languages/Language.php

   1 <?php
   2 /**
   3  * Internationalisation code
   4  *
   5  * @file
   6  * @ingroup Language
   7  */
   8
   9 /**
  10  * @defgroup Language Language
  11  */
  12
  13 if ( !defined( 'MEDIAWIKI' ) ) {
  14         echo "This file is part of MediaWiki, it is not a valid entry point.\n";
  15         exit( 1 );
  16 }
  17
  18 # Read language names
  19 global $wgLanguageNames;
  20 require_once( dirname( __FILE__ ) . '/Names.php' );
  21
  22 global $wgInputEncoding, $wgOutputEncoding;
  23
  24 /**
  25  * These are always UTF-8, they exist only for backwards compatibility
  26  */
  27 $wgInputEncoding    = 'UTF-8';
  28 $wgOutputEncoding       = 'UTF-8';
  29
  30 if ( function_exists( 'mb_strtoupper' ) ) {
  31         mb_internal_encoding( 'UTF-8' );
  32 }
  33
  34 /**
  35  * a fake language converter
  36  *
  37  * @ingroup Language
  38  */
  39 class FakeConverter {
  40         var $mLang;
  41         function __construct( $langobj ) { $this->mLang = $langobj; }
  42         function autoConvertToAllVariants( $text ) { return array( $this->mLang->getCode() => $text ); }
  43         function convert( $t ) { return $t; }
  44         function convertTitle( $t ) { return $t->getPrefixedText(); }
  45         function getVariants() { return array( $this->mLang->getCode() ); }
  46         function getPreferredVariant() { return $this->mLang->getCode(); }
  47         function getConvRuleTitle() { return false; }
  48         function findVariantLink( &$l, &$n, $ignoreOtherCond = false ) { }
  49         function getExtraHashOptions() { return ''; }
  50         function getParsedTitle() { return ''; }
  51         function markNoConversion( $text, $noParse = false ) { return $text; }
  52         function convertCategoryKey( $key ) { return $key; }
  53         function convertLinkToAllVariants( $text ) { return autoConvertToAllVariants( $text ); }
  54         function armourMath( $text ) { return $text; }
  55 }
  56
  57 /**
  58  * Internationalisation code
  59  * @ingroup Language
  60  */
  61 class Language {
  62         var $mConverter, $mVariants, $mCode, $mLoaded = false;
  63         var $mMagicExtensions = array(), $mMagicHookDone = false;
  64
  65         var $mNamespaceIds, $namespaceNames, $namespaceAliases;
  66         var $dateFormatStrings = array();
  67         var $mExtendedSpecialPageAliases;
  68
  69         /**
  70          * ReplacementArray object caches
  71          */
  72         var $transformData = array();
  73
  74         static public $dataCache;
  75         static public $mLangObjCache = array();
  76
  77         static public $mWeekdayMsgs = array(
  78                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  79                 'friday', 'saturday'
  80         );
  81
  82         static public $mWeekdayAbbrevMsgs = array(
  83                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
  84         );
  85
  86         static public $mMonthMsgs = array(
  87                 'january', 'february', 'march', 'april', 'may_long', 'june',
  88                 'july', 'august', 'september', 'october', 'november',
  89                 'december'
  90         );
  91         static public $mMonthGenMsgs = array(
  92                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
  93                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
  94                 'december-gen'
  95         );
  96         static public $mMonthAbbrevMsgs = array(
  97                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
  98                 'sep', 'oct', 'nov', 'dec'
  99         );
 100
 101         static public $mIranianCalendarMonthMsgs = array(
 102                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 103                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 104                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 105                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 106         );
 107
 108         static public $mHebrewCalendarMonthMsgs = array(
 109                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 110                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 111                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 112                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 113                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 114         );
 115
 116         static public $mHebrewCalendarMonthGenMsgs = array(
 117                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 118                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 119                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 120                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 121                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 122         );
 123
 124         static public $mHijriCalendarMonthMsgs = array(
 125                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 126                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 127                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 128                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 129         );
 130
 131         /**
 132          * Get a cached language object for a given language code
 133          */
 134         static function factory( $code ) {
 135                 if ( !isset( self::$mLangObjCache[$code] ) ) {
 136                         if ( count( self::$mLangObjCache ) > 10 ) {
 137                                 // Don't keep a billion objects around, that's stupid.
 138                                 self::$mLangObjCache = array();
 139                         }
 140                         self::$mLangObjCache[$code] = self::newFromCode( $code );
 141                 }
 142                 return self::$mLangObjCache[$code];
 143         }
 144
 145         /**
 146          * Create a language object for a given language code
 147          */
 148         protected static function newFromCode( $code ) {
 149                 global $IP;
 150                 static $recursionLevel = 0;
 151                 if ( $code == 'en' ) {
 152                         $class = 'Language';
 153                 } else {
 154                         $class = 'Language' . str_replace( '-', '_', ucfirst( $code ) );
 155                         // Preload base classes to work around APC/PHP5 bug
 156                         if ( file_exists( "$IP/languages/classes/$class.deps.php" ) ) {
 157                                 include_once( "$IP/languages/classes/$class.deps.php" );
 158                         }
 159                         if ( file_exists( "$IP/languages/classes/$class.php" ) ) {
 160                                 include_once( "$IP/languages/classes/$class.php" );
 161                         }
 162                 }
 163
 164                 if ( $recursionLevel > 5 ) {
 165                         throw new MWException( "Language fallback loop detected when creating class $class\n" );
 166                 }
 167
 168                 if ( !class_exists( $class ) ) {
 169                         $fallback = Language::getFallbackFor( $code );
 170                         ++$recursionLevel;
 171                         $lang = Language::newFromCode( $fallback );
 172                         --$recursionLevel;
 173                         $lang->setCode( $code );
 174                 } else {
 175                         $lang = new $class;
 176                 }
 177                 return $lang;
 178         }
 179
 180         /**
 181          * Get the LocalisationCache instance
 182          */
 183         public static function getLocalisationCache() {
 184                 if ( is_null( self::$dataCache ) ) {
 185                         global $wgLocalisationCacheConf;
 186                         $class = $wgLocalisationCacheConf['class'];
 187                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 188                 }
 189                 return self::$dataCache;
 190         }
 191
 192         function __construct() {
 193                 $this->mConverter = new FakeConverter( $this );
 194                 // Set the code to the name of the descendant
 195                 if ( get_class( $this ) == 'Language' ) {
 196                         $this->mCode = 'en';
 197                 } else {
 198                         $this->mCode = str_replace( '_', '-', strtolower( substr( get_class( $this ), 8 ) ) );
 199                 }
 200                 self::getLocalisationCache();
 201         }
 202
 203         /**
 204          * Reduce memory usage
 205          */
 206         function __destruct() {
 207                 foreach ( $this as $name => $value ) {
 208                         unset( $this->$name );
 209                 }
 210         }
 211
 212         /**
 213          * Hook which will be called if this is the content language.
 214          * Descendants can use this to register hook functions or modify globals
 215          */
 216         function initContLang() { }
 217
 218         /**
 219          * @deprecated Use User::getDefaultOptions()
 220          * @return array
 221          */
 222         function getDefaultUserOptions() {
 223                 wfDeprecated( __METHOD__ );
 224                 return User::getDefaultOptions();
 225         }
 226
 227         function getFallbackLanguageCode() {
 228                 if ( $this->mCode === 'en' ) {
 229                         return false;
 230                 } else {
 231                         return self::$dataCache->getItem( $this->mCode, 'fallback' );
 232                 }
 233         }
 234
 235         /**
 236          * Exports $wgBookstoreListEn
 237          * @return array
 238          */
 239         function getBookstoreList() {
 240                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 241         }
 242
 243         /**
 244          * @return array
 245          */
 246         function getNamespaces() {
 247                 if ( is_null( $this->namespaceNames ) ) {
 248                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 249
 250                         $this->namespaceNames = self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 251                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 252
 253                         $this->namespaceNames = $wgExtraNamespaces + $this->namespaceNames + $validNamespaces;
 254
 255                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 256                         if ( $wgMetaNamespaceTalk ) {
 257                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 258                         } else {
 259                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 260                                 $this->namespaceNames[NS_PROJECT_TALK] =
 261                                         $this->fixVariableInNamespace( $talk );
 262                         }
 263
 264                         # Sometimes a language will be localised but not actually exist on this wiki.
 265                         foreach( $this->namespaceNames as $key => $text ) {
 266                                 if ( !isset( $validNamespaces[$key] ) ) {
 267                                         unset( $this->namespaceNames[$key] );
 268                                 }
 269                         }
 270
 271                         # The above mixing may leave namespaces out of canonical order.
 272                         # Re-order by namespace ID number...
 273                         ksort( $this->namespaceNames );
 274                 }
 275                 return $this->namespaceNames;
 276         }
 277
 278         /**
 279          * A convenience function that returns the same thing as
 280          * getNamespaces() except with the array values changed to ' '
 281          * where it found '_', useful for producing output to be displayed
 282          * e.g. in <select> forms.
 283          *
 284          * @return array
 285          */
 286         function getFormattedNamespaces() {
 287                 $ns = $this->getNamespaces();
 288                 foreach ( $ns as $k => $v ) {
 289                         $ns[$k] = strtr( $v, '_', ' ' );
 290                 }
 291                 return $ns;
 292         }
 293
 294         /**
 295          * Get a namespace value by key
 296          * <code>
 297          * $mw_ns = $wgContLang->getNsText( NS_MEDIAWIKI );
 298          * echo $mw_ns; // prints 'MediaWiki'
 299          * </code>
 300          *
 301          * @param $index Int: the array key of the namespace to return
 302          * @return mixed, string if the namespace value exists, otherwise false
 303          */
 304         function getNsText( $index ) {
 305                 $ns = $this->getNamespaces();
 306                 return isset( $ns[$index] ) ? $ns[$index] : false;
 307         }
 308
 309         /**
 310          * A convenience function that returns the same thing as
 311          * getNsText() except with '_' changed to ' ', useful for
 312          * producing output.
 313          *
 314          * @return array
 315          */
 316         function getFormattedNsText( $index ) {
 317                 $ns = $this->getNsText( $index );
 318                 return strtr( $ns, '_', ' ' );
 319         }
 320
 321         /**
 322          * Get a namespace key by value, case insensitive.
 323          * Only matches namespace names for the current language, not the
 324          * canonical ones defined in Namespace.php.
 325          *
 326          * @param $text String
 327          * @return mixed An integer if $text is a valid value otherwise false
 328          */
 329         function getLocalNsIndex( $text ) {
 330                 $lctext = $this->lc( $text );
 331                 $ids = $this->getNamespaceIds();
 332                 return isset( $ids[$lctext] ) ? $ids[$lctext] : false;
 333         }
 334
 335         function getNamespaceAliases() {
 336                 if ( is_null( $this->namespaceAliases ) ) {
 337                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 338                         if ( !$aliases ) {
 339                                 $aliases = array();
 340                         } else {
 341                                 foreach ( $aliases as $name => $index ) {
 342                                         if ( $index === NS_PROJECT_TALK ) {
 343                                                 unset( $aliases[$name] );
 344                                                 $name = $this->fixVariableInNamespace( $name );
 345                                                 $aliases[$name] = $index;
 346                                         }
 347                                 }
 348                         }
 349                         $this->namespaceAliases = $aliases;
 350                 }
 351                 return $this->namespaceAliases;
 352         }
 353
 354         function getNamespaceIds() {
 355                 if ( is_null( $this->mNamespaceIds ) ) {
 356                         global $wgNamespaceAliases;
 357                         # Put namespace names and aliases into a hashtable.
 358                         # If this is too slow, then we should arrange it so that it is done
 359                         # before caching. The catch is that at pre-cache time, the above
 360                         # class-specific fixup hasn't been done.
 361                         $this->mNamespaceIds = array();
 362                         foreach ( $this->getNamespaces() as $index => $name ) {
 363                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 364                         }
 365                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 366                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 367                         }
 368                         if ( $wgNamespaceAliases ) {
 369                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 370                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 371                                 }
 372                         }
 373                 }
 374                 return $this->mNamespaceIds;
 375         }
 376
 377
 378         /**
 379          * Get a namespace key by value, case insensitive.  Canonical namespace
 380          * names override custom ones defined for the current language.
 381          *
 382          * @param $text String
 383          * @return mixed An integer if $text is a valid value otherwise false
 384          */
 385         function getNsIndex( $text ) {
 386                 $lctext = $this->lc( $text );
 387                 if ( ( $ns = MWNamespace::getCanonicalIndex( $lctext ) ) !== null ) {
 388                         return $ns;
 389                 }
 390                 $ids = $this->getNamespaceIds();
 391                 return isset( $ids[$lctext] ) ? $ids[$lctext] : false;
 392         }
 393
 394         /**
 395          * short names for language variants used for language conversion links.
 396          *
 397          * @param $code String
 398          * @return string
 399          */
 400         function getVariantname( $code ) {
 401                 return $this->getMessageFromDB( "variantname-$code" );
 402         }
 403
 404         function specialPage( $name ) {
 405                 $aliases = $this->getSpecialPageAliases();
 406                 if ( isset( $aliases[$name][0] ) ) {
 407                         $name = $aliases[$name][0];
 408                 }
 409                 return $this->getNsText( NS_SPECIAL ) . ':' . $name;
 410         }
 411
 412         function getQuickbarSettings() {
 413                 return array(
 414                         $this->getMessage( 'qbsettings-none' ),
 415                         $this->getMessage( 'qbsettings-fixedleft' ),
 416                         $this->getMessage( 'qbsettings-fixedright' ),
 417                         $this->getMessage( 'qbsettings-floatingleft' ),
 418                         $this->getMessage( 'qbsettings-floatingright' )
 419                 );
 420         }
 421
 422         function getMathNames() {
 423                 return self::$dataCache->getItem( $this->mCode, 'mathNames' );
 424         }
 425
 426         function getDatePreferences() {
 427                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 428         }
 429
 430         function getDateFormats() {
 431                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 432         }
 433
 434         function getDefaultDateFormat() {
 435                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 436                 if ( $df === 'dmy or mdy' ) {
 437                         global $wgAmericanDates;
 438                         return $wgAmericanDates ? 'mdy' : 'dmy';
 439                 } else {
 440                         return $df;
 441                 }
 442         }
 443
 444         function getDatePreferenceMigrationMap() {
 445                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 446         }
 447
 448         function getImageFile( $image ) {
 449                 return self::$dataCache->getSubitem( $this->mCode, 'imageFiles', $image );
 450         }
 451
 452         function getDefaultUserOptionOverrides() {
 453                 return self::$dataCache->getItem( $this->mCode, 'defaultUserOptionOverrides' );
 454         }
 455
 456         function getExtraUserToggles() {
 457                 return self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 458         }
 459
 460         function getUserToggle( $tog ) {
 461                 return $this->getMessageFromDB( "tog-$tog" );
 462         }
 463
 464         /**
 465          * Get language names, indexed by code.
 466          * If $customisedOnly is true, only returns codes with a messages file
 467          */
 468         public static function getLanguageNames( $customisedOnly = false ) {
 469                 global $wgLanguageNames, $wgExtraLanguageNames;
 470                 $allNames = $wgExtraLanguageNames + $wgLanguageNames;
 471                 if ( !$customisedOnly ) {
 472                         return $allNames;
 473                 }
 474
 475                 global $IP;
 476                 $names = array();
 477                 $dir = opendir( "$IP/languages/messages" );
 478                 while ( false !== ( $file = readdir( $dir ) ) ) {
 479                         $code = self::getCodeFromFileName( $file, 'Messages' );
 480                         if ( $code && isset( $allNames[$code] ) ) {
 481                                 $names[$code] = $allNames[$code];
 482                         }
 483                 }
 484                 closedir( $dir );
 485                 return $names;
 486         }
 487
 488         /**
 489          * Get a message from the MediaWiki namespace.
 490          *
 491          * @param $msg String: message name
 492          * @return string
 493          */
 494         function getMessageFromDB( $msg ) {
 495                 return wfMsgExt( $msg, array( 'parsemag', 'language' => $this ) );
 496         }
 497
 498         function getLanguageName( $code ) {
 499                 $names = self::getLanguageNames();
 500                 if ( !array_key_exists( $code, $names ) ) {
 501                         return '';
 502                 }
 503                 return $names[$code];
 504         }
 505
 506         function getMonthName( $key ) {
 507                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 508         }
 509
 510         function getMonthNameGen( $key ) {
 511                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 512         }
 513
 514         function getMonthAbbreviation( $key ) {
 515                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 516         }
 517
 518         function getWeekdayName( $key ) {
 519                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
 520         }
 521
 522         function getWeekdayAbbreviation( $key ) {
 523                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
 524         }
 525
 526         function getIranianCalendarMonthName( $key ) {
 527                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
 528         }
 529
 530         function getHebrewCalendarMonthName( $key ) {
 531                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
 532         }
 533
 534         function getHebrewCalendarMonthNameGen( $key ) {
 535                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
 536         }
 537
 538         function getHijriCalendarMonthName( $key ) {
 539                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
 540         }
 541
 542         /**
 543          * Used by date() and time() to adjust the time output.
 544          *
 545          * @param $ts Int the time in date('YmdHis') format
 546          * @param $tz Mixed: adjust the time by this amount (default false, mean we
 547          *            get user timecorrection setting)
 548          * @return int
 549          */
 550         function userAdjust( $ts, $tz = false ) {
 551                 global $wgUser, $wgLocalTZoffset;
 552
 553                 if ( $tz === false ) {
 554                         $tz = $wgUser->getOption( 'timecorrection' );
 555                 }
 556
 557                 $data = explode( '|', $tz, 3 );
 558
 559                 if ( $data[0] == 'ZoneInfo' ) {
 560                         if ( function_exists( 'timezone_open' ) && @timezone_open( $data[2] ) !== false ) {
 561                                 $date = date_create( $ts, timezone_open( 'UTC' ) );
 562                                 date_timezone_set( $date, timezone_open( $data[2] ) );
 563                                 $date = date_format( $date, 'YmdHis' );
 564                                 return $date;
 565                         }
 566                         # Unrecognized timezone, default to 'Offset' with the stored offset.
 567                         $data[0] = 'Offset';
 568                 }
 569
 570                 $minDiff = 0;
 571                 if ( $data[0] == 'System' || $tz == '' ) {
 572                         #  Global offset in minutes.
 573                         if ( isset( $wgLocalTZoffset ) ) {
 574                                 $minDiff = $wgLocalTZoffset;
 575                         }
 576                 } else if ( $data[0] == 'Offset' ) {
 577                         $minDiff = intval( $data[1] );
 578                 } else {
 579                         $data = explode( ':', $tz );
 580                         if ( count( $data ) == 2 ) {
 581                                 $data[0] = intval( $data[0] );
 582                                 $data[1] = intval( $data[1] );
 583                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
 584                                 if ( $data[0] < 0 ) {
 585                                         $minDiff = -$minDiff;
 586                                 }
 587                         } else {
 588                                 $minDiff = intval( $data[0] ) * 60;
 589                         }
 590                 }
 591
 592                 # No difference ? Return time unchanged
 593                 if ( 0 == $minDiff ) {
 594                         return $ts;
 595                 }
 596
 597                 wfSuppressWarnings(); // E_STRICT system time bitching
 598                 # Generate an adjusted date; take advantage of the fact that mktime
 599                 # will normalize out-of-range values so we don't have to split $minDiff
 600                 # into hours and minutes.
 601                 $t = mktime( (
 602                   (int)substr( $ts, 8, 2 ) ), # Hours
 603                   (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
 604                   (int)substr( $ts, 12, 2 ), # Seconds
 605                   (int)substr( $ts, 4, 2 ), # Month
 606                   (int)substr( $ts, 6, 2 ), # Day
 607                   (int)substr( $ts, 0, 4 ) ); # Year
 608
 609                 $date = date( 'YmdHis', $t );
 610                 wfRestoreWarnings();
 611
 612                 return $date;
 613         }
 614
 615         /**
 616          * This is a workalike of PHP's date() function, but with better
 617          * internationalisation, a reduced set of format characters, and a better
 618          * escaping format.
 619          *
 620          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrU. See the
 621          * PHP manual for definitions. There are a number of extensions, which
 622          * start with "x":
 623          *
 624          *    xn   Do not translate digits of the next numeric format character
 625          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
 626          *    xr   Use roman numerals for the next numeric format character
 627          *    xh   Use hebrew numerals for the next numeric format character
 628          *    xx   Literal x
 629          *    xg   Genitive month name
 630          *
 631          *    xij  j (day number) in Iranian calendar
 632          *    xiF  F (month name) in Iranian calendar
 633          *    xin  n (month number) in Iranian calendar
 634          *    xiY  Y (full year) in Iranian calendar
 635          *
 636          *    xjj  j (day number) in Hebrew calendar
 637          *    xjF  F (month name) in Hebrew calendar
 638          *    xjt  t (days in month) in Hebrew calendar
 639          *    xjx  xg (genitive month name) in Hebrew calendar
 640          *    xjn  n (month number) in Hebrew calendar
 641          *    xjY  Y (full year) in Hebrew calendar
 642          *
 643          *    xmj  j (day number) in Hijri calendar
 644          *    xmF  F (month name) in Hijri calendar
 645          *    xmn  n (month number) in Hijri calendar
 646          *    xmY  Y (full year) in Hijri calendar
 647          *
 648          *    xkY  Y (full year) in Thai solar calendar. Months and days are
 649          *                       identical to the Gregorian calendar
 650          *    xoY  Y (full year) in Minguo calendar or Juche year.
 651          *                       Months and days are identical to the
 652          *                       Gregorian calendar
 653          *    xtY  Y (full year) in Japanese nengo. Months and days are
 654          *                       identical to the Gregorian calendar
 655          *
 656          * Characters enclosed in double quotes will be considered literal (with
 657          * the quotes themselves removed). Unmatched quotes will be considered
 658          * literal quotes. Example:
 659          *
 660          * "The month is" F       => The month is January
 661          * i's"                   => 20'11"
 662          *
 663          * Backslash escaping is also supported.
 664          *
 665          * Input timestamp is assumed to be pre-normalized to the desired local
 666          * time zone, if any.
 667          *
 668          * @param $format String
 669          * @param $ts String: 14-character timestamp
 670          *      YYYYMMDDHHMMSS
 671          *      01234567890123
 672          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
 673          */
 674         function sprintfDate( $format, $ts ) {
 675                 $s = '';
 676                 $raw = false;
 677                 $roman = false;
 678                 $hebrewNum = false;
 679                 $unix = false;
 680                 $rawToggle = false;
 681                 $iranian = false;
 682                 $hebrew = false;
 683                 $hijri = false;
 684                 $thai = false;
 685                 $minguo = false;
 686                 $tenno = false;
 687                 for ( $p = 0; $p < strlen( $format ); $p++ ) {
 688                         $num = false;
 689                         $code = $format[$p];
 690                         if ( $code == 'x' && $p < strlen( $format ) - 1 ) {
 691                                 $code .= $format[++$p];
 692                         }
 693
 694                         if ( ( $code === 'xi' || $code == 'xj' || $code == 'xk' || $code == 'xm' || $code == 'xo' || $code == 'xt' ) && $p < strlen( $format ) - 1 ) {
 695                                 $code .= $format[++$p];
 696                         }
 697
 698                         switch ( $code ) {
 699                                 case 'xx':
 700                                         $s .= 'x';
 701                                         break;
 702                                 case 'xn':
 703                                         $raw = true;
 704                                         break;
 705                                 case 'xN':
 706                                         $rawToggle = !$rawToggle;
 707                                         break;
 708                                 case 'xr':
 709                                         $roman = true;
 710                                         break;
 711                                 case 'xh':
 712                                         $hebrewNum = true;
 713                                         break;
 714                                 case 'xg':
 715                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
 716                                         break;
 717                                 case 'xjx':
 718                                         if ( !$hebrew ) $hebrew = self::tsToHebrew( $ts );
 719                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
 720                                         break;
 721                                 case 'd':
 722                                         $num = substr( $ts, 6, 2 );
 723                                         break;
 724                                 case 'D':
 725                                         if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
 726                                         $s .= $this->getWeekdayAbbreviation( gmdate( 'w', $unix ) + 1 );
 727                                         break;
 728                                 case 'j':
 729                                         $num = intval( substr( $ts, 6, 2 ) );
 730                                         break;
 731                                 case 'xij':
 732                                         if ( !$iranian ) {
 733                                                 $iranian = self::tsToIranian( $ts );
 734                                         }
 735                                         $num = $iranian[2];
 736                                         break;
 737                                 case 'xmj':
 738                                         if ( !$hijri ) {
 739                                                 $hijri = self::tsToHijri( $ts );
 740                                         }
 741                                         $num = $hijri[2];
 742                                         break;
 743                                 case 'xjj':
 744                                         if ( !$hebrew ) {
 745                                                 $hebrew = self::tsToHebrew( $ts );
 746                                         }
 747                                         $num = $hebrew[2];
 748                                         break;
 749                                 case 'l':
 750                                         if ( !$unix ) {
 751                                                 $unix = wfTimestamp( TS_UNIX, $ts );
 752                                         }
 753                                         $s .= $this->getWeekdayName( gmdate( 'w', $unix ) + 1 );
 754                                         break;
 755                                 case 'N':
 756                                         if ( !$unix ) {
 757                                                 $unix = wfTimestamp( TS_UNIX, $ts );
 758                                         }
 759                                         $w = gmdate( 'w', $unix );
 760                                         $num = $w ? $w : 7;
 761                                         break;
 762                                 case 'w':
 763                                         if ( !$unix ) {
 764                                                 $unix = wfTimestamp( TS_UNIX, $ts );
 765                                         }
 766                                         $num = gmdate( 'w', $unix );
 767                                         break;
 768                                 case 'z':
 769                                         if ( !$unix ) {
 770                                                 $unix = wfTimestamp( TS_UNIX, $ts );
 771                                         }
 772                                         $num = gmdate( 'z', $unix );
 773                                         break;
 774                                 case 'W':
 775                                         if ( !$unix ) {
 776                                                 $unix = wfTimestamp( TS_UNIX, $ts );
 777                                         }
 778                                         $num = gmdate( 'W', $unix );
 779                                         break;
 780                                 case 'F':
 781                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
 782                                         break;
 783                                 case 'xiF':
 784                                         if ( !$iranian ) {
 785                                                 $iranian = self::tsToIranian( $ts );
 786                                         }
 787                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
 788                                         break;
 789                                 case 'xmF':
 790                                         if ( !$hijri ) {
 791                                                 $hijri = self::tsToHijri( $ts );
 792                                         }
 793                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
 794                                         break;
 795                                 case 'xjF':
 796                                         if ( !$hebrew ) {
 797                                                 $hebrew = self::tsToHebrew( $ts );
 798                                         }
 799                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
 800                                         break;
 801                                 case 'm':
 802                                         $num = substr( $ts, 4, 2 );
 803                                         break;
 804                                 case 'M':
 805                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
 806                                         break;
 807                                 case 'n':
 808                                         $num = intval( substr( $ts, 4, 2 ) );
 809                                         break;
 810                                 case 'xin':
 811                                         if ( !$iranian ) {
 812                                                 $iranian = self::tsToIranian( $ts );
 813                                         }
 814                                         $num = $iranian[1];
 815                                         break;
 816                                 case 'xmn':
 817                                         if ( !$hijri ) {
 818                                                 $hijri = self::tsToHijri ( $ts );
 819                                         }
 820                                         $num = $hijri[1];
 821                                         break;
 822                                 case 'xjn':
 823                                         if ( !$hebrew ) {
 824                                                 $hebrew = self::tsToHebrew( $ts );
 825                                         }
 826                                         $num = $hebrew[1];
 827                                         break;
 828                                 case 't':
 829                                         if ( !$unix ) {
 830                                                 $unix = wfTimestamp( TS_UNIX, $ts );
 831                                         }
 832                                         $num = gmdate( 't', $unix );
 833                                         break;
 834                                 case 'xjt':
 835                                         if ( !$hebrew ) {
 836                                                 $hebrew = self::tsToHebrew( $ts );
 837                                         }
 838                                         $num = $hebrew[3];
 839                                         break;
 840                                 case 'L':
 841                                         if ( !$unix ) {
 842                                                 $unix = wfTimestamp( TS_UNIX, $ts );
 843                                         }
 844                                         $num = gmdate( 'L', $unix );
 845                                         break;
 846                                 case 'o':
 847                                         if ( !$unix ) {
 848                                                 $unix = wfTimestamp( TS_UNIX, $ts );
 849                                         }
 850                                         $num = date( 'o', $unix );
 851                                         break;
 852                                 case 'Y':
 853                                         $num = substr( $ts, 0, 4 );
 854                                         break;
 855                                 case 'xiY':
 856                                         if ( !$iranian ) {
 857                                                 $iranian = self::tsToIranian( $ts );
 858                                         }
 859                                         $num = $iranian[0];
 860                                         break;
 861                                 case 'xmY':
 862                                         if ( !$hijri ) {
 863                                                 $hijri = self::tsToHijri( $ts );
 864                                         }
 865                                         $num = $hijri[0];
 866                                         break;
 867                                 case 'xjY':
 868                                         if ( !$hebrew ) {
 869                                                 $hebrew = self::tsToHebrew( $ts );
 870                                         }
 871                                         $num = $hebrew[0];
 872                                         break;
 873                                 case 'xkY':
 874                                         if ( !$thai ) {
 875                                                 $thai = self::tsToYear( $ts, 'thai' );
 876                                         }
 877                                         $num = $thai[0];
 878                                         break;
 879                                 case 'xoY':
 880                                         if ( !$minguo ) {
 881                                                 $minguo = self::tsToYear( $ts, 'minguo' );
 882                                         }
 883                                         $num = $minguo[0];
 884                                         break;
 885                                 case 'xtY':
 886                                         if ( !$tenno ) {
 887                                                 $tenno = self::tsToYear( $ts, 'tenno' );
 888                                         }
 889                                         $num = $tenno[0];
 890                                         break;
 891                                 case 'y':
 892                                         $num = substr( $ts, 2, 2 );
 893                                         break;
 894                                 case 'a':
 895                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
 896                                         break;
 897                                 case 'A':
 898                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
 899                                         break;
 900                                 case 'g':
 901                                         $h = substr( $ts, 8, 2 );
 902                                         $num = $h % 12 ? $h % 12 : 12;
 903                                         break;
 904                                 case 'G':
 905                                         $num = intval( substr( $ts, 8, 2 ) );
 906                                         break;
 907                                 case 'h':
 908                                         $h = substr( $ts, 8, 2 );
 909                                         $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
 910                                         break;
 911                                 case 'H':
 912                                         $num = substr( $ts, 8, 2 );
 913                                         break;
 914                                 case 'i':
 915                                         $num = substr( $ts, 10, 2 );
 916                                         break;
 917                                 case 's':
 918                                         $num = substr( $ts, 12, 2 );
 919                                         break;
 920                                 case 'c':
 921                                         if ( !$unix ) {
 922                                                 $unix = wfTimestamp( TS_UNIX, $ts );
 923                                         }
 924                                         $s .= gmdate( 'c', $unix );
 925                                         break;
 926                                 case 'r':
 927                                         if ( !$unix ) {
 928                                                 $unix = wfTimestamp( TS_UNIX, $ts );
 929                                         }
 930                                         $s .= gmdate( 'r', $unix );
 931                                         break;
 932                                 case 'U':
 933                                         if ( !$unix ) {
 934                                                 $unix = wfTimestamp( TS_UNIX, $ts );
 935                                         }
 936                                         $num = $unix;
 937                                         break;
 938                                 case '\\':
 939                                         # Backslash escaping
 940                                         if ( $p < strlen( $format ) - 1 ) {
 941                                                 $s .= $format[++$p];
 942                                         } else {
 943                                                 $s .= '\\';
 944                                         }
 945                                         break;
 946                                 case '"':
 947                                         # Quoted literal
 948                                         if ( $p < strlen( $format ) - 1 ) {
 949                                                 $endQuote = strpos( $format, '"', $p + 1 );
 950                                                 if ( $endQuote === false ) {
 951                                                         # No terminating quote, assume literal "
 952                                                         $s .= '"';
 953                                                 } else {
 954                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
 955                                                         $p = $endQuote;
 956                                                 }
 957                                         } else {
 958                                                 # Quote at end of string, assume literal "
 959                                                 $s .= '"';
 960                                         }
 961                                         break;
 962                                 default:
 963                                         $s .= $format[$p];
 964                         }
 965                         if ( $num !== false ) {
 966                                 if ( $rawToggle || $raw ) {
 967                                         $s .= $num;
 968                                         $raw = false;
 969                                 } elseif ( $roman ) {
 970                                         $s .= self::romanNumeral( $num );
 971                                         $roman = false;
 972                                 } elseif ( $hebrewNum ) {
 973                                         $s .= self::hebrewNumeral( $num );
 974                                         $hebrewNum = false;
 975                                 } else {
 976                                         $s .= $this->formatNum( $num, true );
 977                                 }
 978                                 $num = false;
 979                         }
 980                 }
 981                 return $s;
 982         }
 983
 984         private static $GREG_DAYS = array( 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 );
 985         private static $IRANIAN_DAYS = array( 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 );
 986         /**
 987          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
 988          * Gregorian dates to Iranian dates. Originally written in C, it
 989          * is released under the terms of GNU Lesser General Public
 990          * License. Conversion to PHP was performed by Niklas Laxström.
 991          *
 992          * Link: http://www.farsiweb.info/jalali/jalali.c
 993          */
 994         private static function tsToIranian( $ts ) {
 995                 $gy = substr( $ts, 0, 4 ) -1600;
 996                 $gm = substr( $ts, 4, 2 ) -1;
 997                 $gd = substr( $ts, 6, 2 ) -1;
 998
 999                 # Days passed from the beginning (including leap years)
1000                 $gDayNo = 365 * $gy
1001                         + floor( ( $gy + 3 ) / 4 )
1002                         - floor( ( $gy + 99 ) / 100 )
1003                         + floor( ( $gy + 399 ) / 400 );
1004
1005
1006                 // Add days of the past months of this year
1007                 for ( $i = 0; $i < $gm; $i++ ) {
1008                         $gDayNo += self::$GREG_DAYS[$i];
1009                 }
1010
1011                 // Leap years
1012                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1013                         $gDayNo++;
1014                 }
1015
1016                 // Days passed in current month
1017                 $gDayNo += $gd;
1018
1019                 $jDayNo = $gDayNo - 79;
1020
1021                 $jNp = floor( $jDayNo / 12053 );
1022                 $jDayNo %= 12053;
1023
1024                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1025                 $jDayNo %= 1461;
1026
1027                 if ( $jDayNo >= 366 ) {
1028                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1029                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1030                 }
1031
1032                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1033                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1034                 }
1035
1036                 $jm = $i + 1;
1037                 $jd = $jDayNo + 1;
1038
1039                 return array( $jy, $jm, $jd );
1040         }
1041
1042         /**
1043          * Converting Gregorian dates to Hijri dates.
1044          *
1045          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1046          *
1047          * @link http://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1048          */
1049         private static function tsToHijri( $ts ) {
1050                 $year = substr( $ts, 0, 4 );
1051                 $month = substr( $ts, 4, 2 );
1052                 $day = substr( $ts, 6, 2 );
1053
1054                 $zyr = $year;
1055                 $zd = $day;
1056                 $zm = $month;
1057                 $zy = $zyr;
1058
1059                 if (
1060                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1061                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1062                 )
1063                 {
1064                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1065                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1066                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1067                                         $zd - 32075;
1068                 } else {
1069                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1070                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1071                 }
1072
1073                 $zl = $zjd -1948440 + 10632;
1074                 $zn = (int)( ( $zl - 1 ) / 10631 );
1075                 $zl = $zl - 10631 * $zn + 354;
1076                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) + ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1077                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) - ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1078                 $zm = (int)( ( 24 * $zl ) / 709 );
1079                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1080                 $zy = 30 * $zn + $zj - 30;
1081
1082                 return array( $zy, $zm, $zd );
1083         }
1084
1085         /**
1086          * Converting Gregorian dates to Hebrew dates.
1087          *
1088          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1089          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1090          * to translate the relevant functions into PHP and release them under
1091          * GNU GPL.
1092          *
1093          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1094          * and Adar II is 14. In a non-leap year, Adar is 6.
1095          */
1096         private static function tsToHebrew( $ts ) {
1097                 # Parse date
1098                 $year = substr( $ts, 0, 4 );
1099                 $month = substr( $ts, 4, 2 );
1100                 $day = substr( $ts, 6, 2 );
1101
1102                 # Calculate Hebrew year
1103                 $hebrewYear = $year + 3760;
1104
1105                 # Month number when September = 1, August = 12
1106                 $month += 4;
1107                 if ( $month > 12 ) {
1108                         # Next year
1109                         $month -= 12;
1110                         $year++;
1111                         $hebrewYear++;
1112                 }
1113
1114                 # Calculate day of year from 1 September
1115                 $dayOfYear = $day;
1116                 for ( $i = 1; $i < $month; $i++ ) {
1117                         if ( $i == 6 ) {
1118                                 # February
1119                                 $dayOfYear += 28;
1120                                 # Check if the year is leap
1121                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1122                                         $dayOfYear++;
1123                                 }
1124                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1125                                 $dayOfYear += 30;
1126                         } else {
1127                                 $dayOfYear += 31;
1128                         }
1129                 }
1130
1131                 # Calculate the start of the Hebrew year
1132                 $start = self::hebrewYearStart( $hebrewYear );
1133
1134                 # Calculate next year's start
1135                 if ( $dayOfYear <= $start ) {
1136                         # Day is before the start of the year - it is the previous year
1137                         # Next year's start
1138                         $nextStart = $start;
1139                         # Previous year
1140                         $year--;
1141                         $hebrewYear--;
1142                         # Add days since previous year's 1 September
1143                         $dayOfYear += 365;
1144                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1145                                 # Leap year
1146                                 $dayOfYear++;
1147                         }
1148                         # Start of the new (previous) year
1149                         $start = self::hebrewYearStart( $hebrewYear );
1150                 } else {
1151                         # Next year's start
1152                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1153                 }
1154
1155                 # Calculate Hebrew day of year
1156                 $hebrewDayOfYear = $dayOfYear - $start;
1157
1158                 # Difference between year's days
1159                 $diff = $nextStart - $start;
1160                 # Add 12 (or 13 for leap years) days to ignore the difference between
1161                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1162                 # difference is only about the year type
1163                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1164                         $diff += 13;
1165                 } else {
1166                         $diff += 12;
1167                 }
1168
1169                 # Check the year pattern, and is leap year
1170                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1171                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1172                 # and non-leap years
1173                 $yearPattern = $diff % 30;
1174                 # Check if leap year
1175                 $isLeap = $diff >= 30;
1176
1177                 # Calculate day in the month from number of day in the Hebrew year
1178                 # Don't check Adar - if the day is not in Adar, we will stop before;
1179                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1180                 $hebrewDay = $hebrewDayOfYear;
1181                 $hebrewMonth = 1;
1182                 $days = 0;
1183                 while ( $hebrewMonth <= 12 ) {
1184                         # Calculate days in this month
1185                         if ( $isLeap && $hebrewMonth == 6 ) {
1186                                 # Adar in a leap year
1187                                 if ( $isLeap ) {
1188                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1189                                         $days = 30;
1190                                         if ( $hebrewDay <= $days ) {
1191                                                 # Day in Adar I
1192                                                 $hebrewMonth = 13;
1193                                         } else {
1194                                                 # Subtract the days of Adar I
1195                                                 $hebrewDay -= $days;
1196                                                 # Try Adar II
1197                                                 $days = 29;
1198                                                 if ( $hebrewDay <= $days ) {
1199                                                         # Day in Adar II
1200                                                         $hebrewMonth = 14;
1201                                                 }
1202                                         }
1203                                 }
1204                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1205                                 # Cheshvan in a complete year (otherwise as the rule below)
1206                                 $days = 30;
1207                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1208                                 # Kislev in an incomplete year (otherwise as the rule below)
1209                                 $days = 29;
1210                         } else {
1211                                 # Odd months have 30 days, even have 29
1212                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1213                         }
1214                         if ( $hebrewDay <= $days ) {
1215                                 # In the current month
1216                                 break;
1217                         } else {
1218                                 # Subtract the days of the current month
1219                                 $hebrewDay -= $days;
1220                                 # Try in the next month
1221                                 $hebrewMonth++;
1222                         }
1223                 }
1224
1225                 return array( $hebrewYear, $hebrewMonth, $hebrewDay, $days );
1226         }
1227
1228         /**
1229          * This calculates the Hebrew year start, as days since 1 September.
1230          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1231          * Used for Hebrew date.
1232          */
1233         private static function hebrewYearStart( $year ) {
1234                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1235                 $b = intval( ( $year - 1 ) % 4 );
1236                 $m = 32.044093161144 + 1.5542417966212 * $a +  $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1237                 if ( $m < 0 ) {
1238                         $m--;
1239                 }
1240                 $Mar = intval( $m );
1241                 if ( $m < 0 ) {
1242                         $m++;
1243                 }
1244                 $m -= $Mar;
1245
1246                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1247                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1248                         $Mar++;
1249                 } else if ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1250                         $Mar += 2;
1251                 } else if ( $c == 2 || $c == 4 || $c == 6 ) {
1252                         $Mar++;
1253                 }
1254
1255                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1256                 return $Mar;
1257         }
1258
1259         /**
1260          * Algorithm to convert Gregorian dates to Thai solar dates,
1261          * Minguo dates or Minguo dates.
1262          *
1263          * Link: http://en.wikipedia.org/wiki/Thai_solar_calendar
1264          *       http://en.wikipedia.org/wiki/Minguo_calendar
1265          *       http://en.wikipedia.org/wiki/Japanese_era_name
1266          *
1267          * @param $ts String: 14-character timestamp
1268          * @param $cName String: calender name
1269          * @return Array: converted year, month, day
1270          */
1271         private static function tsToYear( $ts, $cName ) {
1272                 $gy = substr( $ts, 0, 4 );
1273                 $gm = substr( $ts, 4, 2 );
1274                 $gd = substr( $ts, 6, 2 );
1275
1276                 if ( !strcmp( $cName, 'thai' ) ) {
1277                         # Thai solar dates
1278                         # Add 543 years to the Gregorian calendar
1279                         # Months and days are identical
1280                         $gy_offset = $gy + 543;
1281                 } else if ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1282                         # Minguo dates
1283                         # Deduct 1911 years from the Gregorian calendar
1284                         # Months and days are identical
1285                         $gy_offset = $gy - 1911;
1286                 } else if ( !strcmp( $cName, 'tenno' ) ) {
1287                         # Nengō dates up to Meiji period
1288                         # Deduct years from the Gregorian calendar
1289                         # depending on the nengo periods
1290                         # Months and days are identical
1291                         if ( ( $gy < 1912 ) || ( ( $gy == 1912 ) && ( $gm < 7 ) ) || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) ) ) {
1292                                 # Meiji period
1293                                 $gy_gannen = $gy - 1868 + 1;
1294                                 $gy_offset = $gy_gannen;
1295                                 if ( $gy_gannen == 1 ) {
1296                                         $gy_offset = '元';
1297                                 }
1298                                 $gy_offset = '明治' . $gy_offset;
1299                         } else if (
1300                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1301                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1302                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1303                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1304                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1305                         )
1306                         {
1307                                 # Taishō period
1308                                 $gy_gannen = $gy - 1912 + 1;
1309                                 $gy_offset = $gy_gannen;
1310                                 if ( $gy_gannen == 1 ) {
1311                                         $gy_offset = '元';
1312                                 }
1313                                 $gy_offset = '大正' . $gy_offset;
1314                         } else if (
1315                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1316                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1317                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1318                         )
1319                         {
1320                                 # Shōwa period
1321                                 $gy_gannen = $gy - 1926 + 1;
1322                                 $gy_offset = $gy_gannen;
1323                                 if ( $gy_gannen == 1 ) {
1324                                         $gy_offset = '元';
1325                                 }
1326                                 $gy_offset = '昭和' . $gy_offset;
1327                         } else {
1328                                 # Heisei period
1329                                 $gy_gannen = $gy - 1989 + 1;
1330                                 $gy_offset = $gy_gannen;
1331                                 if ( $gy_gannen == 1 ) {
1332                                         $gy_offset = '元';
1333                                 }
1334                                 $gy_offset = '平成' . $gy_offset;
1335                         }
1336                 } else {
1337                         $gy_offset = $gy;
1338                 }
1339
1340                 return array( $gy_offset, $gm, $gd );
1341         }
1342
1343         /**
1344          * Roman number formatting up to 3000
1345          */
1346         static function romanNumeral( $num ) {
1347                 static $table = array(
1348                         array( '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ),
1349                         array( '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ),
1350                         array( '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ),
1351                         array( '', 'M', 'MM', 'MMM' )
1352                 );
1353
1354                 $num = intval( $num );
1355                 if ( $num > 3000 || $num <= 0 ) {
1356                         return $num;
1357                 }
1358
1359                 $s = '';
1360                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1361                         if ( $num >= $pow10 ) {
1362                                 $s .= $table[$i][floor( $num / $pow10 )];
1363                         }
1364                         $num = $num % $pow10;
1365                 }
1366                 return $s;
1367         }
1368
1369         /**
1370          * Hebrew Gematria number formatting up to 9999
1371          */
1372         static function hebrewNumeral( $num ) {
1373                 static $table = array(
1374                         array( '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ),
1375                         array( '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ),
1376                         array( '', 'ק', 'ר', 'ש', 'ת', 'תק', 'תר', 'תש', 'תת', 'תתק', 'תתר' ),
1377                         array( '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' )
1378                 );
1379
1380                 $num = intval( $num );
1381                 if ( $num > 9999 || $num <= 0 ) {
1382                         return $num;
1383                 }
1384
1385                 $s = '';
1386                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1387                         if ( $num >= $pow10 ) {
1388                                 if ( $num == 15 || $num == 16 ) {
1389                                         $s .= $table[0][9] . $table[0][$num - 9];
1390                                         $num = 0;
1391                                 } else {
1392                                         $s .= $table[$i][intval( ( $num / $pow10 ) )];
1393                                         if ( $pow10 == 1000 ) {
1394                                                 $s .= "'";
1395                                         }
1396                                 }
1397                         }
1398                         $num = $num % $pow10;
1399                 }
1400                 if ( strlen( $s ) == 2 ) {
1401                         $str = $s . "'";
1402                 } else  {
1403                         $str = substr( $s, 0, strlen( $s ) - 2 ) . '"';
1404                         $str .= substr( $s, strlen( $s ) - 2, 2 );
1405                 }
1406                 $start = substr( $str, 0, strlen( $str ) - 2 );
1407                 $end = substr( $str, strlen( $str ) - 2 );
1408                 switch( $end ) {
1409                         case 'כ':
1410                                 $str = $start . 'ך';
1411                                 break;
1412                         case 'מ':
1413                                 $str = $start . 'ם';
1414                                 break;
1415                         case 'נ':
1416                                 $str = $start . 'ן';
1417                                 break;
1418                         case 'פ':
1419                                 $str = $start . 'ף';
1420                                 break;
1421                         case 'צ':
1422                                 $str = $start . 'ץ';
1423                                 break;
1424                 }
1425                 return $str;
1426         }
1427
1428         /**
1429          * This is meant to be used by time(), date(), and timeanddate() to get
1430          * the date preference they're supposed to use, it should be used in
1431          * all children.
1432          *
1433          *<code>
1434          * function timeanddate([...], $format = true) {
1435          *      $datePreference = $this->dateFormat($format);
1436          * [...]
1437          * }
1438          *</code>
1439          *
1440          * @param $usePrefs Mixed: if true, the user's preference is used
1441          *                         if false, the site/language default is used
1442          *                         if int/string, assumed to be a format.
1443          * @return string
1444          */
1445         function dateFormat( $usePrefs = true ) {
1446                 global $wgUser;
1447
1448                 if ( is_bool( $usePrefs ) ) {
1449                         if ( $usePrefs ) {
1450                                 $datePreference = $wgUser->getDatePreference();
1451                         } else {
1452                                 $datePreference = (string)User::getDefaultOption( 'date' );
1453                         }
1454                 } else {
1455                         $datePreference = (string)$usePrefs;
1456                 }
1457
1458                 // return int
1459                 if ( $datePreference == '' ) {
1460                         return 'default';
1461                 }
1462
1463                 return $datePreference;
1464         }
1465
1466         /**
1467          * Get a format string for a given type and preference
1468          * @param $type May be date, time or both
1469          * @param $pref The format name as it appears in Messages*.php
1470          */
1471         function getDateFormatString( $type, $pref ) {
1472                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
1473                         if ( $pref == 'default' ) {
1474                                 $pref = $this->getDefaultDateFormat();
1475                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
1476                         } else {
1477                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
1478                                 if ( is_null( $df ) ) {
1479                                         $pref = $this->getDefaultDateFormat();
1480                                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
1481                                 }
1482                         }
1483                         $this->dateFormatStrings[$type][$pref] = $df;
1484                 }
1485                 return $this->dateFormatStrings[$type][$pref];
1486         }
1487
1488         /**
1489          * @param $ts Mixed: the time format which needs to be turned into a
1490          *            date('YmdHis') format with wfTimestamp(TS_MW,$ts)
1491          * @param $adj Bool: whether to adjust the time output according to the
1492          *             user configured offset ($timecorrection)
1493          * @param $format Mixed: true to use user's date format preference
1494          * @param $timecorrection String: the time offset as returned by
1495          *                        validateTimeZone() in Special:Preferences
1496          * @return string
1497          */
1498         function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
1499                 if ( $adj ) {
1500                         $ts = $this->userAdjust( $ts, $timecorrection );
1501                 }
1502                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
1503                 return $this->sprintfDate( $df, $ts );
1504         }
1505
1506         /**
1507          * @param $ts Mixed: the time format which needs to be turned into a
1508          *            date('YmdHis') format with wfTimestamp(TS_MW,$ts)
1509          * @param $adj Bool: whether to adjust the time output according to the
1510          *             user configured offset ($timecorrection)
1511          * @param $format Mixed: true to use user's date format preference
1512          * @param $timecorrection String: the time offset as returned by
1513          *                        validateTimeZone() in Special:Preferences
1514          * @return string
1515          */
1516         function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
1517                 if ( $adj ) {
1518                         $ts = $this->userAdjust( $ts, $timecorrection );
1519                 }
1520                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
1521                 return $this->sprintfDate( $df, $ts );
1522         }
1523
1524         /**
1525          * @param $ts Mixed: the time format which needs to be turned into a
1526          *            date('YmdHis') format with wfTimestamp(TS_MW,$ts)
1527          * @param $adj Bool: whether to adjust the time output according to the
1528          *             user configured offset ($timecorrection)
1529          * @param $format Mixed: what format to return, if it's false output the
1530          *                default one (default true)
1531          * @param $timecorrection String: the time offset as returned by
1532          *                        validateTimeZone() in Special:Preferences
1533          * @return string
1534          */
1535         function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
1536                 $ts = wfTimestamp( TS_MW, $ts );
1537                 if ( $adj ) {
1538                         $ts = $this->userAdjust( $ts, $timecorrection );
1539                 }
1540                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
1541                 return $this->sprintfDate( $df, $ts );
1542         }
1543
1544         function getMessage( $key ) {
1545                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
1546         }
1547
1548         function getAllMessages() {
1549                 return self::$dataCache->getItem( $this->mCode, 'messages' );
1550         }
1551
1552         function iconv( $in, $out, $string ) {
1553                 # This is a wrapper for iconv in all languages except esperanto,
1554                 # which does some nasty x-conversions beforehand
1555
1556                 # Even with //IGNORE iconv can whine about illegal characters in
1557                 # *input* string. We just ignore those too.
1558                 # REF: http://bugs.php.net/bug.php?id=37166
1559                 # REF: https://bugzilla.wikimedia.org/show_bug.cgi?id=16885
1560                 wfSuppressWarnings();
1561                 $text = iconv( $in, $out . '//IGNORE', $string );
1562                 wfRestoreWarnings();
1563                 return $text;
1564         }
1565
1566         // callback functions for uc(), lc(), ucwords(), ucwordbreaks()
1567         function ucwordbreaksCallbackAscii( $matches ) {
1568                 return $this->ucfirst( $matches[1] );
1569         }
1570
1571         function ucwordbreaksCallbackMB( $matches ) {
1572                 return mb_strtoupper( $matches[0] );
1573         }
1574
1575         function ucCallback( $matches ) {
1576                 list( $wikiUpperChars ) = self::getCaseMaps();
1577                 return strtr( $matches[1], $wikiUpperChars );
1578         }
1579
1580         function lcCallback( $matches ) {
1581                 list( , $wikiLowerChars ) = self::getCaseMaps();
1582                 return strtr( $matches[1], $wikiLowerChars );
1583         }
1584
1585         function ucwordsCallbackMB( $matches ) {
1586                 return mb_strtoupper( $matches[0] );
1587         }
1588
1589         function ucwordsCallbackWiki( $matches ) {
1590                 list( $wikiUpperChars ) = self::getCaseMaps();
1591                 return strtr( $matches[0], $wikiUpperChars );
1592         }
1593
1594         /**
1595          * Make a string's first character uppercase
1596          */
1597         function ucfirst( $str ) {
1598                 $o = ord( $str );
1599                 if ( $o < 96 ) { // if already uppercase...
1600                         return $str;
1601                 } elseif ( $o < 128 ) {
1602                         return ucfirst( $str ); // use PHP's ucfirst()
1603                 } else {
1604                         // fall back to more complex logic in case of multibyte strings
1605                         return $this->uc( $str, true );
1606                 }
1607         }
1608
1609         /**
1610          * Convert a string to uppercase
1611          */
1612         function uc( $str, $first = false ) {
1613                 if ( function_exists( 'mb_strtoupper' ) ) {
1614                         if ( $first ) {
1615                                 if ( $this->isMultibyte( $str ) ) {
1616                                         return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
1617                                 } else {
1618                                         return ucfirst( $str );
1619                                 }
1620                         } else {
1621                                 return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
1622                         }
1623                 } else {
1624                         if ( $this->isMultibyte( $str ) ) {
1625                                 $x = $first ? '^' : '';
1626                                 return preg_replace_callback(
1627                                         "/$x([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/",
1628                                         array( $this, 'ucCallback' ),
1629                                         $str
1630                                 );
1631                         } else {
1632                                 return $first ? ucfirst( $str ) : strtoupper( $str );
1633                         }
1634                 }
1635         }
1636
1637         function lcfirst( $str ) {
1638                 $o = ord( $str );
1639                 if ( !$o ) {
1640                         return strval( $str );
1641                 } elseif ( $o >= 128 ) {
1642                         return $this->lc( $str, true );
1643                 } elseif ( $o > 96 ) {
1644                         return $str;
1645                 } else {
1646                         $str[0] = strtolower( $str[0] );
1647                         return $str;
1648                 }
1649         }
1650
1651         function lc( $str, $first = false ) {
1652                 if ( function_exists( 'mb_strtolower' ) ) {
1653                         if ( $first ) {
1654                                 if ( $this->isMultibyte( $str ) ) {
1655                                         return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
1656                                 } else {
1657                                         return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
1658                                 }
1659                         } else {
1660                                 return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
1661                         }
1662                 } else {
1663                         if ( $this->isMultibyte( $str ) ) {
1664                                 $x = $first ? '^' : '';
1665                                 return preg_replace_callback(
1666                                         "/$x([A-Z]|[\\xc0-\\xff][\\x80-\\xbf]*)/",
1667                                         array( $this, 'lcCallback' ),
1668                                         $str
1669                                 );
1670                         } else {
1671                                 return $first ? strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 ) : strtolower( $str );
1672                         }
1673                 }
1674         }
1675
1676         function isMultibyte( $str ) {
1677                 return (bool)preg_match( '/[\x80-\xff]/', $str );
1678         }
1679
1680         function ucwords( $str ) {
1681                 if ( $this->isMultibyte( $str ) ) {
1682                         $str = $this->lc( $str );
1683
1684                         // regexp to find first letter in each word (i.e. after each space)
1685                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
1686
1687                         // function to use to capitalize a single char
1688                         if ( function_exists( 'mb_strtoupper' ) ) {
1689                                 return preg_replace_callback(
1690                                         $replaceRegexp,
1691                                         array( $this, 'ucwordsCallbackMB' ),
1692                                         $str
1693                                 );
1694                         } else {
1695                                 return preg_replace_callback(
1696                                         $replaceRegexp,
1697                                         array( $this, 'ucwordsCallbackWiki' ),
1698                                         $str
1699                                 );
1700                         }
1701                 } else {
1702                         return ucwords( strtolower( $str ) );
1703                 }
1704         }
1705
1706         # capitalize words at word breaks
1707         function ucwordbreaks( $str ) {
1708                 if ( $this->isMultibyte( $str ) ) {
1709                         $str = $this->lc( $str );
1710
1711                         // since \b doesn't work for UTF-8, we explicitely define word break chars
1712                         $breaks = "[ \-\(\)\}\{\.,\?!]";
1713
1714                         // find first letter after word break
1715                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
1716
1717                         if ( function_exists( 'mb_strtoupper' ) ) {
1718                                 return preg_replace_callback(
1719                                         $replaceRegexp,
1720                                         array( $this, 'ucwordbreaksCallbackMB' ),
1721                                         $str
1722                                 );
1723                         } else {
1724                                 return preg_replace_callback(
1725                                         $replaceRegexp,
1726                                         array( $this, 'ucwordsCallbackWiki' ),
1727                                         $str
1728                                 );
1729                         }
1730                 } else {
1731                         return preg_replace_callback(
1732                                 '/\b([\w\x80-\xff]+)\b/',
1733                                 array( $this, 'ucwordbreaksCallbackAscii' ),
1734                                 $str
1735                         );
1736                 }
1737         }
1738
1739         /**
1740          * Return a case-folded representation of $s
1741          *
1742          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
1743          * and $s2 are the same except for the case of their characters. It is not
1744          * necessary for the value returned to make sense when displayed.
1745          *
1746          * Do *not* perform any other normalisation in this function. If a caller
1747          * uses this function when it should be using a more general normalisation
1748          * function, then fix the caller.
1749          */
1750         function caseFold( $s ) {
1751                 return $this->uc( $s );
1752         }
1753
1754         function checkTitleEncoding( $s ) {
1755                 if ( is_array( $s ) ) {
1756                         wfDebugDieBacktrace( 'Given array to checkTitleEncoding.' );
1757                 }
1758                 # Check for non-UTF-8 URLs
1759                 $ishigh = preg_match( '/[\x80-\xff]/', $s );
1760                 if ( !$ishigh ) {
1761                         return $s;
1762                 }
1763
1764                 $isutf8 = preg_match( '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
1765                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})+$/', $s );
1766                 if ( $isutf8 ) {
1767                         return $s;
1768                 }
1769
1770                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
1771         }
1772
1773         function fallback8bitEncoding() {
1774                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
1775         }
1776
1777         /**
1778          * Most writing systems use whitespace to break up words.
1779          * Some languages such as Chinese don't conventionally do this,
1780          * which requires special handling when breaking up words for
1781          * searching etc.
1782          */
1783         function hasWordBreaks() {
1784                 return true;
1785         }
1786
1787         /**
1788          * Some languages such as Chinese require word segmentation,
1789          * Specify such segmentation when overridden in derived class.
1790          *
1791          * @param $string String
1792          * @return String
1793          */
1794         function segmentByWord( $string ) {
1795                 return $string;
1796         }
1797
1798         /**
1799          * Some languages have special punctuation need to be normalized.
1800          * Make such changes here.
1801          *
1802          * @param $string String
1803          * @return String
1804          */
1805         function normalizeForSearch( $string ) {
1806                 return self::convertDoubleWidth( $string );
1807         }
1808
1809         /**
1810          * convert double-width roman characters to single-width.
1811          * range: ff00-ff5f ~= 0020-007f
1812          */
1813         protected static function convertDoubleWidth( $string ) {
1814                 static $full = null;
1815                 static $half = null;
1816
1817                 if ( $full === null ) {
1818                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
1819                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
1820                         $full = str_split( $fullWidth, 3 );
1821                         $half = str_split( $halfWidth );
1822                 }
1823
1824                 $string = str_replace( $full, $half, $string );
1825                 return $string;
1826         }
1827
1828         protected static function insertSpace( $string, $pattern ) {
1829                 $string = preg_replace( $pattern, " $1 ", $string );
1830                 $string = preg_replace( '/ +/', ' ', $string );
1831                 return $string;
1832         }
1833
1834         function convertForSearchResult( $termsArray ) {
1835                 # some languages, e.g. Chinese, need to do a conversion
1836                 # in order for search results to be displayed correctly
1837                 return $termsArray;
1838         }
1839
1840         /**
1841          * Get the first character of a string.
1842          *
1843          * @param $s string
1844          * @return string
1845          */
1846         function firstChar( $s ) {
1847                 $matches = array();
1848                 preg_match(
1849                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
1850                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
1851                         $s,
1852                         $matches
1853                 );
1854
1855                 if ( isset( $matches[1] ) ) {
1856                         if ( strlen( $matches[1] ) != 3 ) {
1857                                 return $matches[1];
1858                         }
1859
1860                         // Break down Hangul syllables to grab the first jamo
1861                         $code = utf8ToCodepoint( $matches[1] );
1862                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
1863                                 return $matches[1];
1864                         } elseif ( $code < 0xb098 ) {
1865                                 return "\xe3\x84\xb1";
1866                         } elseif ( $code < 0xb2e4 ) {
1867                                 return "\xe3\x84\xb4";
1868                         } elseif ( $code < 0xb77c ) {
1869                                 return "\xe3\x84\xb7";
1870                         } elseif ( $code < 0xb9c8 ) {
1871                                 return "\xe3\x84\xb9";
1872                         } elseif ( $code < 0xbc14 ) {
1873                                 return "\xe3\x85\x81";
1874                         } elseif ( $code < 0xc0ac ) {
1875                                 return "\xe3\x85\x82";
1876                         } elseif ( $code < 0xc544 ) {
1877                                 return "\xe3\x85\x85";
1878                         } elseif ( $code < 0xc790 ) {
1879                                 return "\xe3\x85\x87";
1880                         } elseif ( $code < 0xcc28 ) {
1881                                 return "\xe3\x85\x88";
1882                         } elseif ( $code < 0xce74 ) {
1883                                 return "\xe3\x85\x8a";
1884                         } elseif ( $code < 0xd0c0 ) {
1885                                 return "\xe3\x85\x8b";
1886                         } elseif ( $code < 0xd30c ) {
1887                                 return "\xe3\x85\x8c";
1888                         } elseif ( $code < 0xd558 ) {
1889                                 return "\xe3\x85\x8d";
1890                         } else {
1891                                 return "\xe3\x85\x8e";
1892                         }
1893                 } else {
1894                         return '';
1895                 }
1896         }
1897
1898         function initEncoding() {
1899                 # Some languages may have an alternate char encoding option
1900                 # (Esperanto X-coding, Japanese furigana conversion, etc)
1901                 # If this language is used as the primary content language,
1902                 # an override to the defaults can be set here on startup.
1903         }
1904
1905         function recodeForEdit( $s ) {
1906                 # For some languages we'll want to explicitly specify
1907                 # which characters make it into the edit box raw
1908                 # or are converted in some way or another.
1909                 # Note that if wgOutputEncoding is different from
1910                 # wgInputEncoding, this text will be further converted
1911                 # to wgOutputEncoding.
1912                 global $wgEditEncoding;
1913                 if ( $wgEditEncoding == '' || $wgEditEncoding == 'UTF-8' ) {
1914                         return $s;
1915                 } else {
1916                         return $this->iconv( 'UTF-8', $wgEditEncoding, $s );
1917                 }
1918         }
1919
1920         function recodeInput( $s ) {
1921                 # Take the previous into account.
1922                 global $wgEditEncoding;
1923                 if ( $wgEditEncoding != '' ) {
1924                         $enc = $wgEditEncoding;
1925                 } else {
1926                         $enc = 'UTF-8';
1927                 }
1928                 if ( $enc == 'UTF-8' ) {
1929                         return $s;
1930                 } else {
1931                         return $this->iconv( $enc, 'UTF-8', $s );
1932                 }
1933         }
1934
1935         /**
1936          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
1937          * also cleans up certain backwards-compatible sequences, converting them
1938          * to the modern Unicode equivalent.
1939          *
1940          * This is language-specific for performance reasons only.
1941          */
1942         function normalize( $s ) {
1943                 global $wgAllUnicodeFixes;
1944                 $s = UtfNormal::cleanUp( $s );
1945                 if ( $wgAllUnicodeFixes ) {
1946                         $s = $this->transformUsingPairFile( 'normalize-ar.ser', $s );
1947                         $s = $this->transformUsingPairFile( 'normalize-ml.ser', $s );
1948                 }
1949
1950                 return $s;
1951         }
1952
1953         /**
1954          * Transform a string using serialized data stored in the given file (which
1955          * must be in the serialized subdirectory of $IP). The file contains pairs
1956          * mapping source characters to destination characters.
1957          *
1958          * The data is cached in process memory. This will go faster if you have the
1959          * FastStringSearch extension.
1960          */
1961         function transformUsingPairFile( $file, $string ) {
1962                 if ( !isset( $this->transformData[$file] ) ) {
1963                         $data = wfGetPrecompiledData( $file );
1964                         if ( $data === false ) {
1965                                 throw new MWException( __METHOD__ . ": The transformation file $file is missing" );
1966                         }
1967                         $this->transformData[$file] = new ReplacementArray( $data );
1968                 }
1969                 return $this->transformData[$file]->replace( $string );
1970         }
1971
1972         /**
1973          * For right-to-left language support
1974          *
1975          * @return bool
1976          */
1977         function isRTL() {
1978                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
1979         }
1980
1981         /**
1982          * Return the correct HTML 'dir' attribute value for this language.
1983          * @return String
1984          */
1985         function getDir() {
1986                 return $this->isRTL() ? 'rtl' : 'ltr';
1987         }
1988
1989         /**
1990          * Return 'left' or 'right' as appropriate alignment for line-start
1991          * for this language's text direction.
1992          *
1993          * Should be equivalent to CSS3 'start' text-align value....
1994          *
1995          * @return String
1996          */
1997         function alignStart() {
1998                 return $this->isRTL() ? 'right' : 'left';
1999         }
2000
2001         /**
2002          * Return 'right' or 'left' as appropriate alignment for line-end
2003          * for this language's text direction.
2004          *
2005          * Should be equivalent to CSS3 'end' text-align value....
2006          *
2007          * @return String
2008          */
2009         function alignEnd() {
2010                 return $this->isRTL() ? 'left' : 'right';
2011         }
2012
2013         /**
2014          * A hidden direction mark (LRM or RLM), depending on the language direction
2015          *
2016          * @return string
2017          */
2018         function getDirMark() {
2019                 return $this->isRTL() ? "\xE2\x80\x8F" : "\xE2\x80\x8E";
2020         }
2021
2022         function capitalizeAllNouns() {
2023                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
2024         }
2025
2026         /**
2027          * An arrow, depending on the language direction
2028          *
2029          * @return string
2030          */
2031         function getArrow() {
2032                 return $this->isRTL() ? '←' : '→';
2033         }
2034
2035         /**
2036          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
2037          *
2038          * @return bool
2039          */
2040         function linkPrefixExtension() {
2041                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
2042         }
2043
2044         function getMagicWords() {
2045                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
2046         }
2047
2048         # Fill a MagicWord object with data from here
2049         function getMagic( $mw ) {
2050                 if ( !$this->mMagicHookDone ) {
2051                         $this->mMagicHookDone = true;
2052                         wfProfileIn( 'LanguageGetMagic' );
2053                         wfRunHooks( 'LanguageGetMagic', array( &$this->mMagicExtensions, $this->getCode() ) );
2054                         wfProfileOut( 'LanguageGetMagic' );
2055                 }
2056                 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
2057                         $rawEntry = $this->mMagicExtensions[$mw->mId];
2058                 } else {
2059                         $magicWords = $this->getMagicWords();
2060                         if ( isset( $magicWords[$mw->mId] ) ) {
2061                                 $rawEntry = $magicWords[$mw->mId];
2062                         } else {
2063                                 $rawEntry = false;
2064                         }
2065                 }
2066
2067                 if ( !is_array( $rawEntry ) ) {
2068                         error_log( "\"$rawEntry\" is not a valid magic thingie for \"$mw->mId\"" );
2069                 } else {
2070                         $mw->mCaseSensitive = $rawEntry[0];
2071                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
2072                 }
2073         }
2074
2075         /**
2076          * Add magic words to the extension array
2077          */
2078         function addMagicWordsByLang( $newWords ) {
2079                 $code = $this->getCode();
2080                 $fallbackChain = array();
2081                 while ( $code && !in_array( $code, $fallbackChain ) ) {
2082                         $fallbackChain[] = $code;
2083                         $code = self::getFallbackFor( $code );
2084                 }
2085                 if ( !in_array( 'en', $fallbackChain ) ) {
2086                         $fallbackChain[] = 'en';
2087                 }
2088                 $fallbackChain = array_reverse( $fallbackChain );
2089                 foreach ( $fallbackChain as $code ) {
2090                         if ( isset( $newWords[$code] ) ) {
2091                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
2092                         }
2093                 }
2094         }
2095
2096         /**
2097          * Get special page names, as an associative array
2098          *   case folded alias => real name
2099          */
2100         function getSpecialPageAliases() {
2101                 // Cache aliases because it may be slow to load them
2102                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
2103                         // Initialise array
2104                         $this->mExtendedSpecialPageAliases =
2105                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
2106                         wfRunHooks( 'LanguageGetSpecialPageAliases',
2107                                 array( &$this->mExtendedSpecialPageAliases, $this->getCode() ) );
2108                 }
2109
2110                 return $this->mExtendedSpecialPageAliases;
2111         }
2112
2113         /**
2114          * Italic is unsuitable for some languages
2115          *
2116          * @param $text String: the text to be emphasized.
2117          * @return string
2118          */
2119         function emphasize( $text ) {
2120                 return "<em>$text</em>";
2121         }
2122
2123          /**
2124           * Normally we output all numbers in plain en_US style, that is
2125           * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
2126           * point twohundredthirtyfive. However this is not sutable for all
2127           * languages, some such as Pakaran want ੨੯੩,੨੯੫.੨੩੫ and others such as
2128           * Icelandic just want to use commas instead of dots, and dots instead
2129           * of commas like "293.291,235".
2130           *
2131           * An example of this function being called:
2132           * <code>
2133           * wfMsg( 'message', $wgLang->formatNum( $num ) )
2134           * </code>
2135           *
2136           * See LanguageGu.php for the Gujarati implementation and
2137           * $separatorTransformTable on MessageIs.php for
2138           * the , => . and . => , implementation.
2139           *
2140           * @todo check if it's viable to use localeconv() for the decimal
2141           *       separator thing.
2142           * @param $number Mixed: the string to be formatted, should be an integer
2143           *        or a floating point number.
2144           * @param $nocommafy Bool: set to true for special numbers like dates
2145           * @return string
2146           */
2147         function formatNum( $number, $nocommafy = false ) {
2148                 global $wgTranslateNumerals;
2149                 if ( !$nocommafy ) {
2150                         $number = $this->commafy( $number );
2151                         $s = $this->separatorTransformTable();
2152                         if ( $s ) {
2153                                 $number = strtr( $number, $s );
2154                         }
2155                 }
2156
2157                 if ( $wgTranslateNumerals ) {
2158                         $s = $this->digitTransformTable();
2159                         if ( $s ) {
2160                                 $number = strtr( $number, $s );
2161                         }
2162                 }
2163
2164                 return $number;
2165         }
2166
2167         function parseFormattedNumber( $number ) {
2168                 $s = $this->digitTransformTable();
2169                 if ( $s ) {
2170                         $number = strtr( $number, array_flip( $s ) );
2171                 }
2172
2173                 $s = $this->separatorTransformTable();
2174                 if ( $s ) {
2175                         $number = strtr( $number, array_flip( $s ) );
2176                 }
2177
2178                 $number = strtr( $number, array( ',' => '' ) );
2179                 return $number;
2180         }
2181
2182         /**
2183          * Adds commas to a given number
2184          *
2185          * @param $_ mixed
2186          * @return string
2187          */
2188         function commafy( $_ ) {
2189                 return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $_ ) ) );
2190         }
2191
2192         function digitTransformTable() {
2193                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
2194         }
2195
2196         function separatorTransformTable() {
2197                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
2198         }
2199
2200         /**
2201          * Take a list of strings and build a locale-friendly comma-separated
2202          * list, using the local comma-separator message.
2203          * The last two strings are chained with an "and".
2204          *
2205          * @param $l Array
2206          * @return string
2207          */
2208         function listToText( $l ) {
2209                 $s = '';
2210                 $m = count( $l ) - 1;
2211                 if ( $m == 1 ) {
2212                         return $l[0] . $this->getMessageFromDB( 'and' ) . $this->getMessageFromDB( 'word-separator' ) . $l[1];
2213                 } else {
2214                         for ( $i = $m; $i >= 0; $i-- ) {
2215                                 if ( $i == $m ) {
2216                                         $s = $l[$i];
2217                                 } else if ( $i == $m - 1 ) {
2218                                         $s = $l[$i] . $this->getMessageFromDB( 'and' ) . $this->getMessageFromDB( 'word-separator' ) . $s;
2219                                 } else {
2220                                         $s = $l[$i] . $this->getMessageFromDB( 'comma-separator' ) . $s;
2221                                 }
2222                         }
2223                         return $s;
2224                 }
2225         }
2226
2227         /**
2228          * Take a list of strings and build a locale-friendly comma-separated
2229          * list, using the local comma-separator message.
2230          * @param $list array of strings to put in a comma list
2231          * @return string
2232          */
2233         function commaList( $list ) {
2234                 return implode(
2235                         $list,
2236                         wfMsgExt(
2237                                 'comma-separator',
2238                                 array( 'parsemag', 'escapenoentities', 'language' => $this )
2239                         )
2240                 );
2241         }
2242
2243         /**
2244          * Take a list of strings and build a locale-friendly semicolon-separated
2245          * list, using the local semicolon-separator message.
2246          * @param $list array of strings to put in a semicolon list
2247          * @return string
2248          */
2249         function semicolonList( $list ) {
2250                 return implode(
2251                         $list,
2252                         wfMsgExt(
2253                                 'semicolon-separator',
2254                                 array( 'parsemag', 'escapenoentities', 'language' => $this )
2255                         )
2256                 );
2257         }
2258
2259         /**
2260          * Same as commaList, but separate it with the pipe instead.
2261          * @param $list array of strings to put in a pipe list
2262          * @return string
2263          */
2264         function pipeList( $list ) {
2265                 return implode(
2266                         $list,
2267                         wfMsgExt(
2268                                 'pipe-separator',
2269                                 array( 'escapenoentities', 'language' => $this )
2270                         )
2271                 );
2272         }
2273
2274         /**
2275          * Truncate a string to a specified length in bytes, appending an optional
2276          * string (e.g. for ellipses)
2277          *
2278          * The database offers limited byte lengths for some columns in the database;
2279          * multi-byte character sets mean we need to ensure that only whole characters
2280          * are included, otherwise broken characters can be passed to the user
2281          *
2282          * If $length is negative, the string will be truncated from the beginning
2283          *
2284          * @param $string String to truncate
2285          * @param $length Int: maximum length (excluding ellipses)
2286          * @param $ellipsis String to append to the truncated text
2287          * @return string
2288          */
2289         function truncate( $string, $length, $ellipsis = '...' ) {
2290                 # Use the localized ellipsis character
2291                 if ( $ellipsis == '...' ) {
2292                         $ellipsis = wfMsgExt( 'ellipsis', array( 'escapenoentities', 'language' => $this ) );
2293                 }
2294                 # Check if there is no need to truncate
2295                 if ( $length == 0 ) {
2296                         return $ellipsis;
2297                 } elseif ( strlen( $string ) <= abs( $length ) ) {
2298                         return $string;
2299                 }
2300                 $stringOriginal = $string;
2301                 if ( $length > 0 ) {
2302                         $string = substr( $string, 0, $length ); // xyz...
2303                         $string = $this->removeBadCharLast( $string );
2304                         $string = $string . $ellipsis;
2305                 } else {
2306                         $string = substr( $string, $length ); // ...xyz
2307                         $string = $this->removeBadCharFirst( $string );
2308                         $string = $ellipsis . $string;
2309                 }
2310                 # Do not truncate if the ellipsis makes the string longer/equal (bug 22181)
2311                 if ( strlen( $string ) < strlen( $stringOriginal ) ) {
2312                         return $string;
2313                 } else {
2314                         return $stringOriginal;
2315                 }
2316         }
2317
2318         /**
2319          * Remove bytes that represent an incomplete Unicode character
2320          * at the end of string (e.g. bytes of the char are missing)
2321          *
2322          * @param $string String
2323          * @return string
2324          */
2325         protected function removeBadCharLast( $string ) {
2326                 $char = ord( $string[strlen( $string ) - 1] );
2327                 $m = array();
2328                 if ( $char >= 0xc0 ) {
2329                         # We got the first byte only of a multibyte char; remove it.
2330                         $string = substr( $string, 0, -1 );
2331                 } elseif ( $char >= 0x80 &&
2332                       preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
2333                                   '[\xf0-\xf7][\x80-\xbf]{1,2})$/', $string, $m ) )
2334                 {
2335                         # We chopped in the middle of a character; remove it
2336                         $string = $m[1];
2337                 }
2338                 return $string;
2339         }
2340
2341         /**
2342          * Remove bytes that represent an incomplete Unicode character
2343          * at the start of string (e.g. bytes of the char are missing)
2344          *
2345          * @param $string String
2346          * @return string
2347          */
2348         protected function removeBadCharFirst( $string ) {
2349                 $char = ord( $string[0] );
2350                 if ( $char >= 0x80 && $char < 0xc0 ) {
2351                         # We chopped in the middle of a character; remove the whole thing
2352                         $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
2353                 }
2354                 return $string;
2355         }
2356
2357         /*
2358          * Truncate a string of valid HTML to a specified length in bytes,
2359          * appending an optional string (e.g. for ellipses), and return valid HTML
2360          *
2361          * This is only intended for styled/linked text, such as HTML with
2362          * tags like <span> and <a>, were the tags are self-contained (valid HTML)
2363          *
2364          * Note: tries to fix broken HTML with MWTidy
2365          *
2366          * @param string $text String to truncate
2367          * @param int $length (zero/positive) Maximum length (excluding ellipses)
2368          * @param string $ellipsis String to append to the truncated text
2369          * @returns string
2370          */
2371         function truncateHtml( $text, $length, $ellipsis = '...' ) {
2372                 # Use the localized ellipsis character
2373                 if ( $ellipsis == '...' ) {
2374                         $ellipsis = wfMsgExt( 'ellipsis', array( 'escapenoentities', 'language' => $this ) );
2375                 }
2376                 # Check if there is no need to truncate
2377                 if ( $length <= 0 ) {
2378                         return $ellipsis; // no text shown, nothing to format
2379                 } elseif ( strlen( $text ) <= $length ) {
2380                         return $text; // string short enough even *with* HTML
2381                 }
2382                 $text = MWTidy::tidy( $text ); // fix tags
2383                 $displayLen = 0; // innerHTML legth so far
2384                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
2385                 $tagType = 0; // 0-open, 1-close
2386                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
2387                 $entityState = 0; // 0-not entity, 1-entity
2388                 $tag = $ret = $ch = '';
2389                 $openTags = array();
2390                 $textLen = strlen( $text );
2391                 for ( $pos = 0; $pos < $textLen; ++$pos ) {
2392                         $ch = $text[$pos];
2393                         $lastCh = $pos ? $text[$pos - 1] : '';
2394                         $ret .= $ch; // add to result string
2395                         if ( $ch == '<' ) {
2396                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
2397                                 $entityState = 0; // for bad HTML
2398                                 $bracketState = 1; // tag started (checking for backslash)
2399                         } elseif ( $ch == '>' ) {
2400                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
2401                                 $entityState = 0; // for bad HTML
2402                                 $bracketState = 0; // out of brackets
2403                         } elseif ( $bracketState == 1 ) {
2404                                 if ( $ch == '/' ) {
2405                                         $tagType = 1; // close tag (e.g. "</span>")
2406                                 } else {
2407                                         $tagType = 0; // open tag (e.g. "<span>")
2408                                         $tag .= $ch;
2409                                 }
2410                                 $bracketState = 2; // building tag name
2411                         } elseif ( $bracketState == 2 ) {
2412                                 if ( $ch != ' ' ) {
2413                                         $tag .= $ch;
2414                                 } else {
2415                                         // Name found (e.g. "<a href=..."), add on tag attributes...
2416                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
2417                                 }
2418                         } elseif ( $bracketState == 0 ) {
2419                                 if ( $entityState ) {
2420                                         if ( $ch == ';' ) {
2421                                                 $entityState = 0;
2422                                                 $displayLen++; // entity is one displayed char
2423                                         }
2424                                 } else {
2425                                         if ( $ch == '&' ) {
2426                                                 $entityState = 1; // entity found, (e.g. "&#160;")
2427                                         } else {
2428                                                 $displayLen++; // this char is displayed
2429                                                 // Add on the other display text after this...
2430                                                 $skipped = $this->truncate_skip(
2431                                                         $ret, $text, "<>&", $pos + 1, $length - $displayLen );
2432                                                 $displayLen += $skipped;
2433                                                 $pos += $skipped;
2434                                         }
2435                                 }
2436                         }
2437                         # Consider truncation once the display length has reached the maximim.
2438                         # Double-check that we're not in the middle of a bracket/entity...
2439                         if ( $displayLen >= $length && $bracketState == 0 && $entityState == 0 ) {
2440                                 if ( !$testingEllipsis ) {
2441                                         $testingEllipsis = true;
2442                                         # Save where we are; we will truncate here unless
2443                                         # the ellipsis actually makes the string longer.
2444                                         $pOpenTags = $openTags; // save state
2445                                         $pRet = $ret; // save state
2446                                 } elseif ( $displayLen > ( $length + strlen( $ellipsis ) ) ) {
2447                                         # Ellipsis won't make string longer/equal, the truncation point was OK.
2448                                         $openTags = $pOpenTags; // reload state
2449                                         $ret = $this->removeBadCharLast( $pRet ); // reload state, multi-byte char fix
2450                                         $ret .= $ellipsis; // add ellipsis
2451                                         break;
2452                                 }
2453                         }
2454                 }
2455                 if ( $displayLen == 0 ) {
2456                         return ''; // no text shown, nothing to format
2457                 }
2458                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags ); // for bad HTML
2459                 while ( count( $openTags ) > 0 ) {
2460                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
2461                 }
2462                 return $ret;
2463         }
2464
2465         // truncateHtml() helper function
2466         // like strcspn() but adds the skipped chars to $ret
2467         private function truncate_skip( &$ret, $text, $search, $start, $len = -1 ) {
2468                 $skipCount = 0;
2469                 if ( $start < strlen( $text ) ) {
2470                         $skipCount = strcspn( $text, $search, $start, $len );
2471                         $ret .= substr( $text, $start, $skipCount );
2472                 }
2473                 return $skipCount;
2474         }
2475
2476         // truncateHtml() helper function
2477         // (a) push or pop $tag from $openTags as needed
2478         // (b) clear $tag value
2479         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
2480                 $tag = ltrim( $tag );
2481                 if ( $tag != '' ) {
2482                         if ( $tagType == 0 && $lastCh != '/' ) {
2483                                 $openTags[] = $tag; // tag opened (didn't close itself)
2484                         } else if ( $tagType == 1 ) {
2485                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
2486                                         array_pop( $openTags ); // tag closed
2487                                 }
2488                         }
2489                         $tag = '';
2490                 }
2491         }
2492
2493         /**
2494          * Grammatical transformations, needed for inflected languages
2495          * Invoked by putting {{grammar:case|word}} in a message
2496          *
2497          * @param $word string
2498          * @param $case string
2499          * @return string
2500          */
2501         function convertGrammar( $word, $case ) {
2502                 global $wgGrammarForms;
2503                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
2504                         return $wgGrammarForms[$this->getCode()][$case][$word];
2505                 }
2506                 return $word;
2507         }
2508
2509         /**
2510          * Provides an alternative text depending on specified gender.
2511          * Usage {{gender:username|masculine|feminine|neutral}}.
2512          * username is optional, in which case the gender of current user is used,
2513          * but only in (some) interface messages; otherwise default gender is used.
2514          * If second or third parameter are not specified, masculine is used.
2515          * These details may be overriden per language.
2516          */
2517         function gender( $gender, $forms ) {
2518                 if ( !count( $forms ) ) {
2519                         return '';
2520                 }
2521                 $forms = $this->preConvertPlural( $forms, 2 );
2522                 if ( $gender === 'male' ) {
2523                         return $forms[0];
2524                 }
2525                 if ( $gender === 'female' ) {
2526                         return $forms[1];
2527                 }
2528                 return isset( $forms[2] ) ? $forms[2] : $forms[0];
2529         }
2530
2531         /**
2532          * Plural form transformations, needed for some languages.
2533          * For example, there are 3 form of plural in Russian and Polish,
2534          * depending on "count mod 10". See [[w:Plural]]
2535          * For English it is pretty simple.
2536          *
2537          * Invoked by putting {{plural:count|wordform1|wordform2}}
2538          * or {{plural:count|wordform1|wordform2|wordform3}}
2539          *
2540          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
2541          *
2542          * @param $count Integer: non-localized number
2543          * @param $forms Array: different plural forms
2544          * @return string Correct form of plural for $count in this language
2545          */
2546         function convertPlural( $count, $forms ) {
2547                 if ( !count( $forms ) ) {
2548                         return '';
2549                 }
2550                 $forms = $this->preConvertPlural( $forms, 2 );
2551
2552                 return ( $count == 1 ) ? $forms[0] : $forms[1];
2553         }
2554
2555         /**
2556          * Checks that convertPlural was given an array and pads it to requested
2557          * amound of forms by copying the last one.
2558          *
2559          * @param $count Integer: How many forms should there be at least
2560          * @param $forms Array of forms given to convertPlural
2561          * @return array Padded array of forms or an exception if not an array
2562          */
2563         protected function preConvertPlural( /* Array */ $forms, $count ) {
2564                 while ( count( $forms ) < $count ) {
2565                         $forms[] = $forms[count( $forms ) - 1];
2566                 }
2567                 return $forms;
2568         }
2569
2570         /**
2571          * For translating of expiry times
2572          * @param $str String: the validated block time in English
2573          * @return Somehow translated block time
2574          * @see LanguageFi.php for example implementation
2575          */
2576         function translateBlockExpiry( $str ) {
2577                 $scBlockExpiryOptions = $this->getMessageFromDB( 'ipboptions' );
2578
2579                 if ( $scBlockExpiryOptions == '-' ) {
2580                         return $str;
2581                 }
2582
2583                 foreach ( explode( ',', $scBlockExpiryOptions ) as $option ) {
2584                         if ( strpos( $option, ':' ) === false ) {
2585                                 continue;
2586                         }
2587                         list( $show, $value ) = explode( ':', $option );
2588                         if ( strcmp( $str, $value ) == 0 ) {
2589                                 return htmlspecialchars( trim( $show ) );
2590                         }
2591                 }
2592
2593                 return $str;
2594         }
2595
2596         /**
2597          * languages like Chinese need to be segmented in order for the diff
2598          * to be of any use
2599          *
2600          * @param $text String
2601          * @return String
2602          */
2603         function segmentForDiff( $text ) {
2604                 return $text;
2605         }
2606
2607         /**
2608          * and unsegment to show the result
2609          *
2610          * @param $text String
2611          * @return String
2612          */
2613         function unsegmentForDiff( $text ) {
2614                 return $text;
2615         }
2616
2617         # convert text to all supported variants
2618         function autoConvertToAllVariants( $text ) {
2619                 return $this->mConverter->autoConvertToAllVariants( $text );
2620         }
2621
2622         # convert text to different variants of a language.
2623         function convert( $text ) {
2624                 return $this->mConverter->convert( $text );
2625         }
2626
2627         # Convert a Title object to a string in the preferred variant
2628         function convertTitle( $title ) {
2629                 return $this->mConverter->convertTitle( $title );
2630         }
2631
2632         # Check if this is a language with variants
2633         function hasVariants() {
2634                 return sizeof( $this->getVariants() ) > 1;
2635         }
2636
2637         # Put custom tags (e.g. -{ }-) around math to prevent conversion
2638         function armourMath( $text ) {
2639                 return $this->mConverter->armourMath( $text );
2640         }
2641
2642         /**
2643          * Perform output conversion on a string, and encode for safe HTML output.
2644          * @param $text String text to be converted
2645          * @param $isTitle Bool whether this conversion is for the article title
2646          * @return string
2647          * @todo this should get integrated somewhere sane
2648          */
2649         function convertHtml( $text, $isTitle = false ) {
2650                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
2651         }
2652
2653         function convertCategoryKey( $key ) {
2654                 return $this->mConverter->convertCategoryKey( $key );
2655         }
2656
2657         /**
2658          * Get the list of variants supported by this langauge
2659          * see sample implementation in LanguageZh.php
2660          *
2661          * @return array an array of language codes
2662          */
2663         function getVariants() {
2664                 return $this->mConverter->getVariants();
2665         }
2666
2667         function getPreferredVariant( $fromUser = true, $fromHeader = false ) {
2668                 return $this->mConverter->getPreferredVariant( $fromUser, $fromHeader );
2669         }
2670
2671         /**
2672          * If a language supports multiple variants, it is
2673          * possible that non-existing link in one variant
2674          * actually exists in another variant. this function
2675          * tries to find it. See e.g. LanguageZh.php
2676          *
2677          * @param $link String: the name of the link
2678          * @param $nt Mixed: the title object of the link
2679          * @param $ignoreOtherCond Boolean: to disable other conditions when
2680          *      we need to transclude a template or update a category's link
2681          * @return null the input parameters may be modified upon return
2682          */
2683         function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
2684                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
2685         }
2686
2687         /**
2688          * If a language supports multiple variants, converts text
2689          * into an array of all possible variants of the text:
2690          *  'variant' => text in that variant
2691          */
2692         function convertLinkToAllVariants( $text ) {
2693                 return $this->mConverter->convertLinkToAllVariants( $text );
2694         }
2695
2696         /**
2697          * returns language specific options used by User::getPageRenderHash()
2698          * for example, the preferred language variant
2699          *
2700          * @return string
2701          */
2702         function getExtraHashOptions() {
2703                 return $this->mConverter->getExtraHashOptions();
2704         }
2705
2706         /**
2707          * For languages that support multiple variants, the title of an
2708          * article may be displayed differently in different variants. this
2709          * function returns the apporiate title defined in the body of the article.
2710          *
2711          * @return string
2712          */
2713         function getParsedTitle() {
2714                 return $this->mConverter->getParsedTitle();
2715         }
2716
2717         /**
2718          * Enclose a string with the "no conversion" tag. This is used by
2719          * various functions in the Parser
2720          *
2721          * @param $text String: text to be tagged for no conversion
2722          * @param $noParse
2723          * @return string the tagged text
2724          */
2725         function markNoConversion( $text, $noParse = false ) {
2726                 return $this->mConverter->markNoConversion( $text, $noParse );
2727         }
2728
2729         /**
2730          * A regular expression to match legal word-trailing characters
2731          * which should be merged onto a link of the form [[foo]]bar.
2732          *
2733          * @return string
2734          */
2735         function linkTrail() {
2736                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
2737         }
2738
2739         function getLangObj() {
2740                 return $this;
2741         }
2742
2743         /**
2744          * Get the RFC 3066 code for this language object
2745          */
2746         function getCode() {
2747                 return $this->mCode;
2748         }
2749
2750         function setCode( $code ) {
2751                 $this->mCode = $code;
2752         }
2753
2754         /**
2755          * Get the name of a file for a certain language code
2756          * @param $prefix string Prepend this to the filename
2757          * @param $code string Language code
2758          * @param $suffix string Append this to the filename
2759          * @return string $prefix . $mangledCode . $suffix
2760          */
2761         static function getFileName( $prefix = 'Language', $code, $suffix = '.php' ) {
2762                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
2763         }
2764
2765         /**
2766          * Get the language code from a file name. Inverse of getFileName()
2767          * @param $filename string $prefix . $languageCode . $suffix
2768          * @param $prefix string Prefix before the language code
2769          * @param $suffix string Suffix after the language code
2770          * @return Language code, or false if $prefix or $suffix isn't found
2771          */
2772         static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
2773                 $m = null;
2774                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
2775                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
2776                 if ( !count( $m ) ) {
2777                         return false;
2778                 }
2779                 return str_replace( '_', '-', strtolower( $m[1] ) );
2780         }
2781
2782         static function getMessagesFileName( $code ) {
2783                 global $IP;
2784                 return self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
2785         }
2786
2787         static function getClassFileName( $code ) {
2788                 global $IP;
2789                 return self::getFileName( "$IP/languages/classes/Language", $code, '.php' );
2790         }
2791
2792         /**
2793          * Get the fallback for a given language
2794          */
2795         static function getFallbackFor( $code ) {
2796                 if ( $code === 'en' ) {
2797                         // Shortcut
2798                         return false;
2799                 } else {
2800                         return self::getLocalisationCache()->getItem( $code, 'fallback' );
2801                 }
2802         }
2803
2804         /**
2805          * Get all messages for a given language
2806          * WARNING: this may take a long time
2807          */
2808         static function getMessagesFor( $code ) {
2809                 return self::getLocalisationCache()->getItem( $code, 'messages' );
2810         }
2811
2812         /**
2813          * Get a message for a given language
2814          */
2815         static function getMessageFor( $key, $code ) {
2816                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
2817         }
2818
2819         function fixVariableInNamespace( $talk ) {
2820                 if ( strpos( $talk, '$1' ) === false ) {
2821                         return $talk;
2822                 }
2823
2824                 global $wgMetaNamespace;
2825                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
2826
2827                 # Allow grammar transformations
2828                 # Allowing full message-style parsing would make simple requests
2829                 # such as action=raw much more expensive than they need to be.
2830                 # This will hopefully cover most cases.
2831                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
2832                         array( &$this, 'replaceGrammarInNamespace' ), $talk );
2833                 return str_replace( ' ', '_', $talk );
2834         }
2835
2836         function replaceGrammarInNamespace( $m ) {
2837                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
2838         }
2839
2840         static function getCaseMaps() {
2841                 static $wikiUpperChars, $wikiLowerChars;
2842                 if ( isset( $wikiUpperChars ) ) {
2843                         return array( $wikiUpperChars, $wikiLowerChars );
2844                 }
2845
2846                 wfProfileIn( __METHOD__ );
2847                 $arr = wfGetPrecompiledData( 'Utf8Case.ser' );
2848                 if ( $arr === false ) {
2849                         throw new MWException(
2850                                 "Utf8Case.ser is missing, please run \"make\" in the serialized directory\n" );
2851                 }
2852                 extract( $arr );
2853                 wfProfileOut( __METHOD__ );
2854                 return array( $wikiUpperChars, $wikiLowerChars );
2855         }
2856
2857         function formatTimePeriod( $seconds ) {
2858                 if ( $seconds < 10 ) {
2859                         return $this->formatNum( sprintf( "%.1f", $seconds ) ) . $this->getMessageFromDB( 'seconds-abbrev' );
2860                 } elseif ( $seconds < 60 ) {
2861                         return $this->formatNum( round( $seconds ) ) . $this->getMessageFromDB( 'seconds-abbrev' );
2862                 } elseif ( $seconds < 3600 ) {
2863                         $minutes = floor( $seconds / 60 );
2864                         $secondsPart = round( fmod( $seconds, 60 ) );
2865                         if ( $secondsPart == 60 ) {
2866                                 $secondsPart = 0;
2867                                 $minutes++;
2868                         }
2869                         return $this->formatNum( $minutes ) . $this->getMessageFromDB( 'minutes-abbrev' ) . ' ' .
2870                                 $this->formatNum( $secondsPart ) . $this->getMessageFromDB( 'seconds-abbrev' );
2871                 } else {
2872                         $hours = floor( $seconds / 3600 );
2873                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
2874                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
2875                         if ( $secondsPart == 60 ) {
2876                                 $secondsPart = 0;
2877                                 $minutes++;
2878                         }
2879                         if ( $minutes == 60 ) {
2880                                 $minutes = 0;
2881                                 $hours++;
2882                         }
2883                         return $this->formatNum( $hours ) . $this->getMessageFromDB( 'hours-abbrev' ) . ' ' .
2884                                 $this->formatNum( $minutes ) . $this->getMessageFromDB( 'minutes-abbrev' ) . ' ' .
2885                                 $this->formatNum( $secondsPart ) . $this->getMessageFromDB( 'seconds-abbrev' );
2886                 }
2887         }
2888
2889         function formatBitrate( $bps ) {
2890                 $units = array( 'bps', 'kbps', 'Mbps', 'Gbps' );
2891                 if ( $bps <= 0 ) {
2892                         return $this->formatNum( $bps ) . $units[0];
2893                 }
2894                 $unitIndex = floor( log10( $bps ) / 3 );
2895                 $mantissa = $bps / pow( 1000, $unitIndex );
2896                 if ( $mantissa < 10 ) {
2897                         $mantissa = round( $mantissa, 1 );
2898                 } else {
2899                         $mantissa = round( $mantissa );
2900                 }
2901                 return $this->formatNum( $mantissa ) . $units[$unitIndex];
2902         }
2903
2904         /**
2905          * Format a size in bytes for output, using an appropriate
2906          * unit (B, KB, MB or GB) according to the magnitude in question
2907          *
2908          * @param $size Size to format
2909          * @return string Plain text (not HTML)
2910          */
2911         function formatSize( $size ) {
2912                 // For small sizes no decimal places necessary
2913                 $round = 0;
2914                 if ( $size > 1024 ) {
2915                         $size = $size / 1024;
2916                         if ( $size > 1024 ) {
2917                                 $size = $size / 1024;
2918                                 // For MB and bigger two decimal places are smarter
2919                                 $round = 2;
2920                                 if ( $size > 1024 ) {
2921                                         $size = $size / 1024;
2922                                         $msg = 'size-gigabytes';
2923                                 } else {
2924                                         $msg = 'size-megabytes';
2925                                 }
2926                         } else {
2927                                 $msg = 'size-kilobytes';
2928                         }
2929                 } else {
2930                         $msg = 'size-bytes';
2931                 }
2932                 $size = round( $size, $round );
2933                 $text = $this->getMessageFromDB( $msg );
2934                 return str_replace( '$1', $this->formatNum( $size ), $text );
2935         }
2936
2937         /**
2938          * Get the conversion rule title, if any.
2939          */
2940         function getConvRuleTitle() {
2941                 return $this->mConverter->getConvRuleTitle();
2942         }
2943
2944         /**
2945          * Given a string, convert it to a (hopefully short) key that can be used
2946          * for efficient sorting.  A binary sort according to the sortkeys
2947          * corresponds to a logical sort of the corresponding strings.  Current
2948          * code expects that a null character should sort before all others, but
2949          * has no other particular expectations (and that one can be changed if
2950          * necessary).
2951          *
2952          * @param string $string UTF-8 string
2953          * @return string Binary sortkey
2954          */
2955         public function convertToSortkey( $string ) {
2956                 # Fake function for now
2957                 return strtoupper( $string );
2958         }
2959
2960         /**
2961          * Given a string, return the logical "first letter" to be used for
2962          * grouping on category pages and so on.  This has to be coordinated
2963          * carefully with convertToSortkey(), or else the sorted list might jump
2964          * back and forth between the same "initial letters" or other pathological
2965          * behavior.  For instance, if you just return the first character, but "a"
2966          * sorts the same as "A" based on convertToSortkey(), then you might get a
2967          * list like
2968          *
2969          * == A ==
2970          * * [[Aardvark]]
2971          *
2972          * == a ==
2973          * * [[antelope]]
2974          *
2975          * == A ==
2976          * * [[Ape]]
2977          *
2978          * etc., assuming for the sake of argument that $wgCapitalLinks is false.
2979          *
2980          * @param string $string UTF-8 string
2981          * @return string UTF-8 string corresponding to the first letter of input
2982          */
2983         public function firstLetterForLists( $string ) {
2984                 if ( $string[0] == "\0" ) {
2985                         $string = substr( $string, 1 );
2986                 }
2987                 return strtoupper( $this->firstChar( $string ) );
2988         }
2989 }