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