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, 'replaceCallback' ), $text );
  65                 // ...and <mw:toc> markers
  66                 $wrappedtext = preg_replace_callback( '/\<\\/?mw:toc\>/',
  67                         array( &$this, 'replaceCallback' ), $wrappedtext );
  68
  69                 // Modify inline Microdata <link> and <meta> elements so they say <html-link> and <html-meta> so
  70                 // we can trick Tidy into not stripping them out by including them in tidy's new-empty-tags config
  71                 $wrappedtext = preg_replace( '!<(link|meta)([^>]*?)(/{0,1}>)!', '<html-$1$2$3', $wrappedtext );
  72
  73                 // Wrap the whole thing in a doctype and body for Tidy.
  74                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"' .
  75                         ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>' .
  76                         '<head><title>test</title></head><body>' . $wrappedtext . '</body></html>';
  77
  78                 return $wrappedtext;
  79         }
  80
  81         /**
  82          * @param $m array
  83          *
  84          * @return string
  85          */
  86         function replaceCallback( $m ) {
  87                 $marker = "{$this->mUniqPrefix}-item-{$this->mMarkerIndex}" . Parser::MARKER_SUFFIX;
  88                 $this->mMarkerIndex++;
  89                 $this->mTokens->setPair( $marker, $m[0] );
  90                 return $marker;
  91         }
  92
  93         /**
  94          * @param $text string
  95          * @return string
  96          */
  97         public function postprocess( $text ) {
  98                 // Revert <html-{link,meta}> back to <{link,meta}>
  99                 $text = preg_replace( '!<html-(link|meta)([^>]*?)(/{0,1}>)!', '<$1$2$3', $text );
 100
 101                 // Restore the contents of placeholder tokens
 102                 $text = $this->mTokens->replace( $text );
 103
 104                 return $text;
 105         }
 106
 107 }
 108
 109 /**
 110  * Class to interact with HTML tidy
 111  *
 112  * Either the external tidy program or the in-process tidy extension
 113  * will be used depending on availability. Override the default
 114  * $wgTidyInternal setting to disable the internal if it's not working.
 115  *
 116  * @ingroup Parser
 117  */
 118 class MWTidy {
 119         /**
 120          * Interface with html tidy, used if $wgUseTidy = true.
 121          * If tidy isn't able to correct the markup, the original will be
 122          * returned in all its glory with a warning comment appended.
 123          *
 124          * @param string $text hideous HTML input
 125          * @return String: corrected HTML output
 126          */
 127         public static function tidy( $text ) {
 128                 global $wgTidyInternal;
 129
 130                 $wrapper = new MWTidyWrapper;
 131                 $wrappedtext = $wrapper->getWrapped( $text );
 132
 133                 $retVal = null;
 134                 if ( $wgTidyInternal ) {
 135                         $correctedtext = self::execInternalTidy( $wrappedtext, false, $retVal );
 136                 } else {
 137                         $correctedtext = self::execExternalTidy( $wrappedtext, false, $retVal );
 138                 }
 139
 140                 if ( $retVal < 0 ) {
 141                         wfDebug( "Possible tidy configuration error!\n" );
 142                         return $text . "\n<!-- Tidy was unable to run -->\n";
 143                 } elseif ( is_null( $correctedtext ) ) {
 144                         wfDebug( "Tidy error detected!\n" );
 145                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 146                 }
 147
 148                 $correctedtext = $wrapper->postprocess( $correctedtext ); // restore any hidden tokens
 149
 150                 return $correctedtext;
 151         }
 152
 153         /**
 154          * Check HTML for errors, used if $wgValidateAllHtml = true.
 155          *
 156          * @param $text String
 157          * @param &$errorStr String: return the error string
 158          * @return Boolean: whether the HTML is valid
 159          */
 160         public static function checkErrors( $text, &$errorStr = null ) {
 161                 global $wgTidyInternal;
 162
 163                 $retval = 0;
 164                 if ( $wgTidyInternal ) {
 165                         $errorStr = self::execInternalTidy( $text, true, $retval );
 166                 } else {
 167                         $errorStr = self::execExternalTidy( $text, true, $retval );
 168                 }
 169
 170                 return ( $retval < 0 && $errorStr == '' ) || $retval == 0;
 171         }
 172
 173         /**
 174          * Spawn an external HTML tidy process and get corrected markup back from it.
 175          * Also called in OutputHandler.php for full page validation
 176          *
 177          * @param string $text HTML to check
 178          * @param $stderr Boolean: Whether to read result from STDERR rather than STDOUT
 179          * @param &$retval int Exit code (-1 on internal error)
 180          * @return mixed String or null
 181          */
 182         private static function execExternalTidy( $text, $stderr = false, &$retval = null ) {
 183                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 184                 wfProfileIn( __METHOD__ );
 185
 186                 $cleansource = '';
 187                 $opts = ' -utf8';
 188
 189                 if ( $stderr ) {
 190                         $descriptorspec = array(
 191                                 0 => array( 'pipe', 'r' ),
 192                                 1 => array( 'file', wfGetNull(), 'a' ),
 193                                 2 => array( 'pipe', 'w' )
 194                         );
 195                 } else {
 196                         $descriptorspec = array(
 197                                 0 => array( 'pipe', 'r' ),
 198                                 1 => array( 'pipe', 'w' ),
 199                                 2 => array( 'file', wfGetNull(), 'a' )
 200                         );
 201                 }
 202
 203                 $readpipe = $stderr ? 2 : 1;
 204                 $pipes = array();
 205
 206                 $process = proc_open(
 207                         "$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes );
 208
 209                 if ( is_resource( $process ) ) {
 210                         // Theoretically, this style of communication could cause a deadlock
 211                         // here. If the stdout buffer fills up, then writes to stdin could
 212                         // block. This doesn't appear to happen with tidy, because tidy only
 213                         // writes to stdout after it's finished reading from stdin. Search
 214                         // for tidyParseStdin and tidySaveStdout in console/tidy.c
 215                         fwrite( $pipes[0], $text );
 216                         fclose( $pipes[0] );
 217                         while ( !feof( $pipes[$readpipe] ) ) {
 218                                 $cleansource .= fgets( $pipes[$readpipe], 1024 );
 219                         }
 220                         fclose( $pipes[$readpipe] );
 221                         $retval = proc_close( $process );
 222                 } else {
 223                         wfWarn( "Unable to start external tidy process" );
 224                         $retval = -1;
 225                 }
 226
 227                 if ( !$stderr && $cleansource == '' && $text != '' ) {
 228                         // Some kind of error happened, so we couldn't get the corrected text.
 229                         // Just give up; we'll use the source text and append a warning.
 230                         $cleansource = null;
 231                 }
 232
 233                 wfProfileOut( __METHOD__ );
 234                 return $cleansource;
 235         }
 236
 237         /**
 238          * Use the HTML tidy extension to use the tidy library in-process,
 239          * saving the overhead of spawning a new process.
 240          *
 241          * @param string $text HTML to check
 242          * @param $stderr Boolean: Whether to read result from error status instead of output
 243          * @param &$retval int Exit code (-1 on internal error)
 244          * @return mixed String or null
 245          */
 246         private static function execInternalTidy( $text, $stderr = false, &$retval = null ) {
 247                 global $wgTidyConf, $wgDebugTidy;
 248                 wfProfileIn( __METHOD__ );
 249
 250                 if ( !class_exists( 'tidy' ) ) {
 251                         wfWarn( "Unable to load internal tidy class." );
 252                         $retval = -1;
 253
 254                         wfProfileOut( __METHOD__ );
 255                         return null;
 256                 }
 257
 258                 $tidy = new tidy;
 259                 $tidy->parseString( $text, $wgTidyConf, 'utf8' );
 260
 261                 if ( $stderr ) {
 262                         $retval = $tidy->getStatus();
 263
 264                         wfProfileOut( __METHOD__ );
 265                         return $tidy->errorBuffer;
 266                 }
 267
 268                 $tidy->cleanRepair();
 269                 $retval = $tidy->getStatus();
 270                 if ( $retval == 2 ) {
 271                         // 2 is magic number for fatal error
 272                         // http://www.php.net/manual/en/function.tidy-get-status.php
 273                         $cleansource = null;
 274                 } else {
 275                         $cleansource = tidy_get_output( $tidy );
 276                         if ( $wgDebugTidy && $retval > 0 ) {
 277                                 $cleansource .= "<!--\nTidy reports:\n" .
 278                                         str_replace( '-->', '--&gt;', $tidy->errorBuffer ) .
 279                                         "\n-->";
 280                         }
 281                 }
 282
 283                 wfProfileOut( __METHOD__ );
 284                 return $cleansource;
 285         }
 286 }