languages/Language.php

   1 <?php
   2 /**
   3  * Internationalisation code.
   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  * @ingroup Language
  22  */
  23
  24 /**
  25  * @defgroup Language Language
  26  */
  27
  28 if ( !defined( 'MEDIAWIKI' ) ) {
  29         echo "This file is part of MediaWiki, it is not a valid entry point.\n";
  30         exit( 1 );
  31 }
  32
  33 # Read language names
  34 global $wgLanguageNames;
  35 require_once( __DIR__ . '/Names.php' );
  36
  37 if ( function_exists( 'mb_strtoupper' ) ) {
  38         mb_internal_encoding( 'UTF-8' );
  39 }
  40
  41 /**
  42  * a fake language converter
  43  *
  44  * @ingroup Language
  45  */
  46 class FakeConverter {
  47
  48         /**
  49          * @var Language
  50          */
  51         public $mLang;
  52         function __construct( $langobj ) { $this->mLang = $langobj; }
  53         function autoConvertToAllVariants( $text ) { return array( $this->mLang->getCode() => $text ); }
  54         function convert( $t ) { return $t; }
  55         function convertTo( $text, $variant ) { return $text; }
  56         function convertTitle( $t ) { return $t->getPrefixedText(); }
  57         function convertNamespace( $ns ) { return $this->mLang->getFormattedNsText( $ns ); }
  58         function getVariants() { return array( $this->mLang->getCode() ); }
  59         function getPreferredVariant() { return $this->mLang->getCode(); }
  60         function getDefaultVariant() { return $this->mLang->getCode(); }
  61         function getURLVariant() { return ''; }
  62         function getConvRuleTitle() { return false; }
  63         function findVariantLink( &$l, &$n, $ignoreOtherCond = false ) { }
  64         function getExtraHashOptions() { return ''; }
  65         function getParsedTitle() { return ''; }
  66         function markNoConversion( $text, $noParse = false ) { return $text; }
  67         function convertCategoryKey( $key ) { return $key; }
  68         function convertLinkToAllVariants( $text ) { return $this->autoConvertToAllVariants( $text ); }
  69         function armourMath( $text ) { return $text; }
  70 }
  71
  72 /**
  73  * Internationalisation code
  74  * @ingroup Language
  75  */
  76 class Language {
  77
  78         /**
  79          * @var LanguageConverter
  80          */
  81         public $mConverter;
  82
  83         public $mVariants, $mCode, $mLoaded = false;
  84         public $mMagicExtensions = array(), $mMagicHookDone = false;
  85         private $mHtmlCode = null;
  86
  87         public $dateFormatStrings = array();
  88         public $mExtendedSpecialPageAliases;
  89
  90         protected $namespaceNames, $mNamespaceIds, $namespaceAliases;
  91
  92         /**
  93          * ReplacementArray object caches
  94          */
  95         public $transformData = array();
  96
  97         /**
  98          * @var LocalisationCache
  99          */
 100         static public $dataCache;
 101
 102         static public $mLangObjCache = array();
 103
 104         static public $mWeekdayMsgs = array(
 105                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
 106                 'friday', 'saturday'
 107         );
 108
 109         static public $mWeekdayAbbrevMsgs = array(
 110                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
 111         );
 112
 113         static public $mMonthMsgs = array(
 114                 'january', 'february', 'march', 'april', 'may_long', 'june',
 115                 'july', 'august', 'september', 'october', 'november',
 116                 'december'
 117         );
 118         static public $mMonthGenMsgs = array(
 119                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
 120                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
 121                 'december-gen'
 122         );
 123         static public $mMonthAbbrevMsgs = array(
 124                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
 125                 'sep', 'oct', 'nov', 'dec'
 126         );
 127
 128         static public $mIranianCalendarMonthMsgs = array(
 129                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 130                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 131                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 132                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 133         );
 134
 135         static public $mHebrewCalendarMonthMsgs = array(
 136                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 137                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 138                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 139                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 140                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 141         );
 142
 143         static public $mHebrewCalendarMonthGenMsgs = array(
 144                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 145                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 146                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 147                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 148                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 149         );
 150
 151         static public $mHijriCalendarMonthMsgs = array(
 152                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 153                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 154                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 155                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 156         );
 157
 158         /**
 159          * @since 1.20
 160          * @var array
 161          */
 162         static public $durationIntervals = array(
 163                 'millennia' => 31557600000,
 164                 'centuries' => 3155760000,
 165                 'decades' => 315576000,
 166                 'years' => 31557600, // 86400 * 365.25
 167                 'weeks' => 604800,
 168                 'days' => 86400,
 169                 'hours' => 3600,
 170                 'minutes' => 60,
 171                 'seconds' => 1,
 172         );
 173
 174         /**
 175          * Get a cached or new language object for a given language code
 176          * @param $code String
 177          * @return Language
 178          */
 179         static function factory( $code ) {
 180                 global $wgLangObjCacheSize;
 181
 182                 // get the language object to process
 183                 $langObj = isset( self::$mLangObjCache[$code] )
 184                         ? self::$mLangObjCache[$code]
 185                         : self::newFromCode( $code );
 186
 187                 // merge the language object in to get it up front in the cache
 188                 self::$mLangObjCache = array_merge( array( $code => $langObj ), self::$mLangObjCache );
 189                 // get rid of the oldest ones in case we have an overflow
 190                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 191
 192                 return $langObj;
 193         }
 194
 195         /**
 196          * Create a language object for a given language code
 197          * @param $code String
 198          * @throws MWException
 199          * @return Language
 200          */
 201         protected static function newFromCode( $code ) {
 202                 // Protect against path traversal below
 203                 if ( !Language::isValidCode( $code )
 204                         || strcspn( $code, ":/\\\000" ) !== strlen( $code ) )
 205                 {
 206                         throw new MWException( "Invalid language code \"$code\"" );
 207                 }
 208
 209                 if ( !Language::isValidBuiltInCode( $code ) ) {
 210                         // It's not possible to customise this code with class files, so
 211                         // just return a Language object. This is to support uselang= hacks.
 212                         $lang = new Language;
 213                         $lang->setCode( $code );
 214                         return $lang;
 215                 }
 216
 217                 // Check if there is a language class for the code
 218                 $class = self::classFromCode( $code );
 219                 self::preloadLanguageClass( $class );
 220                 if ( MWInit::classExists( $class ) ) {
 221                         $lang = new $class;
 222                         return $lang;
 223                 }
 224
 225                 // Keep trying the fallback list until we find an existing class
 226                 $fallbacks = Language::getFallbacksFor( $code );
 227                 foreach ( $fallbacks as $fallbackCode ) {
 228                         if ( !Language::isValidBuiltInCode( $fallbackCode ) ) {
 229                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 230                         }
 231
 232                         $class = self::classFromCode( $fallbackCode );
 233                         self::preloadLanguageClass( $class );
 234                         if ( MWInit::classExists( $class ) ) {
 235                                 $lang = Language::newFromCode( $fallbackCode );
 236                                 $lang->setCode( $code );
 237                                 return $lang;
 238                         }
 239                 }
 240
 241                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 242         }
 243
 244         /**
 245          * Returns true if a language code string is of a valid form, whether or
 246          * not it exists. This includes codes which are used solely for
 247          * customisation via the MediaWiki namespace.
 248          *
 249          * @param $code string
 250          *
 251          * @return bool
 252          */
 253         public static function isValidCode( $code ) {
 254                 return
 255                         // People think language codes are html safe, so enforce it.
 256                         // Ideally we should only allow a-zA-Z0-9-
 257                         // but, .+ and other chars are often used for {{int:}} hacks
 258                         // see bugs 37564, 37587, 36938
 259                         strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 260                         && !preg_match( Title::getTitleInvalidRegex(), $code );
 261         }
 262
 263         /**
 264          * Returns true if a language code is of a valid form for the purposes of
 265          * internal customisation of MediaWiki, via Messages*.php.
 266          *
 267          * @param $code string
 268          *
 269          * @throws MWException
 270          * @since 1.18
 271          * @return bool
 272          */
 273         public static function isValidBuiltInCode( $code ) {
 274
 275                 if ( !is_string( $code ) ) {
 276                         $type = gettype( $code );
 277                         if ( $type === 'object' ) {
 278                                 $addmsg = " of class " . get_class( $code );
 279                         } else {
 280                                 $addmsg = '';
 281                         }
 282                         throw new MWException( __METHOD__ . " must be passed a string, $type given$addmsg" );
 283                 }
 284
 285                 return (bool)preg_match( '/^[a-z0-9-]+$/i', $code );
 286         }
 287
 288         /**
 289          * @param $code
 290          * @return String Name of the language class
 291          */
 292         public static function classFromCode( $code ) {
 293                 if ( $code == 'en' ) {
 294                         return 'Language';
 295                 } else {
 296                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
 297                 }
 298         }
 299
 300         /**
 301          * Includes language class files
 302          *
 303          * @param $class string Name of the language class
 304          */
 305         public static function preloadLanguageClass( $class ) {
 306                 global $IP;
 307
 308                 if ( $class === 'Language' ) {
 309                         return;
 310                 }
 311
 312                 if ( !defined( 'MW_COMPILED' ) ) {
 313                         if ( file_exists( "$IP/languages/classes/$class.php" ) ) {
 314                                 include_once( "$IP/languages/classes/$class.php" );
 315                         }
 316                 }
 317         }
 318
 319         /**
 320          * Get the LocalisationCache instance
 321          *
 322          * @return LocalisationCache
 323          */
 324         public static function getLocalisationCache() {
 325                 if ( is_null( self::$dataCache ) ) {
 326                         global $wgLocalisationCacheConf;
 327                         $class = $wgLocalisationCacheConf['class'];
 328                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 329                 }
 330                 return self::$dataCache;
 331         }
 332
 333         function __construct() {
 334                 $this->mConverter = new FakeConverter( $this );
 335                 // Set the code to the name of the descendant
 336                 if ( get_class( $this ) == 'Language' ) {
 337                         $this->mCode = 'en';
 338                 } else {
 339                         $this->mCode = str_replace( '_', '-', strtolower( substr( get_class( $this ), 8 ) ) );
 340                 }
 341                 self::getLocalisationCache();
 342         }
 343
 344         /**
 345          * Reduce memory usage
 346          */
 347         function __destruct() {
 348                 foreach ( $this as $name => $value ) {
 349                         unset( $this->$name );
 350                 }
 351         }
 352
 353         /**
 354          * Hook which will be called if this is the content language.
 355          * Descendants can use this to register hook functions or modify globals
 356          */
 357         function initContLang() { }
 358
 359         /**
 360          * Same as getFallbacksFor for current language.
 361          * @return array|bool
 362          * @deprecated in 1.19
 363          */
 364         function getFallbackLanguageCode() {
 365                 wfDeprecated( __METHOD__, '1.19' );
 366                 return self::getFallbackFor( $this->mCode );
 367         }
 368
 369         /**
 370          * @return array
 371          * @since 1.19
 372          */
 373         function getFallbackLanguages() {
 374                 return self::getFallbacksFor( $this->mCode );
 375         }
 376
 377         /**
 378          * Exports $wgBookstoreListEn
 379          * @return array
 380          */
 381         function getBookstoreList() {
 382                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 383         }
 384
 385         /**
 386          * @return array
 387          */
 388         public function getNamespaces() {
 389                 if ( is_null( $this->namespaceNames ) ) {
 390                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 391
 392                         $this->namespaceNames = self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 393                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 394
 395                         $this->namespaceNames = $wgExtraNamespaces + $this->namespaceNames + $validNamespaces;
 396
 397                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 398                         if ( $wgMetaNamespaceTalk ) {
 399                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 400                         } else {
 401                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 402                                 $this->namespaceNames[NS_PROJECT_TALK] =
 403                                         $this->fixVariableInNamespace( $talk );
 404                         }
 405
 406                         # Sometimes a language will be localised but not actually exist on this wiki.
 407                         foreach ( $this->namespaceNames as $key => $text ) {
 408                                 if ( !isset( $validNamespaces[$key] ) ) {
 409                                         unset( $this->namespaceNames[$key] );
 410                                 }
 411                         }
 412
 413                         # The above mixing may leave namespaces out of canonical order.
 414                         # Re-order by namespace ID number...
 415                         ksort( $this->namespaceNames );
 416
 417                         wfRunHooks( 'LanguageGetNamespaces', array( &$this->namespaceNames ) );
 418                 }
 419                 return $this->namespaceNames;
 420         }
 421
 422         /**
 423          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 424          * @param $namespaces Array of namespaces (id => name)
 425          */
 426         public function setNamespaces( array $namespaces ) {
 427                 $this->namespaceNames = $namespaces;
 428                 $this->mNamespaceIds = null;
 429         }
 430
 431         /**
 432          * Resets all of the namespace caches. Mainly used for testing
 433          */
 434         public function resetNamespaces( ) {
 435                 $this->namespaceNames = null;
 436                 $this->mNamespaceIds = null;
 437                 $this->namespaceAliases = null;
 438         }
 439
 440         /**
 441          * A convenience function that returns the same thing as
 442          * getNamespaces() except with the array values changed to ' '
 443          * where it found '_', useful for producing output to be displayed
 444          * e.g. in <select> forms.
 445          *
 446          * @return array
 447          */
 448         function getFormattedNamespaces() {
 449                 $ns = $this->getNamespaces();
 450                 foreach ( $ns as $k => $v ) {
 451                         $ns[$k] = strtr( $v, '_', ' ' );
 452                 }
 453                 return $ns;
 454         }
 455
 456         /**
 457          * Get a namespace value by key
 458          * <code>
 459          * $mw_ns = $wgContLang->getNsText( NS_MEDIAWIKI );
 460          * echo $mw_ns; // prints 'MediaWiki'
 461          * </code>
 462          *
 463          * @param $index Int: the array key of the namespace to return
 464          * @return mixed, string if the namespace value exists, otherwise false
 465          */
 466         function getNsText( $index ) {
 467                 $ns = $this->getNamespaces();
 468                 return isset( $ns[$index] ) ? $ns[$index] : false;
 469         }
 470
 471         /**
 472          * A convenience function that returns the same thing as
 473          * getNsText() except with '_' changed to ' ', useful for
 474          * producing output.
 475          *
 476          * <code>
 477          * $mw_ns = $wgContLang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 478          * echo $mw_ns; // prints 'MediaWiki talk'
 479          * </code>
 480          *
 481          * @param int $index The array key of the namespace to return
 482          * @return string Namespace name without underscores (empty string if namespace does not exist)
 483          */
 484         function getFormattedNsText( $index ) {
 485                 $ns = $this->getNsText( $index );
 486                 return strtr( $ns, '_', ' ' );
 487         }
 488
 489         /**
 490          * Returns gender-dependent namespace alias if available.
 491          * @param $index Int: namespace index
 492          * @param $gender String: gender key (male, female... )
 493          * @return String
 494          * @since 1.18
 495          */
 496         function getGenderNsText( $index, $gender ) {
 497                 global $wgExtraGenderNamespaces;
 498
 499                 $ns = $wgExtraGenderNamespaces + self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 500                 return isset( $ns[$index][$gender] ) ? $ns[$index][$gender] : $this->getNsText( $index );
 501         }
 502
 503         /**
 504          * Whether this language makes distinguishes genders for example in
 505          * namespaces.
 506          * @return bool
 507          * @since 1.18
 508          */
 509         function needsGenderDistinction() {
 510                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 511                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 512                         // $wgExtraGenderNamespaces overrides everything
 513                         return true;
 514                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 515                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 516                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 517                         return false;
 518                 } else {
 519                         // Check what is in i18n files
 520                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 521                         return count( $aliases ) > 0;
 522                 }
 523         }
 524
 525         /**
 526          * Get a namespace key by value, case insensitive.
 527          * Only matches namespace names for the current language, not the
 528          * canonical ones defined in Namespace.php.
 529          *
 530          * @param $text String
 531          * @return mixed An integer if $text is a valid value otherwise false
 532          */
 533         function getLocalNsIndex( $text ) {
 534                 $lctext = $this->lc( $text );
 535                 $ids = $this->getNamespaceIds();
 536                 return isset( $ids[$lctext] ) ? $ids[$lctext] : false;
 537         }
 538
 539         /**
 540          * @return array
 541          */
 542         function getNamespaceAliases() {
 543                 if ( is_null( $this->namespaceAliases ) ) {
 544                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 545                         if ( !$aliases ) {
 546                                 $aliases = array();
 547                         } else {
 548                                 foreach ( $aliases as $name => $index ) {
 549                                         if ( $index === NS_PROJECT_TALK ) {
 550                                                 unset( $aliases[$name] );
 551                                                 $name = $this->fixVariableInNamespace( $name );
 552                                                 $aliases[$name] = $index;
 553                                         }
 554                                 }
 555                         }
 556
 557                         global $wgExtraGenderNamespaces;
 558                         $genders = $wgExtraGenderNamespaces + (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 559                         foreach ( $genders as $index => $forms ) {
 560                                 foreach ( $forms as $alias ) {
 561                                         $aliases[$alias] = $index;
 562                                 }
 563                         }
 564
 565                         $this->namespaceAliases = $aliases;
 566                 }
 567                 return $this->namespaceAliases;
 568         }
 569
 570         /**
 571          * @return array
 572          */
 573         function getNamespaceIds() {
 574                 if ( is_null( $this->mNamespaceIds ) ) {
 575                         global $wgNamespaceAliases;
 576                         # Put namespace names and aliases into a hashtable.
 577                         # If this is too slow, then we should arrange it so that it is done
 578                         # before caching. The catch is that at pre-cache time, the above
 579                         # class-specific fixup hasn't been done.
 580                         $this->mNamespaceIds = array();
 581                         foreach ( $this->getNamespaces() as $index => $name ) {
 582                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 583                         }
 584                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 585                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 586                         }
 587                         if ( $wgNamespaceAliases ) {
 588                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 589                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 590                                 }
 591                         }
 592                 }
 593                 return $this->mNamespaceIds;
 594         }
 595
 596         /**
 597          * Get a namespace key by value, case insensitive.  Canonical namespace
 598          * names override custom ones defined for the current language.
 599          *
 600          * @param $text String
 601          * @return mixed An integer if $text is a valid value otherwise false
 602          */
 603         function getNsIndex( $text ) {
 604                 $lctext = $this->lc( $text );
 605                 $ns = MWNamespace::getCanonicalIndex( $lctext );
 606                 if ( $ns !== null ) {
 607                         return $ns;
 608                 }
 609                 $ids = $this->getNamespaceIds();
 610                 return isset( $ids[$lctext] ) ? $ids[$lctext] : false;
 611         }
 612
 613         /**
 614          * short names for language variants used for language conversion links.
 615          *
 616          * @param $code String
 617          * @param $usemsg bool Use the "variantname-xyz" message if it exists
 618          * @return string
 619          */
 620         function getVariantname( $code, $usemsg = true ) {
 621                 $msg = "variantname-$code";
 622                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 623                         return $this->getMessageFromDB( $msg );
 624                 }
 625                 $name = self::fetchLanguageName( $code );
 626                 if ( $name ) {
 627                         return $name; # if it's defined as a language name, show that
 628                 } else {
 629                         # otherwise, output the language code
 630                         return $code;
 631                 }
 632         }
 633
 634         /**
 635          * @param $name string
 636          * @return string
 637          */
 638         function specialPage( $name ) {
 639                 $aliases = $this->getSpecialPageAliases();
 640                 if ( isset( $aliases[$name][0] ) ) {
 641                         $name = $aliases[$name][0];
 642                 }
 643                 return $this->getNsText( NS_SPECIAL ) . ':' . $name;
 644         }
 645
 646         /**
 647          * @return array
 648          */
 649         function getQuickbarSettings() {
 650                 return array(
 651                         $this->getMessage( 'qbsettings-none' ),
 652                         $this->getMessage( 'qbsettings-fixedleft' ),
 653                         $this->getMessage( 'qbsettings-fixedright' ),
 654                         $this->getMessage( 'qbsettings-floatingleft' ),
 655                         $this->getMessage( 'qbsettings-floatingright' ),
 656                         $this->getMessage( 'qbsettings-directionality' )
 657                 );
 658         }
 659
 660         /**
 661          * @return array
 662          */
 663         function getDatePreferences() {
 664                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 665         }
 666
 667         /**
 668          * @return array
 669          */
 670         function getDateFormats() {
 671                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 672         }
 673
 674         /**
 675          * @return array|string
 676          */
 677         function getDefaultDateFormat() {
 678                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 679                 if ( $df === 'dmy or mdy' ) {
 680                         global $wgAmericanDates;
 681                         return $wgAmericanDates ? 'mdy' : 'dmy';
 682                 } else {
 683                         return $df;
 684                 }
 685         }
 686
 687         /**
 688          * @return array
 689          */
 690         function getDatePreferenceMigrationMap() {
 691                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 692         }
 693
 694         /**
 695          * @param  $image
 696          * @return array|null
 697          */
 698         function getImageFile( $image ) {
 699                 return self::$dataCache->getSubitem( $this->mCode, 'imageFiles', $image );
 700         }
 701
 702         /**
 703          * @return array
 704          */
 705         function getExtraUserToggles() {
 706                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 707         }
 708
 709         /**
 710          * @param  $tog
 711          * @return string
 712          */
 713         function getUserToggle( $tog ) {
 714                 return $this->getMessageFromDB( "tog-$tog" );
 715         }
 716
 717         /**
 718          * Get native language names, indexed by code.
 719          * Only those defined in MediaWiki, no other data like CLDR.
 720          * If $customisedOnly is true, only returns codes with a messages file
 721          *
 722          * @param $customisedOnly bool
 723          *
 724          * @return array
 725          * @deprecated in 1.20, use fetchLanguageNames()
 726          */
 727         public static function getLanguageNames( $customisedOnly = false ) {
 728                 return self::fetchLanguageNames( null, $customisedOnly ? 'mwfile' : 'mw' );
 729         }
 730
 731         /**
 732          * Get translated language names. This is done on best effort and
 733          * by default this is exactly the same as Language::getLanguageNames.
 734          * The CLDR extension provides translated names.
 735          * @param $code String Language code.
 736          * @return Array language code => language name
 737          * @since 1.18.0
 738          * @deprecated in 1.20, use fetchLanguageNames()
 739          */
 740         public static function getTranslatedLanguageNames( $code ) {
 741                 return self::fetchLanguageNames( $code, 'all' );
 742         }
 743
 744         /**
 745          * Get an array of language names, indexed by code.
 746          * @param $inLanguage null|string: Code of language in which to return the names
 747          *              Use null for autonyms (native names)
 748          * @param $include string:
 749          *              'all' all available languages
 750          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 751          *              'mwfile' only if the language is in 'mw' *and* has a message file
 752          * @return array: language code => language name
 753          * @since 1.20
 754          */
 755         public static function fetchLanguageNames( $inLanguage = null, $include = 'mw' ) {
 756                 global $wgExtraLanguageNames;
 757                 static $coreLanguageNames;
 758
 759                 if ( $coreLanguageNames === null ) {
 760                         include( MWInit::compiledPath( 'languages/Names.php' ) );
 761                 }
 762
 763                 $names = array();
 764
 765                 if ( $inLanguage ) {
 766                         # TODO: also include when $inLanguage is null, when this code is more efficient
 767                         wfRunHooks( 'LanguageGetTranslatedLanguageNames', array( &$names, $inLanguage ) );
 768                 }
 769
 770                 $mwNames = $wgExtraLanguageNames + $coreLanguageNames;
 771                 foreach ( $mwNames as $mwCode => $mwName ) {
 772                         # - Prefer own MediaWiki native name when not using the hook
 773                         # - For other names just add if not added through the hook
 774                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 775                                 $names[$mwCode] = $mwName;
 776                         }
 777                 }
 778
 779                 if ( $include === 'all' ) {
 780                         return $names;
 781                 }
 782
 783                 $returnMw = array();
 784                 $coreCodes = array_keys( $mwNames );
 785                 foreach ( $coreCodes as $coreCode ) {
 786                         $returnMw[$coreCode] = $names[$coreCode];
 787                 }
 788
 789                 if ( $include === 'mwfile' ) {
 790                         $namesMwFile = array();
 791                         # We do this using a foreach over the codes instead of a directory
 792                         # loop so that messages files in extensions will work correctly.
 793                         foreach ( $returnMw as $code => $value ) {
 794                                 if ( is_readable( self::getMessagesFileName( $code ) ) ) {
 795                                         $namesMwFile[$code] = $names[$code];
 796                                 }
 797                         }
 798                         return $namesMwFile;
 799                 }
 800                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 801                 return $returnMw;
 802         }
 803
 804         /**
 805          * @param $code string: The code of the language for which to get the name
 806          * @param $inLanguage null|string: Code of language in which to return the name (null for autonyms)
 807          * @param $include string: 'all', 'mw' or 'mwfile'; see fetchLanguageNames()
 808          * @return string: Language name or empty
 809          * @since 1.20
 810          */
 811         public static function fetchLanguageName( $code, $inLanguage = null, $include = 'all' ) {
 812                 $array = self::fetchLanguageNames( $inLanguage, $include );
 813                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 814         }
 815
 816         /**
 817          * Get a message from the MediaWiki namespace.
 818          *
 819          * @param $msg String: message name
 820          * @return string
 821          */
 822         function getMessageFromDB( $msg ) {
 823                 return wfMessage( $msg )->inLanguage( $this )->text();
 824         }
 825
 826         /**
 827          * Get the native language name of $code.
 828          * Only if defined in MediaWiki, no other data like CLDR.
 829          * @param $code string
 830          * @return string
 831          * @deprecated in 1.20, use fetchLanguageName()
 832          */
 833         function getLanguageName( $code ) {
 834                 return self::fetchLanguageName( $code );
 835         }
 836
 837         /**
 838          * @param $key string
 839          * @return string
 840          */
 841         function getMonthName( $key ) {
 842                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 843         }
 844
 845         /**
 846          * @return array
 847          */
 848         function getMonthNamesArray() {
 849                 $monthNames = array( '' );
 850                 for ( $i = 1; $i < 13; $i++ ) {
 851                         $monthNames[] = $this->getMonthName( $i );
 852                 }
 853                 return $monthNames;
 854         }
 855
 856         /**
 857          * @param $key string
 858          * @return string
 859          */
 860         function getMonthNameGen( $key ) {
 861                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 862         }
 863
 864         /**
 865          * @param $key string
 866          * @return string
 867          */
 868         function getMonthAbbreviation( $key ) {
 869                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 870         }
 871
 872         /**
 873          * @return array
 874          */
 875         function getMonthAbbreviationsArray() {
 876                 $monthNames = array( '' );
 877                 for ( $i = 1; $i < 13; $i++ ) {
 878                         $monthNames[] = $this->getMonthAbbreviation( $i );
 879                 }
 880                 return $monthNames;
 881         }
 882
 883         /**
 884          * @param $key string
 885          * @return string
 886          */
 887         function getWeekdayName( $key ) {
 888                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
 889         }
 890
 891         /**
 892          * @param $key string
 893          * @return string
 894          */
 895         function getWeekdayAbbreviation( $key ) {
 896                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
 897         }
 898
 899         /**
 900          * @param $key string
 901          * @return string
 902          */
 903         function getIranianCalendarMonthName( $key ) {
 904                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
 905         }
 906
 907         /**
 908          * @param $key string
 909          * @return string
 910          */
 911         function getHebrewCalendarMonthName( $key ) {
 912                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
 913         }
 914
 915         /**
 916          * @param $key string
 917          * @return string
 918          */
 919         function getHebrewCalendarMonthNameGen( $key ) {
 920                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
 921         }
 922
 923         /**
 924          * @param $key string
 925          * @return string
 926          */
 927         function getHijriCalendarMonthName( $key ) {
 928                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
 929         }
 930
 931         /**
 932          * This is a workalike of PHP's date() function, but with better
 933          * internationalisation, a reduced set of format characters, and a better
 934          * escaping format.
 935          *
 936          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrU. See the
 937          * PHP manual for definitions. There are a number of extensions, which
 938          * start with "x":
 939          *
 940          *    xn   Do not translate digits of the next numeric format character
 941          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
 942          *    xr   Use roman numerals for the next numeric format character
 943          *    xh   Use hebrew numerals for the next numeric format character
 944          *    xx   Literal x
 945          *    xg   Genitive month name
 946          *
 947          *    xij  j (day number) in Iranian calendar
 948          *    xiF  F (month name) in Iranian calendar
 949          *    xin  n (month number) in Iranian calendar
 950          *    xiy  y (two digit year) in Iranian calendar
 951          *    xiY  Y (full year) in Iranian calendar
 952          *
 953          *    xjj  j (day number) in Hebrew calendar
 954          *    xjF  F (month name) in Hebrew calendar
 955          *    xjt  t (days in month) in Hebrew calendar
 956          *    xjx  xg (genitive month name) in Hebrew calendar
 957          *    xjn  n (month number) in Hebrew calendar
 958          *    xjY  Y (full year) in Hebrew calendar
 959          *
 960          *    xmj  j (day number) in Hijri calendar
 961          *    xmF  F (month name) in Hijri calendar
 962          *    xmn  n (month number) in Hijri calendar
 963          *    xmY  Y (full year) in Hijri calendar
 964          *
 965          *    xkY  Y (full year) in Thai solar calendar. Months and days are
 966          *                       identical to the Gregorian calendar
 967          *    xoY  Y (full year) in Minguo calendar or Juche year.
 968          *                       Months and days are identical to the
 969          *                       Gregorian calendar
 970          *    xtY  Y (full year) in Japanese nengo. Months and days are
 971          *                       identical to the Gregorian calendar
 972          *
 973          * Characters enclosed in double quotes will be considered literal (with
 974          * the quotes themselves removed). Unmatched quotes will be considered
 975          * literal quotes. Example:
 976          *
 977          * "The month is" F       => The month is January
 978          * i's"                   => 20'11"
 979          *
 980          * Backslash escaping is also supported.
 981          *
 982          * Input timestamp is assumed to be pre-normalized to the desired local
 983          * time zone, if any.
 984          *
 985          * @param $format String
 986          * @param $ts String: 14-character timestamp
 987          *      YYYYMMDDHHMMSS
 988          *      01234567890123
 989          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
 990          *
 991          * @return string
 992          */
 993         function sprintfDate( $format, $ts ) {
 994                 $s = '';
 995                 $raw = false;
 996                 $roman = false;
 997                 $hebrewNum = false;
 998                 $unix = false;
 999                 $rawToggle = false;
1000                 $iranian = false;
1001                 $hebrew = false;
1002                 $hijri = false;
1003                 $thai = false;
1004                 $minguo = false;
1005                 $tenno = false;
1006                 for ( $p = 0; $p < strlen( $format ); $p++ ) {
1007                         $num = false;
1008                         $code = $format[$p];
1009                         if ( $code == 'x' && $p < strlen( $format ) - 1 ) {
1010                                 $code .= $format[++$p];
1011                         }
1012
1013                         if ( ( $code === 'xi' || $code == 'xj' || $code == 'xk' || $code == 'xm' || $code == 'xo' || $code == 'xt' ) && $p < strlen( $format ) - 1 ) {
1014                                 $code .= $format[++$p];
1015                         }
1016
1017                         switch ( $code ) {
1018                                 case 'xx':
1019                                         $s .= 'x';
1020                                         break;
1021                                 case 'xn':
1022                                         $raw = true;
1023                                         break;
1024                                 case 'xN':
1025                                         $rawToggle = !$rawToggle;
1026                                         break;
1027                                 case 'xr':
1028                                         $roman = true;
1029                                         break;
1030                                 case 'xh':
1031                                         $hebrewNum = true;
1032                                         break;
1033                                 case 'xg':
1034                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1035                                         break;
1036                                 case 'xjx':
1037                                         if ( !$hebrew ) $hebrew = self::tsToHebrew( $ts );
1038                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1039                                         break;
1040                                 case 'd':
1041                                         $num = substr( $ts, 6, 2 );
1042                                         break;
1043                                 case 'D':
1044                                         if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
1045                                         $s .= $this->getWeekdayAbbreviation( gmdate( 'w', $unix ) + 1 );
1046                                         break;
1047                                 case 'j':
1048                                         $num = intval( substr( $ts, 6, 2 ) );
1049                                         break;
1050                                 case 'xij':
1051                                         if ( !$iranian ) {
1052                                                 $iranian = self::tsToIranian( $ts );
1053                                         }
1054                                         $num = $iranian[2];
1055                                         break;
1056                                 case 'xmj':
1057                                         if ( !$hijri ) {
1058                                                 $hijri = self::tsToHijri( $ts );
1059                                         }
1060                                         $num = $hijri[2];
1061                                         break;
1062                                 case 'xjj':
1063                                         if ( !$hebrew ) {
1064                                                 $hebrew = self::tsToHebrew( $ts );
1065                                         }
1066                                         $num = $hebrew[2];
1067                                         break;
1068                                 case 'l':
1069                                         if ( !$unix ) {
1070                                                 $unix = wfTimestamp( TS_UNIX, $ts );
1071                                         }
1072                                         $s .= $this->getWeekdayName( gmdate( 'w', $unix ) + 1 );
1073                                         break;
1074                                 case 'N':
1075                                         if ( !$unix ) {
1076                                                 $unix = wfTimestamp( TS_UNIX, $ts );
1077                                         }
1078                                         $w = gmdate( 'w', $unix );
1079                                         $num = $w ? $w : 7;
1080                                         break;
1081                                 case 'w':
1082                                         if ( !$unix ) {
1083                                                 $unix = wfTimestamp( TS_UNIX, $ts );
1084                                         }
1085                                         $num = gmdate( 'w', $unix );
1086                                         break;
1087                                 case 'z':
1088                                         if ( !$unix ) {
1089                                                 $unix = wfTimestamp( TS_UNIX, $ts );
1090                                         }
1091                                         $num = gmdate( 'z', $unix );
1092                                         break;
1093                                 case 'W':
1094                                         if ( !$unix ) {
1095                                                 $unix = wfTimestamp( TS_UNIX, $ts );
1096                                         }
1097                                         $num = gmdate( 'W', $unix );
1098                                         break;
1099                                 case 'F':
1100                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1101                                         break;
1102                                 case 'xiF':
1103                                         if ( !$iranian ) {
1104                                                 $iranian = self::tsToIranian( $ts );
1105                                         }
1106                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1107                                         break;
1108                                 case 'xmF':
1109                                         if ( !$hijri ) {
1110                                                 $hijri = self::tsToHijri( $ts );
1111                                         }
1112                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1113                                         break;
1114                                 case 'xjF':
1115                                         if ( !$hebrew ) {
1116                                                 $hebrew = self::tsToHebrew( $ts );
1117                                         }
1118                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1119                                         break;
1120                                 case 'm':
1121                                         $num = substr( $ts, 4, 2 );
1122                                         break;
1123                                 case 'M':
1124                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1125                                         break;
1126                                 case 'n':
1127                                         $num = intval( substr( $ts, 4, 2 ) );
1128                                         break;
1129                                 case 'xin':
1130                                         if ( !$iranian ) {
1131                                                 $iranian = self::tsToIranian( $ts );
1132                                         }
1133                                         $num = $iranian[1];
1134                                         break;
1135                                 case 'xmn':
1136                                         if ( !$hijri ) {
1137                                                 $hijri = self::tsToHijri ( $ts );
1138                                         }
1139                                         $num = $hijri[1];
1140                                         break;
1141                                 case 'xjn':
1142                                         if ( !$hebrew ) {
1143                                                 $hebrew = self::tsToHebrew( $ts );
1144                                         }
1145                                         $num = $hebrew[1];
1146                                         break;
1147                                 case 't':
1148                                         if ( !$unix ) {
1149                                                 $unix = wfTimestamp( TS_UNIX, $ts );
1150                                         }
1151                                         $num = gmdate( 't', $unix );
1152                                         break;
1153                                 case 'xjt':
1154                                         if ( !$hebrew ) {
1155                                                 $hebrew = self::tsToHebrew( $ts );
1156                                         }
1157                                         $num = $hebrew[3];
1158                                         break;
1159                                 case 'L':
1160                                         if ( !$unix ) {
1161                                                 $unix = wfTimestamp( TS_UNIX, $ts );
1162                                         }
1163                                         $num = gmdate( 'L', $unix );
1164                                         break;
1165                                 case 'o':
1166                                         if ( !$unix ) {
1167                                                 $unix = wfTimestamp( TS_UNIX, $ts );
1168                                         }
1169                                         $num = gmdate( 'o', $unix );
1170                                         break;
1171                                 case 'Y':
1172                                         $num = substr( $ts, 0, 4 );
1173                                         break;
1174                                 case 'xiY':
1175                                         if ( !$iranian ) {
1176                                                 $iranian = self::tsToIranian( $ts );
1177                                         }
1178                                         $num = $iranian[0];
1179                                         break;
1180                                 case 'xmY':
1181                                         if ( !$hijri ) {
1182                                                 $hijri = self::tsToHijri( $ts );
1183                                         }
1184                                         $num = $hijri[0];
1185                                         break;
1186                                 case 'xjY':
1187                                         if ( !$hebrew ) {
1188                                                 $hebrew = self::tsToHebrew( $ts );
1189                                         }
1190                                         $num = $hebrew[0];
1191                                         break;
1192                                 case 'xkY':
1193                                         if ( !$thai ) {
1194                                                 $thai = self::tsToYear( $ts, 'thai' );
1195                                         }
1196                                         $num = $thai[0];
1197                                         break;
1198                                 case 'xoY':
1199                                         if ( !$minguo ) {
1200                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1201                                         }
1202                                         $num = $minguo[0];
1203                                         break;
1204                                 case 'xtY':
1205                                         if ( !$tenno ) {
1206                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1207                                         }
1208                                         $num = $tenno[0];
1209                                         break;
1210                                 case 'y':
1211                                         $num = substr( $ts, 2, 2 );
1212                                         break;
1213                                 case 'xiy':
1214                                         if ( !$iranian ) {
1215                                                 $iranian = self::tsToIranian( $ts );
1216                                         }
1217                                         $num = substr( $iranian[0], -2 );
1218                                         break;
1219                                 case 'a':
1220                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1221                                         break;
1222                                 case 'A':
1223                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1224                                         break;
1225                                 case 'g':
1226                                         $h = substr( $ts, 8, 2 );
1227                                         $num = $h % 12 ? $h % 12 : 12;
1228                                         break;
1229                                 case 'G':
1230                                         $num = intval( substr( $ts, 8, 2 ) );
1231                                         break;
1232                                 case 'h':
1233                                         $h = substr( $ts, 8, 2 );
1234                                         $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
1235                                         break;
1236                                 case 'H':
1237                                         $num = substr( $ts, 8, 2 );
1238                                         break;
1239                                 case 'i':
1240                                         $num = substr( $ts, 10, 2 );
1241                                         break;
1242                                 case 's':
1243                                         $num = substr( $ts, 12, 2 );
1244                                         break;
1245                                 case 'c':
1246                                         if ( !$unix ) {
1247                                                 $unix = wfTimestamp( TS_UNIX, $ts );
1248                                         }
1249                                         $s .= gmdate( 'c', $unix );
1250                                         break;
1251                                 case 'r':
1252                                         if ( !$unix ) {
1253                                                 $unix = wfTimestamp( TS_UNIX, $ts );
1254                                         }
1255                                         $s .= gmdate( 'r', $unix );
1256                                         break;
1257                                 case 'U':
1258                                         if ( !$unix ) {
1259                                                 $unix = wfTimestamp( TS_UNIX, $ts );
1260                                         }
1261                                         $num = $unix;
1262                                         break;
1263                                 case '\\':
1264                                         # Backslash escaping
1265                                         if ( $p < strlen( $format ) - 1 ) {
1266                                                 $s .= $format[++$p];
1267                                         } else {
1268                                                 $s .= '\\';
1269                                         }
1270                                         break;
1271                                 case '"':
1272                                         # Quoted literal
1273                                         if ( $p < strlen( $format ) - 1 ) {
1274                                                 $endQuote = strpos( $format, '"', $p + 1 );
1275                                                 if ( $endQuote === false ) {
1276                                                         # No terminating quote, assume literal "
1277                                                         $s .= '"';
1278                                                 } else {
1279                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1280                                                         $p = $endQuote;
1281                                                 }
1282                                         } else {
1283                                                 # Quote at end of string, assume literal "
1284                                                 $s .= '"';
1285                                         }
1286                                         break;
1287                                 default:
1288                                         $s .= $format[$p];
1289                         }
1290                         if ( $num !== false ) {
1291                                 if ( $rawToggle || $raw ) {
1292                                         $s .= $num;
1293                                         $raw = false;
1294                                 } elseif ( $roman ) {
1295                                         $s .= Language::romanNumeral( $num );
1296                                         $roman = false;
1297                                 } elseif ( $hebrewNum ) {
1298                                         $s .= self::hebrewNumeral( $num );
1299                                         $hebrewNum = false;
1300                                 } else {
1301                                         $s .= $this->formatNum( $num, true );
1302                                 }
1303                         }
1304                 }
1305                 return $s;
1306         }
1307
1308         private static $GREG_DAYS = array( 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 );
1309         private static $IRANIAN_DAYS = array( 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 );
1310
1311         /**
1312          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1313          * Gregorian dates to Iranian dates. Originally written in C, it
1314          * is released under the terms of GNU Lesser General Public
1315          * License. Conversion to PHP was performed by Niklas Laxström.
1316          *
1317          * Link: http://www.farsiweb.info/jalali/jalali.c
1318          *
1319          * @param $ts string
1320          *
1321          * @return string
1322          */
1323         private static function tsToIranian( $ts ) {
1324                 $gy = substr( $ts, 0, 4 ) -1600;
1325                 $gm = substr( $ts, 4, 2 ) -1;
1326                 $gd = substr( $ts, 6, 2 ) -1;
1327
1328                 # Days passed from the beginning (including leap years)
1329                 $gDayNo = 365 * $gy
1330                         + floor( ( $gy + 3 ) / 4 )
1331                         - floor( ( $gy + 99 ) / 100 )
1332                         + floor( ( $gy + 399 ) / 400 );
1333
1334                 // Add days of the past months of this year
1335                 for ( $i = 0; $i < $gm; $i++ ) {
1336                         $gDayNo += self::$GREG_DAYS[$i];
1337                 }
1338
1339                 // Leap years
1340                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1341                         $gDayNo++;
1342                 }
1343
1344                 // Days passed in current month
1345                 $gDayNo += (int)$gd;
1346
1347                 $jDayNo = $gDayNo - 79;
1348
1349                 $jNp = floor( $jDayNo / 12053 );
1350                 $jDayNo %= 12053;
1351
1352                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1353                 $jDayNo %= 1461;
1354
1355                 if ( $jDayNo >= 366 ) {
1356                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1357                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1358                 }
1359
1360                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1361                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1362                 }
1363
1364                 $jm = $i + 1;
1365                 $jd = $jDayNo + 1;
1366
1367                 return array( $jy, $jm, $jd );
1368         }
1369
1370         /**
1371          * Converting Gregorian dates to Hijri dates.
1372          *
1373          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1374          *
1375          * @see http://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1376          *
1377          * @param $ts string
1378          *
1379          * @return string
1380          */
1381         private static function tsToHijri( $ts ) {
1382                 $year = substr( $ts, 0, 4 );
1383                 $month = substr( $ts, 4, 2 );
1384                 $day = substr( $ts, 6, 2 );
1385
1386                 $zyr = $year;
1387                 $zd = $day;
1388                 $zm = $month;
1389                 $zy = $zyr;
1390
1391                 if (
1392                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1393                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1394                 )
1395                 {
1396                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1397                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1398                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1399                                         $zd - 32075;
1400                 } else {
1401                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1402                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1403                 }
1404
1405                 $zl = $zjd -1948440 + 10632;
1406                 $zn = (int)( ( $zl - 1 ) / 10631 );
1407                 $zl = $zl - 10631 * $zn + 354;
1408                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) + ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1409                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) - ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1410                 $zm = (int)( ( 24 * $zl ) / 709 );
1411                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1412                 $zy = 30 * $zn + $zj - 30;
1413
1414                 return array( $zy, $zm, $zd );
1415         }
1416
1417         /**
1418          * Converting Gregorian dates to Hebrew dates.
1419          *
1420          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1421          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1422          * to translate the relevant functions into PHP and release them under
1423          * GNU GPL.
1424          *
1425          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1426          * and Adar II is 14. In a non-leap year, Adar is 6.
1427          *
1428          * @param $ts string
1429          *
1430          * @return string
1431          */
1432         private static function tsToHebrew( $ts ) {
1433                 # Parse date
1434                 $year = substr( $ts, 0, 4 );
1435                 $month = substr( $ts, 4, 2 );
1436                 $day = substr( $ts, 6, 2 );
1437
1438                 # Calculate Hebrew year
1439                 $hebrewYear = $year + 3760;
1440
1441                 # Month number when September = 1, August = 12
1442                 $month += 4;
1443                 if ( $month > 12 ) {
1444                         # Next year
1445                         $month -= 12;
1446                         $year++;
1447                         $hebrewYear++;
1448                 }
1449
1450                 # Calculate day of year from 1 September
1451                 $dayOfYear = $day;
1452                 for ( $i = 1; $i < $month; $i++ ) {
1453                         if ( $i == 6 ) {
1454                                 # February
1455                                 $dayOfYear += 28;
1456                                 # Check if the year is leap
1457                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1458                                         $dayOfYear++;
1459                                 }
1460                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1461                                 $dayOfYear += 30;
1462                         } else {
1463                                 $dayOfYear += 31;
1464                         }
1465                 }
1466
1467                 # Calculate the start of the Hebrew year
1468                 $start = self::hebrewYearStart( $hebrewYear );
1469
1470                 # Calculate next year's start
1471                 if ( $dayOfYear <= $start ) {
1472                         # Day is before the start of the year - it is the previous year
1473                         # Next year's start
1474                         $nextStart = $start;
1475                         # Previous year
1476                         $year--;
1477                         $hebrewYear--;
1478                         # Add days since previous year's 1 September
1479                         $dayOfYear += 365;
1480                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1481                                 # Leap year
1482                                 $dayOfYear++;
1483                         }
1484                         # Start of the new (previous) year
1485                         $start = self::hebrewYearStart( $hebrewYear );
1486                 } else {
1487                         # Next year's start
1488                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1489                 }
1490
1491                 # Calculate Hebrew day of year
1492                 $hebrewDayOfYear = $dayOfYear - $start;
1493
1494                 # Difference between year's days
1495                 $diff = $nextStart - $start;
1496                 # Add 12 (or 13 for leap years) days to ignore the difference between
1497                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1498                 # difference is only about the year type
1499                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1500                         $diff += 13;
1501                 } else {
1502                         $diff += 12;
1503                 }
1504
1505                 # Check the year pattern, and is leap year
1506                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1507                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1508                 # and non-leap years
1509                 $yearPattern = $diff % 30;
1510                 # Check if leap year
1511                 $isLeap = $diff >= 30;
1512
1513                 # Calculate day in the month from number of day in the Hebrew year
1514                 # Don't check Adar - if the day is not in Adar, we will stop before;
1515                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1516                 $hebrewDay = $hebrewDayOfYear;
1517                 $hebrewMonth = 1;
1518                 $days = 0;
1519                 while ( $hebrewMonth <= 12 ) {
1520                         # Calculate days in this month
1521                         if ( $isLeap && $hebrewMonth == 6 ) {
1522                                 # Adar in a leap year
1523                                 if ( $isLeap ) {
1524                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1525                                         $days = 30;
1526                                         if ( $hebrewDay <= $days ) {
1527                                                 # Day in Adar I
1528                                                 $hebrewMonth = 13;
1529                                         } else {
1530                                                 # Subtract the days of Adar I
1531                                                 $hebrewDay -= $days;
1532                                                 # Try Adar II
1533                                                 $days = 29;
1534                                                 if ( $hebrewDay <= $days ) {
1535                                                         # Day in Adar II
1536                                                         $hebrewMonth = 14;
1537                                                 }
1538                                         }
1539                                 }
1540                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1541                                 # Cheshvan in a complete year (otherwise as the rule below)
1542                                 $days = 30;
1543                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1544                                 # Kislev in an incomplete year (otherwise as the rule below)
1545                                 $days = 29;
1546                         } else {
1547                                 # Odd months have 30 days, even have 29
1548                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1549                         }
1550                         if ( $hebrewDay <= $days ) {
1551                                 # In the current month
1552                                 break;
1553                         } else {
1554                                 # Subtract the days of the current month
1555                                 $hebrewDay -= $days;
1556                                 # Try in the next month
1557                                 $hebrewMonth++;
1558                         }
1559                 }
1560
1561                 return array( $hebrewYear, $hebrewMonth, $hebrewDay, $days );
1562         }
1563
1564         /**
1565          * This calculates the Hebrew year start, as days since 1 September.
1566          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1567          * Used for Hebrew date.
1568          *
1569          * @param $year int
1570          *
1571          * @return string
1572          */
1573         private static function hebrewYearStart( $year ) {
1574                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1575                 $b = intval( ( $year - 1 ) % 4 );
1576                 $m = 32.044093161144 + 1.5542417966212 * $a +  $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1577                 if ( $m < 0 ) {
1578                         $m--;
1579                 }
1580                 $Mar = intval( $m );
1581                 if ( $m < 0 ) {
1582                         $m++;
1583                 }
1584                 $m -= $Mar;
1585
1586                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1587                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1588                         $Mar++;
1589                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1590                         $Mar += 2;
1591                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1592                         $Mar++;
1593                 }
1594
1595                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1596                 return $Mar;
1597         }
1598
1599         /**
1600          * Algorithm to convert Gregorian dates to Thai solar dates,
1601          * Minguo dates or Minguo dates.
1602          *
1603          * Link: http://en.wikipedia.org/wiki/Thai_solar_calendar
1604          *       http://en.wikipedia.org/wiki/Minguo_calendar
1605          *       http://en.wikipedia.org/wiki/Japanese_era_name
1606          *
1607          * @param $ts String: 14-character timestamp
1608          * @param $cName String: calender name
1609          * @return Array: converted year, month, day
1610          */
1611         private static function tsToYear( $ts, $cName ) {
1612                 $gy = substr( $ts, 0, 4 );
1613                 $gm = substr( $ts, 4, 2 );
1614                 $gd = substr( $ts, 6, 2 );
1615
1616                 if ( !strcmp( $cName, 'thai' ) ) {
1617                         # Thai solar dates
1618                         # Add 543 years to the Gregorian calendar
1619                         # Months and days are identical
1620                         $gy_offset = $gy + 543;
1621                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1622                         # Minguo dates
1623                         # Deduct 1911 years from the Gregorian calendar
1624                         # Months and days are identical
1625                         $gy_offset = $gy - 1911;
1626                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1627                         # Nengō dates up to Meiji period
1628                         # Deduct years from the Gregorian calendar
1629                         # depending on the nengo periods
1630                         # Months and days are identical
1631                         if ( ( $gy < 1912 ) || ( ( $gy == 1912 ) && ( $gm < 7 ) ) || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) ) ) {
1632                                 # Meiji period
1633                                 $gy_gannen = $gy - 1868 + 1;
1634                                 $gy_offset = $gy_gannen;
1635                                 if ( $gy_gannen == 1 ) {
1636                                         $gy_offset = '元';
1637                                 }
1638                                 $gy_offset = '明治' . $gy_offset;
1639                         } elseif (
1640                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1641                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1642                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1643                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1644                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1645                         )
1646                         {
1647                                 # Taishō period
1648                                 $gy_gannen = $gy - 1912 + 1;
1649                                 $gy_offset = $gy_gannen;
1650                                 if ( $gy_gannen == 1 ) {
1651                                         $gy_offset = '元';
1652                                 }
1653                                 $gy_offset = '大正' . $gy_offset;
1654                         } elseif (
1655                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1656                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1657                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1658                         )
1659                         {
1660                                 # Shōwa period
1661                                 $gy_gannen = $gy - 1926 + 1;
1662                                 $gy_offset = $gy_gannen;
1663                                 if ( $gy_gannen == 1 ) {
1664                                         $gy_offset = '元';
1665                                 }
1666                                 $gy_offset = '昭和' . $gy_offset;
1667                         } else {
1668                                 # Heisei period
1669                                 $gy_gannen = $gy - 1989 + 1;
1670                                 $gy_offset = $gy_gannen;
1671                                 if ( $gy_gannen == 1 ) {
1672                                         $gy_offset = '元';
1673                                 }
1674                                 $gy_offset = '平成' . $gy_offset;
1675                         }
1676                 } else {
1677                         $gy_offset = $gy;
1678                 }
1679
1680                 return array( $gy_offset, $gm, $gd );
1681         }
1682
1683         /**
1684          * Roman number formatting up to 10000
1685          *
1686          * @param $num int
1687          *
1688          * @return string
1689          */
1690         static function romanNumeral( $num ) {
1691                 static $table = array(
1692                         array( '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ),
1693                         array( '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ),
1694                         array( '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ),
1695                         array( '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM', 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' )
1696                 );
1697
1698                 $num = intval( $num );
1699                 if ( $num > 10000 || $num <= 0 ) {
1700                         return $num;
1701                 }
1702
1703                 $s = '';
1704                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1705                         if ( $num >= $pow10 ) {
1706                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
1707                         }
1708                         $num = $num % $pow10;
1709                 }
1710                 return $s;
1711         }
1712
1713         /**
1714          * Hebrew Gematria number formatting up to 9999
1715          *
1716          * @param $num int
1717          *
1718          * @return string
1719          */
1720         static function hebrewNumeral( $num ) {
1721                 static $table = array(
1722                         array( '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ),
1723                         array( '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ),
1724                         array( '', 'ק', 'ר', 'ש', 'ת', 'תק', 'תר', 'תש', 'תת', 'תתק', 'תתר' ),
1725                         array( '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' )
1726                 );
1727
1728                 $num = intval( $num );
1729                 if ( $num > 9999 || $num <= 0 ) {
1730                         return $num;
1731                 }
1732
1733                 $s = '';
1734                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1735                         if ( $num >= $pow10 ) {
1736                                 if ( $num == 15 || $num == 16 ) {
1737                                         $s .= $table[0][9] . $table[0][$num - 9];
1738                                         $num = 0;
1739                                 } else {
1740                                         $s .= $table[$i][intval( ( $num / $pow10 ) )];
1741                                         if ( $pow10 == 1000 ) {
1742                                                 $s .= "'";
1743                                         }
1744                                 }
1745                         }
1746                         $num = $num % $pow10;
1747                 }
1748                 if ( strlen( $s ) == 2 ) {
1749                         $str = $s . "'";
1750                 } else  {
1751                         $str = substr( $s, 0, strlen( $s ) - 2 ) . '"';
1752                         $str .= substr( $s, strlen( $s ) - 2, 2 );
1753                 }
1754                 $start = substr( $str, 0, strlen( $str ) - 2 );
1755                 $end = substr( $str, strlen( $str ) - 2 );
1756                 switch( $end ) {
1757                         case 'כ':
1758                                 $str = $start . 'ך';
1759                                 break;
1760                         case 'מ':
1761                                 $str = $start . 'ם';
1762                                 break;
1763                         case 'נ':
1764                                 $str = $start . 'ן';
1765                                 break;
1766                         case 'פ':
1767                                 $str = $start . 'ף';
1768                                 break;
1769                         case 'צ':
1770                                 $str = $start . 'ץ';
1771                                 break;
1772                 }
1773                 return $str;
1774         }
1775
1776         /**
1777          * Used by date() and time() to adjust the time output.
1778          *
1779          * @param $ts Int the time in date('YmdHis') format
1780          * @param $tz Mixed: adjust the time by this amount (default false, mean we
1781          *            get user timecorrection setting)
1782          * @return int
1783          */
1784         function userAdjust( $ts, $tz = false ) {
1785                 global $wgUser, $wgLocalTZoffset;
1786
1787                 if ( $tz === false ) {
1788                         $tz = $wgUser->getOption( 'timecorrection' );
1789                 }
1790
1791                 $data = explode( '|', $tz, 3 );
1792
1793                 if ( $data[0] == 'ZoneInfo' ) {
1794                         wfSuppressWarnings();
1795                         $userTZ = timezone_open( $data[2] );
1796                         wfRestoreWarnings();
1797                         if ( $userTZ !== false ) {
1798                                 $date = date_create( $ts, timezone_open( 'UTC' ) );
1799                                 date_timezone_set( $date, $userTZ );
1800                                 $date = date_format( $date, 'YmdHis' );
1801                                 return $date;
1802                         }
1803                         # Unrecognized timezone, default to 'Offset' with the stored offset.
1804                         $data[0] = 'Offset';
1805                 }
1806
1807                 $minDiff = 0;
1808                 if ( $data[0] == 'System' || $tz == '' ) {
1809                         #  Global offset in minutes.
1810                         if ( isset( $wgLocalTZoffset ) ) {
1811                                 $minDiff = $wgLocalTZoffset;
1812                         }
1813                 } elseif ( $data[0] == 'Offset' ) {
1814                         $minDiff = intval( $data[1] );
1815                 } else {
1816                         $data = explode( ':', $tz );
1817                         if ( count( $data ) == 2 ) {
1818                                 $data[0] = intval( $data[0] );
1819                                 $data[1] = intval( $data[1] );
1820                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
1821                                 if ( $data[0] < 0 ) {
1822                                         $minDiff = -$minDiff;
1823                                 }
1824                         } else {
1825                                 $minDiff = intval( $data[0] ) * 60;
1826                         }
1827                 }
1828
1829                 # No difference ? Return time unchanged
1830                 if ( 0 == $minDiff ) {
1831                         return $ts;
1832                 }
1833
1834                 wfSuppressWarnings(); // E_STRICT system time bitching
1835                 # Generate an adjusted date; take advantage of the fact that mktime
1836                 # will normalize out-of-range values so we don't have to split $minDiff
1837                 # into hours and minutes.
1838                 $t = mktime( (
1839                   (int)substr( $ts, 8, 2 ) ), # Hours
1840                   (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
1841                   (int)substr( $ts, 12, 2 ), # Seconds
1842                   (int)substr( $ts, 4, 2 ), # Month
1843                   (int)substr( $ts, 6, 2 ), # Day
1844                   (int)substr( $ts, 0, 4 ) ); # Year
1845
1846                 $date = date( 'YmdHis', $t );
1847                 wfRestoreWarnings();
1848
1849                 return $date;
1850         }
1851
1852         /**
1853          * This is meant to be used by time(), date(), and timeanddate() to get
1854          * the date preference they're supposed to use, it should be used in
1855          * all children.
1856          *
1857          *<code>
1858          * function timeanddate([...], $format = true) {
1859          *      $datePreference = $this->dateFormat($format);
1860          * [...]
1861          * }
1862          *</code>
1863          *
1864          * @param $usePrefs Mixed: if true, the user's preference is used
1865          *                         if false, the site/language default is used
1866          *                         if int/string, assumed to be a format.
1867          * @return string
1868          */
1869         function dateFormat( $usePrefs = true ) {
1870                 global $wgUser;
1871
1872                 if ( is_bool( $usePrefs ) ) {
1873                         if ( $usePrefs ) {
1874                                 $datePreference = $wgUser->getDatePreference();
1875                         } else {
1876                                 $datePreference = (string)User::getDefaultOption( 'date' );
1877                         }
1878                 } else {
1879                         $datePreference = (string)$usePrefs;
1880                 }
1881
1882                 // return int
1883                 if ( $datePreference == '' ) {
1884                         return 'default';
1885                 }
1886
1887                 return $datePreference;
1888         }
1889
1890         /**
1891          * Get a format string for a given type and preference
1892          * @param $type string May be date, time or both
1893          * @param $pref string The format name as it appears in Messages*.php
1894          *
1895          * @return string
1896          */
1897         function getDateFormatString( $type, $pref ) {
1898                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
1899                         if ( $pref == 'default' ) {
1900                                 $pref = $this->getDefaultDateFormat();
1901                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
1902                         } else {
1903                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
1904                                 if ( is_null( $df ) ) {
1905                                         $pref = $this->getDefaultDateFormat();
1906                                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
1907                                 }
1908                         }
1909                         $this->dateFormatStrings[$type][$pref] = $df;
1910                 }
1911                 return $this->dateFormatStrings[$type][$pref];
1912         }
1913
1914         /**
1915          * @param $ts Mixed: the time format which needs to be turned into a
1916          *            date('YmdHis') format with wfTimestamp(TS_MW,$ts)
1917          * @param $adj Bool: whether to adjust the time output according to the
1918          *             user configured offset ($timecorrection)
1919          * @param $format Mixed: true to use user's date format preference
1920          * @param $timecorrection String|bool the time offset as returned by
1921          *                        validateTimeZone() in Special:Preferences
1922          * @return string
1923          */
1924         function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
1925                 $ts = wfTimestamp( TS_MW, $ts );
1926                 if ( $adj ) {
1927                         $ts = $this->userAdjust( $ts, $timecorrection );
1928                 }
1929                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
1930                 return $this->sprintfDate( $df, $ts );
1931         }
1932
1933         /**
1934          * @param $ts Mixed: the time format which needs to be turned into a
1935          *            date('YmdHis') format with wfTimestamp(TS_MW,$ts)
1936          * @param $adj Bool: whether to adjust the time output according to the
1937          *             user configured offset ($timecorrection)
1938          * @param $format Mixed: true to use user's date format preference
1939          * @param $timecorrection String|bool the time offset as returned by
1940          *                        validateTimeZone() in Special:Preferences
1941          * @return string
1942          */
1943         function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
1944                 $ts = wfTimestamp( TS_MW, $ts );
1945                 if ( $adj ) {
1946                         $ts = $this->userAdjust( $ts, $timecorrection );
1947                 }
1948                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
1949                 return $this->sprintfDate( $df, $ts );
1950         }
1951
1952         /**
1953          * @param $ts Mixed: the time format which needs to be turned into a
1954          *            date('YmdHis') format with wfTimestamp(TS_MW,$ts)
1955          * @param $adj Bool: whether to adjust the time output according to the
1956          *             user configured offset ($timecorrection)
1957          * @param $format Mixed: what format to return, if it's false output the
1958          *                default one (default true)
1959          * @param $timecorrection String|bool the time offset as returned by
1960          *                        validateTimeZone() in Special:Preferences
1961          * @return string
1962          */
1963         function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
1964                 $ts = wfTimestamp( TS_MW, $ts );
1965                 if ( $adj ) {
1966                         $ts = $this->userAdjust( $ts, $timecorrection );
1967                 }
1968                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
1969                 return $this->sprintfDate( $df, $ts );
1970         }
1971
1972         /**
1973          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
1974          *
1975          * @since 1.20
1976          *
1977          * @param integer $seconds The amount of seconds.
1978          * @param array $chosenIntervals The intervals to enable.
1979          *
1980          * @return string
1981          */
1982         public function formatDuration( $seconds, array $chosenIntervals = array() ) {
1983                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
1984
1985                 $segments = array();
1986
1987                 foreach ( $intervals as $intervalName => $intervalValue ) {
1988                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
1989                         $segments[] = $message->inLanguage( $this )->escaped();
1990                 }
1991
1992                 return $this->listToText( $segments );
1993         }
1994
1995         /**
1996          * Takes a number of seconds and returns an array with a set of corresponding intervals.
1997          * For example 65 will be turned into array( minutes => 1, seconds => 5 ).
1998          *
1999          * @since 1.20
2000          *
2001          * @param integer $seconds The amount of seconds.
2002          * @param array $chosenIntervals The intervals to enable.
2003          *
2004          * @return array
2005          */
2006         public function getDurationIntervals( $seconds, array $chosenIntervals = array() ) {
2007                 if ( empty( $chosenIntervals ) ) {
2008                         $chosenIntervals = array( 'millennia', 'centuries', 'decades', 'years', 'days', 'hours', 'minutes', 'seconds' );
2009                 }
2010
2011                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2012                 $sortedNames = array_keys( $intervals );
2013                 $smallestInterval = array_pop( $sortedNames );
2014
2015                 $segments = array();
2016
2017                 foreach ( $intervals as $name => $length ) {
2018                         $value = floor( $seconds / $length );
2019
2020                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2021                                 $seconds -= $value * $length;
2022                                 $segments[$name] = $value;
2023                         }
2024                 }
2025
2026                 return $segments;
2027         }
2028
2029         /**
2030          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2031          *
2032          * @param $type String: can be 'date', 'time' or 'both'
2033          * @param $ts Mixed: the time format which needs to be turned into a
2034          *            date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2035          * @param $user User object used to get preferences for timezone and format
2036          * @param $options Array, can contain the following keys:
2037          *        - 'timecorrection': time correction, can have the following values:
2038          *             - true: use user's preference
2039          *             - false: don't use time correction
2040          *             - integer: value of time correction in minutes
2041          *        - 'format': format to use, can have the following values:
2042          *             - true: use user's preference
2043          *             - false: use default preference
2044          *             - string: format to use
2045          * @since 1.19
2046          * @return String
2047          */
2048         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2049                 $ts = wfTimestamp( TS_MW, $ts );
2050                 $options += array( 'timecorrection' => true, 'format' => true );
2051                 if ( $options['timecorrection'] !== false ) {
2052                         if ( $options['timecorrection'] === true ) {
2053                                 $offset = $user->getOption( 'timecorrection' );
2054                         } else {
2055                                 $offset = $options['timecorrection'];
2056                         }
2057                         $ts = $this->userAdjust( $ts, $offset );
2058                 }
2059                 if ( $options['format'] === true ) {
2060                         $format = $user->getDatePreference();
2061                 } else {
2062                         $format = $options['format'];
2063                 }
2064                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2065                 return $this->sprintfDate( $df, $ts );
2066         }
2067
2068         /**
2069          * Get the formatted date for the given timestamp and formatted for
2070          * the given user.
2071          *
2072          * @param $ts Mixed: the time format which needs to be turned into a
2073          *            date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2074          * @param $user User object used to get preferences for timezone and format
2075          * @param $options Array, can contain the following keys:
2076          *        - 'timecorrection': time correction, can have the following values:
2077          *             - true: use user's preference
2078          *             - false: don't use time correction
2079          *             - integer: value of time correction in minutes
2080          *        - 'format': format to use, can have the following values:
2081          *             - true: use user's preference
2082          *             - false: use default preference
2083          *             - string: format to use
2084          * @since 1.19
2085          * @return String
2086          */
2087         public function userDate( $ts, User $user, array $options = array() ) {
2088                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2089         }
2090
2091         /**
2092          * Get the formatted time for the given timestamp and formatted for
2093          * the given user.
2094          *
2095          * @param $ts Mixed: the time format which needs to be turned into a
2096          *            date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2097          * @param $user User object used to get preferences for timezone and format
2098          * @param $options Array, can contain the following keys:
2099          *        - 'timecorrection': time correction, can have the following values:
2100          *             - true: use user's preference
2101          *             - false: don't use time correction
2102          *             - integer: value of time correction in minutes
2103          *        - 'format': format to use, can have the following values:
2104          *             - true: use user's preference
2105          *             - false: use default preference
2106          *             - string: format to use
2107          * @since 1.19
2108          * @return String
2109          */
2110         public function userTime( $ts, User $user, array $options = array() ) {
2111                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2112         }
2113
2114         /**
2115          * Get the formatted date and time for the given timestamp and formatted for
2116          * the given user.
2117          *
2118          * @param $ts Mixed: the time format which needs to be turned into a
2119          *            date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2120          * @param $user User object used to get preferences for timezone and format
2121          * @param $options Array, can contain the following keys:
2122          *        - 'timecorrection': time correction, can have the following values:
2123          *             - true: use user's preference
2124          *             - false: don't use time correction
2125          *             - integer: value of time correction in minutes
2126          *        - 'format': format to use, can have the following values:
2127          *             - true: use user's preference
2128          *             - false: use default preference
2129          *             - string: format to use
2130          * @since 1.19
2131          * @return String
2132          */
2133         public function userTimeAndDate( $ts, User $user, array $options = array() ) {
2134                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2135         }
2136
2137         /**
2138          * @param $key string
2139          * @return array|null
2140          */
2141         function getMessage( $key ) {
2142                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2143         }
2144
2145         /**
2146          * @return array
2147          */
2148         function getAllMessages() {
2149                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2150         }
2151
2152         /**
2153          * @param $in
2154          * @param $out
2155          * @param $string
2156          * @return string
2157          */
2158         function iconv( $in, $out, $string ) {
2159                 # This is a wrapper for iconv in all languages except esperanto,
2160                 # which does some nasty x-conversions beforehand
2161
2162                 # Even with //IGNORE iconv can whine about illegal characters in
2163                 # *input* string. We just ignore those too.
2164                 # REF: http://bugs.php.net/bug.php?id=37166
2165                 # REF: https://bugzilla.wikimedia.org/show_bug.cgi?id=16885
2166                 wfSuppressWarnings();
2167                 $text = iconv( $in, $out . '//IGNORE', $string );
2168                 wfRestoreWarnings();
2169                 return $text;
2170         }
2171
2172         // callback functions for uc(), lc(), ucwords(), ucwordbreaks()
2173
2174         /**
2175          * @param $matches array
2176          * @return mixed|string
2177          */
2178         function ucwordbreaksCallbackAscii( $matches ) {
2179                 return $this->ucfirst( $matches[1] );
2180         }
2181
2182         /**
2183          * @param $matches array
2184          * @return string
2185          */
2186         function ucwordbreaksCallbackMB( $matches ) {
2187                 return mb_strtoupper( $matches[0] );
2188         }
2189
2190         /**
2191          * @param $matches array
2192          * @return string
2193          */
2194         function ucCallback( $matches ) {
2195                 list( $wikiUpperChars ) = self::getCaseMaps();
2196                 return strtr( $matches[1], $wikiUpperChars );
2197         }
2198
2199         /**
2200          * @param $matches array
2201          * @return string
2202          */
2203         function lcCallback( $matches ) {
2204                 list( , $wikiLowerChars ) = self::getCaseMaps();
2205                 return strtr( $matches[1], $wikiLowerChars );
2206         }
2207
2208         /**
2209          * @param $matches array
2210          * @return string
2211          */
2212         function ucwordsCallbackMB( $matches ) {
2213                 return mb_strtoupper( $matches[0] );
2214         }
2215
2216         /**
2217          * @param $matches array
2218          * @return string
2219          */
2220         function ucwordsCallbackWiki( $matches ) {
2221                 list( $wikiUpperChars ) = self::getCaseMaps();
2222                 return strtr( $matches[0], $wikiUpperChars );
2223         }
2224
2225         /**
2226          * Make a string's first character uppercase
2227          *
2228          * @param $str string
2229          *
2230          * @return string
2231          */
2232         function ucfirst( $str ) {
2233                 $o = ord( $str );
2234                 if ( $o < 96 ) { // if already uppercase...
2235                         return $str;
2236                 } elseif ( $o < 128 ) {
2237                         return ucfirst( $str ); // use PHP's ucfirst()
2238                 } else {
2239                         // fall back to more complex logic in case of multibyte strings
2240                         return $this->uc( $str, true );
2241                 }
2242         }
2243
2244         /**
2245          * Convert a string to uppercase
2246          *
2247          * @param $str string
2248          * @param $first bool
2249          *
2250          * @return string
2251          */
2252         function uc( $str, $first = false ) {
2253                 if ( function_exists( 'mb_strtoupper' ) ) {
2254                         if ( $first ) {
2255                                 if ( $this->isMultibyte( $str ) ) {
2256                                         return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2257                                 } else {
2258                                         return ucfirst( $str );
2259                                 }
2260                         } else {
2261                                 return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2262                         }
2263                 } else {
2264                         if ( $this->isMultibyte( $str ) ) {
2265                                 $x = $first ? '^' : '';
2266                                 return preg_replace_callback(
2267                                         "/$x([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/",
2268                                         array( $this, 'ucCallback' ),
2269                                         $str
2270                                 );
2271                         } else {
2272                                 return $first ? ucfirst( $str ) : strtoupper( $str );
2273                         }
2274                 }
2275         }
2276
2277         /**
2278          * @param $str string
2279          * @return mixed|string
2280          */
2281         function lcfirst( $str ) {
2282                 $o = ord( $str );
2283                 if ( !$o ) {
2284                         return strval( $str );
2285                 } elseif ( $o >= 128 ) {
2286                         return $this->lc( $str, true );
2287                 } elseif ( $o > 96 ) {
2288                         return $str;
2289                 } else {
2290                         $str[0] = strtolower( $str[0] );
2291                         return $str;
2292                 }
2293         }
2294
2295         /**
2296          * @param $str string
2297          * @param $first bool
2298          * @return mixed|string
2299          */
2300         function lc( $str, $first = false ) {
2301                 if ( function_exists( 'mb_strtolower' ) ) {
2302                         if ( $first ) {
2303                                 if ( $this->isMultibyte( $str ) ) {
2304                                         return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2305                                 } else {
2306                                         return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2307                                 }
2308                         } else {
2309                                 return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2310                         }
2311                 } else {
2312                         if ( $this->isMultibyte( $str ) ) {
2313                                 $x = $first ? '^' : '';
2314                                 return preg_replace_callback(
2315                                         "/$x([A-Z]|[\\xc0-\\xff][\\x80-\\xbf]*)/",
2316                                         array( $this, 'lcCallback' ),
2317                                         $str
2318                                 );
2319                         } else {
2320                                 return $first ? strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 ) : strtolower( $str );
2321                         }
2322                 }
2323         }
2324
2325         /**
2326          * @param $str string
2327          * @return bool
2328          */
2329         function isMultibyte( $str ) {
2330                 return (bool)preg_match( '/[\x80-\xff]/', $str );
2331         }
2332
2333         /**
2334          * @param $str string
2335          * @return mixed|string
2336          */
2337         function ucwords( $str ) {
2338                 if ( $this->isMultibyte( $str ) ) {
2339                         $str = $this->lc( $str );
2340
2341                         // regexp to find first letter in each word (i.e. after each space)
2342                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2343
2344                         // function to use to capitalize a single char
2345                         if ( function_exists( 'mb_strtoupper' ) ) {
2346                                 return preg_replace_callback(
2347                                         $replaceRegexp,
2348                                         array( $this, 'ucwordsCallbackMB' ),
2349                                         $str
2350                                 );
2351                         } else {
2352                                 return preg_replace_callback(
2353                                         $replaceRegexp,
2354                                         array( $this, 'ucwordsCallbackWiki' ),
2355                                         $str
2356                                 );
2357                         }
2358                 } else {
2359                         return ucwords( strtolower( $str ) );
2360                 }
2361         }
2362
2363         /**
2364          * capitalize words at word breaks
2365          *
2366          * @param $str string
2367          * @return mixed
2368          */
2369         function ucwordbreaks( $str ) {
2370                 if ( $this->isMultibyte( $str ) ) {
2371                         $str = $this->lc( $str );
2372
2373                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2374                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2375
2376                         // find first letter after word break
2377                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2378
2379                         if ( function_exists( 'mb_strtoupper' ) ) {
2380                                 return preg_replace_callback(
2381                                         $replaceRegexp,
2382                                         array( $this, 'ucwordbreaksCallbackMB' ),
2383                                         $str
2384                                 );
2385                         } else {
2386                                 return preg_replace_callback(
2387                                         $replaceRegexp,
2388                                         array( $this, 'ucwordsCallbackWiki' ),
2389                                         $str
2390                                 );
2391                         }
2392                 } else {
2393                         return preg_replace_callback(
2394                                 '/\b([\w\x80-\xff]+)\b/',
2395                                 array( $this, 'ucwordbreaksCallbackAscii' ),
2396                                 $str
2397                         );
2398                 }
2399         }
2400
2401         /**
2402          * Return a case-folded representation of $s
2403          *
2404          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2405          * and $s2 are the same except for the case of their characters. It is not
2406          * necessary for the value returned to make sense when displayed.
2407          *
2408          * Do *not* perform any other normalisation in this function. If a caller
2409          * uses this function when it should be using a more general normalisation
2410          * function, then fix the caller.
2411          *
2412          * @param $s string
2413          *
2414          * @return string
2415          */
2416         function caseFold( $s ) {
2417                 return $this->uc( $s );
2418         }
2419
2420         /**
2421          * @param $s string
2422          * @return string
2423          */
2424         function checkTitleEncoding( $s ) {
2425                 if ( is_array( $s ) ) {
2426                         wfDebugDieBacktrace( 'Given array to checkTitleEncoding.' );
2427                 }
2428                 if ( StringUtils::isUtf8( $s ) ) {
2429                         return $s;
2430                 }
2431
2432                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2433         }
2434
2435         /**
2436          * @return array
2437          */
2438         function fallback8bitEncoding() {
2439                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2440         }
2441
2442         /**
2443          * Most writing systems use whitespace to break up words.
2444          * Some languages such as Chinese don't conventionally do this,
2445          * which requires special handling when breaking up words for
2446          * searching etc.
2447          *
2448          * @return bool
2449          */
2450         function hasWordBreaks() {
2451                 return true;
2452         }
2453
2454         /**
2455          * Some languages such as Chinese require word segmentation,
2456          * Specify such segmentation when overridden in derived class.
2457          *
2458          * @param $string String
2459          * @return String
2460          */
2461         function segmentByWord( $string ) {
2462                 return $string;
2463         }
2464
2465         /**
2466          * Some languages have special punctuation need to be normalized.
2467          * Make such changes here.
2468          *
2469          * @param $string String
2470          * @return String
2471          */
2472         function normalizeForSearch( $string ) {
2473                 return self::convertDoubleWidth( $string );
2474         }
2475
2476         /**
2477          * convert double-width roman characters to single-width.
2478          * range: ff00-ff5f ~= 0020-007f
2479          *
2480          * @param $string string
2481          *
2482          * @return string
2483          */
2484         protected static function convertDoubleWidth( $string ) {
2485                 static $full = null;
2486                 static $half = null;
2487
2488                 if ( $full === null ) {
2489                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2490                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2491                         $full = str_split( $fullWidth, 3 );
2492                         $half = str_split( $halfWidth );
2493                 }
2494
2495                 $string = str_replace( $full, $half, $string );
2496                 return $string;
2497         }
2498
2499         /**
2500          * @param $string string
2501          * @param $pattern string
2502          * @return string
2503          */
2504         protected static function insertSpace( $string, $pattern ) {
2505                 $string = preg_replace( $pattern, " $1 ", $string );
2506                 $string = preg_replace( '/ +/', ' ', $string );
2507                 return $string;
2508         }
2509
2510         /**
2511          * @param $termsArray array
2512          * @return array
2513          */
2514         function convertForSearchResult( $termsArray ) {
2515                 # some languages, e.g. Chinese, need to do a conversion
2516                 # in order for search results to be displayed correctly
2517                 return $termsArray;
2518         }
2519
2520         /**
2521          * Get the first character of a string.
2522          *
2523          * @param $s string
2524          * @return string
2525          */
2526         function firstChar( $s ) {
2527                 $matches = array();
2528                 preg_match(
2529                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2530                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2531                         $s,
2532                         $matches
2533                 );
2534
2535                 if ( isset( $matches[1] ) ) {
2536                         if ( strlen( $matches[1] ) != 3 ) {
2537                                 return $matches[1];
2538                         }
2539
2540                         // Break down Hangul syllables to grab the first jamo
2541                         $code = utf8ToCodepoint( $matches[1] );
2542                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2543                                 return $matches[1];
2544                         } elseif ( $code < 0xb098 ) {
2545                                 return "\xe3\x84\xb1";
2546                         } elseif ( $code < 0xb2e4 ) {
2547                                 return "\xe3\x84\xb4";
2548                         } elseif ( $code < 0xb77c ) {
2549                                 return "\xe3\x84\xb7";
2550                         } elseif ( $code < 0xb9c8 ) {
2551                                 return "\xe3\x84\xb9";
2552                         } elseif ( $code < 0xbc14 ) {
2553                                 return "\xe3\x85\x81";
2554                         } elseif ( $code < 0xc0ac ) {
2555                                 return "\xe3\x85\x82";
2556                         } elseif ( $code < 0xc544 ) {
2557                                 return "\xe3\x85\x85";
2558                         } elseif ( $code < 0xc790 ) {
2559                                 return "\xe3\x85\x87";
2560                         } elseif ( $code < 0xcc28 ) {
2561                                 return "\xe3\x85\x88";
2562                         } elseif ( $code < 0xce74 ) {
2563                                 return "\xe3\x85\x8a";
2564                         } elseif ( $code < 0xd0c0 ) {
2565                                 return "\xe3\x85\x8b";
2566                         } elseif ( $code < 0xd30c ) {
2567                                 return "\xe3\x85\x8c";
2568                         } elseif ( $code < 0xd558 ) {
2569                                 return "\xe3\x85\x8d";
2570                         } else {
2571                                 return "\xe3\x85\x8e";
2572                         }
2573                 } else {
2574                         return '';
2575                 }
2576         }
2577
2578         function initEncoding() {
2579                 # Some languages may have an alternate char encoding option
2580                 # (Esperanto X-coding, Japanese furigana conversion, etc)
2581                 # If this language is used as the primary content language,
2582                 # an override to the defaults can be set here on startup.
2583         }
2584
2585         /**
2586          * @param $s string
2587          * @return string
2588          */
2589         function recodeForEdit( $s ) {
2590                 # For some languages we'll want to explicitly specify
2591                 # which characters make it into the edit box raw
2592                 # or are converted in some way or another.
2593                 global $wgEditEncoding;
2594                 if ( $wgEditEncoding == '' || $wgEditEncoding == 'UTF-8' ) {
2595                         return $s;
2596                 } else {
2597                         return $this->iconv( 'UTF-8', $wgEditEncoding, $s );
2598                 }
2599         }
2600
2601         /**
2602          * @param $s string
2603          * @return string
2604          */
2605         function recodeInput( $s ) {
2606                 # Take the previous into account.
2607                 global $wgEditEncoding;
2608                 if ( $wgEditEncoding != '' ) {
2609                         $enc = $wgEditEncoding;
2610                 } else {
2611                         $enc = 'UTF-8';
2612                 }
2613                 if ( $enc == 'UTF-8' ) {
2614                         return $s;
2615                 } else {
2616                         return $this->iconv( $enc, 'UTF-8', $s );
2617                 }
2618         }
2619
2620         /**
2621          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
2622          * also cleans up certain backwards-compatible sequences, converting them
2623          * to the modern Unicode equivalent.
2624          *
2625          * This is language-specific for performance reasons only.
2626          *
2627          * @param $s string
2628          *
2629          * @return string
2630          */
2631         function normalize( $s ) {
2632                 global $wgAllUnicodeFixes;
2633                 $s = UtfNormal::cleanUp( $s );
2634                 if ( $wgAllUnicodeFixes ) {
2635                         $s = $this->transformUsingPairFile( 'normalize-ar.ser', $s );
2636                         $s = $this->transformUsingPairFile( 'normalize-ml.ser', $s );
2637                 }
2638
2639                 return $s;
2640         }
2641
2642         /**
2643          * Transform a string using serialized data stored in the given file (which
2644          * must be in the serialized subdirectory of $IP). The file contains pairs
2645          * mapping source characters to destination characters.
2646          *
2647          * The data is cached in process memory. This will go faster if you have the
2648          * FastStringSearch extension.
2649          *
2650          * @param $file string
2651          * @param $string string
2652          *
2653          * @throws MWException
2654          * @return string
2655          */
2656         function transformUsingPairFile( $file, $string ) {
2657                 if ( !isset( $this->transformData[$file] ) ) {
2658                         $data = wfGetPrecompiledData( $file );
2659                         if ( $data === false ) {
2660                                 throw new MWException( __METHOD__ . ": The transformation file $file is missing" );
2661                         }
2662                         $this->transformData[$file] = new ReplacementArray( $data );
2663                 }
2664                 return $this->transformData[$file]->replace( $string );
2665         }
2666
2667         /**
2668          * For right-to-left language support
2669          *
2670          * @return bool
2671          */
2672         function isRTL() {
2673                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
2674         }
2675
2676         /**
2677          * Return the correct HTML 'dir' attribute value for this language.
2678          * @return String
2679          */
2680         function getDir() {
2681                 return $this->isRTL() ? 'rtl' : 'ltr';
2682         }
2683
2684         /**
2685          * Return 'left' or 'right' as appropriate alignment for line-start
2686          * for this language's text direction.
2687          *
2688          * Should be equivalent to CSS3 'start' text-align value....
2689          *
2690          * @return String
2691          */
2692         function alignStart() {
2693                 return $this->isRTL() ? 'right' : 'left';
2694         }
2695
2696         /**
2697          * Return 'right' or 'left' as appropriate alignment for line-end
2698          * for this language's text direction.
2699          *
2700          * Should be equivalent to CSS3 'end' text-align value....
2701          *
2702          * @return String
2703          */
2704         function alignEnd() {
2705                 return $this->isRTL() ? 'left' : 'right';
2706         }
2707
2708         /**
2709          * A hidden direction mark (LRM or RLM), depending on the language direction.
2710          * Unlike getDirMark(), this function returns the character as an HTML entity.
2711          * This function should be used when the output is guaranteed to be HTML,
2712          * because it makes the output HTML source code more readable. When
2713          * the output is plain text or can be escaped, getDirMark() should be used.
2714          *
2715          * @param $opposite Boolean Get the direction mark opposite to your language
2716          * @return string
2717          * @since 1.20
2718          */
2719         function getDirMarkEntity( $opposite = false ) {
2720                 if ( $opposite ) { return $this->isRTL() ? '&lrm;' : '&rlm;'; }
2721                 return $this->isRTL() ? '&rlm;' : '&lrm;';
2722         }
2723
2724         /**
2725          * A hidden direction mark (LRM or RLM), depending on the language direction.
2726          * This function produces them as invisible Unicode characters and
2727          * the output may be hard to read and debug, so it should only be used
2728          * when the output is plain text or can be escaped. When the output is
2729          * HTML, use getDirMarkEntity() instead.
2730          *
2731          * @param $opposite Boolean Get the direction mark opposite to your language
2732          * @return string
2733          */
2734         function getDirMark( $opposite = false ) {
2735                 $lrm = "\xE2\x80\x8E"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
2736                 $rlm = "\xE2\x80\x8F"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
2737                 if ( $opposite ) { return $this->isRTL() ? $lrm : $rlm; }
2738                 return $this->isRTL() ? $rlm : $lrm;
2739         }
2740
2741         /**
2742          * @return array
2743          */
2744         function capitalizeAllNouns() {
2745                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
2746         }
2747
2748         /**
2749          * An arrow, depending on the language direction.
2750          *
2751          * @param $direction String: the direction of the arrow: forwards (default), backwards, left, right, up, down.
2752          * @return string
2753          */
2754         function getArrow( $direction = 'forwards' ) {
2755                 switch ( $direction ) {
2756                 case 'forwards':
2757                         return $this->isRTL() ? '←' : '→';
2758                 case 'backwards':
2759                         return $this->isRTL() ? '→' : '←';
2760                 case 'left':
2761                         return '←';
2762                 case 'right':
2763                         return '→';
2764                 case 'up':
2765                         return '↑';
2766                 case 'down':
2767                         return '↓';
2768                 }
2769         }
2770
2771         /**
2772          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
2773          *
2774          * @return bool
2775          */
2776         function linkPrefixExtension() {
2777                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
2778         }
2779
2780         /**
2781          * @return array
2782          */
2783         function getMagicWords() {
2784                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
2785         }
2786
2787         protected function doMagicHook() {
2788                 if ( $this->mMagicHookDone ) {
2789                         return;
2790                 }
2791                 $this->mMagicHookDone = true;
2792                 wfProfileIn( 'LanguageGetMagic' );
2793                 wfRunHooks( 'LanguageGetMagic', array( &$this->mMagicExtensions, $this->getCode() ) );
2794                 wfProfileOut( 'LanguageGetMagic' );
2795         }
2796
2797         /**
2798          * Fill a MagicWord object with data from here
2799          *
2800          * @param $mw
2801          */
2802         function getMagic( $mw ) {
2803                 $this->doMagicHook();
2804
2805                 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
2806                         $rawEntry = $this->mMagicExtensions[$mw->mId];
2807                 } else {
2808                         $magicWords = $this->getMagicWords();
2809                         if ( isset( $magicWords[$mw->mId] ) ) {
2810                                 $rawEntry = $magicWords[$mw->mId];
2811                         } else {
2812                                 $rawEntry = false;
2813                         }
2814                 }
2815
2816                 if ( !is_array( $rawEntry ) ) {
2817                         error_log( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
2818                 } else {
2819                         $mw->mCaseSensitive = $rawEntry[0];
2820                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
2821                 }
2822         }
2823
2824         /**
2825          * Add magic words to the extension array
2826          *
2827          * @param $newWords array
2828          */
2829         function addMagicWordsByLang( $newWords ) {
2830                 $fallbackChain = $this->getFallbackLanguages();
2831                 $fallbackChain = array_reverse( $fallbackChain );
2832                 foreach ( $fallbackChain as $code ) {
2833                         if ( isset( $newWords[$code] ) ) {
2834                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
2835                         }
2836                 }
2837         }
2838
2839         /**
2840          * Get special page names, as an associative array
2841          *   case folded alias => real name
2842          */
2843         function getSpecialPageAliases() {
2844                 // Cache aliases because it may be slow to load them
2845                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
2846                         // Initialise array
2847                         $this->mExtendedSpecialPageAliases =
2848                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
2849                         wfRunHooks( 'LanguageGetSpecialPageAliases',
2850                                 array( &$this->mExtendedSpecialPageAliases, $this->getCode() ) );
2851                 }
2852
2853                 return $this->mExtendedSpecialPageAliases;
2854         }
2855
2856         /**
2857          * Italic is unsuitable for some languages
2858          *
2859          * @param $text String: the text to be emphasized.
2860          * @return string
2861          */
2862         function emphasize( $text ) {
2863                 return "<em>$text</em>";
2864         }
2865
2866         /**
2867          * Normally we output all numbers in plain en_US style, that is
2868          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
2869          * point twohundredthirtyfive. However this is not suitable for all
2870          * languages, some such as Pakaran want ੨੯੩,੨੯੫.੨੩੫ and others such as
2871          * Icelandic just want to use commas instead of dots, and dots instead
2872          * of commas like "293.291,235".
2873          *
2874          * An example of this function being called:
2875          * <code>
2876          * wfMessage( 'message' )->numParams( $num )->text()
2877          * </code>
2878          *
2879          * See LanguageGu.php for the Gujarati implementation and
2880          * $separatorTransformTable on MessageIs.php for
2881          * the , => . and . => , implementation.
2882          *
2883          * @todo check if it's viable to use localeconv() for the decimal
2884          *       separator thing.
2885          * @param $number Mixed: the string to be formatted, should be an integer
2886          *        or a floating point number.
2887          * @param $nocommafy Bool: set to true for special numbers like dates
2888          * @return string
2889          */
2890         public function formatNum( $number, $nocommafy = false ) {
2891                 global $wgTranslateNumerals;
2892                 if ( !$nocommafy ) {
2893                         $number = $this->commafy( $number );
2894                         $s = $this->separatorTransformTable();
2895                         if ( $s ) {
2896                                 $number = strtr( $number, $s );
2897                         }
2898                 }
2899
2900                 if ( $wgTranslateNumerals ) {
2901                         $s = $this->digitTransformTable();
2902                         if ( $s ) {
2903                                 $number = strtr( $number, $s );
2904                         }
2905                 }
2906
2907                 return $number;
2908         }
2909
2910         /**
2911          * @param $number string
2912          * @return string
2913          */
2914         function parseFormattedNumber( $number ) {
2915                 $s = $this->digitTransformTable();
2916                 if ( $s ) {
2917                         $number = strtr( $number, array_flip( $s ) );
2918                 }
2919
2920                 $s = $this->separatorTransformTable();
2921                 if ( $s ) {
2922                         $number = strtr( $number, array_flip( $s ) );
2923                 }
2924
2925                 $number = strtr( $number, array( ',' => '' ) );
2926                 return $number;
2927         }
2928
2929         /**
2930          * Adds commas to a given number
2931          * @since 1.19
2932          * @param $number mixed
2933          * @return string
2934          */
2935         function commafy( $number ) {
2936                 $digitGroupingPattern = $this->digitGroupingPattern();
2937                 if ( $number === null ) {
2938                         return '';
2939                 }
2940
2941                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
2942                         // default grouping is at thousands,  use the same for ###,###,### pattern too.
2943                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
2944                 } else {
2945                         // Ref: http://cldr.unicode.org/translation/number-patterns
2946                         $sign = "";
2947                         if ( intval( $number ) < 0 ) {
2948                                 // For negative numbers apply the algorithm like positive number and add sign.
2949                                 $sign =  "-";
2950                                 $number = substr( $number, 1 );
2951                         }
2952                         $integerPart = array();
2953                         $decimalPart = array();
2954                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
2955                         preg_match( "/\d+/", $number, $integerPart );
2956                         preg_match( "/\.\d*/", $number, $decimalPart );
2957                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0]:"";
2958                         if ( $groupedNumber  === $number ) {
2959                                 // the string does not have any number part. Eg: .12345
2960                                 return $sign . $groupedNumber;
2961                         }
2962                         $start = $end = strlen( $integerPart[0] );
2963                         while ( $start > 0 ) {
2964                                 $match = $matches[0][$numMatches -1] ;
2965                                 $matchLen = strlen( $match );
2966                                 $start = $end - $matchLen;
2967                                 if ( $start < 0 ) {
2968                                         $start = 0;
2969                                 }
2970                                 $groupedNumber = substr( $number , $start, $end -$start ) . $groupedNumber ;
2971                                 $end = $start;
2972                                 if ( $numMatches > 1 ) {
2973                                         // use the last pattern for the rest of the number
2974                                         $numMatches--;
2975                                 }
2976                                 if ( $start > 0 ) {
2977                                         $groupedNumber = "," . $groupedNumber;
2978                                 }
2979                         }
2980                         return $sign . $groupedNumber;
2981                 }
2982         }
2983
2984         /**
2985          * @return String
2986          */
2987         function digitGroupingPattern() {
2988                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
2989         }
2990
2991         /**
2992          * @return array
2993          */
2994         function digitTransformTable() {
2995                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
2996         }
2997
2998         /**
2999          * @return array
3000          */
3001         function separatorTransformTable() {
3002                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3003         }
3004
3005         /**
3006          * Take a list of strings and build a locale-friendly comma-separated
3007          * list, using the local comma-separator message.
3008          * The last two strings are chained with an "and".
3009          * NOTE: This function will only work with standard numeric array keys (0, 1, 2…)
3010          *
3011          * @param $l Array
3012          * @return string
3013          */
3014         function listToText( array $l ) {
3015                 $s = '';
3016                 $m = count( $l ) - 1;
3017
3018                 if ( $m === 0 ) {
3019                         return $l[0];
3020                 } elseif ( $m === 1 ) {
3021                         return $l[0] . $this->getMessageFromDB( 'and' ) . $this->getMessageFromDB( 'word-separator' ) . $l[1];
3022                 } else {
3023                         for ( $i = $m; $i >= 0; $i-- ) {
3024                                 if ( $i == $m ) {
3025                                         $s = $l[$i];
3026                                 } elseif ( $i == $m - 1 ) {
3027                                         $s = $l[$i] . $this->getMessageFromDB( 'and' ) . $this->getMessageFromDB( 'word-separator' ) . $s;
3028                                 } else {
3029                                         $s = $l[$i] . $this->getMessageFromDB( 'comma-separator' ) . $s;
3030                                 }
3031                         }
3032                         return $s;
3033                 }
3034         }
3035
3036         /**
3037          * Take a list of strings and build a locale-friendly comma-separated
3038          * list, using the local comma-separator message.
3039          * @param $list array of strings to put in a comma list
3040          * @return string
3041          */
3042         function commaList( array $list ) {
3043                 return implode(
3044                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3045                         $list
3046                 );
3047         }
3048
3049         /**
3050          * Take a list of strings and build a locale-friendly semicolon-separated
3051          * list, using the local semicolon-separator message.
3052          * @param $list array of strings to put in a semicolon list
3053          * @return string
3054          */
3055         function semicolonList( array $list ) {
3056                 return implode(
3057                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3058                         $list
3059                 );
3060         }
3061
3062         /**
3063          * Same as commaList, but separate it with the pipe instead.
3064          * @param $list array of strings to put in a pipe list
3065          * @return string
3066          */
3067         function pipeList( array $list ) {
3068                 return implode(
3069                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3070                         $list
3071                 );
3072         }
3073
3074         /**
3075          * Truncate a string to a specified length in bytes, appending an optional
3076          * string (e.g. for ellipses)
3077          *
3078          * The database offers limited byte lengths for some columns in the database;
3079          * multi-byte character sets mean we need to ensure that only whole characters
3080          * are included, otherwise broken characters can be passed to the user
3081          *
3082          * If $length is negative, the string will be truncated from the beginning
3083          *
3084          * @param $string String to truncate
3085          * @param $length Int: maximum length (including ellipses)
3086          * @param $ellipsis String to append to the truncated text
3087          * @param $adjustLength Boolean: Subtract length of ellipsis from $length.
3088          *      $adjustLength was introduced in 1.18, before that behaved as if false.
3089          * @return string
3090          */
3091         function truncate( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3092                 # Use the localized ellipsis character
3093                 if ( $ellipsis == '...' ) {
3094                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3095                 }
3096                 # Check if there is no need to truncate
3097                 if ( $length == 0 ) {
3098                         return $ellipsis; // convention
3099                 } elseif ( strlen( $string ) <= abs( $length ) ) {
3100                         return $string; // no need to truncate
3101                 }
3102                 $stringOriginal = $string;
3103                 # If ellipsis length is >= $length then we can't apply $adjustLength
3104                 if ( $adjustLength && strlen( $ellipsis ) >= abs( $length ) ) {
3105                         $string = $ellipsis; // this can be slightly unexpected
3106                 # Otherwise, truncate and add ellipsis...
3107                 } else {
3108                         $eLength = $adjustLength ? strlen( $ellipsis ) : 0;
3109                         if ( $length > 0 ) {
3110                                 $length -= $eLength;
3111                                 $string = substr( $string, 0, $length ); // xyz...
3112                                 $string = $this->removeBadCharLast( $string );
3113                                 $string = $string . $ellipsis;
3114                         } else {
3115                                 $length += $eLength;
3116                                 $string = substr( $string, $length ); // ...xyz
3117                                 $string = $this->removeBadCharFirst( $string );
3118                                 $string = $ellipsis . $string;
3119                         }
3120                 }
3121                 # Do not truncate if the ellipsis makes the string longer/equal (bug 22181).
3122                 # This check is *not* redundant if $adjustLength, due to the single case where
3123                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3124                 if ( strlen( $string ) < strlen( $stringOriginal ) ) {
3125                         return $string;
3126                 } else {
3127                         return $stringOriginal;
3128                 }
3129         }
3130
3131         /**
3132          * Remove bytes that represent an incomplete Unicode character
3133          * at the end of string (e.g. bytes of the char are missing)
3134          *
3135          * @param $string String
3136          * @return string
3137          */
3138         protected function removeBadCharLast( $string ) {
3139                 if ( $string != '' ) {
3140                         $char = ord( $string[strlen( $string ) - 1] );
3141                         $m = array();
3142                         if ( $char >= 0xc0 ) {
3143                                 # We got the first byte only of a multibyte char; remove it.
3144                                 $string = substr( $string, 0, -1 );
3145                         } elseif ( $char >= 0x80 &&
3146                                   preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3147                                                           '[\xf0-\xf7][\x80-\xbf]{1,2})$/', $string, $m ) )
3148                         {
3149                                 # We chopped in the middle of a character; remove it
3150                                 $string = $m[1];
3151                         }
3152                 }
3153                 return $string;
3154         }
3155
3156         /**
3157          * Remove bytes that represent an incomplete Unicode character
3158          * at the start of string (e.g. bytes of the char are missing)
3159          *
3160          * @param $string String
3161          * @return string
3162          */
3163         protected function removeBadCharFirst( $string ) {
3164                 if ( $string != '' ) {
3165                         $char = ord( $string[0] );
3166                         if ( $char >= 0x80 && $char < 0xc0 ) {
3167                                 # We chopped in the middle of a character; remove the whole thing
3168                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3169                         }
3170                 }
3171                 return $string;
3172         }
3173
3174         /**
3175          * Truncate a string of valid HTML to a specified length in bytes,
3176          * appending an optional string (e.g. for ellipses), and return valid HTML
3177          *
3178          * This is only intended for styled/linked text, such as HTML with
3179          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3180          * Also, this will not detect things like "display:none" CSS.
3181          *
3182          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3183          *
3184          * @param string $text HTML string to truncate
3185          * @param int $length (zero/positive) Maximum length (including ellipses)
3186          * @param string $ellipsis String to append to the truncated text
3187          * @return string
3188          */
3189         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3190                 # Use the localized ellipsis character
3191                 if ( $ellipsis == '...' ) {
3192                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3193                 }
3194                 # Check if there is clearly no need to truncate
3195                 if ( $length <= 0 ) {
3196                         return $ellipsis; // no text shown, nothing to format (convention)
3197                 } elseif ( strlen( $text ) <= $length ) {
3198                         return $text; // string short enough even *with* HTML (short-circuit)
3199                 }
3200
3201                 $dispLen = 0; // innerHTML legth so far
3202                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3203                 $tagType = 0; // 0-open, 1-close
3204                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3205                 $entityState = 0; // 0-not entity, 1-entity
3206                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3207                 $openTags = array(); // open tag stack
3208                 $maybeState = null; // possible truncation state
3209
3210                 $textLen = strlen( $text );
3211                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3212                 for ( $pos = 0; true; ++$pos ) {
3213                         # Consider truncation once the display length has reached the maximim.
3214                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3215                         # Check that we're not in the middle of a bracket/entity...
3216                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3217                                 if ( !$testingEllipsis ) {
3218                                         $testingEllipsis = true;
3219                                         # Save where we are; we will truncate here unless there turn out to
3220                                         # be so few remaining characters that truncation is not necessary.
3221                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3222                                                 $maybeState = array( $ret, $openTags ); // save state
3223                                         }
3224                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3225                                         # String in fact does need truncation, the truncation point was OK.
3226                                         list( $ret, $openTags ) = $maybeState; // reload state
3227                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3228                                         $ret .= $ellipsis; // add ellipsis
3229                                         break;
3230                                 }
3231                         }
3232                         if ( $pos >= $textLen ) break; // extra iteration just for above checks
3233
3234                         # Read the next char...
3235                         $ch = $text[$pos];
3236                         $lastCh = $pos ? $text[$pos - 1] : '';
3237                         $ret .= $ch; // add to result string
3238                         if ( $ch == '<' ) {
3239                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3240                                 $entityState = 0; // for bad HTML
3241                                 $bracketState = 1; // tag started (checking for backslash)
3242                         } elseif ( $ch == '>' ) {
3243                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3244                                 $entityState = 0; // for bad HTML
3245                                 $bracketState = 0; // out of brackets
3246                         } elseif ( $bracketState == 1 ) {
3247                                 if ( $ch == '/' ) {
3248                                         $tagType = 1; // close tag (e.g. "</span>")
3249                                 } else {
3250                                         $tagType = 0; // open tag (e.g. "<span>")
3251                                         $tag .= $ch;
3252                                 }
3253                                 $bracketState = 2; // building tag name
3254                         } elseif ( $bracketState == 2 ) {
3255                                 if ( $ch != ' ' ) {
3256                                         $tag .= $ch;
3257                                 } else {
3258                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3259                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3260                                 }
3261                         } elseif ( $bracketState == 0 ) {
3262                                 if ( $entityState ) {
3263                                         if ( $ch == ';' ) {
3264                                                 $entityState = 0;
3265                                                 $dispLen++; // entity is one displayed char
3266                                         }
3267                                 } else {
3268                                         if ( $neLength == 0 && !$maybeState ) {
3269                                                 // Save state without $ch. We want to *hit* the first
3270                                                 // display char (to get tags) but not *use* it if truncating.
3271                                                 $maybeState = array( substr( $ret, 0, -1 ), $openTags );
3272                                         }
3273                                         if ( $ch == '&' ) {
3274                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3275                                         } else {
3276                                                 $dispLen++; // this char is displayed
3277                                                 // Add the next $max display text chars after this in one swoop...
3278                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3279                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3280                                                 $dispLen += $skipped;
3281                                                 $pos += $skipped;
3282                                         }
3283                                 }
3284                         }
3285                 }
3286                 // Close the last tag if left unclosed by bad HTML
3287                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3288                 while ( count( $openTags ) > 0 ) {
3289                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3290                 }
3291                 return $ret;
3292         }
3293
3294         /**
3295          * truncateHtml() helper function
3296          * like strcspn() but adds the skipped chars to $ret
3297          *
3298          * @param $ret
3299          * @param $text
3300          * @param $search
3301          * @param $start
3302          * @param $len
3303          * @return int
3304          */
3305         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3306                 if ( $len === null ) {
3307                         $len = -1; // -1 means "no limit" for strcspn
3308                 } elseif ( $len < 0 ) {
3309                         $len = 0; // sanity
3310                 }
3311                 $skipCount = 0;
3312                 if ( $start < strlen( $text ) ) {
3313                         $skipCount = strcspn( $text, $search, $start, $len );
3314                         $ret .= substr( $text, $start, $skipCount );
3315                 }
3316                 return $skipCount;
3317         }
3318
3319         /**
3320          * truncateHtml() helper function
3321          * (a) push or pop $tag from $openTags as needed
3322          * (b) clear $tag value
3323          * @param &$tag string Current HTML tag name we are looking at
3324          * @param $tagType int (0-open tag, 1-close tag)
3325          * @param $lastCh string Character before the '>' that ended this tag
3326          * @param &$openTags array Open tag stack (not accounting for $tag)
3327          */
3328         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3329                 $tag = ltrim( $tag );
3330                 if ( $tag != '' ) {
3331                         if ( $tagType == 0 && $lastCh != '/' ) {
3332                                 $openTags[] = $tag; // tag opened (didn't close itself)
3333                         } elseif ( $tagType == 1 ) {
3334                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3335                                         array_pop( $openTags ); // tag closed
3336                                 }
3337                         }
3338                         $tag = '';
3339                 }
3340         }
3341
3342         /**
3343          * Grammatical transformations, needed for inflected languages
3344          * Invoked by putting {{grammar:case|word}} in a message
3345          *
3346          * @param $word string
3347          * @param $case string
3348          * @return string
3349          */
3350         function convertGrammar( $word, $case ) {
3351                 global $wgGrammarForms;
3352                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3353                         return $wgGrammarForms[$this->getCode()][$case][$word];
3354                 }
3355                 return $word;
3356         }
3357         /**
3358          * Get the grammar forms for the content language
3359          * @return array of grammar forms
3360          * @since 1.20
3361          */
3362         function getGrammarForms() {
3363                 global $wgGrammarForms;
3364                 if ( isset( $wgGrammarForms[$this->getCode()] ) && is_array( $wgGrammarForms[$this->getCode()] ) ) {
3365                          return $wgGrammarForms[$this->getCode()];
3366                 }
3367                 return array();
3368         }
3369         /**
3370          * Provides an alternative text depending on specified gender.
3371          * Usage {{gender:username|masculine|feminine|neutral}}.
3372          * username is optional, in which case the gender of current user is used,
3373          * but only in (some) interface messages; otherwise default gender is used.
3374          *
3375          * If no forms are given, an empty string is returned. If only one form is
3376          * given, it will be returned unconditionally. These details are implied by
3377          * the caller and cannot be overridden in subclasses.
3378          *
3379          * If more than one form is given, the default is to use the neutral one
3380          * if it is specified, and to use the masculine one otherwise. These
3381          * details can be overridden in subclasses.
3382          *
3383          * @param $gender string
3384          * @param $forms array
3385          *
3386          * @return string
3387          */
3388         function gender( $gender, $forms ) {
3389                 if ( !count( $forms ) ) {
3390                         return '';
3391                 }
3392                 $forms = $this->preConvertPlural( $forms, 2 );
3393                 if ( $gender === 'male' ) {
3394                         return $forms[0];
3395                 }
3396                 if ( $gender === 'female' ) {
3397                         return $forms[1];
3398                 }
3399                 return isset( $forms[2] ) ? $forms[2] : $forms[0];
3400         }
3401
3402         /**
3403          * Plural form transformations, needed for some languages.
3404          * For example, there are 3 form of plural in Russian and Polish,
3405          * depending on "count mod 10". See [[w:Plural]]
3406          * For English it is pretty simple.
3407          *
3408          * Invoked by putting {{plural:count|wordform1|wordform2}}
3409          * or {{plural:count|wordform1|wordform2|wordform3}}
3410          *
3411          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3412          *
3413          * @param $count Integer: non-localized number
3414          * @param $forms Array: different plural forms
3415          * @return string Correct form of plural for $count in this language
3416          */
3417         function convertPlural( $count, $forms ) {
3418                 if ( !count( $forms ) ) {
3419                         return '';
3420                 }
3421
3422                 // Handle explicit 0= and 1= forms
3423                 foreach ( $forms as $index => $form ) {
3424                         if ( isset( $form[1] ) && $form[1] === '=' ) {
3425                                 if ( $form[0] === (string) $count ) {
3426                                         return substr( $form, 2 );
3427                                 }
3428                                 unset( $forms[$index] );
3429                         }
3430                 }
3431                 $forms = array_values( $forms );
3432
3433                 $pluralForm = $this->getPluralForm( $count );
3434                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3435                 return $forms[$pluralForm];
3436         }
3437
3438         /**
3439          * Checks that convertPlural was given an array and pads it to requested
3440          * amount of forms by copying the last one.
3441          *
3442          * @param $count Integer: How many forms should there be at least
3443          * @param $forms Array of forms given to convertPlural
3444          * @return array Padded array of forms or an exception if not an array
3445          */
3446         protected function preConvertPlural( /* Array */ $forms, $count ) {
3447                 while ( count( $forms ) < $count ) {
3448                         $forms[] = $forms[count( $forms ) - 1];
3449                 }
3450                 return $forms;
3451         }
3452
3453         /**
3454          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
3455          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
3456          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
3457          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
3458          * match up with it.
3459          *
3460          * @param $str String: the validated block duration in English
3461          * @return string Somehow translated block duration
3462          * @see LanguageFi.php for example implementation
3463          */
3464         function translateBlockExpiry( $str ) {
3465                 $duration = SpecialBlock::getSuggestedDurations( $this );
3466                 foreach ( $duration as $show => $value ) {
3467                         if ( strcmp( $str, $value ) == 0 ) {
3468                                 return htmlspecialchars( trim( $show ) );
3469                         }
3470                 }
3471
3472                 // Since usually only infinite or indefinite is only on list, so try
3473                 // equivalents if still here.
3474                 $indefs = array( 'infinite', 'infinity', 'indefinite' );
3475                 if ( in_array( $str, $indefs ) ) {
3476                         foreach ( $indefs as $val ) {
3477                                 $show = array_search( $val, $duration, true );
3478                                 if ( $show !== false ) {
3479                                         return htmlspecialchars( trim( $show ) );
3480                                 }
3481                         }
3482                 }
3483
3484                 // If all else fails, return a standard duration or timestamp description.
3485                 $time = strtotime( $str, 0 );
3486                 if ( $time === false ) { // Unknown format. Return it as-is in case.
3487                         return $str;
3488                 } elseif ( $time !== strtotime( $str, 1 ) ) { // It's a relative timestamp.
3489                         // $time is relative to 0 so it's a duration length.
3490                         return $this->formatDuration( $time );
3491                 } else { // It's an absolute timestamp.
3492                         if ( $time === 0 ) {
3493                                 // wfTimestamp() handles 0 as current time instead of epoch.
3494                                 return $this->timeanddate( '19700101000000' );
3495                         } else {
3496                                 return $this->timeanddate( $time );
3497                         }
3498                 }
3499         }
3500
3501         /**
3502          * languages like Chinese need to be segmented in order for the diff
3503          * to be of any use
3504          *
3505          * @param $text String
3506          * @return String
3507          */
3508         public function segmentForDiff( $text ) {
3509                 return $text;
3510         }
3511
3512         /**
3513          * and unsegment to show the result
3514          *
3515          * @param $text String
3516          * @return String
3517          */
3518         public function unsegmentForDiff( $text ) {
3519                 return $text;
3520         }
3521
3522         /**
3523          * Return the LanguageConverter used in the Language
3524          *
3525          * @since 1.19
3526          * @return LanguageConverter
3527          */
3528         public function getConverter() {
3529                 return $this->mConverter;
3530         }
3531
3532         /**
3533          * convert text to all supported variants
3534          *
3535          * @param $text string
3536          * @return array
3537          */
3538         public function autoConvertToAllVariants( $text ) {
3539                 return $this->mConverter->autoConvertToAllVariants( $text );
3540         }
3541
3542         /**
3543          * convert text to different variants of a language.
3544          *
3545          * @param $text string
3546          * @return string
3547          */
3548         public function convert( $text ) {
3549                 return $this->mConverter->convert( $text );
3550         }
3551
3552         /**
3553          * Convert a Title object to a string in the preferred variant
3554          *
3555          * @param $title Title
3556          * @return string
3557          */
3558         public function convertTitle( $title ) {
3559                 return $this->mConverter->convertTitle( $title );
3560         }
3561
3562         /**
3563          * Convert a namespace index to a string in the preferred variant
3564          *
3565          * @param $ns int
3566          * @return string
3567          */
3568         public function convertNamespace( $ns ) {
3569                 return $this->mConverter->convertNamespace( $ns );
3570         }
3571
3572         /**
3573          * Check if this is a language with variants
3574          *
3575          * @return bool
3576          */
3577         public function hasVariants() {
3578                 return sizeof( $this->getVariants() ) > 1;
3579         }
3580
3581         /**
3582          * Check if the language has the specific variant
3583          *
3584          * @since 1.19
3585          * @param $variant string
3586          * @return bool
3587          */
3588         public function hasVariant( $variant ) {
3589                 return (bool)$this->mConverter->validateVariant( $variant );
3590         }
3591
3592         /**
3593          * Put custom tags (e.g. -{ }-) around math to prevent conversion
3594          *
3595          * @param $text string
3596          * @return string
3597          */
3598         public function armourMath( $text ) {
3599                 return $this->mConverter->armourMath( $text );
3600         }
3601
3602         /**
3603          * Perform output conversion on a string, and encode for safe HTML output.
3604          * @param $text String text to be converted
3605          * @param $isTitle Bool whether this conversion is for the article title
3606          * @return string
3607          * @todo this should get integrated somewhere sane
3608          */
3609         public function convertHtml( $text, $isTitle = false ) {
3610                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
3611         }
3612
3613         /**
3614          * @param $key string
3615          * @return string
3616          */
3617         public function convertCategoryKey( $key ) {
3618                 return $this->mConverter->convertCategoryKey( $key );
3619         }
3620
3621         /**
3622          * Get the list of variants supported by this language
3623          * see sample implementation in LanguageZh.php
3624          *
3625          * @return array an array of language codes
3626          */
3627         public function getVariants() {
3628                 return $this->mConverter->getVariants();
3629         }
3630
3631         /**
3632          * @return string
3633          */
3634         public function getPreferredVariant() {
3635                 return $this->mConverter->getPreferredVariant();
3636         }
3637
3638         /**
3639          * @return string
3640          */
3641         public function getDefaultVariant() {
3642                 return $this->mConverter->getDefaultVariant();
3643         }
3644
3645         /**
3646          * @return string
3647          */
3648         public function getURLVariant() {
3649                 return $this->mConverter->getURLVariant();
3650         }
3651
3652         /**
3653          * If a language supports multiple variants, it is
3654          * possible that non-existing link in one variant
3655          * actually exists in another variant. this function
3656          * tries to find it. See e.g. LanguageZh.php
3657          *
3658          * @param $link String: the name of the link
3659          * @param $nt Mixed: the title object of the link
3660          * @param $ignoreOtherCond Boolean: to disable other conditions when
3661          *      we need to transclude a template or update a category's link
3662          * @return null the input parameters may be modified upon return
3663          */
3664         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
3665                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
3666         }
3667
3668         /**
3669          * If a language supports multiple variants, converts text
3670          * into an array of all possible variants of the text:
3671          *  'variant' => text in that variant
3672          *
3673          * @deprecated since 1.17 Use autoConvertToAllVariants()
3674          *
3675          * @param $text string
3676          *
3677          * @return string
3678          */
3679         public function convertLinkToAllVariants( $text ) {
3680                 return $this->mConverter->convertLinkToAllVariants( $text );
3681         }
3682
3683         /**
3684          * returns language specific options used by User::getPageRenderHash()
3685          * for example, the preferred language variant
3686          *
3687          * @return string
3688          */
3689         function getExtraHashOptions() {
3690                 return $this->mConverter->getExtraHashOptions();
3691         }
3692
3693         /**
3694          * For languages that support multiple variants, the title of an
3695          * article may be displayed differently in different variants. this
3696          * function returns the apporiate title defined in the body of the article.
3697          *
3698          * @return string
3699          */
3700         public function getParsedTitle() {
3701                 return $this->mConverter->getParsedTitle();
3702         }
3703
3704         /**
3705          * Prepare external link text for conversion. When the text is
3706          * a URL, it shouldn't be converted, and it'll be wrapped in
3707          * the "raw" tag (-{R| }-) to prevent conversion.
3708          *
3709          * This function is called "markNoConversion" for historical
3710          * reasons.
3711          *
3712          * @param $text String: text to be used for external link
3713          * @param $noParse bool: wrap it without confirming it's a real URL first
3714          * @return string the tagged text
3715          */
3716         public function markNoConversion( $text, $noParse = false ) {
3717                 // Excluding protocal-relative URLs may avoid many false positives.
3718                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
3719                         return $this->mConverter->markNoConversion( $text );
3720                 } else {
3721                         return $text;
3722                 }
3723         }
3724
3725         /**
3726          * A regular expression to match legal word-trailing characters
3727          * which should be merged onto a link of the form [[foo]]bar.
3728          *
3729          * @return string
3730          */
3731         public function linkTrail() {
3732                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
3733         }
3734
3735         /**
3736          * @return Language
3737          */
3738         function getLangObj() {
3739                 return $this;
3740         }
3741
3742         /**
3743          * Get the RFC 3066 code for this language object
3744          *
3745          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
3746          * htmlspecialchars() or similar
3747          *
3748          * @return string
3749          */
3750         public function getCode() {
3751                 return $this->mCode;
3752         }
3753
3754         /**
3755          * Get the code in Bcp47 format which we can use
3756          * inside of html lang="" tags.
3757          *
3758          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
3759          * htmlspecialchars() or similar.
3760          *
3761          * @since 1.19
3762          * @return string
3763          */
3764         public function getHtmlCode() {
3765                 if ( is_null( $this->mHtmlCode ) ) {
3766                         $this->mHtmlCode = wfBCP47( $this->getCode() );
3767                 }
3768                 return $this->mHtmlCode;
3769         }
3770
3771         /**
3772          * @param $code string
3773          */
3774         public function setCode( $code ) {
3775                 $this->mCode = $code;
3776                 // Ensure we don't leave an incorrect html code lying around
3777                 $this->mHtmlCode = null;
3778         }
3779
3780         /**
3781          * Get the name of a file for a certain language code
3782          * @param $prefix string Prepend this to the filename
3783          * @param $code string Language code
3784          * @param $suffix string Append this to the filename
3785          * @throws MWException
3786          * @return string $prefix . $mangledCode . $suffix
3787          */
3788         public static function getFileName( $prefix = 'Language', $code, $suffix = '.php' ) {
3789                 // Protect against path traversal
3790                 if ( !Language::isValidCode( $code )
3791                         || strcspn( $code, ":/\\\000" ) !== strlen( $code ) )
3792                 {
3793                         throw new MWException( "Invalid language code \"$code\"" );
3794                 }
3795
3796                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
3797         }
3798
3799         /**
3800          * Get the language code from a file name. Inverse of getFileName()
3801          * @param $filename string $prefix . $languageCode . $suffix
3802          * @param $prefix string Prefix before the language code
3803          * @param $suffix string Suffix after the language code
3804          * @return string Language code, or false if $prefix or $suffix isn't found
3805          */
3806         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
3807                 $m = null;
3808                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
3809                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
3810                 if ( !count( $m ) ) {
3811                         return false;
3812                 }
3813                 return str_replace( '_', '-', strtolower( $m[1] ) );
3814         }
3815
3816         /**
3817          * @param $code string
3818          * @return string
3819          */
3820         public static function getMessagesFileName( $code ) {
3821                 global $IP;
3822                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
3823                 wfRunHooks( 'Language::getMessagesFileName', array( $code, &$file ) );
3824                 return $file;
3825         }
3826
3827         /**
3828          * @param $code string
3829          * @return string
3830          */
3831         public static function getClassFileName( $code ) {
3832                 global $IP;
3833                 return self::getFileName( "$IP/languages/classes/Language", $code, '.php' );
3834         }
3835
3836         /**
3837          * Get the first fallback for a given language.
3838          *
3839          * @param $code string
3840          *
3841          * @return bool|string
3842          */
3843         public static function getFallbackFor( $code ) {
3844                 if ( $code === 'en' || !Language::isValidBuiltInCode( $code ) ) {
3845                         return false;
3846                 } else {
3847                         $fallbacks = self::getFallbacksFor( $code );
3848                         $first = array_shift( $fallbacks );
3849                         return $first;
3850                 }
3851         }
3852
3853         /**
3854          * Get the ordered list of fallback languages.
3855          *
3856          * @since 1.19
3857          * @param $code string Language code
3858          * @return array
3859          */
3860         public static function getFallbacksFor( $code ) {
3861                 if ( $code === 'en' || !Language::isValidBuiltInCode( $code ) ) {
3862                         return array();
3863                 } else {
3864                         $v = self::getLocalisationCache()->getItem( $code, 'fallback' );
3865                         $v = array_map( 'trim', explode( ',', $v ) );
3866                         if ( $v[count( $v ) - 1] !== 'en' ) {
3867                                 $v[] = 'en';
3868                         }
3869                         return $v;
3870                 }
3871         }
3872
3873         /**
3874          * Get all messages for a given language
3875          * WARNING: this may take a long time. If you just need all message *keys*
3876          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
3877          *
3878          * @param $code string
3879          *
3880          * @return array
3881          */
3882         public static function getMessagesFor( $code ) {
3883                 return self::getLocalisationCache()->getItem( $code, 'messages' );
3884         }
3885
3886         /**
3887          * Get a message for a given language
3888          *
3889          * @param $key string
3890          * @param $code string
3891          *
3892          * @return string
3893          */
3894         public static function getMessageFor( $key, $code ) {
3895                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
3896         }
3897
3898         /**
3899          * Get all message keys for a given language. This is a faster alternative to
3900          * array_keys( Language::getMessagesFor( $code ) )
3901          *
3902          * @since 1.19
3903          * @param $code string Language code
3904          * @return array of message keys (strings)
3905          */
3906         public static function getMessageKeysFor( $code ) {
3907                 return self::getLocalisationCache()->getSubItemList( $code, 'messages' );
3908         }
3909
3910         /**
3911          * @param $talk
3912          * @return mixed
3913          */
3914         function fixVariableInNamespace( $talk ) {
3915                 if ( strpos( $talk, '$1' ) === false ) {
3916                         return $talk;
3917                 }
3918
3919                 global $wgMetaNamespace;
3920                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
3921
3922                 # Allow grammar transformations
3923                 # Allowing full message-style parsing would make simple requests
3924                 # such as action=raw much more expensive than they need to be.
3925                 # This will hopefully cover most cases.
3926                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
3927                         array( &$this, 'replaceGrammarInNamespace' ), $talk );
3928                 return str_replace( ' ', '_', $talk );
3929         }
3930
3931         /**
3932          * @param $m string
3933          * @return string
3934          */
3935         function replaceGrammarInNamespace( $m ) {
3936                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
3937         }
3938
3939         /**
3940          * @throws MWException
3941          * @return array
3942          */
3943         static function getCaseMaps() {
3944                 static $wikiUpperChars, $wikiLowerChars;
3945                 if ( isset( $wikiUpperChars ) ) {
3946                         return array( $wikiUpperChars, $wikiLowerChars );
3947                 }
3948
3949                 wfProfileIn( __METHOD__ );
3950                 $arr = wfGetPrecompiledData( 'Utf8Case.ser' );
3951                 if ( $arr === false ) {
3952                         throw new MWException(
3953                                 "Utf8Case.ser is missing, please run \"make\" in the serialized directory\n" );
3954                 }
3955                 $wikiUpperChars = $arr['wikiUpperChars'];
3956                 $wikiLowerChars = $arr['wikiLowerChars'];
3957                 wfProfileOut( __METHOD__ );
3958                 return array( $wikiUpperChars, $wikiLowerChars );
3959         }
3960
3961         /**
3962          * Decode an expiry (block, protection, etc) which has come from the DB
3963          *
3964          * @todo FIXME: why are we returnings DBMS-dependent strings???
3965          *
3966          * @param $expiry String: Database expiry String
3967          * @param $format Bool|Int true to process using language functions, or TS_ constant
3968          *     to return the expiry in a given timestamp
3969          * @return String
3970          * @since 1.18
3971          */
3972         public function formatExpiry( $expiry, $format = true ) {
3973                 static $infinity, $infinityMsg;
3974                 if ( $infinity === null ) {
3975                         $infinityMsg = wfMessage( 'infiniteblock' );
3976                         $infinity = wfGetDB( DB_SLAVE )->getInfinity();
3977                 }
3978
3979                 if ( $expiry == '' || $expiry == $infinity ) {
3980                         return $format === true
3981                                 ? $infinityMsg
3982                                 : $infinity;
3983                 } else {
3984                         return $format === true
3985                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
3986                                 : wfTimestamp( $format, $expiry );
3987                 }
3988         }
3989
3990         /**
3991          * @todo Document
3992          * @param $seconds int|float
3993          * @param $format Array Optional
3994          *              If $format['avoid'] == 'avoidseconds' - don't mention seconds if $seconds >= 1 hour
3995          *              If $format['avoid'] == 'avoidminutes' - don't mention seconds/minutes if $seconds > 48 hours
3996          *              If $format['noabbrevs'] is true - use 'seconds' and friends instead of 'seconds-abbrev' and friends
3997          *              For backwards compatibility, $format may also be one of the strings 'avoidseconds' or 'avoidminutes'
3998          * @return string
3999          */
4000         function formatTimePeriod( $seconds, $format = array() ) {
4001                 if ( !is_array( $format ) ) {
4002                         $format = array( 'avoid' => $format ); // For backwards compatibility
4003                 }
4004                 if ( !isset( $format['avoid'] ) ) {
4005                         $format['avoid'] = false;
4006                 }
4007                 if ( !isset( $format['noabbrevs' ] ) ) {
4008                         $format['noabbrevs'] = false;
4009                 }
4010                 $secondsMsg = wfMessage(
4011                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4012                 $minutesMsg = wfMessage(
4013                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4014                 $hoursMsg = wfMessage(
4015                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4016                 $daysMsg = wfMessage(
4017                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4018
4019                 if ( round( $seconds * 10 ) < 100 ) {
4020                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4021                         $s = $secondsMsg->params( $s )->text();
4022                 } elseif ( round( $seconds ) < 60 ) {
4023                         $s = $this->formatNum( round( $seconds ) );
4024                         $s = $secondsMsg->params( $s )->text();
4025                 } elseif ( round( $seconds ) < 3600 ) {
4026                         $minutes = floor( $seconds / 60 );
4027                         $secondsPart = round( fmod( $seconds, 60 ) );
4028                         if ( $secondsPart == 60 ) {
4029                                 $secondsPart = 0;
4030                                 $minutes++;
4031                         }
4032                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4033                         $s .= ' ';
4034                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4035                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4036                         $hours = floor( $seconds / 3600 );
4037                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4038                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4039                         if ( $secondsPart == 60 ) {
4040                                 $secondsPart = 0;
4041                                 $minutes++;
4042                         }
4043                         if ( $minutes == 60 ) {
4044                                 $minutes = 0;
4045                                 $hours++;
4046                         }
4047                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4048                         $s .= ' ';
4049                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4050                         if ( !in_array( $format['avoid'], array( 'avoidseconds', 'avoidminutes' ) ) ) {
4051                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4052                         }
4053                 } else {
4054                         $days = floor( $seconds / 86400 );
4055                         if ( $format['avoid'] === 'avoidminutes' ) {
4056                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4057                                 if ( $hours == 24 ) {
4058                                         $hours = 0;
4059                                         $days++;
4060                                 }
4061                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4062                                 $s .= ' ';
4063                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4064                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4065                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4066                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4067                                 if ( $minutes == 60 ) {
4068                                         $minutes = 0;
4069                                         $hours++;
4070                                 }
4071                                 if ( $hours == 24 ) {
4072                                         $hours = 0;
4073                                         $days++;
4074                                 }
4075                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4076                                 $s .= ' ';
4077                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4078                                 $s .= ' ';
4079                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4080                         } else {
4081                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4082                                 $s .= ' ';
4083                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4084                         }
4085                 }
4086                 return $s;
4087         }
4088
4089         /**
4090          * Format a bitrate for output, using an appropriate
4091          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to the magnitude in question
4092          *
4093          * This use base 1000. For base 1024 use formatSize(), for another base
4094          * see formatComputingNumbers()
4095          *
4096          * @param $bps int
4097          * @return string
4098          */
4099         function formatBitrate( $bps ) {
4100                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4101         }
4102
4103         /**
4104          * @param $size int Size of the unit
4105          * @param $boundary int Size boundary (1000, or 1024 in most cases)
4106          * @param $messageKey string Message key to be uesd
4107          * @return string
4108          */
4109         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4110                 if ( $size <= 0 ) {
4111                         return str_replace( '$1', $this->formatNum( $size ),
4112                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4113                         );
4114                 }
4115                 $sizes = array( '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' );
4116                 $index = 0;
4117
4118                 $maxIndex = count( $sizes ) - 1;
4119                 while ( $size >= $boundary && $index < $maxIndex ) {
4120                         $index++;
4121                         $size /= $boundary;
4122                 }
4123
4124                 // For small sizes no decimal places necessary
4125                 $round = 0;
4126                 if ( $index > 1 ) {
4127                         // For MB and bigger two decimal places are smarter
4128                         $round = 2;
4129                 }
4130                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4131
4132                 $size = round( $size, $round );
4133                 $text = $this->getMessageFromDB( $msg );
4134                 return str_replace( '$1', $this->formatNum( $size ), $text );
4135         }
4136
4137         /**
4138          * Format a size in bytes for output, using an appropriate
4139          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4140          *
4141          * This method use base 1024. For base 1000 use formatBitrate(), for
4142          * another base see formatComputingNumbers()
4143          *
4144          * @param $size int Size to format
4145          * @return string Plain text (not HTML)
4146          */
4147         function formatSize( $size ) {
4148                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4149         }
4150
4151         /**
4152          * Make a list item, used by various special pages
4153          *
4154          * @param $page String Page link
4155          * @param $details String Text between brackets
4156          * @param $oppositedm Boolean Add the direction mark opposite to your
4157          *                    language, to display text properly
4158          * @return String
4159          */
4160         function specialList( $page, $details, $oppositedm = true ) {
4161                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) .
4162                         $this->getDirMark();
4163                 $details = $details ? $dirmark . $this->getMessageFromDB( 'word-separator' ) .
4164                         wfMessage( 'parentheses' )->rawParams( $details )->inLanguage( $this )->escaped() : '';
4165                 return $page . $details;
4166         }
4167
4168         /**
4169          * Generate (prev x| next x) (20|50|100...) type links for paging
4170          *
4171          * @param $title Title object to link
4172          * @param $offset Integer offset parameter
4173          * @param $limit Integer limit parameter
4174          * @param $query array|String optional URL query parameter string
4175          * @param $atend Bool optional param for specified if this is the last page
4176          * @return String
4177          */
4178         public function viewPrevNext( Title $title, $offset, $limit, array $query = array(), $atend = false ) {
4179                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4180
4181                 # Make 'previous' link
4182                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4183                 if ( $offset > 0 ) {
4184                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4185                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4186                 } else {
4187                         $plink = htmlspecialchars( $prev );
4188                 }
4189
4190                 # Make 'next' link
4191                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4192                 if ( $atend ) {
4193                         $nlink = htmlspecialchars( $next );
4194                 } else {
4195                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4196                                 $query, $next, 'prevn-title', 'mw-nextlink' );
4197                 }
4198
4199                 # Make links to set number of items per page
4200                 $numLinks = array();
4201                 foreach ( array( 20, 50, 100, 250, 500 ) as $num ) {
4202                         $numLinks[] = $this->numLink( $title, $offset, $num,
4203                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4204                 }
4205
4206                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4207                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4208         }
4209
4210         /**
4211          * Helper function for viewPrevNext() that generates links
4212          *
4213          * @param $title Title object to link
4214          * @param $offset Integer offset parameter
4215          * @param $limit Integer limit parameter
4216          * @param $query Array extra query parameters
4217          * @param $link String text to use for the link; will be escaped
4218          * @param $tooltipMsg String name of the message to use as tooltip
4219          * @param $class String value of the "class" attribute of the link
4220          * @return String HTML fragment
4221          */
4222         private function numLink( Title $title, $offset, $limit, array $query, $link, $tooltipMsg, $class ) {
4223                 $query = array( 'limit' => $limit, 'offset' => $offset ) + $query;
4224                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4225                 return Html::element( 'a', array( 'href' => $title->getLocalURL( $query ),
4226                         'title' => $tooltip, 'class' => $class ), $link );
4227         }
4228
4229         /**
4230          * Get the conversion rule title, if any.
4231          *
4232          * @return string
4233          */
4234         public function getConvRuleTitle() {
4235                 return $this->mConverter->getConvRuleTitle();
4236         }
4237
4238         /**
4239          * Get the compiled plural rules for the language
4240          * @since 1.20
4241          * @return array Associative array with plural form, and plural rule as key-value pairs
4242          */
4243         public function getCompiledPluralRules() {
4244                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4245                 $fallbacks = Language::getFallbacksFor( $this->mCode );
4246                 if ( !$pluralRules ) {
4247                         foreach ( $fallbacks as $fallbackCode ) {
4248                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4249                                 if ( $pluralRules ) {
4250                                         break;
4251                                 }
4252                         }
4253                 }
4254                 return $pluralRules;
4255         }
4256
4257         /**
4258          * Get the plural rules for the language
4259          * @since 1.20
4260          * @return array Associative array with plural form, and plural rule as key-value pairs
4261          */
4262         public function getPluralRules() {
4263                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4264                 $fallbacks = Language::getFallbacksFor( $this->mCode );
4265                 if ( !$pluralRules ) {
4266                         foreach ( $fallbacks as $fallbackCode ) {
4267                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4268                                 if ( $pluralRules ) {
4269                                         break;
4270                                 }
4271                         }
4272                 }
4273                 return $pluralRules;
4274         }
4275
4276         /**
4277          * Find the plural form matching to the given number
4278          * It return the form index.
4279          * @return int The index of the plural form
4280          */
4281         private function getPluralForm( $number ) {
4282                 $pluralRules = $this->getCompiledPluralRules();
4283                 $form = CLDRPluralRuleEvaluator::evaluateCompiled( $number, $pluralRules );
4284                 return $form;
4285         }
4286
4287 }