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