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
  55         const EDITSECTION_REGEX = '#<(?:mw:)?editsection page="(.*?)" section="(.*?)"(?:/>|>(.*?)(</(?:mw:)?editsection>))#';
  56
  57         function __construct( $text = '', $languageLinks = array(), $categoryLinks = array(),
  58                 $containsOldMagic = false, $titletext = '' )
  59         {
  60                 $this->mText = $text;
  61                 $this->mLanguageLinks = $languageLinks;
  62                 $this->mCategories = $categoryLinks;
  63                 $this->mContainsOldMagic = $containsOldMagic;
  64                 $this->mTitleText = $titletext;
  65         }
  66
  67         function getText() {
  68                 if ( $this->mEditSectionTokens ) {
  69                         return preg_replace_callback( ParserOutput::EDITSECTION_REGEX,
  70                                 array( &$this, 'replaceEditSectionLinksCallback' ), $this->mText );
  71                 }
  72                 return preg_replace( ParserOutput::EDITSECTION_REGEX, '', $this->mText );
  73         }
  74
  75         /**
  76          * callback used by getText to replace editsection tokens
  77          * @private
  78          * @param $m
  79          * @throws MWException
  80          * @return mixed
  81          */
  82         function replaceEditSectionLinksCallback( $m ) {
  83                 global $wgOut, $wgLang;
  84                 $args = array(
  85                         htmlspecialchars_decode($m[1]),
  86                         htmlspecialchars_decode($m[2]),
  87                         isset($m[4]) ? $m[3] : null,
  88                 );
  89                 $args[0] = Title::newFromText( $args[0] );
  90                 if ( !is_object($args[0]) ) {
  91                         throw new MWException("Bad parser output text.");
  92                 }
  93                 $args[] = $wgLang->getCode();
  94                 $skin = $wgOut->getSkin();
  95                 return call_user_func_array( array( $skin, 'doEditSectionLink' ), $args );
  96         }
  97
  98         function &getLanguageLinks()         { return $this->mLanguageLinks; }
  99         function getInterwikiLinks()         { return $this->mInterwikiLinks; }
 100         function getCategoryLinks()          { return array_keys( $this->mCategories ); }
 101         function &getCategories()            { return $this->mCategories; }
 102         function getTitleText()              { return $this->mTitleText; }
 103         function getSections()               { return $this->mSections; }
 104         function getEditSectionTokens()      { return $this->mEditSectionTokens; }
 105         function &getLinks()                 { return $this->mLinks; }
 106         function &getTemplates()             { return $this->mTemplates; }
 107         function &getTemplateIds()           { return $this->mTemplateIds; }
 108         function &getImages()                { return $this->mImages; }
 109         function &getFileSearchOptions()     { return $this->mFileSearchOptions; }
 110         function &getExternalLinks()         { return $this->mExternalLinks; }
 111         function getNoGallery()              { return $this->mNoGallery; }
 112         function getHeadItems()              { return $this->mHeadItems; }
 113         function getModules()                { return $this->mModules; }
 114         function getModuleScripts()          { return $this->mModuleScripts; }
 115         function getModuleStyles()           { return $this->mModuleStyles; }
 116         function getModuleMessages()         { return $this->mModuleMessages; }
 117         function getOutputHooks()            { return (array)$this->mOutputHooks; }
 118         function getWarnings()               { return array_keys( $this->mWarnings ); }
 119         function getIndexPolicy()            { return $this->mIndexPolicy; }
 120         function getTOCHTML()                { return $this->mTOCHTML; }
 121         function getTimestamp()              { return $this->mTimestamp; }
 122
 123         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
 124         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
 125         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategories, $cl ); }
 126
 127         function setTitleText( $t )          { return wfSetVar( $this->mTitleText, $t ); }
 128         function setSections( $toc )         { return wfSetVar( $this->mSections, $toc ); }
 129         function setEditSectionTokens( $t )  { return wfSetVar( $this->mEditSectionTokens, $t ); }
 130         function setIndexPolicy( $policy )   { return wfSetVar( $this->mIndexPolicy, $policy ); }
 131         function setTOCHTML( $tochtml )      { return wfSetVar( $this->mTOCHTML, $tochtml ); }
 132         function setTimestamp( $timestamp )  { return wfSetVar( $this->mTimestamp, $timestamp ); }
 133
 134         function addCategory( $c, $sort )    { $this->mCategories[$c] = $sort; }
 135         function addLanguageLink( $t )       { $this->mLanguageLinks[] = $t; }
 136         function addWarning( $s )            { $this->mWarnings[$s] = 1; }
 137
 138         function addOutputHook( $hook, $data = false ) {
 139                 $this->mOutputHooks[] = array( $hook, $data );
 140         }
 141
 142         function setNewSection( $value ) {
 143                 $this->mNewSection = (bool)$value;
 144         }
 145         function hideNewSection ( $value ) {
 146                 $this->mHideNewSection = (bool)$value;
 147         }
 148         function getHideNewSection () {
 149                 return (bool)$this->mHideNewSection;
 150         }
 151         function getNewSection() {
 152                 return (bool)$this->mNewSection;
 153         }
 154
 155         /**
 156          * Checks, if a url is pointing to the own server
 157          *
 158          * @param $internal String the server to check against
 159          * @param $url String the url to check
 160          * @return bool
 161          */
 162         static function isLinkInternal( $internal, $url ) {
 163                 return (bool)preg_match( '/^' .
 164                         # If server is proto relative, check also for http/https links
 165                         ( substr( $internal, 0, 2 ) === '//' ? '(?:https?:)?' : '' ) .
 166                         preg_quote( $internal, '/' ) .
 167                         # check for query/path/anchor or end of link in each case
 168                         '(?:[\?\/\#]|$)/i',
 169                         $url
 170                 );
 171         }
 172
 173         function addExternalLink( $url ) {
 174                 # We don't register links pointing to our own server, unless... :-)
 175                 global $wgServer, $wgRegisterInternalExternals;
 176
 177                 $registerExternalLink = true;
 178                 if( !$wgRegisterInternalExternals ) {
 179                         $registerExternalLink = !self::isLinkInternal( $wgServer, $url );
 180                 }
 181                 if( $registerExternalLink ) {
 182                         $this->mExternalLinks[$url] = 1;
 183                 }
 184         }
 185
 186         /**
 187          * Record a local or interwiki inline link for saving in future link tables.
 188          *
 189          * @param $title Title object
 190          * @param $id Mixed: optional known page_id so we can skip the lookup
 191          */
 192         function addLink( $title, $id = null ) {
 193                 if ( $title->isExternal() ) {
 194                         // Don't record interwikis in pagelinks
 195                         $this->addInterwikiLink( $title );
 196                         return;
 197                 }
 198                 $ns = $title->getNamespace();
 199                 $dbk = $title->getDBkey();
 200                 if ( $ns == NS_MEDIA ) {
 201                         // Normalize this pseudo-alias if it makes it down here...
 202                         $ns = NS_FILE;
 203                 } elseif( $ns == NS_SPECIAL ) {
 204                         // We don't record Special: links currently
 205                         // It might actually be wise to, but we'd need to do some normalization.
 206                         return;
 207                 } elseif( $dbk === '' ) {
 208                         // Don't record self links -  [[#Foo]]
 209                         return;
 210                 }
 211                 if ( !isset( $this->mLinks[$ns] ) ) {
 212                         $this->mLinks[$ns] = array();
 213                 }
 214                 if ( is_null( $id ) ) {
 215                         $id = $title->getArticleID();
 216                 }
 217                 $this->mLinks[$ns][$dbk] = $id;
 218         }
 219
 220         /**
 221          * Register a file dependency for this output
 222          * @param $name string Title dbKey
 223          * @param $timestamp string MW timestamp of file creation (or false if non-existing)
 224          * @param $sha1 string base 36 SHA-1 of file (or false if non-existing)
 225          * @return void
 226          */
 227         function addImage( $name, $timestamp = null, $sha1 = null ) {
 228                 $this->mImages[$name] = 1;
 229                 if ( $timestamp !== null && $sha1 !== null ) {
 230                         $this->mFileSearchOptions[$name] = array( 'time' => $timestamp, 'sha1' => $sha1 );
 231                 }
 232         }
 233
 234         /**
 235          * Register a template dependency for this output
 236          * @param $title Title
 237          * @param $page_id
 238          * @param $rev_id
 239          * @return void
 240          */
 241         function addTemplate( $title, $page_id, $rev_id ) {
 242                 $ns = $title->getNamespace();
 243                 $dbk = $title->getDBkey();
 244                 if ( !isset( $this->mTemplates[$ns] ) ) {
 245                         $this->mTemplates[$ns] = array();
 246                 }
 247                 $this->mTemplates[$ns][$dbk] = $page_id;
 248                 if ( !isset( $this->mTemplateIds[$ns] ) ) {
 249                         $this->mTemplateIds[$ns] = array();
 250                 }
 251                 $this->mTemplateIds[$ns][$dbk] = $rev_id; // For versioning
 252         }
 253
 254         /**
 255          * @param $title Title object, must be an interwiki link
 256          * @throws MWException if given invalid input
 257          */
 258         function addInterwikiLink( $title ) {
 259                 $prefix = $title->getInterwiki();
 260                 if( $prefix == '' ) {
 261                         throw new MWException( 'Non-interwiki link passed, internal parser error.' );
 262                 }
 263                 if (!isset($this->mInterwikiLinks[$prefix])) {
 264                         $this->mInterwikiLinks[$prefix] = array();
 265                 }
 266                 $this->mInterwikiLinks[$prefix][$title->getDBkey()] = 1;
 267         }
 268
 269         /**
 270          * Add some text to the "<head>".
 271          * If $tag is set, the section with that tag will only be included once
 272          * in a given page.
 273          */
 274         function addHeadItem( $section, $tag = false ) {
 275                 if ( $tag !== false ) {
 276                         $this->mHeadItems[$tag] = $section;
 277                 } else {
 278                         $this->mHeadItems[] = $section;
 279                 }
 280         }
 281
 282         public function addModules( $modules ) {
 283                 $this->mModules = array_merge( $this->mModules, (array) $modules );
 284         }
 285
 286         public function addModuleScripts( $modules ) {
 287                 $this->mModuleScripts = array_merge( $this->mModuleScripts, (array)$modules );
 288         }
 289
 290         public function addModuleStyles( $modules ) {
 291                 $this->mModuleStyles = array_merge( $this->mModuleStyles, (array)$modules );
 292         }
 293
 294         public function addModuleMessages( $modules ) {
 295                 $this->mModuleMessages = array_merge( $this->mModuleMessages, (array)$modules );
 296         }
 297
 298         /**
 299          * Copy items from the OutputPage object into this one
 300          *
 301          * @param $out OutputPage object
 302          */
 303         public function addOutputPageMetadata( OutputPage $out ) {
 304                 $this->addModules( $out->getModules() );
 305                 $this->addModuleScripts( $out->getModuleScripts() );
 306                 $this->addModuleStyles( $out->getModuleStyles() );
 307                 $this->addModuleMessages( $out->getModuleMessages() );
 308
 309                 $this->mHeadItems = array_merge( $this->mHeadItems, $out->getHeadItemsArray() );
 310         }
 311
 312         /**
 313          * Override the title to be used for display
 314          * -- this is assumed to have been validated
 315          * (check equal normalisation, etc.)
 316          *
 317          * @param $text String: desired title text
 318          */
 319         public function setDisplayTitle( $text ) {
 320                 $this->setTitleText( $text );
 321                 $this->setProperty( 'displaytitle', $text );
 322         }
 323
 324         /**
 325          * Get the title to be used for display
 326          *
 327          * @return String
 328          */
 329         public function getDisplayTitle() {
 330                 $t = $this->getTitleText();
 331                 if( $t === '' ) {
 332                         return false;
 333                 }
 334                 return $t;
 335         }
 336
 337         /**
 338          * Fairly generic flag setter thingy.
 339          */
 340         public function setFlag( $flag ) {
 341                 $this->mFlags[$flag] = true;
 342         }
 343
 344         public function getFlag( $flag ) {
 345                 return isset( $this->mFlags[$flag] );
 346         }
 347
 348         /**
 349          * Set a property to be cached in the DB
 350          */
 351         public function setProperty( $name, $value ) {
 352                 $this->mProperties[$name] = $value;
 353         }
 354
 355         public function getProperty( $name ){
 356                 return isset( $this->mProperties[$name] ) ? $this->mProperties[$name] : false;
 357         }
 358
 359         public function getProperties() {
 360                 if ( !isset( $this->mProperties ) ) {
 361                         $this->mProperties = array();
 362                 }
 363                 return $this->mProperties;
 364         }
 365
 366
 367         /**
 368          * Returns the options from its ParserOptions which have been taken
 369          * into account to produce this output or false if not available.
 370          * @return mixed Array
 371          */
 372          public function getUsedOptions() {
 373                 if ( !isset( $this->mAccessedOptions ) ) {
 374                         return array();
 375                 }
 376                 return array_keys( $this->mAccessedOptions );
 377          }
 378
 379          /**
 380           * Callback passed by the Parser to the ParserOptions to keep track of which options are used.
 381           * @access private
 382           */
 383          function recordOption( $option ) {
 384                  $this->mAccessedOptions[$option] = true;
 385          }
 386
 387         /**
 388          * Adds an update job to the output. Any update jobs added to the output will eventually bexecuted in order to
 389          * store any secondary information extracted from the page's content.
 390          *
 391          * @since 1.20
 392          *
 393          * @param DataUpdate $update
 394          */
 395         public function addSecondaryDataUpdate( DataUpdate $update ) {
 396                 $this->mSecondaryDataUpdates[] = $update;
 397         }
 398
 399         /**
 400          * Returns any DataUpdate jobs to be executed in order to store secondary information
 401          * extracted from the page's content, including a LinksUpdate object for all links stored in
 402          * this ParserOutput object.
 403          *
 404          * @note: Avoid using this method directly, use ContentHandler::getSecondaryDataUpdates() instead! The content
 405          *        handler may provide additional update objects.
 406          *
 407          * @since 1.20
 408          *
 409          * @param $title Title The title of the page we're updating. If not given, a title object will be created
 410          *                      based on $this->getTitleText()
 411          * @param $recursive Boolean: queue jobs for recursive updates?
 412          *
 413          * @return Array. An array of instances of DataUpdate
 414          */
 415         public function getSecondaryDataUpdates( Title $title = null, $recursive = true ) {
 416                 if ( is_null( $title ) ) {
 417                         $title = Title::newFromText( $this->getTitleText() );
 418                 }
 419
 420                 $linksUpdate = new LinksUpdate( $title, $this, $recursive );
 421
 422                 if ( $this->mSecondaryDataUpdates === array() ) {
 423                         return array( $linksUpdate );
 424                 } else {
 425                         $updates = array_merge( $this->mSecondaryDataUpdates, array( $linksUpdate ) );
 426                 }
 427
 428                 return $updates;
 429          }
 430
 431 }