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