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