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                         $text = $this->mText;
  71
  72                         // If there's old output with misplaced editsections links cached, mangle it to put them in
  73                         // the right position. We can assume that there is no '</hN>' inside header tags, making this
  74                         // possible to do with a regex.
  75                         $text = preg_replace(
  76                                 //            [ this part is like EDITSECTION_REGEX, but with non-capturing groups                           ]
  77                                 //                                                                                  note the space here ------v
  78                                 '#(<[hH](\d)>)(<(?:mw:)?editsection page="(?:.*?)" section="(?:.*?)"(?:/>|>(?:.*?)(?:</(?:mw:)?editsection>))) ([\s\S]*?)(</[hH]\2>)#',
  79                                 // swap the order of content and editsection link - $2 is ignored since it's the number in hN's tag name
  80                                 '$1$4 $3$5',
  81                                 $text
  82                         );
  83
  84                         return preg_replace_callback( ParserOutput::EDITSECTION_REGEX,
  85                                 array( &$this, 'replaceEditSectionLinksCallback' ), $text );
  86                 }
  87                 return preg_replace( ParserOutput::EDITSECTION_REGEX, '', $this->mText );
  88         }
  89
  90         /**
  91          * callback used by getText to replace editsection tokens
  92          * @private
  93          * @param $m
  94          * @throws MWException
  95          * @return mixed
  96          */
  97         function replaceEditSectionLinksCallback( $m ) {
  98                 global $wgOut, $wgLang;
  99                 $args = array(
 100                         htmlspecialchars_decode( $m[1] ),
 101                         htmlspecialchars_decode( $m[2] ),
 102                         isset( $m[4] ) ? $m[3] : null,
 103                 );
 104                 $args[0] = Title::newFromText( $args[0] );
 105                 if ( !is_object( $args[0] ) ) {
 106                         throw new MWException( "Bad parser output text." );
 107                 }
 108                 $args[] = $wgLang->getCode();
 109                 $skin = $wgOut->getSkin();
 110                 return call_user_func_array( array( $skin, 'doEditSectionLink' ), $args );
 111         }
 112
 113         function &getLanguageLinks()         { return $this->mLanguageLinks; }
 114         function getInterwikiLinks()         { return $this->mInterwikiLinks; }
 115         function getCategoryLinks()          { return array_keys( $this->mCategories ); }
 116         function &getCategories()            { return $this->mCategories; }
 117         function getTitleText()              { return $this->mTitleText; }
 118         function getSections()               { return $this->mSections; }
 119         function getEditSectionTokens()      { return $this->mEditSectionTokens; }
 120         function &getLinks()                 { return $this->mLinks; }
 121         function &getTemplates()             { return $this->mTemplates; }
 122         function &getTemplateIds()           { return $this->mTemplateIds; }
 123         function &getImages()                { return $this->mImages; }
 124         function &getFileSearchOptions()     { return $this->mFileSearchOptions; }
 125         function &getExternalLinks()         { return $this->mExternalLinks; }
 126         function getNoGallery()              { return $this->mNoGallery; }
 127         function getHeadItems()              { return $this->mHeadItems; }
 128         function getModules()                { return $this->mModules; }
 129         function getModuleScripts()          { return $this->mModuleScripts; }
 130         function getModuleStyles()           { return $this->mModuleStyles; }
 131         function getModuleMessages()         { return $this->mModuleMessages; }
 132         function getOutputHooks()            { return (array)$this->mOutputHooks; }
 133         function getWarnings()               { return array_keys( $this->mWarnings ); }
 134         function getIndexPolicy()            { return $this->mIndexPolicy; }
 135         function getTOCHTML()                { return $this->mTOCHTML; }
 136         function getTimestamp()              { return $this->mTimestamp; }
 137
 138         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
 139         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
 140         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategories, $cl ); }
 141
 142         function setTitleText( $t )          { return wfSetVar( $this->mTitleText, $t ); }
 143         function setSections( $toc )         { return wfSetVar( $this->mSections, $toc ); }
 144         function setEditSectionTokens( $t )  { return wfSetVar( $this->mEditSectionTokens, $t ); }
 145         function setIndexPolicy( $policy )   { return wfSetVar( $this->mIndexPolicy, $policy ); }
 146         function setTOCHTML( $tochtml )      { return wfSetVar( $this->mTOCHTML, $tochtml ); }
 147         function setTimestamp( $timestamp )  { return wfSetVar( $this->mTimestamp, $timestamp ); }
 148
 149         function addCategory( $c, $sort )    { $this->mCategories[$c] = $sort; }
 150         function addLanguageLink( $t )       { $this->mLanguageLinks[] = $t; }
 151         function addWarning( $s )            { $this->mWarnings[$s] = 1; }
 152
 153         function addOutputHook( $hook, $data = false ) {
 154                 $this->mOutputHooks[] = array( $hook, $data );
 155         }
 156
 157         function setNewSection( $value ) {
 158                 $this->mNewSection = (bool)$value;
 159         }
 160         function hideNewSection( $value ) {
 161                 $this->mHideNewSection = (bool)$value;
 162         }
 163         function getHideNewSection() {
 164                 return (bool)$this->mHideNewSection;
 165         }
 166         function getNewSection() {
 167                 return (bool)$this->mNewSection;
 168         }
 169
 170         /**
 171          * Checks, if a url is pointing to the own server
 172          *
 173          * @param string $internal the server to check against
 174          * @param string $url the url to check
 175          * @return bool
 176          */
 177         static function isLinkInternal( $internal, $url ) {
 178                 return (bool)preg_match( '/^' .
 179                         # If server is proto relative, check also for http/https links
 180                         ( substr( $internal, 0, 2 ) === '//' ? '(?:https?:)?' : '' ) .
 181                         preg_quote( $internal, '/' ) .
 182                         # check for query/path/anchor or end of link in each case
 183                         '(?:[\?\/\#]|$)/i',
 184                         $url
 185                 );
 186         }
 187
 188         function addExternalLink( $url ) {
 189                 # We don't register links pointing to our own server, unless... :-)
 190                 global $wgServer, $wgRegisterInternalExternals;
 191
 192                 $registerExternalLink = true;
 193                 if ( !$wgRegisterInternalExternals ) {
 194                         $registerExternalLink = !self::isLinkInternal( $wgServer, $url );
 195                 }
 196                 if ( $registerExternalLink ) {
 197                         $this->mExternalLinks[$url] = 1;
 198                 }
 199         }
 200
 201         /**
 202          * Record a local or interwiki inline link for saving in future link tables.
 203          *
 204          * @param $title Title object
 205          * @param $id Mixed: optional known page_id so we can skip the lookup
 206          */
 207         function addLink( Title $title, $id = null ) {
 208                 if ( $title->isExternal() ) {
 209                         // Don't record interwikis in pagelinks
 210                         $this->addInterwikiLink( $title );
 211                         return;
 212                 }
 213                 $ns = $title->getNamespace();
 214                 $dbk = $title->getDBkey();
 215                 if ( $ns == NS_MEDIA ) {
 216                         // Normalize this pseudo-alias if it makes it down here...
 217                         $ns = NS_FILE;
 218                 } elseif ( $ns == NS_SPECIAL ) {
 219                         // We don't record Special: links currently
 220                         // It might actually be wise to, but we'd need to do some normalization.
 221                         return;
 222                 } elseif ( $dbk === '' ) {
 223                         // Don't record self links -  [[#Foo]]
 224                         return;
 225                 }
 226                 if ( !isset( $this->mLinks[$ns] ) ) {
 227                         $this->mLinks[$ns] = array();
 228                 }
 229                 if ( is_null( $id ) ) {
 230                         $id = $title->getArticleID();
 231                 }
 232                 $this->mLinks[$ns][$dbk] = $id;
 233         }
 234
 235         /**
 236          * Register a file dependency for this output
 237          * @param string $name Title dbKey
 238          * @param string $timestamp MW timestamp of file creation (or false if non-existing)
 239          * @param string $sha1 base 36 SHA-1 of file (or false if non-existing)
 240          * @return void
 241          */
 242         function addImage( $name, $timestamp = null, $sha1 = null ) {
 243                 $this->mImages[$name] = 1;
 244                 if ( $timestamp !== null && $sha1 !== null ) {
 245                         $this->mFileSearchOptions[$name] = array( 'time' => $timestamp, 'sha1' => $sha1 );
 246                 }
 247         }
 248
 249         /**
 250          * Register a template dependency for this output
 251          * @param $title Title
 252          * @param $page_id
 253          * @param $rev_id
 254          * @return void
 255          */
 256         function addTemplate( $title, $page_id, $rev_id ) {
 257                 $ns = $title->getNamespace();
 258                 $dbk = $title->getDBkey();
 259                 if ( !isset( $this->mTemplates[$ns] ) ) {
 260                         $this->mTemplates[$ns] = array();
 261                 }
 262                 $this->mTemplates[$ns][$dbk] = $page_id;
 263                 if ( !isset( $this->mTemplateIds[$ns] ) ) {
 264                         $this->mTemplateIds[$ns] = array();
 265                 }
 266                 $this->mTemplateIds[$ns][$dbk] = $rev_id; // For versioning
 267         }
 268
 269         /**
 270          * @param $title Title object, must be an interwiki link
 271          * @throws MWException if given invalid input
 272          */
 273         function addInterwikiLink( $title ) {
 274                 $prefix = $title->getInterwiki();
 275                 if ( $prefix == '' ) {
 276                         throw new MWException( 'Non-interwiki link passed, internal parser error.' );
 277                 }
 278                 if ( !isset( $this->mInterwikiLinks[$prefix] ) ) {
 279                         $this->mInterwikiLinks[$prefix] = array();
 280                 }
 281                 $this->mInterwikiLinks[$prefix][$title->getDBkey()] = 1;
 282         }
 283
 284         /**
 285          * Add some text to the "<head>".
 286          * If $tag is set, the section with that tag will only be included once
 287          * in a given page.
 288          */
 289         function addHeadItem( $section, $tag = false ) {
 290                 if ( $tag !== false ) {
 291                         $this->mHeadItems[$tag] = $section;
 292                 } else {
 293                         $this->mHeadItems[] = $section;
 294                 }
 295         }
 296
 297         public function addModules( $modules ) {
 298                 $this->mModules = array_merge( $this->mModules, (array) $modules );
 299         }
 300
 301         public function addModuleScripts( $modules ) {
 302                 $this->mModuleScripts = array_merge( $this->mModuleScripts, (array)$modules );
 303         }
 304
 305         public function addModuleStyles( $modules ) {
 306                 $this->mModuleStyles = array_merge( $this->mModuleStyles, (array)$modules );
 307         }
 308
 309         public function addModuleMessages( $modules ) {
 310                 $this->mModuleMessages = array_merge( $this->mModuleMessages, (array)$modules );
 311         }
 312
 313         /**
 314          * Copy items from the OutputPage object into this one
 315          *
 316          * @param $out OutputPage object
 317          */
 318         public function addOutputPageMetadata( OutputPage $out ) {
 319                 $this->addModules( $out->getModules() );
 320                 $this->addModuleScripts( $out->getModuleScripts() );
 321                 $this->addModuleStyles( $out->getModuleStyles() );
 322                 $this->addModuleMessages( $out->getModuleMessages() );
 323
 324                 $this->mHeadItems = array_merge( $this->mHeadItems, $out->getHeadItemsArray() );
 325         }
 326
 327         /**
 328          * Override the title to be used for display
 329          * -- this is assumed to have been validated
 330          * (check equal normalisation, etc.)
 331          *
 332          * @param string $text desired title text
 333          */
 334         public function setDisplayTitle( $text ) {
 335                 $this->setTitleText( $text );
 336                 $this->setProperty( 'displaytitle', $text );
 337         }
 338
 339         /**
 340          * Get the title to be used for display
 341          *
 342          * @return String
 343          */
 344         public function getDisplayTitle() {
 345                 $t = $this->getTitleText();
 346                 if ( $t === '' ) {
 347                         return false;
 348                 }
 349                 return $t;
 350         }
 351
 352         /**
 353          * Fairly generic flag setter thingy.
 354          */
 355         public function setFlag( $flag ) {
 356                 $this->mFlags[$flag] = true;
 357         }
 358
 359         public function getFlag( $flag ) {
 360                 return isset( $this->mFlags[$flag] );
 361         }
 362
 363         /**
 364          * Set a property to be stored in the page_props database table.
 365          *
 366          * page_props is a key value store indexed by the page ID. This allows
 367          * the parser to set a property on a page which can then be quickly
 368          * retrieved given the page ID or via a DB join when given the page
 369          * title.
 370          *
 371          * setProperty() is thus used to propagate properties from the parsed
 372          * page to request contexts other than a page view of the currently parsed
 373          * article.
 374          *
 375          * Some applications examples:
 376          *
 377          *   * To implement hidden categories, hiding pages from category listings
 378          *     by storing a property.
 379          *
 380          *   * Overriding the displayed article title.
 381          *   @see ParserOutput::setDisplayTitle()
 382          *
 383          *   * To implement image tagging, for example displaying an icon on an
 384          *     image thumbnail to indicate that it is listed for deletion on
 385          *     Wikimedia Commons.
 386          *     This is not actually implemented, yet but would be pretty cool.
 387          *
 388          * @note: Do not use setProperty() to set a property which is only used
 389          * in a context where the ParserOutput object itself is already available,
 390          * for example a normal page view. There is no need to save such a property
 391          * in the database since it the text is already parsed. You can just hook
 392          * OutputPageParserOutput and get your data out of the ParserOutput object.
 393          *
 394          * If you are writing an extension where you want to set a property in the
 395          * parser which is used by an OutputPageParserOutput hook, you have to
 396          * associate the extension data directly with the ParserOutput object.
 397          * Since MediaWiki 1.21, you can use setExtensionData() to do this:
 398          *
 399          * @par Example:
 400          * @code
 401          *    $parser->getOutput()->setExtensionData( 'my_ext_foo', '...' );
 402          * @endcode
 403          *
 404          * And then later, in OutputPageParserOutput or similar:
 405          *
 406          * @par Example:
 407          * @code
 408          *    $output->getExtensionData( 'my_ext_foo' );
 409          * @endcode
 410          *
 411          * In MediaWiki 1.20 and older, you have to use a custom member variable
 412          * within the ParserOutput object:
 413          *
 414          * @par Example:
 415          * @code
 416          *    $parser->getOutput()->my_ext_foo = '...';
 417          * @endcode
 418          *
 419          */
 420         public function setProperty( $name, $value ) {
 421                 $this->mProperties[$name] = $value;
 422         }
 423
 424         public function getProperty( $name ) {
 425                 return isset( $this->mProperties[$name] ) ? $this->mProperties[$name] : false;
 426         }
 427
 428         public function getProperties() {
 429                 if ( !isset( $this->mProperties ) ) {
 430                         $this->mProperties = array();
 431                 }
 432                 return $this->mProperties;
 433         }
 434
 435         /**
 436          * Returns the options from its ParserOptions which have been taken
 437          * into account to produce this output or false if not available.
 438          * @return mixed Array
 439          */
 440         public function getUsedOptions() {
 441                 if ( !isset( $this->mAccessedOptions ) ) {
 442                         return array();
 443                 }
 444                 return array_keys( $this->mAccessedOptions );
 445         }
 446
 447         /**
 448          * Callback passed by the Parser to the ParserOptions to keep track of which options are used.
 449          * @access private
 450          */
 451         function recordOption( $option ) {
 452                 $this->mAccessedOptions[$option] = true;
 453         }
 454
 455         /**
 456          * Adds an update job to the output. Any update jobs added to the output will eventually bexecuted in order to
 457          * store any secondary information extracted from the page's content.
 458          *
 459          * @since 1.20
 460          *
 461          * @param DataUpdate $update
 462          */
 463         public function addSecondaryDataUpdate( DataUpdate $update ) {
 464                 $this->mSecondaryDataUpdates[] = $update;
 465         }
 466
 467         /**
 468          * Returns any DataUpdate jobs to be executed in order to store secondary information
 469          * extracted from the page's content, including a LinksUpdate object for all links stored in
 470          * this ParserOutput object.
 471          *
 472          * @note: Avoid using this method directly, use ContentHandler::getSecondaryDataUpdates() instead! The content
 473          *        handler may provide additional update objects.
 474          *
 475          * @since 1.20
 476          *
 477          * @param $title Title The title of the page we're updating. If not given, a title object will be created
 478          *                      based on $this->getTitleText()
 479          * @param $recursive Boolean: queue jobs for recursive updates?
 480          *
 481          * @return Array. An array of instances of DataUpdate
 482          */
 483         public function getSecondaryDataUpdates( Title $title = null, $recursive = true ) {
 484                 if ( is_null( $title ) ) {
 485                         $title = Title::newFromText( $this->getTitleText() );
 486                 }
 487
 488                 $linksUpdate = new LinksUpdate( $title, $this, $recursive );
 489
 490                 return array_merge( $this->mSecondaryDataUpdates, array( $linksUpdate ) );
 491         }
 492
 493         /**
 494          * Attaches arbitrary data to this ParserObject. This can be used to store some information in
 495          * the ParserOutput object for later use during page output. The data will be cached along with
 496          * the ParserOutput object, but unlike data set using setProperty(), it is not recorded in the
 497          * database.
 498          *
 499          * This method is provided to overcome the unsafe practice of attaching extra information to a
 500          * ParserObject by directly assigning member variables.
 501          *
 502          * To use setExtensionData() to pass extension information from a hook inside the parser to a
 503          * hook in the page output, use this in the parser hook:
 504          *
 505          * @par Example:
 506          * @code
 507          *    $parser->getOutput()->setExtensionData( 'my_ext_foo', '...' );
 508          * @endcode
 509          *
 510          * And then later, in OutputPageParserOutput or similar:
 511          *
 512          * @par Example:
 513          * @code
 514          *    $output->getExtensionData( 'my_ext_foo' );
 515          * @endcode
 516          *
 517          * In MediaWiki 1.20 and older, you have to use a custom member variable
 518          * within the ParserOutput object:
 519          *
 520          * @par Example:
 521          * @code
 522          *    $parser->getOutput()->my_ext_foo = '...';
 523          * @endcode
 524          *
 525          * @since 1.21
 526          *
 527          * @param string $key The key for accessing the data. Extensions should take care to avoid
 528          *               conflicts in naming keys. It is suggested to use the extension's name as a
 529          *               prefix.
 530          *
 531          * @param mixed $value The value to set. Setting a value to null is equivalent to removing
 532          *              the value.
 533          */
 534         public function setExtensionData( $key, $value ) {
 535                 if ( $value === null ) {
 536                         unset( $this->mExtensionData[$key] );
 537                 } else {
 538                         $this->mExtensionData[$key] = $value;
 539                 }
 540         }
 541
 542         /**
 543          * Gets extensions data previously attached to this ParserOutput using setExtensionData().
 544          * Typically, such data would be set while parsing the page, e.g. by a parser function.
 545          *
 546          * @since 1.21
 547          *
 548          * @param string $key The key to look up.
 549          *
 550          * @return mixed The value previously set for the given key using setExtensionData( $key ),
 551          *         or null if no value was set for this key.
 552          */
 553         public function getExtensionData( $key ) {
 554                 if ( isset( $this->mExtensionData[$key] ) ) {
 555                         return $this->mExtensionData[$key];
 556                 }
 557
 558                 return null;
 559         }
 560
 561 }