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
  48         public function execute() {
  49                 $user = $this->getUser();
  50                 $params = $this->extractRequestParams();
  51
  52                 if ( $user->isBot() ) { // sanity
  53                         $this->dieWithError( 'apierror-botsnotsupported' );
  54                 }
  55
  56                 $cache = ObjectCache::getLocalClusterInstance();
  57                 $page = $this->getTitleOrPageId( $params );
  58                 $title = $page->getTitle();
  59
  60                 if ( !ContentHandler::getForModelID( $params['contentmodel'] )
  61                         ->isSupportedFormat( $params['contentformat'] )
  62                 ) {
  63                         $this->dieWithError(
  64                                 [ 'apierror-badformat-generic', $params['contentformat'], $params['contentmodel'] ],
  65                                 'badmodelformat'
  66                         );
  67                 }
  68
  69                 $this->requireAtLeastOneParameter( $params, 'stashedtexthash', 'text' );
  70
  71                 $text = null;
  72                 $textHash = null;
  73                 if ( strlen( $params['stashedtexthash'] ) ) {
  74                         // Load from cache since the client indicates the text is the same as last stash
  75                         $textHash = $params['stashedtexthash'];
  76                         $textKey = $cache->makeKey( 'stashedit', 'text', $textHash );
  77                         $text = $cache->get( $textKey );
  78                         if ( !is_string( $text ) ) {
  79                                 $this->dieWithError( 'apierror-stashedit-missingtext', 'missingtext' );
  80                         }
  81                 } elseif ( $params['text'] !== null ) {
  82                         // Trim and fix newlines so the key SHA1's match (see WebRequest::getText())
  83                         $text = rtrim( str_replace( "\r\n", "\n", $params['text'] ) );
  84                         $textHash = sha1( $text );
  85                 } else {
  86                         $this->dieWithError( [
  87                                 'apierror-missingparam-at-least-one-of',
  88                                 Message::listParam( [ '<var>stashedtexthash</var>', '<var>text</var>' ] ),
  89                                 2,
  90                         ], 'missingparam' );
  91                 }
  92
  93                 $textContent = ContentHandler::makeContent(
  94                         $text, $title, $params['contentmodel'], $params['contentformat'] );
  95
  96                 $page = WikiPage::factory( $title );
  97                 if ( $page->exists() ) {
  98                         // Page exists: get the merged content with the proposed change
  99                         $baseRev = Revision::newFromPageId( $page->getId(), $params['baserevid'] );
 100                         if ( !$baseRev ) {
 101                                 $this->dieWithError( [ 'apierror-nosuchrevid', $params['baserevid'] ] );
 102                         }
 103                         $currentRev = $page->getRevision();
 104                         if ( !$currentRev ) {
 105                                 $this->dieWithError( [ 'apierror-missingrev-pageid', $page->getId() ], 'missingrev' );
 106                         }
 107                         // Merge in the new version of the section to get the proposed version
 108                         $editContent = $page->replaceSectionAtRev(
 109                                 $params['section'],
 110                                 $textContent,
 111                                 $params['sectiontitle'],
 112                                 $baseRev->getId()
 113                         );
 114                         if ( !$editContent ) {
 115                                 $this->dieWithError( 'apierror-sectionreplacefailed', 'replacefailed' );
 116                         }
 117                         if ( $currentRev->getId() == $baseRev->getId() ) {
 118                                 // Base revision was still the latest; nothing to merge
 119                                 $content = $editContent;
 120                         } else {
 121                                 // Merge the edit into the current version
 122                                 $baseContent = $baseRev->getContent();
 123                                 $currentContent = $currentRev->getContent();
 124                                 if ( !$baseContent || !$currentContent ) {
 125                                         $this->dieWithError( [ 'apierror-missingcontent-pageid', $page->getId() ], 'missingrev' );
 126                                 }
 127                                 $handler = ContentHandler::getForModelID( $baseContent->getModel() );
 128                                 $content = $handler->merge3( $baseContent, $editContent, $currentContent );
 129                         }
 130                 } else {
 131                         // New pages: use the user-provided content model
 132                         $content = $textContent;
 133                 }
 134
 135                 if ( !$content ) { // merge3() failed
 136                         $this->getResult()->addValue( null,
 137                                 $this->getModuleName(), [ 'status' => 'editconflict' ] );
 138                         return;
 139                 }
 140
 141                 // The user will abort the AJAX request by pressing "save", so ignore that
 142                 ignore_user_abort( true );
 143
 144                 if ( $user->pingLimiter( 'stashedit' ) ) {
 145                         $status = 'ratelimited';
 146                 } else {
 147                         $status = self::parseAndStash( $page, $content, $user, $params['summary'] );
 148                         $textKey = $cache->makeKey( 'stashedit', 'text', $textHash );
 149                         $cache->set( $textKey, $text, self::MAX_CACHE_TTL );
 150                 }
 151
 152                 $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
 153                 $stats->increment( "editstash.cache_stores.$status" );
 154
 155                 $this->getResult()->addValue(
 156                         null,
 157                         $this->getModuleName(),
 158                         [
 159                                 'status' => $status,
 160                                 'texthash' => $textHash
 161                         ]
 162                 );
 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                 $cache = ObjectCache::getLocalClusterInstance();
 175                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 176
 177                 $title = $page->getTitle();
 178                 $key = self::getStashKey( $title, self::getContentHash( $content ), $user );
 179
 180                 // Use the master DB for fast blocking locks
 181                 $dbw = wfGetDB( DB_MASTER );
 182                 if ( !$dbw->lock( $key, __METHOD__, 1 ) ) {
 183                         // De-duplicate requests on the same key
 184                         return self::ERROR_BUSY;
 185                 }
 186                 /** @noinspection PhpUnusedLocalVariableInspection */
 187                 $unlocker = new ScopedCallback( function () use ( $dbw, $key ) {
 188                         $dbw->unlock( $key, __METHOD__ );
 189                 } );
 190
 191                 $cutoffTime = time() - self::PRESUME_FRESH_TTL_SEC;
 192
 193                 // Reuse any freshly build matching edit stash cache
 194                 $editInfo = $cache->get( $key );
 195                 if ( $editInfo && wfTimestamp( TS_UNIX, $editInfo->timestamp ) >= $cutoffTime ) {
 196                         $alreadyCached = true;
 197                 } else {
 198                         $format = $content->getDefaultFormat();
 199                         $editInfo = $page->prepareContentForEdit( $content, null, $user, $format, false );
 200                         $alreadyCached = false;
 201                 }
 202
 203                 if ( $editInfo && $editInfo->output ) {
 204                         // Let extensions add ParserOutput metadata or warm other caches
 205                         Hooks::run( 'ParserOutputStashForEdit',
 206                                 [ $page, $content, $editInfo->output, $summary, $user ] );
 207
 208                         if ( $alreadyCached ) {
 209                                 $logger->debug( "Already cached parser output for key '$key' ('$title')." );
 210                                 return self::ERROR_NONE;
 211                         }
 212
 213                         list( $stashInfo, $ttl, $code ) = self::buildStashValue(
 214                                 $editInfo->pstContent,
 215                                 $editInfo->output,
 216                                 $editInfo->timestamp,
 217                                 $user
 218                         );
 219
 220                         if ( $stashInfo ) {
 221                                 $ok = $cache->set( $key, $stashInfo, $ttl );
 222                                 if ( $ok ) {
 223                                         $logger->debug( "Cached parser output for key '$key' ('$title')." );
 224                                         return self::ERROR_NONE;
 225                                 } else {
 226                                         $logger->error( "Failed to cache parser output for key '$key' ('$title')." );
 227                                         return self::ERROR_CACHE;
 228                                 }
 229                         } else {
 230                                 $logger->info( "Uncacheable parser output for key '$key' ('$title') [$code]." );
 231                                 return self::ERROR_UNCACHEABLE;
 232                         }
 233                 }
 234
 235                 return self::ERROR_PARSE;
 236         }
 237
 238         /**
 239          * Check that a prepared edit is in cache and still up-to-date
 240          *
 241          * This method blocks if the prepared edit is already being rendered,
 242          * waiting until rendering finishes before doing final validity checks.
 243          *
 244          * The cache is rejected if template or file changes are detected.
 245          * Note that foreign template or file transclusions are not checked.
 246          *
 247          * The result is a map (pstContent,output,timestamp) with fields
 248          * extracted directly from WikiPage::prepareContentForEdit().
 249          *
 250          * @param Title $title
 251          * @param Content $content
 252          * @param User $user User to get parser options from
 253          * @return stdClass|bool Returns false on cache miss
 254          */
 255         public static function checkCache( Title $title, Content $content, User $user ) {
 256                 if ( $user->isBot() ) {
 257                         return false; // bots never stash - don't pollute stats
 258                 }
 259
 260                 $cache = ObjectCache::getLocalClusterInstance();
 261                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 262                 $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
 263
 264                 $key = self::getStashKey( $title, self::getContentHash( $content ), $user );
 265                 $editInfo = $cache->get( $key );
 266                 if ( !is_object( $editInfo ) ) {
 267                         $start = microtime( true );
 268                         // We ignore user aborts and keep parsing. Block on any prior parsing
 269                         // so as to use its results and make use of the time spent parsing.
 270                         // Skip this logic if there no master connection in case this method
 271                         // is called on an HTTP GET request for some reason.
 272                         $lb = MediaWikiServices::getInstance()->getDBLoadBalancer();
 273                         $dbw = $lb->getAnyOpenConnection( $lb->getWriterIndex() );
 274                         if ( $dbw && $dbw->lock( $key, __METHOD__, 30 ) ) {
 275                                 $editInfo = $cache->get( $key );
 276                                 $dbw->unlock( $key, __METHOD__ );
 277                         }
 278
 279                         $timeMs = 1000 * max( 0, microtime( true ) - $start );
 280                         $stats->timing( 'editstash.lock_wait_time', $timeMs );
 281                 }
 282
 283                 if ( !is_object( $editInfo ) || !$editInfo->output ) {
 284                         $stats->increment( 'editstash.cache_misses.no_stash' );
 285                         $logger->debug( "Empty cache for key '$key' ('$title'); user '{$user->getName()}'." );
 286                         return false;
 287                 }
 288
 289                 $age = time() - wfTimestamp( TS_UNIX, $editInfo->output->getCacheTime() );
 290                 if ( $age <= self::PRESUME_FRESH_TTL_SEC ) {
 291                         // Assume nothing changed in this time
 292                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 293                         $logger->debug( "Timestamp-based cache hit for key '$key' (age: $age sec)." );
 294                 } elseif ( isset( $editInfo->edits ) && $editInfo->edits === $user->getEditCount() ) {
 295                         // Logged-in user made no local upload/template edits in the meantime
 296                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 297                         $logger->debug( "Edit count based cache hit for key '$key' (age: $age sec)." );
 298                 } elseif ( $user->isAnon()
 299                         && self::lastEditTime( $user ) < $editInfo->output->getCacheTime()
 300                 ) {
 301                         // Logged-out user made no local upload/template edits in the meantime
 302                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 303                         $logger->debug( "Edit check based cache hit for key '$key' (age: $age sec)." );
 304                 } else {
 305                         // User may have changed included content
 306                         $editInfo = false;
 307                 }
 308
 309                 if ( !$editInfo ) {
 310                         $stats->increment( 'editstash.cache_misses.proven_stale' );
 311                         $logger->info( "Stale cache for key '$key'; old key with outside edits. (age: $age sec)" );
 312                 } elseif ( $editInfo->output->getFlag( 'vary-revision' ) ) {
 313                         // This can be used for the initial parse, e.g. for filters or doEditContent(),
 314                         // but a second parse will be triggered in doEditUpdates(). This is not optimal.
 315                         $logger->info( "Cache for key '$key' ('$title') has vary_revision." );
 316                 } elseif ( $editInfo->output->getFlag( 'vary-revision-id' ) ) {
 317                         // Similar to the above if we didn't guess the ID correctly.
 318                         $logger->info( "Cache for key '$key' ('$title') has vary_revision_id." );
 319                 }
 320
 321                 return $editInfo;
 322         }
 323
 324         /**
 325          * @param User $user
 326          * @return string|null TS_MW timestamp or null
 327          */
 328         private static function lastEditTime( User $user ) {
 329                 $time = wfGetDB( DB_REPLICA )->selectField(
 330                         'recentchanges',
 331                         'MAX(rc_timestamp)',
 332                         [ 'rc_user_text' => $user->getName() ],
 333                         __METHOD__
 334                 );
 335
 336                 return wfTimestampOrNull( TS_MW, $time );
 337         }
 338
 339         /**
 340          * Get hash of the content, factoring in model/format
 341          *
 342          * @param Content $content
 343          * @return string
 344          */
 345         private static function getContentHash( Content $content ) {
 346                 return sha1( implode( "\n", [
 347                         $content->getModel(),
 348                         $content->getDefaultFormat(),
 349                         $content->serialize( $content->getDefaultFormat() )
 350                 ] ) );
 351         }
 352
 353         /**
 354          * Get the temporary prepared edit stash key for a user
 355          *
 356          * This key can be used for caching prepared edits provided:
 357          *   - a) The $user was used for PST options
 358          *   - b) The parser output was made from the PST using cannonical matching options
 359          *
 360          * @param Title $title
 361          * @param string $contentHash Result of getContentHash()
 362          * @param User $user User to get parser options from
 363          * @return string
 364          */
 365         private static function getStashKey( Title $title, $contentHash, User $user ) {
 366                 return ObjectCache::getLocalClusterInstance()->makeKey(
 367                         'prepared-edit',
 368                         md5( $title->getPrefixedDBkey() ),
 369                         // Account for the edit model/text
 370                         $contentHash,
 371                         // Account for user name related variables like signatures
 372                         md5( $user->getId() . "\n" . $user->getName() )
 373                 );
 374         }
 375
 376         /**
 377          * Build a value to store in memcached based on the PST content and parser output
 378          *
 379          * This makes a simple version of WikiPage::prepareContentForEdit() as stash info
 380          *
 381          * @param Content $pstContent Pre-Save transformed content
 382          * @param ParserOutput $parserOutput
 383          * @param string $timestamp TS_MW
 384          * @param User $user
 385          * @return array (stash info array, TTL in seconds, info code) or (null, 0, info code)
 386          */
 387         private static function buildStashValue(
 388                 Content $pstContent, ParserOutput $parserOutput, $timestamp, User $user
 389         ) {
 390                 // If an item is renewed, mind the cache TTL determined by config and parser functions.
 391                 // Put an upper limit on the TTL for sanity to avoid extreme template/file staleness.
 392                 $since = time() - wfTimestamp( TS_UNIX, $parserOutput->getTimestamp() );
 393                 $ttl = min( $parserOutput->getCacheExpiry() - $since, self::MAX_CACHE_TTL );
 394                 if ( $ttl <= 0 ) {
 395                         return [ null, 0, 'no_ttl' ];
 396                 }
 397
 398                 // Only store what is actually needed
 399                 $stashInfo = (object)[
 400                         'pstContent' => $pstContent,
 401                         'output'     => $parserOutput,
 402                         'timestamp'  => $timestamp,
 403                         'edits'      => $user->getEditCount()
 404                 ];
 405
 406                 return [ $stashInfo, $ttl, 'ok' ];
 407         }
 408
 409         public function getAllowedParams() {
 410                 return [
 411                         'title' => [
 412                                 ApiBase::PARAM_TYPE => 'string',
 413                                 ApiBase::PARAM_REQUIRED => true
 414                         ],
 415                         'section' => [
 416                                 ApiBase::PARAM_TYPE => 'string',
 417                         ],
 418                         'sectiontitle' => [
 419                                 ApiBase::PARAM_TYPE => 'string'
 420                         ],
 421                         'text' => [
 422                                 ApiBase::PARAM_TYPE => 'text',
 423                                 ApiBase::PARAM_DFLT => null
 424                         ],
 425                         'stashedtexthash' => [
 426                                 ApiBase::PARAM_TYPE => 'string',
 427                                 ApiBase::PARAM_DFLT => null
 428                         ],
 429                         'summary' => [
 430                                 ApiBase::PARAM_TYPE => 'string',
 431                         ],
 432                         'contentmodel' => [
 433                                 ApiBase::PARAM_TYPE => ContentHandler::getContentModels(),
 434                                 ApiBase::PARAM_REQUIRED => true
 435                         ],
 436                         'contentformat' => [
 437                                 ApiBase::PARAM_TYPE => ContentHandler::getAllContentFormats(),
 438                                 ApiBase::PARAM_REQUIRED => true
 439                         ],
 440                         'baserevid' => [
 441                                 ApiBase::PARAM_TYPE => 'integer',
 442                                 ApiBase::PARAM_REQUIRED => true
 443                         ]
 444                 ];
 445         }
 446
 447         public function needsToken() {
 448                 return 'csrf';
 449         }
 450
 451         public function mustBePosted() {
 452                 return true;
 453         }
 454
 455         public function isWriteMode() {
 456                 return true;
 457         }
 458
 459         public function isInternal() {
 460                 return true;
 461         }
 462 }