includes/tidy/RaggettWrapper.php

   1 <?php
   2 namespace MediaWiki\Tidy;
   3
   4 use ReplacementArray;
   5 use ParserOutput;
   6 use Parser;
   7
   8 /**
   9  * Class used to hide mw:editsection tokens from Tidy so that it doesn't break them
  10  * or break on them. This is a bit of a hack for now, but hopefully in the future
  11  * we may create a real postprocessor or something that will replace this.
  12  * It's called wrapper because for now it basically takes over MWTidy::tidy's task
  13  * of wrapping the text in a xhtml block
  14  *
  15  * This re-uses some of the parser's UNIQ tricks, though some of it is private so it's
  16  * duplicated. Perhaps we should create an abstract marker hiding class.
  17  *
  18  * @ingroup Parser
  19  */
  20 class RaggettWrapper {
  21
  22         /**
  23          * @var ReplacementArray
  24          */
  25         protected $mTokens;
  26
  27         protected $mMarkerIndex;
  28
  29         public function __construct() {
  30                 $this->mTokens = null;
  31         }
  32
  33         /**
  34          * @param string $text
  35          * @return string
  36          */
  37         public function getWrapped( $text ) {
  38                 $this->mTokens = new ReplacementArray;
  39                 $this->mMarkerIndex = 0;
  40
  41                 // Replace <mw:editsection> elements with placeholders
  42                 $wrappedtext = preg_replace_callback( ParserOutput::EDITSECTION_REGEX,
  43                         array( &$this, 'replaceCallback' ), $text );
  44                 // ...and <mw:toc> markers
  45                 $wrappedtext = preg_replace_callback( '/\<\\/?mw:toc\>/',
  46                         array( &$this, 'replaceCallback' ), $wrappedtext );
  47                 // ... and <math> tags
  48                 $wrappedtext = preg_replace_callback( '/\<math(.*?)\<\\/math\>/s',
  49                         array( &$this, 'replaceCallback' ), $wrappedtext );
  50                 // Modify inline Microdata <link> and <meta> elements so they say <html-link> and <html-meta> so
  51                 // we can trick Tidy into not stripping them out by including them in tidy's new-empty-tags config
  52                 $wrappedtext = preg_replace( '!<(link|meta)([^>]*?)(/{0,1}>)!', '<html-$1$2$3', $wrappedtext );
  53
  54                 // Wrap the whole thing in a doctype and body for Tidy.
  55                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"' .
  56                         ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>' .
  57                         '<head><title>test</title></head><body>' . $wrappedtext . '</body></html>';
  58
  59                 return $wrappedtext;
  60         }
  61
  62         /**
  63          * @param array $m
  64          *
  65          * @return string
  66          */
  67         public function replaceCallback( $m ) {
  68                 $marker = Parser::MARKER_PREFIX . "-item-{$this->mMarkerIndex}" . Parser::MARKER_SUFFIX;
  69                 $this->mMarkerIndex++;
  70                 $this->mTokens->setPair( $marker, $m[0] );
  71                 return $marker;
  72         }
  73
  74         /**
  75          * @param string $text
  76          * @return string
  77          */
  78         public function postprocess( $text ) {
  79                 // Revert <html-{link,meta}> back to <{link,meta}>
  80                 $text = preg_replace( '!<html-(link|meta)([^>]*?)(/{0,1}>)!', '<$1$2$3', $text );
  81
  82                 // Restore the contents of placeholder tokens
  83                 $text = $this->mTokens->replace( $text );
  84
  85                 return $text;
  86         }
  87
  88 }
  89 ?>