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 Cdb\Reader as CdbReader;
  24 use Cdb\Writer as CdbWriter;
  25 use CLDRPluralRuleParser\Evaluator;
  26 use CLDRPluralRuleParser\Error as CLDRPluralRuleError;
  27 use MediaWiki\MediaWikiServices;
  28
  29 /**
  30  * Class for caching the contents of localisation files, Messages*.php
  31  * and *.i18n.php.
  32  *
  33  * An instance of this class is available using Language::getLocalisationCache().
  34  *
  35  * The values retrieved from here are merged, containing items from extension
  36  * files, core messages files and the language fallback sequence (e.g. zh-cn ->
  37  * zh-hans -> en ). Some common errors are corrected, for example namespace
  38  * names with spaces instead of underscores, but heavyweight processing, such
  39  * as grammatical transformation, is done by the caller.
  40  */
  41 class LocalisationCache {
  42         const VERSION = 4;
  43
  44         /** Configuration associative array */
  45         private $conf;
  46
  47         /**
  48          * True if recaching should only be done on an explicit call to recache().
  49          * Setting this reduces the overhead of cache freshness checking, which
  50          * requires doing a stat() for every extension i18n file.
  51          */
  52         private $manualRecache = false;
  53
  54         /**
  55          * True to treat all files as expired until they are regenerated by this object.
  56          */
  57         private $forceRecache = false;
  58
  59         /**
  60          * The cache data. 3-d array, where the first key is the language code,
  61          * the second key is the item key e.g. 'messages', and the third key is
  62          * an item specific subkey index. Some items are not arrays and so for those
  63          * items, there are no subkeys.
  64          */
  65         protected $data = [];
  66
  67         /**
  68          * The persistent store object. An instance of LCStore.
  69          *
  70          * @var LCStore
  71          */
  72         private $store;
  73
  74         /**
  75          * A 2-d associative array, code/key, where presence indicates that the item
  76          * is loaded. Value arbitrary.
  77          *
  78          * For split items, if set, this indicates that all of the subitems have been
  79          * loaded.
  80          */
  81         private $loadedItems = [];
  82
  83         /**
  84          * A 3-d associative array, code/key/subkey, where presence indicates that
  85          * the subitem is loaded. Only used for the split items, i.e. messages.
  86          */
  87         private $loadedSubitems = [];
  88
  89         /**
  90          * An array where presence of a key indicates that that language has been
  91          * initialised. Initialisation includes checking for cache expiry and doing
  92          * any necessary updates.
  93          */
  94         private $initialisedLangs = [];
  95
  96         /**
  97          * An array mapping non-existent pseudo-languages to fallback languages. This
  98          * is filled by initShallowFallback() when data is requested from a language
  99          * that lacks a Messages*.php file.
 100          */
 101         private $shallowFallbacks = [];
 102
 103         /**
 104          * An array where the keys are codes that have been recached by this instance.
 105          */
 106         private $recachedLangs = [];
 107
 108         /**
 109          * All item keys
 110          */
 111         static public $allKeys = [
 112                 'fallback', 'namespaceNames', 'bookstoreList',
 113                 'magicWords', 'messages', 'rtl', 'capitalizeAllNouns', 'digitTransformTable',
 114                 'separatorTransformTable', 'fallback8bitEncoding', 'linkPrefixExtension',
 115                 'linkTrail', 'linkPrefixCharset', 'namespaceAliases',
 116                 'dateFormats', 'datePreferences', 'datePreferenceMigrationMap',
 117                 'defaultDateFormat', 'extraUserToggles', 'specialPageAliases',
 118                 'imageFiles', 'preloadedMessages', 'namespaceGenderAliases',
 119                 'digitGroupingPattern', 'pluralRules', 'pluralRuleTypes', 'compiledPluralRules',
 120         ];
 121
 122         /**
 123          * Keys for items which consist of associative arrays, which may be merged
 124          * by a fallback sequence.
 125          */
 126         static public $mergeableMapKeys = [ 'messages', 'namespaceNames',
 127                 'namespaceAliases', 'dateFormats', 'imageFiles', 'preloadedMessages'
 128         ];
 129
 130         /**
 131          * Keys for items which are a numbered array.
 132          */
 133         static public $mergeableListKeys = [ 'extraUserToggles' ];
 134
 135         /**
 136          * Keys for items which contain an array of arrays of equivalent aliases
 137          * for each subitem. The aliases may be merged by a fallback sequence.
 138          */
 139         static public $mergeableAliasListKeys = [ 'specialPageAliases' ];
 140
 141         /**
 142          * Keys for items which contain an associative array, and may be merged if
 143          * the primary value contains the special array key "inherit". That array
 144          * key is removed after the first merge.
 145          */
 146         static public $optionalMergeKeys = [ 'bookstoreList' ];
 147
 148         /**
 149          * Keys for items that are formatted like $magicWords
 150          */
 151         static public $magicWordKeys = [ 'magicWords' ];
 152
 153         /**
 154          * Keys for items where the subitems are stored in the backend separately.
 155          */
 156         static public $splitKeys = [ 'messages' ];
 157
 158         /**
 159          * Keys which are loaded automatically by initLanguage()
 160          */
 161         static public $preloadedKeys = [ 'dateFormats', 'namespaceNames' ];
 162
 163         /**
 164          * Associative array of cached plural rules. The key is the language code,
 165          * the value is an array of plural rules for that language.
 166          */
 167         private $pluralRules = null;
 168
 169         /**
 170          * Associative array of cached plural rule types. The key is the language
 171          * code, the value is an array of plural rule types for that language. For
 172          * example, $pluralRuleTypes['ar'] = ['zero', 'one', 'two', 'few', 'many'].
 173          * The index for each rule type matches the index for the rule in
 174          * $pluralRules, thus allowing correlation between the two. The reason we
 175          * don't just use the type names as the keys in $pluralRules is because
 176          * Language::convertPlural applies the rules based on numeric order (or
 177          * explicit numeric parameter), not based on the name of the rule type. For
 178          * example, {{plural:count|wordform1|wordform2|wordform3}}, rather than
 179          * {{plural:count|one=wordform1|two=wordform2|many=wordform3}}.
 180          */
 181         private $pluralRuleTypes = null;
 182
 183         private $mergeableKeys = null;
 184
 185         /**
 186          * Constructor.
 187          * For constructor parameters, see the documentation in DefaultSettings.php
 188          * for $wgLocalisationCacheConf.
 189          *
 190          * @param array $conf
 191          * @throws MWException
 192          */
 193         function __construct( $conf ) {
 194                 global $wgCacheDirectory;
 195
 196                 $this->conf = $conf;
 197                 $storeConf = [];
 198                 if ( !empty( $conf['storeClass'] ) ) {
 199                         $storeClass = $conf['storeClass'];
 200                 } else {
 201                         switch ( $conf['store'] ) {
 202                                 case 'files':
 203                                 case 'file':
 204                                         $storeClass = 'LCStoreCDB';
 205                                         break;
 206                                 case 'db':
 207                                         $storeClass = 'LCStoreDB';
 208                                         break;
 209                                 case 'array':
 210                                         $storeClass = 'LCStoreStaticArray';
 211                                         break;
 212                                 case 'detect':
 213                                         if ( !empty( $conf['storeDirectory'] ) ) {
 214                                                 $storeClass = 'LCStoreCDB';
 215                                         } else {
 216                                                 $cacheDir = $wgCacheDirectory ?: wfTempDir();
 217                                                 if ( $cacheDir ) {
 218                                                         $storeConf['directory'] = $cacheDir;
 219                                                         $storeClass = 'LCStoreCDB';
 220                                                 } else {
 221                                                         $storeClass = 'LCStoreDB';
 222                                                 }
 223                                         }
 224                                         break;
 225                                 default:
 226                                         throw new MWException(
 227                                                 'Please set $wgLocalisationCacheConf[\'store\'] to something sensible.' );
 228                         }
 229                 }
 230
 231                 wfDebugLog( 'caches', get_class( $this ) . ": using store $storeClass" );
 232                 if ( !empty( $conf['storeDirectory'] ) ) {
 233                         $storeConf['directory'] = $conf['storeDirectory'];
 234                 }
 235
 236                 $this->store = new $storeClass( $storeConf );
 237                 foreach ( [ 'manualRecache', 'forceRecache' ] as $var ) {
 238                         if ( isset( $conf[$var] ) ) {
 239                                 $this->$var = $conf[$var];
 240                         }
 241                 }
 242         }
 243
 244         /**
 245          * Returns true if the given key is mergeable, that is, if it is an associative
 246          * array which can be merged through a fallback sequence.
 247          * @param string $key
 248          * @return bool
 249          */
 250         public function isMergeableKey( $key ) {
 251                 if ( $this->mergeableKeys === null ) {
 252                         $this->mergeableKeys = array_flip( array_merge(
 253                                 self::$mergeableMapKeys,
 254                                 self::$mergeableListKeys,
 255                                 self::$mergeableAliasListKeys,
 256                                 self::$optionalMergeKeys,
 257                                 self::$magicWordKeys
 258                         ) );
 259                 }
 260
 261                 return isset( $this->mergeableKeys[$key] );
 262         }
 263
 264         /**
 265          * Get a cache item.
 266          *
 267          * Warning: this may be slow for split items (messages), since it will
 268          * need to fetch all of the subitems from the cache individually.
 269          * @param string $code
 270          * @param string $key
 271          * @return mixed
 272          */
 273         public function getItem( $code, $key ) {
 274                 if ( !isset( $this->loadedItems[$code][$key] ) ) {
 275                         $this->loadItem( $code, $key );
 276                 }
 277
 278                 if ( $key === 'fallback' && isset( $this->shallowFallbacks[$code] ) ) {
 279                         return $this->shallowFallbacks[$code];
 280                 }
 281
 282                 return $this->data[$code][$key];
 283         }
 284
 285         /**
 286          * Get a subitem, for instance a single message for a given language.
 287          * @param string $code
 288          * @param string $key
 289          * @param string $subkey
 290          * @return mixed|null
 291          */
 292         public function getSubitem( $code, $key, $subkey ) {
 293                 if ( !isset( $this->loadedSubitems[$code][$key][$subkey] ) &&
 294                         !isset( $this->loadedItems[$code][$key] )
 295                 ) {
 296                         $this->loadSubitem( $code, $key, $subkey );
 297                 }
 298
 299                 if ( isset( $this->data[$code][$key][$subkey] ) ) {
 300                         return $this->data[$code][$key][$subkey];
 301                 } else {
 302                         return null;
 303                 }
 304         }
 305
 306         /**
 307          * Get the list of subitem keys for a given item.
 308          *
 309          * This is faster than array_keys($lc->getItem(...)) for the items listed in
 310          * self::$splitKeys.
 311          *
 312          * Will return null if the item is not found, or false if the item is not an
 313          * array.
 314          * @param string $code
 315          * @param string $key
 316          * @return bool|null|string
 317          */
 318         public function getSubitemList( $code, $key ) {
 319                 if ( in_array( $key, self::$splitKeys ) ) {
 320                         return $this->getSubitem( $code, 'list', $key );
 321                 } else {
 322                         $item = $this->getItem( $code, $key );
 323                         if ( is_array( $item ) ) {
 324                                 return array_keys( $item );
 325                         } else {
 326                                 return false;
 327                         }
 328                 }
 329         }
 330
 331         /**
 332          * Load an item into the cache.
 333          * @param string $code
 334          * @param string $key
 335          */
 336         protected function loadItem( $code, $key ) {
 337                 if ( !isset( $this->initialisedLangs[$code] ) ) {
 338                         $this->initLanguage( $code );
 339                 }
 340
 341                 // Check to see if initLanguage() loaded it for us
 342                 if ( isset( $this->loadedItems[$code][$key] ) ) {
 343                         return;
 344                 }
 345
 346                 if ( isset( $this->shallowFallbacks[$code] ) ) {
 347                         $this->loadItem( $this->shallowFallbacks[$code], $key );
 348
 349                         return;
 350                 }
 351
 352                 if ( in_array( $key, self::$splitKeys ) ) {
 353                         $subkeyList = $this->getSubitem( $code, 'list', $key );
 354                         foreach ( $subkeyList as $subkey ) {
 355                                 if ( isset( $this->data[$code][$key][$subkey] ) ) {
 356                                         continue;
 357                                 }
 358                                 $this->data[$code][$key][$subkey] = $this->getSubitem( $code, $key, $subkey );
 359                         }
 360                 } else {
 361                         $this->data[$code][$key] = $this->store->get( $code, $key );
 362                 }
 363
 364                 $this->loadedItems[$code][$key] = true;
 365         }
 366
 367         /**
 368          * Load a subitem into the cache
 369          * @param string $code
 370          * @param string $key
 371          * @param string $subkey
 372          */
 373         protected function loadSubitem( $code, $key, $subkey ) {
 374                 if ( !in_array( $key, self::$splitKeys ) ) {
 375                         $this->loadItem( $code, $key );
 376
 377                         return;
 378                 }
 379
 380                 if ( !isset( $this->initialisedLangs[$code] ) ) {
 381                         $this->initLanguage( $code );
 382                 }
 383
 384                 // Check to see if initLanguage() loaded it for us
 385                 if ( isset( $this->loadedItems[$code][$key] ) ||
 386                         isset( $this->loadedSubitems[$code][$key][$subkey] )
 387                 ) {
 388                         return;
 389                 }
 390
 391                 if ( isset( $this->shallowFallbacks[$code] ) ) {
 392                         $this->loadSubitem( $this->shallowFallbacks[$code], $key, $subkey );
 393
 394                         return;
 395                 }
 396
 397                 $value = $this->store->get( $code, "$key:$subkey" );
 398                 $this->data[$code][$key][$subkey] = $value;
 399                 $this->loadedSubitems[$code][$key][$subkey] = true;
 400         }
 401
 402         /**
 403          * Returns true if the cache identified by $code is missing or expired.
 404          *
 405          * @param string $code
 406          *
 407          * @return bool
 408          */
 409         public function isExpired( $code ) {
 410                 if ( $this->forceRecache && !isset( $this->recachedLangs[$code] ) ) {
 411                         wfDebug( __METHOD__ . "($code): forced reload\n" );
 412
 413                         return true;
 414                 }
 415
 416                 $deps = $this->store->get( $code, 'deps' );
 417                 $keys = $this->store->get( $code, 'list' );
 418                 $preload = $this->store->get( $code, 'preload' );
 419                 // Different keys may expire separately for some stores
 420                 if ( $deps === null || $keys === null || $preload === null ) {
 421                         wfDebug( __METHOD__ . "($code): cache missing, need to make one\n" );
 422
 423                         return true;
 424                 }
 425
 426                 foreach ( $deps as $dep ) {
 427                         // Because we're unserializing stuff from cache, we
 428                         // could receive objects of classes that don't exist
 429                         // anymore (e.g. uninstalled extensions)
 430                         // When this happens, always expire the cache
 431                         if ( !$dep instanceof CacheDependency || $dep->isExpired() ) {
 432                                 wfDebug( __METHOD__ . "($code): cache for $code expired due to " .
 433                                         get_class( $dep ) . "\n" );
 434
 435                                 return true;
 436                         }
 437                 }
 438
 439                 return false;
 440         }
 441
 442         /**
 443          * Initialise a language in this object. Rebuild the cache if necessary.
 444          * @param string $code
 445          * @throws MWException
 446          */
 447         protected function initLanguage( $code ) {
 448                 if ( isset( $this->initialisedLangs[$code] ) ) {
 449                         return;
 450                 }
 451
 452                 $this->initialisedLangs[$code] = true;
 453
 454                 # If the code is of the wrong form for a Messages*.php file, do a shallow fallback
 455                 if ( !Language::isValidBuiltInCode( $code ) ) {
 456                         $this->initShallowFallback( $code, 'en' );
 457
 458                         return;
 459                 }
 460
 461                 # Recache the data if necessary
 462                 if ( !$this->manualRecache && $this->isExpired( $code ) ) {
 463                         if ( Language::isSupportedLanguage( $code ) ) {
 464                                 $this->recache( $code );
 465                         } elseif ( $code === 'en' ) {
 466                                 throw new MWException( 'MessagesEn.php is missing.' );
 467                         } else {
 468                                 $this->initShallowFallback( $code, 'en' );
 469                         }
 470
 471                         return;
 472                 }
 473
 474                 # Preload some stuff
 475                 $preload = $this->getItem( $code, 'preload' );
 476                 if ( $preload === null ) {
 477                         if ( $this->manualRecache ) {
 478                                 // No Messages*.php file. Do shallow fallback to en.
 479                                 if ( $code === 'en' ) {
 480                                         throw new MWException( 'No localisation cache found for English. ' .
 481                                                 'Please run maintenance/rebuildLocalisationCache.php.' );
 482                                 }
 483                                 $this->initShallowFallback( $code, 'en' );
 484
 485                                 return;
 486                         } else {
 487                                 throw new MWException( 'Invalid or missing localisation cache.' );
 488                         }
 489                 }
 490                 $this->data[$code] = $preload;
 491                 foreach ( $preload as $key => $item ) {
 492                         if ( in_array( $key, self::$splitKeys ) ) {
 493                                 foreach ( $item as $subkey => $subitem ) {
 494                                         $this->loadedSubitems[$code][$key][$subkey] = true;
 495                                 }
 496                         } else {
 497                                 $this->loadedItems[$code][$key] = true;
 498                         }
 499                 }
 500         }
 501
 502         /**
 503          * Create a fallback from one language to another, without creating a
 504          * complete persistent cache.
 505          * @param string $primaryCode
 506          * @param string $fallbackCode
 507          */
 508         public function initShallowFallback( $primaryCode, $fallbackCode ) {
 509                 $this->data[$primaryCode] =& $this->data[$fallbackCode];
 510                 $this->loadedItems[$primaryCode] =& $this->loadedItems[$fallbackCode];
 511                 $this->loadedSubitems[$primaryCode] =& $this->loadedSubitems[$fallbackCode];
 512                 $this->shallowFallbacks[$primaryCode] = $fallbackCode;
 513         }
 514
 515         /**
 516          * Read a PHP file containing localisation data.
 517          * @param string $_fileName
 518          * @param string $_fileType
 519          * @throws MWException
 520          * @return array
 521          */
 522         protected function readPHPFile( $_fileName, $_fileType ) {
 523                 // Disable APC caching
 524                 MediaWiki\suppressWarnings();
 525                 $_apcEnabled = ini_set( 'apc.cache_by_default', '0' );
 526                 MediaWiki\restoreWarnings();
 527
 528                 include $_fileName;
 529
 530                 MediaWiki\suppressWarnings();
 531                 ini_set( 'apc.cache_by_default', $_apcEnabled );
 532                 MediaWiki\restoreWarnings();
 533
 534                 if ( $_fileType == 'core' || $_fileType == 'extension' ) {
 535                         $data = compact( self::$allKeys );
 536                 } elseif ( $_fileType == 'aliases' ) {
 537                         $data = compact( 'aliases' );
 538                 } else {
 539                         throw new MWException( __METHOD__ . ": Invalid file type: $_fileType" );
 540                 }
 541
 542                 return $data;
 543         }
 544
 545         /**
 546          * Read a JSON file containing localisation messages.
 547          * @param string $fileName Name of file to read
 548          * @throws MWException If there is a syntax error in the JSON file
 549          * @return array Array with a 'messages' key, or empty array if the file doesn't exist
 550          */
 551         public function readJSONFile( $fileName ) {
 552
 553                 if ( !is_readable( $fileName ) ) {
 554                         return [];
 555                 }
 556
 557                 $json = file_get_contents( $fileName );
 558                 if ( $json === false ) {
 559                         return [];
 560                 }
 561
 562                 $data = FormatJson::decode( $json, true );
 563                 if ( $data === null ) {
 564
 565                         throw new MWException( __METHOD__ . ": Invalid JSON file: $fileName" );
 566                 }
 567
 568                 // Remove keys starting with '@', they're reserved for metadata and non-message data
 569                 foreach ( $data as $key => $unused ) {
 570                         if ( $key === '' || $key[0] === '@' ) {
 571                                 unset( $data[$key] );
 572                         }
 573                 }
 574
 575                 // The JSON format only supports messages, none of the other variables, so wrap the data
 576                 return [ 'messages' => $data ];
 577         }
 578
 579         /**
 580          * Get the compiled plural rules for a given language from the XML files.
 581          * @since 1.20
 582          * @param string $code
 583          * @return array|null
 584          */
 585         public function getCompiledPluralRules( $code ) {
 586                 $rules = $this->getPluralRules( $code );
 587                 if ( $rules === null ) {
 588                         return null;
 589                 }
 590                 try {
 591                         $compiledRules = Evaluator::compile( $rules );
 592                 } catch ( CLDRPluralRuleError $e ) {
 593                         wfDebugLog( 'l10n', $e->getMessage() );
 594
 595                         return [];
 596                 }
 597
 598                 return $compiledRules;
 599         }
 600
 601         /**
 602          * Get the plural rules for a given language from the XML files.
 603          * Cached.
 604          * @since 1.20
 605          * @param string $code
 606          * @return array|null
 607          */
 608         public function getPluralRules( $code ) {
 609                 if ( $this->pluralRules === null ) {
 610                         $this->loadPluralFiles();
 611                 }
 612                 if ( !isset( $this->pluralRules[$code] ) ) {
 613                         return null;
 614                 } else {
 615                         return $this->pluralRules[$code];
 616                 }
 617         }
 618
 619         /**
 620          * Get the plural rule types for a given language from the XML files.
 621          * Cached.
 622          * @since 1.22
 623          * @param string $code
 624          * @return array|null
 625          */
 626         public function getPluralRuleTypes( $code ) {
 627                 if ( $this->pluralRuleTypes === null ) {
 628                         $this->loadPluralFiles();
 629                 }
 630                 if ( !isset( $this->pluralRuleTypes[$code] ) ) {
 631                         return null;
 632                 } else {
 633                         return $this->pluralRuleTypes[$code];
 634                 }
 635         }
 636
 637         /**
 638          * Load the plural XML files.
 639          */
 640         protected function loadPluralFiles() {
 641                 global $IP;
 642                 $cldrPlural = "$IP/languages/data/plurals.xml";
 643                 $mwPlural = "$IP/languages/data/plurals-mediawiki.xml";
 644                 // Load CLDR plural rules
 645                 $this->loadPluralFile( $cldrPlural );
 646                 if ( file_exists( $mwPlural ) ) {
 647                         // Override or extend
 648                         $this->loadPluralFile( $mwPlural );
 649                 }
 650         }
 651
 652         /**
 653          * Load a plural XML file with the given filename, compile the relevant
 654          * rules, and save the compiled rules in a process-local cache.
 655          *
 656          * @param string $fileName
 657          * @throws MWException
 658          */
 659         protected function loadPluralFile( $fileName ) {
 660                 // Use file_get_contents instead of DOMDocument::load (T58439)
 661                 $xml = file_get_contents( $fileName );
 662                 if ( !$xml ) {
 663                         throw new MWException( "Unable to read plurals file $fileName" );
 664                 }
 665                 $doc = new DOMDocument;
 666                 $doc->loadXML( $xml );
 667                 $rulesets = $doc->getElementsByTagName( "pluralRules" );
 668                 foreach ( $rulesets as $ruleset ) {
 669                         $codes = $ruleset->getAttribute( 'locales' );
 670                         $rules = [];
 671                         $ruleTypes = [];
 672                         $ruleElements = $ruleset->getElementsByTagName( "pluralRule" );
 673                         foreach ( $ruleElements as $elt ) {
 674                                 $ruleType = $elt->getAttribute( 'count' );
 675                                 if ( $ruleType === 'other' ) {
 676                                         // Don't record "other" rules, which have an empty condition
 677                                         continue;
 678                                 }
 679                                 $rules[] = $elt->nodeValue;
 680                                 $ruleTypes[] = $ruleType;
 681                         }
 682                         foreach ( explode( ' ', $codes ) as $code ) {
 683                                 $this->pluralRules[$code] = $rules;
 684                                 $this->pluralRuleTypes[$code] = $ruleTypes;
 685                         }
 686                 }
 687         }
 688
 689         /**
 690          * Read the data from the source files for a given language, and register
 691          * the relevant dependencies in the $deps array. If the localisation
 692          * exists, the data array is returned, otherwise false is returned.
 693          *
 694          * @param string $code
 695          * @param array $deps
 696          * @return array
 697          */
 698         protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
 699                 global $IP;
 700
 701                 // This reads in the PHP i18n file with non-messages l10n data
 702                 $fileName = Language::getMessagesFileName( $code );
 703                 if ( !file_exists( $fileName ) ) {
 704                         $data = [];
 705                 } else {
 706                         $deps[] = new FileDependency( $fileName );
 707                         $data = $this->readPHPFile( $fileName, 'core' );
 708                 }
 709
 710                 # Load CLDR plural rules for JavaScript
 711                 $data['pluralRules'] = $this->getPluralRules( $code );
 712                 # And for PHP
 713                 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
 714                 # Load plural rule types
 715                 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
 716
 717                 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
 718                 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
 719
 720                 return $data;
 721         }
 722
 723         /**
 724          * Merge two localisation values, a primary and a fallback, overwriting the
 725          * primary value in place.
 726          * @param string $key
 727          * @param mixed $value
 728          * @param mixed $fallbackValue
 729          */
 730         protected function mergeItem( $key, &$value, $fallbackValue ) {
 731                 if ( !is_null( $value ) ) {
 732                         if ( !is_null( $fallbackValue ) ) {
 733                                 if ( in_array( $key, self::$mergeableMapKeys ) ) {
 734                                         $value = $value + $fallbackValue;
 735                                 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
 736                                         $value = array_unique( array_merge( $fallbackValue, $value ) );
 737                                 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
 738                                         $value = array_merge_recursive( $value, $fallbackValue );
 739                                 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
 740                                         if ( !empty( $value['inherit'] ) ) {
 741                                                 $value = array_merge( $fallbackValue, $value );
 742                                         }
 743
 744                                         if ( isset( $value['inherit'] ) ) {
 745                                                 unset( $value['inherit'] );
 746                                         }
 747                                 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
 748                                         $this->mergeMagicWords( $value, $fallbackValue );
 749                                 }
 750                         }
 751                 } else {
 752                         $value = $fallbackValue;
 753                 }
 754         }
 755
 756         /**
 757          * @param mixed $value
 758          * @param mixed $fallbackValue
 759          */
 760         protected function mergeMagicWords( &$value, $fallbackValue ) {
 761                 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
 762                         if ( !isset( $value[$magicName] ) ) {
 763                                 $value[$magicName] = $fallbackInfo;
 764                         } else {
 765                                 $oldSynonyms = array_slice( $fallbackInfo, 1 );
 766                                 $newSynonyms = array_slice( $value[$magicName], 1 );
 767                                 $synonyms = array_values( array_unique( array_merge(
 768                                         $newSynonyms, $oldSynonyms ) ) );
 769                                 $value[$magicName] = array_merge( [ $fallbackInfo[0] ], $synonyms );
 770                         }
 771                 }
 772         }
 773
 774         /**
 775          * Given an array mapping language code to localisation value, such as is
 776          * found in extension *.i18n.php files, iterate through a fallback sequence
 777          * to merge the given data with an existing primary value.
 778          *
 779          * Returns true if any data from the extension array was used, false
 780          * otherwise.
 781          * @param array $codeSequence
 782          * @param string $key
 783          * @param mixed $value
 784          * @param mixed $fallbackValue
 785          * @return bool
 786          */
 787         protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
 788                 $used = false;
 789                 foreach ( $codeSequence as $code ) {
 790                         if ( isset( $fallbackValue[$code] ) ) {
 791                                 $this->mergeItem( $key, $value, $fallbackValue[$code] );
 792                                 $used = true;
 793                         }
 794                 }
 795
 796                 return $used;
 797         }
 798
 799         /**
 800          * Gets the combined list of messages dirs from
 801          * core and extensions
 802          *
 803          * @since 1.25
 804          * @return array
 805          */
 806         public function getMessagesDirs() {
 807                 global $IP;
 808
 809                 $config = MediaWikiServices::getInstance()->getMainConfig();
 810                 $messagesDirs = $config->get( 'MessagesDirs' );
 811                 return [
 812                         'core' => "$IP/languages/i18n",
 813                         'api' => "$IP/includes/api/i18n",
 814                         'oojs-ui' => "$IP/resources/lib/oojs-ui/i18n",
 815                 ] + $messagesDirs;
 816         }
 817
 818         /**
 819          * Load localisation data for a given language for both core and extensions
 820          * and save it to the persistent cache store and the process cache
 821          * @param string $code
 822          * @throws MWException
 823          */
 824         public function recache( $code ) {
 825                 global $wgExtensionMessagesFiles;
 826
 827                 if ( !$code ) {
 828                         throw new MWException( "Invalid language code requested" );
 829                 }
 830                 $this->recachedLangs[$code] = true;
 831
 832                 # Initial values
 833                 $initialData = array_fill_keys( self::$allKeys, null );
 834                 $coreData = $initialData;
 835                 $deps = [];
 836
 837                 # Load the primary localisation from the source file
 838                 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
 839                 if ( $data === false ) {
 840                         wfDebug( __METHOD__ . ": no localisation file for $code, using fallback to en\n" );
 841                         $coreData['fallback'] = 'en';
 842                 } else {
 843                         wfDebug( __METHOD__ . ": got localisation for $code from source\n" );
 844
 845                         # Merge primary localisation
 846                         foreach ( $data as $key => $value ) {
 847                                 $this->mergeItem( $key, $coreData[$key], $value );
 848                         }
 849                 }
 850
 851                 # Fill in the fallback if it's not there already
 852                 if ( is_null( $coreData['fallback'] ) ) {
 853                         $coreData['fallback'] = $code === 'en' ? false : 'en';
 854                 }
 855                 if ( $coreData['fallback'] === false ) {
 856                         $coreData['fallbackSequence'] = [];
 857                 } else {
 858                         $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
 859                         $len = count( $coreData['fallbackSequence'] );
 860
 861                         # Ensure that the sequence ends at en
 862                         if ( $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
 863                                 $coreData['fallbackSequence'][] = 'en';
 864                         }
 865                 }
 866
 867                 $codeSequence = array_merge( [ $code ], $coreData['fallbackSequence'] );
 868                 $messageDirs = $this->getMessagesDirs();
 869
 870                 # Load non-JSON localisation data for extensions
 871                 $extensionData = array_fill_keys( $codeSequence, $initialData );
 872                 foreach ( $wgExtensionMessagesFiles as $extension => $fileName ) {
 873                         if ( isset( $messageDirs[$extension] ) ) {
 874                                 # This extension has JSON message data; skip the PHP shim
 875                                 continue;
 876                         }
 877
 878                         $data = $this->readPHPFile( $fileName, 'extension' );
 879                         $used = false;
 880
 881                         foreach ( $data as $key => $item ) {
 882                                 foreach ( $codeSequence as $csCode ) {
 883                                         if ( isset( $item[$csCode] ) ) {
 884                                                 $this->mergeItem( $key, $extensionData[$csCode][$key], $item[$csCode] );
 885                                                 $used = true;
 886                                         }
 887                                 }
 888                         }
 889
 890                         if ( $used ) {
 891                                 $deps[] = new FileDependency( $fileName );
 892                         }
 893                 }
 894
 895                 # Load the localisation data for each fallback, then merge it into the full array
 896                 $allData = $initialData;
 897                 foreach ( $codeSequence as $csCode ) {
 898                         $csData = $initialData;
 899
 900                         # Load core messages and the extension localisations.
 901                         foreach ( $messageDirs as $dirs ) {
 902                                 foreach ( (array)$dirs as $dir ) {
 903                                         $fileName = "$dir/$csCode.json";
 904                                         $data = $this->readJSONFile( $fileName );
 905
 906                                         foreach ( $data as $key => $item ) {
 907                                                 $this->mergeItem( $key, $csData[$key], $item );
 908                                         }
 909
 910                                         $deps[] = new FileDependency( $fileName );
 911                                 }
 912                         }
 913
 914                         # Merge non-JSON extension data
 915                         if ( isset( $extensionData[$csCode] ) ) {
 916                                 foreach ( $extensionData[$csCode] as $key => $item ) {
 917                                         $this->mergeItem( $key, $csData[$key], $item );
 918                                 }
 919                         }
 920
 921                         if ( $csCode === $code ) {
 922                                 # Merge core data into extension data
 923                                 foreach ( $coreData as $key => $item ) {
 924                                         $this->mergeItem( $key, $csData[$key], $item );
 925                                 }
 926                         } else {
 927                                 # Load the secondary localisation from the source file to
 928                                 # avoid infinite cycles on cyclic fallbacks
 929                                 $fbData = $this->readSourceFilesAndRegisterDeps( $csCode, $deps );
 930                                 if ( $fbData !== false ) {
 931                                         # Only merge the keys that make sense to merge
 932                                         foreach ( self::$allKeys as $key ) {
 933                                                 if ( !isset( $fbData[$key] ) ) {
 934                                                         continue;
 935                                                 }
 936
 937                                                 if ( is_null( $coreData[$key] ) || $this->isMergeableKey( $key ) ) {
 938                                                         $this->mergeItem( $key, $csData[$key], $fbData[$key] );
 939                                                 }
 940                                         }
 941                                 }
 942                         }
 943
 944                         # Allow extensions an opportunity to adjust the data for this
 945                         # fallback
 946                         Hooks::run( 'LocalisationCacheRecacheFallback', [ $this, $csCode, &$csData ] );
 947
 948                         # Merge the data for this fallback into the final array
 949                         if ( $csCode === $code ) {
 950                                 $allData = $csData;
 951                         } else {
 952                                 foreach ( self::$allKeys as $key ) {
 953                                         if ( !isset( $csData[$key] ) ) {
 954                                                 continue;
 955                                         }
 956
 957                                         if ( is_null( $allData[$key] ) || $this->isMergeableKey( $key ) ) {
 958                                                 $this->mergeItem( $key, $allData[$key], $csData[$key] );
 959                                         }
 960                                 }
 961                         }
 962                 }
 963
 964                 # Add cache dependencies for any referenced globals
 965                 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
 966                 // The 'MessagesDirs' config setting is used in LocalisationCache::getMessagesDirs().
 967                 // We use the key 'wgMessagesDirs' for historical reasons.
 968                 $deps['wgMessagesDirs'] = new MainConfigDependency( 'MessagesDirs' );
 969                 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
 970
 971                 # Add dependencies to the cache entry
 972                 $allData['deps'] = $deps;
 973
 974                 # Replace spaces with underscores in namespace names
 975                 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
 976
 977                 # And do the same for special page aliases. $page is an array.
 978                 foreach ( $allData['specialPageAliases'] as &$page ) {
 979                         $page = str_replace( ' ', '_', $page );
 980                 }
 981                 # Decouple the reference to prevent accidental damage
 982                 unset( $page );
 983
 984                 # If there were no plural rules, return an empty array
 985                 if ( $allData['pluralRules'] === null ) {
 986                         $allData['pluralRules'] = [];
 987                 }
 988                 if ( $allData['compiledPluralRules'] === null ) {
 989                         $allData['compiledPluralRules'] = [];
 990                 }
 991                 # If there were no plural rule types, return an empty array
 992                 if ( $allData['pluralRuleTypes'] === null ) {
 993                         $allData['pluralRuleTypes'] = [];
 994                 }
 995
 996                 # Set the list keys
 997                 $allData['list'] = [];
 998                 foreach ( self::$splitKeys as $key ) {
 999                         $allData['list'][$key] = array_keys( $allData[$key] );
1000                 }
1001                 # Run hooks
1002                 $purgeBlobs = true;
1003                 Hooks::run( 'LocalisationCacheRecache', [ $this, $code, &$allData, &$purgeBlobs ] );
1004
1005                 if ( is_null( $allData['namespaceNames'] ) ) {
1006                         throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
1007                                 'Check that your languages/messages/MessagesEn.php file is intact.' );
1008                 }
1009
1010                 # Set the preload key
1011                 $allData['preload'] = $this->buildPreload( $allData );
1012
1013                 # Save to the process cache and register the items loaded
1014                 $this->data[$code] = $allData;
1015                 foreach ( $allData as $key => $item ) {
1016                         $this->loadedItems[$code][$key] = true;
1017                 }
1018
1019                 # Save to the persistent cache
1020                 $this->store->startWrite( $code );
1021                 foreach ( $allData as $key => $value ) {
1022                         if ( in_array( $key, self::$splitKeys ) ) {
1023                                 foreach ( $value as $subkey => $subvalue ) {
1024                                         $this->store->set( "$key:$subkey", $subvalue );
1025                                 }
1026                         } else {
1027                                 $this->store->set( $key, $value );
1028                         }
1029                 }
1030                 $this->store->finishWrite();
1031
1032                 # Clear out the MessageBlobStore
1033                 # HACK: If using a null (i.e. disabled) storage backend, we
1034                 # can't write to the MessageBlobStore either
1035                 if ( $purgeBlobs && !$this->store instanceof LCStoreNull ) {
1036                         $blobStore = new MessageBlobStore();
1037                         $blobStore->clear();
1038                 }
1039
1040         }
1041
1042         /**
1043          * Build the preload item from the given pre-cache data.
1044          *
1045          * The preload item will be loaded automatically, improving performance
1046          * for the commonly-requested items it contains.
1047          * @param array $data
1048          * @return array
1049          */
1050         protected function buildPreload( $data ) {
1051                 $preload = [ 'messages' => [] ];
1052                 foreach ( self::$preloadedKeys as $key ) {
1053                         $preload[$key] = $data[$key];
1054                 }
1055
1056                 foreach ( $data['preloadedMessages'] as $subkey ) {
1057                         if ( isset( $data['messages'][$subkey] ) ) {
1058                                 $subitem = $data['messages'][$subkey];
1059                         } else {
1060                                 $subitem = null;
1061                         }
1062                         $preload['messages'][$subkey] = $subitem;
1063                 }
1064
1065                 return $preload;
1066         }
1067
1068         /**
1069          * Unload the data for a given language from the object cache.
1070          * Reduces memory usage.
1071          * @param string $code
1072          */
1073         public function unload( $code ) {
1074                 unset( $this->data[$code] );
1075                 unset( $this->loadedItems[$code] );
1076                 unset( $this->loadedSubitems[$code] );
1077                 unset( $this->initialisedLangs[$code] );
1078                 unset( $this->shallowFallbacks[$code] );
1079
1080                 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1081                         if ( $fbCode === $code ) {
1082                                 $this->unload( $shallowCode );
1083                         }
1084                 }
1085         }
1086
1087         /**
1088          * Unload all data
1089          */
1090         public function unloadAll() {
1091                 foreach ( $this->initialisedLangs as $lang => $unused ) {
1092                         $this->unload( $lang );
1093                 }
1094         }
1095
1096         /**
1097          * Disable the storage backend
1098          */
1099         public function disableBackend() {
1100                 $this->store = new LCStoreNull;
1101                 $this->manualRecache = false;
1102         }
1103
1104 }