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  */
  19 class RaggettWrapper {
  20
  21         /**
  22          * @var array
  23          */
  24         protected $mTokens;
  25
  26         /**
  27          * @var int
  28          */
  29         protected $mMarkerIndex;
  30
  31         /**
  32          * @param string $text
  33          * @return string
  34          */
  35         public function getWrapped( $text ) {
  36                 $this->mTokens = [];
  37                 $this->mMarkerIndex = 0;
  38
  39                 // Replace <mw:editsection> elements with placeholders
  40                 $wrappedtext = preg_replace_callback( ParserOutput::EDITSECTION_REGEX,
  41                         [ $this, 'replaceCallback' ], $text );
  42                 // ...and <mw:toc> markers
  43                 $wrappedtext = preg_replace_callback( '/\<\\/?mw:toc\>/',
  44                         [ $this, 'replaceCallback' ], $wrappedtext );
  45                 // ... and <math> tags
  46                 $wrappedtext = preg_replace_callback( '/\<math(.*?)\<\\/math\>/s',
  47                         [ $this, 'replaceCallback' ], $wrappedtext );
  48                 // Modify inline Microdata <link> and <meta> elements so they say <html-link> and <html-meta> so
  49                 // we can trick Tidy into not stripping them out by including them in tidy's new-empty-tags config
  50                 $wrappedtext = preg_replace( '!<(link|meta)([^>]*?)(/{0,1}>)!', '<html-$1$2$3', $wrappedtext );
  51                 // Similar for inline <style> tags, but those aren't empty.
  52                 $wrappedtext = preg_replace_callback( '!<style([^>]*)>(.*?)</style>!s', function ( $m ) {
  53                         return '<html-style' . $m[1] . '>'
  54                                 . $this->replaceCallback( [ $m[2] ] )
  55                                 . '</html-style>';
  56                 }, $wrappedtext );
  57
  58                 // Preserve empty li elements (T49673) by abusing Tidy's datafld hack
  59                 // The whitespace class is as in TY_(InitMap)
  60                 $wrappedtext = preg_replace( "!<li>([ \r\n\t\f]*)</li>!",
  61                         '<li datafld="" class="mw-empty-elt">\1</li>', $wrappedtext );
  62
  63                 // Wrap the whole thing in a doctype and body for Tidy.
  64                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"' .
  65                         ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>' .
  66                         '<head><title>test</title></head><body>' . $wrappedtext . '</body></html>';
  67
  68                 return $wrappedtext;
  69         }
  70
  71         /**
  72          * @param array $m
  73          * @return string
  74          */
  75         private function replaceCallback( array $m ) {
  76                 $marker = Parser::MARKER_PREFIX . "-item-{$this->mMarkerIndex}" . Parser::MARKER_SUFFIX;
  77                 $this->mMarkerIndex++;
  78                 $this->mTokens[$marker] = $m[0];
  79                 return $marker;
  80         }
  81
  82         /**
  83          * @param string $text
  84          * @return string
  85          */
  86         public function postprocess( $text ) {
  87                 // Revert <html-{link,meta,style}> back to <{link,meta,style}>
  88                 $text = preg_replace( '!<html-(link|meta)([^>]*?)(/{0,1}>)!', '<$1$2$3', $text );
  89                 $text = preg_replace( '!<(/?)html-(style)([^>]*)>!', '<$1$2$3>', $text );
  90
  91                 // Remove datafld
  92                 $text = str_replace( '<li datafld=""', '<li', $text );
  93
  94                 // Restore the contents of placeholder tokens
  95                 $text = strtr( $text, $this->mTokens );
  96
  97                 return $text;
  98         }
  99
 100 }