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