includes/Revision/RenderedRevision.php

   1 <?php
   2 /**
   3  * This file is part of MediaWiki.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 namespace MediaWiki\Revision;
  24
  25 use InvalidArgumentException;
  26 use LogicException;
  27 use ParserOptions;
  28 use ParserOutput;
  29 use Psr\Log\LoggerInterface;
  30 use Psr\Log\NullLogger;
  31 use Revision;
  32 use Title;
  33 use User;
  34 use Content;
  35 use Wikimedia\Assert\Assert;
  36
  37 /**
  38  * RenderedRevision represents the rendered representation of a revision. It acts as a lazy provider
  39  * of ParserOutput objects for the revision's individual slots, as well as a combined ParserOutput
  40  * of all slots.
  41  *
  42  * @since 1.32
  43  */
  44 class RenderedRevision implements SlotRenderingProvider {
  45
  46         /**
  47          * @var Title
  48          */
  49         private $title;
  50
  51         /** @var RevisionRecord */
  52         private $revision;
  53
  54         /**
  55          * @var ParserOptions
  56          */
  57         private $options;
  58
  59         /**
  60          * @var int Audience to check when accessing content.
  61          */
  62         private $audience = RevisionRecord::FOR_PUBLIC;
  63
  64         /**
  65          * @var User|null The user to use for audience checks during content access.
  66          */
  67         private $forUser = null;
  68
  69         /**
  70          * @var ParserOutput|null The combined ParserOutput for the revision,
  71          *      initialized lazily by getRevisionParserOutput().
  72          */
  73         private $revisionOutput = null;
  74
  75         /**
  76          * @var ParserOutput[] The ParserOutput for each slot,
  77          *      initialized lazily by getSlotParserOutput().
  78          */
  79         private $slotsOutput = [];
  80
  81         /**
  82          * @var callable Callback for combining slot output into revision output.
  83          *      Signature: function ( RenderedRevision $this ): ParserOutput.
  84          */
  85         private $combineOutput;
  86
  87         /**
  88          * @var LoggerInterface For profiling ParserOutput re-use.
  89          */
  90         private $saveParseLogger;
  91
  92         /**
  93          * @note Application logic should not instantiate RenderedRevision instances directly,
  94          * but should use a RevisionRenderer instead.
  95          *
  96          * @param Title $title
  97          * @param RevisionRecord $revision The revision to render. The content for rendering will be
  98          *        taken from this RevisionRecord. However, if the RevisionRecord is not complete
  99          *        according isReadyForInsertion(), but a revision ID is known, the parser may load
 100          *        the revision from the database if it needs revision meta data to handle magic
 101          *        words like {{REVISIONUSER}}.
 102          * @param ParserOptions $options
 103          * @param callable $combineOutput Callback for combining slot output into revision output.
 104          *        Signature: function ( RenderedRevision $this ): ParserOutput.
 105          * @param int $audience Use RevisionRecord::FOR_PUBLIC, FOR_THIS_USER, or RAW.
 106          * @param User|null $forUser Required if $audience is FOR_THIS_USER.
 107          */
 108         public function __construct(
 109                 Title $title,
 110                 RevisionRecord $revision,
 111                 ParserOptions $options,
 112                 callable $combineOutput,
 113                 $audience = RevisionRecord::FOR_PUBLIC,
 114                 User $forUser = null
 115         ) {
 116                 $this->title = $title;
 117                 $this->options = $options;
 118
 119                 $this->setRevisionInternal( $revision );
 120
 121                 $this->combineOutput = $combineOutput;
 122                 $this->saveParseLogger = new NullLogger();
 123
 124                 if ( $audience === RevisionRecord::FOR_THIS_USER && !$forUser ) {
 125                         throw new InvalidArgumentException(
 126                                 'User must be specified when setting audience to FOR_THIS_USER'
 127                         );
 128                 }
 129
 130                 $this->audience = $audience;
 131                 $this->forUser = $forUser;
 132         }
 133
 134         /**
 135          * @param LoggerInterface $saveParseLogger
 136          */
 137         public function setSaveParseLogger( LoggerInterface $saveParseLogger ) {
 138                 $this->saveParseLogger = $saveParseLogger;
 139         }
 140
 141         /**
 142          * @return bool Whether the revision's content has been hidden from unprivileged users.
 143          */
 144         public function isContentDeleted() {
 145                 return $this->revision->isDeleted( RevisionRecord::DELETED_TEXT );
 146         }
 147
 148         /**
 149          * @return RevisionRecord
 150          */
 151         public function getRevision() {
 152                 return $this->revision;
 153         }
 154
 155         /**
 156          * @return ParserOptions
 157          */
 158         public function getOptions() {
 159                 return $this->options;
 160         }
 161
 162         /**
 163          * Sets a ParserOutput to be returned by getRevisionParserOutput().
 164          *
 165          * @note For internal use by RevisionRenderer only! This method may be modified
 166          * or removed without notice per the deprecation policy.
 167          *
 168          * @internal
 169          *
 170          * @param ParserOutput $output
 171          */
 172         public function setRevisionParserOutput( ParserOutput $output ) {
 173                 $this->revisionOutput = $output;
 174
 175                 // If there is only one slot, we assume that the combined output is identical
 176                 // with the main slot's output. This is intended to prevent a redundant re-parse of
 177                 // the content in case getSlotParserOutput( SlotRecord::MAIN ) is called, for instance
 178                 // from ContentHandler::getSecondaryDataUpdates.
 179                 if ( $this->revision->getSlotRoles() === [ SlotRecord::MAIN ] ) {
 180                         $this->slotsOutput[ SlotRecord::MAIN ] = $output;
 181                 }
 182         }
 183
 184         /**
 185          * @param array $hints Hints given as an associative array. Known keys:
 186          *      - 'generate-html' => bool: Whether the caller is interested in output HTML (as opposed
 187          *        to just meta-data). Default is to generate HTML.
 188          *
 189          * @return ParserOutput
 190          */
 191         public function getRevisionParserOutput( array $hints = [] ) {
 192                 $withHtml = $hints['generate-html'] ?? true;
 193
 194                 if ( !$this->revisionOutput
 195                         || ( $withHtml && !$this->revisionOutput->hasText() )
 196                 ) {
 197                         $output = call_user_func( $this->combineOutput, $this, $hints );
 198
 199                         Assert::postcondition(
 200                                 $output instanceof ParserOutput,
 201                                 'Callback did not return a ParserOutput object!'
 202                         );
 203
 204                         $this->revisionOutput = $output;
 205                 }
 206
 207                 return $this->revisionOutput;
 208         }
 209
 210         /**
 211          * @param string $role
 212          * @param array $hints Hints given as an associative array. Known keys:
 213          *      - 'generate-html' => bool: Whether the caller is interested in output HTML (as opposed
 214          *        to just meta-data). Default is to generate HTML.
 215          *
 216          * @throws SuppressedDataException if the content is not accessible for the audience
 217          *         specified in the constructor.
 218          * @return ParserOutput
 219          */
 220         public function getSlotParserOutput( $role, array $hints = [] ) {
 221                 $withHtml = $hints['generate-html'] ?? true;
 222
 223                 if ( !isset( $this->slotsOutput[ $role ] )
 224                         || ( $withHtml && !$this->slotsOutput[ $role ]->hasText() )
 225                 ) {
 226                         $content = $this->revision->getContent( $role, $this->audience, $this->forUser );
 227
 228                         if ( !$content ) {
 229                                 throw new SuppressedDataException(
 230                                         'Access to the content has been suppressed for this audience'
 231                                 );
 232                         } else {
 233                                 // XXX: allow SlotRoleHandler to control the ParserOutput?
 234                                 $output = $this->getSlotParserOutputUncached( $content, $withHtml );
 235
 236                                 if ( $withHtml && !$output->hasText() ) {
 237                                         throw new LogicException(
 238                                                 'HTML generation was requested, but '
 239                                                 . get_class( $content )
 240                                                 . '::getParserOutput() returns a ParserOutput with no text set.'
 241                                         );
 242                                 }
 243
 244                                 // Detach watcher, to ensure option use is not recorded in the wrong ParserOutput.
 245                                 $this->options->registerWatcher( null );
 246                         }
 247
 248                         $this->slotsOutput[ $role ] = $output;
 249                 }
 250
 251                 return $this->slotsOutput[$role];
 252         }
 253
 254         /**
 255          * @note This method exist to make duplicate parses easier to see during profiling
 256          * @param Content $content
 257          * @param bool $withHtml
 258          * @return ParserOutput
 259          */
 260         private function getSlotParserOutputUncached( Content $content, $withHtml ) {
 261                 return $content->getParserOutput(
 262                         $this->title,
 263                         $this->revision->getId(),
 264                         $this->options,
 265                         $withHtml
 266                 );
 267         }
 268
 269         /**
 270          * Updates the RevisionRecord after the revision has been saved. This can be used to discard
 271          * and cached ParserOutput so parser functions like {{REVISIONTIMESTAMP}} or {{REVISIONID}}
 272          * are re-evaluated.
 273          *
 274          * @note There should be no need to call this for null-edits.
 275          *
 276          * @param RevisionRecord $rev
 277          */
 278         public function updateRevision( RevisionRecord $rev ) {
 279                 if ( $rev->getId() === $this->revision->getId() ) {
 280                         return;
 281                 }
 282
 283                 if ( $this->revision->getId() ) {
 284                         throw new LogicException( 'RenderedRevision already has a revision with ID '
 285                                 . $this->revision->getId(), ', can\'t update to revision with ID ' . $rev->getId() );
 286                 }
 287
 288                 if ( !$this->revision->getSlots()->hasSameContent( $rev->getSlots() ) ) {
 289                         throw new LogicException( 'Cannot update to a revision with different content!' );
 290                 }
 291
 292                 $this->setRevisionInternal( $rev );
 293
 294                 $this->pruneRevisionSensitiveOutput( $this->revision->getId() );
 295         }
 296
 297         /**
 298          * Prune any output that depends on the revision ID.
 299          *
 300          * @param int|bool  $actualRevId The actual rev id, to check the used speculative rev ID
 301          *        against, or false to not purge on vary-revision-id, or true to purge on
 302          *        vary-revision-id unconditionally.
 303          */
 304         private function pruneRevisionSensitiveOutput( $actualRevId ) {
 305                 if ( $this->revisionOutput ) {
 306                         if ( $this->outputVariesOnRevisionMetaData( $this->revisionOutput, $actualRevId ) ) {
 307                                 $this->revisionOutput = null;
 308                         }
 309                 } else {
 310                         $this->saveParseLogger->debug( __METHOD__ . ": no prepared revision output...\n" );
 311                 }
 312
 313                 foreach ( $this->slotsOutput as $role => $output ) {
 314                         if ( $this->outputVariesOnRevisionMetaData( $output, $actualRevId ) ) {
 315                                 unset( $this->slotsOutput[$role] );
 316                         }
 317                 }
 318         }
 319
 320         /**
 321          * @param RevisionRecord $revision
 322          */
 323         private function setRevisionInternal( RevisionRecord $revision ) {
 324                 $this->revision = $revision;
 325
 326                 // Force the parser to use  $this->revision to resolve magic words like {{REVISIONUSER}}
 327                 // if the revision is either known to be complete, or it doesn't have a revision ID set.
 328                 // If it's incomplete and we have a revision ID, the parser can do better by loading
 329                 // the revision from the database if needed to handle a magic word.
 330                 //
 331                 // The following considerations inform the logic described above:
 332                 //
 333                 // 1) If we have a saved revision already loaded, we want the parser to use it, instead of
 334                 // loading it again.
 335                 //
 336                 // 2) If the revision is a fake that wraps some kind of synthetic content, such as an
 337                 // error message from Article, it should be used directly and things like {{REVISIONUSER}}
 338                 // should not expected to work, since there may not even be an actual revision to
 339                 // refer to.
 340                 //
 341                 // 3) If the revision is a fake constructed around a Title, a Content object, and
 342                 // a revision ID, to provide backwards compatibility to code that has access to those
 343                 // but not to a complete RevisionRecord for rendering, then we want the Parser to
 344                 // load the actual revision from the database when it encounters a magic word like
 345                 // {{REVISIONUSER}}, but we don't want to load that revision ahead of time just in case.
 346                 //
 347                 // 4) Previewing an edit to a template should use the submitted unsaved
 348                 // MutableRevisionRecord for self-transclusions in the template's documentation (see T7278).
 349                 // That revision would be complete except for the ID field.
 350                 //
 351                 // 5) Pre-save transform would provide a RevisionRecord that has all meta-data but is
 352                 // incomplete due to not yet having content set. However, since it doesn't have a revision
 353                 // ID either, the below code would still force it to be used, allowing
 354                 // {{subst::REVISIONUSER}} to function as expected.
 355
 356                 if ( $this->revision->isReadyForInsertion() || !$this->revision->getId() ) {
 357                         $title = $this->title;
 358                         $oldCallback = $this->options->getCurrentRevisionCallback();
 359                         $this->options->setCurrentRevisionCallback(
 360                                 function ( Title $parserTitle, $parser = false ) use ( $title, $oldCallback ) {
 361                                         if ( $title->equals( $parserTitle ) ) {
 362                                                 $legacyRevision = new Revision( $this->revision );
 363                                                 return $legacyRevision;
 364                                         } else {
 365                                                 return call_user_func( $oldCallback, $parserTitle, $parser );
 366                                         }
 367                                 }
 368                         );
 369                 }
 370         }
 371
 372         /**
 373          * @param ParserOutput $out
 374          * @param int|bool  $actualRevId The actual rev id, to check the used speculative rev ID
 375          *        against, or false to not purge on vary-revision-id, or true to purge on
 376          *        vary-revision-id unconditionally.
 377          * @return bool
 378          */
 379         private function outputVariesOnRevisionMetaData( ParserOutput $out, $actualRevId ) {
 380                 $method = __METHOD__;
 381
 382                 if ( $out->getFlag( 'vary-revision' ) ) {
 383                         // XXX: Would be just keep the output if the speculative revision ID was correct,
 384                         // but that can go wrong for some edge cases, like {{PAGEID}} during page creation.
 385                         // For that specific case, it would perhaps nice to have a vary-page flag.
 386                         $this->saveParseLogger->info(
 387                                 "$method: Prepared output has vary-revision...\n"
 388                         );
 389                         return true;
 390                 } elseif ( $out->getFlag( 'vary-revision-id' )
 391                         && $actualRevId !== false
 392                         && ( $actualRevId === true || $out->getSpeculativeRevIdUsed() !== $actualRevId )
 393                 ) {
 394                         $this->saveParseLogger->info(
 395                                 "$method: Prepared output has vary-revision-id with wrong ID...\n"
 396                         );
 397                         return true;
 398                 } elseif ( $out->getFlag( 'vary-revision-exists' ) ) {
 399                         // If {{REVISIONID}} resolved to '', it now needs to resolve to '-'.
 400                         // Note that edit stashing always uses '-', which can be used for both
 401                         // edit filter checks and canonical parser cache.
 402                         $this->saveParseLogger->info(
 403                                 "$method: Prepared output has vary-revision-exists...\n"
 404                         );
 405                         return true;
 406                 } else {
 407                         // NOTE: In the original fix for T135261, the output was discarded if 'vary-user' was
 408                         // set for a null-edit. The reason was that the original rendering in that case was
 409                         // targeting the user making the null-edit, not the user who made the original edit,
 410                         // causing {{REVISIONUSER}} to return the wrong name.
 411                         // This case is now expected to be handled by the code in RevisionRenderer that
 412                         // constructs the ParserOptions: For a null-edit, setCurrentRevisionCallback is called
 413                         // with the old, existing revision.
 414
 415                         wfDebug( "$method: Keeping prepared output...\n" );
 416                         return false;
 417                 }
 418         }
 419
 420 }