includes/tidy/RaggettWrapper.php

   1 <?php
   2 namespace MediaWiki\Tidy;
   3
   4 use ParserOutput;
   5 use Parser;
   6
   7 /**
   8  * Class used to hide mw:editsection tokens from Tidy so that it doesn't break them
   9  * or break on them. This is a bit of a hack for now, but hopefully in the future
  10  * we may create a real postprocessor or something that will replace this.
  11  * It's called wrapper because for now it basically takes over MWTidy::tidy's task
  12  * of wrapping the text in a xhtml block
  13  *
  14  * This re-uses some of the parser's UNIQ tricks, though some of it is private so it's
  15  * duplicated. Perhaps we should create an abstract marker hiding class.
  16  *
  17  * @ingroup Parser
  18  * @deprecated since 1.32
  19  */
  20 class RaggettWrapper {
  21
  22         /**
  23          * @var array
  24          */
  25         protected $mTokens;
  26
  27         /**
  28          * @var int
  29          */
  30         protected $mMarkerIndex;
  31
  32         /**
  33          * @param string $text
  34          * @return string
  35          */
  36         public function getWrapped( $text ) {
  37                 $this->mTokens = [];
  38                 $this->mMarkerIndex = 0;
  39
  40                 // Replace <mw:editsection> elements with placeholders
  41                 $wrappedtext = preg_replace_callback( ParserOutput::EDITSECTION_REGEX,
  42                         [ $this, 'replaceCallback' ], $text );
  43                 // ...and <mw:toc> markers
  44                 $wrappedtext = preg_replace_callback( '/\<\\/?mw:toc\>/',
  45                         [ $this, 'replaceCallback' ], $wrappedtext );
  46                 // ... and <math> tags
  47                 $wrappedtext = preg_replace_callback( '/\<math(.*?)\<\\/math\>/s',
  48                         [ $this, 'replaceCallback' ], $wrappedtext );
  49                 // Modify inline Microdata <link> and <meta> elements so they say <html-link> and <html-meta> so
  50                 // we can trick Tidy into not stripping them out by including them in tidy's new-empty-tags config
  51                 $wrappedtext = preg_replace( '!<(link|meta)([^>]*?)(/{0,1}>)!', '<html-$1$2$3', $wrappedtext );
  52                 // Similar for inline <style> tags, but those aren't empty.
  53                 $wrappedtext = preg_replace_callback( '!<style([^>]*)>(.*?)</style>!s', function ( $m ) {
  54                         return '<html-style' . $m[1] . '>'
  55                                 . $this->replaceCallback( [ $m[2] ] )
  56                                 . '</html-style>';
  57                 }, $wrappedtext );
  58
  59                 // Preserve empty li elements (T49673) by abusing Tidy's datafld hack
  60                 // The whitespace class is as in TY_(InitMap)
  61                 $wrappedtext = preg_replace( "!<li>([ \r\n\t\f]*)</li>!",
  62                         '<li datafld="" class="mw-empty-elt">\1</li>', $wrappedtext );
  63
  64                 // Wrap the whole thing in a doctype and body for Tidy.
  65                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"' .
  66                         ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>' .
  67                         '<head><title>test</title></head><body>' . $wrappedtext . '</body></html>';
  68
  69                 return $wrappedtext;
  70         }
  71
  72         /**
  73          * @param array $m
  74          * @return string
  75          */
  76         private function replaceCallback( array $m ) {
  77                 $marker = Parser::MARKER_PREFIX . "-item-{$this->mMarkerIndex}" . Parser::MARKER_SUFFIX;
  78                 $this->mMarkerIndex++;
  79                 $this->mTokens[$marker] = $m[0];
  80                 return $marker;
  81         }
  82
  83         /**
  84          * @param string $text
  85          * @return string
  86          */
  87         public function postprocess( $text ) {
  88                 // Revert <html-{link,meta,style}> back to <{link,meta,style}>
  89                 $text = preg_replace( '!<html-(link|meta)([^>]*?)(/{0,1}>)!', '<$1$2$3', $text );
  90                 $text = preg_replace( '!<(/?)html-(style)([^>]*)>!', '<$1$2$3>', $text );
  91
  92                 // Remove datafld
  93                 $text = str_replace( '<li datafld=""', '<li', $text );
  94
  95                 // Restore the contents of placeholder tokens
  96                 $text = strtr( $text, $this->mTokens );
  97
  98                 return $text;
  99         }
 100
 101 }