includes/parser/StripState.php

   1 <?php
   2 /**
   3  * Holder for stripped items when parsing wiki markup.
   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  * @todo document, briefly.
  26  * @ingroup Parser
  27  */
  28 class StripState {
  29         protected $prefix;
  30         protected $data;
  31         protected $regex;
  32
  33         protected $parser;
  34
  35         protected $circularRefGuard;
  36         protected $recursionLevel = 0;
  37         protected $highestRecursionLevel = 0;
  38         protected $expandSize = 0;
  39
  40         const UNSTRIP_RECURSION_LIMIT = 20;
  41         const UNSTRIP_SIZE_LIMIT = 5000000;
  42
  43         /**
  44          * @param Parser|null $parser
  45          */
  46         public function __construct( Parser $parser = null ) {
  47                 $this->data = [
  48                         'nowiki' => [],
  49                         'general' => []
  50                 ];
  51                 $this->regex = '/' . Parser::MARKER_PREFIX . "([^\x7f<>&'\"]+)" . Parser::MARKER_SUFFIX . '/';
  52                 $this->circularRefGuard = [];
  53                 $this->parser = $parser;
  54         }
  55
  56         /**
  57          * Add a nowiki strip item
  58          * @param string $marker
  59          * @param string $value
  60          */
  61         public function addNoWiki( $marker, $value ) {
  62                 $this->addItem( 'nowiki', $marker, $value );
  63         }
  64
  65         /**
  66          * @param string $marker
  67          * @param string $value
  68          */
  69         public function addGeneral( $marker, $value ) {
  70                 $this->addItem( 'general', $marker, $value );
  71         }
  72
  73         /**
  74          * @throws MWException
  75          * @param string $type
  76          * @param string $marker
  77          * @param string $value
  78          */
  79         protected function addItem( $type, $marker, $value ) {
  80                 if ( !preg_match( $this->regex, $marker, $m ) ) {
  81                         throw new MWException( "Invalid marker: $marker" );
  82                 }
  83
  84                 $this->data[$type][$m[1]] = $value;
  85         }
  86
  87         /**
  88          * @param string $text
  89          * @return mixed
  90          */
  91         public function unstripGeneral( $text ) {
  92                 return $this->unstripType( 'general', $text );
  93         }
  94
  95         /**
  96          * @param string $text
  97          * @return mixed
  98          */
  99         public function unstripNoWiki( $text ) {
 100                 return $this->unstripType( 'nowiki', $text );
 101         }
 102
 103         /**
 104          * @param string $text
 105          * @return mixed
 106          */
 107         public function unstripBoth( $text ) {
 108                 $text = $this->unstripType( 'general', $text );
 109                 $text = $this->unstripType( 'nowiki', $text );
 110                 return $text;
 111         }
 112
 113         /**
 114          * @param string $type
 115          * @param string $text
 116          * @return mixed
 117          */
 118         protected function unstripType( $type, $text ) {
 119                 // Shortcut
 120                 if ( !count( $this->data[$type] ) ) {
 121                         return $text;
 122                 }
 123
 124                 $callback = function ( $m ) use ( $type ) {
 125                         $marker = $m[1];
 126                         if ( isset( $this->data[$type][$marker] ) ) {
 127                                 if ( isset( $this->circularRefGuard[$marker] ) ) {
 128                                         return $this->getWarning( 'parser-unstrip-loop-warning' );
 129                                 }
 130
 131                                 if ( $this->recursionLevel > $this->highestRecursionLevel ) {
 132                                         $this->highestRecursionLevel = $this->recursionLevel;
 133                                 }
 134                                 if ( $this->recursionLevel >= self::UNSTRIP_RECURSION_LIMIT ) {
 135                                         return $this->getLimitationWarning( 'unstrip-depth',
 136                                                 self::UNSTRIP_RECURSION_LIMIT );
 137                                 }
 138
 139                                 $value = $this->data[$type][$marker];
 140                                 if ( $value instanceof Closure ) {
 141                                         $value = $value();
 142                                 }
 143
 144                                 $this->expandSize += strlen( $value );
 145                                 if ( $this->expandSize > self::UNSTRIP_SIZE_LIMIT ) {
 146                                         return $this->getLimitationWarning( 'unstrip-size',
 147                                                 self::UNSTRIP_SIZE_LIMIT );
 148                                 }
 149
 150                                 $this->circularRefGuard[$marker] = true;
 151                                 $this->recursionLevel++;
 152                                 $ret = $this->unstripType( $type, $value );
 153                                 $this->recursionLevel--;
 154                                 unset( $this->circularRefGuard[$marker] );
 155
 156                                 return $ret;
 157                         } else {
 158                                 return $m[0];
 159                         }
 160                 };
 161
 162                 $text = preg_replace_callback( $this->regex, $callback, $text );
 163                 return $text;
 164         }
 165
 166         /**
 167          * Get warning HTML and register a limitation warning with the parser
 168          *
 169          * @param string $type
 170          * @param int $max
 171          * @return string
 172          */
 173         private function getLimitationWarning( $type, $max = '' ) {
 174                 if ( $this->parser ) {
 175                         $this->parser->limitationWarn( $type, $max );
 176                 }
 177                 return $this->getWarning( "$type-warning", $max );
 178         }
 179
 180         /**
 181          * Get warning HTML
 182          *
 183          * @param string $message
 184          * @param int $max
 185          * @return string
 186          */
 187         private function getWarning( $message, $max = '' ) {
 188                 return '<span class="error">' .
 189                         wfMessage( $message )
 190                                 ->numParams( $max )->inContentLanguage()->text() .
 191                         '</span>';
 192         }
 193
 194         /**
 195          * Get an array of parameters to pass to ParserOutput::setLimitReportData()
 196          *
 197          * @unstable Should only be called by Parser
 198          * @return array
 199          */
 200         public function getLimitReport() {
 201                 return [
 202                         [ 'limitreport-unstrip-depth',
 203                                 [
 204                                         $this->highestRecursionLevel,
 205                                         self::UNSTRIP_RECURSION_LIMIT
 206                                 ],
 207                         ],
 208                         [ 'limitreport-unstrip-size',
 209                                 [
 210                                         $this->expandSize,
 211                                         self::UNSTRIP_SIZE_LIMIT
 212                                 ],
 213                         ]
 214                 ];
 215         }
 216
 217         /**
 218          * Get a StripState object which is sufficient to unstrip the given text.
 219          * It will contain the minimum subset of strip items necessary.
 220          *
 221          * @param string $text
 222          *
 223          * @return StripState
 224          */
 225         public function getSubState( $text ) {
 226                 $subState = new StripState;
 227                 $pos = 0;
 228                 while ( true ) {
 229                         $startPos = strpos( $text, Parser::MARKER_PREFIX, $pos );
 230                         $endPos = strpos( $text, Parser::MARKER_SUFFIX, $pos );
 231                         if ( $startPos === false || $endPos === false ) {
 232                                 break;
 233                         }
 234
 235                         $endPos += strlen( Parser::MARKER_SUFFIX );
 236                         $marker = substr( $text, $startPos, $endPos - $startPos );
 237                         if ( !preg_match( $this->regex, $marker, $m ) ) {
 238                                 continue;
 239                         }
 240
 241                         $key = $m[1];
 242                         if ( isset( $this->data['nowiki'][$key] ) ) {
 243                                 $subState->data['nowiki'][$key] = $this->data['nowiki'][$key];
 244                         } elseif ( isset( $this->data['general'][$key] ) ) {
 245                                 $subState->data['general'][$key] = $this->data['general'][$key];
 246                         }
 247                         $pos = $endPos;
 248                 }
 249                 return $subState;
 250         }
 251
 252         /**
 253          * Merge another StripState object into this one. The strip marker keys
 254          * will not be preserved. The strings in the $texts array will have their
 255          * strip markers rewritten, the resulting array of strings will be returned.
 256          *
 257          * @param StripState $otherState
 258          * @param array $texts
 259          * @return array
 260          */
 261         public function merge( $otherState, $texts ) {
 262                 $mergePrefix = wfRandomString( 16 );
 263
 264                 foreach ( $otherState->data as $type => $items ) {
 265                         foreach ( $items as $key => $value ) {
 266                                 $this->data[$type]["$mergePrefix-$key"] = $value;
 267                         }
 268                 }
 269
 270                 $callback = function ( $m ) use ( $mergePrefix ) {
 271                         $key = $m[1];
 272                         return Parser::MARKER_PREFIX . $mergePrefix . '-' . $key . Parser::MARKER_SUFFIX;
 273                 };
 274                 $texts = preg_replace_callback( $otherState->regex, $callback, $texts );
 275                 return $texts;
 276         }
 277
 278         /**
 279          * Remove any strip markers found in the given text.
 280          *
 281          * @param string $text
 282          * @return string
 283          */
 284         public function killMarkers( $text ) {
 285                 return preg_replace( $this->regex, '', $text );
 286         }
 287 }