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