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