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
  27 /**
  28  * Class for caching the contents of localisation files, Messages*.php
  29  * and *.i18n.php.
  30  *
  31  * An instance of this class is available using Language::getLocalisationCache().
  32  *
  33  * The values retrieved from here are merged, containing items from extension
  34  * files, core messages files and the language fallback sequence (e.g. zh-cn ->
  35  * zh-hans -> en ). Some common errors are corrected, for example namespace
  36  * names with spaces instead of underscores, but heavyweight processing, such
  37  * as grammatical transformation, is done by the caller.
  38  */
  39 class LocalisationCache {
  40         const VERSION = 4;
  41
  42         /** Configuration associative array */
  43         private $conf;
  44
  45         /**
  46          * True if recaching should only be done on an explicit call to recache().
  47          * Setting this reduces the overhead of cache freshness checking, which
  48          * requires doing a stat() for every extension i18n file.
  49          */
  50         private $manualRecache = false;
  51
  52         /**
  53          * True to treat all files as expired until they are regenerated by this object.
  54          */
  55         private $forceRecache = false;
  56
  57         /**
  58          * The cache data. 3-d array, where the first key is the language code,
  59          * the second key is the item key e.g. 'messages', and the third key is
  60          * an item specific subkey index. Some items are not arrays and so for those
  61          * items, there are no subkeys.
  62          */
  63         protected $data = [];
  64
  65         /**
  66          * The persistent store object. An instance of LCStore.
  67          *
  68          * @var LCStore
  69          */
  70         private $store;
  71
  72         /**
  73          * A 2-d associative array, code/key, where presence indicates that the item
  74          * is loaded. Value arbitrary.
  75          *
  76          * For split items, if set, this indicates that all of the subitems have been
  77          * loaded.
  78          */
  79         private $loadedItems = [];
  80
  81         /**
  82          * A 3-d associative array, code/key/subkey, where presence indicates that
  83          * the subitem is loaded. Only used for the split items, i.e. messages.
  84          */
  85         private $loadedSubitems = [];
  86
  87         /**
  88          * An array where presence of a key indicates that that language has been
  89          * initialised. Initialisation includes checking for cache expiry and doing
  90          * any necessary updates.
  91          */
  92         private $initialisedLangs = [];
  93
  94         /**
  95          * An array mapping non-existent pseudo-languages to fallback languages. This
  96          * is filled by initShallowFallback() when data is requested from a language
  97          * that lacks a Messages*.php file.
  98          */
  99         private $shallowFallbacks = [];
 100
 101         /**
 102          * An array where the keys are codes that have been recached by this instance.
 103          */
 104         private $recachedLangs = [];
 105
 106         /**
 107          * All item keys
 108          */
 109         static public $allKeys = [
 110                 'fallback', 'namespaceNames', 'bookstoreList',
 111                 'magicWords', 'messages', 'rtl', 'capitalizeAllNouns', 'digitTransformTable',
 112                 'separatorTransformTable', 'fallback8bitEncoding', 'linkPrefixExtension',
 113                 'linkTrail', 'linkPrefixCharset', 'namespaceAliases',
 114                 'dateFormats', 'datePreferences', 'datePreferenceMigrationMap',
 115                 'defaultDateFormat', 'extraUserToggles', 'specialPageAliases',
 116                 'imageFiles', 'preloadedMessages', 'namespaceGenderAliases',
 117                 'digitGroupingPattern', 'pluralRules', 'pluralRuleTypes', 'compiledPluralRules',
 118         ];
 119
 120         /**
 121          * Keys for items which consist of associative arrays, which may be merged
 122          * by a fallback sequence.
 123          */
 124         static public $mergeableMapKeys = [ 'messages', 'namespaceNames',
 125                 'namespaceAliases', 'dateFormats', 'imageFiles', 'preloadedMessages'
 126         ];
 127
 128         /**
 129          * Keys for items which are a numbered array.
 130          */
 131         static public $mergeableListKeys = [ 'extraUserToggles' ];
 132
 133         /**
 134          * Keys for items which contain an array of arrays of equivalent aliases
 135          * for each subitem. The aliases may be merged by a fallback sequence.
 136          */
 137         static public $mergeableAliasListKeys = [ 'specialPageAliases' ];
 138
 139         /**
 140          * Keys for items which contain an associative array, and may be merged if
 141          * the primary value contains the special array key "inherit". That array
 142          * key is removed after the first merge.
 143          */
 144         static public $optionalMergeKeys = [ 'bookstoreList' ];
 145
 146         /**
 147          * Keys for items that are formatted like $magicWords
 148          */
 149         static public $magicWordKeys = [ 'magicWords' ];
 150
 151         /**
 152          * Keys for items where the subitems are stored in the backend separately.
 153          */
 154         static public $splitKeys = [ 'messages' ];
 155
 156         /**
 157          * Keys which are loaded automatically by initLanguage()
 158          */
 159         static public $preloadedKeys = [ 'dateFormats', 'namespaceNames' ];
 160
 161         /**
 162          * Associative array of cached plural rules. The key is the language code,
 163          * the value is an array of plural rules for that language.
 164          */
 165         private $pluralRules = null;
 166
 167         /**
 168          * Associative array of cached plural rule types. The key is the language
 169          * code, the value is an array of plural rule types for that language. For
 170          * example, $pluralRuleTypes['ar'] = ['zero', 'one', 'two', 'few', 'many'].
 171          * The index for each rule type matches the index for the rule in
 172          * $pluralRules, thus allowing correlation between the two. The reason we
 173          * don't just use the type names as the keys in $pluralRules is because
 174          * Language::convertPlural applies the rules based on numeric order (or
 175          * explicit numeric parameter), not based on the name of the rule type. For
 176          * example, {{plural:count|wordform1|wordform2|wordform3}}, rather than
 177          * {{plural:count|one=wordform1|two=wordform2|many=wordform3}}.
 178          */
 179         private $pluralRuleTypes = null;
 180
 181         private $mergeableKeys = null;
 182
 183         /**
 184          * Constructor.
 185          * For constructor parameters, see the documentation in DefaultSettings.php
 186          * for $wgLocalisationCacheConf.
 187          *
 188          * @param array $conf
 189          * @throws MWException
 190          */
 191         function __construct( $conf ) {
 192                 global $wgCacheDirectory;
 193
 194                 $this->conf = $conf;
 195                 $storeConf = [];
 196                 if ( !empty( $conf['storeClass'] ) ) {
 197                         $storeClass = $conf['storeClass'];
 198                 } else {
 199                         switch ( $conf['store'] ) {
 200                                 case 'files':
 201                                 case 'file':
 202                                         $storeClass = 'LCStoreCDB';
 203                                         break;
 204                                 case 'db':
 205                                         $storeClass = 'LCStoreDB';
 206                                         break;
 207                                 case 'array':
 208                                         $storeClass = 'LCStoreStaticArray';
 209                                         break;
 210                                 case 'detect':
 211                                         if ( !empty( $conf['storeDirectory'] ) ) {
 212                                                 $storeClass = 'LCStoreCDB';
 213                                         } else {
 214                                                 $cacheDir = $wgCacheDirectory ?: wfTempDir();
 215                                                 if ( $cacheDir ) {
 216                                                         $storeConf['directory'] = $cacheDir;
 217                                                         $storeClass = 'LCStoreCDB';
 218                                                 } else {
 219                                                         $storeClass = 'LCStoreDB';
 220                                                 }
 221                                         }
 222                                         break;
 223                                 default:
 224                                         throw new MWException(
 225                                                 'Please set $wgLocalisationCacheConf[\'store\'] to something sensible.' );
 226                         }
 227                 }
 228
 229                 wfDebugLog( 'caches', get_class( $this ) . ": 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
 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
 551                 if ( !is_readable( $fileName ) ) {
 552                         return [];
 553                 }
 554
 555                 $json = file_get_contents( $fileName );
 556                 if ( $json === false ) {
 557                         return [];
 558                 }
 559
 560                 $data = FormatJson::decode( $json, true );
 561                 if ( $data === null ) {
 562
 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 $wgMessagesDirs, $IP;
 806                 return [
 807                         'core' => "$IP/languages/i18n",
 808                         'api' => "$IP/includes/api/i18n",
 809                         'oojs-ui' => "$IP/resources/lib/oojs-ui/i18n",
 810                 ] + $wgMessagesDirs;
 811         }
 812
 813         /**
 814          * Load localisation data for a given language for both core and extensions
 815          * and save it to the persistent cache store and the process cache
 816          * @param string $code
 817          * @throws MWException
 818          */
 819         public function recache( $code ) {
 820                 global $wgExtensionMessagesFiles;
 821
 822                 if ( !$code ) {
 823                         throw new MWException( "Invalid language code requested" );
 824                 }
 825                 $this->recachedLangs[$code] = true;
 826
 827                 # Initial values
 828                 $initialData = array_fill_keys( self::$allKeys, null );
 829                 $coreData = $initialData;
 830                 $deps = [];
 831
 832                 # Load the primary localisation from the source file
 833                 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
 834                 if ( $data === false ) {
 835                         wfDebug( __METHOD__ . ": no localisation file for $code, using fallback to en\n" );
 836                         $coreData['fallback'] = 'en';
 837                 } else {
 838                         wfDebug( __METHOD__ . ": got localisation for $code from source\n" );
 839
 840                         # Merge primary localisation
 841                         foreach ( $data as $key => $value ) {
 842                                 $this->mergeItem( $key, $coreData[$key], $value );
 843                         }
 844                 }
 845
 846                 # Fill in the fallback if it's not there already
 847                 if ( is_null( $coreData['fallback'] ) ) {
 848                         $coreData['fallback'] = $code === 'en' ? false : 'en';
 849                 }
 850                 if ( $coreData['fallback'] === false ) {
 851                         $coreData['fallbackSequence'] = [];
 852                 } else {
 853                         $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
 854                         $len = count( $coreData['fallbackSequence'] );
 855
 856                         # Ensure that the sequence ends at en
 857                         if ( $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
 858                                 $coreData['fallbackSequence'][] = 'en';
 859                         }
 860                 }
 861
 862                 $codeSequence = array_merge( [ $code ], $coreData['fallbackSequence'] );
 863                 $messageDirs = $this->getMessagesDirs();
 864
 865                 # Load non-JSON localisation data for extensions
 866                 $extensionData = array_fill_keys( $codeSequence, $initialData );
 867                 foreach ( $wgExtensionMessagesFiles as $extension => $fileName ) {
 868                         if ( isset( $messageDirs[$extension] ) ) {
 869                                 # This extension has JSON message data; skip the PHP shim
 870                                 continue;
 871                         }
 872
 873                         $data = $this->readPHPFile( $fileName, 'extension' );
 874                         $used = false;
 875
 876                         foreach ( $data as $key => $item ) {
 877                                 foreach ( $codeSequence as $csCode ) {
 878                                         if ( isset( $item[$csCode] ) ) {
 879                                                 $this->mergeItem( $key, $extensionData[$csCode][$key], $item[$csCode] );
 880                                                 $used = true;
 881                                         }
 882                                 }
 883                         }
 884
 885                         if ( $used ) {
 886                                 $deps[] = new FileDependency( $fileName );
 887                         }
 888                 }
 889
 890                 # Load the localisation data for each fallback, then merge it into the full array
 891                 $allData = $initialData;
 892                 foreach ( $codeSequence as $csCode ) {
 893                         $csData = $initialData;
 894
 895                         # Load core messages and the extension localisations.
 896                         foreach ( $messageDirs as $dirs ) {
 897                                 foreach ( (array)$dirs as $dir ) {
 898                                         $fileName = "$dir/$csCode.json";
 899                                         $data = $this->readJSONFile( $fileName );
 900
 901                                         foreach ( $data as $key => $item ) {
 902                                                 $this->mergeItem( $key, $csData[$key], $item );
 903                                         }
 904
 905                                         $deps[] = new FileDependency( $fileName );
 906                                 }
 907                         }
 908
 909                         # Merge non-JSON extension data
 910                         if ( isset( $extensionData[$csCode] ) ) {
 911                                 foreach ( $extensionData[$csCode] as $key => $item ) {
 912                                         $this->mergeItem( $key, $csData[$key], $item );
 913                                 }
 914                         }
 915
 916                         if ( $csCode === $code ) {
 917                                 # Merge core data into extension data
 918                                 foreach ( $coreData as $key => $item ) {
 919                                         $this->mergeItem( $key, $csData[$key], $item );
 920                                 }
 921                         } else {
 922                                 # Load the secondary localisation from the source file to
 923                                 # avoid infinite cycles on cyclic fallbacks
 924                                 $fbData = $this->readSourceFilesAndRegisterDeps( $csCode, $deps );
 925                                 if ( $fbData !== false ) {
 926                                         # Only merge the keys that make sense to merge
 927                                         foreach ( self::$allKeys as $key ) {
 928                                                 if ( !isset( $fbData[$key] ) ) {
 929                                                         continue;
 930                                                 }
 931
 932                                                 if ( is_null( $coreData[$key] ) || $this->isMergeableKey( $key ) ) {
 933                                                         $this->mergeItem( $key, $csData[$key], $fbData[$key] );
 934                                                 }
 935                                         }
 936                                 }
 937                         }
 938
 939                         # Allow extensions an opportunity to adjust the data for this
 940                         # fallback
 941                         Hooks::run( 'LocalisationCacheRecacheFallback', [ $this, $csCode, &$csData ] );
 942
 943                         # Merge the data for this fallback into the final array
 944                         if ( $csCode === $code ) {
 945                                 $allData = $csData;
 946                         } else {
 947                                 foreach ( self::$allKeys as $key ) {
 948                                         if ( !isset( $csData[$key] ) ) {
 949                                                 continue;
 950                                         }
 951
 952                                         if ( is_null( $allData[$key] ) || $this->isMergeableKey( $key ) ) {
 953                                                 $this->mergeItem( $key, $allData[$key], $csData[$key] );
 954                                         }
 955                                 }
 956                         }
 957                 }
 958
 959                 # Add cache dependencies for any referenced globals
 960                 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
 961                 // $wgMessagesDirs is used in LocalisationCache::getMessagesDirs()
 962                 $deps['wgMessagesDirs'] = new GlobalDependency( 'wgMessagesDirs' );
 963                 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
 964
 965                 # Add dependencies to the cache entry
 966                 $allData['deps'] = $deps;
 967
 968                 # Replace spaces with underscores in namespace names
 969                 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
 970
 971                 # And do the same for special page aliases. $page is an array.
 972                 foreach ( $allData['specialPageAliases'] as &$page ) {
 973                         $page = str_replace( ' ', '_', $page );
 974                 }
 975                 # Decouple the reference to prevent accidental damage
 976                 unset( $page );
 977
 978                 # If there were no plural rules, return an empty array
 979                 if ( $allData['pluralRules'] === null ) {
 980                         $allData['pluralRules'] = [];
 981                 }
 982                 if ( $allData['compiledPluralRules'] === null ) {
 983                         $allData['compiledPluralRules'] = [];
 984                 }
 985                 # If there were no plural rule types, return an empty array
 986                 if ( $allData['pluralRuleTypes'] === null ) {
 987                         $allData['pluralRuleTypes'] = [];
 988                 }
 989
 990                 # Set the list keys
 991                 $allData['list'] = [];
 992                 foreach ( self::$splitKeys as $key ) {
 993                         $allData['list'][$key] = array_keys( $allData[$key] );
 994                 }
 995                 # Run hooks
 996                 $purgeBlobs = true;
 997                 Hooks::run( 'LocalisationCacheRecache', [ $this, $code, &$allData, &$purgeBlobs ] );
 998
 999                 if ( is_null( $allData['namespaceNames'] ) ) {
1000                         throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
1001                                 'Check that your languages/messages/MessagesEn.php file is intact.' );
1002                 }
1003
1004                 # Set the preload key
1005                 $allData['preload'] = $this->buildPreload( $allData );
1006
1007                 # Save to the process cache and register the items loaded
1008                 $this->data[$code] = $allData;
1009                 foreach ( $allData as $key => $item ) {
1010                         $this->loadedItems[$code][$key] = true;
1011                 }
1012
1013                 # Save to the persistent cache
1014                 $this->store->startWrite( $code );
1015                 foreach ( $allData as $key => $value ) {
1016                         if ( in_array( $key, self::$splitKeys ) ) {
1017                                 foreach ( $value as $subkey => $subvalue ) {
1018                                         $this->store->set( "$key:$subkey", $subvalue );
1019                                 }
1020                         } else {
1021                                 $this->store->set( $key, $value );
1022                         }
1023                 }
1024                 $this->store->finishWrite();
1025
1026                 # Clear out the MessageBlobStore
1027                 # HACK: If using a null (i.e. disabled) storage backend, we
1028                 # can't write to the MessageBlobStore either
1029                 if ( $purgeBlobs && !$this->store instanceof LCStoreNull ) {
1030                         $blobStore = new MessageBlobStore();
1031                         $blobStore->clear();
1032                 }
1033
1034         }
1035
1036         /**
1037          * Build the preload item from the given pre-cache data.
1038          *
1039          * The preload item will be loaded automatically, improving performance
1040          * for the commonly-requested items it contains.
1041          * @param array $data
1042          * @return array
1043          */
1044         protected function buildPreload( $data ) {
1045                 $preload = [ 'messages' => [] ];
1046                 foreach ( self::$preloadedKeys as $key ) {
1047                         $preload[$key] = $data[$key];
1048                 }
1049
1050                 foreach ( $data['preloadedMessages'] as $subkey ) {
1051                         if ( isset( $data['messages'][$subkey] ) ) {
1052                                 $subitem = $data['messages'][$subkey];
1053                         } else {
1054                                 $subitem = null;
1055                         }
1056                         $preload['messages'][$subkey] = $subitem;
1057                 }
1058
1059                 return $preload;
1060         }
1061
1062         /**
1063          * Unload the data for a given language from the object cache.
1064          * Reduces memory usage.
1065          * @param string $code
1066          */
1067         public function unload( $code ) {
1068                 unset( $this->data[$code] );
1069                 unset( $this->loadedItems[$code] );
1070                 unset( $this->loadedSubitems[$code] );
1071                 unset( $this->initialisedLangs[$code] );
1072                 unset( $this->shallowFallbacks[$code] );
1073
1074                 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1075                         if ( $fbCode === $code ) {
1076                                 $this->unload( $shallowCode );
1077                         }
1078                 }
1079         }
1080
1081         /**
1082          * Unload all data
1083          */
1084         public function unloadAll() {
1085                 foreach ( $this->initialisedLangs as $lang => $unused ) {
1086                         $this->unload( $lang );
1087                 }
1088         }
1089
1090         /**
1091          * Disable the storage backend
1092          */
1093         public function disableBackend() {
1094                 $this->store = new LCStoreNull;
1095                 $this->manualRecache = false;
1096         }
1097
1098 }