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://phabricator.wikimedia.org/diffusion/GUJS/ .
 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          * @throws MWException
 182          * @return Language
 183          */
 184         static function factory( $code ) {
 185                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 186
 187                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 188                         $code = $wgDummyLanguageCodes[$code];
 189                 }
 190
 191                 // get the language object to process
 192                 $langObj = isset( self::$mLangObjCache[$code] )
 193                         ? self::$mLangObjCache[$code]
 194                         : self::newFromCode( $code );
 195
 196                 // merge the language object in to get it up front in the cache
 197                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 198                 // get rid of the oldest ones in case we have an overflow
 199                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 200
 201                 return $langObj;
 202         }
 203
 204         /**
 205          * Create a language object for a given language code
 206          * @param string $code
 207          * @param bool $fallback Whether we're going through language fallback chain
 208          * @throws MWException
 209          * @return Language
 210          */
 211         protected static function newFromCode( $code, $fallback = false ) {
 212                 if ( !self::isValidCode( $code ) ) {
 213                         throw new MWException( "Invalid language code \"$code\"" );
 214                 }
 215
 216                 if ( !self::isValidBuiltInCode( $code ) ) {
 217                         // It's not possible to customise this code with class files, so
 218                         // just return a Language object. This is to support uselang= hacks.
 219                         $lang = new Language;
 220                         $lang->setCode( $code );
 221                         return $lang;
 222                 }
 223
 224                 // Check if there is a language class for the code
 225                 $class = self::classFromCode( $code, $fallback );
 226                 if ( class_exists( $class ) ) {
 227                         $lang = new $class;
 228                         return $lang;
 229                 }
 230
 231                 // Keep trying the fallback list until we find an existing class
 232                 $fallbacks = self::getFallbacksFor( $code );
 233                 foreach ( $fallbacks as $fallbackCode ) {
 234                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 235                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 236                         }
 237
 238                         $class = self::classFromCode( $fallbackCode );
 239                         if ( class_exists( $class ) ) {
 240                                 $lang = new $class;
 241                                 $lang->setCode( $code );
 242                                 return $lang;
 243                         }
 244                 }
 245
 246                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 247         }
 248
 249         /**
 250          * Checks whether any localisation is available for that language tag
 251          * in MediaWiki (MessagesXx.php exists).
 252          *
 253          * @param string $code Language tag (in lower case)
 254          * @return bool Whether language is supported
 255          * @since 1.21
 256          */
 257         public static function isSupportedLanguage( $code ) {
 258                 if ( !self::isValidBuiltInCode( $code ) ) {
 259                         return false;
 260                 }
 261
 262                 if ( $code === 'qqq' ) {
 263                         return false;
 264                 }
 265
 266                 return is_readable( self::getMessagesFileName( $code ) ) ||
 267                         is_readable( self::getJsonMessagesFileName( $code ) );
 268         }
 269
 270         /**
 271          * Returns true if a language code string is a well-formed language tag
 272          * according to RFC 5646.
 273          * This function only checks well-formedness; it doesn't check that
 274          * language, script or variant codes actually exist in the repositories.
 275          *
 276          * Based on regexes by Mark Davis of the Unicode Consortium:
 277          * http://unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 278          *
 279          * @param string $code
 280          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 281          *
 282          * @return bool
 283          * @since 1.21
 284          */
 285         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 286                 $alpha = '[a-z]';
 287                 $digit = '[0-9]';
 288                 $alphanum = '[a-z0-9]';
 289                 $x = 'x'; # private use singleton
 290                 $singleton = '[a-wy-z]'; # other singleton
 291                 $s = $lenient ? '[-_]' : '-';
 292
 293                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 294                 $script = "$alpha{4}"; # ISO 15924
 295                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 296                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 297                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 298                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 299
 300                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 301                 # Since these are limited, this is safe even later changes to the registry --
 302                 # the only oddity is that it might change the type of the tag, and thus
 303                 # the results from the capturing groups.
 304                 # https://www.iana.org/assignments/language-subtag-registry
 305
 306                 $grandfathered = "en{$s}GB{$s}oed"
 307                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 308                         . "|no{$s}(?:bok|nyn)"
 309                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 310                         . "|zh{$s}min{$s}nan";
 311
 312                 $variantList = "$variant(?:$s$variant)*";
 313                 $extensionList = "$extension(?:$s$extension)*";
 314
 315                 $langtag = "(?:($language)"
 316                         . "(?:$s$script)?"
 317                         . "(?:$s$region)?"
 318                         . "(?:$s$variantList)?"
 319                         . "(?:$s$extensionList)?"
 320                         . "(?:$s$privateUse)?)";
 321
 322                 # The final breakdown, with capturing groups for each of these components
 323                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 324
 325                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 326
 327                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 328         }
 329
 330         /**
 331          * Returns true if a language code string is of a valid form, whether or
 332          * not it exists. This includes codes which are used solely for
 333          * customisation via the MediaWiki namespace.
 334          *
 335          * @param string $code
 336          *
 337          * @return bool
 338          */
 339         public static function isValidCode( $code ) {
 340                 static $cache = [];
 341                 if ( !isset( $cache[$code] ) ) {
 342                         // People think language codes are html safe, so enforce it.
 343                         // Ideally we should only allow a-zA-Z0-9-
 344                         // but, .+ and other chars are often used for {{int:}} hacks
 345                         // see bugs T39564, T39587, T38938
 346                         $cache[$code] =
 347                                 // Protect against path traversal
 348                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 349                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 350                 }
 351                 return $cache[$code];
 352         }
 353
 354         /**
 355          * Returns true if a language code is of a valid form for the purposes of
 356          * internal customisation of MediaWiki, via Messages*.php or *.json.
 357          *
 358          * @param string $code
 359          *
 360          * @throws MWException
 361          * @since 1.18
 362          * @return bool
 363          */
 364         public static function isValidBuiltInCode( $code ) {
 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, $wgUsePigLatinVariant;
 831
 832                 // If passed an invalid language code to use, fallback to en
 833                 if ( $inLanguage !== null && !self::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                 if ( $wgUsePigLatinVariant ) {
 846                         // Pig Latin (for variant development)
 847                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 848                 }
 849
 850                 foreach ( $mwNames as $mwCode => $mwName ) {
 851                         # - Prefer own MediaWiki native name when not using the hook
 852                         # - For other names just add if not added through the hook
 853                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 854                                 $names[$mwCode] = $mwName;
 855                         }
 856                 }
 857
 858                 if ( $include === 'all' ) {
 859                         ksort( $names );
 860                         return $names;
 861                 }
 862
 863                 $returnMw = [];
 864                 $coreCodes = array_keys( $mwNames );
 865                 foreach ( $coreCodes as $coreCode ) {
 866                         $returnMw[$coreCode] = $names[$coreCode];
 867                 }
 868
 869                 if ( $include === 'mwfile' ) {
 870                         $namesMwFile = [];
 871                         # We do this using a foreach over the codes instead of a directory
 872                         # loop so that messages files in extensions will work correctly.
 873                         foreach ( $returnMw as $code => $value ) {
 874                                 if ( is_readable( self::getMessagesFileName( $code ) )
 875                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 876                                 ) {
 877                                         $namesMwFile[$code] = $names[$code];
 878                                 }
 879                         }
 880
 881                         ksort( $namesMwFile );
 882                         return $namesMwFile;
 883                 }
 884
 885                 ksort( $returnMw );
 886                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 887                 return $returnMw;
 888         }
 889
 890         /**
 891          * @param string $code The code of the language for which to get the name
 892          * @param null|string $inLanguage Code of language in which to return the name (null for autonyms)
 893          * @param string $include 'all', 'mw' or 'mwfile'; see fetchLanguageNames()
 894          * @return string Language name or empty
 895          * @since 1.20
 896          */
 897         public static function fetchLanguageName( $code, $inLanguage = null, $include = 'all' ) {
 898                 $code = strtolower( $code );
 899                 $array = self::fetchLanguageNames( $inLanguage, $include );
 900                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 901         }
 902
 903         /**
 904          * Get a message from the MediaWiki namespace.
 905          *
 906          * @param string $msg Message name
 907          * @return string
 908          */
 909         public function getMessageFromDB( $msg ) {
 910                 return $this->msg( $msg )->text();
 911         }
 912
 913         /**
 914          * Get message object in this language. Only for use inside this class.
 915          *
 916          * @param string $msg Message name
 917          * @return Message
 918          */
 919         protected function msg( $msg ) {
 920                 return wfMessage( $msg )->inLanguage( $this );
 921         }
 922
 923         /**
 924          * @param string $key
 925          * @return string
 926          */
 927         public function getMonthName( $key ) {
 928                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 929         }
 930
 931         /**
 932          * @return array
 933          */
 934         public function getMonthNamesArray() {
 935                 $monthNames = [ '' ];
 936                 for ( $i = 1; $i < 13; $i++ ) {
 937                         $monthNames[] = $this->getMonthName( $i );
 938                 }
 939                 return $monthNames;
 940         }
 941
 942         /**
 943          * @param string $key
 944          * @return string
 945          */
 946         public function getMonthNameGen( $key ) {
 947                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 948         }
 949
 950         /**
 951          * @param string $key
 952          * @return string
 953          */
 954         public function getMonthAbbreviation( $key ) {
 955                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 956         }
 957
 958         /**
 959          * @return array
 960          */
 961         public function getMonthAbbreviationsArray() {
 962                 $monthNames = [ '' ];
 963                 for ( $i = 1; $i < 13; $i++ ) {
 964                         $monthNames[] = $this->getMonthAbbreviation( $i );
 965                 }
 966                 return $monthNames;
 967         }
 968
 969         /**
 970          * @param string $key
 971          * @return string
 972          */
 973         public function getWeekdayName( $key ) {
 974                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
 975         }
 976
 977         /**
 978          * @param string $key
 979          * @return string
 980          */
 981         function getWeekdayAbbreviation( $key ) {
 982                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
 983         }
 984
 985         /**
 986          * @param string $key
 987          * @return string
 988          */
 989         function getIranianCalendarMonthName( $key ) {
 990                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
 991         }
 992
 993         /**
 994          * @param string $key
 995          * @return string
 996          */
 997         function getHebrewCalendarMonthName( $key ) {
 998                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
 999         }
1000
1001         /**
1002          * @param string $key
1003          * @return string
1004          */
1005         function getHebrewCalendarMonthNameGen( $key ) {
1006                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1007         }
1008
1009         /**
1010          * @param string $key
1011          * @return string
1012          */
1013         function getHijriCalendarMonthName( $key ) {
1014                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1015         }
1016
1017         /**
1018          * Pass through result from $dateTimeObj->format()
1019          * @param DateTime|bool|null &$dateTimeObj
1020          * @param string $ts
1021          * @param DateTimeZone|bool|null $zone
1022          * @param string $code
1023          * @return string
1024          */
1025         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1026                 if ( !$dateTimeObj ) {
1027                         $dateTimeObj = DateTime::createFromFormat(
1028                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1029                         );
1030                 }
1031                 return $dateTimeObj->format( $code );
1032         }
1033
1034         /**
1035          * This is a workalike of PHP's date() function, but with better
1036          * internationalisation, a reduced set of format characters, and a better
1037          * escaping format.
1038          *
1039          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1040          * the PHP manual for definitions. There are a number of extensions, which
1041          * start with "x":
1042          *
1043          *    xn   Do not translate digits of the next numeric format character
1044          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1045          *    xr   Use roman numerals for the next numeric format character
1046          *    xh   Use hebrew numerals for the next numeric format character
1047          *    xx   Literal x
1048          *    xg   Genitive month name
1049          *
1050          *    xij  j (day number) in Iranian calendar
1051          *    xiF  F (month name) in Iranian calendar
1052          *    xin  n (month number) in Iranian calendar
1053          *    xiy  y (two digit year) in Iranian calendar
1054          *    xiY  Y (full year) in Iranian calendar
1055          *    xit  t (days in month) in Iranian calendar
1056          *    xiz  z (day of the year) in Iranian calendar
1057          *
1058          *    xjj  j (day number) in Hebrew calendar
1059          *    xjF  F (month name) in Hebrew calendar
1060          *    xjt  t (days in month) in Hebrew calendar
1061          *    xjx  xg (genitive month name) in Hebrew calendar
1062          *    xjn  n (month number) in Hebrew calendar
1063          *    xjY  Y (full year) in Hebrew calendar
1064          *
1065          *    xmj  j (day number) in Hijri calendar
1066          *    xmF  F (month name) in Hijri calendar
1067          *    xmn  n (month number) in Hijri calendar
1068          *    xmY  Y (full year) in Hijri calendar
1069          *
1070          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1071          *                       identical to the Gregorian calendar
1072          *    xoY  Y (full year) in Minguo calendar or Juche year.
1073          *                       Months and days are identical to the
1074          *                       Gregorian calendar
1075          *    xtY  Y (full year) in Japanese nengo. Months and days are
1076          *                       identical to the Gregorian calendar
1077          *
1078          * Characters enclosed in double quotes will be considered literal (with
1079          * the quotes themselves removed). Unmatched quotes will be considered
1080          * literal quotes. Example:
1081          *
1082          * "The month is" F       => The month is January
1083          * i's"                   => 20'11"
1084          *
1085          * Backslash escaping is also supported.
1086          *
1087          * Input timestamp is assumed to be pre-normalized to the desired local
1088          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1089          * $ts is UTC if $zone is not given.
1090          *
1091          * @param string $format
1092          * @param string $ts 14-character timestamp
1093          *      YYYYMMDDHHMMSS
1094          *      01234567890123
1095          * @param DateTimeZone $zone Timezone of $ts
1096          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1097          * Only makes sense if $ts is the current time.
1098          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1099          *
1100          * @throws MWException
1101          * @return string
1102          */
1103         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1104                 $s = '';
1105                 $raw = false;
1106                 $roman = false;
1107                 $hebrewNum = false;
1108                 $dateTimeObj = false;
1109                 $rawToggle = false;
1110                 $iranian = false;
1111                 $hebrew = false;
1112                 $hijri = false;
1113                 $thai = false;
1114                 $minguo = false;
1115                 $tenno = false;
1116
1117                 $usedSecond = false;
1118                 $usedMinute = false;
1119                 $usedHour = false;
1120                 $usedAMPM = false;
1121                 $usedDay = false;
1122                 $usedWeek = false;
1123                 $usedMonth = false;
1124                 $usedYear = false;
1125                 $usedISOYear = false;
1126                 $usedIsLeapYear = false;
1127
1128                 $usedHebrewMonth = false;
1129                 $usedIranianMonth = false;
1130                 $usedHijriMonth = false;
1131                 $usedHebrewYear = false;
1132                 $usedIranianYear = false;
1133                 $usedHijriYear = false;
1134                 $usedTennoYear = false;
1135
1136                 if ( strlen( $ts ) !== 14 ) {
1137                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1138                 }
1139
1140                 if ( !ctype_digit( $ts ) ) {
1141                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1142                 }
1143
1144                 $formatLength = strlen( $format );
1145                 for ( $p = 0; $p < $formatLength; $p++ ) {
1146                         $num = false;
1147                         $code = $format[$p];
1148                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1149                                 $code .= $format[++$p];
1150                         }
1151
1152                         if ( ( $code === 'xi'
1153                                         || $code === 'xj'
1154                                         || $code === 'xk'
1155                                         || $code === 'xm'
1156                                         || $code === 'xo'
1157                                         || $code === 'xt' )
1158                                 && $p < $formatLength - 1 ) {
1159                                 $code .= $format[++$p];
1160                         }
1161
1162                         switch ( $code ) {
1163                                 case 'xx':
1164                                         $s .= 'x';
1165                                         break;
1166                                 case 'xn':
1167                                         $raw = true;
1168                                         break;
1169                                 case 'xN':
1170                                         $rawToggle = !$rawToggle;
1171                                         break;
1172                                 case 'xr':
1173                                         $roman = true;
1174                                         break;
1175                                 case 'xh':
1176                                         $hebrewNum = true;
1177                                         break;
1178                                 case 'xg':
1179                                         $usedMonth = true;
1180                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1181                                         break;
1182                                 case 'xjx':
1183                                         $usedHebrewMonth = true;
1184                                         if ( !$hebrew ) {
1185                                                 $hebrew = self::tsToHebrew( $ts );
1186                                         }
1187                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1188                                         break;
1189                                 case 'd':
1190                                         $usedDay = true;
1191                                         $num = substr( $ts, 6, 2 );
1192                                         break;
1193                                 case 'D':
1194                                         $usedDay = true;
1195                                         $s .= $this->getWeekdayAbbreviation(
1196                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1197                                         );
1198                                         break;
1199                                 case 'j':
1200                                         $usedDay = true;
1201                                         $num = intval( substr( $ts, 6, 2 ) );
1202                                         break;
1203                                 case 'xij':
1204                                         $usedDay = true;
1205                                         if ( !$iranian ) {
1206                                                 $iranian = self::tsToIranian( $ts );
1207                                         }
1208                                         $num = $iranian[2];
1209                                         break;
1210                                 case 'xmj':
1211                                         $usedDay = true;
1212                                         if ( !$hijri ) {
1213                                                 $hijri = self::tsToHijri( $ts );
1214                                         }
1215                                         $num = $hijri[2];
1216                                         break;
1217                                 case 'xjj':
1218                                         $usedDay = true;
1219                                         if ( !$hebrew ) {
1220                                                 $hebrew = self::tsToHebrew( $ts );
1221                                         }
1222                                         $num = $hebrew[2];
1223                                         break;
1224                                 case 'l':
1225                                         $usedDay = true;
1226                                         $s .= $this->getWeekdayName(
1227                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1228                                         );
1229                                         break;
1230                                 case 'F':
1231                                         $usedMonth = true;
1232                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1233                                         break;
1234                                 case 'xiF':
1235                                         $usedIranianMonth = true;
1236                                         if ( !$iranian ) {
1237                                                 $iranian = self::tsToIranian( $ts );
1238                                         }
1239                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1240                                         break;
1241                                 case 'xmF':
1242                                         $usedHijriMonth = true;
1243                                         if ( !$hijri ) {
1244                                                 $hijri = self::tsToHijri( $ts );
1245                                         }
1246                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1247                                         break;
1248                                 case 'xjF':
1249                                         $usedHebrewMonth = true;
1250                                         if ( !$hebrew ) {
1251                                                 $hebrew = self::tsToHebrew( $ts );
1252                                         }
1253                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1254                                         break;
1255                                 case 'm':
1256                                         $usedMonth = true;
1257                                         $num = substr( $ts, 4, 2 );
1258                                         break;
1259                                 case 'M':
1260                                         $usedMonth = true;
1261                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1262                                         break;
1263                                 case 'n':
1264                                         $usedMonth = true;
1265                                         $num = intval( substr( $ts, 4, 2 ) );
1266                                         break;
1267                                 case 'xin':
1268                                         $usedIranianMonth = true;
1269                                         if ( !$iranian ) {
1270                                                 $iranian = self::tsToIranian( $ts );
1271                                         }
1272                                         $num = $iranian[1];
1273                                         break;
1274                                 case 'xmn':
1275                                         $usedHijriMonth = true;
1276                                         if ( !$hijri ) {
1277                                                 $hijri = self::tsToHijri( $ts );
1278                                         }
1279                                         $num = $hijri[1];
1280                                         break;
1281                                 case 'xjn':
1282                                         $usedHebrewMonth = true;
1283                                         if ( !$hebrew ) {
1284                                                 $hebrew = self::tsToHebrew( $ts );
1285                                         }
1286                                         $num = $hebrew[1];
1287                                         break;
1288                                 case 'xjt':
1289                                         $usedHebrewMonth = true;
1290                                         if ( !$hebrew ) {
1291                                                 $hebrew = self::tsToHebrew( $ts );
1292                                         }
1293                                         $num = $hebrew[3];
1294                                         break;
1295                                 case 'Y':
1296                                         $usedYear = true;
1297                                         $num = substr( $ts, 0, 4 );
1298                                         break;
1299                                 case 'xiY':
1300                                         $usedIranianYear = true;
1301                                         if ( !$iranian ) {
1302                                                 $iranian = self::tsToIranian( $ts );
1303                                         }
1304                                         $num = $iranian[0];
1305                                         break;
1306                                 case 'xmY':
1307                                         $usedHijriYear = true;
1308                                         if ( !$hijri ) {
1309                                                 $hijri = self::tsToHijri( $ts );
1310                                         }
1311                                         $num = $hijri[0];
1312                                         break;
1313                                 case 'xjY':
1314                                         $usedHebrewYear = true;
1315                                         if ( !$hebrew ) {
1316                                                 $hebrew = self::tsToHebrew( $ts );
1317                                         }
1318                                         $num = $hebrew[0];
1319                                         break;
1320                                 case 'xkY':
1321                                         $usedYear = true;
1322                                         if ( !$thai ) {
1323                                                 $thai = self::tsToYear( $ts, 'thai' );
1324                                         }
1325                                         $num = $thai[0];
1326                                         break;
1327                                 case 'xoY':
1328                                         $usedYear = true;
1329                                         if ( !$minguo ) {
1330                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1331                                         }
1332                                         $num = $minguo[0];
1333                                         break;
1334                                 case 'xtY':
1335                                         $usedTennoYear = true;
1336                                         if ( !$tenno ) {
1337                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1338                                         }
1339                                         $num = $tenno[0];
1340                                         break;
1341                                 case 'y':
1342                                         $usedYear = true;
1343                                         $num = substr( $ts, 2, 2 );
1344                                         break;
1345                                 case 'xiy':
1346                                         $usedIranianYear = true;
1347                                         if ( !$iranian ) {
1348                                                 $iranian = self::tsToIranian( $ts );
1349                                         }
1350                                         $num = substr( $iranian[0], -2 );
1351                                         break;
1352                                 case 'xit':
1353                                         $usedIranianYear = true;
1354                                         if ( !$iranian ) {
1355                                                 $iranian = self::tsToIranian( $ts );
1356                                         }
1357                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1358                                         break;
1359                                 case 'xiz':
1360                                         $usedIranianYear = true;
1361                                         if ( !$iranian ) {
1362                                                 $iranian = self::tsToIranian( $ts );
1363                                         }
1364                                         $num = $iranian[3];
1365                                         break;
1366                                 case 'a':
1367                                         $usedAMPM = true;
1368                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1369                                         break;
1370                                 case 'A':
1371                                         $usedAMPM = true;
1372                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1373                                         break;
1374                                 case 'g':
1375                                         $usedHour = true;
1376                                         $h = substr( $ts, 8, 2 );
1377                                         $num = $h % 12 ? $h % 12 : 12;
1378                                         break;
1379                                 case 'G':
1380                                         $usedHour = true;
1381                                         $num = intval( substr( $ts, 8, 2 ) );
1382                                         break;
1383                                 case 'h':
1384                                         $usedHour = true;
1385                                         $h = substr( $ts, 8, 2 );
1386                                         $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
1387                                         break;
1388                                 case 'H':
1389                                         $usedHour = true;
1390                                         $num = substr( $ts, 8, 2 );
1391                                         break;
1392                                 case 'i':
1393                                         $usedMinute = true;
1394                                         $num = substr( $ts, 10, 2 );
1395                                         break;
1396                                 case 's':
1397                                         $usedSecond = true;
1398                                         $num = substr( $ts, 12, 2 );
1399                                         break;
1400                                 case 'c':
1401                                 case 'r':
1402                                         $usedSecond = true;
1403                                         // fall through
1404                                 case 'e':
1405                                 case 'O':
1406                                 case 'P':
1407                                 case 'T':
1408                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1409                                         break;
1410                                 case 'w':
1411                                 case 'N':
1412                                 case 'z':
1413                                         $usedDay = true;
1414                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1415                                         break;
1416                                 case 'W':
1417                                         $usedWeek = true;
1418                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1419                                         break;
1420                                 case 't':
1421                                         $usedMonth = true;
1422                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1423                                         break;
1424                                 case 'L':
1425                                         $usedIsLeapYear = true;
1426                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1427                                         break;
1428                                 case 'o':
1429                                         $usedISOYear = true;
1430                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1431                                         break;
1432                                 case 'U':
1433                                         $usedSecond = true;
1434                                         // fall through
1435                                 case 'I':
1436                                 case 'Z':
1437                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1438                                         break;
1439                                 case '\\':
1440                                         # Backslash escaping
1441                                         if ( $p < $formatLength - 1 ) {
1442                                                 $s .= $format[++$p];
1443                                         } else {
1444                                                 $s .= '\\';
1445                                         }
1446                                         break;
1447                                 case '"':
1448                                         # Quoted literal
1449                                         if ( $p < $formatLength - 1 ) {
1450                                                 $endQuote = strpos( $format, '"', $p + 1 );
1451                                                 if ( $endQuote === false ) {
1452                                                         # No terminating quote, assume literal "
1453                                                         $s .= '"';
1454                                                 } else {
1455                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1456                                                         $p = $endQuote;
1457                                                 }
1458                                         } else {
1459                                                 # Quote at end of string, assume literal "
1460                                                 $s .= '"';
1461                                         }
1462                                         break;
1463                                 default:
1464                                         $s .= $format[$p];
1465                         }
1466                         if ( $num !== false ) {
1467                                 if ( $rawToggle || $raw ) {
1468                                         $s .= $num;
1469                                         $raw = false;
1470                                 } elseif ( $roman ) {
1471                                         $s .= self::romanNumeral( $num );
1472                                         $roman = false;
1473                                 } elseif ( $hebrewNum ) {
1474                                         $s .= self::hebrewNumeral( $num );
1475                                         $hebrewNum = false;
1476                                 } else {
1477                                         $s .= $this->formatNum( $num, true );
1478                                 }
1479                         }
1480                 }
1481
1482                 if ( $ttl === 'unused' ) {
1483                         // No need to calculate the TTL, the caller wont use it anyway.
1484                 } elseif ( $usedSecond ) {
1485                         $ttl = 1;
1486                 } elseif ( $usedMinute ) {
1487                         $ttl = 60 - substr( $ts, 12, 2 );
1488                 } elseif ( $usedHour ) {
1489                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1490                 } elseif ( $usedAMPM ) {
1491                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1492                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1493                 } elseif (
1494                         $usedDay ||
1495                         $usedHebrewMonth ||
1496                         $usedIranianMonth ||
1497                         $usedHijriMonth ||
1498                         $usedHebrewYear ||
1499                         $usedIranianYear ||
1500                         $usedHijriYear ||
1501                         $usedTennoYear
1502                 ) {
1503                         // @todo Someone who understands the non-Gregorian calendars
1504                         // should write proper logic for them so that they don't need purged every day.
1505                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1506                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1507                 } else {
1508                         $possibleTtls = [];
1509                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1510                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1511                         if ( $usedWeek ) {
1512                                 $possibleTtls[] =
1513                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1514                                         $timeRemainingInDay;
1515                         } elseif ( $usedISOYear ) {
1516                                 // December 28th falls on the last ISO week of the year, every year.
1517                                 // The last ISO week of a year can be 52 or 53.
1518                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1519                                         'Ymd',
1520                                         substr( $ts, 0, 4 ) . '1228',
1521                                         $zone ?: new DateTimeZone( 'UTC' )
1522                                 )->format( 'W' );
1523                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1524                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1525                                 $timeRemainingInWeek =
1526                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1527                                         + $timeRemainingInDay;
1528                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1529                         }
1530
1531                         if ( $usedMonth ) {
1532                                 $possibleTtls[] =
1533                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1534                                                 substr( $ts, 6, 2 ) ) * 86400
1535                                         + $timeRemainingInDay;
1536                         } elseif ( $usedYear ) {
1537                                 $possibleTtls[] =
1538                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1539                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1540                                         + $timeRemainingInDay;
1541                         } elseif ( $usedIsLeapYear ) {
1542                                 $year = substr( $ts, 0, 4 );
1543                                 $timeRemainingInYear =
1544                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1545                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1546                                         + $timeRemainingInDay;
1547                                 $mod = $year % 4;
1548                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1549                                         // this isn't a leap year. see when the next one starts
1550                                         $nextCandidate = $year - $mod + 4;
1551                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1552                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1553                                                         $timeRemainingInYear;
1554                                         } else {
1555                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1556                                                         $timeRemainingInYear;
1557                                         }
1558                                 } else {
1559                                         // this is a leap year, so the next year isn't
1560                                         $possibleTtls[] = $timeRemainingInYear;
1561                                 }
1562                         }
1563
1564                         if ( $possibleTtls ) {
1565                                 $ttl = min( $possibleTtls );
1566                         }
1567                 }
1568
1569                 return $s;
1570         }
1571
1572         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1573         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1574
1575         /**
1576          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1577          * Gregorian dates to Iranian dates. Originally written in C, it
1578          * is released under the terms of GNU Lesser General Public
1579          * License. Conversion to PHP was performed by Niklas Laxström.
1580          *
1581          * Link: http://www.farsiweb.info/jalali/jalali.c
1582          *
1583          * @param string $ts
1584          *
1585          * @return int[]
1586          */
1587         private static function tsToIranian( $ts ) {
1588                 $gy = substr( $ts, 0, 4 ) - 1600;
1589                 $gm = substr( $ts, 4, 2 ) - 1;
1590                 $gd = substr( $ts, 6, 2 ) - 1;
1591
1592                 # Days passed from the beginning (including leap years)
1593                 $gDayNo = 365 * $gy
1594                         + floor( ( $gy + 3 ) / 4 )
1595                         - floor( ( $gy + 99 ) / 100 )
1596                         + floor( ( $gy + 399 ) / 400 );
1597
1598                 // Add days of the past months of this year
1599                 for ( $i = 0; $i < $gm; $i++ ) {
1600                         $gDayNo += self::$GREG_DAYS[$i];
1601                 }
1602
1603                 // Leap years
1604                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1605                         $gDayNo++;
1606                 }
1607
1608                 // Days passed in current month
1609                 $gDayNo += (int)$gd;
1610
1611                 $jDayNo = $gDayNo - 79;
1612
1613                 $jNp = floor( $jDayNo / 12053 );
1614                 $jDayNo %= 12053;
1615
1616                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1617                 $jDayNo %= 1461;
1618
1619                 if ( $jDayNo >= 366 ) {
1620                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1621                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1622                 }
1623
1624                 $jz = $jDayNo;
1625
1626                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1627                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1628                 }
1629
1630                 $jm = $i + 1;
1631                 $jd = $jDayNo + 1;
1632
1633                 return [ $jy, $jm, $jd, $jz ];
1634         }
1635
1636         /**
1637          * Converting Gregorian dates to Hijri dates.
1638          *
1639          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1640          *
1641          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1642          *
1643          * @param string $ts
1644          *
1645          * @return int[]
1646          */
1647         private static function tsToHijri( $ts ) {
1648                 $year = substr( $ts, 0, 4 );
1649                 $month = substr( $ts, 4, 2 );
1650                 $day = substr( $ts, 6, 2 );
1651
1652                 $zyr = $year;
1653                 $zd = $day;
1654                 $zm = $month;
1655                 $zy = $zyr;
1656
1657                 if (
1658                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1659                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1660                 ) {
1661                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1662                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1663                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1664                                         $zd - 32075;
1665                 } else {
1666                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1667                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1668                 }
1669
1670                 $zl = $zjd - 1948440 + 10632;
1671                 $zn = (int)( ( $zl - 1 ) / 10631 );
1672                 $zl = $zl - 10631 * $zn + 354;
1673                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1674                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1675                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1676                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1677                 $zm = (int)( ( 24 * $zl ) / 709 );
1678                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1679                 $zy = 30 * $zn + $zj - 30;
1680
1681                 return [ $zy, $zm, $zd ];
1682         }
1683
1684         /**
1685          * Converting Gregorian dates to Hebrew dates.
1686          *
1687          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1688          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1689          * to translate the relevant functions into PHP and release them under
1690          * GNU GPL.
1691          *
1692          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1693          * and Adar II is 14. In a non-leap year, Adar is 6.
1694          *
1695          * @param string $ts
1696          *
1697          * @return int[]
1698          */
1699         private static function tsToHebrew( $ts ) {
1700                 # Parse date
1701                 $year = substr( $ts, 0, 4 );
1702                 $month = substr( $ts, 4, 2 );
1703                 $day = substr( $ts, 6, 2 );
1704
1705                 # Calculate Hebrew year
1706                 $hebrewYear = $year + 3760;
1707
1708                 # Month number when September = 1, August = 12
1709                 $month += 4;
1710                 if ( $month > 12 ) {
1711                         # Next year
1712                         $month -= 12;
1713                         $year++;
1714                         $hebrewYear++;
1715                 }
1716
1717                 # Calculate day of year from 1 September
1718                 $dayOfYear = $day;
1719                 for ( $i = 1; $i < $month; $i++ ) {
1720                         if ( $i == 6 ) {
1721                                 # February
1722                                 $dayOfYear += 28;
1723                                 # Check if the year is leap
1724                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1725                                         $dayOfYear++;
1726                                 }
1727                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1728                                 $dayOfYear += 30;
1729                         } else {
1730                                 $dayOfYear += 31;
1731                         }
1732                 }
1733
1734                 # Calculate the start of the Hebrew year
1735                 $start = self::hebrewYearStart( $hebrewYear );
1736
1737                 # Calculate next year's start
1738                 if ( $dayOfYear <= $start ) {
1739                         # Day is before the start of the year - it is the previous year
1740                         # Next year's start
1741                         $nextStart = $start;
1742                         # Previous year
1743                         $year--;
1744                         $hebrewYear--;
1745                         # Add days since previous year's 1 September
1746                         $dayOfYear += 365;
1747                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1748                                 # Leap year
1749                                 $dayOfYear++;
1750                         }
1751                         # Start of the new (previous) year
1752                         $start = self::hebrewYearStart( $hebrewYear );
1753                 } else {
1754                         # Next year's start
1755                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1756                 }
1757
1758                 # Calculate Hebrew day of year
1759                 $hebrewDayOfYear = $dayOfYear - $start;
1760
1761                 # Difference between year's days
1762                 $diff = $nextStart - $start;
1763                 # Add 12 (or 13 for leap years) days to ignore the difference between
1764                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1765                 # difference is only about the year type
1766                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1767                         $diff += 13;
1768                 } else {
1769                         $diff += 12;
1770                 }
1771
1772                 # Check the year pattern, and is leap year
1773                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1774                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1775                 # and non-leap years
1776                 $yearPattern = $diff % 30;
1777                 # Check if leap year
1778                 $isLeap = $diff >= 30;
1779
1780                 # Calculate day in the month from number of day in the Hebrew year
1781                 # Don't check Adar - if the day is not in Adar, we will stop before;
1782                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1783                 $hebrewDay = $hebrewDayOfYear;
1784                 $hebrewMonth = 1;
1785                 $days = 0;
1786                 while ( $hebrewMonth <= 12 ) {
1787                         # Calculate days in this month
1788                         if ( $isLeap && $hebrewMonth == 6 ) {
1789                                 # Adar in a leap year
1790                                 if ( $isLeap ) {
1791                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1792                                         $days = 30;
1793                                         if ( $hebrewDay <= $days ) {
1794                                                 # Day in Adar I
1795                                                 $hebrewMonth = 13;
1796                                         } else {
1797                                                 # Subtract the days of Adar I
1798                                                 $hebrewDay -= $days;
1799                                                 # Try Adar II
1800                                                 $days = 29;
1801                                                 if ( $hebrewDay <= $days ) {
1802                                                         # Day in Adar II
1803                                                         $hebrewMonth = 14;
1804                                                 }
1805                                         }
1806                                 }
1807                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1808                                 # Cheshvan in a complete year (otherwise as the rule below)
1809                                 $days = 30;
1810                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1811                                 # Kislev in an incomplete year (otherwise as the rule below)
1812                                 $days = 29;
1813                         } else {
1814                                 # Odd months have 30 days, even have 29
1815                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1816                         }
1817                         if ( $hebrewDay <= $days ) {
1818                                 # In the current month
1819                                 break;
1820                         } else {
1821                                 # Subtract the days of the current month
1822                                 $hebrewDay -= $days;
1823                                 # Try in the next month
1824                                 $hebrewMonth++;
1825                         }
1826                 }
1827
1828                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1829         }
1830
1831         /**
1832          * This calculates the Hebrew year start, as days since 1 September.
1833          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1834          * Used for Hebrew date.
1835          *
1836          * @param int $year
1837          *
1838          * @return string
1839          */
1840         private static function hebrewYearStart( $year ) {
1841                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1842                 $b = intval( ( $year - 1 ) % 4 );
1843                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1844                 if ( $m < 0 ) {
1845                         $m--;
1846                 }
1847                 $Mar = intval( $m );
1848                 if ( $m < 0 ) {
1849                         $m++;
1850                 }
1851                 $m -= $Mar;
1852
1853                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1854                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1855                         $Mar++;
1856                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1857                         $Mar += 2;
1858                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1859                         $Mar++;
1860                 }
1861
1862                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1863                 return $Mar;
1864         }
1865
1866         /**
1867          * Algorithm to convert Gregorian dates to Thai solar dates,
1868          * Minguo dates or Minguo dates.
1869          *
1870          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1871          *       https://en.wikipedia.org/wiki/Minguo_calendar
1872          *       https://en.wikipedia.org/wiki/Japanese_era_name
1873          *
1874          * @param string $ts 14-character timestamp
1875          * @param string $cName Calender name
1876          * @return array Converted year, month, day
1877          */
1878         private static function tsToYear( $ts, $cName ) {
1879                 $gy = substr( $ts, 0, 4 );
1880                 $gm = substr( $ts, 4, 2 );
1881                 $gd = substr( $ts, 6, 2 );
1882
1883                 if ( !strcmp( $cName, 'thai' ) ) {
1884                         # Thai solar dates
1885                         # Add 543 years to the Gregorian calendar
1886                         # Months and days are identical
1887                         $gy_offset = $gy + 543;
1888                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1889                         # Minguo dates
1890                         # Deduct 1911 years from the Gregorian calendar
1891                         # Months and days are identical
1892                         $gy_offset = $gy - 1911;
1893                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1894                         # Nengō dates up to Meiji period
1895                         # Deduct years from the Gregorian calendar
1896                         # depending on the nengo periods
1897                         # Months and days are identical
1898                         if ( ( $gy < 1912 )
1899                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1900                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1901                         ) {
1902                                 # Meiji period
1903                                 $gy_gannen = $gy - 1868 + 1;
1904                                 $gy_offset = $gy_gannen;
1905                                 if ( $gy_gannen == 1 ) {
1906                                         $gy_offset = '元';
1907                                 }
1908                                 $gy_offset = '明治' . $gy_offset;
1909                         } elseif (
1910                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1911                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1912                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1913                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1914                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1915                         ) {
1916                                 # Taishō period
1917                                 $gy_gannen = $gy - 1912 + 1;
1918                                 $gy_offset = $gy_gannen;
1919                                 if ( $gy_gannen == 1 ) {
1920                                         $gy_offset = '元';
1921                                 }
1922                                 $gy_offset = '大正' . $gy_offset;
1923                         } elseif (
1924                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1925                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1926                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1927                         ) {
1928                                 # Shōwa period
1929                                 $gy_gannen = $gy - 1926 + 1;
1930                                 $gy_offset = $gy_gannen;
1931                                 if ( $gy_gannen == 1 ) {
1932                                         $gy_offset = '元';
1933                                 }
1934                                 $gy_offset = '昭和' . $gy_offset;
1935                         } else {
1936                                 # Heisei period
1937                                 $gy_gannen = $gy - 1989 + 1;
1938                                 $gy_offset = $gy_gannen;
1939                                 if ( $gy_gannen == 1 ) {
1940                                         $gy_offset = '元';
1941                                 }
1942                                 $gy_offset = '平成' . $gy_offset;
1943                         }
1944                 } else {
1945                         $gy_offset = $gy;
1946                 }
1947
1948                 return [ $gy_offset, $gm, $gd ];
1949         }
1950
1951         /**
1952          * Gets directionality of the first strongly directional codepoint, for embedBidi()
1953          *
1954          * This is the rule the BIDI algorithm uses to determine the directionality of
1955          * paragraphs ( http://unicode.org/reports/tr9/#The_Paragraph_Level ) and
1956          * FSI isolates ( http://unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
1957          *
1958          * TODO: Does not handle BIDI control characters inside the text.
1959          * TODO: Does not handle unallocated characters.
1960          *
1961          * @param string $text Text to test
1962          * @return null|string Directionality ('ltr' or 'rtl') or null
1963          */
1964         private static function strongDirFromContent( $text = '' ) {
1965                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
1966                         return null;
1967                 }
1968                 if ( $matches[1] === '' ) {
1969                         return 'rtl';
1970                 }
1971                 return 'ltr';
1972         }
1973
1974         /**
1975          * Roman number formatting up to 10000
1976          *
1977          * @param int $num
1978          *
1979          * @return string
1980          */
1981         static function romanNumeral( $num ) {
1982                 static $table = [
1983                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
1984                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
1985                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
1986                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
1987                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
1988                 ];
1989
1990                 $num = intval( $num );
1991                 if ( $num > 10000 || $num <= 0 ) {
1992                         return $num;
1993                 }
1994
1995                 $s = '';
1996                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1997                         if ( $num >= $pow10 ) {
1998                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
1999                         }
2000                         $num = $num % $pow10;
2001                 }
2002                 return $s;
2003         }
2004
2005         /**
2006          * Hebrew Gematria number formatting up to 9999
2007          *
2008          * @param int $num
2009          *
2010          * @return string
2011          */
2012         static function hebrewNumeral( $num ) {
2013                 static $table = [
2014                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2015                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2016                         [ '',
2017                                 [ 'ק' ],
2018                                 [ 'ר' ],
2019                                 [ 'ש' ],
2020                                 [ 'ת' ],
2021                                 [ 'ת', 'ק' ],
2022                                 [ 'ת', 'ר' ],
2023                                 [ 'ת', 'ש' ],
2024                                 [ 'ת', 'ת' ],
2025                                 [ 'ת', 'ת', 'ק' ],
2026                                 [ 'ת', 'ת', 'ר' ],
2027                         ],
2028                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2029                 ];
2030
2031                 $num = intval( $num );
2032                 if ( $num > 9999 || $num <= 0 ) {
2033                         return $num;
2034                 }
2035
2036                 // Round thousands have special notations
2037                 if ( $num === 1000 ) {
2038                         return "א' אלף";
2039                 } elseif ( $num % 1000 === 0 ) {
2040                         return $table[0][$num / 1000] . "' אלפים";
2041                 }
2042
2043                 $letters = [];
2044
2045                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2046                         if ( $num >= $pow10 ) {
2047                                 if ( $num === 15 || $num === 16 ) {
2048                                         $letters[] = $table[0][9];
2049                                         $letters[] = $table[0][$num - 9];
2050                                         $num = 0;
2051                                 } else {
2052                                         $letters = array_merge(
2053                                                 $letters,
2054                                                 (array)$table[$i][intval( $num / $pow10 )]
2055                                         );
2056
2057                                         if ( $pow10 === 1000 ) {
2058                                                 $letters[] = "'";
2059                                         }
2060                                 }
2061                         }
2062
2063                         $num = $num % $pow10;
2064                 }
2065
2066                 $preTransformLength = count( $letters );
2067                 if ( $preTransformLength === 1 ) {
2068                         // Add geresh (single quote) to one-letter numbers
2069                         $letters[] = "'";
2070                 } else {
2071                         $lastIndex = $preTransformLength - 1;
2072                         $letters[$lastIndex] = str_replace(
2073                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2074                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2075                                 $letters[$lastIndex]
2076                         );
2077
2078                         // Add gershayim (double quote) to multiple-letter numbers,
2079                         // but exclude numbers with only one letter after the thousands
2080                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2081                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2082                                 $letters[] = "'";
2083                         } else {
2084                                 array_splice( $letters, -1, 0, '"' );
2085                         }
2086                 }
2087
2088                 return implode( $letters );
2089         }
2090
2091         /**
2092          * Used by date() and time() to adjust the time output.
2093          *
2094          * @param string $ts The time in date('YmdHis') format
2095          * @param mixed $tz Adjust the time by this amount (default false, mean we
2096          *   get user timecorrection setting)
2097          * @return int
2098          */
2099         public function userAdjust( $ts, $tz = false ) {
2100                 global $wgUser, $wgLocalTZoffset;
2101
2102                 if ( $tz === false ) {
2103                         $tz = $wgUser->getOption( 'timecorrection' );
2104                 }
2105
2106                 $data = explode( '|', $tz, 3 );
2107
2108                 if ( $data[0] == 'ZoneInfo' ) {
2109                         try {
2110                                 $userTZ = new DateTimeZone( $data[2] );
2111                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2112                                 $date->setTimezone( $userTZ );
2113                                 return $date->format( 'YmdHis' );
2114                         } catch ( Exception $e ) {
2115                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2116                                 $data[0] = 'Offset';
2117                         }
2118                 }
2119
2120                 if ( $data[0] == 'System' || $tz == '' ) {
2121                         # Global offset in minutes.
2122                         $minDiff = $wgLocalTZoffset;
2123                 } elseif ( $data[0] == 'Offset' ) {
2124                         $minDiff = intval( $data[1] );
2125                 } else {
2126                         $data = explode( ':', $tz );
2127                         if ( count( $data ) == 2 ) {
2128                                 $data[0] = intval( $data[0] );
2129                                 $data[1] = intval( $data[1] );
2130                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2131                                 if ( $data[0] < 0 ) {
2132                                         $minDiff = -$minDiff;
2133                                 }
2134                         } else {
2135                                 $minDiff = intval( $data[0] ) * 60;
2136                         }
2137                 }
2138
2139                 # No difference ? Return time unchanged
2140                 if ( 0 == $minDiff ) {
2141                         return $ts;
2142                 }
2143
2144                 MediaWiki\suppressWarnings(); // E_STRICT system time bitching
2145                 # Generate an adjusted date; take advantage of the fact that mktime
2146                 # will normalize out-of-range values so we don't have to split $minDiff
2147                 # into hours and minutes.
2148                 $t = mktime( (
2149                         (int)substr( $ts, 8, 2 ) ), # Hours
2150                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2151                         (int)substr( $ts, 12, 2 ), # Seconds
2152                         (int)substr( $ts, 4, 2 ), # Month
2153                         (int)substr( $ts, 6, 2 ), # Day
2154                         (int)substr( $ts, 0, 4 ) ); # Year
2155
2156                 $date = date( 'YmdHis', $t );
2157                 MediaWiki\restoreWarnings();
2158
2159                 return $date;
2160         }
2161
2162         /**
2163          * This is meant to be used by time(), date(), and timeanddate() to get
2164          * the date preference they're supposed to use, it should be used in
2165          * all children.
2166          *
2167          *     function timeanddate([...], $format = true) {
2168          *       $datePreference = $this->dateFormat($format);
2169          *       [...]
2170          *     }
2171          *
2172          * @param int|string|bool $usePrefs If true, the user's preference is used
2173          *   if false, the site/language default is used
2174          *   if int/string, assumed to be a format.
2175          * @return string
2176          */
2177         function dateFormat( $usePrefs = true ) {
2178                 global $wgUser;
2179
2180                 if ( is_bool( $usePrefs ) ) {
2181                         if ( $usePrefs ) {
2182                                 $datePreference = $wgUser->getDatePreference();
2183                         } else {
2184                                 $datePreference = (string)User::getDefaultOption( 'date' );
2185                         }
2186                 } else {
2187                         $datePreference = (string)$usePrefs;
2188                 }
2189
2190                 // return int
2191                 if ( $datePreference == '' ) {
2192                         return 'default';
2193                 }
2194
2195                 return $datePreference;
2196         }
2197
2198         /**
2199          * Get a format string for a given type and preference
2200          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2201          * @param string $pref The format name as it appears in Messages*.php under
2202          *  $datePreferences.
2203          *
2204          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2205          *
2206          * @return string
2207          */
2208         function getDateFormatString( $type, $pref ) {
2209                 $wasDefault = false;
2210                 if ( $pref == 'default' ) {
2211                         $wasDefault = true;
2212                         $pref = $this->getDefaultDateFormat();
2213                 }
2214
2215                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2216                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2217
2218                         if ( $type === 'pretty' && $df === null ) {
2219                                 $df = $this->getDateFormatString( 'date', $pref );
2220                         }
2221
2222                         if ( !$wasDefault && $df === null ) {
2223                                 $pref = $this->getDefaultDateFormat();
2224                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2225                         }
2226
2227                         $this->dateFormatStrings[$type][$pref] = $df;
2228                 }
2229                 return $this->dateFormatStrings[$type][$pref];
2230         }
2231
2232         /**
2233          * @param string $ts The time format which needs to be turned into a
2234          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2235          * @param bool $adj Whether to adjust the time output according to the
2236          *   user configured offset ($timecorrection)
2237          * @param mixed $format True to use user's date format preference
2238          * @param string|bool $timecorrection The time offset as returned by
2239          *   validateTimeZone() in Special:Preferences
2240          * @return string
2241          */
2242         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2243                 $ts = wfTimestamp( TS_MW, $ts );
2244                 if ( $adj ) {
2245                         $ts = $this->userAdjust( $ts, $timecorrection );
2246                 }
2247                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2248                 return $this->sprintfDate( $df, $ts );
2249         }
2250
2251         /**
2252          * @param string $ts The time format which needs to be turned into a
2253          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2254          * @param bool $adj Whether to adjust the time output according to the
2255          *   user configured offset ($timecorrection)
2256          * @param mixed $format True to use user's date format preference
2257          * @param string|bool $timecorrection The time offset as returned by
2258          *   validateTimeZone() in Special:Preferences
2259          * @return string
2260          */
2261         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2262                 $ts = wfTimestamp( TS_MW, $ts );
2263                 if ( $adj ) {
2264                         $ts = $this->userAdjust( $ts, $timecorrection );
2265                 }
2266                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2267                 return $this->sprintfDate( $df, $ts );
2268         }
2269
2270         /**
2271          * @param string $ts The time format which needs to be turned into a
2272          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2273          * @param bool $adj Whether to adjust the time output according to the
2274          *   user configured offset ($timecorrection)
2275          * @param mixed $format What format to return, if it's false output the
2276          *   default one (default true)
2277          * @param string|bool $timecorrection The time offset as returned by
2278          *   validateTimeZone() in Special:Preferences
2279          * @return string
2280          */
2281         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2282                 $ts = wfTimestamp( TS_MW, $ts );
2283                 if ( $adj ) {
2284                         $ts = $this->userAdjust( $ts, $timecorrection );
2285                 }
2286                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2287                 return $this->sprintfDate( $df, $ts );
2288         }
2289
2290         /**
2291          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2292          *
2293          * @since 1.20
2294          *
2295          * @param int $seconds The amount of seconds.
2296          * @param array $chosenIntervals The intervals to enable.
2297          *
2298          * @return string
2299          */
2300         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2301                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2302
2303                 $segments = [];
2304
2305                 foreach ( $intervals as $intervalName => $intervalValue ) {
2306                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2307                         // duration-years, duration-decades, duration-centuries, duration-millennia
2308                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2309                         $segments[] = $message->inLanguage( $this )->escaped();
2310                 }
2311
2312                 return $this->listToText( $segments );
2313         }
2314
2315         /**
2316          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2317          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2318          *
2319          * @since 1.20
2320          *
2321          * @param int $seconds The amount of seconds.
2322          * @param array $chosenIntervals The intervals to enable.
2323          *
2324          * @return array
2325          */
2326         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2327                 if ( empty( $chosenIntervals ) ) {
2328                         $chosenIntervals = [
2329                                 'millennia',
2330                                 'centuries',
2331                                 'decades',
2332                                 'years',
2333                                 'days',
2334                                 'hours',
2335                                 'minutes',
2336                                 'seconds'
2337                         ];
2338                 }
2339
2340                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2341                 $sortedNames = array_keys( $intervals );
2342                 $smallestInterval = array_pop( $sortedNames );
2343
2344                 $segments = [];
2345
2346                 foreach ( $intervals as $name => $length ) {
2347                         $value = floor( $seconds / $length );
2348
2349                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2350                                 $seconds -= $value * $length;
2351                                 $segments[$name] = $value;
2352                         }
2353                 }
2354
2355                 return $segments;
2356         }
2357
2358         /**
2359          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2360          *
2361          * @param string $type Can be 'date', 'time' or 'both'
2362          * @param string $ts The time format which needs to be turned into a
2363          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2364          * @param User $user User object used to get preferences for timezone and format
2365          * @param array $options Array, can contain the following keys:
2366          *   - 'timecorrection': time correction, can have the following values:
2367          *     - true: use user's preference
2368          *     - false: don't use time correction
2369          *     - int: value of time correction in minutes
2370          *   - 'format': format to use, can have the following values:
2371          *     - true: use user's preference
2372          *     - false: use default preference
2373          *     - string: format to use
2374          * @since 1.19
2375          * @return string
2376          */
2377         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2378                 $ts = wfTimestamp( TS_MW, $ts );
2379                 $options += [ 'timecorrection' => true, 'format' => true ];
2380                 if ( $options['timecorrection'] !== false ) {
2381                         if ( $options['timecorrection'] === true ) {
2382                                 $offset = $user->getOption( 'timecorrection' );
2383                         } else {
2384                                 $offset = $options['timecorrection'];
2385                         }
2386                         $ts = $this->userAdjust( $ts, $offset );
2387                 }
2388                 if ( $options['format'] === true ) {
2389                         $format = $user->getDatePreference();
2390                 } else {
2391                         $format = $options['format'];
2392                 }
2393                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2394                 return $this->sprintfDate( $df, $ts );
2395         }
2396
2397         /**
2398          * Get the formatted date for the given timestamp and formatted for
2399          * the given user.
2400          *
2401          * @param mixed $ts Mixed: the time format which needs to be turned into a
2402          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2403          * @param User $user User object used to get preferences for timezone and format
2404          * @param array $options Array, can contain the following keys:
2405          *   - 'timecorrection': time correction, can have the following values:
2406          *     - true: use user's preference
2407          *     - false: don't use time correction
2408          *     - int: value of time correction in minutes
2409          *   - 'format': format to use, can have the following values:
2410          *     - true: use user's preference
2411          *     - false: use default preference
2412          *     - string: format to use
2413          * @since 1.19
2414          * @return string
2415          */
2416         public function userDate( $ts, User $user, array $options = [] ) {
2417                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2418         }
2419
2420         /**
2421          * Get the formatted time for the given timestamp and formatted for
2422          * the given user.
2423          *
2424          * @param mixed $ts The time format which needs to be turned into a
2425          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2426          * @param User $user User object used to get preferences for timezone and format
2427          * @param array $options Array, can contain the following keys:
2428          *   - 'timecorrection': time correction, can have the following values:
2429          *     - true: use user's preference
2430          *     - false: don't use time correction
2431          *     - int: value of time correction in minutes
2432          *   - 'format': format to use, can have the following values:
2433          *     - true: use user's preference
2434          *     - false: use default preference
2435          *     - string: format to use
2436          * @since 1.19
2437          * @return string
2438          */
2439         public function userTime( $ts, User $user, array $options = [] ) {
2440                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2441         }
2442
2443         /**
2444          * Get the formatted date and time for the given timestamp and formatted for
2445          * the given user.
2446          *
2447          * @param mixed $ts The time format which needs to be turned into a
2448          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2449          * @param User $user User object used to get preferences for timezone and format
2450          * @param array $options Array, can contain the following keys:
2451          *   - 'timecorrection': time correction, can have the following values:
2452          *     - true: use user's preference
2453          *     - false: don't use time correction
2454          *     - int: value of time correction in minutes
2455          *   - 'format': format to use, can have the following values:
2456          *     - true: use user's preference
2457          *     - false: use default preference
2458          *     - string: format to use
2459          * @since 1.19
2460          * @return string
2461          */
2462         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2463                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2464         }
2465
2466         /**
2467          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2468          *
2469          * Determine the difference between the timestamp and the current time, and
2470          * generate a readable timestamp by returning "<N> <units> ago", where the
2471          * largest possible unit is used.
2472          *
2473          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2474          *
2475          * @param MWTimestamp $time
2476          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2477          * @param User|null $user User the timestamp is being generated for
2478          *  (or null to use main context's user)
2479          * @return string Formatted timestamp
2480          */
2481         public function getHumanTimestamp(
2482                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2483         ) {
2484                 if ( $relativeTo === null ) {
2485                         $relativeTo = new MWTimestamp();
2486                 }
2487                 if ( $user === null ) {
2488                         $user = RequestContext::getMain()->getUser();
2489                 }
2490
2491                 // Adjust for the user's timezone.
2492                 $offsetThis = $time->offsetForUser( $user );
2493                 $offsetRel = $relativeTo->offsetForUser( $user );
2494
2495                 $ts = '';
2496                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2497                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2498                 }
2499
2500                 // Reset the timezone on the objects.
2501                 $time->timestamp->sub( $offsetThis );
2502                 $relativeTo->timestamp->sub( $offsetRel );
2503
2504                 return $ts;
2505         }
2506
2507         /**
2508          * Convert an MWTimestamp into a pretty human-readable timestamp using
2509          * the given user preferences and relative base time.
2510          *
2511          * @see Language::getHumanTimestamp
2512          * @param MWTimestamp $ts Timestamp to prettify
2513          * @param MWTimestamp $relativeTo Base timestamp
2514          * @param User $user User preferences to use
2515          * @return string Human timestamp
2516          * @since 1.26
2517          */
2518         private function getHumanTimestampInternal(
2519                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2520         ) {
2521                 $diff = $ts->diff( $relativeTo );
2522                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2523                         (int)$relativeTo->timestamp->format( 'w' ) );
2524                 $days = $diff->days ?: (int)$diffDay;
2525                 if ( $diff->invert || $days > 5
2526                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2527                 ) {
2528                         // Timestamps are in different years: use full timestamp
2529                         // Also do full timestamp for future dates
2530                         /**
2531                          * @todo FIXME: Add better handling of future timestamps.
2532                          */
2533                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2534                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2535                 } elseif ( $days > 5 ) {
2536                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2537                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2538                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2539                 } elseif ( $days > 1 ) {
2540                         // Timestamp within the past week: show the day of the week and time
2541                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2542                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2543                         // Messages:
2544                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2545                         $ts = wfMessage( "$weekday-at" )
2546                                 ->inLanguage( $this )
2547                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2548                                 ->text();
2549                 } elseif ( $days == 1 ) {
2550                         // Timestamp was yesterday: say 'yesterday' and the time.
2551                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2552                         $ts = wfMessage( 'yesterday-at' )
2553                                 ->inLanguage( $this )
2554                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2555                                 ->text();
2556                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2557                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2558                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2559                         $ts = wfMessage( 'today-at' )
2560                                 ->inLanguage( $this )
2561                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2562                                 ->text();
2563
2564                 // From here on in, the timestamp was soon enough ago so that we can simply say
2565                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2566                 } elseif ( $diff->h == 1 ) {
2567                         // Less than 90 minutes, but more than an hour ago.
2568                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2569                 } elseif ( $diff->i >= 1 ) {
2570                         // A few minutes ago.
2571                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2572                 } elseif ( $diff->s >= 30 ) {
2573                         // Less than a minute, but more than 30 sec ago.
2574                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2575                 } else {
2576                         // Less than 30 seconds ago.
2577                         $ts = wfMessage( 'just-now' )->text();
2578                 }
2579
2580                 return $ts;
2581         }
2582
2583         /**
2584          * @param string $key
2585          * @return string|null
2586          */
2587         public function getMessage( $key ) {
2588                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2589         }
2590
2591         /**
2592          * @return array
2593          */
2594         function getAllMessages() {
2595                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2596         }
2597
2598         /**
2599          * @param string $in
2600          * @param string $out
2601          * @param string $string
2602          * @return string
2603          */
2604         public function iconv( $in, $out, $string ) {
2605                 # Even with //IGNORE iconv can whine about illegal characters in
2606                 # *input* string. We just ignore those too.
2607                 # REF: https://bugs.php.net/bug.php?id=37166
2608                 # REF: https://phabricator.wikimedia.org/T18885
2609                 MediaWiki\suppressWarnings();
2610                 $text = iconv( $in, $out . '//IGNORE', $string );
2611                 MediaWiki\restoreWarnings();
2612                 return $text;
2613         }
2614
2615         // callback functions for ucwords(), ucwordbreaks()
2616
2617         /**
2618          * @param array $matches
2619          * @return mixed|string
2620          */
2621         function ucwordbreaksCallbackAscii( $matches ) {
2622                 return $this->ucfirst( $matches[1] );
2623         }
2624
2625         /**
2626          * @param array $matches
2627          * @return string
2628          */
2629         function ucwordbreaksCallbackMB( $matches ) {
2630                 return mb_strtoupper( $matches[0] );
2631         }
2632
2633         /**
2634          * @param array $matches
2635          * @return string
2636          */
2637         function ucwordsCallbackMB( $matches ) {
2638                 return mb_strtoupper( $matches[0] );
2639         }
2640
2641         /**
2642          * Make a string's first character uppercase
2643          *
2644          * @param string $str
2645          *
2646          * @return string
2647          */
2648         public function ucfirst( $str ) {
2649                 $o = ord( $str );
2650                 if ( $o < 96 ) { // if already uppercase...
2651                         return $str;
2652                 } elseif ( $o < 128 ) {
2653                         return ucfirst( $str ); // use PHP's ucfirst()
2654                 } else {
2655                         // fall back to more complex logic in case of multibyte strings
2656                         return $this->uc( $str, true );
2657                 }
2658         }
2659
2660         /**
2661          * Convert a string to uppercase
2662          *
2663          * @param string $str
2664          * @param bool $first
2665          *
2666          * @return string
2667          */
2668         public function uc( $str, $first = false ) {
2669                 if ( $first ) {
2670                         if ( $this->isMultibyte( $str ) ) {
2671                                 return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2672                         } else {
2673                                 return ucfirst( $str );
2674                         }
2675                 } else {
2676                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2677                 }
2678         }
2679
2680         /**
2681          * @param string $str
2682          * @return mixed|string
2683          */
2684         function lcfirst( $str ) {
2685                 $o = ord( $str );
2686                 if ( !$o ) {
2687                         return strval( $str );
2688                 } elseif ( $o >= 128 ) {
2689                         return $this->lc( $str, true );
2690                 } elseif ( $o > 96 ) {
2691                         return $str;
2692                 } else {
2693                         $str[0] = strtolower( $str[0] );
2694                         return $str;
2695                 }
2696         }
2697
2698         /**
2699          * @param string $str
2700          * @param bool $first
2701          * @return mixed|string
2702          */
2703         function lc( $str, $first = false ) {
2704                 if ( $first ) {
2705                         if ( $this->isMultibyte( $str ) ) {
2706                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2707                         } else {
2708                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2709                         }
2710                 } else {
2711                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2712                 }
2713         }
2714
2715         /**
2716          * @param string $str
2717          * @return bool
2718          */
2719         function isMultibyte( $str ) {
2720                 return strlen( $str ) !== mb_strlen( $str );
2721         }
2722
2723         /**
2724          * @param string $str
2725          * @return mixed|string
2726          */
2727         function ucwords( $str ) {
2728                 if ( $this->isMultibyte( $str ) ) {
2729                         $str = $this->lc( $str );
2730
2731                         // regexp to find first letter in each word (i.e. after each space)
2732                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2733
2734                         // function to use to capitalize a single char
2735                         return preg_replace_callback(
2736                                 $replaceRegexp,
2737                                 [ $this, 'ucwordsCallbackMB' ],
2738                                 $str
2739                         );
2740                 } else {
2741                         return ucwords( strtolower( $str ) );
2742                 }
2743         }
2744
2745         /**
2746          * capitalize words at word breaks
2747          *
2748          * @param string $str
2749          * @return mixed
2750          */
2751         function ucwordbreaks( $str ) {
2752                 if ( $this->isMultibyte( $str ) ) {
2753                         $str = $this->lc( $str );
2754
2755                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2756                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2757
2758                         // find first letter after word break
2759                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2760                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2761
2762                         return preg_replace_callback(
2763                                 $replaceRegexp,
2764                                 [ $this, 'ucwordbreaksCallbackMB' ],
2765                                 $str
2766                         );
2767                 } else {
2768                         return preg_replace_callback(
2769                                 '/\b([\w\x80-\xff]+)\b/',
2770                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2771                                 $str
2772                         );
2773                 }
2774         }
2775
2776         /**
2777          * Return a case-folded representation of $s
2778          *
2779          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2780          * and $s2 are the same except for the case of their characters. It is not
2781          * necessary for the value returned to make sense when displayed.
2782          *
2783          * Do *not* perform any other normalisation in this function. If a caller
2784          * uses this function when it should be using a more general normalisation
2785          * function, then fix the caller.
2786          *
2787          * @param string $s
2788          *
2789          * @return string
2790          */
2791         function caseFold( $s ) {
2792                 return $this->uc( $s );
2793         }
2794
2795         /**
2796          * @param string $s
2797          * @return string
2798          * @throws MWException
2799          */
2800         function checkTitleEncoding( $s ) {
2801                 if ( is_array( $s ) ) {
2802                         throw new MWException( 'Given array to checkTitleEncoding.' );
2803                 }
2804                 if ( StringUtils::isUtf8( $s ) ) {
2805                         return $s;
2806                 }
2807
2808                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2809         }
2810
2811         /**
2812          * @return array
2813          */
2814         function fallback8bitEncoding() {
2815                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2816         }
2817
2818         /**
2819          * Most writing systems use whitespace to break up words.
2820          * Some languages such as Chinese don't conventionally do this,
2821          * which requires special handling when breaking up words for
2822          * searching etc.
2823          *
2824          * @return bool
2825          */
2826         function hasWordBreaks() {
2827                 return true;
2828         }
2829
2830         /**
2831          * Some languages such as Chinese require word segmentation,
2832          * Specify such segmentation when overridden in derived class.
2833          *
2834          * @param string $string
2835          * @return string
2836          */
2837         function segmentByWord( $string ) {
2838                 return $string;
2839         }
2840
2841         /**
2842          * Some languages have special punctuation need to be normalized.
2843          * Make such changes here.
2844          *
2845          * @param string $string
2846          * @return string
2847          */
2848         function normalizeForSearch( $string ) {
2849                 return self::convertDoubleWidth( $string );
2850         }
2851
2852         /**
2853          * convert double-width roman characters to single-width.
2854          * range: ff00-ff5f ~= 0020-007f
2855          *
2856          * @param string $string
2857          *
2858          * @return string
2859          */
2860         protected static function convertDoubleWidth( $string ) {
2861                 static $full = null;
2862                 static $half = null;
2863
2864                 if ( $full === null ) {
2865                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2866                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2867                         $full = str_split( $fullWidth, 3 );
2868                         $half = str_split( $halfWidth );
2869                 }
2870
2871                 $string = str_replace( $full, $half, $string );
2872                 return $string;
2873         }
2874
2875         /**
2876          * @param string $string
2877          * @param string $pattern
2878          * @return string
2879          */
2880         protected static function insertSpace( $string, $pattern ) {
2881                 $string = preg_replace( $pattern, " $1 ", $string );
2882                 $string = preg_replace( '/ +/', ' ', $string );
2883                 return $string;
2884         }
2885
2886         /**
2887          * @param array $termsArray
2888          * @return array
2889          */
2890         function convertForSearchResult( $termsArray ) {
2891                 # some languages, e.g. Chinese, need to do a conversion
2892                 # in order for search results to be displayed correctly
2893                 return $termsArray;
2894         }
2895
2896         /**
2897          * Get the first character of a string.
2898          *
2899          * @param string $s
2900          * @return string
2901          */
2902         function firstChar( $s ) {
2903                 $matches = [];
2904                 preg_match(
2905                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2906                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2907                         $s,
2908                         $matches
2909                 );
2910
2911                 if ( isset( $matches[1] ) ) {
2912                         if ( strlen( $matches[1] ) != 3 ) {
2913                                 return $matches[1];
2914                         }
2915
2916                         // Break down Hangul syllables to grab the first jamo
2917                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2918                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2919                                 return $matches[1];
2920                         } elseif ( $code < 0xb098 ) {
2921                                 return "\xe3\x84\xb1";
2922                         } elseif ( $code < 0xb2e4 ) {
2923                                 return "\xe3\x84\xb4";
2924                         } elseif ( $code < 0xb77c ) {
2925                                 return "\xe3\x84\xb7";
2926                         } elseif ( $code < 0xb9c8 ) {
2927                                 return "\xe3\x84\xb9";
2928                         } elseif ( $code < 0xbc14 ) {
2929                                 return "\xe3\x85\x81";
2930                         } elseif ( $code < 0xc0ac ) {
2931                                 return "\xe3\x85\x82";
2932                         } elseif ( $code < 0xc544 ) {
2933                                 return "\xe3\x85\x85";
2934                         } elseif ( $code < 0xc790 ) {
2935                                 return "\xe3\x85\x87";
2936                         } elseif ( $code < 0xcc28 ) {
2937                                 return "\xe3\x85\x88";
2938                         } elseif ( $code < 0xce74 ) {
2939                                 return "\xe3\x85\x8a";
2940                         } elseif ( $code < 0xd0c0 ) {
2941                                 return "\xe3\x85\x8b";
2942                         } elseif ( $code < 0xd30c ) {
2943                                 return "\xe3\x85\x8c";
2944                         } elseif ( $code < 0xd558 ) {
2945                                 return "\xe3\x85\x8d";
2946                         } else {
2947                                 return "\xe3\x85\x8e";
2948                         }
2949                 } else {
2950                         return '';
2951                 }
2952         }
2953
2954         /**
2955          * @deprecated No-op since 1.28
2956          */
2957         function initEncoding() {
2958                 // No-op.
2959         }
2960
2961         /**
2962          * @param string $s
2963          * @return string
2964          * @deprecated No-op since 1.28
2965          */
2966         function recodeForEdit( $s ) {
2967                 return $s;
2968         }
2969
2970         /**
2971          * @param string $s
2972          * @return string
2973          * @deprecated No-op since 1.28
2974          */
2975         function recodeInput( $s ) {
2976                 return $s;
2977         }
2978
2979         /**
2980          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
2981          * also cleans up certain backwards-compatible sequences, converting them
2982          * to the modern Unicode equivalent.
2983          *
2984          * This is language-specific for performance reasons only.
2985          *
2986          * @param string $s
2987          *
2988          * @return string
2989          */
2990         function normalize( $s ) {
2991                 global $wgAllUnicodeFixes;
2992                 $s = UtfNormal\Validator::cleanUp( $s );
2993                 if ( $wgAllUnicodeFixes ) {
2994                         $s = $this->transformUsingPairFile( 'normalize-ar.ser', $s );
2995                         $s = $this->transformUsingPairFile( 'normalize-ml.ser', $s );
2996                 }
2997
2998                 return $s;
2999         }
3000
3001         /**
3002          * Transform a string using serialized data stored in the given file (which
3003          * must be in the serialized subdirectory of $IP). The file contains pairs
3004          * mapping source characters to destination characters.
3005          *
3006          * The data is cached in process memory. This will go faster if you have the
3007          * FastStringSearch extension.
3008          *
3009          * @param string $file
3010          * @param string $string
3011          *
3012          * @throws MWException
3013          * @return string
3014          */
3015         function transformUsingPairFile( $file, $string ) {
3016                 if ( !isset( $this->transformData[$file] ) ) {
3017                         $data = wfGetPrecompiledData( $file );
3018                         if ( $data === false ) {
3019                                 throw new MWException( __METHOD__ . ": The transformation file $file is missing" );
3020                         }
3021                         $this->transformData[$file] = new ReplacementArray( $data );
3022                 }
3023                 return $this->transformData[$file]->replace( $string );
3024         }
3025
3026         /**
3027          * For right-to-left language support
3028          *
3029          * @return bool
3030          */
3031         function isRTL() {
3032                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3033         }
3034
3035         /**
3036          * Return the correct HTML 'dir' attribute value for this language.
3037          * @return string
3038          */
3039         function getDir() {
3040                 return $this->isRTL() ? 'rtl' : 'ltr';
3041         }
3042
3043         /**
3044          * Return 'left' or 'right' as appropriate alignment for line-start
3045          * for this language's text direction.
3046          *
3047          * Should be equivalent to CSS3 'start' text-align value....
3048          *
3049          * @return string
3050          */
3051         function alignStart() {
3052                 return $this->isRTL() ? 'right' : 'left';
3053         }
3054
3055         /**
3056          * Return 'right' or 'left' as appropriate alignment for line-end
3057          * for this language's text direction.
3058          *
3059          * Should be equivalent to CSS3 'end' text-align value....
3060          *
3061          * @return string
3062          */
3063         function alignEnd() {
3064                 return $this->isRTL() ? 'left' : 'right';
3065         }
3066
3067         /**
3068          * A hidden direction mark (LRM or RLM), depending on the language direction.
3069          * Unlike getDirMark(), this function returns the character as an HTML entity.
3070          * This function should be used when the output is guaranteed to be HTML,
3071          * because it makes the output HTML source code more readable. When
3072          * the output is plain text or can be escaped, getDirMark() should be used.
3073          *
3074          * @param bool $opposite Get the direction mark opposite to your language
3075          * @return string
3076          * @since 1.20
3077          */
3078         function getDirMarkEntity( $opposite = false ) {
3079                 if ( $opposite ) {
3080                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3081                 }
3082                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3083         }
3084
3085         /**
3086          * A hidden direction mark (LRM or RLM), depending on the language direction.
3087          * This function produces them as invisible Unicode characters and
3088          * the output may be hard to read and debug, so it should only be used
3089          * when the output is plain text or can be escaped. When the output is
3090          * HTML, use getDirMarkEntity() instead.
3091          *
3092          * @param bool $opposite Get the direction mark opposite to your language
3093          * @return string
3094          */
3095         function getDirMark( $opposite = false ) {
3096                 $lrm = "\xE2\x80\x8E"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3097                 $rlm = "\xE2\x80\x8F"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3098                 if ( $opposite ) {
3099                         return $this->isRTL() ? $lrm : $rlm;
3100                 }
3101                 return $this->isRTL() ? $rlm : $lrm;
3102         }
3103
3104         /**
3105          * @return array
3106          */
3107         function capitalizeAllNouns() {
3108                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3109         }
3110
3111         /**
3112          * An arrow, depending on the language direction.
3113          *
3114          * @param string $direction The direction of the arrow: forwards (default),
3115          *   backwards, left, right, up, down.
3116          * @return string
3117          */
3118         function getArrow( $direction = 'forwards' ) {
3119                 switch ( $direction ) {
3120                         case 'forwards':
3121                                 return $this->isRTL() ? '←' : '→';
3122                         case 'backwards':
3123                                 return $this->isRTL() ? '→' : '←';
3124                         case 'left':
3125                                 return '←';
3126                         case 'right':
3127                                 return '→';
3128                         case 'up':
3129                                 return '↑';
3130                         case 'down':
3131                                 return '↓';
3132                 }
3133         }
3134
3135         /**
3136          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3137          *
3138          * @return bool
3139          */
3140         function linkPrefixExtension() {
3141                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3142         }
3143
3144         /**
3145          * Get all magic words from cache.
3146          * @return array
3147          */
3148         function getMagicWords() {
3149                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3150         }
3151
3152         /**
3153          * Run the LanguageGetMagic hook once.
3154          */
3155         protected function doMagicHook() {
3156                 if ( $this->mMagicHookDone ) {
3157                         return;
3158                 }
3159                 $this->mMagicHookDone = true;
3160                 Hooks::run( 'LanguageGetMagic', [ &$this->mMagicExtensions, $this->getCode() ] );
3161         }
3162
3163         /**
3164          * Fill a MagicWord object with data from here
3165          *
3166          * @param MagicWord $mw
3167          */
3168         function getMagic( $mw ) {
3169                 // Saves a function call
3170                 if ( !$this->mMagicHookDone ) {
3171                         $this->doMagicHook();
3172                 }
3173
3174                 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
3175                         $rawEntry = $this->mMagicExtensions[$mw->mId];
3176                 } else {
3177                         $rawEntry = self::$dataCache->getSubitem(
3178                                 $this->mCode, 'magicWords', $mw->mId );
3179                 }
3180
3181                 if ( !is_array( $rawEntry ) ) {
3182                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3183                 } else {
3184                         $mw->mCaseSensitive = $rawEntry[0];
3185                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3186                 }
3187         }
3188
3189         /**
3190          * Add magic words to the extension array
3191          *
3192          * @param array $newWords
3193          */
3194         function addMagicWordsByLang( $newWords ) {
3195                 $fallbackChain = $this->getFallbackLanguages();
3196                 $fallbackChain = array_reverse( $fallbackChain );
3197                 foreach ( $fallbackChain as $code ) {
3198                         if ( isset( $newWords[$code] ) ) {
3199                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3200                         }
3201                 }
3202         }
3203
3204         /**
3205          * Get special page names, as an associative array
3206          *   canonical name => array of valid names, including aliases
3207          * @return array
3208          */
3209         function getSpecialPageAliases() {
3210                 // Cache aliases because it may be slow to load them
3211                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3212                         // Initialise array
3213                         $this->mExtendedSpecialPageAliases =
3214                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3215                         Hooks::run( 'LanguageGetSpecialPageAliases',
3216                                 [ &$this->mExtendedSpecialPageAliases, $this->getCode() ] );
3217                 }
3218
3219                 return $this->mExtendedSpecialPageAliases;
3220         }
3221
3222         /**
3223          * Italic is unsuitable for some languages
3224          *
3225          * @param string $text The text to be emphasized.
3226          * @return string
3227          */
3228         function emphasize( $text ) {
3229                 return "<em>$text</em>";
3230         }
3231
3232         /**
3233          * Normally we output all numbers in plain en_US style, that is
3234          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3235          * point twohundredthirtyfive. However this is not suitable for all
3236          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3237          * Icelandic just want to use commas instead of dots, and dots instead
3238          * of commas like "293.291,235".
3239          *
3240          * An example of this function being called:
3241          * <code>
3242          * wfMessage( 'message' )->numParams( $num )->text()
3243          * </code>
3244          *
3245          * See $separatorTransformTable on MessageIs.php for
3246          * the , => . and . => , implementation.
3247          *
3248          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3249          * @param int|float $number The string to be formatted, should be an integer
3250          *   or a floating point number.
3251          * @param bool $nocommafy Set to true for special numbers like dates
3252          * @return string
3253          */
3254         public function formatNum( $number, $nocommafy = false ) {
3255                 global $wgTranslateNumerals;
3256                 if ( !$nocommafy ) {
3257                         $number = $this->commafy( $number );
3258                         $s = $this->separatorTransformTable();
3259                         if ( $s ) {
3260                                 $number = strtr( $number, $s );
3261                         }
3262                 }
3263
3264                 if ( $wgTranslateNumerals ) {
3265                         $s = $this->digitTransformTable();
3266                         if ( $s ) {
3267                                 $number = strtr( $number, $s );
3268                         }
3269                 }
3270
3271                 return (string)$number;
3272         }
3273
3274         /**
3275          * Front-end for non-commafied formatNum
3276          *
3277          * @param int|float $number The string to be formatted, should be an integer
3278          *        or a floating point number.
3279          * @since 1.21
3280          * @return string
3281          */
3282         public function formatNumNoSeparators( $number ) {
3283                 return $this->formatNum( $number, true );
3284         }
3285
3286         /**
3287          * @param string $number
3288          * @return string
3289          */
3290         public function parseFormattedNumber( $number ) {
3291                 $s = $this->digitTransformTable();
3292                 if ( $s ) {
3293                         // eliminate empty array values such as ''. (T66347)
3294                         $s = array_filter( $s );
3295                         $number = strtr( $number, array_flip( $s ) );
3296                 }
3297
3298                 $s = $this->separatorTransformTable();
3299                 if ( $s ) {
3300                         // eliminate empty array values such as ''. (T66347)
3301                         $s = array_filter( $s );
3302                         $number = strtr( $number, array_flip( $s ) );
3303                 }
3304
3305                 $number = strtr( $number, [ ',' => '' ] );
3306                 return $number;
3307         }
3308
3309         /**
3310          * Adds commas to a given number
3311          * @since 1.19
3312          * @param mixed $number
3313          * @return string
3314          */
3315         function commafy( $number ) {
3316                 $digitGroupingPattern = $this->digitGroupingPattern();
3317                 if ( $number === null ) {
3318                         return '';
3319                 }
3320
3321                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3322                         // default grouping is at thousands,  use the same for ###,###,### pattern too.
3323                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3324                 } else {
3325                         // Ref: http://cldr.unicode.org/translation/number-patterns
3326                         $sign = "";
3327                         if ( intval( $number ) < 0 ) {
3328                                 // For negative numbers apply the algorithm like positive number and add sign.
3329                                 $sign = "-";
3330                                 $number = substr( $number, 1 );
3331                         }
3332                         $integerPart = [];
3333                         $decimalPart = [];
3334                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3335                         preg_match( "/\d+/", $number, $integerPart );
3336                         preg_match( "/\.\d*/", $number, $decimalPart );
3337                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3338                         if ( $groupedNumber === $number ) {
3339                                 // the string does not have any number part. Eg: .12345
3340                                 return $sign . $groupedNumber;
3341                         }
3342                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3343                         while ( $start > 0 ) {
3344                                 $match = $matches[0][$numMatches - 1];
3345                                 $matchLen = strlen( $match );
3346                                 $start = $end - $matchLen;
3347                                 if ( $start < 0 ) {
3348                                         $start = 0;
3349                                 }
3350                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3351                                 $end = $start;
3352                                 if ( $numMatches > 1 ) {
3353                                         // use the last pattern for the rest of the number
3354                                         $numMatches--;
3355                                 }
3356                                 if ( $start > 0 ) {
3357                                         $groupedNumber = "," . $groupedNumber;
3358                                 }
3359                         }
3360                         return $sign . $groupedNumber;
3361                 }
3362         }
3363
3364         /**
3365          * @return string
3366          */
3367         function digitGroupingPattern() {
3368                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3369         }
3370
3371         /**
3372          * @return array
3373          */
3374         function digitTransformTable() {
3375                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3376         }
3377
3378         /**
3379          * @return array
3380          */
3381         function separatorTransformTable() {
3382                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3383         }
3384
3385         /**
3386          * Take a list of strings and build a locale-friendly comma-separated
3387          * list, using the local comma-separator message.
3388          * The last two strings are chained with an "and".
3389          * NOTE: This function will only work with standard numeric array keys (0, 1, 2…)
3390          *
3391          * @param string[] $l
3392          * @return string
3393          */
3394         function listToText( array $l ) {
3395                 $m = count( $l ) - 1;
3396                 if ( $m < 0 ) {
3397                         return '';
3398                 }
3399                 if ( $m > 0 ) {
3400                         $and = $this->msg( 'and' )->escaped();
3401                         $space = $this->msg( 'word-separator' )->escaped();
3402                         if ( $m > 1 ) {
3403                                 $comma = $this->msg( 'comma-separator' )->escaped();
3404                         }
3405                 }
3406                 $s = $l[$m];
3407                 for ( $i = $m - 1; $i >= 0; $i-- ) {
3408                         if ( $i == $m - 1 ) {
3409                                 $s = $l[$i] . $and . $space . $s;
3410                         } else {
3411                                 $s = $l[$i] . $comma . $s;
3412                         }
3413                 }
3414                 return $s;
3415         }
3416
3417         /**
3418          * Take a list of strings and build a locale-friendly comma-separated
3419          * list, using the local comma-separator message.
3420          * @param string[] $list Array of strings to put in a comma list
3421          * @return string
3422          */
3423         function commaList( array $list ) {
3424                 return implode(
3425                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3426                         $list
3427                 );
3428         }
3429
3430         /**
3431          * Take a list of strings and build a locale-friendly semicolon-separated
3432          * list, using the local semicolon-separator message.
3433          * @param string[] $list Array of strings to put in a semicolon list
3434          * @return string
3435          */
3436         function semicolonList( array $list ) {
3437                 return implode(
3438                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3439                         $list
3440                 );
3441         }
3442
3443         /**
3444          * Same as commaList, but separate it with the pipe instead.
3445          * @param string[] $list Array of strings to put in a pipe list
3446          * @return string
3447          */
3448         function pipeList( array $list ) {
3449                 return implode(
3450                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3451                         $list
3452                 );
3453         }
3454
3455         /**
3456          * Truncate a string to a specified length in bytes, appending an optional
3457          * string (e.g. for ellipses)
3458          *
3459          * The database offers limited byte lengths for some columns in the database;
3460          * multi-byte character sets mean we need to ensure that only whole characters
3461          * are included, otherwise broken characters can be passed to the user
3462          *
3463          * If $length is negative, the string will be truncated from the beginning
3464          *
3465          * @param string $string String to truncate
3466          * @param int $length Maximum length (including ellipses)
3467          * @param string $ellipsis String to append to the truncated text
3468          * @param bool $adjustLength Subtract length of ellipsis from $length.
3469          *      $adjustLength was introduced in 1.18, before that behaved as if false.
3470          * @return string
3471          */
3472         function truncate( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3473                 # Use the localized ellipsis character
3474                 if ( $ellipsis == '...' ) {
3475                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3476                 }
3477                 # Check if there is no need to truncate
3478                 if ( $length == 0 ) {
3479                         return $ellipsis; // convention
3480                 } elseif ( strlen( $string ) <= abs( $length ) ) {
3481                         return $string; // no need to truncate
3482                 }
3483                 $stringOriginal = $string;
3484                 # If ellipsis length is >= $length then we can't apply $adjustLength
3485                 if ( $adjustLength && strlen( $ellipsis ) >= abs( $length ) ) {
3486                         $string = $ellipsis; // this can be slightly unexpected
3487                 # Otherwise, truncate and add ellipsis...
3488                 } else {
3489                         $eLength = $adjustLength ? strlen( $ellipsis ) : 0;
3490                         if ( $length > 0 ) {
3491                                 $length -= $eLength;
3492                                 $string = substr( $string, 0, $length ); // xyz...
3493                                 $string = $this->removeBadCharLast( $string );
3494                                 $string = rtrim( $string );
3495                                 $string = $string . $ellipsis;
3496                         } else {
3497                                 $length += $eLength;
3498                                 $string = substr( $string, $length ); // ...xyz
3499                                 $string = $this->removeBadCharFirst( $string );
3500                                 $string = ltrim( $string );
3501                                 $string = $ellipsis . $string;
3502                         }
3503                 }
3504                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3505                 # This check is *not* redundant if $adjustLength, due to the single case where
3506                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3507                 if ( strlen( $string ) < strlen( $stringOriginal ) ) {
3508                         return $string;
3509                 } else {
3510                         return $stringOriginal;
3511                 }
3512         }
3513
3514         /**
3515          * Remove bytes that represent an incomplete Unicode character
3516          * at the end of string (e.g. bytes of the char are missing)
3517          *
3518          * @param string $string
3519          * @return string
3520          */
3521         protected function removeBadCharLast( $string ) {
3522                 if ( $string != '' ) {
3523                         $char = ord( $string[strlen( $string ) - 1] );
3524                         $m = [];
3525                         if ( $char >= 0xc0 ) {
3526                                 # We got the first byte only of a multibyte char; remove it.
3527                                 $string = substr( $string, 0, -1 );
3528                         } elseif ( $char >= 0x80 &&
3529                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3530                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3531                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3532                         ) {
3533                                 # We chopped in the middle of a character; remove it
3534                                 $string = $m[1];
3535                         }
3536                 }
3537                 return $string;
3538         }
3539
3540         /**
3541          * Remove bytes that represent an incomplete Unicode character
3542          * at the start of string (e.g. bytes of the char are missing)
3543          *
3544          * @param string $string
3545          * @return string
3546          */
3547         protected function removeBadCharFirst( $string ) {
3548                 if ( $string != '' ) {
3549                         $char = ord( $string[0] );
3550                         if ( $char >= 0x80 && $char < 0xc0 ) {
3551                                 # We chopped in the middle of a character; remove the whole thing
3552                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3553                         }
3554                 }
3555                 return $string;
3556         }
3557
3558         /**
3559          * Truncate a string of valid HTML to a specified length in bytes,
3560          * appending an optional string (e.g. for ellipses), and return valid HTML
3561          *
3562          * This is only intended for styled/linked text, such as HTML with
3563          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3564          * Also, this will not detect things like "display:none" CSS.
3565          *
3566          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3567          *
3568          * @param string $text HTML string to truncate
3569          * @param int $length (zero/positive) Maximum length (including ellipses)
3570          * @param string $ellipsis String to append to the truncated text
3571          * @return string
3572          */
3573         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3574                 # Use the localized ellipsis character
3575                 if ( $ellipsis == '...' ) {
3576                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3577                 }
3578                 # Check if there is clearly no need to truncate
3579                 if ( $length <= 0 ) {
3580                         return $ellipsis; // no text shown, nothing to format (convention)
3581                 } elseif ( strlen( $text ) <= $length ) {
3582                         return $text; // string short enough even *with* HTML (short-circuit)
3583                 }
3584
3585                 $dispLen = 0; // innerHTML legth so far
3586                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3587                 $tagType = 0; // 0-open, 1-close
3588                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3589                 $entityState = 0; // 0-not entity, 1-entity
3590                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3591                 $openTags = []; // open tag stack
3592                 $maybeState = null; // possible truncation state
3593
3594                 $textLen = strlen( $text );
3595                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3596                 for ( $pos = 0; true; ++$pos ) {
3597                         # Consider truncation once the display length has reached the maximim.
3598                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3599                         # Check that we're not in the middle of a bracket/entity...
3600                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3601                                 if ( !$testingEllipsis ) {
3602                                         $testingEllipsis = true;
3603                                         # Save where we are; we will truncate here unless there turn out to
3604                                         # be so few remaining characters that truncation is not necessary.
3605                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3606                                                 $maybeState = [ $ret, $openTags ]; // save state
3607                                         }
3608                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3609                                         # String in fact does need truncation, the truncation point was OK.
3610                                         list( $ret, $openTags ) = $maybeState; // reload state
3611                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3612                                         $ret .= $ellipsis; // add ellipsis
3613                                         break;
3614                                 }
3615                         }
3616                         if ( $pos >= $textLen ) {
3617                                 break; // extra iteration just for above checks
3618                         }
3619
3620                         # Read the next char...
3621                         $ch = $text[$pos];
3622                         $lastCh = $pos ? $text[$pos - 1] : '';
3623                         $ret .= $ch; // add to result string
3624                         if ( $ch == '<' ) {
3625                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3626                                 $entityState = 0; // for bad HTML
3627                                 $bracketState = 1; // tag started (checking for backslash)
3628                         } elseif ( $ch == '>' ) {
3629                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3630                                 $entityState = 0; // for bad HTML
3631                                 $bracketState = 0; // out of brackets
3632                         } elseif ( $bracketState == 1 ) {
3633                                 if ( $ch == '/' ) {
3634                                         $tagType = 1; // close tag (e.g. "</span>")
3635                                 } else {
3636                                         $tagType = 0; // open tag (e.g. "<span>")
3637                                         $tag .= $ch;
3638                                 }
3639                                 $bracketState = 2; // building tag name
3640                         } elseif ( $bracketState == 2 ) {
3641                                 if ( $ch != ' ' ) {
3642                                         $tag .= $ch;
3643                                 } else {
3644                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3645                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3646                                 }
3647                         } elseif ( $bracketState == 0 ) {
3648                                 if ( $entityState ) {
3649                                         if ( $ch == ';' ) {
3650                                                 $entityState = 0;
3651                                                 $dispLen++; // entity is one displayed char
3652                                         }
3653                                 } else {
3654                                         if ( $neLength == 0 && !$maybeState ) {
3655                                                 // Save state without $ch. We want to *hit* the first
3656                                                 // display char (to get tags) but not *use* it if truncating.
3657                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3658                                         }
3659                                         if ( $ch == '&' ) {
3660                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3661                                         } else {
3662                                                 $dispLen++; // this char is displayed
3663                                                 // Add the next $max display text chars after this in one swoop...
3664                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3665                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3666                                                 $dispLen += $skipped;
3667                                                 $pos += $skipped;
3668                                         }
3669                                 }
3670                         }
3671                 }
3672                 // Close the last tag if left unclosed by bad HTML
3673                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3674                 while ( count( $openTags ) > 0 ) {
3675                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3676                 }
3677                 return $ret;
3678         }
3679
3680         /**
3681          * truncateHtml() helper function
3682          * like strcspn() but adds the skipped chars to $ret
3683          *
3684          * @param string $ret
3685          * @param string $text
3686          * @param string $search
3687          * @param int $start
3688          * @param null|int $len
3689          * @return int
3690          */
3691         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3692                 if ( $len === null ) {
3693                         $len = -1; // -1 means "no limit" for strcspn
3694                 } elseif ( $len < 0 ) {
3695                         $len = 0; // sanity
3696                 }
3697                 $skipCount = 0;
3698                 if ( $start < strlen( $text ) ) {
3699                         $skipCount = strcspn( $text, $search, $start, $len );
3700                         $ret .= substr( $text, $start, $skipCount );
3701                 }
3702                 return $skipCount;
3703         }
3704
3705         /**
3706          * truncateHtml() helper function
3707          * (a) push or pop $tag from $openTags as needed
3708          * (b) clear $tag value
3709          * @param string &$tag Current HTML tag name we are looking at
3710          * @param int $tagType (0-open tag, 1-close tag)
3711          * @param string $lastCh Character before the '>' that ended this tag
3712          * @param array &$openTags Open tag stack (not accounting for $tag)
3713          */
3714         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3715                 $tag = ltrim( $tag );
3716                 if ( $tag != '' ) {
3717                         if ( $tagType == 0 && $lastCh != '/' ) {
3718                                 $openTags[] = $tag; // tag opened (didn't close itself)
3719                         } elseif ( $tagType == 1 ) {
3720                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3721                                         array_pop( $openTags ); // tag closed
3722                                 }
3723                         }
3724                         $tag = '';
3725                 }
3726         }
3727
3728         /**
3729          * Grammatical transformations, needed for inflected languages
3730          * Invoked by putting {{grammar:case|word}} in a message
3731          *
3732          * @param string $word
3733          * @param string $case
3734          * @return string
3735          */
3736         function convertGrammar( $word, $case ) {
3737                 global $wgGrammarForms;
3738                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3739                         return $wgGrammarForms[$this->getCode()][$case][$word];
3740                 }
3741
3742                 $grammarTransformations = $this->getGrammarTransformations();
3743
3744                 if ( isset( $grammarTransformations[$case] ) ) {
3745                         $forms = $grammarTransformations[$case];
3746
3747                         // Some names of grammar rules are aliases for other rules.
3748                         // In such cases the value is a string rather than object,
3749                         // so load the actual rules.
3750                         if ( is_string( $forms ) ) {
3751                                 $forms = $grammarTransformations[$forms];
3752                         }
3753
3754                         foreach ( array_values( $forms ) as $rule ) {
3755                                 $form = $rule[0];
3756
3757                                 if ( $form === '@metadata' ) {
3758                                         continue;
3759                                 }
3760
3761                                 $replacement = $rule[1];
3762
3763                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3764                                 $patternMatches = preg_match( $regex, $word );
3765
3766                                 if ( $patternMatches === false ) {
3767                                         wfLogWarning(
3768                                                 'An error occurred while processing grammar. ' .
3769                                                 "Word: '$word'. Regex: /$form/."
3770                                         );
3771                                 } elseif ( $patternMatches === 1 ) {
3772                                         $word = preg_replace( $regex, $replacement, $word );
3773
3774                                         break;
3775                                 }
3776                         }
3777                 }
3778
3779                 return $word;
3780         }
3781
3782         /**
3783          * Get the grammar forms for the content language
3784          * @return array Array of grammar forms
3785          * @since 1.20
3786          */
3787         function getGrammarForms() {
3788                 global $wgGrammarForms;
3789                 if ( isset( $wgGrammarForms[$this->getCode()] )
3790                         && is_array( $wgGrammarForms[$this->getCode()] )
3791                 ) {
3792                         return $wgGrammarForms[$this->getCode()];
3793                 }
3794
3795                 return [];
3796         }
3797
3798         /**
3799          * Get the grammar transformations data for the language.
3800          * Used like grammar forms, with {{GRAMMAR}} and cases,
3801          * but uses pairs of regexes and replacements instead of code.
3802          *
3803          * @return array[] Array of grammar transformations.
3804          * @throws MWException
3805          * @since 1.28
3806          */
3807         public function getGrammarTransformations() {
3808                 $languageCode = $this->getCode();
3809
3810                 if ( self::$grammarTransformations === null ) {
3811                         self::$grammarTransformations = new MapCacheLRU( 10 );
3812                 }
3813
3814                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3815                         return self::$grammarTransformations->get( $languageCode );
3816                 }
3817
3818                 $data = [];
3819
3820                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3821                 if ( is_readable( $grammarDataFile ) ) {
3822                         $data = FormatJson::decode(
3823                                 file_get_contents( $grammarDataFile ),
3824                                 true
3825                         );
3826
3827                         if ( $data === null ) {
3828                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3829                         }
3830
3831                         self::$grammarTransformations->set( $languageCode, $data );
3832                 }
3833
3834                 return $data;
3835         }
3836
3837         /**
3838          * Provides an alternative text depending on specified gender.
3839          * Usage {{gender:username|masculine|feminine|unknown}}.
3840          * username is optional, in which case the gender of current user is used,
3841          * but only in (some) interface messages; otherwise default gender is used.
3842          *
3843          * If no forms are given, an empty string is returned. If only one form is
3844          * given, it will be returned unconditionally. These details are implied by
3845          * the caller and cannot be overridden in subclasses.
3846          *
3847          * If three forms are given, the default is to use the third (unknown) form.
3848          * If fewer than three forms are given, the default is to use the first (masculine) form.
3849          * These details can be overridden in subclasses.
3850          *
3851          * @param string $gender
3852          * @param array $forms
3853          *
3854          * @return string
3855          */
3856         function gender( $gender, $forms ) {
3857                 if ( !count( $forms ) ) {
3858                         return '';
3859                 }
3860                 $forms = $this->preConvertPlural( $forms, 2 );
3861                 if ( $gender === 'male' ) {
3862                         return $forms[0];
3863                 }
3864                 if ( $gender === 'female' ) {
3865                         return $forms[1];
3866                 }
3867                 return isset( $forms[2] ) ? $forms[2] : $forms[0];
3868         }
3869
3870         /**
3871          * Plural form transformations, needed for some languages.
3872          * For example, there are 3 form of plural in Russian and Polish,
3873          * depending on "count mod 10". See [[w:Plural]]
3874          * For English it is pretty simple.
3875          *
3876          * Invoked by putting {{plural:count|wordform1|wordform2}}
3877          * or {{plural:count|wordform1|wordform2|wordform3}}
3878          *
3879          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3880          *
3881          * @param int $count Non-localized number
3882          * @param array $forms Different plural forms
3883          * @return string Correct form of plural for $count in this language
3884          */
3885         function convertPlural( $count, $forms ) {
3886                 // Handle explicit n=pluralform cases
3887                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3888                 if ( is_string( $forms ) ) {
3889                         return $forms;
3890                 }
3891                 if ( !count( $forms ) ) {
3892                         return '';
3893                 }
3894
3895                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3896                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3897                 return $forms[$pluralForm];
3898         }
3899
3900         /**
3901          * Handles explicit plural forms for Language::convertPlural()
3902          *
3903          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
3904          * If an explicitly defined plural form matches the $count, then
3905          * string value returned, otherwise array returned for further consideration
3906          * by CLDR rules or overridden convertPlural().
3907          *
3908          * @since 1.23
3909          *
3910          * @param int $count Non-localized number
3911          * @param array $forms Different plural forms
3912          *
3913          * @return array|string
3914          */
3915         protected function handleExplicitPluralForms( $count, array $forms ) {
3916                 foreach ( $forms as $index => $form ) {
3917                         if ( preg_match( '/\d+=/i', $form ) ) {
3918                                 $pos = strpos( $form, '=' );
3919                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
3920                                         return substr( $form, $pos + 1 );
3921                                 }
3922                                 unset( $forms[$index] );
3923                         }
3924                 }
3925                 return array_values( $forms );
3926         }
3927
3928         /**
3929          * Checks that convertPlural was given an array and pads it to requested
3930          * amount of forms by copying the last one.
3931          *
3932          * @param array $forms Array of forms given to convertPlural
3933          * @param int $count How many forms should there be at least
3934          * @return array Padded array of forms or an exception if not an array
3935          */
3936         protected function preConvertPlural( /* Array */ $forms, $count ) {
3937                 while ( count( $forms ) < $count ) {
3938                         $forms[] = $forms[count( $forms ) - 1];
3939                 }
3940                 return $forms;
3941         }
3942
3943         /**
3944          * Wraps argument with unicode control characters for directionality safety
3945          *
3946          * This solves the problem where directionality-neutral characters at the edge of
3947          * the argument string get interpreted with the wrong directionality from the
3948          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
3949          *
3950          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
3951          * directionality of the argument string, using the BIDI algorithm's own "First
3952          * strong directional codepoint" rule. Essentially, this works round the fact that
3953          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
3954          * direction inference). The latter is cleaner but still not widely supported.
3955          *
3956          * @param string $text Text to wrap
3957          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
3958          */
3959         public function embedBidi( $text = '' ) {
3960                 $dir = self::strongDirFromContent( $text );
3961                 if ( $dir === 'ltr' ) {
3962                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
3963                         return self::$lre . $text . self::$pdf;
3964                 }
3965                 if ( $dir === 'rtl' ) {
3966                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
3967                         return self::$rle . $text . self::$pdf;
3968                 }
3969                 // No strong directionality: do not wrap
3970                 return $text;
3971         }
3972
3973         /**
3974          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
3975          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
3976          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
3977          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
3978          * match up with it.
3979          *
3980          * @param string $str The validated block duration in English
3981          * @param User $user User object to use timezone from or null for $wgUser
3982          * @param int $now Current timestamp, for formatting relative block durations
3983          * @return string Somehow translated block duration
3984          * @see LanguageFi.php for example implementation
3985          */
3986         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
3987                 $duration = SpecialBlock::getSuggestedDurations( $this );
3988                 foreach ( $duration as $show => $value ) {
3989                         if ( strcmp( $str, $value ) == 0 ) {
3990                                 return htmlspecialchars( trim( $show ) );
3991                         }
3992                 }
3993
3994                 if ( wfIsInfinity( $str ) ) {
3995                         foreach ( $duration as $show => $value ) {
3996                                 if ( wfIsInfinity( $value ) ) {
3997                                         return htmlspecialchars( trim( $show ) );
3998                                 }
3999                         }
4000                 }
4001
4002                 // If all else fails, return a standard duration or timestamp description.
4003                 $time = strtotime( $str, $now );
4004                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4005                         return $str;
4006                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4007                         // The result differs based on current time, so the difference
4008                         // is a fixed duration length.
4009                         return $this->formatDuration( $time - $now );
4010                 } else { // It's an absolute timestamp.
4011                         if ( $time === 0 ) {
4012                                 // wfTimestamp() handles 0 as current time instead of epoch.
4013                                 $time = '19700101000000';
4014                         }
4015                         if ( $user ) {
4016                                 return $this->userTimeAndDate( $time, $user );
4017                         }
4018                         return $this->timeanddate( $time );
4019                 }
4020         }
4021
4022         /**
4023          * languages like Chinese need to be segmented in order for the diff
4024          * to be of any use
4025          *
4026          * @param string $text
4027          * @return string
4028          */
4029         public function segmentForDiff( $text ) {
4030                 return $text;
4031         }
4032
4033         /**
4034          * and unsegment to show the result
4035          *
4036          * @param string $text
4037          * @return string
4038          */
4039         public function unsegmentForDiff( $text ) {
4040                 return $text;
4041         }
4042
4043         /**
4044          * Return the LanguageConverter used in the Language
4045          *
4046          * @since 1.19
4047          * @return LanguageConverter
4048          */
4049         public function getConverter() {
4050                 return $this->mConverter;
4051         }
4052
4053         /**
4054          * convert text to all supported variants
4055          *
4056          * @param string $text
4057          * @return array
4058          */
4059         public function autoConvertToAllVariants( $text ) {
4060                 return $this->mConverter->autoConvertToAllVariants( $text );
4061         }
4062
4063         /**
4064          * convert text to different variants of a language.
4065          *
4066          * @param string $text
4067          * @return string
4068          */
4069         public function convert( $text ) {
4070                 return $this->mConverter->convert( $text );
4071         }
4072
4073         /**
4074          * Convert a Title object to a string in the preferred variant
4075          *
4076          * @param Title $title
4077          * @return string
4078          */
4079         public function convertTitle( $title ) {
4080                 return $this->mConverter->convertTitle( $title );
4081         }
4082
4083         /**
4084          * Convert a namespace index to a string in the preferred variant
4085          *
4086          * @param int $ns
4087          * @return string
4088          */
4089         public function convertNamespace( $ns ) {
4090                 return $this->mConverter->convertNamespace( $ns );
4091         }
4092
4093         /**
4094          * Check if this is a language with variants
4095          *
4096          * @return bool
4097          */
4098         public function hasVariants() {
4099                 return count( $this->getVariants() ) > 1;
4100         }
4101
4102         /**
4103          * Check if the language has the specific variant
4104          *
4105          * @since 1.19
4106          * @param string $variant
4107          * @return bool
4108          */
4109         public function hasVariant( $variant ) {
4110                 return (bool)$this->mConverter->validateVariant( $variant );
4111         }
4112
4113         /**
4114          * Perform output conversion on a string, and encode for safe HTML output.
4115          * @param string $text Text to be converted
4116          * @param bool $isTitle Whether this conversion is for the article title
4117          * @return string
4118          * @todo this should get integrated somewhere sane
4119          */
4120         public function convertHtml( $text, $isTitle = false ) {
4121                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
4122         }
4123
4124         /**
4125          * @param string $key
4126          * @return string
4127          */
4128         public function convertCategoryKey( $key ) {
4129                 return $this->mConverter->convertCategoryKey( $key );
4130         }
4131
4132         /**
4133          * Get the list of variants supported by this language
4134          * see sample implementation in LanguageZh.php
4135          *
4136          * @return string[] An array of language codes
4137          */
4138         public function getVariants() {
4139                 return $this->mConverter->getVariants();
4140         }
4141
4142         /**
4143          * @return string
4144          */
4145         public function getPreferredVariant() {
4146                 return $this->mConverter->getPreferredVariant();
4147         }
4148
4149         /**
4150          * @return string
4151          */
4152         public function getDefaultVariant() {
4153                 return $this->mConverter->getDefaultVariant();
4154         }
4155
4156         /**
4157          * @return string
4158          */
4159         public function getURLVariant() {
4160                 return $this->mConverter->getURLVariant();
4161         }
4162
4163         /**
4164          * If a language supports multiple variants, it is
4165          * possible that non-existing link in one variant
4166          * actually exists in another variant. this function
4167          * tries to find it. See e.g. LanguageZh.php
4168          * The input parameters may be modified upon return
4169          *
4170          * @param string &$link The name of the link
4171          * @param Title &$nt The title object of the link
4172          * @param bool $ignoreOtherCond To disable other conditions when
4173          *   we need to transclude a template or update a category's link
4174          */
4175         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4176                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4177         }
4178
4179         /**
4180          * returns language specific options used by User::getPageRenderHash()
4181          * for example, the preferred language variant
4182          *
4183          * @return string
4184          */
4185         function getExtraHashOptions() {
4186                 return $this->mConverter->getExtraHashOptions();
4187         }
4188
4189         /**
4190          * For languages that support multiple variants, the title of an
4191          * article may be displayed differently in different variants. this
4192          * function returns the apporiate title defined in the body of the article.
4193          *
4194          * @return string
4195          */
4196         public function getParsedTitle() {
4197                 return $this->mConverter->getParsedTitle();
4198         }
4199
4200         /**
4201          * Refresh the cache of conversion tables when
4202          * MediaWiki:Conversiontable* is updated.
4203          *
4204          * @param Title $title The Title of the page being updated
4205          */
4206         public function updateConversionTable( Title $title ) {
4207                 $this->mConverter->updateConversionTable( $title );
4208         }
4209
4210         /**
4211          * Prepare external link text for conversion. When the text is
4212          * a URL, it shouldn't be converted, and it'll be wrapped in
4213          * the "raw" tag (-{R| }-) to prevent conversion.
4214          *
4215          * This function is called "markNoConversion" for historical
4216          * reasons.
4217          *
4218          * @param string $text Text to be used for external link
4219          * @param bool $noParse Wrap it without confirming it's a real URL first
4220          * @return string The tagged text
4221          */
4222         public function markNoConversion( $text, $noParse = false ) {
4223                 // Excluding protocal-relative URLs may avoid many false positives.
4224                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
4225                         return $this->mConverter->markNoConversion( $text );
4226                 } else {
4227                         return $text;
4228                 }
4229         }
4230
4231         /**
4232          * A regular expression to match legal word-trailing characters
4233          * which should be merged onto a link of the form [[foo]]bar.
4234          *
4235          * @return string
4236          */
4237         public function linkTrail() {
4238                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4239         }
4240
4241         /**
4242          * A regular expression character set to match legal word-prefixing
4243          * characters which should be merged onto a link of the form foo[[bar]].
4244          *
4245          * @return string
4246          */
4247         public function linkPrefixCharset() {
4248                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4249         }
4250
4251         /**
4252          * Get the "parent" language which has a converter to convert a "compatible" language
4253          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4254          *
4255          * @return Language|null
4256          * @since 1.22
4257          */
4258         public function getParentLanguage() {
4259                 if ( $this->mParentLanguage !== false ) {
4260                         return $this->mParentLanguage;
4261                 }
4262
4263                 $code = explode( '-', $this->getCode() )[0];
4264                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4265                         $this->mParentLanguage = null;
4266                         return null;
4267                 }
4268                 $lang = self::factory( $code );
4269                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4270                         $this->mParentLanguage = null;
4271                         return null;
4272                 }
4273
4274                 $this->mParentLanguage = $lang;
4275                 return $lang;
4276         }
4277
4278         /**
4279          * Compare with an other language object
4280          *
4281          * @since 1.28
4282          * @param Language $lang
4283          * @return bool
4284          */
4285         public function equals( Language $lang ) {
4286                 return $lang->getCode() === $this->mCode;
4287         }
4288
4289         /**
4290          * Get the internal language code for this language object
4291          *
4292          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4293          * htmlspecialchars() or similar
4294          *
4295          * @return string
4296          */
4297         public function getCode() {
4298                 return $this->mCode;
4299         }
4300
4301         /**
4302          * Get the code in BCP 47 format which we can use
4303          * inside of html lang="" tags.
4304          *
4305          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4306          * htmlspecialchars() or similar.
4307          *
4308          * @since 1.19
4309          * @return string
4310          */
4311         public function getHtmlCode() {
4312                 if ( is_null( $this->mHtmlCode ) ) {
4313                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4314                 }
4315                 return $this->mHtmlCode;
4316         }
4317
4318         /**
4319          * @param string $code
4320          */
4321         public function setCode( $code ) {
4322                 $this->mCode = $code;
4323                 // Ensure we don't leave incorrect cached data lying around
4324                 $this->mHtmlCode = null;
4325                 $this->mParentLanguage = false;
4326         }
4327
4328         /**
4329          * Get the language code from a file name. Inverse of getFileName()
4330          * @param string $filename $prefix . $languageCode . $suffix
4331          * @param string $prefix Prefix before the language code
4332          * @param string $suffix Suffix after the language code
4333          * @return string Language code, or false if $prefix or $suffix isn't found
4334          */
4335         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4336                 $m = null;
4337                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4338                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4339                 if ( !count( $m ) ) {
4340                         return false;
4341                 }
4342                 return str_replace( '_', '-', strtolower( $m[1] ) );
4343         }
4344
4345         /**
4346          * @param string $code
4347          * @param bool $fallback Whether we're going through language fallback chain
4348          * @return string Name of the language class
4349          */
4350         public static function classFromCode( $code, $fallback = true ) {
4351                 if ( $fallback && $code == 'en' ) {
4352                         return 'Language';
4353                 } else {
4354                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4355                 }
4356         }
4357
4358         /**
4359          * Get the name of a file for a certain language code
4360          * @param string $prefix Prepend this to the filename
4361          * @param string $code Language code
4362          * @param string $suffix Append this to the filename
4363          * @throws MWException
4364          * @return string $prefix . $mangledCode . $suffix
4365          */
4366         public static function getFileName( $prefix = 'Language', $code, $suffix = '.php' ) {
4367                 if ( !self::isValidBuiltInCode( $code ) ) {
4368                         throw new MWException( "Invalid language code \"$code\"" );
4369                 }
4370
4371                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4372         }
4373
4374         /**
4375          * @param string $code
4376          * @return string
4377          */
4378         public static function getMessagesFileName( $code ) {
4379                 global $IP;
4380                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4381                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4382                 return $file;
4383         }
4384
4385         /**
4386          * @param string $code
4387          * @return string
4388          * @throws MWException
4389          * @since 1.23
4390          */
4391         public static function getJsonMessagesFileName( $code ) {
4392                 global $IP;
4393
4394                 if ( !self::isValidBuiltInCode( $code ) ) {
4395                         throw new MWException( "Invalid language code \"$code\"" );
4396                 }
4397
4398                 return "$IP/languages/i18n/$code.json";
4399         }
4400
4401         /**
4402          * Get the first fallback for a given language.
4403          *
4404          * @param string $code
4405          *
4406          * @return bool|string
4407          */
4408         public static function getFallbackFor( $code ) {
4409                 $fallbacks = self::getFallbacksFor( $code );
4410                 if ( $fallbacks ) {
4411                         return $fallbacks[0];
4412                 }
4413                 return false;
4414         }
4415
4416         /**
4417          * Get the ordered list of fallback languages.
4418          *
4419          * @since 1.19
4420          * @param string $code Language code
4421          * @return array Non-empty array, ending in "en"
4422          */
4423         public static function getFallbacksFor( $code ) {
4424                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4425                         return [];
4426                 }
4427                 // For unknown languages, fallbackSequence returns an empty array,
4428                 // hardcode fallback to 'en' in that case.
4429                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4430         }
4431
4432         /**
4433          * Get the ordered list of fallback languages, ending with the fallback
4434          * language chain for the site language.
4435          *
4436          * @since 1.22
4437          * @param string $code Language code
4438          * @return array Array( fallbacks, site fallbacks )
4439          */
4440         public static function getFallbacksIncludingSiteLanguage( $code ) {
4441                 global $wgLanguageCode;
4442
4443                 // Usually, we will only store a tiny number of fallback chains, so we
4444                 // keep them in static memory.
4445                 $cacheKey = "{$code}-{$wgLanguageCode}";
4446
4447                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4448                         $fallbacks = self::getFallbacksFor( $code );
4449
4450                         // Append the site's fallback chain, including the site language itself
4451                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4452                         array_unshift( $siteFallbacks, $wgLanguageCode );
4453
4454                         // Eliminate any languages already included in the chain
4455                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4456
4457                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4458                 }
4459                 return self::$fallbackLanguageCache[$cacheKey];
4460         }
4461
4462         /**
4463          * Get all messages for a given language
4464          * WARNING: this may take a long time. If you just need all message *keys*
4465          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4466          *
4467          * @param string $code
4468          *
4469          * @return array
4470          */
4471         public static function getMessagesFor( $code ) {
4472                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4473         }
4474
4475         /**
4476          * Get a message for a given language
4477          *
4478          * @param string $key
4479          * @param string $code
4480          *
4481          * @return string
4482          */
4483         public static function getMessageFor( $key, $code ) {
4484                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4485         }
4486
4487         /**
4488          * Get all message keys for a given language. This is a faster alternative to
4489          * array_keys( Language::getMessagesFor( $code ) )
4490          *
4491          * @since 1.19
4492          * @param string $code Language code
4493          * @return array Array of message keys (strings)
4494          */
4495         public static function getMessageKeysFor( $code ) {
4496                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4497         }
4498
4499         /**
4500          * @param string $talk
4501          * @return mixed
4502          */
4503         function fixVariableInNamespace( $talk ) {
4504                 if ( strpos( $talk, '$1' ) === false ) {
4505                         return $talk;
4506                 }
4507
4508                 global $wgMetaNamespace;
4509                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4510
4511                 # Allow grammar transformations
4512                 # Allowing full message-style parsing would make simple requests
4513                 # such as action=raw much more expensive than they need to be.
4514                 # This will hopefully cover most cases.
4515                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4516                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4517                 return str_replace( ' ', '_', $talk );
4518         }
4519
4520         /**
4521          * @param string $m
4522          * @return string
4523          */
4524         function replaceGrammarInNamespace( $m ) {
4525                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4526         }
4527
4528         /**
4529          * Decode an expiry (block, protection, etc) which has come from the DB
4530          *
4531          * @param string $expiry Database expiry String
4532          * @param bool|int $format True to process using language functions, or TS_ constant
4533          *     to return the expiry in a given timestamp
4534          * @param string $infinity If $format is not true, use this string for infinite expiry
4535          * @return string
4536          * @since 1.18
4537          */
4538         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4539                 static $dbInfinity;
4540                 if ( $dbInfinity === null ) {
4541                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4542                 }
4543
4544                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4545                         return $format === true
4546                                 ? $this->getMessageFromDB( 'infiniteblock' )
4547                                 : $infinity;
4548                 } else {
4549                         return $format === true
4550                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4551                                 : wfTimestamp( $format, $expiry );
4552                 }
4553         }
4554
4555         /**
4556          * Formats a time given in seconds into a string representation of that time.
4557          *
4558          * @param int|float $seconds
4559          * @param array $format An optional argument that formats the returned string in different ways:
4560          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4561          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4562          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4563          *     and friends.
4564          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4565          *     or 'avoidminutes'.
4566          * @return string
4567          */
4568         function formatTimePeriod( $seconds, $format = [] ) {
4569                 if ( !is_array( $format ) ) {
4570                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4571                 }
4572                 if ( !isset( $format['avoid'] ) ) {
4573                         $format['avoid'] = false;
4574                 }
4575                 if ( !isset( $format['noabbrevs'] ) ) {
4576                         $format['noabbrevs'] = false;
4577                 }
4578                 $secondsMsg = wfMessage(
4579                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4580                 $minutesMsg = wfMessage(
4581                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4582                 $hoursMsg = wfMessage(
4583                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4584                 $daysMsg = wfMessage(
4585                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4586
4587                 if ( round( $seconds * 10 ) < 100 ) {
4588                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4589                         $s = $secondsMsg->params( $s )->text();
4590                 } elseif ( round( $seconds ) < 60 ) {
4591                         $s = $this->formatNum( round( $seconds ) );
4592                         $s = $secondsMsg->params( $s )->text();
4593                 } elseif ( round( $seconds ) < 3600 ) {
4594                         $minutes = floor( $seconds / 60 );
4595                         $secondsPart = round( fmod( $seconds, 60 ) );
4596                         if ( $secondsPart == 60 ) {
4597                                 $secondsPart = 0;
4598                                 $minutes++;
4599                         }
4600                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4601                         $s .= ' ';
4602                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4603                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4604                         $hours = floor( $seconds / 3600 );
4605                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4606                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4607                         if ( $secondsPart == 60 ) {
4608                                 $secondsPart = 0;
4609                                 $minutes++;
4610                         }
4611                         if ( $minutes == 60 ) {
4612                                 $minutes = 0;
4613                                 $hours++;
4614                         }
4615                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4616                         $s .= ' ';
4617                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4618                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4619                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4620                         }
4621                 } else {
4622                         $days = floor( $seconds / 86400 );
4623                         if ( $format['avoid'] === 'avoidminutes' ) {
4624                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4625                                 if ( $hours == 24 ) {
4626                                         $hours = 0;
4627                                         $days++;
4628                                 }
4629                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4630                                 $s .= ' ';
4631                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4632                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4633                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4634                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4635                                 if ( $minutes == 60 ) {
4636                                         $minutes = 0;
4637                                         $hours++;
4638                                 }
4639                                 if ( $hours == 24 ) {
4640                                         $hours = 0;
4641                                         $days++;
4642                                 }
4643                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4644                                 $s .= ' ';
4645                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4646                                 $s .= ' ';
4647                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4648                         } else {
4649                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4650                                 $s .= ' ';
4651                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4652                         }
4653                 }
4654                 return $s;
4655         }
4656
4657         /**
4658          * Format a bitrate for output, using an appropriate
4659          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4660          *   the magnitude in question.
4661          *
4662          * This use base 1000. For base 1024 use formatSize(), for another base
4663          * see formatComputingNumbers().
4664          *
4665          * @param int $bps
4666          * @return string
4667          */
4668         function formatBitrate( $bps ) {
4669                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4670         }
4671
4672         /**
4673          * @param int $size Size of the unit
4674          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4675          * @param string $messageKey Message key to be uesd
4676          * @return string
4677          */
4678         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4679                 if ( $size <= 0 ) {
4680                         return str_replace( '$1', $this->formatNum( $size ),
4681                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4682                         );
4683                 }
4684                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4685                 $index = 0;
4686
4687                 $maxIndex = count( $sizes ) - 1;
4688                 while ( $size >= $boundary && $index < $maxIndex ) {
4689                         $index++;
4690                         $size /= $boundary;
4691                 }
4692
4693                 // For small sizes no decimal places necessary
4694                 $round = 0;
4695                 if ( $index > 1 ) {
4696                         // For MB and bigger two decimal places are smarter
4697                         $round = 2;
4698                 }
4699                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4700
4701                 $size = round( $size, $round );
4702                 $text = $this->getMessageFromDB( $msg );
4703                 return str_replace( '$1', $this->formatNum( $size ), $text );
4704         }
4705
4706         /**
4707          * Format a size in bytes for output, using an appropriate
4708          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4709          *
4710          * This method use base 1024. For base 1000 use formatBitrate(), for
4711          * another base see formatComputingNumbers()
4712          *
4713          * @param int $size Size to format
4714          * @return string Plain text (not HTML)
4715          */
4716         function formatSize( $size ) {
4717                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4718         }
4719
4720         /**
4721          * Make a list item, used by various special pages
4722          *
4723          * @param string $page Page link
4724          * @param string $details HTML safe text between brackets
4725          * @param bool $oppositedm Add the direction mark opposite to your
4726          *   language, to display text properly
4727          * @return HTML escaped string
4728          */
4729         function specialList( $page, $details, $oppositedm = true ) {
4730                 if ( !$details ) {
4731                         return $page;
4732                 }
4733
4734                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4735                 return
4736                         $page .
4737                         $dirmark .
4738                         $this->msg( 'word-separator' )->escaped() .
4739                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4740         }
4741
4742         /**
4743          * Generate (prev x| next x) (20|50|100...) type links for paging
4744          *
4745          * @param Title $title Title object to link
4746          * @param int $offset
4747          * @param int $limit
4748          * @param array $query Optional URL query parameter string
4749          * @param bool $atend Optional param for specified if this is the last page
4750          * @return string
4751          */
4752         public function viewPrevNext( Title $title, $offset, $limit,
4753                 array $query = [], $atend = false
4754         ) {
4755                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4756
4757                 # Make 'previous' link
4758                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4759                 if ( $offset > 0 ) {
4760                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4761                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4762                 } else {
4763                         $plink = htmlspecialchars( $prev );
4764                 }
4765
4766                 # Make 'next' link
4767                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4768                 if ( $atend ) {
4769                         $nlink = htmlspecialchars( $next );
4770                 } else {
4771                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4772                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4773                 }
4774
4775                 # Make links to set number of items per page
4776                 $numLinks = [];
4777                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4778                         $numLinks[] = $this->numLink( $title, $offset, $num,
4779                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4780                 }
4781
4782                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4783                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4784         }
4785
4786         /**
4787          * Helper function for viewPrevNext() that generates links
4788          *
4789          * @param Title $title Title object to link
4790          * @param int $offset
4791          * @param int $limit
4792          * @param array $query Extra query parameters
4793          * @param string $link Text to use for the link; will be escaped
4794          * @param string $tooltipMsg Name of the message to use as tooltip
4795          * @param string $class Value of the "class" attribute of the link
4796          * @return string HTML fragment
4797          */
4798         private function numLink( Title $title, $offset, $limit, array $query, $link,
4799                 $tooltipMsg, $class
4800         ) {
4801                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4802                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4803                         ->numParams( $limit )->text();
4804
4805                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4806                         'title' => $tooltip, 'class' => $class ], $link );
4807         }
4808
4809         /**
4810          * Get the conversion rule title, if any.
4811          *
4812          * @return string
4813          */
4814         public function getConvRuleTitle() {
4815                 return $this->mConverter->getConvRuleTitle();
4816         }
4817
4818         /**
4819          * Get the compiled plural rules for the language
4820          * @since 1.20
4821          * @return array Associative array with plural form, and plural rule as key-value pairs
4822          */
4823         public function getCompiledPluralRules() {
4824                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4825                 $fallbacks = self::getFallbacksFor( $this->mCode );
4826                 if ( !$pluralRules ) {
4827                         foreach ( $fallbacks as $fallbackCode ) {
4828                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4829                                 if ( $pluralRules ) {
4830                                         break;
4831                                 }
4832                         }
4833                 }
4834                 return $pluralRules;
4835         }
4836
4837         /**
4838          * Get the plural rules for the language
4839          * @since 1.20
4840          * @return array Associative array with plural form number and plural rule as key-value pairs
4841          */
4842         public function getPluralRules() {
4843                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4844                 $fallbacks = self::getFallbacksFor( $this->mCode );
4845                 if ( !$pluralRules ) {
4846                         foreach ( $fallbacks as $fallbackCode ) {
4847                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4848                                 if ( $pluralRules ) {
4849                                         break;
4850                                 }
4851                         }
4852                 }
4853                 return $pluralRules;
4854         }
4855
4856         /**
4857          * Get the plural rule types for the language
4858          * @since 1.22
4859          * @return array Associative array with plural form number and plural rule type as key-value pairs
4860          */
4861         public function getPluralRuleTypes() {
4862                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4863                 $fallbacks = self::getFallbacksFor( $this->mCode );
4864                 if ( !$pluralRuleTypes ) {
4865                         foreach ( $fallbacks as $fallbackCode ) {
4866                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4867                                 if ( $pluralRuleTypes ) {
4868                                         break;
4869                                 }
4870                         }
4871                 }
4872                 return $pluralRuleTypes;
4873         }
4874
4875         /**
4876          * Find the index number of the plural rule appropriate for the given number
4877          * @param int $number
4878          * @return int The index number of the plural rule
4879          */
4880         public function getPluralRuleIndexNumber( $number ) {
4881                 $pluralRules = $this->getCompiledPluralRules();
4882                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4883                 return $form;
4884         }
4885
4886         /**
4887          * Find the plural rule type appropriate for the given number
4888          * For example, if the language is set to Arabic, getPluralType(5) should
4889          * return 'few'.
4890          * @since 1.22
4891          * @param int $number
4892          * @return string The name of the plural rule type, e.g. one, two, few, many
4893          */
4894         public function getPluralRuleType( $number ) {
4895                 $index = $this->getPluralRuleIndexNumber( $number );
4896                 $pluralRuleTypes = $this->getPluralRuleTypes();
4897                 if ( isset( $pluralRuleTypes[$index] ) ) {
4898                         return $pluralRuleTypes[$index];
4899                 } else {
4900                         return 'other';
4901                 }
4902         }
4903 }