includes/parser/Tidy.php

   1 <?php
   2 /**
   3  * HTML validation and correction
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Parser
  22  */
  23
  24 /**
  25  * Class used to hide mw:editsection tokens from Tidy so that it doesn't break them
  26  * or break on them. This is a bit of a hack for now, but hopefully in the future
  27  * we may create a real postprocessor or something that will replace this.
  28  * It's called wrapper because for now it basically takes over MWTidy::tidy's task
  29  * of wrapping the text in a xhtml block
  30  *
  31  * This re-uses some of the parser's UNIQ tricks, though some of it is private so it's
  32  * duplicated. Perhaps we should create an abstract marker hiding class.
  33  *
  34  * @ingroup Parser
  35  */
  36 class MWTidyWrapper {
  37
  38         /**
  39          * @var ReplacementArray
  40          */
  41         protected $mTokens;
  42
  43         protected $mUniqPrefix;
  44
  45         protected $mMarkerIndex;
  46
  47         public function __construct() {
  48                 $this->mTokens = null;
  49                 $this->mUniqPrefix = null;
  50         }
  51
  52         /**
  53          * @param $text string
  54          * @return string
  55          */
  56         public function getWrapped( $text ) {
  57                 $this->mTokens = new ReplacementArray;
  58                 $this->mUniqPrefix = "\x7fUNIQ" .
  59                         dechex( mt_rand( 0, 0x7fffffff ) ) . dechex( mt_rand( 0, 0x7fffffff ) );
  60                 $this->mMarkerIndex = 0;
  61
  62                 // Replace <mw:editsection> elements with placeholders
  63                 $wrappedtext = preg_replace_callback( ParserOutput::EDITSECTION_REGEX,
  64                         array( &$this, 'replaceEditSectionLinksCallback' ), $text );
  65
  66                 // Modify inline Microdata <link> and <meta> elements so they say <html-link> and <html-meta> so
  67                 // we can trick Tidy into not stripping them out by including them in tidy's new-empty-tags config
  68                 $wrappedtext = preg_replace( '!<(link|meta)([^>]*?)(/{0,1}>)!', '<html-$1$2$3', $wrappedtext );
  69
  70                 // Wrap the whole thing in a doctype and body for Tidy.
  71                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
  72                         ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
  73                         '<head><title>test</title></head><body>'.$wrappedtext.'</body></html>';
  74
  75                 return $wrappedtext;
  76         }
  77
  78         /**
  79          * @param $m array
  80          *
  81          * @return string
  82          */
  83         function replaceEditSectionLinksCallback( $m ) {
  84                 $marker = "{$this->mUniqPrefix}-item-{$this->mMarkerIndex}" . Parser::MARKER_SUFFIX;
  85                 $this->mMarkerIndex++;
  86                 $this->mTokens->setPair( $marker, $m[0] );
  87                 return $marker;
  88         }
  89
  90         /**
  91          * @param $text string
  92          * @return string
  93          */
  94         public function postprocess( $text ) {
  95                 // Revert <html-{link,meta}> back to <{link,meta}>
  96                 $text = preg_replace( '!<html-(link|meta)([^>]*?)(/{0,1}>)!', '<$1$2$3', $text );
  97
  98                 // Restore the contents of placeholder tokens
  99                 $text = $this->mTokens->replace( $text );
 100
 101                 return $text;
 102         }
 103
 104 }
 105
 106 /**
 107  * Class to interact with HTML tidy
 108  *
 109  * Either the external tidy program or the in-process tidy extension
 110  * will be used depending on availability. Override the default
 111  * $wgTidyInternal setting to disable the internal if it's not working.
 112  *
 113  * @ingroup Parser
 114  */
 115 class MWTidy {
 116         /**
 117          * Interface with html tidy, used if $wgUseTidy = true.
 118          * If tidy isn't able to correct the markup, the original will be
 119          * returned in all its glory with a warning comment appended.
 120          *
 121          * @param $text String: hideous HTML input
 122          * @return String: corrected HTML output
 123          */
 124         public static function tidy( $text ) {
 125                 global $wgTidyInternal;
 126
 127                 $wrapper = new MWTidyWrapper;
 128                 $wrappedtext = $wrapper->getWrapped( $text );
 129
 130                 $retVal = null;
 131                 if ( $wgTidyInternal ) {
 132                         $correctedtext = self::execInternalTidy( $wrappedtext, false, $retVal );
 133                 } else {
 134                         $correctedtext = self::execExternalTidy( $wrappedtext, false, $retVal );
 135                 }
 136
 137                 if ( $retVal < 0 ) {
 138                         wfDebug( "Possible tidy configuration error!\n" );
 139                         return $text . "\n<!-- Tidy was unable to run -->\n";
 140                 } elseif ( is_null( $correctedtext ) ) {
 141                         wfDebug( "Tidy error detected!\n" );
 142                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 143                 }
 144
 145                 $correctedtext = $wrapper->postprocess( $correctedtext ); // restore any hidden tokens
 146
 147                 return $correctedtext;
 148         }
 149
 150         /**
 151          * Check HTML for errors, used if $wgValidateAllHtml = true.
 152          *
 153          * @param $text String
 154          * @param &$errorStr String: return the error string
 155          * @return Boolean: whether the HTML is valid
 156          */
 157         public static function checkErrors( $text, &$errorStr = null ) {
 158                 global $wgTidyInternal;
 159
 160                 $retval = 0;
 161                 if( $wgTidyInternal ) {
 162                         $errorStr = self::execInternalTidy( $text, true, $retval );
 163                 } else {
 164                         $errorStr = self::execExternalTidy( $text, true, $retval );
 165                 }
 166
 167                 return ( $retval < 0 && $errorStr == '' ) || $retval == 0;
 168         }
 169
 170         /**
 171          * Spawn an external HTML tidy process and get corrected markup back from it.
 172          * Also called in OutputHandler.php for full page validation
 173          *
 174          * @param $text String: HTML to check
 175          * @param $stderr Boolean: Whether to read result from STDERR rather than STDOUT
 176          * @param &$retval int Exit code (-1 on internal error)
 177          * @return mixed String or null
 178          */
 179         private static function execExternalTidy( $text, $stderr = false, &$retval = null ) {
 180                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 181                 wfProfileIn( __METHOD__ );
 182
 183                 $cleansource = '';
 184                 $opts = ' -utf8';
 185
 186                 if ( $stderr ) {
 187                         $descriptorspec = array(
 188                                 0 => array( 'pipe', 'r' ),
 189                                 1 => array( 'file', wfGetNull(), 'a' ),
 190                                 2 => array( 'pipe', 'w' )
 191                         );
 192                 } else {
 193                         $descriptorspec = array(
 194                                 0 => array( 'pipe', 'r' ),
 195                                 1 => array( 'pipe', 'w' ),
 196                                 2 => array( 'file', wfGetNull(), 'a' )
 197                         );
 198                 }
 199
 200                 $readpipe = $stderr ? 2 : 1;
 201                 $pipes = array();
 202
 203                 $process = proc_open(
 204                         "$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes );
 205
 206                 if ( is_resource( $process ) ) {
 207                         // Theoretically, this style of communication could cause a deadlock
 208                         // here. If the stdout buffer fills up, then writes to stdin could
 209                         // block. This doesn't appear to happen with tidy, because tidy only
 210                         // writes to stdout after it's finished reading from stdin. Search
 211                         // for tidyParseStdin and tidySaveStdout in console/tidy.c
 212                         fwrite( $pipes[0], $text );
 213                         fclose( $pipes[0] );
 214                         while ( !feof( $pipes[$readpipe] ) ) {
 215                                 $cleansource .= fgets( $pipes[$readpipe], 1024 );
 216                         }
 217                         fclose( $pipes[$readpipe] );
 218                         $retval = proc_close( $process );
 219                 } else {
 220                         wfWarn( "Unable to start external tidy process" );
 221                         $retval = -1;
 222                 }
 223
 224                 if ( !$stderr && $cleansource == '' && $text != '' ) {
 225                         // Some kind of error happened, so we couldn't get the corrected text.
 226                         // Just give up; we'll use the source text and append a warning.
 227                         $cleansource = null;
 228                 }
 229
 230                 wfProfileOut( __METHOD__ );
 231                 return $cleansource;
 232         }
 233
 234         /**
 235          * Use the HTML tidy extension to use the tidy library in-process,
 236          * saving the overhead of spawning a new process.
 237          *
 238          * @param $text String: HTML to check
 239          * @param $stderr Boolean: Whether to read result from error status instead of output
 240          * @param &$retval int Exit code (-1 on internal error)
 241          * @return mixed String or null
 242          */
 243         private static function execInternalTidy( $text, $stderr = false, &$retval = null ) {
 244                 global $wgTidyConf, $wgDebugTidy;
 245                 wfProfileIn( __METHOD__ );
 246
 247                 if ( !MWInit::classExists( 'tidy' ) ) {
 248                         wfWarn( "Unable to load internal tidy class." );
 249                         $retval = -1;
 250
 251                         wfProfileOut( __METHOD__ );
 252                         return null;
 253                 }
 254
 255                 $tidy = new tidy;
 256                 $tidy->parseString( $text, $wgTidyConf, 'utf8' );
 257
 258                 if ( $stderr ) {
 259                         $retval = $tidy->getStatus();
 260
 261                         wfProfileOut( __METHOD__ );
 262                         return $tidy->errorBuffer;
 263                 } else {
 264                         $tidy->cleanRepair();
 265                         $retval = $tidy->getStatus();
 266                         if ( $retval == 2 ) {
 267                                 // 2 is magic number for fatal error
 268                                 // http://www.php.net/manual/en/function.tidy-get-status.php
 269                                 $cleansource = null;
 270                         } else {
 271                                 $cleansource = tidy_get_output( $tidy );
 272                                 if ( $wgDebugTidy && $retval > 0 ) {
 273                                         $cleansource .= "<!--\nTidy reports:\n" .
 274                                                 str_replace( '-->', '--&gt;', $tidy->errorBuffer ) .
 275                                                 "\n-->";
 276                                 }
 277                         }
 278
 279                         wfProfileOut( __METHOD__ );
 280                         return $cleansource;
 281                 }
 282         }
 283 }