapes into the actual character now, warning if invalid. |
| 289 | if ($command eq 'E') { |
| 290 | if (/^\d+$/) { |
| 291 | return chr; |
| 292 | } else { |
| 293 | return $ESCAPES{$_} if defined $ESCAPES{$_}; |
| 294 | my ($file, $line) = $seq->file_line; |
| 295 | warn "$file:$line: Unknown escape: E<$_>\n"; |
| 296 | return "E<$_>"; |
| 297 | } |
| 298 | } |
| 299 | |
| 300 | # For all the other formatting codes, empty content produces no output. |
| 301 | return if $_ eq ''; |
| 302 | |
| 303 | # For S<>, compress all internal whitespace and then map spaces to \01. |
| 304 | # When we output the text, we'll map this back. |
| 305 | if ($command eq 'S') { |
| 306 | s/\s+/ /g; |
| 307 | tr/ /\01/; |
| 308 | return $_; |
| 309 | } |
| 310 | |
| 311 | # Anything else needs to get dispatched to another method. |
| 312 | if ($command eq 'B') { return $self->seq_b ($_) } |
| 313 | elsif ($command eq 'C') { return $self->seq_c ($_) } |
| 314 | elsif ($command eq 'F') { return $self->seq_f ($_) } |
| 315 | elsif ($command eq 'I') { return $self->seq_i ($_) } |
| 316 | elsif ($command eq 'L') { return $self->seq_l ($_, $seq) } |
| 317 | else { |
| 318 | my ($file, $line) = $seq->file_line; |
| 319 | warn "$file:$line: Unknown formatting code: $command<$_>\n"; |
| 320 | } |
| 321 | } |
| 322 | |
| 323 | # Called for each paragraph that's actually part of the POD. We take |
| 324 | # advantage of this opportunity to untabify the input. Also, if given the |
| 325 | # code option, we may see paragraphs that aren't part of the POD and need to |
| 326 | # output them directly. |
| 327 | sub preprocess_paragraph { |
| 328 | my $self = shift; |
| 329 | local $_ = shift; |
| 330 | 1 while s/^(.*?)(\t+)/$1 . ' ' x (length ($2) * 8 - length ($1) % 8)/me; |
| 331 | $self->output_code ($_) if $self->cutting; |
| 332 | $_; |
| 333 | } |
| 334 | |
| 335 | |
| 336 | ############################################################################## |
| 337 | # Command paragraphs |
| 338 | ############################################################################## |
| 339 | |
| 340 | # All command paragraphs take the paragraph and the line number. |
| 341 | |
| 342 | # First level heading. |
| 343 | sub cmd_head1 { |
| 344 | my ($self, $text, $line) = @_; |
| 345 | $self->heading ($text, $line, 0, '===='); |
| 346 | } |
| 347 | |
| 348 | # Second level heading. |
| 349 | sub cmd_head2 { |
| 350 | my ($self, $text, $line) = @_; |
| 351 | $self->heading ($text, $line, $$self{indent} / 2, '== '); |
| 352 | } |
| 353 | |
| 354 | # Third level heading. |
| 355 | sub cmd_head3 { |
| 356 | my ($self, $text, $line) = @_; |
| 357 | $self->heading ($text, $line, $$self{indent} * 2 / 3 + 0.5, '= '); |
| 358 | } |
| 359 | |
| 360 | # Third level heading. |
| 361 | sub cmd_head4 { |
| 362 | my ($self, $text, $line) = @_; |
| 363 | $self->heading ($text, $line, $$self{indent} * 3 / 4 + 0.5, '- '); |
| 364 | } |
| 365 | |
| 366 | # Start a list. |
| 367 | sub cmd_over { |
| 368 | my $self = shift; |
| 369 | local $_ = shift; |
| 370 | $self->item ("\n\n") if defined $$self{ITEM}; |
| 371 | unless (/^[-+]?\d+\s+$/) { $_ = $$self{indent} } |
| 372 | push (@{ $$self{INDENTS} }, $$self{MARGIN}); |
| 373 | $$self{MARGIN} += ($_ + 0); |
| 374 | } |
| 375 | |
| 376 | # End a list. |
| 377 | sub cmd_back { |
| 378 | my ($self, $text, $line, $paragraph) = @_; |
| 379 | $self->item ("\n\n") if defined $$self{ITEM}; |
| 380 | $$self{MARGIN} = pop @{ $$self{INDENTS} }; |
| 381 | unless (defined $$self{MARGIN}) { |
| 382 | my $file; |
| 383 | ($file, $line) = $paragraph->file_line; |
| 384 | warn "$file:$line: Unmatched =back\n"; |
| 385 | $$self{MARGIN} = $$self{indent}; |
| 386 | } |
| 387 | } |
| 388 | |
| 389 | # An individual list item. |
| 390 | sub cmd_item { |
| 391 | my $self = shift; |
| 392 | if (defined $$self{ITEM}) { $self->item } |
| 393 | local $_ = shift; |
| 394 | s/\s+$//; |
| 395 | $$self{ITEM} = $_ ? $self->interpolate ($_) : '*'; |
| 396 | } |
| 397 | |
| 398 | # Begin a block for a particular translator. Setting VERBATIM triggers |
| 399 | # special handling in textblock(). |
| 400 | sub cmd_begin { |
| 401 | my $self = shift; |
| 402 | local $_ = shift; |
| 403 | my ($kind) = /^(\S+)/ or return; |
| 404 | if ($kind eq 'text') { |
| 405 | $$self{VERBATIM} = 1; |
| 406 | } else { |
| 407 | $$self{EXCLUDE} = 1; |
| 408 | } |
| 409 | } |
| 410 | |
| 411 | # End a block for a particular translator. We assume that all =begin/=end |
| 412 | # pairs are properly closed. |
| 413 | sub cmd_end { |
| 414 | my $self = shift; |
| 415 | $$self{EXCLUDE} = 0; |
| 416 | $$self{VERBATIM} = 0; |
| 417 | } |
| 418 | |
| 419 | # One paragraph for a particular translator. Ignore it unless it's intended |
| 420 | # for text, in which case we treat it as a verbatim text block. |
| 421 | sub cmd_for { |
| 422 | my $self = shift; |
| 423 | local $_ = shift; |
| 424 | my $line = shift; |
| 425 | return unless s/^text\b[ \t]*\n?//; |
| 426 | $self->verbatim ($_, $line); |
| 427 | } |
| 428 | |
| 429 | |
| 430 | ############################################################################## |
| 431 | # Formatting codes |
| 432 | ############################################################################## |
| 433 | |
| 434 | # The simple ones. These are here mostly so that subclasses can override them |
| 435 | # and do more complicated things. |
| 436 | sub seq_b { return $_[0]{alt} ? "``$_[1]''" : $_[1] } |
| 437 | sub seq_f { return $_[0]{alt} ? "\"$_[1]\"" : $_[1] } |
| 438 | sub seq_i { return '*' . $_[1] . '*' } |
| 439 | |
| 440 | # Apply a whole bunch of messy heuristics to not quote things that don't |
| 441 | # benefit from being quoted. These originally come from Barrie Slaymaker and |
| 442 | # largely duplicate code in Pod::Man. |
| 443 | sub seq_c { |
| 444 | my $self = shift; |
| 445 | local $_ = shift; |
| 446 | |
| 447 | # A regex that matches the portion of a variable reference that's the |
| 448 | # array or hash index, separated out just because we want to use it in |
| 449 | # several places in the following regex. |
| 450 | my $index = '(?: \[.*\] | \{.*\} )?'; |
| 451 | |
| 452 | # Check for things that we don't want to quote, and if we find any of |
| 453 | # them, return the string with just a font change and no quoting. |
| 454 | m{ |
| 455 | ^\s* |
| 456 | (?: |
| 457 | ( [\'\`\"] ) .* \1 # already quoted |
| 458 | | \` .* \' # `quoted' |
| 459 | | \$+ [\#^]? \S $index # special ($^Foo, $") |
| 460 | | [\$\@%&*]+ \#? [:\'\w]+ $index # plain var or func |
| 461 | | [\$\@%&*]* [:\'\w]+ (?: -> )? \(\s*[^\s,]\s*\) # 0/1-arg func call |
| 462 | | [+-]? ( \d[\d.]* | \.\d+ ) (?: [eE][+-]?\d+ )? # a number |
| 463 | | 0x [a-fA-F\d]+ # a hex constant |
| 464 | ) |
| 465 | \s*\z |
| 466 | }xo && return $_; |
| 467 | |
| 468 | # If we didn't return, go ahead and quote the text. |
| 469 | return $$self{alt} ? "``$_''" : "$$self{LQUOTE}$_$$self{RQUOTE}"; |
| 470 | } |
| 471 | |
| 472 | # Handle links. Since this is plain text, we can't actually make any real |
| 473 | # links, so this is all to figure out what text we print out. Most of the |
| 474 | # work is done by Pod::ParseLink. |
| 475 | sub seq_l { |
| 476 | my ($self, $link, $seq) = @_; |
| 477 | my ($text, $type) = (parselink ($link))[1,4]; |
| 478 | my ($file, $line) = $seq->file_line; |
| 479 | $text = $self->interpolate ($text, $line); |
| 480 | $text = '<' . $text . '>' if $type eq 'url'; |
| 481 | return $text || ''; |
| 482 | } |
| 483 | |
| 484 | |
| 485 | ############################################################################## |
| 486 | # Header handling |
| 487 | ############################################################################## |
| 488 | |
| 489 | # The common code for handling all headers. Takes the interpolated header |
| 490 | # text, the line number, the indentation, and the surrounding marker for the |
| 491 | # alt formatting method. |
| 492 | sub heading { |
| 493 | my ($self, $text, $line, $indent, $marker) = @_; |
| 494 | $self->item ("\n\n") if defined $$self{ITEM}; |
| 495 | $text =~ s/\s+$//; |
| 496 | $text = $self->interpolate ($text, $line); |
| 497 | if ($$self{alt}) { |
| 498 | my $closemark = reverse (split (//, $marker)); |
| 499 | $self->output ("\n" . "$marker $text $closemark" . "\n\n"); |
| 500 | } else { |
| 501 | $text .= "\n" if $$self{loose}; |
| 502 | $self->output (' ' x $indent . $text . "\n"); |
| 503 | } |
| 504 | } |
| 505 | |
| 506 | |
| 507 | ############################################################################## |
| 508 | # List handling |
| 509 | ############################################################################## |
| 510 | |
| 511 | # This method is called whenever an =item command is complete (in other words, |
| 512 | # we've seen its associated paragraph or know for certain that it doesn't have |
| 513 | # one). It gets the paragraph associated with the item as an argument. If |
| 514 | # that argument is empty, just output the item tag; if it contains a newline, |
| 515 | # output the item tag followed by the newline. Otherwise, see if there's |
| 516 | # enough room for us to output the item tag in the margin of the text or if we |
| 517 | # have to put it on a separate line. |
| 518 | sub item { |
| 519 | my $self = shift; |
| 520 | local $_ = shift; |
| 521 | my $tag = $$self{ITEM}; |
| 522 | unless (defined $tag) { |
| 523 | carp "Item called without tag"; |
| 524 | return; |
| 525 | } |
| 526 | undef $$self{ITEM}; |
| 527 | my $indent = $$self{INDENTS}[-1]; |
| 528 | unless (defined $indent) { $indent = $$self{indent} } |
| 529 | my $space = ' ' x $indent; |
| 530 | $space =~ s/^ /:/ if $$self{alt}; |
| 531 | if (!$_ || /^\s+$/ || ($$self{MARGIN} - $indent < length ($tag) + 1)) { |
| 532 | my $margin = $$self{MARGIN}; |
| 533 | $$self{MARGIN} = $indent; |
| 534 | my $output = $self->reformat ($tag); |
| 535 | $output =~ s/\n*$/\n/; |
| 536 | |
| 537 | # If the text is just whitespace, we have an empty item paragraph; |
| 538 | # this can result from =over/=item/=back without any intermixed |
| 539 | # paragraphs. Insert some whitespace to keep the =item from merging |
| 540 | # into the next paragraph. |
| 541 | $output .= "\n" if $_ && $_ =~ /^\s*$/; |
| 542 | |
| 543 | $self->output ($output); |
| 544 | $$self{MARGIN} = $margin; |
| 545 | $self->output ($self->reformat ($_)) if $_ && /\S/; |
| 546 | } else { |
| 547 | $_ = $self->reformat ($_); |
| 548 | s/^ /:/ if ($$self{alt} && $indent > 0); |
| 549 | my $tagspace = ' ' x length $tag; |
| 550 | s/^($space)$tagspace/$1$tag/ or warn "Bizarre space in item"; |
| 551 | $self->output ($_); |
| 552 | } |
| 553 | } |
| 554 | |
| 555 | |
| 556 | ############################################################################## |
| 557 | # Output formatting |
| 558 | ############################################################################## |
| 559 | |
| 560 | # Wrap a line, indenting by the current left margin. We can't use Text::Wrap |
| 561 | # because it plays games with tabs. We can't use formline, even though we'd |
| 562 | # really like to, because it screws up non-printing characters. So we have to |
| 563 | # do the wrapping ourselves. |
| 564 | sub wrap { |
| 565 | my $self = shift; |
| 566 | local $_ = shift; |
| 567 | my $output = ''; |
| 568 | my $spaces = ' ' x $$self{MARGIN}; |
| 569 | my $width = $$self{width} - $$self{MARGIN}; |
| 570 | while (length > $width) { |
| 571 | if (s/^([^\n]{0,$width})\s+// || s/^([^\n]{$width})//) { |
| 572 | $output .= $spaces . $1 . "\n"; |
| 573 | } else { |
| 574 | last; |
| 575 | } |
| 576 | } |
| 577 | $output .= $spaces . $_; |
| 578 | $output =~ s/\s+$/\n\n/; |
| 579 | $output; |
| 580 | } |
| 581 | |
| 582 | # Reformat a paragraph of text for the current margin. Takes the text to |
| 583 | # reformat and returns the formatted text. |
| 584 | sub reformat { |
| 585 | my $self = shift; |
| 586 | local $_ = shift; |
| 587 | |
| 588 | # If we're trying to preserve two spaces after sentences, do some munging |
| 589 | # to support that. Otherwise, smash all repeated whitespace. |
| 590 | if ($$self{sentence}) { |
| 591 | s/ +$//mg; |
| 592 | s/\.\n/. \n/g; |
| 593 | s/\n/ /g; |
| 594 | s/ +/ /g; |
| 595 | } else { |
| 596 | s/\s+/ /g; |
| 597 | } |
| 598 | $self->wrap ($_); |
| 599 | } |
| 600 | |
| 601 | # Output text to the output device. |
| 602 | sub output { $_[1] =~ tr/\01/ /; print { $_[0]->output_handle } $_[1] } |
| 603 | |
| 604 | # Output a block of code (something that isn't part of the POD text). Called |
| 605 | # by preprocess_paragraph only if we were given the code option. Exists here |
| 606 | # only so that it can be overridden by subclasses. |
| 607 | sub output_code { $_[0]->output ($_[1]) } |
| 608 | |
| 609 | |
| 610 | ############################################################################## |
| 611 | # Backwards compatibility |
| 612 | ############################################################################## |
| 613 | |
| 614 | # The old Pod::Text module did everything in a pod2text() function. This |
| 615 | # tries to provide the same interface for legacy applications. |
| 616 | sub pod2text { |
| 617 | my @args; |
| 618 | |
| 619 | # This is really ugly; I hate doing option parsing in the middle of a |
| 620 | # module. But the old Pod::Text module supported passing flags to its |
| 621 | # entry function, so handle -a and -<number>. |
| 622 | while ($_[0] =~ /^-/) { |
| 623 | my $flag = shift; |
| 624 | if ($flag eq '-a') { push (@args, alt => 1) } |
| 625 | elsif ($flag =~ /^-(\d+)$/) { push (@args, width => $1) } |
| 626 | else { |
| 627 | unshift (@_, $flag); |
| 628 | last; |
| 629 | } |
| 630 | } |
| 631 | |
| 632 | # Now that we know what arguments we're using, create the parser. |
| 633 | my $parser = Pod::Text->new (@args); |
| 634 | |
| 635 | # If two arguments were given, the second argument is going to be a file |
| 636 | # handle. That means we want to call parse_from_filehandle(), which means |
| 637 | # we need to turn the first argument into a file handle. Magic open will |
| 638 | # handle the <&STDIN case automagically. |
| 639 | if (defined $_[1]) { |
| 640 | my @fhs = @_; |
| 641 | local *IN; |
| 642 | unless (open (IN, $fhs[0])) { |
| 643 | croak ("Can't open $fhs[0] for reading: $!\n"); |
| 644 | return; |
| 645 | } |
| 646 | $fhs[0] = \*IN; |
| 647 | return $parser->parse_from_filehandle (@fhs); |
| 648 | } else { |
| 649 | return $parser->parse_from_file (@_); |
| 650 | } |
| 651 | } |
| 652 | |
| 653 | |
| 654 | ############################################################################## |
| 655 | # Module return value and documentation |
| 656 | ############################################################################## |
| 657 | |
| 658 | 1; |
| 659 | __END__ |
| 660 | |
| 661 | =head1 NAME |
| 662 | |
| 663 | Pod::Text - Convert POD data to formatted ASCII text |
| 664 | |
| 665 | =head1 SYNOPSIS |
| 666 | |
| 667 | use Pod::Text; |
| 668 | my $parser = Pod::Text->new (sentence => 0, width => 78); |
| 669 | |
| 670 | # Read POD from STDIN and write to STDOUT. |
| 671 | $parser->parse_from_filehandle; |
| 672 | |
| 673 | # Read POD from file.pod and write to file.txt. |
| 674 | $parser->parse_from_file ('file.pod', 'file.txt'); |
| 675 | |
| 676 | =head1 DESCRIPTION |
| 677 | |
| 678 | Pod::Text is a module that can convert documentation in the POD format (the |
| 679 | preferred language for documenting Perl) into formatted ASCII. It uses no |
| 680 | special formatting controls or codes whatsoever, and its output is therefore |
| 681 | suitable for nearly any device. |
| 682 | |
| 683 | As a derived class from Pod::Parser, Pod::Text supports the same methods and |
| 684 | interfaces. See L<Pod::Parser> for all the details; briefly, one creates a |
| 685 | new parser with C<< Pod::Text->new() >> and then calls either |
| 686 | parse_from_filehandle() or parse_from_file(). |
| 687 | |
| 688 | new() can take options, in the form of key/value pairs, that control the |
| 689 | behavior of the parser. The currently recognized options are: |
| 690 | |
| 691 | =over 4 |
| 692 | |
| 693 | =item alt |
| 694 | |
| 695 | If set to a true value, selects an alternate output format that, among other |
| 696 | things, uses a different heading style and marks C<=item> entries with a |
| 697 | colon in the left margin. Defaults to false. |
| 698 | |
| 699 | =item code |
| 700 | |
| 701 | If set to a true value, the non-POD parts of the input file will be included |
| 702 | in the output. Useful for viewing code documented with POD blocks with the |
| 703 | POD rendered and the code left intact. |
| 704 | |
| 705 | =item indent |
| 706 | |
| 707 | The number of spaces to indent regular text, and the default indentation for |
| 708 | C<=over> blocks. Defaults to 4. |
| 709 | |
| 710 | =item loose |
| 711 | |
| 712 | If set to a true value, a blank line is printed after a C<=head1> heading. |
| 713 | If set to false (the default), no blank line is printed after C<=head1>, |
| 714 | although one is still printed after C<=head2>. This is the default because |
| 715 | it's the expected formatting for manual pages; if you're formatting |
| 716 | arbitrary text documents, setting this to true may result in more pleasing |
| 717 | output. |
| 718 | |
| 719 | =item quotes |
| 720 | |
| 721 | Sets the quote marks used to surround CE<lt>> text. If the value is a |
| 722 | single character, it is used as both the left and right quote; if it is two |
| 723 | characters, the first character is used as the left quote and the second as |
| 724 | the right quoted; and if it is four characters, the first two are used as |
| 725 | the left quote and the second two as the right quote. |
| 726 | |
| 727 | This may also be set to the special value C<none>, in which case no quote |
| 728 | marks are added around CE<lt>> text. |
| 729 | |
| 730 | =item sentence |
| 731 | |
| 732 | If set to a true value, Pod::Text will assume that each sentence ends in two |
| 733 | spaces, and will try to preserve that spacing. If set to false, all |
| 734 | consecutive whitespace in non-verbatim paragraphs is compressed into a |
| 735 | single space. Defaults to true. |
| 736 | |
| 737 | =item width |
| 738 | |
| 739 | The column at which to wrap text on the right-hand side. Defaults to 76. |
| 740 | |
| 741 | =back |
| 742 | |
| 743 | The standard Pod::Parser method parse_from_filehandle() takes up to two |
| 744 | arguments, the first being the file handle to read POD from and the second |
| 745 | being the file handle to write the formatted output to. The first defaults |
| 746 | to STDIN if not given, and the second defaults to STDOUT. The method |
| 747 | parse_from_file() is almost identical, except that its two arguments are the |
| 748 | input and output disk files instead. See L<Pod::Parser> for the specific |
| 749 | details. |
| 750 | |
| 751 | =head1 DIAGNOSTICS |
| 752 | |
| 753 | =over 4 |
| 754 | |
| 755 | =item Bizarre space in item |
| 756 | |
| 757 | =item Item called without tag |
| 758 | |
| 759 | (W) Something has gone wrong in internal C<=item> processing. These |
| 760 | messages indicate a bug in Pod::Text; you should never see them. |
| 761 | |
| 762 | =item Can't open %s for reading: %s |
| 763 | |
| 764 | (F) Pod::Text was invoked via the compatibility mode pod2text() interface |
| 765 | and the input file it was given could not be opened. |
| 766 | |
| 767 | =item Invalid quote specification "%s" |
| 768 | |
| 769 | (F) The quote specification given (the quotes option to the constructor) was |
| 770 | invalid. A quote specification must be one, two, or four characters long. |
| 771 | |
| 772 | =item %s:%d: Unknown command paragraph: %s |
| 773 | |
| 774 | (W) The POD source contained a non-standard command paragraph (something of |
| 775 | the form C<=command args>) that Pod::Man didn't know about. It was ignored. |
| 776 | |
| 777 | =item %s:%d: Unknown escape: %s |
| 778 | |
| 779 | (W) The POD source contained an C<EE<lt>E<gt>> escape that Pod::Text didn't |
| 780 | know about. |
| 781 | |
| 782 | =item %s:%d: Unknown formatting code: %s |
| 783 | |
| 784 | (W) The POD source contained a non-standard formatting code (something of |
| 785 | the form C<XE<lt>E<gt>>) that Pod::Text didn't know about. |
| 786 | |
| 787 | =item %s:%d: Unmatched =back |
| 788 | |
| 789 | (W) Pod::Text encountered a C<=back> command that didn't correspond to an |
| 790 | C<=over> command. |
| 791 | |
| 792 | =back |
| 793 | |
| 794 | =head1 RESTRICTIONS |
| 795 | |
| 796 | Embedded Ctrl-As (octal 001) in the input will be mapped to spaces on |
| 797 | output, due to an internal implementation detail. |
| 798 | |
| 799 | =head1 NOTES |
| 800 | |
| 801 | This is a replacement for an earlier Pod::Text module written by Tom |
| 802 | Christiansen. It has a revamped interface, since it now uses Pod::Parser, |
| 803 | but an interface roughly compatible with the old Pod::Text::pod2text() |
| 804 | function is still available. Please change to the new calling convention, |
| 805 | though. |
| 806 | |
| 807 | The original Pod::Text contained code to do formatting via termcap |
| 808 | sequences, although it wasn't turned on by default and it was problematic to |
| 809 | get it to work at all. This rewrite doesn't even try to do that, but a |
| 810 | subclass of it does. Look for L<Pod::Text::Termcap>. |
| 811 | |
| 812 | =head1 SEE ALSO |
| 813 | |
| 814 | L<Pod::Parser>, L<Pod::Text::Termcap>, L<pod2text(1)> |
| 815 | |
| 816 | The current version of this module is always available from its web site at |
| 817 | L<http://www.eyrie.org/~eagle/software/podlators/>. It is also part of the |
| 818 | Perl core distribution as of 5.6.0. |
| 819 | |
| 820 | =head1 AUTHOR |
| 821 | |
| 822 | Russ Allbery <rra@stanford.edu>, based I<very> heavily on the original |
| 823 | Pod::Text by Tom Christiansen <tchrist@mox.perl.com> and its conversion to |
| 824 | Pod::Parser by Brad Appleton <bradapp@enteract.com>. |
| 825 | |
| 826 | =head1 COPYRIGHT AND LICENSE |
| 827 | |
| 828 | Copyright 1999, 2000, 2001, 2002 by Russ Allbery <rra@stanford.edu>. |
| 829 | |
| 830 | This program is free software; you may redistribute it and/or modify it |
| 831 | under the same terms as Perl itself. |
| 832 | |
| 833 | =cut |