aleph1.co.uk Git - yaffs-website/blob - vendor/michelf/php-markdown/Michelf/MarkdownExtra.php

   1 <?php
   2 /**
   3  * Markdown Extra - A text-to-HTML conversion tool for web writers
   4  *
   5  * @package   php-markdown
   6  * @author    Michel Fortin <michel.fortin@michelf.com>
   7  * @copyright 2004-2016 Michel Fortin <https://michelf.com/projects/php-markdown/>
   8  * @copyright (Original Markdown) 2004-2006 John Gruber <https://daringfireball.net/projects/markdown/>
   9  */
  10
  11 namespace Michelf;
  12
  13 /**
  14  * Markdown Extra Parser Class
  15  */
  16 class MarkdownExtra extends \Michelf\Markdown {
  17         /**
  18          * Configuration variables
  19          */
  20
  21         /**
  22          * Prefix for footnote ids.
  23          * @var string
  24          */
  25         public $fn_id_prefix = "";
  26
  27         /**
  28          * Optional title attribute for footnote links and backlinks.
  29          * @var string
  30          */
  31         public $fn_link_title     = "";
  32         public $fn_backlink_title = "";
  33
  34         /**
  35          * Optional class attribute for footnote links and backlinks.
  36          * @var string
  37          */
  38         public $fn_link_class     = "footnote-ref";
  39         public $fn_backlink_class = "footnote-backref";
  40
  41         /**
  42          * Content to be displayed within footnote backlinks. The default is '↩';
  43          * the U+FE0E on the end is a Unicode variant selector used to prevent iOS
  44          * from displaying the arrow character as an emoji.
  45          * @var string
  46          */
  47         public $fn_backlink_html = '&#8617;&#xFE0E;';
  48
  49         /**
  50          * Class name for table cell alignment (%% replaced left/center/right)
  51          * For instance: 'go-%%' becomes 'go-left' or 'go-right' or 'go-center'
  52          * If empty, the align attribute is used instead of a class name.
  53          * @var string
  54          */
  55         public $table_align_class_tmpl = '';
  56
  57         /**
  58          * Optional class prefix for fenced code block.
  59          * @var string
  60          */
  61         public $code_class_prefix = "";
  62
  63         /**
  64          * Class attribute for code blocks goes on the `code` tag;
  65          * setting this to true will put attributes on the `pre` tag instead.
  66          * @var boolean
  67          */
  68         public $code_attr_on_pre = false;
  69
  70         /**
  71          * Predefined abbreviations.
  72          * @var array
  73          */
  74         public $predef_abbr = array();
  75
  76         /**
  77          * Parser implementation
  78          */
  79
  80         /**
  81          * Constructor function. Initialize the parser object.
  82          * @return void
  83          */
  84         public function __construct() {
  85                 // Add extra escapable characters before parent constructor
  86                 // initialize the table.
  87                 $this->escape_chars .= ':|';
  88
  89                 // Insert extra document, block, and span transformations.
  90                 // Parent constructor will do the sorting.
  91                 $this->document_gamut += array(
  92                         "doFencedCodeBlocks" => 5,
  93                         "stripFootnotes"     => 15,
  94                         "stripAbbreviations" => 25,
  95                         "appendFootnotes"    => 50,
  96                 );
  97                 $this->block_gamut += array(
  98                         "doFencedCodeBlocks" => 5,
  99                         "doTables"           => 15,
 100                         "doDefLists"         => 45,
 101                 );
 102                 $this->span_gamut += array(
 103                         "doFootnotes"        => 5,
 104                         "doAbbreviations"    => 70,
 105                 );
 106
 107                 $this->enhanced_ordered_list = true;
 108                 parent::__construct();
 109         }
 110
 111
 112         /**
 113          * Extra variables used during extra transformations.
 114          * @var array
 115          */
 116         protected $footnotes = array();
 117         protected $footnotes_ordered = array();
 118         protected $footnotes_ref_count = array();
 119         protected $footnotes_numbers = array();
 120         protected $abbr_desciptions = array();
 121         /** @var @string */
 122         protected $abbr_word_re = '';
 123
 124         /**
 125          * Give the current footnote number.
 126          * @var integer
 127          */
 128         protected $footnote_counter = 1;
 129
 130         /**
 131          * Setting up Extra-specific variables.
 132          */
 133         protected function setup() {
 134                 parent::setup();
 135
 136                 $this->footnotes = array();
 137                 $this->footnotes_ordered = array();
 138                 $this->footnotes_ref_count = array();
 139                 $this->footnotes_numbers = array();
 140                 $this->abbr_desciptions = array();
 141                 $this->abbr_word_re = '';
 142                 $this->footnote_counter = 1;
 143
 144                 foreach ($this->predef_abbr as $abbr_word => $abbr_desc) {
 145                         if ($this->abbr_word_re)
 146                                 $this->abbr_word_re .= '|';
 147                         $this->abbr_word_re .= preg_quote($abbr_word);
 148                         $this->abbr_desciptions[$abbr_word] = trim($abbr_desc);
 149                 }
 150         }
 151
 152         /**
 153          * Clearing Extra-specific variables.
 154          */
 155         protected function teardown() {
 156                 $this->footnotes = array();
 157                 $this->footnotes_ordered = array();
 158                 $this->footnotes_ref_count = array();
 159                 $this->footnotes_numbers = array();
 160                 $this->abbr_desciptions = array();
 161                 $this->abbr_word_re = '';
 162
 163                 parent::teardown();
 164         }
 165
 166
 167         /**
 168          * Extra attribute parser
 169          */
 170
 171         /**
 172          * Expression to use to catch attributes (includes the braces)
 173          * @var string
 174          */
 175         protected $id_class_attr_catch_re = '\{((?>[ ]*[#.a-z][-_:a-zA-Z0-9=]+){1,})[ ]*\}';
 176
 177         /**
 178          * Expression to use when parsing in a context when no capture is desired
 179          * @var string
 180          */
 181         protected $id_class_attr_nocatch_re = '\{(?>[ ]*[#.a-z][-_:a-zA-Z0-9=]+){1,}[ ]*\}';
 182
 183         /**
 184          * Parse attributes caught by the $this->id_class_attr_catch_re expression
 185          * and return the HTML-formatted list of attributes.
 186          *
 187          * Currently supported attributes are .class and #id.
 188          *
 189          * In addition, this method also supports supplying a default Id value,
 190          * which will be used to populate the id attribute in case it was not
 191          * overridden.
 192          * @param  string $tag_name
 193          * @param  string $attr
 194          * @param  mixed  $defaultIdValue
 195          * @param  array  $classes
 196          * @return string
 197          */
 198         protected function doExtraAttributes($tag_name, $attr, $defaultIdValue = null, $classes = array()) {
 199                 if (empty($attr) && !$defaultIdValue && empty($classes)) return "";
 200
 201                 // Split on components
 202                 preg_match_all('/[#.a-z][-_:a-zA-Z0-9=]+/', $attr, $matches);
 203                 $elements = $matches[0];
 204
 205                 // Handle classes and IDs (only first ID taken into account)
 206                 $attributes = array();
 207                 $id = false;
 208                 foreach ($elements as $element) {
 209                         if ($element{0} == '.') {
 210                                 $classes[] = substr($element, 1);
 211                         } else if ($element{0} == '#') {
 212                                 if ($id === false) $id = substr($element, 1);
 213                         } else if (strpos($element, '=') > 0) {
 214                                 $parts = explode('=', $element, 2);
 215                                 $attributes[] = $parts[0] . '="' . $parts[1] . '"';
 216                         }
 217                 }
 218
 219                 if (!$id) $id = $defaultIdValue;
 220
 221                 // Compose attributes as string
 222                 $attr_str = "";
 223                 if (!empty($id)) {
 224                         $attr_str .= ' id="'.$this->encodeAttribute($id) .'"';
 225                 }
 226                 if (!empty($classes)) {
 227                         $attr_str .= ' class="'. implode(" ", $classes) . '"';
 228                 }
 229                 if (!$this->no_markup && !empty($attributes)) {
 230                         $attr_str .= ' '.implode(" ", $attributes);
 231                 }
 232                 return $attr_str;
 233         }
 234
 235         /**
 236          * Strips link definitions from text, stores the URLs and titles in
 237          * hash references.
 238          * @param  string $text
 239          * @return string
 240          */
 241         protected function stripLinkDefinitions($text) {
 242                 $less_than_tab = $this->tab_width - 1;
 243
 244                 // Link defs are in the form: ^[id]: url "optional title"
 245                 $text = preg_replace_callback('{
 246                                                         ^[ ]{0,'.$less_than_tab.'}\[(.+)\][ ]?: # id = $1
 247                                                           [ ]*
 248                                                           \n?                           # maybe *one* newline
 249                                                           [ ]*
 250                                                         (?:
 251                                                           <(.+?)>                       # url = $2
 252                                                         |
 253                                                           (\S+?)                        # url = $3
 254                                                         )
 255                                                           [ ]*
 256                                                           \n?                           # maybe one newline
 257                                                           [ ]*
 258                                                         (?:
 259                                                                 (?<=\s)                 # lookbehind for whitespace
 260                                                                 ["(]
 261                                                                 (.*?)                   # title = $4
 262                                                                 [")]
 263                                                                 [ ]*
 264                                                         )?      # title is optional
 265                                         (?:[ ]* '.$this->id_class_attr_catch_re.' )?  # $5 = extra id & class attr
 266                                                         (?:\n+|\Z)
 267                         }xm',
 268                         array($this, '_stripLinkDefinitions_callback'),
 269                         $text);
 270                 return $text;
 271         }
 272
 273         /**
 274          * Strip link definition callback
 275          * @param  array $matches
 276          * @return string
 277          */
 278         protected function _stripLinkDefinitions_callback($matches) {
 279                 $link_id = strtolower($matches[1]);
 280                 $url = $matches[2] == '' ? $matches[3] : $matches[2];
 281                 $this->urls[$link_id] = $url;
 282                 $this->titles[$link_id] =& $matches[4];
 283                 $this->ref_attr[$link_id] = $this->doExtraAttributes("", $dummy =& $matches[5]);
 284                 return ''; // String that will replace the block
 285         }
 286
 287
 288         /**
 289          * HTML block parser
 290          */
 291
 292         /**
 293          * Tags that are always treated as block tags
 294          * @var string
 295          */
 296         protected $block_tags_re = 'p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|address|form|fieldset|iframe|hr|legend|article|section|nav|aside|hgroup|header|footer|figcaption|figure';
 297
 298         /**
 299          * Tags treated as block tags only if the opening tag is alone on its line
 300          * @var string
 301          */
 302         protected $context_block_tags_re = 'script|noscript|style|ins|del|iframe|object|source|track|param|math|svg|canvas|audio|video';
 303
 304         /**
 305          * Tags where markdown="1" default to span mode:
 306          * @var string
 307          */
 308         protected $contain_span_tags_re = 'p|h[1-6]|li|dd|dt|td|th|legend|address';
 309
 310         /**
 311          * Tags which must not have their contents modified, no matter where
 312          * they appear
 313          * @var string
 314          */
 315         protected $clean_tags_re = 'script|style|math|svg';
 316
 317         /**
 318          * Tags that do not need to be closed.
 319          * @var string
 320          */
 321         protected $auto_close_tags_re = 'hr|img|param|source|track';
 322
 323         /**
 324          * Hashify HTML Blocks and "clean tags".
 325          *
 326          * We only want to do this for block-level HTML tags, such as headers,
 327          * lists, and tables. That's because we still want to wrap <p>s around
 328          * "paragraphs" that are wrapped in non-block-level tags, such as anchors,
 329          * phrase emphasis, and spans. The list of tags we're looking for is
 330          * hard-coded.
 331          *
 332          * This works by calling _HashHTMLBlocks_InMarkdown, which then calls
 333          * _HashHTMLBlocks_InHTML when it encounter block tags. When the markdown="1"
 334          * attribute is found within a tag, _HashHTMLBlocks_InHTML calls back
 335          *  _HashHTMLBlocks_InMarkdown to handle the Markdown syntax within the tag.
 336          * These two functions are calling each other. It's recursive!
 337          * @param  string $text
 338          * @return string
 339          */
 340         protected function hashHTMLBlocks($text) {
 341                 if ($this->no_markup) {
 342                         return $text;
 343                 }
 344
 345                 // Call the HTML-in-Markdown hasher.
 346                 list($text, ) = $this->_hashHTMLBlocks_inMarkdown($text);
 347
 348                 return $text;
 349         }
 350
 351         /**
 352          * Parse markdown text, calling _HashHTMLBlocks_InHTML for block tags.
 353          *
 354          * *   $indent is the number of space to be ignored when checking for code
 355          *     blocks. This is important because if we don't take the indent into
 356          *     account, something like this (which looks right) won't work as expected:
 357          *
 358          *     <div>
 359          *         <div markdown="1">
 360          *         Hello World.  <-- Is this a Markdown code block or text?
 361          *         </div>  <-- Is this a Markdown code block or a real tag?
 362          *     <div>
 363          *
 364          *     If you don't like this, just don't indent the tag on which
 365          *     you apply the markdown="1" attribute.
 366          *
 367          * *   If $enclosing_tag_re is not empty, stops at the first unmatched closing
 368          *     tag with that name. Nested tags supported.
 369          *
 370          * *   If $span is true, text inside must treated as span. So any double
 371          *     newline will be replaced by a single newline so that it does not create
 372          *     paragraphs.
 373          *
 374          * Returns an array of that form: ( processed text , remaining text )
 375          *
 376          * @param  string  $text
 377          * @param  integer $indent
 378          * @param  string  $enclosing_tag_re
 379          * @param  boolean $span
 380          * @return array
 381          */
 382         protected function _hashHTMLBlocks_inMarkdown($text, $indent = 0,
 383                                                                                 $enclosing_tag_re = '', $span = false)
 384         {
 385
 386                 if ($text === '') return array('', '');
 387
 388                 // Regex to check for the presense of newlines around a block tag.
 389                 $newline_before_re = '/(?:^\n?|\n\n)*$/';
 390                 $newline_after_re =
 391                         '{
 392                                 ^                                               # Start of text following the tag.
 393                                 (?>[ ]*<!--.*?-->)?             # Optional comment.
 394                                 [ ]*\n                                  # Must be followed by newline.
 395                         }xs';
 396
 397                 // Regex to match any tag.
 398                 $block_tag_re =
 399                         '{
 400                                 (                                       # $2: Capture whole tag.
 401                                         </?                                     # Any opening or closing tag.
 402                                                 (?>                             # Tag name.
 403                                                         ' . $this->block_tags_re . '                    |
 404                                                         ' . $this->context_block_tags_re . '    |
 405                                                         ' . $this->clean_tags_re . '            |
 406                                                         (?!\s)'.$enclosing_tag_re . '
 407                                                 )
 408                                                 (?:
 409                                                         (?=[\s"\'/a-zA-Z0-9])   # Allowed characters after tag name.
 410                                                         (?>
 411                                                                 ".*?"           |       # Double quotes (can contain `>`)
 412                                                                 \'.*?\'         |       # Single quotes (can contain `>`)
 413                                                                 .+?                             # Anything but quotes and `>`.
 414                                                         )*?
 415                                                 )?
 416                                         >                                       # End of tag.
 417                                 |
 418                                         <!--    .*?     -->     # HTML Comment
 419                                 |
 420                                         <\?.*?\?> | <%.*?%>     # Processing instruction
 421                                 |
 422                                         <!\[CDATA\[.*?\]\]>     # CData Block
 423                                 ' . ( !$span ? ' # If not in span.
 424                                 |
 425                                         # Indented code block
 426                                         (?: ^[ ]*\n | ^ | \n[ ]*\n )
 427                                         [ ]{' . ($indent + 4) . '}[^\n]* \n
 428                                         (?>
 429                                                 (?: [ ]{' . ($indent + 4) . '}[^\n]* | [ ]* ) \n
 430                                         )*
 431                                 |
 432                                         # Fenced code block marker
 433                                         (?<= ^ | \n )
 434                                         [ ]{0,' . ($indent + 3) . '}(?:~{3,}|`{3,})
 435                                         [ ]*
 436                                         (?: \.?[-_:a-zA-Z0-9]+ )? # standalone class name
 437                                         [ ]*
 438                                         (?: ' . $this->id_class_attr_nocatch_re . ' )? # extra attributes
 439                                         [ ]*
 440                                         (?= \n )
 441                                 ' : '' ) . ' # End (if not is span).
 442                                 |
 443                                         # Code span marker
 444                                         # Note, this regex needs to go after backtick fenced
 445                                         # code blocks but it should also be kept outside of the
 446                                         # "if not in span" condition adding backticks to the parser
 447                                         `+
 448                                 )
 449                         }xs';
 450
 451
 452                 $depth = 0;             // Current depth inside the tag tree.
 453                 $parsed = "";   // Parsed text that will be returned.
 454
 455                 // Loop through every tag until we find the closing tag of the parent
 456                 // or loop until reaching the end of text if no parent tag specified.
 457                 do {
 458                         // Split the text using the first $tag_match pattern found.
 459                         // Text before  pattern will be first in the array, text after
 460                         // pattern will be at the end, and between will be any catches made
 461                         // by the pattern.
 462                         $parts = preg_split($block_tag_re, $text, 2,
 463                                                                 PREG_SPLIT_DELIM_CAPTURE);
 464
 465                         // If in Markdown span mode, add a empty-string span-level hash
 466                         // after each newline to prevent triggering any block element.
 467                         if ($span) {
 468                                 $void = $this->hashPart("", ':');
 469                                 $newline = "\n$void";
 470                                 $parts[0] = $void . str_replace("\n", $newline, $parts[0]) . $void;
 471                         }
 472
 473                         $parsed .= $parts[0]; // Text before current tag.
 474
 475                         // If end of $text has been reached. Stop loop.
 476                         if (count($parts) < 3) {
 477                                 $text = "";
 478                                 break;
 479                         }
 480
 481                         $tag  = $parts[1]; // Tag to handle.
 482                         $text = $parts[2]; // Remaining text after current tag.
 483                         $tag_re = preg_quote($tag); // For use in a regular expression.
 484
 485                         // Check for: Fenced code block marker.
 486                         // Note: need to recheck the whole tag to disambiguate backtick
 487                         // fences from code spans
 488                         if (preg_match('{^\n?([ ]{0,' . ($indent + 3) . '})(~{3,}|`{3,})[ ]*(?:\.?[-_:a-zA-Z0-9]+)?[ ]*(?:' . $this->id_class_attr_nocatch_re . ')?[ ]*\n?$}', $tag, $capture)) {
 489                                 // Fenced code block marker: find matching end marker.
 490                                 $fence_indent = strlen($capture[1]); // use captured indent in re
 491                                 $fence_re = $capture[2]; // use captured fence in re
 492                                 if (preg_match('{^(?>.*\n)*?[ ]{' . ($fence_indent) . '}' . $fence_re . '[ ]*(?:\n|$)}', $text,
 493                                         $matches))
 494                                 {
 495                                         // End marker found: pass text unchanged until marker.
 496                                         $parsed .= $tag . $matches[0];
 497                                         $text = substr($text, strlen($matches[0]));
 498                                 }
 499                                 else {
 500                                         // No end marker: just skip it.
 501                                         $parsed .= $tag;
 502                                 }
 503                         }
 504                         // Check for: Indented code block.
 505                         else if ($tag{0} == "\n" || $tag{0} == " ") {
 506                                 // Indented code block: pass it unchanged, will be handled
 507                                 // later.
 508                                 $parsed .= $tag;
 509                         }
 510                         // Check for: Code span marker
 511                         // Note: need to check this after backtick fenced code blocks
 512                         else if ($tag{0} == "`") {
 513                                 // Find corresponding end marker.
 514                                 $tag_re = preg_quote($tag);
 515                                 if (preg_match('{^(?>.+?|\n(?!\n))*?(?<!`)' . $tag_re . '(?!`)}',
 516                                         $text, $matches))
 517                                 {
 518                                         // End marker found: pass text unchanged until marker.
 519                                         $parsed .= $tag . $matches[0];
 520                                         $text = substr($text, strlen($matches[0]));
 521                                 }
 522                                 else {
 523                                         // Unmatched marker: just skip it.
 524                                         $parsed .= $tag;
 525                                 }
 526                         }
 527                         // Check for: Opening Block level tag or
 528                         //            Opening Context Block tag (like ins and del)
 529                         //               used as a block tag (tag is alone on it's line).
 530                         else if (preg_match('{^<(?:' . $this->block_tags_re . ')\b}', $tag) ||
 531                                 (       preg_match('{^<(?:' . $this->context_block_tags_re . ')\b}', $tag) &&
 532                                         preg_match($newline_before_re, $parsed) &&
 533                                         preg_match($newline_after_re, $text)    )
 534                                 )
 535                         {
 536                                 // Need to parse tag and following text using the HTML parser.
 537                                 list($block_text, $text) =
 538                                         $this->_hashHTMLBlocks_inHTML($tag . $text, "hashBlock", true);
 539
 540                                 // Make sure it stays outside of any paragraph by adding newlines.
 541                                 $parsed .= "\n\n$block_text\n\n";
 542                         }
 543                         // Check for: Clean tag (like script, math)
 544                         //            HTML Comments, processing instructions.
 545                         else if (preg_match('{^<(?:' . $this->clean_tags_re . ')\b}', $tag) ||
 546                                 $tag{1} == '!' || $tag{1} == '?')
 547                         {
 548                                 // Need to parse tag and following text using the HTML parser.
 549                                 // (don't check for markdown attribute)
 550                                 list($block_text, $text) =
 551                                         $this->_hashHTMLBlocks_inHTML($tag . $text, "hashClean", false);
 552
 553                                 $parsed .= $block_text;
 554                         }
 555                         // Check for: Tag with same name as enclosing tag.
 556                         else if ($enclosing_tag_re !== '' &&
 557                                 // Same name as enclosing tag.
 558                                 preg_match('{^</?(?:' . $enclosing_tag_re . ')\b}', $tag))
 559                         {
 560                                 // Increase/decrease nested tag count.
 561                                 if ($tag{1} == '/')                                             $depth--;
 562                                 else if ($tag{strlen($tag)-2} != '/')   $depth++;
 563
 564                                 if ($depth < 0) {
 565                                         // Going out of parent element. Clean up and break so we
 566                                         // return to the calling function.
 567                                         $text = $tag . $text;
 568                                         break;
 569                                 }
 570
 571                                 $parsed .= $tag;
 572                         }
 573                         else {
 574                                 $parsed .= $tag;
 575                         }
 576                 } while ($depth >= 0);
 577
 578                 return array($parsed, $text);
 579         }
 580
 581         /**
 582          * Parse HTML, calling _HashHTMLBlocks_InMarkdown for block tags.
 583          *
 584          * *   Calls $hash_method to convert any blocks.
 585          * *   Stops when the first opening tag closes.
 586          * *   $md_attr indicate if the use of the `markdown="1"` attribute is allowed.
 587          *     (it is not inside clean tags)
 588          *
 589          * Returns an array of that form: ( processed text , remaining text )
 590          * @param  string $text
 591          * @param  string $hash_method
 592          * @param  string $md_attr
 593          * @return array
 594          */
 595         protected function _hashHTMLBlocks_inHTML($text, $hash_method, $md_attr) {
 596                 if ($text === '') return array('', '');
 597
 598                 // Regex to match `markdown` attribute inside of a tag.
 599                 $markdown_attr_re = '
 600                         {
 601                                 \s*                     # Eat whitespace before the `markdown` attribute
 602                                 markdown
 603                                 \s*=\s*
 604                                 (?>
 605                                         (["\'])         # $1: quote delimiter
 606                                         (.*?)           # $2: attribute value
 607                                         \1                      # matching delimiter
 608                                 |
 609                                         ([^\s>]*)       # $3: unquoted attribute value
 610                                 )
 611                                 ()                              # $4: make $3 always defined (avoid warnings)
 612                         }xs';
 613
 614                 // Regex to match any tag.
 615                 $tag_re = '{
 616                                 (                                       # $2: Capture whole tag.
 617                                         </?                                     # Any opening or closing tag.
 618                                                 [\w:$]+                 # Tag name.
 619                                                 (?:
 620                                                         (?=[\s"\'/a-zA-Z0-9])   # Allowed characters after tag name.
 621                                                         (?>
 622                                                                 ".*?"           |       # Double quotes (can contain `>`)
 623                                                                 \'.*?\'         |       # Single quotes (can contain `>`)
 624                                                                 .+?                             # Anything but quotes and `>`.
 625                                                         )*?
 626                                                 )?
 627                                         >                                       # End of tag.
 628                                 |
 629                                         <!--    .*?     -->     # HTML Comment
 630                                 |
 631                                         <\?.*?\?> | <%.*?%>     # Processing instruction
 632                                 |
 633                                         <!\[CDATA\[.*?\]\]>     # CData Block
 634                                 )
 635                         }xs';
 636
 637                 $original_text = $text;         // Save original text in case of faliure.
 638
 639                 $depth          = 0;    // Current depth inside the tag tree.
 640                 $block_text     = "";   // Temporary text holder for current text.
 641                 $parsed         = "";   // Parsed text that will be returned.
 642
 643                 // Get the name of the starting tag.
 644                 // (This pattern makes $base_tag_name_re safe without quoting.)
 645                 if (preg_match('/^<([\w:$]*)\b/', $text, $matches))
 646                         $base_tag_name_re = $matches[1];
 647
 648                 // Loop through every tag until we find the corresponding closing tag.
 649                 do {
 650                         // Split the text using the first $tag_match pattern found.
 651                         // Text before  pattern will be first in the array, text after
 652                         // pattern will be at the end, and between will be any catches made
 653                         // by the pattern.
 654                         $parts = preg_split($tag_re, $text, 2, PREG_SPLIT_DELIM_CAPTURE);
 655
 656                         if (count($parts) < 3) {
 657                                 // End of $text reached with unbalenced tag(s).
 658                                 // In that case, we return original text unchanged and pass the
 659                                 // first character as filtered to prevent an infinite loop in the
 660                                 // parent function.
 661                                 return array($original_text{0}, substr($original_text, 1));
 662                         }
 663
 664                         $block_text .= $parts[0]; // Text before current tag.
 665                         $tag         = $parts[1]; // Tag to handle.
 666                         $text        = $parts[2]; // Remaining text after current tag.
 667
 668                         // Check for: Auto-close tag (like <hr/>)
 669                         //                       Comments and Processing Instructions.
 670                         if (preg_match('{^</?(?:' . $this->auto_close_tags_re . ')\b}', $tag) ||
 671                                 $tag{1} == '!' || $tag{1} == '?')
 672                         {
 673                                 // Just add the tag to the block as if it was text.
 674                                 $block_text .= $tag;
 675                         }
 676                         else {
 677                                 // Increase/decrease nested tag count. Only do so if
 678                                 // the tag's name match base tag's.
 679                                 if (preg_match('{^</?' . $base_tag_name_re . '\b}', $tag)) {
 680                                         if ($tag{1} == '/')                                             $depth--;
 681                                         else if ($tag{strlen($tag)-2} != '/')   $depth++;
 682                                 }
 683
 684                                 // Check for `markdown="1"` attribute and handle it.
 685                                 if ($md_attr &&
 686                                         preg_match($markdown_attr_re, $tag, $attr_m) &&
 687                                         preg_match('/^1|block|span$/', $attr_m[2] . $attr_m[3]))
 688                                 {
 689                                         // Remove `markdown` attribute from opening tag.
 690                                         $tag = preg_replace($markdown_attr_re, '', $tag);
 691
 692                                         // Check if text inside this tag must be parsed in span mode.
 693                                         $this->mode = $attr_m[2] . $attr_m[3];
 694                                         $span_mode = $this->mode == 'span' || $this->mode != 'block' &&
 695                                                 preg_match('{^<(?:' . $this->contain_span_tags_re . ')\b}', $tag);
 696
 697                                         // Calculate indent before tag.
 698                                         if (preg_match('/(?:^|\n)( *?)(?! ).*?$/', $block_text, $matches)) {
 699                                                 $strlen = $this->utf8_strlen;
 700                                                 $indent = $strlen($matches[1], 'UTF-8');
 701                                         } else {
 702                                                 $indent = 0;
 703                                         }
 704
 705                                         // End preceding block with this tag.
 706                                         $block_text .= $tag;
 707                                         $parsed .= $this->$hash_method($block_text);
 708
 709                                         // Get enclosing tag name for the ParseMarkdown function.
 710                                         // (This pattern makes $tag_name_re safe without quoting.)
 711                                         preg_match('/^<([\w:$]*)\b/', $tag, $matches);
 712                                         $tag_name_re = $matches[1];
 713
 714                                         // Parse the content using the HTML-in-Markdown parser.
 715                                         list ($block_text, $text)
 716                                                 = $this->_hashHTMLBlocks_inMarkdown($text, $indent,
 717                                                         $tag_name_re, $span_mode);
 718
 719                                         // Outdent markdown text.
 720                                         if ($indent > 0) {
 721                                                 $block_text = preg_replace("/^[ ]{1,$indent}/m", "",
 722                                                                                                         $block_text);
 723                                         }
 724
 725                                         // Append tag content to parsed text.
 726                                         if (!$span_mode)        $parsed .= "\n\n$block_text\n\n";
 727                                         else                            $parsed .= "$block_text";
 728
 729                                         // Start over with a new block.
 730                                         $block_text = "";
 731                                 }
 732                                 else $block_text .= $tag;
 733                         }
 734
 735                 } while ($depth > 0);
 736
 737                 // Hash last block text that wasn't processed inside the loop.
 738                 $parsed .= $this->$hash_method($block_text);
 739
 740                 return array($parsed, $text);
 741         }
 742
 743         /**
 744          * Called whenever a tag must be hashed when a function inserts a "clean" tag
 745          * in $text, it passes through this function and is automaticaly escaped,
 746          * blocking invalid nested overlap.
 747          * @param  string $text
 748          * @return string
 749          */
 750         protected function hashClean($text) {
 751                 return $this->hashPart($text, 'C');
 752         }
 753
 754         /**
 755          * Turn Markdown link shortcuts into XHTML <a> tags.
 756          * @param  string $text
 757          * @return string
 758          */
 759         protected function doAnchors($text) {
 760                 if ($this->in_anchor) {
 761                         return $text;
 762                 }
 763                 $this->in_anchor = true;
 764
 765                 // First, handle reference-style links: [link text] [id]
 766                 $text = preg_replace_callback('{
 767                         (                                       # wrap whole match in $1
 768                           \[
 769                                 (' . $this->nested_brackets_re . ')     # link text = $2
 770                           \]
 771
 772                           [ ]?                          # one optional space
 773                           (?:\n[ ]*)?           # one optional newline followed by spaces
 774
 775                           \[
 776                                 (.*?)           # id = $3
 777                           \]
 778                         )
 779                         }xs',
 780                         array($this, '_doAnchors_reference_callback'), $text);
 781
 782                 // Next, inline-style links: [link text](url "optional title")
 783                 $text = preg_replace_callback('{
 784                         (                               # wrap whole match in $1
 785                           \[
 786                                 (' . $this->nested_brackets_re . ')     # link text = $2
 787                           \]
 788                           \(                    # literal paren
 789                                 [ \n]*
 790                                 (?:
 791                                         <(.+?)> # href = $3
 792                                 |
 793                                         (' . $this->nested_url_parenthesis_re . ')      # href = $4
 794                                 )
 795                                 [ \n]*
 796                                 (                       # $5
 797                                   ([\'"])       # quote char = $6
 798                                   (.*?)         # Title = $7
 799                                   \6            # matching quote
 800                                   [ \n]*        # ignore any spaces/tabs between closing quote and )
 801                                 )?                      # title is optional
 802                           \)
 803                           (?:[ ]? ' . $this->id_class_attr_catch_re . ' )?       # $8 = id/class attributes
 804                         )
 805                         }xs',
 806                         array($this, '_doAnchors_inline_callback'), $text);
 807
 808                 // Last, handle reference-style shortcuts: [link text]
 809                 // These must come last in case you've also got [link text][1]
 810                 // or [link text](/foo)
 811                 $text = preg_replace_callback('{
 812                         (                                       # wrap whole match in $1
 813                           \[
 814                                 ([^\[\]]+)              # link text = $2; can\'t contain [ or ]
 815                           \]
 816                         )
 817                         }xs',
 818                         array($this, '_doAnchors_reference_callback'), $text);
 819
 820                 $this->in_anchor = false;
 821                 return $text;
 822         }
 823
 824         /**
 825          * Callback for reference anchors
 826          * @param  array $matches
 827          * @return string
 828          */
 829         protected function _doAnchors_reference_callback($matches) {
 830                 $whole_match =  $matches[1];
 831                 $link_text   =  $matches[2];
 832                 $link_id     =& $matches[3];
 833
 834                 if ($link_id == "") {
 835                         // for shortcut links like [this][] or [this].
 836                         $link_id = $link_text;
 837                 }
 838
 839                 // lower-case and turn embedded newlines into spaces
 840                 $link_id = strtolower($link_id);
 841                 $link_id = preg_replace('{[ ]?\n}', ' ', $link_id);
 842
 843                 if (isset($this->urls[$link_id])) {
 844                         $url = $this->urls[$link_id];
 845                         $url = $this->encodeURLAttribute($url);
 846
 847                         $result = "<a href=\"$url\"";
 848                         if ( isset( $this->titles[$link_id] ) ) {
 849                                 $title = $this->titles[$link_id];
 850                                 $title = $this->encodeAttribute($title);
 851                                 $result .=  " title=\"$title\"";
 852                         }
 853                         if (isset($this->ref_attr[$link_id]))
 854                                 $result .= $this->ref_attr[$link_id];
 855
 856                         $link_text = $this->runSpanGamut($link_text);
 857                         $result .= ">$link_text</a>";
 858                         $result = $this->hashPart($result);
 859                 }
 860                 else {
 861                         $result = $whole_match;
 862                 }
 863                 return $result;
 864         }
 865
 866         /**
 867          * Callback for inline anchors
 868          * @param  array $matches
 869          * @return string
 870          */
 871         protected function _doAnchors_inline_callback($matches) {
 872                 $whole_match    =  $matches[1];
 873                 $link_text              =  $this->runSpanGamut($matches[2]);
 874                 $url                    =  $matches[3] == '' ? $matches[4] : $matches[3];
 875                 $title                  =& $matches[7];
 876                 $attr  = $this->doExtraAttributes("a", $dummy =& $matches[8]);
 877
 878                 // if the URL was of the form <s p a c e s> it got caught by the HTML
 879                 // tag parser and hashed. Need to reverse the process before using the URL.
 880                 $unhashed = $this->unhash($url);
 881                 if ($unhashed != $url)
 882                         $url = preg_replace('/^<(.*)>$/', '\1', $unhashed);
 883
 884                 $url = $this->encodeURLAttribute($url);
 885
 886                 $result = "<a href=\"$url\"";
 887                 if (isset($title)) {
 888                         $title = $this->encodeAttribute($title);
 889                         $result .=  " title=\"$title\"";
 890                 }
 891                 $result .= $attr;
 892
 893                 $link_text = $this->runSpanGamut($link_text);
 894                 $result .= ">$link_text</a>";
 895
 896                 return $this->hashPart($result);
 897         }
 898
 899         /**
 900          * Turn Markdown image shortcuts into <img> tags.
 901          * @param  string $text
 902          * @return string
 903          */
 904         protected function doImages($text) {
 905                 // First, handle reference-style labeled images: ![alt text][id]
 906                 $text = preg_replace_callback('{
 907                         (                               # wrap whole match in $1
 908                           !\[
 909                                 (' . $this->nested_brackets_re . ')             # alt text = $2
 910                           \]
 911
 912                           [ ]?                          # one optional space
 913                           (?:\n[ ]*)?           # one optional newline followed by spaces
 914
 915                           \[
 916                                 (.*?)           # id = $3
 917                           \]
 918
 919                         )
 920                         }xs',
 921                         array($this, '_doImages_reference_callback'), $text);
 922
 923                 // Next, handle inline images:  ![alt text](url "optional title")
 924                 // Don't forget: encode * and _
 925                 $text = preg_replace_callback('{
 926                         (                               # wrap whole match in $1
 927                           !\[
 928                                 (' . $this->nested_brackets_re . ')             # alt text = $2
 929                           \]
 930                           \s?                   # One optional whitespace character
 931                           \(                    # literal paren
 932                                 [ \n]*
 933                                 (?:
 934                                         <(\S*)> # src url = $3
 935                                 |
 936                                         (' . $this->nested_url_parenthesis_re . ')      # src url = $4
 937                                 )
 938                                 [ \n]*
 939                                 (                       # $5
 940                                   ([\'"])       # quote char = $6
 941                                   (.*?)         # title = $7
 942                                   \6            # matching quote
 943                                   [ \n]*
 944                                 )?                      # title is optional
 945                           \)
 946                           (?:[ ]? ' . $this->id_class_attr_catch_re . ' )?       # $8 = id/class attributes
 947                         )
 948                         }xs',
 949                         array($this, '_doImages_inline_callback'), $text);
 950
 951                 return $text;
 952         }
 953
 954         /**
 955          * Callback for referenced images
 956          * @param  array $matches
 957          * @return string
 958          */
 959         protected function _doImages_reference_callback($matches) {
 960                 $whole_match = $matches[1];
 961                 $alt_text    = $matches[2];
 962                 $link_id     = strtolower($matches[3]);
 963
 964                 if ($link_id == "") {
 965                         $link_id = strtolower($alt_text); // for shortcut links like ![this][].
 966                 }
 967
 968                 $alt_text = $this->encodeAttribute($alt_text);
 969                 if (isset($this->urls[$link_id])) {
 970                         $url = $this->encodeURLAttribute($this->urls[$link_id]);
 971                         $result = "<img src=\"$url\" alt=\"$alt_text\"";
 972                         if (isset($this->titles[$link_id])) {
 973                                 $title = $this->titles[$link_id];
 974                                 $title = $this->encodeAttribute($title);
 975                                 $result .=  " title=\"$title\"";
 976                         }
 977                         if (isset($this->ref_attr[$link_id]))
 978                                 $result .= $this->ref_attr[$link_id];
 979                         $result .= $this->empty_element_suffix;
 980                         $result = $this->hashPart($result);
 981                 }
 982                 else {
 983                         // If there's no such link ID, leave intact:
 984                         $result = $whole_match;
 985                 }
 986
 987                 return $result;
 988         }
 989
 990         /**
 991          * Callback for inline images
 992          * @param  array $matches
 993          * @return string
 994          */
 995         protected function _doImages_inline_callback($matches) {
 996                 $whole_match    = $matches[1];
 997                 $alt_text               = $matches[2];
 998                 $url                    = $matches[3] == '' ? $matches[4] : $matches[3];
 999                 $title                  =& $matches[7];
1000                 $attr  = $this->doExtraAttributes("img", $dummy =& $matches[8]);
1001
1002                 $alt_text = $this->encodeAttribute($alt_text);
1003                 $url = $this->encodeURLAttribute($url);
1004                 $result = "<img src=\"$url\" alt=\"$alt_text\"";
1005                 if (isset($title)) {
1006                         $title = $this->encodeAttribute($title);
1007                         $result .=  " title=\"$title\""; // $title already quoted
1008                 }
1009                 $result .= $attr;
1010                 $result .= $this->empty_element_suffix;
1011
1012                 return $this->hashPart($result);
1013         }
1014
1015         /**
1016          * Process markdown headers. Redefined to add ID and class attribute support.
1017          * @param  string $text
1018          * @return string
1019          */
1020         protected function doHeaders($text) {
1021                 // Setext-style headers:
1022                 //  Header 1  {#header1}
1023                 //        ========
1024                 //
1025                 //        Header 2  {#header2 .class1 .class2}
1026                 //        --------
1027                 //
1028                 $text = preg_replace_callback(
1029                         '{
1030                                 (^.+?)                                                          # $1: Header text
1031                                 (?:[ ]+ ' . $this->id_class_attr_catch_re . ' )?         # $3 = id/class attributes
1032                                 [ ]*\n(=+|-+)[ ]*\n+                            # $3: Header footer
1033                         }mx',
1034                         array($this, '_doHeaders_callback_setext'), $text);
1035
1036                 // atx-style headers:
1037                 //      # Header 1        {#header1}
1038                 //      ## Header 2       {#header2}
1039                 //      ## Header 2 with closing hashes ##  {#header3.class1.class2}
1040                 //      ...
1041                 //      ###### Header 6   {.class2}
1042                 //
1043                 $text = preg_replace_callback('{
1044                                 ^(\#{1,6})      # $1 = string of #\'s
1045                                 [ ]*
1046                                 (.+?)           # $2 = Header text
1047                                 [ ]*
1048                                 \#*                     # optional closing #\'s (not counted)
1049                                 (?:[ ]+ ' . $this->id_class_attr_catch_re . ' )?         # $3 = id/class attributes
1050                                 [ ]*
1051                                 \n+
1052                         }xm',
1053                         array($this, '_doHeaders_callback_atx'), $text);
1054
1055                 return $text;
1056         }
1057
1058         /**
1059          * Callback for setext headers
1060          * @param  array $matches
1061          * @return string
1062          */
1063         protected function _doHeaders_callback_setext($matches) {
1064                 if ($matches[3] == '-' && preg_match('{^- }', $matches[1])) {
1065                         return $matches[0];
1066                 }
1067
1068                 $level = $matches[3]{0} == '=' ? 1 : 2;
1069
1070                 $defaultId = is_callable($this->header_id_func) ? call_user_func($this->header_id_func, $matches[1]) : null;
1071
1072                 $attr  = $this->doExtraAttributes("h$level", $dummy =& $matches[2], $defaultId);
1073                 $block = "<h$level$attr>" . $this->runSpanGamut($matches[1]) . "</h$level>";
1074                 return "\n" . $this->hashBlock($block) . "\n\n";
1075         }
1076
1077         /**
1078          * Callback for atx headers
1079          * @param  array $matches
1080          * @return string
1081          */
1082         protected function _doHeaders_callback_atx($matches) {
1083                 $level = strlen($matches[1]);
1084
1085                 $defaultId = is_callable($this->header_id_func) ? call_user_func($this->header_id_func, $matches[2]) : null;
1086                 $attr  = $this->doExtraAttributes("h$level", $dummy =& $matches[3], $defaultId);
1087                 $block = "<h$level$attr>" . $this->runSpanGamut($matches[2]) . "</h$level>";
1088                 return "\n" . $this->hashBlock($block) . "\n\n";
1089         }
1090
1091         /**
1092          * Form HTML tables.
1093          * @param  string $text
1094          * @return string
1095          */
1096         protected function doTables($text) {
1097                 $less_than_tab = $this->tab_width - 1;
1098                 // Find tables with leading pipe.
1099                 //
1100                 //      | Header 1 | Header 2
1101                 //      | -------- | --------
1102                 //      | Cell 1   | Cell 2
1103                 //      | Cell 3   | Cell 4
1104                 $text = preg_replace_callback('
1105                         {
1106                                 ^                                                       # Start of a line
1107                                 [ ]{0,' . $less_than_tab . '}   # Allowed whitespace.
1108                                 [|]                                                     # Optional leading pipe (present)
1109                                 (.+) \n                                         # $1: Header row (at least one pipe)
1110
1111                                 [ ]{0,' . $less_than_tab . '}   # Allowed whitespace.
1112                                 [|] ([ ]*[-:]+[-| :]*) \n       # $2: Header underline
1113
1114                                 (                                                       # $3: Cells
1115                                         (?>
1116                                                 [ ]*                            # Allowed whitespace.
1117                                                 [|] .* \n                       # Row content.
1118                                         )*
1119                                 )
1120                                 (?=\n|\Z)                                       # Stop at final double newline.
1121                         }xm',
1122                         array($this, '_doTable_leadingPipe_callback'), $text);
1123
1124                 // Find tables without leading pipe.
1125                 //
1126                 //      Header 1 | Header 2
1127                 //      -------- | --------
1128                 //      Cell 1   | Cell 2
1129                 //      Cell 3   | Cell 4
1130                 $text = preg_replace_callback('
1131                         {
1132                                 ^                                                       # Start of a line
1133                                 [ ]{0,' . $less_than_tab . '}   # Allowed whitespace.
1134                                 (\S.*[|].*) \n                          # $1: Header row (at least one pipe)
1135
1136                                 [ ]{0,' . $less_than_tab . '}   # Allowed whitespace.
1137                                 ([-:]+[ ]*[|][-| :]*) \n        # $2: Header underline
1138
1139                                 (                                                       # $3: Cells
1140                                         (?>
1141                                                 .* [|] .* \n            # Row content
1142                                         )*
1143                                 )
1144                                 (?=\n|\Z)                                       # Stop at final double newline.
1145                         }xm',
1146                         array($this, '_DoTable_callback'), $text);
1147
1148                 return $text;
1149         }
1150
1151         /**
1152          * Callback for removing the leading pipe for each row
1153          * @param  array $matches
1154          * @return string
1155          */
1156         protected function _doTable_leadingPipe_callback($matches) {
1157                 $head           = $matches[1];
1158                 $underline      = $matches[2];
1159                 $content        = $matches[3];
1160
1161                 $content        = preg_replace('/^ *[|]/m', '', $content);
1162
1163                 return $this->_doTable_callback(array($matches[0], $head, $underline, $content));
1164         }
1165
1166         /**
1167          * Make the align attribute in a table
1168          * @param  string $alignname
1169          * @return string
1170          */
1171         protected function _doTable_makeAlignAttr($alignname)
1172         {
1173                 if (empty($this->table_align_class_tmpl)) {
1174                         return " align=\"$alignname\"";
1175                 }
1176
1177                 $classname = str_replace('%%', $alignname, $this->table_align_class_tmpl);
1178                 return " class=\"$classname\"";
1179         }
1180
1181         /**
1182          * Calback for processing tables
1183          * @param  array $matches
1184          * @return string
1185          */
1186         protected function _doTable_callback($matches) {
1187                 $head           = $matches[1];
1188                 $underline      = $matches[2];
1189                 $content        = $matches[3];
1190
1191                 // Remove any tailing pipes for each line.
1192                 $head           = preg_replace('/[|] *$/m', '', $head);
1193                 $underline      = preg_replace('/[|] *$/m', '', $underline);
1194                 $content        = preg_replace('/[|] *$/m', '', $content);
1195
1196                 // Reading alignement from header underline.
1197                 $separators     = preg_split('/ *[|] */', $underline);
1198                 foreach ($separators as $n => $s) {
1199                         if (preg_match('/^ *-+: *$/', $s))
1200                                 $attr[$n] = $this->_doTable_makeAlignAttr('right');
1201                         else if (preg_match('/^ *:-+: *$/', $s))
1202                                 $attr[$n] = $this->_doTable_makeAlignAttr('center');
1203                         else if (preg_match('/^ *:-+ *$/', $s))
1204                                 $attr[$n] = $this->_doTable_makeAlignAttr('left');
1205                         else
1206                                 $attr[$n] = '';
1207                 }
1208
1209                 // Parsing span elements, including code spans, character escapes,
1210                 // and inline HTML tags, so that pipes inside those gets ignored.
1211                 $head           = $this->parseSpan($head);
1212                 $headers        = preg_split('/ *[|] */', $head);
1213                 $col_count      = count($headers);
1214                 $attr       = array_pad($attr, $col_count, '');
1215
1216                 // Write column headers.
1217                 $text = "<table>\n";
1218                 $text .= "<thead>\n";
1219                 $text .= "<tr>\n";
1220                 foreach ($headers as $n => $header)
1221                         $text .= "  <th$attr[$n]>" . $this->runSpanGamut(trim($header)) . "</th>\n";
1222                 $text .= "</tr>\n";
1223                 $text .= "</thead>\n";
1224
1225                 // Split content by row.
1226                 $rows = explode("\n", trim($content, "\n"));
1227
1228                 $text .= "<tbody>\n";
1229                 foreach ($rows as $row) {
1230                         // Parsing span elements, including code spans, character escapes,
1231                         // and inline HTML tags, so that pipes inside those gets ignored.
1232                         $row = $this->parseSpan($row);
1233
1234                         // Split row by cell.
1235                         $row_cells = preg_split('/ *[|] */', $row, $col_count);
1236                         $row_cells = array_pad($row_cells, $col_count, '');
1237
1238                         $text .= "<tr>\n";
1239                         foreach ($row_cells as $n => $cell)
1240                                 $text .= "  <td$attr[$n]>" . $this->runSpanGamut(trim($cell)) . "</td>\n";
1241                         $text .= "</tr>\n";
1242                 }
1243                 $text .= "</tbody>\n";
1244                 $text .= "</table>";
1245
1246                 return $this->hashBlock($text) . "\n";
1247         }
1248
1249         /**
1250          * Form HTML definition lists.
1251          * @param  string $text
1252          * @return string
1253          */
1254         protected function doDefLists($text) {
1255                 $less_than_tab = $this->tab_width - 1;
1256
1257                 // Re-usable pattern to match any entire dl list:
1258                 $whole_list_re = '(?>
1259                         (                                                               # $1 = whole list
1260                           (                                                             # $2
1261                                 [ ]{0,' . $less_than_tab . '}
1262                                 ((?>.*\S.*\n)+)                         # $3 = defined term
1263                                 \n?
1264                                 [ ]{0,' . $less_than_tab . '}:[ ]+ # colon starting definition
1265                           )
1266                           (?s:.+?)
1267                           (                                                             # $4
1268                                   \z
1269                                 |
1270                                   \n{2,}
1271                                   (?=\S)
1272                                   (?!                                           # Negative lookahead for another term
1273                                         [ ]{0,' . $less_than_tab . '}
1274                                         (?: \S.*\n )+?                  # defined term
1275                                         \n?
1276                                         [ ]{0,' . $less_than_tab . '}:[ ]+ # colon starting definition
1277                                   )
1278                                   (?!                                           # Negative lookahead for another definition
1279                                         [ ]{0,' . $less_than_tab . '}:[ ]+ # colon starting definition
1280                                   )
1281                           )
1282                         )
1283                 )'; // mx
1284
1285                 $text = preg_replace_callback('{
1286                                 (?>\A\n?|(?<=\n\n))
1287                                 ' . $whole_list_re . '
1288                         }mx',
1289                         array($this, '_doDefLists_callback'), $text);
1290
1291                 return $text;
1292         }
1293
1294         /**
1295          * Callback for processing definition lists
1296          * @param  array $matches
1297          * @return string
1298          */
1299         protected function _doDefLists_callback($matches) {
1300                 // Re-usable patterns to match list item bullets and number markers:
1301                 $list = $matches[1];
1302
1303                 // Turn double returns into triple returns, so that we can make a
1304                 // paragraph for the last item in a list, if necessary:
1305                 $result = trim($this->processDefListItems($list));
1306                 $result = "<dl>\n" . $result . "\n</dl>";
1307                 return $this->hashBlock($result) . "\n\n";
1308         }
1309
1310         /**
1311          * Process the contents of a single definition list, splitting it
1312          * into individual term and definition list items.
1313          * @param  string $list_str
1314          * @return string
1315          */
1316         protected function processDefListItems($list_str) {
1317
1318                 $less_than_tab = $this->tab_width - 1;
1319
1320                 // Trim trailing blank lines:
1321                 $list_str = preg_replace("/\n{2,}\\z/", "\n", $list_str);
1322
1323                 // Process definition terms.
1324                 $list_str = preg_replace_callback('{
1325                         (?>\A\n?|\n\n+)                                         # leading line
1326                         (                                                                       # definition terms = $1
1327                                 [ ]{0,' . $less_than_tab . '}   # leading whitespace
1328                                 (?!\:[ ]|[ ])                                   # negative lookahead for a definition
1329                                                                                                 #   mark (colon) or more whitespace.
1330                                 (?> \S.* \n)+?                                  # actual term (not whitespace).
1331                         )
1332                         (?=\n?[ ]{0,3}:[ ])                                     # lookahead for following line feed
1333                                                                                                 #   with a definition mark.
1334                         }xm',
1335                         array($this, '_processDefListItems_callback_dt'), $list_str);
1336
1337                 // Process actual definitions.
1338                 $list_str = preg_replace_callback('{
1339                         \n(\n+)?                                                        # leading line = $1
1340                         (                                                                       # marker space = $2
1341                                 [ ]{0,' . $less_than_tab . '}   # whitespace before colon
1342                                 \:[ ]+                                                  # definition mark (colon)
1343                         )
1344                         ((?s:.+?))                                                      # definition text = $3
1345                         (?= \n+                                                         # stop at next definition mark,
1346                                 (?:                                                             # next term or end of text
1347                                         [ ]{0,' . $less_than_tab . '} \:[ ]     |
1348                                         <dt> | \z
1349                                 )
1350                         )
1351                         }xm',
1352                         array($this, '_processDefListItems_callback_dd'), $list_str);
1353
1354                 return $list_str;
1355         }
1356
1357         /**
1358          * Callback for <dt> elements in definition lists
1359          * @param  array $matches
1360          * @return string
1361          */
1362         protected function _processDefListItems_callback_dt($matches) {
1363                 $terms = explode("\n", trim($matches[1]));
1364                 $text = '';
1365                 foreach ($terms as $term) {
1366                         $term = $this->runSpanGamut(trim($term));
1367                         $text .= "\n<dt>" . $term . "</dt>";
1368                 }
1369                 return $text . "\n";
1370         }
1371
1372         /**
1373          * Callback for <dd> elements in definition lists
1374          * @param  array $matches
1375          * @return string
1376          */
1377         protected function _processDefListItems_callback_dd($matches) {
1378                 $leading_line   = $matches[1];
1379                 $marker_space   = $matches[2];
1380                 $def                    = $matches[3];
1381
1382                 if ($leading_line || preg_match('/\n{2,}/', $def)) {
1383                         // Replace marker with the appropriate whitespace indentation
1384                         $def = str_repeat(' ', strlen($marker_space)) . $def;
1385                         $def = $this->runBlockGamut($this->outdent($def . "\n\n"));
1386                         $def = "\n". $def ."\n";
1387                 }
1388                 else {
1389                         $def = rtrim($def);
1390                         $def = $this->runSpanGamut($this->outdent($def));
1391                 }
1392
1393                 return "\n<dd>" . $def . "</dd>\n";
1394         }
1395
1396         /**
1397          * Adding the fenced code block syntax to regular Markdown:
1398          *
1399          * ~~~
1400          * Code block
1401          * ~~~
1402          *
1403          * @param  string $text
1404          * @return string
1405          */
1406         protected function doFencedCodeBlocks($text) {
1407
1408                 $less_than_tab = $this->tab_width;
1409
1410                 $text = preg_replace_callback('{
1411                                 (?:\n|\A)
1412                                 # 1: Opening marker
1413                                 (
1414                                         (?:~{3,}|`{3,}) # 3 or more tildes/backticks.
1415                                 )
1416                                 [ ]*
1417                                 (?:
1418                                         \.?([-_:a-zA-Z0-9]+) # 2: standalone class name
1419                                 )?
1420                                 [ ]*
1421                                 (?:
1422                                         ' . $this->id_class_attr_catch_re . ' # 3: Extra attributes
1423                                 )?
1424                                 [ ]* \n # Whitespace and newline following marker.
1425
1426                                 # 4: Content
1427                                 (
1428                                         (?>
1429                                                 (?!\1 [ ]* \n)  # Not a closing marker.
1430                                                 .*\n+
1431                                         )+
1432                                 )
1433
1434                                 # Closing marker.
1435                                 \1 [ ]* (?= \n )
1436                         }xm',
1437                         array($this, '_doFencedCodeBlocks_callback'), $text);
1438
1439                 return $text;
1440         }
1441
1442         /**
1443          * Callback to process fenced code blocks
1444          * @param  array $matches
1445          * @return string
1446          */
1447         protected function _doFencedCodeBlocks_callback($matches) {
1448                 $classname =& $matches[2];
1449                 $attrs     =& $matches[3];
1450                 $codeblock = $matches[4];
1451
1452                 if ($this->code_block_content_func) {
1453                         $codeblock = call_user_func($this->code_block_content_func, $codeblock, $classname);
1454                 } else {
1455                         $codeblock = htmlspecialchars($codeblock, ENT_NOQUOTES);
1456                 }
1457
1458                 $codeblock = preg_replace_callback('/^\n+/',
1459                         array($this, '_doFencedCodeBlocks_newlines'), $codeblock);
1460
1461                 $classes = array();
1462                 if ($classname != "") {
1463                         if ($classname{0} == '.')
1464                                 $classname = substr($classname, 1);
1465                         $classes[] = $this->code_class_prefix . $classname;
1466                 }
1467                 $attr_str = $this->doExtraAttributes($this->code_attr_on_pre ? "pre" : "code", $attrs, null, $classes);
1468                 $pre_attr_str  = $this->code_attr_on_pre ? $attr_str : '';
1469                 $code_attr_str = $this->code_attr_on_pre ? '' : $attr_str;
1470                 $codeblock  = "<pre$pre_attr_str><code$code_attr_str>$codeblock</code></pre>";
1471
1472                 return "\n\n".$this->hashBlock($codeblock)."\n\n";
1473         }
1474
1475         /**
1476          * Replace new lines in fenced code blocks
1477          * @param  array $matches
1478          * @return string
1479          */
1480         protected function _doFencedCodeBlocks_newlines($matches) {
1481                 return str_repeat("<br$this->empty_element_suffix",
1482                         strlen($matches[0]));
1483         }
1484
1485         /**
1486          * Redefining emphasis markers so that emphasis by underscore does not
1487          * work in the middle of a word.
1488          * @var array
1489          */
1490         protected $em_relist = array(
1491                 ''  => '(?:(?<!\*)\*(?!\*)|(?<![a-zA-Z0-9_])_(?!_))(?![\.,:;]?\s)',
1492                 '*' => '(?<![\s*])\*(?!\*)',
1493                 '_' => '(?<![\s_])_(?![a-zA-Z0-9_])',
1494         );
1495         protected $strong_relist = array(
1496                 ''   => '(?:(?<!\*)\*\*(?!\*)|(?<![a-zA-Z0-9_])__(?!_))(?![\.,:;]?\s)',
1497                 '**' => '(?<![\s*])\*\*(?!\*)',
1498                 '__' => '(?<![\s_])__(?![a-zA-Z0-9_])',
1499         );
1500         protected $em_strong_relist = array(
1501                 ''    => '(?:(?<!\*)\*\*\*(?!\*)|(?<![a-zA-Z0-9_])___(?!_))(?![\.,:;]?\s)',
1502                 '***' => '(?<![\s*])\*\*\*(?!\*)',
1503                 '___' => '(?<![\s_])___(?![a-zA-Z0-9_])',
1504         );
1505
1506         /**
1507          * Parse text into paragraphs
1508          * @param  string $text String to process in paragraphs
1509          * @param  boolean $wrap_in_p Whether paragraphs should be wrapped in <p> tags
1510          * @return string       HTML output
1511          */
1512         protected function formParagraphs($text, $wrap_in_p = true) {
1513                 // Strip leading and trailing lines:
1514                 $text = preg_replace('/\A\n+|\n+\z/', '', $text);
1515
1516                 $grafs = preg_split('/\n{2,}/', $text, -1, PREG_SPLIT_NO_EMPTY);
1517
1518                 // Wrap <p> tags and unhashify HTML blocks
1519                 foreach ($grafs as $key => $value) {
1520                         $value = trim($this->runSpanGamut($value));
1521
1522                         // Check if this should be enclosed in a paragraph.
1523                         // Clean tag hashes & block tag hashes are left alone.
1524                         $is_p = $wrap_in_p && !preg_match('/^B\x1A[0-9]+B|^C\x1A[0-9]+C$/', $value);
1525
1526                         if ($is_p) {
1527                                 $value = "<p>$value</p>";
1528                         }
1529                         $grafs[$key] = $value;
1530                 }
1531
1532                 // Join grafs in one text, then unhash HTML tags.
1533                 $text = implode("\n\n", $grafs);
1534
1535                 // Finish by removing any tag hashes still present in $text.
1536                 $text = $this->unhash($text);
1537
1538                 return $text;
1539         }
1540
1541
1542         /**
1543          * Footnotes - Strips link definitions from text, stores the URLs and
1544          * titles in hash references.
1545          * @param  string $text
1546          * @return string
1547          */
1548         protected function stripFootnotes($text) {
1549                 $less_than_tab = $this->tab_width - 1;
1550
1551                 // Link defs are in the form: [^id]: url "optional title"
1552                 $text = preg_replace_callback('{
1553                         ^[ ]{0,' . $less_than_tab . '}\[\^(.+?)\][ ]?:  # note_id = $1
1554                           [ ]*
1555                           \n?                                   # maybe *one* newline
1556                         (                                               # text = $2 (no blank lines allowed)
1557                                 (?:
1558                                         .+                              # actual text
1559                                 |
1560                                         \n                              # newlines but
1561                                         (?!\[.+?\][ ]?:\s)# negative lookahead for footnote or link definition marker.
1562                                         (?!\n+[ ]{0,3}\S)# ensure line is not blank and followed
1563                                                                         # by non-indented content
1564                                 )*
1565                         )
1566                         }xm',
1567                         array($this, '_stripFootnotes_callback'),
1568                         $text);
1569                 return $text;
1570         }
1571
1572         /**
1573          * Callback for stripping footnotes
1574          * @param  array $matches
1575          * @return string
1576          */
1577         protected function _stripFootnotes_callback($matches) {
1578                 $note_id = $this->fn_id_prefix . $matches[1];
1579                 $this->footnotes[$note_id] = $this->outdent($matches[2]);
1580                 return ''; // String that will replace the block
1581         }
1582
1583         /**
1584          * Replace footnote references in $text [^id] with a special text-token
1585          * which will be replaced by the actual footnote marker in appendFootnotes.
1586          * @param  string $text
1587          * @return string
1588          */
1589         protected function doFootnotes($text) {
1590                 if (!$this->in_anchor) {
1591                         $text = preg_replace('{\[\^(.+?)\]}', "F\x1Afn:\\1\x1A:", $text);
1592                 }
1593                 return $text;
1594         }
1595
1596         /**
1597          * Append footnote list to text
1598          * @param  string $text
1599          * @return string
1600          */
1601         protected function appendFootnotes($text) {
1602                 $text = preg_replace_callback('{F\x1Afn:(.*?)\x1A:}',
1603                         array($this, '_appendFootnotes_callback'), $text);
1604
1605                 if (!empty($this->footnotes_ordered)) {
1606                         $text .= "\n\n";
1607                         $text .= "<div class=\"footnotes\">\n";
1608                         $text .= "<hr" . $this->empty_element_suffix . "\n";
1609                         $text .= "<ol>\n\n";
1610
1611                         $attr = "";
1612                         if ($this->fn_backlink_class != "") {
1613                                 $class = $this->fn_backlink_class;
1614                                 $class = $this->encodeAttribute($class);
1615                                 $attr .= " class=\"$class\"";
1616                         }
1617                         if ($this->fn_backlink_title != "") {
1618                                 $title = $this->fn_backlink_title;
1619                                 $title = $this->encodeAttribute($title);
1620                                 $attr .= " title=\"$title\"";
1621                         }
1622                         $backlink_text = $this->fn_backlink_html;
1623                         $num = 0;
1624
1625                         while (!empty($this->footnotes_ordered)) {
1626                                 $footnote = reset($this->footnotes_ordered);
1627                                 $note_id = key($this->footnotes_ordered);
1628                                 unset($this->footnotes_ordered[$note_id]);
1629                                 $ref_count = $this->footnotes_ref_count[$note_id];
1630                                 unset($this->footnotes_ref_count[$note_id]);
1631                                 unset($this->footnotes[$note_id]);
1632
1633                                 $footnote .= "\n"; // Need to append newline before parsing.
1634                                 $footnote = $this->runBlockGamut("$footnote\n");
1635                                 $footnote = preg_replace_callback('{F\x1Afn:(.*?)\x1A:}',
1636                                         array($this, '_appendFootnotes_callback'), $footnote);
1637
1638                                 $attr = str_replace("%%", ++$num, $attr);
1639                                 $note_id = $this->encodeAttribute($note_id);
1640
1641                                 // Prepare backlink, multiple backlinks if multiple references
1642                                 $backlink = "<a href=\"#fnref:$note_id\"$attr>$backlink_text</a>";
1643                                 for ($ref_num = 2; $ref_num <= $ref_count; ++$ref_num) {
1644                                         $backlink .= " <a href=\"#fnref$ref_num:$note_id\"$attr>$backlink_text</a>";
1645                                 }
1646                                 // Add backlink to last paragraph; create new paragraph if needed.
1647                                 if (preg_match('{</p>$}', $footnote)) {
1648                                         $footnote = substr($footnote, 0, -4) . "&#160;$backlink</p>";
1649                                 } else {
1650                                         $footnote .= "\n\n<p>$backlink</p>";
1651                                 }
1652
1653                                 $text .= "<li id=\"fn:$note_id\">\n";
1654                                 $text .= $footnote . "\n";
1655                                 $text .= "</li>\n\n";
1656                         }
1657
1658                         $text .= "</ol>\n";
1659                         $text .= "</div>";
1660                 }
1661                 return $text;
1662         }
1663
1664         /**
1665          * Callback for appending footnotes
1666          * @param  array $matches
1667          * @return string
1668          */
1669         protected function _appendFootnotes_callback($matches) {
1670                 $node_id = $this->fn_id_prefix . $matches[1];
1671
1672                 // Create footnote marker only if it has a corresponding footnote *and*
1673                 // the footnote hasn't been used by another marker.
1674                 if (isset($this->footnotes[$node_id])) {
1675                         $num =& $this->footnotes_numbers[$node_id];
1676                         if (!isset($num)) {
1677                                 // Transfer footnote content to the ordered list and give it its
1678                                 // number
1679                                 $this->footnotes_ordered[$node_id] = $this->footnotes[$node_id];
1680                                 $this->footnotes_ref_count[$node_id] = 1;
1681                                 $num = $this->footnote_counter++;
1682                                 $ref_count_mark = '';
1683                         } else {
1684                                 $ref_count_mark = $this->footnotes_ref_count[$node_id] += 1;
1685                         }
1686
1687                         $attr = "";
1688                         if ($this->fn_link_class != "") {
1689                                 $class = $this->fn_link_class;
1690                                 $class = $this->encodeAttribute($class);
1691                                 $attr .= " class=\"$class\"";
1692                         }
1693                         if ($this->fn_link_title != "") {
1694                                 $title = $this->fn_link_title;
1695                                 $title = $this->encodeAttribute($title);
1696                                 $attr .= " title=\"$title\"";
1697                         }
1698
1699                         $attr = str_replace("%%", $num, $attr);
1700                         $node_id = $this->encodeAttribute($node_id);
1701
1702                         return
1703                                 "<sup id=\"fnref$ref_count_mark:$node_id\">".
1704                                 "<a href=\"#fn:$node_id\"$attr>$num</a>".
1705                                 "</sup>";
1706                 }
1707
1708                 return "[^" . $matches[1] . "]";
1709         }
1710
1711
1712         /**
1713          * Abbreviations - strips abbreviations from text, stores titles in hash
1714          * references.
1715          * @param  string $text
1716          * @return string
1717          */
1718         protected function stripAbbreviations($text) {
1719                 $less_than_tab = $this->tab_width - 1;
1720
1721                 // Link defs are in the form: [id]*: url "optional title"
1722                 $text = preg_replace_callback('{
1723                         ^[ ]{0,' . $less_than_tab . '}\*\[(.+?)\][ ]?:  # abbr_id = $1
1724                         (.*)                                    # text = $2 (no blank lines allowed)
1725                         }xm',
1726                         array($this, '_stripAbbreviations_callback'),
1727                         $text);
1728                 return $text;
1729         }
1730
1731         /**
1732          * Callback for stripping abbreviations
1733          * @param  array $matches
1734          * @return string
1735          */
1736         protected function _stripAbbreviations_callback($matches) {
1737                 $abbr_word = $matches[1];
1738                 $abbr_desc = $matches[2];
1739                 if ($this->abbr_word_re) {
1740                         $this->abbr_word_re .= '|';
1741                 }
1742                 $this->abbr_word_re .= preg_quote($abbr_word);
1743                 $this->abbr_desciptions[$abbr_word] = trim($abbr_desc);
1744                 return ''; // String that will replace the block
1745         }
1746
1747         /**
1748          * Find defined abbreviations in text and wrap them in <abbr> elements.
1749          * @param  string $text
1750          * @return string
1751          */
1752         protected function doAbbreviations($text) {
1753                 if ($this->abbr_word_re) {
1754                         // cannot use the /x modifier because abbr_word_re may
1755                         // contain significant spaces:
1756                         $text = preg_replace_callback('{' .
1757                                 '(?<![\w\x1A])' .
1758                                 '(?:' . $this->abbr_word_re . ')' .
1759                                 '(?![\w\x1A])' .
1760                                 '}',
1761                                 array($this, '_doAbbreviations_callback'), $text);
1762                 }
1763                 return $text;
1764         }
1765
1766         /**
1767          * Callback for processing abbreviations
1768          * @param  array $matches
1769          * @return string
1770          */
1771         protected function _doAbbreviations_callback($matches) {
1772                 $abbr = $matches[0];
1773                 if (isset($this->abbr_desciptions[$abbr])) {
1774                         $desc = $this->abbr_desciptions[$abbr];
1775                         if (empty($desc)) {
1776                                 return $this->hashPart("<abbr>$abbr</abbr>");
1777                         } else {
1778                                 $desc = $this->encodeAttribute($desc);
1779                                 return $this->hashPart("<abbr title=\"$desc\">$abbr</abbr>");
1780                         }
1781                 } else {
1782                         return $matches[0];
1783                 }
1784         }
1785 }