languages/Language.php

   1 <?php
   2 /**
   3  * Internationalisation code.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Language
  22  */
  23
  24 /**
  25  * @defgroup Language Language
  26  */
  27
  28 if ( !defined( 'MEDIAWIKI' ) ) {
  29         echo "This file is part of MediaWiki, it is not a valid entry point.\n";
  30         exit( 1 );
  31 }
  32
  33 if ( function_exists( 'mb_strtoupper' ) ) {
  34         mb_internal_encoding( 'UTF-8' );
  35 }
  36
  37 use CLDRPluralRuleParser\Evaluator;
  38
  39 /**
  40  * Internationalisation code
  41  * @ingroup Language
  42  */
  43 class Language {
  44         /**
  45          * @var LanguageConverter
  46          */
  47         public $mConverter;
  48
  49         public $mVariants, $mCode, $mLoaded = false;
  50         public $mMagicExtensions = array(), $mMagicHookDone = false;
  51         private $mHtmlCode = null, $mParentLanguage = false;
  52
  53         public $dateFormatStrings = array();
  54         public $mExtendedSpecialPageAliases;
  55
  56         protected $namespaceNames, $mNamespaceIds, $namespaceAliases;
  57
  58         /**
  59          * ReplacementArray object caches
  60          */
  61         public $transformData = array();
  62
  63         /**
  64          * @var LocalisationCache
  65          */
  66         static public $dataCache;
  67
  68         static public $mLangObjCache = array();
  69
  70         static public $mWeekdayMsgs = array(
  71                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  72                 'friday', 'saturday'
  73         );
  74
  75         static public $mWeekdayAbbrevMsgs = array(
  76                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
  77         );
  78
  79         static public $mMonthMsgs = array(
  80                 'january', 'february', 'march', 'april', 'may_long', 'june',
  81                 'july', 'august', 'september', 'october', 'november',
  82                 'december'
  83         );
  84         static public $mMonthGenMsgs = array(
  85                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
  86                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
  87                 'december-gen'
  88         );
  89         static public $mMonthAbbrevMsgs = array(
  90                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
  91                 'sep', 'oct', 'nov', 'dec'
  92         );
  93
  94         static public $mIranianCalendarMonthMsgs = array(
  95                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
  96                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
  97                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
  98                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
  99         );
 100
 101         static public $mHebrewCalendarMonthMsgs = array(
 102                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 103                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 104                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 105                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 106                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 107         );
 108
 109         static public $mHebrewCalendarMonthGenMsgs = array(
 110                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 111                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 112                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 113                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 114                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 115         );
 116
 117         static public $mHijriCalendarMonthMsgs = array(
 118                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 119                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 120                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 121                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 122         );
 123
 124         /**
 125          * @since 1.20
 126          * @var array
 127          */
 128         static public $durationIntervals = array(
 129                 'millennia' => 31556952000,
 130                 'centuries' => 3155695200,
 131                 'decades' => 315569520,
 132                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 133                 'weeks' => 604800,
 134                 'days' => 86400,
 135                 'hours' => 3600,
 136                 'minutes' => 60,
 137                 'seconds' => 1,
 138         );
 139
 140         /**
 141          * Cache for language fallbacks.
 142          * @see Language::getFallbacksIncludingSiteLanguage
 143          * @since 1.21
 144          * @var array
 145          */
 146         static private $fallbackLanguageCache = array();
 147
 148         /**
 149          * Cache for language names
 150          * @var MapCacheLRU|null
 151          */
 152         static private $languageNameCache;
 153
 154         /**
 155          * Unicode directional formatting characters, for embedBidi()
 156          */
 157         static private $lre = "\xE2\x80\xAA"; // U+202A LEFT-TO-RIGHT EMBEDDING
 158         static private $rle = "\xE2\x80\xAB"; // U+202B RIGHT-TO-LEFT EMBEDDING
 159         static private $pdf = "\xE2\x80\xAC"; // U+202C POP DIRECTIONAL FORMATTING
 160
 161         /**
 162          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 163          * - in group 1 if it is LTR
 164          * - in group 2 if it is RTL
 165          * Does not match if there is no strong directionality codepoint.
 166          *
 167          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 168          *
 169          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 170          * https://git.wikimedia.org/summary/unicodejs.git .
 171          */
 172         // @codingStandardsIgnoreStart
 173         // @codeCoverageIgnoreStart
 174         static private $strongDirRegex = '/(?:([\x{41}-\x{5a}\x{61}-\x{7a}\x{aa}\x{b5}\x{ba}\x{c0}-\x{d6}\x{d8}-\x{f6}\x{f8}-\x{2b8}\x{2bb}-\x{2c1}\x{2d0}\x{2d1}\x{2e0}-\x{2e4}\x{2ee}\x{370}-\x{373}\x{376}\x{377}\x{37a}-\x{37d}\x{37f}\x{386}\x{388}-\x{38a}\x{38c}\x{38e}-\x{3a1}\x{3a3}-\x{3f5}\x{3f7}-\x{482}\x{48a}-\x{52f}\x{531}-\x{556}\x{559}-\x{55f}\x{561}-\x{587}\x{589}\x{903}-\x{939}\x{93b}\x{93d}-\x{940}\x{949}-\x{94c}\x{94e}-\x{950}\x{958}-\x{961}\x{964}-\x{980}\x{982}\x{983}\x{985}-\x{98c}\x{98f}\x{990}\x{993}-\x{9a8}\x{9aa}-\x{9b0}\x{9b2}\x{9b6}-\x{9b9}\x{9bd}-\x{9c0}\x{9c7}\x{9c8}\x{9cb}\x{9cc}\x{9ce}\x{9d7}\x{9dc}\x{9dd}\x{9df}-\x{9e1}\x{9e6}-\x{9f1}\x{9f4}-\x{9fa}\x{a03}\x{a05}-\x{a0a}\x{a0f}\x{a10}\x{a13}-\x{a28}\x{a2a}-\x{a30}\x{a32}\x{a33}\x{a35}\x{a36}\x{a38}\x{a39}\x{a3e}-\x{a40}\x{a59}-\x{a5c}\x{a5e}\x{a66}-\x{a6f}\x{a72}-\x{a74}\x{a83}\x{a85}-\x{a8d}\x{a8f}-\x{a91}\x{a93}-\x{aa8}\x{aaa}-\x{ab0}\x{ab2}\x{ab3}\x{ab5}-\x{ab9}\x{abd}-\x{ac0}\x{ac9}\x{acb}\x{acc}\x{ad0}\x{ae0}\x{ae1}\x{ae6}-\x{af0}\x{af9}\x{b02}\x{b03}\x{b05}-\x{b0c}\x{b0f}\x{b10}\x{b13}-\x{b28}\x{b2a}-\x{b30}\x{b32}\x{b33}\x{b35}-\x{b39}\x{b3d}\x{b3e}\x{b40}\x{b47}\x{b48}\x{b4b}\x{b4c}\x{b57}\x{b5c}\x{b5d}\x{b5f}-\x{b61}\x{b66}-\x{b77}\x{b83}\x{b85}-\x{b8a}\x{b8e}-\x{b90}\x{b92}-\x{b95}\x{b99}\x{b9a}\x{b9c}\x{b9e}\x{b9f}\x{ba3}\x{ba4}\x{ba8}-\x{baa}\x{bae}-\x{bb9}\x{bbe}\x{bbf}\x{bc1}\x{bc2}\x{bc6}-\x{bc8}\x{bca}-\x{bcc}\x{bd0}\x{bd7}\x{be6}-\x{bf2}\x{c01}-\x{c03}\x{c05}-\x{c0c}\x{c0e}-\x{c10}\x{c12}-\x{c28}\x{c2a}-\x{c39}\x{c3d}\x{c41}-\x{c44}\x{c58}-\x{c5a}\x{c60}\x{c61}\x{c66}-\x{c6f}\x{c7f}\x{c82}\x{c83}\x{c85}-\x{c8c}\x{c8e}-\x{c90}\x{c92}-\x{ca8}\x{caa}-\x{cb3}\x{cb5}-\x{cb9}\x{cbd}-\x{cc4}\x{cc6}-\x{cc8}\x{cca}\x{ccb}\x{cd5}\x{cd6}\x{cde}\x{ce0}\x{ce1}\x{ce6}-\x{cef}\x{cf1}\x{cf2}\x{d02}\x{d03}\x{d05}-\x{d0c}\x{d0e}-\x{d10}\x{d12}-\x{d3a}\x{d3d}-\x{d40}\x{d46}-\x{d48}\x{d4a}-\x{d4c}\x{d4e}\x{d57}\x{d5f}-\x{d61}\x{d66}-\x{d75}\x{d79}-\x{d7f}\x{d82}\x{d83}\x{d85}-\x{d96}\x{d9a}-\x{db1}\x{db3}-\x{dbb}\x{dbd}\x{dc0}-\x{dc6}\x{dcf}-\x{dd1}\x{dd8}-\x{ddf}\x{de6}-\x{def}\x{df2}-\x{df4}\x{e01}-\x{e30}\x{e32}\x{e33}\x{e40}-\x{e46}\x{e4f}-\x{e5b}\x{e81}\x{e82}\x{e84}\x{e87}\x{e88}\x{e8a}\x{e8d}\x{e94}-\x{e97}\x{e99}-\x{e9f}\x{ea1}-\x{ea3}\x{ea5}\x{ea7}\x{eaa}\x{eab}\x{ead}-\x{eb0}\x{eb2}\x{eb3}\x{ebd}\x{ec0}-\x{ec4}\x{ec6}\x{ed0}-\x{ed9}\x{edc}-\x{edf}\x{f00}-\x{f17}\x{f1a}-\x{f34}\x{f36}\x{f38}\x{f3e}-\x{f47}\x{f49}-\x{f6c}\x{f7f}\x{f85}\x{f88}-\x{f8c}\x{fbe}-\x{fc5}\x{fc7}-\x{fcc}\x{fce}-\x{fda}\x{1000}-\x{102c}\x{1031}\x{1038}\x{103b}\x{103c}\x{103f}-\x{1057}\x{105a}-\x{105d}\x{1061}-\x{1070}\x{1075}-\x{1081}\x{1083}\x{1084}\x{1087}-\x{108c}\x{108e}-\x{109c}\x{109e}-\x{10c5}\x{10c7}\x{10cd}\x{10d0}-\x{1248}\x{124a}-\x{124d}\x{1250}-\x{1256}\x{1258}\x{125a}-\x{125d}\x{1260}-\x{1288}\x{128a}-\x{128d}\x{1290}-\x{12b0}\x{12b2}-\x{12b5}\x{12b8}-\x{12be}\x{12c0}\x{12c2}-\x{12c5}\x{12c8}-\x{12d6}\x{12d8}-\x{1310}\x{1312}-\x{1315}\x{1318}-\x{135a}\x{1360}-\x{137c}\x{1380}-\x{138f}\x{13a0}-\x{13f5}\x{13f8}-\x{13fd}\x{1401}-\x{167f}\x{1681}-\x{169a}\x{16a0}-\x{16f8}\x{1700}-\x{170c}\x{170e}-\x{1711}\x{1720}-\x{1731}\x{1735}\x{1736}\x{1740}-\x{1751}\x{1760}-\x{176c}\x{176e}-\x{1770}\x{1780}-\x{17b3}\x{17b6}\x{17be}-\x{17c5}\x{17c7}\x{17c8}\x{17d4}-\x{17da}\x{17dc}\x{17e0}-\x{17e9}\x{1810}-\x{1819}\x{1820}-\x{1877}\x{1880}-\x{18a8}\x{18aa}\x{18b0}-\x{18f5}\x{1900}-\x{191e}\x{1923}-\x{1926}\x{1929}-\x{192b}\x{1930}\x{1931}\x{1933}-\x{1938}\x{1946}-\x{196d}\x{1970}-\x{1974}\x{1980}-\x{19ab}\x{19b0}-\x{19c9}\x{19d0}-\x{19da}\x{1a00}-\x{1a16}\x{1a19}\x{1a1a}\x{1a1e}-\x{1a55}\x{1a57}\x{1a61}\x{1a63}\x{1a64}\x{1a6d}-\x{1a72}\x{1a80}-\x{1a89}\x{1a90}-\x{1a99}\x{1aa0}-\x{1aad}\x{1b04}-\x{1b33}\x{1b35}\x{1b3b}\x{1b3d}-\x{1b41}\x{1b43}-\x{1b4b}\x{1b50}-\x{1b6a}\x{1b74}-\x{1b7c}\x{1b82}-\x{1ba1}\x{1ba6}\x{1ba7}\x{1baa}\x{1bae}-\x{1be5}\x{1be7}\x{1bea}-\x{1bec}\x{1bee}\x{1bf2}\x{1bf3}\x{1bfc}-\x{1c2b}\x{1c34}\x{1c35}\x{1c3b}-\x{1c49}\x{1c4d}-\x{1c7f}\x{1cc0}-\x{1cc7}\x{1cd3}\x{1ce1}\x{1ce9}-\x{1cec}\x{1cee}-\x{1cf3}\x{1cf5}\x{1cf6}\x{1d00}-\x{1dbf}\x{1e00}-\x{1f15}\x{1f18}-\x{1f1d}\x{1f20}-\x{1f45}\x{1f48}-\x{1f4d}\x{1f50}-\x{1f57}\x{1f59}\x{1f5b}\x{1f5d}\x{1f5f}-\x{1f7d}\x{1f80}-\x{1fb4}\x{1fb6}-\x{1fbc}\x{1fbe}\x{1fc2}-\x{1fc4}\x{1fc6}-\x{1fcc}\x{1fd0}-\x{1fd3}\x{1fd6}-\x{1fdb}\x{1fe0}-\x{1fec}\x{1ff2}-\x{1ff4}\x{1ff6}-\x{1ffc}\x{200e}\x{2071}\x{207f}\x{2090}-\x{209c}\x{2102}\x{2107}\x{210a}-\x{2113}\x{2115}\x{2119}-\x{211d}\x{2124}\x{2126}\x{2128}\x{212a}-\x{212d}\x{212f}-\x{2139}\x{213c}-\x{213f}\x{2145}-\x{2149}\x{214e}\x{214f}\x{2160}-\x{2188}\x{2336}-\x{237a}\x{2395}\x{249c}-\x{24e9}\x{26ac}\x{2800}-\x{28ff}\x{2c00}-\x{2c2e}\x{2c30}-\x{2c5e}\x{2c60}-\x{2ce4}\x{2ceb}-\x{2cee}\x{2cf2}\x{2cf3}\x{2d00}-\x{2d25}\x{2d27}\x{2d2d}\x{2d30}-\x{2d67}\x{2d6f}\x{2d70}\x{2d80}-\x{2d96}\x{2da0}-\x{2da6}\x{2da8}-\x{2dae}\x{2db0}-\x{2db6}\x{2db8}-\x{2dbe}\x{2dc0}-\x{2dc6}\x{2dc8}-\x{2dce}\x{2dd0}-\x{2dd6}\x{2dd8}-\x{2dde}\x{3005}-\x{3007}\x{3021}-\x{3029}\x{302e}\x{302f}\x{3031}-\x{3035}\x{3038}-\x{303c}\x{3041}-\x{3096}\x{309d}-\x{309f}\x{30a1}-\x{30fa}\x{30fc}-\x{30ff}\x{3105}-\x{312d}\x{3131}-\x{318e}\x{3190}-\x{31ba}\x{31f0}-\x{321c}\x{3220}-\x{324f}\x{3260}-\x{327b}\x{327f}-\x{32b0}\x{32c0}-\x{32cb}\x{32d0}-\x{32fe}\x{3300}-\x{3376}\x{337b}-\x{33dd}\x{33e0}-\x{33fe}\x{3400}-\x{4db5}\x{4e00}-\x{9fd5}\x{a000}-\x{a48c}\x{a4d0}-\x{a60c}\x{a610}-\x{a62b}\x{a640}-\x{a66e}\x{a680}-\x{a69d}\x{a6a0}-\x{a6ef}\x{a6f2}-\x{a6f7}\x{a722}-\x{a787}\x{a789}-\x{a7ad}\x{a7b0}-\x{a7b7}\x{a7f7}-\x{a801}\x{a803}-\x{a805}\x{a807}-\x{a80a}\x{a80c}-\x{a824}\x{a827}\x{a830}-\x{a837}\x{a840}-\x{a873}\x{a880}-\x{a8c3}\x{a8ce}-\x{a8d9}\x{a8f2}-\x{a8fd}\x{a900}-\x{a925}\x{a92e}-\x{a946}\x{a952}\x{a953}\x{a95f}-\x{a97c}\x{a983}-\x{a9b2}\x{a9b4}\x{a9b5}\x{a9ba}\x{a9bb}\x{a9bd}-\x{a9cd}\x{a9cf}-\x{a9d9}\x{a9de}-\x{a9e4}\x{a9e6}-\x{a9fe}\x{aa00}-\x{aa28}\x{aa2f}\x{aa30}\x{aa33}\x{aa34}\x{aa40}-\x{aa42}\x{aa44}-\x{aa4b}\x{aa4d}\x{aa50}-\x{aa59}\x{aa5c}-\x{aa7b}\x{aa7d}-\x{aaaf}\x{aab1}\x{aab5}\x{aab6}\x{aab9}-\x{aabd}\x{aac0}\x{aac2}\x{aadb}-\x{aaeb}\x{aaee}-\x{aaf5}\x{ab01}-\x{ab06}\x{ab09}-\x{ab0e}\x{ab11}-\x{ab16}\x{ab20}-\x{ab26}\x{ab28}-\x{ab2e}\x{ab30}-\x{ab65}\x{ab70}-\x{abe4}\x{abe6}\x{abe7}\x{abe9}-\x{abec}\x{abf0}-\x{abf9}\x{ac00}-\x{d7a3}\x{d7b0}-\x{d7c6}\x{d7cb}-\x{d7fb}\x{e000}-\x{fa6d}\x{fa70}-\x{fad9}\x{fb00}-\x{fb06}\x{fb13}-\x{fb17}\x{ff21}-\x{ff3a}\x{ff41}-\x{ff5a}\x{ff66}-\x{ffbe}\x{ffc2}-\x{ffc7}\x{ffca}-\x{ffcf}\x{ffd2}-\x{ffd7}\x{ffda}-\x{ffdc}\x{10000}-\x{1000b}\x{1000d}-\x{10026}\x{10028}-\x{1003a}\x{1003c}\x{1003d}\x{1003f}-\x{1004d}\x{10050}-\x{1005d}\x{10080}-\x{100fa}\x{10100}\x{10102}\x{10107}-\x{10133}\x{10137}-\x{1013f}\x{101d0}-\x{101fc}\x{10280}-\x{1029c}\x{102a0}-\x{102d0}\x{10300}-\x{10323}\x{10330}-\x{1034a}\x{10350}-\x{10375}\x{10380}-\x{1039d}\x{1039f}-\x{103c3}\x{103c8}-\x{103d5}\x{10400}-\x{1049d}\x{104a0}-\x{104a9}\x{10500}-\x{10527}\x{10530}-\x{10563}\x{1056f}\x{10600}-\x{10736}\x{10740}-\x{10755}\x{10760}-\x{10767}\x{11000}\x{11002}-\x{11037}\x{11047}-\x{1104d}\x{11066}-\x{1106f}\x{11082}-\x{110b2}\x{110b7}\x{110b8}\x{110bb}-\x{110c1}\x{110d0}-\x{110e8}\x{110f0}-\x{110f9}\x{11103}-\x{11126}\x{1112c}\x{11136}-\x{11143}\x{11150}-\x{11172}\x{11174}-\x{11176}\x{11182}-\x{111b5}\x{111bf}-\x{111c9}\x{111cd}\x{111d0}-\x{111df}\x{111e1}-\x{111f4}\x{11200}-\x{11211}\x{11213}-\x{1122e}\x{11232}\x{11233}\x{11235}\x{11238}-\x{1123d}\x{11280}-\x{11286}\x{11288}\x{1128a}-\x{1128d}\x{1128f}-\x{1129d}\x{1129f}-\x{112a9}\x{112b0}-\x{112de}\x{112e0}-\x{112e2}\x{112f0}-\x{112f9}\x{11302}\x{11303}\x{11305}-\x{1130c}\x{1130f}\x{11310}\x{11313}-\x{11328}\x{1132a}-\x{11330}\x{11332}\x{11333}\x{11335}-\x{11339}\x{1133d}-\x{1133f}\x{11341}-\x{11344}\x{11347}\x{11348}\x{1134b}-\x{1134d}\x{11350}\x{11357}\x{1135d}-\x{11363}\x{11480}-\x{114b2}\x{114b9}\x{114bb}-\x{114be}\x{114c1}\x{114c4}-\x{114c7}\x{114d0}-\x{114d9}\x{11580}-\x{115b1}\x{115b8}-\x{115bb}\x{115be}\x{115c1}-\x{115db}\x{11600}-\x{11632}\x{1163b}\x{1163c}\x{1163e}\x{11641}-\x{11644}\x{11650}-\x{11659}\x{11680}-\x{116aa}\x{116ac}\x{116ae}\x{116af}\x{116b6}\x{116c0}-\x{116c9}\x{11700}-\x{11719}\x{11720}\x{11721}\x{11726}\x{11730}-\x{1173f}\x{118a0}-\x{118f2}\x{118ff}\x{11ac0}-\x{11af8}\x{12000}-\x{12399}\x{12400}-\x{1246e}\x{12470}-\x{12474}\x{12480}-\x{12543}\x{13000}-\x{1342e}\x{14400}-\x{14646}\x{16800}-\x{16a38}\x{16a40}-\x{16a5e}\x{16a60}-\x{16a69}\x{16a6e}\x{16a6f}\x{16ad0}-\x{16aed}\x{16af5}\x{16b00}-\x{16b2f}\x{16b37}-\x{16b45}\x{16b50}-\x{16b59}\x{16b5b}-\x{16b61}\x{16b63}-\x{16b77}\x{16b7d}-\x{16b8f}\x{16f00}-\x{16f44}\x{16f50}-\x{16f7e}\x{16f93}-\x{16f9f}\x{1b000}\x{1b001}\x{1bc00}-\x{1bc6a}\x{1bc70}-\x{1bc7c}\x{1bc80}-\x{1bc88}\x{1bc90}-\x{1bc99}\x{1bc9c}\x{1bc9f}\x{1d000}-\x{1d0f5}\x{1d100}-\x{1d126}\x{1d129}-\x{1d166}\x{1d16a}-\x{1d172}\x{1d183}\x{1d184}\x{1d18c}-\x{1d1a9}\x{1d1ae}-\x{1d1e8}\x{1d360}-\x{1d371}\x{1d400}-\x{1d454}\x{1d456}-\x{1d49c}\x{1d49e}\x{1d49f}\x{1d4a2}\x{1d4a5}\x{1d4a6}\x{1d4a9}-\x{1d4ac}\x{1d4ae}-\x{1d4b9}\x{1d4bb}\x{1d4bd}-\x{1d4c3}\x{1d4c5}-\x{1d505}\x{1d507}-\x{1d50a}\x{1d50d}-\x{1d514}\x{1d516}-\x{1d51c}\x{1d51e}-\x{1d539}\x{1d53b}-\x{1d53e}\x{1d540}-\x{1d544}\x{1d546}\x{1d54a}-\x{1d550}\x{1d552}-\x{1d6a5}\x{1d6a8}-\x{1d6da}\x{1d6dc}-\x{1d714}\x{1d716}-\x{1d74e}\x{1d750}-\x{1d788}\x{1d78a}-\x{1d7c2}\x{1d7c4}-\x{1d7cb}\x{1d800}-\x{1d9ff}\x{1da37}-\x{1da3a}\x{1da6d}-\x{1da74}\x{1da76}-\x{1da83}\x{1da85}-\x{1da8b}\x{1f110}-\x{1f12e}\x{1f130}-\x{1f169}\x{1f170}-\x{1f19a}\x{1f1e6}-\x{1f202}\x{1f210}-\x{1f23a}\x{1f240}-\x{1f248}\x{1f250}\x{1f251}\x{20000}-\x{2a6d6}\x{2a700}-\x{2b734}\x{2b740}-\x{2b81d}\x{2b820}-\x{2cea1}\x{2f800}-\x{2fa1d}\x{f0000}-\x{ffffd}\x{100000}-\x{10fffd}])|([\x{590}\x{5be}\x{5c0}\x{5c3}\x{5c6}\x{5c8}-\x{5ff}\x{7c0}-\x{7ea}\x{7f4}\x{7f5}\x{7fa}-\x{815}\x{81a}\x{824}\x{828}\x{82e}-\x{858}\x{85c}-\x{89f}\x{200f}\x{fb1d}\x{fb1f}-\x{fb28}\x{fb2a}-\x{fb4f}\x{10800}-\x{1091e}\x{10920}-\x{10a00}\x{10a04}\x{10a07}-\x{10a0b}\x{10a10}-\x{10a37}\x{10a3b}-\x{10a3e}\x{10a40}-\x{10ae4}\x{10ae7}-\x{10b38}\x{10b40}-\x{10e5f}\x{10e7f}-\x{10fff}\x{1e800}-\x{1e8cf}\x{1e8d7}-\x{1edff}\x{1ef00}-\x{1efff}\x{608}\x{60b}\x{60d}\x{61b}-\x{64a}\x{66d}-\x{66f}\x{671}-\x{6d5}\x{6e5}\x{6e6}\x{6ee}\x{6ef}\x{6fa}-\x{710}\x{712}-\x{72f}\x{74b}-\x{7a5}\x{7b1}-\x{7bf}\x{8a0}-\x{8e2}\x{fb50}-\x{fd3d}\x{fd40}-\x{fdcf}\x{fdf0}-\x{fdfc}\x{fdfe}\x{fdff}\x{fe70}-\x{fefe}\x{1ee00}-\x{1eeef}\x{1eef2}-\x{1eeff}]))/u';
 175         // @codeCoverageIgnoreEnd
 176         // @codingStandardsIgnoreEnd
 177
 178         /**
 179          * Get a cached or new language object for a given language code
 180          * @param string $code
 181          * @return Language
 182          */
 183         static function factory( $code ) {
 184                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 185
 186                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 187                         $code = $wgDummyLanguageCodes[$code];
 188                 }
 189
 190                 // get the language object to process
 191                 $langObj = isset( self::$mLangObjCache[$code] )
 192                         ? self::$mLangObjCache[$code]
 193                         : self::newFromCode( $code );
 194
 195                 // merge the language object in to get it up front in the cache
 196                 self::$mLangObjCache = array_merge( array( $code => $langObj ), self::$mLangObjCache );
 197                 // get rid of the oldest ones in case we have an overflow
 198                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 199
 200                 return $langObj;
 201         }
 202
 203         /**
 204          * Create a language object for a given language code
 205          * @param string $code
 206          * @throws MWException
 207          * @return Language
 208          */
 209         protected static function newFromCode( $code ) {
 210                 // Protect against path traversal below
 211                 if ( !Language::isValidCode( $code )
 212                         || strcspn( $code, ":/\\\000" ) !== strlen( $code )
 213                 ) {
 214                         throw new MWException( "Invalid language code \"$code\"" );
 215                 }
 216
 217                 if ( !Language::isValidBuiltInCode( $code ) ) {
 218                         // It's not possible to customise this code with class files, so
 219                         // just return a Language object. This is to support uselang= hacks.
 220                         $lang = new Language;
 221                         $lang->setCode( $code );
 222                         return $lang;
 223                 }
 224
 225                 // Check if there is a language class for the code
 226                 $class = self::classFromCode( $code );
 227                 self::preloadLanguageClass( $class );
 228                 if ( class_exists( $class ) ) {
 229                         $lang = new $class;
 230                         return $lang;
 231                 }
 232
 233                 // Keep trying the fallback list until we find an existing class
 234                 $fallbacks = Language::getFallbacksFor( $code );
 235                 foreach ( $fallbacks as $fallbackCode ) {
 236                         if ( !Language::isValidBuiltInCode( $fallbackCode ) ) {
 237                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 238                         }
 239
 240                         $class = self::classFromCode( $fallbackCode );
 241                         self::preloadLanguageClass( $class );
 242                         if ( class_exists( $class ) ) {
 243                                 $lang = Language::newFromCode( $fallbackCode );
 244                                 $lang->setCode( $code );
 245                                 return $lang;
 246                         }
 247                 }
 248
 249                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 250         }
 251
 252         /**
 253          * Checks whether any localisation is available for that language tag
 254          * in MediaWiki (MessagesXx.php exists).
 255          *
 256          * @param string $code Language tag (in lower case)
 257          * @return bool Whether language is supported
 258          * @since 1.21
 259          */
 260         public static function isSupportedLanguage( $code ) {
 261                 if ( !self::isValidBuiltInCode( $code ) ) {
 262                         return false;
 263                 }
 264
 265                 if ( $code === 'qqq' ) {
 266                         return false;
 267                 }
 268
 269                 return is_readable( self::getMessagesFileName( $code ) ) ||
 270                         is_readable( self::getJsonMessagesFileName( $code ) );
 271         }
 272
 273         /**
 274          * Returns true if a language code string is a well-formed language tag
 275          * according to RFC 5646.
 276          * This function only checks well-formedness; it doesn't check that
 277          * language, script or variant codes actually exist in the repositories.
 278          *
 279          * Based on regexes by Mark Davis of the Unicode Consortium:
 280          * http://unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 281          *
 282          * @param string $code
 283          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 284          *
 285          * @return bool
 286          * @since 1.21
 287          */
 288         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 289                 $alpha = '[a-z]';
 290                 $digit = '[0-9]';
 291                 $alphanum = '[a-z0-9]';
 292                 $x = 'x'; # private use singleton
 293                 $singleton = '[a-wy-z]'; # other singleton
 294                 $s = $lenient ? '[-_]' : '-';
 295
 296                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 297                 $script = "$alpha{4}"; # ISO 15924
 298                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 299                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 300                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 301                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 302
 303                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 304                 # Since these are limited, this is safe even later changes to the registry --
 305                 # the only oddity is that it might change the type of the tag, and thus
 306                 # the results from the capturing groups.
 307                 # http://www.iana.org/assignments/language-subtag-registry
 308
 309                 $grandfathered = "en{$s}GB{$s}oed"
 310                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 311                         . "|no{$s}(?:bok|nyn)"
 312                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 313                         . "|zh{$s}min{$s}nan";
 314
 315                 $variantList = "$variant(?:$s$variant)*";
 316                 $extensionList = "$extension(?:$s$extension)*";
 317
 318                 $langtag = "(?:($language)"
 319                         . "(?:$s$script)?"
 320                         . "(?:$s$region)?"
 321                         . "(?:$s$variantList)?"
 322                         . "(?:$s$extensionList)?"
 323                         . "(?:$s$privateUse)?)";
 324
 325                 # The final breakdown, with capturing groups for each of these components
 326                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 327
 328                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 329
 330                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 331         }
 332
 333         /**
 334          * Returns true if a language code string is of a valid form, whether or
 335          * not it exists. This includes codes which are used solely for
 336          * customisation via the MediaWiki namespace.
 337          *
 338          * @param string $code
 339          *
 340          * @return bool
 341          */
 342         public static function isValidCode( $code ) {
 343                 static $cache = array();
 344                 if ( isset( $cache[$code] ) ) {
 345                         return $cache[$code];
 346                 }
 347                 // People think language codes are html safe, so enforce it.
 348                 // Ideally we should only allow a-zA-Z0-9-
 349                 // but, .+ and other chars are often used for {{int:}} hacks
 350                 // see bugs 37564, 37587, 36938
 351                 $cache[$code] =
 352                         strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 353                         && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 354
 355                 return $cache[$code];
 356         }
 357
 358         /**
 359          * Returns true if a language code is of a valid form for the purposes of
 360          * internal customisation of MediaWiki, via Messages*.php or *.json.
 361          *
 362          * @param string $code
 363          *
 364          * @throws MWException
 365          * @since 1.18
 366          * @return bool
 367          */
 368         public static function isValidBuiltInCode( $code ) {
 369
 370                 if ( !is_string( $code ) ) {
 371                         if ( is_object( $code ) ) {
 372                                 $addmsg = " of class " . get_class( $code );
 373                         } else {
 374                                 $addmsg = '';
 375                         }
 376                         $type = gettype( $code );
 377                         throw new MWException( __METHOD__ . " must be passed a string, $type given$addmsg" );
 378                 }
 379
 380                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 381         }
 382
 383         /**
 384          * Returns true if a language code is an IETF tag known to MediaWiki.
 385          *
 386          * @param string $tag
 387          *
 388          * @since 1.21
 389          * @return bool
 390          */
 391         public static function isKnownLanguageTag( $tag ) {
 392                 static $coreLanguageNames;
 393
 394                 // Quick escape for invalid input to avoid exceptions down the line
 395                 // when code tries to process tags which are not valid at all.
 396                 if ( !self::isValidBuiltInCode( $tag ) ) {
 397                         return false;
 398                 }
 399
 400                 if ( $coreLanguageNames === null ) {
 401                         global $IP;
 402                         include "$IP/languages/Names.php";
 403                 }
 404
 405                 if ( isset( $coreLanguageNames[$tag] )
 406                         || self::fetchLanguageName( $tag, $tag ) !== ''
 407                 ) {
 408                         return true;
 409                 }
 410
 411                 return false;
 412         }
 413
 414         /**
 415          * @param string $code
 416          * @return string Name of the language class
 417          */
 418         public static function classFromCode( $code ) {
 419                 if ( $code == 'en' ) {
 420                         return 'Language';
 421                 } else {
 422                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
 423                 }
 424         }
 425
 426         /**
 427          * Includes language class files
 428          *
 429          * @param string $class Name of the language class
 430          */
 431         public static function preloadLanguageClass( $class ) {
 432                 global $IP;
 433
 434                 if ( $class === 'Language' ) {
 435                         return;
 436                 }
 437
 438                 if ( file_exists( "$IP/languages/classes/$class.php" ) ) {
 439                         include_once "$IP/languages/classes/$class.php";
 440                 }
 441         }
 442
 443         /**
 444          * Get the LocalisationCache instance
 445          *
 446          * @return LocalisationCache
 447          */
 448         public static function getLocalisationCache() {
 449                 if ( is_null( self::$dataCache ) ) {
 450                         global $wgLocalisationCacheConf;
 451                         $class = $wgLocalisationCacheConf['class'];
 452                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 453                 }
 454                 return self::$dataCache;
 455         }
 456
 457         function __construct() {
 458                 $this->mConverter = new FakeConverter( $this );
 459                 // Set the code to the name of the descendant
 460                 if ( get_class( $this ) == 'Language' ) {
 461                         $this->mCode = 'en';
 462                 } else {
 463                         $this->mCode = str_replace( '_', '-', strtolower( substr( get_class( $this ), 8 ) ) );
 464                 }
 465                 self::getLocalisationCache();
 466         }
 467
 468         /**
 469          * Reduce memory usage
 470          */
 471         function __destruct() {
 472                 foreach ( $this as $name => $value ) {
 473                         unset( $this->$name );
 474                 }
 475         }
 476
 477         /**
 478          * Hook which will be called if this is the content language.
 479          * Descendants can use this to register hook functions or modify globals
 480          */
 481         function initContLang() {
 482         }
 483
 484         /**
 485          * @return array
 486          * @since 1.19
 487          */
 488         function getFallbackLanguages() {
 489                 return self::getFallbacksFor( $this->mCode );
 490         }
 491
 492         /**
 493          * Exports $wgBookstoreListEn
 494          * @return array
 495          */
 496         function getBookstoreList() {
 497                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 498         }
 499
 500         /**
 501          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 502          * available in localised form, it will be included in English.
 503          *
 504          * @return array
 505          */
 506         public function getNamespaces() {
 507                 if ( is_null( $this->namespaceNames ) ) {
 508                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 509
 510                         $this->namespaceNames = self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 511                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 512
 513                         $this->namespaceNames = $wgExtraNamespaces + $this->namespaceNames + $validNamespaces;
 514
 515                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 516                         if ( $wgMetaNamespaceTalk ) {
 517                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 518                         } else {
 519                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 520                                 $this->namespaceNames[NS_PROJECT_TALK] =
 521                                         $this->fixVariableInNamespace( $talk );
 522                         }
 523
 524                         # Sometimes a language will be localised but not actually exist on this wiki.
 525                         foreach ( $this->namespaceNames as $key => $text ) {
 526                                 if ( !isset( $validNamespaces[$key] ) ) {
 527                                         unset( $this->namespaceNames[$key] );
 528                                 }
 529                         }
 530
 531                         # The above mixing may leave namespaces out of canonical order.
 532                         # Re-order by namespace ID number...
 533                         ksort( $this->namespaceNames );
 534
 535                         Hooks::run( 'LanguageGetNamespaces', array( &$this->namespaceNames ) );
 536                 }
 537
 538                 return $this->namespaceNames;
 539         }
 540
 541         /**
 542          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 543          * @param array $namespaces Array of namespaces (id => name)
 544          */
 545         public function setNamespaces( array $namespaces ) {
 546                 $this->namespaceNames = $namespaces;
 547                 $this->mNamespaceIds = null;
 548         }
 549
 550         /**
 551          * Resets all of the namespace caches. Mainly used for testing
 552          */
 553         public function resetNamespaces() {
 554                 $this->namespaceNames = null;
 555                 $this->mNamespaceIds = null;
 556                 $this->namespaceAliases = null;
 557         }
 558
 559         /**
 560          * A convenience function that returns getNamespaces() with spaces instead of underscores
 561          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 562          *
 563          * @return array
 564          */
 565         function getFormattedNamespaces() {
 566                 $ns = $this->getNamespaces();
 567                 foreach ( $ns as $k => $v ) {
 568                         $ns[$k] = strtr( $v, '_', ' ' );
 569                 }
 570                 return $ns;
 571         }
 572
 573         /**
 574          * Get a namespace value by key
 575          *
 576          * <code>
 577          * $mw_ns = $wgContLang->getNsText( NS_MEDIAWIKI );
 578          * echo $mw_ns; // prints 'MediaWiki'
 579          * </code>
 580          *
 581          * @param int $index The array key of the namespace to return
 582          * @return string|bool String if the namespace value exists, otherwise false
 583          */
 584         function getNsText( $index ) {
 585                 $ns = $this->getNamespaces();
 586                 return isset( $ns[$index] ) ? $ns[$index] : false;
 587         }
 588
 589         /**
 590          * A convenience function that returns the same thing as
 591          * getNsText() except with '_' changed to ' ', useful for
 592          * producing output.
 593          *
 594          * <code>
 595          * $mw_ns = $wgContLang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 596          * echo $mw_ns; // prints 'MediaWiki talk'
 597          * </code>
 598          *
 599          * @param int $index The array key of the namespace to return
 600          * @return string Namespace name without underscores (empty string if namespace does not exist)
 601          */
 602         function getFormattedNsText( $index ) {
 603                 $ns = $this->getNsText( $index );
 604                 return strtr( $ns, '_', ' ' );
 605         }
 606
 607         /**
 608          * Returns gender-dependent namespace alias if available.
 609          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 610          * @param int $index Namespace index
 611          * @param string $gender Gender key (male, female... )
 612          * @return string
 613          * @since 1.18
 614          */
 615         function getGenderNsText( $index, $gender ) {
 616                 global $wgExtraGenderNamespaces;
 617
 618                 $ns = $wgExtraGenderNamespaces +
 619                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 620
 621                 return isset( $ns[$index][$gender] ) ? $ns[$index][$gender] : $this->getNsText( $index );
 622         }
 623
 624         /**
 625          * Whether this language uses gender-dependent namespace aliases.
 626          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 627          * @return bool
 628          * @since 1.18
 629          */
 630         function needsGenderDistinction() {
 631                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 632                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 633                         // $wgExtraGenderNamespaces overrides everything
 634                         return true;
 635                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 636                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 637                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 638                         return false;
 639                 } else {
 640                         // Check what is in i18n files
 641                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 642                         return count( $aliases ) > 0;
 643                 }
 644         }
 645
 646         /**
 647          * Get a namespace key by value, case insensitive.
 648          * Only matches namespace names for the current language, not the
 649          * canonical ones defined in Namespace.php.
 650          *
 651          * @param string $text
 652          * @return int|bool An integer if $text is a valid value otherwise false
 653          */
 654         function getLocalNsIndex( $text ) {
 655                 $lctext = $this->lc( $text );
 656                 $ids = $this->getNamespaceIds();
 657                 return isset( $ids[$lctext] ) ? $ids[$lctext] : false;
 658         }
 659
 660         /**
 661          * @return array
 662          */
 663         function getNamespaceAliases() {
 664                 if ( is_null( $this->namespaceAliases ) ) {
 665                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 666                         if ( !$aliases ) {
 667                                 $aliases = array();
 668                         } else {
 669                                 foreach ( $aliases as $name => $index ) {
 670                                         if ( $index === NS_PROJECT_TALK ) {
 671                                                 unset( $aliases[$name] );
 672                                                 $name = $this->fixVariableInNamespace( $name );
 673                                                 $aliases[$name] = $index;
 674                                         }
 675                                 }
 676                         }
 677
 678                         global $wgExtraGenderNamespaces;
 679                         $genders = $wgExtraGenderNamespaces +
 680                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 681                         foreach ( $genders as $index => $forms ) {
 682                                 foreach ( $forms as $alias ) {
 683                                         $aliases[$alias] = $index;
 684                                 }
 685                         }
 686
 687                         # Also add converted namespace names as aliases, to avoid confusion.
 688                         $convertedNames = array();
 689                         foreach ( $this->getVariants() as $variant ) {
 690                                 if ( $variant === $this->mCode ) {
 691                                         continue;
 692                                 }
 693                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 694                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 695                                 }
 696                         }
 697
 698                         $this->namespaceAliases = $aliases + $convertedNames;
 699                 }
 700
 701                 return $this->namespaceAliases;
 702         }
 703
 704         /**
 705          * @return array
 706          */
 707         function getNamespaceIds() {
 708                 if ( is_null( $this->mNamespaceIds ) ) {
 709                         global $wgNamespaceAliases;
 710                         # Put namespace names and aliases into a hashtable.
 711                         # If this is too slow, then we should arrange it so that it is done
 712                         # before caching. The catch is that at pre-cache time, the above
 713                         # class-specific fixup hasn't been done.
 714                         $this->mNamespaceIds = array();
 715                         foreach ( $this->getNamespaces() as $index => $name ) {
 716                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 717                         }
 718                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 719                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 720                         }
 721                         if ( $wgNamespaceAliases ) {
 722                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 723                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 724                                 }
 725                         }
 726                 }
 727                 return $this->mNamespaceIds;
 728         }
 729
 730         /**
 731          * Get a namespace key by value, case insensitive.  Canonical namespace
 732          * names override custom ones defined for the current language.
 733          *
 734          * @param string $text
 735          * @return int|bool An integer if $text is a valid value otherwise false
 736          */
 737         function getNsIndex( $text ) {
 738                 $lctext = $this->lc( $text );
 739                 $ns = MWNamespace::getCanonicalIndex( $lctext );
 740                 if ( $ns !== null ) {
 741                         return $ns;
 742                 }
 743                 $ids = $this->getNamespaceIds();
 744                 return isset( $ids[$lctext] ) ? $ids[$lctext] : false;
 745         }
 746
 747         /**
 748          * short names for language variants used for language conversion links.
 749          *
 750          * @param string $code
 751          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 752          * @return string
 753          */
 754         function getVariantname( $code, $usemsg = true ) {
 755                 $msg = "variantname-$code";
 756                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 757                         return $this->getMessageFromDB( $msg );
 758                 }
 759                 $name = self::fetchLanguageName( $code );
 760                 if ( $name ) {
 761                         return $name; # if it's defined as a language name, show that
 762                 } else {
 763                         # otherwise, output the language code
 764                         return $code;
 765                 }
 766         }
 767
 768         /**
 769          * @deprecated since 1.24, doesn't handle conflicting aliases. Use
 770          *   SpecialPageFactory::getLocalNameFor instead.
 771          * @param string $name
 772          * @return string
 773          */
 774         function specialPage( $name ) {
 775                 $aliases = $this->getSpecialPageAliases();
 776                 if ( isset( $aliases[$name][0] ) ) {
 777                         $name = $aliases[$name][0];
 778                 }
 779                 return $this->getNsText( NS_SPECIAL ) . ':' . $name;
 780         }
 781
 782         /**
 783          * @return array
 784          */
 785         function getDatePreferences() {
 786                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 787         }
 788
 789         /**
 790          * @return array
 791          */
 792         function getDateFormats() {
 793                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 794         }
 795
 796         /**
 797          * @return array|string
 798          */
 799         function getDefaultDateFormat() {
 800                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 801                 if ( $df === 'dmy or mdy' ) {
 802                         global $wgAmericanDates;
 803                         return $wgAmericanDates ? 'mdy' : 'dmy';
 804                 } else {
 805                         return $df;
 806                 }
 807         }
 808
 809         /**
 810          * @return array
 811          */
 812         function getDatePreferenceMigrationMap() {
 813                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 814         }
 815
 816         /**
 817          * @param string $image
 818          * @return array|null
 819          */
 820         function getImageFile( $image ) {
 821                 return self::$dataCache->getSubitem( $this->mCode, 'imageFiles', $image );
 822         }
 823
 824         /**
 825          * @return array
 826          * @since 1.24
 827          */
 828         function getImageFiles() {
 829                 return self::$dataCache->getItem( $this->mCode, 'imageFiles' );
 830         }
 831
 832         /**
 833          * @return array
 834          */
 835         function getExtraUserToggles() {
 836                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 837         }
 838
 839         /**
 840          * @param string $tog
 841          * @return string
 842          */
 843         function getUserToggle( $tog ) {
 844                 return $this->getMessageFromDB( "tog-$tog" );
 845         }
 846
 847         /**
 848          * Get native language names, indexed by code.
 849          * Only those defined in MediaWiki, no other data like CLDR.
 850          * If $customisedOnly is true, only returns codes with a messages file
 851          *
 852          * @param bool $customisedOnly
 853          *
 854          * @return array
 855          * @deprecated since 1.20, use fetchLanguageNames()
 856          */
 857         public static function getLanguageNames( $customisedOnly = false ) {
 858                 return self::fetchLanguageNames( null, $customisedOnly ? 'mwfile' : 'mw' );
 859         }
 860
 861         /**
 862          * Get translated language names. This is done on best effort and
 863          * by default this is exactly the same as Language::getLanguageNames.
 864          * The CLDR extension provides translated names.
 865          * @param string $code Language code.
 866          * @return array Language code => language name
 867          * @since 1.18.0
 868          * @deprecated since 1.20, use fetchLanguageNames()
 869          */
 870         public static function getTranslatedLanguageNames( $code ) {
 871                 return self::fetchLanguageNames( $code, 'all' );
 872         }
 873
 874         /**
 875          * Get an array of language names, indexed by code.
 876          * @param null|string $inLanguage Code of language in which to return the names
 877          *              Use null for autonyms (native names)
 878          * @param string $include One of:
 879          *              'all' all available languages
 880          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 881          *              'mwfile' only if the language is in 'mw' *and* has a message file
 882          * @return array Language code => language name
 883          * @since 1.20
 884          */
 885         public static function fetchLanguageNames( $inLanguage = null, $include = 'mw' ) {
 886                 $cacheKey = $inLanguage === null ? 'null' : $inLanguage;
 887                 $cacheKey .= ":$include";
 888                 if ( self::$languageNameCache === null ) {
 889                         self::$languageNameCache = new MapCacheLRU( 20 );
 890                 }
 891                 if ( self::$languageNameCache->has( $cacheKey ) ) {
 892                         $ret = self::$languageNameCache->get( $cacheKey );
 893                 } else {
 894                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 895                         self::$languageNameCache->set( $cacheKey, $ret );
 896                 }
 897                 return $ret;
 898         }
 899
 900         /**
 901          * Uncached helper for fetchLanguageNames
 902          * @param null|string $inLanguage Code of language in which to return the names
 903          *              Use null for autonyms (native names)
 904          * @param string $include One of:
 905          *              'all' all available languages
 906          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 907          *              'mwfile' only if the language is in 'mw' *and* has a message file
 908          * @return array Language code => language name
 909          */
 910         private static function fetchLanguageNamesUncached( $inLanguage = null, $include = 'mw' ) {
 911                 global $wgExtraLanguageNames;
 912                 static $coreLanguageNames;
 913
 914                 if ( $coreLanguageNames === null ) {
 915                         global $IP;
 916                         include "$IP/languages/Names.php";
 917                 }
 918
 919                 // If passed an invalid language code to use, fallback to en
 920                 if ( $inLanguage !== null && !Language::isValidCode( $inLanguage ) ) {
 921                         $inLanguage = 'en';
 922                 }
 923
 924                 $names = array();
 925
 926                 if ( $inLanguage ) {
 927                         # TODO: also include when $inLanguage is null, when this code is more efficient
 928                         Hooks::run( 'LanguageGetTranslatedLanguageNames', array( &$names, $inLanguage ) );
 929                 }
 930
 931                 $mwNames = $wgExtraLanguageNames + $coreLanguageNames;
 932                 foreach ( $mwNames as $mwCode => $mwName ) {
 933                         # - Prefer own MediaWiki native name when not using the hook
 934                         # - For other names just add if not added through the hook
 935                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 936                                 $names[$mwCode] = $mwName;
 937                         }
 938                 }
 939
 940                 if ( $include === 'all' ) {
 941                         ksort( $names );
 942                         return $names;
 943                 }
 944
 945                 $returnMw = array();
 946                 $coreCodes = array_keys( $mwNames );
 947                 foreach ( $coreCodes as $coreCode ) {
 948                         $returnMw[$coreCode] = $names[$coreCode];
 949                 }
 950
 951                 if ( $include === 'mwfile' ) {
 952                         $namesMwFile = array();
 953                         # We do this using a foreach over the codes instead of a directory
 954                         # loop so that messages files in extensions will work correctly.
 955                         foreach ( $returnMw as $code => $value ) {
 956                                 if ( is_readable( self::getMessagesFileName( $code ) )
 957                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 958                                 ) {
 959                                         $namesMwFile[$code] = $names[$code];
 960                                 }
 961                         }
 962
 963                         ksort( $namesMwFile );
 964                         return $namesMwFile;
 965                 }
 966
 967                 ksort( $returnMw );
 968                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 969                 return $returnMw;
 970         }
 971
 972         /**
 973          * @param string $code The code of the language for which to get the name
 974          * @param null|string $inLanguage Code of language in which to return the name (null for autonyms)
 975          * @param string $include 'all', 'mw' or 'mwfile'; see fetchLanguageNames()
 976          * @return string Language name or empty
 977          * @since 1.20
 978          */
 979         public static function fetchLanguageName( $code, $inLanguage = null, $include = 'all' ) {
 980                 $code = strtolower( $code );
 981                 $array = self::fetchLanguageNames( $inLanguage, $include );
 982                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 983         }
 984
 985         /**
 986          * Get a message from the MediaWiki namespace.
 987          *
 988          * @param string $msg Message name
 989          * @return string
 990          */
 991         function getMessageFromDB( $msg ) {
 992                 return $this->msg( $msg )->text();
 993         }
 994
 995         /**
 996          * Get message object in this language. Only for use inside this class.
 997          *
 998          * @param string $msg Message name
 999          * @return Message
1000          */
1001         protected function msg( $msg ) {
1002                 return wfMessage( $msg )->inLanguage( $this );
1003         }
1004
1005         /**
1006          * Get the native language name of $code.
1007          * Only if defined in MediaWiki, no other data like CLDR.
1008          * @param string $code
1009          * @return string
1010          * @deprecated since 1.20, use fetchLanguageName()
1011          */
1012         function getLanguageName( $code ) {
1013                 return self::fetchLanguageName( $code );
1014         }
1015
1016         /**
1017          * @param string $key
1018          * @return string
1019          */
1020         function getMonthName( $key ) {
1021                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
1022         }
1023
1024         /**
1025          * @return array
1026          */
1027         function getMonthNamesArray() {
1028                 $monthNames = array( '' );
1029                 for ( $i = 1; $i < 13; $i++ ) {
1030                         $monthNames[] = $this->getMonthName( $i );
1031                 }
1032                 return $monthNames;
1033         }
1034
1035         /**
1036          * @param string $key
1037          * @return string
1038          */
1039         function getMonthNameGen( $key ) {
1040                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
1041         }
1042
1043         /**
1044          * @param string $key
1045          * @return string
1046          */
1047         function getMonthAbbreviation( $key ) {
1048                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
1049         }
1050
1051         /**
1052          * @return array
1053          */
1054         function getMonthAbbreviationsArray() {
1055                 $monthNames = array( '' );
1056                 for ( $i = 1; $i < 13; $i++ ) {
1057                         $monthNames[] = $this->getMonthAbbreviation( $i );
1058                 }
1059                 return $monthNames;
1060         }
1061
1062         /**
1063          * @param string $key
1064          * @return string
1065          */
1066         function getWeekdayName( $key ) {
1067                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1068         }
1069
1070         /**
1071          * @param string $key
1072          * @return string
1073          */
1074         function getWeekdayAbbreviation( $key ) {
1075                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1076         }
1077
1078         /**
1079          * @param string $key
1080          * @return string
1081          */
1082         function getIranianCalendarMonthName( $key ) {
1083                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1084         }
1085
1086         /**
1087          * @param string $key
1088          * @return string
1089          */
1090         function getHebrewCalendarMonthName( $key ) {
1091                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1092         }
1093
1094         /**
1095          * @param string $key
1096          * @return string
1097          */
1098         function getHebrewCalendarMonthNameGen( $key ) {
1099                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1100         }
1101
1102         /**
1103          * @param string $key
1104          * @return string
1105          */
1106         function getHijriCalendarMonthName( $key ) {
1107                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1108         }
1109
1110         /**
1111          * Pass through result from $dateTimeObj->format()
1112          * @param DateTime|bool|null &$dateTimeObj
1113          * @param string $ts
1114          * @param DateTimeZone|bool|null $zone
1115          * @param string $code
1116          * @return string
1117          */
1118         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1119                 if ( !$dateTimeObj ) {
1120                         $dateTimeObj = DateTime::createFromFormat(
1121                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1122                         );
1123                 }
1124                 return $dateTimeObj->format( $code );
1125         }
1126
1127         /**
1128          * This is a workalike of PHP's date() function, but with better
1129          * internationalisation, a reduced set of format characters, and a better
1130          * escaping format.
1131          *
1132          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1133          * the PHP manual for definitions. There are a number of extensions, which
1134          * start with "x":
1135          *
1136          *    xn   Do not translate digits of the next numeric format character
1137          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1138          *    xr   Use roman numerals for the next numeric format character
1139          *    xh   Use hebrew numerals for the next numeric format character
1140          *    xx   Literal x
1141          *    xg   Genitive month name
1142          *
1143          *    xij  j (day number) in Iranian calendar
1144          *    xiF  F (month name) in Iranian calendar
1145          *    xin  n (month number) in Iranian calendar
1146          *    xiy  y (two digit year) in Iranian calendar
1147          *    xiY  Y (full year) in Iranian calendar
1148          *
1149          *    xjj  j (day number) in Hebrew calendar
1150          *    xjF  F (month name) in Hebrew calendar
1151          *    xjt  t (days in month) in Hebrew calendar
1152          *    xjx  xg (genitive month name) in Hebrew calendar
1153          *    xjn  n (month number) in Hebrew calendar
1154          *    xjY  Y (full year) in Hebrew calendar
1155          *
1156          *    xmj  j (day number) in Hijri calendar
1157          *    xmF  F (month name) in Hijri calendar
1158          *    xmn  n (month number) in Hijri calendar
1159          *    xmY  Y (full year) in Hijri calendar
1160          *
1161          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1162          *                       identical to the Gregorian calendar
1163          *    xoY  Y (full year) in Minguo calendar or Juche year.
1164          *                       Months and days are identical to the
1165          *                       Gregorian calendar
1166          *    xtY  Y (full year) in Japanese nengo. Months and days are
1167          *                       identical to the Gregorian calendar
1168          *
1169          * Characters enclosed in double quotes will be considered literal (with
1170          * the quotes themselves removed). Unmatched quotes will be considered
1171          * literal quotes. Example:
1172          *
1173          * "The month is" F       => The month is January
1174          * i's"                   => 20'11"
1175          *
1176          * Backslash escaping is also supported.
1177          *
1178          * Input timestamp is assumed to be pre-normalized to the desired local
1179          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1180          * $ts is UTC if $zone is not given.
1181          *
1182          * @param string $format
1183          * @param string $ts 14-character timestamp
1184          *      YYYYMMDDHHMMSS
1185          *      01234567890123
1186          * @param DateTimeZone $zone Timezone of $ts
1187          * @param[out] int $ttl The amount of time (in seconds) the output may be cached for.
1188          * Only makes sense if $ts is the current time.
1189          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1190          *
1191          * @throws MWException
1192          * @return string
1193          */
1194         function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = null ) {
1195                 $s = '';
1196                 $raw = false;
1197                 $roman = false;
1198                 $hebrewNum = false;
1199                 $dateTimeObj = false;
1200                 $rawToggle = false;
1201                 $iranian = false;
1202                 $hebrew = false;
1203                 $hijri = false;
1204                 $thai = false;
1205                 $minguo = false;
1206                 $tenno = false;
1207
1208                 $usedSecond = false;
1209                 $usedMinute = false;
1210                 $usedHour = false;
1211                 $usedAMPM = false;
1212                 $usedDay = false;
1213                 $usedWeek = false;
1214                 $usedMonth = false;
1215                 $usedYear = false;
1216                 $usedISOYear = false;
1217                 $usedIsLeapYear = false;
1218
1219                 $usedHebrewMonth = false;
1220                 $usedIranianMonth = false;
1221                 $usedHijriMonth = false;
1222                 $usedHebrewYear = false;
1223                 $usedIranianYear = false;
1224                 $usedHijriYear = false;
1225                 $usedTennoYear = false;
1226
1227                 if ( strlen( $ts ) !== 14 ) {
1228                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1229                 }
1230
1231                 if ( !ctype_digit( $ts ) ) {
1232                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1233                 }
1234
1235                 $formatLength = strlen( $format );
1236                 for ( $p = 0; $p < $formatLength; $p++ ) {
1237                         $num = false;
1238                         $code = $format[$p];
1239                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1240                                 $code .= $format[++$p];
1241                         }
1242
1243                         if ( ( $code === 'xi'
1244                                         || $code === 'xj'
1245                                         || $code === 'xk'
1246                                         || $code === 'xm'
1247                                         || $code === 'xo'
1248                                         || $code === 'xt' )
1249                                 && $p < $formatLength - 1 ) {
1250                                 $code .= $format[++$p];
1251                         }
1252
1253                         switch ( $code ) {
1254                                 case 'xx':
1255                                         $s .= 'x';
1256                                         break;
1257                                 case 'xn':
1258                                         $raw = true;
1259                                         break;
1260                                 case 'xN':
1261                                         $rawToggle = !$rawToggle;
1262                                         break;
1263                                 case 'xr':
1264                                         $roman = true;
1265                                         break;
1266                                 case 'xh':
1267                                         $hebrewNum = true;
1268                                         break;
1269                                 case 'xg':
1270                                         $usedMonth = true;
1271                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1272                                         break;
1273                                 case 'xjx':
1274                                         $usedHebrewMonth = true;
1275                                         if ( !$hebrew ) {
1276                                                 $hebrew = self::tsToHebrew( $ts );
1277                                         }
1278                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1279                                         break;
1280                                 case 'd':
1281                                         $usedDay = true;
1282                                         $num = substr( $ts, 6, 2 );
1283                                         break;
1284                                 case 'D':
1285                                         $usedDay = true;
1286                                         $s .= $this->getWeekdayAbbreviation(
1287                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1288                                         );
1289                                         break;
1290                                 case 'j':
1291                                         $usedDay = true;
1292                                         $num = intval( substr( $ts, 6, 2 ) );
1293                                         break;
1294                                 case 'xij':
1295                                         $usedDay = true;
1296                                         if ( !$iranian ) {
1297                                                 $iranian = self::tsToIranian( $ts );
1298                                         }
1299                                         $num = $iranian[2];
1300                                         break;
1301                                 case 'xmj':
1302                                         $usedDay = true;
1303                                         if ( !$hijri ) {
1304                                                 $hijri = self::tsToHijri( $ts );
1305                                         }
1306                                         $num = $hijri[2];
1307                                         break;
1308                                 case 'xjj':
1309                                         $usedDay = true;
1310                                         if ( !$hebrew ) {
1311                                                 $hebrew = self::tsToHebrew( $ts );
1312                                         }
1313                                         $num = $hebrew[2];
1314                                         break;
1315                                 case 'l':
1316                                         $usedDay = true;
1317                                         $s .= $this->getWeekdayName(
1318                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1319                                         );
1320                                         break;
1321                                 case 'F':
1322                                         $usedMonth = true;
1323                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1324                                         break;
1325                                 case 'xiF':
1326                                         $usedIranianMonth = true;
1327                                         if ( !$iranian ) {
1328                                                 $iranian = self::tsToIranian( $ts );
1329                                         }
1330                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1331                                         break;
1332                                 case 'xmF':
1333                                         $usedHijriMonth = true;
1334                                         if ( !$hijri ) {
1335                                                 $hijri = self::tsToHijri( $ts );
1336                                         }
1337                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1338                                         break;
1339                                 case 'xjF':
1340                                         $usedHebrewMonth = true;
1341                                         if ( !$hebrew ) {
1342                                                 $hebrew = self::tsToHebrew( $ts );
1343                                         }
1344                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1345                                         break;
1346                                 case 'm':
1347                                         $usedMonth = true;
1348                                         $num = substr( $ts, 4, 2 );
1349                                         break;
1350                                 case 'M':
1351                                         $usedMonth = true;
1352                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1353                                         break;
1354                                 case 'n':
1355                                         $usedMonth = true;
1356                                         $num = intval( substr( $ts, 4, 2 ) );
1357                                         break;
1358                                 case 'xin':
1359                                         $usedIranianMonth = true;
1360                                         if ( !$iranian ) {
1361                                                 $iranian = self::tsToIranian( $ts );
1362                                         }
1363                                         $num = $iranian[1];
1364                                         break;
1365                                 case 'xmn':
1366                                         $usedHijriMonth = true;
1367                                         if ( !$hijri ) {
1368                                                 $hijri = self::tsToHijri( $ts );
1369                                         }
1370                                         $num = $hijri[1];
1371                                         break;
1372                                 case 'xjn':
1373                                         $usedHebrewMonth = true;
1374                                         if ( !$hebrew ) {
1375                                                 $hebrew = self::tsToHebrew( $ts );
1376                                         }
1377                                         $num = $hebrew[1];
1378                                         break;
1379                                 case 'xjt':
1380                                         $usedHebrewMonth = true;
1381                                         if ( !$hebrew ) {
1382                                                 $hebrew = self::tsToHebrew( $ts );
1383                                         }
1384                                         $num = $hebrew[3];
1385                                         break;
1386                                 case 'Y':
1387                                         $usedYear = true;
1388                                         $num = substr( $ts, 0, 4 );
1389                                         break;
1390                                 case 'xiY':
1391                                         $usedIranianYear = true;
1392                                         if ( !$iranian ) {
1393                                                 $iranian = self::tsToIranian( $ts );
1394                                         }
1395                                         $num = $iranian[0];
1396                                         break;
1397                                 case 'xmY':
1398                                         $usedHijriYear = true;
1399                                         if ( !$hijri ) {
1400                                                 $hijri = self::tsToHijri( $ts );
1401                                         }
1402                                         $num = $hijri[0];
1403                                         break;
1404                                 case 'xjY':
1405                                         $usedHebrewYear = true;
1406                                         if ( !$hebrew ) {
1407                                                 $hebrew = self::tsToHebrew( $ts );
1408                                         }
1409                                         $num = $hebrew[0];
1410                                         break;
1411                                 case 'xkY':
1412                                         $usedYear = true;
1413                                         if ( !$thai ) {
1414                                                 $thai = self::tsToYear( $ts, 'thai' );
1415                                         }
1416                                         $num = $thai[0];
1417                                         break;
1418                                 case 'xoY':
1419                                         $usedYear = true;
1420                                         if ( !$minguo ) {
1421                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1422                                         }
1423                                         $num = $minguo[0];
1424                                         break;
1425                                 case 'xtY':
1426                                         $usedTennoYear = true;
1427                                         if ( !$tenno ) {
1428                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1429                                         }
1430                                         $num = $tenno[0];
1431                                         break;
1432                                 case 'y':
1433                                         $usedYear = true;
1434                                         $num = substr( $ts, 2, 2 );
1435                                         break;
1436                                 case 'xiy':
1437                                         $usedIranianYear = true;
1438                                         if ( !$iranian ) {
1439                                                 $iranian = self::tsToIranian( $ts );
1440                                         }
1441                                         $num = substr( $iranian[0], -2 );
1442                                         break;
1443                                 case 'a':
1444                                         $usedAMPM = true;
1445                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1446                                         break;
1447                                 case 'A':
1448                                         $usedAMPM = true;
1449                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1450                                         break;
1451                                 case 'g':
1452                                         $usedHour = true;
1453                                         $h = substr( $ts, 8, 2 );
1454                                         $num = $h % 12 ? $h % 12 : 12;
1455                                         break;
1456                                 case 'G':
1457                                         $usedHour = true;
1458                                         $num = intval( substr( $ts, 8, 2 ) );
1459                                         break;
1460                                 case 'h':
1461                                         $usedHour = true;
1462                                         $h = substr( $ts, 8, 2 );
1463                                         $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
1464                                         break;
1465                                 case 'H':
1466                                         $usedHour = true;
1467                                         $num = substr( $ts, 8, 2 );
1468                                         break;
1469                                 case 'i':
1470                                         $usedMinute = true;
1471                                         $num = substr( $ts, 10, 2 );
1472                                         break;
1473                                 case 's':
1474                                         $usedSecond = true;
1475                                         $num = substr( $ts, 12, 2 );
1476                                         break;
1477                                 case 'c':
1478                                 case 'r':
1479                                         $usedSecond = true;
1480                                         // fall through
1481                                 case 'e':
1482                                 case 'O':
1483                                 case 'P':
1484                                 case 'T':
1485                                         $s .= Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1486                                         break;
1487                                 case 'w':
1488                                 case 'N':
1489                                 case 'z':
1490                                         $usedDay = true;
1491                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1492                                         break;
1493                                 case 'W':
1494                                         $usedWeek = true;
1495                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1496                                         break;
1497                                 case 't':
1498                                         $usedMonth = true;
1499                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1500                                         break;
1501                                 case 'L':
1502                                         $usedIsLeapYear = true;
1503                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1504                                         break;
1505                                 case 'o':
1506                                         $usedISOYear = true;
1507                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1508                                         break;
1509                                 case 'U':
1510                                         $usedSecond = true;
1511                                         // fall through
1512                                 case 'I':
1513                                 case 'Z':
1514                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1515                                         break;
1516                                 case '\\':
1517                                         # Backslash escaping
1518                                         if ( $p < $formatLength - 1 ) {
1519                                                 $s .= $format[++$p];
1520                                         } else {
1521                                                 $s .= '\\';
1522                                         }
1523                                         break;
1524                                 case '"':
1525                                         # Quoted literal
1526                                         if ( $p < $formatLength - 1 ) {
1527                                                 $endQuote = strpos( $format, '"', $p + 1 );
1528                                                 if ( $endQuote === false ) {
1529                                                         # No terminating quote, assume literal "
1530                                                         $s .= '"';
1531                                                 } else {
1532                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1533                                                         $p = $endQuote;
1534                                                 }
1535                                         } else {
1536                                                 # Quote at end of string, assume literal "
1537                                                 $s .= '"';
1538                                         }
1539                                         break;
1540                                 default:
1541                                         $s .= $format[$p];
1542                         }
1543                         if ( $num !== false ) {
1544                                 if ( $rawToggle || $raw ) {
1545                                         $s .= $num;
1546                                         $raw = false;
1547                                 } elseif ( $roman ) {
1548                                         $s .= Language::romanNumeral( $num );
1549                                         $roman = false;
1550                                 } elseif ( $hebrewNum ) {
1551                                         $s .= self::hebrewNumeral( $num );
1552                                         $hebrewNum = false;
1553                                 } else {
1554                                         $s .= $this->formatNum( $num, true );
1555                                 }
1556                         }
1557                 }
1558
1559                 if ( $usedSecond ) {
1560                         $ttl = 1;
1561                 } elseif ( $usedMinute ) {
1562                         $ttl = 60 - substr( $ts, 12, 2 );
1563                 } elseif ( $usedHour ) {
1564                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1565                 } elseif ( $usedAMPM ) {
1566                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1567                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1568                 } elseif (
1569                         $usedDay ||
1570                         $usedHebrewMonth ||
1571                         $usedIranianMonth ||
1572                         $usedHijriMonth ||
1573                         $usedHebrewYear ||
1574                         $usedIranianYear ||
1575                         $usedHijriYear ||
1576                         $usedTennoYear
1577                 ) {
1578                         // @todo Someone who understands the non-Gregorian calendars
1579                         // should write proper logic for them so that they don't need purged every day.
1580                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1581                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1582                 } else {
1583                         $possibleTtls = array();
1584                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1585                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1586                         if ( $usedWeek ) {
1587                                 $possibleTtls[] =
1588                                         ( 7 - Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1589                                         $timeRemainingInDay;
1590                         } elseif ( $usedISOYear ) {
1591                                 // December 28th falls on the last ISO week of the year, every year.
1592                                 // The last ISO week of a year can be 52 or 53.
1593                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1594                                         'Ymd',
1595                                         substr( $ts, 0, 4 ) . '1228',
1596                                         $zone ?: new DateTimeZone( 'UTC' )
1597                                 )->format( 'W' );
1598                                 $currentISOWeek = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1599                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1600                                 $timeRemainingInWeek =
1601                                         ( 7 - Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1602                                         + $timeRemainingInDay;
1603                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1604                         }
1605
1606                         if ( $usedMonth ) {
1607                                 $possibleTtls[] =
1608                                         ( Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1609                                                 substr( $ts, 6, 2 ) ) * 86400
1610                                         + $timeRemainingInDay;
1611                         } elseif ( $usedYear ) {
1612                                 $possibleTtls[] =
1613                                         ( Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1614                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1615                                         + $timeRemainingInDay;
1616                         } elseif ( $usedIsLeapYear ) {
1617                                 $year = substr( $ts, 0, 4 );
1618                                 $timeRemainingInYear =
1619                                         ( Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1620                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1621                                         + $timeRemainingInDay;
1622                                 $mod = $year % 4;
1623                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1624                                         // this isn't a leap year. see when the next one starts
1625                                         $nextCandidate = $year - $mod + 4;
1626                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1627                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1628                                                         $timeRemainingInYear;
1629                                         } else {
1630                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1631                                                         $timeRemainingInYear;
1632                                         }
1633                                 } else {
1634                                         // this is a leap year, so the next year isn't
1635                                         $possibleTtls[] = $timeRemainingInYear;
1636                                 }
1637                         }
1638
1639                         if ( $possibleTtls ) {
1640                                 $ttl = min( $possibleTtls );
1641                         }
1642                 }
1643
1644                 return $s;
1645         }
1646
1647         private static $GREG_DAYS = array( 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 );
1648         private static $IRANIAN_DAYS = array( 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 );
1649
1650         /**
1651          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1652          * Gregorian dates to Iranian dates. Originally written in C, it
1653          * is released under the terms of GNU Lesser General Public
1654          * License. Conversion to PHP was performed by Niklas Laxström.
1655          *
1656          * Link: http://www.farsiweb.info/jalali/jalali.c
1657          *
1658          * @param string $ts
1659          *
1660          * @return string
1661          */
1662         private static function tsToIranian( $ts ) {
1663                 $gy = substr( $ts, 0, 4 ) -1600;
1664                 $gm = substr( $ts, 4, 2 ) -1;
1665                 $gd = substr( $ts, 6, 2 ) -1;
1666
1667                 # Days passed from the beginning (including leap years)
1668                 $gDayNo = 365 * $gy
1669                         + floor( ( $gy + 3 ) / 4 )
1670                         - floor( ( $gy + 99 ) / 100 )
1671                         + floor( ( $gy + 399 ) / 400 );
1672
1673                 // Add days of the past months of this year
1674                 for ( $i = 0; $i < $gm; $i++ ) {
1675                         $gDayNo += self::$GREG_DAYS[$i];
1676                 }
1677
1678                 // Leap years
1679                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1680                         $gDayNo++;
1681                 }
1682
1683                 // Days passed in current month
1684                 $gDayNo += (int)$gd;
1685
1686                 $jDayNo = $gDayNo - 79;
1687
1688                 $jNp = floor( $jDayNo / 12053 );
1689                 $jDayNo %= 12053;
1690
1691                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1692                 $jDayNo %= 1461;
1693
1694                 if ( $jDayNo >= 366 ) {
1695                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1696                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1697                 }
1698
1699                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1700                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1701                 }
1702
1703                 $jm = $i + 1;
1704                 $jd = $jDayNo + 1;
1705
1706                 return array( $jy, $jm, $jd );
1707         }
1708
1709         /**
1710          * Converting Gregorian dates to Hijri dates.
1711          *
1712          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1713          *
1714          * @see http://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1715          *
1716          * @param string $ts
1717          *
1718          * @return string
1719          */
1720         private static function tsToHijri( $ts ) {
1721                 $year = substr( $ts, 0, 4 );
1722                 $month = substr( $ts, 4, 2 );
1723                 $day = substr( $ts, 6, 2 );
1724
1725                 $zyr = $year;
1726                 $zd = $day;
1727                 $zm = $month;
1728                 $zy = $zyr;
1729
1730                 if (
1731                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1732                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1733                 ) {
1734                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1735                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1736                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1737                                         $zd - 32075;
1738                 } else {
1739                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1740                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1741                 }
1742
1743                 $zl = $zjd -1948440 + 10632;
1744                 $zn = (int)( ( $zl - 1 ) / 10631 );
1745                 $zl = $zl - 10631 * $zn + 354;
1746                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1747                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1748                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1749                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1750                 $zm = (int)( ( 24 * $zl ) / 709 );
1751                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1752                 $zy = 30 * $zn + $zj - 30;
1753
1754                 return array( $zy, $zm, $zd );
1755         }
1756
1757         /**
1758          * Converting Gregorian dates to Hebrew dates.
1759          *
1760          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1761          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1762          * to translate the relevant functions into PHP and release them under
1763          * GNU GPL.
1764          *
1765          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1766          * and Adar II is 14. In a non-leap year, Adar is 6.
1767          *
1768          * @param string $ts
1769          *
1770          * @return string
1771          */
1772         private static function tsToHebrew( $ts ) {
1773                 # Parse date
1774                 $year = substr( $ts, 0, 4 );
1775                 $month = substr( $ts, 4, 2 );
1776                 $day = substr( $ts, 6, 2 );
1777
1778                 # Calculate Hebrew year
1779                 $hebrewYear = $year + 3760;
1780
1781                 # Month number when September = 1, August = 12
1782                 $month += 4;
1783                 if ( $month > 12 ) {
1784                         # Next year
1785                         $month -= 12;
1786                         $year++;
1787                         $hebrewYear++;
1788                 }
1789
1790                 # Calculate day of year from 1 September
1791                 $dayOfYear = $day;
1792                 for ( $i = 1; $i < $month; $i++ ) {
1793                         if ( $i == 6 ) {
1794                                 # February
1795                                 $dayOfYear += 28;
1796                                 # Check if the year is leap
1797                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1798                                         $dayOfYear++;
1799                                 }
1800                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1801                                 $dayOfYear += 30;
1802                         } else {
1803                                 $dayOfYear += 31;
1804                         }
1805                 }
1806
1807                 # Calculate the start of the Hebrew year
1808                 $start = self::hebrewYearStart( $hebrewYear );
1809
1810                 # Calculate next year's start
1811                 if ( $dayOfYear <= $start ) {
1812                         # Day is before the start of the year - it is the previous year
1813                         # Next year's start
1814                         $nextStart = $start;
1815                         # Previous year
1816                         $year--;
1817                         $hebrewYear--;
1818                         # Add days since previous year's 1 September
1819                         $dayOfYear += 365;
1820                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1821                                 # Leap year
1822                                 $dayOfYear++;
1823                         }
1824                         # Start of the new (previous) year
1825                         $start = self::hebrewYearStart( $hebrewYear );
1826                 } else {
1827                         # Next year's start
1828                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1829                 }
1830
1831                 # Calculate Hebrew day of year
1832                 $hebrewDayOfYear = $dayOfYear - $start;
1833
1834                 # Difference between year's days
1835                 $diff = $nextStart - $start;
1836                 # Add 12 (or 13 for leap years) days to ignore the difference between
1837                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1838                 # difference is only about the year type
1839                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1840                         $diff += 13;
1841                 } else {
1842                         $diff += 12;
1843                 }
1844
1845                 # Check the year pattern, and is leap year
1846                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1847                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1848                 # and non-leap years
1849                 $yearPattern = $diff % 30;
1850                 # Check if leap year
1851                 $isLeap = $diff >= 30;
1852
1853                 # Calculate day in the month from number of day in the Hebrew year
1854                 # Don't check Adar - if the day is not in Adar, we will stop before;
1855                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1856                 $hebrewDay = $hebrewDayOfYear;
1857                 $hebrewMonth = 1;
1858                 $days = 0;
1859                 while ( $hebrewMonth <= 12 ) {
1860                         # Calculate days in this month
1861                         if ( $isLeap && $hebrewMonth == 6 ) {
1862                                 # Adar in a leap year
1863                                 if ( $isLeap ) {
1864                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1865                                         $days = 30;
1866                                         if ( $hebrewDay <= $days ) {
1867                                                 # Day in Adar I
1868                                                 $hebrewMonth = 13;
1869                                         } else {
1870                                                 # Subtract the days of Adar I
1871                                                 $hebrewDay -= $days;
1872                                                 # Try Adar II
1873                                                 $days = 29;
1874                                                 if ( $hebrewDay <= $days ) {
1875                                                         # Day in Adar II
1876                                                         $hebrewMonth = 14;
1877                                                 }
1878                                         }
1879                                 }
1880                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1881                                 # Cheshvan in a complete year (otherwise as the rule below)
1882                                 $days = 30;
1883                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1884                                 # Kislev in an incomplete year (otherwise as the rule below)
1885                                 $days = 29;
1886                         } else {
1887                                 # Odd months have 30 days, even have 29
1888                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1889                         }
1890                         if ( $hebrewDay <= $days ) {
1891                                 # In the current month
1892                                 break;
1893                         } else {
1894                                 # Subtract the days of the current month
1895                                 $hebrewDay -= $days;
1896                                 # Try in the next month
1897                                 $hebrewMonth++;
1898                         }
1899                 }
1900
1901                 return array( $hebrewYear, $hebrewMonth, $hebrewDay, $days );
1902         }
1903
1904         /**
1905          * This calculates the Hebrew year start, as days since 1 September.
1906          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1907          * Used for Hebrew date.
1908          *
1909          * @param int $year
1910          *
1911          * @return string
1912          */
1913         private static function hebrewYearStart( $year ) {
1914                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1915                 $b = intval( ( $year - 1 ) % 4 );
1916                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1917                 if ( $m < 0 ) {
1918                         $m--;
1919                 }
1920                 $Mar = intval( $m );
1921                 if ( $m < 0 ) {
1922                         $m++;
1923                 }
1924                 $m -= $Mar;
1925
1926                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1927                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1928                         $Mar++;
1929                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1930                         $Mar += 2;
1931                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1932                         $Mar++;
1933                 }
1934
1935                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1936                 return $Mar;
1937         }
1938
1939         /**
1940          * Algorithm to convert Gregorian dates to Thai solar dates,
1941          * Minguo dates or Minguo dates.
1942          *
1943          * Link: http://en.wikipedia.org/wiki/Thai_solar_calendar
1944          *       http://en.wikipedia.org/wiki/Minguo_calendar
1945          *       http://en.wikipedia.org/wiki/Japanese_era_name
1946          *
1947          * @param string $ts 14-character timestamp
1948          * @param string $cName Calender name
1949          * @return array Converted year, month, day
1950          */
1951         private static function tsToYear( $ts, $cName ) {
1952                 $gy = substr( $ts, 0, 4 );
1953                 $gm = substr( $ts, 4, 2 );
1954                 $gd = substr( $ts, 6, 2 );
1955
1956                 if ( !strcmp( $cName, 'thai' ) ) {
1957                         # Thai solar dates
1958                         # Add 543 years to the Gregorian calendar
1959                         # Months and days are identical
1960                         $gy_offset = $gy + 543;
1961                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1962                         # Minguo dates
1963                         # Deduct 1911 years from the Gregorian calendar
1964                         # Months and days are identical
1965                         $gy_offset = $gy - 1911;
1966                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1967                         # Nengō dates up to Meiji period
1968                         # Deduct years from the Gregorian calendar
1969                         # depending on the nengo periods
1970                         # Months and days are identical
1971                         if ( ( $gy < 1912 )
1972                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1973                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1974                         ) {
1975                                 # Meiji period
1976                                 $gy_gannen = $gy - 1868 + 1;
1977                                 $gy_offset = $gy_gannen;
1978                                 if ( $gy_gannen == 1 ) {
1979                                         $gy_offset = '元';
1980                                 }
1981                                 $gy_offset = '明治' . $gy_offset;
1982                         } elseif (
1983                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1984                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1985                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1986                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1987                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1988                         ) {
1989                                 # Taishō period
1990                                 $gy_gannen = $gy - 1912 + 1;
1991                                 $gy_offset = $gy_gannen;
1992                                 if ( $gy_gannen == 1 ) {
1993                                         $gy_offset = '元';
1994                                 }
1995                                 $gy_offset = '大正' . $gy_offset;
1996                         } elseif (
1997                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1998                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1999                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
2000                         ) {
2001                                 # Shōwa period
2002                                 $gy_gannen = $gy - 1926 + 1;
2003                                 $gy_offset = $gy_gannen;
2004                                 if ( $gy_gannen == 1 ) {
2005                                         $gy_offset = '元';
2006                                 }
2007                                 $gy_offset = '昭和' . $gy_offset;
2008                         } else {
2009                                 # Heisei period
2010                                 $gy_gannen = $gy - 1989 + 1;
2011                                 $gy_offset = $gy_gannen;
2012                                 if ( $gy_gannen == 1 ) {
2013                                         $gy_offset = '元';
2014                                 }
2015                                 $gy_offset = '平成' . $gy_offset;
2016                         }
2017                 } else {
2018                         $gy_offset = $gy;
2019                 }
2020
2021                 return array( $gy_offset, $gm, $gd );
2022         }
2023
2024         /**
2025          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2026          *
2027          * This is the rule the BIDI algorithm uses to determine the directionality of
2028          * paragraphs ( http://unicode.org/reports/tr9/#The_Paragraph_Level ) and
2029          * FSI isolates ( http://unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2030          *
2031          * TODO: Does not handle BIDI control characters inside the text.
2032          * TODO: Does not handle unallocated characters.
2033          *
2034          * @param string $text Text to test
2035          * @return null|string Directionality ('ltr' or 'rtl') or null
2036          */
2037         private static function strongDirFromContent( $text = '' ) {
2038                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2039                         return null;
2040                 }
2041                 if ( $matches[1] === '' ) {
2042                         return 'rtl';
2043                 }
2044                 return 'ltr';
2045         }
2046
2047         /**
2048          * Roman number formatting up to 10000
2049          *
2050          * @param int $num
2051          *
2052          * @return string
2053          */
2054         static function romanNumeral( $num ) {
2055                 static $table = array(
2056                         array( '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ),
2057                         array( '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ),
2058                         array( '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ),
2059                         array( '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2060                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' )
2061                 );
2062
2063                 $num = intval( $num );
2064                 if ( $num > 10000 || $num <= 0 ) {
2065                         return $num;
2066                 }
2067
2068                 $s = '';
2069                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2070                         if ( $num >= $pow10 ) {
2071                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2072                         }
2073                         $num = $num % $pow10;
2074                 }
2075                 return $s;
2076         }
2077
2078         /**
2079          * Hebrew Gematria number formatting up to 9999
2080          *
2081          * @param int $num
2082          *
2083          * @return string
2084          */
2085         static function hebrewNumeral( $num ) {
2086                 static $table = array(
2087                         array( '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ),
2088                         array( '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ),
2089                         array( '',
2090                                 array( 'ק' ),
2091                                 array( 'ר' ),
2092                                 array( 'ש' ),
2093                                 array( 'ת' ),
2094                                 array( 'ת', 'ק' ),
2095                                 array( 'ת', 'ר' ),
2096                                 array( 'ת', 'ש' ),
2097                                 array( 'ת', 'ת' ),
2098                                 array( 'ת', 'ת', 'ק' ),
2099                                 array( 'ת', 'ת', 'ר' ),
2100                         ),
2101                         array( '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' )
2102                 );
2103
2104                 $num = intval( $num );
2105                 if ( $num > 9999 || $num <= 0 ) {
2106                         return $num;
2107                 }
2108
2109                 // Round thousands have special notations
2110                 if ( $num === 1000 ) {
2111                         return "א' אלף";
2112                 } elseif ( $num % 1000 === 0 ) {
2113                         return $table[0][$num / 1000] . "' אלפים";
2114                 }
2115
2116                 $letters = array();
2117
2118                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2119                         if ( $num >= $pow10 ) {
2120                                 if ( $num === 15 || $num === 16 ) {
2121                                         $letters[] = $table[0][9];
2122                                         $letters[] = $table[0][$num - 9];
2123                                         $num = 0;
2124                                 } else {
2125                                         $letters = array_merge(
2126                                                 $letters,
2127                                                 (array)$table[$i][intval( $num / $pow10 )]
2128                                         );
2129
2130                                         if ( $pow10 === 1000 ) {
2131                                                 $letters[] = "'";
2132                                         }
2133                                 }
2134                         }
2135
2136                         $num = $num % $pow10;
2137                 }
2138
2139                 $preTransformLength = count( $letters );
2140                 if ( $preTransformLength === 1 ) {
2141                         // Add geresh (single quote) to one-letter numbers
2142                         $letters[] = "'";
2143                 } else {
2144                         $lastIndex = $preTransformLength - 1;
2145                         $letters[$lastIndex] = str_replace(
2146                                 array( 'כ', 'מ', 'נ', 'פ', 'צ' ),
2147                                 array( 'ך', 'ם', 'ן', 'ף', 'ץ' ),
2148                                 $letters[$lastIndex]
2149                         );
2150
2151                         // Add gershayim (double quote) to multiple-letter numbers,
2152                         // but exclude numbers with only one letter after the thousands
2153                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2154                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2155                                 $letters[] = "'";
2156                         } else {
2157                                 array_splice( $letters, -1, 0, '"' );
2158                         }
2159                 }
2160
2161                 return implode( $letters );
2162         }
2163
2164         /**
2165          * Used by date() and time() to adjust the time output.
2166          *
2167          * @param string $ts The time in date('YmdHis') format
2168          * @param mixed $tz Adjust the time by this amount (default false, mean we
2169          *   get user timecorrection setting)
2170          * @return int
2171          */
2172         function userAdjust( $ts, $tz = false ) {
2173                 global $wgUser, $wgLocalTZoffset;
2174
2175                 if ( $tz === false ) {
2176                         $tz = $wgUser->getOption( 'timecorrection' );
2177                 }
2178
2179                 $data = explode( '|', $tz, 3 );
2180
2181                 if ( $data[0] == 'ZoneInfo' ) {
2182                         MediaWiki\suppressWarnings();
2183                         $userTZ = timezone_open( $data[2] );
2184                         MediaWiki\restoreWarnings();
2185                         if ( $userTZ !== false ) {
2186                                 $date = date_create( $ts, timezone_open( 'UTC' ) );
2187                                 date_timezone_set( $date, $userTZ );
2188                                 $date = date_format( $date, 'YmdHis' );
2189                                 return $date;
2190                         }
2191                         # Unrecognized timezone, default to 'Offset' with the stored offset.
2192                         $data[0] = 'Offset';
2193                 }
2194
2195                 if ( $data[0] == 'System' || $tz == '' ) {
2196                         # Global offset in minutes.
2197                         $minDiff = $wgLocalTZoffset;
2198                 } elseif ( $data[0] == 'Offset' ) {
2199                         $minDiff = intval( $data[1] );
2200                 } else {
2201                         $data = explode( ':', $tz );
2202                         if ( count( $data ) == 2 ) {
2203                                 $data[0] = intval( $data[0] );
2204                                 $data[1] = intval( $data[1] );
2205                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2206                                 if ( $data[0] < 0 ) {
2207                                         $minDiff = -$minDiff;
2208                                 }
2209                         } else {
2210                                 $minDiff = intval( $data[0] ) * 60;
2211                         }
2212                 }
2213
2214                 # No difference ? Return time unchanged
2215                 if ( 0 == $minDiff ) {
2216                         return $ts;
2217                 }
2218
2219                 MediaWiki\suppressWarnings(); // E_STRICT system time bitching
2220                 # Generate an adjusted date; take advantage of the fact that mktime
2221                 # will normalize out-of-range values so we don't have to split $minDiff
2222                 # into hours and minutes.
2223                 $t = mktime( (
2224                         (int)substr( $ts, 8, 2 ) ), # Hours
2225                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2226                         (int)substr( $ts, 12, 2 ), # Seconds
2227                         (int)substr( $ts, 4, 2 ), # Month
2228                         (int)substr( $ts, 6, 2 ), # Day
2229                         (int)substr( $ts, 0, 4 ) ); # Year
2230
2231                 $date = date( 'YmdHis', $t );
2232                 MediaWiki\restoreWarnings();
2233
2234                 return $date;
2235         }
2236
2237         /**
2238          * This is meant to be used by time(), date(), and timeanddate() to get
2239          * the date preference they're supposed to use, it should be used in
2240          * all children.
2241          *
2242          *<code>
2243          * function timeanddate([...], $format = true) {
2244          *      $datePreference = $this->dateFormat($format);
2245          * [...]
2246          * }
2247          *</code>
2248          *
2249          * @param int|string|bool $usePrefs If true, the user's preference is used
2250          *   if false, the site/language default is used
2251          *   if int/string, assumed to be a format.
2252          * @return string
2253          */
2254         function dateFormat( $usePrefs = true ) {
2255                 global $wgUser;
2256
2257                 if ( is_bool( $usePrefs ) ) {
2258                         if ( $usePrefs ) {
2259                                 $datePreference = $wgUser->getDatePreference();
2260                         } else {
2261                                 $datePreference = (string)User::getDefaultOption( 'date' );
2262                         }
2263                 } else {
2264                         $datePreference = (string)$usePrefs;
2265                 }
2266
2267                 // return int
2268                 if ( $datePreference == '' ) {
2269                         return 'default';
2270                 }
2271
2272                 return $datePreference;
2273         }
2274
2275         /**
2276          * Get a format string for a given type and preference
2277          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2278          * @param string $pref The format name as it appears in Messages*.php under
2279          *  $datePreferences.
2280          *
2281          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2282          *
2283          * @return string
2284          */
2285         function getDateFormatString( $type, $pref ) {
2286                 $wasDefault = false;
2287                 if ( $pref == 'default' ) {
2288                         $wasDefault = true;
2289                         $pref = $this->getDefaultDateFormat();
2290                 }
2291
2292                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2293                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2294
2295                         if ( $type === 'pretty' && $df === null ) {
2296                                 $df = $this->getDateFormatString( 'date', $pref );
2297                         }
2298
2299                         if ( !$wasDefault && $df === null ) {
2300                                 $pref = $this->getDefaultDateFormat();
2301                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2302                         }
2303
2304                         $this->dateFormatStrings[$type][$pref] = $df;
2305                 }
2306                 return $this->dateFormatStrings[$type][$pref];
2307         }
2308
2309         /**
2310          * @param string $ts The time format which needs to be turned into a
2311          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2312          * @param bool $adj Whether to adjust the time output according to the
2313          *   user configured offset ($timecorrection)
2314          * @param mixed $format True to use user's date format preference
2315          * @param string|bool $timecorrection The time offset as returned by
2316          *   validateTimeZone() in Special:Preferences
2317          * @return string
2318          */
2319         function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2320                 $ts = wfTimestamp( TS_MW, $ts );
2321                 if ( $adj ) {
2322                         $ts = $this->userAdjust( $ts, $timecorrection );
2323                 }
2324                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2325                 return $this->sprintfDate( $df, $ts );
2326         }
2327
2328         /**
2329          * @param string $ts The time format which needs to be turned into a
2330          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2331          * @param bool $adj Whether to adjust the time output according to the
2332          *   user configured offset ($timecorrection)
2333          * @param mixed $format True to use user's date format preference
2334          * @param string|bool $timecorrection The time offset as returned by
2335          *   validateTimeZone() in Special:Preferences
2336          * @return string
2337          */
2338         function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2339                 $ts = wfTimestamp( TS_MW, $ts );
2340                 if ( $adj ) {
2341                         $ts = $this->userAdjust( $ts, $timecorrection );
2342                 }
2343                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2344                 return $this->sprintfDate( $df, $ts );
2345         }
2346
2347         /**
2348          * @param string $ts The time format which needs to be turned into a
2349          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2350          * @param bool $adj Whether to adjust the time output according to the
2351          *   user configured offset ($timecorrection)
2352          * @param mixed $format What format to return, if it's false output the
2353          *   default one (default true)
2354          * @param string|bool $timecorrection The time offset as returned by
2355          *   validateTimeZone() in Special:Preferences
2356          * @return string
2357          */
2358         function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2359                 $ts = wfTimestamp( TS_MW, $ts );
2360                 if ( $adj ) {
2361                         $ts = $this->userAdjust( $ts, $timecorrection );
2362                 }
2363                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2364                 return $this->sprintfDate( $df, $ts );
2365         }
2366
2367         /**
2368          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2369          *
2370          * @since 1.20
2371          *
2372          * @param int $seconds The amount of seconds.
2373          * @param array $chosenIntervals The intervals to enable.
2374          *
2375          * @return string
2376          */
2377         public function formatDuration( $seconds, array $chosenIntervals = array() ) {
2378                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2379
2380                 $segments = array();
2381
2382                 foreach ( $intervals as $intervalName => $intervalValue ) {
2383                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2384                         // duration-years, duration-decades, duration-centuries, duration-millennia
2385                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2386                         $segments[] = $message->inLanguage( $this )->escaped();
2387                 }
2388
2389                 return $this->listToText( $segments );
2390         }
2391
2392         /**
2393          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2394          * For example 65 will be turned into array( minutes => 1, seconds => 5 ).
2395          *
2396          * @since 1.20
2397          *
2398          * @param int $seconds The amount of seconds.
2399          * @param array $chosenIntervals The intervals to enable.
2400          *
2401          * @return array
2402          */
2403         public function getDurationIntervals( $seconds, array $chosenIntervals = array() ) {
2404                 if ( empty( $chosenIntervals ) ) {
2405                         $chosenIntervals = array(
2406                                 'millennia',
2407                                 'centuries',
2408                                 'decades',
2409                                 'years',
2410                                 'days',
2411                                 'hours',
2412                                 'minutes',
2413                                 'seconds'
2414                         );
2415                 }
2416
2417                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2418                 $sortedNames = array_keys( $intervals );
2419                 $smallestInterval = array_pop( $sortedNames );
2420
2421                 $segments = array();
2422
2423                 foreach ( $intervals as $name => $length ) {
2424                         $value = floor( $seconds / $length );
2425
2426                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2427                                 $seconds -= $value * $length;
2428                                 $segments[$name] = $value;
2429                         }
2430                 }
2431
2432                 return $segments;
2433         }
2434
2435         /**
2436          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2437          *
2438          * @param string $type Can be 'date', 'time' or 'both'
2439          * @param string $ts The time format which needs to be turned into a
2440          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2441          * @param User $user User object used to get preferences for timezone and format
2442          * @param array $options Array, can contain the following keys:
2443          *   - 'timecorrection': time correction, can have the following values:
2444          *     - true: use user's preference
2445          *     - false: don't use time correction
2446          *     - int: value of time correction in minutes
2447          *   - 'format': format to use, can have the following values:
2448          *     - true: use user's preference
2449          *     - false: use default preference
2450          *     - string: format to use
2451          * @since 1.19
2452          * @return string
2453          */
2454         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2455                 $ts = wfTimestamp( TS_MW, $ts );
2456                 $options += array( 'timecorrection' => true, 'format' => true );
2457                 if ( $options['timecorrection'] !== false ) {
2458                         if ( $options['timecorrection'] === true ) {
2459                                 $offset = $user->getOption( 'timecorrection' );
2460                         } else {
2461                                 $offset = $options['timecorrection'];
2462                         }
2463                         $ts = $this->userAdjust( $ts, $offset );
2464                 }
2465                 if ( $options['format'] === true ) {
2466                         $format = $user->getDatePreference();
2467                 } else {
2468                         $format = $options['format'];
2469                 }
2470                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2471                 return $this->sprintfDate( $df, $ts );
2472         }
2473
2474         /**
2475          * Get the formatted date for the given timestamp and formatted for
2476          * the given user.
2477          *
2478          * @param mixed $ts Mixed: the time format which needs to be turned into a
2479          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2480          * @param User $user User object used to get preferences for timezone and format
2481          * @param array $options Array, can contain the following keys:
2482          *   - 'timecorrection': time correction, can have the following values:
2483          *     - true: use user's preference
2484          *     - false: don't use time correction
2485          *     - int: value of time correction in minutes
2486          *   - 'format': format to use, can have the following values:
2487          *     - true: use user's preference
2488          *     - false: use default preference
2489          *     - string: format to use
2490          * @since 1.19
2491          * @return string
2492          */
2493         public function userDate( $ts, User $user, array $options = array() ) {
2494                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2495         }
2496
2497         /**
2498          * Get the formatted time for the given timestamp and formatted for
2499          * the given user.
2500          *
2501          * @param mixed $ts The time format which needs to be turned into a
2502          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2503          * @param User $user User object used to get preferences for timezone and format
2504          * @param array $options Array, can contain the following keys:
2505          *   - 'timecorrection': time correction, can have the following values:
2506          *     - true: use user's preference
2507          *     - false: don't use time correction
2508          *     - int: value of time correction in minutes
2509          *   - 'format': format to use, can have the following values:
2510          *     - true: use user's preference
2511          *     - false: use default preference
2512          *     - string: format to use
2513          * @since 1.19
2514          * @return string
2515          */
2516         public function userTime( $ts, User $user, array $options = array() ) {
2517                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2518         }
2519
2520         /**
2521          * Get the formatted date and time for the given timestamp and formatted for
2522          * the given user.
2523          *
2524          * @param mixed $ts The time format which needs to be turned into a
2525          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2526          * @param User $user User object used to get preferences for timezone and format
2527          * @param array $options Array, can contain the following keys:
2528          *   - 'timecorrection': time correction, can have the following values:
2529          *     - true: use user's preference
2530          *     - false: don't use time correction
2531          *     - int: value of time correction in minutes
2532          *   - 'format': format to use, can have the following values:
2533          *     - true: use user's preference
2534          *     - false: use default preference
2535          *     - string: format to use
2536          * @since 1.19
2537          * @return string
2538          */
2539         public function userTimeAndDate( $ts, User $user, array $options = array() ) {
2540                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2541         }
2542
2543         /**
2544          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2545          *
2546          * Determine the difference between the timestamp and the current time, and
2547          * generate a readable timestamp by returning "<N> <units> ago", where the
2548          * largest possible unit is used.
2549          *
2550          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2551          *
2552          * @param MWTimestamp $time
2553          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2554          * @param User|null $user User the timestamp is being generated for
2555          *  (or null to use main context's user)
2556          * @return string Formatted timestamp
2557          */
2558         public function getHumanTimestamp(
2559                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2560         ) {
2561                 if ( $relativeTo === null ) {
2562                         $relativeTo = new MWTimestamp();
2563                 }
2564                 if ( $user === null ) {
2565                         $user = RequestContext::getMain()->getUser();
2566                 }
2567
2568                 // Adjust for the user's timezone.
2569                 $offsetThis = $time->offsetForUser( $user );
2570                 $offsetRel = $relativeTo->offsetForUser( $user );
2571
2572                 $ts = '';
2573                 if ( Hooks::run( 'GetHumanTimestamp', array( &$ts, $time, $relativeTo, $user, $this ) ) ) {
2574                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2575                 }
2576
2577                 // Reset the timezone on the objects.
2578                 $time->timestamp->sub( $offsetThis );
2579                 $relativeTo->timestamp->sub( $offsetRel );
2580
2581                 return $ts;
2582         }
2583
2584         /**
2585          * Convert an MWTimestamp into a pretty human-readable timestamp using
2586          * the given user preferences and relative base time.
2587          *
2588          * @see Language::getHumanTimestamp
2589          * @param MWTimestamp $ts Timestamp to prettify
2590          * @param MWTimestamp $relativeTo Base timestamp
2591          * @param User $user User preferences to use
2592          * @return string Human timestamp
2593          * @since 1.26
2594          */
2595         private function getHumanTimestampInternal(
2596                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2597         ) {
2598                 $diff = $ts->diff( $relativeTo );
2599                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2600                         (int)$relativeTo->timestamp->format( 'w' ) );
2601                 $days = $diff->days ?: (int)$diffDay;
2602                 if ( $diff->invert || $days > 5
2603                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2604                 ) {
2605                         // Timestamps are in different years: use full timestamp
2606                         // Also do full timestamp for future dates
2607                         /**
2608                          * @todo FIXME: Add better handling of future timestamps.
2609                          */
2610                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2611                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2612                 } elseif ( $days > 5 ) {
2613                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2614                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2615                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2616                 } elseif ( $days > 1 ) {
2617                         // Timestamp within the past week: show the day of the week and time
2618                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2619                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2620                         // Messages:
2621                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2622                         $ts = wfMessage( "$weekday-at" )
2623                                 ->inLanguage( $this )
2624                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2625                                 ->text();
2626                 } elseif ( $days == 1 ) {
2627                         // Timestamp was yesterday: say 'yesterday' and the time.
2628                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2629                         $ts = wfMessage( 'yesterday-at' )
2630                                 ->inLanguage( $this )
2631                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2632                                 ->text();
2633                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2634                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2635                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2636                         $ts = wfMessage( 'today-at' )
2637                                 ->inLanguage( $this )
2638                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2639                                 ->text();
2640
2641                 // From here on in, the timestamp was soon enough ago so that we can simply say
2642                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2643                 } elseif ( $diff->h == 1 ) {
2644                         // Less than 90 minutes, but more than an hour ago.
2645                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2646                 } elseif ( $diff->i >= 1 ) {
2647                         // A few minutes ago.
2648                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2649                 } elseif ( $diff->s >= 30 ) {
2650                         // Less than a minute, but more than 30 sec ago.
2651                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2652                 } else {
2653                         // Less than 30 seconds ago.
2654                         $ts = wfMessage( 'just-now' )->text();
2655                 }
2656
2657                 return $ts;
2658         }
2659
2660         /**
2661          * @param string $key
2662          * @return array|null
2663          */
2664         function getMessage( $key ) {
2665                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2666         }
2667
2668         /**
2669          * @return array
2670          */
2671         function getAllMessages() {
2672                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2673         }
2674
2675         /**
2676          * @param string $in
2677          * @param string $out
2678          * @param string $string
2679          * @return string
2680          */
2681         function iconv( $in, $out, $string ) {
2682                 # This is a wrapper for iconv in all languages except esperanto,
2683                 # which does some nasty x-conversions beforehand
2684
2685                 # Even with //IGNORE iconv can whine about illegal characters in
2686                 # *input* string. We just ignore those too.
2687                 # REF: http://bugs.php.net/bug.php?id=37166
2688                 # REF: https://phabricator.wikimedia.org/T18885
2689                 MediaWiki\suppressWarnings();
2690                 $text = iconv( $in, $out . '//IGNORE', $string );
2691                 MediaWiki\restoreWarnings();
2692                 return $text;
2693         }
2694
2695         // callback functions for uc(), lc(), ucwords(), ucwordbreaks()
2696
2697         /**
2698          * @param array $matches
2699          * @return mixed|string
2700          */
2701         function ucwordbreaksCallbackAscii( $matches ) {
2702                 return $this->ucfirst( $matches[1] );
2703         }
2704
2705         /**
2706          * @param array $matches
2707          * @return string
2708          */
2709         function ucwordbreaksCallbackMB( $matches ) {
2710                 return mb_strtoupper( $matches[0] );
2711         }
2712
2713         /**
2714          * @param array $matches
2715          * @return string
2716          */
2717         function ucCallback( $matches ) {
2718                 list( $wikiUpperChars ) = self::getCaseMaps();
2719                 return strtr( $matches[1], $wikiUpperChars );
2720         }
2721
2722         /**
2723          * @param array $matches
2724          * @return string
2725          */
2726         function lcCallback( $matches ) {
2727                 list( , $wikiLowerChars ) = self::getCaseMaps();
2728                 return strtr( $matches[1], $wikiLowerChars );
2729         }
2730
2731         /**
2732          * @param array $matches
2733          * @return string
2734          */
2735         function ucwordsCallbackMB( $matches ) {
2736                 return mb_strtoupper( $matches[0] );
2737         }
2738
2739         /**
2740          * @param array $matches
2741          * @return string
2742          */
2743         function ucwordsCallbackWiki( $matches ) {
2744                 list( $wikiUpperChars ) = self::getCaseMaps();
2745                 return strtr( $matches[0], $wikiUpperChars );
2746         }
2747
2748         /**
2749          * Make a string's first character uppercase
2750          *
2751          * @param string $str
2752          *
2753          * @return string
2754          */
2755         function ucfirst( $str ) {
2756                 $o = ord( $str );
2757                 if ( $o < 96 ) { // if already uppercase...
2758                         return $str;
2759                 } elseif ( $o < 128 ) {
2760                         return ucfirst( $str ); // use PHP's ucfirst()
2761                 } else {
2762                         // fall back to more complex logic in case of multibyte strings
2763                         return $this->uc( $str, true );
2764                 }
2765         }
2766
2767         /**
2768          * Convert a string to uppercase
2769          *
2770          * @param string $str
2771          * @param bool $first
2772          *
2773          * @return string
2774          */
2775         function uc( $str, $first = false ) {
2776                 if ( function_exists( 'mb_strtoupper' ) ) {
2777                         if ( $first ) {
2778                                 if ( $this->isMultibyte( $str ) ) {
2779                                         return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2780                                 } else {
2781                                         return ucfirst( $str );
2782                                 }
2783                         } else {
2784                                 return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2785                         }
2786                 } else {
2787                         if ( $this->isMultibyte( $str ) ) {
2788                                 $x = $first ? '^' : '';
2789                                 return preg_replace_callback(
2790                                         "/$x([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/",
2791                                         array( $this, 'ucCallback' ),
2792                                         $str
2793                                 );
2794                         } else {
2795                                 return $first ? ucfirst( $str ) : strtoupper( $str );
2796                         }
2797                 }
2798         }
2799
2800         /**
2801          * @param string $str
2802          * @return mixed|string
2803          */
2804         function lcfirst( $str ) {
2805                 $o = ord( $str );
2806                 if ( !$o ) {
2807                         return strval( $str );
2808                 } elseif ( $o >= 128 ) {
2809                         return $this->lc( $str, true );
2810                 } elseif ( $o > 96 ) {
2811                         return $str;
2812                 } else {
2813                         $str[0] = strtolower( $str[0] );
2814                         return $str;
2815                 }
2816         }
2817
2818         /**
2819          * @param string $str
2820          * @param bool $first
2821          * @return mixed|string
2822          */
2823         function lc( $str, $first = false ) {
2824                 if ( function_exists( 'mb_strtolower' ) ) {
2825                         if ( $first ) {
2826                                 if ( $this->isMultibyte( $str ) ) {
2827                                         return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2828                                 } else {
2829                                         return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2830                                 }
2831                         } else {
2832                                 return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2833                         }
2834                 } else {
2835                         if ( $this->isMultibyte( $str ) ) {
2836                                 $x = $first ? '^' : '';
2837                                 return preg_replace_callback(
2838                                         "/$x([A-Z]|[\\xc0-\\xff][\\x80-\\xbf]*)/",
2839                                         array( $this, 'lcCallback' ),
2840                                         $str
2841                                 );
2842                         } else {
2843                                 return $first ? strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 ) : strtolower( $str );
2844                         }
2845                 }
2846         }
2847
2848         /**
2849          * @param string $str
2850          * @return bool
2851          */
2852         function isMultibyte( $str ) {
2853                 return strlen( $str ) !== mb_strlen( $str );
2854         }
2855
2856         /**
2857          * @param string $str
2858          * @return mixed|string
2859          */
2860         function ucwords( $str ) {
2861                 if ( $this->isMultibyte( $str ) ) {
2862                         $str = $this->lc( $str );
2863
2864                         // regexp to find first letter in each word (i.e. after each space)
2865                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2866
2867                         // function to use to capitalize a single char
2868                         if ( function_exists( 'mb_strtoupper' ) ) {
2869                                 return preg_replace_callback(
2870                                         $replaceRegexp,
2871                                         array( $this, 'ucwordsCallbackMB' ),
2872                                         $str
2873                                 );
2874                         } else {
2875                                 return preg_replace_callback(
2876                                         $replaceRegexp,
2877                                         array( $this, 'ucwordsCallbackWiki' ),
2878                                         $str
2879                                 );
2880                         }
2881                 } else {
2882                         return ucwords( strtolower( $str ) );
2883                 }
2884         }
2885
2886         /**
2887          * capitalize words at word breaks
2888          *
2889          * @param string $str
2890          * @return mixed
2891          */
2892         function ucwordbreaks( $str ) {
2893                 if ( $this->isMultibyte( $str ) ) {
2894                         $str = $this->lc( $str );
2895
2896                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2897                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2898
2899                         // find first letter after word break
2900                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2901                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2902
2903                         if ( function_exists( 'mb_strtoupper' ) ) {
2904                                 return preg_replace_callback(
2905                                         $replaceRegexp,
2906                                         array( $this, 'ucwordbreaksCallbackMB' ),
2907                                         $str
2908                                 );
2909                         } else {
2910                                 return preg_replace_callback(
2911                                         $replaceRegexp,
2912                                         array( $this, 'ucwordsCallbackWiki' ),
2913                                         $str
2914                                 );
2915                         }
2916                 } else {
2917                         return preg_replace_callback(
2918                                 '/\b([\w\x80-\xff]+)\b/',
2919                                 array( $this, 'ucwordbreaksCallbackAscii' ),
2920                                 $str
2921                         );
2922                 }
2923         }
2924
2925         /**
2926          * Return a case-folded representation of $s
2927          *
2928          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2929          * and $s2 are the same except for the case of their characters. It is not
2930          * necessary for the value returned to make sense when displayed.
2931          *
2932          * Do *not* perform any other normalisation in this function. If a caller
2933          * uses this function when it should be using a more general normalisation
2934          * function, then fix the caller.
2935          *
2936          * @param string $s
2937          *
2938          * @return string
2939          */
2940         function caseFold( $s ) {
2941                 return $this->uc( $s );
2942         }
2943
2944         /**
2945          * @param string $s
2946          * @return string
2947          */
2948         function checkTitleEncoding( $s ) {
2949                 if ( is_array( $s ) ) {
2950                         throw new MWException( 'Given array to checkTitleEncoding.' );
2951                 }
2952                 if ( StringUtils::isUtf8( $s ) ) {
2953                         return $s;
2954                 }
2955
2956                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2957         }
2958
2959         /**
2960          * @return array
2961          */
2962         function fallback8bitEncoding() {
2963                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2964         }
2965
2966         /**
2967          * Most writing systems use whitespace to break up words.
2968          * Some languages such as Chinese don't conventionally do this,
2969          * which requires special handling when breaking up words for
2970          * searching etc.
2971          *
2972          * @return bool
2973          */
2974         function hasWordBreaks() {
2975                 return true;
2976         }
2977
2978         /**
2979          * Some languages such as Chinese require word segmentation,
2980          * Specify such segmentation when overridden in derived class.
2981          *
2982          * @param string $string
2983          * @return string
2984          */
2985         function segmentByWord( $string ) {
2986                 return $string;
2987         }
2988
2989         /**
2990          * Some languages have special punctuation need to be normalized.
2991          * Make such changes here.
2992          *
2993          * @param string $string
2994          * @return string
2995          */
2996         function normalizeForSearch( $string ) {
2997                 return self::convertDoubleWidth( $string );
2998         }
2999
3000         /**
3001          * convert double-width roman characters to single-width.
3002          * range: ff00-ff5f ~= 0020-007f
3003          *
3004          * @param string $string
3005          *
3006          * @return string
3007          */
3008         protected static function convertDoubleWidth( $string ) {
3009                 static $full = null;
3010                 static $half = null;
3011
3012                 if ( $full === null ) {
3013                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
3014                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
3015                         $full = str_split( $fullWidth, 3 );
3016                         $half = str_split( $halfWidth );
3017                 }
3018
3019                 $string = str_replace( $full, $half, $string );
3020                 return $string;
3021         }
3022
3023         /**
3024          * @param string $string
3025          * @param string $pattern
3026          * @return string
3027          */
3028         protected static function insertSpace( $string, $pattern ) {
3029                 $string = preg_replace( $pattern, " $1 ", $string );
3030                 $string = preg_replace( '/ +/', ' ', $string );
3031                 return $string;
3032         }
3033
3034         /**
3035          * @param array $termsArray
3036          * @return array
3037          */
3038         function convertForSearchResult( $termsArray ) {
3039                 # some languages, e.g. Chinese, need to do a conversion
3040                 # in order for search results to be displayed correctly
3041                 return $termsArray;
3042         }
3043
3044         /**
3045          * Get the first character of a string.
3046          *
3047          * @param string $s
3048          * @return string
3049          */
3050         function firstChar( $s ) {
3051                 $matches = array();
3052                 preg_match(
3053                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
3054                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
3055                         $s,
3056                         $matches
3057                 );
3058
3059                 if ( isset( $matches[1] ) ) {
3060                         if ( strlen( $matches[1] ) != 3 ) {
3061                                 return $matches[1];
3062                         }
3063
3064                         // Break down Hangul syllables to grab the first jamo
3065                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
3066                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
3067                                 return $matches[1];
3068                         } elseif ( $code < 0xb098 ) {
3069                                 return "\xe3\x84\xb1";
3070                         } elseif ( $code < 0xb2e4 ) {
3071                                 return "\xe3\x84\xb4";
3072                         } elseif ( $code < 0xb77c ) {
3073                                 return "\xe3\x84\xb7";
3074                         } elseif ( $code < 0xb9c8 ) {
3075                                 return "\xe3\x84\xb9";
3076                         } elseif ( $code < 0xbc14 ) {
3077                                 return "\xe3\x85\x81";
3078                         } elseif ( $code < 0xc0ac ) {
3079                                 return "\xe3\x85\x82";
3080                         } elseif ( $code < 0xc544 ) {
3081                                 return "\xe3\x85\x85";
3082                         } elseif ( $code < 0xc790 ) {
3083                                 return "\xe3\x85\x87";
3084                         } elseif ( $code < 0xcc28 ) {
3085                                 return "\xe3\x85\x88";
3086                         } elseif ( $code < 0xce74 ) {
3087                                 return "\xe3\x85\x8a";
3088                         } elseif ( $code < 0xd0c0 ) {
3089                                 return "\xe3\x85\x8b";
3090                         } elseif ( $code < 0xd30c ) {
3091                                 return "\xe3\x85\x8c";
3092                         } elseif ( $code < 0xd558 ) {
3093                                 return "\xe3\x85\x8d";
3094                         } else {
3095                                 return "\xe3\x85\x8e";
3096                         }
3097                 } else {
3098                         return '';
3099                 }
3100         }
3101
3102         function initEncoding() {
3103                 # Some languages may have an alternate char encoding option
3104                 # (Esperanto X-coding, Japanese furigana conversion, etc)
3105                 # If this language is used as the primary content language,
3106                 # an override to the defaults can be set here on startup.
3107         }
3108
3109         /**
3110          * @param string $s
3111          * @return string
3112          */
3113         function recodeForEdit( $s ) {
3114                 # For some languages we'll want to explicitly specify
3115                 # which characters make it into the edit box raw
3116                 # or are converted in some way or another.
3117                 global $wgEditEncoding;
3118                 if ( $wgEditEncoding == '' || $wgEditEncoding == 'UTF-8' ) {
3119                         return $s;
3120                 } else {
3121                         return $this->iconv( 'UTF-8', $wgEditEncoding, $s );
3122                 }
3123         }
3124
3125         /**
3126          * @param string $s
3127          * @return string
3128          */
3129         function recodeInput( $s ) {
3130                 # Take the previous into account.
3131                 global $wgEditEncoding;
3132                 if ( $wgEditEncoding != '' ) {
3133                         $enc = $wgEditEncoding;
3134                 } else {
3135                         $enc = 'UTF-8';
3136                 }
3137                 if ( $enc == 'UTF-8' ) {
3138                         return $s;
3139                 } else {
3140                         return $this->iconv( $enc, 'UTF-8', $s );
3141                 }
3142         }
3143
3144         /**
3145          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3146          * also cleans up certain backwards-compatible sequences, converting them
3147          * to the modern Unicode equivalent.
3148          *
3149          * This is language-specific for performance reasons only.
3150          *
3151          * @param string $s
3152          *
3153          * @return string
3154          */
3155         function normalize( $s ) {
3156                 global $wgAllUnicodeFixes;
3157                 $s = UtfNormal\Validator::cleanUp( $s );
3158                 if ( $wgAllUnicodeFixes ) {
3159                         $s = $this->transformUsingPairFile( 'normalize-ar.ser', $s );
3160                         $s = $this->transformUsingPairFile( 'normalize-ml.ser', $s );
3161                 }
3162
3163                 return $s;
3164         }
3165
3166         /**
3167          * Transform a string using serialized data stored in the given file (which
3168          * must be in the serialized subdirectory of $IP). The file contains pairs
3169          * mapping source characters to destination characters.
3170          *
3171          * The data is cached in process memory. This will go faster if you have the
3172          * FastStringSearch extension.
3173          *
3174          * @param string $file
3175          * @param string $string
3176          *
3177          * @throws MWException
3178          * @return string
3179          */
3180         function transformUsingPairFile( $file, $string ) {
3181                 if ( !isset( $this->transformData[$file] ) ) {
3182                         $data = wfGetPrecompiledData( $file );
3183                         if ( $data === false ) {
3184                                 throw new MWException( __METHOD__ . ": The transformation file $file is missing" );
3185                         }
3186                         $this->transformData[$file] = new ReplacementArray( $data );
3187                 }
3188                 return $this->transformData[$file]->replace( $string );
3189         }
3190
3191         /**
3192          * For right-to-left language support
3193          *
3194          * @return bool
3195          */
3196         function isRTL() {
3197                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3198         }
3199
3200         /**
3201          * Return the correct HTML 'dir' attribute value for this language.
3202          * @return string
3203          */
3204         function getDir() {
3205                 return $this->isRTL() ? 'rtl' : 'ltr';
3206         }
3207
3208         /**
3209          * Return 'left' or 'right' as appropriate alignment for line-start
3210          * for this language's text direction.
3211          *
3212          * Should be equivalent to CSS3 'start' text-align value....
3213          *
3214          * @return string
3215          */
3216         function alignStart() {
3217                 return $this->isRTL() ? 'right' : 'left';
3218         }
3219
3220         /**
3221          * Return 'right' or 'left' as appropriate alignment for line-end
3222          * for this language's text direction.
3223          *
3224          * Should be equivalent to CSS3 'end' text-align value....
3225          *
3226          * @return string
3227          */
3228         function alignEnd() {
3229                 return $this->isRTL() ? 'left' : 'right';
3230         }
3231
3232         /**
3233          * A hidden direction mark (LRM or RLM), depending on the language direction.
3234          * Unlike getDirMark(), this function returns the character as an HTML entity.
3235          * This function should be used when the output is guaranteed to be HTML,
3236          * because it makes the output HTML source code more readable. When
3237          * the output is plain text or can be escaped, getDirMark() should be used.
3238          *
3239          * @param bool $opposite Get the direction mark opposite to your language
3240          * @return string
3241          * @since 1.20
3242          */
3243         function getDirMarkEntity( $opposite = false ) {
3244                 if ( $opposite ) {
3245                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3246                 }
3247                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3248         }
3249
3250         /**
3251          * A hidden direction mark (LRM or RLM), depending on the language direction.
3252          * This function produces them as invisible Unicode characters and
3253          * the output may be hard to read and debug, so it should only be used
3254          * when the output is plain text or can be escaped. When the output is
3255          * HTML, use getDirMarkEntity() instead.
3256          *
3257          * @param bool $opposite Get the direction mark opposite to your language
3258          * @return string
3259          */
3260         function getDirMark( $opposite = false ) {
3261                 $lrm = "\xE2\x80\x8E"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3262                 $rlm = "\xE2\x80\x8F"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3263                 if ( $opposite ) {
3264                         return $this->isRTL() ? $lrm : $rlm;
3265                 }
3266                 return $this->isRTL() ? $rlm : $lrm;
3267         }
3268
3269         /**
3270          * @return array
3271          */
3272         function capitalizeAllNouns() {
3273                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3274         }
3275
3276         /**
3277          * An arrow, depending on the language direction.
3278          *
3279          * @param string $direction The direction of the arrow: forwards (default),
3280          *   backwards, left, right, up, down.
3281          * @return string
3282          */
3283         function getArrow( $direction = 'forwards' ) {
3284                 switch ( $direction ) {
3285                 case 'forwards':
3286                         return $this->isRTL() ? '←' : '→';
3287                 case 'backwards':
3288                         return $this->isRTL() ? '→' : '←';
3289                 case 'left':
3290                         return '←';
3291                 case 'right':
3292                         return '→';
3293                 case 'up':
3294                         return '↑';
3295                 case 'down':
3296                         return '↓';
3297                 }
3298         }
3299
3300         /**
3301          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3302          *
3303          * @return bool
3304          */
3305         function linkPrefixExtension() {
3306                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3307         }
3308
3309         /**
3310          * Get all magic words from cache.
3311          * @return array
3312          */
3313         function getMagicWords() {
3314                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3315         }
3316
3317         /**
3318          * Run the LanguageGetMagic hook once.
3319          */
3320         protected function doMagicHook() {
3321                 if ( $this->mMagicHookDone ) {
3322                         return;
3323                 }
3324                 $this->mMagicHookDone = true;
3325                 Hooks::run( 'LanguageGetMagic', array( &$this->mMagicExtensions, $this->getCode() ) );
3326         }
3327
3328         /**
3329          * Fill a MagicWord object with data from here
3330          *
3331          * @param MagicWord $mw
3332          */
3333         function getMagic( $mw ) {
3334                 // Saves a function call
3335                 if ( !$this->mMagicHookDone ) {
3336                         $this->doMagicHook();
3337                 }
3338
3339                 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
3340                         $rawEntry = $this->mMagicExtensions[$mw->mId];
3341                 } else {
3342                         $rawEntry = self::$dataCache->getSubitem(
3343                                 $this->mCode, 'magicWords', $mw->mId );
3344                 }
3345
3346                 if ( !is_array( $rawEntry ) ) {
3347                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3348                 } else {
3349                         $mw->mCaseSensitive = $rawEntry[0];
3350                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3351                 }
3352         }
3353
3354         /**
3355          * Add magic words to the extension array
3356          *
3357          * @param array $newWords
3358          */
3359         function addMagicWordsByLang( $newWords ) {
3360                 $fallbackChain = $this->getFallbackLanguages();
3361                 $fallbackChain = array_reverse( $fallbackChain );
3362                 foreach ( $fallbackChain as $code ) {
3363                         if ( isset( $newWords[$code] ) ) {
3364                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3365                         }
3366                 }
3367         }
3368
3369         /**
3370          * Get special page names, as an associative array
3371          *   canonical name => array of valid names, including aliases
3372          * @return array
3373          */
3374         function getSpecialPageAliases() {
3375                 // Cache aliases because it may be slow to load them
3376                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3377                         // Initialise array
3378                         $this->mExtendedSpecialPageAliases =
3379                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3380                         Hooks::run( 'LanguageGetSpecialPageAliases',
3381                                 array( &$this->mExtendedSpecialPageAliases, $this->getCode() ) );
3382                 }
3383
3384                 return $this->mExtendedSpecialPageAliases;
3385         }
3386
3387         /**
3388          * Italic is unsuitable for some languages
3389          *
3390          * @param string $text The text to be emphasized.
3391          * @return string
3392          */
3393         function emphasize( $text ) {
3394                 return "<em>$text</em>";
3395         }
3396
3397         /**
3398          * Normally we output all numbers in plain en_US style, that is
3399          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3400          * point twohundredthirtyfive. However this is not suitable for all
3401          * languages, some such as Punjabi want ੨੯੩,੨੯੫.੨੩੫ and others such as
3402          * Icelandic just want to use commas instead of dots, and dots instead
3403          * of commas like "293.291,235".
3404          *
3405          * An example of this function being called:
3406          * <code>
3407          * wfMessage( 'message' )->numParams( $num )->text()
3408          * </code>
3409          *
3410          * See $separatorTransformTable on MessageIs.php for
3411          * the , => . and . => , implementation.
3412          *
3413          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3414          * @param int|float $number The string to be formatted, should be an integer
3415          *   or a floating point number.
3416          * @param bool $nocommafy Set to true for special numbers like dates
3417          * @return string
3418          */
3419         public function formatNum( $number, $nocommafy = false ) {
3420                 global $wgTranslateNumerals;
3421                 if ( !$nocommafy ) {
3422                         $number = $this->commafy( $number );
3423                         $s = $this->separatorTransformTable();
3424                         if ( $s ) {
3425                                 $number = strtr( $number, $s );
3426                         }
3427                 }
3428
3429                 if ( $wgTranslateNumerals ) {
3430                         $s = $this->digitTransformTable();
3431                         if ( $s ) {
3432                                 $number = strtr( $number, $s );
3433                         }
3434                 }
3435
3436                 return $number;
3437         }
3438
3439         /**
3440          * Front-end for non-commafied formatNum
3441          *
3442          * @param int|float $number The string to be formatted, should be an integer
3443          *        or a floating point number.
3444          * @since 1.21
3445          * @return string
3446          */
3447         public function formatNumNoSeparators( $number ) {
3448                 return $this->formatNum( $number, true );
3449         }
3450
3451         /**
3452          * @param string $number
3453          * @return string
3454          */
3455         public function parseFormattedNumber( $number ) {
3456                 $s = $this->digitTransformTable();
3457                 if ( $s ) {
3458                         // eliminate empty array values such as ''. (bug 64347)
3459                         $s = array_filter( $s );
3460                         $number = strtr( $number, array_flip( $s ) );
3461                 }
3462
3463                 $s = $this->separatorTransformTable();
3464                 if ( $s ) {
3465                         // eliminate empty array values such as ''. (bug 64347)
3466                         $s = array_filter( $s );
3467                         $number = strtr( $number, array_flip( $s ) );
3468                 }
3469
3470                 $number = strtr( $number, array( ',' => '' ) );
3471                 return $number;
3472         }
3473
3474         /**
3475          * Adds commas to a given number
3476          * @since 1.19
3477          * @param mixed $number
3478          * @return string
3479          */
3480         function commafy( $number ) {
3481                 $digitGroupingPattern = $this->digitGroupingPattern();
3482                 if ( $number === null ) {
3483                         return '';
3484                 }
3485
3486                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3487                         // default grouping is at thousands,  use the same for ###,###,### pattern too.
3488                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3489                 } else {
3490                         // Ref: http://cldr.unicode.org/translation/number-patterns
3491                         $sign = "";
3492                         if ( intval( $number ) < 0 ) {
3493                                 // For negative numbers apply the algorithm like positive number and add sign.
3494                                 $sign = "-";
3495                                 $number = substr( $number, 1 );
3496                         }
3497                         $integerPart = array();
3498                         $decimalPart = array();
3499                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3500                         preg_match( "/\d+/", $number, $integerPart );
3501                         preg_match( "/\.\d*/", $number, $decimalPart );
3502                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3503                         if ( $groupedNumber === $number ) {
3504                                 // the string does not have any number part. Eg: .12345
3505                                 return $sign . $groupedNumber;
3506                         }
3507                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3508                         while ( $start > 0 ) {
3509                                 $match = $matches[0][$numMatches - 1];
3510                                 $matchLen = strlen( $match );
3511                                 $start = $end - $matchLen;
3512                                 if ( $start < 0 ) {
3513                                         $start = 0;
3514                                 }
3515                                 $groupedNumber = substr( $number, $start, $end -$start ) . $groupedNumber;
3516                                 $end = $start;
3517                                 if ( $numMatches > 1 ) {
3518                                         // use the last pattern for the rest of the number
3519                                         $numMatches--;
3520                                 }
3521                                 if ( $start > 0 ) {
3522                                         $groupedNumber = "," . $groupedNumber;
3523                                 }
3524                         }
3525                         return $sign . $groupedNumber;
3526                 }
3527         }
3528
3529         /**
3530          * @return string
3531          */
3532         function digitGroupingPattern() {
3533                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3534         }
3535
3536         /**
3537          * @return array
3538          */
3539         function digitTransformTable() {
3540                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3541         }
3542
3543         /**
3544          * @return array
3545          */
3546         function separatorTransformTable() {
3547                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3548         }
3549
3550         /**
3551          * Take a list of strings and build a locale-friendly comma-separated
3552          * list, using the local comma-separator message.
3553          * The last two strings are chained with an "and".
3554          * NOTE: This function will only work with standard numeric array keys (0, 1, 2…)
3555          *
3556          * @param string[] $l
3557          * @return string
3558          */
3559         function listToText( array $l ) {
3560                 $m = count( $l ) - 1;
3561                 if ( $m < 0 ) {
3562                         return '';
3563                 }
3564                 if ( $m > 0 ) {
3565                         $and = $this->msg( 'and' )->escaped();
3566                         $space = $this->msg( 'word-separator' )->escaped();
3567                         if ( $m > 1 ) {
3568                                 $comma = $this->msg( 'comma-separator' )->escaped();
3569                         }
3570                 }
3571                 $s = $l[$m];
3572                 for ( $i = $m - 1; $i >= 0; $i-- ) {
3573                         if ( $i == $m - 1 ) {
3574                                 $s = $l[$i] . $and . $space . $s;
3575                         } else {
3576                                 $s = $l[$i] . $comma . $s;
3577                         }
3578                 }
3579                 return $s;
3580         }
3581
3582         /**
3583          * Take a list of strings and build a locale-friendly comma-separated
3584          * list, using the local comma-separator message.
3585          * @param string[] $list Array of strings to put in a comma list
3586          * @return string
3587          */
3588         function commaList( array $list ) {
3589                 return implode(
3590                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3591                         $list
3592                 );
3593         }
3594
3595         /**
3596          * Take a list of strings and build a locale-friendly semicolon-separated
3597          * list, using the local semicolon-separator message.
3598          * @param string[] $list Array of strings to put in a semicolon list
3599          * @return string
3600          */
3601         function semicolonList( array $list ) {
3602                 return implode(
3603                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3604                         $list
3605                 );
3606         }
3607
3608         /**
3609          * Same as commaList, but separate it with the pipe instead.
3610          * @param string[] $list Array of strings to put in a pipe list
3611          * @return string
3612          */
3613         function pipeList( array $list ) {
3614                 return implode(
3615                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3616                         $list
3617                 );
3618         }
3619
3620         /**
3621          * Truncate a string to a specified length in bytes, appending an optional
3622          * string (e.g. for ellipses)
3623          *
3624          * The database offers limited byte lengths for some columns in the database;
3625          * multi-byte character sets mean we need to ensure that only whole characters
3626          * are included, otherwise broken characters can be passed to the user
3627          *
3628          * If $length is negative, the string will be truncated from the beginning
3629          *
3630          * @param string $string String to truncate
3631          * @param int $length Maximum length (including ellipses)
3632          * @param string $ellipsis String to append to the truncated text
3633          * @param bool $adjustLength Subtract length of ellipsis from $length.
3634          *      $adjustLength was introduced in 1.18, before that behaved as if false.
3635          * @return string
3636          */
3637         function truncate( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3638                 # Use the localized ellipsis character
3639                 if ( $ellipsis == '...' ) {
3640                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3641                 }
3642                 # Check if there is no need to truncate
3643                 if ( $length == 0 ) {
3644                         return $ellipsis; // convention
3645                 } elseif ( strlen( $string ) <= abs( $length ) ) {
3646                         return $string; // no need to truncate
3647                 }
3648                 $stringOriginal = $string;
3649                 # If ellipsis length is >= $length then we can't apply $adjustLength
3650                 if ( $adjustLength && strlen( $ellipsis ) >= abs( $length ) ) {
3651                         $string = $ellipsis; // this can be slightly unexpected
3652                 # Otherwise, truncate and add ellipsis...
3653                 } else {
3654                         $eLength = $adjustLength ? strlen( $ellipsis ) : 0;
3655                         if ( $length > 0 ) {
3656                                 $length -= $eLength;
3657                                 $string = substr( $string, 0, $length ); // xyz...
3658                                 $string = $this->removeBadCharLast( $string );
3659                                 $string = rtrim( $string );
3660                                 $string = $string . $ellipsis;
3661                         } else {
3662                                 $length += $eLength;
3663                                 $string = substr( $string, $length ); // ...xyz
3664                                 $string = $this->removeBadCharFirst( $string );
3665                                 $string = ltrim( $string );
3666                                 $string = $ellipsis . $string;
3667                         }
3668                 }
3669                 # Do not truncate if the ellipsis makes the string longer/equal (bug 22181).
3670                 # This check is *not* redundant if $adjustLength, due to the single case where
3671                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3672                 if ( strlen( $string ) < strlen( $stringOriginal ) ) {
3673                         return $string;
3674                 } else {
3675                         return $stringOriginal;
3676                 }
3677         }
3678
3679         /**
3680          * Remove bytes that represent an incomplete Unicode character
3681          * at the end of string (e.g. bytes of the char are missing)
3682          *
3683          * @param string $string
3684          * @return string
3685          */
3686         protected function removeBadCharLast( $string ) {
3687                 if ( $string != '' ) {
3688                         $char = ord( $string[strlen( $string ) - 1] );
3689                         $m = array();
3690                         if ( $char >= 0xc0 ) {
3691                                 # We got the first byte only of a multibyte char; remove it.
3692                                 $string = substr( $string, 0, -1 );
3693                         } elseif ( $char >= 0x80 &&
3694                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3695                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/', $string, $m )
3696                         ) {
3697                                 # We chopped in the middle of a character; remove it
3698                                 $string = $m[1];
3699                         }
3700                 }
3701                 return $string;
3702         }
3703
3704         /**
3705          * Remove bytes that represent an incomplete Unicode character
3706          * at the start of string (e.g. bytes of the char are missing)
3707          *
3708          * @param string $string
3709          * @return string
3710          */
3711         protected function removeBadCharFirst( $string ) {
3712                 if ( $string != '' ) {
3713                         $char = ord( $string[0] );
3714                         if ( $char >= 0x80 && $char < 0xc0 ) {
3715                                 # We chopped in the middle of a character; remove the whole thing
3716                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3717                         }
3718                 }
3719                 return $string;
3720         }
3721
3722         /**
3723          * Truncate a string of valid HTML to a specified length in bytes,
3724          * appending an optional string (e.g. for ellipses), and return valid HTML
3725          *
3726          * This is only intended for styled/linked text, such as HTML with
3727          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3728          * Also, this will not detect things like "display:none" CSS.
3729          *
3730          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3731          *
3732          * @param string $text HTML string to truncate
3733          * @param int $length (zero/positive) Maximum length (including ellipses)
3734          * @param string $ellipsis String to append to the truncated text
3735          * @return string
3736          */
3737         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3738                 # Use the localized ellipsis character
3739                 if ( $ellipsis == '...' ) {
3740                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3741                 }
3742                 # Check if there is clearly no need to truncate
3743                 if ( $length <= 0 ) {
3744                         return $ellipsis; // no text shown, nothing to format (convention)
3745                 } elseif ( strlen( $text ) <= $length ) {
3746                         return $text; // string short enough even *with* HTML (short-circuit)
3747                 }
3748
3749                 $dispLen = 0; // innerHTML legth so far
3750                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3751                 $tagType = 0; // 0-open, 1-close
3752                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3753                 $entityState = 0; // 0-not entity, 1-entity
3754                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3755                 $openTags = array(); // open tag stack
3756                 $maybeState = null; // possible truncation state
3757
3758                 $textLen = strlen( $text );
3759                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3760                 for ( $pos = 0; true; ++$pos ) {
3761                         # Consider truncation once the display length has reached the maximim.
3762                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3763                         # Check that we're not in the middle of a bracket/entity...
3764                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3765                                 if ( !$testingEllipsis ) {
3766                                         $testingEllipsis = true;
3767                                         # Save where we are; we will truncate here unless there turn out to
3768                                         # be so few remaining characters that truncation is not necessary.
3769                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3770                                                 $maybeState = array( $ret, $openTags ); // save state
3771                                         }
3772                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3773                                         # String in fact does need truncation, the truncation point was OK.
3774                                         list( $ret, $openTags ) = $maybeState; // reload state
3775                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3776                                         $ret .= $ellipsis; // add ellipsis
3777                                         break;
3778                                 }
3779                         }
3780                         if ( $pos >= $textLen ) {
3781                                 break; // extra iteration just for above checks
3782                         }
3783
3784                         # Read the next char...
3785                         $ch = $text[$pos];
3786                         $lastCh = $pos ? $text[$pos - 1] : '';
3787                         $ret .= $ch; // add to result string
3788                         if ( $ch == '<' ) {
3789                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3790                                 $entityState = 0; // for bad HTML
3791                                 $bracketState = 1; // tag started (checking for backslash)
3792                         } elseif ( $ch == '>' ) {
3793                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3794                                 $entityState = 0; // for bad HTML
3795                                 $bracketState = 0; // out of brackets
3796                         } elseif ( $bracketState == 1 ) {
3797                                 if ( $ch == '/' ) {
3798                                         $tagType = 1; // close tag (e.g. "</span>")
3799                                 } else {
3800                                         $tagType = 0; // open tag (e.g. "<span>")
3801                                         $tag .= $ch;
3802                                 }
3803                                 $bracketState = 2; // building tag name
3804                         } elseif ( $bracketState == 2 ) {
3805                                 if ( $ch != ' ' ) {
3806                                         $tag .= $ch;
3807                                 } else {
3808                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3809                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3810                                 }
3811                         } elseif ( $bracketState == 0 ) {
3812                                 if ( $entityState ) {
3813                                         if ( $ch == ';' ) {
3814                                                 $entityState = 0;
3815                                                 $dispLen++; // entity is one displayed char
3816                                         }
3817                                 } else {
3818                                         if ( $neLength == 0 && !$maybeState ) {
3819                                                 // Save state without $ch. We want to *hit* the first
3820                                                 // display char (to get tags) but not *use* it if truncating.
3821                                                 $maybeState = array( substr( $ret, 0, -1 ), $openTags );
3822                                         }
3823                                         if ( $ch == '&' ) {
3824                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3825                                         } else {
3826                                                 $dispLen++; // this char is displayed
3827                                                 // Add the next $max display text chars after this in one swoop...
3828                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3829                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3830                                                 $dispLen += $skipped;
3831                                                 $pos += $skipped;
3832                                         }
3833                                 }
3834                         }
3835                 }
3836                 // Close the last tag if left unclosed by bad HTML
3837                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3838                 while ( count( $openTags ) > 0 ) {
3839                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3840                 }
3841                 return $ret;
3842         }
3843
3844         /**
3845          * truncateHtml() helper function
3846          * like strcspn() but adds the skipped chars to $ret
3847          *
3848          * @param string $ret
3849          * @param string $text
3850          * @param string $search
3851          * @param int $start
3852          * @param null|int $len
3853          * @return int
3854          */
3855         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3856                 if ( $len === null ) {
3857                         $len = -1; // -1 means "no limit" for strcspn
3858                 } elseif ( $len < 0 ) {
3859                         $len = 0; // sanity
3860                 }
3861                 $skipCount = 0;
3862                 if ( $start < strlen( $text ) ) {
3863                         $skipCount = strcspn( $text, $search, $start, $len );
3864                         $ret .= substr( $text, $start, $skipCount );
3865                 }
3866                 return $skipCount;
3867         }
3868
3869         /**
3870          * truncateHtml() helper function
3871          * (a) push or pop $tag from $openTags as needed
3872          * (b) clear $tag value
3873          * @param string &$tag Current HTML tag name we are looking at
3874          * @param int $tagType (0-open tag, 1-close tag)
3875          * @param string $lastCh Character before the '>' that ended this tag
3876          * @param array &$openTags Open tag stack (not accounting for $tag)
3877          */
3878         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3879                 $tag = ltrim( $tag );
3880                 if ( $tag != '' ) {
3881                         if ( $tagType == 0 && $lastCh != '/' ) {
3882                                 $openTags[] = $tag; // tag opened (didn't close itself)
3883                         } elseif ( $tagType == 1 ) {
3884                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3885                                         array_pop( $openTags ); // tag closed
3886                                 }
3887                         }
3888                         $tag = '';
3889                 }
3890         }
3891
3892         /**
3893          * Grammatical transformations, needed for inflected languages
3894          * Invoked by putting {{grammar:case|word}} in a message
3895          *
3896          * @param string $word
3897          * @param string $case
3898          * @return string
3899          */
3900         function convertGrammar( $word, $case ) {
3901                 global $wgGrammarForms;
3902                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3903                         return $wgGrammarForms[$this->getCode()][$case][$word];
3904                 }
3905
3906                 return $word;
3907         }
3908         /**
3909          * Get the grammar forms for the content language
3910          * @return array Array of grammar forms
3911          * @since 1.20
3912          */
3913         function getGrammarForms() {
3914                 global $wgGrammarForms;
3915                 if ( isset( $wgGrammarForms[$this->getCode()] )
3916                         && is_array( $wgGrammarForms[$this->getCode()] )
3917                 ) {
3918                         return $wgGrammarForms[$this->getCode()];
3919                 }
3920
3921                 return array();
3922         }
3923         /**
3924          * Provides an alternative text depending on specified gender.
3925          * Usage {{gender:username|masculine|feminine|unknown}}.
3926          * username is optional, in which case the gender of current user is used,
3927          * but only in (some) interface messages; otherwise default gender is used.
3928          *
3929          * If no forms are given, an empty string is returned. If only one form is
3930          * given, it will be returned unconditionally. These details are implied by
3931          * the caller and cannot be overridden in subclasses.
3932          *
3933          * If three forms are given, the default is to use the third (unknown) form.
3934          * If fewer than three forms are given, the default is to use the first (masculine) form.
3935          * These details can be overridden in subclasses.
3936          *
3937          * @param string $gender
3938          * @param array $forms
3939          *
3940          * @return string
3941          */
3942         function gender( $gender, $forms ) {
3943                 if ( !count( $forms ) ) {
3944                         return '';
3945                 }
3946                 $forms = $this->preConvertPlural( $forms, 2 );
3947                 if ( $gender === 'male' ) {
3948                         return $forms[0];
3949                 }
3950                 if ( $gender === 'female' ) {
3951                         return $forms[1];
3952                 }
3953                 return isset( $forms[2] ) ? $forms[2] : $forms[0];
3954         }
3955
3956         /**
3957          * Plural form transformations, needed for some languages.
3958          * For example, there are 3 form of plural in Russian and Polish,
3959          * depending on "count mod 10". See [[w:Plural]]
3960          * For English it is pretty simple.
3961          *
3962          * Invoked by putting {{plural:count|wordform1|wordform2}}
3963          * or {{plural:count|wordform1|wordform2|wordform3}}
3964          *
3965          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3966          *
3967          * @param int $count Non-localized number
3968          * @param array $forms Different plural forms
3969          * @return string Correct form of plural for $count in this language
3970          */
3971         function convertPlural( $count, $forms ) {
3972                 // Handle explicit n=pluralform cases
3973                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3974                 if ( is_string( $forms ) ) {
3975                         return $forms;
3976                 }
3977                 if ( !count( $forms ) ) {
3978                         return '';
3979                 }
3980
3981                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3982                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3983                 return $forms[$pluralForm];
3984         }
3985
3986         /**
3987          * Handles explicit plural forms for Language::convertPlural()
3988          *
3989          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
3990          * If an explicitly defined plural form matches the $count, then
3991          * string value returned, otherwise array returned for further consideration
3992          * by CLDR rules or overridden convertPlural().
3993          *
3994          * @since 1.23
3995          *
3996          * @param int $count Non-localized number
3997          * @param array $forms Different plural forms
3998          *
3999          * @return array|string
4000          */
4001         protected function handleExplicitPluralForms( $count, array $forms ) {
4002                 foreach ( $forms as $index => $form ) {
4003                         if ( preg_match( '/\d+=/i', $form ) ) {
4004                                 $pos = strpos( $form, '=' );
4005                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4006                                         return substr( $form, $pos + 1 );
4007                                 }
4008                                 unset( $forms[$index] );
4009                         }
4010                 }
4011                 return array_values( $forms );
4012         }
4013
4014         /**
4015          * Checks that convertPlural was given an array and pads it to requested
4016          * amount of forms by copying the last one.
4017          *
4018          * @param array $forms Array of forms given to convertPlural
4019          * @param int $count How many forms should there be at least
4020          * @return array Padded array of forms or an exception if not an array
4021          */
4022         protected function preConvertPlural( /* Array */ $forms, $count ) {
4023                 while ( count( $forms ) < $count ) {
4024                         $forms[] = $forms[count( $forms ) - 1];
4025                 }
4026                 return $forms;
4027         }
4028
4029         /**
4030          * Wraps argument with unicode control characters for directionality safety
4031          *
4032          * This solves the problem where directionality-neutral characters at the edge of
4033          * the argument string get interpreted with the wrong directionality from the
4034          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4035          *
4036          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4037          * directionality of the argument string, using the BIDI algorithm's own "First
4038          * strong directional codepoint" rule. Essentially, this works round the fact that
4039          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4040          * direction inference). The latter is cleaner but still not widely supported.
4041          *
4042          * @param string $text Text to wrap
4043          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4044          */
4045         public function embedBidi( $text = '' ) {
4046                 $dir = Language::strongDirFromContent( $text );
4047                 if ( $dir === 'ltr' ) {
4048                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4049                         return self::$lre . $text . self::$pdf;
4050                 }
4051                 if ( $dir === 'rtl' ) {
4052                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4053                         return self::$rle . $text . self::$pdf;
4054                 }
4055                 // No strong directionality: do not wrap
4056                 return $text;
4057         }
4058
4059         /**
4060          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4061          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4062          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4063          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4064          * match up with it.
4065          *
4066          * @param string $str The validated block duration in English
4067          * @return string Somehow translated block duration
4068          * @see LanguageFi.php for example implementation
4069          */
4070         function translateBlockExpiry( $str ) {
4071                 $duration = SpecialBlock::getSuggestedDurations( $this );
4072                 foreach ( $duration as $show => $value ) {
4073                         if ( strcmp( $str, $value ) == 0 ) {
4074                                 return htmlspecialchars( trim( $show ) );
4075                         }
4076                 }
4077
4078                 if ( wfIsInfinity( $str ) ) {
4079                         foreach ( $duration as $show => $value ) {
4080                                 if ( wfIsInfinity( $value ) ) {
4081                                         return htmlspecialchars( trim( $show ) );
4082                                 }
4083                         }
4084                 }
4085
4086                 // If all else fails, return a standard duration or timestamp description.
4087                 $time = strtotime( $str, 0 );
4088                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4089                         return $str;
4090                 } elseif ( $time !== strtotime( $str, 1 ) ) { // It's a relative timestamp.
4091                         // $time is relative to 0 so it's a duration length.
4092                         return $this->formatDuration( $time );
4093                 } else { // It's an absolute timestamp.
4094                         if ( $time === 0 ) {
4095                                 // wfTimestamp() handles 0 as current time instead of epoch.
4096                                 return $this->timeanddate( '19700101000000' );
4097                         } else {
4098                                 return $this->timeanddate( $time );
4099                         }
4100                 }
4101         }
4102
4103         /**
4104          * languages like Chinese need to be segmented in order for the diff
4105          * to be of any use
4106          *
4107          * @param string $text
4108          * @return string
4109          */
4110         public function segmentForDiff( $text ) {
4111                 return $text;
4112         }
4113
4114         /**
4115          * and unsegment to show the result
4116          *
4117          * @param string $text
4118          * @return string
4119          */
4120         public function unsegmentForDiff( $text ) {
4121                 return $text;
4122         }
4123
4124         /**
4125          * Return the LanguageConverter used in the Language
4126          *
4127          * @since 1.19
4128          * @return LanguageConverter
4129          */
4130         public function getConverter() {
4131                 return $this->mConverter;
4132         }
4133
4134         /**
4135          * convert text to all supported variants
4136          *
4137          * @param string $text
4138          * @return array
4139          */
4140         public function autoConvertToAllVariants( $text ) {
4141                 return $this->mConverter->autoConvertToAllVariants( $text );
4142         }
4143
4144         /**
4145          * convert text to different variants of a language.
4146          *
4147          * @param string $text
4148          * @return string
4149          */
4150         public function convert( $text ) {
4151                 return $this->mConverter->convert( $text );
4152         }
4153
4154         /**
4155          * Convert a Title object to a string in the preferred variant
4156          *
4157          * @param Title $title
4158          * @return string
4159          */
4160         public function convertTitle( $title ) {
4161                 return $this->mConverter->convertTitle( $title );
4162         }
4163
4164         /**
4165          * Convert a namespace index to a string in the preferred variant
4166          *
4167          * @param int $ns
4168          * @return string
4169          */
4170         public function convertNamespace( $ns ) {
4171                 return $this->mConverter->convertNamespace( $ns );
4172         }
4173
4174         /**
4175          * Check if this is a language with variants
4176          *
4177          * @return bool
4178          */
4179         public function hasVariants() {
4180                 return count( $this->getVariants() ) > 1;
4181         }
4182
4183         /**
4184          * Check if the language has the specific variant
4185          *
4186          * @since 1.19
4187          * @param string $variant
4188          * @return bool
4189          */
4190         public function hasVariant( $variant ) {
4191                 return (bool)$this->mConverter->validateVariant( $variant );
4192         }
4193
4194         /**
4195          * Put custom tags (e.g. -{ }-) around math to prevent conversion
4196          *
4197          * @param string $text
4198          * @return string
4199          * @deprecated since 1.22 is no longer used
4200          */
4201         public function armourMath( $text ) {
4202                 return $this->mConverter->armourMath( $text );
4203         }
4204
4205         /**
4206          * Perform output conversion on a string, and encode for safe HTML output.
4207          * @param string $text Text to be converted
4208          * @param bool $isTitle Whether this conversion is for the article title
4209          * @return string
4210          * @todo this should get integrated somewhere sane
4211          */
4212         public function convertHtml( $text, $isTitle = false ) {
4213                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
4214         }
4215
4216         /**
4217          * @param string $key
4218          * @return string
4219          */
4220         public function convertCategoryKey( $key ) {
4221                 return $this->mConverter->convertCategoryKey( $key );
4222         }
4223
4224         /**
4225          * Get the list of variants supported by this language
4226          * see sample implementation in LanguageZh.php
4227          *
4228          * @return array An array of language codes
4229          */
4230         public function getVariants() {
4231                 return $this->mConverter->getVariants();
4232         }
4233
4234         /**
4235          * @return string
4236          */
4237         public function getPreferredVariant() {
4238                 return $this->mConverter->getPreferredVariant();
4239         }
4240
4241         /**
4242          * @return string
4243          */
4244         public function getDefaultVariant() {
4245                 return $this->mConverter->getDefaultVariant();
4246         }
4247
4248         /**
4249          * @return string
4250          */
4251         public function getURLVariant() {
4252                 return $this->mConverter->getURLVariant();
4253         }
4254
4255         /**
4256          * If a language supports multiple variants, it is
4257          * possible that non-existing link in one variant
4258          * actually exists in another variant. this function
4259          * tries to find it. See e.g. LanguageZh.php
4260          * The input parameters may be modified upon return
4261          *
4262          * @param string &$link The name of the link
4263          * @param Title &$nt The title object of the link
4264          * @param bool $ignoreOtherCond To disable other conditions when
4265          *   we need to transclude a template or update a category's link
4266          */
4267         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4268                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4269         }
4270
4271         /**
4272          * returns language specific options used by User::getPageRenderHash()
4273          * for example, the preferred language variant
4274          *
4275          * @return string
4276          */
4277         function getExtraHashOptions() {
4278                 return $this->mConverter->getExtraHashOptions();
4279         }
4280
4281         /**
4282          * For languages that support multiple variants, the title of an
4283          * article may be displayed differently in different variants. this
4284          * function returns the apporiate title defined in the body of the article.
4285          *
4286          * @return string
4287          */
4288         public function getParsedTitle() {
4289                 return $this->mConverter->getParsedTitle();
4290         }
4291
4292         /**
4293          * Prepare external link text for conversion. When the text is
4294          * a URL, it shouldn't be converted, and it'll be wrapped in
4295          * the "raw" tag (-{R| }-) to prevent conversion.
4296          *
4297          * This function is called "markNoConversion" for historical
4298          * reasons.
4299          *
4300          * @param string $text Text to be used for external link
4301          * @param bool $noParse Wrap it without confirming it's a real URL first
4302          * @return string The tagged text
4303          */
4304         public function markNoConversion( $text, $noParse = false ) {
4305                 // Excluding protocal-relative URLs may avoid many false positives.
4306                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
4307                         return $this->mConverter->markNoConversion( $text );
4308                 } else {
4309                         return $text;
4310                 }
4311         }
4312
4313         /**
4314          * A regular expression to match legal word-trailing characters
4315          * which should be merged onto a link of the form [[foo]]bar.
4316          *
4317          * @return string
4318          */
4319         public function linkTrail() {
4320                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4321         }
4322
4323         /**
4324          * A regular expression character set to match legal word-prefixing
4325          * characters which should be merged onto a link of the form foo[[bar]].
4326          *
4327          * @return string
4328          */
4329         public function linkPrefixCharset() {
4330                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4331         }
4332
4333         /**
4334          * @deprecated since 1.24, will be removed in 1.25
4335          * @return Language
4336          */
4337         function getLangObj() {
4338                 wfDeprecated( __METHOD__, '1.24' );
4339                 return $this;
4340         }
4341
4342         /**
4343          * Get the "parent" language which has a converter to convert a "compatible" language
4344          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4345          *
4346          * @return Language|null
4347          * @since 1.22
4348          */
4349         public function getParentLanguage() {
4350                 if ( $this->mParentLanguage !== false ) {
4351                         return $this->mParentLanguage;
4352                 }
4353
4354                 $pieces = explode( '-', $this->getCode() );
4355                 $code = $pieces[0];
4356                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4357                         $this->mParentLanguage = null;
4358                         return null;
4359                 }
4360                 $lang = Language::factory( $code );
4361                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4362                         $this->mParentLanguage = null;
4363                         return null;
4364                 }
4365
4366                 $this->mParentLanguage = $lang;
4367                 return $lang;
4368         }
4369
4370         /**
4371          * Get the RFC 3066 code for this language object
4372          *
4373          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4374          * htmlspecialchars() or similar
4375          *
4376          * @return string
4377          */
4378         public function getCode() {
4379                 return $this->mCode;
4380         }
4381
4382         /**
4383          * Get the code in Bcp47 format which we can use
4384          * inside of html lang="" tags.
4385          *
4386          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4387          * htmlspecialchars() or similar.
4388          *
4389          * @since 1.19
4390          * @return string
4391          */
4392         public function getHtmlCode() {
4393                 if ( is_null( $this->mHtmlCode ) ) {
4394                         $this->mHtmlCode = wfBCP47( $this->getCode() );
4395                 }
4396                 return $this->mHtmlCode;
4397         }
4398
4399         /**
4400          * @param string $code
4401          */
4402         public function setCode( $code ) {
4403                 $this->mCode = $code;
4404                 // Ensure we don't leave incorrect cached data lying around
4405                 $this->mHtmlCode = null;
4406                 $this->mParentLanguage = false;
4407         }
4408
4409         /**
4410          * Get the name of a file for a certain language code
4411          * @param string $prefix Prepend this to the filename
4412          * @param string $code Language code
4413          * @param string $suffix Append this to the filename
4414          * @throws MWException
4415          * @return string $prefix . $mangledCode . $suffix
4416          */
4417         public static function getFileName( $prefix = 'Language', $code, $suffix = '.php' ) {
4418                 if ( !self::isValidBuiltInCode( $code ) ) {
4419                         throw new MWException( "Invalid language code \"$code\"" );
4420                 }
4421
4422                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4423         }
4424
4425         /**
4426          * Get the language code from a file name. Inverse of getFileName()
4427          * @param string $filename $prefix . $languageCode . $suffix
4428          * @param string $prefix Prefix before the language code
4429          * @param string $suffix Suffix after the language code
4430          * @return string Language code, or false if $prefix or $suffix isn't found
4431          */
4432         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4433                 $m = null;
4434                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4435                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4436                 if ( !count( $m ) ) {
4437                         return false;
4438                 }
4439                 return str_replace( '_', '-', strtolower( $m[1] ) );
4440         }
4441
4442         /**
4443          * @param string $code
4444          * @return string
4445          */
4446         public static function getMessagesFileName( $code ) {
4447                 global $IP;
4448                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4449                 Hooks::run( 'Language::getMessagesFileName', array( $code, &$file ) );
4450                 return $file;
4451         }
4452
4453         /**
4454          * @param string $code
4455          * @return string
4456          * @since 1.23
4457          */
4458         public static function getJsonMessagesFileName( $code ) {
4459                 global $IP;
4460
4461                 if ( !self::isValidBuiltInCode( $code ) ) {
4462                         throw new MWException( "Invalid language code \"$code\"" );
4463                 }
4464
4465                 return "$IP/languages/i18n/$code.json";
4466         }
4467
4468         /**
4469          * @param string $code
4470          * @return string
4471          */
4472         public static function getClassFileName( $code ) {
4473                 global $IP;
4474                 return self::getFileName( "$IP/languages/classes/Language", $code, '.php' );
4475         }
4476
4477         /**
4478          * Get the first fallback for a given language.
4479          *
4480          * @param string $code
4481          *
4482          * @return bool|string
4483          */
4484         public static function getFallbackFor( $code ) {
4485                 if ( $code === 'en' || !Language::isValidBuiltInCode( $code ) ) {
4486                         return false;
4487                 } else {
4488                         $fallbacks = self::getFallbacksFor( $code );
4489                         return $fallbacks[0];
4490                 }
4491         }
4492
4493         /**
4494          * Get the ordered list of fallback languages.
4495          *
4496          * @since 1.19
4497          * @param string $code Language code
4498          * @return array Non-empty array, ending in "en"
4499          */
4500         public static function getFallbacksFor( $code ) {
4501                 if ( $code === 'en' || !Language::isValidBuiltInCode( $code ) ) {
4502                         return array();
4503                 }
4504                 // For unknown languages, fallbackSequence returns an empty array,
4505                 // hardcode fallback to 'en' in that case.
4506                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: array( 'en' );
4507         }
4508
4509         /**
4510          * Get the ordered list of fallback languages, ending with the fallback
4511          * language chain for the site language.
4512          *
4513          * @since 1.22
4514          * @param string $code Language code
4515          * @return array Array( fallbacks, site fallbacks )
4516          */
4517         public static function getFallbacksIncludingSiteLanguage( $code ) {
4518                 global $wgLanguageCode;
4519
4520                 // Usually, we will only store a tiny number of fallback chains, so we
4521                 // keep them in static memory.
4522                 $cacheKey = "{$code}-{$wgLanguageCode}";
4523
4524                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4525                         $fallbacks = self::getFallbacksFor( $code );
4526
4527                         // Append the site's fallback chain, including the site language itself
4528                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4529                         array_unshift( $siteFallbacks, $wgLanguageCode );
4530
4531                         // Eliminate any languages already included in the chain
4532                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4533
4534                         self::$fallbackLanguageCache[$cacheKey] = array( $fallbacks, $siteFallbacks );
4535                 }
4536                 return self::$fallbackLanguageCache[$cacheKey];
4537         }
4538
4539         /**
4540          * Get all messages for a given language
4541          * WARNING: this may take a long time. If you just need all message *keys*
4542          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4543          *
4544          * @param string $code
4545          *
4546          * @return array
4547          */
4548         public static function getMessagesFor( $code ) {
4549                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4550         }
4551
4552         /**
4553          * Get a message for a given language
4554          *
4555          * @param string $key
4556          * @param string $code
4557          *
4558          * @return string
4559          */
4560         public static function getMessageFor( $key, $code ) {
4561                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4562         }
4563
4564         /**
4565          * Get all message keys for a given language. This is a faster alternative to
4566          * array_keys( Language::getMessagesFor( $code ) )
4567          *
4568          * @since 1.19
4569          * @param string $code Language code
4570          * @return array Array of message keys (strings)
4571          */
4572         public static function getMessageKeysFor( $code ) {
4573                 return self::getLocalisationCache()->getSubItemList( $code, 'messages' );
4574         }
4575
4576         /**
4577          * @param string $talk
4578          * @return mixed
4579          */
4580         function fixVariableInNamespace( $talk ) {
4581                 if ( strpos( $talk, '$1' ) === false ) {
4582                         return $talk;
4583                 }
4584
4585                 global $wgMetaNamespace;
4586                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4587
4588                 # Allow grammar transformations
4589                 # Allowing full message-style parsing would make simple requests
4590                 # such as action=raw much more expensive than they need to be.
4591                 # This will hopefully cover most cases.
4592                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4593                         array( &$this, 'replaceGrammarInNamespace' ), $talk );
4594                 return str_replace( ' ', '_', $talk );
4595         }
4596
4597         /**
4598          * @param string $m
4599          * @return string
4600          */
4601         function replaceGrammarInNamespace( $m ) {
4602                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4603         }
4604
4605         /**
4606          * @throws MWException
4607          * @return array
4608          */
4609         static function getCaseMaps() {
4610                 static $wikiUpperChars, $wikiLowerChars;
4611                 if ( isset( $wikiUpperChars ) ) {
4612                         return array( $wikiUpperChars, $wikiLowerChars );
4613                 }
4614
4615                 $arr = wfGetPrecompiledData( 'Utf8Case.ser' );
4616                 if ( $arr === false ) {
4617                         throw new MWException(
4618                                 "Utf8Case.ser is missing, please run \"make\" in the serialized directory\n" );
4619                 }
4620                 $wikiUpperChars = $arr['wikiUpperChars'];
4621                 $wikiLowerChars = $arr['wikiLowerChars'];
4622                 return array( $wikiUpperChars, $wikiLowerChars );
4623         }
4624
4625         /**
4626          * Decode an expiry (block, protection, etc) which has come from the DB
4627          *
4628          * @param string $expiry Database expiry String
4629          * @param bool|int $format True to process using language functions, or TS_ constant
4630          *     to return the expiry in a given timestamp
4631          * @param string $inifinity If $format is not true, use this string for infinite expiry
4632          * @return string
4633          * @since 1.18
4634          */
4635         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4636                 static $dbInfinity;
4637                 if ( $dbInfinity === null ) {
4638                         $dbInfinity = wfGetDB( DB_SLAVE )->getInfinity();
4639                 }
4640
4641                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4642                         return $format === true
4643                                 ? $this->getMessageFromDB( 'infiniteblock' )
4644                                 : $infinity;
4645                 } else {
4646                         return $format === true
4647                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4648                                 : wfTimestamp( $format, $expiry );
4649                 }
4650         }
4651
4652         /**
4653          * @todo Document
4654          * @param int|float $seconds
4655          * @param array $format Optional
4656          *   If $format['avoid'] === 'avoidseconds': don't mention seconds if $seconds >= 1 hour.
4657          *   If $format['avoid'] === 'avoidminutes': don't mention seconds/minutes if $seconds > 48 hours.
4658          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4659          *     and friends.
4660          *   For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4661          *     or 'avoidminutes'.
4662          * @return string
4663          */
4664         function formatTimePeriod( $seconds, $format = array() ) {
4665                 if ( !is_array( $format ) ) {
4666                         $format = array( 'avoid' => $format ); // For backwards compatibility
4667                 }
4668                 if ( !isset( $format['avoid'] ) ) {
4669                         $format['avoid'] = false;
4670                 }
4671                 if ( !isset( $format['noabbrevs'] ) ) {
4672                         $format['noabbrevs'] = false;
4673                 }
4674                 $secondsMsg = wfMessage(
4675                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4676                 $minutesMsg = wfMessage(
4677                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4678                 $hoursMsg = wfMessage(
4679                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4680                 $daysMsg = wfMessage(
4681                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4682
4683                 if ( round( $seconds * 10 ) < 100 ) {
4684                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4685                         $s = $secondsMsg->params( $s )->text();
4686                 } elseif ( round( $seconds ) < 60 ) {
4687                         $s = $this->formatNum( round( $seconds ) );
4688                         $s = $secondsMsg->params( $s )->text();
4689                 } elseif ( round( $seconds ) < 3600 ) {
4690                         $minutes = floor( $seconds / 60 );
4691                         $secondsPart = round( fmod( $seconds, 60 ) );
4692                         if ( $secondsPart == 60 ) {
4693                                 $secondsPart = 0;
4694                                 $minutes++;
4695                         }
4696                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4697                         $s .= ' ';
4698                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4699                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4700                         $hours = floor( $seconds / 3600 );
4701                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4702                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4703                         if ( $secondsPart == 60 ) {
4704                                 $secondsPart = 0;
4705                                 $minutes++;
4706                         }
4707                         if ( $minutes == 60 ) {
4708                                 $minutes = 0;
4709                                 $hours++;
4710                         }
4711                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4712                         $s .= ' ';
4713                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4714                         if ( !in_array( $format['avoid'], array( 'avoidseconds', 'avoidminutes' ) ) ) {
4715                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4716                         }
4717                 } else {
4718                         $days = floor( $seconds / 86400 );
4719                         if ( $format['avoid'] === 'avoidminutes' ) {
4720                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4721                                 if ( $hours == 24 ) {
4722                                         $hours = 0;
4723                                         $days++;
4724                                 }
4725                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4726                                 $s .= ' ';
4727                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4728                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4729                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4730                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4731                                 if ( $minutes == 60 ) {
4732                                         $minutes = 0;
4733                                         $hours++;
4734                                 }
4735                                 if ( $hours == 24 ) {
4736                                         $hours = 0;
4737                                         $days++;
4738                                 }
4739                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4740                                 $s .= ' ';
4741                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4742                                 $s .= ' ';
4743                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4744                         } else {
4745                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4746                                 $s .= ' ';
4747                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4748                         }
4749                 }
4750                 return $s;
4751         }
4752
4753         /**
4754          * Format a bitrate for output, using an appropriate
4755          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4756          *   the magnitude in question.
4757          *
4758          * This use base 1000. For base 1024 use formatSize(), for another base
4759          * see formatComputingNumbers().
4760          *
4761          * @param int $bps
4762          * @return string
4763          */
4764         function formatBitrate( $bps ) {
4765                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4766         }
4767
4768         /**
4769          * @param int $size Size of the unit
4770          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4771          * @param string $messageKey Message key to be uesd
4772          * @return string
4773          */
4774         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4775                 if ( $size <= 0 ) {
4776                         return str_replace( '$1', $this->formatNum( $size ),
4777                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4778                         );
4779                 }
4780                 $sizes = array( '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' );
4781                 $index = 0;
4782
4783                 $maxIndex = count( $sizes ) - 1;
4784                 while ( $size >= $boundary && $index < $maxIndex ) {
4785                         $index++;
4786                         $size /= $boundary;
4787                 }
4788
4789                 // For small sizes no decimal places necessary
4790                 $round = 0;
4791                 if ( $index > 1 ) {
4792                         // For MB and bigger two decimal places are smarter
4793                         $round = 2;
4794                 }
4795                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4796
4797                 $size = round( $size, $round );
4798                 $text = $this->getMessageFromDB( $msg );
4799                 return str_replace( '$1', $this->formatNum( $size ), $text );
4800         }
4801
4802         /**
4803          * Format a size in bytes for output, using an appropriate
4804          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4805          *
4806          * This method use base 1024. For base 1000 use formatBitrate(), for
4807          * another base see formatComputingNumbers()
4808          *
4809          * @param int $size Size to format
4810          * @return string Plain text (not HTML)
4811          */
4812         function formatSize( $size ) {
4813                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4814         }
4815
4816         /**
4817          * Make a list item, used by various special pages
4818          *
4819          * @param string $page Page link
4820          * @param string $details HTML safe text between brackets
4821          * @param bool $oppositedm Add the direction mark opposite to your
4822          *   language, to display text properly
4823          * @return HTML escaped string
4824          */
4825         function specialList( $page, $details, $oppositedm = true ) {
4826                 if ( !$details ) {
4827                         return $page;
4828                 }
4829
4830                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4831                 return
4832                         $page .
4833                         $dirmark .
4834                         $this->msg( 'word-separator' )->escaped() .
4835                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4836         }
4837
4838         /**
4839          * Generate (prev x| next x) (20|50|100...) type links for paging
4840          *
4841          * @param Title $title Title object to link
4842          * @param int $offset
4843          * @param int $limit
4844          * @param array $query Optional URL query parameter string
4845          * @param bool $atend Optional param for specified if this is the last page
4846          * @return string
4847          */
4848         public function viewPrevNext( Title $title, $offset, $limit,
4849                 array $query = array(), $atend = false
4850         ) {
4851                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4852
4853                 # Make 'previous' link
4854                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4855                 if ( $offset > 0 ) {
4856                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4857                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4858                 } else {
4859                         $plink = htmlspecialchars( $prev );
4860                 }
4861
4862                 # Make 'next' link
4863                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4864                 if ( $atend ) {
4865                         $nlink = htmlspecialchars( $next );
4866                 } else {
4867                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4868                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4869                 }
4870
4871                 # Make links to set number of items per page
4872                 $numLinks = array();
4873                 foreach ( array( 20, 50, 100, 250, 500 ) as $num ) {
4874                         $numLinks[] = $this->numLink( $title, $offset, $num,
4875                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4876                 }
4877
4878                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4879                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4880         }
4881
4882         /**
4883          * Helper function for viewPrevNext() that generates links
4884          *
4885          * @param Title $title Title object to link
4886          * @param int $offset
4887          * @param int $limit
4888          * @param array $query Extra query parameters
4889          * @param string $link Text to use for the link; will be escaped
4890          * @param string $tooltipMsg Name of the message to use as tooltip
4891          * @param string $class Value of the "class" attribute of the link
4892          * @return string HTML fragment
4893          */
4894         private function numLink( Title $title, $offset, $limit, array $query, $link,
4895                 $tooltipMsg, $class
4896         ) {
4897                 $query = array( 'limit' => $limit, 'offset' => $offset ) + $query;
4898                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4899                         ->numParams( $limit )->text();
4900
4901                 return Html::element( 'a', array( 'href' => $title->getLocalURL( $query ),
4902                         'title' => $tooltip, 'class' => $class ), $link );
4903         }
4904
4905         /**
4906          * Get the conversion rule title, if any.
4907          *
4908          * @return string
4909          */
4910         public function getConvRuleTitle() {
4911                 return $this->mConverter->getConvRuleTitle();
4912         }
4913
4914         /**
4915          * Get the compiled plural rules for the language
4916          * @since 1.20
4917          * @return array Associative array with plural form, and plural rule as key-value pairs
4918          */
4919         public function getCompiledPluralRules() {
4920                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4921                 $fallbacks = Language::getFallbacksFor( $this->mCode );
4922                 if ( !$pluralRules ) {
4923                         foreach ( $fallbacks as $fallbackCode ) {
4924                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4925                                 if ( $pluralRules ) {
4926                                         break;
4927                                 }
4928                         }
4929                 }
4930                 return $pluralRules;
4931         }
4932
4933         /**
4934          * Get the plural rules for the language
4935          * @since 1.20
4936          * @return array Associative array with plural form number and plural rule as key-value pairs
4937          */
4938         public function getPluralRules() {
4939                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4940                 $fallbacks = Language::getFallbacksFor( $this->mCode );
4941                 if ( !$pluralRules ) {
4942                         foreach ( $fallbacks as $fallbackCode ) {
4943                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4944                                 if ( $pluralRules ) {
4945                                         break;
4946                                 }
4947                         }
4948                 }
4949                 return $pluralRules;
4950         }
4951
4952         /**
4953          * Get the plural rule types for the language
4954          * @since 1.22
4955          * @return array Associative array with plural form number and plural rule type as key-value pairs
4956          */
4957         public function getPluralRuleTypes() {
4958                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4959                 $fallbacks = Language::getFallbacksFor( $this->mCode );
4960                 if ( !$pluralRuleTypes ) {
4961                         foreach ( $fallbacks as $fallbackCode ) {
4962                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4963                                 if ( $pluralRuleTypes ) {
4964                                         break;
4965                                 }
4966                         }
4967                 }
4968                 return $pluralRuleTypes;
4969         }
4970
4971         /**
4972          * Find the index number of the plural rule appropriate for the given number
4973          * @param int $number
4974          * @return int The index number of the plural rule
4975          */
4976         public function getPluralRuleIndexNumber( $number ) {
4977                 $pluralRules = $this->getCompiledPluralRules();
4978                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4979                 return $form;
4980         }
4981
4982         /**
4983          * Find the plural rule type appropriate for the given number
4984          * For example, if the language is set to Arabic, getPluralType(5) should
4985          * return 'few'.
4986          * @since 1.22
4987          * @param int $number
4988          * @return string The name of the plural rule type, e.g. one, two, few, many
4989          */
4990         public function getPluralRuleType( $number ) {
4991                 $index = $this->getPluralRuleIndexNumber( $number );
4992                 $pluralRuleTypes = $this->getPluralRuleTypes();
4993                 if ( isset( $pluralRuleTypes[$index] ) ) {
4994                         return $pluralRuleTypes[$index];
4995                 } else {
4996                         return 'other';
4997                 }
4998         }
4999 }