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          * For constructor parameters, see the documentation in DefaultSettings.php
 187          * for $wgLocalisationCacheConf.
 188          *
 189          * @param array $conf
 190          * @throws MWException
 191          */
 192         function __construct( $conf ) {
 193                 global $wgCacheDirectory;
 194
 195                 $this->conf = $conf;
 196                 $storeConf = [];
 197                 if ( !empty( $conf['storeClass'] ) ) {
 198                         $storeClass = $conf['storeClass'];
 199                 } else {
 200                         switch ( $conf['store'] ) {
 201                                 case 'files':
 202                                 case 'file':
 203                                         $storeClass = 'LCStoreCDB';
 204                                         break;
 205                                 case 'db':
 206                                         $storeClass = 'LCStoreDB';
 207                                         break;
 208                                 case 'array':
 209                                         $storeClass = 'LCStoreStaticArray';
 210                                         break;
 211                                 case 'detect':
 212                                         if ( !empty( $conf['storeDirectory'] ) ) {
 213                                                 $storeClass = 'LCStoreCDB';
 214                                         } elseif ( $wgCacheDirectory ) {
 215                                                 $storeConf['directory'] = $wgCacheDirectory;
 216                                                 $storeClass = 'LCStoreCDB';
 217                                         } else {
 218                                                 $storeClass = 'LCStoreDB';
 219                                         }
 220                                         break;
 221                                 default:
 222                                         throw new MWException(
 223                                                 'Please set $wgLocalisationCacheConf[\'store\'] to something sensible.'
 224                                         );
 225                         }
 226                 }
 227
 228                 wfDebugLog( 'caches', static::class . ": using store $storeClass" );
 229                 if ( !empty( $conf['storeDirectory'] ) ) {
 230                         $storeConf['directory'] = $conf['storeDirectory'];
 231                 }
 232
 233                 $this->store = new $storeClass( $storeConf );
 234                 foreach ( [ 'manualRecache', 'forceRecache' ] as $var ) {
 235                         if ( isset( $conf[$var] ) ) {
 236                                 $this->$var = $conf[$var];
 237                         }
 238                 }
 239         }
 240
 241         /**
 242          * Returns true if the given key is mergeable, that is, if it is an associative
 243          * array which can be merged through a fallback sequence.
 244          * @param string $key
 245          * @return bool
 246          */
 247         public function isMergeableKey( $key ) {
 248                 if ( $this->mergeableKeys === null ) {
 249                         $this->mergeableKeys = array_flip( array_merge(
 250                                 self::$mergeableMapKeys,
 251                                 self::$mergeableListKeys,
 252                                 self::$mergeableAliasListKeys,
 253                                 self::$optionalMergeKeys,
 254                                 self::$magicWordKeys
 255                         ) );
 256                 }
 257
 258                 return isset( $this->mergeableKeys[$key] );
 259         }
 260
 261         /**
 262          * Get a cache item.
 263          *
 264          * Warning: this may be slow for split items (messages), since it will
 265          * need to fetch all of the subitems from the cache individually.
 266          * @param string $code
 267          * @param string $key
 268          * @return mixed
 269          */
 270         public function getItem( $code, $key ) {
 271                 if ( !isset( $this->loadedItems[$code][$key] ) ) {
 272                         $this->loadItem( $code, $key );
 273                 }
 274
 275                 if ( $key === 'fallback' && isset( $this->shallowFallbacks[$code] ) ) {
 276                         return $this->shallowFallbacks[$code];
 277                 }
 278
 279                 return $this->data[$code][$key];
 280         }
 281
 282         /**
 283          * Get a subitem, for instance a single message for a given language.
 284          * @param string $code
 285          * @param string $key
 286          * @param string $subkey
 287          * @return mixed|null
 288          */
 289         public function getSubitem( $code, $key, $subkey ) {
 290                 if ( !isset( $this->loadedSubitems[$code][$key][$subkey] ) &&
 291                         !isset( $this->loadedItems[$code][$key] )
 292                 ) {
 293                         $this->loadSubitem( $code, $key, $subkey );
 294                 }
 295
 296                 if ( isset( $this->data[$code][$key][$subkey] ) ) {
 297                         return $this->data[$code][$key][$subkey];
 298                 } else {
 299                         return null;
 300                 }
 301         }
 302
 303         /**
 304          * Get the list of subitem keys for a given item.
 305          *
 306          * This is faster than array_keys($lc->getItem(...)) for the items listed in
 307          * self::$splitKeys.
 308          *
 309          * Will return null if the item is not found, or false if the item is not an
 310          * array.
 311          * @param string $code
 312          * @param string $key
 313          * @return bool|null|string|string[]
 314          */
 315         public function getSubitemList( $code, $key ) {
 316                 if ( in_array( $key, self::$splitKeys ) ) {
 317                         return $this->getSubitem( $code, 'list', $key );
 318                 } else {
 319                         $item = $this->getItem( $code, $key );
 320                         if ( is_array( $item ) ) {
 321                                 return array_keys( $item );
 322                         } else {
 323                                 return false;
 324                         }
 325                 }
 326         }
 327
 328         /**
 329          * Load an item into the cache.
 330          * @param string $code
 331          * @param string $key
 332          */
 333         protected function loadItem( $code, $key ) {
 334                 if ( !isset( $this->initialisedLangs[$code] ) ) {
 335                         $this->initLanguage( $code );
 336                 }
 337
 338                 // Check to see if initLanguage() loaded it for us
 339                 if ( isset( $this->loadedItems[$code][$key] ) ) {
 340                         return;
 341                 }
 342
 343                 if ( isset( $this->shallowFallbacks[$code] ) ) {
 344                         $this->loadItem( $this->shallowFallbacks[$code], $key );
 345
 346                         return;
 347                 }
 348
 349                 if ( in_array( $key, self::$splitKeys ) ) {
 350                         $subkeyList = $this->getSubitem( $code, 'list', $key );
 351                         foreach ( $subkeyList as $subkey ) {
 352                                 if ( isset( $this->data[$code][$key][$subkey] ) ) {
 353                                         continue;
 354                                 }
 355                                 $this->data[$code][$key][$subkey] = $this->getSubitem( $code, $key, $subkey );
 356                         }
 357                 } else {
 358                         $this->data[$code][$key] = $this->store->get( $code, $key );
 359                 }
 360
 361                 $this->loadedItems[$code][$key] = true;
 362         }
 363
 364         /**
 365          * Load a subitem into the cache
 366          * @param string $code
 367          * @param string $key
 368          * @param string $subkey
 369          */
 370         protected function loadSubitem( $code, $key, $subkey ) {
 371                 if ( !in_array( $key, self::$splitKeys ) ) {
 372                         $this->loadItem( $code, $key );
 373
 374                         return;
 375                 }
 376
 377                 if ( !isset( $this->initialisedLangs[$code] ) ) {
 378                         $this->initLanguage( $code );
 379                 }
 380
 381                 // Check to see if initLanguage() loaded it for us
 382                 if ( isset( $this->loadedItems[$code][$key] ) ||
 383                         isset( $this->loadedSubitems[$code][$key][$subkey] )
 384                 ) {
 385                         return;
 386                 }
 387
 388                 if ( isset( $this->shallowFallbacks[$code] ) ) {
 389                         $this->loadSubitem( $this->shallowFallbacks[$code], $key, $subkey );
 390
 391                         return;
 392                 }
 393
 394                 $value = $this->store->get( $code, "$key:$subkey" );
 395                 $this->data[$code][$key][$subkey] = $value;
 396                 $this->loadedSubitems[$code][$key][$subkey] = true;
 397         }
 398
 399         /**
 400          * Returns true if the cache identified by $code is missing or expired.
 401          *
 402          * @param string $code
 403          *
 404          * @return bool
 405          */
 406         public function isExpired( $code ) {
 407                 if ( $this->forceRecache && !isset( $this->recachedLangs[$code] ) ) {
 408                         wfDebug( __METHOD__ . "($code): forced reload\n" );
 409
 410                         return true;
 411                 }
 412
 413                 $deps = $this->store->get( $code, 'deps' );
 414                 $keys = $this->store->get( $code, 'list' );
 415                 $preload = $this->store->get( $code, 'preload' );
 416                 // Different keys may expire separately for some stores
 417                 if ( $deps === null || $keys === null || $preload === null ) {
 418                         wfDebug( __METHOD__ . "($code): cache missing, need to make one\n" );
 419
 420                         return true;
 421                 }
 422
 423                 foreach ( $deps as $dep ) {
 424                         // Because we're unserializing stuff from cache, we
 425                         // could receive objects of classes that don't exist
 426                         // anymore (e.g. uninstalled extensions)
 427                         // When this happens, always expire the cache
 428                         if ( !$dep instanceof CacheDependency || $dep->isExpired() ) {
 429                                 wfDebug( __METHOD__ . "($code): cache for $code expired due to " .
 430                                         get_class( $dep ) . "\n" );
 431
 432                                 return true;
 433                         }
 434                 }
 435
 436                 return false;
 437         }
 438
 439         /**
 440          * Initialise a language in this object. Rebuild the cache if necessary.
 441          * @param string $code
 442          * @throws MWException
 443          */
 444         protected function initLanguage( $code ) {
 445                 if ( isset( $this->initialisedLangs[$code] ) ) {
 446                         return;
 447                 }
 448
 449                 $this->initialisedLangs[$code] = true;
 450
 451                 # If the code is of the wrong form for a Messages*.php file, do a shallow fallback
 452                 if ( !Language::isValidBuiltInCode( $code ) ) {
 453                         $this->initShallowFallback( $code, 'en' );
 454
 455                         return;
 456                 }
 457
 458                 # Recache the data if necessary
 459                 if ( !$this->manualRecache && $this->isExpired( $code ) ) {
 460                         if ( Language::isSupportedLanguage( $code ) ) {
 461                                 $this->recache( $code );
 462                         } elseif ( $code === 'en' ) {
 463                                 throw new MWException( 'MessagesEn.php is missing.' );
 464                         } else {
 465                                 $this->initShallowFallback( $code, 'en' );
 466                         }
 467
 468                         return;
 469                 }
 470
 471                 # Preload some stuff
 472                 $preload = $this->getItem( $code, 'preload' );
 473                 if ( $preload === null ) {
 474                         if ( $this->manualRecache ) {
 475                                 // No Messages*.php file. Do shallow fallback to en.
 476                                 if ( $code === 'en' ) {
 477                                         throw new MWException( 'No localisation cache found for English. ' .
 478                                                 'Please run maintenance/rebuildLocalisationCache.php.' );
 479                                 }
 480                                 $this->initShallowFallback( $code, 'en' );
 481
 482                                 return;
 483                         } else {
 484                                 throw new MWException( 'Invalid or missing localisation cache.' );
 485                         }
 486                 }
 487                 $this->data[$code] = $preload;
 488                 foreach ( $preload as $key => $item ) {
 489                         if ( in_array( $key, self::$splitKeys ) ) {
 490                                 foreach ( $item as $subkey => $subitem ) {
 491                                         $this->loadedSubitems[$code][$key][$subkey] = true;
 492                                 }
 493                         } else {
 494                                 $this->loadedItems[$code][$key] = true;
 495                         }
 496                 }
 497         }
 498
 499         /**
 500          * Create a fallback from one language to another, without creating a
 501          * complete persistent cache.
 502          * @param string $primaryCode
 503          * @param string $fallbackCode
 504          */
 505         public function initShallowFallback( $primaryCode, $fallbackCode ) {
 506                 $this->data[$primaryCode] =& $this->data[$fallbackCode];
 507                 $this->loadedItems[$primaryCode] =& $this->loadedItems[$fallbackCode];
 508                 $this->loadedSubitems[$primaryCode] =& $this->loadedSubitems[$fallbackCode];
 509                 $this->shallowFallbacks[$primaryCode] = $fallbackCode;
 510         }
 511
 512         /**
 513          * Read a PHP file containing localisation data.
 514          * @param string $_fileName
 515          * @param string $_fileType
 516          * @throws MWException
 517          * @return array
 518          */
 519         protected function readPHPFile( $_fileName, $_fileType ) {
 520                 // Disable APC caching
 521                 MediaWiki\suppressWarnings();
 522                 $_apcEnabled = ini_set( 'apc.cache_by_default', '0' );
 523                 MediaWiki\restoreWarnings();
 524
 525                 include $_fileName;
 526
 527                 MediaWiki\suppressWarnings();
 528                 ini_set( 'apc.cache_by_default', $_apcEnabled );
 529                 MediaWiki\restoreWarnings();
 530
 531                 if ( $_fileType == 'core' || $_fileType == 'extension' ) {
 532                         $data = compact( self::$allKeys );
 533                 } elseif ( $_fileType == 'aliases' ) {
 534                         $data = compact( 'aliases' );
 535                 } else {
 536                         throw new MWException( __METHOD__ . ": Invalid file type: $_fileType" );
 537                 }
 538
 539                 return $data;
 540         }
 541
 542         /**
 543          * Read a JSON file containing localisation messages.
 544          * @param string $fileName Name of file to read
 545          * @throws MWException If there is a syntax error in the JSON file
 546          * @return array Array with a 'messages' key, or empty array if the file doesn't exist
 547          */
 548         public function readJSONFile( $fileName ) {
 549                 if ( !is_readable( $fileName ) ) {
 550                         return [];
 551                 }
 552
 553                 $json = file_get_contents( $fileName );
 554                 if ( $json === false ) {
 555                         return [];
 556                 }
 557
 558                 $data = FormatJson::decode( $json, true );
 559                 if ( $data === null ) {
 560                         throw new MWException( __METHOD__ . ": Invalid JSON file: $fileName" );
 561                 }
 562
 563                 // Remove keys starting with '@', they're reserved for metadata and non-message data
 564                 foreach ( $data as $key => $unused ) {
 565                         if ( $key === '' || $key[0] === '@' ) {
 566                                 unset( $data[$key] );
 567                         }
 568                 }
 569
 570                 // The JSON format only supports messages, none of the other variables, so wrap the data
 571                 return [ 'messages' => $data ];
 572         }
 573
 574         /**
 575          * Get the compiled plural rules for a given language from the XML files.
 576          * @since 1.20
 577          * @param string $code
 578          * @return array|null
 579          */
 580         public function getCompiledPluralRules( $code ) {
 581                 $rules = $this->getPluralRules( $code );
 582                 if ( $rules === null ) {
 583                         return null;
 584                 }
 585                 try {
 586                         $compiledRules = Evaluator::compile( $rules );
 587                 } catch ( CLDRPluralRuleError $e ) {
 588                         wfDebugLog( 'l10n', $e->getMessage() );
 589
 590                         return [];
 591                 }
 592
 593                 return $compiledRules;
 594         }
 595
 596         /**
 597          * Get the plural rules for a given language from the XML files.
 598          * Cached.
 599          * @since 1.20
 600          * @param string $code
 601          * @return array|null
 602          */
 603         public function getPluralRules( $code ) {
 604                 if ( $this->pluralRules === null ) {
 605                         $this->loadPluralFiles();
 606                 }
 607                 if ( !isset( $this->pluralRules[$code] ) ) {
 608                         return null;
 609                 } else {
 610                         return $this->pluralRules[$code];
 611                 }
 612         }
 613
 614         /**
 615          * Get the plural rule types for a given language from the XML files.
 616          * Cached.
 617          * @since 1.22
 618          * @param string $code
 619          * @return array|null
 620          */
 621         public function getPluralRuleTypes( $code ) {
 622                 if ( $this->pluralRuleTypes === null ) {
 623                         $this->loadPluralFiles();
 624                 }
 625                 if ( !isset( $this->pluralRuleTypes[$code] ) ) {
 626                         return null;
 627                 } else {
 628                         return $this->pluralRuleTypes[$code];
 629                 }
 630         }
 631
 632         /**
 633          * Load the plural XML files.
 634          */
 635         protected function loadPluralFiles() {
 636                 global $IP;
 637                 $cldrPlural = "$IP/languages/data/plurals.xml";
 638                 $mwPlural = "$IP/languages/data/plurals-mediawiki.xml";
 639                 // Load CLDR plural rules
 640                 $this->loadPluralFile( $cldrPlural );
 641                 if ( file_exists( $mwPlural ) ) {
 642                         // Override or extend
 643                         $this->loadPluralFile( $mwPlural );
 644                 }
 645         }
 646
 647         /**
 648          * Load a plural XML file with the given filename, compile the relevant
 649          * rules, and save the compiled rules in a process-local cache.
 650          *
 651          * @param string $fileName
 652          * @throws MWException
 653          */
 654         protected function loadPluralFile( $fileName ) {
 655                 // Use file_get_contents instead of DOMDocument::load (T58439)
 656                 $xml = file_get_contents( $fileName );
 657                 if ( !$xml ) {
 658                         throw new MWException( "Unable to read plurals file $fileName" );
 659                 }
 660                 $doc = new DOMDocument;
 661                 $doc->loadXML( $xml );
 662                 $rulesets = $doc->getElementsByTagName( "pluralRules" );
 663                 foreach ( $rulesets as $ruleset ) {
 664                         $codes = $ruleset->getAttribute( 'locales' );
 665                         $rules = [];
 666                         $ruleTypes = [];
 667                         $ruleElements = $ruleset->getElementsByTagName( "pluralRule" );
 668                         foreach ( $ruleElements as $elt ) {
 669                                 $ruleType = $elt->getAttribute( 'count' );
 670                                 if ( $ruleType === 'other' ) {
 671                                         // Don't record "other" rules, which have an empty condition
 672                                         continue;
 673                                 }
 674                                 $rules[] = $elt->nodeValue;
 675                                 $ruleTypes[] = $ruleType;
 676                         }
 677                         foreach ( explode( ' ', $codes ) as $code ) {
 678                                 $this->pluralRules[$code] = $rules;
 679                                 $this->pluralRuleTypes[$code] = $ruleTypes;
 680                         }
 681                 }
 682         }
 683
 684         /**
 685          * Read the data from the source files for a given language, and register
 686          * the relevant dependencies in the $deps array. If the localisation
 687          * exists, the data array is returned, otherwise false is returned.
 688          *
 689          * @param string $code
 690          * @param array $deps
 691          * @return array
 692          */
 693         protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
 694                 global $IP;
 695
 696                 // This reads in the PHP i18n file with non-messages l10n data
 697                 $fileName = Language::getMessagesFileName( $code );
 698                 if ( !file_exists( $fileName ) ) {
 699                         $data = [];
 700                 } else {
 701                         $deps[] = new FileDependency( $fileName );
 702                         $data = $this->readPHPFile( $fileName, 'core' );
 703                 }
 704
 705                 # Load CLDR plural rules for JavaScript
 706                 $data['pluralRules'] = $this->getPluralRules( $code );
 707                 # And for PHP
 708                 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
 709                 # Load plural rule types
 710                 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
 711
 712                 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
 713                 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
 714
 715                 return $data;
 716         }
 717
 718         /**
 719          * Merge two localisation values, a primary and a fallback, overwriting the
 720          * primary value in place.
 721          * @param string $key
 722          * @param mixed $value
 723          * @param mixed $fallbackValue
 724          */
 725         protected function mergeItem( $key, &$value, $fallbackValue ) {
 726                 if ( !is_null( $value ) ) {
 727                         if ( !is_null( $fallbackValue ) ) {
 728                                 if ( in_array( $key, self::$mergeableMapKeys ) ) {
 729                                         $value = $value + $fallbackValue;
 730                                 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
 731                                         $value = array_unique( array_merge( $fallbackValue, $value ) );
 732                                 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
 733                                         $value = array_merge_recursive( $value, $fallbackValue );
 734                                 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
 735                                         if ( !empty( $value['inherit'] ) ) {
 736                                                 $value = array_merge( $fallbackValue, $value );
 737                                         }
 738
 739                                         if ( isset( $value['inherit'] ) ) {
 740                                                 unset( $value['inherit'] );
 741                                         }
 742                                 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
 743                                         $this->mergeMagicWords( $value, $fallbackValue );
 744                                 }
 745                         }
 746                 } else {
 747                         $value = $fallbackValue;
 748                 }
 749         }
 750
 751         /**
 752          * @param mixed $value
 753          * @param mixed $fallbackValue
 754          */
 755         protected function mergeMagicWords( &$value, $fallbackValue ) {
 756                 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
 757                         if ( !isset( $value[$magicName] ) ) {
 758                                 $value[$magicName] = $fallbackInfo;
 759                         } else {
 760                                 $oldSynonyms = array_slice( $fallbackInfo, 1 );
 761                                 $newSynonyms = array_slice( $value[$magicName], 1 );
 762                                 $synonyms = array_values( array_unique( array_merge(
 763                                         $newSynonyms, $oldSynonyms ) ) );
 764                                 $value[$magicName] = array_merge( [ $fallbackInfo[0] ], $synonyms );
 765                         }
 766                 }
 767         }
 768
 769         /**
 770          * Given an array mapping language code to localisation value, such as is
 771          * found in extension *.i18n.php files, iterate through a fallback sequence
 772          * to merge the given data with an existing primary value.
 773          *
 774          * Returns true if any data from the extension array was used, false
 775          * otherwise.
 776          * @param array $codeSequence
 777          * @param string $key
 778          * @param mixed $value
 779          * @param mixed $fallbackValue
 780          * @return bool
 781          */
 782         protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
 783                 $used = false;
 784                 foreach ( $codeSequence as $code ) {
 785                         if ( isset( $fallbackValue[$code] ) ) {
 786                                 $this->mergeItem( $key, $value, $fallbackValue[$code] );
 787                                 $used = true;
 788                         }
 789                 }
 790
 791                 return $used;
 792         }
 793
 794         /**
 795          * Gets the combined list of messages dirs from
 796          * core and extensions
 797          *
 798          * @since 1.25
 799          * @return array
 800          */
 801         public function getMessagesDirs() {
 802                 global $IP;
 803
 804                 $config = MediaWikiServices::getInstance()->getMainConfig();
 805                 $messagesDirs = $config->get( 'MessagesDirs' );
 806                 return [
 807                         'core' => "$IP/languages/i18n",
 808                         'api' => "$IP/includes/api/i18n",
 809                         'oojs-ui' => "$IP/resources/lib/oojs-ui/i18n",
 810                 ] + $messagesDirs;
 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                 // The 'MessagesDirs' config setting is used in LocalisationCache::getMessagesDirs().
 962                 // We use the key 'wgMessagesDirs' for historical reasons.
 963                 $deps['wgMessagesDirs'] = new MainConfigDependency( 'MessagesDirs' );
 964                 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
 965
 966                 # Add dependencies to the cache entry
 967                 $allData['deps'] = $deps;
 968
 969                 # Replace spaces with underscores in namespace names
 970                 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
 971
 972                 # And do the same for special page aliases. $page is an array.
 973                 foreach ( $allData['specialPageAliases'] as &$page ) {
 974                         $page = str_replace( ' ', '_', $page );
 975                 }
 976                 # Decouple the reference to prevent accidental damage
 977                 unset( $page );
 978
 979                 # If there were no plural rules, return an empty array
 980                 if ( $allData['pluralRules'] === null ) {
 981                         $allData['pluralRules'] = [];
 982                 }
 983                 if ( $allData['compiledPluralRules'] === null ) {
 984                         $allData['compiledPluralRules'] = [];
 985                 }
 986                 # If there were no plural rule types, return an empty array
 987                 if ( $allData['pluralRuleTypes'] === null ) {
 988                         $allData['pluralRuleTypes'] = [];
 989                 }
 990
 991                 # Set the list keys
 992                 $allData['list'] = [];
 993                 foreach ( self::$splitKeys as $key ) {
 994                         $allData['list'][$key] = array_keys( $allData[$key] );
 995                 }
 996                 # Run hooks
 997                 $purgeBlobs = true;
 998                 Hooks::run( 'LocalisationCacheRecache', [ $this, $code, &$allData, &$purgeBlobs ] );
 999
1000                 if ( is_null( $allData['namespaceNames'] ) ) {
1001                         throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
1002                                 'Check that your languages/messages/MessagesEn.php file is intact.' );
1003                 }
1004
1005                 # Set the preload key
1006                 $allData['preload'] = $this->buildPreload( $allData );
1007
1008                 # Save to the process cache and register the items loaded
1009                 $this->data[$code] = $allData;
1010                 foreach ( $allData as $key => $item ) {
1011                         $this->loadedItems[$code][$key] = true;
1012                 }
1013
1014                 # Save to the persistent cache
1015                 $this->store->startWrite( $code );
1016                 foreach ( $allData as $key => $value ) {
1017                         if ( in_array( $key, self::$splitKeys ) ) {
1018                                 foreach ( $value as $subkey => $subvalue ) {
1019                                         $this->store->set( "$key:$subkey", $subvalue );
1020                                 }
1021                         } else {
1022                                 $this->store->set( $key, $value );
1023                         }
1024                 }
1025                 $this->store->finishWrite();
1026
1027                 # Clear out the MessageBlobStore
1028                 # HACK: If using a null (i.e. disabled) storage backend, we
1029                 # can't write to the MessageBlobStore either
1030                 if ( $purgeBlobs && !$this->store instanceof LCStoreNull ) {
1031                         $blobStore = new MessageBlobStore();
1032                         $blobStore->clear();
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 }