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 int 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 appropriate MIME type for a given content format,
 315          * or null if no MIME type is known for this format.
 316          *
 317          * MIME types can be registered in the global array $wgContentFormatMimeTypes.
 318          *
 319          * @static
 320          * @param $id int The content format id, as given by a CONTENT_FORMAT_XXX
 321          *    constant or returned by Revision::getContentFormat().
 322          *
 323          * @return string|null The content format's MIME type.
 324          */
 325         public static function getContentFormatMimeType( $id ) {
 326                 global $wgContentFormatMimeTypes;
 327
 328                 if ( !isset( $wgContentFormatMimeTypes[ $id ] ) ) {
 329                         return null;
 330                 }
 331
 332                 return $wgContentFormatMimeTypes[ $id ];
 333         }
 334
 335         /**
 336          * Returns the content format if for a given MIME type,
 337          * or null if no format ID if known for this MIME type.
 338          *
 339          * Mime types can be registered in the global array $wgContentFormatMimeTypes.
 340          *
 341          * @static
 342          * @param $mime string the MIME type
 343          *
 344          * @return int|null The format ID, as defined by a CONTENT_FORMAT_XXX constant
 345          */
 346         public static function getContentFormatID( $mime ) {
 347                 global $wgContentFormatMimeTypes;
 348
 349                 static $format_ids = null;
 350
 351                 if ( $format_ids === null ) {
 352                         $format_ids = array_flip( $wgContentFormatMimeTypes );
 353                 }
 354
 355                 if ( !isset( $format_ids[ $mime ] ) ) {
 356                         return null;
 357                 }
 358
 359                 return $format_ids[ $mime ];
 360         }
 361
 362         /**
 363          * Returns the symbolic name for a given content model.
 364          *
 365          * @param $id int The content model ID, as given by a CONTENT_MODEL_XXX
 366          *    constant or returned by Revision::getContentModel().
 367          *
 368          * @return string The content model's symbolic name.
 369          * @throws MWException if the model id isn't known.
 370          */
 371         public static function getContentModelName( $id ) {
 372                 $handler = self::getForModelID( $id );
 373                 return $handler->getModelName();
 374         }
 375
 376
 377         /**
 378          * Returns the localized name for a given content model.
 379          *
 380          * Model names are localized using system messages. Message keys
 381          * have the form content-model-$name, where $name is getContentModelName( $id ).
 382          *
 383          * @static
 384          * @param $id int The content model ID, as given by a CONTENT_MODEL_XXX
 385          *    constant or returned by Revision::getContentModel().
 386          * @todo also accept a symbolic name instead of a numeric id
 387          *
 388          * @return string The content format's localized name.
 389          * @throws MWException if the model id isn't known.
 390          */
 391         public static function getLocalizedName( $id ) {
 392                 $name = self::getContentModelName( $id );
 393                 $key = "content-model-$name";
 394
 395                 if ( wfEmptyMsg( $key ) ) return $name;
 396                 else return wfMsg( $key );
 397         }
 398
 399         // ------------------------------------------------------------------------
 400
 401         protected $mModelID;
 402         protected $mModelName;
 403         protected $mSupportedFormats;
 404
 405         /**
 406          * Constructor, initializing the ContentHandler instance with its model ID
 407          * and a list of supported formats. Values for the parameters are typically
 408          * provided as literals by subclass's constructors.
 409          *
 410          * @param $modelId int (use CONTENT_MODEL_XXX constants).
 411          * @param $formats array List for supported serialization formats
 412          *    (typically as MIME types)
 413          */
 414         public function __construct( $modelId, $formats ) {
 415                 $this->mModelID = $modelId;
 416                 $this->mSupportedFormats = $formats;
 417
 418                 $this->mModelName = preg_replace( '/(Content)?Handler$/', '', get_class( $this ) );
 419                 $this->mModelName = preg_replace( '/[_\\\\]/', '', $this->mModelName );
 420                 $this->mModelName = strtolower( $this->mModelName );
 421         }
 422
 423         /**
 424          * Serializes a Content object of the type supported by this ContentHandler.
 425          *
 426          * @since WD.1
 427          *
 428          * @abstract
 429          * @param $content Content The Content object to serialize
 430          * @param $format null The desired serialization format
 431          * @return string Serialized form of the content
 432          */
 433         public abstract function serializeContent( Content $content, $format = null );
 434
 435         /**
 436          * Unserializes a Content object of the type supported by this ContentHandler.
 437          *
 438          * @since WD.1
 439          *
 440          * @abstract
 441          * @param $blob string serialized form of the content
 442          * @param $format null the format used for serialization
 443          * @return Content the Content object created by deserializing $blob
 444          */
 445         public abstract function unserializeContent( $blob, $format = null );
 446
 447         /**
 448          * Creates an empty Content object of the type supported by this
 449          * ContentHandler.
 450          *
 451          * @since WD.1
 452          *
 453          * @return Content
 454          */
 455         public abstract function makeEmptyContent();
 456
 457         /**
 458          * Returns the model id that identifies the content model this
 459          * ContentHandler can handle. Use with the CONTENT_MODEL_XXX constants.
 460          *
 461          * @since WD.1
 462          *
 463          * @return int The model ID
 464          */
 465         public function getModelID() {
 466                 return $this->mModelID;
 467         }
 468
 469         /**
 470          * Returns the content model's symbolic name.
 471          *
 472          * The symbolic name is is this object's class name in lower case with the trailing "ContentHandler"
 473          * and and special characters removed.
 474          *
 475          * @since WD.1
 476          *
 477          * @return String The content model's name
 478          */
 479         public function getModelName() {
 480                 return $this->mModelName;
 481         }
 482
 483         /**
 484          * Throws an MWException if $model_id is not the ID of the content model
 485          * supported by this ContentHandler.
 486          *
 487          * @since WD.1
 488          *
 489          * @param $model_id int The model to check
 490          *
 491          * @throws MWException
 492          */
 493         protected function checkModelID( $model_id ) {
 494                 if ( $model_id !== $this->mModelID ) {
 495                         $model_name = ContentHandler::getContentModelName( $model_id );
 496                         $own_model_name = ContentHandler::getContentModelName( $this->mModelID );
 497
 498                         throw new MWException( "Bad content model: " .
 499                                 "expected {$this->mModelID} ($own_model_name) " .
 500                                 "but got $model_id ($model_name)." );
 501                 }
 502         }
 503
 504         /**
 505          * Returns a list of serialization formats supported by the
 506          * serializeContent() and unserializeContent() methods of this
 507          * ContentHandler.
 508          *
 509          * @since WD.1
 510          *
 511          * @return array of serialization formats as MIME type like strings
 512          */
 513         public function getSupportedFormats() {
 514                 return $this->mSupportedFormats;
 515         }
 516
 517         /**
 518          * The format used for serialization/deserialization by default by this
 519          * ContentHandler.
 520          *
 521          * This default implementation will return the first element of the array
 522          * of formats that was passed to the constructor.
 523          *
 524          * @since WD.1
 525          *
 526          * @return string the name of the default serialization format as a MIME type
 527          */
 528         public function getDefaultFormat() {
 529                 return $this->mSupportedFormats[0];
 530         }
 531
 532         /**
 533          * Returns true if $format is a serialization format supported by this
 534          * ContentHandler, and false otherwise.
 535          *
 536          * Note that if $format is null, this method always returns true, because
 537          * null means "use the default format".
 538          *
 539          * @since WD.1
 540          *
 541          * @param $format string the serialization format to check
 542          * @return bool
 543          */
 544         public function isSupportedFormat( $format ) {
 545
 546                 if ( !$format ) {
 547                         return true; // this means "use the default"
 548                 }
 549
 550                 return in_array( $format, $this->mSupportedFormats );
 551         }
 552
 553         /**
 554          * Throws an MWException if isSupportedFormat( $format ) is not true.
 555          * Convenient for checking whether a format provided as a parameter is
 556          * actually supported.
 557          *
 558          * @param $format string the serialization format to check
 559          *
 560          * @throws MWException
 561          */
 562         protected function checkFormat( $format ) {
 563                 if ( !$this->isSupportedFormat( $format ) ) {
 564                         throw new MWException(
 565                                 "Format $format is not supported for content model "
 566                                 . $this->getModelID()
 567                         );
 568                 }
 569         }
 570
 571         /**
 572          * Returns true if the content is consistent with the database, that is if
 573          * saving it to the database would not violate any global constraints.
 574          *
 575          * Content needs to be valid using this method before it can be saved.
 576          *
 577          * This default implementation always returns true.
 578          *
 579          * @since WD.1
 580          *
 581          * @param $content \Content
 582          *
 583          * @return boolean
 584          */
 585         public function isConsistentWithDatabase( Content $content ) {
 586                 return true;
 587         }
 588
 589         /**
 590          * Returns overrides for action handlers.
 591          * Classes listed here will be used instead of the default one when
 592          * (and only when) $wgActions[$action] === true. This allows subclasses
 593          * to override the default action handlers.
 594          *
 595          * @since WD.1
 596          *
 597          * @return Array
 598          */
 599         public function getActionOverrides() {
 600                 return array();
 601         }
 602
 603         /**
 604          * Factory for creating an appropriate DifferenceEngine for this content model.
 605          *
 606          * @since WD.1
 607          *
 608          * @param $context IContextSource context to use, anything else will be
 609          *    ignored
 610          * @param $old Integer Old ID we want to show and diff with.
 611          * @param $new int|string String either 'prev' or 'next'.
 612          * @param $rcid Integer ??? FIXME (default 0)
 613          * @param $refreshCache boolean If set, refreshes the diff cache
 614          * @param $unhide boolean If set, allow viewing deleted revs
 615          *
 616          * @return DifferenceEngine
 617          */
 618         public function createDifferenceEngine( IContextSource $context,
 619                 $old = 0, $new = 0,
 620                 $rcid = 0, # FIXME: use everywhere!
 621                 $refreshCache = false, $unhide = false
 622         ) {
 623                 $this->checkModelID( $context->getTitle()->getContentModel() );
 624
 625                 $diffEngineClass = $this->getDiffEngineClass();
 626
 627                 return new $diffEngineClass( $context, $old, $new, $rcid, $refreshCache, $unhide );
 628         }
 629
 630         /**
 631          * Returns the name of the diff engine to use.
 632          *
 633          * @since WD.1
 634          *
 635          * @return string
 636          */
 637         protected function getDiffEngineClass() {
 638                 return 'DifferenceEngine';
 639         }
 640
 641         /**
 642          * Attempts to merge differences between three versions.
 643          * Returns a new Content object for a clean merge and false for failure or
 644          * a conflict.
 645          *
 646          * This default implementation always returns false.
 647          *
 648          * @since WD.1
 649          *
 650          * @param $oldContent Content|string  String
 651          * @param $myContent Content|string   String
 652          * @param $yourContent Content|string String
 653          *
 654          * @return Content|Bool
 655          */
 656         public function merge3( Content $oldContent, Content $myContent, Content $yourContent ) {
 657                 return false;
 658         }
 659
 660         /**
 661          * Return an applicable auto-summary if one exists for the given edit.
 662          *
 663          * @since WD.1
 664          *
 665          * @param $oldContent Content|null: the previous text of the page.
 666          * @param $newContent Content|null: The submitted text of the page.
 667          * @param $flags int Bit mask: a bit mask of flags submitted for the edit.
 668          *
 669          * @return string An appropriate auto-summary, or an empty string.
 670          */
 671         public function getAutosummary( Content $oldContent = null, Content $newContent = null, $flags ) {
 672                 global $wgContLang;
 673
 674                 // Decide what kind of auto-summary is needed.
 675
 676                 // Redirect auto-summaries
 677
 678                 /**
 679                  * @var $ot Title
 680                  * @var $rt Title
 681                  */
 682
 683                 $ot = !is_null( $oldContent ) ? $oldContent->getRedirectTarget() : null;
 684                 $rt = !is_null( $newContent ) ? $newContent->getRedirectTarget() : null;
 685
 686                 if ( is_object( $rt ) ) {
 687                         if ( !is_object( $ot )
 688                                 || !$rt->equals( $ot )
 689                                 || $ot->getFragment() != $rt->getFragment() )
 690                         {
 691                                 $truncatedtext = $newContent->getTextForSummary(
 692                                         250
 693                                                 - strlen( wfMsgForContent( 'autoredircomment' ) )
 694                                                 - strlen( $rt->getFullText() ) );
 695
 696                                 return wfMsgForContent( 'autoredircomment', $rt->getFullText(), $truncatedtext );
 697                         }
 698                 }
 699
 700                 // New page auto-summaries
 701                 if ( $flags & EDIT_NEW && $newContent->getSize() > 0 ) {
 702                         // If they're making a new article, give its text, truncated, in
 703                         // the summary.
 704
 705                         $truncatedtext = $newContent->getTextForSummary(
 706                                 200 - strlen( wfMsgForContent( 'autosumm-new' ) ) );
 707
 708                         return wfMsgForContent( 'autosumm-new', $truncatedtext );
 709                 }
 710
 711                 // Blanking auto-summaries
 712                 if ( !empty( $oldContent ) && $oldContent->getSize() > 0 && $newContent->getSize() == 0 ) {
 713                         return wfMsgForContent( 'autosumm-blank' );
 714                 } elseif ( !empty( $oldContent )
 715                         && $oldContent->getSize() > 10 * $newContent->getSize()
 716                         && $newContent->getSize() < 500 )
 717                 {
 718                         // Removing more than 90% of the article
 719
 720                         $truncatedtext = $newContent->getTextForSummary(
 721                                 200 - strlen( wfMsgForContent( 'autosumm-replace' ) ) );
 722
 723                         return wfMsgForContent( 'autosumm-replace', $truncatedtext );
 724                 }
 725
 726                 // If we reach this point, there's no applicable auto-summary for our
 727                 // case, so our auto-summary is empty.
 728
 729                 return '';
 730         }
 731
 732         /**
 733          * Auto-generates a deletion reason
 734          *
 735          * @since WD.1
 736          *
 737          * @param $title Title: the page's title
 738          * @param &$hasHistory Boolean: whether the page has a history
 739          * @return mixed String containing deletion reason or empty string, or
 740          *    boolean false if no revision occurred
 741          *
 742          * @XXX &$hasHistory is extremely ugly, it's here because
 743          * WikiPage::getAutoDeleteReason() and Article::getReason()
 744          * have it / want it.
 745          */
 746         public function getAutoDeleteReason( Title $title, &$hasHistory ) {
 747                 $dbw = wfGetDB( DB_MASTER );
 748
 749                 // Get the last revision
 750                 $rev = Revision::newFromTitle( $title );
 751
 752                 if ( is_null( $rev ) ) {
 753                         return false;
 754                 }
 755
 756                 // Get the article's contents
 757                 $content = $rev->getContent();
 758                 $blank = false;
 759
 760                 $this->checkModelID( $content->getModel() );
 761
 762                 // If the page is blank, use the text from the previous revision,
 763                 // which can only be blank if there's a move/import/protect dummy
 764                 // revision involved
 765                 if ( $content->getSize() == 0 ) {
 766                         $prev = $rev->getPrevious();
 767
 768                         if ( $prev )    {
 769                                 $content = $prev->getContent();
 770                                 $blank = true;
 771                         }
 772                 }
 773
 774                 // Find out if there was only one contributor
 775                 // Only scan the last 20 revisions
 776                 $res = $dbw->select( 'revision', 'rev_user_text',
 777                         array(
 778                                 'rev_page' => $title->getArticleID(),
 779                                 $dbw->bitAnd( 'rev_deleted', Revision::DELETED_USER ) . ' = 0'
 780                         ),
 781                         __METHOD__,
 782                         array( 'LIMIT' => 20 )
 783                 );
 784
 785                 if ( $res === false ) {
 786                         // This page has no revisions, which is very weird
 787                         return false;
 788                 }
 789
 790                 $hasHistory = ( $res->numRows() > 1 );
 791                 $row = $dbw->fetchObject( $res );
 792
 793                 if ( $row ) { // $row is false if the only contributor is hidden
 794                         $onlyAuthor = $row->rev_user_text;
 795                         // Try to find a second contributor
 796                         foreach ( $res as $row ) {
 797                                 if ( $row->rev_user_text != $onlyAuthor ) { // Bug 22999
 798                                         $onlyAuthor = false;
 799                                         break;
 800                                 }
 801                         }
 802                 } else {
 803                         $onlyAuthor = false;
 804                 }
 805
 806                 // Generate the summary with a '$1' placeholder
 807                 if ( $blank ) {
 808                         // The current revision is blank and the one before is also
 809                         // blank. It's just not our lucky day
 810                         $reason = wfMsgForContent( 'exbeforeblank', '$1' );
 811                 } else {
 812                         if ( $onlyAuthor ) {
 813                                 $reason = wfMsgForContent( 'excontentauthor', '$1', $onlyAuthor );
 814                         } else {
 815                                 $reason = wfMsgForContent( 'excontent', '$1' );
 816                         }
 817                 }
 818
 819                 if ( $reason == '-' ) {
 820                         // Allow these UI messages to be blanked out cleanly
 821                         return '';
 822                 }
 823
 824                 // Max content length = max comment length - length of the comment (excl. $1)
 825                 $text = $content->getTextForSummary( 255 - ( strlen( $reason ) - 2 ) );
 826
 827                 // Now replace the '$1' placeholder
 828                 $reason = str_replace( '$1', $text, $reason );
 829
 830                 return $reason;
 831         }
 832
 833         /**
 834          * Parse the Content object and generate a ParserOutput from the result.
 835          * $result->getText() can be used to obtain the generated HTML. If no HTML
 836          * is needed, $generateHtml can be set to false; in that case,
 837          * $result->getText() may return null.
 838          *
 839          * @param $content Content the content to render
 840          * @param $title Title The page title to use as a context for rendering
 841          * @param $revId null|int The revision being rendered (optional)
 842          * @param $options null|ParserOptions Any parser options
 843          * @param $generateHtml Boolean Whether to generate HTML (default: true). If false,
 844          *        the result of calling getText() on the ParserOutput object returned by
 845          *        this method is undefined.
 846          *
 847          * @since WD.1
 848          *
 849          * @return ParserOutput
 850          */
 851         public abstract function getParserOutput( Content $content, Title $title,
 852                 $revId = null,
 853                 ParserOptions $options = null, $generateHtml = true );
 854         # TODO: make RenderOutput and RenderOptions base classes
 855
 856         /**
 857          * Returns a list of DataUpdate objects for recording information about this
 858          * Content in some secondary data store. If the optional second argument,
 859          * $old, is given, the updates may model only the changes that need to be
 860          * made to replace information about the old content with information about
 861          * the new content.
 862          *
 863          * This default implementation calls
 864          * $this->getParserOutput( $content, $title, null, null, false ),
 865          * and then calls getSecondaryDataUpdates( $title, $recursive ) on the
 866          * resulting ParserOutput object.
 867          *
 868          * Subclasses may implement this to determine the necessary updates more
 869          * efficiently, or make use of information about the old content.
 870          *
 871          * @param $content Content The content for determining the necessary updates
 872          * @param $title Title The context for determining the necessary updates
 873          * @param $old Content|null An optional Content object representing the
 874          *    previous content, i.e. the content being replaced by this Content
 875          *    object.
 876          * @param $recursive boolean Whether to include recursive updates (default:
 877          *    false).
 878          * @param $parserOutput ParserOutput|null Optional ParserOutput object.
 879          *    Provide if you have one handy, to avoid re-parsing of the content.
 880          *
 881          * @return Array. A list of DataUpdate objects for putting information
 882          *    about this content object somewhere.
 883          *
 884          * @since WD.1
 885          */
 886         public function getSecondaryDataUpdates( Content $content, Title $title,
 887                 Content $old = null,
 888                 $recursive = true, ParserOutput $parserOutput = null
 889         ) {
 890                 if ( !$parserOutput ) {
 891                         $parserOutput = $this->getParserOutput( $content, $title, null, null, false );
 892                 }
 893
 894                 return $parserOutput->getSecondaryDataUpdates( $title, $recursive );
 895         }
 896
 897
 898         /**
 899          * Get the Content object that needs to be saved in order to undo all revisions
 900          * between $undo and $undoafter. Revisions must belong to the same page,
 901          * must exist and must not be deleted.
 902          *
 903          * @since WD.1
 904          *
 905          * @param $current Revision The current text
 906          * @param $undo Revision The revision to undo
 907          * @param $undoafter Revision Must be an earlier revision than $undo
 908          *
 909          * @return mixed String on success, false on failure
 910          */
 911         public function getUndoContent( Revision $current, Revision $undo, Revision $undoafter ) {
 912                 $cur_content = $current->getContent();
 913
 914                 if ( empty( $cur_content ) ) {
 915                         return false; // no page
 916                 }
 917
 918                 $undo_content = $undo->getContent();
 919                 $undoafter_content = $undoafter->getContent();
 920
 921                 $this->checkModelID( $cur_content->getModel() );
 922                 $this->checkModelID( $undo_content->getModel() );
 923                 $this->checkModelID( $undoafter_content->getModel() );
 924
 925                 if ( $cur_content->equals( $undo_content ) ) {
 926                         // No use doing a merge if it's just a straight revert.
 927                         return $undoafter_content;
 928                 }
 929
 930                 $undone_content = $this->merge3( $undo_content, $undoafter_content, $cur_content );
 931
 932                 return $undone_content;
 933         }
 934
 935         /**
 936          * Returns true for content models that support caching using the
 937          * ParserCache mechanism. See WikiPage::isParserCacheUser().
 938          *
 939          * @since WD.1
 940          *
 941          * @return bool
 942          */
 943         public function isParserCacheSupported() {
 944                 return true;
 945         }
 946
 947         /**
 948          * Returns a list of updates to perform when the given content is deleted.
 949          * The necessary updates may be taken from the Content object, or depend on
 950          * the current state of the database.
 951          *
 952          * @since WD.1
 953          *
 954          * @param $content \Content the Content object for deletion
 955          * @param $title \Title the title of the deleted page
 956          * @param $parserOutput null|\ParserOutput optional parser output object
 957          *    for efficient access to meta-information about the content object.
 958          *    Provide if you have one handy.
 959          *
 960          * @return array A list of DataUpdate instances that will clean up the
 961          *    database after deletion.
 962          */
 963         public function getDeletionUpdates( Content $content, Title $title,
 964                 ParserOutput $parserOutput = null )
 965         {
 966                 return array(
 967                         new LinksDeletionUpdate( $title ),
 968                 );
 969         }
 970
 971         /**
 972          * Returns true if this content model supports sections.
 973          *
 974          * This default implementation returns false.
 975          *
 976          * @return boolean whether sections are supported.
 977          */
 978         public function supportsSections() {
 979                 return false;
 980         }
 981 }
 982
 983 /**
 984  * @since WD.1
 985  */
 986 abstract class TextContentHandler extends ContentHandler {
 987
 988         public function __construct( $modelId, $formats ) {
 989                 parent::__construct( $modelId, $formats );
 990         }
 991
 992         /**
 993          * Returns the content's text as-is.
 994          *
 995          * @param $content Content
 996          * @param $format string|null
 997          * @return mixed
 998          */
 999         public function serializeContent( Content $content, $format = null ) {
1000                 $this->checkFormat( $format );
1001                 return $content->getNativeData();
1002         }
1003
1004         /**
1005          * Attempts to merge differences between three versions. Returns a new
1006          * Content object for a clean merge and false for failure or a conflict.
1007          *
1008          * All three Content objects passed as parameters must have the same
1009          * content model.
1010          *
1011          * This text-based implementation uses wfMerge().
1012          *
1013          * @param $oldContent \Content|string  String
1014          * @param $myContent \Content|string   String
1015          * @param $yourContent \Content|string String
1016          *
1017          * @return Content|Bool
1018          */
1019         public function merge3( Content $oldContent, Content $myContent, Content $yourContent ) {
1020                 $this->checkModelID( $oldContent->getModel() );
1021                 $this->checkModelID( $myContent->getModel() );
1022                 $this->checkModelID( $yourContent->getModel() );
1023
1024                 $format = $this->getDefaultFormat();
1025
1026                 $old = $this->serializeContent( $oldContent, $format );
1027                 $mine = $this->serializeContent( $myContent, $format );
1028                 $yours = $this->serializeContent( $yourContent, $format );
1029
1030                 $ok = wfMerge( $old, $mine, $yours, $result );
1031
1032                 if ( !$ok ) {
1033                         return false;
1034                 }
1035
1036                 if ( !$result ) {
1037                         return $this->makeEmptyContent();
1038                 }
1039
1040                 $mergedContent = $this->unserializeContent( $result, $format );
1041                 return $mergedContent;
1042         }
1043
1044         /**
1045          * Returns a generic ParserOutput object, wrapping the HTML returned by
1046          * getHtml().
1047          *
1048          * @param $content Content The content to render
1049          * @param $title Title Context title for parsing
1050          * @param $revId int|null Revision ID (for {{REVISIONID}})
1051          * @param $options ParserOptions|null Parser options
1052          * @param $generateHtml bool Whether or not to generate HTML
1053          *
1054          * @return ParserOutput representing the HTML form of the text
1055          */
1056         public function getParserOutput( Content $content, Title $title,
1057                 $revId = null,
1058                 ParserOptions $options = null, $generateHtml = true
1059         ) {
1060                 $this->checkModelID( $content->getModel() );
1061
1062                 # Generic implementation, relying on $this->getHtml()
1063
1064                 if ( $generateHtml ) {
1065                         $html = $this->getHtml( $content );
1066                 } else {
1067                         $html = '';
1068                 }
1069
1070                 $po = new ParserOutput( $html );
1071                 return $po;
1072         }
1073
1074         /**
1075          * Generates an HTML version of the content, for display. Used by
1076          * getParserOutput() to construct a ParserOutput object.
1077          *
1078          * This default implementation just calls getHighlightHtml(). Content
1079          * models that have another mapping to HTML (as is the case for markup
1080          * languages like wikitext) should override this method to generate the
1081          * appropriate HTML.
1082          *
1083          * @param $content Content The content to render
1084          *
1085          * @return string An HTML representation of the content
1086          */
1087         protected function getHtml( Content $content ) {
1088                 $this->checkModelID( $content->getModel() );
1089
1090                 return $this->getHighlightHtml( $content );
1091         }
1092
1093         /**
1094          * Generates a syntax-highlighted version the content, as HTML.
1095          * Used by the default implementation of getHtml().
1096          *
1097          * @param $content Content the content to render
1098          *
1099          * @return string an HTML representation of the content's markup
1100          */
1101         protected function getHighlightHtml( Content $content ) {
1102                 $this->checkModelID( $content->getModel() );
1103
1104                 # TODO: make Highlighter interface, use highlighter here, if available
1105                 return htmlspecialchars( $content->getNativeData() );
1106         }
1107
1108
1109 }
1110
1111 /**
1112  * @since WD.1
1113  */
1114 class WikitextContentHandler extends TextContentHandler {
1115
1116         public function __construct( $modelId = CONTENT_MODEL_WIKITEXT ) {
1117                 parent::__construct( $modelId, array( CONTENT_FORMAT_WIKITEXT ) );
1118         }
1119
1120         public function unserializeContent( $text, $format = null ) {
1121                 $this->checkFormat( $format );
1122
1123                 return new WikitextContent( $text );
1124         }
1125
1126         public function makeEmptyContent() {
1127                 return new WikitextContent( '' );
1128         }
1129
1130         /**
1131          * Returns a ParserOutput object resulting from parsing the content's text
1132          * using $wgParser.
1133          *
1134          * @since    WD.1
1135          *
1136          * @param $content Content the content to render
1137          * @param $title \Title
1138          * @param $revId null
1139          * @param $options null|ParserOptions
1140          * @param $generateHtml bool
1141          *
1142          * @internal param \IContextSource|null $context
1143          * @return ParserOutput representing the HTML form of the text
1144          */
1145         public function getParserOutput( Content $content, Title $title,
1146                 $revId = null,
1147                 ParserOptions $options = null, $generateHtml = true
1148         ) {
1149                 global $wgParser;
1150
1151                 $this->checkModelID( $content->getModel() );
1152
1153                 if ( !$options ) {
1154                         $options = new ParserOptions();
1155                 }
1156
1157                 $po = $wgParser->parse( $content->getNativeData(), $title, $options, true, true, $revId );
1158                 return $po;
1159         }
1160
1161         protected function getHtml( Content $content ) {
1162                 throw new MWException(
1163                         "getHtml() not implemented for wikitext. "
1164                         . "Use getParserOutput()->getText()."
1165                 );
1166         }
1167
1168         /**
1169          * Returns true because wikitext supports sections.
1170          *
1171          * @return boolean whether sections are supported.
1172          */
1173         public function supportsSections() {
1174                 return true;
1175         }
1176 }
1177
1178 # XXX: make ScriptContentHandler base class, do highlighting stuff there?
1179
1180 /**
1181  * @since WD.1
1182  */
1183 class JavaScriptContentHandler extends TextContentHandler {
1184
1185         public function __construct( $modelId = CONTENT_MODEL_JAVASCRIPT ) {
1186                 parent::__construct( $modelId, array( CONTENT_FORMAT_JAVASCRIPT ) );
1187         }
1188
1189         public function unserializeContent( $text, $format = null ) {
1190                 $this->checkFormat( $format );
1191
1192                 return new JavaScriptContent( $text );
1193         }
1194
1195         public function makeEmptyContent() {
1196                 return new JavaScriptContent( '' );
1197         }
1198
1199         protected function getHtml( Content $content ) {
1200                 $html = "";
1201                 $html .= "<pre class=\"mw-code mw-js\" dir=\"ltr\">\n";
1202                 $html .= $this->getHighlightHtml( $content );
1203                 $html .= "\n</pre>\n";
1204
1205                 return $html;
1206         }
1207 }
1208
1209 /**
1210  * @since WD.1
1211  */
1212 class CssContentHandler extends TextContentHandler {
1213
1214         public function __construct( $modelId = CONTENT_MODEL_CSS ) {
1215                 parent::__construct( $modelId, array( CONTENT_FORMAT_CSS ) );
1216         }
1217
1218         public function unserializeContent( $text, $format = null ) {
1219                 $this->checkFormat( $format );
1220
1221                 return new CssContent( $text );
1222         }
1223
1224         public function makeEmptyContent() {
1225                 return new CssContent( '' );
1226         }
1227
1228
1229         protected function getHtml( Content $content ) {
1230                 $html = "";
1231                 $html .= "<pre class=\"mw-code mw-css\" dir=\"ltr\">\n";
1232                 $html .= $this->getHighlightHtml( $content );
1233                 $html .= "\n</pre>\n";
1234
1235                 return $html;
1236         }
1237 }