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