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