includes/cache/localisation/LocalisationCache.php

   1 <?php
   2 /**
   3  * Cache of the contents of localisation files.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 use CLDRPluralRuleParser\Evaluator;
  24 use CLDRPluralRuleParser\Error as CLDRPluralRuleError;
  25 use MediaWiki\Config\ServiceOptions;
  26 use Psr\Log\LoggerInterface;
  27
  28 /**
  29  * Class for caching the contents of localisation files, Messages*.php
  30  * and *.i18n.php.
  31  *
  32  * An instance of this class is available using MediaWikiServices.
  33  *
  34  * The values retrieved from here are merged, containing items from extension
  35  * files, core messages files and the language fallback sequence (e.g. zh-cn ->
  36  * zh-hans -> en ). Some common errors are corrected, for example namespace
  37  * names with spaces instead of underscores, but heavyweight processing, such
  38  * as grammatical transformation, is done by the caller.
  39  */
  40 class LocalisationCache {
  41         const VERSION = 4;
  42
  43         /** @var ServiceOptions */
  44         private $options;
  45
  46         /**
  47          * True if recaching should only be done on an explicit call to recache().
  48          * Setting this reduces the overhead of cache freshness checking, which
  49          * requires doing a stat() for every extension i18n file.
  50          */
  51         private $manualRecache = false;
  52
  53         /**
  54          * The cache data. 3-d array, where the first key is the language code,
  55          * the second key is the item key e.g. 'messages', and the third key is
  56          * an item specific subkey index. Some items are not arrays and so for those
  57          * items, there are no subkeys.
  58          */
  59         protected $data = [];
  60
  61         /**
  62          * The persistent store object. An instance of LCStore.
  63          *
  64          * @var LCStore
  65          */
  66         private $store;
  67
  68         /**
  69          * @var LoggerInterface
  70          */
  71         private $logger;
  72
  73         /** @var callable[] See comment for parameter in constructor */
  74         private $clearStoreCallbacks;
  75
  76         /**
  77          * A 2-d associative array, code/key, where presence indicates that the item
  78          * is loaded. Value arbitrary.
  79          *
  80          * For split items, if set, this indicates that all of the subitems have been
  81          * loaded.
  82          *
  83          */
  84         private $loadedItems = [];
  85
  86         /**
  87          * A 3-d associative array, code/key/subkey, where presence indicates that
  88          * the subitem is loaded. Only used for the split items, i.e. messages.
  89          */
  90         private $loadedSubitems = [];
  91
  92         /**
  93          * An array where presence of a key indicates that that language has been
  94          * initialised. Initialisation includes checking for cache expiry and doing
  95          * any necessary updates.
  96          */
  97         private $initialisedLangs = [];
  98
  99         /**
 100          * An array mapping non-existent pseudo-languages to fallback languages. This
 101          * is filled by initShallowFallback() when data is requested from a language
 102          * that lacks a Messages*.php file.
 103          */
 104         private $shallowFallbacks = [];
 105
 106         /**
 107          * An array where the keys are codes that have been recached by this instance.
 108          */
 109         private $recachedLangs = [];
 110
 111         /**
 112          * All item keys
 113          */
 114         public static $allKeys = [
 115                 'fallback', 'namespaceNames', 'bookstoreList',
 116                 'magicWords', 'messages', 'rtl', 'capitalizeAllNouns',
 117                 'digitTransformTable', 'separatorTransformTable',
 118                 'minimumGroupingDigits', 'fallback8bitEncoding',
 119                 'linkPrefixExtension', 'linkTrail', 'linkPrefixCharset',
 120                 'namespaceAliases', 'dateFormats', 'datePreferences',
 121                 'datePreferenceMigrationMap', 'defaultDateFormat',
 122                 'specialPageAliases', 'imageFiles', 'preloadedMessages',
 123                 'namespaceGenderAliases', 'digitGroupingPattern', 'pluralRules',
 124                 'pluralRuleTypes', 'compiledPluralRules',
 125         ];
 126
 127         /**
 128          * Keys for items which consist of associative arrays, which may be merged
 129          * by a fallback sequence.
 130          */
 131         public static $mergeableMapKeys = [ 'messages', 'namespaceNames',
 132                 'namespaceAliases', 'dateFormats', 'imageFiles', 'preloadedMessages'
 133         ];
 134
 135         /**
 136          * Keys for items which are a numbered array.
 137          */
 138         public static $mergeableListKeys = [];
 139
 140         /**
 141          * Keys for items which contain an array of arrays of equivalent aliases
 142          * for each subitem. The aliases may be merged by a fallback sequence.
 143          */
 144         public static $mergeableAliasListKeys = [ 'specialPageAliases' ];
 145
 146         /**
 147          * Keys for items which contain an associative array, and may be merged if
 148          * the primary value contains the special array key "inherit". That array
 149          * key is removed after the first merge.
 150          */
 151         public static $optionalMergeKeys = [ 'bookstoreList' ];
 152
 153         /**
 154          * Keys for items that are formatted like $magicWords
 155          */
 156         public static $magicWordKeys = [ 'magicWords' ];
 157
 158         /**
 159          * Keys for items where the subitems are stored in the backend separately.
 160          */
 161         public static $splitKeys = [ 'messages' ];
 162
 163         /**
 164          * Keys which are loaded automatically by initLanguage()
 165          */
 166         public static $preloadedKeys = [ 'dateFormats', 'namespaceNames' ];
 167
 168         /**
 169          * Associative array of cached plural rules. The key is the language code,
 170          * the value is an array of plural rules for that language.
 171          */
 172         private $pluralRules = null;
 173
 174         /**
 175          * Associative array of cached plural rule types. The key is the language
 176          * code, the value is an array of plural rule types for that language. For
 177          * example, $pluralRuleTypes['ar'] = ['zero', 'one', 'two', 'few', 'many'].
 178          * The index for each rule type matches the index for the rule in
 179          * $pluralRules, thus allowing correlation between the two. The reason we
 180          * don't just use the type names as the keys in $pluralRules is because
 181          * Language::convertPlural applies the rules based on numeric order (or
 182          * explicit numeric parameter), not based on the name of the rule type. For
 183          * example, {{plural:count|wordform1|wordform2|wordform3}}, rather than
 184          * {{plural:count|one=wordform1|two=wordform2|many=wordform3}}.
 185          */
 186         private $pluralRuleTypes = null;
 187
 188         private $mergeableKeys = null;
 189
 190         /**
 191          * Return a suitable LCStore as specified by the given configuration.
 192          *
 193          * @since 1.34
 194          * @param array $conf In the format of $wgLocalisationCacheConf
 195          * @param string|false|null $fallbackCacheDir In case 'storeDirectory' isn't specified
 196          * @return LCStore
 197          */
 198         public static function getStoreFromConf( array $conf, $fallbackCacheDir ) : LCStore {
 199                 $storeArg = [];
 200                 $storeArg['directory'] =
 201                         $conf['storeDirectory'] ?: $fallbackCacheDir;
 202
 203                 if ( !empty( $conf['storeClass'] ) ) {
 204                         $storeClass = $conf['storeClass'];
 205                 } elseif ( $conf['store'] === 'files' || $conf['store'] === 'file' ||
 206                         ( $conf['store'] === 'detect' && $storeArg['directory'] )
 207                 ) {
 208                         $storeClass = LCStoreCDB::class;
 209                 } elseif ( $conf['store'] === 'db' || $conf['store'] === 'detect' ) {
 210                         $storeClass = LCStoreDB::class;
 211                         $storeArg['server'] = $conf['storeServer'] ?? [];
 212                 } elseif ( $conf['store'] === 'array' ) {
 213                         $storeClass = LCStoreStaticArray::class;
 214                 } else {
 215                         throw new MWException(
 216                                 'Please set $wgLocalisationCacheConf[\'store\'] to something sensible.'
 217                         );
 218                 }
 219
 220                 return new $storeClass( $storeArg );
 221         }
 222
 223         /**
 224          * @todo Make this a const when HHVM support is dropped (T192166)
 225          *
 226          * @var array
 227          * @since 1.34
 228          */
 229         public static $constructorOptions = [
 230                 // True to treat all files as expired until they are regenerated by this object.
 231                 'forceRecache',
 232                 'manualRecache',
 233                 'ExtensionMessagesFiles',
 234                 'MessagesDirs',
 235         ];
 236
 237         /**
 238          * For constructor parameters, see the documentation in DefaultSettings.php
 239          * for $wgLocalisationCacheConf.
 240          *
 241          * Do not construct this directly. Use MediaWikiServices.
 242          *
 243          * @param ServiceOptions $options
 244          * @param LCStore $store What backend to use for storage
 245          * @param LoggerInterface $logger
 246          * @param callable[] $clearStoreCallbacks To be called whenever the cache is cleared. Can be
 247          *   used to clear other caches that depend on this one, such as ResourceLoader's
 248          *   MessageBlobStore.
 249          * @throws MWException
 250          */
 251         function __construct(
 252                 ServiceOptions $options,
 253                 LCStore $store,
 254                 LoggerInterface $logger,
 255                 array $clearStoreCallbacks = []
 256         ) {
 257                 $options->assertRequiredOptions( self::$constructorOptions );
 258
 259                 $this->options = $options;
 260                 $this->store = $store;
 261                 $this->logger = $logger;
 262                 $this->clearStoreCallbacks = $clearStoreCallbacks;
 263
 264                 // Keep this separate from $this->options so it can be mutable
 265                 $this->manualRecache = $options->get( 'manualRecache' );
 266         }
 267
 268         /**
 269          * Returns true if the given key is mergeable, that is, if it is an associative
 270          * array which can be merged through a fallback sequence.
 271          * @param string $key
 272          * @return bool
 273          */
 274         public function isMergeableKey( $key ) {
 275                 if ( $this->mergeableKeys === null ) {
 276                         $this->mergeableKeys = array_flip( array_merge(
 277                                 self::$mergeableMapKeys,
 278                                 self::$mergeableListKeys,
 279                                 self::$mergeableAliasListKeys,
 280                                 self::$optionalMergeKeys,
 281                                 self::$magicWordKeys
 282                         ) );
 283                 }
 284
 285                 return isset( $this->mergeableKeys[$key] );
 286         }
 287
 288         /**
 289          * Get a cache item.
 290          *
 291          * Warning: this may be slow for split items (messages), since it will
 292          * need to fetch all of the subitems from the cache individually.
 293          * @param string $code
 294          * @param string $key
 295          * @return mixed
 296          */
 297         public function getItem( $code, $key ) {
 298                 if ( !isset( $this->loadedItems[$code][$key] ) ) {
 299                         $this->loadItem( $code, $key );
 300                 }
 301
 302                 if ( $key === 'fallback' && isset( $this->shallowFallbacks[$code] ) ) {
 303                         return $this->shallowFallbacks[$code];
 304                 }
 305
 306                 return $this->data[$code][$key];
 307         }
 308
 309         /**
 310          * Get a subitem, for instance a single message for a given language.
 311          * @param string $code
 312          * @param string $key
 313          * @param string $subkey
 314          * @return mixed|null
 315          */
 316         public function getSubitem( $code, $key, $subkey ) {
 317                 if ( !isset( $this->loadedSubitems[$code][$key][$subkey] ) &&
 318                         !isset( $this->loadedItems[$code][$key] )
 319                 ) {
 320                         $this->loadSubitem( $code, $key, $subkey );
 321                 }
 322
 323                 return $this->data[$code][$key][$subkey] ?? null;
 324         }
 325
 326         /**
 327          * Get the list of subitem keys for a given item.
 328          *
 329          * This is faster than array_keys($lc->getItem(...)) for the items listed in
 330          * self::$splitKeys.
 331          *
 332          * Will return null if the item is not found, or false if the item is not an
 333          * array.
 334          * @param string $code
 335          * @param string $key
 336          * @return bool|null|string|string[]
 337          */
 338         public function getSubitemList( $code, $key ) {
 339                 if ( in_array( $key, self::$splitKeys ) ) {
 340                         return $this->getSubitem( $code, 'list', $key );
 341                 } else {
 342                         $item = $this->getItem( $code, $key );
 343                         if ( is_array( $item ) ) {
 344                                 return array_keys( $item );
 345                         } else {
 346                                 return false;
 347                         }
 348                 }
 349         }
 350
 351         /**
 352          * Load an item into the cache.
 353          * @param string $code
 354          * @param string $key
 355          */
 356         protected function loadItem( $code, $key ) {
 357                 if ( !isset( $this->initialisedLangs[$code] ) ) {
 358                         $this->initLanguage( $code );
 359                 }
 360
 361                 // Check to see if initLanguage() loaded it for us
 362                 if ( isset( $this->loadedItems[$code][$key] ) ) {
 363                         return;
 364                 }
 365
 366                 if ( isset( $this->shallowFallbacks[$code] ) ) {
 367                         $this->loadItem( $this->shallowFallbacks[$code], $key );
 368
 369                         return;
 370                 }
 371
 372                 if ( in_array( $key, self::$splitKeys ) ) {
 373                         $subkeyList = $this->getSubitem( $code, 'list', $key );
 374                         foreach ( $subkeyList as $subkey ) {
 375                                 if ( isset( $this->data[$code][$key][$subkey] ) ) {
 376                                         continue;
 377                                 }
 378                                 $this->data[$code][$key][$subkey] = $this->getSubitem( $code, $key, $subkey );
 379                         }
 380                 } else {
 381                         $this->data[$code][$key] = $this->store->get( $code, $key );
 382                 }
 383
 384                 $this->loadedItems[$code][$key] = true;
 385         }
 386
 387         /**
 388          * Load a subitem into the cache
 389          * @param string $code
 390          * @param string $key
 391          * @param string $subkey
 392          */
 393         protected function loadSubitem( $code, $key, $subkey ) {
 394                 if ( !in_array( $key, self::$splitKeys ) ) {
 395                         $this->loadItem( $code, $key );
 396
 397                         return;
 398                 }
 399
 400                 if ( !isset( $this->initialisedLangs[$code] ) ) {
 401                         $this->initLanguage( $code );
 402                 }
 403
 404                 // Check to see if initLanguage() loaded it for us
 405                 if ( isset( $this->loadedItems[$code][$key] ) ||
 406                         isset( $this->loadedSubitems[$code][$key][$subkey] )
 407                 ) {
 408                         return;
 409                 }
 410
 411                 if ( isset( $this->shallowFallbacks[$code] ) ) {
 412                         $this->loadSubitem( $this->shallowFallbacks[$code], $key, $subkey );
 413
 414                         return;
 415                 }
 416
 417                 $value = $this->store->get( $code, "$key:$subkey" );
 418                 $this->data[$code][$key][$subkey] = $value;
 419                 $this->loadedSubitems[$code][$key][$subkey] = true;
 420         }
 421
 422         /**
 423          * Returns true if the cache identified by $code is missing or expired.
 424          *
 425          * @param string $code
 426          *
 427          * @return bool
 428          */
 429         public function isExpired( $code ) {
 430                 if ( $this->options->get( 'forceRecache' ) && !isset( $this->recachedLangs[$code] ) ) {
 431                         $this->logger->debug( __METHOD__ . "($code): forced reload" );
 432
 433                         return true;
 434                 }
 435
 436                 $deps = $this->store->get( $code, 'deps' );
 437                 $keys = $this->store->get( $code, 'list' );
 438                 $preload = $this->store->get( $code, 'preload' );
 439                 // Different keys may expire separately for some stores
 440                 if ( $deps === null || $keys === null || $preload === null ) {
 441                         $this->logger->debug( __METHOD__ . "($code): cache missing, need to make one" );
 442
 443                         return true;
 444                 }
 445
 446                 foreach ( $deps as $dep ) {
 447                         // Because we're unserializing stuff from cache, we
 448                         // could receive objects of classes that don't exist
 449                         // anymore (e.g. uninstalled extensions)
 450                         // When this happens, always expire the cache
 451                         if ( !$dep instanceof CacheDependency || $dep->isExpired() ) {
 452                                 $this->logger->debug( __METHOD__ . "($code): cache for $code expired due to " .
 453                                         get_class( $dep ) );
 454
 455                                 return true;
 456                         }
 457                 }
 458
 459                 return false;
 460         }
 461
 462         /**
 463          * Initialise a language in this object. Rebuild the cache if necessary.
 464          * @param string $code
 465          * @throws MWException
 466          */
 467         protected function initLanguage( $code ) {
 468                 if ( isset( $this->initialisedLangs[$code] ) ) {
 469                         return;
 470                 }
 471
 472                 $this->initialisedLangs[$code] = true;
 473
 474                 # If the code is of the wrong form for a Messages*.php file, do a shallow fallback
 475                 if ( !Language::isValidBuiltInCode( $code ) ) {
 476                         $this->initShallowFallback( $code, 'en' );
 477
 478                         return;
 479                 }
 480
 481                 # Recache the data if necessary
 482                 if ( !$this->manualRecache && $this->isExpired( $code ) ) {
 483                         if ( Language::isSupportedLanguage( $code ) ) {
 484                                 $this->recache( $code );
 485                         } elseif ( $code === 'en' ) {
 486                                 throw new MWException( 'MessagesEn.php is missing.' );
 487                         } else {
 488                                 $this->initShallowFallback( $code, 'en' );
 489                         }
 490
 491                         return;
 492                 }
 493
 494                 # Preload some stuff
 495                 $preload = $this->getItem( $code, 'preload' );
 496                 if ( $preload === null ) {
 497                         if ( $this->manualRecache ) {
 498                                 // No Messages*.php file. Do shallow fallback to en.
 499                                 if ( $code === 'en' ) {
 500                                         throw new MWException( 'No localisation cache found for English. ' .
 501                                                 'Please run maintenance/rebuildLocalisationCache.php.' );
 502                                 }
 503                                 $this->initShallowFallback( $code, 'en' );
 504
 505                                 return;
 506                         } else {
 507                                 throw new MWException( 'Invalid or missing localisation cache.' );
 508                         }
 509                 }
 510                 $this->data[$code] = $preload;
 511                 foreach ( $preload as $key => $item ) {
 512                         if ( in_array( $key, self::$splitKeys ) ) {
 513                                 foreach ( $item as $subkey => $subitem ) {
 514                                         $this->loadedSubitems[$code][$key][$subkey] = true;
 515                                 }
 516                         } else {
 517                                 $this->loadedItems[$code][$key] = true;
 518                         }
 519                 }
 520         }
 521
 522         /**
 523          * Create a fallback from one language to another, without creating a
 524          * complete persistent cache.
 525          * @param string $primaryCode
 526          * @param string $fallbackCode
 527          */
 528         public function initShallowFallback( $primaryCode, $fallbackCode ) {
 529                 $this->data[$primaryCode] =& $this->data[$fallbackCode];
 530                 $this->loadedItems[$primaryCode] =& $this->loadedItems[$fallbackCode];
 531                 $this->loadedSubitems[$primaryCode] =& $this->loadedSubitems[$fallbackCode];
 532                 $this->shallowFallbacks[$primaryCode] = $fallbackCode;
 533         }
 534
 535         /**
 536          * Read a PHP file containing localisation data.
 537          * @param string $_fileName
 538          * @param string $_fileType
 539          * @throws MWException
 540          * @return array
 541          */
 542         protected function readPHPFile( $_fileName, $_fileType ) {
 543                 include $_fileName;
 544
 545                 $data = [];
 546                 if ( $_fileType == 'core' || $_fileType == 'extension' ) {
 547                         foreach ( self::$allKeys as $key ) {
 548                                 // Not all keys are set in language files, so
 549                                 // check they exist first
 550                                 if ( isset( $$key ) ) {
 551                                         $data[$key] = $$key;
 552                                 }
 553                         }
 554                 } elseif ( $_fileType == 'aliases' ) {
 555                         if ( isset( $aliases ) ) {
 556                                 $data['aliases'] = $aliases;
 557                         }
 558                 } else {
 559                         throw new MWException( __METHOD__ . ": Invalid file type: $_fileType" );
 560                 }
 561
 562                 return $data;
 563         }
 564
 565         /**
 566          * Read a JSON file containing localisation messages.
 567          * @param string $fileName Name of file to read
 568          * @throws MWException If there is a syntax error in the JSON file
 569          * @return array Array with a 'messages' key, or empty array if the file doesn't exist
 570          */
 571         public function readJSONFile( $fileName ) {
 572                 if ( !is_readable( $fileName ) ) {
 573                         return [];
 574                 }
 575
 576                 $json = file_get_contents( $fileName );
 577                 if ( $json === false ) {
 578                         return [];
 579                 }
 580
 581                 $data = FormatJson::decode( $json, true );
 582                 if ( $data === null ) {
 583                         throw new MWException( __METHOD__ . ": Invalid JSON file: $fileName" );
 584                 }
 585
 586                 // Remove keys starting with '@', they're reserved for metadata and non-message data
 587                 foreach ( $data as $key => $unused ) {
 588                         if ( $key === '' || $key[0] === '@' ) {
 589                                 unset( $data[$key] );
 590                         }
 591                 }
 592
 593                 // The JSON format only supports messages, none of the other variables, so wrap the data
 594                 return [ 'messages' => $data ];
 595         }
 596
 597         /**
 598          * Get the compiled plural rules for a given language from the XML files.
 599          * @since 1.20
 600          * @param string $code
 601          * @return array|null
 602          */
 603         public function getCompiledPluralRules( $code ) {
 604                 $rules = $this->getPluralRules( $code );
 605                 if ( $rules === null ) {
 606                         return null;
 607                 }
 608                 try {
 609                         $compiledRules = Evaluator::compile( $rules );
 610                 } catch ( CLDRPluralRuleError $e ) {
 611                         $this->logger->debug( $e->getMessage() );
 612
 613                         return [];
 614                 }
 615
 616                 return $compiledRules;
 617         }
 618
 619         /**
 620          * Get the plural rules for a given language from the XML files.
 621          * Cached.
 622          * @since 1.20
 623          * @param string $code
 624          * @return array|null
 625          */
 626         public function getPluralRules( $code ) {
 627                 if ( $this->pluralRules === null ) {
 628                         $this->loadPluralFiles();
 629                 }
 630                 return $this->pluralRules[$code] ?? null;
 631         }
 632
 633         /**
 634          * Get the plural rule types for a given language from the XML files.
 635          * Cached.
 636          * @since 1.22
 637          * @param string $code
 638          * @return array|null
 639          */
 640         public function getPluralRuleTypes( $code ) {
 641                 if ( $this->pluralRuleTypes === null ) {
 642                         $this->loadPluralFiles();
 643                 }
 644                 return $this->pluralRuleTypes[$code] ?? null;
 645         }
 646
 647         /**
 648          * Load the plural XML files.
 649          */
 650         protected function loadPluralFiles() {
 651                 global $IP;
 652                 $cldrPlural = "$IP/languages/data/plurals.xml";
 653                 $mwPlural = "$IP/languages/data/plurals-mediawiki.xml";
 654                 // Load CLDR plural rules
 655                 $this->loadPluralFile( $cldrPlural );
 656                 if ( file_exists( $mwPlural ) ) {
 657                         // Override or extend
 658                         $this->loadPluralFile( $mwPlural );
 659                 }
 660         }
 661
 662         /**
 663          * Load a plural XML file with the given filename, compile the relevant
 664          * rules, and save the compiled rules in a process-local cache.
 665          *
 666          * @param string $fileName
 667          * @throws MWException
 668          */
 669         protected function loadPluralFile( $fileName ) {
 670                 // Use file_get_contents instead of DOMDocument::load (T58439)
 671                 $xml = file_get_contents( $fileName );
 672                 if ( !$xml ) {
 673                         throw new MWException( "Unable to read plurals file $fileName" );
 674                 }
 675                 $doc = new DOMDocument;
 676                 $doc->loadXML( $xml );
 677                 $rulesets = $doc->getElementsByTagName( "pluralRules" );
 678                 foreach ( $rulesets as $ruleset ) {
 679                         $codes = $ruleset->getAttribute( 'locales' );
 680                         $rules = [];
 681                         $ruleTypes = [];
 682                         $ruleElements = $ruleset->getElementsByTagName( "pluralRule" );
 683                         foreach ( $ruleElements as $elt ) {
 684                                 $ruleType = $elt->getAttribute( 'count' );
 685                                 if ( $ruleType === 'other' ) {
 686                                         // Don't record "other" rules, which have an empty condition
 687                                         continue;
 688                                 }
 689                                 $rules[] = $elt->nodeValue;
 690                                 $ruleTypes[] = $ruleType;
 691                         }
 692                         foreach ( explode( ' ', $codes ) as $code ) {
 693                                 $this->pluralRules[$code] = $rules;
 694                                 $this->pluralRuleTypes[$code] = $ruleTypes;
 695                         }
 696                 }
 697         }
 698
 699         /**
 700          * Read the data from the source files for a given language, and register
 701          * the relevant dependencies in the $deps array. If the localisation
 702          * exists, the data array is returned, otherwise false is returned.
 703          *
 704          * @param string $code
 705          * @param array &$deps
 706          * @return array
 707          */
 708         protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
 709                 global $IP;
 710
 711                 // This reads in the PHP i18n file with non-messages l10n data
 712                 $fileName = Language::getMessagesFileName( $code );
 713                 if ( !file_exists( $fileName ) ) {
 714                         $data = [];
 715                 } else {
 716                         $deps[] = new FileDependency( $fileName );
 717                         $data = $this->readPHPFile( $fileName, 'core' );
 718                 }
 719
 720                 # Load CLDR plural rules for JavaScript
 721                 $data['pluralRules'] = $this->getPluralRules( $code );
 722                 # And for PHP
 723                 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
 724                 # Load plural rule types
 725                 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
 726
 727                 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
 728                 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
 729
 730                 return $data;
 731         }
 732
 733         /**
 734          * Merge two localisation values, a primary and a fallback, overwriting the
 735          * primary value in place.
 736          * @param string $key
 737          * @param mixed &$value
 738          * @param mixed $fallbackValue
 739          */
 740         protected function mergeItem( $key, &$value, $fallbackValue ) {
 741                 if ( !is_null( $value ) ) {
 742                         if ( !is_null( $fallbackValue ) ) {
 743                                 if ( in_array( $key, self::$mergeableMapKeys ) ) {
 744                                         $value = $value + $fallbackValue;
 745                                 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
 746                                         // @phan-suppress-next-line PhanTypeMismatchArgumentInternal
 747                                         $value = array_unique( array_merge( $fallbackValue, $value ) );
 748                                 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
 749                                         $value = array_merge_recursive( $value, $fallbackValue );
 750                                 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
 751                                         if ( !empty( $value['inherit'] ) ) {
 752                                                 $value = array_merge( $fallbackValue, $value );
 753                                         }
 754
 755                                         if ( isset( $value['inherit'] ) ) {
 756                                                 unset( $value['inherit'] );
 757                                         }
 758                                 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
 759                                         $this->mergeMagicWords( $value, $fallbackValue );
 760                                 }
 761                         }
 762                 } else {
 763                         $value = $fallbackValue;
 764                 }
 765         }
 766
 767         /**
 768          * @param mixed &$value
 769          * @param mixed $fallbackValue
 770          */
 771         protected function mergeMagicWords( &$value, $fallbackValue ) {
 772                 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
 773                         if ( !isset( $value[$magicName] ) ) {
 774                                 $value[$magicName] = $fallbackInfo;
 775                         } else {
 776                                 $oldSynonyms = array_slice( $fallbackInfo, 1 );
 777                                 $newSynonyms = array_slice( $value[$magicName], 1 );
 778                                 $synonyms = array_values( array_unique( array_merge(
 779                                         $newSynonyms, $oldSynonyms ) ) );
 780                                 $value[$magicName] = array_merge( [ $fallbackInfo[0] ], $synonyms );
 781                         }
 782                 }
 783         }
 784
 785         /**
 786          * Given an array mapping language code to localisation value, such as is
 787          * found in extension *.i18n.php files, iterate through a fallback sequence
 788          * to merge the given data with an existing primary value.
 789          *
 790          * Returns true if any data from the extension array was used, false
 791          * otherwise.
 792          * @param array $codeSequence
 793          * @param string $key
 794          * @param mixed &$value
 795          * @param mixed $fallbackValue
 796          * @return bool
 797          */
 798         protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
 799                 $used = false;
 800                 foreach ( $codeSequence as $code ) {
 801                         if ( isset( $fallbackValue[$code] ) ) {
 802                                 $this->mergeItem( $key, $value, $fallbackValue[$code] );
 803                                 $used = true;
 804                         }
 805                 }
 806
 807                 return $used;
 808         }
 809
 810         /**
 811          * Gets the combined list of messages dirs from
 812          * core and extensions
 813          *
 814          * @since 1.25
 815          * @return array
 816          */
 817         public function getMessagesDirs() {
 818                 global $IP;
 819
 820                 return [
 821                         'core' => "$IP/languages/i18n",
 822                         'exif' => "$IP/languages/i18n/exif",
 823                         'api' => "$IP/includes/api/i18n",
 824                         'oojs-ui' => "$IP/resources/lib/ooui/i18n",
 825                 ] + $this->options->get( 'MessagesDirs' );
 826         }
 827
 828         /**
 829          * Load localisation data for a given language for both core and extensions
 830          * and save it to the persistent cache store and the process cache
 831          * @param string $code
 832          * @throws MWException
 833          */
 834         public function recache( $code ) {
 835                 if ( !$code ) {
 836                         throw new MWException( "Invalid language code requested" );
 837                 }
 838                 $this->recachedLangs[ $code ] = true;
 839
 840                 # Initial values
 841                 $initialData = array_fill_keys( self::$allKeys, null );
 842                 $coreData = $initialData;
 843                 $deps = [];
 844
 845                 # Load the primary localisation from the source file
 846                 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
 847                 $this->logger->debug( __METHOD__ . ": got localisation for $code from source" );
 848
 849                 # Merge primary localisation
 850                 foreach ( $data as $key => $value ) {
 851                         $this->mergeItem( $key, $coreData[ $key ], $value );
 852                 }
 853
 854                 # Fill in the fallback if it's not there already
 855                 if ( ( is_null( $coreData['fallback'] ) || $coreData['fallback'] === false ) && $code === 'en' ) {
 856                         $coreData['fallback'] = false;
 857                         $coreData['originalFallbackSequence'] = $coreData['fallbackSequence'] = [];
 858                 } else {
 859                         if ( !is_null( $coreData['fallback'] ) ) {
 860                                 $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
 861                         } else {
 862                                 $coreData['fallbackSequence'] = [];
 863                         }
 864                         $len = count( $coreData['fallbackSequence'] );
 865
 866                         # Before we add the 'en' fallback for messages, keep a copy of
 867                         # the original fallback sequence
 868                         $coreData['originalFallbackSequence'] = $coreData['fallbackSequence'];
 869
 870                         # Ensure that the sequence ends at 'en' for messages
 871                         if ( !$len || $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
 872                                 $coreData['fallbackSequence'][] = 'en';
 873                         }
 874                 }
 875
 876                 $codeSequence = array_merge( [ $code ], $coreData['fallbackSequence'] );
 877                 $messageDirs = $this->getMessagesDirs();
 878
 879                 # Load non-JSON localisation data for extensions
 880                 $extensionData = array_fill_keys( $codeSequence, $initialData );
 881                 foreach ( $this->options->get( 'ExtensionMessagesFiles' ) as $extension => $fileName ) {
 882                         if ( isset( $messageDirs[$extension] ) ) {
 883                                 # This extension has JSON message data; skip the PHP shim
 884                                 continue;
 885                         }
 886
 887                         $data = $this->readPHPFile( $fileName, 'extension' );
 888                         $used = false;
 889
 890                         foreach ( $data as $key => $item ) {
 891                                 foreach ( $codeSequence as $csCode ) {
 892                                         if ( isset( $item[$csCode] ) ) {
 893                                                 $this->mergeItem( $key, $extensionData[$csCode][$key], $item[$csCode] );
 894                                                 $used = true;
 895                                         }
 896                                 }
 897                         }
 898
 899                         if ( $used ) {
 900                                 $deps[] = new FileDependency( $fileName );
 901                         }
 902                 }
 903
 904                 # Load the localisation data for each fallback, then merge it into the full array
 905                 $allData = $initialData;
 906                 foreach ( $codeSequence as $csCode ) {
 907                         $csData = $initialData;
 908
 909                         # Load core messages and the extension localisations.
 910                         foreach ( $messageDirs as $dirs ) {
 911                                 foreach ( (array)$dirs as $dir ) {
 912                                         $fileName = "$dir/$csCode.json";
 913                                         $data = $this->readJSONFile( $fileName );
 914
 915                                         foreach ( $data as $key => $item ) {
 916                                                 $this->mergeItem( $key, $csData[$key], $item );
 917                                         }
 918
 919                                         $deps[] = new FileDependency( $fileName );
 920                                 }
 921                         }
 922
 923                         # Merge non-JSON extension data
 924                         if ( isset( $extensionData[$csCode] ) ) {
 925                                 foreach ( $extensionData[$csCode] as $key => $item ) {
 926                                         $this->mergeItem( $key, $csData[$key], $item );
 927                                 }
 928                         }
 929
 930                         if ( $csCode === $code ) {
 931                                 # Merge core data into extension data
 932                                 foreach ( $coreData as $key => $item ) {
 933                                         $this->mergeItem( $key, $csData[$key], $item );
 934                                 }
 935                         } else {
 936                                 # Load the secondary localisation from the source file to
 937                                 # avoid infinite cycles on cyclic fallbacks
 938                                 $fbData = $this->readSourceFilesAndRegisterDeps( $csCode, $deps );
 939                                 # Only merge the keys that make sense to merge
 940                                 foreach ( self::$allKeys as $key ) {
 941                                         if ( !isset( $fbData[ $key ] ) ) {
 942                                                 continue;
 943                                         }
 944
 945                                         if ( is_null( $coreData[ $key ] ) || $this->isMergeableKey( $key ) ) {
 946                                                 $this->mergeItem( $key, $csData[ $key ], $fbData[ $key ] );
 947                                         }
 948                                 }
 949                         }
 950
 951                         # Allow extensions an opportunity to adjust the data for this
 952                         # fallback
 953                         Hooks::run( 'LocalisationCacheRecacheFallback', [ $this, $csCode, &$csData ] );
 954
 955                         # Merge the data for this fallback into the final array
 956                         if ( $csCode === $code ) {
 957                                 $allData = $csData;
 958                         } else {
 959                                 foreach ( self::$allKeys as $key ) {
 960                                         if ( !isset( $csData[$key] ) ) {
 961                                                 continue;
 962                                         }
 963
 964                                         if ( is_null( $allData[$key] ) || $this->isMergeableKey( $key ) ) {
 965                                                 $this->mergeItem( $key, $allData[$key], $csData[$key] );
 966                                         }
 967                                 }
 968                         }
 969                 }
 970
 971                 # Add cache dependencies for any referenced globals
 972                 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
 973                 // The 'MessagesDirs' config setting is used in LocalisationCache::getMessagesDirs().
 974                 // We use the key 'wgMessagesDirs' for historical reasons.
 975                 $deps['wgMessagesDirs'] = new MainConfigDependency( 'MessagesDirs' );
 976                 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
 977
 978                 # Add dependencies to the cache entry
 979                 $allData['deps'] = $deps;
 980
 981                 # Replace spaces with underscores in namespace names
 982                 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
 983
 984                 # And do the same for special page aliases. $page is an array.
 985                 foreach ( $allData['specialPageAliases'] as &$page ) {
 986                         $page = str_replace( ' ', '_', $page );
 987                 }
 988                 # Decouple the reference to prevent accidental damage
 989                 unset( $page );
 990
 991                 # If there were no plural rules, return an empty array
 992                 if ( $allData['pluralRules'] === null ) {
 993                         $allData['pluralRules'] = [];
 994                 }
 995                 if ( $allData['compiledPluralRules'] === null ) {
 996                         $allData['compiledPluralRules'] = [];
 997                 }
 998                 # If there were no plural rule types, return an empty array
 999                 if ( $allData['pluralRuleTypes'] === null ) {
1000                         $allData['pluralRuleTypes'] = [];
1001                 }
1002
1003                 # Set the list keys
1004                 $allData['list'] = [];
1005                 foreach ( self::$splitKeys as $key ) {
1006                         $allData['list'][$key] = array_keys( $allData[$key] );
1007                 }
1008                 # Run hooks
1009                 $unused = true; // Used to be $purgeBlobs, removed in 1.34
1010                 Hooks::run( 'LocalisationCacheRecache', [ $this, $code, &$allData, &$unused ] );
1011
1012                 if ( is_null( $allData['namespaceNames'] ) ) {
1013                         throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
1014                                 'Check that your languages/messages/MessagesEn.php file is intact.' );
1015                 }
1016
1017                 # Set the preload key
1018                 $allData['preload'] = $this->buildPreload( $allData );
1019
1020                 # Save to the process cache and register the items loaded
1021                 $this->data[$code] = $allData;
1022                 foreach ( $allData as $key => $item ) {
1023                         $this->loadedItems[$code][$key] = true;
1024                 }
1025
1026                 # Save to the persistent cache
1027                 $this->store->startWrite( $code );
1028                 foreach ( $allData as $key => $value ) {
1029                         if ( in_array( $key, self::$splitKeys ) ) {
1030                                 foreach ( $value as $subkey => $subvalue ) {
1031                                         $this->store->set( "$key:$subkey", $subvalue );
1032                                 }
1033                         } else {
1034                                 $this->store->set( $key, $value );
1035                         }
1036                 }
1037                 $this->store->finishWrite();
1038
1039                 # Clear out the MessageBlobStore
1040                 # HACK: If using a null (i.e. disabled) storage backend, we
1041                 # can't write to the MessageBlobStore either
1042                 if ( !$this->store instanceof LCStoreNull ) {
1043                         foreach ( $this->clearStoreCallbacks as $callback ) {
1044                                 $callback();
1045                         }
1046                 }
1047         }
1048
1049         /**
1050          * Build the preload item from the given pre-cache data.
1051          *
1052          * The preload item will be loaded automatically, improving performance
1053          * for the commonly-requested items it contains.
1054          * @param array $data
1055          * @return array
1056          */
1057         protected function buildPreload( $data ) {
1058                 $preload = [ 'messages' => [] ];
1059                 foreach ( self::$preloadedKeys as $key ) {
1060                         $preload[$key] = $data[$key];
1061                 }
1062
1063                 foreach ( $data['preloadedMessages'] as $subkey ) {
1064                         $subitem = $data['messages'][$subkey] ?? null;
1065                         $preload['messages'][$subkey] = $subitem;
1066                 }
1067
1068                 return $preload;
1069         }
1070
1071         /**
1072          * Unload the data for a given language from the object cache.
1073          * Reduces memory usage.
1074          * @param string $code
1075          */
1076         public function unload( $code ) {
1077                 unset( $this->data[$code] );
1078                 unset( $this->loadedItems[$code] );
1079                 unset( $this->loadedSubitems[$code] );
1080                 unset( $this->initialisedLangs[$code] );
1081                 unset( $this->shallowFallbacks[$code] );
1082
1083                 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1084                         if ( $fbCode === $code ) {
1085                                 $this->unload( $shallowCode );
1086                         }
1087                 }
1088         }
1089
1090         /**
1091          * Unload all data
1092          */
1093         public function unloadAll() {
1094                 foreach ( $this->initialisedLangs as $lang => $unused ) {
1095                         $this->unload( $lang );
1096                 }
1097         }
1098
1099         /**
1100          * Disable the storage backend
1101          */
1102         public function disableBackend() {
1103                 $this->store = new LCStoreNull;
1104                 $this->manualRecache = false;
1105         }
1106 }