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