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