includes/parser/ParserOutput.php

   1 <?php
   2
   3 /**
   4  * Output of the PHP parser.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Parser
  23  */
  24 class ParserOutput extends CacheTime {
  25         var $mText,                       # The output text
  26                 $mLanguageLinks,              # List of the full text of language links, in the order they appear
  27                 $mCategories,                 # Map of category names to sort keys
  28                 $mTitleText,                  # title text of the chosen language variant
  29                 $mLinks = array(),            # 2-D map of NS/DBK to ID for the links in the document. ID=zero for broken.
  30                 $mTemplates = array(),        # 2-D map of NS/DBK to ID for the template references. ID=zero for broken.
  31                 $mTemplateIds = array(),      # 2-D map of NS/DBK to rev ID for the template references. ID=zero for broken.
  32                 $mImages = array(),           # DB keys of the images used, in the array key only
  33                 $mFileSearchOptions = array(), # DB keys of the images used mapped to sha1 and MW timestamp
  34                 $mExternalLinks = array(),    # External link URLs, in the key only
  35                 $mInterwikiLinks = array(),   # 2-D map of prefix/DBK (in keys only) for the inline interwiki links in the document.
  36                 $mNewSection = false,         # Show a new section link?
  37                 $mHideNewSection = false,     # Hide the new section link?
  38                 $mNoGallery = false,          # No gallery on category page? (__NOGALLERY__)
  39                 $mHeadItems = array(),        # Items to put in the <head> section
  40                 $mModules = array(),          # Modules to be loaded by the resource loader
  41                 $mModuleScripts = array(),    # Modules of which only the JS will be loaded by the resource loader
  42                 $mModuleStyles = array(),     # Modules of which only the CSSS will be loaded by the resource loader
  43                 $mModuleMessages = array(),   # Modules of which only the messages will be loaded by the resource loader
  44                 $mOutputHooks = array(),      # Hook tags as per $wgParserOutputHooks
  45                 $mWarnings = array(),         # Warning text to be returned to the user. Wikitext formatted, in the key only
  46                 $mSections = array(),         # Table of contents
  47                 $mEditSectionTokens = false,  # prefix/suffix markers if edit sections were output as tokens
  48                 $mProperties = array(),       # Name/value pairs to be cached in the DB
  49                 $mTOCHTML = '',               # HTML of the TOC
  50                 $mTimestamp;                  # Timestamp of the revision
  51                 private $mIndexPolicy = '';       # 'index' or 'noindex'?  Any other value will result in no change.
  52                 private $mAccessedOptions = array(); # List of ParserOptions (stored in the keys)
  53                 private $mSecondaryDataUpdates = array(); # List of DataUpdate, used to save info from the page somewhere else.
  54                 private $mExtensionData = array(); # extra data used by extensions
  55
  56         const EDITSECTION_REGEX = '#<(?:mw:)?editsection page="(.*?)" section="(.*?)"(?:/>|>(.*?)(</(?:mw:)?editsection>))#';
  57
  58         function __construct( $text = '', $languageLinks = array(), $categoryLinks = array(),
  59                 $containsOldMagic = false, $titletext = '' )
  60         {
  61                 $this->mText = $text;
  62                 $this->mLanguageLinks = $languageLinks;
  63                 $this->mCategories = $categoryLinks;
  64                 $this->mContainsOldMagic = $containsOldMagic;
  65                 $this->mTitleText = $titletext;
  66         }
  67
  68         function getText() {
  69                 if ( $this->mEditSectionTokens ) {
  70                         return preg_replace_callback( ParserOutput::EDITSECTION_REGEX,
  71                                 array( &$this, 'replaceEditSectionLinksCallback' ), $this->mText );
  72                 }
  73                 return preg_replace( ParserOutput::EDITSECTION_REGEX, '', $this->mText );
  74         }
  75
  76         /**
  77          * callback used by getText to replace editsection tokens
  78          * @private
  79          * @param $m
  80          * @throws MWException
  81          * @return mixed
  82          */
  83         function replaceEditSectionLinksCallback( $m ) {
  84                 global $wgOut, $wgLang;
  85                 $args = array(
  86                         htmlspecialchars_decode( $m[1] ),
  87                         htmlspecialchars_decode( $m[2] ),
  88                         isset( $m[4] ) ? $m[3] : null,
  89                 );
  90                 $args[0] = Title::newFromText( $args[0] );
  91                 if ( !is_object( $args[0] ) ) {
  92                         throw new MWException( "Bad parser output text." );
  93                 }
  94                 $args[] = $wgLang->getCode();
  95                 $skin = $wgOut->getSkin();
  96                 return call_user_func_array( array( $skin, 'doEditSectionLink' ), $args );
  97         }
  98
  99         function &getLanguageLinks()         { return $this->mLanguageLinks; }
 100         function getInterwikiLinks()         { return $this->mInterwikiLinks; }
 101         function getCategoryLinks()          { return array_keys( $this->mCategories ); }
 102         function &getCategories()            { return $this->mCategories; }
 103         function getTitleText()              { return $this->mTitleText; }
 104         function getSections()               { return $this->mSections; }
 105         function getEditSectionTokens()      { return $this->mEditSectionTokens; }
 106         function &getLinks()                 { return $this->mLinks; }
 107         function &getTemplates()             { return $this->mTemplates; }
 108         function &getTemplateIds()           { return $this->mTemplateIds; }
 109         function &getImages()                { return $this->mImages; }
 110         function &getFileSearchOptions()     { return $this->mFileSearchOptions; }
 111         function &getExternalLinks()         { return $this->mExternalLinks; }
 112         function getNoGallery()              { return $this->mNoGallery; }
 113         function getHeadItems()              { return $this->mHeadItems; }
 114         function getModules()                { return $this->mModules; }
 115         function getModuleScripts()          { return $this->mModuleScripts; }
 116         function getModuleStyles()           { return $this->mModuleStyles; }
 117         function getModuleMessages()         { return $this->mModuleMessages; }
 118         function getOutputHooks()            { return (array)$this->mOutputHooks; }
 119         function getWarnings()               { return array_keys( $this->mWarnings ); }
 120         function getIndexPolicy()            { return $this->mIndexPolicy; }
 121         function getTOCHTML()                { return $this->mTOCHTML; }
 122         function getTimestamp()              { return $this->mTimestamp; }
 123
 124         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
 125         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
 126         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategories, $cl ); }
 127
 128         function setTitleText( $t )          { return wfSetVar( $this->mTitleText, $t ); }
 129         function setSections( $toc )         { return wfSetVar( $this->mSections, $toc ); }
 130         function setEditSectionTokens( $t )  { return wfSetVar( $this->mEditSectionTokens, $t ); }
 131         function setIndexPolicy( $policy )   { return wfSetVar( $this->mIndexPolicy, $policy ); }
 132         function setTOCHTML( $tochtml )      { return wfSetVar( $this->mTOCHTML, $tochtml ); }
 133         function setTimestamp( $timestamp )  { return wfSetVar( $this->mTimestamp, $timestamp ); }
 134
 135         function addCategory( $c, $sort )    { $this->mCategories[$c] = $sort; }
 136         function addLanguageLink( $t )       { $this->mLanguageLinks[] = $t; }
 137         function addWarning( $s )            { $this->mWarnings[$s] = 1; }
 138
 139         function addOutputHook( $hook, $data = false ) {
 140                 $this->mOutputHooks[] = array( $hook, $data );
 141         }
 142
 143         function setNewSection( $value ) {
 144                 $this->mNewSection = (bool)$value;
 145         }
 146         function hideNewSection( $value ) {
 147                 $this->mHideNewSection = (bool)$value;
 148         }
 149         function getHideNewSection() {
 150                 return (bool)$this->mHideNewSection;
 151         }
 152         function getNewSection() {
 153                 return (bool)$this->mNewSection;
 154         }
 155
 156         /**
 157          * Checks, if a url is pointing to the own server
 158          *
 159          * @param string $internal the server to check against
 160          * @param string $url the url to check
 161          * @return bool
 162          */
 163         static function isLinkInternal( $internal, $url ) {
 164                 return (bool)preg_match( '/^' .
 165                         # If server is proto relative, check also for http/https links
 166                         ( substr( $internal, 0, 2 ) === '//' ? '(?:https?:)?' : '' ) .
 167                         preg_quote( $internal, '/' ) .
 168                         # check for query/path/anchor or end of link in each case
 169                         '(?:[\?\/\#]|$)/i',
 170                         $url
 171                 );
 172         }
 173
 174         function addExternalLink( $url ) {
 175                 # We don't register links pointing to our own server, unless... :-)
 176                 global $wgServer, $wgRegisterInternalExternals;
 177
 178                 $registerExternalLink = true;
 179                 if ( !$wgRegisterInternalExternals ) {
 180                         $registerExternalLink = !self::isLinkInternal( $wgServer, $url );
 181                 }
 182                 if ( $registerExternalLink ) {
 183                         $this->mExternalLinks[$url] = 1;
 184                 }
 185         }
 186
 187         /**
 188          * Record a local or interwiki inline link for saving in future link tables.
 189          *
 190          * @param $title Title object
 191          * @param $id Mixed: optional known page_id so we can skip the lookup
 192          */
 193         function addLink( Title $title, $id = null ) {
 194                 if ( $title->isExternal() ) {
 195                         // Don't record interwikis in pagelinks
 196                         $this->addInterwikiLink( $title );
 197                         return;
 198                 }
 199                 $ns = $title->getNamespace();
 200                 $dbk = $title->getDBkey();
 201                 if ( $ns == NS_MEDIA ) {
 202                         // Normalize this pseudo-alias if it makes it down here...
 203                         $ns = NS_FILE;
 204                 } elseif ( $ns == NS_SPECIAL ) {
 205                         // We don't record Special: links currently
 206                         // It might actually be wise to, but we'd need to do some normalization.
 207                         return;
 208                 } elseif ( $dbk === '' ) {
 209                         // Don't record self links -  [[#Foo]]
 210                         return;
 211                 }
 212                 if ( !isset( $this->mLinks[$ns] ) ) {
 213                         $this->mLinks[$ns] = array();
 214                 }
 215                 if ( is_null( $id ) ) {
 216                         $id = $title->getArticleID();
 217                 }
 218                 $this->mLinks[$ns][$dbk] = $id;
 219         }
 220
 221         /**
 222          * Register a file dependency for this output
 223          * @param string $name Title dbKey
 224          * @param string $timestamp MW timestamp of file creation (or false if non-existing)
 225          * @param string $sha1 base 36 SHA-1 of file (or false if non-existing)
 226          * @return void
 227          */
 228         function addImage( $name, $timestamp = null, $sha1 = null ) {
 229                 $this->mImages[$name] = 1;
 230                 if ( $timestamp !== null && $sha1 !== null ) {
 231                         $this->mFileSearchOptions[$name] = array( 'time' => $timestamp, 'sha1' => $sha1 );
 232                 }
 233         }
 234
 235         /**
 236          * Register a template dependency for this output
 237          * @param $title Title
 238          * @param $page_id
 239          * @param $rev_id
 240          * @return void
 241          */
 242         function addTemplate( $title, $page_id, $rev_id ) {
 243                 $ns = $title->getNamespace();
 244                 $dbk = $title->getDBkey();
 245                 if ( !isset( $this->mTemplates[$ns] ) ) {
 246                         $this->mTemplates[$ns] = array();
 247                 }
 248                 $this->mTemplates[$ns][$dbk] = $page_id;
 249                 if ( !isset( $this->mTemplateIds[$ns] ) ) {
 250                         $this->mTemplateIds[$ns] = array();
 251                 }
 252                 $this->mTemplateIds[$ns][$dbk] = $rev_id; // For versioning
 253         }
 254
 255         /**
 256          * @param $title Title object, must be an interwiki link
 257          * @throws MWException if given invalid input
 258          */
 259         function addInterwikiLink( $title ) {
 260                 $prefix = $title->getInterwiki();
 261                 if ( $prefix == '' ) {
 262                         throw new MWException( 'Non-interwiki link passed, internal parser error.' );
 263                 }
 264                 if ( !isset( $this->mInterwikiLinks[$prefix] ) ) {
 265                         $this->mInterwikiLinks[$prefix] = array();
 266                 }
 267                 $this->mInterwikiLinks[$prefix][$title->getDBkey()] = 1;
 268         }
 269
 270         /**
 271          * Add some text to the "<head>".
 272          * If $tag is set, the section with that tag will only be included once
 273          * in a given page.
 274          */
 275         function addHeadItem( $section, $tag = false ) {
 276                 if ( $tag !== false ) {
 277                         $this->mHeadItems[$tag] = $section;
 278                 } else {
 279                         $this->mHeadItems[] = $section;
 280                 }
 281         }
 282
 283         public function addModules( $modules ) {
 284                 $this->mModules = array_merge( $this->mModules, (array) $modules );
 285         }
 286
 287         public function addModuleScripts( $modules ) {
 288                 $this->mModuleScripts = array_merge( $this->mModuleScripts, (array)$modules );
 289         }
 290
 291         public function addModuleStyles( $modules ) {
 292                 $this->mModuleStyles = array_merge( $this->mModuleStyles, (array)$modules );
 293         }
 294
 295         public function addModuleMessages( $modules ) {
 296                 $this->mModuleMessages = array_merge( $this->mModuleMessages, (array)$modules );
 297         }
 298
 299         /**
 300          * Copy items from the OutputPage object into this one
 301          *
 302          * @param $out OutputPage object
 303          */
 304         public function addOutputPageMetadata( OutputPage $out ) {
 305                 $this->addModules( $out->getModules() );
 306                 $this->addModuleScripts( $out->getModuleScripts() );
 307                 $this->addModuleStyles( $out->getModuleStyles() );
 308                 $this->addModuleMessages( $out->getModuleMessages() );
 309
 310                 $this->mHeadItems = array_merge( $this->mHeadItems, $out->getHeadItemsArray() );
 311         }
 312
 313         /**
 314          * Override the title to be used for display
 315          * -- this is assumed to have been validated
 316          * (check equal normalisation, etc.)
 317          *
 318          * @param string $text desired title text
 319          */
 320         public function setDisplayTitle( $text ) {
 321                 $this->setTitleText( $text );
 322                 $this->setProperty( 'displaytitle', $text );
 323         }
 324
 325         /**
 326          * Get the title to be used for display
 327          *
 328          * @return String
 329          */
 330         public function getDisplayTitle() {
 331                 $t = $this->getTitleText();
 332                 if ( $t === '' ) {
 333                         return false;
 334                 }
 335                 return $t;
 336         }
 337
 338         /**
 339          * Fairly generic flag setter thingy.
 340          */
 341         public function setFlag( $flag ) {
 342                 $this->mFlags[$flag] = true;
 343         }
 344
 345         public function getFlag( $flag ) {
 346                 return isset( $this->mFlags[$flag] );
 347         }
 348
 349         /**
 350          * Set a property to be stored in the page_props database table.
 351          *
 352          * page_props is a key value store indexed by the page ID. This allows
 353          * the parser to set a property on a page which can then be quickly
 354          * retrieved given the page ID or via a DB join when given the page
 355          * title.
 356          *
 357          * setProperty() is thus used to propagate properties from the parsed
 358          * page to request contexts other than a page view of the currently parsed
 359          * article.
 360          *
 361          * Some applications examples:
 362          *
 363          *   * To implement hidden categories, hiding pages from category listings
 364          *     by storing a property.
 365          *
 366          *   * Overriding the displayed article title.
 367          *   @see ParserOutput::setDisplayTitle()
 368          *
 369          *   * To implement image tagging, for example displaying an icon on an
 370          *     image thumbnail to indicate that it is listed for deletion on
 371          *     Wikimedia Commons.
 372          *     This is not actually implemented, yet but would be pretty cool.
 373          *
 374          * @note: Do not use setProperty() to set a property which is only used
 375          * in a context where the ParserOutput object itself is already available,
 376          * for example a normal page view. There is no need to save such a property
 377          * in the database since it the text is already parsed. You can just hook
 378          * OutputPageParserOutput and get your data out of the ParserOutput object.
 379          *
 380          * If you are writing an extension where you want to set a property in the
 381          * parser which is used by an OutputPageParserOutput hook, you have to
 382          * associate the extension data directly with the ParserOutput object.
 383          * Since MediaWiki 1.21, you can use setExtensionData() to do this:
 384          *
 385          * @par Example:
 386          * @code
 387          *    $parser->getOutput()->setExtensionData( 'my_ext_foo', '...' );
 388          * @endcode
 389          *
 390          * And then later, in OutputPageParserOutput or similar:
 391          *
 392          * @par Example:
 393          * @code
 394          *    $output->getExtensionData( 'my_ext_foo' );
 395          * @endcode
 396          *
 397          * In MediaWiki 1.20 and older, you have to use a custom member variable
 398          * within the ParserOutput object:
 399          *
 400          * @par Example:
 401          * @code
 402          *    $parser->getOutput()->my_ext_foo = '...';
 403          * @endcode
 404          *
 405          */
 406         public function setProperty( $name, $value ) {
 407                 $this->mProperties[$name] = $value;
 408         }
 409
 410         public function getProperty( $name ) {
 411                 return isset( $this->mProperties[$name] ) ? $this->mProperties[$name] : false;
 412         }
 413
 414         public function getProperties() {
 415                 if ( !isset( $this->mProperties ) ) {
 416                         $this->mProperties = array();
 417                 }
 418                 return $this->mProperties;
 419         }
 420
 421         /**
 422          * Returns the options from its ParserOptions which have been taken
 423          * into account to produce this output or false if not available.
 424          * @return mixed Array
 425          */
 426         public function getUsedOptions() {
 427                 if ( !isset( $this->mAccessedOptions ) ) {
 428                         return array();
 429                 }
 430                 return array_keys( $this->mAccessedOptions );
 431         }
 432
 433         /**
 434          * Callback passed by the Parser to the ParserOptions to keep track of which options are used.
 435          * @access private
 436          */
 437         function recordOption( $option ) {
 438                 $this->mAccessedOptions[$option] = true;
 439         }
 440
 441         /**
 442          * Adds an update job to the output. Any update jobs added to the output will eventually bexecuted in order to
 443          * store any secondary information extracted from the page's content.
 444          *
 445          * @since 1.20
 446          *
 447          * @param DataUpdate $update
 448          */
 449         public function addSecondaryDataUpdate( DataUpdate $update ) {
 450                 $this->mSecondaryDataUpdates[] = $update;
 451         }
 452
 453         /**
 454          * Returns any DataUpdate jobs to be executed in order to store secondary information
 455          * extracted from the page's content, including a LinksUpdate object for all links stored in
 456          * this ParserOutput object.
 457          *
 458          * @note: Avoid using this method directly, use ContentHandler::getSecondaryDataUpdates() instead! The content
 459          *        handler may provide additional update objects.
 460          *
 461          * @since 1.20
 462          *
 463          * @param $title Title The title of the page we're updating. If not given, a title object will be created
 464          *                      based on $this->getTitleText()
 465          * @param $recursive Boolean: queue jobs for recursive updates?
 466          *
 467          * @return Array. An array of instances of DataUpdate
 468          */
 469         public function getSecondaryDataUpdates( Title $title = null, $recursive = true ) {
 470                 if ( is_null( $title ) ) {
 471                         $title = Title::newFromText( $this->getTitleText() );
 472                 }
 473
 474                 $linksUpdate = new LinksUpdate( $title, $this, $recursive );
 475
 476                 return array_merge( $this->mSecondaryDataUpdates, array( $linksUpdate ) );
 477         }
 478
 479         /**
 480          * Attaches arbitrary data to this ParserObject. This can be used to store some information in
 481          * the ParserOutput object for later use during page output. The data will be cached along with
 482          * the ParserOutput object, but unlike data set using setProperty(), it is not recorded in the
 483          * database.
 484          *
 485          * This method is provided to overcome the unsafe practice of attaching extra information to a
 486          * ParserObject by directly assigning member variables.
 487          *
 488          * To use setExtensionData() to pass extension information from a hook inside the parser to a
 489          * hook in the page output, use this in the parser hook:
 490          *
 491          * @par Example:
 492          * @code
 493          *    $parser->getOutput()->setExtensionData( 'my_ext_foo', '...' );
 494          * @endcode
 495          *
 496          * And then later, in OutputPageParserOutput or similar:
 497          *
 498          * @par Example:
 499          * @code
 500          *    $output->getExtensionData( 'my_ext_foo' );
 501          * @endcode
 502          *
 503          * In MediaWiki 1.20 and older, you have to use a custom member variable
 504          * within the ParserOutput object:
 505          *
 506          * @par Example:
 507          * @code
 508          *    $parser->getOutput()->my_ext_foo = '...';
 509          * @endcode
 510          *
 511          * @since 1.21
 512          *
 513          * @param string $key The key for accessing the data. Extensions should take care to avoid
 514          *               conflicts in naming keys. It is suggested to use the extension's name as a
 515          *               prefix.
 516          *
 517          * @param mixed $value The value to set. Setting a value to null is equivalent to removing
 518          *              the value.
 519          */
 520         public function setExtensionData( $key, $value ) {
 521                 if ( $value === null ) {
 522                         unset( $this->mExtensionData[$key] );
 523                 } else {
 524                         $this->mExtensionData[$key] = $value;
 525                 }
 526         }
 527
 528         /**
 529          * Gets extensions data previously attached to this ParserOutput using setExtensionData().
 530          * Typically, such data would be set while parsing the page, e.g. by a parser function.
 531          *
 532          * @since 1.21
 533          *
 534          * @param string $key The key to look up.
 535          *
 536          * @return mixed The value previously set for the given key using setExtensionData( $key ),
 537          *         or null if no value was set for this key.
 538          */
 539         public function getExtensionData( $key ) {
 540                 if ( isset( $this->mExtensionData[$key] ) ) {
 541                         return $this->mExtensionData[$key];
 542                 }
 543
 544                 return null;
 545         }
 546
 547 }