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