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 MediaWiki\Languages\LanguageNameUtils;
  27 use Psr\Log\LoggerInterface;
  28
  29 /**
  30  * Class for caching the contents of localisation files, Messages*.php
  31  * and *.i18n.php.
  32  *
  33  * An instance of this class is available using MediaWikiServices.
  34  *
  35  * The values retrieved from here are merged, containing items from extension
  36  * files, core messages files and the language fallback sequence (e.g. zh-cn ->
  37  * zh-hans -> en ). Some common errors are corrected, for example namespace
  38  * names with spaces instead of underscores, but heavyweight processing, such
  39  * as grammatical transformation, is done by the caller.
  40  */
  41 class LocalisationCache {
  42         const VERSION = 4;
  43
  44         /** @var ServiceOptions */
  45         private $options;
  46
  47         /**
  48          * True if recaching should only be done on an explicit call to recache().
  49          * Setting this reduces the overhead of cache freshness checking, which
  50          * requires doing a stat() for every extension i18n file.
  51          */
  52         private $manualRecache = false;
  53
  54         /**
  55          * The cache data. 3-d array, where the first key is the language code,
  56          * the second key is the item key e.g. 'messages', and the third key is
  57          * an item specific subkey index. Some items are not arrays and so for those
  58          * items, there are no subkeys.
  59          */
  60         protected $data = [];
  61
  62         /**
  63          * The persistent store object. An instance of LCStore.
  64          *
  65          * @var LCStore
  66          */
  67         private $store;
  68
  69         /**
  70          * @var LoggerInterface
  71          */
  72         private $logger;
  73
  74         /** @var callable[] See comment for parameter in constructor */
  75         private $clearStoreCallbacks;
  76
  77         /** @var LanguageNameUtils */
  78         private $langNameUtils;
  79
  80         /**
  81          * A 2-d associative array, code/key, where presence indicates that the item
  82          * is loaded. Value arbitrary.
  83          *
  84          * For split items, if set, this indicates that all of the subitems have been
  85          * loaded.
  86          */
  87         private $loadedItems = [];
  88
  89         /**
  90          * A 3-d associative array, code/key/subkey, where presence indicates that
  91          * the subitem is loaded. Only used for the split items, i.e. messages.
  92          */
  93         private $loadedSubitems = [];
  94
  95         /**
  96          * An array where presence of a key indicates that that language has been
  97          * initialised. Initialisation includes checking for cache expiry and doing
  98          * any necessary updates.
  99          */
 100         private $initialisedLangs = [];
 101
 102         /**
 103          * An array mapping non-existent pseudo-languages to fallback languages. This
 104          * is filled by initShallowFallback() when data is requested from a language
 105          * that lacks a Messages*.php file.
 106          */
 107         private $shallowFallbacks = [];
 108
 109         /**
 110          * An array where the keys are codes that have been recached by this instance.
 111          */
 112         private $recachedLangs = [];
 113
 114         /**
 115          * All item keys
 116          */
 117         public static $allKeys = [
 118                 'fallback', 'namespaceNames', 'bookstoreList',
 119                 'magicWords', 'messages', 'rtl', 'capitalizeAllNouns',
 120                 'digitTransformTable', 'separatorTransformTable',
 121                 'minimumGroupingDigits', 'fallback8bitEncoding',
 122                 'linkPrefixExtension', 'linkTrail', 'linkPrefixCharset',
 123                 'namespaceAliases', 'dateFormats', 'datePreferences',
 124                 'datePreferenceMigrationMap', 'defaultDateFormat',
 125                 'specialPageAliases', 'imageFiles', 'preloadedMessages',
 126                 'namespaceGenderAliases', 'digitGroupingPattern', 'pluralRules',
 127                 'pluralRuleTypes', 'compiledPluralRules',
 128         ];
 129
 130         /**
 131          * Keys for items which consist of associative arrays, which may be merged
 132          * by a fallback sequence.
 133          */
 134         public static $mergeableMapKeys = [ 'messages', 'namespaceNames',
 135                 'namespaceAliases', 'dateFormats', 'imageFiles', 'preloadedMessages'
 136         ];
 137
 138         /**
 139          * Keys for items which are a numbered array.
 140          */
 141         public static $mergeableListKeys = [];
 142
 143         /**
 144          * Keys for items which contain an array of arrays of equivalent aliases
 145          * for each subitem. The aliases may be merged by a fallback sequence.
 146          */
 147         public static $mergeableAliasListKeys = [ 'specialPageAliases' ];
 148
 149         /**
 150          * Keys for items which contain an associative array, and may be merged if
 151          * the primary value contains the special array key "inherit". That array
 152          * key is removed after the first merge.
 153          */
 154         public static $optionalMergeKeys = [ 'bookstoreList' ];
 155
 156         /**
 157          * Keys for items that are formatted like $magicWords
 158          */
 159         public static $magicWordKeys = [ 'magicWords' ];
 160
 161         /**
 162          * Keys for items where the subitems are stored in the backend separately.
 163          */
 164         public static $splitKeys = [ 'messages' ];
 165
 166         /**
 167          * Keys which are loaded automatically by initLanguage()
 168          */
 169         public static $preloadedKeys = [ 'dateFormats', 'namespaceNames' ];
 170
 171         /**
 172          * Associative array of cached plural rules. The key is the language code,
 173          * the value is an array of plural rules for that language.
 174          */
 175         private $pluralRules = null;
 176
 177         /**
 178          * Associative array of cached plural rule types. The key is the language
 179          * code, the value is an array of plural rule types for that language. For
 180          * example, $pluralRuleTypes['ar'] = ['zero', 'one', 'two', 'few', 'many'].
 181          * The index for each rule type matches the index for the rule in
 182          * $pluralRules, thus allowing correlation between the two. The reason we
 183          * don't just use the type names as the keys in $pluralRules is because
 184          * Language::convertPlural applies the rules based on numeric order (or
 185          * explicit numeric parameter), not based on the name of the rule type. For
 186          * example, {{plural:count|wordform1|wordform2|wordform3}}, rather than
 187          * {{plural:count|one=wordform1|two=wordform2|many=wordform3}}.
 188          */
 189         private $pluralRuleTypes = null;
 190
 191         private $mergeableKeys = null;
 192
 193         /**
 194          * Return a suitable LCStore as specified by the given configuration.
 195          *
 196          * @param array $conf In the format of $wgLocalisationCacheConf
 197          * @param string|false|null $fallbackCacheDir In case 'storeDirectory' isn't specified
 198          * @return LCStore
 199          */
 200         public static function getStoreFromConf( array $conf, $fallbackCacheDir ) : LCStore {
 201                 $storeArg = [];
 202                 $storeArg['directory'] =
 203                         $conf['storeDirectory'] ?: $fallbackCacheDir;
 204
 205                 if ( !empty( $conf['storeClass'] ) ) {
 206                         $storeClass = $conf['storeClass'];
 207                 } elseif ( $conf['store'] === 'files' || $conf['store'] === 'file' ||
 208                         ( $conf['store'] === 'detect' && $storeArg['directory'] )
 209                 ) {
 210                         $storeClass = LCStoreCDB::class;
 211                 } elseif ( $conf['store'] === 'db' || $conf['store'] === 'detect' ) {
 212                         $storeClass = LCStoreDB::class;
 213                         $storeArg['server'] = $conf['storeServer'] ?? [];
 214                 } elseif ( $conf['store'] === 'array' ) {
 215                         $storeClass = LCStoreStaticArray::class;
 216                 } else {
 217                         throw new MWException(
 218                                 'Please set $wgLocalisationCacheConf[\'store\'] to something sensible.'
 219                         );
 220                 }
 221
 222                 return new $storeClass( $storeArg );
 223         }
 224
 225         /**
 226          * @todo Make this a const when HHVM support is dropped (T192166)
 227          *
 228          * @var array
 229          * @since 1.34
 230          */
 231         public static $constructorOptions = [
 232                 // True to treat all files as expired until they are regenerated by this object.
 233                 'forceRecache',
 234                 'manualRecache',
 235                 'ExtensionMessagesFiles',
 236                 'MessagesDirs',
 237         ];
 238
 239         /**
 240          * For constructor parameters, see the documentation in DefaultSettings.php
 241          * for $wgLocalisationCacheConf.
 242          *
 243          * Do not construct this directly. Use MediaWikiServices.
 244          *
 245          * @param ServiceOptions $options
 246          * @param LCStore $store What backend to use for storage
 247          * @param LoggerInterface $logger
 248          * @param callable[] $clearStoreCallbacks To be called whenever the cache is cleared. Can be
 249          *   used to clear other caches that depend on this one, such as ResourceLoader's
 250          *   MessageBlobStore.
 251          * @param LanguageNameUtils $langNameUtils
 252          * @throws MWException
 253          */
 254         function __construct(
 255                 ServiceOptions $options,
 256                 LCStore $store,
 257                 LoggerInterface $logger,
 258                 array $clearStoreCallbacks,
 259                 LanguageNameUtils $langNameUtils
 260         ) {
 261                 $options->assertRequiredOptions( self::$constructorOptions );
 262
 263                 $this->options = $options;
 264                 $this->store = $store;
 265                 $this->logger = $logger;
 266                 $this->clearStoreCallbacks = $clearStoreCallbacks;
 267                 $this->langNameUtils = $langNameUtils;
 268
 269                 // Keep this separate from $this->options so it can be mutable
 270                 $this->manualRecache = $options->get( 'manualRecache' );
 271         }
 272
 273         /**
 274          * Returns true if the given key is mergeable, that is, if it is an associative
 275          * array which can be merged through a fallback sequence.
 276          * @param string $key
 277          * @return bool
 278          */
 279         public function isMergeableKey( $key ) {
 280                 if ( $this->mergeableKeys === null ) {
 281                         $this->mergeableKeys = array_flip( array_merge(
 282                                 self::$mergeableMapKeys,
 283                                 self::$mergeableListKeys,
 284                                 self::$mergeableAliasListKeys,
 285                                 self::$optionalMergeKeys,
 286                                 self::$magicWordKeys
 287                         ) );
 288                 }
 289
 290                 return isset( $this->mergeableKeys[$key] );
 291         }
 292
 293         /**
 294          * Get a cache item.
 295          *
 296          * Warning: this may be slow for split items (messages), since it will
 297          * need to fetch all of the subitems from the cache individually.
 298          * @param string $code
 299          * @param string $key
 300          * @return mixed
 301          */
 302         public function getItem( $code, $key ) {
 303                 if ( !isset( $this->loadedItems[$code][$key] ) ) {
 304                         $this->loadItem( $code, $key );
 305                 }
 306
 307                 if ( $key === 'fallback' && isset( $this->shallowFallbacks[$code] ) ) {
 308                         return $this->shallowFallbacks[$code];
 309                 }
 310
 311                 return $this->data[$code][$key];
 312         }
 313
 314         /**
 315          * Get a subitem, for instance a single message for a given language.
 316          * @param string $code
 317          * @param string $key
 318          * @param string $subkey
 319          * @return mixed|null
 320          */
 321         public function getSubitem( $code, $key, $subkey ) {
 322                 if ( !isset( $this->loadedSubitems[$code][$key][$subkey] ) &&
 323                         !isset( $this->loadedItems[$code][$key] )
 324                 ) {
 325                         $this->loadSubitem( $code, $key, $subkey );
 326                 }
 327
 328                 return $this->data[$code][$key][$subkey] ?? null;
 329         }
 330
 331         /**
 332          * Get the list of subitem keys for a given item.
 333          *
 334          * This is faster than array_keys($lc->getItem(...)) for the items listed in
 335          * self::$splitKeys.
 336          *
 337          * Will return null if the item is not found, or false if the item is not an
 338          * array.
 339          * @param string $code
 340          * @param string $key
 341          * @return bool|null|string|string[]
 342          */
 343         public function getSubitemList( $code, $key ) {
 344                 if ( in_array( $key, self::$splitKeys ) ) {
 345                         return $this->getSubitem( $code, 'list', $key );
 346                 } else {
 347                         $item = $this->getItem( $code, $key );
 348                         if ( is_array( $item ) ) {
 349                                 return array_keys( $item );
 350                         } else {
 351                                 return false;
 352                         }
 353                 }
 354         }
 355
 356         /**
 357          * Load an item into the cache.
 358          * @param string $code
 359          * @param string $key
 360          */
 361         protected function loadItem( $code, $key ) {
 362                 if ( !isset( $this->initialisedLangs[$code] ) ) {
 363                         $this->initLanguage( $code );
 364                 }
 365
 366                 // Check to see if initLanguage() loaded it for us
 367                 if ( isset( $this->loadedItems[$code][$key] ) ) {
 368                         return;
 369                 }
 370
 371                 if ( isset( $this->shallowFallbacks[$code] ) ) {
 372                         $this->loadItem( $this->shallowFallbacks[$code], $key );
 373
 374                         return;
 375                 }
 376
 377                 if ( in_array( $key, self::$splitKeys ) ) {
 378                         $subkeyList = $this->getSubitem( $code, 'list', $key );
 379                         foreach ( $subkeyList as $subkey ) {
 380                                 if ( isset( $this->data[$code][$key][$subkey] ) ) {
 381                                         continue;
 382                                 }
 383                                 $this->data[$code][$key][$subkey] = $this->getSubitem( $code, $key, $subkey );
 384                         }
 385                 } else {
 386                         $this->data[$code][$key] = $this->store->get( $code, $key );
 387                 }
 388
 389                 $this->loadedItems[$code][$key] = true;
 390         }
 391
 392         /**
 393          * Load a subitem into the cache
 394          * @param string $code
 395          * @param string $key
 396          * @param string $subkey
 397          */
 398         protected function loadSubitem( $code, $key, $subkey ) {
 399                 if ( !in_array( $key, self::$splitKeys ) ) {
 400                         $this->loadItem( $code, $key );
 401
 402                         return;
 403                 }
 404
 405                 if ( !isset( $this->initialisedLangs[$code] ) ) {
 406                         $this->initLanguage( $code );
 407                 }
 408
 409                 // Check to see if initLanguage() loaded it for us
 410                 if ( isset( $this->loadedItems[$code][$key] ) ||
 411                         isset( $this->loadedSubitems[$code][$key][$subkey] )
 412                 ) {
 413                         return;
 414                 }
 415
 416                 if ( isset( $this->shallowFallbacks[$code] ) ) {
 417                         $this->loadSubitem( $this->shallowFallbacks[$code], $key, $subkey );
 418
 419                         return;
 420                 }
 421
 422                 $value = $this->store->get( $code, "$key:$subkey" );
 423                 $this->data[$code][$key][$subkey] = $value;
 424                 $this->loadedSubitems[$code][$key][$subkey] = true;
 425         }
 426
 427         /**
 428          * Returns true if the cache identified by $code is missing or expired.
 429          *
 430          * @param string $code
 431          *
 432          * @return bool
 433          */
 434         public function isExpired( $code ) {
 435                 if ( $this->options->get( 'forceRecache' ) && !isset( $this->recachedLangs[$code] ) ) {
 436                         $this->logger->debug( __METHOD__ . "($code): forced reload" );
 437
 438                         return true;
 439                 }
 440
 441                 $deps = $this->store->get( $code, 'deps' );
 442                 $keys = $this->store->get( $code, 'list' );
 443                 $preload = $this->store->get( $code, 'preload' );
 444                 // Different keys may expire separately for some stores
 445                 if ( $deps === null || $keys === null || $preload === null ) {
 446                         $this->logger->debug( __METHOD__ . "($code): cache missing, need to make one" );
 447
 448                         return true;
 449                 }
 450
 451                 foreach ( $deps as $dep ) {
 452                         // Because we're unserializing stuff from cache, we
 453                         // could receive objects of classes that don't exist
 454                         // anymore (e.g. uninstalled extensions)
 455                         // When this happens, always expire the cache
 456                         if ( !$dep instanceof CacheDependency || $dep->isExpired() ) {
 457                                 $this->logger->debug( __METHOD__ . "($code): cache for $code expired due to " .
 458                                         get_class( $dep ) );
 459
 460                                 return true;
 461                         }
 462                 }
 463
 464                 return false;
 465         }
 466
 467         /**
 468          * Initialise a language in this object. Rebuild the cache if necessary.
 469          * @param string $code
 470          * @throws MWException
 471          */
 472         protected function initLanguage( $code ) {
 473                 if ( isset( $this->initialisedLangs[$code] ) ) {
 474                         return;
 475                 }
 476
 477                 $this->initialisedLangs[$code] = true;
 478
 479                 # If the code is of the wrong form for a Messages*.php file, do a shallow fallback
 480                 if ( !$this->langNameUtils->isValidBuiltInCode( $code ) ) {
 481                         $this->initShallowFallback( $code, 'en' );
 482
 483                         return;
 484                 }
 485
 486                 # Recache the data if necessary
 487                 if ( !$this->manualRecache && $this->isExpired( $code ) ) {
 488                         if ( $this->langNameUtils->isSupportedLanguage( $code ) ) {
 489                                 $this->recache( $code );
 490                         } elseif ( $code === 'en' ) {
 491                                 throw new MWException( 'MessagesEn.php is missing.' );
 492                         } else {
 493                                 $this->initShallowFallback( $code, 'en' );
 494                         }
 495
 496                         return;
 497                 }
 498
 499                 # Preload some stuff
 500                 $preload = $this->getItem( $code, 'preload' );
 501                 if ( $preload === null ) {
 502                         if ( $this->manualRecache ) {
 503                                 // No Messages*.php file. Do shallow fallback to en.
 504                                 if ( $code === 'en' ) {
 505                                         throw new MWException( 'No localisation cache found for English. ' .
 506                                                 'Please run maintenance/rebuildLocalisationCache.php.' );
 507                                 }
 508                                 $this->initShallowFallback( $code, 'en' );
 509
 510                                 return;
 511                         } else {
 512                                 throw new MWException( 'Invalid or missing localisation cache.' );
 513                         }
 514                 }
 515                 $this->data[$code] = $preload;
 516                 foreach ( $preload as $key => $item ) {
 517                         if ( in_array( $key, self::$splitKeys ) ) {
 518                                 foreach ( $item as $subkey => $subitem ) {
 519                                         $this->loadedSubitems[$code][$key][$subkey] = true;
 520                                 }
 521                         } else {
 522                                 $this->loadedItems[$code][$key] = true;
 523                         }
 524                 }
 525         }
 526
 527         /**
 528          * Create a fallback from one language to another, without creating a
 529          * complete persistent cache.
 530          * @param string $primaryCode
 531          * @param string $fallbackCode
 532          */
 533         public function initShallowFallback( $primaryCode, $fallbackCode ) {
 534                 $this->data[$primaryCode] =& $this->data[$fallbackCode];
 535                 $this->loadedItems[$primaryCode] =& $this->loadedItems[$fallbackCode];
 536                 $this->loadedSubitems[$primaryCode] =& $this->loadedSubitems[$fallbackCode];
 537                 $this->shallowFallbacks[$primaryCode] = $fallbackCode;
 538         }
 539
 540         /**
 541          * Read a PHP file containing localisation data.
 542          * @param string $_fileName
 543          * @param string $_fileType
 544          * @throws MWException
 545          * @return array
 546          */
 547         protected function readPHPFile( $_fileName, $_fileType ) {
 548                 // Disable APC caching
 549                 Wikimedia\suppressWarnings();
 550                 $_apcEnabled = ini_set( 'apc.cache_by_default', '0' );
 551                 Wikimedia\restoreWarnings();
 552
 553                 include $_fileName;
 554
 555                 Wikimedia\suppressWarnings();
 556                 ini_set( 'apc.cache_by_default', $_apcEnabled );
 557                 Wikimedia\restoreWarnings();
 558
 559                 $data = [];
 560                 if ( $_fileType == 'core' || $_fileType == 'extension' ) {
 561                         foreach ( self::$allKeys as $key ) {
 562                                 // Not all keys are set in language files, so
 563                                 // check they exist first
 564                                 if ( isset( $$key ) ) {
 565                                         $data[$key] = $$key;
 566                                 }
 567                         }
 568                 } elseif ( $_fileType == 'aliases' ) {
 569                         if ( isset( $aliases ) ) {
 570                                 $data['aliases'] = $aliases;
 571                         }
 572                 } else {
 573                         throw new MWException( __METHOD__ . ": Invalid file type: $_fileType" );
 574                 }
 575
 576                 return $data;
 577         }
 578
 579         /**
 580          * Read a JSON file containing localisation messages.
 581          * @param string $fileName Name of file to read
 582          * @throws MWException If there is a syntax error in the JSON file
 583          * @return array Array with a 'messages' key, or empty array if the file doesn't exist
 584          */
 585         public function readJSONFile( $fileName ) {
 586                 if ( !is_readable( $fileName ) ) {
 587                         return [];
 588                 }
 589
 590                 $json = file_get_contents( $fileName );
 591                 if ( $json === false ) {
 592                         return [];
 593                 }
 594
 595                 $data = FormatJson::decode( $json, true );
 596                 if ( $data === null ) {
 597                         throw new MWException( __METHOD__ . ": Invalid JSON file: $fileName" );
 598                 }
 599
 600                 // Remove keys starting with '@', they're reserved for metadata and non-message data
 601                 foreach ( $data as $key => $unused ) {
 602                         if ( $key === '' || $key[0] === '@' ) {
 603                                 unset( $data[$key] );
 604                         }
 605                 }
 606
 607                 // The JSON format only supports messages, none of the other variables, so wrap the data
 608                 return [ 'messages' => $data ];
 609         }
 610
 611         /**
 612          * Get the compiled plural rules for a given language from the XML files.
 613          * @since 1.20
 614          * @param string $code
 615          * @return array|null
 616          */
 617         public function getCompiledPluralRules( $code ) {
 618                 $rules = $this->getPluralRules( $code );
 619                 if ( $rules === null ) {
 620                         return null;
 621                 }
 622                 try {
 623                         $compiledRules = Evaluator::compile( $rules );
 624                 } catch ( CLDRPluralRuleError $e ) {
 625                         $this->logger->debug( $e->getMessage() );
 626
 627                         return [];
 628                 }
 629
 630                 return $compiledRules;
 631         }
 632
 633         /**
 634          * Get the plural rules for a given language from the XML files.
 635          * Cached.
 636          * @since 1.20
 637          * @param string $code
 638          * @return array|null
 639          */
 640         public function getPluralRules( $code ) {
 641                 if ( $this->pluralRules === null ) {
 642                         $this->loadPluralFiles();
 643                 }
 644                 return $this->pluralRules[$code] ?? null;
 645         }
 646
 647         /**
 648          * Get the plural rule types for a given language from the XML files.
 649          * Cached.
 650          * @since 1.22
 651          * @param string $code
 652          * @return array|null
 653          */
 654         public function getPluralRuleTypes( $code ) {
 655                 if ( $this->pluralRuleTypes === null ) {
 656                         $this->loadPluralFiles();
 657                 }
 658                 return $this->pluralRuleTypes[$code] ?? null;
 659         }
 660
 661         /**
 662          * Load the plural XML files.
 663          */
 664         protected function loadPluralFiles() {
 665                 global $IP;
 666                 $cldrPlural = "$IP/languages/data/plurals.xml";
 667                 $mwPlural = "$IP/languages/data/plurals-mediawiki.xml";
 668                 // Load CLDR plural rules
 669                 $this->loadPluralFile( $cldrPlural );
 670                 if ( file_exists( $mwPlural ) ) {
 671                         // Override or extend
 672                         $this->loadPluralFile( $mwPlural );
 673                 }
 674         }
 675
 676         /**
 677          * Load a plural XML file with the given filename, compile the relevant
 678          * rules, and save the compiled rules in a process-local cache.
 679          *
 680          * @param string $fileName
 681          * @throws MWException
 682          */
 683         protected function loadPluralFile( $fileName ) {
 684                 // Use file_get_contents instead of DOMDocument::load (T58439)
 685                 $xml = file_get_contents( $fileName );
 686                 if ( !$xml ) {
 687                         throw new MWException( "Unable to read plurals file $fileName" );
 688                 }
 689                 $doc = new DOMDocument;
 690                 $doc->loadXML( $xml );
 691                 $rulesets = $doc->getElementsByTagName( "pluralRules" );
 692                 foreach ( $rulesets as $ruleset ) {
 693                         $codes = $ruleset->getAttribute( 'locales' );
 694                         $rules = [];
 695                         $ruleTypes = [];
 696                         $ruleElements = $ruleset->getElementsByTagName( "pluralRule" );
 697                         foreach ( $ruleElements as $elt ) {
 698                                 $ruleType = $elt->getAttribute( 'count' );
 699                                 if ( $ruleType === 'other' ) {
 700                                         // Don't record "other" rules, which have an empty condition
 701                                         continue;
 702                                 }
 703                                 $rules[] = $elt->nodeValue;
 704                                 $ruleTypes[] = $ruleType;
 705                         }
 706                         foreach ( explode( ' ', $codes ) as $code ) {
 707                                 $this->pluralRules[$code] = $rules;
 708                                 $this->pluralRuleTypes[$code] = $ruleTypes;
 709                         }
 710                 }
 711         }
 712
 713         /**
 714          * Read the data from the source files for a given language, and register
 715          * the relevant dependencies in the $deps array. If the localisation
 716          * exists, the data array is returned, otherwise false is returned.
 717          *
 718          * @param string $code
 719          * @param array &$deps
 720          * @return array
 721          */
 722         protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
 723                 global $IP;
 724
 725                 // This reads in the PHP i18n file with non-messages l10n data
 726                 $fileName = $this->langNameUtils->getMessagesFileName( $code );
 727                 if ( !file_exists( $fileName ) ) {
 728                         $data = [];
 729                 } else {
 730                         $deps[] = new FileDependency( $fileName );
 731                         $data = $this->readPHPFile( $fileName, 'core' );
 732                 }
 733
 734                 # Load CLDR plural rules for JavaScript
 735                 $data['pluralRules'] = $this->getPluralRules( $code );
 736                 # And for PHP
 737                 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
 738                 # Load plural rule types
 739                 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
 740
 741                 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
 742                 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
 743
 744                 return $data;
 745         }
 746
 747         /**
 748          * Merge two localisation values, a primary and a fallback, overwriting the
 749          * primary value in place.
 750          * @param string $key
 751          * @param mixed &$value
 752          * @param mixed $fallbackValue
 753          */
 754         protected function mergeItem( $key, &$value, $fallbackValue ) {
 755                 if ( !is_null( $value ) ) {
 756                         if ( !is_null( $fallbackValue ) ) {
 757                                 if ( in_array( $key, self::$mergeableMapKeys ) ) {
 758                                         $value = $value + $fallbackValue;
 759                                 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
 760                                         $value = array_unique( array_merge( $fallbackValue, $value ) );
 761                                 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
 762                                         $value = array_merge_recursive( $value, $fallbackValue );
 763                                 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
 764                                         if ( !empty( $value['inherit'] ) ) {
 765                                                 $value = array_merge( $fallbackValue, $value );
 766                                         }
 767
 768                                         if ( isset( $value['inherit'] ) ) {
 769                                                 unset( $value['inherit'] );
 770                                         }
 771                                 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
 772                                         $this->mergeMagicWords( $value, $fallbackValue );
 773                                 }
 774                         }
 775                 } else {
 776                         $value = $fallbackValue;
 777                 }
 778         }
 779
 780         /**
 781          * @param mixed &$value
 782          * @param mixed $fallbackValue
 783          */
 784         protected function mergeMagicWords( &$value, $fallbackValue ) {
 785                 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
 786                         if ( !isset( $value[$magicName] ) ) {
 787                                 $value[$magicName] = $fallbackInfo;
 788                         } else {
 789                                 $oldSynonyms = array_slice( $fallbackInfo, 1 );
 790                                 $newSynonyms = array_slice( $value[$magicName], 1 );
 791                                 $synonyms = array_values( array_unique( array_merge(
 792                                         $newSynonyms, $oldSynonyms ) ) );
 793                                 $value[$magicName] = array_merge( [ $fallbackInfo[0] ], $synonyms );
 794                         }
 795                 }
 796         }
 797
 798         /**
 799          * Given an array mapping language code to localisation value, such as is
 800          * found in extension *.i18n.php files, iterate through a fallback sequence
 801          * to merge the given data with an existing primary value.
 802          *
 803          * Returns true if any data from the extension array was used, false
 804          * otherwise.
 805          * @param array $codeSequence
 806          * @param string $key
 807          * @param mixed &$value
 808          * @param mixed $fallbackValue
 809          * @return bool
 810          */
 811         protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
 812                 $used = false;
 813                 foreach ( $codeSequence as $code ) {
 814                         if ( isset( $fallbackValue[$code] ) ) {
 815                                 $this->mergeItem( $key, $value, $fallbackValue[$code] );
 816                                 $used = true;
 817                         }
 818                 }
 819
 820                 return $used;
 821         }
 822
 823         /**
 824          * Gets the combined list of messages dirs from
 825          * core and extensions
 826          *
 827          * @since 1.25
 828          * @return array
 829          */
 830         public function getMessagesDirs() {
 831                 global $IP;
 832
 833                 return [
 834                         'core' => "$IP/languages/i18n",
 835                         'exif' => "$IP/languages/i18n/exif",
 836                         'api' => "$IP/includes/api/i18n",
 837                         'oojs-ui' => "$IP/resources/lib/ooui/i18n",
 838                 ] + $this->options->get( 'MessagesDirs' );
 839         }
 840
 841         /**
 842          * Load localisation data for a given language for both core and extensions
 843          * and save it to the persistent cache store and the process cache
 844          * @param string $code
 845          * @throws MWException
 846          */
 847         public function recache( $code ) {
 848                 if ( !$code ) {
 849                         throw new MWException( "Invalid language code requested" );
 850                 }
 851                 $this->recachedLangs[$code] = true;
 852
 853                 # Initial values
 854                 $initialData = array_fill_keys( self::$allKeys, null );
 855                 $coreData = $initialData;
 856                 $deps = [];
 857
 858                 # Load the primary localisation from the source file
 859                 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
 860                 if ( $data === false ) {
 861                         $this->logger->debug( __METHOD__ . ": no localisation file for $code, using fallback to en" );
 862                         $coreData['fallback'] = 'en';
 863                 } else {
 864                         $this->logger->debug( __METHOD__ . ": got localisation for $code from source" );
 865
 866                         # Merge primary localisation
 867                         foreach ( $data as $key => $value ) {
 868                                 $this->mergeItem( $key, $coreData[$key], $value );
 869                         }
 870                 }
 871
 872                 # Fill in the fallback if it's not there already
 873                 if ( ( is_null( $coreData['fallback'] ) || $coreData['fallback'] === false ) && $code === 'en' ) {
 874                         $coreData['fallback'] = false;
 875                         $coreData['originalFallbackSequence'] = $coreData['fallbackSequence'] = [];
 876                 } else {
 877                         if ( !is_null( $coreData['fallback'] ) ) {
 878                                 $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
 879                         } else {
 880                                 $coreData['fallbackSequence'] = [];
 881                         }
 882                         $len = count( $coreData['fallbackSequence'] );
 883
 884                         # Before we add the 'en' fallback for messages, keep a copy of
 885                         # the original fallback sequence
 886                         $coreData['originalFallbackSequence'] = $coreData['fallbackSequence'];
 887
 888                         # Ensure that the sequence ends at 'en' for messages
 889                         if ( !$len || $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
 890                                 $coreData['fallbackSequence'][] = 'en';
 891                         }
 892                 }
 893
 894                 $codeSequence = array_merge( [ $code ], $coreData['fallbackSequence'] );
 895                 $messageDirs = $this->getMessagesDirs();
 896
 897                 # Load non-JSON localisation data for extensions
 898                 $extensionData = array_fill_keys( $codeSequence, $initialData );
 899                 foreach ( $this->options->get( 'ExtensionMessagesFiles' ) as $extension => $fileName ) {
 900                         if ( isset( $messageDirs[$extension] ) ) {
 901                                 # This extension has JSON message data; skip the PHP shim
 902                                 continue;
 903                         }
 904
 905                         $data = $this->readPHPFile( $fileName, 'extension' );
 906                         $used = false;
 907
 908                         foreach ( $data as $key => $item ) {
 909                                 foreach ( $codeSequence as $csCode ) {
 910                                         if ( isset( $item[$csCode] ) ) {
 911                                                 $this->mergeItem( $key, $extensionData[$csCode][$key], $item[$csCode] );
 912                                                 $used = true;
 913                                         }
 914                                 }
 915                         }
 916
 917                         if ( $used ) {
 918                                 $deps[] = new FileDependency( $fileName );
 919                         }
 920                 }
 921
 922                 # Load the localisation data for each fallback, then merge it into the full array
 923                 $allData = $initialData;
 924                 foreach ( $codeSequence as $csCode ) {
 925                         $csData = $initialData;
 926
 927                         # Load core messages and the extension localisations.
 928                         foreach ( $messageDirs as $dirs ) {
 929                                 foreach ( (array)$dirs as $dir ) {
 930                                         $fileName = "$dir/$csCode.json";
 931                                         $data = $this->readJSONFile( $fileName );
 932
 933                                         foreach ( $data as $key => $item ) {
 934                                                 $this->mergeItem( $key, $csData[$key], $item );
 935                                         }
 936
 937                                         $deps[] = new FileDependency( $fileName );
 938                                 }
 939                         }
 940
 941                         # Merge non-JSON extension data
 942                         if ( isset( $extensionData[$csCode] ) ) {
 943                                 foreach ( $extensionData[$csCode] as $key => $item ) {
 944                                         $this->mergeItem( $key, $csData[$key], $item );
 945                                 }
 946                         }
 947
 948                         if ( $csCode === $code ) {
 949                                 # Merge core data into extension data
 950                                 foreach ( $coreData as $key => $item ) {
 951                                         $this->mergeItem( $key, $csData[$key], $item );
 952                                 }
 953                         } else {
 954                                 # Load the secondary localisation from the source file to
 955                                 # avoid infinite cycles on cyclic fallbacks
 956                                 $fbData = $this->readSourceFilesAndRegisterDeps( $csCode, $deps );
 957                                 if ( $fbData !== false ) {
 958                                         # Only merge the keys that make sense to merge
 959                                         foreach ( self::$allKeys as $key ) {
 960                                                 if ( !isset( $fbData[$key] ) ) {
 961                                                         continue;
 962                                                 }
 963
 964                                                 if ( is_null( $coreData[$key] ) || $this->isMergeableKey( $key ) ) {
 965                                                         $this->mergeItem( $key, $csData[$key], $fbData[$key] );
 966                                                 }
 967                                         }
 968                                 }
 969                         }
 970
 971                         # Allow extensions an opportunity to adjust the data for this
 972                         # fallback
 973                         Hooks::run( 'LocalisationCacheRecacheFallback', [ $this, $csCode, &$csData ] );
 974
 975                         # Merge the data for this fallback into the final array
 976                         if ( $csCode === $code ) {
 977                                 $allData = $csData;
 978                         } else {
 979                                 foreach ( self::$allKeys as $key ) {
 980                                         if ( !isset( $csData[$key] ) ) {
 981                                                 continue;
 982                                         }
 983
 984                                         if ( is_null( $allData[$key] ) || $this->isMergeableKey( $key ) ) {
 985                                                 $this->mergeItem( $key, $allData[$key], $csData[$key] );
 986                                         }
 987                                 }
 988                         }
 989                 }
 990
 991                 # Add cache dependencies for any referenced globals
 992                 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
 993                 // The 'MessagesDirs' config setting is used in LocalisationCache::getMessagesDirs().
 994                 // We use the key 'wgMessagesDirs' for historical reasons.
 995                 $deps['wgMessagesDirs'] = new MainConfigDependency( 'MessagesDirs' );
 996                 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
 997
 998                 # Add dependencies to the cache entry
 999                 $allData['deps'] = $deps;
1000
1001                 # Replace spaces with underscores in namespace names
1002                 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
1003
1004                 # And do the same for special page aliases. $page is an array.
1005                 foreach ( $allData['specialPageAliases'] as &$page ) {
1006                         $page = str_replace( ' ', '_', $page );
1007                 }
1008                 # Decouple the reference to prevent accidental damage
1009                 unset( $page );
1010
1011                 # If there were no plural rules, return an empty array
1012                 if ( $allData['pluralRules'] === null ) {
1013                         $allData['pluralRules'] = [];
1014                 }
1015                 if ( $allData['compiledPluralRules'] === null ) {
1016                         $allData['compiledPluralRules'] = [];
1017                 }
1018                 # If there were no plural rule types, return an empty array
1019                 if ( $allData['pluralRuleTypes'] === null ) {
1020                         $allData['pluralRuleTypes'] = [];
1021                 }
1022
1023                 # Set the list keys
1024                 $allData['list'] = [];
1025                 foreach ( self::$splitKeys as $key ) {
1026                         $allData['list'][$key] = array_keys( $allData[$key] );
1027                 }
1028                 # Run hooks
1029                 $unused = true; // Used to be $purgeBlobs, removed in 1.34
1030                 Hooks::run( 'LocalisationCacheRecache', [ $this, $code, &$allData, &$unused ] );
1031
1032                 if ( is_null( $allData['namespaceNames'] ) ) {
1033                         throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
1034                                 'Check that your languages/messages/MessagesEn.php file is intact.' );
1035                 }
1036
1037                 # Set the preload key
1038                 $allData['preload'] = $this->buildPreload( $allData );
1039
1040                 # Save to the process cache and register the items loaded
1041                 $this->data[$code] = $allData;
1042                 foreach ( $allData as $key => $item ) {
1043                         $this->loadedItems[$code][$key] = true;
1044                 }
1045
1046                 # Save to the persistent cache
1047                 $this->store->startWrite( $code );
1048                 foreach ( $allData as $key => $value ) {
1049                         if ( in_array( $key, self::$splitKeys ) ) {
1050                                 foreach ( $value as $subkey => $subvalue ) {
1051                                         $this->store->set( "$key:$subkey", $subvalue );
1052                                 }
1053                         } else {
1054                                 $this->store->set( $key, $value );
1055                         }
1056                 }
1057                 $this->store->finishWrite();
1058
1059                 # Clear out the MessageBlobStore
1060                 # HACK: If using a null (i.e. disabled) storage backend, we
1061                 # can't write to the MessageBlobStore either
1062                 if ( !$this->store instanceof LCStoreNull ) {
1063                         foreach ( $this->clearStoreCallbacks as $callback ) {
1064                                 $callback();
1065                         }
1066                 }
1067         }
1068
1069         /**
1070          * Build the preload item from the given pre-cache data.
1071          *
1072          * The preload item will be loaded automatically, improving performance
1073          * for the commonly-requested items it contains.
1074          * @param array $data
1075          * @return array
1076          */
1077         protected function buildPreload( $data ) {
1078                 $preload = [ 'messages' => [] ];
1079                 foreach ( self::$preloadedKeys as $key ) {
1080                         $preload[$key] = $data[$key];
1081                 }
1082
1083                 foreach ( $data['preloadedMessages'] as $subkey ) {
1084                         $subitem = $data['messages'][$subkey] ?? null;
1085                         $preload['messages'][$subkey] = $subitem;
1086                 }
1087
1088                 return $preload;
1089         }
1090
1091         /**
1092          * Unload the data for a given language from the object cache.
1093          * Reduces memory usage.
1094          * @param string $code
1095          */
1096         public function unload( $code ) {
1097                 unset( $this->data[$code] );
1098                 unset( $this->loadedItems[$code] );
1099                 unset( $this->loadedSubitems[$code] );
1100                 unset( $this->initialisedLangs[$code] );
1101                 unset( $this->shallowFallbacks[$code] );
1102
1103                 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1104                         if ( $fbCode === $code ) {
1105                                 $this->unload( $shallowCode );
1106                         }
1107                 }
1108         }
1109
1110         /**
1111          * Unload all data
1112          */
1113         public function unloadAll() {
1114                 foreach ( $this->initialisedLangs as $lang => $unused ) {
1115                         $this->unload( $lang );
1116                 }
1117         }
1118
1119         /**
1120          * Disable the storage backend
1121          */
1122         public function disableBackend() {
1123                 $this->store = new LCStoreNull;
1124                 $this->manualRecache = false;
1125         }
1126 }