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 translated language names. This is done on best effort and
 807          * by default this is exactly the same as Language::getLanguageNames.
 808          * The CLDR extension provides translated names.
 809          * @param string $code Language code.
 810          * @return array Language code => language name
 811          * @since 1.18.0
 812          * @deprecated since 1.20, use fetchLanguageNames()
 813          */
 814         public static function getTranslatedLanguageNames( $code ) {
 815                 return self::fetchLanguageNames( $code, 'all' );
 816         }
 817
 818         /**
 819          * Get an array of language names, indexed by code.
 820          * @param null|string $inLanguage Code of language in which to return the names
 821          *              Use null for autonyms (native names)
 822          * @param string $include One of:
 823          *              'all' all available languages
 824          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 825          *              'mwfile' only if the language is in 'mw' *and* has a message file
 826          * @return array Language code => language name
 827          * @since 1.20
 828          */
 829         public static function fetchLanguageNames( $inLanguage = null, $include = 'mw' ) {
 830                 $cacheKey = $inLanguage === null ? 'null' : $inLanguage;
 831                 $cacheKey .= ":$include";
 832                 if ( self::$languageNameCache === null ) {
 833                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 834                 }
 835
 836                 $ret = self::$languageNameCache->get( $cacheKey );
 837                 if ( !$ret ) {
 838                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 839                         self::$languageNameCache->set( $cacheKey, $ret );
 840                 }
 841                 return $ret;
 842         }
 843
 844         /**
 845          * Uncached helper for fetchLanguageNames
 846          * @param null|string $inLanguage Code of language in which to return the names
 847          *              Use null for autonyms (native names)
 848          * @param string $include One of:
 849          *              'all' all available languages
 850          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 851          *              'mwfile' only if the language is in 'mw' *and* has a message file
 852          * @return array Language code => language name
 853          */
 854         private static function fetchLanguageNamesUncached( $inLanguage = null, $include = 'mw' ) {
 855                 global $wgExtraLanguageNames;
 856
 857                 // If passed an invalid language code to use, fallback to en
 858                 if ( $inLanguage !== null && !Language::isValidCode( $inLanguage ) ) {
 859                         $inLanguage = 'en';
 860                 }
 861
 862                 $names = [];
 863
 864                 if ( $inLanguage ) {
 865                         # TODO: also include when $inLanguage is null, when this code is more efficient
 866                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 867                 }
 868
 869                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 870                 foreach ( $mwNames as $mwCode => $mwName ) {
 871                         # - Prefer own MediaWiki native name when not using the hook
 872                         # - For other names just add if not added through the hook
 873                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 874                                 $names[$mwCode] = $mwName;
 875                         }
 876                 }
 877
 878                 if ( $include === 'all' ) {
 879                         ksort( $names );
 880                         return $names;
 881                 }
 882
 883                 $returnMw = [];
 884                 $coreCodes = array_keys( $mwNames );
 885                 foreach ( $coreCodes as $coreCode ) {
 886                         $returnMw[$coreCode] = $names[$coreCode];
 887                 }
 888
 889                 if ( $include === 'mwfile' ) {
 890                         $namesMwFile = [];
 891                         # We do this using a foreach over the codes instead of a directory
 892                         # loop so that messages files in extensions will work correctly.
 893                         foreach ( $returnMw as $code => $value ) {
 894                                 if ( is_readable( self::getMessagesFileName( $code ) )
 895                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 896                                 ) {
 897                                         $namesMwFile[$code] = $names[$code];
 898                                 }
 899                         }
 900
 901                         ksort( $namesMwFile );
 902                         return $namesMwFile;
 903                 }
 904
 905                 ksort( $returnMw );
 906                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 907                 return $returnMw;
 908         }
 909
 910         /**
 911          * @param string $code The code of the language for which to get the name
 912          * @param null|string $inLanguage Code of language in which to return the name (null for autonyms)
 913          * @param string $include 'all', 'mw' or 'mwfile'; see fetchLanguageNames()
 914          * @return string Language name or empty
 915          * @since 1.20
 916          */
 917         public static function fetchLanguageName( $code, $inLanguage = null, $include = 'all' ) {
 918                 $code = strtolower( $code );
 919                 $array = self::fetchLanguageNames( $inLanguage, $include );
 920                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 921         }
 922
 923         /**
 924          * Get a message from the MediaWiki namespace.
 925          *
 926          * @param string $msg Message name
 927          * @return string
 928          */
 929         function getMessageFromDB( $msg ) {
 930                 return $this->msg( $msg )->text();
 931         }
 932
 933         /**
 934          * Get message object in this language. Only for use inside this class.
 935          *
 936          * @param string $msg Message name
 937          * @return Message
 938          */
 939         protected function msg( $msg ) {
 940                 return wfMessage( $msg )->inLanguage( $this );
 941         }
 942
 943         /**
 944          * Get the native language name of $code.
 945          * Only if defined in MediaWiki, no other data like CLDR.
 946          * @param string $code
 947          * @return string
 948          * @deprecated since 1.20, use fetchLanguageName()
 949          */
 950         function getLanguageName( $code ) {
 951                 return self::fetchLanguageName( $code );
 952         }
 953
 954         /**
 955          * @param string $key
 956          * @return string
 957          */
 958         function getMonthName( $key ) {
 959                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 960         }
 961
 962         /**
 963          * @return array
 964          */
 965         function getMonthNamesArray() {
 966                 $monthNames = [ '' ];
 967                 for ( $i = 1; $i < 13; $i++ ) {
 968                         $monthNames[] = $this->getMonthName( $i );
 969                 }
 970                 return $monthNames;
 971         }
 972
 973         /**
 974          * @param string $key
 975          * @return string
 976          */
 977         function getMonthNameGen( $key ) {
 978                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 979         }
 980
 981         /**
 982          * @param string $key
 983          * @return string
 984          */
 985         function getMonthAbbreviation( $key ) {
 986                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 987         }
 988
 989         /**
 990          * @return array
 991          */
 992         function getMonthAbbreviationsArray() {
 993                 $monthNames = [ '' ];
 994                 for ( $i = 1; $i < 13; $i++ ) {
 995                         $monthNames[] = $this->getMonthAbbreviation( $i );
 996                 }
 997                 return $monthNames;
 998         }
 999
1000         /**
1001          * @param string $key
1002          * @return string
1003          */
1004         function getWeekdayName( $key ) {
1005                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1006         }
1007
1008         /**
1009          * @param string $key
1010          * @return string
1011          */
1012         function getWeekdayAbbreviation( $key ) {
1013                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1014         }
1015
1016         /**
1017          * @param string $key
1018          * @return string
1019          */
1020         function getIranianCalendarMonthName( $key ) {
1021                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1022         }
1023
1024         /**
1025          * @param string $key
1026          * @return string
1027          */
1028         function getHebrewCalendarMonthName( $key ) {
1029                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1030         }
1031
1032         /**
1033          * @param string $key
1034          * @return string
1035          */
1036         function getHebrewCalendarMonthNameGen( $key ) {
1037                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1038         }
1039
1040         /**
1041          * @param string $key
1042          * @return string
1043          */
1044         function getHijriCalendarMonthName( $key ) {
1045                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1046         }
1047
1048         /**
1049          * Pass through result from $dateTimeObj->format()
1050          * @param DateTime|bool|null &$dateTimeObj
1051          * @param string $ts
1052          * @param DateTimeZone|bool|null $zone
1053          * @param string $code
1054          * @return string
1055          */
1056         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1057                 if ( !$dateTimeObj ) {
1058                         $dateTimeObj = DateTime::createFromFormat(
1059                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1060                         );
1061                 }
1062                 return $dateTimeObj->format( $code );
1063         }
1064
1065         /**
1066          * This is a workalike of PHP's date() function, but with better
1067          * internationalisation, a reduced set of format characters, and a better
1068          * escaping format.
1069          *
1070          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1071          * the PHP manual for definitions. There are a number of extensions, which
1072          * start with "x":
1073          *
1074          *    xn   Do not translate digits of the next numeric format character
1075          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1076          *    xr   Use roman numerals for the next numeric format character
1077          *    xh   Use hebrew numerals for the next numeric format character
1078          *    xx   Literal x
1079          *    xg   Genitive month name
1080          *
1081          *    xij  j (day number) in Iranian calendar
1082          *    xiF  F (month name) in Iranian calendar
1083          *    xin  n (month number) in Iranian calendar
1084          *    xiy  y (two digit year) in Iranian calendar
1085          *    xiY  Y (full year) in Iranian calendar
1086          *
1087          *    xjj  j (day number) in Hebrew calendar
1088          *    xjF  F (month name) in Hebrew calendar
1089          *    xjt  t (days in month) in Hebrew calendar
1090          *    xjx  xg (genitive month name) in Hebrew calendar
1091          *    xjn  n (month number) in Hebrew calendar
1092          *    xjY  Y (full year) in Hebrew calendar
1093          *
1094          *    xmj  j (day number) in Hijri calendar
1095          *    xmF  F (month name) in Hijri calendar
1096          *    xmn  n (month number) in Hijri calendar
1097          *    xmY  Y (full year) in Hijri calendar
1098          *
1099          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1100          *                       identical to the Gregorian calendar
1101          *    xoY  Y (full year) in Minguo calendar or Juche year.
1102          *                       Months and days are identical to the
1103          *                       Gregorian calendar
1104          *    xtY  Y (full year) in Japanese nengo. Months and days are
1105          *                       identical to the Gregorian calendar
1106          *
1107          * Characters enclosed in double quotes will be considered literal (with
1108          * the quotes themselves removed). Unmatched quotes will be considered
1109          * literal quotes. Example:
1110          *
1111          * "The month is" F       => The month is January
1112          * i's"                   => 20'11"
1113          *
1114          * Backslash escaping is also supported.
1115          *
1116          * Input timestamp is assumed to be pre-normalized to the desired local
1117          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1118          * $ts is UTC if $zone is not given.
1119          *
1120          * @param string $format
1121          * @param string $ts 14-character timestamp
1122          *      YYYYMMDDHHMMSS
1123          *      01234567890123
1124          * @param DateTimeZone $zone Timezone of $ts
1125          * @param[out] int $ttl The amount of time (in seconds) the output may be cached for.
1126          * Only makes sense if $ts is the current time.
1127          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1128          *
1129          * @throws MWException
1130          * @return string
1131          */
1132         function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = null ) {
1133                 $s = '';
1134                 $raw = false;
1135                 $roman = false;
1136                 $hebrewNum = false;
1137                 $dateTimeObj = false;
1138                 $rawToggle = false;
1139                 $iranian = false;
1140                 $hebrew = false;
1141                 $hijri = false;
1142                 $thai = false;
1143                 $minguo = false;
1144                 $tenno = false;
1145
1146                 $usedSecond = false;
1147                 $usedMinute = false;
1148                 $usedHour = false;
1149                 $usedAMPM = false;
1150                 $usedDay = false;
1151                 $usedWeek = false;
1152                 $usedMonth = false;
1153                 $usedYear = false;
1154                 $usedISOYear = false;
1155                 $usedIsLeapYear = false;
1156
1157                 $usedHebrewMonth = false;
1158                 $usedIranianMonth = false;
1159                 $usedHijriMonth = false;
1160                 $usedHebrewYear = false;
1161                 $usedIranianYear = false;
1162                 $usedHijriYear = false;
1163                 $usedTennoYear = false;
1164
1165                 if ( strlen( $ts ) !== 14 ) {
1166                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1167                 }
1168
1169                 if ( !ctype_digit( $ts ) ) {
1170                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1171                 }
1172
1173                 $formatLength = strlen( $format );
1174                 for ( $p = 0; $p < $formatLength; $p++ ) {
1175                         $num = false;
1176                         $code = $format[$p];
1177                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1178                                 $code .= $format[++$p];
1179                         }
1180
1181                         if ( ( $code === 'xi'
1182                                         || $code === 'xj'
1183                                         || $code === 'xk'
1184                                         || $code === 'xm'
1185                                         || $code === 'xo'
1186                                         || $code === 'xt' )
1187                                 && $p < $formatLength - 1 ) {
1188                                 $code .= $format[++$p];
1189                         }
1190
1191                         switch ( $code ) {
1192                                 case 'xx':
1193                                         $s .= 'x';
1194                                         break;
1195                                 case 'xn':
1196                                         $raw = true;
1197                                         break;
1198                                 case 'xN':
1199                                         $rawToggle = !$rawToggle;
1200                                         break;
1201                                 case 'xr':
1202                                         $roman = true;
1203                                         break;
1204                                 case 'xh':
1205                                         $hebrewNum = true;
1206                                         break;
1207                                 case 'xg':
1208                                         $usedMonth = true;
1209                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1210                                         break;
1211                                 case 'xjx':
1212                                         $usedHebrewMonth = true;
1213                                         if ( !$hebrew ) {
1214                                                 $hebrew = self::tsToHebrew( $ts );
1215                                         }
1216                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1217                                         break;
1218                                 case 'd':
1219                                         $usedDay = true;
1220                                         $num = substr( $ts, 6, 2 );
1221                                         break;
1222                                 case 'D':
1223                                         $usedDay = true;
1224                                         $s .= $this->getWeekdayAbbreviation(
1225                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1226                                         );
1227                                         break;
1228                                 case 'j':
1229                                         $usedDay = true;
1230                                         $num = intval( substr( $ts, 6, 2 ) );
1231                                         break;
1232                                 case 'xij':
1233                                         $usedDay = true;
1234                                         if ( !$iranian ) {
1235                                                 $iranian = self::tsToIranian( $ts );
1236                                         }
1237                                         $num = $iranian[2];
1238                                         break;
1239                                 case 'xmj':
1240                                         $usedDay = true;
1241                                         if ( !$hijri ) {
1242                                                 $hijri = self::tsToHijri( $ts );
1243                                         }
1244                                         $num = $hijri[2];
1245                                         break;
1246                                 case 'xjj':
1247                                         $usedDay = true;
1248                                         if ( !$hebrew ) {
1249                                                 $hebrew = self::tsToHebrew( $ts );
1250                                         }
1251                                         $num = $hebrew[2];
1252                                         break;
1253                                 case 'l':
1254                                         $usedDay = true;
1255                                         $s .= $this->getWeekdayName(
1256                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1257                                         );
1258                                         break;
1259                                 case 'F':
1260                                         $usedMonth = true;
1261                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1262                                         break;
1263                                 case 'xiF':
1264                                         $usedIranianMonth = true;
1265                                         if ( !$iranian ) {
1266                                                 $iranian = self::tsToIranian( $ts );
1267                                         }
1268                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1269                                         break;
1270                                 case 'xmF':
1271                                         $usedHijriMonth = true;
1272                                         if ( !$hijri ) {
1273                                                 $hijri = self::tsToHijri( $ts );
1274                                         }
1275                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1276                                         break;
1277                                 case 'xjF':
1278                                         $usedHebrewMonth = true;
1279                                         if ( !$hebrew ) {
1280                                                 $hebrew = self::tsToHebrew( $ts );
1281                                         }
1282                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1283                                         break;
1284                                 case 'm':
1285                                         $usedMonth = true;
1286                                         $num = substr( $ts, 4, 2 );
1287                                         break;
1288                                 case 'M':
1289                                         $usedMonth = true;
1290                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1291                                         break;
1292                                 case 'n':
1293                                         $usedMonth = true;
1294                                         $num = intval( substr( $ts, 4, 2 ) );
1295                                         break;
1296                                 case 'xin':
1297                                         $usedIranianMonth = true;
1298                                         if ( !$iranian ) {
1299                                                 $iranian = self::tsToIranian( $ts );
1300                                         }
1301                                         $num = $iranian[1];
1302                                         break;
1303                                 case 'xmn':
1304                                         $usedHijriMonth = true;
1305                                         if ( !$hijri ) {
1306                                                 $hijri = self::tsToHijri( $ts );
1307                                         }
1308                                         $num = $hijri[1];
1309                                         break;
1310                                 case 'xjn':
1311                                         $usedHebrewMonth = true;
1312                                         if ( !$hebrew ) {
1313                                                 $hebrew = self::tsToHebrew( $ts );
1314                                         }
1315                                         $num = $hebrew[1];
1316                                         break;
1317                                 case 'xjt':
1318                                         $usedHebrewMonth = true;
1319                                         if ( !$hebrew ) {
1320                                                 $hebrew = self::tsToHebrew( $ts );
1321                                         }
1322                                         $num = $hebrew[3];
1323                                         break;
1324                                 case 'Y':
1325                                         $usedYear = true;
1326                                         $num = substr( $ts, 0, 4 );
1327                                         break;
1328                                 case 'xiY':
1329                                         $usedIranianYear = true;
1330                                         if ( !$iranian ) {
1331                                                 $iranian = self::tsToIranian( $ts );
1332                                         }
1333                                         $num = $iranian[0];
1334                                         break;
1335                                 case 'xmY':
1336                                         $usedHijriYear = true;
1337                                         if ( !$hijri ) {
1338                                                 $hijri = self::tsToHijri( $ts );
1339                                         }
1340                                         $num = $hijri[0];
1341                                         break;
1342                                 case 'xjY':
1343                                         $usedHebrewYear = true;
1344                                         if ( !$hebrew ) {
1345                                                 $hebrew = self::tsToHebrew( $ts );
1346                                         }
1347                                         $num = $hebrew[0];
1348                                         break;
1349                                 case 'xkY':
1350                                         $usedYear = true;
1351                                         if ( !$thai ) {
1352                                                 $thai = self::tsToYear( $ts, 'thai' );
1353                                         }
1354                                         $num = $thai[0];
1355                                         break;
1356                                 case 'xoY':
1357                                         $usedYear = true;
1358                                         if ( !$minguo ) {
1359                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1360                                         }
1361                                         $num = $minguo[0];
1362                                         break;
1363                                 case 'xtY':
1364                                         $usedTennoYear = true;
1365                                         if ( !$tenno ) {
1366                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1367                                         }
1368                                         $num = $tenno[0];
1369                                         break;
1370                                 case 'y':
1371                                         $usedYear = true;
1372                                         $num = substr( $ts, 2, 2 );
1373                                         break;
1374                                 case 'xiy':
1375                                         $usedIranianYear = true;
1376                                         if ( !$iranian ) {
1377                                                 $iranian = self::tsToIranian( $ts );
1378                                         }
1379                                         $num = substr( $iranian[0], -2 );
1380                                         break;
1381                                 case 'a':
1382                                         $usedAMPM = true;
1383                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1384                                         break;
1385                                 case 'A':
1386                                         $usedAMPM = true;
1387                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1388                                         break;
1389                                 case 'g':
1390                                         $usedHour = true;
1391                                         $h = substr( $ts, 8, 2 );
1392                                         $num = $h % 12 ? $h % 12 : 12;
1393                                         break;
1394                                 case 'G':
1395                                         $usedHour = true;
1396                                         $num = intval( substr( $ts, 8, 2 ) );
1397                                         break;
1398                                 case 'h':
1399                                         $usedHour = true;
1400                                         $h = substr( $ts, 8, 2 );
1401                                         $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
1402                                         break;
1403                                 case 'H':
1404                                         $usedHour = true;
1405                                         $num = substr( $ts, 8, 2 );
1406                                         break;
1407                                 case 'i':
1408                                         $usedMinute = true;
1409                                         $num = substr( $ts, 10, 2 );
1410                                         break;
1411                                 case 's':
1412                                         $usedSecond = true;
1413                                         $num = substr( $ts, 12, 2 );
1414                                         break;
1415                                 case 'c':
1416                                 case 'r':
1417                                         $usedSecond = true;
1418                                         // fall through
1419                                 case 'e':
1420                                 case 'O':
1421                                 case 'P':
1422                                 case 'T':
1423                                         $s .= Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1424                                         break;
1425                                 case 'w':
1426                                 case 'N':
1427                                 case 'z':
1428                                         $usedDay = true;
1429                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1430                                         break;
1431                                 case 'W':
1432                                         $usedWeek = true;
1433                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1434                                         break;
1435                                 case 't':
1436                                         $usedMonth = true;
1437                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1438                                         break;
1439                                 case 'L':
1440                                         $usedIsLeapYear = true;
1441                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1442                                         break;
1443                                 case 'o':
1444                                         $usedISOYear = true;
1445                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1446                                         break;
1447                                 case 'U':
1448                                         $usedSecond = true;
1449                                         // fall through
1450                                 case 'I':
1451                                 case 'Z':
1452                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1453                                         break;
1454                                 case '\\':
1455                                         # Backslash escaping
1456                                         if ( $p < $formatLength - 1 ) {
1457                                                 $s .= $format[++$p];
1458                                         } else {
1459                                                 $s .= '\\';
1460                                         }
1461                                         break;
1462                                 case '"':
1463                                         # Quoted literal
1464                                         if ( $p < $formatLength - 1 ) {
1465                                                 $endQuote = strpos( $format, '"', $p + 1 );
1466                                                 if ( $endQuote === false ) {
1467                                                         # No terminating quote, assume literal "
1468                                                         $s .= '"';
1469                                                 } else {
1470                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1471                                                         $p = $endQuote;
1472                                                 }
1473                                         } else {
1474                                                 # Quote at end of string, assume literal "
1475                                                 $s .= '"';
1476                                         }
1477                                         break;
1478                                 default:
1479                                         $s .= $format[$p];
1480                         }
1481                         if ( $num !== false ) {
1482                                 if ( $rawToggle || $raw ) {
1483                                         $s .= $num;
1484                                         $raw = false;
1485                                 } elseif ( $roman ) {
1486                                         $s .= Language::romanNumeral( $num );
1487                                         $roman = false;
1488                                 } elseif ( $hebrewNum ) {
1489                                         $s .= self::hebrewNumeral( $num );
1490                                         $hebrewNum = false;
1491                                 } else {
1492                                         $s .= $this->formatNum( $num, true );
1493                                 }
1494                         }
1495                 }
1496
1497                 if ( $usedSecond ) {
1498                         $ttl = 1;
1499                 } elseif ( $usedMinute ) {
1500                         $ttl = 60 - substr( $ts, 12, 2 );
1501                 } elseif ( $usedHour ) {
1502                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1503                 } elseif ( $usedAMPM ) {
1504                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1505                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1506                 } elseif (
1507                         $usedDay ||
1508                         $usedHebrewMonth ||
1509                         $usedIranianMonth ||
1510                         $usedHijriMonth ||
1511                         $usedHebrewYear ||
1512                         $usedIranianYear ||
1513                         $usedHijriYear ||
1514                         $usedTennoYear
1515                 ) {
1516                         // @todo Someone who understands the non-Gregorian calendars
1517                         // should write proper logic for them so that they don't need purged every day.
1518                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1519                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1520                 } else {
1521                         $possibleTtls = [];
1522                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1523                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1524                         if ( $usedWeek ) {
1525                                 $possibleTtls[] =
1526                                         ( 7 - Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1527                                         $timeRemainingInDay;
1528                         } elseif ( $usedISOYear ) {
1529                                 // December 28th falls on the last ISO week of the year, every year.
1530                                 // The last ISO week of a year can be 52 or 53.
1531                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1532                                         'Ymd',
1533                                         substr( $ts, 0, 4 ) . '1228',
1534                                         $zone ?: new DateTimeZone( 'UTC' )
1535                                 )->format( 'W' );
1536                                 $currentISOWeek = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1537                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1538                                 $timeRemainingInWeek =
1539                                         ( 7 - Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1540                                         + $timeRemainingInDay;
1541                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1542                         }
1543
1544                         if ( $usedMonth ) {
1545                                 $possibleTtls[] =
1546                                         ( Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1547                                                 substr( $ts, 6, 2 ) ) * 86400
1548                                         + $timeRemainingInDay;
1549                         } elseif ( $usedYear ) {
1550                                 $possibleTtls[] =
1551                                         ( Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1552                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1553                                         + $timeRemainingInDay;
1554                         } elseif ( $usedIsLeapYear ) {
1555                                 $year = substr( $ts, 0, 4 );
1556                                 $timeRemainingInYear =
1557                                         ( Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1558                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1559                                         + $timeRemainingInDay;
1560                                 $mod = $year % 4;
1561                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1562                                         // this isn't a leap year. see when the next one starts
1563                                         $nextCandidate = $year - $mod + 4;
1564                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1565                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1566                                                         $timeRemainingInYear;
1567                                         } else {
1568                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1569                                                         $timeRemainingInYear;
1570                                         }
1571                                 } else {
1572                                         // this is a leap year, so the next year isn't
1573                                         $possibleTtls[] = $timeRemainingInYear;
1574                                 }
1575                         }
1576
1577                         if ( $possibleTtls ) {
1578                                 $ttl = min( $possibleTtls );
1579                         }
1580                 }
1581
1582                 return $s;
1583         }
1584
1585         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1586         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1587
1588         /**
1589          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1590          * Gregorian dates to Iranian dates. Originally written in C, it
1591          * is released under the terms of GNU Lesser General Public
1592          * License. Conversion to PHP was performed by Niklas Laxström.
1593          *
1594          * Link: http://www.farsiweb.info/jalali/jalali.c
1595          *
1596          * @param string $ts
1597          *
1598          * @return string
1599          */
1600         private static function tsToIranian( $ts ) {
1601                 $gy = substr( $ts, 0, 4 ) -1600;
1602                 $gm = substr( $ts, 4, 2 ) -1;
1603                 $gd = substr( $ts, 6, 2 ) -1;
1604
1605                 # Days passed from the beginning (including leap years)
1606                 $gDayNo = 365 * $gy
1607                         + floor( ( $gy + 3 ) / 4 )
1608                         - floor( ( $gy + 99 ) / 100 )
1609                         + floor( ( $gy + 399 ) / 400 );
1610
1611                 // Add days of the past months of this year
1612                 for ( $i = 0; $i < $gm; $i++ ) {
1613                         $gDayNo += self::$GREG_DAYS[$i];
1614                 }
1615
1616                 // Leap years
1617                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1618                         $gDayNo++;
1619                 }
1620
1621                 // Days passed in current month
1622                 $gDayNo += (int)$gd;
1623
1624                 $jDayNo = $gDayNo - 79;
1625
1626                 $jNp = floor( $jDayNo / 12053 );
1627                 $jDayNo %= 12053;
1628
1629                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1630                 $jDayNo %= 1461;
1631
1632                 if ( $jDayNo >= 366 ) {
1633                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1634                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1635                 }
1636
1637                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1638                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1639                 }
1640
1641                 $jm = $i + 1;
1642                 $jd = $jDayNo + 1;
1643
1644                 return [ $jy, $jm, $jd ];
1645         }
1646
1647         /**
1648          * Converting Gregorian dates to Hijri dates.
1649          *
1650          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1651          *
1652          * @see http://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1653          *
1654          * @param string $ts
1655          *
1656          * @return string
1657          */
1658         private static function tsToHijri( $ts ) {
1659                 $year = substr( $ts, 0, 4 );
1660                 $month = substr( $ts, 4, 2 );
1661                 $day = substr( $ts, 6, 2 );
1662
1663                 $zyr = $year;
1664                 $zd = $day;
1665                 $zm = $month;
1666                 $zy = $zyr;
1667
1668                 if (
1669                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1670                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1671                 ) {
1672                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1673                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1674                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1675                                         $zd - 32075;
1676                 } else {
1677                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1678                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1679                 }
1680
1681                 $zl = $zjd -1948440 + 10632;
1682                 $zn = (int)( ( $zl - 1 ) / 10631 );
1683                 $zl = $zl - 10631 * $zn + 354;
1684                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1685                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1686                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1687                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1688                 $zm = (int)( ( 24 * $zl ) / 709 );
1689                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1690                 $zy = 30 * $zn + $zj - 30;
1691
1692                 return [ $zy, $zm, $zd ];
1693         }
1694
1695         /**
1696          * Converting Gregorian dates to Hebrew dates.
1697          *
1698          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1699          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1700          * to translate the relevant functions into PHP and release them under
1701          * GNU GPL.
1702          *
1703          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1704          * and Adar II is 14. In a non-leap year, Adar is 6.
1705          *
1706          * @param string $ts
1707          *
1708          * @return string
1709          */
1710         private static function tsToHebrew( $ts ) {
1711                 # Parse date
1712                 $year = substr( $ts, 0, 4 );
1713                 $month = substr( $ts, 4, 2 );
1714                 $day = substr( $ts, 6, 2 );
1715
1716                 # Calculate Hebrew year
1717                 $hebrewYear = $year + 3760;
1718
1719                 # Month number when September = 1, August = 12
1720                 $month += 4;
1721                 if ( $month > 12 ) {
1722                         # Next year
1723                         $month -= 12;
1724                         $year++;
1725                         $hebrewYear++;
1726                 }
1727
1728                 # Calculate day of year from 1 September
1729                 $dayOfYear = $day;
1730                 for ( $i = 1; $i < $month; $i++ ) {
1731                         if ( $i == 6 ) {
1732                                 # February
1733                                 $dayOfYear += 28;
1734                                 # Check if the year is leap
1735                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1736                                         $dayOfYear++;
1737                                 }
1738                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1739                                 $dayOfYear += 30;
1740                         } else {
1741                                 $dayOfYear += 31;
1742                         }
1743                 }
1744
1745                 # Calculate the start of the Hebrew year
1746                 $start = self::hebrewYearStart( $hebrewYear );
1747
1748                 # Calculate next year's start
1749                 if ( $dayOfYear <= $start ) {
1750                         # Day is before the start of the year - it is the previous year
1751                         # Next year's start
1752                         $nextStart = $start;
1753                         # Previous year
1754                         $year--;
1755                         $hebrewYear--;
1756                         # Add days since previous year's 1 September
1757                         $dayOfYear += 365;
1758                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1759                                 # Leap year
1760                                 $dayOfYear++;
1761                         }
1762                         # Start of the new (previous) year
1763                         $start = self::hebrewYearStart( $hebrewYear );
1764                 } else {
1765                         # Next year's start
1766                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1767                 }
1768
1769                 # Calculate Hebrew day of year
1770                 $hebrewDayOfYear = $dayOfYear - $start;
1771
1772                 # Difference between year's days
1773                 $diff = $nextStart - $start;
1774                 # Add 12 (or 13 for leap years) days to ignore the difference between
1775                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1776                 # difference is only about the year type
1777                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1778                         $diff += 13;
1779                 } else {
1780                         $diff += 12;
1781                 }
1782
1783                 # Check the year pattern, and is leap year
1784                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1785                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1786                 # and non-leap years
1787                 $yearPattern = $diff % 30;
1788                 # Check if leap year
1789                 $isLeap = $diff >= 30;
1790
1791                 # Calculate day in the month from number of day in the Hebrew year
1792                 # Don't check Adar - if the day is not in Adar, we will stop before;
1793                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1794                 $hebrewDay = $hebrewDayOfYear;
1795                 $hebrewMonth = 1;
1796                 $days = 0;
1797                 while ( $hebrewMonth <= 12 ) {
1798                         # Calculate days in this month
1799                         if ( $isLeap && $hebrewMonth == 6 ) {
1800                                 # Adar in a leap year
1801                                 if ( $isLeap ) {
1802                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1803                                         $days = 30;
1804                                         if ( $hebrewDay <= $days ) {
1805                                                 # Day in Adar I
1806                                                 $hebrewMonth = 13;
1807                                         } else {
1808                                                 # Subtract the days of Adar I
1809                                                 $hebrewDay -= $days;
1810                                                 # Try Adar II
1811                                                 $days = 29;
1812                                                 if ( $hebrewDay <= $days ) {
1813                                                         # Day in Adar II
1814                                                         $hebrewMonth = 14;
1815                                                 }
1816                                         }
1817                                 }
1818                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1819                                 # Cheshvan in a complete year (otherwise as the rule below)
1820                                 $days = 30;
1821                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1822                                 # Kislev in an incomplete year (otherwise as the rule below)
1823                                 $days = 29;
1824                         } else {
1825                                 # Odd months have 30 days, even have 29
1826                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1827                         }
1828                         if ( $hebrewDay <= $days ) {
1829                                 # In the current month
1830                                 break;
1831                         } else {
1832                                 # Subtract the days of the current month
1833                                 $hebrewDay -= $days;
1834                                 # Try in the next month
1835                                 $hebrewMonth++;
1836                         }
1837                 }
1838
1839                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1840         }
1841
1842         /**
1843          * This calculates the Hebrew year start, as days since 1 September.
1844          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1845          * Used for Hebrew date.
1846          *
1847          * @param int $year
1848          *
1849          * @return string
1850          */
1851         private static function hebrewYearStart( $year ) {
1852                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1853                 $b = intval( ( $year - 1 ) % 4 );
1854                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1855                 if ( $m < 0 ) {
1856                         $m--;
1857                 }
1858                 $Mar = intval( $m );
1859                 if ( $m < 0 ) {
1860                         $m++;
1861                 }
1862                 $m -= $Mar;
1863
1864                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1865                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1866                         $Mar++;
1867                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1868                         $Mar += 2;
1869                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1870                         $Mar++;
1871                 }
1872
1873                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1874                 return $Mar;
1875         }
1876
1877         /**
1878          * Algorithm to convert Gregorian dates to Thai solar dates,
1879          * Minguo dates or Minguo dates.
1880          *
1881          * Link: http://en.wikipedia.org/wiki/Thai_solar_calendar
1882          *       http://en.wikipedia.org/wiki/Minguo_calendar
1883          *       http://en.wikipedia.org/wiki/Japanese_era_name
1884          *
1885          * @param string $ts 14-character timestamp
1886          * @param string $cName Calender name
1887          * @return array Converted year, month, day
1888          */
1889         private static function tsToYear( $ts, $cName ) {
1890                 $gy = substr( $ts, 0, 4 );
1891                 $gm = substr( $ts, 4, 2 );
1892                 $gd = substr( $ts, 6, 2 );
1893
1894                 if ( !strcmp( $cName, 'thai' ) ) {
1895                         # Thai solar dates
1896                         # Add 543 years to the Gregorian calendar
1897                         # Months and days are identical
1898                         $gy_offset = $gy + 543;
1899                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1900                         # Minguo dates
1901                         # Deduct 1911 years from the Gregorian calendar
1902                         # Months and days are identical
1903                         $gy_offset = $gy - 1911;
1904                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1905                         # Nengō dates up to Meiji period
1906                         # Deduct years from the Gregorian calendar
1907                         # depending on the nengo periods
1908                         # Months and days are identical
1909                         if ( ( $gy < 1912 )
1910                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1911                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1912                         ) {
1913                                 # Meiji period
1914                                 $gy_gannen = $gy - 1868 + 1;
1915                                 $gy_offset = $gy_gannen;
1916                                 if ( $gy_gannen == 1 ) {
1917                                         $gy_offset = '元';
1918                                 }
1919                                 $gy_offset = '明治' . $gy_offset;
1920                         } elseif (
1921                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1922                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1923                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1924                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1925                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1926                         ) {
1927                                 # Taishō period
1928                                 $gy_gannen = $gy - 1912 + 1;
1929                                 $gy_offset = $gy_gannen;
1930                                 if ( $gy_gannen == 1 ) {
1931                                         $gy_offset = '元';
1932                                 }
1933                                 $gy_offset = '大正' . $gy_offset;
1934                         } elseif (
1935                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1936                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1937                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1938                         ) {
1939                                 # Shōwa period
1940                                 $gy_gannen = $gy - 1926 + 1;
1941                                 $gy_offset = $gy_gannen;
1942                                 if ( $gy_gannen == 1 ) {
1943                                         $gy_offset = '元';
1944                                 }
1945                                 $gy_offset = '昭和' . $gy_offset;
1946                         } else {
1947                                 # Heisei period
1948                                 $gy_gannen = $gy - 1989 + 1;
1949                                 $gy_offset = $gy_gannen;
1950                                 if ( $gy_gannen == 1 ) {
1951                                         $gy_offset = '元';
1952                                 }
1953                                 $gy_offset = '平成' . $gy_offset;
1954                         }
1955                 } else {
1956                         $gy_offset = $gy;
1957                 }
1958
1959                 return [ $gy_offset, $gm, $gd ];
1960         }
1961
1962         /**
1963          * Gets directionality of the first strongly directional codepoint, for embedBidi()
1964          *
1965          * This is the rule the BIDI algorithm uses to determine the directionality of
1966          * paragraphs ( http://unicode.org/reports/tr9/#The_Paragraph_Level ) and
1967          * FSI isolates ( http://unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
1968          *
1969          * TODO: Does not handle BIDI control characters inside the text.
1970          * TODO: Does not handle unallocated characters.
1971          *
1972          * @param string $text Text to test
1973          * @return null|string Directionality ('ltr' or 'rtl') or null
1974          */
1975         private static function strongDirFromContent( $text = '' ) {
1976                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
1977                         return null;
1978                 }
1979                 if ( $matches[1] === '' ) {
1980                         return 'rtl';
1981                 }
1982                 return 'ltr';
1983         }
1984
1985         /**
1986          * Roman number formatting up to 10000
1987          *
1988          * @param int $num
1989          *
1990          * @return string
1991          */
1992         static function romanNumeral( $num ) {
1993                 static $table = [
1994                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
1995                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
1996                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
1997                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
1998                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
1999                 ];
2000
2001                 $num = intval( $num );
2002                 if ( $num > 10000 || $num <= 0 ) {
2003                         return $num;
2004                 }
2005
2006                 $s = '';
2007                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2008                         if ( $num >= $pow10 ) {
2009                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2010                         }
2011                         $num = $num % $pow10;
2012                 }
2013                 return $s;
2014         }
2015
2016         /**
2017          * Hebrew Gematria number formatting up to 9999
2018          *
2019          * @param int $num
2020          *
2021          * @return string
2022          */
2023         static function hebrewNumeral( $num ) {
2024                 static $table = [
2025                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2026                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2027                         [ '',
2028                                 [ 'ק' ],
2029                                 [ 'ר' ],
2030                                 [ 'ש' ],
2031                                 [ 'ת' ],
2032                                 [ 'ת', 'ק' ],
2033                                 [ 'ת', 'ר' ],
2034                                 [ 'ת', 'ש' ],
2035                                 [ 'ת', 'ת' ],
2036                                 [ 'ת', 'ת', 'ק' ],
2037                                 [ 'ת', 'ת', 'ר' ],
2038                         ],
2039                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2040                 ];
2041
2042                 $num = intval( $num );
2043                 if ( $num > 9999 || $num <= 0 ) {
2044                         return $num;
2045                 }
2046
2047                 // Round thousands have special notations
2048                 if ( $num === 1000 ) {
2049                         return "א' אלף";
2050                 } elseif ( $num % 1000 === 0 ) {
2051                         return $table[0][$num / 1000] . "' אלפים";
2052                 }
2053
2054                 $letters = [];
2055
2056                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2057                         if ( $num >= $pow10 ) {
2058                                 if ( $num === 15 || $num === 16 ) {
2059                                         $letters[] = $table[0][9];
2060                                         $letters[] = $table[0][$num - 9];
2061                                         $num = 0;
2062                                 } else {
2063                                         $letters = array_merge(
2064                                                 $letters,
2065                                                 (array)$table[$i][intval( $num / $pow10 )]
2066                                         );
2067
2068                                         if ( $pow10 === 1000 ) {
2069                                                 $letters[] = "'";
2070                                         }
2071                                 }
2072                         }
2073
2074                         $num = $num % $pow10;
2075                 }
2076
2077                 $preTransformLength = count( $letters );
2078                 if ( $preTransformLength === 1 ) {
2079                         // Add geresh (single quote) to one-letter numbers
2080                         $letters[] = "'";
2081                 } else {
2082                         $lastIndex = $preTransformLength - 1;
2083                         $letters[$lastIndex] = str_replace(
2084                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2085                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2086                                 $letters[$lastIndex]
2087                         );
2088
2089                         // Add gershayim (double quote) to multiple-letter numbers,
2090                         // but exclude numbers with only one letter after the thousands
2091                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2092                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2093                                 $letters[] = "'";
2094                         } else {
2095                                 array_splice( $letters, -1, 0, '"' );
2096                         }
2097                 }
2098
2099                 return implode( $letters );
2100         }
2101
2102         /**
2103          * Used by date() and time() to adjust the time output.
2104          *
2105          * @param string $ts The time in date('YmdHis') format
2106          * @param mixed $tz Adjust the time by this amount (default false, mean we
2107          *   get user timecorrection setting)
2108          * @return int
2109          */
2110         function userAdjust( $ts, $tz = false ) {
2111                 global $wgUser, $wgLocalTZoffset;
2112
2113                 if ( $tz === false ) {
2114                         $tz = $wgUser->getOption( 'timecorrection' );
2115                 }
2116
2117                 $data = explode( '|', $tz, 3 );
2118
2119                 if ( $data[0] == 'ZoneInfo' ) {
2120                         MediaWiki\suppressWarnings();
2121                         $userTZ = timezone_open( $data[2] );
2122                         MediaWiki\restoreWarnings();
2123                         if ( $userTZ !== false ) {
2124                                 $date = date_create( $ts, timezone_open( 'UTC' ) );
2125                                 date_timezone_set( $date, $userTZ );
2126                                 $date = date_format( $date, 'YmdHis' );
2127                                 return $date;
2128                         }
2129                         # Unrecognized timezone, default to 'Offset' with the stored offset.
2130                         $data[0] = 'Offset';
2131                 }
2132
2133                 if ( $data[0] == 'System' || $tz == '' ) {
2134                         # Global offset in minutes.
2135                         $minDiff = $wgLocalTZoffset;
2136                 } elseif ( $data[0] == 'Offset' ) {
2137                         $minDiff = intval( $data[1] );
2138                 } else {
2139                         $data = explode( ':', $tz );
2140                         if ( count( $data ) == 2 ) {
2141                                 $data[0] = intval( $data[0] );
2142                                 $data[1] = intval( $data[1] );
2143                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2144                                 if ( $data[0] < 0 ) {
2145                                         $minDiff = -$minDiff;
2146                                 }
2147                         } else {
2148                                 $minDiff = intval( $data[0] ) * 60;
2149                         }
2150                 }
2151
2152                 # No difference ? Return time unchanged
2153                 if ( 0 == $minDiff ) {
2154                         return $ts;
2155                 }
2156
2157                 MediaWiki\suppressWarnings(); // E_STRICT system time bitching
2158                 # Generate an adjusted date; take advantage of the fact that mktime
2159                 # will normalize out-of-range values so we don't have to split $minDiff
2160                 # into hours and minutes.
2161                 $t = mktime( (
2162                         (int)substr( $ts, 8, 2 ) ), # Hours
2163                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2164                         (int)substr( $ts, 12, 2 ), # Seconds
2165                         (int)substr( $ts, 4, 2 ), # Month
2166                         (int)substr( $ts, 6, 2 ), # Day
2167                         (int)substr( $ts, 0, 4 ) ); # Year
2168
2169                 $date = date( 'YmdHis', $t );
2170                 MediaWiki\restoreWarnings();
2171
2172                 return $date;
2173         }
2174
2175         /**
2176          * This is meant to be used by time(), date(), and timeanddate() to get
2177          * the date preference they're supposed to use, it should be used in
2178          * all children.
2179          *
2180          *<code>
2181          * function timeanddate([...], $format = true) {
2182          *      $datePreference = $this->dateFormat($format);
2183          * [...]
2184          * }
2185          *</code>
2186          *
2187          * @param int|string|bool $usePrefs If true, the user's preference is used
2188          *   if false, the site/language default is used
2189          *   if int/string, assumed to be a format.
2190          * @return string
2191          */
2192         function dateFormat( $usePrefs = true ) {
2193                 global $wgUser;
2194
2195                 if ( is_bool( $usePrefs ) ) {
2196                         if ( $usePrefs ) {
2197                                 $datePreference = $wgUser->getDatePreference();
2198                         } else {
2199                                 $datePreference = (string)User::getDefaultOption( 'date' );
2200                         }
2201                 } else {
2202                         $datePreference = (string)$usePrefs;
2203                 }
2204
2205                 // return int
2206                 if ( $datePreference == '' ) {
2207                         return 'default';
2208                 }
2209
2210                 return $datePreference;
2211         }
2212
2213         /**
2214          * Get a format string for a given type and preference
2215          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2216          * @param string $pref The format name as it appears in Messages*.php under
2217          *  $datePreferences.
2218          *
2219          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2220          *
2221          * @return string
2222          */
2223         function getDateFormatString( $type, $pref ) {
2224                 $wasDefault = false;
2225                 if ( $pref == 'default' ) {
2226                         $wasDefault = true;
2227                         $pref = $this->getDefaultDateFormat();
2228                 }
2229
2230                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2231                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2232
2233                         if ( $type === 'pretty' && $df === null ) {
2234                                 $df = $this->getDateFormatString( 'date', $pref );
2235                         }
2236
2237                         if ( !$wasDefault && $df === null ) {
2238                                 $pref = $this->getDefaultDateFormat();
2239                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2240                         }
2241
2242                         $this->dateFormatStrings[$type][$pref] = $df;
2243                 }
2244                 return $this->dateFormatStrings[$type][$pref];
2245         }
2246
2247         /**
2248          * @param string $ts The time format which needs to be turned into a
2249          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2250          * @param bool $adj Whether to adjust the time output according to the
2251          *   user configured offset ($timecorrection)
2252          * @param mixed $format True to use user's date format preference
2253          * @param string|bool $timecorrection The time offset as returned by
2254          *   validateTimeZone() in Special:Preferences
2255          * @return string
2256          */
2257         function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2258                 $ts = wfTimestamp( TS_MW, $ts );
2259                 if ( $adj ) {
2260                         $ts = $this->userAdjust( $ts, $timecorrection );
2261                 }
2262                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2263                 return $this->sprintfDate( $df, $ts );
2264         }
2265
2266         /**
2267          * @param string $ts The time format which needs to be turned into a
2268          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2269          * @param bool $adj Whether to adjust the time output according to the
2270          *   user configured offset ($timecorrection)
2271          * @param mixed $format True to use user's date format preference
2272          * @param string|bool $timecorrection The time offset as returned by
2273          *   validateTimeZone() in Special:Preferences
2274          * @return string
2275          */
2276         function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2277                 $ts = wfTimestamp( TS_MW, $ts );
2278                 if ( $adj ) {
2279                         $ts = $this->userAdjust( $ts, $timecorrection );
2280                 }
2281                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2282                 return $this->sprintfDate( $df, $ts );
2283         }
2284
2285         /**
2286          * @param string $ts The time format which needs to be turned into a
2287          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2288          * @param bool $adj Whether to adjust the time output according to the
2289          *   user configured offset ($timecorrection)
2290          * @param mixed $format What format to return, if it's false output the
2291          *   default one (default true)
2292          * @param string|bool $timecorrection The time offset as returned by
2293          *   validateTimeZone() in Special:Preferences
2294          * @return string
2295          */
2296         function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2297                 $ts = wfTimestamp( TS_MW, $ts );
2298                 if ( $adj ) {
2299                         $ts = $this->userAdjust( $ts, $timecorrection );
2300                 }
2301                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2302                 return $this->sprintfDate( $df, $ts );
2303         }
2304
2305         /**
2306          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2307          *
2308          * @since 1.20
2309          *
2310          * @param int $seconds The amount of seconds.
2311          * @param array $chosenIntervals The intervals to enable.
2312          *
2313          * @return string
2314          */
2315         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2316                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2317
2318                 $segments = [];
2319
2320                 foreach ( $intervals as $intervalName => $intervalValue ) {
2321                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2322                         // duration-years, duration-decades, duration-centuries, duration-millennia
2323                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2324                         $segments[] = $message->inLanguage( $this )->escaped();
2325                 }
2326
2327                 return $this->listToText( $segments );
2328         }
2329
2330         /**
2331          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2332          * For example 65 will be turned into array( minutes => 1, seconds => 5 ).
2333          *
2334          * @since 1.20
2335          *
2336          * @param int $seconds The amount of seconds.
2337          * @param array $chosenIntervals The intervals to enable.
2338          *
2339          * @return array
2340          */
2341         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2342                 if ( empty( $chosenIntervals ) ) {
2343                         $chosenIntervals = [
2344                                 'millennia',
2345                                 'centuries',
2346                                 'decades',
2347                                 'years',
2348                                 'days',
2349                                 'hours',
2350                                 'minutes',
2351                                 'seconds'
2352                         ];
2353                 }
2354
2355                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2356                 $sortedNames = array_keys( $intervals );
2357                 $smallestInterval = array_pop( $sortedNames );
2358
2359                 $segments = [];
2360
2361                 foreach ( $intervals as $name => $length ) {
2362                         $value = floor( $seconds / $length );
2363
2364                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2365                                 $seconds -= $value * $length;
2366                                 $segments[$name] = $value;
2367                         }
2368                 }
2369
2370                 return $segments;
2371         }
2372
2373         /**
2374          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2375          *
2376          * @param string $type Can be 'date', 'time' or 'both'
2377          * @param string $ts The time format which needs to be turned into a
2378          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2379          * @param User $user User object used to get preferences for timezone and format
2380          * @param array $options Array, can contain the following keys:
2381          *   - 'timecorrection': time correction, can have the following values:
2382          *     - true: use user's preference
2383          *     - false: don't use time correction
2384          *     - int: value of time correction in minutes
2385          *   - 'format': format to use, can have the following values:
2386          *     - true: use user's preference
2387          *     - false: use default preference
2388          *     - string: format to use
2389          * @since 1.19
2390          * @return string
2391          */
2392         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2393                 $ts = wfTimestamp( TS_MW, $ts );
2394                 $options += [ 'timecorrection' => true, 'format' => true ];
2395                 if ( $options['timecorrection'] !== false ) {
2396                         if ( $options['timecorrection'] === true ) {
2397                                 $offset = $user->getOption( 'timecorrection' );
2398                         } else {
2399                                 $offset = $options['timecorrection'];
2400                         }
2401                         $ts = $this->userAdjust( $ts, $offset );
2402                 }
2403                 if ( $options['format'] === true ) {
2404                         $format = $user->getDatePreference();
2405                 } else {
2406                         $format = $options['format'];
2407                 }
2408                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2409                 return $this->sprintfDate( $df, $ts );
2410         }
2411
2412         /**
2413          * Get the formatted date for the given timestamp and formatted for
2414          * the given user.
2415          *
2416          * @param mixed $ts Mixed: the time format which needs to be turned into a
2417          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2418          * @param User $user User object used to get preferences for timezone and format
2419          * @param array $options Array, can contain the following keys:
2420          *   - 'timecorrection': time correction, can have the following values:
2421          *     - true: use user's preference
2422          *     - false: don't use time correction
2423          *     - int: value of time correction in minutes
2424          *   - 'format': format to use, can have the following values:
2425          *     - true: use user's preference
2426          *     - false: use default preference
2427          *     - string: format to use
2428          * @since 1.19
2429          * @return string
2430          */
2431         public function userDate( $ts, User $user, array $options = [] ) {
2432                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2433         }
2434
2435         /**
2436          * Get the formatted time for the given timestamp and formatted for
2437          * the given user.
2438          *
2439          * @param mixed $ts The time format which needs to be turned into a
2440          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2441          * @param User $user User object used to get preferences for timezone and format
2442          * @param array $options Array, can contain the following keys:
2443          *   - 'timecorrection': time correction, can have the following values:
2444          *     - true: use user's preference
2445          *     - false: don't use time correction
2446          *     - int: value of time correction in minutes
2447          *   - 'format': format to use, can have the following values:
2448          *     - true: use user's preference
2449          *     - false: use default preference
2450          *     - string: format to use
2451          * @since 1.19
2452          * @return string
2453          */
2454         public function userTime( $ts, User $user, array $options = [] ) {
2455                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2456         }
2457
2458         /**
2459          * Get the formatted date and time for the given timestamp and formatted for
2460          * the given user.
2461          *
2462          * @param mixed $ts The time format which needs to be turned into a
2463          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2464          * @param User $user User object used to get preferences for timezone and format
2465          * @param array $options Array, can contain the following keys:
2466          *   - 'timecorrection': time correction, can have the following values:
2467          *     - true: use user's preference
2468          *     - false: don't use time correction
2469          *     - int: value of time correction in minutes
2470          *   - 'format': format to use, can have the following values:
2471          *     - true: use user's preference
2472          *     - false: use default preference
2473          *     - string: format to use
2474          * @since 1.19
2475          * @return string
2476          */
2477         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2478                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2479         }
2480
2481         /**
2482          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2483          *
2484          * Determine the difference between the timestamp and the current time, and
2485          * generate a readable timestamp by returning "<N> <units> ago", where the
2486          * largest possible unit is used.
2487          *
2488          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2489          *
2490          * @param MWTimestamp $time
2491          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2492          * @param User|null $user User the timestamp is being generated for
2493          *  (or null to use main context's user)
2494          * @return string Formatted timestamp
2495          */
2496         public function getHumanTimestamp(
2497                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2498         ) {
2499                 if ( $relativeTo === null ) {
2500                         $relativeTo = new MWTimestamp();
2501                 }
2502                 if ( $user === null ) {
2503                         $user = RequestContext::getMain()->getUser();
2504                 }
2505
2506                 // Adjust for the user's timezone.
2507                 $offsetThis = $time->offsetForUser( $user );
2508                 $offsetRel = $relativeTo->offsetForUser( $user );
2509
2510                 $ts = '';
2511                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2512                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2513                 }
2514
2515                 // Reset the timezone on the objects.
2516                 $time->timestamp->sub( $offsetThis );
2517                 $relativeTo->timestamp->sub( $offsetRel );
2518
2519                 return $ts;
2520         }
2521
2522         /**
2523          * Convert an MWTimestamp into a pretty human-readable timestamp using
2524          * the given user preferences and relative base time.
2525          *
2526          * @see Language::getHumanTimestamp
2527          * @param MWTimestamp $ts Timestamp to prettify
2528          * @param MWTimestamp $relativeTo Base timestamp
2529          * @param User $user User preferences to use
2530          * @return string Human timestamp
2531          * @since 1.26
2532          */
2533         private function getHumanTimestampInternal(
2534                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2535         ) {
2536                 $diff = $ts->diff( $relativeTo );
2537                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2538                         (int)$relativeTo->timestamp->format( 'w' ) );
2539                 $days = $diff->days ?: (int)$diffDay;
2540                 if ( $diff->invert || $days > 5
2541                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2542                 ) {
2543                         // Timestamps are in different years: use full timestamp
2544                         // Also do full timestamp for future dates
2545                         /**
2546                          * @todo FIXME: Add better handling of future timestamps.
2547                          */
2548                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2549                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2550                 } elseif ( $days > 5 ) {
2551                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2552                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2553                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2554                 } elseif ( $days > 1 ) {
2555                         // Timestamp within the past week: show the day of the week and time
2556                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2557                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2558                         // Messages:
2559                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2560                         $ts = wfMessage( "$weekday-at" )
2561                                 ->inLanguage( $this )
2562                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2563                                 ->text();
2564                 } elseif ( $days == 1 ) {
2565                         // Timestamp was yesterday: say 'yesterday' and the time.
2566                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2567                         $ts = wfMessage( 'yesterday-at' )
2568                                 ->inLanguage( $this )
2569                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2570                                 ->text();
2571                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2572                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2573                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2574                         $ts = wfMessage( 'today-at' )
2575                                 ->inLanguage( $this )
2576                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2577                                 ->text();
2578
2579                 // From here on in, the timestamp was soon enough ago so that we can simply say
2580                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2581                 } elseif ( $diff->h == 1 ) {
2582                         // Less than 90 minutes, but more than an hour ago.
2583                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2584                 } elseif ( $diff->i >= 1 ) {
2585                         // A few minutes ago.
2586                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2587                 } elseif ( $diff->s >= 30 ) {
2588                         // Less than a minute, but more than 30 sec ago.
2589                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2590                 } else {
2591                         // Less than 30 seconds ago.
2592                         $ts = wfMessage( 'just-now' )->text();
2593                 }
2594
2595                 return $ts;
2596         }
2597
2598         /**
2599          * @param string $key
2600          * @return array|null
2601          */
2602         function getMessage( $key ) {
2603                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2604         }
2605
2606         /**
2607          * @return array
2608          */
2609         function getAllMessages() {
2610                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2611         }
2612
2613         /**
2614          * @param string $in
2615          * @param string $out
2616          * @param string $string
2617          * @return string
2618          */
2619         function iconv( $in, $out, $string ) {
2620                 # This is a wrapper for iconv in all languages except esperanto,
2621                 # which does some nasty x-conversions beforehand
2622
2623                 # Even with //IGNORE iconv can whine about illegal characters in
2624                 # *input* string. We just ignore those too.
2625                 # REF: http://bugs.php.net/bug.php?id=37166
2626                 # REF: https://phabricator.wikimedia.org/T18885
2627                 MediaWiki\suppressWarnings();
2628                 $text = iconv( $in, $out . '//IGNORE', $string );
2629                 MediaWiki\restoreWarnings();
2630                 return $text;
2631         }
2632
2633         // callback functions for uc(), lc(), ucwords(), ucwordbreaks()
2634
2635         /**
2636          * @param array $matches
2637          * @return mixed|string
2638          */
2639         function ucwordbreaksCallbackAscii( $matches ) {
2640                 return $this->ucfirst( $matches[1] );
2641         }
2642
2643         /**
2644          * @param array $matches
2645          * @return string
2646          */
2647         function ucwordbreaksCallbackMB( $matches ) {
2648                 return mb_strtoupper( $matches[0] );
2649         }
2650
2651         /**
2652          * @param array $matches
2653          * @return string
2654          */
2655         function ucCallback( $matches ) {
2656                 list( $wikiUpperChars ) = self::getCaseMaps();
2657                 return strtr( $matches[1], $wikiUpperChars );
2658         }
2659
2660         /**
2661          * @param array $matches
2662          * @return string
2663          */
2664         function lcCallback( $matches ) {
2665                 list( , $wikiLowerChars ) = self::getCaseMaps();
2666                 return strtr( $matches[1], $wikiLowerChars );
2667         }
2668
2669         /**
2670          * @param array $matches
2671          * @return string
2672          */
2673         function ucwordsCallbackMB( $matches ) {
2674                 return mb_strtoupper( $matches[0] );
2675         }
2676
2677         /**
2678          * @param array $matches
2679          * @return string
2680          */
2681         function ucwordsCallbackWiki( $matches ) {
2682                 list( $wikiUpperChars ) = self::getCaseMaps();
2683                 return strtr( $matches[0], $wikiUpperChars );
2684         }
2685
2686         /**
2687          * Make a string's first character uppercase
2688          *
2689          * @param string $str
2690          *
2691          * @return string
2692          */
2693         function ucfirst( $str ) {
2694                 $o = ord( $str );
2695                 if ( $o < 96 ) { // if already uppercase...
2696                         return $str;
2697                 } elseif ( $o < 128 ) {
2698                         return ucfirst( $str ); // use PHP's ucfirst()
2699                 } else {
2700                         // fall back to more complex logic in case of multibyte strings
2701                         return $this->uc( $str, true );
2702                 }
2703         }
2704
2705         /**
2706          * Convert a string to uppercase
2707          *
2708          * @param string $str
2709          * @param bool $first
2710          *
2711          * @return string
2712          */
2713         function uc( $str, $first = false ) {
2714                 if ( function_exists( 'mb_strtoupper' ) ) {
2715                         if ( $first ) {
2716                                 if ( $this->isMultibyte( $str ) ) {
2717                                         return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2718                                 } else {
2719                                         return ucfirst( $str );
2720                                 }
2721                         } else {
2722                                 return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2723                         }
2724                 } else {
2725                         if ( $this->isMultibyte( $str ) ) {
2726                                 $x = $first ? '^' : '';
2727                                 return preg_replace_callback(
2728                                         "/$x([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/",
2729                                         [ $this, 'ucCallback' ],
2730                                         $str
2731                                 );
2732                         } else {
2733                                 return $first ? ucfirst( $str ) : strtoupper( $str );
2734                         }
2735                 }
2736         }
2737
2738         /**
2739          * @param string $str
2740          * @return mixed|string
2741          */
2742         function lcfirst( $str ) {
2743                 $o = ord( $str );
2744                 if ( !$o ) {
2745                         return strval( $str );
2746                 } elseif ( $o >= 128 ) {
2747                         return $this->lc( $str, true );
2748                 } elseif ( $o > 96 ) {
2749                         return $str;
2750                 } else {
2751                         $str[0] = strtolower( $str[0] );
2752                         return $str;
2753                 }
2754         }
2755
2756         /**
2757          * @param string $str
2758          * @param bool $first
2759          * @return mixed|string
2760          */
2761         function lc( $str, $first = false ) {
2762                 if ( function_exists( 'mb_strtolower' ) ) {
2763                         if ( $first ) {
2764                                 if ( $this->isMultibyte( $str ) ) {
2765                                         return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2766                                 } else {
2767                                         return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2768                                 }
2769                         } else {
2770                                 return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2771                         }
2772                 } else {
2773                         if ( $this->isMultibyte( $str ) ) {
2774                                 $x = $first ? '^' : '';
2775                                 return preg_replace_callback(
2776                                         "/$x([A-Z]|[\\xc0-\\xff][\\x80-\\xbf]*)/",
2777                                         [ $this, 'lcCallback' ],
2778                                         $str
2779                                 );
2780                         } else {
2781                                 return $first ? strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 ) : strtolower( $str );
2782                         }
2783                 }
2784         }
2785
2786         /**
2787          * @param string $str
2788          * @return bool
2789          */
2790         function isMultibyte( $str ) {
2791                 return strlen( $str ) !== mb_strlen( $str );
2792         }
2793
2794         /**
2795          * @param string $str
2796          * @return mixed|string
2797          */
2798         function ucwords( $str ) {
2799                 if ( $this->isMultibyte( $str ) ) {
2800                         $str = $this->lc( $str );
2801
2802                         // regexp to find first letter in each word (i.e. after each space)
2803                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2804
2805                         // function to use to capitalize a single char
2806                         if ( function_exists( 'mb_strtoupper' ) ) {
2807                                 return preg_replace_callback(
2808                                         $replaceRegexp,
2809                                         [ $this, 'ucwordsCallbackMB' ],
2810                                         $str
2811                                 );
2812                         } else {
2813                                 return preg_replace_callback(
2814                                         $replaceRegexp,
2815                                         [ $this, 'ucwordsCallbackWiki' ],
2816                                         $str
2817                                 );
2818                         }
2819                 } else {
2820                         return ucwords( strtolower( $str ) );
2821                 }
2822         }
2823
2824         /**
2825          * capitalize words at word breaks
2826          *
2827          * @param string $str
2828          * @return mixed
2829          */
2830         function ucwordbreaks( $str ) {
2831                 if ( $this->isMultibyte( $str ) ) {
2832                         $str = $this->lc( $str );
2833
2834                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2835                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2836
2837                         // find first letter after word break
2838                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2839                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2840
2841                         if ( function_exists( 'mb_strtoupper' ) ) {
2842                                 return preg_replace_callback(
2843                                         $replaceRegexp,
2844                                         [ $this, 'ucwordbreaksCallbackMB' ],
2845                                         $str
2846                                 );
2847                         } else {
2848                                 return preg_replace_callback(
2849                                         $replaceRegexp,
2850                                         [ $this, 'ucwordsCallbackWiki' ],
2851                                         $str
2852                                 );
2853                         }
2854                 } else {
2855                         return preg_replace_callback(
2856                                 '/\b([\w\x80-\xff]+)\b/',
2857                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2858                                 $str
2859                         );
2860                 }
2861         }
2862
2863         /**
2864          * Return a case-folded representation of $s
2865          *
2866          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2867          * and $s2 are the same except for the case of their characters. It is not
2868          * necessary for the value returned to make sense when displayed.
2869          *
2870          * Do *not* perform any other normalisation in this function. If a caller
2871          * uses this function when it should be using a more general normalisation
2872          * function, then fix the caller.
2873          *
2874          * @param string $s
2875          *
2876          * @return string
2877          */
2878         function caseFold( $s ) {
2879                 return $this->uc( $s );
2880         }
2881
2882         /**
2883          * @param string $s
2884          * @return string
2885          */
2886         function checkTitleEncoding( $s ) {
2887                 if ( is_array( $s ) ) {
2888                         throw new MWException( 'Given array to checkTitleEncoding.' );
2889                 }
2890                 if ( StringUtils::isUtf8( $s ) ) {
2891                         return $s;
2892                 }
2893
2894                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2895         }
2896
2897         /**
2898          * @return array
2899          */
2900         function fallback8bitEncoding() {
2901                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2902         }
2903
2904         /**
2905          * Most writing systems use whitespace to break up words.
2906          * Some languages such as Chinese don't conventionally do this,
2907          * which requires special handling when breaking up words for
2908          * searching etc.
2909          *
2910          * @return bool
2911          */
2912         function hasWordBreaks() {
2913                 return true;
2914         }
2915
2916         /**
2917          * Some languages such as Chinese require word segmentation,
2918          * Specify such segmentation when overridden in derived class.
2919          *
2920          * @param string $string
2921          * @return string
2922          */
2923         function segmentByWord( $string ) {
2924                 return $string;
2925         }
2926
2927         /**
2928          * Some languages have special punctuation need to be normalized.
2929          * Make such changes here.
2930          *
2931          * @param string $string
2932          * @return string
2933          */
2934         function normalizeForSearch( $string ) {
2935                 return self::convertDoubleWidth( $string );
2936         }
2937
2938         /**
2939          * convert double-width roman characters to single-width.
2940          * range: ff00-ff5f ~= 0020-007f
2941          *
2942          * @param string $string
2943          *
2944          * @return string
2945          */
2946         protected static function convertDoubleWidth( $string ) {
2947                 static $full = null;
2948                 static $half = null;
2949
2950                 if ( $full === null ) {
2951                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2952                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2953                         $full = str_split( $fullWidth, 3 );
2954                         $half = str_split( $halfWidth );
2955                 }
2956
2957                 $string = str_replace( $full, $half, $string );
2958                 return $string;
2959         }
2960
2961         /**
2962          * @param string $string
2963          * @param string $pattern
2964          * @return string
2965          */
2966         protected static function insertSpace( $string, $pattern ) {
2967                 $string = preg_replace( $pattern, " $1 ", $string );
2968                 $string = preg_replace( '/ +/', ' ', $string );
2969                 return $string;
2970         }
2971
2972         /**
2973          * @param array $termsArray
2974          * @return array
2975          */
2976         function convertForSearchResult( $termsArray ) {
2977                 # some languages, e.g. Chinese, need to do a conversion
2978                 # in order for search results to be displayed correctly
2979                 return $termsArray;
2980         }
2981
2982         /**
2983          * Get the first character of a string.
2984          *
2985          * @param string $s
2986          * @return string
2987          */
2988         function firstChar( $s ) {
2989                 $matches = [];
2990                 preg_match(
2991                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2992                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2993                         $s,
2994                         $matches
2995                 );
2996
2997                 if ( isset( $matches[1] ) ) {
2998                         if ( strlen( $matches[1] ) != 3 ) {
2999                                 return $matches[1];
3000                         }
3001
3002                         // Break down Hangul syllables to grab the first jamo
3003                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
3004                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
3005                                 return $matches[1];
3006                         } elseif ( $code < 0xb098 ) {
3007                                 return "\xe3\x84\xb1";
3008                         } elseif ( $code < 0xb2e4 ) {
3009                                 return "\xe3\x84\xb4";
3010                         } elseif ( $code < 0xb77c ) {
3011                                 return "\xe3\x84\xb7";
3012                         } elseif ( $code < 0xb9c8 ) {
3013                                 return "\xe3\x84\xb9";
3014                         } elseif ( $code < 0xbc14 ) {
3015                                 return "\xe3\x85\x81";
3016                         } elseif ( $code < 0xc0ac ) {
3017                                 return "\xe3\x85\x82";
3018                         } elseif ( $code < 0xc544 ) {
3019                                 return "\xe3\x85\x85";
3020                         } elseif ( $code < 0xc790 ) {
3021                                 return "\xe3\x85\x87";
3022                         } elseif ( $code < 0xcc28 ) {
3023                                 return "\xe3\x85\x88";
3024                         } elseif ( $code < 0xce74 ) {
3025                                 return "\xe3\x85\x8a";
3026                         } elseif ( $code < 0xd0c0 ) {
3027                                 return "\xe3\x85\x8b";
3028                         } elseif ( $code < 0xd30c ) {
3029                                 return "\xe3\x85\x8c";
3030                         } elseif ( $code < 0xd558 ) {
3031                                 return "\xe3\x85\x8d";
3032                         } else {
3033                                 return "\xe3\x85\x8e";
3034                         }
3035                 } else {
3036                         return '';
3037                 }
3038         }
3039
3040         function initEncoding() {
3041                 # Some languages may have an alternate char encoding option
3042                 # (Esperanto X-coding, Japanese furigana conversion, etc)
3043                 # If this language is used as the primary content language,
3044                 # an override to the defaults can be set here on startup.
3045         }
3046
3047         /**
3048          * @param string $s
3049          * @return string
3050          */
3051         function recodeForEdit( $s ) {
3052                 # For some languages we'll want to explicitly specify
3053                 # which characters make it into the edit box raw
3054                 # or are converted in some way or another.
3055                 global $wgEditEncoding;
3056                 if ( $wgEditEncoding == '' || $wgEditEncoding == 'UTF-8' ) {
3057                         return $s;
3058                 } else {
3059                         return $this->iconv( 'UTF-8', $wgEditEncoding, $s );
3060                 }
3061         }
3062
3063         /**
3064          * @param string $s
3065          * @return string
3066          */
3067         function recodeInput( $s ) {
3068                 # Take the previous into account.
3069                 global $wgEditEncoding;
3070                 if ( $wgEditEncoding != '' ) {
3071                         $enc = $wgEditEncoding;
3072                 } else {
3073                         $enc = 'UTF-8';
3074                 }
3075                 if ( $enc == 'UTF-8' ) {
3076                         return $s;
3077                 } else {
3078                         return $this->iconv( $enc, 'UTF-8', $s );
3079                 }
3080         }
3081
3082         /**
3083          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3084          * also cleans up certain backwards-compatible sequences, converting them
3085          * to the modern Unicode equivalent.
3086          *
3087          * This is language-specific for performance reasons only.
3088          *
3089          * @param string $s
3090          *
3091          * @return string
3092          */
3093         function normalize( $s ) {
3094                 global $wgAllUnicodeFixes;
3095                 $s = UtfNormal\Validator::cleanUp( $s );
3096                 if ( $wgAllUnicodeFixes ) {
3097                         $s = $this->transformUsingPairFile( 'normalize-ar.ser', $s );
3098                         $s = $this->transformUsingPairFile( 'normalize-ml.ser', $s );
3099                 }
3100
3101                 return $s;
3102         }
3103
3104         /**
3105          * Transform a string using serialized data stored in the given file (which
3106          * must be in the serialized subdirectory of $IP). The file contains pairs
3107          * mapping source characters to destination characters.
3108          *
3109          * The data is cached in process memory. This will go faster if you have the
3110          * FastStringSearch extension.
3111          *
3112          * @param string $file
3113          * @param string $string
3114          *
3115          * @throws MWException
3116          * @return string
3117          */
3118         function transformUsingPairFile( $file, $string ) {
3119                 if ( !isset( $this->transformData[$file] ) ) {
3120                         $data = wfGetPrecompiledData( $file );
3121                         if ( $data === false ) {
3122                                 throw new MWException( __METHOD__ . ": The transformation file $file is missing" );
3123                         }
3124                         $this->transformData[$file] = new ReplacementArray( $data );
3125                 }
3126                 return $this->transformData[$file]->replace( $string );
3127         }
3128
3129         /**
3130          * For right-to-left language support
3131          *
3132          * @return bool
3133          */
3134         function isRTL() {
3135                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3136         }
3137
3138         /**
3139          * Return the correct HTML 'dir' attribute value for this language.
3140          * @return string
3141          */
3142         function getDir() {
3143                 return $this->isRTL() ? 'rtl' : 'ltr';
3144         }
3145
3146         /**
3147          * Return 'left' or 'right' as appropriate alignment for line-start
3148          * for this language's text direction.
3149          *
3150          * Should be equivalent to CSS3 'start' text-align value....
3151          *
3152          * @return string
3153          */
3154         function alignStart() {
3155                 return $this->isRTL() ? 'right' : 'left';
3156         }
3157
3158         /**
3159          * Return 'right' or 'left' as appropriate alignment for line-end
3160          * for this language's text direction.
3161          *
3162          * Should be equivalent to CSS3 'end' text-align value....
3163          *
3164          * @return string
3165          */
3166         function alignEnd() {
3167                 return $this->isRTL() ? 'left' : 'right';
3168         }
3169
3170         /**
3171          * A hidden direction mark (LRM or RLM), depending on the language direction.
3172          * Unlike getDirMark(), this function returns the character as an HTML entity.
3173          * This function should be used when the output is guaranteed to be HTML,
3174          * because it makes the output HTML source code more readable. When
3175          * the output is plain text or can be escaped, getDirMark() should be used.
3176          *
3177          * @param bool $opposite Get the direction mark opposite to your language
3178          * @return string
3179          * @since 1.20
3180          */
3181         function getDirMarkEntity( $opposite = false ) {
3182                 if ( $opposite ) {
3183                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3184                 }
3185                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3186         }
3187
3188         /**
3189          * A hidden direction mark (LRM or RLM), depending on the language direction.
3190          * This function produces them as invisible Unicode characters and
3191          * the output may be hard to read and debug, so it should only be used
3192          * when the output is plain text or can be escaped. When the output is
3193          * HTML, use getDirMarkEntity() instead.
3194          *
3195          * @param bool $opposite Get the direction mark opposite to your language
3196          * @return string
3197          */
3198         function getDirMark( $opposite = false ) {
3199                 $lrm = "\xE2\x80\x8E"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3200                 $rlm = "\xE2\x80\x8F"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3201                 if ( $opposite ) {
3202                         return $this->isRTL() ? $lrm : $rlm;
3203                 }
3204                 return $this->isRTL() ? $rlm : $lrm;
3205         }
3206
3207         /**
3208          * @return array
3209          */
3210         function capitalizeAllNouns() {
3211                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3212         }
3213
3214         /**
3215          * An arrow, depending on the language direction.
3216          *
3217          * @param string $direction The direction of the arrow: forwards (default),
3218          *   backwards, left, right, up, down.
3219          * @return string
3220          */
3221         function getArrow( $direction = 'forwards' ) {
3222                 switch ( $direction ) {
3223                 case 'forwards':
3224                         return $this->isRTL() ? '←' : '→';
3225                 case 'backwards':
3226                         return $this->isRTL() ? '→' : '←';
3227                 case 'left':
3228                         return '←';
3229                 case 'right':
3230                         return '→';
3231                 case 'up':
3232                         return '↑';
3233                 case 'down':
3234                         return '↓';
3235                 }
3236         }
3237
3238         /**
3239          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3240          *
3241          * @return bool
3242          */
3243         function linkPrefixExtension() {
3244                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3245         }
3246
3247         /**
3248          * Get all magic words from cache.
3249          * @return array
3250          */
3251         function getMagicWords() {
3252                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3253         }
3254
3255         /**
3256          * Run the LanguageGetMagic hook once.
3257          */
3258         protected function doMagicHook() {
3259                 if ( $this->mMagicHookDone ) {
3260                         return;
3261                 }
3262                 $this->mMagicHookDone = true;
3263                 Hooks::run( 'LanguageGetMagic', [ &$this->mMagicExtensions, $this->getCode() ] );
3264         }
3265
3266         /**
3267          * Fill a MagicWord object with data from here
3268          *
3269          * @param MagicWord $mw
3270          */
3271         function getMagic( $mw ) {
3272                 // Saves a function call
3273                 if ( !$this->mMagicHookDone ) {
3274                         $this->doMagicHook();
3275                 }
3276
3277                 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
3278                         $rawEntry = $this->mMagicExtensions[$mw->mId];
3279                 } else {
3280                         $rawEntry = self::$dataCache->getSubitem(
3281                                 $this->mCode, 'magicWords', $mw->mId );
3282                 }
3283
3284                 if ( !is_array( $rawEntry ) ) {
3285                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3286                 } else {
3287                         $mw->mCaseSensitive = $rawEntry[0];
3288                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3289                 }
3290         }
3291
3292         /**
3293          * Add magic words to the extension array
3294          *
3295          * @param array $newWords
3296          */
3297         function addMagicWordsByLang( $newWords ) {
3298                 $fallbackChain = $this->getFallbackLanguages();
3299                 $fallbackChain = array_reverse( $fallbackChain );
3300                 foreach ( $fallbackChain as $code ) {
3301                         if ( isset( $newWords[$code] ) ) {
3302                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3303                         }
3304                 }
3305         }
3306
3307         /**
3308          * Get special page names, as an associative array
3309          *   canonical name => array of valid names, including aliases
3310          * @return array
3311          */
3312         function getSpecialPageAliases() {
3313                 // Cache aliases because it may be slow to load them
3314                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3315                         // Initialise array
3316                         $this->mExtendedSpecialPageAliases =
3317                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3318                         Hooks::run( 'LanguageGetSpecialPageAliases',
3319                                 [ &$this->mExtendedSpecialPageAliases, $this->getCode() ] );
3320                 }
3321
3322                 return $this->mExtendedSpecialPageAliases;
3323         }
3324
3325         /**
3326          * Italic is unsuitable for some languages
3327          *
3328          * @param string $text The text to be emphasized.
3329          * @return string
3330          */
3331         function emphasize( $text ) {
3332                 return "<em>$text</em>";
3333         }
3334
3335         /**
3336          * Normally we output all numbers in plain en_US style, that is
3337          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3338          * point twohundredthirtyfive. However this is not suitable for all
3339          * languages, some such as Punjabi want ੨੯੩,੨੯੫.੨੩੫ and others such as
3340          * Icelandic just want to use commas instead of dots, and dots instead
3341          * of commas like "293.291,235".
3342          *
3343          * An example of this function being called:
3344          * <code>
3345          * wfMessage( 'message' )->numParams( $num )->text()
3346          * </code>
3347          *
3348          * See $separatorTransformTable on MessageIs.php for
3349          * the , => . and . => , implementation.
3350          *
3351          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3352          * @param int|float $number The string to be formatted, should be an integer
3353          *   or a floating point number.
3354          * @param bool $nocommafy Set to true for special numbers like dates
3355          * @return string
3356          */
3357         public function formatNum( $number, $nocommafy = false ) {
3358                 global $wgTranslateNumerals;
3359                 if ( !$nocommafy ) {
3360                         $number = $this->commafy( $number );
3361                         $s = $this->separatorTransformTable();
3362                         if ( $s ) {
3363                                 $number = strtr( $number, $s );
3364                         }
3365                 }
3366
3367                 if ( $wgTranslateNumerals ) {
3368                         $s = $this->digitTransformTable();
3369                         if ( $s ) {
3370                                 $number = strtr( $number, $s );
3371                         }
3372                 }
3373
3374                 return $number;
3375         }
3376
3377         /**
3378          * Front-end for non-commafied formatNum
3379          *
3380          * @param int|float $number The string to be formatted, should be an integer
3381          *        or a floating point number.
3382          * @since 1.21
3383          * @return string
3384          */
3385         public function formatNumNoSeparators( $number ) {
3386                 return $this->formatNum( $number, true );
3387         }
3388
3389         /**
3390          * @param string $number
3391          * @return string
3392          */
3393         public function parseFormattedNumber( $number ) {
3394                 $s = $this->digitTransformTable();
3395                 if ( $s ) {
3396                         // eliminate empty array values such as ''. (bug 64347)
3397                         $s = array_filter( $s );
3398                         $number = strtr( $number, array_flip( $s ) );
3399                 }
3400
3401                 $s = $this->separatorTransformTable();
3402                 if ( $s ) {
3403                         // eliminate empty array values such as ''. (bug 64347)
3404                         $s = array_filter( $s );
3405                         $number = strtr( $number, array_flip( $s ) );
3406                 }
3407
3408                 $number = strtr( $number, [ ',' => '' ] );
3409                 return $number;
3410         }
3411
3412         /**
3413          * Adds commas to a given number
3414          * @since 1.19
3415          * @param mixed $number
3416          * @return string
3417          */
3418         function commafy( $number ) {
3419                 $digitGroupingPattern = $this->digitGroupingPattern();
3420                 if ( $number === null ) {
3421                         return '';
3422                 }
3423
3424                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3425                         // default grouping is at thousands,  use the same for ###,###,### pattern too.
3426                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3427                 } else {
3428                         // Ref: http://cldr.unicode.org/translation/number-patterns
3429                         $sign = "";
3430                         if ( intval( $number ) < 0 ) {
3431                                 // For negative numbers apply the algorithm like positive number and add sign.
3432                                 $sign = "-";
3433                                 $number = substr( $number, 1 );
3434                         }
3435                         $integerPart = [];
3436                         $decimalPart = [];
3437                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3438                         preg_match( "/\d+/", $number, $integerPart );
3439                         preg_match( "/\.\d*/", $number, $decimalPart );
3440                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3441                         if ( $groupedNumber === $number ) {
3442                                 // the string does not have any number part. Eg: .12345
3443                                 return $sign . $groupedNumber;
3444                         }
3445                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3446                         while ( $start > 0 ) {
3447                                 $match = $matches[0][$numMatches - 1];
3448                                 $matchLen = strlen( $match );
3449                                 $start = $end - $matchLen;
3450                                 if ( $start < 0 ) {
3451                                         $start = 0;
3452                                 }
3453                                 $groupedNumber = substr( $number, $start, $end -$start ) . $groupedNumber;
3454                                 $end = $start;
3455                                 if ( $numMatches > 1 ) {
3456                                         // use the last pattern for the rest of the number
3457                                         $numMatches--;
3458                                 }
3459                                 if ( $start > 0 ) {
3460                                         $groupedNumber = "," . $groupedNumber;
3461                                 }
3462                         }
3463                         return $sign . $groupedNumber;
3464                 }
3465         }
3466
3467         /**
3468          * @return string
3469          */
3470         function digitGroupingPattern() {
3471                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3472         }
3473
3474         /**
3475          * @return array
3476          */
3477         function digitTransformTable() {
3478                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3479         }
3480
3481         /**
3482          * @return array
3483          */
3484         function separatorTransformTable() {
3485                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3486         }
3487
3488         /**
3489          * Take a list of strings and build a locale-friendly comma-separated
3490          * list, using the local comma-separator message.
3491          * The last two strings are chained with an "and".
3492          * NOTE: This function will only work with standard numeric array keys (0, 1, 2…)
3493          *
3494          * @param string[] $l
3495          * @return string
3496          */
3497         function listToText( array $l ) {
3498                 $m = count( $l ) - 1;
3499                 if ( $m < 0 ) {
3500                         return '';
3501                 }
3502                 if ( $m > 0 ) {
3503                         $and = $this->msg( 'and' )->escaped();
3504                         $space = $this->msg( 'word-separator' )->escaped();
3505                         if ( $m > 1 ) {
3506                                 $comma = $this->msg( 'comma-separator' )->escaped();
3507                         }
3508                 }
3509                 $s = $l[$m];
3510                 for ( $i = $m - 1; $i >= 0; $i-- ) {
3511                         if ( $i == $m - 1 ) {
3512                                 $s = $l[$i] . $and . $space . $s;
3513                         } else {
3514                                 $s = $l[$i] . $comma . $s;
3515                         }
3516                 }
3517                 return $s;
3518         }
3519
3520         /**
3521          * Take a list of strings and build a locale-friendly comma-separated
3522          * list, using the local comma-separator message.
3523          * @param string[] $list Array of strings to put in a comma list
3524          * @return string
3525          */
3526         function commaList( array $list ) {
3527                 return implode(
3528                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3529                         $list
3530                 );
3531         }
3532
3533         /**
3534          * Take a list of strings and build a locale-friendly semicolon-separated
3535          * list, using the local semicolon-separator message.
3536          * @param string[] $list Array of strings to put in a semicolon list
3537          * @return string
3538          */
3539         function semicolonList( array $list ) {
3540                 return implode(
3541                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3542                         $list
3543                 );
3544         }
3545
3546         /**
3547          * Same as commaList, but separate it with the pipe instead.
3548          * @param string[] $list Array of strings to put in a pipe list
3549          * @return string
3550          */
3551         function pipeList( array $list ) {
3552                 return implode(
3553                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3554                         $list
3555                 );
3556         }
3557
3558         /**
3559          * Truncate a string to a specified length in bytes, appending an optional
3560          * string (e.g. for ellipses)
3561          *
3562          * The database offers limited byte lengths for some columns in the database;
3563          * multi-byte character sets mean we need to ensure that only whole characters
3564          * are included, otherwise broken characters can be passed to the user
3565          *
3566          * If $length is negative, the string will be truncated from the beginning
3567          *
3568          * @param string $string String to truncate
3569          * @param int $length Maximum length (including ellipses)
3570          * @param string $ellipsis String to append to the truncated text
3571          * @param bool $adjustLength Subtract length of ellipsis from $length.
3572          *      $adjustLength was introduced in 1.18, before that behaved as if false.
3573          * @return string
3574          */
3575         function truncate( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3576                 # Use the localized ellipsis character
3577                 if ( $ellipsis == '...' ) {
3578                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3579                 }
3580                 # Check if there is no need to truncate
3581                 if ( $length == 0 ) {
3582                         return $ellipsis; // convention
3583                 } elseif ( strlen( $string ) <= abs( $length ) ) {
3584                         return $string; // no need to truncate
3585                 }
3586                 $stringOriginal = $string;
3587                 # If ellipsis length is >= $length then we can't apply $adjustLength
3588                 if ( $adjustLength && strlen( $ellipsis ) >= abs( $length ) ) {
3589                         $string = $ellipsis; // this can be slightly unexpected
3590                 # Otherwise, truncate and add ellipsis...
3591                 } else {
3592                         $eLength = $adjustLength ? strlen( $ellipsis ) : 0;
3593                         if ( $length > 0 ) {
3594                                 $length -= $eLength;
3595                                 $string = substr( $string, 0, $length ); // xyz...
3596                                 $string = $this->removeBadCharLast( $string );
3597                                 $string = rtrim( $string );
3598                                 $string = $string . $ellipsis;
3599                         } else {
3600                                 $length += $eLength;
3601                                 $string = substr( $string, $length ); // ...xyz
3602                                 $string = $this->removeBadCharFirst( $string );
3603                                 $string = ltrim( $string );
3604                                 $string = $ellipsis . $string;
3605                         }
3606                 }
3607                 # Do not truncate if the ellipsis makes the string longer/equal (bug 22181).
3608                 # This check is *not* redundant if $adjustLength, due to the single case where
3609                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3610                 if ( strlen( $string ) < strlen( $stringOriginal ) ) {
3611                         return $string;
3612                 } else {
3613                         return $stringOriginal;
3614                 }
3615         }
3616
3617         /**
3618          * Remove bytes that represent an incomplete Unicode character
3619          * at the end of string (e.g. bytes of the char are missing)
3620          *
3621          * @param string $string
3622          * @return string
3623          */
3624         protected function removeBadCharLast( $string ) {
3625                 if ( $string != '' ) {
3626                         $char = ord( $string[strlen( $string ) - 1] );
3627                         $m = [];
3628                         if ( $char >= 0xc0 ) {
3629                                 # We got the first byte only of a multibyte char; remove it.
3630                                 $string = substr( $string, 0, -1 );
3631                         } elseif ( $char >= 0x80 &&
3632                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3633                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3634                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3635                         ) {
3636                                 # We chopped in the middle of a character; remove it
3637                                 $string = $m[1];
3638                         }
3639                 }
3640                 return $string;
3641         }
3642
3643         /**
3644          * Remove bytes that represent an incomplete Unicode character
3645          * at the start of string (e.g. bytes of the char are missing)
3646          *
3647          * @param string $string
3648          * @return string
3649          */
3650         protected function removeBadCharFirst( $string ) {
3651                 if ( $string != '' ) {
3652                         $char = ord( $string[0] );
3653                         if ( $char >= 0x80 && $char < 0xc0 ) {
3654                                 # We chopped in the middle of a character; remove the whole thing
3655                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3656                         }
3657                 }
3658                 return $string;
3659         }
3660
3661         /**
3662          * Truncate a string of valid HTML to a specified length in bytes,
3663          * appending an optional string (e.g. for ellipses), and return valid HTML
3664          *
3665          * This is only intended for styled/linked text, such as HTML with
3666          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3667          * Also, this will not detect things like "display:none" CSS.
3668          *
3669          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3670          *
3671          * @param string $text HTML string to truncate
3672          * @param int $length (zero/positive) Maximum length (including ellipses)
3673          * @param string $ellipsis String to append to the truncated text
3674          * @return string
3675          */
3676         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3677                 # Use the localized ellipsis character
3678                 if ( $ellipsis == '...' ) {
3679                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3680                 }
3681                 # Check if there is clearly no need to truncate
3682                 if ( $length <= 0 ) {
3683                         return $ellipsis; // no text shown, nothing to format (convention)
3684                 } elseif ( strlen( $text ) <= $length ) {
3685                         return $text; // string short enough even *with* HTML (short-circuit)
3686                 }
3687
3688                 $dispLen = 0; // innerHTML legth so far
3689                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3690                 $tagType = 0; // 0-open, 1-close
3691                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3692                 $entityState = 0; // 0-not entity, 1-entity
3693                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3694                 $openTags = []; // open tag stack
3695                 $maybeState = null; // possible truncation state
3696
3697                 $textLen = strlen( $text );
3698                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3699                 for ( $pos = 0; true; ++$pos ) {
3700                         # Consider truncation once the display length has reached the maximim.
3701                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3702                         # Check that we're not in the middle of a bracket/entity...
3703                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3704                                 if ( !$testingEllipsis ) {
3705                                         $testingEllipsis = true;
3706                                         # Save where we are; we will truncate here unless there turn out to
3707                                         # be so few remaining characters that truncation is not necessary.
3708                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3709                                                 $maybeState = [ $ret, $openTags ]; // save state
3710                                         }
3711                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3712                                         # String in fact does need truncation, the truncation point was OK.
3713                                         list( $ret, $openTags ) = $maybeState; // reload state
3714                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3715                                         $ret .= $ellipsis; // add ellipsis
3716                                         break;
3717                                 }
3718                         }
3719                         if ( $pos >= $textLen ) {
3720                                 break; // extra iteration just for above checks
3721                         }
3722
3723                         # Read the next char...
3724                         $ch = $text[$pos];
3725                         $lastCh = $pos ? $text[$pos - 1] : '';
3726                         $ret .= $ch; // add to result string
3727                         if ( $ch == '<' ) {
3728                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3729                                 $entityState = 0; // for bad HTML
3730                                 $bracketState = 1; // tag started (checking for backslash)
3731                         } elseif ( $ch == '>' ) {
3732                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3733                                 $entityState = 0; // for bad HTML
3734                                 $bracketState = 0; // out of brackets
3735                         } elseif ( $bracketState == 1 ) {
3736                                 if ( $ch == '/' ) {
3737                                         $tagType = 1; // close tag (e.g. "</span>")
3738                                 } else {
3739                                         $tagType = 0; // open tag (e.g. "<span>")
3740                                         $tag .= $ch;
3741                                 }
3742                                 $bracketState = 2; // building tag name
3743                         } elseif ( $bracketState == 2 ) {
3744                                 if ( $ch != ' ' ) {
3745                                         $tag .= $ch;
3746                                 } else {
3747                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3748                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3749                                 }
3750                         } elseif ( $bracketState == 0 ) {
3751                                 if ( $entityState ) {
3752                                         if ( $ch == ';' ) {
3753                                                 $entityState = 0;
3754                                                 $dispLen++; // entity is one displayed char
3755                                         }
3756                                 } else {
3757                                         if ( $neLength == 0 && !$maybeState ) {
3758                                                 // Save state without $ch. We want to *hit* the first
3759                                                 // display char (to get tags) but not *use* it if truncating.
3760                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3761                                         }
3762                                         if ( $ch == '&' ) {
3763                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3764                                         } else {
3765                                                 $dispLen++; // this char is displayed
3766                                                 // Add the next $max display text chars after this in one swoop...
3767                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3768                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3769                                                 $dispLen += $skipped;
3770                                                 $pos += $skipped;
3771                                         }
3772                                 }
3773                         }
3774                 }
3775                 // Close the last tag if left unclosed by bad HTML
3776                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3777                 while ( count( $openTags ) > 0 ) {
3778                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3779                 }
3780                 return $ret;
3781         }
3782
3783         /**
3784          * truncateHtml() helper function
3785          * like strcspn() but adds the skipped chars to $ret
3786          *
3787          * @param string $ret
3788          * @param string $text
3789          * @param string $search
3790          * @param int $start
3791          * @param null|int $len
3792          * @return int
3793          */
3794         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3795                 if ( $len === null ) {
3796                         $len = -1; // -1 means "no limit" for strcspn
3797                 } elseif ( $len < 0 ) {
3798                         $len = 0; // sanity
3799                 }
3800                 $skipCount = 0;
3801                 if ( $start < strlen( $text ) ) {
3802                         $skipCount = strcspn( $text, $search, $start, $len );
3803                         $ret .= substr( $text, $start, $skipCount );
3804                 }
3805                 return $skipCount;
3806         }
3807
3808         /**
3809          * truncateHtml() helper function
3810          * (a) push or pop $tag from $openTags as needed
3811          * (b) clear $tag value
3812          * @param string &$tag Current HTML tag name we are looking at
3813          * @param int $tagType (0-open tag, 1-close tag)
3814          * @param string $lastCh Character before the '>' that ended this tag
3815          * @param array &$openTags Open tag stack (not accounting for $tag)
3816          */
3817         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3818                 $tag = ltrim( $tag );
3819                 if ( $tag != '' ) {
3820                         if ( $tagType == 0 && $lastCh != '/' ) {
3821                                 $openTags[] = $tag; // tag opened (didn't close itself)
3822                         } elseif ( $tagType == 1 ) {
3823                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3824                                         array_pop( $openTags ); // tag closed
3825                                 }
3826                         }
3827                         $tag = '';
3828                 }
3829         }
3830
3831         /**
3832          * Grammatical transformations, needed for inflected languages
3833          * Invoked by putting {{grammar:case|word}} in a message
3834          *
3835          * @param string $word
3836          * @param string $case
3837          * @return string
3838          */
3839         function convertGrammar( $word, $case ) {
3840                 global $wgGrammarForms;
3841                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3842                         return $wgGrammarForms[$this->getCode()][$case][$word];
3843                 }
3844
3845                 return $word;
3846         }
3847         /**
3848          * Get the grammar forms for the content language
3849          * @return array Array of grammar forms
3850          * @since 1.20
3851          */
3852         function getGrammarForms() {
3853                 global $wgGrammarForms;
3854                 if ( isset( $wgGrammarForms[$this->getCode()] )
3855                         && is_array( $wgGrammarForms[$this->getCode()] )
3856                 ) {
3857                         return $wgGrammarForms[$this->getCode()];
3858                 }
3859
3860                 return [];
3861         }
3862         /**
3863          * Provides an alternative text depending on specified gender.
3864          * Usage {{gender:username|masculine|feminine|unknown}}.
3865          * username is optional, in which case the gender of current user is used,
3866          * but only in (some) interface messages; otherwise default gender is used.
3867          *
3868          * If no forms are given, an empty string is returned. If only one form is
3869          * given, it will be returned unconditionally. These details are implied by
3870          * the caller and cannot be overridden in subclasses.
3871          *
3872          * If three forms are given, the default is to use the third (unknown) form.
3873          * If fewer than three forms are given, the default is to use the first (masculine) form.
3874          * These details can be overridden in subclasses.
3875          *
3876          * @param string $gender
3877          * @param array $forms
3878          *
3879          * @return string
3880          */
3881         function gender( $gender, $forms ) {
3882                 if ( !count( $forms ) ) {
3883                         return '';
3884                 }
3885                 $forms = $this->preConvertPlural( $forms, 2 );
3886                 if ( $gender === 'male' ) {
3887                         return $forms[0];
3888                 }
3889                 if ( $gender === 'female' ) {
3890                         return $forms[1];
3891                 }
3892                 return isset( $forms[2] ) ? $forms[2] : $forms[0];
3893         }
3894
3895         /**
3896          * Plural form transformations, needed for some languages.
3897          * For example, there are 3 form of plural in Russian and Polish,
3898          * depending on "count mod 10". See [[w:Plural]]
3899          * For English it is pretty simple.
3900          *
3901          * Invoked by putting {{plural:count|wordform1|wordform2}}
3902          * or {{plural:count|wordform1|wordform2|wordform3}}
3903          *
3904          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3905          *
3906          * @param int $count Non-localized number
3907          * @param array $forms Different plural forms
3908          * @return string Correct form of plural for $count in this language
3909          */
3910         function convertPlural( $count, $forms ) {
3911                 // Handle explicit n=pluralform cases
3912                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3913                 if ( is_string( $forms ) ) {
3914                         return $forms;
3915                 }
3916                 if ( !count( $forms ) ) {
3917                         return '';
3918                 }
3919
3920                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3921                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3922                 return $forms[$pluralForm];
3923         }
3924
3925         /**
3926          * Handles explicit plural forms for Language::convertPlural()
3927          *
3928          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
3929          * If an explicitly defined plural form matches the $count, then
3930          * string value returned, otherwise array returned for further consideration
3931          * by CLDR rules or overridden convertPlural().
3932          *
3933          * @since 1.23
3934          *
3935          * @param int $count Non-localized number
3936          * @param array $forms Different plural forms
3937          *
3938          * @return array|string
3939          */
3940         protected function handleExplicitPluralForms( $count, array $forms ) {
3941                 foreach ( $forms as $index => $form ) {
3942                         if ( preg_match( '/\d+=/i', $form ) ) {
3943                                 $pos = strpos( $form, '=' );
3944                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
3945                                         return substr( $form, $pos + 1 );
3946                                 }
3947                                 unset( $forms[$index] );
3948                         }
3949                 }
3950                 return array_values( $forms );
3951         }
3952
3953         /**
3954          * Checks that convertPlural was given an array and pads it to requested
3955          * amount of forms by copying the last one.
3956          *
3957          * @param array $forms Array of forms given to convertPlural
3958          * @param int $count How many forms should there be at least
3959          * @return array Padded array of forms or an exception if not an array
3960          */
3961         protected function preConvertPlural( /* Array */ $forms, $count ) {
3962                 while ( count( $forms ) < $count ) {
3963                         $forms[] = $forms[count( $forms ) - 1];
3964                 }
3965                 return $forms;
3966         }
3967
3968         /**
3969          * Wraps argument with unicode control characters for directionality safety
3970          *
3971          * This solves the problem where directionality-neutral characters at the edge of
3972          * the argument string get interpreted with the wrong directionality from the
3973          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
3974          *
3975          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
3976          * directionality of the argument string, using the BIDI algorithm's own "First
3977          * strong directional codepoint" rule. Essentially, this works round the fact that
3978          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
3979          * direction inference). The latter is cleaner but still not widely supported.
3980          *
3981          * @param string $text Text to wrap
3982          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
3983          */
3984         public function embedBidi( $text = '' ) {
3985                 $dir = Language::strongDirFromContent( $text );
3986                 if ( $dir === 'ltr' ) {
3987                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
3988                         return self::$lre . $text . self::$pdf;
3989                 }
3990                 if ( $dir === 'rtl' ) {
3991                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
3992                         return self::$rle . $text . self::$pdf;
3993                 }
3994                 // No strong directionality: do not wrap
3995                 return $text;
3996         }
3997
3998         /**
3999          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4000          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4001          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4002          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4003          * match up with it.
4004          *
4005          * @param string $str The validated block duration in English
4006          * @return string Somehow translated block duration
4007          * @see LanguageFi.php for example implementation
4008          */
4009         function translateBlockExpiry( $str ) {
4010                 $duration = SpecialBlock::getSuggestedDurations( $this );
4011                 foreach ( $duration as $show => $value ) {
4012                         if ( strcmp( $str, $value ) == 0 ) {
4013                                 return htmlspecialchars( trim( $show ) );
4014                         }
4015                 }
4016
4017                 if ( wfIsInfinity( $str ) ) {
4018                         foreach ( $duration as $show => $value ) {
4019                                 if ( wfIsInfinity( $value ) ) {
4020                                         return htmlspecialchars( trim( $show ) );
4021                                 }
4022                         }
4023                 }
4024
4025                 // If all else fails, return a standard duration or timestamp description.
4026                 $time = strtotime( $str, 0 );
4027                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4028                         return $str;
4029                 } elseif ( $time !== strtotime( $str, 1 ) ) { // It's a relative timestamp.
4030                         // $time is relative to 0 so it's a duration length.
4031                         return $this->formatDuration( $time );
4032                 } else { // It's an absolute timestamp.
4033                         if ( $time === 0 ) {
4034                                 // wfTimestamp() handles 0 as current time instead of epoch.
4035                                 return $this->timeanddate( '19700101000000' );
4036                         } else {
4037                                 return $this->timeanddate( $time );
4038                         }
4039                 }
4040         }
4041
4042         /**
4043          * languages like Chinese need to be segmented in order for the diff
4044          * to be of any use
4045          *
4046          * @param string $text
4047          * @return string
4048          */
4049         public function segmentForDiff( $text ) {
4050                 return $text;
4051         }
4052
4053         /**
4054          * and unsegment to show the result
4055          *
4056          * @param string $text
4057          * @return string
4058          */
4059         public function unsegmentForDiff( $text ) {
4060                 return $text;
4061         }
4062
4063         /**
4064          * Return the LanguageConverter used in the Language
4065          *
4066          * @since 1.19
4067          * @return LanguageConverter
4068          */
4069         public function getConverter() {
4070                 return $this->mConverter;
4071         }
4072
4073         /**
4074          * convert text to all supported variants
4075          *
4076          * @param string $text
4077          * @return array
4078          */
4079         public function autoConvertToAllVariants( $text ) {
4080                 return $this->mConverter->autoConvertToAllVariants( $text );
4081         }
4082
4083         /**
4084          * convert text to different variants of a language.
4085          *
4086          * @param string $text
4087          * @return string
4088          */
4089         public function convert( $text ) {
4090                 return $this->mConverter->convert( $text );
4091         }
4092
4093         /**
4094          * Convert a Title object to a string in the preferred variant
4095          *
4096          * @param Title $title
4097          * @return string
4098          */
4099         public function convertTitle( $title ) {
4100                 return $this->mConverter->convertTitle( $title );
4101         }
4102
4103         /**
4104          * Convert a namespace index to a string in the preferred variant
4105          *
4106          * @param int $ns
4107          * @return string
4108          */
4109         public function convertNamespace( $ns ) {
4110                 return $this->mConverter->convertNamespace( $ns );
4111         }
4112
4113         /**
4114          * Check if this is a language with variants
4115          *
4116          * @return bool
4117          */
4118         public function hasVariants() {
4119                 return count( $this->getVariants() ) > 1;
4120         }
4121
4122         /**
4123          * Check if the language has the specific variant
4124          *
4125          * @since 1.19
4126          * @param string $variant
4127          * @return bool
4128          */
4129         public function hasVariant( $variant ) {
4130                 return (bool)$this->mConverter->validateVariant( $variant );
4131         }
4132
4133         /**
4134          * Perform output conversion on a string, and encode for safe HTML output.
4135          * @param string $text Text to be converted
4136          * @param bool $isTitle Whether this conversion is for the article title
4137          * @return string
4138          * @todo this should get integrated somewhere sane
4139          */
4140         public function convertHtml( $text, $isTitle = false ) {
4141                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
4142         }
4143
4144         /**
4145          * @param string $key
4146          * @return string
4147          */
4148         public function convertCategoryKey( $key ) {
4149                 return $this->mConverter->convertCategoryKey( $key );
4150         }
4151
4152         /**
4153          * Get the list of variants supported by this language
4154          * see sample implementation in LanguageZh.php
4155          *
4156          * @return array An array of language codes
4157          */
4158         public function getVariants() {
4159                 return $this->mConverter->getVariants();
4160         }
4161
4162         /**
4163          * @return string
4164          */
4165         public function getPreferredVariant() {
4166                 return $this->mConverter->getPreferredVariant();
4167         }
4168
4169         /**
4170          * @return string
4171          */
4172         public function getDefaultVariant() {
4173                 return $this->mConverter->getDefaultVariant();
4174         }
4175
4176         /**
4177          * @return string
4178          */
4179         public function getURLVariant() {
4180                 return $this->mConverter->getURLVariant();
4181         }
4182
4183         /**
4184          * If a language supports multiple variants, it is
4185          * possible that non-existing link in one variant
4186          * actually exists in another variant. this function
4187          * tries to find it. See e.g. LanguageZh.php
4188          * The input parameters may be modified upon return
4189          *
4190          * @param string &$link The name of the link
4191          * @param Title &$nt The title object of the link
4192          * @param bool $ignoreOtherCond To disable other conditions when
4193          *   we need to transclude a template or update a category's link
4194          */
4195         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4196                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4197         }
4198
4199         /**
4200          * returns language specific options used by User::getPageRenderHash()
4201          * for example, the preferred language variant
4202          *
4203          * @return string
4204          */
4205         function getExtraHashOptions() {
4206                 return $this->mConverter->getExtraHashOptions();
4207         }
4208
4209         /**
4210          * For languages that support multiple variants, the title of an
4211          * article may be displayed differently in different variants. this
4212          * function returns the apporiate title defined in the body of the article.
4213          *
4214          * @return string
4215          */
4216         public function getParsedTitle() {
4217                 return $this->mConverter->getParsedTitle();
4218         }
4219
4220         /**
4221          * Refresh the cache of conversion tables when
4222          * MediaWiki:Conversiontable* is updated.
4223          *
4224          * @param Title $titleobj The Title of the page being updated
4225          */
4226         public function updateConversionTable( Title $title ) {
4227                 $this->mConverter->updateConversionTable( $title );
4228         }
4229
4230         /**
4231          * Prepare external link text for conversion. When the text is
4232          * a URL, it shouldn't be converted, and it'll be wrapped in
4233          * the "raw" tag (-{R| }-) to prevent conversion.
4234          *
4235          * This function is called "markNoConversion" for historical
4236          * reasons.
4237          *
4238          * @param string $text Text to be used for external link
4239          * @param bool $noParse Wrap it without confirming it's a real URL first
4240          * @return string The tagged text
4241          */
4242         public function markNoConversion( $text, $noParse = false ) {
4243                 // Excluding protocal-relative URLs may avoid many false positives.
4244                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
4245                         return $this->mConverter->markNoConversion( $text );
4246                 } else {
4247                         return $text;
4248                 }
4249         }
4250
4251         /**
4252          * A regular expression to match legal word-trailing characters
4253          * which should be merged onto a link of the form [[foo]]bar.
4254          *
4255          * @return string
4256          */
4257         public function linkTrail() {
4258                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4259         }
4260
4261         /**
4262          * A regular expression character set to match legal word-prefixing
4263          * characters which should be merged onto a link of the form foo[[bar]].
4264          *
4265          * @return string
4266          */
4267         public function linkPrefixCharset() {
4268                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4269         }
4270
4271         /**
4272          * @deprecated since 1.24, will be removed in 1.25
4273          * @return Language
4274          */
4275         function getLangObj() {
4276                 wfDeprecated( __METHOD__, '1.24' );
4277                 return $this;
4278         }
4279
4280         /**
4281          * Get the "parent" language which has a converter to convert a "compatible" language
4282          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4283          *
4284          * @return Language|null
4285          * @since 1.22
4286          */
4287         public function getParentLanguage() {
4288                 if ( $this->mParentLanguage !== false ) {
4289                         return $this->mParentLanguage;
4290                 }
4291
4292                 $code = explode( '-', $this->getCode() )[0];
4293                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4294                         $this->mParentLanguage = null;
4295                         return null;
4296                 }
4297                 $lang = Language::factory( $code );
4298                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4299                         $this->mParentLanguage = null;
4300                         return null;
4301                 }
4302
4303                 $this->mParentLanguage = $lang;
4304                 return $lang;
4305         }
4306
4307         /**
4308          * Get the internal language code for this language object
4309          *
4310          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4311          * htmlspecialchars() or similar
4312          *
4313          * @return string
4314          */
4315         public function getCode() {
4316                 return $this->mCode;
4317         }
4318
4319         /**
4320          * Get the code in BCP 47 format which we can use
4321          * inside of html lang="" tags.
4322          *
4323          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4324          * htmlspecialchars() or similar.
4325          *
4326          * @since 1.19
4327          * @return string
4328          */
4329         public function getHtmlCode() {
4330                 if ( is_null( $this->mHtmlCode ) ) {
4331                         $this->mHtmlCode = wfBCP47( $this->getCode() );
4332                 }
4333                 return $this->mHtmlCode;
4334         }
4335
4336         /**
4337          * @param string $code
4338          */
4339         public function setCode( $code ) {
4340                 $this->mCode = $code;
4341                 // Ensure we don't leave incorrect cached data lying around
4342                 $this->mHtmlCode = null;
4343                 $this->mParentLanguage = false;
4344         }
4345
4346         /**
4347          * Get the language code from a file name. Inverse of getFileName()
4348          * @param string $filename $prefix . $languageCode . $suffix
4349          * @param string $prefix Prefix before the language code
4350          * @param string $suffix Suffix after the language code
4351          * @return string Language code, or false if $prefix or $suffix isn't found
4352          */
4353         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4354                 $m = null;
4355                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4356                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4357                 if ( !count( $m ) ) {
4358                         return false;
4359                 }
4360                 return str_replace( '_', '-', strtolower( $m[1] ) );
4361         }
4362
4363         /**
4364          * @param string $code
4365          * @return string Name of the language class
4366          */
4367         public static function classFromCode( $code ) {
4368                 if ( $code == 'en' ) {
4369                         return 'Language';
4370                 } else {
4371                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4372                 }
4373         }
4374
4375         /**
4376          * Get the name of a file for a certain language code
4377          * @param string $prefix Prepend this to the filename
4378          * @param string $code Language code
4379          * @param string $suffix Append this to the filename
4380          * @throws MWException
4381          * @return string $prefix . $mangledCode . $suffix
4382          */
4383         public static function getFileName( $prefix = 'Language', $code, $suffix = '.php' ) {
4384                 if ( !self::isValidBuiltInCode( $code ) ) {
4385                         throw new MWException( "Invalid language code \"$code\"" );
4386                 }
4387
4388                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4389         }
4390
4391         /**
4392          * @param string $code
4393          * @return string
4394          */
4395         public static function getMessagesFileName( $code ) {
4396                 global $IP;
4397                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4398                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4399                 return $file;
4400         }
4401
4402         /**
4403          * @param string $code
4404          * @return string
4405          * @since 1.23
4406          */
4407         public static function getJsonMessagesFileName( $code ) {
4408                 global $IP;
4409
4410                 if ( !self::isValidBuiltInCode( $code ) ) {
4411                         throw new MWException( "Invalid language code \"$code\"" );
4412                 }
4413
4414                 return "$IP/languages/i18n/$code.json";
4415         }
4416
4417         /**
4418          * Get the first fallback for a given language.
4419          *
4420          * @param string $code
4421          *
4422          * @return bool|string
4423          */
4424         public static function getFallbackFor( $code ) {
4425                 $fallbacks = self::getFallbacksFor( $code );
4426                 if ( $fallbacks ) {
4427                         return $fallbacks[0];
4428                 }
4429                 return false;
4430         }
4431
4432         /**
4433          * Get the ordered list of fallback languages.
4434          *
4435          * @since 1.19
4436          * @param string $code Language code
4437          * @return array Non-empty array, ending in "en"
4438          */
4439         public static function getFallbacksFor( $code ) {
4440                 if ( $code === 'en' || !Language::isValidBuiltInCode( $code ) ) {
4441                         return [];
4442                 }
4443                 // For unknown languages, fallbackSequence returns an empty array,
4444                 // hardcode fallback to 'en' in that case.
4445                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4446         }
4447
4448         /**
4449          * Get the ordered list of fallback languages, ending with the fallback
4450          * language chain for the site language.
4451          *
4452          * @since 1.22
4453          * @param string $code Language code
4454          * @return array Array( fallbacks, site fallbacks )
4455          */
4456         public static function getFallbacksIncludingSiteLanguage( $code ) {
4457                 global $wgLanguageCode;
4458
4459                 // Usually, we will only store a tiny number of fallback chains, so we
4460                 // keep them in static memory.
4461                 $cacheKey = "{$code}-{$wgLanguageCode}";
4462
4463                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4464                         $fallbacks = self::getFallbacksFor( $code );
4465
4466                         // Append the site's fallback chain, including the site language itself
4467                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4468                         array_unshift( $siteFallbacks, $wgLanguageCode );
4469
4470                         // Eliminate any languages already included in the chain
4471                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4472
4473                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4474                 }
4475                 return self::$fallbackLanguageCache[$cacheKey];
4476         }
4477
4478         /**
4479          * Get all messages for a given language
4480          * WARNING: this may take a long time. If you just need all message *keys*
4481          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4482          *
4483          * @param string $code
4484          *
4485          * @return array
4486          */
4487         public static function getMessagesFor( $code ) {
4488                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4489         }
4490
4491         /**
4492          * Get a message for a given language
4493          *
4494          * @param string $key
4495          * @param string $code
4496          *
4497          * @return string
4498          */
4499         public static function getMessageFor( $key, $code ) {
4500                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4501         }
4502
4503         /**
4504          * Get all message keys for a given language. This is a faster alternative to
4505          * array_keys( Language::getMessagesFor( $code ) )
4506          *
4507          * @since 1.19
4508          * @param string $code Language code
4509          * @return array Array of message keys (strings)
4510          */
4511         public static function getMessageKeysFor( $code ) {
4512                 return self::getLocalisationCache()->getSubItemList( $code, 'messages' );
4513         }
4514
4515         /**
4516          * @param string $talk
4517          * @return mixed
4518          */
4519         function fixVariableInNamespace( $talk ) {
4520                 if ( strpos( $talk, '$1' ) === false ) {
4521                         return $talk;
4522                 }
4523
4524                 global $wgMetaNamespace;
4525                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4526
4527                 # Allow grammar transformations
4528                 # Allowing full message-style parsing would make simple requests
4529                 # such as action=raw much more expensive than they need to be.
4530                 # This will hopefully cover most cases.
4531                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4532                         [ &$this, 'replaceGrammarInNamespace' ], $talk );
4533                 return str_replace( ' ', '_', $talk );
4534         }
4535
4536         /**
4537          * @param string $m
4538          * @return string
4539          */
4540         function replaceGrammarInNamespace( $m ) {
4541                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4542         }
4543
4544         /**
4545          * @throws MWException
4546          * @return array
4547          */
4548         static function getCaseMaps() {
4549                 static $wikiUpperChars, $wikiLowerChars;
4550                 if ( isset( $wikiUpperChars ) ) {
4551                         return [ $wikiUpperChars, $wikiLowerChars ];
4552                 }
4553
4554                 $arr = wfGetPrecompiledData( 'Utf8Case.ser' );
4555                 if ( $arr === false ) {
4556                         throw new MWException(
4557                                 "Utf8Case.ser is missing, please run \"make\" in the serialized directory\n" );
4558                 }
4559                 $wikiUpperChars = $arr['wikiUpperChars'];
4560                 $wikiLowerChars = $arr['wikiLowerChars'];
4561                 return [ $wikiUpperChars, $wikiLowerChars ];
4562         }
4563
4564         /**
4565          * Decode an expiry (block, protection, etc) which has come from the DB
4566          *
4567          * @param string $expiry Database expiry String
4568          * @param bool|int $format True to process using language functions, or TS_ constant
4569          *     to return the expiry in a given timestamp
4570          * @param string $inifinity If $format is not true, use this string for infinite expiry
4571          * @return string
4572          * @since 1.18
4573          */
4574         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4575                 static $dbInfinity;
4576                 if ( $dbInfinity === null ) {
4577                         $dbInfinity = wfGetDB( DB_SLAVE )->getInfinity();
4578                 }
4579
4580                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4581                         return $format === true
4582                                 ? $this->getMessageFromDB( 'infiniteblock' )
4583                                 : $infinity;
4584                 } else {
4585                         return $format === true
4586                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4587                                 : wfTimestamp( $format, $expiry );
4588                 }
4589         }
4590
4591         /**
4592          * @todo Document
4593          * @param int|float $seconds
4594          * @param array $format Optional
4595          *   If $format['avoid'] === 'avoidseconds': don't mention seconds if $seconds >= 1 hour.
4596          *   If $format['avoid'] === 'avoidminutes': don't mention seconds/minutes if $seconds > 48 hours.
4597          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4598          *     and friends.
4599          *   For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4600          *     or 'avoidminutes'.
4601          * @return string
4602          */
4603         function formatTimePeriod( $seconds, $format = [] ) {
4604                 if ( !is_array( $format ) ) {
4605                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4606                 }
4607                 if ( !isset( $format['avoid'] ) ) {
4608                         $format['avoid'] = false;
4609                 }
4610                 if ( !isset( $format['noabbrevs'] ) ) {
4611                         $format['noabbrevs'] = false;
4612                 }
4613                 $secondsMsg = wfMessage(
4614                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4615                 $minutesMsg = wfMessage(
4616                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4617                 $hoursMsg = wfMessage(
4618                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4619                 $daysMsg = wfMessage(
4620                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4621
4622                 if ( round( $seconds * 10 ) < 100 ) {
4623                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4624                         $s = $secondsMsg->params( $s )->text();
4625                 } elseif ( round( $seconds ) < 60 ) {
4626                         $s = $this->formatNum( round( $seconds ) );
4627                         $s = $secondsMsg->params( $s )->text();
4628                 } elseif ( round( $seconds ) < 3600 ) {
4629                         $minutes = floor( $seconds / 60 );
4630                         $secondsPart = round( fmod( $seconds, 60 ) );
4631                         if ( $secondsPart == 60 ) {
4632                                 $secondsPart = 0;
4633                                 $minutes++;
4634                         }
4635                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4636                         $s .= ' ';
4637                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4638                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4639                         $hours = floor( $seconds / 3600 );
4640                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4641                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4642                         if ( $secondsPart == 60 ) {
4643                                 $secondsPart = 0;
4644                                 $minutes++;
4645                         }
4646                         if ( $minutes == 60 ) {
4647                                 $minutes = 0;
4648                                 $hours++;
4649                         }
4650                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4651                         $s .= ' ';
4652                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4653                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4654                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4655                         }
4656                 } else {
4657                         $days = floor( $seconds / 86400 );
4658                         if ( $format['avoid'] === 'avoidminutes' ) {
4659                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4660                                 if ( $hours == 24 ) {
4661                                         $hours = 0;
4662                                         $days++;
4663                                 }
4664                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4665                                 $s .= ' ';
4666                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4667                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4668                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4669                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4670                                 if ( $minutes == 60 ) {
4671                                         $minutes = 0;
4672                                         $hours++;
4673                                 }
4674                                 if ( $hours == 24 ) {
4675                                         $hours = 0;
4676                                         $days++;
4677                                 }
4678                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4679                                 $s .= ' ';
4680                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4681                                 $s .= ' ';
4682                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4683                         } else {
4684                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4685                                 $s .= ' ';
4686                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4687                         }
4688                 }
4689                 return $s;
4690         }
4691
4692         /**
4693          * Format a bitrate for output, using an appropriate
4694          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4695          *   the magnitude in question.
4696          *
4697          * This use base 1000. For base 1024 use formatSize(), for another base
4698          * see formatComputingNumbers().
4699          *
4700          * @param int $bps
4701          * @return string
4702          */
4703         function formatBitrate( $bps ) {
4704                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4705         }
4706
4707         /**
4708          * @param int $size Size of the unit
4709          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4710          * @param string $messageKey Message key to be uesd
4711          * @return string
4712          */
4713         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4714                 if ( $size <= 0 ) {
4715                         return str_replace( '$1', $this->formatNum( $size ),
4716                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4717                         );
4718                 }
4719                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4720                 $index = 0;
4721
4722                 $maxIndex = count( $sizes ) - 1;
4723                 while ( $size >= $boundary && $index < $maxIndex ) {
4724                         $index++;
4725                         $size /= $boundary;
4726                 }
4727
4728                 // For small sizes no decimal places necessary
4729                 $round = 0;
4730                 if ( $index > 1 ) {
4731                         // For MB and bigger two decimal places are smarter
4732                         $round = 2;
4733                 }
4734                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4735
4736                 $size = round( $size, $round );
4737                 $text = $this->getMessageFromDB( $msg );
4738                 return str_replace( '$1', $this->formatNum( $size ), $text );
4739         }
4740
4741         /**
4742          * Format a size in bytes for output, using an appropriate
4743          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4744          *
4745          * This method use base 1024. For base 1000 use formatBitrate(), for
4746          * another base see formatComputingNumbers()
4747          *
4748          * @param int $size Size to format
4749          * @return string Plain text (not HTML)
4750          */
4751         function formatSize( $size ) {
4752                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4753         }
4754
4755         /**
4756          * Make a list item, used by various special pages
4757          *
4758          * @param string $page Page link
4759          * @param string $details HTML safe text between brackets
4760          * @param bool $oppositedm Add the direction mark opposite to your
4761          *   language, to display text properly
4762          * @return HTML escaped string
4763          */
4764         function specialList( $page, $details, $oppositedm = true ) {
4765                 if ( !$details ) {
4766                         return $page;
4767                 }
4768
4769                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4770                 return
4771                         $page .
4772                         $dirmark .
4773                         $this->msg( 'word-separator' )->escaped() .
4774                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4775         }
4776
4777         /**
4778          * Generate (prev x| next x) (20|50|100...) type links for paging
4779          *
4780          * @param Title $title Title object to link
4781          * @param int $offset
4782          * @param int $limit
4783          * @param array $query Optional URL query parameter string
4784          * @param bool $atend Optional param for specified if this is the last page
4785          * @return string
4786          */
4787         public function viewPrevNext( Title $title, $offset, $limit,
4788                 array $query = [], $atend = false
4789         ) {
4790                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4791
4792                 # Make 'previous' link
4793                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4794                 if ( $offset > 0 ) {
4795                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4796                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4797                 } else {
4798                         $plink = htmlspecialchars( $prev );
4799                 }
4800
4801                 # Make 'next' link
4802                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4803                 if ( $atend ) {
4804                         $nlink = htmlspecialchars( $next );
4805                 } else {
4806                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4807                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4808                 }
4809
4810                 # Make links to set number of items per page
4811                 $numLinks = [];
4812                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4813                         $numLinks[] = $this->numLink( $title, $offset, $num,
4814                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4815                 }
4816
4817                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4818                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4819         }
4820
4821         /**
4822          * Helper function for viewPrevNext() that generates links
4823          *
4824          * @param Title $title Title object to link
4825          * @param int $offset
4826          * @param int $limit
4827          * @param array $query Extra query parameters
4828          * @param string $link Text to use for the link; will be escaped
4829          * @param string $tooltipMsg Name of the message to use as tooltip
4830          * @param string $class Value of the "class" attribute of the link
4831          * @return string HTML fragment
4832          */
4833         private function numLink( Title $title, $offset, $limit, array $query, $link,
4834                 $tooltipMsg, $class
4835         ) {
4836                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4837                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4838                         ->numParams( $limit )->text();
4839
4840                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4841                         'title' => $tooltip, 'class' => $class ], $link );
4842         }
4843
4844         /**
4845          * Get the conversion rule title, if any.
4846          *
4847          * @return string
4848          */
4849         public function getConvRuleTitle() {
4850                 return $this->mConverter->getConvRuleTitle();
4851         }
4852
4853         /**
4854          * Get the compiled plural rules for the language
4855          * @since 1.20
4856          * @return array Associative array with plural form, and plural rule as key-value pairs
4857          */
4858         public function getCompiledPluralRules() {
4859                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4860                 $fallbacks = Language::getFallbacksFor( $this->mCode );
4861                 if ( !$pluralRules ) {
4862                         foreach ( $fallbacks as $fallbackCode ) {
4863                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4864                                 if ( $pluralRules ) {
4865                                         break;
4866                                 }
4867                         }
4868                 }
4869                 return $pluralRules;
4870         }
4871
4872         /**
4873          * Get the plural rules for the language
4874          * @since 1.20
4875          * @return array Associative array with plural form number and plural rule as key-value pairs
4876          */
4877         public function getPluralRules() {
4878                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4879                 $fallbacks = Language::getFallbacksFor( $this->mCode );
4880                 if ( !$pluralRules ) {
4881                         foreach ( $fallbacks as $fallbackCode ) {
4882                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4883                                 if ( $pluralRules ) {
4884                                         break;
4885                                 }
4886                         }
4887                 }
4888                 return $pluralRules;
4889         }
4890
4891         /**
4892          * Get the plural rule types for the language
4893          * @since 1.22
4894          * @return array Associative array with plural form number and plural rule type as key-value pairs
4895          */
4896         public function getPluralRuleTypes() {
4897                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4898                 $fallbacks = Language::getFallbacksFor( $this->mCode );
4899                 if ( !$pluralRuleTypes ) {
4900                         foreach ( $fallbacks as $fallbackCode ) {
4901                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4902                                 if ( $pluralRuleTypes ) {
4903                                         break;
4904                                 }
4905                         }
4906                 }
4907                 return $pluralRuleTypes;
4908         }
4909
4910         /**
4911          * Find the index number of the plural rule appropriate for the given number
4912          * @param int $number
4913          * @return int The index number of the plural rule
4914          */
4915         public function getPluralRuleIndexNumber( $number ) {
4916                 $pluralRules = $this->getCompiledPluralRules();
4917                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4918                 return $form;
4919         }
4920
4921         /**
4922          * Find the plural rule type appropriate for the given number
4923          * For example, if the language is set to Arabic, getPluralType(5) should
4924          * return 'few'.
4925          * @since 1.22
4926          * @param int $number
4927          * @return string The name of the plural rule type, e.g. one, two, few, many
4928          */
4929         public function getPluralRuleType( $number ) {
4930                 $index = $this->getPluralRuleIndexNumber( $number );
4931                 $pluralRuleTypes = $this->getPluralRuleTypes();
4932                 if ( isset( $pluralRuleTypes[$index] ) ) {
4933                         return $pluralRuleTypes[$index];
4934                 } else {
4935                         return 'other';
4936                 }
4937         }
4938 }