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                 $data = [];
 531                 if ( $_fileType == 'core' || $_fileType == 'extension' ) {
 532                         foreach ( self::$allKeys as $key ) {
 533                                 // Not all keys are set in language files, so
 534                                 // check they exist first
 535                                 if ( isset( $$key ) ) {
 536                                         $data[$key] = $$key;
 537                                 }
 538                         }
 539                 } elseif ( $_fileType == 'aliases' ) {
 540                         if ( isset( $aliases ) ) {
 541                                 /** @suppress PhanUndeclaredVariable */
 542                                 $data['aliases'] = $aliases;
 543                         }
 544                 } else {
 545                         throw new MWException( __METHOD__ . ": Invalid file type: $_fileType" );
 546                 }
 547
 548                 return $data;
 549         }
 550
 551         /**
 552          * Read a JSON file containing localisation messages.
 553          * @param string $fileName Name of file to read
 554          * @throws MWException If there is a syntax error in the JSON file
 555          * @return array Array with a 'messages' key, or empty array if the file doesn't exist
 556          */
 557         public function readJSONFile( $fileName ) {
 558                 if ( !is_readable( $fileName ) ) {
 559                         return [];
 560                 }
 561
 562                 $json = file_get_contents( $fileName );
 563                 if ( $json === false ) {
 564                         return [];
 565                 }
 566
 567                 $data = FormatJson::decode( $json, true );
 568                 if ( $data === null ) {
 569                         throw new MWException( __METHOD__ . ": Invalid JSON file: $fileName" );
 570                 }
 571
 572                 // Remove keys starting with '@', they're reserved for metadata and non-message data
 573                 foreach ( $data as $key => $unused ) {
 574                         if ( $key === '' || $key[0] === '@' ) {
 575                                 unset( $data[$key] );
 576                         }
 577                 }
 578
 579                 // The JSON format only supports messages, none of the other variables, so wrap the data
 580                 return [ 'messages' => $data ];
 581         }
 582
 583         /**
 584          * Get the compiled plural rules for a given language from the XML files.
 585          * @since 1.20
 586          * @param string $code
 587          * @return array|null
 588          */
 589         public function getCompiledPluralRules( $code ) {
 590                 $rules = $this->getPluralRules( $code );
 591                 if ( $rules === null ) {
 592                         return null;
 593                 }
 594                 try {
 595                         $compiledRules = Evaluator::compile( $rules );
 596                 } catch ( CLDRPluralRuleError $e ) {
 597                         wfDebugLog( 'l10n', $e->getMessage() );
 598
 599                         return [];
 600                 }
 601
 602                 return $compiledRules;
 603         }
 604
 605         /**
 606          * Get the plural rules for a given language from the XML files.
 607          * Cached.
 608          * @since 1.20
 609          * @param string $code
 610          * @return array|null
 611          */
 612         public function getPluralRules( $code ) {
 613                 if ( $this->pluralRules === null ) {
 614                         $this->loadPluralFiles();
 615                 }
 616                 if ( !isset( $this->pluralRules[$code] ) ) {
 617                         return null;
 618                 } else {
 619                         return $this->pluralRules[$code];
 620                 }
 621         }
 622
 623         /**
 624          * Get the plural rule types for a given language from the XML files.
 625          * Cached.
 626          * @since 1.22
 627          * @param string $code
 628          * @return array|null
 629          */
 630         public function getPluralRuleTypes( $code ) {
 631                 if ( $this->pluralRuleTypes === null ) {
 632                         $this->loadPluralFiles();
 633                 }
 634                 if ( !isset( $this->pluralRuleTypes[$code] ) ) {
 635                         return null;
 636                 } else {
 637                         return $this->pluralRuleTypes[$code];
 638                 }
 639         }
 640
 641         /**
 642          * Load the plural XML files.
 643          */
 644         protected function loadPluralFiles() {
 645                 global $IP;
 646                 $cldrPlural = "$IP/languages/data/plurals.xml";
 647                 $mwPlural = "$IP/languages/data/plurals-mediawiki.xml";
 648                 // Load CLDR plural rules
 649                 $this->loadPluralFile( $cldrPlural );
 650                 if ( file_exists( $mwPlural ) ) {
 651                         // Override or extend
 652                         $this->loadPluralFile( $mwPlural );
 653                 }
 654         }
 655
 656         /**
 657          * Load a plural XML file with the given filename, compile the relevant
 658          * rules, and save the compiled rules in a process-local cache.
 659          *
 660          * @param string $fileName
 661          * @throws MWException
 662          */
 663         protected function loadPluralFile( $fileName ) {
 664                 // Use file_get_contents instead of DOMDocument::load (T58439)
 665                 $xml = file_get_contents( $fileName );
 666                 if ( !$xml ) {
 667                         throw new MWException( "Unable to read plurals file $fileName" );
 668                 }
 669                 $doc = new DOMDocument;
 670                 $doc->loadXML( $xml );
 671                 $rulesets = $doc->getElementsByTagName( "pluralRules" );
 672                 foreach ( $rulesets as $ruleset ) {
 673                         $codes = $ruleset->getAttribute( 'locales' );
 674                         $rules = [];
 675                         $ruleTypes = [];
 676                         $ruleElements = $ruleset->getElementsByTagName( "pluralRule" );
 677                         foreach ( $ruleElements as $elt ) {
 678                                 $ruleType = $elt->getAttribute( 'count' );
 679                                 if ( $ruleType === 'other' ) {
 680                                         // Don't record "other" rules, which have an empty condition
 681                                         continue;
 682                                 }
 683                                 $rules[] = $elt->nodeValue;
 684                                 $ruleTypes[] = $ruleType;
 685                         }
 686                         foreach ( explode( ' ', $codes ) as $code ) {
 687                                 $this->pluralRules[$code] = $rules;
 688                                 $this->pluralRuleTypes[$code] = $ruleTypes;
 689                         }
 690                 }
 691         }
 692
 693         /**
 694          * Read the data from the source files for a given language, and register
 695          * the relevant dependencies in the $deps array. If the localisation
 696          * exists, the data array is returned, otherwise false is returned.
 697          *
 698          * @param string $code
 699          * @param array &$deps
 700          * @return array
 701          */
 702         protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
 703                 global $IP;
 704
 705                 // This reads in the PHP i18n file with non-messages l10n data
 706                 $fileName = Language::getMessagesFileName( $code );
 707                 if ( !file_exists( $fileName ) ) {
 708                         $data = [];
 709                 } else {
 710                         $deps[] = new FileDependency( $fileName );
 711                         $data = $this->readPHPFile( $fileName, 'core' );
 712                 }
 713
 714                 # Load CLDR plural rules for JavaScript
 715                 $data['pluralRules'] = $this->getPluralRules( $code );
 716                 # And for PHP
 717                 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
 718                 # Load plural rule types
 719                 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
 720
 721                 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
 722                 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
 723
 724                 return $data;
 725         }
 726
 727         /**
 728          * Merge two localisation values, a primary and a fallback, overwriting the
 729          * primary value in place.
 730          * @param string $key
 731          * @param mixed &$value
 732          * @param mixed $fallbackValue
 733          */
 734         protected function mergeItem( $key, &$value, $fallbackValue ) {
 735                 if ( !is_null( $value ) ) {
 736                         if ( !is_null( $fallbackValue ) ) {
 737                                 if ( in_array( $key, self::$mergeableMapKeys ) ) {
 738                                         $value = $value + $fallbackValue;
 739                                 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
 740                                         $value = array_unique( array_merge( $fallbackValue, $value ) );
 741                                 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
 742                                         $value = array_merge_recursive( $value, $fallbackValue );
 743                                 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
 744                                         if ( !empty( $value['inherit'] ) ) {
 745                                                 $value = array_merge( $fallbackValue, $value );
 746                                         }
 747
 748                                         if ( isset( $value['inherit'] ) ) {
 749                                                 unset( $value['inherit'] );
 750                                         }
 751                                 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
 752                                         $this->mergeMagicWords( $value, $fallbackValue );
 753                                 }
 754                         }
 755                 } else {
 756                         $value = $fallbackValue;
 757                 }
 758         }
 759
 760         /**
 761          * @param mixed &$value
 762          * @param mixed $fallbackValue
 763          */
 764         protected function mergeMagicWords( &$value, $fallbackValue ) {
 765                 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
 766                         if ( !isset( $value[$magicName] ) ) {
 767                                 $value[$magicName] = $fallbackInfo;
 768                         } else {
 769                                 $oldSynonyms = array_slice( $fallbackInfo, 1 );
 770                                 $newSynonyms = array_slice( $value[$magicName], 1 );
 771                                 $synonyms = array_values( array_unique( array_merge(
 772                                         $newSynonyms, $oldSynonyms ) ) );
 773                                 $value[$magicName] = array_merge( [ $fallbackInfo[0] ], $synonyms );
 774                         }
 775                 }
 776         }
 777
 778         /**
 779          * Given an array mapping language code to localisation value, such as is
 780          * found in extension *.i18n.php files, iterate through a fallback sequence
 781          * to merge the given data with an existing primary value.
 782          *
 783          * Returns true if any data from the extension array was used, false
 784          * otherwise.
 785          * @param array $codeSequence
 786          * @param string $key
 787          * @param mixed &$value
 788          * @param mixed $fallbackValue
 789          * @return bool
 790          */
 791         protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
 792                 $used = false;
 793                 foreach ( $codeSequence as $code ) {
 794                         if ( isset( $fallbackValue[$code] ) ) {
 795                                 $this->mergeItem( $key, $value, $fallbackValue[$code] );
 796                                 $used = true;
 797                         }
 798                 }
 799
 800                 return $used;
 801         }
 802
 803         /**
 804          * Gets the combined list of messages dirs from
 805          * core and extensions
 806          *
 807          * @since 1.25
 808          * @return array
 809          */
 810         public function getMessagesDirs() {
 811                 global $IP;
 812
 813                 $config = MediaWikiServices::getInstance()->getMainConfig();
 814                 $messagesDirs = $config->get( 'MessagesDirs' );
 815                 return [
 816                         'core' => "$IP/languages/i18n",
 817                         'api' => "$IP/includes/api/i18n",
 818                         'oojs-ui' => "$IP/resources/lib/oojs-ui/i18n",
 819                 ] + $messagesDirs;
 820         }
 821
 822         /**
 823          * Load localisation data for a given language for both core and extensions
 824          * and save it to the persistent cache store and the process cache
 825          * @param string $code
 826          * @throws MWException
 827          */
 828         public function recache( $code ) {
 829                 global $wgExtensionMessagesFiles;
 830
 831                 if ( !$code ) {
 832                         throw new MWException( "Invalid language code requested" );
 833                 }
 834                 $this->recachedLangs[$code] = true;
 835
 836                 # Initial values
 837                 $initialData = array_fill_keys( self::$allKeys, null );
 838                 $coreData = $initialData;
 839                 $deps = [];
 840
 841                 # Load the primary localisation from the source file
 842                 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
 843                 if ( $data === false ) {
 844                         wfDebug( __METHOD__ . ": no localisation file for $code, using fallback to en\n" );
 845                         $coreData['fallback'] = 'en';
 846                 } else {
 847                         wfDebug( __METHOD__ . ": got localisation for $code from source\n" );
 848
 849                         # Merge primary localisation
 850                         foreach ( $data as $key => $value ) {
 851                                 $this->mergeItem( $key, $coreData[$key], $value );
 852                         }
 853                 }
 854
 855                 # Fill in the fallback if it's not there already
 856                 if ( is_null( $coreData['fallback'] ) ) {
 857                         $coreData['fallback'] = $code === 'en' ? false : 'en';
 858                 }
 859                 if ( $coreData['fallback'] === false ) {
 860                         $coreData['fallbackSequence'] = [];
 861                 } else {
 862                         $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
 863                         $len = count( $coreData['fallbackSequence'] );
 864
 865                         # Ensure that the sequence ends at en
 866                         if ( $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
 867                                 $coreData['fallbackSequence'][] = 'en';
 868                         }
 869                 }
 870
 871                 $codeSequence = array_merge( [ $code ], $coreData['fallbackSequence'] );
 872                 $messageDirs = $this->getMessagesDirs();
 873
 874                 # Load non-JSON localisation data for extensions
 875                 $extensionData = array_fill_keys( $codeSequence, $initialData );
 876                 foreach ( $wgExtensionMessagesFiles as $extension => $fileName ) {
 877                         if ( isset( $messageDirs[$extension] ) ) {
 878                                 # This extension has JSON message data; skip the PHP shim
 879                                 continue;
 880                         }
 881
 882                         $data = $this->readPHPFile( $fileName, 'extension' );
 883                         $used = false;
 884
 885                         foreach ( $data as $key => $item ) {
 886                                 foreach ( $codeSequence as $csCode ) {
 887                                         if ( isset( $item[$csCode] ) ) {
 888                                                 $this->mergeItem( $key, $extensionData[$csCode][$key], $item[$csCode] );
 889                                                 $used = true;
 890                                         }
 891                                 }
 892                         }
 893
 894                         if ( $used ) {
 895                                 $deps[] = new FileDependency( $fileName );
 896                         }
 897                 }
 898
 899                 # Load the localisation data for each fallback, then merge it into the full array
 900                 $allData = $initialData;
 901                 foreach ( $codeSequence as $csCode ) {
 902                         $csData = $initialData;
 903
 904                         # Load core messages and the extension localisations.
 905                         foreach ( $messageDirs as $dirs ) {
 906                                 foreach ( (array)$dirs as $dir ) {
 907                                         $fileName = "$dir/$csCode.json";
 908                                         $data = $this->readJSONFile( $fileName );
 909
 910                                         foreach ( $data as $key => $item ) {
 911                                                 $this->mergeItem( $key, $csData[$key], $item );
 912                                         }
 913
 914                                         $deps[] = new FileDependency( $fileName );
 915                                 }
 916                         }
 917
 918                         # Merge non-JSON extension data
 919                         if ( isset( $extensionData[$csCode] ) ) {
 920                                 foreach ( $extensionData[$csCode] as $key => $item ) {
 921                                         $this->mergeItem( $key, $csData[$key], $item );
 922                                 }
 923                         }
 924
 925                         if ( $csCode === $code ) {
 926                                 # Merge core data into extension data
 927                                 foreach ( $coreData as $key => $item ) {
 928                                         $this->mergeItem( $key, $csData[$key], $item );
 929                                 }
 930                         } else {
 931                                 # Load the secondary localisation from the source file to
 932                                 # avoid infinite cycles on cyclic fallbacks
 933                                 $fbData = $this->readSourceFilesAndRegisterDeps( $csCode, $deps );
 934                                 if ( $fbData !== false ) {
 935                                         # Only merge the keys that make sense to merge
 936                                         foreach ( self::$allKeys as $key ) {
 937                                                 if ( !isset( $fbData[$key] ) ) {
 938                                                         continue;
 939                                                 }
 940
 941                                                 if ( is_null( $coreData[$key] ) || $this->isMergeableKey( $key ) ) {
 942                                                         $this->mergeItem( $key, $csData[$key], $fbData[$key] );
 943                                                 }
 944                                         }
 945                                 }
 946                         }
 947
 948                         # Allow extensions an opportunity to adjust the data for this
 949                         # fallback
 950                         Hooks::run( 'LocalisationCacheRecacheFallback', [ $this, $csCode, &$csData ] );
 951
 952                         # Merge the data for this fallback into the final array
 953                         if ( $csCode === $code ) {
 954                                 $allData = $csData;
 955                         } else {
 956                                 foreach ( self::$allKeys as $key ) {
 957                                         if ( !isset( $csData[$key] ) ) {
 958                                                 continue;
 959                                         }
 960
 961                                         if ( is_null( $allData[$key] ) || $this->isMergeableKey( $key ) ) {
 962                                                 $this->mergeItem( $key, $allData[$key], $csData[$key] );
 963                                         }
 964                                 }
 965                         }
 966                 }
 967
 968                 # Add cache dependencies for any referenced globals
 969                 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
 970                 // The 'MessagesDirs' config setting is used in LocalisationCache::getMessagesDirs().
 971                 // We use the key 'wgMessagesDirs' for historical reasons.
 972                 $deps['wgMessagesDirs'] = new MainConfigDependency( 'MessagesDirs' );
 973                 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
 974
 975                 # Add dependencies to the cache entry
 976                 $allData['deps'] = $deps;
 977
 978                 # Replace spaces with underscores in namespace names
 979                 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
 980
 981                 # And do the same for special page aliases. $page is an array.
 982                 foreach ( $allData['specialPageAliases'] as &$page ) {
 983                         $page = str_replace( ' ', '_', $page );
 984                 }
 985                 # Decouple the reference to prevent accidental damage
 986                 unset( $page );
 987
 988                 # If there were no plural rules, return an empty array
 989                 if ( $allData['pluralRules'] === null ) {
 990                         $allData['pluralRules'] = [];
 991                 }
 992                 if ( $allData['compiledPluralRules'] === null ) {
 993                         $allData['compiledPluralRules'] = [];
 994                 }
 995                 # If there were no plural rule types, return an empty array
 996                 if ( $allData['pluralRuleTypes'] === null ) {
 997                         $allData['pluralRuleTypes'] = [];
 998                 }
 999
1000                 # Set the list keys
1001                 $allData['list'] = [];
1002                 foreach ( self::$splitKeys as $key ) {
1003                         $allData['list'][$key] = array_keys( $allData[$key] );
1004                 }
1005                 # Run hooks
1006                 $purgeBlobs = true;
1007                 Hooks::run( 'LocalisationCacheRecache', [ $this, $code, &$allData, &$purgeBlobs ] );
1008
1009                 if ( is_null( $allData['namespaceNames'] ) ) {
1010                         throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
1011                                 'Check that your languages/messages/MessagesEn.php file is intact.' );
1012                 }
1013
1014                 # Set the preload key
1015                 $allData['preload'] = $this->buildPreload( $allData );
1016
1017                 # Save to the process cache and register the items loaded
1018                 $this->data[$code] = $allData;
1019                 foreach ( $allData as $key => $item ) {
1020                         $this->loadedItems[$code][$key] = true;
1021                 }
1022
1023                 # Save to the persistent cache
1024                 $this->store->startWrite( $code );
1025                 foreach ( $allData as $key => $value ) {
1026                         if ( in_array( $key, self::$splitKeys ) ) {
1027                                 foreach ( $value as $subkey => $subvalue ) {
1028                                         $this->store->set( "$key:$subkey", $subvalue );
1029                                 }
1030                         } else {
1031                                 $this->store->set( $key, $value );
1032                         }
1033                 }
1034                 $this->store->finishWrite();
1035
1036                 # Clear out the MessageBlobStore
1037                 # HACK: If using a null (i.e. disabled) storage backend, we
1038                 # can't write to the MessageBlobStore either
1039                 if ( $purgeBlobs && !$this->store instanceof LCStoreNull ) {
1040                         $blobStore = new MessageBlobStore();
1041                         $blobStore->clear();
1042                 }
1043         }
1044
1045         /**
1046          * Build the preload item from the given pre-cache data.
1047          *
1048          * The preload item will be loaded automatically, improving performance
1049          * for the commonly-requested items it contains.
1050          * @param array $data
1051          * @return array
1052          */
1053         protected function buildPreload( $data ) {
1054                 $preload = [ 'messages' => [] ];
1055                 foreach ( self::$preloadedKeys as $key ) {
1056                         $preload[$key] = $data[$key];
1057                 }
1058
1059                 foreach ( $data['preloadedMessages'] as $subkey ) {
1060                         if ( isset( $data['messages'][$subkey] ) ) {
1061                                 $subitem = $data['messages'][$subkey];
1062                         } else {
1063                                 $subitem = null;
1064                         }
1065                         $preload['messages'][$subkey] = $subitem;
1066                 }
1067
1068                 return $preload;
1069         }
1070
1071         /**
1072          * Unload the data for a given language from the object cache.
1073          * Reduces memory usage.
1074          * @param string $code
1075          */
1076         public function unload( $code ) {
1077                 unset( $this->data[$code] );
1078                 unset( $this->loadedItems[$code] );
1079                 unset( $this->loadedSubitems[$code] );
1080                 unset( $this->initialisedLangs[$code] );
1081                 unset( $this->shallowFallbacks[$code] );
1082
1083                 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1084                         if ( $fbCode === $code ) {
1085                                 $this->unload( $shallowCode );
1086                         }
1087                 }
1088         }
1089
1090         /**
1091          * Unload all data
1092          */
1093         public function unloadAll() {
1094                 foreach ( $this->initialisedLangs as $lang => $unused ) {
1095                         $this->unload( $lang );
1096                 }
1097         }
1098
1099         /**
1100          * Disable the storage backend
1101          */
1102         public function disableBackend() {
1103                 $this->store = new LCStoreNull;
1104                 $this->manualRecache = false;
1105         }
1106
1107 }