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,
 151                                 $editInfo->output,
 152                                 $editInfo->timestamp,
 153                                 $user
 154                         );
 155
 156                         if ( $stashInfo ) {
 157                                 $ok = $cache->set( $key, $stashInfo, $ttl );
 158                                 if ( $ok ) {
 159                                         $logger->debug( "Cached parser output for key '$key'." );
 160                                         return self::ERROR_NONE;
 161                                 } else {
 162                                         $logger->error( "Failed to cache parser output for key '$key'." );
 163                                         return self::ERROR_CACHE;
 164                                 }
 165                         } else {
 166                                 $logger->info( "Uncacheable parser output for key '$key'." );
 167                                 return self::ERROR_UNCACHEABLE;
 168                         }
 169                 }
 170
 171                 return self::ERROR_PARSE;
 172         }
 173
 174         /**
 175          * Attempt to cache PST content and corresponding parser output in passing
 176          *
 177          * This method can be called when the output was already generated for other
 178          * reasons. Parsing should not be done just to call this method, however.
 179          * $pstOpts must be that of the user doing the edit preview. If $pOpts does
 180          * not match the options of WikiPage::makeParserOptions( 'canonical' ), this
 181          * will do nothing. Provided the values are cacheable, they will be stored
 182          * in memcached so that final edit submission might make use of them.
 183          *
 184          * @param Page|Article|WikiPage $page Page title
 185          * @param Content $content Proposed page content
 186          * @param Content $pstContent The result of preSaveTransform() on $content
 187          * @param ParserOutput $pOut The result of getParserOutput() on $pstContent
 188          * @param ParserOptions $pstOpts Options for $pstContent (MUST be for prospective author)
 189          * @param ParserOptions $pOpts Options for $pOut
 190          * @param string $timestamp TS_MW timestamp of parser output generation
 191          * @return bool Success
 192          */
 193         public static function stashEditFromPreview(
 194                 Page $page, Content $content, Content $pstContent, ParserOutput $pOut,
 195                 ParserOptions $pstOpts, ParserOptions $pOpts, $timestamp
 196         ) {
 197                 $cache = ObjectCache::getLocalClusterInstance();
 198                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 199
 200                 // getIsPreview() controls parser function behavior that references things
 201                 // like user/revision that don't exists yet. The user/text should already
 202                 // be set correctly by callers, just double check the preview flag.
 203                 if ( !$pOpts->getIsPreview() ) {
 204                         return false; // sanity
 205                 } elseif ( $pOpts->getIsSectionPreview() ) {
 206                         return false; // short-circuit (need the full content)
 207                 }
 208
 209                 // PST parser options are for the user (handles signatures, etc...)
 210                 $user = $pstOpts->getUser();
 211                 // Get a key based on the source text, format, and user preferences
 212                 $key = self::getStashKey( $page->getTitle(), $content, $user );
 213
 214                 // Parser output options must match cannonical options.
 215                 // Treat some options as matching that are different but don't matter.
 216                 $canonicalPOpts = $page->makeParserOptions( 'canonical' );
 217                 $canonicalPOpts->setIsPreview( true ); // force match
 218                 $canonicalPOpts->setTimestamp( $pOpts->getTimestamp() ); // force match
 219                 if ( !$pOpts->matches( $canonicalPOpts ) ) {
 220                         $logger->info( "Uncacheable preview output for key '$key' (options)." );
 221                         return false;
 222                 }
 223
 224                 // Set the time the output was generated
 225                 $pOut->setCacheTime( wfTimestampNow() );
 226
 227                 // Build a value to cache with a proper TTL
 228                 list( $stashInfo, $ttl ) = self::buildStashValue( $pstContent, $pOut, $timestamp, $user );
 229                 if ( !$stashInfo ) {
 230                         $logger->info( "Uncacheable parser output for key '$key' (rev/TTL)." );
 231                         return false;
 232                 }
 233
 234                 $ok = $cache->set( $key, $stashInfo, $ttl );
 235                 if ( !$ok ) {
 236                         $logger->error( "Failed to cache preview parser output for key '$key'." );
 237                 } else {
 238                         $logger->debug( "Cached preview output for key '$key'." );
 239                 }
 240
 241                 return $ok;
 242         }
 243
 244         /**
 245          * Check that a prepared edit is in cache and still up-to-date
 246          *
 247          * This method blocks if the prepared edit is already being rendered,
 248          * waiting until rendering finishes before doing final validity checks.
 249          *
 250          * The cache is rejected if template or file changes are detected.
 251          * Note that foreign template or file transclusions are not checked.
 252          *
 253          * The result is a map (pstContent,output,timestamp) with fields
 254          * extracted directly from WikiPage::prepareContentForEdit().
 255          *
 256          * @param Title $title
 257          * @param Content $content
 258          * @param User $user User to get parser options from
 259          * @return stdClass|bool Returns false on cache miss
 260          */
 261         public static function checkCache( Title $title, Content $content, User $user ) {
 262                 $cache = ObjectCache::getLocalClusterInstance();
 263                 $logger = LoggerFactory::getInstance( 'StashEdit' );
 264                 $stats = RequestContext::getMain()->getStats();
 265
 266                 $key = self::getStashKey( $title, $content, $user );
 267                 $editInfo = $cache->get( $key );
 268                 if ( !is_object( $editInfo ) ) {
 269                         $start = microtime( true );
 270                         // We ignore user aborts and keep parsing. Block on any prior parsing
 271                         // so as to use its results and make use of the time spent parsing.
 272                         // Skip this logic if there no master connection in case this method
 273                         // is called on an HTTP GET request for some reason.
 274                         $lb = wfGetLB();
 275                         $dbw = $lb->getAnyOpenConnection( $lb->getWriterIndex() );
 276                         if ( $dbw && $dbw->lock( $key, __METHOD__, 30 ) ) {
 277                                 $editInfo = $cache->get( $key );
 278                                 $dbw->unlock( $key, __METHOD__ );
 279                         }
 280
 281                         $timeMs = 1000 * max( 0, microtime( true ) - $start );
 282                         $stats->timing( 'editstash.lock_wait_time', $timeMs );
 283                 }
 284
 285                 if ( !is_object( $editInfo ) || !$editInfo->output ) {
 286                         $stats->increment( 'editstash.cache_misses.no_stash' );
 287                         $logger->debug( "No cache value for key '$key'." );
 288                         return false;
 289                 }
 290
 291                 $age = time() - wfTimestamp( TS_UNIX, $editInfo->output->getCacheTime() );
 292                 if ( $age <= self::PRESUME_FRESH_TTL_SEC ) {
 293                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 294                         $logger->debug( "Timestamp-based cache hit for key '$key' (age: $age sec)." );
 295                         return $editInfo; // assume nothing changed
 296                 } elseif ( isset( $editInfo->edits ) && $editInfo->edits === $user->getEditCount() ) {
 297                         $stats->increment( 'editstash.cache_hits.presumed_fresh' );
 298                         $logger->debug( "Edit count based cache hit for key '$key' (age: $age sec)." );
 299                         return $editInfo; // use made no local upload/template edits in the meantime
 300                 }
 301
 302                 $dbr = wfGetDB( DB_SLAVE );
 303
 304                 $templates = []; // conditions to find changes/creations
 305                 $templateUses = 0; // expected existing templates
 306                 foreach ( $editInfo->output->getTemplateIds() as $ns => $stuff ) {
 307                         foreach ( $stuff as $dbkey => $revId ) {
 308                                 $templates[(string)$ns][$dbkey] = (int)$revId;
 309                                 ++$templateUses;
 310                         }
 311                 }
 312                 // Check that no templates used in the output changed...
 313                 if ( count( $templates ) ) {
 314                         $res = $dbr->select(
 315                                 'page',
 316                                 [ 'ns' => 'page_namespace', 'dbk' => 'page_title', 'page_latest' ],
 317                                 $dbr->makeWhereFrom2d( $templates, 'page_namespace', 'page_title' ),
 318                                 __METHOD__
 319                         );
 320                         $changed = false;
 321                         foreach ( $res as $row ) {
 322                                 $changed = $changed || ( $row->page_latest != $templates[$row->ns][$row->dbk] );
 323                         }
 324
 325                         if ( $changed || $res->numRows() != $templateUses ) {
 326                                 $stats->increment( 'editstash.cache_misses.proven_stale' );
 327                                 $logger->info( "Stale cache for key '$key'; template changed. (age: $age sec)" );
 328                                 return false;
 329                         }
 330                 }
 331
 332                 $files = []; // conditions to find changes/creations
 333                 foreach ( $editInfo->output->getFileSearchOptions() as $name => $options ) {
 334                         $files[$name] = (string)$options['sha1'];
 335                 }
 336                 // Check that no files used in the output changed...
 337                 if ( count( $files ) ) {
 338                         $res = $dbr->select(
 339                                 'image',
 340                                 [ 'name' => 'img_name', 'img_sha1' ],
 341                                 [ 'img_name' => array_keys( $files ) ],
 342                                 __METHOD__
 343                         );
 344                         $changed = false;
 345                         foreach ( $res as $row ) {
 346                                 $changed = $changed || ( $row->img_sha1 != $files[$row->name] );
 347                         }
 348
 349                         if ( $changed || $res->numRows() != count( $files ) ) {
 350                                 $stats->increment( 'editstash.cache_misses.proven_stale' );
 351                                 $logger->info( "Stale cache for key '$key'; file changed. (age: $age sec)" );
 352                                 return false;
 353                         }
 354                 }
 355
 356                 $stats->increment( 'editstash.cache_hits.proven_fresh' );
 357                 $logger->debug( "Verified cache hit for key '$key' (age: $age sec)." );
 358
 359                 return $editInfo;
 360         }
 361
 362         /**
 363          * Get the temporary prepared edit stash key for a user
 364          *
 365          * This key can be used for caching prepared edits provided:
 366          *   - a) The $user was used for PST options
 367          *   - b) The parser output was made from the PST using cannonical matching options
 368          *
 369          * @param Title $title
 370          * @param Content $content
 371          * @param User $user User to get parser options from
 372          * @return string
 373          */
 374         protected static function getStashKey( Title $title, Content $content, User $user ) {
 375                 $hash = sha1( implode( ':', [
 376                         $content->getModel(),
 377                         $content->getDefaultFormat(),
 378                         sha1( $content->serialize( $content->getDefaultFormat() ) ),
 379                         $user->getId() ?: md5( $user->getName() ), // account for user parser options
 380                         $user->getId() ? $user->getDBTouched() : '-' // handle preference change races
 381                 ] ) );
 382
 383                 return wfMemcKey( 'prepared-edit', md5( $title->getPrefixedDBkey() ), $hash );
 384         }
 385
 386         /**
 387          * Build a value to store in memcached based on the PST content and parser output
 388          *
 389          * This makes a simple version of WikiPage::prepareContentForEdit() as stash info
 390          *
 391          * @param Content $pstContent
 392          * @param ParserOutput $parserOutput
 393          * @param string $timestamp TS_MW
 394          * @param User $user
 395          * @return array (stash info array, TTL in seconds) or (null, 0)
 396          */
 397         protected static function buildStashValue(
 398                 Content $pstContent, ParserOutput $parserOutput, $timestamp, User $user
 399         ) {
 400                 // If an item is renewed, mind the cache TTL determined by config and parser functions.
 401                 // Put an upper limit on the TTL for sanity to avoid extreme template/file staleness.
 402                 $since = time() - wfTimestamp( TS_UNIX, $parserOutput->getTimestamp() );
 403                 $ttl = min( $parserOutput->getCacheExpiry() - $since, 5 * 60 );
 404
 405                 if ( $ttl > 0 && !$parserOutput->getFlag( 'vary-revision' ) ) {
 406                         // Only store what is actually needed
 407                         $stashInfo = (object)[
 408                                 'pstContent' => $pstContent,
 409                                 'output'     => $parserOutput,
 410                                 'timestamp'  => $timestamp,
 411                                 'edits'      => $user->getEditCount()
 412                         ];
 413                         return [ $stashInfo, $ttl ];
 414                 }
 415
 416                 return [ null, 0 ];
 417         }
 418
 419         public function getAllowedParams() {
 420                 return [
 421                         'title' => [
 422                                 ApiBase::PARAM_TYPE => 'string',
 423                                 ApiBase::PARAM_REQUIRED => true
 424                         ],
 425                         'section' => [
 426                                 ApiBase::PARAM_TYPE => 'string',
 427                         ],
 428                         'sectiontitle' => [
 429                                 ApiBase::PARAM_TYPE => 'string'
 430                         ],
 431                         'text' => [
 432                                 ApiBase::PARAM_TYPE => 'text',
 433                                 ApiBase::PARAM_REQUIRED => true
 434                         ],
 435                         'contentmodel' => [
 436                                 ApiBase::PARAM_TYPE => ContentHandler::getContentModels(),
 437                                 ApiBase::PARAM_REQUIRED => true
 438                         ],
 439                         'contentformat' => [
 440                                 ApiBase::PARAM_TYPE => ContentHandler::getAllContentFormats(),
 441                                 ApiBase::PARAM_REQUIRED => true
 442                         ],
 443                         'baserevid' => [
 444                                 ApiBase::PARAM_TYPE => 'integer',
 445                                 ApiBase::PARAM_REQUIRED => true
 446                         ]
 447                 ];
 448         }
 449
 450         public function needsToken() {
 451                 return 'csrf';
 452         }
 453
 454         public function mustBePosted() {
 455                 return true;
 456         }
 457
 458         public function isWriteMode() {
 459                 return true;
 460         }
 461
 462         public function isInternal() {
 463                 return true;
 464         }
 465 }