includes/ContentHandler.php

   1 <?php
   2
   3 /**
   4  * Exception representing a failure to serialize or unserialize a content object.
   5  */
   6 class MWContentSerializationException extends MWException {
   7
   8 }
   9
  10 /**
  11  * A content handler knows how do deal with a specific type of content on a wiki
  12  * page. Content is stored in the database in a serialized form (using a
  13  * serialization format a.k.a. MIME type) and is unserialized into its native
  14  * PHP representation (the content model), which is wrapped in an instance of
  15  * the appropriate subclass of Content.
  16  *
  17  * ContentHandler instances are stateless singletons that serve, among other
  18  * things, as a factory for Content objects. Generally, there is one subclass
  19  * of ContentHandler and one subclass of Content for every type of content model.
  20  *
  21  * Some content types have a flat model, that is, their native representation
  22  * is the same as their serialized form. Examples would be JavaScript and CSS
  23  * code. As of now, this also applies to wikitext (MediaWiki's default content
  24  * type), but wikitext content may be represented by a DOM or AST structure in
  25  * the future.
  26  *
  27  * @since 1.WD
  28  */
  29 abstract class ContentHandler {
  30
  31         /**
  32          * Convenience function for getting flat text from a Content object. This
  33          * should only be used in the context of backwards compatibility with code
  34          * that is not yet able to handle Content objects!
  35          *
  36          * If $content is null, this method returns the empty string.
  37          *
  38          * If $content is an instance of TextContent, this method returns the flat
  39          * text as returned by $content->getNativeData().
  40          *
  41          * If $content is not a TextContent object, the behavior of this method
  42          * depends on the global $wgContentHandlerTextFallback:
  43          * - If $wgContentHandlerTextFallback is 'fail' and $content is not a
  44          *   TextContent object, an MWException is thrown.
  45          * - If $wgContentHandlerTextFallback is 'serialize' and $content is not a
  46          *   TextContent object, $content->serialize() is called to get a string
  47          *   form of the content.
  48          * - If $wgContentHandlerTextFallback is 'ignore' and $content is not a
  49          *   TextContent object, this method returns null.
  50          * - otherwise, the behaviour is undefined.
  51          *
  52          * @since WD.1
  53          *
  54          * @static
  55          * @param $content Content|null
  56          * @return null|string the textual form of $content, if available
  57          * @throws MWException if $content is not an instance of TextContent and
  58          *   $wgContentHandlerTextFallback was set to 'fail'.
  59          */
  60         public static function getContentText( Content $content = null ) {
  61                 global $wgContentHandlerTextFallback;
  62
  63                 if ( is_null( $content ) ) {
  64                         return '';
  65                 }
  66
  67                 if ( $content instanceof TextContent ) {
  68                         return $content->getNativeData();
  69                 }
  70
  71                 if ( $wgContentHandlerTextFallback == 'fail' ) {
  72                         throw new MWException(
  73                                 "Attempt to get text from Content with model " .
  74                                 $content->getModel()
  75                         );
  76                 }
  77
  78                 if ( $wgContentHandlerTextFallback == 'serialize' ) {
  79                         return $content->serialize();
  80                 }
  81
  82                 return null;
  83         }
  84
  85         /**
  86          * Convenience function for creating a Content object from a given textual
  87          * representation.
  88          *
  89          * $text will be deserialized into a Content object of the model specified
  90          * by $modelId (or, if that is not given, $title->getContentModel()) using
  91          * the given format.
  92          *
  93          * @since WD.1
  94          *
  95          * @static
  96          *
  97          * @param $text string the textual representation, will be
  98          *    unserialized to create the Content object
  99          * @param $title null|Title the title of the page this text belongs to.
 100          *    Required if $modelId is not provided.
 101          * @param $modelId null|string the model to deserialize to. If not provided,
 102          *    $title->getContentModel() is used.
 103          * @param $format null|string the format to use for deserialization. If not
 104          *    given, the model's default format is used.
 105          *
 106          * @return Content a Content object representing $text
 107          *
 108          * @throw MWException if $model or $format is not supported or if $text can
 109          *    not be unserialized using $format.
 110          */
 111         public static function makeContent( $text, Title $title = null,
 112                 $modelId = null, $format = null )
 113         {
 114                 if ( is_null( $modelId ) ) {
 115                         if ( is_null( $title ) ) {
 116                                 throw new MWException( "Must provide a Title object or a content model ID." );
 117                         }
 118
 119                         $modelId = $title->getContentModel();
 120                 }
 121
 122                 $handler = ContentHandler::getForModelID( $modelId );
 123                 return $handler->unserializeContent( $text, $format );
 124         }
 125
 126         /**
 127          * Returns the name of the default content model to be used for the page
 128          * with the given title.
 129          *
 130          * Note: There should rarely be need to call this method directly.
 131          * To determine the actual content model for a given page, use
 132          * Title::getContentModel().
 133          *
 134          * Which model is to be used by default for the page is determined based
 135          * on several factors:
 136          * - The global setting $wgNamespaceContentModels specifies a content model
 137          *   per namespace.
 138          * - The hook DefaultModelFor may be used to override the page's default
 139          *   model.
 140          * - Pages in NS_MEDIAWIKI and NS_USER default to the CSS or JavaScript
 141          *   model if they end in .js or .css, respectively.
 142          * - Pages in NS_MEDIAWIKI default to the wikitext model otherwise.
 143          * - The hook TitleIsCssOrJsPage may be used to force a page to use the CSS
 144          *   or JavaScript model if they end in .js or .css, respectively.
 145          * - The hook TitleIsWikitextPage may be used to force a page to use the
 146          *   wikitext model.
 147          *
 148          * If none of the above applies, the wikitext model is used.
 149          *
 150          * Note: this is used by, and may thus not use, Title::getContentModel()
 151          *
 152          * @since WD.1
 153          *
 154          * @static
 155          * @param $title Title
 156          * @return null|string default model name for the page given by $title
 157          */
 158         public static function getDefaultModelFor( Title $title ) {
 159                 global $wgNamespaceContentModels;
 160
 161                 // NOTE: this method must not rely on $title->getContentModel() directly or indirectly,
 162                 //       because it is used to initialize the mContentModel member.
 163
 164                 $ns = $title->getNamespace();
 165
 166                 $ext = false;
 167                 $m = null;
 168                 $model = null;
 169
 170                 if ( !empty( $wgNamespaceContentModels[ $ns ] ) ) {
 171                         $model = $wgNamespaceContentModels[ $ns ];
 172                 }
 173
 174                 // Hook can determine default model
 175                 if ( !wfRunHooks( 'ContentHandlerDefaultModelFor', array( $title, &$model ) ) ) {
 176                         if ( !is_null( $model ) ) {
 177                                 return $model;
 178                         }
 179                 }
 180
 181                 // Could this page contain custom CSS or JavaScript, based on the title?
 182                 $isCssOrJsPage = NS_MEDIAWIKI == $ns && preg_match( '!\.(css|js)$!u', $title->getText(), $m );
 183                 if ( $isCssOrJsPage ) {
 184                         $ext = $m[1];
 185                 }
 186
 187                 // Hook can force JS/CSS
 188                 wfRunHooks( 'TitleIsCssOrJsPage', array( $title, &$isCssOrJsPage ) );
 189
 190                 // Is this a .css subpage of a user page?
 191                 $isJsCssSubpage = NS_USER == $ns
 192                         && !$isCssOrJsPage
 193                         && preg_match( "/\\/.*\\.(js|css)$/", $title->getText(), $m );
 194                 if ( $isJsCssSubpage ) {
 195                         $ext = $m[1];
 196                 }
 197
 198                 // Is this wikitext, according to $wgNamespaceContentModels or the DefaultModelFor hook?
 199                 $isWikitext = is_null( $model ) || $model == CONTENT_MODEL_WIKITEXT;
 200                 $isWikitext = $isWikitext && !$isCssOrJsPage && !$isJsCssSubpage;
 201
 202                 // Hook can override $isWikitext
 203                 wfRunHooks( 'TitleIsWikitextPage', array( $title, &$isWikitext ) );
 204
 205                 if ( !$isWikitext ) {
 206                         switch ( $ext ) {
 207                                 case 'js':
 208                                         return CONTENT_MODEL_JAVASCRIPT;
 209                                 case 'css':
 210                                         return CONTENT_MODEL_CSS;
 211                                 default:
 212                                         return is_null( $model ) ? CONTENT_MODEL_TEXT : $model;
 213                         }
 214                 }
 215
 216                 // We established that it must be wikitext
 217
 218                 return CONTENT_MODEL_WIKITEXT;
 219         }
 220
 221         /**
 222          * Returns the appropriate ContentHandler singleton for the given title.
 223          *
 224          * @since WD.1
 225          *
 226          * @static
 227          * @param $title Title
 228          * @return ContentHandler
 229          */
 230         public static function getForTitle( Title $title ) {
 231                 $modelId = $title->getContentModel();
 232                 return ContentHandler::getForModelID( $modelId );
 233         }
 234
 235         /**
 236          * Returns the appropriate ContentHandler singleton for the given Content
 237          * object.
 238          *
 239          * @since WD.1
 240          *
 241          * @static
 242          * @param $content Content
 243          * @return ContentHandler
 244          */
 245         public static function getForContent( Content $content ) {
 246                 $modelId = $content->getModel();
 247                 return ContentHandler::getForModelID( $modelId );
 248         }
 249
 250         /**
 251          * @var Array A Cache of ContentHandler instances by model id
 252          */
 253         static $handlers;
 254
 255         /**
 256          * Returns the ContentHandler singleton for the given model ID. Use the
 257          * CONTENT_MODEL_XXX constants to identify the desired content model.
 258          *
 259          * ContentHandler singletons are taken from the global $wgContentHandlers
 260          * array. Keys in that array are model names, the values are either
 261          * ContentHandler singleton objects, or strings specifying the appropriate
 262          * subclass of ContentHandler.
 263          *
 264          * If a class name is encountered when looking up the singleton for a given
 265          * model name, the class is instantiated and the class name is replaced by
 266          * the resulting singleton in $wgContentHandlers.
 267          *
 268          * If no ContentHandler is defined for the desired $modelId, the
 269          * ContentHandler may be provided by the ContentHandlerForModelID hook.
 270          * If no ContentHandler can be determined, an MWException is raised.
 271          *
 272          * @since WD.1
 273          *
 274          * @static
 275          * @param $modelId String The ID of the content model for which to get a
 276          *    handler. Use CONTENT_MODEL_XXX constants.
 277          * @return ContentHandler The ContentHandler singleton for handling the
 278          *    model given by $modelId
 279          * @throws MWException if no handler is known for $modelId.
 280          */
 281         public static function getForModelID( $modelId ) {
 282                 global $wgContentHandlers;
 283
 284                 if ( isset( ContentHandler::$handlers[$modelId] ) ) {
 285                         return ContentHandler::$handlers[$modelId];
 286                 }
 287
 288                 if ( empty( $wgContentHandlers[$modelId] ) ) {
 289                         $handler = null;
 290
 291                         wfRunHooks( 'ContentHandlerForModelID', array( $modelId, &$handler ) );
 292
 293                         if ( $handler === null ) {
 294                                 throw new MWException( "No handler for model #$modelId registered in \$wgContentHandlers" );
 295                         }
 296
 297                         if ( !( $handler instanceof ContentHandler ) ) {
 298                                 throw new MWException( "ContentHandlerForModelID must supply a ContentHandler instance" );
 299                         }
 300                 } else {
 301                         $class = $wgContentHandlers[$modelId];
 302                         $handler = new $class( $modelId );
 303
 304                         if ( !( $handler instanceof ContentHandler ) ) {
 305                                 throw new MWException( "$class from \$wgContentHandlers is not compatible with ContentHandler" );
 306                         }
 307                 }
 308
 309                 ContentHandler::$handlers[$modelId] = $handler;
 310                 return ContentHandler::$handlers[$modelId];
 311         }
 312
 313         /**
 314          * Returns the localized name for a given content model.
 315          *
 316          * Model names are localized using system messages. Message keys
 317          * have the form content-model-$name, where $name is getContentModelName( $id ).
 318          *
 319          * @static
 320          * @param $name String The content model ID, as given by a CONTENT_MODEL_XXX
 321          *    constant or returned by Revision::getContentModel().
 322          *
 323          * @return string The content format's localized name.
 324          * @throws MWException if the model id isn't known.
 325          */
 326         public static function getLocalizedName( $name ) {
 327                 $key = "content-model-$name";
 328
 329                 if ( wfEmptyMsg( $key ) ) return $name;
 330                 else return wfMsg( $key );
 331         }
 332
 333         public static function getAllContentFormats() {
 334                 global $wgContentHandlers;
 335
 336                 $formats = array();
 337
 338                 foreach ( $wgContentHandlers as $model => $class ) {
 339                         $handler = ContentHandler::getForModelID( $model );
 340                         $formats = array_merge( $formats, $handler->getSupportedFormats() );
 341                 }
 342
 343                 $formats = array_unique( $formats );
 344                 return $formats;
 345         }
 346
 347         // ------------------------------------------------------------------------
 348
 349         protected $mModelID;
 350         protected $mSupportedFormats;
 351
 352         /**
 353          * Constructor, initializing the ContentHandler instance with its model ID
 354          * and a list of supported formats. Values for the parameters are typically
 355          * provided as literals by subclass's constructors.
 356          *
 357          * @param $modelId String (use CONTENT_MODEL_XXX constants).
 358          * @param $formats array List for supported serialization formats
 359          *    (typically as MIME types)
 360          */
 361         public function __construct( $modelId, $formats ) {
 362                 $this->mModelID = $modelId;
 363                 $this->mSupportedFormats = $formats;
 364
 365                 $this->mModelName = preg_replace( '/(Content)?Handler$/', '', get_class( $this ) );
 366                 $this->mModelName = preg_replace( '/[_\\\\]/', '', $this->mModelName );
 367                 $this->mModelName = strtolower( $this->mModelName );
 368         }
 369
 370         /**
 371          * Serializes a Content object of the type supported by this ContentHandler.
 372          *
 373          * @since WD.1
 374          *
 375          * @abstract
 376          * @param $content Content The Content object to serialize
 377          * @param $format null|String The desired serialization format
 378          * @return string Serialized form of the content
 379          */
 380         public abstract function serializeContent( Content $content, $format = null );
 381
 382         /**
 383          * Unserializes a Content object of the type supported by this ContentHandler.
 384          *
 385          * @since WD.1
 386          *
 387          * @abstract
 388          * @param $blob string serialized form of the content
 389          * @param $format null|String the format used for serialization
 390          * @return Content the Content object created by deserializing $blob
 391          */
 392         public abstract function unserializeContent( $blob, $format = null );
 393
 394         /**
 395          * Creates an empty Content object of the type supported by this
 396          * ContentHandler.
 397          *
 398          * @since WD.1
 399          *
 400          * @return Content
 401          */
 402         public abstract function makeEmptyContent();
 403
 404         /**
 405          * Returns the model id that identifies the content model this
 406          * ContentHandler can handle. Use with the CONTENT_MODEL_XXX constants.
 407          *
 408          * @since WD.1
 409          *
 410          * @return String The model ID
 411          */
 412         public function getModelID() {
 413                 return $this->mModelID;
 414         }
 415
 416         /**
 417          * Throws an MWException if $model_id is not the ID of the content model
 418          * supported by this ContentHandler.
 419          *
 420          * @since WD.1
 421          *
 422          * @param String $model_id The model to check
 423          *
 424          * @throws MWException
 425          */
 426         protected function checkModelID( $model_id ) {
 427                 if ( $model_id !== $this->mModelID ) {
 428                         throw new MWException( "Bad content model: " .
 429                                 "expected {$this->mModelID} " .
 430                                 "but got $model_id." );
 431                 }
 432         }
 433
 434         /**
 435          * Returns a list of serialization formats supported by the
 436          * serializeContent() and unserializeContent() methods of this
 437          * ContentHandler.
 438          *
 439          * @since WD.1
 440          *
 441          * @return array of serialization formats as MIME type like strings
 442          */
 443         public function getSupportedFormats() {
 444                 return $this->mSupportedFormats;
 445         }
 446
 447         /**
 448          * The format used for serialization/deserialization by default by this
 449          * ContentHandler.
 450          *
 451          * This default implementation will return the first element of the array
 452          * of formats that was passed to the constructor.
 453          *
 454          * @since WD.1
 455          *
 456          * @return string the name of the default serialization format as a MIME type
 457          */
 458         public function getDefaultFormat() {
 459                 return $this->mSupportedFormats[0];
 460         }
 461
 462         /**
 463          * Returns true if $format is a serialization format supported by this
 464          * ContentHandler, and false otherwise.
 465          *
 466          * Note that if $format is null, this method always returns true, because
 467          * null means "use the default format".
 468          *
 469          * @since WD.1
 470          *
 471          * @param $format string the serialization format to check
 472          * @return bool
 473          */
 474         public function isSupportedFormat( $format ) {
 475
 476                 if ( !$format ) {
 477                         return true; // this means "use the default"
 478                 }
 479
 480                 return in_array( $format, $this->mSupportedFormats );
 481         }
 482
 483         /**
 484          * Throws an MWException if isSupportedFormat( $format ) is not true.
 485          * Convenient for checking whether a format provided as a parameter is
 486          * actually supported.
 487          *
 488          * @param $format string the serialization format to check
 489          *
 490          * @throws MWException
 491          */
 492         protected function checkFormat( $format ) {
 493                 if ( !$this->isSupportedFormat( $format ) ) {
 494                         throw new MWException(
 495                                 "Format $format is not supported for content model "
 496                                 . $this->getModelID()
 497                         );
 498                 }
 499         }
 500
 501         /**
 502          * Returns true if the content is consistent with the database, that is if
 503          * saving it to the database would not violate any global constraints.
 504          *
 505          * Content needs to be valid using this method before it can be saved.
 506          *
 507          * This default implementation always returns true.
 508          *
 509          * @since WD.1
 510          *
 511          * @param $content \Content
 512          *
 513          * @return boolean
 514          */
 515         public function isConsistentWithDatabase( Content $content ) {
 516                 return true;
 517         }
 518
 519         /**
 520          * Returns overrides for action handlers.
 521          * Classes listed here will be used instead of the default one when
 522          * (and only when) $wgActions[$action] === true. This allows subclasses
 523          * to override the default action handlers.
 524          *
 525          * @since WD.1
 526          *
 527          * @return Array
 528          */
 529         public function getActionOverrides() {
 530                 return array();
 531         }
 532
 533         /**
 534          * Factory for creating an appropriate DifferenceEngine for this content model.
 535          *
 536          * @since WD.1
 537          *
 538          * @param $context IContextSource context to use, anything else will be
 539          *    ignored
 540          * @param $old Integer Old ID we want to show and diff with.
 541          * @param $new int|string String either 'prev' or 'next'.
 542          * @param $rcid Integer ??? FIXME (default 0)
 543          * @param $refreshCache boolean If set, refreshes the diff cache
 544          * @param $unhide boolean If set, allow viewing deleted revs
 545          *
 546          * @return DifferenceEngine
 547          */
 548         public function createDifferenceEngine( IContextSource $context,
 549                 $old = 0, $new = 0,
 550                 $rcid = 0, # FIXME: use everywhere!
 551                 $refreshCache = false, $unhide = false
 552         ) {
 553                 $this->checkModelID( $context->getTitle()->getContentModel() );
 554
 555                 $diffEngineClass = $this->getDiffEngineClass();
 556
 557                 return new $diffEngineClass( $context, $old, $new, $rcid, $refreshCache, $unhide );
 558         }
 559
 560         /**
 561          * Get the language in which the content of the given page is written.
 562          *
 563          * This default implementation returns $wgContLang->getCode().
 564          *
 565          * Note that a page's language must be permanent and cacheable, that is, it must not depend
 566          * on user preferences, request parameters or session state.
 567          *
 568          * Also note that the page language may or may not depend on the actual content of the page,
 569          * that is, this method may load the content in order to determine the language.
 570          *
 571          * @since 1.WD
 572          *
 573          * @param Title        $title the page to determine the language for.
 574          * @param Content|null $content the page's content, if you have it handy, to avoid reloading it.
 575          *
 576          * @return Language the page's language code
 577          */
 578         public function getPageLanguage( Title $title, Content $content = null ) {
 579                 global $wgContLang;
 580                 return $wgContLang;
 581         }
 582
 583         /**
 584          * Returns the name of the diff engine to use.
 585          *
 586          * @since WD.1
 587          *
 588          * @return string
 589          */
 590         protected function getDiffEngineClass() {
 591                 return 'DifferenceEngine';
 592         }
 593
 594         /**
 595          * Attempts to merge differences between three versions.
 596          * Returns a new Content object for a clean merge and false for failure or
 597          * a conflict.
 598          *
 599          * This default implementation always returns false.
 600          *
 601          * @since WD.1
 602          *
 603          * @param $oldContent Content|string  String
 604          * @param $myContent Content|string   String
 605          * @param $yourContent Content|string String
 606          *
 607          * @return Content|Bool
 608          */
 609         public function merge3( Content $oldContent, Content $myContent, Content $yourContent ) {
 610                 return false;
 611         }
 612
 613         /**
 614          * Return an applicable auto-summary if one exists for the given edit.
 615          *
 616          * @since WD.1
 617          *
 618          * @param $oldContent Content|null: the previous text of the page.
 619          * @param $newContent Content|null: The submitted text of the page.
 620          * @param $flags int Bit mask: a bit mask of flags submitted for the edit.
 621          *
 622          * @return string An appropriate auto-summary, or an empty string.
 623          */
 624         public function getAutosummary( Content $oldContent = null, Content $newContent = null, $flags ) {
 625                 global $wgContLang;
 626
 627                 // Decide what kind of auto-summary is needed.
 628
 629                 // Redirect auto-summaries
 630
 631                 /**
 632                  * @var $ot Title
 633                  * @var $rt Title
 634                  */
 635
 636                 $ot = !is_null( $oldContent ) ? $oldContent->getRedirectTarget() : null;
 637                 $rt = !is_null( $newContent ) ? $newContent->getRedirectTarget() : null;
 638
 639                 if ( is_object( $rt ) ) {
 640                         if ( !is_object( $ot )
 641                                 || !$rt->equals( $ot )
 642                                 || $ot->getFragment() != $rt->getFragment() )
 643                         {
 644                                 $truncatedtext = $newContent->getTextForSummary(
 645                                         250
 646                                                 - strlen( wfMsgForContent( 'autoredircomment' ) )
 647                                                 - strlen( $rt->getFullText() ) );
 648
 649                                 return wfMsgForContent( 'autoredircomment', $rt->getFullText(), $truncatedtext );
 650                         }
 651                 }
 652
 653                 // New page auto-summaries
 654                 if ( $flags & EDIT_NEW && $newContent->getSize() > 0 ) {
 655                         // If they're making a new article, give its text, truncated, in
 656                         // the summary.
 657
 658                         $truncatedtext = $newContent->getTextForSummary(
 659                                 200 - strlen( wfMsgForContent( 'autosumm-new' ) ) );
 660
 661                         return wfMsgForContent( 'autosumm-new', $truncatedtext );
 662                 }
 663
 664                 // Blanking auto-summaries
 665                 if ( !empty( $oldContent ) && $oldContent->getSize() > 0 && $newContent->getSize() == 0 ) {
 666                         return wfMsgForContent( 'autosumm-blank' );
 667                 } elseif ( !empty( $oldContent )
 668                         && $oldContent->getSize() > 10 * $newContent->getSize()
 669                         && $newContent->getSize() < 500 )
 670                 {
 671                         // Removing more than 90% of the article
 672
 673                         $truncatedtext = $newContent->getTextForSummary(
 674                                 200 - strlen( wfMsgForContent( 'autosumm-replace' ) ) );
 675
 676                         return wfMsgForContent( 'autosumm-replace', $truncatedtext );
 677                 }
 678
 679                 // If we reach this point, there's no applicable auto-summary for our
 680                 // case, so our auto-summary is empty.
 681
 682                 return '';
 683         }
 684
 685         /**
 686          * Auto-generates a deletion reason
 687          *
 688          * @since WD.1
 689          *
 690          * @param $title Title: the page's title
 691          * @param &$hasHistory Boolean: whether the page has a history
 692          * @return mixed String containing deletion reason or empty string, or
 693          *    boolean false if no revision occurred
 694          *
 695          * @XXX &$hasHistory is extremely ugly, it's here because
 696          * WikiPage::getAutoDeleteReason() and Article::getReason()
 697          * have it / want it.
 698          */
 699         public function getAutoDeleteReason( Title $title, &$hasHistory ) {
 700                 $dbw = wfGetDB( DB_MASTER );
 701
 702                 // Get the last revision
 703                 $rev = Revision::newFromTitle( $title );
 704
 705                 if ( is_null( $rev ) ) {
 706                         return false;
 707                 }
 708
 709                 // Get the article's contents
 710                 $content = $rev->getContent();
 711                 $blank = false;
 712
 713                 $this->checkModelID( $content->getModel() );
 714
 715                 // If the page is blank, use the text from the previous revision,
 716                 // which can only be blank if there's a move/import/protect dummy
 717                 // revision involved
 718                 if ( $content->getSize() == 0 ) {
 719                         $prev = $rev->getPrevious();
 720
 721                         if ( $prev )    {
 722                                 $content = $prev->getContent();
 723                                 $blank = true;
 724                         }
 725                 }
 726
 727                 // Find out if there was only one contributor
 728                 // Only scan the last 20 revisions
 729                 $res = $dbw->select( 'revision', 'rev_user_text',
 730                         array(
 731                                 'rev_page' => $title->getArticleID(),
 732                                 $dbw->bitAnd( 'rev_deleted', Revision::DELETED_USER ) . ' = 0'
 733                         ),
 734                         __METHOD__,
 735                         array( 'LIMIT' => 20 )
 736                 );
 737
 738                 if ( $res === false ) {
 739                         // This page has no revisions, which is very weird
 740                         return false;
 741                 }
 742
 743                 $hasHistory = ( $res->numRows() > 1 );
 744                 $row = $dbw->fetchObject( $res );
 745
 746                 if ( $row ) { // $row is false if the only contributor is hidden
 747                         $onlyAuthor = $row->rev_user_text;
 748                         // Try to find a second contributor
 749                         foreach ( $res as $row ) {
 750                                 if ( $row->rev_user_text != $onlyAuthor ) { // Bug 22999
 751                                         $onlyAuthor = false;
 752                                         break;
 753                                 }
 754                         }
 755                 } else {
 756                         $onlyAuthor = false;
 757                 }
 758
 759                 // Generate the summary with a '$1' placeholder
 760                 if ( $blank ) {
 761                         // The current revision is blank and the one before is also
 762                         // blank. It's just not our lucky day
 763                         $reason = wfMsgForContent( 'exbeforeblank', '$1' );
 764                 } else {
 765                         if ( $onlyAuthor ) {
 766                                 $reason = wfMsgForContent( 'excontentauthor', '$1', $onlyAuthor );
 767                         } else {
 768                                 $reason = wfMsgForContent( 'excontent', '$1' );
 769                         }
 770                 }
 771
 772                 if ( $reason == '-' ) {
 773                         // Allow these UI messages to be blanked out cleanly
 774                         return '';
 775                 }
 776
 777                 // Max content length = max comment length - length of the comment (excl. $1)
 778                 $text = $content->getTextForSummary( 255 - ( strlen( $reason ) - 2 ) );
 779
 780                 // Now replace the '$1' placeholder
 781                 $reason = str_replace( '$1', $text, $reason );
 782
 783                 return $reason;
 784         }
 785
 786         /**
 787          * Parse the Content object and generate a ParserOutput from the result.
 788          * $result->getText() can be used to obtain the generated HTML. If no HTML
 789          * is needed, $generateHtml can be set to false; in that case,
 790          * $result->getText() may return null.
 791          *
 792          * @param $content Content the content to render
 793          * @param $title Title The page title to use as a context for rendering
 794          * @param $revId null|int The revision being rendered (optional)
 795          * @param $options null|ParserOptions Any parser options
 796          * @param $generateHtml Boolean Whether to generate HTML (default: true). If false,
 797          *        the result of calling getText() on the ParserOutput object returned by
 798          *        this method is undefined.
 799          *
 800          * @since WD.1
 801          *
 802          * @return ParserOutput
 803          */
 804         public abstract function getParserOutput( Content $content, Title $title,
 805                 $revId = null,
 806                 ParserOptions $options = null, $generateHtml = true );
 807         # TODO: make RenderOutput and RenderOptions base classes
 808
 809         /**
 810          * Returns a list of DataUpdate objects for recording information about this
 811          * Content in some secondary data store. If the optional second argument,
 812          * $old, is given, the updates may model only the changes that need to be
 813          * made to replace information about the old content with information about
 814          * the new content.
 815          *
 816          * This default implementation calls
 817          * $this->getParserOutput( $content, $title, null, null, false ),
 818          * and then calls getSecondaryDataUpdates( $title, $recursive ) on the
 819          * resulting ParserOutput object.
 820          *
 821          * Subclasses may implement this to determine the necessary updates more
 822          * efficiently, or make use of information about the old content.
 823          *
 824          * @param $content Content The content for determining the necessary updates
 825          * @param $title Title The context for determining the necessary updates
 826          * @param $old Content|null An optional Content object representing the
 827          *    previous content, i.e. the content being replaced by this Content
 828          *    object.
 829          * @param $recursive boolean Whether to include recursive updates (default:
 830          *    false).
 831          * @param $parserOutput ParserOutput|null Optional ParserOutput object.
 832          *    Provide if you have one handy, to avoid re-parsing of the content.
 833          *
 834          * @return Array. A list of DataUpdate objects for putting information
 835          *    about this content object somewhere.
 836          *
 837          * @since WD.1
 838          */
 839         public function getSecondaryDataUpdates( Content $content, Title $title,
 840                 Content $old = null,
 841                 $recursive = true, ParserOutput $parserOutput = null
 842         ) {
 843                 if ( !$parserOutput ) {
 844                         $parserOutput = $this->getParserOutput( $content, $title, null, null, false );
 845                 }
 846
 847                 return $parserOutput->getSecondaryDataUpdates( $title, $recursive );
 848         }
 849
 850
 851         /**
 852          * Get the Content object that needs to be saved in order to undo all revisions
 853          * between $undo and $undoafter. Revisions must belong to the same page,
 854          * must exist and must not be deleted.
 855          *
 856          * @since WD.1
 857          *
 858          * @param $current Revision The current text
 859          * @param $undo Revision The revision to undo
 860          * @param $undoafter Revision Must be an earlier revision than $undo
 861          *
 862          * @return mixed String on success, false on failure
 863          */
 864         public function getUndoContent( Revision $current, Revision $undo, Revision $undoafter ) {
 865                 $cur_content = $current->getContent();
 866
 867                 if ( empty( $cur_content ) ) {
 868                         return false; // no page
 869                 }
 870
 871                 $undo_content = $undo->getContent();
 872                 $undoafter_content = $undoafter->getContent();
 873
 874                 $this->checkModelID( $cur_content->getModel() );
 875                 $this->checkModelID( $undo_content->getModel() );
 876                 $this->checkModelID( $undoafter_content->getModel() );
 877
 878                 if ( $cur_content->equals( $undo_content ) ) {
 879                         // No use doing a merge if it's just a straight revert.
 880                         return $undoafter_content;
 881                 }
 882
 883                 $undone_content = $this->merge3( $undo_content, $undoafter_content, $cur_content );
 884
 885                 return $undone_content;
 886         }
 887
 888         /**
 889          * Returns true for content models that support caching using the
 890          * ParserCache mechanism. See WikiPage::isParserCacheUser().
 891          *
 892          * @since WD.1
 893          *
 894          * @return bool
 895          */
 896         public function isParserCacheSupported() {
 897                 return true;
 898         }
 899
 900         /**
 901          * Returns a list of updates to perform when the given content is deleted.
 902          * The necessary updates may be taken from the Content object, or depend on
 903          * the current state of the database.
 904          *
 905          * @since WD.1
 906          *
 907          * @param $content \Content the Content object for deletion
 908          * @param $title \Title the title of the deleted page
 909          * @param $parserOutput null|\ParserOutput optional parser output object
 910          *    for efficient access to meta-information about the content object.
 911          *    Provide if you have one handy.
 912          *
 913          * @return array A list of DataUpdate instances that will clean up the
 914          *    database after deletion.
 915          */
 916         public function getDeletionUpdates( Content $content, Title $title,
 917                 ParserOutput $parserOutput = null )
 918         {
 919                 return array(
 920                         new LinksDeletionUpdate( $title ),
 921                 );
 922         }
 923
 924         /**
 925          * Returns true if this content model supports sections.
 926          *
 927          * This default implementation returns false.
 928          *
 929          * @return boolean whether sections are supported.
 930          */
 931         public function supportsSections() {
 932                 return false;
 933         }
 934 }
 935
 936 /**
 937  * @since WD.1
 938  */
 939 abstract class TextContentHandler extends ContentHandler {
 940
 941         public function __construct( $modelId, $formats ) {
 942                 parent::__construct( $modelId, $formats );
 943         }
 944
 945         /**
 946          * Returns the content's text as-is.
 947          *
 948          * @param $content Content
 949          * @param $format string|null
 950          * @return mixed
 951          */
 952         public function serializeContent( Content $content, $format = null ) {
 953                 $this->checkFormat( $format );
 954                 return $content->getNativeData();
 955         }
 956
 957         /**
 958          * Attempts to merge differences between three versions. Returns a new
 959          * Content object for a clean merge and false for failure or a conflict.
 960          *
 961          * All three Content objects passed as parameters must have the same
 962          * content model.
 963          *
 964          * This text-based implementation uses wfMerge().
 965          *
 966          * @param $oldContent \Content|string  String
 967          * @param $myContent \Content|string   String
 968          * @param $yourContent \Content|string String
 969          *
 970          * @return Content|Bool
 971          */
 972         public function merge3( Content $oldContent, Content $myContent, Content $yourContent ) {
 973                 $this->checkModelID( $oldContent->getModel() );
 974                 $this->checkModelID( $myContent->getModel() );
 975                 $this->checkModelID( $yourContent->getModel() );
 976
 977                 $format = $this->getDefaultFormat();
 978
 979                 $old = $this->serializeContent( $oldContent, $format );
 980                 $mine = $this->serializeContent( $myContent, $format );
 981                 $yours = $this->serializeContent( $yourContent, $format );
 982
 983                 $ok = wfMerge( $old, $mine, $yours, $result );
 984
 985                 if ( !$ok ) {
 986                         return false;
 987                 }
 988
 989                 if ( !$result ) {
 990                         return $this->makeEmptyContent();
 991                 }
 992
 993                 $mergedContent = $this->unserializeContent( $result, $format );
 994                 return $mergedContent;
 995         }
 996
 997         /**
 998          * Returns a generic ParserOutput object, wrapping the HTML returned by
 999          * getHtml().
1000          *
1001          * @param $content Content The content to render
1002          * @param $title Title Context title for parsing
1003          * @param $revId int|null Revision ID (for {{REVISIONID}})
1004          * @param $options ParserOptions|null Parser options
1005          * @param $generateHtml bool Whether or not to generate HTML
1006          *
1007          * @return ParserOutput representing the HTML form of the text
1008          */
1009         public function getParserOutput( Content $content, Title $title,
1010                 $revId = null,
1011                 ParserOptions $options = null, $generateHtml = true
1012         ) {
1013                 $this->checkModelID( $content->getModel() );
1014
1015                 # Generic implementation, relying on $this->getHtml()
1016
1017                 if ( $generateHtml ) {
1018                         $html = $this->getHtml( $content );
1019                 } else {
1020                         $html = '';
1021                 }
1022
1023                 $po = new ParserOutput( $html );
1024                 return $po;
1025         }
1026
1027         /**
1028          * Generates an HTML version of the content, for display. Used by
1029          * getParserOutput() to construct a ParserOutput object.
1030          *
1031          * This default implementation just calls getHighlightHtml(). Content
1032          * models that have another mapping to HTML (as is the case for markup
1033          * languages like wikitext) should override this method to generate the
1034          * appropriate HTML.
1035          *
1036          * @param $content Content The content to render
1037          *
1038          * @return string An HTML representation of the content
1039          */
1040         protected function getHtml( Content $content ) {
1041                 $this->checkModelID( $content->getModel() );
1042
1043                 return $this->getHighlightHtml( $content );
1044         }
1045
1046         /**
1047          * Generates a syntax-highlighted version the content, as HTML.
1048          * Used by the default implementation of getHtml().
1049          *
1050          * @param $content Content the content to render
1051          *
1052          * @return string an HTML representation of the content's markup
1053          */
1054         protected function getHighlightHtml( Content $content ) {
1055                 $this->checkModelID( $content->getModel() );
1056
1057                 # TODO: make Highlighter interface, use highlighter here, if available
1058                 return htmlspecialchars( $content->getNativeData() );
1059         }
1060
1061
1062 }
1063
1064 /**
1065  * @since WD.1
1066  */
1067 class WikitextContentHandler extends TextContentHandler {
1068
1069         public function __construct( $modelId = CONTENT_MODEL_WIKITEXT ) {
1070                 parent::__construct( $modelId, array( CONTENT_FORMAT_WIKITEXT ) );
1071         }
1072
1073         public function unserializeContent( $text, $format = null ) {
1074                 $this->checkFormat( $format );
1075
1076                 return new WikitextContent( $text );
1077         }
1078
1079         public function makeEmptyContent() {
1080                 return new WikitextContent( '' );
1081         }
1082
1083         /**
1084          * Returns a ParserOutput object resulting from parsing the content's text
1085          * using $wgParser.
1086          *
1087          * @since    WD.1
1088          *
1089          * @param $content Content the content to render
1090          * @param $title \Title
1091          * @param $revId null
1092          * @param $options null|ParserOptions
1093          * @param $generateHtml bool
1094          *
1095          * @internal param \IContextSource|null $context
1096          * @return ParserOutput representing the HTML form of the text
1097          */
1098         public function getParserOutput( Content $content, Title $title,
1099                 $revId = null,
1100                 ParserOptions $options = null, $generateHtml = true
1101         ) {
1102                 global $wgParser;
1103
1104                 $this->checkModelID( $content->getModel() );
1105
1106                 if ( !$options ) {
1107                         $options = new ParserOptions();
1108                 }
1109
1110                 $po = $wgParser->parse( $content->getNativeData(), $title, $options, true, true, $revId );
1111                 return $po;
1112         }
1113
1114         protected function getHtml( Content $content ) {
1115                 throw new MWException(
1116                         "getHtml() not implemented for wikitext. "
1117                         . "Use getParserOutput()->getText()."
1118                 );
1119         }
1120
1121         /**
1122          * Returns true because wikitext supports sections.
1123          *
1124          * @return boolean whether sections are supported.
1125          */
1126         public function supportsSections() {
1127                 return true;
1128         }
1129 }
1130
1131 # XXX: make ScriptContentHandler base class, do highlighting stuff there?
1132
1133 /**
1134  * @since WD.1
1135  */
1136 class JavaScriptContentHandler extends TextContentHandler {
1137
1138         public function __construct( $modelId = CONTENT_MODEL_JAVASCRIPT ) {
1139                 parent::__construct( $modelId, array( CONTENT_FORMAT_JAVASCRIPT ) );
1140         }
1141
1142         public function unserializeContent( $text, $format = null ) {
1143                 $this->checkFormat( $format );
1144
1145                 return new JavaScriptContent( $text );
1146         }
1147
1148         public function makeEmptyContent() {
1149                 return new JavaScriptContent( '' );
1150         }
1151
1152         /**
1153          * Returns the english language, because JS is english, and should be handled as such.
1154          *
1155          * @return Language wfGetLangObj( 'en' )
1156          *
1157          * @see ContentHandler::getPageLanguage()
1158          */
1159         public function getPageLanguage( Title $title, Content $content = null ) {
1160                 return wfGetLangObj( 'en' );
1161         }
1162
1163         protected function getHtml( Content $content ) {
1164                 $html = "";
1165                 $html .= "<pre class=\"mw-code mw-js\" dir=\"ltr\">\n";
1166                 $html .= $this->getHighlightHtml( $content );
1167                 $html .= "\n</pre>\n";
1168
1169                 return $html;
1170         }
1171 }
1172
1173 /**
1174  * @since WD.1
1175  */
1176 class CssContentHandler extends TextContentHandler {
1177
1178         public function __construct( $modelId = CONTENT_MODEL_CSS ) {
1179                 parent::__construct( $modelId, array( CONTENT_FORMAT_CSS ) );
1180         }
1181
1182         public function unserializeContent( $text, $format = null ) {
1183                 $this->checkFormat( $format );
1184
1185                 return new CssContent( $text );
1186         }
1187
1188         public function makeEmptyContent() {
1189                 return new CssContent( '' );
1190         }
1191
1192         /**
1193          * Returns the english language, because CSS is english, and should be handled as such.
1194          *
1195          * @return Language wfGetLangObj( 'en' )
1196          *
1197          * @see ContentHandler::getPageLanguage()
1198          */
1199         public function getPageLanguage( Title $title, Content $content = null ) {
1200                 return wfGetLangObj( 'en' );
1201         }
1202
1203         protected function getHtml( Content $content ) {
1204                 $html = "";
1205                 $html .= "<pre class=\"mw-code mw-css\" dir=\"ltr\">\n";
1206                 $html .= $this->getHighlightHtml( $content );
1207                 $html .= "\n</pre>\n";
1208
1209                 return $html;
1210         }
1211 }