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  */
  20
  21 use MediaWiki\Logger\LoggerFactory;
  22 use MediaWiki\MediaWikiServices;
  23 use Wikimedia\ScopedCallback;
  24
  25 /**
  26  * Prepare an edit in shared cache so that it can be reused on edit
  27  *
  28  * This endpoint can be called via AJAX as the user focuses on the edit
  29  * summary box. By the time of submission, the parse may have already
  30  * finished, and can be immediately used on page save. Certain parser
  31  * functions like {{REVISIONID}} or {{CURRENTTIME}} may cause the cache
  32  * to not be used on edit. Template and files used are check for changes
  33  * since the output was generated. The cache TTL is also kept low for sanity.
  34  *
  35  * @ingroup API
  36  * @since 1.25
  37  */
  38 class ApiStashEdit extends ApiBase {
  39         const ERROR_NONE = 'stashed';
  40         const ERROR_PARSE = 'error_parse';
  41         const ERROR_CACHE = 'error_cache';
  42         const ERROR_UNCACHEABLE = 'uncacheable';
  43         const ERROR_BUSY = 'busy';
  44
  45         const PRESUME_FRESH_TTL_SEC = 30;
  46         const MAX_CACHE_TTL = 300; // 5 minutes
  47         const MAX_SIGNATURE_TTL = 60;
  48
  49         const MAX_CACHE_RECENT = 2;
  50
  51         public function execute() {
  52                 $user = $this->getUser();
  53                 $params = $this->extractRequestParams();
  54
  55                 if ( $user->isBot() ) { // sanity
  56                         $this->dieWithError( 'apierror-botsnotsupported' );
  57                 }
  58
  59                 $cache = ObjectCache::getLocalClusterInstance();
  60                 $page = $this->getTitleOrPageId( $params );
  61                 $title = $page->getTitle();
  62
  63                 if ( !ContentHandler::getForModelID( $params['contentmodel'] )
  64                         ->isSupportedFormat( $params['contentformat'] )
  65                 ) {
  66                         $this->dieWithError(
  67                                 [ 'apierror-badformat-generic', $params['contentformat'], $params['contentmodel'] ],
  68                                 'badmodelformat'
  69                         );
  70                 }
  71
  72                 $this->requireOnlyOneParameter( $params, 'stashedtexthash', 'text' );
  73
  74                 $text = null;
  75                 $textHash = null;
  76                 if ( $params['stashedtexthash'] !== null ) {
  77                         // Load from cache since the client indicates the text is the same as last stash
  78                         $textHash = $params['stashedtexthash'];
  79                         if ( !preg_match( '/^[0-9a-f]{40}$/', $textHash ) ) {
  80                                 $this->dieWithError( 'apierror-stashedit-missingtext', 'missingtext' );
  81                         }
  82                         $textKey = $cache->makeKey( 'stashedit', 'text', $textHash );
  83                         $text = $cache->get( $textKey );
  84                         if ( !is_string( $text ) ) {
  85                                 $this->dieWithError( 'apierror-stashedit-missingtext', 'missingtext' );
  86                         }
  87                 } else {
  88                         // 'text' was passed.  Trim and fix newlines so the key SHA1's
  89                         // match (see WebRequest::getText())
  90                         $text = rtrim( str_replace( "\r\n", "\n", $params['text'] ) );
  91                         $textHash = sha1( $text );
  92                 }
  93
  94                 $textContent = ContentHandler::makeContent(
  95                         $text, $title, $params['contentmodel'], $params['contentformat'] );
  96
  97                 $page = WikiPage::factory( $title );
  98                 if ( $page->exists() ) {
  99                         // Page exists: get the merged content with the proposed change
 100                         $baseRev = Revision::newFromPageId( $page->getId(), $params['baserevid'] );
 101                         if ( !$baseRev ) {
 102                                 $this->dieWithError( [ 'apierror-nosuchrevid', $params['baserevid'] ] );
 103                         }
 104                         $currentRev = $page->getRevision();
 105                         if ( !$currentRev ) {
 106                                 $this->dieWithError( [ 'apierror-missingrev-pageid', $page->getId() ], 'missingrev' );
 107                         }
 108                         // Merge in the new version of the section to get the proposed version
 109                         $editContent = $page->replaceSectionAtRev(
 110                                 $params['section'],
 111                                 $textContent,
 112                                 $params['sectiontitle'],
 113                                 $baseRev->getId()
 114                         );
 115                         if ( !$editContent ) {
 116                                 $this->dieWithError( 'apierror-sectionreplacefailed', 'replacefailed' );
 117                         }
 118                         if ( $currentRev->getId() == $baseRev->getId() ) {
 119                                 // Base revision was still the latest; nothing to merge
 120                                 $content = $editContent;
 121                         } else {
 122                                 // Merge the edit into the current version
 123                                 $baseContent = $baseRev->getContent();
 124                                 $currentContent = $currentRev->getContent();
 125                                 if ( !$baseContent || !$currentContent ) {
 126                                         $this->dieWithError( [ 'apierror-missingcontent-pageid', $page->getId() ], 'missingrev' );
 127                                 }
 128                                 $handler = ContentHandler::getForModelID( $baseContent->getModel() );
 129                                 $content = $handler->merge3( $baseContent, $editContent, $currentContent );
 130                         }
 131                 } else {
 132                         // New pages: use the user-provided content model
 133                         $content = $textContent;
 134                 }
 135
 136                 if ( !$content ) { // merge3() failed
 137                         $this->getResult()->addValue( null,
 138                                 $this->getModuleName(), [ 'status' => 'editconflict' ] );
 139                         return;
 140                 }
 141
 142                 // The user will abort the AJAX request by pressing "save", so ignore that
 143                 ignore_user_abort( true );
 144
 145                 if ( $user->pingLimiter( 'stashedit' ) ) {
 146                         $status = 'ratelimited';
 147                 } else {
 148                         $status = self::parseAndStash( $page, $content, $user, $params['summary'] );
 149                         $textKey = $cache->makeKey( 'stashedit', 'text', $textHash );
 150                         $cache->set( $textKey, $text, self::MAX_CACHE_TTL );
 151                 }
 152
 153                 $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
 154                 $stats->increment( "editstash.cache_stores.$status" );
 155
 156                 $ret = [ 'status' => $status ];
 157                 // If we were rate-limited, we still return the pre-existing valid hash if one was passed
 158                 if ( $status !== 'ratelimited' || $params['stashedtexthash'] !== null ) {
 159                         $ret['texthash'] = $textHash;
 160                 }
 161
 162                 $this->getResult()->addValue( null, $this->getModuleName(), $ret );
 163         }
 164
 165         /**
 166          * @param WikiPage $page
 167          * @param Content $content Edit content
 168          * @param User $user
 169          * @param string $summary Edit summary
 170          * @return string ApiStashEdit::ERROR_* constant
 171          * @since 1.25
 172          */
 173         public static function parseAndStash( WikiPage $page, Content $content, User $user, $summary ) {
 174                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 175
 176                 $title = $page->getTitle();
 177                 $key = self::getStashKey( $title, self::getContentHash( $content ), $user );
 178                 $fname = __METHOD__;
 179
 180                 // Use the master DB to allow for fast blocking locks on the "save path" where this
 181                 // value might actually be used to complete a page edit. If the edit submission request
 182                 // happens before this edit stash requests finishes, then the submission will block until
 183                 // the stash request finishes parsing. For the lock acquisition below, there is not much
 184                 // need to duplicate parsing of the same content/user/summary bundle, so try to avoid
 185                 // blocking at all here.
 186                 $dbw = wfGetDB( DB_MASTER );
 187                 if ( !$dbw->lock( $key, $fname, 0 ) ) {
 188                         // De-duplicate requests on the same key
 189                         return self::ERROR_BUSY;
 190                 }
 191                 /** @noinspection PhpUnusedLocalVariableInspection */
 192                 $unlocker = new ScopedCallback( function () use ( $dbw, $key, $fname ) {
 193                         $dbw->unlock( $key, $fname );
 194                 } );
 195
 196                 $cutoffTime = time() - self::PRESUME_FRESH_TTL_SEC;
 197
 198                 // Reuse any freshly build matching edit stash cache
 199                 $editInfo = self::getStashValue( $key );
 200                 if ( $editInfo && wfTimestamp( TS_UNIX, $editInfo->timestamp ) >= $cutoffTime ) {
 201                         $alreadyCached = true;
 202                 } else {
 203                         $format = $content->getDefaultFormat();
 204                         $editInfo = $page->prepareContentForEdit( $content, null, $user, $format, false );
 205                         $alreadyCached = false;
 206                 }
 207
 208                 if ( $editInfo && $editInfo->output ) {
 209                         // Let extensions add ParserOutput metadata or warm other caches
 210                         Hooks::run( 'ParserOutputStashForEdit',
 211                                 [ $page, $content, $editInfo->output, $summary, $user ] );
 212
 213                         $titleStr = (string)$title;
 214                         if ( $alreadyCached ) {
 215                                 $logger->debug( "Already cached parser output for key '{cachekey}' ('{title}').",
 216                                         [ 'cachekey' => $key, 'title' => $titleStr ] );
 217                                 return self::ERROR_NONE;
 218                         }
 219
 220                         $code = self::storeStashValue(
 221                                 $key,
 222                                 $editInfo->pstContent,
 223                                 $editInfo->output,
 224                                 $editInfo->timestamp,
 225                                 $user
 226                         );
 227
 228                         if ( $code === true ) {
 229                                 $logger->debug( "Cached parser output for key '{cachekey}' ('{title}').",
 230                                         [ 'cachekey' => $key, 'title' => $titleStr ] );
 231                                 return self::ERROR_NONE;
 232                         } elseif ( $code === 'uncacheable' ) {
 233                                 $logger->info(
 234                                         "Uncacheable parser output for key '{cachekey}' ('{title}') [{code}].",
 235                                         [ 'cachekey' => $key, 'title' => $titleStr, 'code' => $code ] );
 236                                 return self::ERROR_UNCACHEABLE;
 237                         } else {
 238                                 $logger->error( "Failed to cache parser output for key '{cachekey}' ('{title}').",
 239                                         [ 'cachekey' => $key, 'title' => $titleStr, 'code' => $code ] );
 240                                 return self::ERROR_CACHE;
 241                         }
 242                 }
 243
 244                 return self::ERROR_PARSE;
 245         }
 246
 247         /**
 248          * Check that a prepared edit is in cache and still up-to-date
 249          *
 250          * This method blocks if the prepared edit is already being rendered,
 251          * waiting until rendering finishes before doing final validity checks.
 252          *
 253          * The cache is rejected if template or file changes are detected.
 254          * Note that foreign template or file transclusions are not checked.
 255          *
 256          * The result is a map (pstContent,output,timestamp) with fields
 257          * extracted directly from WikiPage::prepareContentForEdit().
 258          *
 259          * @param Title $title
 260          * @param Content $content
 261          * @param User $user User to get parser options from
 262          * @return stdClass|bool Returns false on cache miss
 263          */
 264         public static function checkCache( Title $title, Content $content, User $user ) {
 265                 if ( $user->isBot() ) {
 266                         return false; // bots never stash - don't pollute stats
 267                 }
 268
 269                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 270                 $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
 271
 272                 $key = self::getStashKey( $title, self::getContentHash( $content ), $user );
 273                 $editInfo = self::getStashValue( $key );
 274                 if ( !is_object( $editInfo ) ) {
 275                         $start = microtime( true );
 276                         // We ignore user aborts and keep parsing. Block on any prior parsing
 277                         // so as to use its results and make use of the time spent parsing.
 278                         // Skip this logic if there no master connection in case this method
 279                         // is called on an HTTP GET request for some reason.
 280                         $lb = MediaWikiServices::getInstance()->getDBLoadBalancer();
 281                         $dbw = $lb->getAnyOpenConnection( $lb->getWriterIndex() );
 282                         if ( $dbw && $dbw->lock( $key, __METHOD__, 30 ) ) {
 283                                 $editInfo = self::getStashValue( $key );
 284                                 $dbw->unlock( $key, __METHOD__ );
 285                         }
 286
 287                         $timeMs = 1000 * max( 0, microtime( true ) - $start );
 288                         $stats->timing( 'editstash.lock_wait_time', $timeMs );
 289                 }
 290
 291                 if ( !is_object( $editInfo ) || !$editInfo->output ) {
 292                         $stats->increment( 'editstash.cache_misses.no_stash' );
 293                         $logger->debug( "Empty cache for key '$key' ('$title'); user '{$user->getName()}'." );
 294                         return false;
 295                 }
 296
 297                 $age = time() - wfTimestamp( TS_UNIX, $editInfo->output->getCacheTime() );
 298                 if ( $age <= self::PRESUME_FRESH_TTL_SEC ) {
 299                         // Assume nothing changed in this time
 300                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 301                         $logger->debug( "Timestamp-based cache hit for key '$key' (age: $age sec)." );
 302                 } elseif ( isset( $editInfo->edits ) && $editInfo->edits === $user->getEditCount() ) {
 303                         // Logged-in user made no local upload/template edits in the meantime
 304                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 305                         $logger->debug( "Edit count based cache hit for key '$key' (age: $age sec)." );
 306                 } elseif ( $user->isAnon()
 307                         && self::lastEditTime( $user ) < $editInfo->output->getCacheTime()
 308                 ) {
 309                         // Logged-out user made no local upload/template edits in the meantime
 310                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 311                         $logger->debug( "Edit check based cache hit for key '$key' (age: $age sec)." );
 312                 } else {
 313                         // User may have changed included content
 314                         $editInfo = false;
 315                 }
 316
 317                 if ( !$editInfo ) {
 318                         $stats->increment( 'editstash.cache_misses.proven_stale' );
 319                         $logger->info( "Stale cache for key '$key'; old key with outside edits. (age: $age sec)" );
 320                 } elseif ( $editInfo->output->getFlag( 'vary-revision' ) ) {
 321                         // This can be used for the initial parse, e.g. for filters or doEditContent(),
 322                         // but a second parse will be triggered in doEditUpdates(). This is not optimal.
 323                         $logger->info( "Cache for key '$key' ('$title') has vary_revision." );
 324                 } elseif ( $editInfo->output->getFlag( 'vary-revision-id' ) ) {
 325                         // Similar to the above if we didn't guess the ID correctly.
 326                         $logger->info( "Cache for key '$key' ('$title') has vary_revision_id." );
 327                 }
 328
 329                 return $editInfo;
 330         }
 331
 332         /**
 333          * @param User $user
 334          * @return string|null TS_MW timestamp or null
 335          */
 336         private static function lastEditTime( User $user ) {
 337                 $db = wfGetDB( DB_REPLICA );
 338                 $actorQuery = ActorMigration::newMigration()->getWhere( $db, 'rc_user', $user, false );
 339                 $time = $db->selectField(
 340                         [ 'recentchanges' ] + $actorQuery['tables'],
 341                         'MAX(rc_timestamp)',
 342                         [ $actorQuery['conds'] ],
 343                         __METHOD__,
 344                         [],
 345                         $actorQuery['joins']
 346                 );
 347
 348                 return wfTimestampOrNull( TS_MW, $time );
 349         }
 350
 351         /**
 352          * Get hash of the content, factoring in model/format
 353          *
 354          * @param Content $content
 355          * @return string
 356          */
 357         private static function getContentHash( Content $content ) {
 358                 return sha1( implode( "\n", [
 359                         $content->getModel(),
 360                         $content->getDefaultFormat(),
 361                         $content->serialize( $content->getDefaultFormat() )
 362                 ] ) );
 363         }
 364
 365         /**
 366          * Get the temporary prepared edit stash key for a user
 367          *
 368          * This key can be used for caching prepared edits provided:
 369          *   - a) The $user was used for PST options
 370          *   - b) The parser output was made from the PST using cannonical matching options
 371          *
 372          * @param Title $title
 373          * @param string $contentHash Result of getContentHash()
 374          * @param User $user User to get parser options from
 375          * @return string
 376          */
 377         private static function getStashKey( Title $title, $contentHash, User $user ) {
 378                 return ObjectCache::getLocalClusterInstance()->makeKey(
 379                         'stashed-edit-info',
 380                         md5( $title->getPrefixedDBkey() ),
 381                         // Account for the edit model/text
 382                         $contentHash,
 383                         // Account for user name related variables like signatures
 384                         md5( $user->getId() . "\n" . $user->getName() )
 385                 );
 386         }
 387
 388         /**
 389          * @param string $uuid
 390          * @return string
 391          */
 392         private static function getStashParserOutputKey( $uuid ) {
 393                 return ObjectCache::getLocalClusterInstance()->makeKey( 'stashed-edit-output', $uuid );
 394         }
 395
 396         /**
 397          * @param string $key
 398          * @return stdClass|bool Object map (pstContent,output,outputID,timestamp,edits) or false
 399          */
 400         private static function getStashValue( $key ) {
 401                 $cache = ObjectCache::getLocalClusterInstance();
 402
 403                 $stashInfo = $cache->get( $key );
 404                 if ( !is_object( $stashInfo ) ) {
 405                         return false;
 406                 }
 407
 408                 $parserOutputKey = self::getStashParserOutputKey( $stashInfo->outputID );
 409                 $parserOutput = $cache->get( $parserOutputKey );
 410                 if ( $parserOutput instanceof ParserOutput ) {
 411                         $stashInfo->output = $parserOutput;
 412
 413                         return $stashInfo;
 414                 }
 415
 416                 return false;
 417         }
 418
 419         /**
 420          * Build a value to store in memcached based on the PST content and parser output
 421          *
 422          * This makes a simple version of WikiPage::prepareContentForEdit() as stash info
 423          *
 424          * @param string $key
 425          * @param Content $pstContent Pre-Save transformed content
 426          * @param ParserOutput $parserOutput
 427          * @param string $timestamp TS_MW
 428          * @param User $user
 429          * @return string|bool True or an error code
 430          */
 431         private static function storeStashValue(
 432                 $key, Content $pstContent, ParserOutput $parserOutput, $timestamp, User $user
 433         ) {
 434                 // If an item is renewed, mind the cache TTL determined by config and parser functions.
 435                 // Put an upper limit on the TTL for sanity to avoid extreme template/file staleness.
 436                 $age = time() - wfTimestamp( TS_UNIX, $parserOutput->getCacheTime() );
 437                 $ttl = min( $parserOutput->getCacheExpiry() - $age, self::MAX_CACHE_TTL );
 438                 // Avoid extremely stale user signature timestamps (T84843)
 439                 if ( $parserOutput->getFlag( 'user-signature' ) ) {
 440                         $ttl = min( $ttl, self::MAX_SIGNATURE_TTL );
 441                 }
 442
 443                 if ( $ttl <= 0 ) {
 444                         return 'uncacheable'; // low TTL due to a tag, magic word, or signature?
 445                 }
 446
 447                 // Store what is actually needed and split the output into another key (T204742)
 448                 $parseroutputID = md5( $key );
 449                 $stashInfo = (object)[
 450                         'pstContent' => $pstContent,
 451                         'outputID'   => $parseroutputID,
 452                         'timestamp'  => $timestamp,
 453                         'edits'      => $user->getEditCount()
 454                 ];
 455
 456                 $cache = ObjectCache::getLocalClusterInstance();
 457                 $ok = $cache->set( $key, $stashInfo, $ttl );
 458                 if ( $ok ) {
 459                         $ok = $cache->set(
 460                                 self::getStashParserOutputKey( $parseroutputID ),
 461                                 $parserOutput,
 462                                 $ttl
 463                         );
 464                 }
 465
 466                 if ( $ok ) {
 467                         // These blobs can waste slots in low cardinality memcached slabs
 468                         self::pruneExcessStashedEntries( $cache, $user, $key );
 469                 }
 470
 471                 return $ok ? true : 'store_error';
 472         }
 473
 474         /**
 475          * @param BagOStuff $cache
 476          * @param User $user
 477          * @param string $newKey
 478          */
 479         private static function pruneExcessStashedEntries( BagOStuff $cache, User $user, $newKey ) {
 480                 $key = $cache->makeKey( 'stash-edit-recent', sha1( $user->getName() ) );
 481
 482                 $keyList = $cache->get( $key ) ?: [];
 483                 if ( count( $keyList ) >= self::MAX_CACHE_RECENT ) {
 484                         $oldestKey = array_shift( $keyList );
 485                         $cache->delete( $oldestKey );
 486                 }
 487
 488                 $keyList[] = $newKey;
 489                 $cache->set( $key, $keyList, 2 * self::MAX_CACHE_TTL );
 490         }
 491
 492         public function getAllowedParams() {
 493                 return [
 494                         'title' => [
 495                                 ApiBase::PARAM_TYPE => 'string',
 496                                 ApiBase::PARAM_REQUIRED => true
 497                         ],
 498                         'section' => [
 499                                 ApiBase::PARAM_TYPE => 'string',
 500                         ],
 501                         'sectiontitle' => [
 502                                 ApiBase::PARAM_TYPE => 'string'
 503                         ],
 504                         'text' => [
 505                                 ApiBase::PARAM_TYPE => 'text',
 506                                 ApiBase::PARAM_DFLT => null
 507                         ],
 508                         'stashedtexthash' => [
 509                                 ApiBase::PARAM_TYPE => 'string',
 510                                 ApiBase::PARAM_DFLT => null
 511                         ],
 512                         'summary' => [
 513                                 ApiBase::PARAM_TYPE => 'string',
 514                         ],
 515                         'contentmodel' => [
 516                                 ApiBase::PARAM_TYPE => ContentHandler::getContentModels(),
 517                                 ApiBase::PARAM_REQUIRED => true
 518                         ],
 519                         'contentformat' => [
 520                                 ApiBase::PARAM_TYPE => ContentHandler::getAllContentFormats(),
 521                                 ApiBase::PARAM_REQUIRED => true
 522                         ],
 523                         'baserevid' => [
 524                                 ApiBase::PARAM_TYPE => 'integer',
 525                                 ApiBase::PARAM_REQUIRED => true
 526                         ]
 527                 ];
 528         }
 529
 530         public function needsToken() {
 531                 return 'csrf';
 532         }
 533
 534         public function mustBePosted() {
 535                 return true;
 536         }
 537
 538         public function isWriteMode() {
 539                 return true;
 540         }
 541
 542         public function isInternal() {
 543                 return true;
 544         }
 545 }