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