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          * @deprecated since WD.1. Always try to use the content object.
  54          *
  55          * @static
  56          * @param $content Content|null
  57          * @return null|string the textual form of $content, if available
  58          * @throws MWException if $content is not an instance of TextContent and
  59          *   $wgContentHandlerTextFallback was set to 'fail'.
  60          */
  61         public static function getContentText( Content $content = null ) {
  62                 global $wgContentHandlerTextFallback;
  63
  64                 if ( is_null( $content ) ) {
  65                         return '';
  66                 }
  67
  68                 if ( $content instanceof TextContent ) {
  69                         return $content->getNativeData();
  70                 }
  71
  72                 if ( $wgContentHandlerTextFallback == 'fail' ) {
  73                         throw new MWException(
  74                                 "Attempt to get text from Content with model " .
  75                                 $content->getModel()
  76                         );
  77                 }
  78
  79                 if ( $wgContentHandlerTextFallback == 'serialize' ) {
  80                         return $content->serialize();
  81                 }
  82
  83                 return null;
  84         }
  85
  86         /**
  87          * Convenience function for creating a Content object from a given textual
  88          * representation.
  89          *
  90          * $text will be deserialized into a Content object of the model specified
  91          * by $modelId (or, if that is not given, $title->getContentModel()) using
  92          * the given format.
  93          *
  94          * @since WD.1
  95          *
  96          * @static
  97          *
  98          * @param $text string the textual representation, will be
  99          *    unserialized to create the Content object
 100          * @param $title null|Title the title of the page this text belongs to.
 101          *    Required if $modelId is not provided.
 102          * @param $modelId null|string the model to deserialize to. If not provided,
 103          *    $title->getContentModel() is used.
 104          * @param $format null|string the format to use for deserialization. If not
 105          *    given, the model's default format is used.
 106          *
 107          * @return Content a Content object representing $text
 108          *
 109          * @throw MWException if $model or $format is not supported or if $text can
 110          *    not be unserialized using $format.
 111          */
 112         public static function makeContent( $text, Title $title = null,
 113                 $modelId = null, $format = null )
 114         {
 115                 if ( is_null( $modelId ) ) {
 116                         if ( is_null( $title ) ) {
 117                                 throw new MWException( "Must provide a Title object or a content model ID." );
 118                         }
 119
 120                         $modelId = $title->getContentModel();
 121                 }
 122
 123                 $handler = ContentHandler::getForModelID( $modelId );
 124                 return $handler->unserializeContent( $text, $format );
 125         }
 126
 127         /**
 128          * Returns the name of the default content model to be used for the page
 129          * with the given title.
 130          *
 131          * Note: There should rarely be need to call this method directly.
 132          * To determine the actual content model for a given page, use
 133          * Title::getContentModel().
 134          *
 135          * Which model is to be used by default for the page is determined based
 136          * on several factors:
 137          * - The global setting $wgNamespaceContentModels specifies a content model
 138          *   per namespace.
 139          * - The hook DefaultModelFor may be used to override the page's default
 140          *   model.
 141          * - Pages in NS_MEDIAWIKI and NS_USER default to the CSS or JavaScript
 142          *   model if they end in .js or .css, respectively.
 143          * - Pages in NS_MEDIAWIKI default to the wikitext model otherwise.
 144          * - The hook TitleIsCssOrJsPage may be used to force a page to use the CSS
 145          *   or JavaScript model if they end in .js or .css, respectively.
 146          * - The hook TitleIsWikitextPage may be used to force a page to use the
 147          *   wikitext model.
 148          *
 149          * If none of the above applies, the wikitext model is used.
 150          *
 151          * Note: this is used by, and may thus not use, Title::getContentModel()
 152          *
 153          * @since WD.1
 154          *
 155          * @static
 156          * @param $title Title
 157          * @return null|string default model name for the page given by $title
 158          */
 159         public static function getDefaultModelFor( Title $title ) {
 160                 global $wgNamespaceContentModels;
 161
 162                 // NOTE: this method must not rely on $title->getContentModel() directly or indirectly,
 163                 //       because it is used to initialize the mContentModel member.
 164
 165                 $ns = $title->getNamespace();
 166
 167                 $ext = false;
 168                 $m = null;
 169                 $model = null;
 170
 171                 if ( !empty( $wgNamespaceContentModels[ $ns ] ) ) {
 172                         $model = $wgNamespaceContentModels[ $ns ];
 173                 }
 174
 175                 // Hook can determine default model
 176                 if ( !wfRunHooks( 'ContentHandlerDefaultModelFor', array( $title, &$model ) ) ) {
 177                         if ( !is_null( $model ) ) {
 178                                 return $model;
 179                         }
 180                 }
 181
 182                 // Could this page contain custom CSS or JavaScript, based on the title?
 183                 $isCssOrJsPage = NS_MEDIAWIKI == $ns && preg_match( '!\.(css|js)$!u', $title->getText(), $m );
 184                 if ( $isCssOrJsPage ) {
 185                         $ext = $m[1];
 186                 }
 187
 188                 // Hook can force JS/CSS
 189                 wfRunHooks( 'TitleIsCssOrJsPage', array( $title, &$isCssOrJsPage ) );
 190
 191                 // Is this a .css subpage of a user page?
 192                 $isJsCssSubpage = NS_USER == $ns
 193                         && !$isCssOrJsPage
 194                         && preg_match( "/\\/.*\\.(js|css)$/", $title->getText(), $m );
 195                 if ( $isJsCssSubpage ) {
 196                         $ext = $m[1];
 197                 }
 198
 199                 // Is this wikitext, according to $wgNamespaceContentModels or the DefaultModelFor hook?
 200                 $isWikitext = is_null( $model ) || $model == CONTENT_MODEL_WIKITEXT;
 201                 $isWikitext = $isWikitext && !$isCssOrJsPage && !$isJsCssSubpage;
 202
 203                 // Hook can override $isWikitext
 204                 wfRunHooks( 'TitleIsWikitextPage', array( $title, &$isWikitext ) );
 205
 206                 if ( !$isWikitext ) {
 207                         switch ( $ext ) {
 208                                 case 'js':
 209                                         return CONTENT_MODEL_JAVASCRIPT;
 210                                 case 'css':
 211                                         return CONTENT_MODEL_CSS;
 212                                 default:
 213                                         return is_null( $model ) ? CONTENT_MODEL_TEXT : $model;
 214                         }
 215                 }
 216
 217                 // We established that it must be wikitext
 218
 219                 return CONTENT_MODEL_WIKITEXT;
 220         }
 221
 222         /**
 223          * Returns the appropriate ContentHandler singleton for the given title.
 224          *
 225          * @since WD.1
 226          *
 227          * @static
 228          * @param $title Title
 229          * @return ContentHandler
 230          */
 231         public static function getForTitle( Title $title ) {
 232                 $modelId = $title->getContentModel();
 233                 return ContentHandler::getForModelID( $modelId );
 234         }
 235
 236         /**
 237          * Returns the appropriate ContentHandler singleton for the given Content
 238          * object.
 239          *
 240          * @since WD.1
 241          *
 242          * @static
 243          * @param $content Content
 244          * @return ContentHandler
 245          */
 246         public static function getForContent( Content $content ) {
 247                 $modelId = $content->getModel();
 248                 return ContentHandler::getForModelID( $modelId );
 249         }
 250
 251         /**
 252          * @var Array A Cache of ContentHandler instances by model id
 253          */
 254         static $handlers;
 255
 256         /**
 257          * Returns the ContentHandler singleton for the given model ID. Use the
 258          * CONTENT_MODEL_XXX constants to identify the desired content model.
 259          *
 260          * ContentHandler singletons are taken from the global $wgContentHandlers
 261          * array. Keys in that array are model names, the values are either
 262          * ContentHandler singleton objects, or strings specifying the appropriate
 263          * subclass of ContentHandler.
 264          *
 265          * If a class name is encountered when looking up the singleton for a given
 266          * model name, the class is instantiated and the class name is replaced by
 267          * the resulting singleton in $wgContentHandlers.
 268          *
 269          * If no ContentHandler is defined for the desired $modelId, the
 270          * ContentHandler may be provided by the ContentHandlerForModelID hook.
 271          * If no ContentHandler can be determined, an MWException is raised.
 272          *
 273          * @since WD.1
 274          *
 275          * @static
 276          * @param $modelId String The ID of the content model for which to get a
 277          *    handler. Use CONTENT_MODEL_XXX constants.
 278          * @return ContentHandler The ContentHandler singleton for handling the
 279          *    model given by $modelId
 280          * @throws MWException if no handler is known for $modelId.
 281          */
 282         public static function getForModelID( $modelId ) {
 283                 global $wgContentHandlers;
 284
 285                 if ( isset( ContentHandler::$handlers[$modelId] ) ) {
 286                         return ContentHandler::$handlers[$modelId];
 287                 }
 288
 289                 if ( empty( $wgContentHandlers[$modelId] ) ) {
 290                         $handler = null;
 291
 292                         wfRunHooks( 'ContentHandlerForModelID', array( $modelId, &$handler ) );
 293
 294                         if ( $handler === null ) {
 295                                 throw new MWException( "No handler for model #$modelId registered in \$wgContentHandlers" );
 296                         }
 297
 298                         if ( !( $handler instanceof ContentHandler ) ) {
 299                                 throw new MWException( "ContentHandlerForModelID must supply a ContentHandler instance" );
 300                         }
 301                 } else {
 302                         $class = $wgContentHandlers[$modelId];
 303                         $handler = new $class( $modelId );
 304
 305                         if ( !( $handler instanceof ContentHandler ) ) {
 306                                 throw new MWException( "$class from \$wgContentHandlers is not compatible with ContentHandler" );
 307                         }
 308                 }
 309
 310                 ContentHandler::$handlers[$modelId] = $handler;
 311                 return ContentHandler::$handlers[$modelId];
 312         }
 313
 314         /**
 315          * Returns the localized name for a given content model.
 316          *
 317          * Model names are localized using system messages. Message keys
 318          * have the form content-model-$name, where $name is getContentModelName( $id ).
 319          *
 320          * @static
 321          * @param $name String The content model ID, as given by a CONTENT_MODEL_XXX
 322          *    constant or returned by Revision::getContentModel().
 323          *
 324          * @return string The content format's localized name.
 325          * @throws MWException if the model id isn't known.
 326          */
 327         public static function getLocalizedName( $name ) {
 328                 $key = "content-model-$name";
 329
 330                 if ( wfEmptyMsg( $key ) ) return $name;
 331                 else return wfMsg( $key );
 332         }
 333
 334         public static function getContentModels() {
 335                 global $wgContentHandlers;
 336
 337                 return array_keys( $wgContentHandlers );
 338         }
 339
 340         public static function getAllContentFormats() {
 341                 global $wgContentHandlers;
 342
 343                 $formats = array();
 344
 345                 foreach ( $wgContentHandlers as $model => $class ) {
 346                         $handler = ContentHandler::getForModelID( $model );
 347                         $formats = array_merge( $formats, $handler->getSupportedFormats() );
 348                 }
 349
 350                 $formats = array_unique( $formats );
 351                 return $formats;
 352         }
 353
 354         // ------------------------------------------------------------------------
 355
 356         protected $mModelID;
 357         protected $mSupportedFormats;
 358
 359         /**
 360          * Constructor, initializing the ContentHandler instance with its model ID
 361          * and a list of supported formats. Values for the parameters are typically
 362          * provided as literals by subclass's constructors.
 363          *
 364          * @param $modelId String (use CONTENT_MODEL_XXX constants).
 365          * @param $formats array List for supported serialization formats
 366          *    (typically as MIME types)
 367          */
 368         public function __construct( $modelId, $formats ) {
 369                 $this->mModelID = $modelId;
 370                 $this->mSupportedFormats = $formats;
 371
 372                 $this->mModelName = preg_replace( '/(Content)?Handler$/', '', get_class( $this ) );
 373                 $this->mModelName = preg_replace( '/[_\\\\]/', '', $this->mModelName );
 374                 $this->mModelName = strtolower( $this->mModelName );
 375         }
 376
 377         /**
 378          * Serializes a Content object of the type supported by this ContentHandler.
 379          *
 380          * @since WD.1
 381          *
 382          * @abstract
 383          * @param $content Content The Content object to serialize
 384          * @param $format null|String The desired serialization format
 385          * @return string Serialized form of the content
 386          */
 387         public abstract function serializeContent( Content $content, $format = null );
 388
 389         /**
 390          * Unserializes a Content object of the type supported by this ContentHandler.
 391          *
 392          * @since WD.1
 393          *
 394          * @abstract
 395          * @param $blob string serialized form of the content
 396          * @param $format null|String the format used for serialization
 397          * @return Content the Content object created by deserializing $blob
 398          */
 399         public abstract function unserializeContent( $blob, $format = null );
 400
 401         /**
 402          * Creates an empty Content object of the type supported by this
 403          * ContentHandler.
 404          *
 405          * @since WD.1
 406          *
 407          * @return Content
 408          */
 409         public abstract function makeEmptyContent();
 410
 411         /**
 412          * Returns the model id that identifies the content model this
 413          * ContentHandler can handle. Use with the CONTENT_MODEL_XXX constants.
 414          *
 415          * @since WD.1
 416          *
 417          * @return String The model ID
 418          */
 419         public function getModelID() {
 420                 return $this->mModelID;
 421         }
 422
 423         /**
 424          * Throws an MWException if $model_id is not the ID of the content model
 425          * supported by this ContentHandler.
 426          *
 427          * @since WD.1
 428          *
 429          * @param String $model_id The model to check
 430          *
 431          * @throws MWException
 432          */
 433         protected function checkModelID( $model_id ) {
 434                 if ( $model_id !== $this->mModelID ) {
 435                         throw new MWException( "Bad content model: " .
 436                                 "expected {$this->mModelID} " .
 437                                 "but got $model_id." );
 438                 }
 439         }
 440
 441         /**
 442          * Returns a list of serialization formats supported by the
 443          * serializeContent() and unserializeContent() methods of this
 444          * ContentHandler.
 445          *
 446          * @since WD.1
 447          *
 448          * @return array of serialization formats as MIME type like strings
 449          */
 450         public function getSupportedFormats() {
 451                 return $this->mSupportedFormats;
 452         }
 453
 454         /**
 455          * The format used for serialization/deserialization by default by this
 456          * ContentHandler.
 457          *
 458          * This default implementation will return the first element of the array
 459          * of formats that was passed to the constructor.
 460          *
 461          * @since WD.1
 462          *
 463          * @return string the name of the default serialization format as a MIME type
 464          */
 465         public function getDefaultFormat() {
 466                 return $this->mSupportedFormats[0];
 467         }
 468
 469         /**
 470          * Returns true if $format is a serialization format supported by this
 471          * ContentHandler, and false otherwise.
 472          *
 473          * Note that if $format is null, this method always returns true, because
 474          * null means "use the default format".
 475          *
 476          * @since WD.1
 477          *
 478          * @param $format string the serialization format to check
 479          * @return bool
 480          */
 481         public function isSupportedFormat( $format ) {
 482
 483                 if ( !$format ) {
 484                         return true; // this means "use the default"
 485                 }
 486
 487                 return in_array( $format, $this->mSupportedFormats );
 488         }
 489
 490         /**
 491          * Throws an MWException if isSupportedFormat( $format ) is not true.
 492          * Convenient for checking whether a format provided as a parameter is
 493          * actually supported.
 494          *
 495          * @param $format string the serialization format to check
 496          *
 497          * @throws MWException
 498          */
 499         protected function checkFormat( $format ) {
 500                 if ( !$this->isSupportedFormat( $format ) ) {
 501                         throw new MWException(
 502                                 "Format $format is not supported for content model "
 503                                 . $this->getModelID()
 504                         );
 505                 }
 506         }
 507
 508         /**
 509          * Returns overrides for action handlers.
 510          * Classes listed here will be used instead of the default one when
 511          * (and only when) $wgActions[$action] === true. This allows subclasses
 512          * to override the default action handlers.
 513          *
 514          * @since WD.1
 515          *
 516          * @return Array
 517          */
 518         public function getActionOverrides() {
 519                 return array();
 520         }
 521
 522         /**
 523          * Factory for creating an appropriate DifferenceEngine for this content model.
 524          *
 525          * @since WD.1
 526          *
 527          * @param $context IContextSource context to use, anything else will be
 528          *    ignored
 529          * @param $old Integer Old ID we want to show and diff with.
 530          * @param $new int|string String either 'prev' or 'next'.
 531          * @param $rcid Integer ??? FIXME (default 0)
 532          * @param $refreshCache boolean If set, refreshes the diff cache
 533          * @param $unhide boolean If set, allow viewing deleted revs
 534          *
 535          * @return DifferenceEngine
 536          */
 537         public function createDifferenceEngine( IContextSource $context,
 538                 $old = 0, $new = 0,
 539                 $rcid = 0, # FIXME: use everywhere!
 540                 $refreshCache = false, $unhide = false
 541         ) {
 542                 $this->checkModelID( $context->getTitle()->getContentModel() );
 543
 544                 $diffEngineClass = $this->getDiffEngineClass();
 545
 546                 return new $diffEngineClass( $context, $old, $new, $rcid, $refreshCache, $unhide );
 547         }
 548
 549         /**
 550          * Get the language in which the content of the given page is written.
 551          *
 552          * This default implementation just returns $wgContLang (except for pages in the MediaWiki namespace)
 553          *
 554          * Note that the pages language is not cacheable, since it may in some cases depend on user settings.
 555          *
 556          * Also note that the page language may or may not depend on the actual content of the page,
 557          * that is, this method may load the content in order to determine the language.
 558          *
 559          * @since 1.WD
 560          *
 561          * @param Title        $title the page to determine the language for.
 562          * @param Content|null $content the page's content, if you have it handy, to avoid reloading it.
 563          *
 564          * @return Language the page's language
 565          */
 566         public function getPageLanguage( Title $title, Content $content = null ) {
 567                 global $wgContLang;
 568
 569                 if ( $title->getNamespace() == NS_MEDIAWIKI ) {
 570                         // Parse mediawiki messages with correct target language
 571                         list( /* $unused */, $lang ) = MessageCache::singleton()->figureMessage( $title->getText() );
 572                         return wfGetLangObj( $lang );
 573                 }
 574
 575                 return $wgContLang;
 576         }
 577
 578         /**
 579          * Get the language in which the content of this page is written when
 580          * viewed by user. Defaults to $this->getPageLanguage(), but if the user
 581          * specified a preferred variant, the variant will be used.
 582          *
 583          * This default implementation just returns $this->getPageLanguage( $title, $content ) unless
 584          * the user specified a preferred variant.
 585          *
 586          * Note that the pages view language is not cacheable, since it depends on user settings.
 587          *
 588          * Also note that the page language may or may not depend on the actual content of the page,
 589          * that is, this method may load the content in order to determine the language.
 590          *
 591          * @since 1.WD
 592          *
 593          * @param Title        $title the page to determine the language for.
 594          * @param Content|null $content the page's content, if you have it handy, to avoid reloading it.
 595          *
 596          * @return Language the page's language for viewing
 597          */
 598         public function getPageViewLanguage( Title $title, Content $content = null ) {
 599                 $pageLang = $this->getPageLanguage( $title, $content );
 600
 601                 if ( $title->getNamespace() !== NS_MEDIAWIKI ) {
 602                         // If the user chooses a variant, the content is actually
 603                         // in a language whose code is the variant code.
 604                         $variant = $pageLang->getPreferredVariant();
 605                         if ( $pageLang->getCode() !== $variant ) {
 606                                 $pageLang = Language::factory( $variant );
 607                         }
 608                 }
 609
 610                 return $pageLang;
 611         }
 612
 613         /**
 614          * Determines whether the content type handled by this ContentHandler
 615          * can be used on the given page.
 616          *
 617          * This default implementation always returns true.
 618          * Subclasses may override this to restrict the use of this content model to specific locations,
 619          * typically based on the namespace or some other aspect of the title, such as a special suffix
 620          * (e.g. ".svg" for SVG content).
 621          *
 622          * @param Title $title the page's title.
 623          *
 624          * @return bool true if content of this kind can be used on the given page, false otherwise.
 625          */
 626         public function canBeUsedOn( Title $title ) {
 627                 return true;
 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( wfMessage( 'autoredircomment' )->inContentLanguage()->text() )
 694                                                 - strlen( $rt->getFullText() ) );
 695
 696                                 return wfMessage( 'autoredircomment', $rt->getFullText() )
 697                                                 ->rawParams( $truncatedtext )->inContentLanguage()->text();
 698                         }
 699                 }
 700
 701                 // New page auto-summaries
 702                 if ( $flags & EDIT_NEW && $newContent->getSize() > 0 ) {
 703                         // If they're making a new article, give its text, truncated, in
 704                         // the summary.
 705
 706                         $truncatedtext = $newContent->getTextForSummary(
 707                                 200 - strlen( wfMessage( 'autosumm-new' )->inContentLanguage()->text() ) );
 708
 709                         return wfMessage( 'autosumm-new' )->rawParams( $truncatedtext )
 710                                         ->inContentLanguage()->text();
 711                 }
 712
 713                 // Blanking auto-summaries
 714                 if ( !empty( $oldContent ) && $oldContent->getSize() > 0 && $newContent->getSize() == 0 ) {
 715                         return wfMessage( 'autosumm-blank' )->inContentLanguage()->text();
 716                 } elseif ( !empty( $oldContent )
 717                         && $oldContent->getSize() > 10 * $newContent->getSize()
 718                         && $newContent->getSize() < 500 )
 719                 {
 720                         // Removing more than 90% of the article
 721
 722                         $truncatedtext = $newContent->getTextForSummary(
 723                                 200 - strlen( wfMessage( 'autosumm-replace' )->inContentLanguage()->text() ) );
 724
 725                         return wfMessage( 'autosumm-replace' )->rawParams( $truncatedtext )
 726                                         ->inContentLanguage()->text();
 727                 }
 728
 729                 // If we reach this point, there's no applicable auto-summary for our
 730                 // case, so our auto-summary is empty.
 731                 return '';
 732         }
 733
 734         /**
 735          * Auto-generates a deletion reason
 736          *
 737          * @since WD.1
 738          *
 739          * @param $title Title: the page's title
 740          * @param &$hasHistory Boolean: whether the page has a history
 741          * @return mixed String containing deletion reason or empty string, or
 742          *    boolean false if no revision occurred
 743          *
 744          * @XXX &$hasHistory is extremely ugly, it's here because
 745          * WikiPage::getAutoDeleteReason() and Article::getReason()
 746          * have it / want it.
 747          */
 748         public function getAutoDeleteReason( Title $title, &$hasHistory ) {
 749                 $dbw = wfGetDB( DB_MASTER );
 750
 751                 // Get the last revision
 752                 $rev = Revision::newFromTitle( $title );
 753
 754                 if ( is_null( $rev ) ) {
 755                         return false;
 756                 }
 757
 758                 // Get the article's contents
 759                 $content = $rev->getContent();
 760                 $blank = false;
 761
 762                 $this->checkModelID( $content->getModel() );
 763
 764                 // If the page is blank, use the text from the previous revision,
 765                 // which can only be blank if there's a move/import/protect dummy
 766                 // revision involved
 767                 if ( $content->getSize() == 0 ) {
 768                         $prev = $rev->getPrevious();
 769
 770                         if ( $prev )    {
 771                                 $content = $prev->getContent();
 772                                 $blank = true;
 773                         }
 774                 }
 775
 776                 // Find out if there was only one contributor
 777                 // Only scan the last 20 revisions
 778                 $res = $dbw->select( 'revision', 'rev_user_text',
 779                         array(
 780                                 'rev_page' => $title->getArticleID(),
 781                                 $dbw->bitAnd( 'rev_deleted', Revision::DELETED_USER ) . ' = 0'
 782                         ),
 783                         __METHOD__,
 784                         array( 'LIMIT' => 20 )
 785                 );
 786
 787                 if ( $res === false ) {
 788                         // This page has no revisions, which is very weird
 789                         return false;
 790                 }
 791
 792                 $hasHistory = ( $res->numRows() > 1 );
 793                 $row = $dbw->fetchObject( $res );
 794
 795                 if ( $row ) { // $row is false if the only contributor is hidden
 796                         $onlyAuthor = $row->rev_user_text;
 797                         // Try to find a second contributor
 798                         foreach ( $res as $row ) {
 799                                 if ( $row->rev_user_text != $onlyAuthor ) { // Bug 22999
 800                                         $onlyAuthor = false;
 801                                         break;
 802                                 }
 803                         }
 804                 } else {
 805                         $onlyAuthor = false;
 806                 }
 807
 808                 // Generate the summary with a '$1' placeholder
 809                 if ( $blank ) {
 810                         // The current revision is blank and the one before is also
 811                         // blank. It's just not our lucky day
 812                         $reason = wfMessage( 'exbeforeblank', '$1' )->inContentLanguage()->text();
 813                 } else {
 814                         if ( $onlyAuthor ) {
 815                                 $reason = wfMessage(
 816                                         'excontentauthor',
 817                                         '$1',
 818                                         $onlyAuthor
 819                                 )->inContentLanguage()->text();
 820                         } else {
 821                                 $reason = wfMessage( 'excontent', '$1' )->inContentLanguage()->text();
 822                         }
 823                 }
 824
 825                 if ( $reason == '-' ) {
 826                         // Allow these UI messages to be blanked out cleanly
 827                         return '';
 828                 }
 829
 830                 // Max content length = max comment length - length of the comment (excl. $1)
 831                 $text = $content->getTextForSummary( 255 - ( strlen( $reason ) - 2 ) );
 832
 833                 // Now replace the '$1' placeholder
 834                 $reason = str_replace( '$1', $text, $reason );
 835
 836                 return $reason;
 837         }
 838
 839         /**
 840          * Get the Content object that needs to be saved in order to undo all revisions
 841          * between $undo and $undoafter. Revisions must belong to the same page,
 842          * must exist and must not be deleted.
 843          *
 844          * @since WD.1
 845          *
 846          * @param $current Revision The current text
 847          * @param $undo Revision The revision to undo
 848          * @param $undoafter Revision Must be an earlier revision than $undo
 849          *
 850          * @return mixed String on success, false on failure
 851          */
 852         public function getUndoContent( Revision $current, Revision $undo, Revision $undoafter ) {
 853                 $cur_content = $current->getContent();
 854
 855                 if ( empty( $cur_content ) ) {
 856                         return false; // no page
 857                 }
 858
 859                 $undo_content = $undo->getContent();
 860                 $undoafter_content = $undoafter->getContent();
 861
 862                 $this->checkModelID( $cur_content->getModel() );
 863                 $this->checkModelID( $undo_content->getModel() );
 864                 $this->checkModelID( $undoafter_content->getModel() );
 865
 866                 if ( $cur_content->equals( $undo_content ) ) {
 867                         // No use doing a merge if it's just a straight revert.
 868                         return $undoafter_content;
 869                 }
 870
 871                 $undone_content = $this->merge3( $undo_content, $undoafter_content, $cur_content );
 872
 873                 return $undone_content;
 874         }
 875
 876         /**
 877          * Returns true for content models that support caching using the
 878          * ParserCache mechanism. See WikiPage::isParserCacheUser().
 879          *
 880          * @since WD.1
 881          *
 882          * @return bool
 883          */
 884         public function isParserCacheSupported() {
 885                 return true;
 886         }
 887
 888         /**
 889          * Returns true if this content model supports sections.
 890          *
 891          * This default implementation returns false.
 892          *
 893          * @return boolean whether sections are supported.
 894          */
 895         public function supportsSections() {
 896                 return false;
 897         }
 898
 899         /**
 900          * Call a legacy hook that uses text instead of Content objects.
 901          * Will log a warning when a matching hook function is registered.
 902          * If the textual representation of the content is changed by the
 903          * hook function, a new Content object is constructed from the new
 904          * text.
 905          *
 906          * @param $event String: event name
 907          * @param $args Array: parameters passed to hook functions
 908          * @param $warn bool: whether to log a warning (default: true). Should generally be true,
 909          *                    may be set to false for testing.
 910          *
 911          * @return Boolean True if no handler aborted the hook
 912          */
 913         public static function runLegacyHooks( $event, $args = array(), $warn = true ) {
 914                 if ( !Hooks::isRegistered( $event ) ) {
 915                         return true; // nothing to do here
 916                 }
 917
 918                 if ( $warn ) {
 919                         wfWarn( "Using obsolete hook $event" );
 920                 }
 921
 922                 // convert Content objects to text
 923                 $contentObjects = array();
 924                 $contentTexts = array();
 925
 926                 foreach ( $args as $k => $v ) {
 927                         if ( $v instanceof Content ) {
 928                                 /* @var Content $v */
 929
 930                                 $contentObjects[$k] = $v;
 931
 932                                 $v = $v->serialize();
 933                                 $contentTexts[ $k ] = $v;
 934                                 $args[ $k ] = $v;
 935                         }
 936                 }
 937
 938                 // call the hook functions
 939                 $ok = wfRunHooks( $event, $args );
 940
 941                 // see if the hook changed the text
 942                 foreach ( $contentTexts as $k => $orig ) {
 943                         /* @var Content $content */
 944
 945                         $modified = $args[ $k ];
 946                         $content = $contentObjects[$k];
 947
 948                         if ( $modified !== $orig ) {
 949                                 // text was changed, create updated Content object
 950                                 $content = $content->getContentHandler()->unserializeContent( $modified );
 951                         }
 952
 953                         $args[ $k ] = $content;
 954                 }
 955
 956                 return $ok;
 957         }
 958 }
 959
 960 /**
 961  * @since WD.1
 962  */
 963 abstract class TextContentHandler extends ContentHandler {
 964
 965         public function __construct( $modelId, $formats ) {
 966                 parent::__construct( $modelId, $formats );
 967         }
 968
 969         /**
 970          * Returns the content's text as-is.
 971          *
 972          * @param $content Content
 973          * @param $format string|null
 974          * @return mixed
 975          */
 976         public function serializeContent( Content $content, $format = null ) {
 977                 $this->checkFormat( $format );
 978                 return $content->getNativeData();
 979         }
 980
 981         /**
 982          * Attempts to merge differences between three versions. Returns a new
 983          * Content object for a clean merge and false for failure or a conflict.
 984          *
 985          * All three Content objects passed as parameters must have the same
 986          * content model.
 987          *
 988          * This text-based implementation uses wfMerge().
 989          *
 990          * @param $oldContent \Content|string  String
 991          * @param $myContent \Content|string   String
 992          * @param $yourContent \Content|string String
 993          *
 994          * @return Content|Bool
 995          */
 996         public function merge3( Content $oldContent, Content $myContent, Content $yourContent ) {
 997                 $this->checkModelID( $oldContent->getModel() );
 998                 $this->checkModelID( $myContent->getModel() );
 999                 $this->checkModelID( $yourContent->getModel() );
1000
1001                 $format = $this->getDefaultFormat();
1002
1003                 $old = $this->serializeContent( $oldContent, $format );
1004                 $mine = $this->serializeContent( $myContent, $format );
1005                 $yours = $this->serializeContent( $yourContent, $format );
1006
1007                 $ok = wfMerge( $old, $mine, $yours, $result );
1008
1009                 if ( !$ok ) {
1010                         return false;
1011                 }
1012
1013                 if ( !$result ) {
1014                         return $this->makeEmptyContent();
1015                 }
1016
1017                 $mergedContent = $this->unserializeContent( $result, $format );
1018                 return $mergedContent;
1019         }
1020
1021 }
1022
1023 /**
1024  * @since WD.1
1025  */
1026 class WikitextContentHandler extends TextContentHandler {
1027
1028         public function __construct( $modelId = CONTENT_MODEL_WIKITEXT ) {
1029                 parent::__construct( $modelId, array( CONTENT_FORMAT_WIKITEXT ) );
1030         }
1031
1032         public function unserializeContent( $text, $format = null ) {
1033                 $this->checkFormat( $format );
1034
1035                 return new WikitextContent( $text );
1036         }
1037
1038         public function makeEmptyContent() {
1039                 return new WikitextContent( '' );
1040         }
1041
1042         /**
1043          * Returns true because wikitext supports sections.
1044          *
1045          * @return boolean whether sections are supported.
1046          */
1047         public function supportsSections() {
1048                 return true;
1049         }
1050 }
1051
1052 # XXX: make ScriptContentHandler base class, do highlighting stuff there?
1053
1054 /**
1055  * @since WD.1
1056  */
1057 class JavaScriptContentHandler extends TextContentHandler {
1058
1059         public function __construct( $modelId = CONTENT_MODEL_JAVASCRIPT ) {
1060                 parent::__construct( $modelId, array( CONTENT_FORMAT_JAVASCRIPT ) );
1061         }
1062
1063         public function unserializeContent( $text, $format = null ) {
1064                 $this->checkFormat( $format );
1065
1066                 return new JavaScriptContent( $text );
1067         }
1068
1069         public function makeEmptyContent() {
1070                 return new JavaScriptContent( '' );
1071         }
1072
1073         /**
1074          * Returns the english language, because JS is english, and should be handled as such.
1075          *
1076          * @return Language wfGetLangObj( 'en' )
1077          *
1078          * @see ContentHandler::getPageLanguage()
1079          */
1080         public function getPageLanguage( Title $title, Content $content = null ) {
1081                 return wfGetLangObj( 'en' );
1082         }
1083
1084         /**
1085          * Returns the english language, because CSS is english, and should be handled as such.
1086          *
1087          * @return Language wfGetLangObj( 'en' )
1088          *
1089          * @see ContentHandler::getPageViewLanguage()
1090          */
1091         public function getPageViewLanguage( Title $title, Content $content = null ) {
1092                 return wfGetLangObj( 'en' );
1093         }
1094 }
1095
1096 /**
1097  * @since WD.1
1098  */
1099 class CssContentHandler extends TextContentHandler {
1100
1101         public function __construct( $modelId = CONTENT_MODEL_CSS ) {
1102                 parent::__construct( $modelId, array( CONTENT_FORMAT_CSS ) );
1103         }
1104
1105         public function unserializeContent( $text, $format = null ) {
1106                 $this->checkFormat( $format );
1107
1108                 return new CssContent( $text );
1109         }
1110
1111         public function makeEmptyContent() {
1112                 return new CssContent( '' );
1113         }
1114
1115         /**
1116          * Returns the english language, because CSS is english, and should be handled as such.
1117          *
1118          * @return Language wfGetLangObj( 'en' )
1119          *
1120          * @see ContentHandler::getPageLanguage()
1121          */
1122         public function getPageLanguage( Title $title, Content $content = null ) {
1123                 return wfGetLangObj( 'en' );
1124         }
1125
1126         /**
1127          * Returns the english language, because CSS is english, and should be handled as such.
1128          *
1129          * @return Language wfGetLangObj( 'en' )
1130          *
1131          * @see ContentHandler::getPageViewLanguage()
1132          */
1133         public function getPageViewLanguage( Title $title, Content $content = null ) {
1134                 return wfGetLangObj( 'en' );
1135         }
1136 }