includes/api/ApiStashEdit.php

   1 <?php
   2 /**
   3  * This program is free software; you can redistribute it and/or modify
   4  * it under the terms of the GNU General Public License as published by
   5  * the Free Software Foundation; either version 2 of the License, or
   6  * (at your option) any later version.
   7  *
   8  * This program is distributed in the hope that it will be useful,
   9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  11  * GNU General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public License along
  14  * with this program; if not, write to the Free Software Foundation, Inc.,
  15  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  16  * http://www.gnu.org/copyleft/gpl.html
  17  *
  18  * @file
  19  * @author Aaron Schulz
  20  */
  21
  22 use MediaWiki\Logger\LoggerFactory;
  23 use MediaWiki\MediaWikiServices;
  24 use Wikimedia\ScopedCallback;
  25
  26 /**
  27  * Prepare an edit in shared cache so that it can be reused on edit
  28  *
  29  * This endpoint can be called via AJAX as the user focuses on the edit
  30  * summary box. By the time of submission, the parse may have already
  31  * finished, and can be immediately used on page save. Certain parser
  32  * functions like {{REVISIONID}} or {{CURRENTTIME}} may cause the cache
  33  * to not be used on edit. Template and files used are check for changes
  34  * since the output was generated. The cache TTL is also kept low for sanity.
  35  *
  36  * @ingroup API
  37  * @since 1.25
  38  */
  39 class ApiStashEdit extends ApiBase {
  40         const ERROR_NONE = 'stashed';
  41         const ERROR_PARSE = 'error_parse';
  42         const ERROR_CACHE = 'error_cache';
  43         const ERROR_UNCACHEABLE = 'uncacheable';
  44         const ERROR_BUSY = 'busy';
  45
  46         const PRESUME_FRESH_TTL_SEC = 30;
  47         const MAX_CACHE_TTL = 300; // 5 minutes
  48
  49         public function execute() {
  50                 $user = $this->getUser();
  51                 $params = $this->extractRequestParams();
  52
  53                 if ( $user->isBot() ) { // sanity
  54                         $this->dieUsage( 'This interface is not supported for bots', 'botsnotsupported' );
  55                 }
  56
  57                 $cache = ObjectCache::getLocalClusterInstance();
  58                 $page = $this->getTitleOrPageId( $params );
  59                 $title = $page->getTitle();
  60
  61                 if ( !ContentHandler::getForModelID( $params['contentmodel'] )
  62                         ->isSupportedFormat( $params['contentformat'] )
  63                 ) {
  64                         $this->dieUsage( 'Unsupported content model/format', 'badmodelformat' );
  65                 }
  66
  67                 $text = null;
  68                 $textHash = null;
  69                 if ( strlen( $params['stashedtexthash'] ) ) {
  70                         // Load from cache since the client indicates the text is the same as last stash
  71                         $textHash = $params['stashedtexthash'];
  72                         $textKey = $cache->makeKey( 'stashedit', 'text', $textHash );
  73                         $text = $cache->get( $textKey );
  74                         if ( !is_string( $text ) ) {
  75                                 $this->dieUsage( 'No stashed text found with the given hash', 'missingtext' );
  76                         }
  77                 } elseif ( $params['text'] !== null ) {
  78                         // Trim and fix newlines so the key SHA1's match (see WebRequest::getText())
  79                         $text = rtrim( str_replace( "\r\n", "\n", $params['text'] ) );
  80                         $textHash = sha1( $text );
  81                 } else {
  82                         $this->dieUsage(
  83                                 'The text or stashedtexthash parameter must be given', 'missingtextparam' );
  84                 }
  85
  86                 $textContent = ContentHandler::makeContent(
  87                         $text, $title, $params['contentmodel'], $params['contentformat'] );
  88
  89                 $page = WikiPage::factory( $title );
  90                 if ( $page->exists() ) {
  91                         // Page exists: get the merged content with the proposed change
  92                         $baseRev = Revision::newFromPageId( $page->getId(), $params['baserevid'] );
  93                         if ( !$baseRev ) {
  94                                 $this->dieUsage( "No revision ID {$params['baserevid']}", 'missingrev' );
  95                         }
  96                         $currentRev = $page->getRevision();
  97                         if ( !$currentRev ) {
  98                                 $this->dieUsage( "No current revision of page ID {$page->getId()}", 'missingrev' );
  99                         }
 100                         // Merge in the new version of the section to get the proposed version
 101                         $editContent = $page->replaceSectionAtRev(
 102                                 $params['section'],
 103                                 $textContent,
 104                                 $params['sectiontitle'],
 105                                 $baseRev->getId()
 106                         );
 107                         if ( !$editContent ) {
 108                                 $this->dieUsage( 'Could not merge updated section.', 'replacefailed' );
 109                         }
 110                         if ( $currentRev->getId() == $baseRev->getId() ) {
 111                                 // Base revision was still the latest; nothing to merge
 112                                 $content = $editContent;
 113                         } else {
 114                                 // Merge the edit into the current version
 115                                 $baseContent = $baseRev->getContent();
 116                                 $currentContent = $currentRev->getContent();
 117                                 if ( !$baseContent || !$currentContent ) {
 118                                         $this->dieUsage( "Missing content for page ID {$page->getId()}", 'missingrev' );
 119                                 }
 120                                 $handler = ContentHandler::getForModelID( $baseContent->getModel() );
 121                                 $content = $handler->merge3( $baseContent, $editContent, $currentContent );
 122                         }
 123                 } else {
 124                         // New pages: use the user-provided content model
 125                         $content = $textContent;
 126                 }
 127
 128                 if ( !$content ) { // merge3() failed
 129                         $this->getResult()->addValue( null,
 130                                 $this->getModuleName(), [ 'status' => 'editconflict' ] );
 131                         return;
 132                 }
 133
 134                 // The user will abort the AJAX request by pressing "save", so ignore that
 135                 ignore_user_abort( true );
 136
 137                 if ( $user->pingLimiter( 'stashedit' ) ) {
 138                         $status = 'ratelimited';
 139                 } else {
 140                         $status = self::parseAndStash( $page, $content, $user, $params['summary'] );
 141                         $textKey = $cache->makeKey( 'stashedit', 'text', $textHash );
 142                         $cache->set( $textKey, $text, self::MAX_CACHE_TTL );
 143                 }
 144
 145                 $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
 146                 $stats->increment( "editstash.cache_stores.$status" );
 147
 148                 $this->getResult()->addValue(
 149                         null,
 150                         $this->getModuleName(),
 151                         [
 152                                 'status' => $status,
 153                                 'texthash' => $textHash
 154                         ]
 155                 );
 156         }
 157
 158         /**
 159          * @param WikiPage $page
 160          * @param Content $content Edit content
 161          * @param User $user
 162          * @param string $summary Edit summary
 163          * @return integer ApiStashEdit::ERROR_* constant
 164          * @since 1.25
 165          */
 166         public static function parseAndStash( WikiPage $page, Content $content, User $user, $summary ) {
 167                 $cache = ObjectCache::getLocalClusterInstance();
 168                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 169
 170                 $title = $page->getTitle();
 171                 $key = self::getStashKey( $title, self::getContentHash( $content ), $user );
 172
 173                 // Use the master DB for fast blocking locks
 174                 $dbw = wfGetDB( DB_MASTER );
 175                 if ( !$dbw->lock( $key, __METHOD__, 1 ) ) {
 176                         // De-duplicate requests on the same key
 177                         return self::ERROR_BUSY;
 178                 }
 179                 /** @noinspection PhpUnusedLocalVariableInspection */
 180                 $unlocker = new ScopedCallback( function () use ( $dbw, $key ) {
 181                         $dbw->unlock( $key, __METHOD__ );
 182                 } );
 183
 184                 $cutoffTime = time() - self::PRESUME_FRESH_TTL_SEC;
 185
 186                 // Reuse any freshly build matching edit stash cache
 187                 $editInfo = $cache->get( $key );
 188                 if ( $editInfo && wfTimestamp( TS_UNIX, $editInfo->timestamp ) >= $cutoffTime ) {
 189                         $alreadyCached = true;
 190                 } else {
 191                         $format = $content->getDefaultFormat();
 192                         $editInfo = $page->prepareContentForEdit( $content, null, $user, $format, false );
 193                         $alreadyCached = false;
 194                 }
 195
 196                 if ( $editInfo && $editInfo->output ) {
 197                         // Let extensions add ParserOutput metadata or warm other caches
 198                         Hooks::run( 'ParserOutputStashForEdit',
 199                                 [ $page, $content, $editInfo->output, $summary, $user ] );
 200
 201                         if ( $alreadyCached ) {
 202                                 $logger->debug( "Already cached parser output for key '$key' ('$title')." );
 203                                 return self::ERROR_NONE;
 204                         }
 205
 206                         list( $stashInfo, $ttl, $code ) = self::buildStashValue(
 207                                 $editInfo->pstContent,
 208                                 $editInfo->output,
 209                                 $editInfo->timestamp,
 210                                 $user
 211                         );
 212
 213                         if ( $stashInfo ) {
 214                                 $ok = $cache->set( $key, $stashInfo, $ttl );
 215                                 if ( $ok ) {
 216                                         $logger->debug( "Cached parser output for key '$key' ('$title')." );
 217                                         return self::ERROR_NONE;
 218                                 } else {
 219                                         $logger->error( "Failed to cache parser output for key '$key' ('$title')." );
 220                                         return self::ERROR_CACHE;
 221                                 }
 222                         } else {
 223                                 $logger->info( "Uncacheable parser output for key '$key' ('$title') [$code]." );
 224                                 return self::ERROR_UNCACHEABLE;
 225                         }
 226                 }
 227
 228                 return self::ERROR_PARSE;
 229         }
 230
 231         /**
 232          * Check that a prepared edit is in cache and still up-to-date
 233          *
 234          * This method blocks if the prepared edit is already being rendered,
 235          * waiting until rendering finishes before doing final validity checks.
 236          *
 237          * The cache is rejected if template or file changes are detected.
 238          * Note that foreign template or file transclusions are not checked.
 239          *
 240          * The result is a map (pstContent,output,timestamp) with fields
 241          * extracted directly from WikiPage::prepareContentForEdit().
 242          *
 243          * @param Title $title
 244          * @param Content $content
 245          * @param User $user User to get parser options from
 246          * @return stdClass|bool Returns false on cache miss
 247          */
 248         public static function checkCache( Title $title, Content $content, User $user ) {
 249                 if ( $user->isBot() ) {
 250                         return false; // bots never stash - don't pollute stats
 251                 }
 252
 253                 $cache = ObjectCache::getLocalClusterInstance();
 254                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 255                 $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
 256
 257                 $key = self::getStashKey( $title, self::getContentHash( $content ), $user );
 258                 $editInfo = $cache->get( $key );
 259                 if ( !is_object( $editInfo ) ) {
 260                         $start = microtime( true );
 261                         // We ignore user aborts and keep parsing. Block on any prior parsing
 262                         // so as to use its results and make use of the time spent parsing.
 263                         // Skip this logic if there no master connection in case this method
 264                         // is called on an HTTP GET request for some reason.
 265                         $lb = MediaWikiServices::getInstance()->getDBLoadBalancer();
 266                         $dbw = $lb->getAnyOpenConnection( $lb->getWriterIndex() );
 267                         if ( $dbw && $dbw->lock( $key, __METHOD__, 30 ) ) {
 268                                 $editInfo = $cache->get( $key );
 269                                 $dbw->unlock( $key, __METHOD__ );
 270                         }
 271
 272                         $timeMs = 1000 * max( 0, microtime( true ) - $start );
 273                         $stats->timing( 'editstash.lock_wait_time', $timeMs );
 274                 }
 275
 276                 if ( !is_object( $editInfo ) || !$editInfo->output ) {
 277                         $stats->increment( 'editstash.cache_misses.no_stash' );
 278                         $logger->debug( "Empty cache for key '$key' ('$title'); user '{$user->getName()}'." );
 279                         return false;
 280                 }
 281
 282                 $age = time() - wfTimestamp( TS_UNIX, $editInfo->output->getCacheTime() );
 283                 if ( $age <= self::PRESUME_FRESH_TTL_SEC ) {
 284                         // Assume nothing changed in this time
 285                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 286                         $logger->debug( "Timestamp-based cache hit for key '$key' (age: $age sec)." );
 287                 } elseif ( isset( $editInfo->edits ) && $editInfo->edits === $user->getEditCount() ) {
 288                         // Logged-in user made no local upload/template edits in the meantime
 289                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 290                         $logger->debug( "Edit count based cache hit for key '$key' (age: $age sec)." );
 291                 } elseif ( $user->isAnon()
 292                         && self::lastEditTime( $user ) < $editInfo->output->getCacheTime()
 293                 ) {
 294                         // Logged-out user made no local upload/template edits in the meantime
 295                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 296                         $logger->debug( "Edit check based cache hit for key '$key' (age: $age sec)." );
 297                 } else {
 298                         // User may have changed included content
 299                         $editInfo = false;
 300                 }
 301
 302                 if ( !$editInfo ) {
 303                         $stats->increment( 'editstash.cache_misses.proven_stale' );
 304                         $logger->info( "Stale cache for key '$key'; old key with outside edits. (age: $age sec)" );
 305                 } elseif ( $editInfo->output->getFlag( 'vary-revision' ) ) {
 306                         // This can be used for the initial parse, e.g. for filters or doEditContent(),
 307                         // but a second parse will be triggered in doEditUpdates(). This is not optimal.
 308                         $logger->info( "Cache for key '$key' ('$title') has vary_revision." );
 309                 } elseif ( $editInfo->output->getFlag( 'vary-revision-id' ) ) {
 310                         // Similar to the above if we didn't guess the ID correctly.
 311                         $logger->info( "Cache for key '$key' ('$title') has vary_revision_id." );
 312                 }
 313
 314                 return $editInfo;
 315         }
 316
 317         /**
 318          * @param User $user
 319          * @return string|null TS_MW timestamp or null
 320          */
 321         private static function lastEditTime( User $user ) {
 322                 $time = wfGetDB( DB_REPLICA )->selectField(
 323                         'recentchanges',
 324                         'MAX(rc_timestamp)',
 325                         [ 'rc_user_text' => $user->getName() ],
 326                         __METHOD__
 327                 );
 328
 329                 return wfTimestampOrNull( TS_MW, $time );
 330         }
 331
 332         /**
 333          * Get hash of the content, factoring in model/format
 334          *
 335          * @param Content $content
 336          * @return string
 337          */
 338         private static function getContentHash( Content $content ) {
 339                 return sha1( implode( "\n", [
 340                         $content->getModel(),
 341                         $content->getDefaultFormat(),
 342                         $content->serialize( $content->getDefaultFormat() )
 343                 ] ) );
 344         }
 345
 346         /**
 347          * Get the temporary prepared edit stash key for a user
 348          *
 349          * This key can be used for caching prepared edits provided:
 350          *   - a) The $user was used for PST options
 351          *   - b) The parser output was made from the PST using cannonical matching options
 352          *
 353          * @param Title $title
 354          * @param string $contentHash Result of getContentHash()
 355          * @param User $user User to get parser options from
 356          * @return string
 357          */
 358         private static function getStashKey( Title $title, $contentHash, User $user ) {
 359                 return ObjectCache::getLocalClusterInstance()->makeKey(
 360                         'prepared-edit',
 361                         md5( $title->getPrefixedDBkey() ),
 362                         // Account for the edit model/text
 363                         $contentHash,
 364                         // Account for user name related variables like signatures
 365                         md5( $user->getId() . "\n" . $user->getName() )
 366                 );
 367         }
 368
 369         /**
 370          * Build a value to store in memcached based on the PST content and parser output
 371          *
 372          * This makes a simple version of WikiPage::prepareContentForEdit() as stash info
 373          *
 374          * @param Content $pstContent Pre-Save transformed content
 375          * @param ParserOutput $parserOutput
 376          * @param string $timestamp TS_MW
 377          * @param User $user
 378          * @return array (stash info array, TTL in seconds, info code) or (null, 0, info code)
 379          */
 380         private static function buildStashValue(
 381                 Content $pstContent, ParserOutput $parserOutput, $timestamp, User $user
 382         ) {
 383                 // If an item is renewed, mind the cache TTL determined by config and parser functions.
 384                 // Put an upper limit on the TTL for sanity to avoid extreme template/file staleness.
 385                 $since = time() - wfTimestamp( TS_UNIX, $parserOutput->getTimestamp() );
 386                 $ttl = min( $parserOutput->getCacheExpiry() - $since, self::MAX_CACHE_TTL );
 387                 if ( $ttl <= 0 ) {
 388                         return [ null, 0, 'no_ttl' ];
 389                 }
 390
 391                 // Only store what is actually needed
 392                 $stashInfo = (object)[
 393                         'pstContent' => $pstContent,
 394                         'output'     => $parserOutput,
 395                         'timestamp'  => $timestamp,
 396                         'edits'      => $user->getEditCount()
 397                 ];
 398
 399                 return [ $stashInfo, $ttl, 'ok' ];
 400         }
 401
 402         public function getAllowedParams() {
 403                 return [
 404                         'title' => [
 405                                 ApiBase::PARAM_TYPE => 'string',
 406                                 ApiBase::PARAM_REQUIRED => true
 407                         ],
 408                         'section' => [
 409                                 ApiBase::PARAM_TYPE => 'string',
 410                         ],
 411                         'sectiontitle' => [
 412                                 ApiBase::PARAM_TYPE => 'string'
 413                         ],
 414                         'text' => [
 415                                 ApiBase::PARAM_TYPE => 'text',
 416                                 ApiBase::PARAM_DFLT => null
 417                         ],
 418                         'stashedtexthash' => [
 419                                 ApiBase::PARAM_TYPE => 'string',
 420                                 ApiBase::PARAM_DFLT => null
 421                         ],
 422                         'summary' => [
 423                                 ApiBase::PARAM_TYPE => 'string',
 424                         ],
 425                         'contentmodel' => [
 426                                 ApiBase::PARAM_TYPE => ContentHandler::getContentModels(),
 427                                 ApiBase::PARAM_REQUIRED => true
 428                         ],
 429                         'contentformat' => [
 430                                 ApiBase::PARAM_TYPE => ContentHandler::getAllContentFormats(),
 431                                 ApiBase::PARAM_REQUIRED => true
 432                         ],
 433                         'baserevid' => [
 434                                 ApiBase::PARAM_TYPE => 'integer',
 435                                 ApiBase::PARAM_REQUIRED => true
 436                         ]
 437                 ];
 438         }
 439
 440         public function needsToken() {
 441                 return 'csrf';
 442         }
 443
 444         public function mustBePosted() {
 445                 return true;
 446         }
 447
 448         public function isWriteMode() {
 449                 return true;
 450         }
 451
 452         public function isInternal() {
 453                 return true;
 454         }
 455 }