includes/api/ApiStashEdit.php

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