languages/Language.php

   1 <?php
   2 /**
   3  * Internationalisation code.
   4  * See https://www.mediawiki.org/wiki/Special:MyLanguage/Localisation for more information.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Language
  23  */
  24
  25 /**
  26  * @defgroup Language Language
  27  */
  28
  29 use CLDRPluralRuleParser\Evaluator;
  30
  31 /**
  32  * Internationalisation code
  33  * @ingroup Language
  34  */
  35 class Language {
  36         /**
  37          * @var LanguageConverter
  38          */
  39         public $mConverter;
  40
  41         public $mVariants, $mCode, $mLoaded = false;
  42         public $mMagicExtensions = [], $mMagicHookDone = false;
  43         private $mHtmlCode = null, $mParentLanguage = false;
  44
  45         public $dateFormatStrings = [];
  46         public $mExtendedSpecialPageAliases;
  47
  48         /** @var array|null */
  49         protected $namespaceNames;
  50         protected $mNamespaceIds, $namespaceAliases;
  51
  52         /**
  53          * ReplacementArray object caches
  54          */
  55         public $transformData = [];
  56
  57         /**
  58          * @var LocalisationCache
  59          */
  60         static public $dataCache;
  61
  62         static public $mLangObjCache = [];
  63
  64         static public $mWeekdayMsgs = [
  65                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  66                 'friday', 'saturday'
  67         ];
  68
  69         static public $mWeekdayAbbrevMsgs = [
  70                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
  71         ];
  72
  73         static public $mMonthMsgs = [
  74                 'january', 'february', 'march', 'april', 'may_long', 'june',
  75                 'july', 'august', 'september', 'october', 'november',
  76                 'december'
  77         ];
  78         static public $mMonthGenMsgs = [
  79                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
  80                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
  81                 'december-gen'
  82         ];
  83         static public $mMonthAbbrevMsgs = [
  84                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
  85                 'sep', 'oct', 'nov', 'dec'
  86         ];
  87
  88         static public $mIranianCalendarMonthMsgs = [
  89                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
  90                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
  91                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
  92                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
  93         ];
  94
  95         static public $mHebrewCalendarMonthMsgs = [
  96                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
  97                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
  98                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
  99                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 100                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 101         ];
 102
 103         static public $mHebrewCalendarMonthGenMsgs = [
 104                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 105                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 106                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 107                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 108                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 109         ];
 110
 111         static public $mHijriCalendarMonthMsgs = [
 112                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 113                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 114                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 115                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 116         ];
 117
 118         /**
 119          * @since 1.20
 120          * @var array
 121          */
 122         static public $durationIntervals = [
 123                 'millennia' => 31556952000,
 124                 'centuries' => 3155695200,
 125                 'decades' => 315569520,
 126                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 127                 'weeks' => 604800,
 128                 'days' => 86400,
 129                 'hours' => 3600,
 130                 'minutes' => 60,
 131                 'seconds' => 1,
 132         ];
 133
 134         /**
 135          * Cache for language fallbacks.
 136          * @see Language::getFallbacksIncludingSiteLanguage
 137          * @since 1.21
 138          * @var array
 139          */
 140         static private $fallbackLanguageCache = [];
 141
 142         /**
 143          * Cache for grammar rules data
 144          * @var MapCacheLRU|null
 145          */
 146         static private $grammarTransformations;
 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          * @param $fallback boolean Whether we're going through language fallback chain
 207          * @throws MWException
 208          * @return Language
 209          */
 210         protected static function newFromCode( $code, $fallback = false ) {
 211                 if ( !Language::isValidCode( $code ) ) {
 212                         throw new MWException( "Invalid language code \"$code\"" );
 213                 }
 214
 215                 if ( !Language::isValidBuiltInCode( $code ) ) {
 216                         // It's not possible to customise this code with class files, so
 217                         // just return a Language object. This is to support uselang= hacks.
 218                         $lang = new Language;
 219                         $lang->setCode( $code );
 220                         return $lang;
 221                 }
 222
 223                 // Check if there is a language class for the code
 224                 $class = self::classFromCode( $code, $fallback );
 225                 if ( class_exists( $class ) ) {
 226                         $lang = new $class;
 227                         return $lang;
 228                 }
 229
 230                 // Keep trying the fallback list until we find an existing class
 231                 $fallbacks = Language::getFallbacksFor( $code );
 232                 foreach ( $fallbacks as $fallbackCode ) {
 233                         if ( !Language::isValidBuiltInCode( $fallbackCode ) ) {
 234                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 235                         }
 236
 237                         $class = self::classFromCode( $fallbackCode );
 238                         if ( class_exists( $class ) ) {
 239                                 $lang = new $class;
 240                                 $lang->setCode( $code );
 241                                 return $lang;
 242                         }
 243                 }
 244
 245                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 246         }
 247
 248         /**
 249          * Checks whether any localisation is available for that language tag
 250          * in MediaWiki (MessagesXx.php exists).
 251          *
 252          * @param string $code Language tag (in lower case)
 253          * @return bool Whether language is supported
 254          * @since 1.21
 255          */
 256         public static function isSupportedLanguage( $code ) {
 257                 if ( !self::isValidBuiltInCode( $code ) ) {
 258                         return false;
 259                 }
 260
 261                 if ( $code === 'qqq' ) {
 262                         return false;
 263                 }
 264
 265                 return is_readable( self::getMessagesFileName( $code ) ) ||
 266                         is_readable( self::getJsonMessagesFileName( $code ) );
 267         }
 268
 269         /**
 270          * Returns true if a language code string is a well-formed language tag
 271          * according to RFC 5646.
 272          * This function only checks well-formedness; it doesn't check that
 273          * language, script or variant codes actually exist in the repositories.
 274          *
 275          * Based on regexes by Mark Davis of the Unicode Consortium:
 276          * http://unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 277          *
 278          * @param string $code
 279          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 280          *
 281          * @return bool
 282          * @since 1.21
 283          */
 284         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 285                 $alpha = '[a-z]';
 286                 $digit = '[0-9]';
 287                 $alphanum = '[a-z0-9]';
 288                 $x = 'x'; # private use singleton
 289                 $singleton = '[a-wy-z]'; # other singleton
 290                 $s = $lenient ? '[-_]' : '-';
 291
 292                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 293                 $script = "$alpha{4}"; # ISO 15924
 294                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 295                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 296                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 297                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 298
 299                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 300                 # Since these are limited, this is safe even later changes to the registry --
 301                 # the only oddity is that it might change the type of the tag, and thus
 302                 # the results from the capturing groups.
 303                 # https://www.iana.org/assignments/language-subtag-registry
 304
 305                 $grandfathered = "en{$s}GB{$s}oed"
 306                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 307                         . "|no{$s}(?:bok|nyn)"
 308                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 309                         . "|zh{$s}min{$s}nan";
 310
 311                 $variantList = "$variant(?:$s$variant)*";
 312                 $extensionList = "$extension(?:$s$extension)*";
 313
 314                 $langtag = "(?:($language)"
 315                         . "(?:$s$script)?"
 316                         . "(?:$s$region)?"
 317                         . "(?:$s$variantList)?"
 318                         . "(?:$s$extensionList)?"
 319                         . "(?:$s$privateUse)?)";
 320
 321                 # The final breakdown, with capturing groups for each of these components
 322                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 323
 324                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 325
 326                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 327         }
 328
 329         /**
 330          * Returns true if a language code string is of a valid form, whether or
 331          * not it exists. This includes codes which are used solely for
 332          * customisation via the MediaWiki namespace.
 333          *
 334          * @param string $code
 335          *
 336          * @return bool
 337          */
 338         public static function isValidCode( $code ) {
 339                 static $cache = [];
 340                 if ( !isset( $cache[$code] ) ) {
 341                         // People think language codes are html safe, so enforce it.
 342                         // Ideally we should only allow a-zA-Z0-9-
 343                         // but, .+ and other chars are often used for {{int:}} hacks
 344                         // see bugs T39564, T39587, T38938
 345                         $cache[$code] =
 346                                 // Protect against path traversal
 347                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 348                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 349                 }
 350                 return $cache[$code];
 351         }
 352
 353         /**
 354          * Returns true if a language code is of a valid form for the purposes of
 355          * internal customisation of MediaWiki, via Messages*.php or *.json.
 356          *
 357          * @param string $code
 358          *
 359          * @throws MWException
 360          * @since 1.18
 361          * @return bool
 362          */
 363         public static function isValidBuiltInCode( $code ) {
 364
 365                 if ( !is_string( $code ) ) {
 366                         if ( is_object( $code ) ) {
 367                                 $addmsg = " of class " . get_class( $code );
 368                         } else {
 369                                 $addmsg = '';
 370                         }
 371                         $type = gettype( $code );
 372                         throw new MWException( __METHOD__ . " must be passed a string, $type given$addmsg" );
 373                 }
 374
 375                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 376         }
 377
 378         /**
 379          * Returns true if a language code is an IETF tag known to MediaWiki.
 380          *
 381          * @param string $tag
 382          *
 383          * @since 1.21
 384          * @return bool
 385          */
 386         public static function isKnownLanguageTag( $tag ) {
 387                 // Quick escape for invalid input to avoid exceptions down the line
 388                 // when code tries to process tags which are not valid at all.
 389                 if ( !self::isValidBuiltInCode( $tag ) ) {
 390                         return false;
 391                 }
 392
 393                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 394                         || self::fetchLanguageName( $tag, $tag ) !== ''
 395                 ) {
 396                         return true;
 397                 }
 398
 399                 return false;
 400         }
 401
 402         /**
 403          * Get the LocalisationCache instance
 404          *
 405          * @return LocalisationCache
 406          */
 407         public static function getLocalisationCache() {
 408                 if ( is_null( self::$dataCache ) ) {
 409                         global $wgLocalisationCacheConf;
 410                         $class = $wgLocalisationCacheConf['class'];
 411                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 412                 }
 413                 return self::$dataCache;
 414         }
 415
 416         function __construct() {
 417                 $this->mConverter = new FakeConverter( $this );
 418                 // Set the code to the name of the descendant
 419                 if ( static::class === 'Language' ) {
 420                         $this->mCode = 'en';
 421                 } else {
 422                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 423                 }
 424                 self::getLocalisationCache();
 425         }
 426
 427         /**
 428          * Reduce memory usage
 429          */
 430         function __destruct() {
 431                 foreach ( $this as $name => $value ) {
 432                         unset( $this->$name );
 433                 }
 434         }
 435
 436         /**
 437          * Hook which will be called if this is the content language.
 438          * Descendants can use this to register hook functions or modify globals
 439          */
 440         function initContLang() {
 441         }
 442
 443         /**
 444          * @return array
 445          * @since 1.19
 446          */
 447         public function getFallbackLanguages() {
 448                 return self::getFallbacksFor( $this->mCode );
 449         }
 450
 451         /**
 452          * Exports $wgBookstoreListEn
 453          * @return array
 454          */
 455         public function getBookstoreList() {
 456                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 457         }
 458
 459         /**
 460          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 461          * available in localised form, it will be included in English.
 462          *
 463          * @return array
 464          */
 465         public function getNamespaces() {
 466                 if ( is_null( $this->namespaceNames ) ) {
 467                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 468
 469                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 470
 471                         $this->namespaceNames = $wgExtraNamespaces +
 472                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 473                         $this->namespaceNames += $validNamespaces;
 474
 475                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 476                         if ( $wgMetaNamespaceTalk ) {
 477                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 478                         } else {
 479                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 480                                 $this->namespaceNames[NS_PROJECT_TALK] =
 481                                         $this->fixVariableInNamespace( $talk );
 482                         }
 483
 484                         # Sometimes a language will be localised but not actually exist on this wiki.
 485                         foreach ( $this->namespaceNames as $key => $text ) {
 486                                 if ( !isset( $validNamespaces[$key] ) ) {
 487                                         unset( $this->namespaceNames[$key] );
 488                                 }
 489                         }
 490
 491                         # The above mixing may leave namespaces out of canonical order.
 492                         # Re-order by namespace ID number...
 493                         ksort( $this->namespaceNames );
 494
 495                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 496                 }
 497
 498                 return $this->namespaceNames;
 499         }
 500
 501         /**
 502          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 503          * @param array $namespaces Array of namespaces (id => name)
 504          */
 505         public function setNamespaces( array $namespaces ) {
 506                 $this->namespaceNames = $namespaces;
 507                 $this->mNamespaceIds = null;
 508         }
 509
 510         /**
 511          * Resets all of the namespace caches. Mainly used for testing
 512          */
 513         public function resetNamespaces() {
 514                 $this->namespaceNames = null;
 515                 $this->mNamespaceIds = null;
 516                 $this->namespaceAliases = null;
 517         }
 518
 519         /**
 520          * A convenience function that returns getNamespaces() with spaces instead of underscores
 521          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 522          *
 523          * @return array
 524          */
 525         public function getFormattedNamespaces() {
 526                 $ns = $this->getNamespaces();
 527                 foreach ( $ns as $k => $v ) {
 528                         $ns[$k] = strtr( $v, '_', ' ' );
 529                 }
 530                 return $ns;
 531         }
 532
 533         /**
 534          * Get a namespace value by key
 535          *
 536          * <code>
 537          * $mw_ns = $wgContLang->getNsText( NS_MEDIAWIKI );
 538          * echo $mw_ns; // prints 'MediaWiki'
 539          * </code>
 540          *
 541          * @param int $index The array key of the namespace to return
 542          * @return string|bool String if the namespace value exists, otherwise false
 543          */
 544         public function getNsText( $index ) {
 545                 $ns = $this->getNamespaces();
 546                 return isset( $ns[$index] ) ? $ns[$index] : false;
 547         }
 548
 549         /**
 550          * A convenience function that returns the same thing as
 551          * getNsText() except with '_' changed to ' ', useful for
 552          * producing output.
 553          *
 554          * <code>
 555          * $mw_ns = $wgContLang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 556          * echo $mw_ns; // prints 'MediaWiki talk'
 557          * </code>
 558          *
 559          * @param int $index The array key of the namespace to return
 560          * @return string Namespace name without underscores (empty string if namespace does not exist)
 561          */
 562         public function getFormattedNsText( $index ) {
 563                 $ns = $this->getNsText( $index );
 564                 return strtr( $ns, '_', ' ' );
 565         }
 566
 567         /**
 568          * Returns gender-dependent namespace alias if available.
 569          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 570          * @param int $index Namespace index
 571          * @param string $gender Gender key (male, female... )
 572          * @return string
 573          * @since 1.18
 574          */
 575         public function getGenderNsText( $index, $gender ) {
 576                 global $wgExtraGenderNamespaces;
 577
 578                 $ns = $wgExtraGenderNamespaces +
 579                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 580
 581                 return isset( $ns[$index][$gender] ) ? $ns[$index][$gender] : $this->getNsText( $index );
 582         }
 583
 584         /**
 585          * Whether this language uses gender-dependent namespace aliases.
 586          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 587          * @return bool
 588          * @since 1.18
 589          */
 590         public function needsGenderDistinction() {
 591                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 592                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 593                         // $wgExtraGenderNamespaces overrides everything
 594                         return true;
 595                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 596                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 597                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 598                         return false;
 599                 } else {
 600                         // Check what is in i18n files
 601                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 602                         return count( $aliases ) > 0;
 603                 }
 604         }
 605
 606         /**
 607          * Get a namespace key by value, case insensitive.
 608          * Only matches namespace names for the current language, not the
 609          * canonical ones defined in Namespace.php.
 610          *
 611          * @param string $text
 612          * @return int|bool An integer if $text is a valid value otherwise false
 613          */
 614         function getLocalNsIndex( $text ) {
 615                 $lctext = $this->lc( $text );
 616                 $ids = $this->getNamespaceIds();
 617                 return isset( $ids[$lctext] ) ? $ids[$lctext] : false;
 618         }
 619
 620         /**
 621          * @return array
 622          */
 623         public function getNamespaceAliases() {
 624                 if ( is_null( $this->namespaceAliases ) ) {
 625                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 626                         if ( !$aliases ) {
 627                                 $aliases = [];
 628                         } else {
 629                                 foreach ( $aliases as $name => $index ) {
 630                                         if ( $index === NS_PROJECT_TALK ) {
 631                                                 unset( $aliases[$name] );
 632                                                 $name = $this->fixVariableInNamespace( $name );
 633                                                 $aliases[$name] = $index;
 634                                         }
 635                                 }
 636                         }
 637
 638                         global $wgExtraGenderNamespaces;
 639                         $genders = $wgExtraGenderNamespaces +
 640                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 641                         foreach ( $genders as $index => $forms ) {
 642                                 foreach ( $forms as $alias ) {
 643                                         $aliases[$alias] = $index;
 644                                 }
 645                         }
 646
 647                         # Also add converted namespace names as aliases, to avoid confusion.
 648                         $convertedNames = [];
 649                         foreach ( $this->getVariants() as $variant ) {
 650                                 if ( $variant === $this->mCode ) {
 651                                         continue;
 652                                 }
 653                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 654                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 655                                 }
 656                         }
 657
 658                         $this->namespaceAliases = $aliases + $convertedNames;
 659                 }
 660
 661                 return $this->namespaceAliases;
 662         }
 663
 664         /**
 665          * @return array
 666          */
 667         public function getNamespaceIds() {
 668                 if ( is_null( $this->mNamespaceIds ) ) {
 669                         global $wgNamespaceAliases;
 670                         # Put namespace names and aliases into a hashtable.
 671                         # If this is too slow, then we should arrange it so that it is done
 672                         # before caching. The catch is that at pre-cache time, the above
 673                         # class-specific fixup hasn't been done.
 674                         $this->mNamespaceIds = [];
 675                         foreach ( $this->getNamespaces() as $index => $name ) {
 676                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 677                         }
 678                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 679                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 680                         }
 681                         if ( $wgNamespaceAliases ) {
 682                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 683                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 684                                 }
 685                         }
 686                 }
 687                 return $this->mNamespaceIds;
 688         }
 689
 690         /**
 691          * Get a namespace key by value, case insensitive.  Canonical namespace
 692          * names override custom ones defined for the current language.
 693          *
 694          * @param string $text
 695          * @return int|bool An integer if $text is a valid value otherwise false
 696          */
 697         public function getNsIndex( $text ) {
 698                 $lctext = $this->lc( $text );
 699                 $ns = MWNamespace::getCanonicalIndex( $lctext );
 700                 if ( $ns !== null ) {
 701                         return $ns;
 702                 }
 703                 $ids = $this->getNamespaceIds();
 704                 return isset( $ids[$lctext] ) ? $ids[$lctext] : false;
 705         }
 706
 707         /**
 708          * short names for language variants used for language conversion links.
 709          *
 710          * @param string $code
 711          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 712          * @return string
 713          */
 714         public function getVariantname( $code, $usemsg = true ) {
 715                 $msg = "variantname-$code";
 716                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 717                         return $this->getMessageFromDB( $msg );
 718                 }
 719                 $name = self::fetchLanguageName( $code );
 720                 if ( $name ) {
 721                         return $name; # if it's defined as a language name, show that
 722                 } else {
 723                         # otherwise, output the language code
 724                         return $code;
 725                 }
 726         }
 727
 728         /**
 729          * @return string[]|bool List of date format preference keys, or false if disabled.
 730          */
 731         public function getDatePreferences() {
 732                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 733         }
 734
 735         /**
 736          * @return array
 737          */
 738         function getDateFormats() {
 739                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 740         }
 741
 742         /**
 743          * @return array|string
 744          */
 745         public function getDefaultDateFormat() {
 746                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 747                 if ( $df === 'dmy or mdy' ) {
 748                         global $wgAmericanDates;
 749                         return $wgAmericanDates ? 'mdy' : 'dmy';
 750                 } else {
 751                         return $df;
 752                 }
 753         }
 754
 755         /**
 756          * @return array
 757          */
 758         public function getDatePreferenceMigrationMap() {
 759                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 760         }
 761
 762         /**
 763          * @param string $image
 764          * @return array|null
 765          */
 766         function getImageFile( $image ) {
 767                 return self::$dataCache->getSubitem( $this->mCode, 'imageFiles', $image );
 768         }
 769
 770         /**
 771          * @return array
 772          * @since 1.24
 773          */
 774         public function getImageFiles() {
 775                 return self::$dataCache->getItem( $this->mCode, 'imageFiles' );
 776         }
 777
 778         /**
 779          * @return array
 780          */
 781         public function getExtraUserToggles() {
 782                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 783         }
 784
 785         /**
 786          * @param string $tog
 787          * @return string
 788          */
 789         function getUserToggle( $tog ) {
 790                 return $this->getMessageFromDB( "tog-$tog" );
 791         }
 792
 793         /**
 794          * Get an array of language names, indexed by code.
 795          * @param null|string $inLanguage Code of language in which to return the names
 796          *              Use null for autonyms (native names)
 797          * @param string $include One of:
 798          *              'all' all available languages
 799          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 800          *              'mwfile' only if the language is in 'mw' *and* has a message file
 801          * @return array Language code => language name
 802          * @since 1.20
 803          */
 804         public static function fetchLanguageNames( $inLanguage = null, $include = 'mw' ) {
 805                 $cacheKey = $inLanguage === null ? 'null' : $inLanguage;
 806                 $cacheKey .= ":$include";
 807                 if ( self::$languageNameCache === null ) {
 808                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 809                 }
 810
 811                 $ret = self::$languageNameCache->get( $cacheKey );
 812                 if ( !$ret ) {
 813                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 814                         self::$languageNameCache->set( $cacheKey, $ret );
 815                 }
 816                 return $ret;
 817         }
 818
 819         /**
 820          * Uncached helper for fetchLanguageNames
 821          * @param null|string $inLanguage Code of language in which to return the names
 822          *              Use null for autonyms (native names)
 823          * @param string $include One of:
 824          *              'all' all available languages
 825          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 826          *              'mwfile' only if the language is in 'mw' *and* has a message file
 827          * @return array Language code => language name
 828          */
 829         private static function fetchLanguageNamesUncached( $inLanguage = null, $include = 'mw' ) {
 830                 global $wgExtraLanguageNames;
 831
 832                 // If passed an invalid language code to use, fallback to en
 833                 if ( $inLanguage !== null && !Language::isValidCode( $inLanguage ) ) {
 834                         $inLanguage = 'en';
 835                 }
 836
 837                 $names = [];
 838
 839                 if ( $inLanguage ) {
 840                         # TODO: also include when $inLanguage is null, when this code is more efficient
 841                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 842                 }
 843
 844                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 845                 foreach ( $mwNames as $mwCode => $mwName ) {
 846                         # - Prefer own MediaWiki native name when not using the hook
 847                         # - For other names just add if not added through the hook
 848                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 849                                 $names[$mwCode] = $mwName;
 850                         }
 851                 }
 852
 853                 if ( $include === 'all' ) {
 854                         ksort( $names );
 855                         return $names;
 856                 }
 857
 858                 $returnMw = [];
 859                 $coreCodes = array_keys( $mwNames );
 860                 foreach ( $coreCodes as $coreCode ) {
 861                         $returnMw[$coreCode] = $names[$coreCode];
 862                 }
 863
 864                 if ( $include === 'mwfile' ) {
 865                         $namesMwFile = [];
 866                         # We do this using a foreach over the codes instead of a directory
 867                         # loop so that messages files in extensions will work correctly.
 868                         foreach ( $returnMw as $code => $value ) {
 869                                 if ( is_readable( self::getMessagesFileName( $code ) )
 870                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 871                                 ) {
 872                                         $namesMwFile[$code] = $names[$code];
 873                                 }
 874                         }
 875
 876                         ksort( $namesMwFile );
 877                         return $namesMwFile;
 878                 }
 879
 880                 ksort( $returnMw );
 881                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 882                 return $returnMw;
 883         }
 884
 885         /**
 886          * @param string $code The code of the language for which to get the name
 887          * @param null|string $inLanguage Code of language in which to return the name (null for autonyms)
 888          * @param string $include 'all', 'mw' or 'mwfile'; see fetchLanguageNames()
 889          * @return string Language name or empty
 890          * @since 1.20
 891          */
 892         public static function fetchLanguageName( $code, $inLanguage = null, $include = 'all' ) {
 893                 $code = strtolower( $code );
 894                 $array = self::fetchLanguageNames( $inLanguage, $include );
 895                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 896         }
 897
 898         /**
 899          * Get a message from the MediaWiki namespace.
 900          *
 901          * @param string $msg Message name
 902          * @return string
 903          */
 904         public function getMessageFromDB( $msg ) {
 905                 return $this->msg( $msg )->text();
 906         }
 907
 908         /**
 909          * Get message object in this language. Only for use inside this class.
 910          *
 911          * @param string $msg Message name
 912          * @return Message
 913          */
 914         protected function msg( $msg ) {
 915                 return wfMessage( $msg )->inLanguage( $this );
 916         }
 917
 918         /**
 919          * @param string $key
 920          * @return string
 921          */
 922         public function getMonthName( $key ) {
 923                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 924         }
 925
 926         /**
 927          * @return array
 928          */
 929         public function getMonthNamesArray() {
 930                 $monthNames = [ '' ];
 931                 for ( $i = 1; $i < 13; $i++ ) {
 932                         $monthNames[] = $this->getMonthName( $i );
 933                 }
 934                 return $monthNames;
 935         }
 936
 937         /**
 938          * @param string $key
 939          * @return string
 940          */
 941         public function getMonthNameGen( $key ) {
 942                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 943         }
 944
 945         /**
 946          * @param string $key
 947          * @return string
 948          */
 949         public function getMonthAbbreviation( $key ) {
 950                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 951         }
 952
 953         /**
 954          * @return array
 955          */
 956         public function getMonthAbbreviationsArray() {
 957                 $monthNames = [ '' ];
 958                 for ( $i = 1; $i < 13; $i++ ) {
 959                         $monthNames[] = $this->getMonthAbbreviation( $i );
 960                 }
 961                 return $monthNames;
 962         }
 963
 964         /**
 965          * @param string $key
 966          * @return string
 967          */
 968         public function getWeekdayName( $key ) {
 969                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
 970         }
 971
 972         /**
 973          * @param string $key
 974          * @return string
 975          */
 976         function getWeekdayAbbreviation( $key ) {
 977                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
 978         }
 979
 980         /**
 981          * @param string $key
 982          * @return string
 983          */
 984         function getIranianCalendarMonthName( $key ) {
 985                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
 986         }
 987
 988         /**
 989          * @param string $key
 990          * @return string
 991          */
 992         function getHebrewCalendarMonthName( $key ) {
 993                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
 994         }
 995
 996         /**
 997          * @param string $key
 998          * @return string
 999          */
1000         function getHebrewCalendarMonthNameGen( $key ) {
1001                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1002         }
1003
1004         /**
1005          * @param string $key
1006          * @return string
1007          */
1008         function getHijriCalendarMonthName( $key ) {
1009                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1010         }
1011
1012         /**
1013          * Pass through result from $dateTimeObj->format()
1014          * @param DateTime|bool|null &$dateTimeObj
1015          * @param string $ts
1016          * @param DateTimeZone|bool|null $zone
1017          * @param string $code
1018          * @return string
1019          */
1020         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1021                 if ( !$dateTimeObj ) {
1022                         $dateTimeObj = DateTime::createFromFormat(
1023                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1024                         );
1025                 }
1026                 return $dateTimeObj->format( $code );
1027         }
1028
1029         /**
1030          * This is a workalike of PHP's date() function, but with better
1031          * internationalisation, a reduced set of format characters, and a better
1032          * escaping format.
1033          *
1034          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1035          * the PHP manual for definitions. There are a number of extensions, which
1036          * start with "x":
1037          *
1038          *    xn   Do not translate digits of the next numeric format character
1039          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1040          *    xr   Use roman numerals for the next numeric format character
1041          *    xh   Use hebrew numerals for the next numeric format character
1042          *    xx   Literal x
1043          *    xg   Genitive month name
1044          *
1045          *    xij  j (day number) in Iranian calendar
1046          *    xiF  F (month name) in Iranian calendar
1047          *    xin  n (month number) in Iranian calendar
1048          *    xiy  y (two digit year) in Iranian calendar
1049          *    xiY  Y (full year) in Iranian calendar
1050          *    xit  t (days in month) in Iranian calendar
1051          *    xiz  z (day of the year) in Iranian calendar
1052          *
1053          *    xjj  j (day number) in Hebrew calendar
1054          *    xjF  F (month name) in Hebrew calendar
1055          *    xjt  t (days in month) in Hebrew calendar
1056          *    xjx  xg (genitive month name) in Hebrew calendar
1057          *    xjn  n (month number) in Hebrew calendar
1058          *    xjY  Y (full year) in Hebrew calendar
1059          *
1060          *    xmj  j (day number) in Hijri calendar
1061          *    xmF  F (month name) in Hijri calendar
1062          *    xmn  n (month number) in Hijri calendar
1063          *    xmY  Y (full year) in Hijri calendar
1064          *
1065          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1066          *                       identical to the Gregorian calendar
1067          *    xoY  Y (full year) in Minguo calendar or Juche year.
1068          *                       Months and days are identical to the
1069          *                       Gregorian calendar
1070          *    xtY  Y (full year) in Japanese nengo. Months and days are
1071          *                       identical to the Gregorian calendar
1072          *
1073          * Characters enclosed in double quotes will be considered literal (with
1074          * the quotes themselves removed). Unmatched quotes will be considered
1075          * literal quotes. Example:
1076          *
1077          * "The month is" F       => The month is January
1078          * i's"                   => 20'11"
1079          *
1080          * Backslash escaping is also supported.
1081          *
1082          * Input timestamp is assumed to be pre-normalized to the desired local
1083          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1084          * $ts is UTC if $zone is not given.
1085          *
1086          * @param string $format
1087          * @param string $ts 14-character timestamp
1088          *      YYYYMMDDHHMMSS
1089          *      01234567890123
1090          * @param DateTimeZone $zone Timezone of $ts
1091          * @param[out] int $ttl The amount of time (in seconds) the output may be cached for.
1092          * Only makes sense if $ts is the current time.
1093          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1094          *
1095          * @throws MWException
1096          * @return string
1097          */
1098         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1099                 $s = '';
1100                 $raw = false;
1101                 $roman = false;
1102                 $hebrewNum = false;
1103                 $dateTimeObj = false;
1104                 $rawToggle = false;
1105                 $iranian = false;
1106                 $hebrew = false;
1107                 $hijri = false;
1108                 $thai = false;
1109                 $minguo = false;
1110                 $tenno = false;
1111
1112                 $usedSecond = false;
1113                 $usedMinute = false;
1114                 $usedHour = false;
1115                 $usedAMPM = false;
1116                 $usedDay = false;
1117                 $usedWeek = false;
1118                 $usedMonth = false;
1119                 $usedYear = false;
1120                 $usedISOYear = false;
1121                 $usedIsLeapYear = false;
1122
1123                 $usedHebrewMonth = false;
1124                 $usedIranianMonth = false;
1125                 $usedHijriMonth = false;
1126                 $usedHebrewYear = false;
1127                 $usedIranianYear = false;
1128                 $usedHijriYear = false;
1129                 $usedTennoYear = false;
1130
1131                 if ( strlen( $ts ) !== 14 ) {
1132                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1133                 }
1134
1135                 if ( !ctype_digit( $ts ) ) {
1136                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1137                 }
1138
1139                 $formatLength = strlen( $format );
1140                 for ( $p = 0; $p < $formatLength; $p++ ) {
1141                         $num = false;
1142                         $code = $format[$p];
1143                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1144                                 $code .= $format[++$p];
1145                         }
1146
1147                         if ( ( $code === 'xi'
1148                                         || $code === 'xj'
1149                                         || $code === 'xk'
1150                                         || $code === 'xm'
1151                                         || $code === 'xo'
1152                                         || $code === 'xt' )
1153                                 && $p < $formatLength - 1 ) {
1154                                 $code .= $format[++$p];
1155                         }
1156
1157                         switch ( $code ) {
1158                                 case 'xx':
1159                                         $s .= 'x';
1160                                         break;
1161                                 case 'xn':
1162                                         $raw = true;
1163                                         break;
1164                                 case 'xN':
1165                                         $rawToggle = !$rawToggle;
1166                                         break;
1167                                 case 'xr':
1168                                         $roman = true;
1169                                         break;
1170                                 case 'xh':
1171                                         $hebrewNum = true;
1172                                         break;
1173                                 case 'xg':
1174                                         $usedMonth = true;
1175                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1176                                         break;
1177                                 case 'xjx':
1178                                         $usedHebrewMonth = true;
1179                                         if ( !$hebrew ) {
1180                                                 $hebrew = self::tsToHebrew( $ts );
1181                                         }
1182                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1183                                         break;
1184                                 case 'd':
1185                                         $usedDay = true;
1186                                         $num = substr( $ts, 6, 2 );
1187                                         break;
1188                                 case 'D':
1189                                         $usedDay = true;
1190                                         $s .= $this->getWeekdayAbbreviation(
1191                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1192                                         );
1193                                         break;
1194                                 case 'j':
1195                                         $usedDay = true;
1196                                         $num = intval( substr( $ts, 6, 2 ) );
1197                                         break;
1198                                 case 'xij':
1199                                         $usedDay = true;
1200                                         if ( !$iranian ) {
1201                                                 $iranian = self::tsToIranian( $ts );
1202                                         }
1203                                         $num = $iranian[2];
1204                                         break;
1205                                 case 'xmj':
1206                                         $usedDay = true;
1207                                         if ( !$hijri ) {
1208                                                 $hijri = self::tsToHijri( $ts );
1209                                         }
1210                                         $num = $hijri[2];
1211                                         break;
1212                                 case 'xjj':
1213                                         $usedDay = true;
1214                                         if ( !$hebrew ) {
1215                                                 $hebrew = self::tsToHebrew( $ts );
1216                                         }
1217                                         $num = $hebrew[2];
1218                                         break;
1219                                 case 'l':
1220                                         $usedDay = true;
1221                                         $s .= $this->getWeekdayName(
1222                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1223                                         );
1224                                         break;
1225                                 case 'F':
1226                                         $usedMonth = true;
1227                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1228                                         break;
1229                                 case 'xiF':
1230                                         $usedIranianMonth = true;
1231                                         if ( !$iranian ) {
1232                                                 $iranian = self::tsToIranian( $ts );
1233                                         }
1234                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1235                                         break;
1236                                 case 'xmF':
1237                                         $usedHijriMonth = true;
1238                                         if ( !$hijri ) {
1239                                                 $hijri = self::tsToHijri( $ts );
1240                                         }
1241                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1242                                         break;
1243                                 case 'xjF':
1244                                         $usedHebrewMonth = true;
1245                                         if ( !$hebrew ) {
1246                                                 $hebrew = self::tsToHebrew( $ts );
1247                                         }
1248                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1249                                         break;
1250                                 case 'm':
1251                                         $usedMonth = true;
1252                                         $num = substr( $ts, 4, 2 );
1253                                         break;
1254                                 case 'M':
1255                                         $usedMonth = true;
1256                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1257                                         break;
1258                                 case 'n':
1259                                         $usedMonth = true;
1260                                         $num = intval( substr( $ts, 4, 2 ) );
1261                                         break;
1262                                 case 'xin':
1263                                         $usedIranianMonth = true;
1264                                         if ( !$iranian ) {
1265                                                 $iranian = self::tsToIranian( $ts );
1266                                         }
1267                                         $num = $iranian[1];
1268                                         break;
1269                                 case 'xmn':
1270                                         $usedHijriMonth = true;
1271                                         if ( !$hijri ) {
1272                                                 $hijri = self::tsToHijri( $ts );
1273                                         }
1274                                         $num = $hijri[1];
1275                                         break;
1276                                 case 'xjn':
1277                                         $usedHebrewMonth = true;
1278                                         if ( !$hebrew ) {
1279                                                 $hebrew = self::tsToHebrew( $ts );
1280                                         }
1281                                         $num = $hebrew[1];
1282                                         break;
1283                                 case 'xjt':
1284                                         $usedHebrewMonth = true;
1285                                         if ( !$hebrew ) {
1286                                                 $hebrew = self::tsToHebrew( $ts );
1287                                         }
1288                                         $num = $hebrew[3];
1289                                         break;
1290                                 case 'Y':
1291                                         $usedYear = true;
1292                                         $num = substr( $ts, 0, 4 );
1293                                         break;
1294                                 case 'xiY':
1295                                         $usedIranianYear = true;
1296                                         if ( !$iranian ) {
1297                                                 $iranian = self::tsToIranian( $ts );
1298                                         }
1299                                         $num = $iranian[0];
1300                                         break;
1301                                 case 'xmY':
1302                                         $usedHijriYear = true;
1303                                         if ( !$hijri ) {
1304                                                 $hijri = self::tsToHijri( $ts );
1305                                         }
1306                                         $num = $hijri[0];
1307                                         break;
1308                                 case 'xjY':
1309                                         $usedHebrewYear = true;
1310                                         if ( !$hebrew ) {
1311                                                 $hebrew = self::tsToHebrew( $ts );
1312                                         }
1313                                         $num = $hebrew[0];
1314                                         break;
1315                                 case 'xkY':
1316                                         $usedYear = true;
1317                                         if ( !$thai ) {
1318                                                 $thai = self::tsToYear( $ts, 'thai' );
1319                                         }
1320                                         $num = $thai[0];
1321                                         break;
1322                                 case 'xoY':
1323                                         $usedYear = true;
1324                                         if ( !$minguo ) {
1325                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1326                                         }
1327                                         $num = $minguo[0];
1328                                         break;
1329                                 case 'xtY':
1330                                         $usedTennoYear = true;
1331                                         if ( !$tenno ) {
1332                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1333                                         }
1334                                         $num = $tenno[0];
1335                                         break;
1336                                 case 'y':
1337                                         $usedYear = true;
1338                                         $num = substr( $ts, 2, 2 );
1339                                         break;
1340                                 case 'xiy':
1341                                         $usedIranianYear = true;
1342                                         if ( !$iranian ) {
1343                                                 $iranian = self::tsToIranian( $ts );
1344                                         }
1345                                         $num = substr( $iranian[0], -2 );
1346                                         break;
1347                                 case 'xit':
1348                                         $usedIranianYear = true;
1349                                         if ( !$iranian ) {
1350                                                 $iranian = self::tsToIranian( $ts );
1351                                         }
1352                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1353                                         break;
1354                                 case 'xiz':
1355                                         $usedIranianYear = true;
1356                                         if ( !$iranian ) {
1357                                                 $iranian = self::tsToIranian( $ts );
1358                                         }
1359                                         $num = $iranian[3];
1360                                         break;
1361                                 case 'a':
1362                                         $usedAMPM = true;
1363                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1364                                         break;
1365                                 case 'A':
1366                                         $usedAMPM = true;
1367                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1368                                         break;
1369                                 case 'g':
1370                                         $usedHour = true;
1371                                         $h = substr( $ts, 8, 2 );
1372                                         $num = $h % 12 ? $h % 12 : 12;
1373                                         break;
1374                                 case 'G':
1375                                         $usedHour = true;
1376                                         $num = intval( substr( $ts, 8, 2 ) );
1377                                         break;
1378                                 case 'h':
1379                                         $usedHour = true;
1380                                         $h = substr( $ts, 8, 2 );
1381                                         $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
1382                                         break;
1383                                 case 'H':
1384                                         $usedHour = true;
1385                                         $num = substr( $ts, 8, 2 );
1386                                         break;
1387                                 case 'i':
1388                                         $usedMinute = true;
1389                                         $num = substr( $ts, 10, 2 );
1390                                         break;
1391                                 case 's':
1392                                         $usedSecond = true;
1393                                         $num = substr( $ts, 12, 2 );
1394                                         break;
1395                                 case 'c':
1396                                 case 'r':
1397                                         $usedSecond = true;
1398                                         // fall through
1399                                 case 'e':
1400                                 case 'O':
1401                                 case 'P':
1402                                 case 'T':
1403                                         $s .= Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1404                                         break;
1405                                 case 'w':
1406                                 case 'N':
1407                                 case 'z':
1408                                         $usedDay = true;
1409                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1410                                         break;
1411                                 case 'W':
1412                                         $usedWeek = true;
1413                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1414                                         break;
1415                                 case 't':
1416                                         $usedMonth = true;
1417                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1418                                         break;
1419                                 case 'L':
1420                                         $usedIsLeapYear = true;
1421                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1422                                         break;
1423                                 case 'o':
1424                                         $usedISOYear = true;
1425                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1426                                         break;
1427                                 case 'U':
1428                                         $usedSecond = true;
1429                                         // fall through
1430                                 case 'I':
1431                                 case 'Z':
1432                                         $num = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1433                                         break;
1434                                 case '\\':
1435                                         # Backslash escaping
1436                                         if ( $p < $formatLength - 1 ) {
1437                                                 $s .= $format[++$p];
1438                                         } else {
1439                                                 $s .= '\\';
1440                                         }
1441                                         break;
1442                                 case '"':
1443                                         # Quoted literal
1444                                         if ( $p < $formatLength - 1 ) {
1445                                                 $endQuote = strpos( $format, '"', $p + 1 );
1446                                                 if ( $endQuote === false ) {
1447                                                         # No terminating quote, assume literal "
1448                                                         $s .= '"';
1449                                                 } else {
1450                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1451                                                         $p = $endQuote;
1452                                                 }
1453                                         } else {
1454                                                 # Quote at end of string, assume literal "
1455                                                 $s .= '"';
1456                                         }
1457                                         break;
1458                                 default:
1459                                         $s .= $format[$p];
1460                         }
1461                         if ( $num !== false ) {
1462                                 if ( $rawToggle || $raw ) {
1463                                         $s .= $num;
1464                                         $raw = false;
1465                                 } elseif ( $roman ) {
1466                                         $s .= Language::romanNumeral( $num );
1467                                         $roman = false;
1468                                 } elseif ( $hebrewNum ) {
1469                                         $s .= self::hebrewNumeral( $num );
1470                                         $hebrewNum = false;
1471                                 } else {
1472                                         $s .= $this->formatNum( $num, true );
1473                                 }
1474                         }
1475                 }
1476
1477                 if ( $ttl === 'unused' ) {
1478                         // No need to calculate the TTL, the caller wont use it anyway.
1479                 } elseif ( $usedSecond ) {
1480                         $ttl = 1;
1481                 } elseif ( $usedMinute ) {
1482                         $ttl = 60 - substr( $ts, 12, 2 );
1483                 } elseif ( $usedHour ) {
1484                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1485                 } elseif ( $usedAMPM ) {
1486                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1487                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1488                 } elseif (
1489                         $usedDay ||
1490                         $usedHebrewMonth ||
1491                         $usedIranianMonth ||
1492                         $usedHijriMonth ||
1493                         $usedHebrewYear ||
1494                         $usedIranianYear ||
1495                         $usedHijriYear ||
1496                         $usedTennoYear
1497                 ) {
1498                         // @todo Someone who understands the non-Gregorian calendars
1499                         // should write proper logic for them so that they don't need purged every day.
1500                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1501                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1502                 } else {
1503                         $possibleTtls = [];
1504                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1505                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1506                         if ( $usedWeek ) {
1507                                 $possibleTtls[] =
1508                                         ( 7 - Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1509                                         $timeRemainingInDay;
1510                         } elseif ( $usedISOYear ) {
1511                                 // December 28th falls on the last ISO week of the year, every year.
1512                                 // The last ISO week of a year can be 52 or 53.
1513                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1514                                         'Ymd',
1515                                         substr( $ts, 0, 4 ) . '1228',
1516                                         $zone ?: new DateTimeZone( 'UTC' )
1517                                 )->format( 'W' );
1518                                 $currentISOWeek = Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1519                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1520                                 $timeRemainingInWeek =
1521                                         ( 7 - Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1522                                         + $timeRemainingInDay;
1523                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1524                         }
1525
1526                         if ( $usedMonth ) {
1527                                 $possibleTtls[] =
1528                                         ( Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1529                                                 substr( $ts, 6, 2 ) ) * 86400
1530                                         + $timeRemainingInDay;
1531                         } elseif ( $usedYear ) {
1532                                 $possibleTtls[] =
1533                                         ( Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1534                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1535                                         + $timeRemainingInDay;
1536                         } elseif ( $usedIsLeapYear ) {
1537                                 $year = substr( $ts, 0, 4 );
1538                                 $timeRemainingInYear =
1539                                         ( Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1540                                                 Language::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1541                                         + $timeRemainingInDay;
1542                                 $mod = $year % 4;
1543                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1544                                         // this isn't a leap year. see when the next one starts
1545                                         $nextCandidate = $year - $mod + 4;
1546                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1547                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1548                                                         $timeRemainingInYear;
1549                                         } else {
1550                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1551                                                         $timeRemainingInYear;
1552                                         }
1553                                 } else {
1554                                         // this is a leap year, so the next year isn't
1555                                         $possibleTtls[] = $timeRemainingInYear;
1556                                 }
1557                         }
1558
1559                         if ( $possibleTtls ) {
1560                                 $ttl = min( $possibleTtls );
1561                         }
1562                 }
1563
1564                 return $s;
1565         }
1566
1567         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1568         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1569
1570         /**
1571          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1572          * Gregorian dates to Iranian dates. Originally written in C, it
1573          * is released under the terms of GNU Lesser General Public
1574          * License. Conversion to PHP was performed by Niklas Laxström.
1575          *
1576          * Link: http://www.farsiweb.info/jalali/jalali.c
1577          *
1578          * @param string $ts
1579          *
1580          * @return int[]
1581          */
1582         private static function tsToIranian( $ts ) {
1583                 $gy = substr( $ts, 0, 4 ) -1600;
1584                 $gm = substr( $ts, 4, 2 ) -1;
1585                 $gd = substr( $ts, 6, 2 ) -1;
1586
1587                 # Days passed from the beginning (including leap years)
1588                 $gDayNo = 365 * $gy
1589                         + floor( ( $gy + 3 ) / 4 )
1590                         - floor( ( $gy + 99 ) / 100 )
1591                         + floor( ( $gy + 399 ) / 400 );
1592
1593                 // Add days of the past months of this year
1594                 for ( $i = 0; $i < $gm; $i++ ) {
1595                         $gDayNo += self::$GREG_DAYS[$i];
1596                 }
1597
1598                 // Leap years
1599                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1600                         $gDayNo++;
1601                 }
1602
1603                 // Days passed in current month
1604                 $gDayNo += (int)$gd;
1605
1606                 $jDayNo = $gDayNo - 79;
1607
1608                 $jNp = floor( $jDayNo / 12053 );
1609                 $jDayNo %= 12053;
1610
1611                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1612                 $jDayNo %= 1461;
1613
1614                 if ( $jDayNo >= 366 ) {
1615                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1616                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1617                 }
1618
1619                 $jz = $jDayNo;
1620
1621                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1622                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1623                 }
1624
1625                 $jm = $i + 1;
1626                 $jd = $jDayNo + 1;
1627
1628                 return [ $jy, $jm, $jd, $jz ];
1629         }
1630
1631         /**
1632          * Converting Gregorian dates to Hijri dates.
1633          *
1634          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1635          *
1636          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1637          *
1638          * @param string $ts
1639          *
1640          * @return int[]
1641          */
1642         private static function tsToHijri( $ts ) {
1643                 $year = substr( $ts, 0, 4 );
1644                 $month = substr( $ts, 4, 2 );
1645                 $day = substr( $ts, 6, 2 );
1646
1647                 $zyr = $year;
1648                 $zd = $day;
1649                 $zm = $month;
1650                 $zy = $zyr;
1651
1652                 if (
1653                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1654                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1655                 ) {
1656                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1657                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1658                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1659                                         $zd - 32075;
1660                 } else {
1661                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1662                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1663                 }
1664
1665                 $zl = $zjd -1948440 + 10632;
1666                 $zn = (int)( ( $zl - 1 ) / 10631 );
1667                 $zl = $zl - 10631 * $zn + 354;
1668                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1669                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1670                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1671                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1672                 $zm = (int)( ( 24 * $zl ) / 709 );
1673                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1674                 $zy = 30 * $zn + $zj - 30;
1675
1676                 return [ $zy, $zm, $zd ];
1677         }
1678
1679         /**
1680          * Converting Gregorian dates to Hebrew dates.
1681          *
1682          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1683          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1684          * to translate the relevant functions into PHP and release them under
1685          * GNU GPL.
1686          *
1687          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1688          * and Adar II is 14. In a non-leap year, Adar is 6.
1689          *
1690          * @param string $ts
1691          *
1692          * @return int[]
1693          */
1694         private static function tsToHebrew( $ts ) {
1695                 # Parse date
1696                 $year = substr( $ts, 0, 4 );
1697                 $month = substr( $ts, 4, 2 );
1698                 $day = substr( $ts, 6, 2 );
1699
1700                 # Calculate Hebrew year
1701                 $hebrewYear = $year + 3760;
1702
1703                 # Month number when September = 1, August = 12
1704                 $month += 4;
1705                 if ( $month > 12 ) {
1706                         # Next year
1707                         $month -= 12;
1708                         $year++;
1709                         $hebrewYear++;
1710                 }
1711
1712                 # Calculate day of year from 1 September
1713                 $dayOfYear = $day;
1714                 for ( $i = 1; $i < $month; $i++ ) {
1715                         if ( $i == 6 ) {
1716                                 # February
1717                                 $dayOfYear += 28;
1718                                 # Check if the year is leap
1719                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1720                                         $dayOfYear++;
1721                                 }
1722                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1723                                 $dayOfYear += 30;
1724                         } else {
1725                                 $dayOfYear += 31;
1726                         }
1727                 }
1728
1729                 # Calculate the start of the Hebrew year
1730                 $start = self::hebrewYearStart( $hebrewYear );
1731
1732                 # Calculate next year's start
1733                 if ( $dayOfYear <= $start ) {
1734                         # Day is before the start of the year - it is the previous year
1735                         # Next year's start
1736                         $nextStart = $start;
1737                         # Previous year
1738                         $year--;
1739                         $hebrewYear--;
1740                         # Add days since previous year's 1 September
1741                         $dayOfYear += 365;
1742                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1743                                 # Leap year
1744                                 $dayOfYear++;
1745                         }
1746                         # Start of the new (previous) year
1747                         $start = self::hebrewYearStart( $hebrewYear );
1748                 } else {
1749                         # Next year's start
1750                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1751                 }
1752
1753                 # Calculate Hebrew day of year
1754                 $hebrewDayOfYear = $dayOfYear - $start;
1755
1756                 # Difference between year's days
1757                 $diff = $nextStart - $start;
1758                 # Add 12 (or 13 for leap years) days to ignore the difference between
1759                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1760                 # difference is only about the year type
1761                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1762                         $diff += 13;
1763                 } else {
1764                         $diff += 12;
1765                 }
1766
1767                 # Check the year pattern, and is leap year
1768                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1769                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1770                 # and non-leap years
1771                 $yearPattern = $diff % 30;
1772                 # Check if leap year
1773                 $isLeap = $diff >= 30;
1774
1775                 # Calculate day in the month from number of day in the Hebrew year
1776                 # Don't check Adar - if the day is not in Adar, we will stop before;
1777                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1778                 $hebrewDay = $hebrewDayOfYear;
1779                 $hebrewMonth = 1;
1780                 $days = 0;
1781                 while ( $hebrewMonth <= 12 ) {
1782                         # Calculate days in this month
1783                         if ( $isLeap && $hebrewMonth == 6 ) {
1784                                 # Adar in a leap year
1785                                 if ( $isLeap ) {
1786                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1787                                         $days = 30;
1788                                         if ( $hebrewDay <= $days ) {
1789                                                 # Day in Adar I
1790                                                 $hebrewMonth = 13;
1791                                         } else {
1792                                                 # Subtract the days of Adar I
1793                                                 $hebrewDay -= $days;
1794                                                 # Try Adar II
1795                                                 $days = 29;
1796                                                 if ( $hebrewDay <= $days ) {
1797                                                         # Day in Adar II
1798                                                         $hebrewMonth = 14;
1799                                                 }
1800                                         }
1801                                 }
1802                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1803                                 # Cheshvan in a complete year (otherwise as the rule below)
1804                                 $days = 30;
1805                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1806                                 # Kislev in an incomplete year (otherwise as the rule below)
1807                                 $days = 29;
1808                         } else {
1809                                 # Odd months have 30 days, even have 29
1810                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1811                         }
1812                         if ( $hebrewDay <= $days ) {
1813                                 # In the current month
1814                                 break;
1815                         } else {
1816                                 # Subtract the days of the current month
1817                                 $hebrewDay -= $days;
1818                                 # Try in the next month
1819                                 $hebrewMonth++;
1820                         }
1821                 }
1822
1823                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1824         }
1825
1826         /**
1827          * This calculates the Hebrew year start, as days since 1 September.
1828          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1829          * Used for Hebrew date.
1830          *
1831          * @param int $year
1832          *
1833          * @return string
1834          */
1835         private static function hebrewYearStart( $year ) {
1836                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1837                 $b = intval( ( $year - 1 ) % 4 );
1838                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1839                 if ( $m < 0 ) {
1840                         $m--;
1841                 }
1842                 $Mar = intval( $m );
1843                 if ( $m < 0 ) {
1844                         $m++;
1845                 }
1846                 $m -= $Mar;
1847
1848                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1849                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1850                         $Mar++;
1851                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1852                         $Mar += 2;
1853                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1854                         $Mar++;
1855                 }
1856
1857                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1858                 return $Mar;
1859         }
1860
1861         /**
1862          * Algorithm to convert Gregorian dates to Thai solar dates,
1863          * Minguo dates or Minguo dates.
1864          *
1865          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1866          *       https://en.wikipedia.org/wiki/Minguo_calendar
1867          *       https://en.wikipedia.org/wiki/Japanese_era_name
1868          *
1869          * @param string $ts 14-character timestamp
1870          * @param string $cName Calender name
1871          * @return array Converted year, month, day
1872          */
1873         private static function tsToYear( $ts, $cName ) {
1874                 $gy = substr( $ts, 0, 4 );
1875                 $gm = substr( $ts, 4, 2 );
1876                 $gd = substr( $ts, 6, 2 );
1877
1878                 if ( !strcmp( $cName, 'thai' ) ) {
1879                         # Thai solar dates
1880                         # Add 543 years to the Gregorian calendar
1881                         # Months and days are identical
1882                         $gy_offset = $gy + 543;
1883                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1884                         # Minguo dates
1885                         # Deduct 1911 years from the Gregorian calendar
1886                         # Months and days are identical
1887                         $gy_offset = $gy - 1911;
1888                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1889                         # Nengō dates up to Meiji period
1890                         # Deduct years from the Gregorian calendar
1891                         # depending on the nengo periods
1892                         # Months and days are identical
1893                         if ( ( $gy < 1912 )
1894                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1895                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1896                         ) {
1897                                 # Meiji period
1898                                 $gy_gannen = $gy - 1868 + 1;
1899                                 $gy_offset = $gy_gannen;
1900                                 if ( $gy_gannen == 1 ) {
1901                                         $gy_offset = '元';
1902                                 }
1903                                 $gy_offset = '明治' . $gy_offset;
1904                         } elseif (
1905                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1906                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1907                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1908                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1909                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1910                         ) {
1911                                 # Taishō period
1912                                 $gy_gannen = $gy - 1912 + 1;
1913                                 $gy_offset = $gy_gannen;
1914                                 if ( $gy_gannen == 1 ) {
1915                                         $gy_offset = '元';
1916                                 }
1917                                 $gy_offset = '大正' . $gy_offset;
1918                         } elseif (
1919                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1920                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1921                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1922                         ) {
1923                                 # Shōwa period
1924                                 $gy_gannen = $gy - 1926 + 1;
1925                                 $gy_offset = $gy_gannen;
1926                                 if ( $gy_gannen == 1 ) {
1927                                         $gy_offset = '元';
1928                                 }
1929                                 $gy_offset = '昭和' . $gy_offset;
1930                         } else {
1931                                 # Heisei period
1932                                 $gy_gannen = $gy - 1989 + 1;
1933                                 $gy_offset = $gy_gannen;
1934                                 if ( $gy_gannen == 1 ) {
1935                                         $gy_offset = '元';
1936                                 }
1937                                 $gy_offset = '平成' . $gy_offset;
1938                         }
1939                 } else {
1940                         $gy_offset = $gy;
1941                 }
1942
1943                 return [ $gy_offset, $gm, $gd ];
1944         }
1945
1946         /**
1947          * Gets directionality of the first strongly directional codepoint, for embedBidi()
1948          *
1949          * This is the rule the BIDI algorithm uses to determine the directionality of
1950          * paragraphs ( http://unicode.org/reports/tr9/#The_Paragraph_Level ) and
1951          * FSI isolates ( http://unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
1952          *
1953          * TODO: Does not handle BIDI control characters inside the text.
1954          * TODO: Does not handle unallocated characters.
1955          *
1956          * @param string $text Text to test
1957          * @return null|string Directionality ('ltr' or 'rtl') or null
1958          */
1959         private static function strongDirFromContent( $text = '' ) {
1960                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
1961                         return null;
1962                 }
1963                 if ( $matches[1] === '' ) {
1964                         return 'rtl';
1965                 }
1966                 return 'ltr';
1967         }
1968
1969         /**
1970          * Roman number formatting up to 10000
1971          *
1972          * @param int $num
1973          *
1974          * @return string
1975          */
1976         static function romanNumeral( $num ) {
1977                 static $table = [
1978                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
1979                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
1980                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
1981                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
1982                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
1983                 ];
1984
1985                 $num = intval( $num );
1986                 if ( $num > 10000 || $num <= 0 ) {
1987                         return $num;
1988                 }
1989
1990                 $s = '';
1991                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1992                         if ( $num >= $pow10 ) {
1993                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
1994                         }
1995                         $num = $num % $pow10;
1996                 }
1997                 return $s;
1998         }
1999
2000         /**
2001          * Hebrew Gematria number formatting up to 9999
2002          *
2003          * @param int $num
2004          *
2005          * @return string
2006          */
2007         static function hebrewNumeral( $num ) {
2008                 static $table = [
2009                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2010                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2011                         [ '',
2012                                 [ 'ק' ],
2013                                 [ 'ר' ],
2014                                 [ 'ש' ],
2015                                 [ 'ת' ],
2016                                 [ 'ת', 'ק' ],
2017                                 [ 'ת', 'ר' ],
2018                                 [ 'ת', 'ש' ],
2019                                 [ 'ת', 'ת' ],
2020                                 [ 'ת', 'ת', 'ק' ],
2021                                 [ 'ת', 'ת', 'ר' ],
2022                         ],
2023                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2024                 ];
2025
2026                 $num = intval( $num );
2027                 if ( $num > 9999 || $num <= 0 ) {
2028                         return $num;
2029                 }
2030
2031                 // Round thousands have special notations
2032                 if ( $num === 1000 ) {
2033                         return "א' אלף";
2034                 } elseif ( $num % 1000 === 0 ) {
2035                         return $table[0][$num / 1000] . "' אלפים";
2036                 }
2037
2038                 $letters = [];
2039
2040                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2041                         if ( $num >= $pow10 ) {
2042                                 if ( $num === 15 || $num === 16 ) {
2043                                         $letters[] = $table[0][9];
2044                                         $letters[] = $table[0][$num - 9];
2045                                         $num = 0;
2046                                 } else {
2047                                         $letters = array_merge(
2048                                                 $letters,
2049                                                 (array)$table[$i][intval( $num / $pow10 )]
2050                                         );
2051
2052                                         if ( $pow10 === 1000 ) {
2053                                                 $letters[] = "'";
2054                                         }
2055                                 }
2056                         }
2057
2058                         $num = $num % $pow10;
2059                 }
2060
2061                 $preTransformLength = count( $letters );
2062                 if ( $preTransformLength === 1 ) {
2063                         // Add geresh (single quote) to one-letter numbers
2064                         $letters[] = "'";
2065                 } else {
2066                         $lastIndex = $preTransformLength - 1;
2067                         $letters[$lastIndex] = str_replace(
2068                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2069                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2070                                 $letters[$lastIndex]
2071                         );
2072
2073                         // Add gershayim (double quote) to multiple-letter numbers,
2074                         // but exclude numbers with only one letter after the thousands
2075                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2076                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2077                                 $letters[] = "'";
2078                         } else {
2079                                 array_splice( $letters, -1, 0, '"' );
2080                         }
2081                 }
2082
2083                 return implode( $letters );
2084         }
2085
2086         /**
2087          * Used by date() and time() to adjust the time output.
2088          *
2089          * @param string $ts The time in date('YmdHis') format
2090          * @param mixed $tz Adjust the time by this amount (default false, mean we
2091          *   get user timecorrection setting)
2092          * @return int
2093          */
2094         public function userAdjust( $ts, $tz = false ) {
2095                 global $wgUser, $wgLocalTZoffset;
2096
2097                 if ( $tz === false ) {
2098                         $tz = $wgUser->getOption( 'timecorrection' );
2099                 }
2100
2101                 $data = explode( '|', $tz, 3 );
2102
2103                 if ( $data[0] == 'ZoneInfo' ) {
2104                         try {
2105                                 $userTZ = new DateTimeZone( $data[2] );
2106                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2107                                 $date->setTimezone( $userTZ );
2108                                 return $date->format( 'YmdHis' );
2109                         } catch ( Exception $e ) {
2110                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2111                                 $data[0] = 'Offset';
2112                         }
2113                 }
2114
2115                 if ( $data[0] == 'System' || $tz == '' ) {
2116                         # Global offset in minutes.
2117                         $minDiff = $wgLocalTZoffset;
2118                 } elseif ( $data[0] == 'Offset' ) {
2119                         $minDiff = intval( $data[1] );
2120                 } else {
2121                         $data = explode( ':', $tz );
2122                         if ( count( $data ) == 2 ) {
2123                                 $data[0] = intval( $data[0] );
2124                                 $data[1] = intval( $data[1] );
2125                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2126                                 if ( $data[0] < 0 ) {
2127                                         $minDiff = -$minDiff;
2128                                 }
2129                         } else {
2130                                 $minDiff = intval( $data[0] ) * 60;
2131                         }
2132                 }
2133
2134                 # No difference ? Return time unchanged
2135                 if ( 0 == $minDiff ) {
2136                         return $ts;
2137                 }
2138
2139                 MediaWiki\suppressWarnings(); // E_STRICT system time bitching
2140                 # Generate an adjusted date; take advantage of the fact that mktime
2141                 # will normalize out-of-range values so we don't have to split $minDiff
2142                 # into hours and minutes.
2143                 $t = mktime( (
2144                         (int)substr( $ts, 8, 2 ) ), # Hours
2145                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2146                         (int)substr( $ts, 12, 2 ), # Seconds
2147                         (int)substr( $ts, 4, 2 ), # Month
2148                         (int)substr( $ts, 6, 2 ), # Day
2149                         (int)substr( $ts, 0, 4 ) ); # Year
2150
2151                 $date = date( 'YmdHis', $t );
2152                 MediaWiki\restoreWarnings();
2153
2154                 return $date;
2155         }
2156
2157         /**
2158          * This is meant to be used by time(), date(), and timeanddate() to get
2159          * the date preference they're supposed to use, it should be used in
2160          * all children.
2161          *
2162          *     function timeanddate([...], $format = true) {
2163          *       $datePreference = $this->dateFormat($format);
2164          *       [...]
2165          *     }
2166          *
2167          * @param int|string|bool $usePrefs If true, the user's preference is used
2168          *   if false, the site/language default is used
2169          *   if int/string, assumed to be a format.
2170          * @return string
2171          */
2172         function dateFormat( $usePrefs = true ) {
2173                 global $wgUser;
2174
2175                 if ( is_bool( $usePrefs ) ) {
2176                         if ( $usePrefs ) {
2177                                 $datePreference = $wgUser->getDatePreference();
2178                         } else {
2179                                 $datePreference = (string)User::getDefaultOption( 'date' );
2180                         }
2181                 } else {
2182                         $datePreference = (string)$usePrefs;
2183                 }
2184
2185                 // return int
2186                 if ( $datePreference == '' ) {
2187                         return 'default';
2188                 }
2189
2190                 return $datePreference;
2191         }
2192
2193         /**
2194          * Get a format string for a given type and preference
2195          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2196          * @param string $pref The format name as it appears in Messages*.php under
2197          *  $datePreferences.
2198          *
2199          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2200          *
2201          * @return string
2202          */
2203         function getDateFormatString( $type, $pref ) {
2204                 $wasDefault = false;
2205                 if ( $pref == 'default' ) {
2206                         $wasDefault = true;
2207                         $pref = $this->getDefaultDateFormat();
2208                 }
2209
2210                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2211                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2212
2213                         if ( $type === 'pretty' && $df === null ) {
2214                                 $df = $this->getDateFormatString( 'date', $pref );
2215                         }
2216
2217                         if ( !$wasDefault && $df === null ) {
2218                                 $pref = $this->getDefaultDateFormat();
2219                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2220                         }
2221
2222                         $this->dateFormatStrings[$type][$pref] = $df;
2223                 }
2224                 return $this->dateFormatStrings[$type][$pref];
2225         }
2226
2227         /**
2228          * @param string $ts The time format which needs to be turned into a
2229          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2230          * @param bool $adj Whether to adjust the time output according to the
2231          *   user configured offset ($timecorrection)
2232          * @param mixed $format True to use user's date format preference
2233          * @param string|bool $timecorrection The time offset as returned by
2234          *   validateTimeZone() in Special:Preferences
2235          * @return string
2236          */
2237         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2238                 $ts = wfTimestamp( TS_MW, $ts );
2239                 if ( $adj ) {
2240                         $ts = $this->userAdjust( $ts, $timecorrection );
2241                 }
2242                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2243                 return $this->sprintfDate( $df, $ts );
2244         }
2245
2246         /**
2247          * @param string $ts The time format which needs to be turned into a
2248          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2249          * @param bool $adj Whether to adjust the time output according to the
2250          *   user configured offset ($timecorrection)
2251          * @param mixed $format True to use user's date format preference
2252          * @param string|bool $timecorrection The time offset as returned by
2253          *   validateTimeZone() in Special:Preferences
2254          * @return string
2255          */
2256         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2257                 $ts = wfTimestamp( TS_MW, $ts );
2258                 if ( $adj ) {
2259                         $ts = $this->userAdjust( $ts, $timecorrection );
2260                 }
2261                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2262                 return $this->sprintfDate( $df, $ts );
2263         }
2264
2265         /**
2266          * @param string $ts The time format which needs to be turned into a
2267          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2268          * @param bool $adj Whether to adjust the time output according to the
2269          *   user configured offset ($timecorrection)
2270          * @param mixed $format What format to return, if it's false output the
2271          *   default one (default true)
2272          * @param string|bool $timecorrection The time offset as returned by
2273          *   validateTimeZone() in Special:Preferences
2274          * @return string
2275          */
2276         public function timeanddate( $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( 'both', $this->dateFormat( $format ) );
2282                 return $this->sprintfDate( $df, $ts );
2283         }
2284
2285         /**
2286          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2287          *
2288          * @since 1.20
2289          *
2290          * @param int $seconds The amount of seconds.
2291          * @param array $chosenIntervals The intervals to enable.
2292          *
2293          * @return string
2294          */
2295         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2296                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2297
2298                 $segments = [];
2299
2300                 foreach ( $intervals as $intervalName => $intervalValue ) {
2301                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2302                         // duration-years, duration-decades, duration-centuries, duration-millennia
2303                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2304                         $segments[] = $message->inLanguage( $this )->escaped();
2305                 }
2306
2307                 return $this->listToText( $segments );
2308         }
2309
2310         /**
2311          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2312          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2313          *
2314          * @since 1.20
2315          *
2316          * @param int $seconds The amount of seconds.
2317          * @param array $chosenIntervals The intervals to enable.
2318          *
2319          * @return array
2320          */
2321         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2322                 if ( empty( $chosenIntervals ) ) {
2323                         $chosenIntervals = [
2324                                 'millennia',
2325                                 'centuries',
2326                                 'decades',
2327                                 'years',
2328                                 'days',
2329                                 'hours',
2330                                 'minutes',
2331                                 'seconds'
2332                         ];
2333                 }
2334
2335                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2336                 $sortedNames = array_keys( $intervals );
2337                 $smallestInterval = array_pop( $sortedNames );
2338
2339                 $segments = [];
2340
2341                 foreach ( $intervals as $name => $length ) {
2342                         $value = floor( $seconds / $length );
2343
2344                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2345                                 $seconds -= $value * $length;
2346                                 $segments[$name] = $value;
2347                         }
2348                 }
2349
2350                 return $segments;
2351         }
2352
2353         /**
2354          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2355          *
2356          * @param string $type Can be 'date', 'time' or 'both'
2357          * @param string $ts The time format which needs to be turned into a
2358          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2359          * @param User $user User object used to get preferences for timezone and format
2360          * @param array $options Array, can contain the following keys:
2361          *   - 'timecorrection': time correction, can have the following values:
2362          *     - true: use user's preference
2363          *     - false: don't use time correction
2364          *     - int: value of time correction in minutes
2365          *   - 'format': format to use, can have the following values:
2366          *     - true: use user's preference
2367          *     - false: use default preference
2368          *     - string: format to use
2369          * @since 1.19
2370          * @return string
2371          */
2372         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2373                 $ts = wfTimestamp( TS_MW, $ts );
2374                 $options += [ 'timecorrection' => true, 'format' => true ];
2375                 if ( $options['timecorrection'] !== false ) {
2376                         if ( $options['timecorrection'] === true ) {
2377                                 $offset = $user->getOption( 'timecorrection' );
2378                         } else {
2379                                 $offset = $options['timecorrection'];
2380                         }
2381                         $ts = $this->userAdjust( $ts, $offset );
2382                 }
2383                 if ( $options['format'] === true ) {
2384                         $format = $user->getDatePreference();
2385                 } else {
2386                         $format = $options['format'];
2387                 }
2388                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2389                 return $this->sprintfDate( $df, $ts );
2390         }
2391
2392         /**
2393          * Get the formatted date for the given timestamp and formatted for
2394          * the given user.
2395          *
2396          * @param mixed $ts Mixed: the time format which needs to be turned into a
2397          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2398          * @param User $user User object used to get preferences for timezone and format
2399          * @param array $options Array, can contain the following keys:
2400          *   - 'timecorrection': time correction, can have the following values:
2401          *     - true: use user's preference
2402          *     - false: don't use time correction
2403          *     - int: value of time correction in minutes
2404          *   - 'format': format to use, can have the following values:
2405          *     - true: use user's preference
2406          *     - false: use default preference
2407          *     - string: format to use
2408          * @since 1.19
2409          * @return string
2410          */
2411         public function userDate( $ts, User $user, array $options = [] ) {
2412                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2413         }
2414
2415         /**
2416          * Get the formatted time for the given timestamp and formatted for
2417          * the given user.
2418          *
2419          * @param mixed $ts The time format which needs to be turned into a
2420          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2421          * @param User $user User object used to get preferences for timezone and format
2422          * @param array $options Array, can contain the following keys:
2423          *   - 'timecorrection': time correction, can have the following values:
2424          *     - true: use user's preference
2425          *     - false: don't use time correction
2426          *     - int: value of time correction in minutes
2427          *   - 'format': format to use, can have the following values:
2428          *     - true: use user's preference
2429          *     - false: use default preference
2430          *     - string: format to use
2431          * @since 1.19
2432          * @return string
2433          */
2434         public function userTime( $ts, User $user, array $options = [] ) {
2435                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2436         }
2437
2438         /**
2439          * Get the formatted date and time for the given timestamp and formatted for
2440          * the given user.
2441          *
2442          * @param mixed $ts The time format which needs to be turned into a
2443          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2444          * @param User $user User object used to get preferences for timezone and format
2445          * @param array $options Array, can contain the following keys:
2446          *   - 'timecorrection': time correction, can have the following values:
2447          *     - true: use user's preference
2448          *     - false: don't use time correction
2449          *     - int: value of time correction in minutes
2450          *   - 'format': format to use, can have the following values:
2451          *     - true: use user's preference
2452          *     - false: use default preference
2453          *     - string: format to use
2454          * @since 1.19
2455          * @return string
2456          */
2457         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2458                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2459         }
2460
2461         /**
2462          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2463          *
2464          * Determine the difference between the timestamp and the current time, and
2465          * generate a readable timestamp by returning "<N> <units> ago", where the
2466          * largest possible unit is used.
2467          *
2468          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2469          *
2470          * @param MWTimestamp $time
2471          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2472          * @param User|null $user User the timestamp is being generated for
2473          *  (or null to use main context's user)
2474          * @return string Formatted timestamp
2475          */
2476         public function getHumanTimestamp(
2477                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2478         ) {
2479                 if ( $relativeTo === null ) {
2480                         $relativeTo = new MWTimestamp();
2481                 }
2482                 if ( $user === null ) {
2483                         $user = RequestContext::getMain()->getUser();
2484                 }
2485
2486                 // Adjust for the user's timezone.
2487                 $offsetThis = $time->offsetForUser( $user );
2488                 $offsetRel = $relativeTo->offsetForUser( $user );
2489
2490                 $ts = '';
2491                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2492                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2493                 }
2494
2495                 // Reset the timezone on the objects.
2496                 $time->timestamp->sub( $offsetThis );
2497                 $relativeTo->timestamp->sub( $offsetRel );
2498
2499                 return $ts;
2500         }
2501
2502         /**
2503          * Convert an MWTimestamp into a pretty human-readable timestamp using
2504          * the given user preferences and relative base time.
2505          *
2506          * @see Language::getHumanTimestamp
2507          * @param MWTimestamp $ts Timestamp to prettify
2508          * @param MWTimestamp $relativeTo Base timestamp
2509          * @param User $user User preferences to use
2510          * @return string Human timestamp
2511          * @since 1.26
2512          */
2513         private function getHumanTimestampInternal(
2514                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2515         ) {
2516                 $diff = $ts->diff( $relativeTo );
2517                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2518                         (int)$relativeTo->timestamp->format( 'w' ) );
2519                 $days = $diff->days ?: (int)$diffDay;
2520                 if ( $diff->invert || $days > 5
2521                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2522                 ) {
2523                         // Timestamps are in different years: use full timestamp
2524                         // Also do full timestamp for future dates
2525                         /**
2526                          * @todo FIXME: Add better handling of future timestamps.
2527                          */
2528                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2529                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2530                 } elseif ( $days > 5 ) {
2531                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2532                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2533                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2534                 } elseif ( $days > 1 ) {
2535                         // Timestamp within the past week: show the day of the week and time
2536                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2537                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2538                         // Messages:
2539                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2540                         $ts = wfMessage( "$weekday-at" )
2541                                 ->inLanguage( $this )
2542                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2543                                 ->text();
2544                 } elseif ( $days == 1 ) {
2545                         // Timestamp was yesterday: say 'yesterday' and the time.
2546                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2547                         $ts = wfMessage( 'yesterday-at' )
2548                                 ->inLanguage( $this )
2549                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2550                                 ->text();
2551                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2552                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2553                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2554                         $ts = wfMessage( 'today-at' )
2555                                 ->inLanguage( $this )
2556                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2557                                 ->text();
2558
2559                 // From here on in, the timestamp was soon enough ago so that we can simply say
2560                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2561                 } elseif ( $diff->h == 1 ) {
2562                         // Less than 90 minutes, but more than an hour ago.
2563                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2564                 } elseif ( $diff->i >= 1 ) {
2565                         // A few minutes ago.
2566                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2567                 } elseif ( $diff->s >= 30 ) {
2568                         // Less than a minute, but more than 30 sec ago.
2569                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2570                 } else {
2571                         // Less than 30 seconds ago.
2572                         $ts = wfMessage( 'just-now' )->text();
2573                 }
2574
2575                 return $ts;
2576         }
2577
2578         /**
2579          * @param string $key
2580          * @return string|null
2581          */
2582         public function getMessage( $key ) {
2583                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2584         }
2585
2586         /**
2587          * @return array
2588          */
2589         function getAllMessages() {
2590                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2591         }
2592
2593         /**
2594          * @param string $in
2595          * @param string $out
2596          * @param string $string
2597          * @return string
2598          */
2599         public function iconv( $in, $out, $string ) {
2600                 # Even with //IGNORE iconv can whine about illegal characters in
2601                 # *input* string. We just ignore those too.
2602                 # REF: https://bugs.php.net/bug.php?id=37166
2603                 # REF: https://phabricator.wikimedia.org/T18885
2604                 MediaWiki\suppressWarnings();
2605                 $text = iconv( $in, $out . '//IGNORE', $string );
2606                 MediaWiki\restoreWarnings();
2607                 return $text;
2608         }
2609
2610         // callback functions for ucwords(), ucwordbreaks()
2611
2612         /**
2613          * @param array $matches
2614          * @return mixed|string
2615          */
2616         function ucwordbreaksCallbackAscii( $matches ) {
2617                 return $this->ucfirst( $matches[1] );
2618         }
2619
2620         /**
2621          * @param array $matches
2622          * @return string
2623          */
2624         function ucwordbreaksCallbackMB( $matches ) {
2625                 return mb_strtoupper( $matches[0] );
2626         }
2627
2628         /**
2629          * @param array $matches
2630          * @return string
2631          */
2632         function ucwordsCallbackMB( $matches ) {
2633                 return mb_strtoupper( $matches[0] );
2634         }
2635
2636         /**
2637          * Make a string's first character uppercase
2638          *
2639          * @param string $str
2640          *
2641          * @return string
2642          */
2643         public function ucfirst( $str ) {
2644                 $o = ord( $str );
2645                 if ( $o < 96 ) { // if already uppercase...
2646                         return $str;
2647                 } elseif ( $o < 128 ) {
2648                         return ucfirst( $str ); // use PHP's ucfirst()
2649                 } else {
2650                         // fall back to more complex logic in case of multibyte strings
2651                         return $this->uc( $str, true );
2652                 }
2653         }
2654
2655         /**
2656          * Convert a string to uppercase
2657          *
2658          * @param string $str
2659          * @param bool $first
2660          *
2661          * @return string
2662          */
2663         public function uc( $str, $first = false ) {
2664                 if ( $first ) {
2665                         if ( $this->isMultibyte( $str ) ) {
2666                                 return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2667                         } else {
2668                                 return ucfirst( $str );
2669                         }
2670                 } else {
2671                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2672                 }
2673         }
2674
2675         /**
2676          * @param string $str
2677          * @return mixed|string
2678          */
2679         function lcfirst( $str ) {
2680                 $o = ord( $str );
2681                 if ( !$o ) {
2682                         return strval( $str );
2683                 } elseif ( $o >= 128 ) {
2684                         return $this->lc( $str, true );
2685                 } elseif ( $o > 96 ) {
2686                         return $str;
2687                 } else {
2688                         $str[0] = strtolower( $str[0] );
2689                         return $str;
2690                 }
2691         }
2692
2693         /**
2694          * @param string $str
2695          * @param bool $first
2696          * @return mixed|string
2697          */
2698         function lc( $str, $first = false ) {
2699                 if ( $first ) {
2700                         if ( $this->isMultibyte( $str ) ) {
2701                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2702                         } else {
2703                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2704                         }
2705                 } else {
2706                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2707                 }
2708         }
2709
2710         /**
2711          * @param string $str
2712          * @return bool
2713          */
2714         function isMultibyte( $str ) {
2715                 return strlen( $str ) !== mb_strlen( $str );
2716         }
2717
2718         /**
2719          * @param string $str
2720          * @return mixed|string
2721          */
2722         function ucwords( $str ) {
2723                 if ( $this->isMultibyte( $str ) ) {
2724                         $str = $this->lc( $str );
2725
2726                         // regexp to find first letter in each word (i.e. after each space)
2727                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2728
2729                         // function to use to capitalize a single char
2730                         return preg_replace_callback(
2731                                 $replaceRegexp,
2732                                 [ $this, 'ucwordsCallbackMB' ],
2733                                 $str
2734                         );
2735                 } else {
2736                         return ucwords( strtolower( $str ) );
2737                 }
2738         }
2739
2740         /**
2741          * capitalize words at word breaks
2742          *
2743          * @param string $str
2744          * @return mixed
2745          */
2746         function ucwordbreaks( $str ) {
2747                 if ( $this->isMultibyte( $str ) ) {
2748                         $str = $this->lc( $str );
2749
2750                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2751                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2752
2753                         // find first letter after word break
2754                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2755                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2756
2757                         return preg_replace_callback(
2758                                 $replaceRegexp,
2759                                 [ $this, 'ucwordbreaksCallbackMB' ],
2760                                 $str
2761                         );
2762                 } else {
2763                         return preg_replace_callback(
2764                                 '/\b([\w\x80-\xff]+)\b/',
2765                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2766                                 $str
2767                         );
2768                 }
2769         }
2770
2771         /**
2772          * Return a case-folded representation of $s
2773          *
2774          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2775          * and $s2 are the same except for the case of their characters. It is not
2776          * necessary for the value returned to make sense when displayed.
2777          *
2778          * Do *not* perform any other normalisation in this function. If a caller
2779          * uses this function when it should be using a more general normalisation
2780          * function, then fix the caller.
2781          *
2782          * @param string $s
2783          *
2784          * @return string
2785          */
2786         function caseFold( $s ) {
2787                 return $this->uc( $s );
2788         }
2789
2790         /**
2791          * @param string $s
2792          * @return string
2793          * @throws MWException
2794          */
2795         function checkTitleEncoding( $s ) {
2796                 if ( is_array( $s ) ) {
2797                         throw new MWException( 'Given array to checkTitleEncoding.' );
2798                 }
2799                 if ( StringUtils::isUtf8( $s ) ) {
2800                         return $s;
2801                 }
2802
2803                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2804         }
2805
2806         /**
2807          * @return array
2808          */
2809         function fallback8bitEncoding() {
2810                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2811         }
2812
2813         /**
2814          * Most writing systems use whitespace to break up words.
2815          * Some languages such as Chinese don't conventionally do this,
2816          * which requires special handling when breaking up words for
2817          * searching etc.
2818          *
2819          * @return bool
2820          */
2821         function hasWordBreaks() {
2822                 return true;
2823         }
2824
2825         /**
2826          * Some languages such as Chinese require word segmentation,
2827          * Specify such segmentation when overridden in derived class.
2828          *
2829          * @param string $string
2830          * @return string
2831          */
2832         function segmentByWord( $string ) {
2833                 return $string;
2834         }
2835
2836         /**
2837          * Some languages have special punctuation need to be normalized.
2838          * Make such changes here.
2839          *
2840          * @param string $string
2841          * @return string
2842          */
2843         function normalizeForSearch( $string ) {
2844                 return self::convertDoubleWidth( $string );
2845         }
2846
2847         /**
2848          * convert double-width roman characters to single-width.
2849          * range: ff00-ff5f ~= 0020-007f
2850          *
2851          * @param string $string
2852          *
2853          * @return string
2854          */
2855         protected static function convertDoubleWidth( $string ) {
2856                 static $full = null;
2857                 static $half = null;
2858
2859                 if ( $full === null ) {
2860                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2861                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2862                         $full = str_split( $fullWidth, 3 );
2863                         $half = str_split( $halfWidth );
2864                 }
2865
2866                 $string = str_replace( $full, $half, $string );
2867                 return $string;
2868         }
2869
2870         /**
2871          * @param string $string
2872          * @param string $pattern
2873          * @return string
2874          */
2875         protected static function insertSpace( $string, $pattern ) {
2876                 $string = preg_replace( $pattern, " $1 ", $string );
2877                 $string = preg_replace( '/ +/', ' ', $string );
2878                 return $string;
2879         }
2880
2881         /**
2882          * @param array $termsArray
2883          * @return array
2884          */
2885         function convertForSearchResult( $termsArray ) {
2886                 # some languages, e.g. Chinese, need to do a conversion
2887                 # in order for search results to be displayed correctly
2888                 return $termsArray;
2889         }
2890
2891         /**
2892          * Get the first character of a string.
2893          *
2894          * @param string $s
2895          * @return string
2896          */
2897         function firstChar( $s ) {
2898                 $matches = [];
2899                 preg_match(
2900                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2901                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2902                         $s,
2903                         $matches
2904                 );
2905
2906                 if ( isset( $matches[1] ) ) {
2907                         if ( strlen( $matches[1] ) != 3 ) {
2908                                 return $matches[1];
2909                         }
2910
2911                         // Break down Hangul syllables to grab the first jamo
2912                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2913                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2914                                 return $matches[1];
2915                         } elseif ( $code < 0xb098 ) {
2916                                 return "\xe3\x84\xb1";
2917                         } elseif ( $code < 0xb2e4 ) {
2918                                 return "\xe3\x84\xb4";
2919                         } elseif ( $code < 0xb77c ) {
2920                                 return "\xe3\x84\xb7";
2921                         } elseif ( $code < 0xb9c8 ) {
2922                                 return "\xe3\x84\xb9";
2923                         } elseif ( $code < 0xbc14 ) {
2924                                 return "\xe3\x85\x81";
2925                         } elseif ( $code < 0xc0ac ) {
2926                                 return "\xe3\x85\x82";
2927                         } elseif ( $code < 0xc544 ) {
2928                                 return "\xe3\x85\x85";
2929                         } elseif ( $code < 0xc790 ) {
2930                                 return "\xe3\x85\x87";
2931                         } elseif ( $code < 0xcc28 ) {
2932                                 return "\xe3\x85\x88";
2933                         } elseif ( $code < 0xce74 ) {
2934                                 return "\xe3\x85\x8a";
2935                         } elseif ( $code < 0xd0c0 ) {
2936                                 return "\xe3\x85\x8b";
2937                         } elseif ( $code < 0xd30c ) {
2938                                 return "\xe3\x85\x8c";
2939                         } elseif ( $code < 0xd558 ) {
2940                                 return "\xe3\x85\x8d";
2941                         } else {
2942                                 return "\xe3\x85\x8e";
2943                         }
2944                 } else {
2945                         return '';
2946                 }
2947         }
2948
2949         /**
2950          * @deprecated No-op since 1.28
2951          */
2952         function initEncoding() {
2953                 // No-op.
2954         }
2955
2956         /**
2957          * @param string $s
2958          * @return string
2959          * @deprecated No-op since 1.28
2960          */
2961         function recodeForEdit( $s ) {
2962                 return $s;
2963         }
2964
2965         /**
2966          * @param string $s
2967          * @return string
2968          * @deprecated No-op since 1.28
2969          */
2970         function recodeInput( $s ) {
2971                 return $s;
2972         }
2973
2974         /**
2975          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
2976          * also cleans up certain backwards-compatible sequences, converting them
2977          * to the modern Unicode equivalent.
2978          *
2979          * This is language-specific for performance reasons only.
2980          *
2981          * @param string $s
2982          *
2983          * @return string
2984          */
2985         function normalize( $s ) {
2986                 global $wgAllUnicodeFixes;
2987                 $s = UtfNormal\Validator::cleanUp( $s );
2988                 if ( $wgAllUnicodeFixes ) {
2989                         $s = $this->transformUsingPairFile( 'normalize-ar.ser', $s );
2990                         $s = $this->transformUsingPairFile( 'normalize-ml.ser', $s );
2991                 }
2992
2993                 return $s;
2994         }
2995
2996         /**
2997          * Transform a string using serialized data stored in the given file (which
2998          * must be in the serialized subdirectory of $IP). The file contains pairs
2999          * mapping source characters to destination characters.
3000          *
3001          * The data is cached in process memory. This will go faster if you have the
3002          * FastStringSearch extension.
3003          *
3004          * @param string $file
3005          * @param string $string
3006          *
3007          * @throws MWException
3008          * @return string
3009          */
3010         function transformUsingPairFile( $file, $string ) {
3011                 if ( !isset( $this->transformData[$file] ) ) {
3012                         $data = wfGetPrecompiledData( $file );
3013                         if ( $data === false ) {
3014                                 throw new MWException( __METHOD__ . ": The transformation file $file is missing" );
3015                         }
3016                         $this->transformData[$file] = new ReplacementArray( $data );
3017                 }
3018                 return $this->transformData[$file]->replace( $string );
3019         }
3020
3021         /**
3022          * For right-to-left language support
3023          *
3024          * @return bool
3025          */
3026         function isRTL() {
3027                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3028         }
3029
3030         /**
3031          * Return the correct HTML 'dir' attribute value for this language.
3032          * @return string
3033          */
3034         function getDir() {
3035                 return $this->isRTL() ? 'rtl' : 'ltr';
3036         }
3037
3038         /**
3039          * Return 'left' or 'right' as appropriate alignment for line-start
3040          * for this language's text direction.
3041          *
3042          * Should be equivalent to CSS3 'start' text-align value....
3043          *
3044          * @return string
3045          */
3046         function alignStart() {
3047                 return $this->isRTL() ? 'right' : 'left';
3048         }
3049
3050         /**
3051          * Return 'right' or 'left' as appropriate alignment for line-end
3052          * for this language's text direction.
3053          *
3054          * Should be equivalent to CSS3 'end' text-align value....
3055          *
3056          * @return string
3057          */
3058         function alignEnd() {
3059                 return $this->isRTL() ? 'left' : 'right';
3060         }
3061
3062         /**
3063          * A hidden direction mark (LRM or RLM), depending on the language direction.
3064          * Unlike getDirMark(), this function returns the character as an HTML entity.
3065          * This function should be used when the output is guaranteed to be HTML,
3066          * because it makes the output HTML source code more readable. When
3067          * the output is plain text or can be escaped, getDirMark() should be used.
3068          *
3069          * @param bool $opposite Get the direction mark opposite to your language
3070          * @return string
3071          * @since 1.20
3072          */
3073         function getDirMarkEntity( $opposite = false ) {
3074                 if ( $opposite ) {
3075                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3076                 }
3077                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3078         }
3079
3080         /**
3081          * A hidden direction mark (LRM or RLM), depending on the language direction.
3082          * This function produces them as invisible Unicode characters and
3083          * the output may be hard to read and debug, so it should only be used
3084          * when the output is plain text or can be escaped. When the output is
3085          * HTML, use getDirMarkEntity() instead.
3086          *
3087          * @param bool $opposite Get the direction mark opposite to your language
3088          * @return string
3089          */
3090         function getDirMark( $opposite = false ) {
3091                 $lrm = "\xE2\x80\x8E"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3092                 $rlm = "\xE2\x80\x8F"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3093                 if ( $opposite ) {
3094                         return $this->isRTL() ? $lrm : $rlm;
3095                 }
3096                 return $this->isRTL() ? $rlm : $lrm;
3097         }
3098
3099         /**
3100          * @return array
3101          */
3102         function capitalizeAllNouns() {
3103                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3104         }
3105
3106         /**
3107          * An arrow, depending on the language direction.
3108          *
3109          * @param string $direction The direction of the arrow: forwards (default),
3110          *   backwards, left, right, up, down.
3111          * @return string
3112          */
3113         function getArrow( $direction = 'forwards' ) {
3114                 switch ( $direction ) {
3115                 case 'forwards':
3116                         return $this->isRTL() ? '←' : '→';
3117                 case 'backwards':
3118                         return $this->isRTL() ? '→' : '←';
3119                 case 'left':
3120                         return '←';
3121                 case 'right':
3122                         return '→';
3123                 case 'up':
3124                         return '↑';
3125                 case 'down':
3126                         return '↓';
3127                 }
3128         }
3129
3130         /**
3131          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3132          *
3133          * @return bool
3134          */
3135         function linkPrefixExtension() {
3136                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3137         }
3138
3139         /**
3140          * Get all magic words from cache.
3141          * @return array
3142          */
3143         function getMagicWords() {
3144                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3145         }
3146
3147         /**
3148          * Run the LanguageGetMagic hook once.
3149          */
3150         protected function doMagicHook() {
3151                 if ( $this->mMagicHookDone ) {
3152                         return;
3153                 }
3154                 $this->mMagicHookDone = true;
3155                 Hooks::run( 'LanguageGetMagic', [ &$this->mMagicExtensions, $this->getCode() ] );
3156         }
3157
3158         /**
3159          * Fill a MagicWord object with data from here
3160          *
3161          * @param MagicWord $mw
3162          */
3163         function getMagic( $mw ) {
3164                 // Saves a function call
3165                 if ( !$this->mMagicHookDone ) {
3166                         $this->doMagicHook();
3167                 }
3168
3169                 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
3170                         $rawEntry = $this->mMagicExtensions[$mw->mId];
3171                 } else {
3172                         $rawEntry = self::$dataCache->getSubitem(
3173                                 $this->mCode, 'magicWords', $mw->mId );
3174                 }
3175
3176                 if ( !is_array( $rawEntry ) ) {
3177                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3178                 } else {
3179                         $mw->mCaseSensitive = $rawEntry[0];
3180                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3181                 }
3182         }
3183
3184         /**
3185          * Add magic words to the extension array
3186          *
3187          * @param array $newWords
3188          */
3189         function addMagicWordsByLang( $newWords ) {
3190                 $fallbackChain = $this->getFallbackLanguages();
3191                 $fallbackChain = array_reverse( $fallbackChain );
3192                 foreach ( $fallbackChain as $code ) {
3193                         if ( isset( $newWords[$code] ) ) {
3194                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3195                         }
3196                 }
3197         }
3198
3199         /**
3200          * Get special page names, as an associative array
3201          *   canonical name => array of valid names, including aliases
3202          * @return array
3203          */
3204         function getSpecialPageAliases() {
3205                 // Cache aliases because it may be slow to load them
3206                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3207                         // Initialise array
3208                         $this->mExtendedSpecialPageAliases =
3209                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3210                         Hooks::run( 'LanguageGetSpecialPageAliases',
3211                                 [ &$this->mExtendedSpecialPageAliases, $this->getCode() ] );
3212                 }
3213
3214                 return $this->mExtendedSpecialPageAliases;
3215         }
3216
3217         /**
3218          * Italic is unsuitable for some languages
3219          *
3220          * @param string $text The text to be emphasized.
3221          * @return string
3222          */
3223         function emphasize( $text ) {
3224                 return "<em>$text</em>";
3225         }
3226
3227         /**
3228          * Normally we output all numbers in plain en_US style, that is
3229          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3230          * point twohundredthirtyfive. However this is not suitable for all
3231          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3232          * Icelandic just want to use commas instead of dots, and dots instead
3233          * of commas like "293.291,235".
3234          *
3235          * An example of this function being called:
3236          * <code>
3237          * wfMessage( 'message' )->numParams( $num )->text()
3238          * </code>
3239          *
3240          * See $separatorTransformTable on MessageIs.php for
3241          * the , => . and . => , implementation.
3242          *
3243          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3244          * @param int|float $number The string to be formatted, should be an integer
3245          *   or a floating point number.
3246          * @param bool $nocommafy Set to true for special numbers like dates
3247          * @return string
3248          */
3249         public function formatNum( $number, $nocommafy = false ) {
3250                 global $wgTranslateNumerals;
3251                 if ( !$nocommafy ) {
3252                         $number = $this->commafy( $number );
3253                         $s = $this->separatorTransformTable();
3254                         if ( $s ) {
3255                                 $number = strtr( $number, $s );
3256                         }
3257                 }
3258
3259                 if ( $wgTranslateNumerals ) {
3260                         $s = $this->digitTransformTable();
3261                         if ( $s ) {
3262                                 $number = strtr( $number, $s );
3263                         }
3264                 }
3265
3266                 return $number;
3267         }
3268
3269         /**
3270          * Front-end for non-commafied formatNum
3271          *
3272          * @param int|float $number The string to be formatted, should be an integer
3273          *        or a floating point number.
3274          * @since 1.21
3275          * @return string
3276          */
3277         public function formatNumNoSeparators( $number ) {
3278                 return $this->formatNum( $number, true );
3279         }
3280
3281         /**
3282          * @param string $number
3283          * @return string
3284          */
3285         public function parseFormattedNumber( $number ) {
3286                 $s = $this->digitTransformTable();
3287                 if ( $s ) {
3288                         // eliminate empty array values such as ''. (T66347)
3289                         $s = array_filter( $s );
3290                         $number = strtr( $number, array_flip( $s ) );
3291                 }
3292
3293                 $s = $this->separatorTransformTable();
3294                 if ( $s ) {
3295                         // eliminate empty array values such as ''. (T66347)
3296                         $s = array_filter( $s );
3297                         $number = strtr( $number, array_flip( $s ) );
3298                 }
3299
3300                 $number = strtr( $number, [ ',' => '' ] );
3301                 return $number;
3302         }
3303
3304         /**
3305          * Adds commas to a given number
3306          * @since 1.19
3307          * @param mixed $number
3308          * @return string
3309          */
3310         function commafy( $number ) {
3311                 $digitGroupingPattern = $this->digitGroupingPattern();
3312                 if ( $number === null ) {
3313                         return '';
3314                 }
3315
3316                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3317                         // default grouping is at thousands,  use the same for ###,###,### pattern too.
3318                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3319                 } else {
3320                         // Ref: http://cldr.unicode.org/translation/number-patterns
3321                         $sign = "";
3322                         if ( intval( $number ) < 0 ) {
3323                                 // For negative numbers apply the algorithm like positive number and add sign.
3324                                 $sign = "-";
3325                                 $number = substr( $number, 1 );
3326                         }
3327                         $integerPart = [];
3328                         $decimalPart = [];
3329                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3330                         preg_match( "/\d+/", $number, $integerPart );
3331                         preg_match( "/\.\d*/", $number, $decimalPart );
3332                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3333                         if ( $groupedNumber === $number ) {
3334                                 // the string does not have any number part. Eg: .12345
3335                                 return $sign . $groupedNumber;
3336                         }
3337                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3338                         while ( $start > 0 ) {
3339                                 $match = $matches[0][$numMatches - 1];
3340                                 $matchLen = strlen( $match );
3341                                 $start = $end - $matchLen;
3342                                 if ( $start < 0 ) {
3343                                         $start = 0;
3344                                 }
3345                                 $groupedNumber = substr( $number, $start, $end -$start ) . $groupedNumber;
3346                                 $end = $start;
3347                                 if ( $numMatches > 1 ) {
3348                                         // use the last pattern for the rest of the number
3349                                         $numMatches--;
3350                                 }
3351                                 if ( $start > 0 ) {
3352                                         $groupedNumber = "," . $groupedNumber;
3353                                 }
3354                         }
3355                         return $sign . $groupedNumber;
3356                 }
3357         }
3358
3359         /**
3360          * @return string
3361          */
3362         function digitGroupingPattern() {
3363                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3364         }
3365
3366         /**
3367          * @return array
3368          */
3369         function digitTransformTable() {
3370                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3371         }
3372
3373         /**
3374          * @return array
3375          */
3376         function separatorTransformTable() {
3377                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3378         }
3379
3380         /**
3381          * Take a list of strings and build a locale-friendly comma-separated
3382          * list, using the local comma-separator message.
3383          * The last two strings are chained with an "and".
3384          * NOTE: This function will only work with standard numeric array keys (0, 1, 2…)
3385          *
3386          * @param string[] $l
3387          * @return string
3388          */
3389         function listToText( array $l ) {
3390                 $m = count( $l ) - 1;
3391                 if ( $m < 0 ) {
3392                         return '';
3393                 }
3394                 if ( $m > 0 ) {
3395                         $and = $this->msg( 'and' )->escaped();
3396                         $space = $this->msg( 'word-separator' )->escaped();
3397                         if ( $m > 1 ) {
3398                                 $comma = $this->msg( 'comma-separator' )->escaped();
3399                         }
3400                 }
3401                 $s = $l[$m];
3402                 for ( $i = $m - 1; $i >= 0; $i-- ) {
3403                         if ( $i == $m - 1 ) {
3404                                 $s = $l[$i] . $and . $space . $s;
3405                         } else {
3406                                 $s = $l[$i] . $comma . $s;
3407                         }
3408                 }
3409                 return $s;
3410         }
3411
3412         /**
3413          * Take a list of strings and build a locale-friendly comma-separated
3414          * list, using the local comma-separator message.
3415          * @param string[] $list Array of strings to put in a comma list
3416          * @return string
3417          */
3418         function commaList( array $list ) {
3419                 return implode(
3420                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3421                         $list
3422                 );
3423         }
3424
3425         /**
3426          * Take a list of strings and build a locale-friendly semicolon-separated
3427          * list, using the local semicolon-separator message.
3428          * @param string[] $list Array of strings to put in a semicolon list
3429          * @return string
3430          */
3431         function semicolonList( array $list ) {
3432                 return implode(
3433                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3434                         $list
3435                 );
3436         }
3437
3438         /**
3439          * Same as commaList, but separate it with the pipe instead.
3440          * @param string[] $list Array of strings to put in a pipe list
3441          * @return string
3442          */
3443         function pipeList( array $list ) {
3444                 return implode(
3445                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3446                         $list
3447                 );
3448         }
3449
3450         /**
3451          * Truncate a string to a specified length in bytes, appending an optional
3452          * string (e.g. for ellipses)
3453          *
3454          * The database offers limited byte lengths for some columns in the database;
3455          * multi-byte character sets mean we need to ensure that only whole characters
3456          * are included, otherwise broken characters can be passed to the user
3457          *
3458          * If $length is negative, the string will be truncated from the beginning
3459          *
3460          * @param string $string String to truncate
3461          * @param int $length Maximum length (including ellipses)
3462          * @param string $ellipsis String to append to the truncated text
3463          * @param bool $adjustLength Subtract length of ellipsis from $length.
3464          *      $adjustLength was introduced in 1.18, before that behaved as if false.
3465          * @return string
3466          */
3467         function truncate( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3468                 # Use the localized ellipsis character
3469                 if ( $ellipsis == '...' ) {
3470                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3471                 }
3472                 # Check if there is no need to truncate
3473                 if ( $length == 0 ) {
3474                         return $ellipsis; // convention
3475                 } elseif ( strlen( $string ) <= abs( $length ) ) {
3476                         return $string; // no need to truncate
3477                 }
3478                 $stringOriginal = $string;
3479                 # If ellipsis length is >= $length then we can't apply $adjustLength
3480                 if ( $adjustLength && strlen( $ellipsis ) >= abs( $length ) ) {
3481                         $string = $ellipsis; // this can be slightly unexpected
3482                 # Otherwise, truncate and add ellipsis...
3483                 } else {
3484                         $eLength = $adjustLength ? strlen( $ellipsis ) : 0;
3485                         if ( $length > 0 ) {
3486                                 $length -= $eLength;
3487                                 $string = substr( $string, 0, $length ); // xyz...
3488                                 $string = $this->removeBadCharLast( $string );
3489                                 $string = rtrim( $string );
3490                                 $string = $string . $ellipsis;
3491                         } else {
3492                                 $length += $eLength;
3493                                 $string = substr( $string, $length ); // ...xyz
3494                                 $string = $this->removeBadCharFirst( $string );
3495                                 $string = ltrim( $string );
3496                                 $string = $ellipsis . $string;
3497                         }
3498                 }
3499                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3500                 # This check is *not* redundant if $adjustLength, due to the single case where
3501                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3502                 if ( strlen( $string ) < strlen( $stringOriginal ) ) {
3503                         return $string;
3504                 } else {
3505                         return $stringOriginal;
3506                 }
3507         }
3508
3509         /**
3510          * Remove bytes that represent an incomplete Unicode character
3511          * at the end of string (e.g. bytes of the char are missing)
3512          *
3513          * @param string $string
3514          * @return string
3515          */
3516         protected function removeBadCharLast( $string ) {
3517                 if ( $string != '' ) {
3518                         $char = ord( $string[strlen( $string ) - 1] );
3519                         $m = [];
3520                         if ( $char >= 0xc0 ) {
3521                                 # We got the first byte only of a multibyte char; remove it.
3522                                 $string = substr( $string, 0, -1 );
3523                         } elseif ( $char >= 0x80 &&
3524                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3525                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3526                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3527                         ) {
3528                                 # We chopped in the middle of a character; remove it
3529                                 $string = $m[1];
3530                         }
3531                 }
3532                 return $string;
3533         }
3534
3535         /**
3536          * Remove bytes that represent an incomplete Unicode character
3537          * at the start of string (e.g. bytes of the char are missing)
3538          *
3539          * @param string $string
3540          * @return string
3541          */
3542         protected function removeBadCharFirst( $string ) {
3543                 if ( $string != '' ) {
3544                         $char = ord( $string[0] );
3545                         if ( $char >= 0x80 && $char < 0xc0 ) {
3546                                 # We chopped in the middle of a character; remove the whole thing
3547                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3548                         }
3549                 }
3550                 return $string;
3551         }
3552
3553         /**
3554          * Truncate a string of valid HTML to a specified length in bytes,
3555          * appending an optional string (e.g. for ellipses), and return valid HTML
3556          *
3557          * This is only intended for styled/linked text, such as HTML with
3558          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3559          * Also, this will not detect things like "display:none" CSS.
3560          *
3561          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3562          *
3563          * @param string $text HTML string to truncate
3564          * @param int $length (zero/positive) Maximum length (including ellipses)
3565          * @param string $ellipsis String to append to the truncated text
3566          * @return string
3567          */
3568         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3569                 # Use the localized ellipsis character
3570                 if ( $ellipsis == '...' ) {
3571                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3572                 }
3573                 # Check if there is clearly no need to truncate
3574                 if ( $length <= 0 ) {
3575                         return $ellipsis; // no text shown, nothing to format (convention)
3576                 } elseif ( strlen( $text ) <= $length ) {
3577                         return $text; // string short enough even *with* HTML (short-circuit)
3578                 }
3579
3580                 $dispLen = 0; // innerHTML legth so far
3581                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3582                 $tagType = 0; // 0-open, 1-close
3583                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3584                 $entityState = 0; // 0-not entity, 1-entity
3585                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3586                 $openTags = []; // open tag stack
3587                 $maybeState = null; // possible truncation state
3588
3589                 $textLen = strlen( $text );
3590                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3591                 for ( $pos = 0; true; ++$pos ) {
3592                         # Consider truncation once the display length has reached the maximim.
3593                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3594                         # Check that we're not in the middle of a bracket/entity...
3595                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3596                                 if ( !$testingEllipsis ) {
3597                                         $testingEllipsis = true;
3598                                         # Save where we are; we will truncate here unless there turn out to
3599                                         # be so few remaining characters that truncation is not necessary.
3600                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3601                                                 $maybeState = [ $ret, $openTags ]; // save state
3602                                         }
3603                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3604                                         # String in fact does need truncation, the truncation point was OK.
3605                                         list( $ret, $openTags ) = $maybeState; // reload state
3606                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3607                                         $ret .= $ellipsis; // add ellipsis
3608                                         break;
3609                                 }
3610                         }
3611                         if ( $pos >= $textLen ) {
3612                                 break; // extra iteration just for above checks
3613                         }
3614
3615                         # Read the next char...
3616                         $ch = $text[$pos];
3617                         $lastCh = $pos ? $text[$pos - 1] : '';
3618                         $ret .= $ch; // add to result string
3619                         if ( $ch == '<' ) {
3620                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3621                                 $entityState = 0; // for bad HTML
3622                                 $bracketState = 1; // tag started (checking for backslash)
3623                         } elseif ( $ch == '>' ) {
3624                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3625                                 $entityState = 0; // for bad HTML
3626                                 $bracketState = 0; // out of brackets
3627                         } elseif ( $bracketState == 1 ) {
3628                                 if ( $ch == '/' ) {
3629                                         $tagType = 1; // close tag (e.g. "</span>")
3630                                 } else {
3631                                         $tagType = 0; // open tag (e.g. "<span>")
3632                                         $tag .= $ch;
3633                                 }
3634                                 $bracketState = 2; // building tag name
3635                         } elseif ( $bracketState == 2 ) {
3636                                 if ( $ch != ' ' ) {
3637                                         $tag .= $ch;
3638                                 } else {
3639                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3640                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3641                                 }
3642                         } elseif ( $bracketState == 0 ) {
3643                                 if ( $entityState ) {
3644                                         if ( $ch == ';' ) {
3645                                                 $entityState = 0;
3646                                                 $dispLen++; // entity is one displayed char
3647                                         }
3648                                 } else {
3649                                         if ( $neLength == 0 && !$maybeState ) {
3650                                                 // Save state without $ch. We want to *hit* the first
3651                                                 // display char (to get tags) but not *use* it if truncating.
3652                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3653                                         }
3654                                         if ( $ch == '&' ) {
3655                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3656                                         } else {
3657                                                 $dispLen++; // this char is displayed
3658                                                 // Add the next $max display text chars after this in one swoop...
3659                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3660                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3661                                                 $dispLen += $skipped;
3662                                                 $pos += $skipped;
3663                                         }
3664                                 }
3665                         }
3666                 }
3667                 // Close the last tag if left unclosed by bad HTML
3668                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3669                 while ( count( $openTags ) > 0 ) {
3670                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3671                 }
3672                 return $ret;
3673         }
3674
3675         /**
3676          * truncateHtml() helper function
3677          * like strcspn() but adds the skipped chars to $ret
3678          *
3679          * @param string $ret
3680          * @param string $text
3681          * @param string $search
3682          * @param int $start
3683          * @param null|int $len
3684          * @return int
3685          */
3686         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3687                 if ( $len === null ) {
3688                         $len = -1; // -1 means "no limit" for strcspn
3689                 } elseif ( $len < 0 ) {
3690                         $len = 0; // sanity
3691                 }
3692                 $skipCount = 0;
3693                 if ( $start < strlen( $text ) ) {
3694                         $skipCount = strcspn( $text, $search, $start, $len );
3695                         $ret .= substr( $text, $start, $skipCount );
3696                 }
3697                 return $skipCount;
3698         }
3699
3700         /**
3701          * truncateHtml() helper function
3702          * (a) push or pop $tag from $openTags as needed
3703          * (b) clear $tag value
3704          * @param string &$tag Current HTML tag name we are looking at
3705          * @param int $tagType (0-open tag, 1-close tag)
3706          * @param string $lastCh Character before the '>' that ended this tag
3707          * @param array &$openTags Open tag stack (not accounting for $tag)
3708          */
3709         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3710                 $tag = ltrim( $tag );
3711                 if ( $tag != '' ) {
3712                         if ( $tagType == 0 && $lastCh != '/' ) {
3713                                 $openTags[] = $tag; // tag opened (didn't close itself)
3714                         } elseif ( $tagType == 1 ) {
3715                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3716                                         array_pop( $openTags ); // tag closed
3717                                 }
3718                         }
3719                         $tag = '';
3720                 }
3721         }
3722
3723         /**
3724          * Grammatical transformations, needed for inflected languages
3725          * Invoked by putting {{grammar:case|word}} in a message
3726          *
3727          * @param string $word
3728          * @param string $case
3729          * @return string
3730          */
3731         function convertGrammar( $word, $case ) {
3732                 global $wgGrammarForms;
3733                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3734                         return $wgGrammarForms[$this->getCode()][$case][$word];
3735                 }
3736
3737                 $grammarTransformations = $this->getGrammarTransformations();
3738
3739                 if ( isset( $grammarTransformations[$case] ) ) {
3740                         $forms = $grammarTransformations[$case];
3741
3742                         // Some names of grammar rules are aliases for other rules.
3743                         // In such cases the value is a string rather than object,
3744                         // so load the actual rules.
3745                         if ( is_string( $forms ) ) {
3746                                 $forms = $grammarTransformations[$forms];
3747                         }
3748
3749                         foreach ( array_values( $forms ) as $rule ) {
3750                                 $form = $rule[0];
3751
3752                                 if ( $form === '@metadata' ) {
3753                                         continue;
3754                                 }
3755
3756                                 $replacement = $rule[1];
3757
3758                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3759                                 $patternMatches = preg_match( $regex, $word );
3760
3761                                 if ( $patternMatches === false ) {
3762                                         wfLogWarning(
3763                                                 'An error occurred while processing grammar. ' .
3764                                                 "Word: '$word'. Regex: /$form/."
3765                                         );
3766                                 } elseif ( $patternMatches === 1 ) {
3767                                         $word = preg_replace( $regex, $replacement, $word );
3768
3769                                         break;
3770                                 }
3771                         }
3772                 }
3773
3774                 return $word;
3775         }
3776
3777         /**
3778          * Get the grammar forms for the content language
3779          * @return array Array of grammar forms
3780          * @since 1.20
3781          */
3782         function getGrammarForms() {
3783                 global $wgGrammarForms;
3784                 if ( isset( $wgGrammarForms[$this->getCode()] )
3785                         && is_array( $wgGrammarForms[$this->getCode()] )
3786                 ) {
3787                         return $wgGrammarForms[$this->getCode()];
3788                 }
3789
3790                 return [];
3791         }
3792
3793         /**
3794          * Get the grammar transformations data for the language.
3795          * Used like grammar forms, with {{GRAMMAR}} and cases,
3796          * but uses pairs of regexes and replacements instead of code.
3797          *
3798          * @return array[] Array of grammar transformations.
3799          * @throws MWException
3800          * @since 1.28
3801          */
3802         public function getGrammarTransformations() {
3803                 $languageCode = $this->getCode();
3804
3805                 if ( self::$grammarTransformations === null ) {
3806                         self::$grammarTransformations = new MapCacheLRU( 10 );
3807                 }
3808
3809                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3810                         return self::$grammarTransformations->get( $languageCode );
3811                 }
3812
3813                 $data = [];
3814
3815                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3816                 if ( is_readable( $grammarDataFile ) ) {
3817                         $data = FormatJson::decode(
3818                                 file_get_contents( $grammarDataFile ),
3819                                 true
3820                         );
3821
3822                         if ( $data === null ) {
3823                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3824                         }
3825
3826                         self::$grammarTransformations->set( $languageCode, $data );
3827                 }
3828
3829                 return $data;
3830         }
3831
3832         /**
3833          * Provides an alternative text depending on specified gender.
3834          * Usage {{gender:username|masculine|feminine|unknown}}.
3835          * username is optional, in which case the gender of current user is used,
3836          * but only in (some) interface messages; otherwise default gender is used.
3837          *
3838          * If no forms are given, an empty string is returned. If only one form is
3839          * given, it will be returned unconditionally. These details are implied by
3840          * the caller and cannot be overridden in subclasses.
3841          *
3842          * If three forms are given, the default is to use the third (unknown) form.
3843          * If fewer than three forms are given, the default is to use the first (masculine) form.
3844          * These details can be overridden in subclasses.
3845          *
3846          * @param string $gender
3847          * @param array $forms
3848          *
3849          * @return string
3850          */
3851         function gender( $gender, $forms ) {
3852                 if ( !count( $forms ) ) {
3853                         return '';
3854                 }
3855                 $forms = $this->preConvertPlural( $forms, 2 );
3856                 if ( $gender === 'male' ) {
3857                         return $forms[0];
3858                 }
3859                 if ( $gender === 'female' ) {
3860                         return $forms[1];
3861                 }
3862                 return isset( $forms[2] ) ? $forms[2] : $forms[0];
3863         }
3864
3865         /**
3866          * Plural form transformations, needed for some languages.
3867          * For example, there are 3 form of plural in Russian and Polish,
3868          * depending on "count mod 10". See [[w:Plural]]
3869          * For English it is pretty simple.
3870          *
3871          * Invoked by putting {{plural:count|wordform1|wordform2}}
3872          * or {{plural:count|wordform1|wordform2|wordform3}}
3873          *
3874          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3875          *
3876          * @param int $count Non-localized number
3877          * @param array $forms Different plural forms
3878          * @return string Correct form of plural for $count in this language
3879          */
3880         function convertPlural( $count, $forms ) {
3881                 // Handle explicit n=pluralform cases
3882                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3883                 if ( is_string( $forms ) ) {
3884                         return $forms;
3885                 }
3886                 if ( !count( $forms ) ) {
3887                         return '';
3888                 }
3889
3890                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3891                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3892                 return $forms[$pluralForm];
3893         }
3894
3895         /**
3896          * Handles explicit plural forms for Language::convertPlural()
3897          *
3898          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
3899          * If an explicitly defined plural form matches the $count, then
3900          * string value returned, otherwise array returned for further consideration
3901          * by CLDR rules or overridden convertPlural().
3902          *
3903          * @since 1.23
3904          *
3905          * @param int $count Non-localized number
3906          * @param array $forms Different plural forms
3907          *
3908          * @return array|string
3909          */
3910         protected function handleExplicitPluralForms( $count, array $forms ) {
3911                 foreach ( $forms as $index => $form ) {
3912                         if ( preg_match( '/\d+=/i', $form ) ) {
3913                                 $pos = strpos( $form, '=' );
3914                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
3915                                         return substr( $form, $pos + 1 );
3916                                 }
3917                                 unset( $forms[$index] );
3918                         }
3919                 }
3920                 return array_values( $forms );
3921         }
3922
3923         /**
3924          * Checks that convertPlural was given an array and pads it to requested
3925          * amount of forms by copying the last one.
3926          *
3927          * @param array $forms Array of forms given to convertPlural
3928          * @param int $count How many forms should there be at least
3929          * @return array Padded array of forms or an exception if not an array
3930          */
3931         protected function preConvertPlural( /* Array */ $forms, $count ) {
3932                 while ( count( $forms ) < $count ) {
3933                         $forms[] = $forms[count( $forms ) - 1];
3934                 }
3935                 return $forms;
3936         }
3937
3938         /**
3939          * Wraps argument with unicode control characters for directionality safety
3940          *
3941          * This solves the problem where directionality-neutral characters at the edge of
3942          * the argument string get interpreted with the wrong directionality from the
3943          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
3944          *
3945          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
3946          * directionality of the argument string, using the BIDI algorithm's own "First
3947          * strong directional codepoint" rule. Essentially, this works round the fact that
3948          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
3949          * direction inference). The latter is cleaner but still not widely supported.
3950          *
3951          * @param string $text Text to wrap
3952          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
3953          */
3954         public function embedBidi( $text = '' ) {
3955                 $dir = Language::strongDirFromContent( $text );
3956                 if ( $dir === 'ltr' ) {
3957                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
3958                         return self::$lre . $text . self::$pdf;
3959                 }
3960                 if ( $dir === 'rtl' ) {
3961                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
3962                         return self::$rle . $text . self::$pdf;
3963                 }
3964                 // No strong directionality: do not wrap
3965                 return $text;
3966         }
3967
3968         /**
3969          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
3970          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
3971          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
3972          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
3973          * match up with it.
3974          *
3975          * @param string $str The validated block duration in English
3976          * @param User $user User object to use timezone from or null for $wgUser
3977          * @param int $now Current timestamp, for formatting relative block durations
3978          * @return string Somehow translated block duration
3979          * @see LanguageFi.php for example implementation
3980          */
3981         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
3982                 $duration = SpecialBlock::getSuggestedDurations( $this );
3983                 foreach ( $duration as $show => $value ) {
3984                         if ( strcmp( $str, $value ) == 0 ) {
3985                                 return htmlspecialchars( trim( $show ) );
3986                         }
3987                 }
3988
3989                 if ( wfIsInfinity( $str ) ) {
3990                         foreach ( $duration as $show => $value ) {
3991                                 if ( wfIsInfinity( $value ) ) {
3992                                         return htmlspecialchars( trim( $show ) );
3993                                 }
3994                         }
3995                 }
3996
3997                 // If all else fails, return a standard duration or timestamp description.
3998                 $time = strtotime( $str, $now );
3999                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4000                         return $str;
4001                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4002                         // The result differs based on current time, so the difference
4003                         // is a fixed duration length.
4004                         return $this->formatDuration( $time - $now );
4005                 } else { // It's an absolute timestamp.
4006                         if ( $time === 0 ) {
4007                                 // wfTimestamp() handles 0 as current time instead of epoch.
4008                                 $time = '19700101000000';
4009                         }
4010                         if ( $user ) {
4011                                 return $this->userTimeAndDate( $time, $user );
4012                         }
4013                         return $this->timeanddate( $time );
4014                 }
4015         }
4016
4017         /**
4018          * languages like Chinese need to be segmented in order for the diff
4019          * to be of any use
4020          *
4021          * @param string $text
4022          * @return string
4023          */
4024         public function segmentForDiff( $text ) {
4025                 return $text;
4026         }
4027
4028         /**
4029          * and unsegment to show the result
4030          *
4031          * @param string $text
4032          * @return string
4033          */
4034         public function unsegmentForDiff( $text ) {
4035                 return $text;
4036         }
4037
4038         /**
4039          * Return the LanguageConverter used in the Language
4040          *
4041          * @since 1.19
4042          * @return LanguageConverter
4043          */
4044         public function getConverter() {
4045                 return $this->mConverter;
4046         }
4047
4048         /**
4049          * convert text to all supported variants
4050          *
4051          * @param string $text
4052          * @return array
4053          */
4054         public function autoConvertToAllVariants( $text ) {
4055                 return $this->mConverter->autoConvertToAllVariants( $text );
4056         }
4057
4058         /**
4059          * convert text to different variants of a language.
4060          *
4061          * @param string $text
4062          * @return string
4063          */
4064         public function convert( $text ) {
4065                 return $this->mConverter->convert( $text );
4066         }
4067
4068         /**
4069          * Convert a Title object to a string in the preferred variant
4070          *
4071          * @param Title $title
4072          * @return string
4073          */
4074         public function convertTitle( $title ) {
4075                 return $this->mConverter->convertTitle( $title );
4076         }
4077
4078         /**
4079          * Convert a namespace index to a string in the preferred variant
4080          *
4081          * @param int $ns
4082          * @return string
4083          */
4084         public function convertNamespace( $ns ) {
4085                 return $this->mConverter->convertNamespace( $ns );
4086         }
4087
4088         /**
4089          * Check if this is a language with variants
4090          *
4091          * @return bool
4092          */
4093         public function hasVariants() {
4094                 return count( $this->getVariants() ) > 1;
4095         }
4096
4097         /**
4098          * Check if the language has the specific variant
4099          *
4100          * @since 1.19
4101          * @param string $variant
4102          * @return bool
4103          */
4104         public function hasVariant( $variant ) {
4105                 return (bool)$this->mConverter->validateVariant( $variant );
4106         }
4107
4108         /**
4109          * Perform output conversion on a string, and encode for safe HTML output.
4110          * @param string $text Text to be converted
4111          * @param bool $isTitle Whether this conversion is for the article title
4112          * @return string
4113          * @todo this should get integrated somewhere sane
4114          */
4115         public function convertHtml( $text, $isTitle = false ) {
4116                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
4117         }
4118
4119         /**
4120          * @param string $key
4121          * @return string
4122          */
4123         public function convertCategoryKey( $key ) {
4124                 return $this->mConverter->convertCategoryKey( $key );
4125         }
4126
4127         /**
4128          * Get the list of variants supported by this language
4129          * see sample implementation in LanguageZh.php
4130          *
4131          * @return string[] An array of language codes
4132          */
4133         public function getVariants() {
4134                 return $this->mConverter->getVariants();
4135         }
4136
4137         /**
4138          * @return string
4139          */
4140         public function getPreferredVariant() {
4141                 return $this->mConverter->getPreferredVariant();
4142         }
4143
4144         /**
4145          * @return string
4146          */
4147         public function getDefaultVariant() {
4148                 return $this->mConverter->getDefaultVariant();
4149         }
4150
4151         /**
4152          * @return string
4153          */
4154         public function getURLVariant() {
4155                 return $this->mConverter->getURLVariant();
4156         }
4157
4158         /**
4159          * If a language supports multiple variants, it is
4160          * possible that non-existing link in one variant
4161          * actually exists in another variant. this function
4162          * tries to find it. See e.g. LanguageZh.php
4163          * The input parameters may be modified upon return
4164          *
4165          * @param string &$link The name of the link
4166          * @param Title &$nt The title object of the link
4167          * @param bool $ignoreOtherCond To disable other conditions when
4168          *   we need to transclude a template or update a category's link
4169          */
4170         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4171                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4172         }
4173
4174         /**
4175          * returns language specific options used by User::getPageRenderHash()
4176          * for example, the preferred language variant
4177          *
4178          * @return string
4179          */
4180         function getExtraHashOptions() {
4181                 return $this->mConverter->getExtraHashOptions();
4182         }
4183
4184         /**
4185          * For languages that support multiple variants, the title of an
4186          * article may be displayed differently in different variants. this
4187          * function returns the apporiate title defined in the body of the article.
4188          *
4189          * @return string
4190          */
4191         public function getParsedTitle() {
4192                 return $this->mConverter->getParsedTitle();
4193         }
4194
4195         /**
4196          * Refresh the cache of conversion tables when
4197          * MediaWiki:Conversiontable* is updated.
4198          *
4199          * @param Title $title The Title of the page being updated
4200          */
4201         public function updateConversionTable( Title $title ) {
4202                 $this->mConverter->updateConversionTable( $title );
4203         }
4204
4205         /**
4206          * Prepare external link text for conversion. When the text is
4207          * a URL, it shouldn't be converted, and it'll be wrapped in
4208          * the "raw" tag (-{R| }-) to prevent conversion.
4209          *
4210          * This function is called "markNoConversion" for historical
4211          * reasons.
4212          *
4213          * @param string $text Text to be used for external link
4214          * @param bool $noParse Wrap it without confirming it's a real URL first
4215          * @return string The tagged text
4216          */
4217         public function markNoConversion( $text, $noParse = false ) {
4218                 // Excluding protocal-relative URLs may avoid many false positives.
4219                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
4220                         return $this->mConverter->markNoConversion( $text );
4221                 } else {
4222                         return $text;
4223                 }
4224         }
4225
4226         /**
4227          * A regular expression to match legal word-trailing characters
4228          * which should be merged onto a link of the form [[foo]]bar.
4229          *
4230          * @return string
4231          */
4232         public function linkTrail() {
4233                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4234         }
4235
4236         /**
4237          * A regular expression character set to match legal word-prefixing
4238          * characters which should be merged onto a link of the form foo[[bar]].
4239          *
4240          * @return string
4241          */
4242         public function linkPrefixCharset() {
4243                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4244         }
4245
4246         /**
4247          * Get the "parent" language which has a converter to convert a "compatible" language
4248          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4249          *
4250          * @return Language|null
4251          * @since 1.22
4252          */
4253         public function getParentLanguage() {
4254                 if ( $this->mParentLanguage !== false ) {
4255                         return $this->mParentLanguage;
4256                 }
4257
4258                 $code = explode( '-', $this->getCode() )[0];
4259                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4260                         $this->mParentLanguage = null;
4261                         return null;
4262                 }
4263                 $lang = Language::factory( $code );
4264                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4265                         $this->mParentLanguage = null;
4266                         return null;
4267                 }
4268
4269                 $this->mParentLanguage = $lang;
4270                 return $lang;
4271         }
4272
4273         /**
4274          * Compare with an other language object
4275          *
4276          * @since 1.28
4277          * @param Language $lang
4278          * @return boolean
4279          */
4280         public function equals( Language $lang ) {
4281                 return $lang->getCode() === $this->mCode;
4282         }
4283
4284         /**
4285          * Get the internal language code for this language object
4286          *
4287          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4288          * htmlspecialchars() or similar
4289          *
4290          * @return string
4291          */
4292         public function getCode() {
4293                 return $this->mCode;
4294         }
4295
4296         /**
4297          * Get the code in BCP 47 format which we can use
4298          * inside of html lang="" tags.
4299          *
4300          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4301          * htmlspecialchars() or similar.
4302          *
4303          * @since 1.19
4304          * @return string
4305          */
4306         public function getHtmlCode() {
4307                 if ( is_null( $this->mHtmlCode ) ) {
4308                         $this->mHtmlCode = wfBCP47( $this->getCode() );
4309                 }
4310                 return $this->mHtmlCode;
4311         }
4312
4313         /**
4314          * @param string $code
4315          */
4316         public function setCode( $code ) {
4317                 $this->mCode = $code;
4318                 // Ensure we don't leave incorrect cached data lying around
4319                 $this->mHtmlCode = null;
4320                 $this->mParentLanguage = false;
4321         }
4322
4323         /**
4324          * Get the language code from a file name. Inverse of getFileName()
4325          * @param string $filename $prefix . $languageCode . $suffix
4326          * @param string $prefix Prefix before the language code
4327          * @param string $suffix Suffix after the language code
4328          * @return string Language code, or false if $prefix or $suffix isn't found
4329          */
4330         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4331                 $m = null;
4332                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4333                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4334                 if ( !count( $m ) ) {
4335                         return false;
4336                 }
4337                 return str_replace( '_', '-', strtolower( $m[1] ) );
4338         }
4339
4340         /**
4341          * @param string $code
4342          * @param boolean $fallback Whether we're going through language fallback chain
4343          * @return string Name of the language class
4344          */
4345         public static function classFromCode( $code, $fallback = true ) {
4346                 if ( $fallback && $code == 'en' ) {
4347                         return 'Language';
4348                 } else {
4349                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4350                 }
4351         }
4352
4353         /**
4354          * Get the name of a file for a certain language code
4355          * @param string $prefix Prepend this to the filename
4356          * @param string $code Language code
4357          * @param string $suffix Append this to the filename
4358          * @throws MWException
4359          * @return string $prefix . $mangledCode . $suffix
4360          */
4361         public static function getFileName( $prefix = 'Language', $code, $suffix = '.php' ) {
4362                 if ( !self::isValidBuiltInCode( $code ) ) {
4363                         throw new MWException( "Invalid language code \"$code\"" );
4364                 }
4365
4366                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4367         }
4368
4369         /**
4370          * @param string $code
4371          * @return string
4372          */
4373         public static function getMessagesFileName( $code ) {
4374                 global $IP;
4375                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4376                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4377                 return $file;
4378         }
4379
4380         /**
4381          * @param string $code
4382          * @return string
4383          * @throws MWException
4384          * @since 1.23
4385          */
4386         public static function getJsonMessagesFileName( $code ) {
4387                 global $IP;
4388
4389                 if ( !self::isValidBuiltInCode( $code ) ) {
4390                         throw new MWException( "Invalid language code \"$code\"" );
4391                 }
4392
4393                 return "$IP/languages/i18n/$code.json";
4394         }
4395
4396         /**
4397          * Get the first fallback for a given language.
4398          *
4399          * @param string $code
4400          *
4401          * @return bool|string
4402          */
4403         public static function getFallbackFor( $code ) {
4404                 $fallbacks = self::getFallbacksFor( $code );
4405                 if ( $fallbacks ) {
4406                         return $fallbacks[0];
4407                 }
4408                 return false;
4409         }
4410
4411         /**
4412          * Get the ordered list of fallback languages.
4413          *
4414          * @since 1.19
4415          * @param string $code Language code
4416          * @return array Non-empty array, ending in "en"
4417          */
4418         public static function getFallbacksFor( $code ) {
4419                 if ( $code === 'en' || !Language::isValidBuiltInCode( $code ) ) {
4420                         return [];
4421                 }
4422                 // For unknown languages, fallbackSequence returns an empty array,
4423                 // hardcode fallback to 'en' in that case.
4424                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4425         }
4426
4427         /**
4428          * Get the ordered list of fallback languages, ending with the fallback
4429          * language chain for the site language.
4430          *
4431          * @since 1.22
4432          * @param string $code Language code
4433          * @return array Array( fallbacks, site fallbacks )
4434          */
4435         public static function getFallbacksIncludingSiteLanguage( $code ) {
4436                 global $wgLanguageCode;
4437
4438                 // Usually, we will only store a tiny number of fallback chains, so we
4439                 // keep them in static memory.
4440                 $cacheKey = "{$code}-{$wgLanguageCode}";
4441
4442                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4443                         $fallbacks = self::getFallbacksFor( $code );
4444
4445                         // Append the site's fallback chain, including the site language itself
4446                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4447                         array_unshift( $siteFallbacks, $wgLanguageCode );
4448
4449                         // Eliminate any languages already included in the chain
4450                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4451
4452                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4453                 }
4454                 return self::$fallbackLanguageCache[$cacheKey];
4455         }
4456
4457         /**
4458          * Get all messages for a given language
4459          * WARNING: this may take a long time. If you just need all message *keys*
4460          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4461          *
4462          * @param string $code
4463          *
4464          * @return array
4465          */
4466         public static function getMessagesFor( $code ) {
4467                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4468         }
4469
4470         /**
4471          * Get a message for a given language
4472          *
4473          * @param string $key
4474          * @param string $code
4475          *
4476          * @return string
4477          */
4478         public static function getMessageFor( $key, $code ) {
4479                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4480         }
4481
4482         /**
4483          * Get all message keys for a given language. This is a faster alternative to
4484          * array_keys( Language::getMessagesFor( $code ) )
4485          *
4486          * @since 1.19
4487          * @param string $code Language code
4488          * @return array Array of message keys (strings)
4489          */
4490         public static function getMessageKeysFor( $code ) {
4491                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4492         }
4493
4494         /**
4495          * @param string $talk
4496          * @return mixed
4497          */
4498         function fixVariableInNamespace( $talk ) {
4499                 if ( strpos( $talk, '$1' ) === false ) {
4500                         return $talk;
4501                 }
4502
4503                 global $wgMetaNamespace;
4504                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4505
4506                 # Allow grammar transformations
4507                 # Allowing full message-style parsing would make simple requests
4508                 # such as action=raw much more expensive than they need to be.
4509                 # This will hopefully cover most cases.
4510                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4511                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4512                 return str_replace( ' ', '_', $talk );
4513         }
4514
4515         /**
4516          * @param string $m
4517          * @return string
4518          */
4519         function replaceGrammarInNamespace( $m ) {
4520                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4521         }
4522
4523         /**
4524          * Decode an expiry (block, protection, etc) which has come from the DB
4525          *
4526          * @param string $expiry Database expiry String
4527          * @param bool|int $format True to process using language functions, or TS_ constant
4528          *     to return the expiry in a given timestamp
4529          * @param string $infinity If $format is not true, use this string for infinite expiry
4530          * @return string
4531          * @since 1.18
4532          */
4533         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4534                 static $dbInfinity;
4535                 if ( $dbInfinity === null ) {
4536                         $dbInfinity = wfGetDB( DB_SLAVE )->getInfinity();
4537                 }
4538
4539                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4540                         return $format === true
4541                                 ? $this->getMessageFromDB( 'infiniteblock' )
4542                                 : $infinity;
4543                 } else {
4544                         return $format === true
4545                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4546                                 : wfTimestamp( $format, $expiry );
4547                 }
4548         }
4549
4550         /**
4551          * Formats a time given in seconds into a string representation of that time.
4552          *
4553          * @param int|float $seconds
4554          * @param array $format An optional argument that formats the returned string in different ways:
4555          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4556          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4557          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4558          *     and friends.
4559          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4560          *     or 'avoidminutes'.
4561          * @return string
4562          */
4563         function formatTimePeriod( $seconds, $format = [] ) {
4564                 if ( !is_array( $format ) ) {
4565                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4566                 }
4567                 if ( !isset( $format['avoid'] ) ) {
4568                         $format['avoid'] = false;
4569                 }
4570                 if ( !isset( $format['noabbrevs'] ) ) {
4571                         $format['noabbrevs'] = false;
4572                 }
4573                 $secondsMsg = wfMessage(
4574                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4575                 $minutesMsg = wfMessage(
4576                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4577                 $hoursMsg = wfMessage(
4578                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4579                 $daysMsg = wfMessage(
4580                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4581
4582                 if ( round( $seconds * 10 ) < 100 ) {
4583                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4584                         $s = $secondsMsg->params( $s )->text();
4585                 } elseif ( round( $seconds ) < 60 ) {
4586                         $s = $this->formatNum( round( $seconds ) );
4587                         $s = $secondsMsg->params( $s )->text();
4588                 } elseif ( round( $seconds ) < 3600 ) {
4589                         $minutes = floor( $seconds / 60 );
4590                         $secondsPart = round( fmod( $seconds, 60 ) );
4591                         if ( $secondsPart == 60 ) {
4592                                 $secondsPart = 0;
4593                                 $minutes++;
4594                         }
4595                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4596                         $s .= ' ';
4597                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4598                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4599                         $hours = floor( $seconds / 3600 );
4600                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4601                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4602                         if ( $secondsPart == 60 ) {
4603                                 $secondsPart = 0;
4604                                 $minutes++;
4605                         }
4606                         if ( $minutes == 60 ) {
4607                                 $minutes = 0;
4608                                 $hours++;
4609                         }
4610                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4611                         $s .= ' ';
4612                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4613                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4614                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4615                         }
4616                 } else {
4617                         $days = floor( $seconds / 86400 );
4618                         if ( $format['avoid'] === 'avoidminutes' ) {
4619                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4620                                 if ( $hours == 24 ) {
4621                                         $hours = 0;
4622                                         $days++;
4623                                 }
4624                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4625                                 $s .= ' ';
4626                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4627                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4628                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4629                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4630                                 if ( $minutes == 60 ) {
4631                                         $minutes = 0;
4632                                         $hours++;
4633                                 }
4634                                 if ( $hours == 24 ) {
4635                                         $hours = 0;
4636                                         $days++;
4637                                 }
4638                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4639                                 $s .= ' ';
4640                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4641                                 $s .= ' ';
4642                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4643                         } else {
4644                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4645                                 $s .= ' ';
4646                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4647                         }
4648                 }
4649                 return $s;
4650         }
4651
4652         /**
4653          * Format a bitrate for output, using an appropriate
4654          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4655          *   the magnitude in question.
4656          *
4657          * This use base 1000. For base 1024 use formatSize(), for another base
4658          * see formatComputingNumbers().
4659          *
4660          * @param int $bps
4661          * @return string
4662          */
4663         function formatBitrate( $bps ) {
4664                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4665         }
4666
4667         /**
4668          * @param int $size Size of the unit
4669          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4670          * @param string $messageKey Message key to be uesd
4671          * @return string
4672          */
4673         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4674                 if ( $size <= 0 ) {
4675                         return str_replace( '$1', $this->formatNum( $size ),
4676                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4677                         );
4678                 }
4679                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4680                 $index = 0;
4681
4682                 $maxIndex = count( $sizes ) - 1;
4683                 while ( $size >= $boundary && $index < $maxIndex ) {
4684                         $index++;
4685                         $size /= $boundary;
4686                 }
4687
4688                 // For small sizes no decimal places necessary
4689                 $round = 0;
4690                 if ( $index > 1 ) {
4691                         // For MB and bigger two decimal places are smarter
4692                         $round = 2;
4693                 }
4694                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4695
4696                 $size = round( $size, $round );
4697                 $text = $this->getMessageFromDB( $msg );
4698                 return str_replace( '$1', $this->formatNum( $size ), $text );
4699         }
4700
4701         /**
4702          * Format a size in bytes for output, using an appropriate
4703          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4704          *
4705          * This method use base 1024. For base 1000 use formatBitrate(), for
4706          * another base see formatComputingNumbers()
4707          *
4708          * @param int $size Size to format
4709          * @return string Plain text (not HTML)
4710          */
4711         function formatSize( $size ) {
4712                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4713         }
4714
4715         /**
4716          * Make a list item, used by various special pages
4717          *
4718          * @param string $page Page link
4719          * @param string $details HTML safe text between brackets
4720          * @param bool $oppositedm Add the direction mark opposite to your
4721          *   language, to display text properly
4722          * @return HTML escaped string
4723          */
4724         function specialList( $page, $details, $oppositedm = true ) {
4725                 if ( !$details ) {
4726                         return $page;
4727                 }
4728
4729                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4730                 return
4731                         $page .
4732                         $dirmark .
4733                         $this->msg( 'word-separator' )->escaped() .
4734                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4735         }
4736
4737         /**
4738          * Generate (prev x| next x) (20|50|100...) type links for paging
4739          *
4740          * @param Title $title Title object to link
4741          * @param int $offset
4742          * @param int $limit
4743          * @param array $query Optional URL query parameter string
4744          * @param bool $atend Optional param for specified if this is the last page
4745          * @return string
4746          */
4747         public function viewPrevNext( Title $title, $offset, $limit,
4748                 array $query = [], $atend = false
4749         ) {
4750                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4751
4752                 # Make 'previous' link
4753                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4754                 if ( $offset > 0 ) {
4755                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4756                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4757                 } else {
4758                         $plink = htmlspecialchars( $prev );
4759                 }
4760
4761                 # Make 'next' link
4762                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4763                 if ( $atend ) {
4764                         $nlink = htmlspecialchars( $next );
4765                 } else {
4766                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4767                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4768                 }
4769
4770                 # Make links to set number of items per page
4771                 $numLinks = [];
4772                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4773                         $numLinks[] = $this->numLink( $title, $offset, $num,
4774                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4775                 }
4776
4777                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4778                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4779         }
4780
4781         /**
4782          * Helper function for viewPrevNext() that generates links
4783          *
4784          * @param Title $title Title object to link
4785          * @param int $offset
4786          * @param int $limit
4787          * @param array $query Extra query parameters
4788          * @param string $link Text to use for the link; will be escaped
4789          * @param string $tooltipMsg Name of the message to use as tooltip
4790          * @param string $class Value of the "class" attribute of the link
4791          * @return string HTML fragment
4792          */
4793         private function numLink( Title $title, $offset, $limit, array $query, $link,
4794                 $tooltipMsg, $class
4795         ) {
4796                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4797                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4798                         ->numParams( $limit )->text();
4799
4800                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4801                         'title' => $tooltip, 'class' => $class ], $link );
4802         }
4803
4804         /**
4805          * Get the conversion rule title, if any.
4806          *
4807          * @return string
4808          */
4809         public function getConvRuleTitle() {
4810                 return $this->mConverter->getConvRuleTitle();
4811         }
4812
4813         /**
4814          * Get the compiled plural rules for the language
4815          * @since 1.20
4816          * @return array Associative array with plural form, and plural rule as key-value pairs
4817          */
4818         public function getCompiledPluralRules() {
4819                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4820                 $fallbacks = Language::getFallbacksFor( $this->mCode );
4821                 if ( !$pluralRules ) {
4822                         foreach ( $fallbacks as $fallbackCode ) {
4823                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4824                                 if ( $pluralRules ) {
4825                                         break;
4826                                 }
4827                         }
4828                 }
4829                 return $pluralRules;
4830         }
4831
4832         /**
4833          * Get the plural rules for the language
4834          * @since 1.20
4835          * @return array Associative array with plural form number and plural rule as key-value pairs
4836          */
4837         public function getPluralRules() {
4838                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4839                 $fallbacks = Language::getFallbacksFor( $this->mCode );
4840                 if ( !$pluralRules ) {
4841                         foreach ( $fallbacks as $fallbackCode ) {
4842                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4843                                 if ( $pluralRules ) {
4844                                         break;
4845                                 }
4846                         }
4847                 }
4848                 return $pluralRules;
4849         }
4850
4851         /**
4852          * Get the plural rule types for the language
4853          * @since 1.22
4854          * @return array Associative array with plural form number and plural rule type as key-value pairs
4855          */
4856         public function getPluralRuleTypes() {
4857                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4858                 $fallbacks = Language::getFallbacksFor( $this->mCode );
4859                 if ( !$pluralRuleTypes ) {
4860                         foreach ( $fallbacks as $fallbackCode ) {
4861                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4862                                 if ( $pluralRuleTypes ) {
4863                                         break;
4864                                 }
4865                         }
4866                 }
4867                 return $pluralRuleTypes;
4868         }
4869
4870         /**
4871          * Find the index number of the plural rule appropriate for the given number
4872          * @param int $number
4873          * @return int The index number of the plural rule
4874          */
4875         public function getPluralRuleIndexNumber( $number ) {
4876                 $pluralRules = $this->getCompiledPluralRules();
4877                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4878                 return $form;
4879         }
4880
4881         /**
4882          * Find the plural rule type appropriate for the given number
4883          * For example, if the language is set to Arabic, getPluralType(5) should
4884          * return 'few'.
4885          * @since 1.22
4886          * @param int $number
4887          * @return string The name of the plural rule type, e.g. one, two, few, many
4888          */
4889         public function getPluralRuleType( $number ) {
4890                 $index = $this->getPluralRuleIndexNumber( $number );
4891                 $pluralRuleTypes = $this->getPluralRuleTypes();
4892                 if ( isset( $pluralRuleTypes[$index] ) ) {
4893                         return $pluralRuleTypes[$index];
4894                 } else {
4895                         return 'other';
4896                 }
4897         }
4898 }