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