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