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
  45         public function execute() {
  46                 $user = $this->getUser();
  47                 $params = $this->extractRequestParams();
  48
  49                 if ( $user->isBot() ) { // sanity
  50                         $this->dieUsage( 'This interface is not supported for bots', 'botsnotsupported' );
  51                 }
  52
  53                 $page = $this->getTitleOrPageId( $params );
  54                 $title = $page->getTitle();
  55
  56                 if ( !ContentHandler::getForModelID( $params['contentmodel'] )
  57                         ->isSupportedFormat( $params['contentformat'] )
  58                 ) {
  59                         $this->dieUsage( 'Unsupported content model/format', 'badmodelformat' );
  60                 }
  61
  62                 // Trim and fix newlines so the key SHA1's match (see RequestContext::getText())
  63                 $text = rtrim( str_replace( "\r\n", "\n", $params['text'] ) );
  64                 $textContent = ContentHandler::makeContent(
  65                         $text, $title, $params['contentmodel'], $params['contentformat'] );
  66
  67                 $page = WikiPage::factory( $title );
  68                 if ( $page->exists() ) {
  69                         // Page exists: get the merged content with the proposed change
  70                         $baseRev = Revision::newFromPageId( $page->getId(), $params['baserevid'] );
  71                         if ( !$baseRev ) {
  72                                 $this->dieUsage( "No revision ID {$params['baserevid']}", 'missingrev' );
  73                         }
  74                         $currentRev = $page->getRevision();
  75                         if ( !$currentRev ) {
  76                                 $this->dieUsage( "No current revision of page ID {$page->getId()}", 'missingrev' );
  77                         }
  78                         // Merge in the new version of the section to get the proposed version
  79                         $editContent = $page->replaceSectionAtRev(
  80                                 $params['section'],
  81                                 $textContent,
  82                                 $params['sectiontitle'],
  83                                 $baseRev->getId()
  84                         );
  85                         if ( !$editContent ) {
  86                                 $this->dieUsage( 'Could not merge updated section.', 'replacefailed' );
  87                         }
  88                         if ( $currentRev->getId() == $baseRev->getId() ) {
  89                                 // Base revision was still the latest; nothing to merge
  90                                 $content = $editContent;
  91                         } else {
  92                                 // Merge the edit into the current version
  93                                 $baseContent = $baseRev->getContent();
  94                                 $currentContent = $currentRev->getContent();
  95                                 if ( !$baseContent || !$currentContent ) {
  96                                         $this->dieUsage( "Missing content for page ID {$page->getId()}", 'missingrev' );
  97                                 }
  98                                 $handler = ContentHandler::getForModelID( $baseContent->getModel() );
  99                                 $content = $handler->merge3( $baseContent, $editContent, $currentContent );
 100                         }
 101                 } else {
 102                         // New pages: use the user-provided content model
 103                         $content = $textContent;
 104                 }
 105
 106                 if ( !$content ) { // merge3() failed
 107                         $this->getResult()->addValue( null,
 108                                 $this->getModuleName(), [ 'status' => 'editconflict' ] );
 109                         return;
 110                 }
 111
 112                 // The user will abort the AJAX request by pressing "save", so ignore that
 113                 ignore_user_abort( true );
 114
 115                 // Use the master DB for fast blocking locks
 116                 $dbw = wfGetDB( DB_MASTER );
 117
 118                 // Get a key based on the source text, format, and user preferences
 119                 $key = self::getStashKey( $title, $content, $user );
 120                 // De-duplicate requests on the same key
 121                 if ( $user->pingLimiter( 'stashedit' ) ) {
 122                         $status = 'ratelimited';
 123                 } elseif ( $dbw->lock( $key, __METHOD__, 1 ) ) {
 124                         $status = self::parseAndStash( $page, $content, $user );
 125                         $dbw->unlock( $key, __METHOD__ );
 126                 } else {
 127                         $status = 'busy';
 128                 }
 129
 130                 $this->getResult()->addValue( null, $this->getModuleName(), [ 'status' => $status ] );
 131         }
 132
 133         /**
 134          * @param WikiPage $page
 135          * @param Content $content
 136          * @param User $user
 137          * @return integer ApiStashEdit::ERROR_* constant
 138          * @since 1.25
 139          */
 140         public static function parseAndStash( WikiPage $page, Content $content, User $user ) {
 141                 $cache = ObjectCache::getLocalClusterInstance();
 142                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 143
 144                 $format = $content->getDefaultFormat();
 145                 $editInfo = $page->prepareContentForEdit( $content, null, $user, $format, false );
 146
 147                 if ( $editInfo && $editInfo->output ) {
 148                         $key = self::getStashKey( $page->getTitle(), $content, $user );
 149
 150                         // Let extensions add ParserOutput metadata or warm other caches
 151                         Hooks::run( 'ParserOutputStashForEdit', [ $page, $content, $editInfo->output ] );
 152
 153                         list( $stashInfo, $ttl ) = self::buildStashValue(
 154                                 $editInfo->pstContent,
 155                                 $editInfo->output,
 156                                 $editInfo->timestamp,
 157                                 $user
 158                         );
 159
 160                         if ( $stashInfo ) {
 161                                 $ok = $cache->set( $key, $stashInfo, $ttl );
 162                                 if ( $ok ) {
 163                                         $logger->debug( "Cached parser output for key '$key'." );
 164                                         return self::ERROR_NONE;
 165                                 } else {
 166                                         $logger->error( "Failed to cache parser output for key '$key'." );
 167                                         return self::ERROR_CACHE;
 168                                 }
 169                         } else {
 170                                 $logger->info( "Uncacheable parser output for key '$key'." );
 171                                 return self::ERROR_UNCACHEABLE;
 172                         }
 173                 }
 174
 175                 return self::ERROR_PARSE;
 176         }
 177
 178         /**
 179          * Attempt to cache PST content and corresponding parser output in passing
 180          *
 181          * This method can be called when the output was already generated for other
 182          * reasons. Parsing should not be done just to call this method, however.
 183          * $pstOpts must be that of the user doing the edit preview. If $pOpts does
 184          * not match the options of WikiPage::makeParserOptions( 'canonical' ), this
 185          * will do nothing. Provided the values are cacheable, they will be stored
 186          * in memcached so that final edit submission might make use of them.
 187          *
 188          * @param Page|Article|WikiPage $page Page title
 189          * @param Content $content Proposed page content
 190          * @param Content $pstContent The result of preSaveTransform() on $content
 191          * @param ParserOutput $pOut The result of getParserOutput() on $pstContent
 192          * @param ParserOptions $pstOpts Options for $pstContent (MUST be for prospective author)
 193          * @param ParserOptions $pOpts Options for $pOut
 194          * @param string $timestamp TS_MW timestamp of parser output generation
 195          * @return bool Success
 196          */
 197         public static function stashEditFromPreview(
 198                 Page $page, Content $content, Content $pstContent, ParserOutput $pOut,
 199                 ParserOptions $pstOpts, ParserOptions $pOpts, $timestamp
 200         ) {
 201                 $cache = ObjectCache::getLocalClusterInstance();
 202                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 203
 204                 // getIsPreview() controls parser function behavior that references things
 205                 // like user/revision that don't exists yet. The user/text should already
 206                 // be set correctly by callers, just double check the preview flag.
 207                 if ( !$pOpts->getIsPreview() ) {
 208                         return false; // sanity
 209                 } elseif ( $pOpts->getIsSectionPreview() ) {
 210                         return false; // short-circuit (need the full content)
 211                 }
 212
 213                 // PST parser options are for the user (handles signatures, etc...)
 214                 $user = $pstOpts->getUser();
 215                 // Get a key based on the source text, format, and user preferences
 216                 $key = self::getStashKey( $page->getTitle(), $content, $user );
 217
 218                 // Parser output options must match cannonical options.
 219                 // Treat some options as matching that are different but don't matter.
 220                 $canonicalPOpts = $page->makeParserOptions( 'canonical' );
 221                 $canonicalPOpts->setIsPreview( true ); // force match
 222                 $canonicalPOpts->setTimestamp( $pOpts->getTimestamp() ); // force match
 223                 if ( !$pOpts->matches( $canonicalPOpts ) ) {
 224                         $logger->info( "Uncacheable preview output for key '$key' (options)." );
 225                         return false;
 226                 }
 227
 228                 // Set the time the output was generated
 229                 $pOut->setCacheTime( wfTimestampNow() );
 230
 231                 // Build a value to cache with a proper TTL
 232                 list( $stashInfo, $ttl ) = self::buildStashValue( $pstContent, $pOut, $timestamp, $user );
 233                 if ( !$stashInfo ) {
 234                         $logger->info( "Uncacheable parser output for key '$key' (rev/TTL)." );
 235                         return false;
 236                 }
 237
 238                 $ok = $cache->set( $key, $stashInfo, $ttl );
 239                 if ( !$ok ) {
 240                         $logger->error( "Failed to cache preview parser output for key '$key'." );
 241                 } else {
 242                         $logger->debug( "Cached preview output for key '$key'." );
 243                 }
 244
 245                 return $ok;
 246         }
 247
 248         /**
 249          * Check that a prepared edit is in cache and still up-to-date
 250          *
 251          * This method blocks if the prepared edit is already being rendered,
 252          * waiting until rendering finishes before doing final validity checks.
 253          *
 254          * The cache is rejected if template or file changes are detected.
 255          * Note that foreign template or file transclusions are not checked.
 256          *
 257          * The result is a map (pstContent,output,timestamp) with fields
 258          * extracted directly from WikiPage::prepareContentForEdit().
 259          *
 260          * @param Title $title
 261          * @param Content $content
 262          * @param User $user User to get parser options from
 263          * @return stdClass|bool Returns false on cache miss
 264          */
 265         public static function checkCache( Title $title, Content $content, User $user ) {
 266                 if ( $user->isBot() ) {
 267                         return false; // bots never stash - don't pollute stats
 268                 }
 269
 270                 $cache = ObjectCache::getLocalClusterInstance();
 271                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 272                 $stats = RequestContext::getMain()->getStats();
 273
 274                 $key = self::getStashKey( $title, $content, $user );
 275                 $editInfo = $cache->get( $key );
 276                 if ( !is_object( $editInfo ) ) {
 277                         $start = microtime( true );
 278                         // We ignore user aborts and keep parsing. Block on any prior parsing
 279                         // so as to use its results and make use of the time spent parsing.
 280                         // Skip this logic if there no master connection in case this method
 281                         // is called on an HTTP GET request for some reason.
 282                         $lb = wfGetLB();
 283                         $dbw = $lb->getAnyOpenConnection( $lb->getWriterIndex() );
 284                         if ( $dbw && $dbw->lock( $key, __METHOD__, 30 ) ) {
 285                                 $editInfo = $cache->get( $key );
 286                                 $dbw->unlock( $key, __METHOD__ );
 287                         }
 288
 289                         $timeMs = 1000 * max( 0, microtime( true ) - $start );
 290                         $stats->timing( 'editstash.lock_wait_time', $timeMs );
 291                 }
 292
 293                 if ( !is_object( $editInfo ) || !$editInfo->output ) {
 294                         $stats->increment( 'editstash.cache_misses.no_stash' );
 295                         $logger->debug( "No cache value for key '$key'." );
 296                         return false;
 297                 }
 298
 299                 $age = time() - wfTimestamp( TS_UNIX, $editInfo->output->getCacheTime() );
 300                 if ( $age <= self::PRESUME_FRESH_TTL_SEC ) {
 301                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 302                         $logger->debug( "Timestamp-based cache hit for key '$key' (age: $age sec)." );
 303                         return $editInfo; // assume nothing changed
 304                 } elseif ( isset( $editInfo->edits ) && $editInfo->edits === $user->getEditCount() ) {
 305                         // Logged-in user made no local upload/template edits in the meantime
 306                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 307                         $logger->debug( "Edit count based cache hit for key '$key' (age: $age sec)." );
 308                         return $editInfo;
 309                 } elseif ( $user->isAnon()
 310                         && self::lastEditTime( $user ) < $editInfo->output->getCacheTime()
 311                 ) {
 312                         // Logged-out user made no local upload/template edits in the meantime
 313                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 314                         $logger->debug( "Edit check based cache hit for key '$key' (age: $age sec)." );
 315                         return $editInfo;
 316                 }
 317
 318                 $dbr = wfGetDB( DB_SLAVE );
 319
 320                 $templates = []; // conditions to find changes/creations
 321                 $templateUses = 0; // expected existing templates
 322                 foreach ( $editInfo->output->getTemplateIds() as $ns => $stuff ) {
 323                         foreach ( $stuff as $dbkey => $revId ) {
 324                                 $templates[(string)$ns][$dbkey] = (int)$revId;
 325                                 ++$templateUses;
 326                         }
 327                 }
 328                 // Check that no templates used in the output changed...
 329                 if ( count( $templates ) ) {
 330                         $res = $dbr->select(
 331                                 'page',
 332                                 [ 'ns' => 'page_namespace', 'dbk' => 'page_title', 'page_latest' ],
 333                                 $dbr->makeWhereFrom2d( $templates, 'page_namespace', 'page_title' ),
 334                                 __METHOD__
 335                         );
 336                         $changed = false;
 337                         foreach ( $res as $row ) {
 338                                 $changed = $changed || ( $row->page_latest != $templates[$row->ns][$row->dbk] );
 339                         }
 340
 341                         if ( $changed || $res->numRows() != $templateUses ) {
 342                                 $stats->increment( 'editstash.cache_misses.proven_stale' );
 343                                 $logger->info( "Stale cache for key '$key'; template changed. (age: $age sec)" );
 344                                 return false;
 345                         }
 346                 }
 347
 348                 $files = []; // conditions to find changes/creations
 349                 foreach ( $editInfo->output->getFileSearchOptions() as $name => $options ) {
 350                         $files[$name] = (string)$options['sha1'];
 351                 }
 352                 // Check that no files used in the output changed...
 353                 if ( count( $files ) ) {
 354                         $res = $dbr->select(
 355                                 'image',
 356                                 [ 'name' => 'img_name', 'img_sha1' ],
 357                                 [ 'img_name' => array_keys( $files ) ],
 358                                 __METHOD__
 359                         );
 360                         $changed = false;
 361                         foreach ( $res as $row ) {
 362                                 $changed = $changed || ( $row->img_sha1 != $files[$row->name] );
 363                         }
 364
 365                         if ( $changed || $res->numRows() != count( $files ) ) {
 366                                 $stats->increment( 'editstash.cache_misses.proven_stale' );
 367                                 $logger->info( "Stale cache for key '$key'; file changed. (age: $age sec)" );
 368                                 return false;
 369                         }
 370                 }
 371
 372                 $stats->increment( 'editstash.cache_hits.proven_fresh' );
 373                 $logger->debug( "Verified cache hit for key '$key' (age: $age sec)." );
 374
 375                 return $editInfo;
 376         }
 377
 378         /**
 379          * @param User $user
 380          * @return string|null TS_MW timestamp or null
 381          */
 382         private static function lastEditTime( User $user ) {
 383                 $time = wfGetDB( DB_SLAVE )->selectField(
 384                         'recentchanges',
 385                         'MAX(rc_timestamp)',
 386                         [ 'rc_user_text' => $user->getName() ],
 387                         __METHOD__
 388                 );
 389
 390                 return wfTimestampOrNull( TS_MW, $time );
 391         }
 392
 393         /**
 394          * Get the temporary prepared edit stash key for a user
 395          *
 396          * This key can be used for caching prepared edits provided:
 397          *   - a) The $user was used for PST options
 398          *   - b) The parser output was made from the PST using cannonical matching options
 399          *
 400          * @param Title $title
 401          * @param Content $content
 402          * @param User $user User to get parser options from
 403          * @return string
 404          */
 405         private static function getStashKey( Title $title, Content $content, User $user ) {
 406                 $hash = sha1( implode( ':', [
 407                         $content->getModel(),
 408                         $content->getDefaultFormat(),
 409                         sha1( $content->serialize( $content->getDefaultFormat() ) ),
 410                         $user->getId() ?: md5( $user->getName() ), // account for user parser options
 411                         $user->getId() ? $user->getDBTouched() : '-' // handle preference change races
 412                 ] ) );
 413
 414                 return wfMemcKey( 'prepared-edit', md5( $title->getPrefixedDBkey() ), $hash );
 415         }
 416
 417         /**
 418          * Build a value to store in memcached based on the PST content and parser output
 419          *
 420          * This makes a simple version of WikiPage::prepareContentForEdit() as stash info
 421          *
 422          * @param Content $pstContent
 423          * @param ParserOutput $parserOutput
 424          * @param string $timestamp TS_MW
 425          * @param User $user
 426          * @return array (stash info array, TTL in seconds) or (null, 0)
 427          */
 428         private static function buildStashValue(
 429                 Content $pstContent, ParserOutput $parserOutput, $timestamp, User $user
 430         ) {
 431                 // If an item is renewed, mind the cache TTL determined by config and parser functions.
 432                 // Put an upper limit on the TTL for sanity to avoid extreme template/file staleness.
 433                 $since = time() - wfTimestamp( TS_UNIX, $parserOutput->getTimestamp() );
 434                 $ttl = min( $parserOutput->getCacheExpiry() - $since, 5 * 60 );
 435
 436                 if ( $ttl > 0 && !$parserOutput->getFlag( 'vary-revision' ) ) {
 437                         // Only store what is actually needed
 438                         $stashInfo = (object)[
 439                                 'pstContent' => $pstContent,
 440                                 'output'     => $parserOutput,
 441                                 'timestamp'  => $timestamp,
 442                                 'edits'      => $user->getEditCount()
 443                         ];
 444                         return [ $stashInfo, $ttl ];
 445                 }
 446
 447                 return [ null, 0 ];
 448         }
 449
 450         public function getAllowedParams() {
 451                 return [
 452                         'title' => [
 453                                 ApiBase::PARAM_TYPE => 'string',
 454                                 ApiBase::PARAM_REQUIRED => true
 455                         ],
 456                         'section' => [
 457                                 ApiBase::PARAM_TYPE => 'string',
 458                         ],
 459                         'sectiontitle' => [
 460                                 ApiBase::PARAM_TYPE => 'string'
 461                         ],
 462                         'text' => [
 463                                 ApiBase::PARAM_TYPE => 'text',
 464                                 ApiBase::PARAM_REQUIRED => true
 465                         ],
 466                         'contentmodel' => [
 467                                 ApiBase::PARAM_TYPE => ContentHandler::getContentModels(),
 468                                 ApiBase::PARAM_REQUIRED => true
 469                         ],
 470                         'contentformat' => [
 471                                 ApiBase::PARAM_TYPE => ContentHandler::getAllContentFormats(),
 472                                 ApiBase::PARAM_REQUIRED => true
 473                         ],
 474                         'baserevid' => [
 475                                 ApiBase::PARAM_TYPE => 'integer',
 476                                 ApiBase::PARAM_REQUIRED => true
 477                         ]
 478                 ];
 479         }
 480
 481         public function needsToken() {
 482                 return 'csrf';
 483         }
 484
 485         public function mustBePosted() {
 486                 return true;
 487         }
 488
 489         public function isWriteMode() {
 490                 return true;
 491         }
 492
 493         public function isInternal() {
 494                 return true;
 495         }
 496 }