Localization update.
[mediawiki.git] / languages / Language.php
blob97f441babc7bece73fe56f45a296feff1740b046
1 <?php
2 /**
3 * @defgroup Language Language
5 * @file
6 * @ingroup Language
7 */
9 if( !defined( 'MEDIAWIKI' ) ) {
10 echo "This file is part of MediaWiki, it is not a valid entry point.\n";
11 exit( 1 );
14 # Read language names
15 global $wgLanguageNames;
16 require_once( dirname(__FILE__) . '/Names.php' ) ;
18 global $wgInputEncoding, $wgOutputEncoding;
20 /**
21 * These are always UTF-8, they exist only for backwards compatibility
23 $wgInputEncoding = "UTF-8";
24 $wgOutputEncoding = "UTF-8";
26 if( function_exists( 'mb_strtoupper' ) ) {
27 mb_internal_encoding('UTF-8');
30 /**
31 * a fake language converter
33 * @ingroup Language
35 class FakeConverter {
36 var $mLang;
37 function FakeConverter($langobj) {$this->mLang = $langobj;}
38 function convert($t, $i) {return $t;}
39 function parserConvert($t, $p) {return $t;}
40 function getVariants() { return array( $this->mLang->getCode() ); }
41 function getPreferredVariant() {return $this->mLang->getCode(); }
42 function findVariantLink(&$l, &$n, $ignoreOtherCond = false) {}
43 function getExtraHashOptions() {return '';}
44 function getParsedTitle() {return '';}
45 function markNoConversion($text, $noParse=false) {return $text;}
46 function convertCategoryKey( $key ) {return $key; }
47 function convertLinkToAllVariants($text){ return array( $this->mLang->getCode() => $text); }
48 function armourMath($text){ return $text; }
51 /**
52 * Internationalisation code
53 * @ingroup Language
55 class Language {
56 var $mConverter, $mVariants, $mCode, $mLoaded = false;
57 var $mMagicExtensions = array(), $mMagicHookDone = false;
59 static public $mLocalisationKeys = array(
60 'fallback', 'namespaceNames', 'mathNames', 'bookstoreList',
61 'magicWords', 'messages', 'rtl', 'digitTransformTable',
62 'separatorTransformTable', 'fallback8bitEncoding', 'linkPrefixExtension',
63 'defaultUserOptionOverrides', 'linkTrail', 'namespaceAliases',
64 'dateFormats', 'datePreferences', 'datePreferenceMigrationMap',
65 'defaultDateFormat', 'extraUserToggles', 'specialPageAliases',
66 'imageFiles'
69 static public $mMergeableMapKeys = array( 'messages', 'namespaceNames', 'mathNames',
70 'dateFormats', 'defaultUserOptionOverrides', 'magicWords', 'imageFiles' );
72 static public $mMergeableListKeys = array( 'extraUserToggles' );
74 static public $mMergeableAliasListKeys = array( 'specialPageAliases' );
76 static public $mLocalisationCache = array();
77 static public $mLangObjCache = array();
79 static public $mWeekdayMsgs = array(
80 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
81 'friday', 'saturday'
84 static public $mWeekdayAbbrevMsgs = array(
85 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
88 static public $mMonthMsgs = array(
89 'january', 'february', 'march', 'april', 'may_long', 'june',
90 'july', 'august', 'september', 'october', 'november',
91 'december'
93 static public $mMonthGenMsgs = array(
94 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
95 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
96 'december-gen'
98 static public $mMonthAbbrevMsgs = array(
99 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
100 'sep', 'oct', 'nov', 'dec'
103 static public $mIranianCalendarMonthMsgs = array(
104 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
105 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
106 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
107 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
110 static public $mHebrewCalendarMonthMsgs = array(
111 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
112 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
113 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
114 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
115 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
118 static public $mHebrewCalendarMonthGenMsgs = array(
119 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
120 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
121 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
122 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
123 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
126 static public $mHijriCalendarMonthMsgs = array(
127 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
128 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
129 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
130 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
134 * Get a cached language object for a given language code
136 static function factory( $code ) {
137 if ( !isset( self::$mLangObjCache[$code] ) ) {
138 if( count( self::$mLangObjCache ) > 10 ) {
139 // Don't keep a billion objects around, that's stupid.
140 self::$mLangObjCache = array();
142 self::$mLangObjCache[$code] = self::newFromCode( $code );
144 return self::$mLangObjCache[$code];
148 * Create a language object for a given language code
150 protected static function newFromCode( $code ) {
151 global $IP;
152 static $recursionLevel = 0;
153 if ( $code == 'en' ) {
154 $class = 'Language';
155 } else {
156 $class = 'Language' . str_replace( '-', '_', ucfirst( $code ) );
157 // Preload base classes to work around APC/PHP5 bug
158 if ( file_exists( "$IP/languages/classes/$class.deps.php" ) ) {
159 include_once("$IP/languages/classes/$class.deps.php");
161 if ( file_exists( "$IP/languages/classes/$class.php" ) ) {
162 include_once("$IP/languages/classes/$class.php");
166 if ( $recursionLevel > 5 ) {
167 throw new MWException( "Language fallback loop detected when creating class $class\n" );
170 if( ! class_exists( $class ) ) {
171 $fallback = Language::getFallbackFor( $code );
172 ++$recursionLevel;
173 $lang = Language::newFromCode( $fallback );
174 --$recursionLevel;
175 $lang->setCode( $code );
176 } else {
177 $lang = new $class;
179 return $lang;
182 function __construct() {
183 $this->mConverter = new FakeConverter($this);
184 // Set the code to the name of the descendant
185 if ( get_class( $this ) == 'Language' ) {
186 $this->mCode = 'en';
187 } else {
188 $this->mCode = str_replace( '_', '-', strtolower( substr( get_class( $this ), 8 ) ) );
193 * Reduce memory usage
195 function __destruct() {
196 foreach ( $this as $name => $value ) {
197 unset( $this->$name );
202 * Hook which will be called if this is the content language.
203 * Descendants can use this to register hook functions or modify globals
205 function initContLang() {}
208 * @deprecated Use User::getDefaultOptions()
209 * @return array
211 function getDefaultUserOptions() {
212 wfDeprecated( __METHOD__ );
213 return User::getDefaultOptions();
216 function getFallbackLanguageCode() {
217 return self::getFallbackFor( $this->mCode );
221 * Exports $wgBookstoreListEn
222 * @return array
224 function getBookstoreList() {
225 $this->load();
226 return $this->bookstoreList;
230 * @return array
232 function getNamespaces() {
233 $this->load();
234 return $this->namespaceNames;
238 * A convenience function that returns the same thing as
239 * getNamespaces() except with the array values changed to ' '
240 * where it found '_', useful for producing output to be displayed
241 * e.g. in <select> forms.
243 * @return array
245 function getFormattedNamespaces() {
246 $ns = $this->getNamespaces();
247 foreach($ns as $k => $v) {
248 $ns[$k] = strtr($v, '_', ' ');
250 return $ns;
254 * Get a namespace value by key
255 * <code>
256 * $mw_ns = $wgContLang->getNsText( NS_MEDIAWIKI );
257 * echo $mw_ns; // prints 'MediaWiki'
258 * </code>
260 * @param $index Int: the array key of the namespace to return
261 * @return mixed, string if the namespace value exists, otherwise false
263 function getNsText( $index ) {
264 $ns = $this->getNamespaces();
265 return isset( $ns[$index] ) ? $ns[$index] : false;
269 * A convenience function that returns the same thing as
270 * getNsText() except with '_' changed to ' ', useful for
271 * producing output.
273 * @return array
275 function getFormattedNsText( $index ) {
276 $ns = $this->getNsText( $index );
277 return strtr($ns, '_', ' ');
281 * Get a namespace key by value, case insensitive.
282 * Only matches namespace names for the current language, not the
283 * canonical ones defined in Namespace.php.
285 * @param $text String
286 * @return mixed An integer if $text is a valid value otherwise false
288 function getLocalNsIndex( $text ) {
289 $this->load();
290 $lctext = $this->lc($text);
291 return isset( $this->mNamespaceIds[$lctext] ) ? $this->mNamespaceIds[$lctext] : false;
295 * Get a namespace key by value, case insensitive. Canonical namespace
296 * names override custom ones defined for the current language.
298 * @param $text String
299 * @return mixed An integer if $text is a valid value otherwise false
301 function getNsIndex( $text ) {
302 $this->load();
303 $lctext = $this->lc($text);
304 if( ( $ns = MWNamespace::getCanonicalIndex( $lctext ) ) !== null ) return $ns;
305 return isset( $this->mNamespaceIds[$lctext] ) ? $this->mNamespaceIds[$lctext] : false;
309 * short names for language variants used for language conversion links.
311 * @param $code String
312 * @return string
314 function getVariantname( $code ) {
315 return $this->getMessageFromDB( "variantname-$code" );
318 function specialPage( $name ) {
319 $aliases = $this->getSpecialPageAliases();
320 if ( isset( $aliases[$name][0] ) ) {
321 $name = $aliases[$name][0];
323 return $this->getNsText( NS_SPECIAL ) . ':' . $name;
326 function getQuickbarSettings() {
327 return array(
328 $this->getMessage( 'qbsettings-none' ),
329 $this->getMessage( 'qbsettings-fixedleft' ),
330 $this->getMessage( 'qbsettings-fixedright' ),
331 $this->getMessage( 'qbsettings-floatingleft' ),
332 $this->getMessage( 'qbsettings-floatingright' )
336 function getMathNames() {
337 $this->load();
338 return $this->mathNames;
341 function getDatePreferences() {
342 $this->load();
343 return $this->datePreferences;
346 function getDateFormats() {
347 $this->load();
348 return $this->dateFormats;
351 function getDefaultDateFormat() {
352 $this->load();
353 return $this->defaultDateFormat;
356 function getDatePreferenceMigrationMap() {
357 $this->load();
358 return $this->datePreferenceMigrationMap;
361 function getImageFile( $image ) {
362 $this->load();
363 return $this->imageFiles[$image];
366 function getDefaultUserOptionOverrides() {
367 $this->load();
368 # XXX - apparently some languageas get empty arrays, didn't get to it yet -- midom
369 if (is_array($this->defaultUserOptionOverrides)) {
370 return $this->defaultUserOptionOverrides;
371 } else {
372 return array();
376 function getExtraUserToggles() {
377 $this->load();
378 return $this->extraUserToggles;
381 function getUserToggle( $tog ) {
382 return $this->getMessageFromDB( "tog-$tog" );
386 * Get language names, indexed by code.
387 * If $customisedOnly is true, only returns codes with a messages file
389 public static function getLanguageNames( $customisedOnly = false ) {
390 global $wgLanguageNames, $wgExtraLanguageNames;
391 $allNames = $wgExtraLanguageNames + $wgLanguageNames;
392 if ( !$customisedOnly ) {
393 return $allNames;
396 global $IP;
397 $names = array();
398 $dir = opendir( "$IP/languages/messages" );
399 while( false !== ( $file = readdir( $dir ) ) ) {
400 $m = array();
401 if( preg_match( '/Messages([A-Z][a-z_]+)\.php$/', $file, $m ) ) {
402 $code = str_replace( '_', '-', strtolower( $m[1] ) );
403 if ( isset( $allNames[$code] ) ) {
404 $names[$code] = $allNames[$code];
408 closedir( $dir );
409 return $names;
413 * Get a message from the MediaWiki namespace.
415 * @param $msg String: message name
416 * @return string
418 function getMessageFromDB( $msg ) {
419 return wfMsgExt( $msg, array( 'parsemag', 'language' => $this ) );
422 function getLanguageName( $code ) {
423 $names = self::getLanguageNames();
424 if ( !array_key_exists( $code, $names ) ) {
425 return '';
427 return $names[$code];
430 function getMonthName( $key ) {
431 return $this->getMessageFromDB( self::$mMonthMsgs[$key-1] );
434 function getMonthNameGen( $key ) {
435 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key-1] );
438 function getMonthAbbreviation( $key ) {
439 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key-1] );
442 function getWeekdayName( $key ) {
443 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key-1] );
446 function getWeekdayAbbreviation( $key ) {
447 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key-1] );
450 function getIranianCalendarMonthName( $key ) {
451 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key-1] );
454 function getHebrewCalendarMonthName( $key ) {
455 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key-1] );
458 function getHebrewCalendarMonthNameGen( $key ) {
459 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key-1] );
462 function getHijriCalendarMonthName( $key ) {
463 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key-1] );
467 * Used by date() and time() to adjust the time output.
469 * @param $ts Int the time in date('YmdHis') format
470 * @param $tz Mixed: adjust the time by this amount (default false, mean we
471 * get user timecorrection setting)
472 * @return int
474 function userAdjust( $ts, $tz = false ) {
475 global $wgUser, $wgLocalTZoffset;
477 if ( $tz === false ) {
478 $tz = $wgUser->getOption( 'timecorrection' );
481 $data = explode( '|', $tz, 3 );
483 if ( $data[0] == 'ZoneInfo' ) {
484 if ( function_exists( 'timezone_open' ) && @timezone_open( $data[2] ) !== false ) {
485 $date = date_create( $ts, timezone_open( 'UTC' ) );
486 date_timezone_set( $date, timezone_open( $data[2] ) );
487 $date = date_format( $date, 'YmdHis' );
488 return $date;
490 # Unrecognized timezone, default to 'Offset' with the stored offset.
491 $data[0] = 'Offset';
494 $minDiff = 0;
495 if ( $data[0] == 'System' || $tz == '' ) {
496 # Global offset in minutes.
497 if( isset($wgLocalTZoffset) ) $minDiff = $wgLocalTZoffset;
498 } else if ( $data[0] == 'Offset' ) {
499 $minDiff = intval( $data[1] );
500 } else {
501 $data = explode( ':', $tz );
502 if( count( $data ) == 2 ) {
503 $data[0] = intval( $data[0] );
504 $data[1] = intval( $data[1] );
505 $minDiff = abs( $data[0] ) * 60 + $data[1];
506 if ( $data[0] < 0 ) $minDiff = -$minDiff;
507 } else {
508 $minDiff = intval( $data[0] ) * 60;
512 # No difference ? Return time unchanged
513 if ( 0 == $minDiff ) return $ts;
515 wfSuppressWarnings(); // E_STRICT system time bitching
516 # Generate an adjusted date; take advantage of the fact that mktime
517 # will normalize out-of-range values so we don't have to split $minDiff
518 # into hours and minutes.
519 $t = mktime( (
520 (int)substr( $ts, 8, 2) ), # Hours
521 (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
522 (int)substr( $ts, 12, 2 ), # Seconds
523 (int)substr( $ts, 4, 2 ), # Month
524 (int)substr( $ts, 6, 2 ), # Day
525 (int)substr( $ts, 0, 4 ) ); #Year
527 $date = date( 'YmdHis', $t );
528 wfRestoreWarnings();
530 return $date;
534 * This is a workalike of PHP's date() function, but with better
535 * internationalisation, a reduced set of format characters, and a better
536 * escaping format.
538 * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrU. See the
539 * PHP manual for definitions. "o" format character is supported since
540 * PHP 5.1.0, previous versions return literal o.
541 * There are a number of extensions, which start with "x":
543 * xn Do not translate digits of the next numeric format character
544 * xN Toggle raw digit (xn) flag, stays set until explicitly unset
545 * xr Use roman numerals for the next numeric format character
546 * xh Use hebrew numerals for the next numeric format character
547 * xx Literal x
548 * xg Genitive month name
550 * xij j (day number) in Iranian calendar
551 * xiF F (month name) in Iranian calendar
552 * xin n (month number) in Iranian calendar
553 * xiY Y (full year) in Iranian calendar
555 * xjj j (day number) in Hebrew calendar
556 * xjF F (month name) in Hebrew calendar
557 * xjt t (days in month) in Hebrew calendar
558 * xjx xg (genitive month name) in Hebrew calendar
559 * xjn n (month number) in Hebrew calendar
560 * xjY Y (full year) in Hebrew calendar
562 * xmj j (day number) in Hijri calendar
563 * xmF F (month name) in Hijri calendar
564 * xmn n (month number) in Hijri calendar
565 * xmY Y (full year) in Hijri calendar
567 * xkY Y (full year) in Thai solar calendar. Months and days are
568 * identical to the Gregorian calendar
570 * Characters enclosed in double quotes will be considered literal (with
571 * the quotes themselves removed). Unmatched quotes will be considered
572 * literal quotes. Example:
574 * "The month is" F => The month is January
575 * i's" => 20'11"
577 * Backslash escaping is also supported.
579 * Input timestamp is assumed to be pre-normalized to the desired local
580 * time zone, if any.
582 * @param $format String
583 * @param $ts String: 14-character timestamp
584 * YYYYMMDDHHMMSS
585 * 01234567890123
586 * @todo emulation of "o" format character for PHP pre 5.1.0
587 * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
589 function sprintfDate( $format, $ts ) {
590 $s = '';
591 $raw = false;
592 $roman = false;
593 $hebrewNum = false;
594 $unix = false;
595 $rawToggle = false;
596 $iranian = false;
597 $hebrew = false;
598 $hijri = false;
599 $thai = false;
600 for ( $p = 0; $p < strlen( $format ); $p++ ) {
601 $num = false;
602 $code = $format[$p];
603 if ( $code == 'x' && $p < strlen( $format ) - 1 ) {
604 $code .= $format[++$p];
607 if ( ( $code === 'xi' || $code == 'xj' || $code == 'xk' || $code == 'xm' ) && $p < strlen( $format ) - 1 ) {
608 $code .= $format[++$p];
611 switch ( $code ) {
612 case 'xx':
613 $s .= 'x';
614 break;
615 case 'xn':
616 $raw = true;
617 break;
618 case 'xN':
619 $rawToggle = !$rawToggle;
620 break;
621 case 'xr':
622 $roman = true;
623 break;
624 case 'xh':
625 $hebrewNum = true;
626 break;
627 case 'xg':
628 $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
629 break;
630 case 'xjx':
631 if ( !$hebrew ) $hebrew = self::tsToHebrew( $ts );
632 $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
633 break;
634 case 'd':
635 $num = substr( $ts, 6, 2 );
636 break;
637 case 'D':
638 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
639 $s .= $this->getWeekdayAbbreviation( gmdate( 'w', $unix ) + 1 );
640 break;
641 case 'j':
642 $num = intval( substr( $ts, 6, 2 ) );
643 break;
644 case 'xij':
645 if ( !$iranian ) $iranian = self::tsToIranian( $ts );
646 $num = $iranian[2];
647 break;
648 case 'xmj':
649 if ( !$hijri ) $hijri = self::tsToHijri( $ts );
650 $num = $hijri[2];
651 break;
652 case 'xjj':
653 if ( !$hebrew ) $hebrew = self::tsToHebrew( $ts );
654 $num = $hebrew[2];
655 break;
656 case 'l':
657 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
658 $s .= $this->getWeekdayName( gmdate( 'w', $unix ) + 1 );
659 break;
660 case 'N':
661 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
662 $w = gmdate( 'w', $unix );
663 $num = $w ? $w : 7;
664 break;
665 case 'w':
666 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
667 $num = gmdate( 'w', $unix );
668 break;
669 case 'z':
670 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
671 $num = gmdate( 'z', $unix );
672 break;
673 case 'W':
674 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
675 $num = gmdate( 'W', $unix );
676 break;
677 case 'F':
678 $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
679 break;
680 case 'xiF':
681 if ( !$iranian ) $iranian = self::tsToIranian( $ts );
682 $s .= $this->getIranianCalendarMonthName( $iranian[1] );
683 break;
684 case 'xmF':
685 if ( !$hijri ) $hijri = self::tsToHijri( $ts );
686 $s .= $this->getHijriCalendarMonthName( $hijri[1] );
687 break;
688 case 'xjF':
689 if ( !$hebrew ) $hebrew = self::tsToHebrew( $ts );
690 $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
691 break;
692 case 'm':
693 $num = substr( $ts, 4, 2 );
694 break;
695 case 'M':
696 $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
697 break;
698 case 'n':
699 $num = intval( substr( $ts, 4, 2 ) );
700 break;
701 case 'xin':
702 if ( !$iranian ) $iranian = self::tsToIranian( $ts );
703 $num = $iranian[1];
704 break;
705 case 'xmn':
706 if ( !$hijri ) $hijri = self::tsToHijri ( $ts );
707 $num = $hijri[1];
708 break;
709 case 'xjn':
710 if ( !$hebrew ) $hebrew = self::tsToHebrew( $ts );
711 $num = $hebrew[1];
712 break;
713 case 't':
714 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
715 $num = gmdate( 't', $unix );
716 break;
717 case 'xjt':
718 if ( !$hebrew ) $hebrew = self::tsToHebrew( $ts );
719 $num = $hebrew[3];
720 break;
721 case 'L':
722 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
723 $num = gmdate( 'L', $unix );
724 break;
725 # 'o' is supported since PHP 5.1.0
726 # return literal if not supported
727 # TODO: emulation for pre 5.1.0 versions
728 case 'o':
729 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
730 if ( version_compare(PHP_VERSION, '5.1.0') === 1 )
731 $num = date( 'o', $unix );
732 else
733 $s .= 'o';
734 break;
735 case 'Y':
736 $num = substr( $ts, 0, 4 );
737 break;
738 case 'xiY':
739 if ( !$iranian ) $iranian = self::tsToIranian( $ts );
740 $num = $iranian[0];
741 break;
742 case 'xmY':
743 if ( !$hijri ) $hijri = self::tsToHijri( $ts );
744 $num = $hijri[0];
745 break;
746 case 'xjY':
747 if ( !$hebrew ) $hebrew = self::tsToHebrew( $ts );
748 $num = $hebrew[0];
749 break;
750 case 'xkY':
751 if ( !$thai ) $thai = self::tsToThai( $ts );
752 $num = $thai[0];
753 break;
754 case 'y':
755 $num = substr( $ts, 2, 2 );
756 break;
757 case 'a':
758 $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
759 break;
760 case 'A':
761 $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
762 break;
763 case 'g':
764 $h = substr( $ts, 8, 2 );
765 $num = $h % 12 ? $h % 12 : 12;
766 break;
767 case 'G':
768 $num = intval( substr( $ts, 8, 2 ) );
769 break;
770 case 'h':
771 $h = substr( $ts, 8, 2 );
772 $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
773 break;
774 case 'H':
775 $num = substr( $ts, 8, 2 );
776 break;
777 case 'i':
778 $num = substr( $ts, 10, 2 );
779 break;
780 case 's':
781 $num = substr( $ts, 12, 2 );
782 break;
783 case 'c':
784 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
785 $s .= gmdate( 'c', $unix );
786 break;
787 case 'r':
788 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
789 $s .= gmdate( 'r', $unix );
790 break;
791 case 'U':
792 if ( !$unix ) $unix = wfTimestamp( TS_UNIX, $ts );
793 $num = $unix;
794 break;
795 case '\\':
796 # Backslash escaping
797 if ( $p < strlen( $format ) - 1 ) {
798 $s .= $format[++$p];
799 } else {
800 $s .= '\\';
802 break;
803 case '"':
804 # Quoted literal
805 if ( $p < strlen( $format ) - 1 ) {
806 $endQuote = strpos( $format, '"', $p + 1 );
807 if ( $endQuote === false ) {
808 # No terminating quote, assume literal "
809 $s .= '"';
810 } else {
811 $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
812 $p = $endQuote;
814 } else {
815 # Quote at end of string, assume literal "
816 $s .= '"';
818 break;
819 default:
820 $s .= $format[$p];
822 if ( $num !== false ) {
823 if ( $rawToggle || $raw ) {
824 $s .= $num;
825 $raw = false;
826 } elseif ( $roman ) {
827 $s .= self::romanNumeral( $num );
828 $roman = false;
829 } elseif( $hebrewNum ) {
830 $s .= self::hebrewNumeral( $num );
831 $hebrewNum = false;
832 } else {
833 $s .= $this->formatNum( $num, true );
835 $num = false;
838 return $s;
841 private static $GREG_DAYS = array( 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 );
842 private static $IRANIAN_DAYS = array( 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 );
844 * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
845 * Gregorian dates to Iranian dates. Originally written in C, it
846 * is released under the terms of GNU Lesser General Public
847 * License. Conversion to PHP was performed by Niklas Laxström.
849 * Link: http://www.farsiweb.info/jalali/jalali.c
851 private static function tsToIranian( $ts ) {
852 $gy = substr( $ts, 0, 4 ) -1600;
853 $gm = substr( $ts, 4, 2 ) -1;
854 $gd = substr( $ts, 6, 2 ) -1;
856 # Days passed from the beginning (including leap years)
857 $gDayNo = 365*$gy
858 + floor(($gy+3) / 4)
859 - floor(($gy+99) / 100)
860 + floor(($gy+399) / 400);
863 // Add days of the past months of this year
864 for( $i = 0; $i < $gm; $i++ ) {
865 $gDayNo += self::$GREG_DAYS[$i];
868 // Leap years
869 if ( $gm > 1 && (($gy%4===0 && $gy%100!==0 || ($gy%400==0)))) {
870 $gDayNo++;
873 // Days passed in current month
874 $gDayNo += $gd;
876 $jDayNo = $gDayNo - 79;
878 $jNp = floor($jDayNo / 12053);
879 $jDayNo %= 12053;
881 $jy = 979 + 33*$jNp + 4*floor($jDayNo/1461);
882 $jDayNo %= 1461;
884 if ( $jDayNo >= 366 ) {
885 $jy += floor(($jDayNo-1)/365);
886 $jDayNo = floor(($jDayNo-1)%365);
889 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
890 $jDayNo -= self::$IRANIAN_DAYS[$i];
893 $jm= $i+1;
894 $jd= $jDayNo+1;
896 return array($jy, $jm, $jd);
899 * Converting Gregorian dates to Hijri dates.
901 * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
903 * @link http://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
905 private static function tsToHijri ( $ts ) {
906 $year = substr( $ts, 0, 4 );
907 $month = substr( $ts, 4, 2 );
908 $day = substr( $ts, 6, 2 );
910 $zyr = $year;
911 $zd=$day;
912 $zm=$month;
913 $zy=$zyr;
917 if (($zy>1582)||(($zy==1582)&&($zm>10))||(($zy==1582)&&($zm==10)&&($zd>14)))
921 $zjd=(int)((1461*($zy + 4800 + (int)( ($zm-14) /12) ))/4) + (int)((367*($zm-2-12*((int)(($zm-14)/12))))/12)-(int)((3*(int)(( ($zy+4900+(int)(($zm-14)/12))/100)))/4)+$zd-32075;
923 else
925 $zjd = 367*$zy-(int)((7*($zy+5001+(int)(($zm-9)/7)))/4)+(int)((275*$zm)/9)+$zd+1729777;
928 $zl=$zjd-1948440+10632;
929 $zn=(int)(($zl-1)/10631);
930 $zl=$zl-10631*$zn+354;
931 $zj=((int)((10985-$zl)/5316))*((int)((50*$zl)/17719))+((int)($zl/5670))*((int)((43*$zl)/15238));
932 $zl=$zl-((int)((30-$zj)/15))*((int)((17719*$zj)/50))-((int)($zj/16))*((int)((15238*$zj)/43))+29;
933 $zm=(int)((24*$zl)/709);
934 $zd=$zl-(int)((709*$zm)/24);
935 $zy=30*$zn+$zj-30;
937 return array ($zy, $zm, $zd);
941 * Converting Gregorian dates to Hebrew dates.
943 * Based on a JavaScript code by Abu Mami and Yisrael Hersch
944 * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
945 * to translate the relevant functions into PHP and release them under
946 * GNU GPL.
948 * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
949 * and Adar II is 14. In a non-leap year, Adar is 6.
951 private static function tsToHebrew( $ts ) {
952 # Parse date
953 $year = substr( $ts, 0, 4 );
954 $month = substr( $ts, 4, 2 );
955 $day = substr( $ts, 6, 2 );
957 # Calculate Hebrew year
958 $hebrewYear = $year + 3760;
960 # Month number when September = 1, August = 12
961 $month += 4;
962 if( $month > 12 ) {
963 # Next year
964 $month -= 12;
965 $year++;
966 $hebrewYear++;
969 # Calculate day of year from 1 September
970 $dayOfYear = $day;
971 for( $i = 1; $i < $month; $i++ ) {
972 if( $i == 6 ) {
973 # February
974 $dayOfYear += 28;
975 # Check if the year is leap
976 if( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
977 $dayOfYear++;
979 } elseif( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
980 $dayOfYear += 30;
981 } else {
982 $dayOfYear += 31;
986 # Calculate the start of the Hebrew year
987 $start = self::hebrewYearStart( $hebrewYear );
989 # Calculate next year's start
990 if( $dayOfYear <= $start ) {
991 # Day is before the start of the year - it is the previous year
992 # Next year's start
993 $nextStart = $start;
994 # Previous year
995 $year--;
996 $hebrewYear--;
997 # Add days since previous year's 1 September
998 $dayOfYear += 365;
999 if( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1000 # Leap year
1001 $dayOfYear++;
1003 # Start of the new (previous) year
1004 $start = self::hebrewYearStart( $hebrewYear );
1005 } else {
1006 # Next year's start
1007 $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1010 # Calculate Hebrew day of year
1011 $hebrewDayOfYear = $dayOfYear - $start;
1013 # Difference between year's days
1014 $diff = $nextStart - $start;
1015 # Add 12 (or 13 for leap years) days to ignore the difference between
1016 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1017 # difference is only about the year type
1018 if( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1019 $diff += 13;
1020 } else {
1021 $diff += 12;
1024 # Check the year pattern, and is leap year
1025 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1026 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1027 # and non-leap years
1028 $yearPattern = $diff % 30;
1029 # Check if leap year
1030 $isLeap = $diff >= 30;
1032 # Calculate day in the month from number of day in the Hebrew year
1033 # Don't check Adar - if the day is not in Adar, we will stop before;
1034 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1035 $hebrewDay = $hebrewDayOfYear;
1036 $hebrewMonth = 1;
1037 $days = 0;
1038 while( $hebrewMonth <= 12 ) {
1039 # Calculate days in this month
1040 if( $isLeap && $hebrewMonth == 6 ) {
1041 # Adar in a leap year
1042 if( $isLeap ) {
1043 # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1044 $days = 30;
1045 if( $hebrewDay <= $days ) {
1046 # Day in Adar I
1047 $hebrewMonth = 13;
1048 } else {
1049 # Subtract the days of Adar I
1050 $hebrewDay -= $days;
1051 # Try Adar II
1052 $days = 29;
1053 if( $hebrewDay <= $days ) {
1054 # Day in Adar II
1055 $hebrewMonth = 14;
1059 } elseif( $hebrewMonth == 2 && $yearPattern == 2 ) {
1060 # Cheshvan in a complete year (otherwise as the rule below)
1061 $days = 30;
1062 } elseif( $hebrewMonth == 3 && $yearPattern == 0 ) {
1063 # Kislev in an incomplete year (otherwise as the rule below)
1064 $days = 29;
1065 } else {
1066 # Odd months have 30 days, even have 29
1067 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1069 if( $hebrewDay <= $days ) {
1070 # In the current month
1071 break;
1072 } else {
1073 # Subtract the days of the current month
1074 $hebrewDay -= $days;
1075 # Try in the next month
1076 $hebrewMonth++;
1080 return array( $hebrewYear, $hebrewMonth, $hebrewDay, $days );
1084 * This calculates the Hebrew year start, as days since 1 September.
1085 * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1086 * Used for Hebrew date.
1088 private static function hebrewYearStart( $year ) {
1089 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1090 $b = intval( ( $year - 1 ) % 4 );
1091 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1092 if( $m < 0 ) {
1093 $m--;
1095 $Mar = intval( $m );
1096 if( $m < 0 ) {
1097 $m++;
1099 $m -= $Mar;
1101 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7);
1102 if( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1103 $Mar++;
1104 } else if( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1105 $Mar += 2;
1106 } else if( $c == 2 || $c == 4 || $c == 6 ) {
1107 $Mar++;
1110 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1111 return $Mar;
1115 * Algorithm to convert Gregorian dates to Thai solar dates.
1117 * Link: http://en.wikipedia.org/wiki/Thai_solar_calendar
1119 * @param $ts String: 14-character timestamp
1120 * @return array converted year, month, day
1122 private static function tsToThai( $ts ) {
1123 $gy = substr( $ts, 0, 4 );
1124 $gm = substr( $ts, 4, 2 );
1125 $gd = substr( $ts, 6, 2 );
1127 # Add 543 years to the Gregorian calendar
1128 # Months and days are identical
1129 $gy_thai = $gy + 543;
1131 return array( $gy_thai, $gm, $gd );
1136 * Roman number formatting up to 3000
1138 static function romanNumeral( $num ) {
1139 static $table = array(
1140 array( '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ),
1141 array( '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ),
1142 array( '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ),
1143 array( '', 'M', 'MM', 'MMM' )
1146 $num = intval( $num );
1147 if ( $num > 3000 || $num <= 0 ) {
1148 return $num;
1151 $s = '';
1152 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1153 if ( $num >= $pow10 ) {
1154 $s .= $table[$i][floor($num / $pow10)];
1156 $num = $num % $pow10;
1158 return $s;
1162 * Hebrew Gematria number formatting up to 9999
1164 static function hebrewNumeral( $num ) {
1165 static $table = array(
1166 array( '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ),
1167 array( '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ),
1168 array( '', 'ק', 'ר', 'ש', 'ת', 'תק', 'תר', 'תש', 'תת', 'תתק', 'תתר' ),
1169 array( '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' )
1172 $num = intval( $num );
1173 if ( $num > 9999 || $num <= 0 ) {
1174 return $num;
1177 $s = '';
1178 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1179 if ( $num >= $pow10 ) {
1180 if ( $num == 15 || $num == 16 ) {
1181 $s .= $table[0][9] . $table[0][$num - 9];
1182 $num = 0;
1183 } else {
1184 $s .= $table[$i][intval( ( $num / $pow10 ) )];
1185 if( $pow10 == 1000 ) {
1186 $s .= "'";
1190 $num = $num % $pow10;
1192 if( strlen( $s ) == 2 ) {
1193 $str = $s . "'";
1194 } else {
1195 $str = substr( $s, 0, strlen( $s ) - 2 ) . '"';
1196 $str .= substr( $s, strlen( $s ) - 2, 2 );
1198 $start = substr( $str, 0, strlen( $str ) - 2 );
1199 $end = substr( $str, strlen( $str ) - 2 );
1200 switch( $end ) {
1201 case 'כ':
1202 $str = $start . 'ך';
1203 break;
1204 case 'מ':
1205 $str = $start . 'ם';
1206 break;
1207 case 'נ':
1208 $str = $start . 'ן';
1209 break;
1210 case 'פ':
1211 $str = $start . 'ף';
1212 break;
1213 case 'צ':
1214 $str = $start . 'ץ';
1215 break;
1217 return $str;
1221 * This is meant to be used by time(), date(), and timeanddate() to get
1222 * the date preference they're supposed to use, it should be used in
1223 * all children.
1225 *<code>
1226 * function timeanddate([...], $format = true) {
1227 * $datePreference = $this->dateFormat($format);
1228 * [...]
1230 *</code>
1232 * @param $usePrefs Mixed: if true, the user's preference is used
1233 * if false, the site/language default is used
1234 * if int/string, assumed to be a format.
1235 * @return string
1237 function dateFormat( $usePrefs = true ) {
1238 global $wgUser;
1240 if( is_bool( $usePrefs ) ) {
1241 if( $usePrefs ) {
1242 $datePreference = $wgUser->getDatePreference();
1243 } else {
1244 $options = User::getDefaultOptions();
1245 $datePreference = (string)$options['date'];
1247 } else {
1248 $datePreference = (string)$usePrefs;
1251 // return int
1252 if( $datePreference == '' ) {
1253 return 'default';
1256 return $datePreference;
1260 * @param $ts Mixed: the time format which needs to be turned into a
1261 * date('YmdHis') format with wfTimestamp(TS_MW,$ts)
1262 * @param $adj Bool: whether to adjust the time output according to the
1263 * user configured offset ($timecorrection)
1264 * @param $format Mixed: true to use user's date format preference
1265 * @param $timecorrection String: the time offset as returned by
1266 * validateTimeZone() in Special:Preferences
1267 * @return string
1269 function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
1270 $this->load();
1271 if ( $adj ) {
1272 $ts = $this->userAdjust( $ts, $timecorrection );
1275 $pref = $this->dateFormat( $format );
1276 if( $pref == 'default' || !isset( $this->dateFormats["$pref date"] ) ) {
1277 $pref = $this->defaultDateFormat;
1279 return $this->sprintfDate( $this->dateFormats["$pref date"], $ts );
1283 * @param $ts Mixed: the time format which needs to be turned into a
1284 * date('YmdHis') format with wfTimestamp(TS_MW,$ts)
1285 * @param $adj Bool: whether to adjust the time output according to the
1286 * user configured offset ($timecorrection)
1287 * @param $format Mixed: true to use user's date format preference
1288 * @param $timecorrection String: the time offset as returned by
1289 * validateTimeZone() in Special:Preferences
1290 * @return string
1292 function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
1293 $this->load();
1294 if ( $adj ) {
1295 $ts = $this->userAdjust( $ts, $timecorrection );
1298 $pref = $this->dateFormat( $format );
1299 if( $pref == 'default' || !isset( $this->dateFormats["$pref time"] ) ) {
1300 $pref = $this->defaultDateFormat;
1302 return $this->sprintfDate( $this->dateFormats["$pref time"], $ts );
1306 * @param $ts Mixed: the time format which needs to be turned into a
1307 * date('YmdHis') format with wfTimestamp(TS_MW,$ts)
1308 * @param $adj Bool: whether to adjust the time output according to the
1309 * user configured offset ($timecorrection)
1310 * @param $format Mixed: what format to return, if it's false output the
1311 * default one (default true)
1312 * @param $timecorrection String: the time offset as returned by
1313 * validateTimeZone() in Special:Preferences
1314 * @return string
1316 function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false) {
1317 $this->load();
1319 $ts = wfTimestamp( TS_MW, $ts );
1321 if ( $adj ) {
1322 $ts = $this->userAdjust( $ts, $timecorrection );
1325 $pref = $this->dateFormat( $format );
1326 if( $pref == 'default' || !isset( $this->dateFormats["$pref both"] ) ) {
1327 $pref = $this->defaultDateFormat;
1330 return $this->sprintfDate( $this->dateFormats["$pref both"], $ts );
1333 function getMessage( $key ) {
1334 $this->load();
1335 return isset( $this->messages[$key] ) ? $this->messages[$key] : null;
1338 function getAllMessages() {
1339 $this->load();
1340 return $this->messages;
1343 function iconv( $in, $out, $string ) {
1344 # For most languages, this is a wrapper for iconv
1345 return iconv( $in, $out . '//IGNORE', $string );
1348 // callback functions for uc(), lc(), ucwords(), ucwordbreaks()
1349 function ucwordbreaksCallbackAscii($matches){
1350 return $this->ucfirst($matches[1]);
1353 function ucwordbreaksCallbackMB($matches){
1354 return mb_strtoupper($matches[0]);
1357 function ucCallback($matches){
1358 list( $wikiUpperChars ) = self::getCaseMaps();
1359 return strtr( $matches[1], $wikiUpperChars );
1362 function lcCallback($matches){
1363 list( , $wikiLowerChars ) = self::getCaseMaps();
1364 return strtr( $matches[1], $wikiLowerChars );
1367 function ucwordsCallbackMB($matches){
1368 return mb_strtoupper($matches[0]);
1371 function ucwordsCallbackWiki($matches){
1372 list( $wikiUpperChars ) = self::getCaseMaps();
1373 return strtr( $matches[0], $wikiUpperChars );
1376 function ucfirst( $str ) {
1377 if ( empty($str) ) return $str;
1378 if ( ord($str[0]) < 128 ) return ucfirst($str);
1379 else return self::uc($str,true); // fall back to more complex logic in case of multibyte strings
1382 function uc( $str, $first = false ) {
1383 if ( function_exists( 'mb_strtoupper' ) ) {
1384 if ( $first ) {
1385 if ( self::isMultibyte( $str ) ) {
1386 return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
1387 } else {
1388 return ucfirst( $str );
1390 } else {
1391 return self::isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
1393 } else {
1394 if ( self::isMultibyte( $str ) ) {
1395 list( $wikiUpperChars ) = $this->getCaseMaps();
1396 $x = $first ? '^' : '';
1397 return preg_replace_callback(
1398 "/$x([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/",
1399 array($this,"ucCallback"),
1400 $str
1402 } else {
1403 return $first ? ucfirst( $str ) : strtoupper( $str );
1408 function lcfirst( $str ) {
1409 if ( empty($str) ) return $str;
1410 if ( is_string( $str ) && ord($str[0]) < 128 ) {
1411 // editing string in place = cool
1412 $str[0]=strtolower($str[0]);
1413 return $str;
1415 else return self::lc( $str, true );
1418 function lc( $str, $first = false ) {
1419 if ( function_exists( 'mb_strtolower' ) )
1420 if ( $first )
1421 if ( self::isMultibyte( $str ) )
1422 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
1423 else
1424 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
1425 else
1426 return self::isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
1427 else
1428 if ( self::isMultibyte( $str ) ) {
1429 list( , $wikiLowerChars ) = self::getCaseMaps();
1430 $x = $first ? '^' : '';
1431 return preg_replace_callback(
1432 "/$x([A-Z]|[\\xc0-\\xff][\\x80-\\xbf]*)/",
1433 array($this,"lcCallback"),
1434 $str
1436 } else
1437 return $first ? strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 ) : strtolower( $str );
1440 function isMultibyte( $str ) {
1441 return (bool)preg_match( '/[\x80-\xff]/', $str );
1444 function ucwords($str) {
1445 if ( self::isMultibyte( $str ) ) {
1446 $str = self::lc($str);
1448 // regexp to find first letter in each word (i.e. after each space)
1449 $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
1451 // function to use to capitalize a single char
1452 if ( function_exists( 'mb_strtoupper' ) )
1453 return preg_replace_callback(
1454 $replaceRegexp,
1455 array($this,"ucwordsCallbackMB"),
1456 $str
1458 else
1459 return preg_replace_callback(
1460 $replaceRegexp,
1461 array($this,"ucwordsCallbackWiki"),
1462 $str
1465 else
1466 return ucwords( strtolower( $str ) );
1469 # capitalize words at word breaks
1470 function ucwordbreaks($str){
1471 if (self::isMultibyte( $str ) ) {
1472 $str = self::lc($str);
1474 // since \b doesn't work for UTF-8, we explicitely define word break chars
1475 $breaks= "[ \-\(\)\}\{\.,\?!]";
1477 // find first letter after word break
1478 $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
1480 if ( function_exists( 'mb_strtoupper' ) )
1481 return preg_replace_callback(
1482 $replaceRegexp,
1483 array($this,"ucwordbreaksCallbackMB"),
1484 $str
1486 else
1487 return preg_replace_callback(
1488 $replaceRegexp,
1489 array($this,"ucwordsCallbackWiki"),
1490 $str
1493 else
1494 return preg_replace_callback(
1495 '/\b([\w\x80-\xff]+)\b/',
1496 array($this,"ucwordbreaksCallbackAscii"),
1497 $str );
1501 * Return a case-folded representation of $s
1503 * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
1504 * and $s2 are the same except for the case of their characters. It is not
1505 * necessary for the value returned to make sense when displayed.
1507 * Do *not* perform any other normalisation in this function. If a caller
1508 * uses this function when it should be using a more general normalisation
1509 * function, then fix the caller.
1511 function caseFold( $s ) {
1512 return $this->uc( $s );
1515 function checkTitleEncoding( $s ) {
1516 if( is_array( $s ) ) {
1517 wfDebugDieBacktrace( 'Given array to checkTitleEncoding.' );
1519 # Check for non-UTF-8 URLs
1520 $ishigh = preg_match( '/[\x80-\xff]/', $s);
1521 if(!$ishigh) return $s;
1523 $isutf8 = preg_match( '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
1524 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})+$/', $s );
1525 if( $isutf8 ) return $s;
1527 return $this->iconv( $this->fallback8bitEncoding(), "utf-8", $s );
1530 function fallback8bitEncoding() {
1531 $this->load();
1532 return $this->fallback8bitEncoding;
1536 * Some languages have special punctuation to strip out
1537 * or characters which need to be converted for MySQL's
1538 * indexing to grok it correctly. Make such changes here.
1540 * @param $string String
1541 * @return String
1543 function stripForSearch( $string ) {
1544 global $wgDBtype;
1545 if ( $wgDBtype != 'mysql' ) {
1546 return $string;
1550 wfProfileIn( __METHOD__ );
1552 // MySQL fulltext index doesn't grok utf-8, so we
1553 // need to fold cases and convert to hex
1554 $out = preg_replace_callback(
1555 "/([\\xc0-\\xff][\\x80-\\xbf]*)/",
1556 array( $this, 'stripForSearchCallback' ),
1557 $this->lc( $string ) );
1559 // And to add insult to injury, the default indexing
1560 // ignores short words... Pad them so we can pass them
1561 // through without reconfiguring the server...
1562 $minLength = $this->minSearchLength();
1563 if( $minLength > 1 ) {
1564 $n = $minLength-1;
1565 $out = preg_replace(
1566 "/\b(\w{1,$n})\b/",
1567 "$1u800",
1568 $out );
1571 // Periods within things like hostnames and IP addresses
1572 // are also important -- we want a search for "example.com"
1573 // or "192.168.1.1" to work sanely.
1575 // MySQL's search seems to ignore them, so you'd match on
1576 // "example.wikipedia.com" and "192.168.83.1" as well.
1577 $out = preg_replace(
1578 "/(\w)\.(\w|\*)/u",
1579 "$1u82e$2",
1580 $out );
1582 wfProfileOut( __METHOD__ );
1583 return $out;
1587 * Armor a case-folded UTF-8 string to get through MySQL's
1588 * fulltext search without being mucked up by funny charset
1589 * settings or anything else of the sort.
1591 protected function stripForSearchCallback( $matches ) {
1592 return 'u8' . bin2hex( $matches[1] );
1596 * Check MySQL server's ft_min_word_len setting so we know
1597 * if we need to pad short words...
1599 protected function minSearchLength() {
1600 if( !isset( $this->minSearchLength ) ) {
1601 $sql = "show global variables like 'ft\\_min\\_word\\_len'";
1602 $dbr = wfGetDB( DB_SLAVE );
1603 $result = $dbr->query( $sql );
1604 $row = $result->fetchObject();
1605 $result->free();
1607 if( $row && $row->Variable_name == 'ft_min_word_len' ) {
1608 $this->minSearchLength = intval( $row->Value );
1609 } else {
1610 $this->minSearchLength = 0;
1613 return $this->minSearchLength;
1616 function convertForSearchResult( $termsArray ) {
1617 # some languages, e.g. Chinese, need to do a conversion
1618 # in order for search results to be displayed correctly
1619 return $termsArray;
1623 * Get the first character of a string.
1625 * @param $s string
1626 * @return string
1628 function firstChar( $s ) {
1629 $matches = array();
1630 preg_match( '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
1631 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/', $s, $matches);
1633 if ( isset( $matches[1] ) ) {
1634 if ( strlen( $matches[1] ) != 3 ) {
1635 return $matches[1];
1638 // Break down Hangul syllables to grab the first jamo
1639 $code = utf8ToCodepoint( $matches[1] );
1640 if ( $code < 0xac00 || 0xd7a4 <= $code) {
1641 return $matches[1];
1642 } elseif ( $code < 0xb098 ) {
1643 return "\xe3\x84\xb1";
1644 } elseif ( $code < 0xb2e4 ) {
1645 return "\xe3\x84\xb4";
1646 } elseif ( $code < 0xb77c ) {
1647 return "\xe3\x84\xb7";
1648 } elseif ( $code < 0xb9c8 ) {
1649 return "\xe3\x84\xb9";
1650 } elseif ( $code < 0xbc14 ) {
1651 return "\xe3\x85\x81";
1652 } elseif ( $code < 0xc0ac ) {
1653 return "\xe3\x85\x82";
1654 } elseif ( $code < 0xc544 ) {
1655 return "\xe3\x85\x85";
1656 } elseif ( $code < 0xc790 ) {
1657 return "\xe3\x85\x87";
1658 } elseif ( $code < 0xcc28 ) {
1659 return "\xe3\x85\x88";
1660 } elseif ( $code < 0xce74 ) {
1661 return "\xe3\x85\x8a";
1662 } elseif ( $code < 0xd0c0 ) {
1663 return "\xe3\x85\x8b";
1664 } elseif ( $code < 0xd30c ) {
1665 return "\xe3\x85\x8c";
1666 } elseif ( $code < 0xd558 ) {
1667 return "\xe3\x85\x8d";
1668 } else {
1669 return "\xe3\x85\x8e";
1671 } else {
1672 return "";
1676 function initEncoding() {
1677 # Some languages may have an alternate char encoding option
1678 # (Esperanto X-coding, Japanese furigana conversion, etc)
1679 # If this language is used as the primary content language,
1680 # an override to the defaults can be set here on startup.
1683 function recodeForEdit( $s ) {
1684 # For some languages we'll want to explicitly specify
1685 # which characters make it into the edit box raw
1686 # or are converted in some way or another.
1687 # Note that if wgOutputEncoding is different from
1688 # wgInputEncoding, this text will be further converted
1689 # to wgOutputEncoding.
1690 global $wgEditEncoding;
1691 if( $wgEditEncoding == '' or
1692 $wgEditEncoding == 'UTF-8' ) {
1693 return $s;
1694 } else {
1695 return $this->iconv( 'UTF-8', $wgEditEncoding, $s );
1699 function recodeInput( $s ) {
1700 # Take the previous into account.
1701 global $wgEditEncoding;
1702 if($wgEditEncoding != "") {
1703 $enc = $wgEditEncoding;
1704 } else {
1705 $enc = 'UTF-8';
1707 if( $enc == 'UTF-8' ) {
1708 return $s;
1709 } else {
1710 return $this->iconv( $enc, 'UTF-8', $s );
1715 * For right-to-left language support
1717 * @return bool
1719 function isRTL() {
1720 $this->load();
1721 return $this->rtl;
1725 * A hidden direction mark (LRM or RLM), depending on the language direction
1727 * @return string
1729 function getDirMark() {
1730 return $this->isRTL() ? "\xE2\x80\x8F" : "\xE2\x80\x8E";
1734 * An arrow, depending on the language direction
1736 * @return string
1738 function getArrow() {
1739 return $this->isRTL() ? '←' : '→';
1743 * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
1745 * @return bool
1747 function linkPrefixExtension() {
1748 $this->load();
1749 return $this->linkPrefixExtension;
1752 function &getMagicWords() {
1753 $this->load();
1754 return $this->magicWords;
1757 # Fill a MagicWord object with data from here
1758 function getMagic( &$mw ) {
1759 if ( !$this->mMagicHookDone ) {
1760 $this->mMagicHookDone = true;
1761 wfRunHooks( 'LanguageGetMagic', array( &$this->mMagicExtensions, $this->getCode() ) );
1763 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
1764 $rawEntry = $this->mMagicExtensions[$mw->mId];
1765 } else {
1766 $magicWords =& $this->getMagicWords();
1767 if ( isset( $magicWords[$mw->mId] ) ) {
1768 $rawEntry = $magicWords[$mw->mId];
1769 } else {
1770 # Fall back to English if local list is incomplete
1771 $magicWords =& Language::getMagicWords();
1772 if ( !isset($magicWords[$mw->mId]) ) {
1773 throw new MWException("Magic word '{$mw->mId}' not found" );
1775 $rawEntry = $magicWords[$mw->mId];
1779 if( !is_array( $rawEntry ) ) {
1780 error_log( "\"$rawEntry\" is not a valid magic thingie for \"$mw->mId\"" );
1781 } else {
1782 $mw->mCaseSensitive = $rawEntry[0];
1783 $mw->mSynonyms = array_slice( $rawEntry, 1 );
1788 * Add magic words to the extension array
1790 function addMagicWordsByLang( $newWords ) {
1791 $code = $this->getCode();
1792 $fallbackChain = array();
1793 while ( $code && !in_array( $code, $fallbackChain ) ) {
1794 $fallbackChain[] = $code;
1795 $code = self::getFallbackFor( $code );
1797 if ( !in_array( 'en', $fallbackChain ) ) {
1798 $fallbackChain[] = 'en';
1800 $fallbackChain = array_reverse( $fallbackChain );
1801 foreach ( $fallbackChain as $code ) {
1802 if ( isset( $newWords[$code] ) ) {
1803 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
1809 * Get special page names, as an associative array
1810 * case folded alias => real name
1812 function getSpecialPageAliases() {
1813 $this->load();
1815 // Cache aliases because it may be slow to load them
1816 if ( !isset( $this->mExtendedSpecialPageAliases ) ) {
1818 // Initialise array
1819 $this->mExtendedSpecialPageAliases = $this->specialPageAliases;
1821 global $wgExtensionAliasesFiles;
1822 foreach ( $wgExtensionAliasesFiles as $file ) {
1824 // Fail fast
1825 if ( !file_exists($file) )
1826 throw new MWException( "Aliases file does not exist: $file" );
1828 $aliases = array();
1829 require($file);
1831 // Check the availability of aliases
1832 if ( !isset($aliases['en']) )
1833 throw new MWException( "Malformed aliases file: $file" );
1835 // Merge all aliases in fallback chain
1836 $code = $this->getCode();
1837 do {
1838 if ( !isset($aliases[$code]) ) continue;
1840 $aliases[$code] = $this->fixSpecialPageAliases( $aliases[$code] );
1841 /* Merge the aliases, THIS will break if there is special page name
1842 * which looks like a numerical key, thanks to PHP...
1843 * See the array_merge_recursive manual entry */
1844 $this->mExtendedSpecialPageAliases = array_merge_recursive(
1845 $this->mExtendedSpecialPageAliases, $aliases[$code] );
1847 } while ( $code = self::getFallbackFor( $code ) );
1850 wfRunHooks( 'LanguageGetSpecialPageAliases',
1851 array( &$this->mExtendedSpecialPageAliases, $this->getCode() ) );
1854 return $this->mExtendedSpecialPageAliases;
1858 * Function to fix special page aliases. Will convert the first letter to
1859 * upper case and spaces to underscores. Can be given a full aliases array,
1860 * in which case it will recursively fix all aliases.
1862 public function fixSpecialPageAliases( $mixed ) {
1863 // Work recursively until in string level
1864 if ( is_array($mixed) ) {
1865 $callback = array( $this, 'fixSpecialPageAliases' );
1866 return array_map( $callback, $mixed );
1868 return str_replace( ' ', '_', $this->ucfirst( $mixed ) );
1872 * Italic is unsuitable for some languages
1874 * @param $text String: the text to be emphasized.
1875 * @return string
1877 function emphasize( $text ) {
1878 return "<em>$text</em>";
1882 * Normally we output all numbers in plain en_US style, that is
1883 * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
1884 * point twohundredthirtyfive. However this is not sutable for all
1885 * languages, some such as Pakaran want ੨੯੩,੨੯੫.੨੩੫ and others such as
1886 * Icelandic just want to use commas instead of dots, and dots instead
1887 * of commas like "293.291,235".
1889 * An example of this function being called:
1890 * <code>
1891 * wfMsg( 'message', $wgLang->formatNum( $num ) )
1892 * </code>
1894 * See LanguageGu.php for the Gujarati implementation and
1895 * $separatorTransformTable on MessageIs.php for
1896 * the , => . and . => , implementation.
1898 * @todo check if it's viable to use localeconv() for the decimal
1899 * separator thing.
1900 * @param $number Mixed: the string to be formatted, should be an integer
1901 * or a floating point number.
1902 * @param $nocommafy Bool: set to true for special numbers like dates
1903 * @return string
1905 function formatNum( $number, $nocommafy = false ) {
1906 global $wgTranslateNumerals;
1907 if (!$nocommafy) {
1908 $number = $this->commafy($number);
1909 $s = $this->separatorTransformTable();
1910 if ($s) { $number = strtr($number, $s); }
1913 if ($wgTranslateNumerals) {
1914 $s = $this->digitTransformTable();
1915 if ($s) { $number = strtr($number, $s); }
1918 return $number;
1921 function parseFormattedNumber( $number ) {
1922 $s = $this->digitTransformTable();
1923 if ($s) { $number = strtr($number, array_flip($s)); }
1925 $s = $this->separatorTransformTable();
1926 if ($s) { $number = strtr($number, array_flip($s)); }
1928 $number = strtr( $number, array (',' => '') );
1929 return $number;
1933 * Adds commas to a given number
1935 * @param $_ mixed
1936 * @return string
1938 function commafy($_) {
1939 return strrev((string)preg_replace('/(\d{3})(?=\d)(?!\d*\.)/','$1,',strrev($_)));
1942 function digitTransformTable() {
1943 $this->load();
1944 return $this->digitTransformTable;
1947 function separatorTransformTable() {
1948 $this->load();
1949 return $this->separatorTransformTable;
1954 * Take a list of strings and build a locale-friendly comma-separated
1955 * list, using the local comma-separator message.
1956 * The last two strings are chained with an "and".
1958 * @param $l Array
1959 * @return string
1961 function listToText( $l ) {
1962 $s = '';
1963 $m = count( $l ) - 1;
1964 if( $m == 1 ) {
1965 return $l[0] . $this->getMessageFromDB( 'and' ) . $this->getMessageFromDB( 'word-separator' ) . $l[1];
1967 else {
1968 for ( $i = $m; $i >= 0; $i-- ) {
1969 if ( $i == $m ) {
1970 $s = $l[$i];
1971 } else if( $i == $m - 1 ) {
1972 $s = $l[$i] . $this->getMessageFromDB( 'and' ) . $this->getMessageFromDB( 'word-separator' ) . $s;
1973 } else {
1974 $s = $l[$i] . $this->getMessageFromDB( 'comma-separator' ) . $s;
1977 return $s;
1982 * Take a list of strings and build a locale-friendly comma-separated
1983 * list, using the local comma-separator message.
1984 * @param $list array of strings to put in a comma list
1985 * @return string
1987 function commaList( $list ) {
1988 return implode(
1989 $list,
1990 wfMsgExt( 'comma-separator', array( 'escapenoentities', 'language' => $this ) ) );
1994 * Same as commaList, but separate it with the pipe instead.
1995 * @param $list array of strings to put in a pipe list
1996 * @return string
1998 function pipeList( $list ) {
1999 return implode(
2000 $list,
2001 wfMsgExt( 'pipe-separator', array( 'escapenoentities', 'language' => $this ) ) );
2005 * Truncate a string to a specified length in bytes, appending an optional
2006 * string (e.g. for ellipses)
2008 * The database offers limited byte lengths for some columns in the database;
2009 * multi-byte character sets mean we need to ensure that only whole characters
2010 * are included, otherwise broken characters can be passed to the user
2012 * If $length is negative, the string will be truncated from the beginning
2014 * @param $string String to truncate
2015 * @param $length Int: maximum length (excluding ellipses)
2016 * @param $ellipsis String to append to the truncated text
2017 * @return string
2019 function truncate( $string, $length, $ellipsis = '...' ) {
2020 # Use the localized ellipsis character
2021 if( $ellipsis == '...' ) {
2022 $ellipsis = wfMsgExt( 'ellipsis', array( 'escapenoentities', 'language' => $this ) );
2025 if( $length == 0 ) {
2026 return $ellipsis;
2028 if ( strlen( $string ) <= abs( $length ) ) {
2029 return $string;
2031 if( $length > 0 ) {
2032 $string = substr( $string, 0, $length );
2033 $char = ord( $string[strlen( $string ) - 1] );
2034 $m = array();
2035 if ($char >= 0xc0) {
2036 # We got the first byte only of a multibyte char; remove it.
2037 $string = substr( $string, 0, -1 );
2038 } elseif( $char >= 0x80 &&
2039 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
2040 '[\xf0-\xf7][\x80-\xbf]{1,2})$/', $string, $m ) ) {
2041 # We chopped in the middle of a character; remove it
2042 $string = $m[1];
2044 return $string . $ellipsis;
2045 } else {
2046 $string = substr( $string, $length );
2047 $char = ord( $string[0] );
2048 if( $char >= 0x80 && $char < 0xc0 ) {
2049 # We chopped in the middle of a character; remove the whole thing
2050 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
2052 return $ellipsis . $string;
2057 * Grammatical transformations, needed for inflected languages
2058 * Invoked by putting {{grammar:case|word}} in a message
2060 * @param $word string
2061 * @param $case string
2062 * @return string
2064 function convertGrammar( $word, $case ) {
2065 global $wgGrammarForms;
2066 if ( isset($wgGrammarForms[$this->getCode()][$case][$word]) ) {
2067 return $wgGrammarForms[$this->getCode()][$case][$word];
2069 return $word;
2073 * Provides an alternative text depending on specified gender.
2074 * Usage {{gender:username|masculine|feminine|neutral}}.
2075 * username is optional, in which case the gender of current user is used,
2076 * but only in (some) interface messages; otherwise default gender is used.
2077 * If second or third parameter are not specified, masculine is used.
2078 * These details may be overriden per language.
2080 function gender( $gender, $forms ) {
2081 if ( !count($forms) ) { return ''; }
2082 $forms = $this->preConvertPlural( $forms, 2 );
2083 if ( $gender === 'male' ) return $forms[0];
2084 if ( $gender === 'female' ) return $forms[1];
2085 return isset($forms[2]) ? $forms[2] : $forms[0];
2089 * Plural form transformations, needed for some languages.
2090 * For example, there are 3 form of plural in Russian and Polish,
2091 * depending on "count mod 10". See [[w:Plural]]
2092 * For English it is pretty simple.
2094 * Invoked by putting {{plural:count|wordform1|wordform2}}
2095 * or {{plural:count|wordform1|wordform2|wordform3}}
2097 * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
2099 * @param $count Integer: non-localized number
2100 * @param $forms Array: different plural forms
2101 * @return string Correct form of plural for $count in this language
2103 function convertPlural( $count, $forms ) {
2104 if ( !count($forms) ) { return ''; }
2105 $forms = $this->preConvertPlural( $forms, 2 );
2107 return ( $count == 1 ) ? $forms[0] : $forms[1];
2111 * Checks that convertPlural was given an array and pads it to requested
2112 * amound of forms by copying the last one.
2114 * @param $count Integer: How many forms should there be at least
2115 * @param $forms Array of forms given to convertPlural
2116 * @return array Padded array of forms or an exception if not an array
2118 protected function preConvertPlural( /* Array */ $forms, $count ) {
2119 while ( count($forms) < $count ) {
2120 $forms[] = $forms[count($forms)-1];
2122 return $forms;
2126 * For translaing of expiry times
2127 * @param $str String: the validated block time in English
2128 * @return Somehow translated block time
2129 * @see LanguageFi.php for example implementation
2131 function translateBlockExpiry( $str ) {
2133 $scBlockExpiryOptions = $this->getMessageFromDB( 'ipboptions' );
2135 if ( $scBlockExpiryOptions == '-') {
2136 return $str;
2139 foreach (explode(',', $scBlockExpiryOptions) as $option) {
2140 if ( strpos($option, ":") === false )
2141 continue;
2142 list($show, $value) = explode(":", $option);
2143 if ( strcmp ( $str, $value) == 0 ) {
2144 return htmlspecialchars( trim( $show ) );
2148 return $str;
2152 * languages like Chinese need to be segmented in order for the diff
2153 * to be of any use
2155 * @param $text String
2156 * @return String
2158 function segmentForDiff( $text ) {
2159 return $text;
2163 * and unsegment to show the result
2165 * @param $text String
2166 * @return String
2168 function unsegmentForDiff( $text ) {
2169 return $text;
2172 # convert text to different variants of a language.
2173 function convert( $text, $isTitle = false) {
2174 return $this->mConverter->convert($text, $isTitle);
2177 # Convert text from within Parser
2178 function parserConvert( $text, &$parser ) {
2179 return $this->mConverter->parserConvert( $text, $parser );
2182 # Check if this is a language with variants
2183 function hasVariants(){
2184 return sizeof($this->getVariants())>1;
2187 # Put custom tags (e.g. -{ }-) around math to prevent conversion
2188 function armourMath($text){
2189 return $this->mConverter->armourMath($text);
2194 * Perform output conversion on a string, and encode for safe HTML output.
2195 * @param $text String
2196 * @param $isTitle Bool -- wtf?
2197 * @return string
2198 * @todo this should get integrated somewhere sane
2200 function convertHtml( $text, $isTitle = false ) {
2201 return htmlspecialchars( $this->convert( $text, $isTitle ) );
2204 function convertCategoryKey( $key ) {
2205 return $this->mConverter->convertCategoryKey( $key );
2209 * get the list of variants supported by this langauge
2210 * see sample implementation in LanguageZh.php
2212 * @return array an array of language codes
2214 function getVariants() {
2215 return $this->mConverter->getVariants();
2219 function getPreferredVariant( $fromUser = true ) {
2220 return $this->mConverter->getPreferredVariant( $fromUser );
2224 * if a language supports multiple variants, it is
2225 * possible that non-existing link in one variant
2226 * actually exists in another variant. this function
2227 * tries to find it. See e.g. LanguageZh.php
2229 * @param $link String: the name of the link
2230 * @param $nt Mixed: the title object of the link
2231 * @param boolean $ignoreOtherCond: to disable other conditions when
2232 * we need to transclude a template or update a category's link
2233 * @return null the input parameters may be modified upon return
2235 function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
2236 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
2240 * If a language supports multiple variants, converts text
2241 * into an array of all possible variants of the text:
2242 * 'variant' => text in that variant
2244 function convertLinkToAllVariants($text){
2245 return $this->mConverter->convertLinkToAllVariants($text);
2250 * returns language specific options used by User::getPageRenderHash()
2251 * for example, the preferred language variant
2253 * @return string
2255 function getExtraHashOptions() {
2256 return $this->mConverter->getExtraHashOptions();
2260 * for languages that support multiple variants, the title of an
2261 * article may be displayed differently in different variants. this
2262 * function returns the apporiate title defined in the body of the article.
2264 * @return string
2266 function getParsedTitle() {
2267 return $this->mConverter->getParsedTitle();
2271 * Enclose a string with the "no conversion" tag. This is used by
2272 * various functions in the Parser
2274 * @param $text String: text to be tagged for no conversion
2275 * @param $noParse
2276 * @return string the tagged text
2278 function markNoConversion( $text, $noParse=false ) {
2279 return $this->mConverter->markNoConversion( $text, $noParse );
2283 * A regular expression to match legal word-trailing characters
2284 * which should be merged onto a link of the form [[foo]]bar.
2286 * @return string
2288 function linkTrail() {
2289 $this->load();
2290 return $this->linkTrail;
2293 function getLangObj() {
2294 return $this;
2298 * Get the RFC 3066 code for this language object
2300 function getCode() {
2301 return $this->mCode;
2304 function setCode( $code ) {
2305 $this->mCode = $code;
2308 static function getFileName( $prefix = 'Language', $code, $suffix = '.php' ) {
2309 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
2312 static function getMessagesFileName( $code ) {
2313 global $IP;
2314 return self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
2317 static function getClassFileName( $code ) {
2318 global $IP;
2319 return self::getFileName( "$IP/languages/classes/Language", $code, '.php' );
2322 static function getLocalisationArray( $code, $disableCache = false ) {
2323 self::loadLocalisation( $code, $disableCache );
2324 return self::$mLocalisationCache[$code];
2328 * Load localisation data for a given code into the static cache
2330 * @return array Dependencies, map of filenames to mtimes
2332 static function loadLocalisation( $code, $disableCache = false ) {
2333 static $recursionGuard = array();
2334 global $wgMemc, $wgEnableSerializedMessages, $wgCheckSerialized;
2336 if ( !$code ) {
2337 throw new MWException( "Invalid language code requested" );
2340 if ( !$disableCache ) {
2341 # Try the per-process cache
2342 if ( isset( self::$mLocalisationCache[$code] ) ) {
2343 return self::$mLocalisationCache[$code]['deps'];
2346 wfProfileIn( __METHOD__ );
2348 # Try the serialized directory
2349 if( $wgEnableSerializedMessages ) {
2350 $cache = wfGetPrecompiledData( self::getFileName( "Messages", $code, '.ser' ) );
2351 if ( $cache ) {
2352 if ( $wgCheckSerialized && self::isLocalisationOutOfDate( $cache ) ) {
2353 $cache = false;
2354 wfDebug( "Language::loadLocalisation(): precompiled data file for $code is out of date\n" );
2355 } else {
2356 self::$mLocalisationCache[$code] = $cache;
2357 wfDebug( "Language::loadLocalisation(): got localisation for $code from precompiled data file\n" );
2358 wfProfileOut( __METHOD__ );
2359 return self::$mLocalisationCache[$code]['deps'];
2364 # Try the global cache
2365 $memcKey = wfMemcKey('localisation', $code );
2366 $fbMemcKey = wfMemcKey('fallback', $cache['fallback'] );
2367 $cache = $wgMemc->get( $memcKey );
2368 if ( $cache ) {
2369 if ( self::isLocalisationOutOfDate( $cache ) ) {
2370 $wgMemc->delete( $memcKey );
2371 $wgMemc->delete( $fbMemcKey );
2372 $cache = false;
2373 wfDebug( "Language::loadLocalisation(): localisation cache for $code had expired\n" );
2374 } else {
2375 self::$mLocalisationCache[$code] = $cache;
2376 wfDebug( "Language::loadLocalisation(): got localisation for $code from cache\n" );
2377 wfProfileOut( __METHOD__ );
2378 return $cache['deps'];
2381 } else {
2382 wfProfileIn( __METHOD__ );
2385 # Default fallback, may be overridden when the messages file is included
2386 if ( $code != 'en' ) {
2387 $fallback = 'en';
2388 } else {
2389 $fallback = false;
2392 # Load the primary localisation from the source file
2393 $filename = self::getMessagesFileName( $code );
2394 if ( !file_exists( $filename ) ) {
2395 wfDebug( "Language::loadLocalisation(): no localisation file for $code, using implicit fallback to en\n" );
2396 $cache = compact( self::$mLocalisationKeys ); // Set correct fallback
2397 $deps = array();
2398 } else {
2399 $deps = array( $filename => filemtime( $filename ) );
2400 require( $filename );
2401 $cache = compact( self::$mLocalisationKeys );
2402 wfDebug( "Language::loadLocalisation(): got localisation for $code from source\n" );
2405 # Load magic word source file
2406 global $IP;
2407 $filename = "$IP/includes/MagicWord.php";
2408 $newDeps = array( $filename => filemtime( $filename ) );
2409 $deps = array_merge( $deps, $newDeps );
2411 if ( !empty( $fallback ) ) {
2412 # Load the fallback localisation, with a circular reference guard
2413 if ( isset( $recursionGuard[$code] ) ) {
2414 throw new MWException( "Error: Circular fallback reference in language code $code" );
2416 $recursionGuard[$code] = true;
2417 $newDeps = self::loadLocalisation( $fallback, $disableCache );
2418 unset( $recursionGuard[$code] );
2420 $secondary = self::$mLocalisationCache[$fallback];
2421 $deps = array_merge( $deps, $newDeps );
2423 # Merge the fallback localisation with the current localisation
2424 foreach ( self::$mLocalisationKeys as $key ) {
2425 if ( isset( $cache[$key] ) ) {
2426 if ( isset( $secondary[$key] ) ) {
2427 if ( in_array( $key, self::$mMergeableMapKeys ) ) {
2428 $cache[$key] = $cache[$key] + $secondary[$key];
2429 } elseif ( in_array( $key, self::$mMergeableListKeys ) ) {
2430 $cache[$key] = array_merge( $secondary[$key], $cache[$key] );
2431 } elseif ( in_array( $key, self::$mMergeableAliasListKeys ) ) {
2432 $cache[$key] = array_merge_recursive( $cache[$key], $secondary[$key] );
2435 } else {
2436 $cache[$key] = $secondary[$key];
2440 # Merge bookstore lists if requested
2441 if ( !empty( $cache['bookstoreList']['inherit'] ) ) {
2442 $cache['bookstoreList'] = array_merge( $cache['bookstoreList'], $secondary['bookstoreList'] );
2444 if ( isset( $cache['bookstoreList']['inherit'] ) ) {
2445 unset( $cache['bookstoreList']['inherit'] );
2449 # Add dependencies to the cache entry
2450 $cache['deps'] = $deps;
2452 # Replace spaces with underscores in namespace names
2453 $cache['namespaceNames'] = str_replace( ' ', '_', $cache['namespaceNames'] );
2455 # And do the same for specialpage aliases. $page is an array.
2456 foreach ( $cache['specialPageAliases'] as &$page ) {
2457 $page = str_replace( ' ', '_', $page );
2459 # Decouple the reference to prevent accidental damage
2460 unset($page);
2462 # Save to both caches
2463 self::$mLocalisationCache[$code] = $cache;
2464 if ( !$disableCache ) {
2465 $wgMemc->set( $memcKey, $cache );
2466 $wgMemc->set( $fbMemcKey, (string) $cache['fallback'] );
2469 wfProfileOut( __METHOD__ );
2470 return $deps;
2474 * Test if a given localisation cache is out of date with respect to the
2475 * source Messages files. This is done automatically for the global cache
2476 * in $wgMemc, but is only done on certain occasions for the serialized
2477 * data file.
2479 * @param $cache mixed Either a language code or a cache array
2481 static function isLocalisationOutOfDate( $cache ) {
2482 if ( !is_array( $cache ) ) {
2483 self::loadLocalisation( $cache );
2484 $cache = self::$mLocalisationCache[$cache];
2486 // At least one language file and the MagicWord file needed
2487 if( count($cache['deps']) < 2 ) {
2488 return true;
2490 $expired = false;
2491 foreach ( $cache['deps'] as $file => $mtime ) {
2492 if ( !file_exists( $file ) || filemtime( $file ) > $mtime ) {
2493 $expired = true;
2494 break;
2497 return $expired;
2501 * Get the fallback for a given language
2503 static function getFallbackFor( $code ) {
2504 // Shortcut
2505 if ( $code === 'en' ) return false;
2507 // Local cache
2508 static $cache = array();
2509 // Quick return
2510 if ( isset($cache[$code]) ) return $cache[$code];
2512 // Try memcache
2513 global $wgMemc;
2514 $memcKey = wfMemcKey( 'fallback', $code );
2515 $fbcode = $wgMemc->get( $memcKey );
2517 if ( is_string($fbcode) ) {
2518 // False is stored as a string to detect failures in memcache properly
2519 if ( $fbcode === '' ) $fbcode = false;
2521 // Update local cache and return
2522 $cache[$code] = $fbcode;
2523 return $fbcode;
2526 // Nothing in caches, load and and update both caches
2527 self::loadLocalisation( $code );
2528 $fbcode = self::$mLocalisationCache[$code]['fallback'];
2530 $cache[$code] = $fbcode;
2531 $wgMemc->set( $memcKey, (string) $fbcode );
2533 return $fbcode;
2536 /**
2537 * Get all messages for a given language
2539 static function getMessagesFor( $code ) {
2540 self::loadLocalisation( $code );
2541 return self::$mLocalisationCache[$code]['messages'];
2544 /**
2545 * Get a message for a given language
2547 static function getMessageFor( $key, $code ) {
2548 self::loadLocalisation( $code );
2549 return isset( self::$mLocalisationCache[$code]['messages'][$key] ) ? self::$mLocalisationCache[$code]['messages'][$key] : null;
2553 * Load localisation data for this object
2555 function load() {
2556 if ( !$this->mLoaded ) {
2557 self::loadLocalisation( $this->getCode() );
2558 $cache =& self::$mLocalisationCache[$this->getCode()];
2559 foreach ( self::$mLocalisationKeys as $key ) {
2560 $this->$key = $cache[$key];
2562 $this->mLoaded = true;
2564 $this->fixUpSettings();
2569 * Do any necessary post-cache-load settings adjustment
2571 function fixUpSettings() {
2572 global $wgExtraNamespaces, $wgMetaNamespace, $wgMetaNamespaceTalk,
2573 $wgNamespaceAliases, $wgAmericanDates;
2574 wfProfileIn( __METHOD__ );
2575 if ( $wgExtraNamespaces ) {
2576 $this->namespaceNames = $wgExtraNamespaces + $this->namespaceNames;
2579 $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
2580 if ( $wgMetaNamespaceTalk ) {
2581 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
2582 } else {
2583 $talk = $this->namespaceNames[NS_PROJECT_TALK];
2584 $this->namespaceNames[NS_PROJECT_TALK] =
2585 $this->fixVariableInNamespace( $talk );
2588 # The above mixing may leave namespaces out of canonical order.
2589 # Re-order by namespace ID number...
2590 ksort( $this->namespaceNames );
2592 # Put namespace names and aliases into a hashtable.
2593 # If this is too slow, then we should arrange it so that it is done
2594 # before caching. The catch is that at pre-cache time, the above
2595 # class-specific fixup hasn't been done.
2596 $this->mNamespaceIds = array();
2597 foreach ( $this->namespaceNames as $index => $name ) {
2598 $this->mNamespaceIds[$this->lc($name)] = $index;
2600 if ( $this->namespaceAliases ) {
2601 foreach ( $this->namespaceAliases as $name => $index ) {
2602 if ( $index === NS_PROJECT_TALK ) {
2603 unset( $this->namespaceAliases[$name] );
2604 $name = $this->fixVariableInNamespace( $name );
2605 $this->namespaceAliases[$name] = $index;
2607 $this->mNamespaceIds[$this->lc($name)] = $index;
2610 if ( $wgNamespaceAliases ) {
2611 foreach ( $wgNamespaceAliases as $name => $index ) {
2612 $this->mNamespaceIds[$this->lc($name)] = $index;
2616 if ( $this->defaultDateFormat == 'dmy or mdy' ) {
2617 $this->defaultDateFormat = $wgAmericanDates ? 'mdy' : 'dmy';
2619 wfProfileOut( __METHOD__ );
2622 function fixVariableInNamespace( $talk ) {
2623 if ( strpos( $talk, '$1' ) === false ) return $talk;
2625 global $wgMetaNamespace;
2626 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
2628 # Allow grammar transformations
2629 # Allowing full message-style parsing would make simple requests
2630 # such as action=raw much more expensive than they need to be.
2631 # This will hopefully cover most cases.
2632 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
2633 array( &$this, 'replaceGrammarInNamespace' ), $talk );
2634 return str_replace( ' ', '_', $talk );
2637 function replaceGrammarInNamespace( $m ) {
2638 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
2641 static function getCaseMaps() {
2642 static $wikiUpperChars, $wikiLowerChars;
2643 if ( isset( $wikiUpperChars ) ) {
2644 return array( $wikiUpperChars, $wikiLowerChars );
2647 wfProfileIn( __METHOD__ );
2648 $arr = wfGetPrecompiledData( 'Utf8Case.ser' );
2649 if ( $arr === false ) {
2650 throw new MWException(
2651 "Utf8Case.ser is missing, please run \"make\" in the serialized directory\n" );
2653 extract( $arr );
2654 wfProfileOut( __METHOD__ );
2655 return array( $wikiUpperChars, $wikiLowerChars );
2658 function formatTimePeriod( $seconds ) {
2659 if ( $seconds < 10 ) {
2660 return $this->formatNum( sprintf( "%.1f", $seconds ) ) . wfMsg( 'seconds-abbrev' );
2661 } elseif ( $seconds < 60 ) {
2662 return $this->formatNum( round( $seconds ) ) . wfMsg( 'seconds-abbrev' );
2663 } elseif ( $seconds < 3600 ) {
2664 return $this->formatNum( floor( $seconds / 60 ) ) . wfMsg( 'minutes-abbrev' ) .
2665 $this->formatNum( round( fmod( $seconds, 60 ) ) ) . wfMsg( 'seconds-abbrev' );
2666 } else {
2667 $hours = floor( $seconds / 3600 );
2668 $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
2669 $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
2670 return $this->formatNum( $hours ) . wfMsg( 'hours-abbrev' ) .
2671 $this->formatNum( $minutes ) . wfMsg( 'minutes-abbrev' ) .
2672 $this->formatNum( $secondsPart ) . wfMsg( 'seconds-abbrev' );
2676 function formatBitrate( $bps ) {
2677 $units = array( 'bps', 'kbps', 'Mbps', 'Gbps' );
2678 if ( $bps <= 0 ) {
2679 return $this->formatNum( $bps ) . $units[0];
2681 $unitIndex = floor( log10( $bps ) / 3 );
2682 $mantissa = $bps / pow( 1000, $unitIndex );
2683 if ( $mantissa < 10 ) {
2684 $mantissa = round( $mantissa, 1 );
2685 } else {
2686 $mantissa = round( $mantissa );
2688 return $this->formatNum( $mantissa ) . $units[$unitIndex];
2692 * Format a size in bytes for output, using an appropriate
2693 * unit (B, KB, MB or GB) according to the magnitude in question
2695 * @param $size Size to format
2696 * @return string Plain text (not HTML)
2698 function formatSize( $size ) {
2699 // For small sizes no decimal places necessary
2700 $round = 0;
2701 if( $size > 1024 ) {
2702 $size = $size / 1024;
2703 if( $size > 1024 ) {
2704 $size = $size / 1024;
2705 // For MB and bigger two decimal places are smarter
2706 $round = 2;
2707 if( $size > 1024 ) {
2708 $size = $size / 1024;
2709 $msg = 'size-gigabytes';
2710 } else {
2711 $msg = 'size-megabytes';
2713 } else {
2714 $msg = 'size-kilobytes';
2716 } else {
2717 $msg = 'size-bytes';
2719 $size = round( $size, $round );
2720 $text = $this->getMessageFromDB( $msg );
2721 return str_replace( '$1', $this->formatNum( $size ), $text );