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         public function execute() {
  44                 $user = $this->getUser();
  45                 $params = $this->extractRequestParams();
  46
  47                 $page = $this->getTitleOrPageId( $params );
  48                 $title = $page->getTitle();
  49
  50                 if ( !ContentHandler::getForModelID( $params['contentmodel'] )
  51                         ->isSupportedFormat( $params['contentformat'] )
  52                 ) {
  53                         $this->dieUsage( 'Unsupported content model/format', 'badmodelformat' );
  54                 }
  55
  56                 // Trim and fix newlines so the key SHA1's match (see RequestContext::getText())
  57                 $text = rtrim( str_replace( "\r\n", "\n", $params['text'] ) );
  58                 $textContent = ContentHandler::makeContent(
  59                         $text, $title, $params['contentmodel'], $params['contentformat'] );
  60
  61                 $page = WikiPage::factory( $title );
  62                 if ( $page->exists() ) {
  63                         // Page exists: get the merged content with the proposed change
  64                         $baseRev = Revision::newFromPageId( $page->getId(), $params['baserevid'] );
  65                         if ( !$baseRev ) {
  66                                 $this->dieUsage( "No revision ID {$params['baserevid']}", 'missingrev' );
  67                         }
  68                         $currentRev = $page->getRevision();
  69                         if ( !$currentRev ) {
  70                                 $this->dieUsage( "No current revision of page ID {$page->getId()}", 'missingrev' );
  71                         }
  72                         // Merge in the new version of the section to get the proposed version
  73                         $editContent = $page->replaceSectionAtRev(
  74                                 $params['section'],
  75                                 $textContent,
  76                                 $params['sectiontitle'],
  77                                 $baseRev->getId()
  78                         );
  79                         if ( !$editContent ) {
  80                                 $this->dieUsage( 'Could not merge updated section.', 'replacefailed' );
  81                         }
  82                         if ( $currentRev->getId() == $baseRev->getId() ) {
  83                                 // Base revision was still the latest; nothing to merge
  84                                 $content = $editContent;
  85                         } else {
  86                                 // Merge the edit into the current version
  87                                 $baseContent = $baseRev->getContent();
  88                                 $currentContent = $currentRev->getContent();
  89                                 if ( !$baseContent || !$currentContent ) {
  90                                         $this->dieUsage( "Missing content for page ID {$page->getId()}", 'missingrev' );
  91                                 }
  92                                 $handler = ContentHandler::getForModelID( $baseContent->getModel() );
  93                                 $content = $handler->merge3( $baseContent, $editContent, $currentContent );
  94                         }
  95                 } else {
  96                         // New pages: use the user-provided content model
  97                         $content = $textContent;
  98                 }
  99
 100                 if ( !$content ) { // merge3() failed
 101                         $this->getResult()->addValue( null,
 102                                 $this->getModuleName(), [ 'status' => 'editconflict' ] );
 103                         return;
 104                 }
 105
 106                 // The user will abort the AJAX request by pressing "save", so ignore that
 107                 ignore_user_abort( true );
 108
 109                 // Use the master DB for fast blocking locks
 110                 $dbw = wfGetDB( DB_MASTER );
 111
 112                 // Get a key based on the source text, format, and user preferences
 113                 $key = self::getStashKey( $title, $content, $user );
 114                 // De-duplicate requests on the same key
 115                 if ( $user->pingLimiter( 'stashedit' ) ) {
 116                         $status = 'ratelimited';
 117                 } elseif ( $dbw->lock( $key, __METHOD__, 1 ) ) {
 118                         $status = self::parseAndStash( $page, $content, $user );
 119                         $dbw->unlock( $key, __METHOD__ );
 120                 } else {
 121                         $status = 'busy';
 122                 }
 123
 124                 $this->getResult()->addValue( null, $this->getModuleName(), [ 'status' => $status ] );
 125         }
 126
 127         /**
 128          * @param WikiPage $page
 129          * @param Content $content
 130          * @param User $user
 131          * @return integer ApiStashEdit::ERROR_* constant
 132          * @since 1.25
 133          */
 134         public static function parseAndStash( WikiPage $page, Content $content, User $user ) {
 135                 $cache = ObjectCache::getLocalClusterInstance();
 136                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 137
 138                 $format = $content->getDefaultFormat();
 139                 $editInfo = $page->prepareContentForEdit( $content, null, $user, $format, false );
 140
 141                 if ( $editInfo && $editInfo->output ) {
 142                         $key = self::getStashKey( $page->getTitle(), $content, $user );
 143
 144                         // Let extensions add ParserOutput metadata or warm other caches
 145                         Hooks::run( 'ParserOutputStashForEdit', [ $page, $content, $editInfo->output ] );
 146
 147                         list( $stashInfo, $ttl ) = self::buildStashValue(
 148                                 $editInfo->pstContent, $editInfo->output, $editInfo->timestamp
 149                         );
 150
 151                         if ( $stashInfo ) {
 152                                 $ok = $cache->set( $key, $stashInfo, $ttl );
 153                                 if ( $ok ) {
 154
 155                                         $logger->debug( "Cached parser output for key '$key'." );
 156                                         return self::ERROR_NONE;
 157                                 } else {
 158                                         $logger->error( "Failed to cache parser output for key '$key'." );
 159                                         return self::ERROR_CACHE;
 160                                 }
 161                         } else {
 162                                 $logger->info( "Uncacheable parser output for key '$key'." );
 163                                 return self::ERROR_UNCACHEABLE;
 164                         }
 165                 }
 166
 167                 return self::ERROR_PARSE;
 168         }
 169
 170         /**
 171          * Attempt to cache PST content and corresponding parser output in passing
 172          *
 173          * This method can be called when the output was already generated for other
 174          * reasons. Parsing should not be done just to call this method, however.
 175          * $pstOpts must be that of the user doing the edit preview. If $pOpts does
 176          * not match the options of WikiPage::makeParserOptions( 'canonical' ), this
 177          * will do nothing. Provided the values are cacheable, they will be stored
 178          * in memcached so that final edit submission might make use of them.
 179          *
 180          * @param Page|Article|WikiPage $page Page title
 181          * @param Content $content Proposed page content
 182          * @param Content $pstContent The result of preSaveTransform() on $content
 183          * @param ParserOutput $pOut The result of getParserOutput() on $pstContent
 184          * @param ParserOptions $pstOpts Options for $pstContent (MUST be for prospective author)
 185          * @param ParserOptions $pOpts Options for $pOut
 186          * @param string $timestamp TS_MW timestamp of parser output generation
 187          * @return bool Success
 188          */
 189         public static function stashEditFromPreview(
 190                 Page $page, Content $content, Content $pstContent, ParserOutput $pOut,
 191                 ParserOptions $pstOpts, ParserOptions $pOpts, $timestamp
 192         ) {
 193                 $cache = ObjectCache::getLocalClusterInstance();
 194                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 195
 196                 // getIsPreview() controls parser function behavior that references things
 197                 // like user/revision that don't exists yet. The user/text should already
 198                 // be set correctly by callers, just double check the preview flag.
 199                 if ( !$pOpts->getIsPreview() ) {
 200                         return false; // sanity
 201                 } elseif ( $pOpts->getIsSectionPreview() ) {
 202                         return false; // short-circuit (need the full content)
 203                 }
 204
 205                 // PST parser options are for the user (handles signatures, etc...)
 206                 $user = $pstOpts->getUser();
 207                 // Get a key based on the source text, format, and user preferences
 208                 $key = self::getStashKey( $page->getTitle(), $content, $user );
 209
 210                 // Parser output options must match cannonical options.
 211                 // Treat some options as matching that are different but don't matter.
 212                 $canonicalPOpts = $page->makeParserOptions( 'canonical' );
 213                 $canonicalPOpts->setIsPreview( true ); // force match
 214                 $canonicalPOpts->setTimestamp( $pOpts->getTimestamp() ); // force match
 215                 if ( !$pOpts->matches( $canonicalPOpts ) ) {
 216                         $logger->info( "Uncacheable preview output for key '$key' (options)." );
 217                         return false;
 218                 }
 219
 220                 // Build a value to cache with a proper TTL
 221                 list( $stashInfo, $ttl ) = self::buildStashValue( $pstContent, $pOut, $timestamp );
 222                 if ( !$stashInfo ) {
 223                         $logger->info( "Uncacheable parser output for key '$key' (rev/TTL)." );
 224                         return false;
 225                 }
 226
 227                 $ok = $cache->set( $key, $stashInfo, $ttl );
 228                 if ( !$ok ) {
 229                         $logger->error( "Failed to cache preview parser output for key '$key'." );
 230                 } else {
 231                         $logger->debug( "Cached preview output for key '$key'." );
 232                 }
 233
 234                 return $ok;
 235         }
 236
 237         /**
 238          * Check that a prepared edit is in cache and still up-to-date
 239          *
 240          * This method blocks if the prepared edit is already being rendered,
 241          * waiting until rendering finishes before doing final validity checks.
 242          *
 243          * The cache is rejected if template or file changes are detected.
 244          * Note that foreign template or file transclusions are not checked.
 245          *
 246          * The result is a map (pstContent,output,timestamp) with fields
 247          * extracted directly from WikiPage::prepareContentForEdit().
 248          *
 249          * @param Title $title
 250          * @param Content $content
 251          * @param User $user User to get parser options from
 252          * @return stdClass|bool Returns false on cache miss
 253          */
 254         public static function checkCache( Title $title, Content $content, User $user ) {
 255                 $cache = ObjectCache::getLocalClusterInstance();
 256                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 257                 $stats = RequestContext::getMain()->getStats();
 258
 259                 $key = self::getStashKey( $title, $content, $user );
 260                 $editInfo = $cache->get( $key );
 261                 if ( !is_object( $editInfo ) ) {
 262                         $start = microtime( true );
 263                         // We ignore user aborts and keep parsing. Block on any prior parsing
 264                         // so as to use its results and make use of the time spent parsing.
 265                         // Skip this logic if there no master connection in case this method
 266                         // is called on an HTTP GET request for some reason.
 267                         $lb = wfGetLB();
 268                         $dbw = $lb->getAnyOpenConnection( $lb->getWriterIndex() );
 269                         if ( $dbw && $dbw->lock( $key, __METHOD__, 30 ) ) {
 270                                 $editInfo = $cache->get( $key );
 271                                 $dbw->unlock( $key, __METHOD__ );
 272                         }
 273
 274                         $timeMs = 1000 * max( 0, microtime( true ) - $start );
 275                         $stats->timing( 'editstash.lock-wait-time', $timeMs );
 276                 }
 277
 278                 if ( !is_object( $editInfo ) || !$editInfo->output ) {
 279                         $stats->increment( 'editstash.cache-misses' );
 280                         $logger->debug( "No cache value for key '$key'." );
 281                         return false;
 282                 }
 283
 284                 $time = wfTimestamp( TS_UNIX, $editInfo->output->getTimestamp() );
 285                 if ( ( time() - $time ) <= 3 ) {
 286                         $stats->increment( 'editstash.cache-hits' );
 287                         $logger->debug( "Timestamp-based cache hit for key '$key'." );
 288                         return $editInfo; // assume nothing changed
 289                 }
 290
 291                 $dbr = wfGetDB( DB_SLAVE );
 292
 293                 $templates = []; // conditions to find changes/creations
 294                 $templateUses = 0; // expected existing templates
 295                 foreach ( $editInfo->output->getTemplateIds() as $ns => $stuff ) {
 296                         foreach ( $stuff as $dbkey => $revId ) {
 297                                 $templates[(string)$ns][$dbkey] = (int)$revId;
 298                                 ++$templateUses;
 299                         }
 300                 }
 301                 // Check that no templates used in the output changed...
 302                 if ( count( $templates ) ) {
 303                         $res = $dbr->select(
 304                                 'page',
 305                                 [ 'ns' => 'page_namespace', 'dbk' => 'page_title', 'page_latest' ],
 306                                 $dbr->makeWhereFrom2d( $templates, 'page_namespace', 'page_title' ),
 307                                 __METHOD__
 308                         );
 309                         $changed = false;
 310                         foreach ( $res as $row ) {
 311                                 $changed = $changed || ( $row->page_latest != $templates[$row->ns][$row->dbk] );
 312                         }
 313
 314                         if ( $changed || $res->numRows() != $templateUses ) {
 315                                 $stats->increment( 'editstash.cache-misses' );
 316                                 $logger->info( "Stale cache for key '$key'; template changed." );
 317                                 return false;
 318                         }
 319                 }
 320
 321                 $files = []; // conditions to find changes/creations
 322                 foreach ( $editInfo->output->getFileSearchOptions() as $name => $options ) {
 323                         $files[$name] = (string)$options['sha1'];
 324                 }
 325                 // Check that no files used in the output changed...
 326                 if ( count( $files ) ) {
 327                         $res = $dbr->select(
 328                                 'image',
 329                                 [ 'name' => 'img_name', 'img_sha1' ],
 330                                 [ 'img_name' => array_keys( $files ) ],
 331                                 __METHOD__
 332                         );
 333                         $changed = false;
 334                         foreach ( $res as $row ) {
 335                                 $changed = $changed || ( $row->img_sha1 != $files[$row->name] );
 336                         }
 337
 338                         if ( $changed || $res->numRows() != count( $files ) ) {
 339                                 $stats->increment( 'editstash.cache-misses' );
 340                                 $logger->info( "Stale cache for key '$key'; file changed." );
 341                                 return false;
 342                         }
 343                 }
 344
 345                 $stats->increment( 'editstash.cache-hits' );
 346                 $logger->debug( "Cache hit for key '$key'." );
 347
 348                 return $editInfo;
 349         }
 350
 351         /**
 352          * Get the temporary prepared edit stash key for a user
 353          *
 354          * This key can be used for caching prepared edits provided:
 355          *   - a) The $user was used for PST options
 356          *   - b) The parser output was made from the PST using cannonical matching options
 357          *
 358          * @param Title $title
 359          * @param Content $content
 360          * @param User $user User to get parser options from
 361          * @return string
 362          */
 363         protected static function getStashKey( Title $title, Content $content, User $user ) {
 364                 $hash = sha1( implode( ':', [
 365                         $content->getModel(),
 366                         $content->getDefaultFormat(),
 367                         sha1( $content->serialize( $content->getDefaultFormat() ) ),
 368                         $user->getId() ?: md5( $user->getName() ), // account for user parser options
 369                         $user->getId() ? $user->getDBTouched() : '-' // handle preference change races
 370                 ] ) );
 371
 372                 return wfMemcKey( 'prepared-edit', md5( $title->getPrefixedDBkey() ), $hash );
 373         }
 374
 375         /**
 376          * Build a value to store in memcached based on the PST content and parser output
 377          *
 378          * This makes a simple version of WikiPage::prepareContentForEdit() as stash info
 379          *
 380          * @param Content $pstContent
 381          * @param ParserOutput $parserOutput
 382          * @param string $timestamp TS_MW
 383          * @return array (stash info array, TTL in seconds) or (null, 0)
 384          */
 385         protected static function buildStashValue(
 386                 Content $pstContent, ParserOutput $parserOutput, $timestamp
 387         ) {
 388                 // If an item is renewed, mind the cache TTL determined by config and parser functions
 389                 $since = time() - wfTimestamp( TS_UNIX, $parserOutput->getTimestamp() );
 390                 $ttl = min( $parserOutput->getCacheExpiry() - $since, 5 * 60 );
 391
 392                 if ( $ttl > 0 && !$parserOutput->getFlag( 'vary-revision' ) ) {
 393                         // Only store what is actually needed
 394                         $stashInfo = (object)[
 395                                 'pstContent' => $pstContent,
 396                                 'output'     => $parserOutput,
 397                                 'timestamp'  => $timestamp
 398                         ];
 399                         return [ $stashInfo, $ttl ];
 400                 }
 401
 402                 return [ null, 0 ];
 403         }
 404
 405         public function getAllowedParams() {
 406                 return [
 407                         'title' => [
 408                                 ApiBase::PARAM_TYPE => 'string',
 409                                 ApiBase::PARAM_REQUIRED => true
 410                         ],
 411                         'section' => [
 412                                 ApiBase::PARAM_TYPE => 'string',
 413                         ],
 414                         'sectiontitle' => [
 415                                 ApiBase::PARAM_TYPE => 'string'
 416                         ],
 417                         'text' => [
 418                                 ApiBase::PARAM_TYPE => 'text',
 419                                 ApiBase::PARAM_REQUIRED => true
 420                         ],
 421                         'contentmodel' => [
 422                                 ApiBase::PARAM_TYPE => ContentHandler::getContentModels(),
 423                                 ApiBase::PARAM_REQUIRED => true
 424                         ],
 425                         'contentformat' => [
 426                                 ApiBase::PARAM_TYPE => ContentHandler::getAllContentFormats(),
 427                                 ApiBase::PARAM_REQUIRED => true
 428                         ],
 429                         'baserevid' => [
 430                                 ApiBase::PARAM_TYPE => 'integer',
 431                                 ApiBase::PARAM_REQUIRED => true
 432                         ]
 433                 ];
 434         }
 435
 436         public function needsToken() {
 437                 return 'csrf';
 438         }
 439
 440         public function mustBePosted() {
 441                 return true;
 442         }
 443
 444         public function isWriteMode() {
 445                 return true;
 446         }
 447
 448         public function isInternal() {
 449                 return true;
 450         }
 451 }