languages/Language.php

   1 <?php
   2 /**
   3  * Internationalisation code.
   4  * See https://www.mediawiki.org/wiki/Special:MyLanguage/Localisation for more information.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Language
  23  */
  24
  25 /**
  26  * @defgroup Language Language
  27  */
  28
  29 use CLDRPluralRuleParser\Evaluator;
  30
  31 /**
  32  * Internationalisation code
  33  * @ingroup Language
  34  */
  35 class Language {
  36         /**
  37          * @var LanguageConverter
  38          */
  39         public $mConverter;
  40
  41         public $mVariants, $mCode, $mLoaded = false;
  42         public $mMagicExtensions = [], $mMagicHookDone = false;
  43         private $mHtmlCode = null, $mParentLanguage = false;
  44
  45         public $dateFormatStrings = [];
  46         public $mExtendedSpecialPageAliases;
  47
  48         /** @var array|null */
  49         protected $namespaceNames;
  50         protected $mNamespaceIds, $namespaceAliases;
  51
  52         /**
  53          * ReplacementArray object caches
  54          */
  55         public $transformData = [];
  56
  57         /**
  58          * @var LocalisationCache
  59          */
  60         static public $dataCache;
  61
  62         static public $mLangObjCache = [];
  63
  64         static public $mWeekdayMsgs = [
  65                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  66                 'friday', 'saturday'
  67         ];
  68
  69         static public $mWeekdayAbbrevMsgs = [
  70                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
  71         ];
  72
  73         static public $mMonthMsgs = [
  74                 'january', 'february', 'march', 'april', 'may_long', 'june',
  75                 'july', 'august', 'september', 'october', 'november',
  76                 'december'
  77         ];
  78         static public $mMonthGenMsgs = [
  79                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
  80                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
  81                 'december-gen'
  82         ];
  83         static public $mMonthAbbrevMsgs = [
  84                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
  85                 'sep', 'oct', 'nov', 'dec'
  86         ];
  87
  88         static public $mIranianCalendarMonthMsgs = [
  89                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
  90                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
  91                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
  92                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
  93         ];
  94
  95         static public $mHebrewCalendarMonthMsgs = [
  96                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
  97                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
  98                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
  99                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 100                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 101         ];
 102
 103         static public $mHebrewCalendarMonthGenMsgs = [
 104                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 105                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 106                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 107                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 108                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 109         ];
 110
 111         static public $mHijriCalendarMonthMsgs = [
 112                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 113                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 114                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 115                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 116         ];
 117
 118         /**
 119          * @since 1.20
 120          * @var array
 121          */
 122         static public $durationIntervals = [
 123                 'millennia' => 31556952000,
 124                 'centuries' => 3155695200,
 125                 'decades' => 315569520,
 126                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 127                 'weeks' => 604800,
 128                 'days' => 86400,
 129                 'hours' => 3600,
 130                 'minutes' => 60,
 131                 'seconds' => 1,
 132         ];
 133
 134         /**
 135          * Cache for language fallbacks.
 136          * @see Language::getFallbacksIncludingSiteLanguage
 137          * @since 1.21
 138          * @var array
 139          */
 140         static private $fallbackLanguageCache = [];
 141
 142         /**
 143          * Cache for grammar rules data
 144          * @var MapCacheLRU|null
 145          */
 146         static private $grammarTransformations;
 147
 148         /**
 149          * Cache for language names
 150          * @var HashBagOStuff|null
 151          */
 152         static private $languageNameCache;
 153
 154         /**
 155          * Unicode directional formatting characters, for embedBidi()
 156          */
 157         static private $lre = "\xE2\x80\xAA"; // U+202A LEFT-TO-RIGHT EMBEDDING
 158         static private $rle = "\xE2\x80\xAB"; // U+202B RIGHT-TO-LEFT EMBEDDING
 159         static private $pdf = "\xE2\x80\xAC"; // U+202C POP DIRECTIONAL FORMATTING
 160
 161         /**
 162          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 163          * - in group 1 if it is LTR
 164          * - in group 2 if it is RTL
 165          * Does not match if there is no strong directionality codepoint.
 166          *
 167          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 168          *
 169          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 170          * https://git.wikimedia.org/summary/unicodejs.git .
 171          */
 172         // @codingStandardsIgnoreStart
 173         // @codeCoverageIgnoreStart
 174         static private $strongDirRegex = '/(?:([\x{41}-\x{5a}\x{61}-\x{7a}\x{aa}\x{b5}\x{ba}\x{c0}-\x{d6}\x{d8}-\x{f6}\x{f8}-\x{2b8}\x{2bb}-\x{2c1}\x{2d0}\x{2d1}\x{2e0}-\x{2e4}\x{2ee}\x{370}-\x{373}\x{376}\x{377}\x{37a}-\x{37d}\x{37f}\x{386}\x{388}-\x{38a}\x{38c}\x{38e}-\x{3a1}\x{3a3}-\x{3f5}\x{3f7}-\x{482}\x{48a}-\x{52f}\x{531}-\x{556}\x{559}-\x{55f}\x{561}-\x{587}\x{589}\x{903}-\x{939}\x{93b}\x{93d}-\x{940}\x{949}-\x{94c}\x{94e}-\x{950}\x{958}-\x{961}\x{964}-\x{980}\x{982}\x{983}\x{985}-\x{98c}\x{98f}\x{990}\x{993}-\x{9a8}\x{9aa}-\x{9b0}\x{9b2}\x{9b6}-\x{9b9}\x{9bd}-\x{9c0}\x{9c7}\x{9c8}\x{9cb}\x{9cc}\x{9ce}\x{9d7}\x{9dc}\x{9dd}\x{9df}-\x{9e1}\x{9e6}-\x{9f1}\x{9f4}-\x{9fa}\x{a03}\x{a05}-\x{a0a}\x{a0f}\x{a10}\x{a13}-\x{a28}\x{a2a}-\x{a30}\x{a32}\x{a33}\x{a35}\x{a36}\x{a38}\x{a39}\x{a3e}-\x{a40}\x{a59}-\x{a5c}\x{a5e}\x{a66}-\x{a6f}\x{a72}-\x{a74}\x{a83}\x{a85}-\x{a8d}\x{a8f}-\x{a91}\x{a93}-\x{aa8}\x{aaa}-\x{ab0}\x{ab2}\x{ab3}\x{ab5}-\x{ab9}\x{abd}-\x{ac0}\x{ac9}\x{acb}\x{acc}\x{ad0}\x{ae0}\x{ae1}\x{ae6}-\x{af0}\x{af9}\x{b02}\x{b03}\x{b05}-\x{b0c}\x{b0f}\x{b10}\x{b13}-\x{b28}\x{b2a}-\x{b30}\x{b32}\x{b33}\x{b35}-\x{b39}\x{b3d}\x{b3e}\x{b40}\x{b47}\x{b48}\x{b4b}\x{b4c}\x{b57}\x{b5c}\x{b5d}\x{b5f}-\x{b61}\x{b66}-\x{b77}\x{b83}\x{b85}-\x{b8a}\x{b8e}-\x{b90}\x{b92}-\x{b95}\x{b99}\x{b9a}\x{b9c}\x{b9e}\x{b9f}\x{ba3}\x{ba4}\x{ba8}-\x{baa}\x{bae}-\x{bb9}\x{bbe}\x{bbf}\x{bc1}\x{bc2}\x{bc6}-\x{bc8}\x{bca}-\x{bcc}\x{bd0}\x{bd7}\x{be6}-\x{bf2}\x{c01}-\x{c03}\x{c05}-\x{c0c}\x{c0e}-\x{c10}\x{c12}-\x{c28}\x{c2a}-\x{c39}\x{c3d}\x{c41}-\x{c44}\x{c58}-\x{c5a}\x{c60}\x{c61}\x{c66}-\x{c6f}\x{c7f}\x{c82}\x{c83}\x{c85}-\x{c8c}\x{c8e}-\x{c90}\x{c92}-\x{ca8}\x{caa}-\x{cb3}\x{cb5}-\x{cb9}\x{cbd}-\x{cc4}\x{cc6}-\x{cc8}\x{cca}\x{ccb}\x{cd5}\x{cd6}\x{cde}\x{ce0}\x{ce1}\x{ce6}-\x{cef}\x{cf1}\x{cf2}\x{d02}\x{d03}\x{d05}-\x{d0c}\x{d0e}-\x{d10}\x{d12}-\x{d3a}\x{d3d}-\x{d40}\x{d46}-\x{d48}\x{d4a}-\x{d4c}\x{d4e}\x{d57}\x{d5f}-\x{d61}\x{d66}-\x{d75}\x{d79}-\x{d7f}\x{d82}\x{d83}\x{d85}-\x{d96}\x{d9a}-\x{db1}\x{db3}-\x{dbb}\x{dbd}\x{dc0}-\x{dc6}\x{dcf}-\x{dd1}\x{dd8}-\x{ddf}\x{de6}-\x{def}\x{df2}-\x{df4}\x{e01}-\x{e30}\x{e32}\x{e33}\x{e40}-\x{e46}\x{e4f}-\x{e5b}\x{e81}\x{e82}\x{e84}\x{e87}\x{e88}\x{e8a}\x{e8d}\x{e94}-\x{e97}\x{e99}-\x{e9f}\x{ea1}-\x{ea3}\x{ea5}\x{ea7}\x{eaa}\x{eab}\x{ead}-\x{eb0}\x{eb2}\x{eb3}\x{ebd}\x{ec0}-\x{ec4}\x{ec6}\x{ed0}-\x{ed9}\x{edc}-\x{edf}\x{f00}-\x{f17}\x{f1a}-\x{f34}\x{f36}\x{f38}\x{f3e}-\x{f47}\x{f49}-\x{f6c}\x{f7f}\x{f85}\x{f88}-\x{f8c}\x{fbe}-\x{fc5}\x{fc7}-\x{fcc}\x{fce}-\x{fda}\x{1000}-\x{102c}\x{1031}\x{1038}\x{103b}\x{103c}\x{103f}-\x{1057}\x{105a}-\x{105d}\x{1061}-\x{1070}\x{1075}-\x{1081}\x{1083}\x{1084}\x{1087}-\x{108c}\x{108e}-\x{109c}\x{109e}-\x{10c5}\x{10c7}\x{10cd}\x{10d0}-\x{1248}\x{124a}-\x{124d}\x{1250}-\x{1256}\x{1258}\x{125a}-\x{125d}\x{1260}-\x{1288}\x{128a}-\x{128d}\x{1290}-\x{12b0}\x{12b2}-\x{12b5}\x{12b8}-\x{12be}\x{12c0}\x{12c2}-\x{12c5}\x{12c8}-\x{12d6}\x{12d8}-\x{1310}\x{1312}-\x{1315}\x{1318}-\x{135a}\x{1360}-\x{137c}\x{1380}-\x{138f}\x{13a0}-\x{13f5}\x{13f8}-\x{13fd}\x{1401}-\x{167f}\x{1681}-\x{169a}\x{16a0}-\x{16f8}\x{1700}-\x{170c}\x{170e}-\x{1711}\x{1720}-\x{1731}\x{1735}\x{1736}\x{1740}-\x{1751}\x{1760}-\x{176c}\x{176e}-\x{1770}\x{1780}-\x{17b3}\x{17b6}\x{17be}-\x{17c5}\x{17c7}\x{17c8}\x{17d4}-\x{17da}\x{17dc}\x{17e0}-\x{17e9}\x{1810}-\x{1819}\x{1820}-\x{1877}\x{1880}-\x{18a8}\x{18aa}\x{18b0}-\x{18f5}\x{1900}-\x{191e}\x{1923}-\x{1926}\x{1929}-\x{192b}\x{1930}\x{1931}\x{1933}-\x{1938}\x{1946}-\x{196d}\x{1970}-\x{1974}\x{1980}-\x{19ab}\x{19b0}-\x{19c9}\x{19d0}-\x{19da}\x{1a00}-\x{1a16}\x{1a19}\x{1a1a}\x{1a1e}-\x{1a55}\x{1a57}\x{1a61}\x{1a63}\x{1a64}\x{1a6d}-\x{1a72}\x{1a80}-\x{1a89}\x{1a90}-\x{1a99}\x{1aa0}-\x{1aad}\x{1b04}-\x{1b33}\x{1b35}\x{1b3b}\x{1b3d}-\x{1b41}\x{1b43}-\x{1b4b}\x{1b50}-\x{1b6a}\x{1b74}-\x{1b7c}\x{1b82}-\x{1ba1}\x{1ba6}\x{1ba7}\x{1baa}\x{1bae}-\x{1be5}\x{1be7}\x{1bea}-\x{1bec}\x{1bee}\x{1bf2}\x{1bf3}\x{1bfc}-\x{1c2b}\x{1c34}\x{1c35}\x{1c3b}-\x{1c49}\x{1c4d}-\x{1c7f}\x{1cc0}-\x{1cc7}\x{1cd3}\x{1ce1}\x{1ce9}-\x{1cec}\x{1cee}-\x{1cf3}\x{1cf5}\x{1cf6}\x{1d00}-\x{1dbf}\x{1e00}-\x{1f15}\x{1f18}-\x{1f1d}\x{1f20}-\x{1f45}\x{1f48}-\x{1f4d}\x{1f50}-\x{1f57}\x{1f59}\x{1f5b}\x{1f5d}\x{1f5f}-\x{1f7d}\x{1f80}-\x{1fb4}\x{1fb6}-\x{1fbc}\x{1fbe}\x{1fc2}-\x{1fc4}\x{1fc6}-\x{1fcc}\x{1fd0}-\x{1fd3}\x{1fd6}-\x{1fdb}\x{1fe0}-\x{1fec}\x{1ff2}-\x{1ff4}\x{1ff6}-\x{1ffc}\x{200e}\x{2071}\x{207f}\x{2090}-\x{209c}\x{2102}\x{2107}\x{210a}-\x{2113}\x{2115}\x{2119}-\x{211d}\x{2124}\x{2126}\x{2128}\x{212a}-\x{212d}\x{212f}-\x{2139}\x{213c}-\x{213f}\x{2145}-\x{2149}\x{214e}\x{214f}\x{2160}-\x{2188}\x{2336}-\x{237a}\x{2395}\x{249c}-\x{24e9}\x{26ac}\x{2800}-\x{28ff}\x{2c00}-\x{2c2e}\x{2c30}-\x{2c5e}\x{2c60}-\x{2ce4}\x{2ceb}-\x{2cee}\x{2cf2}\x{2cf3}\x{2d00}-\x{2d25}\x{2d27}\x{2d2d}\x{2d30}-\x{2d67}\x{2d6f}\x{2d70}\x{2d80}-\x{2d96}\x{2da0}-\x{2da6}\x{2da8}-\x{2dae}\x{2db0}-\x{2db6}\x{2db8}-\x{2dbe}\x{2dc0}-\x{2dc6}\x{2dc8}-\x{2dce}\x{2dd0}-\x{2dd6}\x{2dd8}-\x{2dde}\x{3005}-\x{3007}\x{3021}-\x{3029}\x{302e}\x{302f}\x{3031}-\x{3035}\x{3038}-\x{303c}\x{3041}-\x{3096}\x{309d}-\x{309f}\x{30a1}-\x{30fa}\x{30fc}-\x{30ff}\x{3105}-\x{312d}\x{3131}-\x{318e}\x{3190}-\x{31ba}\x{31f0}-\x{321c}\x{3220}-\x{324f}\x{3260}-\x{327b}\x{327f}-\x{32b0}\x{32c0}-\x{32cb}\x{32d0}-\x{32fe}\x{3300}-\x{3376}\x{337b}-\x{33dd}\x{33e0}-\x{33fe}\x{3400}-\x{4db5}\x{4e00}-\x{9fd5}\x{a000}-\x{a48c}\x{a4d0}-\x{a60c}\x{a610}-\x{a62b}\x{a640}-\x{a66e}\x{a680}-\x{a69d}\x{a6a0}-\x{a6ef}\x{a6f2}-\x{a6f7}\x{a722}-\x{a787}\x{a789}-\x{a7ad}\x{a7b0}-\x{a7b7}\x{a7f7}-\x{a801}\x{a803}-\x{a805}\x{a807}-\x{a80a}\x{a80c}-\x{a824}\x{a827}\x{a830}-\x{a837}\x{a840}-\x{a873}\x{a880}-\x{a8c3}\x{a8ce}-\x{a8d9}\x{a8f2}-\x{a8fd}\x{a900}-\x{a925}\x{a92e}-\x{a946}\x{a952}\x{a953}\x{a95f}-\x{a97c}\x{a983}-\x{a9b2}\x{a9b4}\x{a9b5}\x{a9ba}\x{a9bb}\x{a9bd}-\x{a9cd}\x{a9cf}-\x{a9d9}\x{a9de}-\x{a9e4}\x{a9e6}-\x{a9fe}\x{aa00}-\x{aa28}\x{aa2f}\x{aa30}\x{aa33}\x{aa34}\x{aa40}-\x{aa42}\x{aa44}-\x{aa4b}\x{aa4d}\x{aa50}-\x{aa59}\x{aa5c}-\x{aa7b}\x{aa7d}-\x{aaaf}\x{aab1}\x{aab5}\x{aab6}\x{aab9}-\x{aabd}\x{aac0}\x{aac2}\x{aadb}-\x{aaeb}\x{aaee}-\x{aaf5}\x{ab01}-\x{ab06}\x{ab09}-\x{ab0e}\x{ab11}-\x{ab16}\x{ab20}-\x{ab26}\x{ab28}-\x{ab2e}\x{ab30}-\x{ab65}\x{ab70}-\x{abe4}\x{abe6}\x{abe7}\x{abe9}-\x{abec}\x{abf0}-\x{abf9}\x{ac00}-\x{d7a3}\x{d7b0}-\x{d7c6}\x{d7cb}-\x{d7fb}\x{e000}-\x{fa6d}\x{fa70}-\x{fad9}\x{fb00}-\x{fb06}\x{fb13}-\x{fb17}\x{ff21}-\x{ff3a}\x{ff41}-\x{ff5a}\x{ff66}-\x{ffbe}\x{ffc2}-\x{ffc7}\x{ffca}-\x{ffcf}\x{ffd2}-\x{ffd7}\x{ffda}-\x{ffdc}\x{10000}-\x{1000b}\x{1000d}-\x{10026}\x{10028}-\x{1003a}\x{1003c}\x{1003d}\x{1003f}-\x{1004d}\x{10050}-\x{1005d}\x{10080}-\x{100fa}\x{10100}\x{10102}\x{10107}-\x{10133}\x{10137}-\x{1013f}\x{101d0}-\x{101fc}\x{10280}-\x{1029c}\x{102a0}-\x{102d0}\x{10300}-\x{10323}\x{10330}-\x{1034a}\x{10350}-\x{10375}\x{10380}-\x{1039d}\x{1039f}-\x{103c3}\x{103c8}-\x{103d5}\x{10400}-\x{1049d}\x{104a0}-\x{104a9}\x{10500}-\x{10527}\x{10530}-\x{10563}\x{1056f}\x{10600}-\x{10736}\x{10740}-\x{10755}\x{10760}-\x{10767}\x{11000}\x{11002}-\x{11037}\x{11047}-\x{1104d}\x{11066}-\x{1106f}\x{11082}-\x{110b2}\x{110b7}\x{110b8}\x{110bb}-\x{110c1}\x{110d0}-\x{110e8}\x{110f0}-\x{110f9}\x{11103}-\x{11126}\x{1112c}\x{11136}-\x{11143}\x{11150}-\x{11172}\x{11174}-\x{11176}\x{11182}-\x{111b5}\x{111bf}-\x{111c9}\x{111cd}\x{111d0}-\x{111df}\x{111e1}-\x{111f4}\x{11200}-\x{11211}\x{11213}-\x{1122e}\x{11232}\x{11233}\x{11235}\x{11238}-\x{1123d}\x{11280}-\x{11286}\x{11288}\x{1128a}-\x{1128d}\x{1128f}-\x{1129d}\x{1129f}-\x{112a9}\x{112b0}-\x{112de}\x{112e0}-\x{112e2}\x{112f0}-\x{112f9}\x{11302}\x{11303}\x{11305}-\x{1130c}\x{1130f}\x{11310}\x{11313}-\x{11328}\x{1132a}-\x{11330}\x{11332}\x{11333}\x{11335}-\x{11339}\x{1133d}-\x{1133f}\x{11341}-\x{11344}\x{11347}\x{11348}\x{1134b}-\x{1134d}\x{11350}\x{11357}\x{1135d}-\x{11363}\x{11480}-\x{114b2}\x{114b9}\x{114bb}-\x{114be}\x{114c1}\x{114c4}-\x{114c7}\x{114d0}-\x{114d9}\x{11580}-\x{115b1}\x{115b8}-\x{115bb}\x{115be}\x{115c1}-\x{115db}\x{11600}-\x{11632}\x{1163b}\x{1163c}\x{1163e}\x{11641}-\x{11644}\x{11650}-\x{11659}\x{11680}-\x{116aa}\x{116ac}\x{116ae}\x{116af}\x{116b6}\x{116c0}-\x{116c9}\x{11700}-\x{11719}\x{11720}\x{11721}\x{11726}\x{11730}-\x{1173f}\x{118a0}-\x{118f2}\x{118ff}\x{11ac0}-\x{11af8}\x{12000}-\x{12399}\x{12400}-\x{1246e}\x{12470}-\x{12474}\x{12480}-\x{12543}\x{13000}-\x{1342e}\x{14400}-\x{14646}\x{16800}-\x{16a38}\x{16a40}-\x{16a5e}\x{16a60}-\x{16a69}\x{16a6e}\x{16a6f}\x{16ad0}-\x{16aed}\x{16af5}\x{16b00}-\x{16b2f}\x{16b37}-\x{16b45}\x{16b50}-\x{16b59}\x{16b5b}-\x{16b61}\x{16b63}-\x{16b77}\x{16b7d}-\x{16b8f}\x{16f00}-\x{16f44}\x{16f50}-\x{16f7e}\x{16f93}-\x{16f9f}\x{1b000}\x{1b001}\x{1bc00}-\x{1bc6a}\x{1bc70}-\x{1bc7c}\x{1bc80}-\x{1bc88}\x{1bc90}-\x{1bc99}\x{1bc9c}\x{1bc9f}\x{1d000}-\x{1d0f5}\x{1d100}-\x{1d126}\x{1d129}-\x{1d166}\x{1d16a}-\x{1d172}\x{1d183}\x{1d184}\x{1d18c}-\x{1d1a9}\x{1d1ae}-\x{1d1e8}\x{1d360}-\x{1d371}\x{1d400}-\x{1d454}\x{1d456}-\x{1d49c}\x{1d49e}\x{1d49f}\x{1d4a2}\x{1d4a5}\x{1d4a6}\x{1d4a9}-\x{1d4ac}\x{1d4ae}-\x{1d4b9}\x{1d4bb}\x{1d4bd}-\x{1d4c3}\x{1d4c5}-\x{1d505}\x{1d507}-\x{1d50a}\x{1d50d}-\x{1d514}\x{1d516}-\x{1d51c}\x{1d51e}-\x{1d539}\x{1d53b}-\x{1d53e}\x{1d540}-\x{1d544}\x{1d546}\x{1d54a}-\x{1d550}\x{1d552}-\x{1d6a5}\x{1d6a8}-\x{1d6da}\x{1d6dc}-\x{1d714}\x{1d716}-\x{1d74e}\x{1d750}-\x{1d788}\x{1d78a}-\x{1d7c2}\x{1d7c4}-\x{1d7cb}\x{1d800}-\x{1d9ff}\x{1da37}-\x{1da3a}\x{1da6d}-\x{1da74}\x{1da76}-\x{1da83}\x{1da85}-\x{1da8b}\x{1f110}-\x{1f12e}\x{1f130}-\x{1f169}\x{1f170}-\x{1f19a}\x{1f1e6}-\x{1f202}\x{1f210}-\x{1f23a}\x{1f240}-\x{1f248}\x{1f250}\x{1f251}\x{20000}-\x{2a6d6}\x{2a700}-\x{2b734}\x{2b740}-\x{2b81d}\x{2b820}-\x{2cea1}\x{2f800}-\x{2fa1d}\x{f0000}-\x{ffffd}\x{100000}-\x{10fffd}])|([\x{590}\x{5be}\x{5c0}\x{5c3}\x{5c6}\x{5c8}-\x{5ff}\x{7c0}-\x{7ea}\x{7f4}\x{7f5}\x{7fa}-\x{815}\x{81a}\x{824}\x{828}\x{82e}-\x{858}\x{85c}-\x{89f}\x{200f}\x{fb1d}\x{fb1f}-\x{fb28}\x{fb2a}-\x{fb4f}\x{10800}-\x{1091e}\x{10920}-\x{10a00}\x{10a04}\x{10a07}-\x{10a0b}\x{10a10}-\x{10a37}\x{10a3b}-\x{10a3e}\x{10a40}-\x{10ae4}\x{10ae7}-\x{10b38}\x{10b40}-\x{10e5f}\x{10e7f}-\x{10fff}\x{1e800}-\x{1e8cf}\x{1e8d7}-\x{1edff}\x{1ef00}-\x{1efff}\x{608}\x{60b}\x{60d}\x{61b}-\x{64a}\x{66d}-\x{66f}\x{671}-\x{6d5}\x{6e5}\x{6e6}\x{6ee}\x{6ef}\x{6fa}-\x{710}\x{712}-\x{72f}\x{74b}-\x{7a5}\x{7b1}-\x{7bf}\x{8a0}-\x{8e2}\x{fb50}-\x{fd3d}\x{fd40}-\x{fdcf}\x{fdf0}-\x{fdfc}\x{fdfe}\x{fdff}\x{fe70}-\x{fefe}\x{1ee00}-\x{1eeef}\x{1eef2}-\x{1eeff}]))/u';
 175         // @codeCoverageIgnoreEnd
 176         // @codingStandardsIgnoreEnd
 177
 178         /**
 179          * Get a cached or new language object for a given language code
 180          * @param string $code
 181          * @return Language
 182          */
 183         static function factory( $code ) {
 184                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 185
 186                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 187                         $code = $wgDummyLanguageCodes[$code];
 188                 }
 189
 190                 // get the language object to process
 191                 $langObj = isset( self::$mLangObjCache[$code] )
 192                         ? self::$mLangObjCache[$code]
 193                         : self::newFromCode( $code );
 194
 195                 // merge the language object in to get it up front in the cache
 196                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 197                 // get rid of the oldest ones in case we have an overflow
 198                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 199
 200                 return $langObj;
 201         }
 202
 203         /**
 204          * Create a language object for a given language code
 205          * @param string $code
 206          * @param bool $fallback Whether we're going through language fallback chain
 207          * @throws MWException
 208          * @return Language
 209          */
 210         protected static function newFromCode( $code, $fallback = false ) {
 211                 if ( !self::isValidCode( $code ) ) {
 212                         throw new MWException( "Invalid language code \"$code\"" );
 213                 }
 214
 215                 if ( !self::isValidBuiltInCode( $code ) ) {
 216                         // It's not possible to customise this code with class files, so
 217                         // just return a Language object. This is to support uselang= hacks.
 218                         $lang = new Language;
 219                         $lang->setCode( $code );
 220                         return $lang;
 221                 }
 222
 223                 // Check if there is a language class for the code
 224                 $class = self::classFromCode( $code, $fallback );
 225                 if ( class_exists( $class ) ) {
 226                         $lang = new $class;
 227                         return $lang;
 228                 }
 229
 230                 // Keep trying the fallback list until we find an existing class
 231                 $fallbacks = self::getFallbacksFor( $code );
 232                 foreach ( $fallbacks as $fallbackCode ) {
 233                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 234                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 235                         }
 236
 237                         $class = self::classFromCode( $fallbackCode );
 238                         if ( class_exists( $class ) ) {
 239                                 $lang = new $class;
 240                                 $lang->setCode( $code );
 241                                 return $lang;
 242                         }
 243                 }
 244
 245                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 246         }
 247
 248         /**
 249          * Checks whether any localisation is available for that language tag
 250          * in MediaWiki (MessagesXx.php exists).
 251          *
 252          * @param string $code Language tag (in lower case)
 253          * @return bool Whether language is supported
 254          * @since 1.21
 255          */
 256         public static function isSupportedLanguage( $code ) {
 257                 if ( !self::isValidBuiltInCode( $code ) ) {
 258                         return false;
 259                 }
 260
 261                 if ( $code === 'qqq' ) {
 262                         return false;
 263                 }
 264
 265                 return is_readable( self::getMessagesFileName( $code ) ) ||
 266                         is_readable( self::getJsonMessagesFileName( $code ) );
 267         }
 268
 269         /**
 270          * Returns true if a language code string is a well-formed language tag
 271          * according to RFC 5646.
 272          * This function only checks well-formedness; it doesn't check that
 273          * language, script or variant codes actually exist in the repositories.
 274          *
 275          * Based on regexes by Mark Davis of the Unicode Consortium:
 276          * http://unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 277          *
 278          * @param string $code
 279          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 280          *
 281          * @return bool
 282          * @since 1.21
 283          */
 284         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 285                 $alpha = '[a-z]';
 286                 $digit = '[0-9]';
 287                 $alphanum = '[a-z0-9]';
 288                 $x = 'x'; # private use singleton
 289                 $singleton = '[a-wy-z]'; # other singleton
 290                 $s = $lenient ? '[-_]' : '-';
 291
 292                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 293                 $script = "$alpha{4}"; # ISO 15924
 294                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 295                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 296                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 297                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 298
 299                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 300                 # Since these are limited, this is safe even later changes to the registry --
 301                 # the only oddity is that it might change the type of the tag, and thus
 302                 # the results from the capturing groups.
 303                 # https://www.iana.org/assignments/language-subtag-registry
 304
 305                 $grandfathered = "en{$s}GB{$s}oed"
 306                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 307                         . "|no{$s}(?:bok|nyn)"
 308                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 309                         . "|zh{$s}min{$s}nan";
 310
 311                 $variantList = "$variant(?:$s$variant)*";
 312                 $extensionList = "$extension(?:$s$extension)*";
 313
 314                 $langtag = "(?:($language)"
 315                         . "(?:$s$script)?"
 316                         . "(?:$s$region)?"
 317                         . "(?:$s$variantList)?"
 318                         . "(?:$s$extensionList)?"
 319                         . "(?:$s$privateUse)?)";
 320
 321                 # The final breakdown, with capturing groups for each of these components
 322                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 323
 324                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 325
 326                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 327         }
 328
 329         /**
 330          * Returns true if a language code string is of a valid form, whether or
 331          * not it exists. This includes codes which are used solely for
 332          * customisation via the MediaWiki namespace.
 333          *
 334          * @param string $code
 335          *
 336          * @return bool
 337          */
 338         public static function isValidCode( $code ) {
 339                 static $cache = [];
 340                 if ( !isset( $cache[$code] ) ) {
 341                         // People think language codes are html safe, so enforce it.
 342                         // Ideally we should only allow a-zA-Z0-9-
 343                         // but, .+ and other chars are often used for {{int:}} hacks
 344                         // see bugs T39564, T39587, T38938
 345                         $cache[$code] =
 346                                 // Protect against path traversal
 347                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 348                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 349                 }
 350                 return $cache[$code];
 351         }
 352
 353         /**
 354          * Returns true if a language code is of a valid form for the purposes of
 355          * internal customisation of MediaWiki, via Messages*.php or *.json.
 356          *
 357          * @param string $code
 358          *
 359          * @throws MWException
 360          * @since 1.18
 361          * @return bool
 362          */
 363         public static function isValidBuiltInCode( $code ) {
 364                 if ( !is_string( $code ) ) {
 365                         if ( is_object( $code ) ) {
 366                                 $addmsg = " of class " . get_class( $code );
 367                         } else {
 368                                 $addmsg = '';
 369                         }
 370                         $type = gettype( $code );
 371                         throw new MWException( __METHOD__ . " must be passed a string, $type given$addmsg" );
 372                 }
 373
 374                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 375         }
 376
 377         /**
 378          * Returns true if a language code is an IETF tag known to MediaWiki.
 379          *
 380          * @param string $tag
 381          *
 382          * @since 1.21
 383          * @return bool
 384          */
 385         public static function isKnownLanguageTag( $tag ) {
 386                 // Quick escape for invalid input to avoid exceptions down the line
 387                 // when code tries to process tags which are not valid at all.
 388                 if ( !self::isValidBuiltInCode( $tag ) ) {
 389                         return false;
 390                 }
 391
 392                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 393                         || self::fetchLanguageName( $tag, $tag ) !== ''
 394                 ) {
 395                         return true;
 396                 }
 397
 398                 return false;
 399         }
 400
 401         /**
 402          * Get the LocalisationCache instance
 403          *
 404          * @return LocalisationCache
 405          */
 406         public static function getLocalisationCache() {
 407                 if ( is_null( self::$dataCache ) ) {
 408                         global $wgLocalisationCacheConf;
 409                         $class = $wgLocalisationCacheConf['class'];
 410                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 411                 }
 412                 return self::$dataCache;
 413         }
 414
 415         function __construct() {
 416                 $this->mConverter = new FakeConverter( $this );
 417                 // Set the code to the name of the descendant
 418                 if ( static::class === 'Language' ) {
 419                         $this->mCode = 'en';
 420                 } else {
 421                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 422                 }
 423                 self::getLocalisationCache();
 424         }
 425
 426         /**
 427          * Reduce memory usage
 428          */
 429         function __destruct() {
 430                 foreach ( $this as $name => $value ) {
 431                         unset( $this->$name );
 432                 }
 433         }
 434
 435         /**
 436          * Hook which will be called if this is the content language.
 437          * Descendants can use this to register hook functions or modify globals
 438          */
 439         function initContLang() {
 440         }
 441
 442         /**
 443          * @return array
 444          * @since 1.19
 445          */
 446         public function getFallbackLanguages() {
 447                 return self::getFallbacksFor( $this->mCode );
 448         }
 449
 450         /**
 451          * Exports $wgBookstoreListEn
 452          * @return array
 453          */
 454         public function getBookstoreList() {
 455                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 456         }
 457
 458         /**
 459          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 460          * available in localised form, it will be included in English.
 461          *
 462          * @return array
 463          */
 464         public function getNamespaces() {
 465                 if ( is_null( $this->namespaceNames ) ) {
 466                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 467
 468                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 469
 470                         $this->namespaceNames = $wgExtraNamespaces +
 471                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 472                         $this->namespaceNames += $validNamespaces;
 473
 474                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 475                         if ( $wgMetaNamespaceTalk ) {
 476                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 477                         } else {
 478                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 479                                 $this->namespaceNames[NS_PROJECT_TALK] =
 480                                         $this->fixVariableInNamespace( $talk );
 481                         }
 482
 483                         # Sometimes a language will be localised but not actually exist on this wiki.
 484                         foreach ( $this->namespaceNames as $key => $text ) {
 485                                 if ( !isset( $validNamespaces[$key] ) ) {
 486                                         unset( $this->namespaceNames[$key] );
 487                                 }
 488                         }
 489
 490                         # The above mixing may leave namespaces out of canonical order.
 491                         # Re-order by namespace ID number...
 492                         ksort( $this->namespaceNames );
 493
 494                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 495                 }
 496
 497                 return $this->namespaceNames;
 498         }
 499
 500         /**
 501          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 502          * @param array $namespaces Array of namespaces (id => name)
 503          */
 504         public function setNamespaces( array $namespaces ) {
 505                 $this->namespaceNames = $namespaces;
 506                 $this->mNamespaceIds = null;
 507         }
 508
 509         /**
 510          * Resets all of the namespace caches. Mainly used for testing
 511          */
 512         public function resetNamespaces() {
 513                 $this->namespaceNames = null;
 514                 $this->mNamespaceIds = null;
 515                 $this->namespaceAliases = null;
 516         }
 517
 518         /**
 519          * A convenience function that returns getNamespaces() with spaces instead of underscores
 520          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 521          *
 522          * @return array
 523          */
 524         public function getFormattedNamespaces() {
 525                 $ns = $this->getNamespaces();
 526                 foreach ( $ns as $k => $v ) {
 527                         $ns[$k] = strtr( $v, '_', ' ' );
 528                 }
 529                 return $ns;
 530         }
 531
 532         /**
 533          * Get a namespace value by key
 534          *
 535          * <code>
 536          * $mw_ns = $wgContLang->getNsText( NS_MEDIAWIKI );
 537          * echo $mw_ns; // prints 'MediaWiki'
 538          * </code>
 539          *
 540          * @param int $index The array key of the namespace to return
 541          * @return string|bool String if the namespace value exists, otherwise false
 542          */
 543         public function getNsText( $index ) {
 544                 $ns = $this->getNamespaces();
 545                 return isset( $ns[$index] ) ? $ns[$index] : false;
 546         }
 547
 548         /**
 549          * A convenience function that returns the same thing as
 550          * getNsText() except with '_' changed to ' ', useful for
 551          * producing output.
 552          *
 553          * <code>
 554          * $mw_ns = $wgContLang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 555          * echo $mw_ns; // prints 'MediaWiki talk'
 556          * </code>
 557          *
 558          * @param int $index The array key of the namespace to return
 559          * @return string Namespace name without underscores (empty string if namespace does not exist)
 560          */
 561         public function getFormattedNsText( $index ) {
 562                 $ns = $this->getNsText( $index );
 563                 return strtr( $ns, '_', ' ' );
 564         }
 565
 566         /**
 567          * Returns gender-dependent namespace alias if available.
 568          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 569          * @param int $index Namespace index
 570          * @param string $gender Gender key (male, female... )
 571          * @return string
 572          * @since 1.18
 573          */
 574         public function getGenderNsText( $index, $gender ) {
 575                 global $wgExtraGenderNamespaces;
 576
 577                 $ns = $wgExtraGenderNamespaces +
 578                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 579
 580                 return isset( $ns[$index][$gender] ) ? $ns[$index][$gender] : $this->getNsText( $index );
 581         }
 582
 583         /**
 584          * Whether this language uses gender-dependent namespace aliases.
 585          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 586          * @return bool
 587          * @since 1.18
 588          */
 589         public function needsGenderDistinction() {
 590                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 591                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 592                         // $wgExtraGenderNamespaces overrides everything
 593                         return true;
 594                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 595                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 596                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 597                         return false;
 598                 } else {
 599                         // Check what is in i18n files
 600                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 601                         return count( $aliases ) > 0;
 602                 }
 603         }
 604
 605         /**
 606          * Get a namespace key by value, case insensitive.
 607          * Only matches namespace names for the current language, not the
 608          * canonical ones defined in Namespace.php.
 609          *
 610          * @param string $text
 611          * @return int|bool An integer if $text is a valid value otherwise false
 612          */
 613         function getLocalNsIndex( $text ) {
 614                 $lctext = $this->lc( $text );
 615                 $ids = $this->getNamespaceIds();
 616                 return isset( $ids[$lctext] ) ? $ids[$lctext] : false;
 617         }
 618
 619         /**
 620          * @return array
 621          */
 622         public function getNamespaceAliases() {
 623                 if ( is_null( $this->namespaceAliases ) ) {
 624                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 625                         if ( !$aliases ) {
 626                                 $aliases = [];
 627                         } else {
 628                                 foreach ( $aliases as $name => $index ) {
 629                                         if ( $index === NS_PROJECT_TALK ) {
 630                                                 unset( $aliases[$name] );
 631                                                 $name = $this->fixVariableInNamespace( $name );
 632                                                 $aliases[$name] = $index;
 633                                         }
 634                                 }
 635                         }
 636
 637                         global $wgExtraGenderNamespaces;
 638                         $genders = $wgExtraGenderNamespaces +
 639                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 640                         foreach ( $genders as $index => $forms ) {
 641                                 foreach ( $forms as $alias ) {
 642                                         $aliases[$alias] = $index;
 643                                 }
 644                         }
 645
 646                         # Also add converted namespace names as aliases, to avoid confusion.
 647                         $convertedNames = [];
 648                         foreach ( $this->getVariants() as $variant ) {
 649                                 if ( $variant === $this->mCode ) {
 650                                         continue;
 651                                 }
 652                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 653                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 654                                 }
 655                         }
 656
 657                         $this->namespaceAliases = $aliases + $convertedNames;
 658                 }
 659
 660                 return $this->namespaceAliases;
 661         }
 662
 663         /**
 664          * @return array
 665          */
 666         public function getNamespaceIds() {
 667                 if ( is_null( $this->mNamespaceIds ) ) {
 668                         global $wgNamespaceAliases;
 669                         # Put namespace names and aliases into a hashtable.
 670                         # If this is too slow, then we should arrange it so that it is done
 671                         # before caching. The catch is that at pre-cache time, the above
 672                         # class-specific fixup hasn't been done.
 673                         $this->mNamespaceIds = [];
 674                         foreach ( $this->getNamespaces() as $index => $name ) {
 675                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 676                         }
 677                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 678                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 679                         }
 680                         if ( $wgNamespaceAliases ) {
 681                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 682                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 683                                 }
 684                         }
 685                 }
 686                 return $this->mNamespaceIds;
 687         }
 688
 689         /**
 690          * Get a namespace key by value, case insensitive.  Canonical namespace
 691          * names override custom ones defined for the current language.
 692          *
 693          * @param string $text
 694          * @return int|bool An integer if $text is a valid value otherwise false
 695          */
 696         public function getNsIndex( $text ) {
 697                 $lctext = $this->lc( $text );
 698                 $ns = MWNamespace::getCanonicalIndex( $lctext );
 699                 if ( $ns !== null ) {
 700                         return $ns;
 701                 }
 702                 $ids = $this->getNamespaceIds();
 703                 return isset( $ids[$lctext] ) ? $ids[$lctext] : false;
 704         }
 705
 706         /**
 707          * short names for language variants used for language conversion links.
 708          *
 709          * @param string $code
 710          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 711          * @return string
 712          */
 713         public function getVariantname( $code, $usemsg = true ) {
 714                 $msg = "variantname-$code";
 715                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 716                         return $this->getMessageFromDB( $msg );
 717                 }
 718                 $name = self::fetchLanguageName( $code );
 719                 if ( $name ) {
 720                         return $name; # if it's defined as a language name, show that
 721                 } else {
 722                         # otherwise, output the language code
 723                         return $code;
 724                 }
 725         }
 726
 727         /**
 728          * @return string[]|bool List of date format preference keys, or false if disabled.
 729          */
 730         public function getDatePreferences() {
 731                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 732         }
 733
 734         /**
 735          * @return array
 736          */
 737         function getDateFormats() {
 738                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 739         }
 740
 741         /**
 742          * @return array|string
 743          */
 744         public function getDefaultDateFormat() {
 745                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 746                 if ( $df === 'dmy or mdy' ) {
 747                         global $wgAmericanDates;
 748                         return $wgAmericanDates ? 'mdy' : 'dmy';
 749                 } else {
 750                         return $df;
 751                 }
 752         }
 753
 754         /**
 755          * @return array
 756          */
 757         public function getDatePreferenceMigrationMap() {
 758                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 759         }
 760
 761         /**
 762          * @param string $image
 763          * @return array|null
 764          */
 765         function getImageFile( $image ) {
 766                 return self::$dataCache->getSubitem( $this->mCode, 'imageFiles', $image );
 767         }
 768
 769         /**
 770          * @return array
 771          * @since 1.24
 772          */
 773         public function getImageFiles() {
 774                 return self::$dataCache->getItem( $this->mCode, 'imageFiles' );
 775         }
 776
 777         /**
 778          * @return array
 779          */
 780         public function getExtraUserToggles() {
 781                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 782         }
 783
 784         /**
 785          * @param string $tog
 786          * @return string
 787          */
 788         function getUserToggle( $tog ) {
 789                 return $this->getMessageFromDB( "tog-$tog" );
 790         }
 791
 792         /**
 793          * Get an array of language names, indexed by code.
 794          * @param null|string $inLanguage Code of language in which to return the names
 795          *              Use null for autonyms (native names)
 796          * @param string $include One of:
 797          *              'all' all available languages
 798          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 799          *              'mwfile' only if the language is in 'mw' *and* has a message file
 800          * @return array Language code => language name
 801          * @since 1.20
 802          */
 803         public static function fetchLanguageNames( $inLanguage = null, $include = 'mw' ) {
 804                 $cacheKey = $inLanguage === null ? 'null' : $inLanguage;
 805                 $cacheKey .= ":$include";
 806                 if ( self::$languageNameCache === null ) {
 807                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 808                 }
 809
 810                 $ret = self::$languageNameCache->get( $cacheKey );
 811                 if ( !$ret ) {
 812                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 813                         self::$languageNameCache->set( $cacheKey, $ret );
 814                 }
 815                 return $ret;
 816         }
 817
 818         /**
 819          * Uncached helper for fetchLanguageNames
 820          * @param null|string $inLanguage Code of language in which to return the names
 821          *              Use null for autonyms (native names)
 822          * @param string $include One of:
 823          *              'all' all available languages
 824          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 825          *              'mwfile' only if the language is in 'mw' *and* has a message file
 826          * @return array Language code => language name
 827          */
 828         private static function fetchLanguageNamesUncached( $inLanguage = null, $include = 'mw' ) {
 829                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 830
 831                 // If passed an invalid language code to use, fallback to en
 832                 if ( $inLanguage !== null && !self::isValidCode( $inLanguage ) ) {
 833                         $inLanguage = 'en';
 834                 }
 835
 836                 $names = [];
 837
 838                 if ( $inLanguage ) {
 839                         # TODO: also include when $inLanguage is null, when this code is more efficient
 840                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 841                 }
 842
 843                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 844                 if ( $wgUsePigLatinVariant ) {
 845                         // Pig Latin (for variant development)
 846                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 847                 }
 848
 849                 foreach ( $mwNames as $mwCode => $mwName ) {
 850                         # - Prefer own MediaWiki native name when not using the hook
 851                         # - For other names just add if not added through the hook
 852                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 853                                 $names[$mwCode] = $mwName;
 854                         }
 855                 }
 856
 857                 if ( $include === 'all' ) {
 858                         ksort( $names );
 859                         return $names;
 860                 }
 861
 862                 $returnMw = [];
 863                 $coreCodes = array_keys( $mwNames );
 864                 foreach ( $coreCodes as $coreCode ) {
 865                         $returnMw[$coreCode] = $names[$coreCode];
 866                 }
 867
 868                 if ( $include === 'mwfile' ) {
 869                         $namesMwFile = [];
 870                         # We do this using a foreach over the codes instead of a directory
 871                         # loop so that messages files in extensions will work correctly.
 872                         foreach ( $returnMw as $code => $value ) {
 873                                 if ( is_readable( self::getMessagesFileName( $code ) )
 874                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 875                                 ) {
 876                                         $namesMwFile[$code] = $names[$code];
 877                                 }
 878                         }
 879
 880                         ksort( $namesMwFile );
 881                         return $namesMwFile;
 882                 }
 883
 884                 ksort( $returnMw );
 885                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 886                 return $returnMw;
 887         }
 888
 889         /**
 890          * @param string $code The code of the language for which to get the name
 891          * @param null|string $inLanguage Code of language in which to return the name (null for autonyms)
 892          * @param string $include 'all', 'mw' or 'mwfile'; see fetchLanguageNames()
 893          * @return string Language name or empty
 894          * @since 1.20
 895          */
 896         public static function fetchLanguageName( $code, $inLanguage = null, $include = 'all' ) {
 897                 $code = strtolower( $code );
 898                 $array = self::fetchLanguageNames( $inLanguage, $include );
 899                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 900         }
 901
 902         /**
 903          * Get a message from the MediaWiki namespace.
 904          *
 905          * @param string $msg Message name
 906          * @return string
 907          */
 908         public function getMessageFromDB( $msg ) {
 909                 return $this->msg( $msg )->text();
 910         }
 911
 912         /**
 913          * Get message object in this language. Only for use inside this class.
 914          *
 915          * @param string $msg Message name
 916          * @return Message
 917          */
 918         protected function msg( $msg ) {
 919                 return wfMessage( $msg )->inLanguage( $this );
 920         }
 921
 922         /**
 923          * @param string $key
 924          * @return string
 925          */
 926         public function getMonthName( $key ) {
 927                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 928         }
 929
 930         /**
 931          * @return array
 932          */
 933         public function getMonthNamesArray() {
 934                 $monthNames = [ '' ];
 935                 for ( $i = 1; $i < 13; $i++ ) {
 936                         $monthNames[] = $this->getMonthName( $i );
 937                 }
 938                 return $monthNames;
 939         }
 940
 941         /**
 942          * @param string $key
 943          * @return string
 944          */
 945         public function getMonthNameGen( $key ) {
 946                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 947         }
 948
 949         /**
 950          * @param string $key
 951          * @return string
 952          */
 953         public function getMonthAbbreviation( $key ) {
 954                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 955         }
 956
 957         /**
 958          * @return array
 959          */
 960         public function getMonthAbbreviationsArray() {
 961                 $monthNames = [ '' ];
 962                 for ( $i = 1; $i < 13; $i++ ) {
 963                         $monthNames[] = $this->getMonthAbbreviation( $i );
 964                 }
 965                 return $monthNames;
 966         }
 967
 968         /**
 969          * @param string $key
 970          * @return string
 971          */
 972         public function getWeekdayName( $key ) {
 973                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
 974         }
 975
 976         /**
 977          * @param string $key
 978          * @return string
 979          */
 980         function getWeekdayAbbreviation( $key ) {
 981                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
 982         }
 983
 984         /**
 985          * @param string $key
 986          * @return string
 987          */
 988         function getIranianCalendarMonthName( $key ) {
 989                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
 990         }
 991
 992         /**
 993          * @param string $key
 994          * @return string
 995          */
 996         function getHebrewCalendarMonthName( $key ) {
 997                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
 998         }
 999
1000         /**
1001          * @param string $key
1002          * @return string
1003          */
1004         function getHebrewCalendarMonthNameGen( $key ) {
1005                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1006         }
1007
1008         /**
1009          * @param string $key
1010          * @return string
1011          */
1012         function getHijriCalendarMonthName( $key ) {
1013                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1014         }
1015
1016         /**
1017          * Pass through result from $dateTimeObj->format()
1018          * @param DateTime|bool|null &$dateTimeObj
1019          * @param string $ts
1020          * @param DateTimeZone|bool|null $zone
1021          * @param string $code
1022          * @return string
1023          */
1024         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1025                 if ( !$dateTimeObj ) {
1026                         $dateTimeObj = DateTime::createFromFormat(
1027                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1028                         );
1029                 }
1030                 return $dateTimeObj->format( $code );
1031         }
1032
1033         /**
1034          * This is a workalike of PHP's date() function, but with better
1035          * internationalisation, a reduced set of format characters, and a better
1036          * escaping format.
1037          *
1038          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1039          * the PHP manual for definitions. There are a number of extensions, which
1040          * start with "x":
1041          *
1042          *    xn   Do not translate digits of the next numeric format character
1043          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1044          *    xr   Use roman numerals for the next numeric format character
1045          *    xh   Use hebrew numerals for the next numeric format character
1046          *    xx   Literal x
1047          *    xg   Genitive month name
1048          *
1049          *    xij  j (day number) in Iranian calendar
1050          *    xiF  F (month name) in Iranian calendar
1051          *    xin  n (month number) in Iranian calendar
1052          *    xiy  y (two digit year) in Iranian calendar
1053          *    xiY  Y (full year) in Iranian calendar
1054          *    xit  t (days in month) in Iranian calendar
1055          *    xiz  z (day of the year) in Iranian calendar
1056          *
1057          *    xjj  j (day number) in Hebrew calendar
1058          *    xjF  F (month name) in Hebrew calendar
1059          *    xjt  t (days in month) in Hebrew calendar
1060          *    xjx  xg (genitive month name) in Hebrew calendar
1061          *    xjn  n (month number) in Hebrew calendar
1062          *    xjY  Y (full year) in Hebrew calendar
1063          *
1064          *    xmj  j (day number) in Hijri calendar
1065          *    xmF  F (month name) in Hijri calendar
1066          *    xmn  n (month number) in Hijri calendar
1067          *    xmY  Y (full year) in Hijri calendar
1068          *
1069          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1070          *                       identical to the Gregorian calendar
1071          *    xoY  Y (full year) in Minguo calendar or Juche year.
1072          *                       Months and days are identical to the
1073          *                       Gregorian calendar
1074          *    xtY  Y (full year) in Japanese nengo. Months and days are
1075          *                       identical to the Gregorian calendar
1076          *
1077          * Characters enclosed in double quotes will be considered literal (with
1078          * the quotes themselves removed). Unmatched quotes will be considered
1079          * literal quotes. Example:
1080          *
1081          * "The month is" F       => The month is January
1082          * i's"                   => 20'11"
1083          *
1084          * Backslash escaping is also supported.
1085          *
1086          * Input timestamp is assumed to be pre-normalized to the desired local
1087          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1088          * $ts is UTC if $zone is not given.
1089          *
1090          * @param string $format
1091          * @param string $ts 14-character timestamp
1092          *      YYYYMMDDHHMMSS
1093          *      01234567890123
1094          * @param DateTimeZone $zone Timezone of $ts
1095          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1096          * Only makes sense if $ts is the current time.
1097          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1098          *
1099          * @throws MWException
1100          * @return string
1101          */
1102         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1103                 $s = '';
1104                 $raw = false;
1105                 $roman = false;
1106                 $hebrewNum = false;
1107                 $dateTimeObj = false;
1108                 $rawToggle = false;
1109                 $iranian = false;
1110                 $hebrew = false;
1111                 $hijri = false;
1112                 $thai = false;
1113                 $minguo = false;
1114                 $tenno = false;
1115
1116                 $usedSecond = false;
1117                 $usedMinute = false;
1118                 $usedHour = false;
1119                 $usedAMPM = false;
1120                 $usedDay = false;
1121                 $usedWeek = false;
1122                 $usedMonth = false;
1123                 $usedYear = false;
1124                 $usedISOYear = false;
1125                 $usedIsLeapYear = false;
1126
1127                 $usedHebrewMonth = false;
1128                 $usedIranianMonth = false;
1129                 $usedHijriMonth = false;
1130                 $usedHebrewYear = false;
1131                 $usedIranianYear = false;
1132                 $usedHijriYear = false;
1133                 $usedTennoYear = false;
1134
1135                 if ( strlen( $ts ) !== 14 ) {
1136                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1137                 }
1138
1139                 if ( !ctype_digit( $ts ) ) {
1140                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1141                 }
1142
1143                 $formatLength = strlen( $format );
1144                 for ( $p = 0; $p < $formatLength; $p++ ) {
1145                         $num = false;
1146                         $code = $format[$p];
1147                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1148                                 $code .= $format[++$p];
1149                         }
1150
1151                         if ( ( $code === 'xi'
1152                                         || $code === 'xj'
1153                                         || $code === 'xk'
1154                                         || $code === 'xm'
1155                                         || $code === 'xo'
1156                                         || $code === 'xt' )
1157                                 && $p < $formatLength - 1 ) {
1158                                 $code .= $format[++$p];
1159                         }
1160
1161                         switch ( $code ) {
1162                                 case 'xx':
1163                                         $s .= 'x';
1164                                         break;
1165                                 case 'xn':
1166                                         $raw = true;
1167                                         break;
1168                                 case 'xN':
1169                                         $rawToggle = !$rawToggle;
1170                                         break;
1171                                 case 'xr':
1172                                         $roman = true;
1173                                         break;
1174                                 case 'xh':
1175                                         $hebrewNum = true;
1176                                         break;
1177                                 case 'xg':
1178                                         $usedMonth = true;
1179                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1180                                         break;
1181                                 case 'xjx':
1182                                         $usedHebrewMonth = true;
1183                                         if ( !$hebrew ) {
1184                                                 $hebrew = self::tsToHebrew( $ts );
1185                                         }
1186                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1187                                         break;
1188                                 case 'd':
1189                                         $usedDay = true;
1190                                         $num = substr( $ts, 6, 2 );
1191                                         break;
1192                                 case 'D':
1193                                         $usedDay = true;
1194                                         $s .= $this->getWeekdayAbbreviation(
1195                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1196                                         );
1197                                         break;
1198                                 case 'j':
1199                                         $usedDay = true;
1200                                         $num = intval( substr( $ts, 6, 2 ) );
1201                                         break;
1202                                 case 'xij':
1203                                         $usedDay = true;
1204                                         if ( !$iranian ) {
1205                                                 $iranian = self::tsToIranian( $ts );
1206                                         }
1207                                         $num = $iranian[2];
1208                                         break;
1209                                 case 'xmj':
1210                                         $usedDay = true;
1211                                         if ( !$hijri ) {
1212                                                 $hijri = self::tsToHijri( $ts );
1213                                         }
1214                                         $num = $hijri[2];
1215                                         break;
1216                                 case 'xjj':
1217                                         $usedDay = true;
1218                                         if ( !$hebrew ) {
1219                                                 $hebrew = self::tsToHebrew( $ts );
1220                                         }
1221                                         $num = $hebrew[2];
1222                                         break;
1223                                 case 'l':
1224                                         $usedDay = true;
1225                                         $s .= $this->getWeekdayName(
1226                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1227                                         );
1228                                         break;
1229                                 case 'F':
1230                                         $usedMonth = true;
1231                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1232                                         break;
1233                                 case 'xiF':
1234                                         $usedIranianMonth = true;
1235                                         if ( !$iranian ) {
1236                                                 $iranian = self::tsToIranian( $ts );
1237                                         }
1238                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1239                                         break;
1240                                 case 'xmF':
1241                                         $usedHijriMonth = true;
1242                                         if ( !$hijri ) {
1243                                                 $hijri = self::tsToHijri( $ts );
1244                                         }
1245                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1246                                         break;
1247                                 case 'xjF':
1248                                         $usedHebrewMonth = true;
1249                                         if ( !$hebrew ) {
1250                                                 $hebrew = self::tsToHebrew( $ts );
1251                                         }
1252                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1253                                         break;
1254                                 case 'm':
1255                                         $usedMonth = true;
1256                                         $num = substr( $ts, 4, 2 );
1257                                         break;
1258                                 case 'M':
1259                                         $usedMonth = true;
1260                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1261                                         break;
1262                                 case 'n':
1263                                         $usedMonth = true;
1264                                         $num = intval( substr( $ts, 4, 2 ) );
1265                                         break;
1266                                 case 'xin':
1267                                         $usedIranianMonth = true;
1268                                         if ( !$iranian ) {
1269                                                 $iranian = self::tsToIranian( $ts );
1270                                         }
1271                                         $num = $iranian[1];
1272                                         break;
1273                                 case 'xmn':
1274                                         $usedHijriMonth = true;
1275                                         if ( !$hijri ) {
1276                                                 $hijri = self::tsToHijri( $ts );
1277                                         }
1278                                         $num = $hijri[1];
1279                                         break;
1280                                 case 'xjn':
1281                                         $usedHebrewMonth = true;
1282                                         if ( !$hebrew ) {
1283                                                 $hebrew = self::tsToHebrew( $ts );
1284                                         }
1285                                         $num = $hebrew[1];
1286                                         break;
1287                                 case 'xjt':
1288                                         $usedHebrewMonth = true;
1289                                         if ( !$hebrew ) {
1290                                                 $hebrew = self::tsToHebrew( $ts );
1291                                         }
1292                                         $num = $hebrew[3];
1293                                         break;
1294                                 case 'Y':
1295                                         $usedYear = true;
1296                                         $num = substr( $ts, 0, 4 );
1297                                         break;
1298                                 case 'xiY':
1299                                         $usedIranianYear = true;
1300                                         if ( !$iranian ) {
1301                                                 $iranian = self::tsToIranian( $ts );
1302                                         }
1303                                         $num = $iranian[0];
1304                                         break;
1305                                 case 'xmY':
1306                                         $usedHijriYear = true;
1307                                         if ( !$hijri ) {
1308                                                 $hijri = self::tsToHijri( $ts );
1309                                         }
1310                                         $num = $hijri[0];
1311                                         break;
1312                                 case 'xjY':
1313                                         $usedHebrewYear = true;
1314                                         if ( !$hebrew ) {
1315                                                 $hebrew = self::tsToHebrew( $ts );
1316                                         }
1317                                         $num = $hebrew[0];
1318                                         break;
1319                                 case 'xkY':
1320                                         $usedYear = true;
1321                                         if ( !$thai ) {
1322                                                 $thai = self::tsToYear( $ts, 'thai' );
1323                                         }
1324                                         $num = $thai[0];
1325                                         break;
1326                                 case 'xoY':
1327                                         $usedYear = true;
1328                                         if ( !$minguo ) {
1329                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1330                                         }
1331                                         $num = $minguo[0];
1332                                         break;
1333                                 case 'xtY':
1334                                         $usedTennoYear = true;
1335                                         if ( !$tenno ) {
1336                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1337                                         }
1338                                         $num = $tenno[0];
1339                                         break;
1340                                 case 'y':
1341                                         $usedYear = true;
1342                                         $num = substr( $ts, 2, 2 );
1343                                         break;
1344                                 case 'xiy':
1345                                         $usedIranianYear = true;
1346                                         if ( !$iranian ) {
1347                                                 $iranian = self::tsToIranian( $ts );
1348                                         }
1349                                         $num = substr( $iranian[0], -2 );
1350                                         break;
1351                                 case 'xit':
1352                                         $usedIranianYear = true;
1353                                         if ( !$iranian ) {
1354                                                 $iranian = self::tsToIranian( $ts );
1355                                         }
1356                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1357                                         break;
1358                                 case 'xiz':
1359                                         $usedIranianYear = true;
1360                                         if ( !$iranian ) {
1361                                                 $iranian = self::tsToIranian( $ts );
1362                                         }
1363                                         $num = $iranian[3];
1364                                         break;
1365                                 case 'a':
1366                                         $usedAMPM = true;
1367                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1368                                         break;
1369                                 case 'A':
1370                                         $usedAMPM = true;
1371                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1372                                         break;
1373                                 case 'g':
1374                                         $usedHour = true;
1375                                         $h = substr( $ts, 8, 2 );
1376                                         $num = $h % 12 ? $h % 12 : 12;
1377                                         break;
1378                                 case 'G':
1379                                         $usedHour = true;
1380                                         $num = intval( substr( $ts, 8, 2 ) );
1381                                         break;
1382                                 case 'h':
1383                                         $usedHour = true;
1384                                         $h = substr( $ts, 8, 2 );
1385                                         $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
1386                                         break;
1387                                 case 'H':
1388                                         $usedHour = true;
1389                                         $num = substr( $ts, 8, 2 );
1390                                         break;
1391                                 case 'i':
1392                                         $usedMinute = true;
1393                                         $num = substr( $ts, 10, 2 );
1394                                         break;
1395                                 case 's':
1396                                         $usedSecond = true;
1397                                         $num = substr( $ts, 12, 2 );
1398                                         break;
1399                                 case 'c':
1400                                 case 'r':
1401                                         $usedSecond = true;
1402                                         // fall through
1403                                 case 'e':
1404                                 case 'O':
1405                                 case 'P':
1406                                 case 'T':
1407                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1408                                         break;
1409                                 case 'w':
1410                                 case 'N':
1411                                 case 'z':
1412                                         $usedDay = true;
1413                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1414                                         break;
1415                                 case 'W':
1416                                         $usedWeek = true;
1417                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1418                                         break;
1419                                 case 't':
1420                                         $usedMonth = true;
1421                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1422                                         break;
1423                                 case 'L':
1424                                         $usedIsLeapYear = true;
1425                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1426                                         break;
1427                                 case 'o':
1428                                         $usedISOYear = true;
1429                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1430                                         break;
1431                                 case 'U':
1432                                         $usedSecond = true;
1433                                         // fall through
1434                                 case 'I':
1435                                 case 'Z':
1436                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1437                                         break;
1438                                 case '\\':
1439                                         # Backslash escaping
1440                                         if ( $p < $formatLength - 1 ) {
1441                                                 $s .= $format[++$p];
1442                                         } else {
1443                                                 $s .= '\\';
1444                                         }
1445                                         break;
1446                                 case '"':
1447                                         # Quoted literal
1448                                         if ( $p < $formatLength - 1 ) {
1449                                                 $endQuote = strpos( $format, '"', $p + 1 );
1450                                                 if ( $endQuote === false ) {
1451                                                         # No terminating quote, assume literal "
1452                                                         $s .= '"';
1453                                                 } else {
1454                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1455                                                         $p = $endQuote;
1456                                                 }
1457                                         } else {
1458                                                 # Quote at end of string, assume literal "
1459                                                 $s .= '"';
1460                                         }
1461                                         break;
1462                                 default:
1463                                         $s .= $format[$p];
1464                         }
1465                         if ( $num !== false ) {
1466                                 if ( $rawToggle || $raw ) {
1467                                         $s .= $num;
1468                                         $raw = false;
1469                                 } elseif ( $roman ) {
1470                                         $s .= self::romanNumeral( $num );
1471                                         $roman = false;
1472                                 } elseif ( $hebrewNum ) {
1473                                         $s .= self::hebrewNumeral( $num );
1474                                         $hebrewNum = false;
1475                                 } else {
1476                                         $s .= $this->formatNum( $num, true );
1477                                 }
1478                         }
1479                 }
1480
1481                 if ( $ttl === 'unused' ) {
1482                         // No need to calculate the TTL, the caller wont use it anyway.
1483                 } elseif ( $usedSecond ) {
1484                         $ttl = 1;
1485                 } elseif ( $usedMinute ) {
1486                         $ttl = 60 - substr( $ts, 12, 2 );
1487                 } elseif ( $usedHour ) {
1488                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1489                 } elseif ( $usedAMPM ) {
1490                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1491                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1492                 } elseif (
1493                         $usedDay ||
1494                         $usedHebrewMonth ||
1495                         $usedIranianMonth ||
1496                         $usedHijriMonth ||
1497                         $usedHebrewYear ||
1498                         $usedIranianYear ||
1499                         $usedHijriYear ||
1500                         $usedTennoYear
1501                 ) {
1502                         // @todo Someone who understands the non-Gregorian calendars
1503                         // should write proper logic for them so that they don't need purged every day.
1504                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1505                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1506                 } else {
1507                         $possibleTtls = [];
1508                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1509                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1510                         if ( $usedWeek ) {
1511                                 $possibleTtls[] =
1512                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1513                                         $timeRemainingInDay;
1514                         } elseif ( $usedISOYear ) {
1515                                 // December 28th falls on the last ISO week of the year, every year.
1516                                 // The last ISO week of a year can be 52 or 53.
1517                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1518                                         'Ymd',
1519                                         substr( $ts, 0, 4 ) . '1228',
1520                                         $zone ?: new DateTimeZone( 'UTC' )
1521                                 )->format( 'W' );
1522                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1523                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1524                                 $timeRemainingInWeek =
1525                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1526                                         + $timeRemainingInDay;
1527                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1528                         }
1529
1530                         if ( $usedMonth ) {
1531                                 $possibleTtls[] =
1532                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1533                                                 substr( $ts, 6, 2 ) ) * 86400
1534                                         + $timeRemainingInDay;
1535                         } elseif ( $usedYear ) {
1536                                 $possibleTtls[] =
1537                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1538                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1539                                         + $timeRemainingInDay;
1540                         } elseif ( $usedIsLeapYear ) {
1541                                 $year = substr( $ts, 0, 4 );
1542                                 $timeRemainingInYear =
1543                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1544                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1545                                         + $timeRemainingInDay;
1546                                 $mod = $year % 4;
1547                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1548                                         // this isn't a leap year. see when the next one starts
1549                                         $nextCandidate = $year - $mod + 4;
1550                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1551                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1552                                                         $timeRemainingInYear;
1553                                         } else {
1554                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1555                                                         $timeRemainingInYear;
1556                                         }
1557                                 } else {
1558                                         // this is a leap year, so the next year isn't
1559                                         $possibleTtls[] = $timeRemainingInYear;
1560                                 }
1561                         }
1562
1563                         if ( $possibleTtls ) {
1564                                 $ttl = min( $possibleTtls );
1565                         }
1566                 }
1567
1568                 return $s;
1569         }
1570
1571         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1572         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1573
1574         /**
1575          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1576          * Gregorian dates to Iranian dates. Originally written in C, it
1577          * is released under the terms of GNU Lesser General Public
1578          * License. Conversion to PHP was performed by Niklas Laxström.
1579          *
1580          * Link: http://www.farsiweb.info/jalali/jalali.c
1581          *
1582          * @param string $ts
1583          *
1584          * @return int[]
1585          */
1586         private static function tsToIranian( $ts ) {
1587                 $gy = substr( $ts, 0, 4 ) - 1600;
1588                 $gm = substr( $ts, 4, 2 ) - 1;
1589                 $gd = substr( $ts, 6, 2 ) - 1;
1590
1591                 # Days passed from the beginning (including leap years)
1592                 $gDayNo = 365 * $gy
1593                         + floor( ( $gy + 3 ) / 4 )
1594                         - floor( ( $gy + 99 ) / 100 )
1595                         + floor( ( $gy + 399 ) / 400 );
1596
1597                 // Add days of the past months of this year
1598                 for ( $i = 0; $i < $gm; $i++ ) {
1599                         $gDayNo += self::$GREG_DAYS[$i];
1600                 }
1601
1602                 // Leap years
1603                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1604                         $gDayNo++;
1605                 }
1606
1607                 // Days passed in current month
1608                 $gDayNo += (int)$gd;
1609
1610                 $jDayNo = $gDayNo - 79;
1611
1612                 $jNp = floor( $jDayNo / 12053 );
1613                 $jDayNo %= 12053;
1614
1615                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1616                 $jDayNo %= 1461;
1617
1618                 if ( $jDayNo >= 366 ) {
1619                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1620                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1621                 }
1622
1623                 $jz = $jDayNo;
1624
1625                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1626                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1627                 }
1628
1629                 $jm = $i + 1;
1630                 $jd = $jDayNo + 1;
1631
1632                 return [ $jy, $jm, $jd, $jz ];
1633         }
1634
1635         /**
1636          * Converting Gregorian dates to Hijri dates.
1637          *
1638          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1639          *
1640          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1641          *
1642          * @param string $ts
1643          *
1644          * @return int[]
1645          */
1646         private static function tsToHijri( $ts ) {
1647                 $year = substr( $ts, 0, 4 );
1648                 $month = substr( $ts, 4, 2 );
1649                 $day = substr( $ts, 6, 2 );
1650
1651                 $zyr = $year;
1652                 $zd = $day;
1653                 $zm = $month;
1654                 $zy = $zyr;
1655
1656                 if (
1657                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1658                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1659                 ) {
1660                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1661                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1662                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1663                                         $zd - 32075;
1664                 } else {
1665                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1666                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1667                 }
1668
1669                 $zl = $zjd - 1948440 + 10632;
1670                 $zn = (int)( ( $zl - 1 ) / 10631 );
1671                 $zl = $zl - 10631 * $zn + 354;
1672                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1673                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1674                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1675                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1676                 $zm = (int)( ( 24 * $zl ) / 709 );
1677                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1678                 $zy = 30 * $zn + $zj - 30;
1679
1680                 return [ $zy, $zm, $zd ];
1681         }
1682
1683         /**
1684          * Converting Gregorian dates to Hebrew dates.
1685          *
1686          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1687          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1688          * to translate the relevant functions into PHP and release them under
1689          * GNU GPL.
1690          *
1691          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1692          * and Adar II is 14. In a non-leap year, Adar is 6.
1693          *
1694          * @param string $ts
1695          *
1696          * @return int[]
1697          */
1698         private static function tsToHebrew( $ts ) {
1699                 # Parse date
1700                 $year = substr( $ts, 0, 4 );
1701                 $month = substr( $ts, 4, 2 );
1702                 $day = substr( $ts, 6, 2 );
1703
1704                 # Calculate Hebrew year
1705                 $hebrewYear = $year + 3760;
1706
1707                 # Month number when September = 1, August = 12
1708                 $month += 4;
1709                 if ( $month > 12 ) {
1710                         # Next year
1711                         $month -= 12;
1712                         $year++;
1713                         $hebrewYear++;
1714                 }
1715
1716                 # Calculate day of year from 1 September
1717                 $dayOfYear = $day;
1718                 for ( $i = 1; $i < $month; $i++ ) {
1719                         if ( $i == 6 ) {
1720                                 # February
1721                                 $dayOfYear += 28;
1722                                 # Check if the year is leap
1723                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1724                                         $dayOfYear++;
1725                                 }
1726                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1727                                 $dayOfYear += 30;
1728                         } else {
1729                                 $dayOfYear += 31;
1730                         }
1731                 }
1732
1733                 # Calculate the start of the Hebrew year
1734                 $start = self::hebrewYearStart( $hebrewYear );
1735
1736                 # Calculate next year's start
1737                 if ( $dayOfYear <= $start ) {
1738                         # Day is before the start of the year - it is the previous year
1739                         # Next year's start
1740                         $nextStart = $start;
1741                         # Previous year
1742                         $year--;
1743                         $hebrewYear--;
1744                         # Add days since previous year's 1 September
1745                         $dayOfYear += 365;
1746                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1747                                 # Leap year
1748                                 $dayOfYear++;
1749                         }
1750                         # Start of the new (previous) year
1751                         $start = self::hebrewYearStart( $hebrewYear );
1752                 } else {
1753                         # Next year's start
1754                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1755                 }
1756
1757                 # Calculate Hebrew day of year
1758                 $hebrewDayOfYear = $dayOfYear - $start;
1759
1760                 # Difference between year's days
1761                 $diff = $nextStart - $start;
1762                 # Add 12 (or 13 for leap years) days to ignore the difference between
1763                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1764                 # difference is only about the year type
1765                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1766                         $diff += 13;
1767                 } else {
1768                         $diff += 12;
1769                 }
1770
1771                 # Check the year pattern, and is leap year
1772                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1773                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1774                 # and non-leap years
1775                 $yearPattern = $diff % 30;
1776                 # Check if leap year
1777                 $isLeap = $diff >= 30;
1778
1779                 # Calculate day in the month from number of day in the Hebrew year
1780                 # Don't check Adar - if the day is not in Adar, we will stop before;
1781                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1782                 $hebrewDay = $hebrewDayOfYear;
1783                 $hebrewMonth = 1;
1784                 $days = 0;
1785                 while ( $hebrewMonth <= 12 ) {
1786                         # Calculate days in this month
1787                         if ( $isLeap && $hebrewMonth == 6 ) {
1788                                 # Adar in a leap year
1789                                 if ( $isLeap ) {
1790                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1791                                         $days = 30;
1792                                         if ( $hebrewDay <= $days ) {
1793                                                 # Day in Adar I
1794                                                 $hebrewMonth = 13;
1795                                         } else {
1796                                                 # Subtract the days of Adar I
1797                                                 $hebrewDay -= $days;
1798                                                 # Try Adar II
1799                                                 $days = 29;
1800                                                 if ( $hebrewDay <= $days ) {
1801                                                         # Day in Adar II
1802                                                         $hebrewMonth = 14;
1803                                                 }
1804                                         }
1805                                 }
1806                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1807                                 # Cheshvan in a complete year (otherwise as the rule below)
1808                                 $days = 30;
1809                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1810                                 # Kislev in an incomplete year (otherwise as the rule below)
1811                                 $days = 29;
1812                         } else {
1813                                 # Odd months have 30 days, even have 29
1814                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1815                         }
1816                         if ( $hebrewDay <= $days ) {
1817                                 # In the current month
1818                                 break;
1819                         } else {
1820                                 # Subtract the days of the current month
1821                                 $hebrewDay -= $days;
1822                                 # Try in the next month
1823                                 $hebrewMonth++;
1824                         }
1825                 }
1826
1827                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1828         }
1829
1830         /**
1831          * This calculates the Hebrew year start, as days since 1 September.
1832          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1833          * Used for Hebrew date.
1834          *
1835          * @param int $year
1836          *
1837          * @return string
1838          */
1839         private static function hebrewYearStart( $year ) {
1840                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1841                 $b = intval( ( $year - 1 ) % 4 );
1842                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1843                 if ( $m < 0 ) {
1844                         $m--;
1845                 }
1846                 $Mar = intval( $m );
1847                 if ( $m < 0 ) {
1848                         $m++;
1849                 }
1850                 $m -= $Mar;
1851
1852                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1853                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1854                         $Mar++;
1855                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1856                         $Mar += 2;
1857                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1858                         $Mar++;
1859                 }
1860
1861                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1862                 return $Mar;
1863         }
1864
1865         /**
1866          * Algorithm to convert Gregorian dates to Thai solar dates,
1867          * Minguo dates or Minguo dates.
1868          *
1869          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1870          *       https://en.wikipedia.org/wiki/Minguo_calendar
1871          *       https://en.wikipedia.org/wiki/Japanese_era_name
1872          *
1873          * @param string $ts 14-character timestamp
1874          * @param string $cName Calender name
1875          * @return array Converted year, month, day
1876          */
1877         private static function tsToYear( $ts, $cName ) {
1878                 $gy = substr( $ts, 0, 4 );
1879                 $gm = substr( $ts, 4, 2 );
1880                 $gd = substr( $ts, 6, 2 );
1881
1882                 if ( !strcmp( $cName, 'thai' ) ) {
1883                         # Thai solar dates
1884                         # Add 543 years to the Gregorian calendar
1885                         # Months and days are identical
1886                         $gy_offset = $gy + 543;
1887                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1888                         # Minguo dates
1889                         # Deduct 1911 years from the Gregorian calendar
1890                         # Months and days are identical
1891                         $gy_offset = $gy - 1911;
1892                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1893                         # Nengō dates up to Meiji period
1894                         # Deduct years from the Gregorian calendar
1895                         # depending on the nengo periods
1896                         # Months and days are identical
1897                         if ( ( $gy < 1912 )
1898                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1899                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1900                         ) {
1901                                 # Meiji period
1902                                 $gy_gannen = $gy - 1868 + 1;
1903                                 $gy_offset = $gy_gannen;
1904                                 if ( $gy_gannen == 1 ) {
1905                                         $gy_offset = '元';
1906                                 }
1907                                 $gy_offset = '明治' . $gy_offset;
1908                         } elseif (
1909                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1910                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1911                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1912                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1913                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1914                         ) {
1915                                 # Taishō period
1916                                 $gy_gannen = $gy - 1912 + 1;
1917                                 $gy_offset = $gy_gannen;
1918                                 if ( $gy_gannen == 1 ) {
1919                                         $gy_offset = '元';
1920                                 }
1921                                 $gy_offset = '大正' . $gy_offset;
1922                         } elseif (
1923                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1924                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1925                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1926                         ) {
1927                                 # Shōwa period
1928                                 $gy_gannen = $gy - 1926 + 1;
1929                                 $gy_offset = $gy_gannen;
1930                                 if ( $gy_gannen == 1 ) {
1931                                         $gy_offset = '元';
1932                                 }
1933                                 $gy_offset = '昭和' . $gy_offset;
1934                         } else {
1935                                 # Heisei period
1936                                 $gy_gannen = $gy - 1989 + 1;
1937                                 $gy_offset = $gy_gannen;
1938                                 if ( $gy_gannen == 1 ) {
1939                                         $gy_offset = '元';
1940                                 }
1941                                 $gy_offset = '平成' . $gy_offset;
1942                         }
1943                 } else {
1944                         $gy_offset = $gy;
1945                 }
1946
1947                 return [ $gy_offset, $gm, $gd ];
1948         }
1949
1950         /**
1951          * Gets directionality of the first strongly directional codepoint, for embedBidi()
1952          *
1953          * This is the rule the BIDI algorithm uses to determine the directionality of
1954          * paragraphs ( http://unicode.org/reports/tr9/#The_Paragraph_Level ) and
1955          * FSI isolates ( http://unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
1956          *
1957          * TODO: Does not handle BIDI control characters inside the text.
1958          * TODO: Does not handle unallocated characters.
1959          *
1960          * @param string $text Text to test
1961          * @return null|string Directionality ('ltr' or 'rtl') or null
1962          */
1963         private static function strongDirFromContent( $text = '' ) {
1964                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
1965                         return null;
1966                 }
1967                 if ( $matches[1] === '' ) {
1968                         return 'rtl';
1969                 }
1970                 return 'ltr';
1971         }
1972
1973         /**
1974          * Roman number formatting up to 10000
1975          *
1976          * @param int $num
1977          *
1978          * @return string
1979          */
1980         static function romanNumeral( $num ) {
1981                 static $table = [
1982                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
1983                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
1984                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
1985                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
1986                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
1987                 ];
1988
1989                 $num = intval( $num );
1990                 if ( $num > 10000 || $num <= 0 ) {
1991                         return $num;
1992                 }
1993
1994                 $s = '';
1995                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1996                         if ( $num >= $pow10 ) {
1997                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
1998                         }
1999                         $num = $num % $pow10;
2000                 }
2001                 return $s;
2002         }
2003
2004         /**
2005          * Hebrew Gematria number formatting up to 9999
2006          *
2007          * @param int $num
2008          *
2009          * @return string
2010          */
2011         static function hebrewNumeral( $num ) {
2012                 static $table = [
2013                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2014                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2015                         [ '',
2016                                 [ 'ק' ],
2017                                 [ 'ר' ],
2018                                 [ 'ש' ],
2019                                 [ 'ת' ],
2020                                 [ 'ת', 'ק' ],
2021                                 [ 'ת', 'ר' ],
2022                                 [ 'ת', 'ש' ],
2023                                 [ 'ת', 'ת' ],
2024                                 [ 'ת', 'ת', 'ק' ],
2025                                 [ 'ת', 'ת', 'ר' ],
2026                         ],
2027                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2028                 ];
2029
2030                 $num = intval( $num );
2031                 if ( $num > 9999 || $num <= 0 ) {
2032                         return $num;
2033                 }
2034
2035                 // Round thousands have special notations
2036                 if ( $num === 1000 ) {
2037                         return "א' אלף";
2038                 } elseif ( $num % 1000 === 0 ) {
2039                         return $table[0][$num / 1000] . "' אלפים";
2040                 }
2041
2042                 $letters = [];
2043
2044                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2045                         if ( $num >= $pow10 ) {
2046                                 if ( $num === 15 || $num === 16 ) {
2047                                         $letters[] = $table[0][9];
2048                                         $letters[] = $table[0][$num - 9];
2049                                         $num = 0;
2050                                 } else {
2051                                         $letters = array_merge(
2052                                                 $letters,
2053                                                 (array)$table[$i][intval( $num / $pow10 )]
2054                                         );
2055
2056                                         if ( $pow10 === 1000 ) {
2057                                                 $letters[] = "'";
2058                                         }
2059                                 }
2060                         }
2061
2062                         $num = $num % $pow10;
2063                 }
2064
2065                 $preTransformLength = count( $letters );
2066                 if ( $preTransformLength === 1 ) {
2067                         // Add geresh (single quote) to one-letter numbers
2068                         $letters[] = "'";
2069                 } else {
2070                         $lastIndex = $preTransformLength - 1;
2071                         $letters[$lastIndex] = str_replace(
2072                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2073                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2074                                 $letters[$lastIndex]
2075                         );
2076
2077                         // Add gershayim (double quote) to multiple-letter numbers,
2078                         // but exclude numbers with only one letter after the thousands
2079                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2080                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2081                                 $letters[] = "'";
2082                         } else {
2083                                 array_splice( $letters, -1, 0, '"' );
2084                         }
2085                 }
2086
2087                 return implode( $letters );
2088         }
2089
2090         /**
2091          * Used by date() and time() to adjust the time output.
2092          *
2093          * @param string $ts The time in date('YmdHis') format
2094          * @param mixed $tz Adjust the time by this amount (default false, mean we
2095          *   get user timecorrection setting)
2096          * @return int
2097          */
2098         public function userAdjust( $ts, $tz = false ) {
2099                 global $wgUser, $wgLocalTZoffset;
2100
2101                 if ( $tz === false ) {
2102                         $tz = $wgUser->getOption( 'timecorrection' );
2103                 }
2104
2105                 $data = explode( '|', $tz, 3 );
2106
2107                 if ( $data[0] == 'ZoneInfo' ) {
2108                         try {
2109                                 $userTZ = new DateTimeZone( $data[2] );
2110                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2111                                 $date->setTimezone( $userTZ );
2112                                 return $date->format( 'YmdHis' );
2113                         } catch ( Exception $e ) {
2114                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2115                                 $data[0] = 'Offset';
2116                         }
2117                 }
2118
2119                 if ( $data[0] == 'System' || $tz == '' ) {
2120                         # Global offset in minutes.
2121                         $minDiff = $wgLocalTZoffset;
2122                 } elseif ( $data[0] == 'Offset' ) {
2123                         $minDiff = intval( $data[1] );
2124                 } else {
2125                         $data = explode( ':', $tz );
2126                         if ( count( $data ) == 2 ) {
2127                                 $data[0] = intval( $data[0] );
2128                                 $data[1] = intval( $data[1] );
2129                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2130                                 if ( $data[0] < 0 ) {
2131                                         $minDiff = -$minDiff;
2132                                 }
2133                         } else {
2134                                 $minDiff = intval( $data[0] ) * 60;
2135                         }
2136                 }
2137
2138                 # No difference ? Return time unchanged
2139                 if ( 0 == $minDiff ) {
2140                         return $ts;
2141                 }
2142
2143                 MediaWiki\suppressWarnings(); // E_STRICT system time bitching
2144                 # Generate an adjusted date; take advantage of the fact that mktime
2145                 # will normalize out-of-range values so we don't have to split $minDiff
2146                 # into hours and minutes.
2147                 $t = mktime( (
2148                         (int)substr( $ts, 8, 2 ) ), # Hours
2149                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2150                         (int)substr( $ts, 12, 2 ), # Seconds
2151                         (int)substr( $ts, 4, 2 ), # Month
2152                         (int)substr( $ts, 6, 2 ), # Day
2153                         (int)substr( $ts, 0, 4 ) ); # Year
2154
2155                 $date = date( 'YmdHis', $t );
2156                 MediaWiki\restoreWarnings();
2157
2158                 return $date;
2159         }
2160
2161         /**
2162          * This is meant to be used by time(), date(), and timeanddate() to get
2163          * the date preference they're supposed to use, it should be used in
2164          * all children.
2165          *
2166          *     function timeanddate([...], $format = true) {
2167          *       $datePreference = $this->dateFormat($format);
2168          *       [...]
2169          *     }
2170          *
2171          * @param int|string|bool $usePrefs If true, the user's preference is used
2172          *   if false, the site/language default is used
2173          *   if int/string, assumed to be a format.
2174          * @return string
2175          */
2176         function dateFormat( $usePrefs = true ) {
2177                 global $wgUser;
2178
2179                 if ( is_bool( $usePrefs ) ) {
2180                         if ( $usePrefs ) {
2181                                 $datePreference = $wgUser->getDatePreference();
2182                         } else {
2183                                 $datePreference = (string)User::getDefaultOption( 'date' );
2184                         }
2185                 } else {
2186                         $datePreference = (string)$usePrefs;
2187                 }
2188
2189                 // return int
2190                 if ( $datePreference == '' ) {
2191                         return 'default';
2192                 }
2193
2194                 return $datePreference;
2195         }
2196
2197         /**
2198          * Get a format string for a given type and preference
2199          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2200          * @param string $pref The format name as it appears in Messages*.php under
2201          *  $datePreferences.
2202          *
2203          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2204          *
2205          * @return string
2206          */
2207         function getDateFormatString( $type, $pref ) {
2208                 $wasDefault = false;
2209                 if ( $pref == 'default' ) {
2210                         $wasDefault = true;
2211                         $pref = $this->getDefaultDateFormat();
2212                 }
2213
2214                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2215                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2216
2217                         if ( $type === 'pretty' && $df === null ) {
2218                                 $df = $this->getDateFormatString( 'date', $pref );
2219                         }
2220
2221                         if ( !$wasDefault && $df === null ) {
2222                                 $pref = $this->getDefaultDateFormat();
2223                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2224                         }
2225
2226                         $this->dateFormatStrings[$type][$pref] = $df;
2227                 }
2228                 return $this->dateFormatStrings[$type][$pref];
2229         }
2230
2231         /**
2232          * @param string $ts The time format which needs to be turned into a
2233          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2234          * @param bool $adj Whether to adjust the time output according to the
2235          *   user configured offset ($timecorrection)
2236          * @param mixed $format True to use user's date format preference
2237          * @param string|bool $timecorrection The time offset as returned by
2238          *   validateTimeZone() in Special:Preferences
2239          * @return string
2240          */
2241         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2242                 $ts = wfTimestamp( TS_MW, $ts );
2243                 if ( $adj ) {
2244                         $ts = $this->userAdjust( $ts, $timecorrection );
2245                 }
2246                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2247                 return $this->sprintfDate( $df, $ts );
2248         }
2249
2250         /**
2251          * @param string $ts The time format which needs to be turned into a
2252          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2253          * @param bool $adj Whether to adjust the time output according to the
2254          *   user configured offset ($timecorrection)
2255          * @param mixed $format True to use user's date format preference
2256          * @param string|bool $timecorrection The time offset as returned by
2257          *   validateTimeZone() in Special:Preferences
2258          * @return string
2259          */
2260         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2261                 $ts = wfTimestamp( TS_MW, $ts );
2262                 if ( $adj ) {
2263                         $ts = $this->userAdjust( $ts, $timecorrection );
2264                 }
2265                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2266                 return $this->sprintfDate( $df, $ts );
2267         }
2268
2269         /**
2270          * @param string $ts The time format which needs to be turned into a
2271          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2272          * @param bool $adj Whether to adjust the time output according to the
2273          *   user configured offset ($timecorrection)
2274          * @param mixed $format What format to return, if it's false output the
2275          *   default one (default true)
2276          * @param string|bool $timecorrection The time offset as returned by
2277          *   validateTimeZone() in Special:Preferences
2278          * @return string
2279          */
2280         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2281                 $ts = wfTimestamp( TS_MW, $ts );
2282                 if ( $adj ) {
2283                         $ts = $this->userAdjust( $ts, $timecorrection );
2284                 }
2285                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2286                 return $this->sprintfDate( $df, $ts );
2287         }
2288
2289         /**
2290          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2291          *
2292          * @since 1.20
2293          *
2294          * @param int $seconds The amount of seconds.
2295          * @param array $chosenIntervals The intervals to enable.
2296          *
2297          * @return string
2298          */
2299         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2300                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2301
2302                 $segments = [];
2303
2304                 foreach ( $intervals as $intervalName => $intervalValue ) {
2305                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2306                         // duration-years, duration-decades, duration-centuries, duration-millennia
2307                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2308                         $segments[] = $message->inLanguage( $this )->escaped();
2309                 }
2310
2311                 return $this->listToText( $segments );
2312         }
2313
2314         /**
2315          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2316          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2317          *
2318          * @since 1.20
2319          *
2320          * @param int $seconds The amount of seconds.
2321          * @param array $chosenIntervals The intervals to enable.
2322          *
2323          * @return array
2324          */
2325         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2326                 if ( empty( $chosenIntervals ) ) {
2327                         $chosenIntervals = [
2328                                 'millennia',
2329                                 'centuries',
2330                                 'decades',
2331                                 'years',
2332                                 'days',
2333                                 'hours',
2334                                 'minutes',
2335                                 'seconds'
2336                         ];
2337                 }
2338
2339                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2340                 $sortedNames = array_keys( $intervals );
2341                 $smallestInterval = array_pop( $sortedNames );
2342
2343                 $segments = [];
2344
2345                 foreach ( $intervals as $name => $length ) {
2346                         $value = floor( $seconds / $length );
2347
2348                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2349                                 $seconds -= $value * $length;
2350                                 $segments[$name] = $value;
2351                         }
2352                 }
2353
2354                 return $segments;
2355         }
2356
2357         /**
2358          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2359          *
2360          * @param string $type Can be 'date', 'time' or 'both'
2361          * @param string $ts The time format which needs to be turned into a
2362          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2363          * @param User $user User object used to get preferences for timezone and format
2364          * @param array $options Array, can contain the following keys:
2365          *   - 'timecorrection': time correction, can have the following values:
2366          *     - true: use user's preference
2367          *     - false: don't use time correction
2368          *     - int: value of time correction in minutes
2369          *   - 'format': format to use, can have the following values:
2370          *     - true: use user's preference
2371          *     - false: use default preference
2372          *     - string: format to use
2373          * @since 1.19
2374          * @return string
2375          */
2376         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2377                 $ts = wfTimestamp( TS_MW, $ts );
2378                 $options += [ 'timecorrection' => true, 'format' => true ];
2379                 if ( $options['timecorrection'] !== false ) {
2380                         if ( $options['timecorrection'] === true ) {
2381                                 $offset = $user->getOption( 'timecorrection' );
2382                         } else {
2383                                 $offset = $options['timecorrection'];
2384                         }
2385                         $ts = $this->userAdjust( $ts, $offset );
2386                 }
2387                 if ( $options['format'] === true ) {
2388                         $format = $user->getDatePreference();
2389                 } else {
2390                         $format = $options['format'];
2391                 }
2392                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2393                 return $this->sprintfDate( $df, $ts );
2394         }
2395
2396         /**
2397          * Get the formatted date for the given timestamp and formatted for
2398          * the given user.
2399          *
2400          * @param mixed $ts Mixed: the time format which needs to be turned into a
2401          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2402          * @param User $user User object used to get preferences for timezone and format
2403          * @param array $options Array, can contain the following keys:
2404          *   - 'timecorrection': time correction, can have the following values:
2405          *     - true: use user's preference
2406          *     - false: don't use time correction
2407          *     - int: value of time correction in minutes
2408          *   - 'format': format to use, can have the following values:
2409          *     - true: use user's preference
2410          *     - false: use default preference
2411          *     - string: format to use
2412          * @since 1.19
2413          * @return string
2414          */
2415         public function userDate( $ts, User $user, array $options = [] ) {
2416                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2417         }
2418
2419         /**
2420          * Get the formatted time for the given timestamp and formatted for
2421          * the given user.
2422          *
2423          * @param mixed $ts The time format which needs to be turned into a
2424          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2425          * @param User $user User object used to get preferences for timezone and format
2426          * @param array $options Array, can contain the following keys:
2427          *   - 'timecorrection': time correction, can have the following values:
2428          *     - true: use user's preference
2429          *     - false: don't use time correction
2430          *     - int: value of time correction in minutes
2431          *   - 'format': format to use, can have the following values:
2432          *     - true: use user's preference
2433          *     - false: use default preference
2434          *     - string: format to use
2435          * @since 1.19
2436          * @return string
2437          */
2438         public function userTime( $ts, User $user, array $options = [] ) {
2439                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2440         }
2441
2442         /**
2443          * Get the formatted date and time for the given timestamp and formatted for
2444          * the given user.
2445          *
2446          * @param mixed $ts The time format which needs to be turned into a
2447          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2448          * @param User $user User object used to get preferences for timezone and format
2449          * @param array $options Array, can contain the following keys:
2450          *   - 'timecorrection': time correction, can have the following values:
2451          *     - true: use user's preference
2452          *     - false: don't use time correction
2453          *     - int: value of time correction in minutes
2454          *   - 'format': format to use, can have the following values:
2455          *     - true: use user's preference
2456          *     - false: use default preference
2457          *     - string: format to use
2458          * @since 1.19
2459          * @return string
2460          */
2461         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2462                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2463         }
2464
2465         /**
2466          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2467          *
2468          * Determine the difference between the timestamp and the current time, and
2469          * generate a readable timestamp by returning "<N> <units> ago", where the
2470          * largest possible unit is used.
2471          *
2472          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2473          *
2474          * @param MWTimestamp $time
2475          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2476          * @param User|null $user User the timestamp is being generated for
2477          *  (or null to use main context's user)
2478          * @return string Formatted timestamp
2479          */
2480         public function getHumanTimestamp(
2481                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2482         ) {
2483                 if ( $relativeTo === null ) {
2484                         $relativeTo = new MWTimestamp();
2485                 }
2486                 if ( $user === null ) {
2487                         $user = RequestContext::getMain()->getUser();
2488                 }
2489
2490                 // Adjust for the user's timezone.
2491                 $offsetThis = $time->offsetForUser( $user );
2492                 $offsetRel = $relativeTo->offsetForUser( $user );
2493
2494                 $ts = '';
2495                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2496                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2497                 }
2498
2499                 // Reset the timezone on the objects.
2500                 $time->timestamp->sub( $offsetThis );
2501                 $relativeTo->timestamp->sub( $offsetRel );
2502
2503                 return $ts;
2504         }
2505
2506         /**
2507          * Convert an MWTimestamp into a pretty human-readable timestamp using
2508          * the given user preferences and relative base time.
2509          *
2510          * @see Language::getHumanTimestamp
2511          * @param MWTimestamp $ts Timestamp to prettify
2512          * @param MWTimestamp $relativeTo Base timestamp
2513          * @param User $user User preferences to use
2514          * @return string Human timestamp
2515          * @since 1.26
2516          */
2517         private function getHumanTimestampInternal(
2518                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2519         ) {
2520                 $diff = $ts->diff( $relativeTo );
2521                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2522                         (int)$relativeTo->timestamp->format( 'w' ) );
2523                 $days = $diff->days ?: (int)$diffDay;
2524                 if ( $diff->invert || $days > 5
2525                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2526                 ) {
2527                         // Timestamps are in different years: use full timestamp
2528                         // Also do full timestamp for future dates
2529                         /**
2530                          * @todo FIXME: Add better handling of future timestamps.
2531                          */
2532                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2533                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2534                 } elseif ( $days > 5 ) {
2535                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2536                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2537                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2538                 } elseif ( $days > 1 ) {
2539                         // Timestamp within the past week: show the day of the week and time
2540                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2541                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2542                         // Messages:
2543                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2544                         $ts = wfMessage( "$weekday-at" )
2545                                 ->inLanguage( $this )
2546                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2547                                 ->text();
2548                 } elseif ( $days == 1 ) {
2549                         // Timestamp was yesterday: say 'yesterday' and the time.
2550                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2551                         $ts = wfMessage( 'yesterday-at' )
2552                                 ->inLanguage( $this )
2553                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2554                                 ->text();
2555                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2556                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2557                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2558                         $ts = wfMessage( 'today-at' )
2559                                 ->inLanguage( $this )
2560                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2561                                 ->text();
2562
2563                 // From here on in, the timestamp was soon enough ago so that we can simply say
2564                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2565                 } elseif ( $diff->h == 1 ) {
2566                         // Less than 90 minutes, but more than an hour ago.
2567                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2568                 } elseif ( $diff->i >= 1 ) {
2569                         // A few minutes ago.
2570                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2571                 } elseif ( $diff->s >= 30 ) {
2572                         // Less than a minute, but more than 30 sec ago.
2573                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2574                 } else {
2575                         // Less than 30 seconds ago.
2576                         $ts = wfMessage( 'just-now' )->text();
2577                 }
2578
2579                 return $ts;
2580         }
2581
2582         /**
2583          * @param string $key
2584          * @return string|null
2585          */
2586         public function getMessage( $key ) {
2587                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2588         }
2589
2590         /**
2591          * @return array
2592          */
2593         function getAllMessages() {
2594                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2595         }
2596
2597         /**
2598          * @param string $in
2599          * @param string $out
2600          * @param string $string
2601          * @return string
2602          */
2603         public function iconv( $in, $out, $string ) {
2604                 # Even with //IGNORE iconv can whine about illegal characters in
2605                 # *input* string. We just ignore those too.
2606                 # REF: https://bugs.php.net/bug.php?id=37166
2607                 # REF: https://phabricator.wikimedia.org/T18885
2608                 MediaWiki\suppressWarnings();
2609                 $text = iconv( $in, $out . '//IGNORE', $string );
2610                 MediaWiki\restoreWarnings();
2611                 return $text;
2612         }
2613
2614         // callback functions for ucwords(), ucwordbreaks()
2615
2616         /**
2617          * @param array $matches
2618          * @return mixed|string
2619          */
2620         function ucwordbreaksCallbackAscii( $matches ) {
2621                 return $this->ucfirst( $matches[1] );
2622         }
2623
2624         /**
2625          * @param array $matches
2626          * @return string
2627          */
2628         function ucwordbreaksCallbackMB( $matches ) {
2629                 return mb_strtoupper( $matches[0] );
2630         }
2631
2632         /**
2633          * @param array $matches
2634          * @return string
2635          */
2636         function ucwordsCallbackMB( $matches ) {
2637                 return mb_strtoupper( $matches[0] );
2638         }
2639
2640         /**
2641          * Make a string's first character uppercase
2642          *
2643          * @param string $str
2644          *
2645          * @return string
2646          */
2647         public function ucfirst( $str ) {
2648                 $o = ord( $str );
2649                 if ( $o < 96 ) { // if already uppercase...
2650                         return $str;
2651                 } elseif ( $o < 128 ) {
2652                         return ucfirst( $str ); // use PHP's ucfirst()
2653                 } else {
2654                         // fall back to more complex logic in case of multibyte strings
2655                         return $this->uc( $str, true );
2656                 }
2657         }
2658
2659         /**
2660          * Convert a string to uppercase
2661          *
2662          * @param string $str
2663          * @param bool $first
2664          *
2665          * @return string
2666          */
2667         public function uc( $str, $first = false ) {
2668                 if ( $first ) {
2669                         if ( $this->isMultibyte( $str ) ) {
2670                                 return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2671                         } else {
2672                                 return ucfirst( $str );
2673                         }
2674                 } else {
2675                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2676                 }
2677         }
2678
2679         /**
2680          * @param string $str
2681          * @return mixed|string
2682          */
2683         function lcfirst( $str ) {
2684                 $o = ord( $str );
2685                 if ( !$o ) {
2686                         return strval( $str );
2687                 } elseif ( $o >= 128 ) {
2688                         return $this->lc( $str, true );
2689                 } elseif ( $o > 96 ) {
2690                         return $str;
2691                 } else {
2692                         $str[0] = strtolower( $str[0] );
2693                         return $str;
2694                 }
2695         }
2696
2697         /**
2698          * @param string $str
2699          * @param bool $first
2700          * @return mixed|string
2701          */
2702         function lc( $str, $first = false ) {
2703                 if ( $first ) {
2704                         if ( $this->isMultibyte( $str ) ) {
2705                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2706                         } else {
2707                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2708                         }
2709                 } else {
2710                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2711                 }
2712         }
2713
2714         /**
2715          * @param string $str
2716          * @return bool
2717          */
2718         function isMultibyte( $str ) {
2719                 return strlen( $str ) !== mb_strlen( $str );
2720         }
2721
2722         /**
2723          * @param string $str
2724          * @return mixed|string
2725          */
2726         function ucwords( $str ) {
2727                 if ( $this->isMultibyte( $str ) ) {
2728                         $str = $this->lc( $str );
2729
2730                         // regexp to find first letter in each word (i.e. after each space)
2731                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2732
2733                         // function to use to capitalize a single char
2734                         return preg_replace_callback(
2735                                 $replaceRegexp,
2736                                 [ $this, 'ucwordsCallbackMB' ],
2737                                 $str
2738                         );
2739                 } else {
2740                         return ucwords( strtolower( $str ) );
2741                 }
2742         }
2743
2744         /**
2745          * capitalize words at word breaks
2746          *
2747          * @param string $str
2748          * @return mixed
2749          */
2750         function ucwordbreaks( $str ) {
2751                 if ( $this->isMultibyte( $str ) ) {
2752                         $str = $this->lc( $str );
2753
2754                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2755                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2756
2757                         // find first letter after word break
2758                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2759                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2760
2761                         return preg_replace_callback(
2762                                 $replaceRegexp,
2763                                 [ $this, 'ucwordbreaksCallbackMB' ],
2764                                 $str
2765                         );
2766                 } else {
2767                         return preg_replace_callback(
2768                                 '/\b([\w\x80-\xff]+)\b/',
2769                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2770                                 $str
2771                         );
2772                 }
2773         }
2774
2775         /**
2776          * Return a case-folded representation of $s
2777          *
2778          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2779          * and $s2 are the same except for the case of their characters. It is not
2780          * necessary for the value returned to make sense when displayed.
2781          *
2782          * Do *not* perform any other normalisation in this function. If a caller
2783          * uses this function when it should be using a more general normalisation
2784          * function, then fix the caller.
2785          *
2786          * @param string $s
2787          *
2788          * @return string
2789          */
2790         function caseFold( $s ) {
2791                 return $this->uc( $s );
2792         }
2793
2794         /**
2795          * @param string $s
2796          * @return string
2797          * @throws MWException
2798          */
2799         function checkTitleEncoding( $s ) {
2800                 if ( is_array( $s ) ) {
2801                         throw new MWException( 'Given array to checkTitleEncoding.' );
2802                 }
2803                 if ( StringUtils::isUtf8( $s ) ) {
2804                         return $s;
2805                 }
2806
2807                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2808         }
2809
2810         /**
2811          * @return array
2812          */
2813         function fallback8bitEncoding() {
2814                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2815         }
2816
2817         /**
2818          * Most writing systems use whitespace to break up words.
2819          * Some languages such as Chinese don't conventionally do this,
2820          * which requires special handling when breaking up words for
2821          * searching etc.
2822          *
2823          * @return bool
2824          */
2825         function hasWordBreaks() {
2826                 return true;
2827         }
2828
2829         /**
2830          * Some languages such as Chinese require word segmentation,
2831          * Specify such segmentation when overridden in derived class.
2832          *
2833          * @param string $string
2834          * @return string
2835          */
2836         function segmentByWord( $string ) {
2837                 return $string;
2838         }
2839
2840         /**
2841          * Some languages have special punctuation need to be normalized.
2842          * Make such changes here.
2843          *
2844          * @param string $string
2845          * @return string
2846          */
2847         function normalizeForSearch( $string ) {
2848                 return self::convertDoubleWidth( $string );
2849         }
2850
2851         /**
2852          * convert double-width roman characters to single-width.
2853          * range: ff00-ff5f ~= 0020-007f
2854          *
2855          * @param string $string
2856          *
2857          * @return string
2858          */
2859         protected static function convertDoubleWidth( $string ) {
2860                 static $full = null;
2861                 static $half = null;
2862
2863                 if ( $full === null ) {
2864                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2865                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2866                         $full = str_split( $fullWidth, 3 );
2867                         $half = str_split( $halfWidth );
2868                 }
2869
2870                 $string = str_replace( $full, $half, $string );
2871                 return $string;
2872         }
2873
2874         /**
2875          * @param string $string
2876          * @param string $pattern
2877          * @return string
2878          */
2879         protected static function insertSpace( $string, $pattern ) {
2880                 $string = preg_replace( $pattern, " $1 ", $string );
2881                 $string = preg_replace( '/ +/', ' ', $string );
2882                 return $string;
2883         }
2884
2885         /**
2886          * @param array $termsArray
2887          * @return array
2888          */
2889         function convertForSearchResult( $termsArray ) {
2890                 # some languages, e.g. Chinese, need to do a conversion
2891                 # in order for search results to be displayed correctly
2892                 return $termsArray;
2893         }
2894
2895         /**
2896          * Get the first character of a string.
2897          *
2898          * @param string $s
2899          * @return string
2900          */
2901         function firstChar( $s ) {
2902                 $matches = [];
2903                 preg_match(
2904                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2905                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2906                         $s,
2907                         $matches
2908                 );
2909
2910                 if ( isset( $matches[1] ) ) {
2911                         if ( strlen( $matches[1] ) != 3 ) {
2912                                 return $matches[1];
2913                         }
2914
2915                         // Break down Hangul syllables to grab the first jamo
2916                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2917                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2918                                 return $matches[1];
2919                         } elseif ( $code < 0xb098 ) {
2920                                 return "\xe3\x84\xb1";
2921                         } elseif ( $code < 0xb2e4 ) {
2922                                 return "\xe3\x84\xb4";
2923                         } elseif ( $code < 0xb77c ) {
2924                                 return "\xe3\x84\xb7";
2925                         } elseif ( $code < 0xb9c8 ) {
2926                                 return "\xe3\x84\xb9";
2927                         } elseif ( $code < 0xbc14 ) {
2928                                 return "\xe3\x85\x81";
2929                         } elseif ( $code < 0xc0ac ) {
2930                                 return "\xe3\x85\x82";
2931                         } elseif ( $code < 0xc544 ) {
2932                                 return "\xe3\x85\x85";
2933                         } elseif ( $code < 0xc790 ) {
2934                                 return "\xe3\x85\x87";
2935                         } elseif ( $code < 0xcc28 ) {
2936                                 return "\xe3\x85\x88";
2937                         } elseif ( $code < 0xce74 ) {
2938                                 return "\xe3\x85\x8a";
2939                         } elseif ( $code < 0xd0c0 ) {
2940                                 return "\xe3\x85\x8b";
2941                         } elseif ( $code < 0xd30c ) {
2942                                 return "\xe3\x85\x8c";
2943                         } elseif ( $code < 0xd558 ) {
2944                                 return "\xe3\x85\x8d";
2945                         } else {
2946                                 return "\xe3\x85\x8e";
2947                         }
2948                 } else {
2949                         return '';
2950                 }
2951         }
2952
2953         /**
2954          * @deprecated No-op since 1.28
2955          */
2956         function initEncoding() {
2957                 // No-op.
2958         }
2959
2960         /**
2961          * @param string $s
2962          * @return string
2963          * @deprecated No-op since 1.28
2964          */
2965         function recodeForEdit( $s ) {
2966                 return $s;
2967         }
2968
2969         /**
2970          * @param string $s
2971          * @return string
2972          * @deprecated No-op since 1.28
2973          */
2974         function recodeInput( $s ) {
2975                 return $s;
2976         }
2977
2978         /**
2979          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
2980          * also cleans up certain backwards-compatible sequences, converting them
2981          * to the modern Unicode equivalent.
2982          *
2983          * This is language-specific for performance reasons only.
2984          *
2985          * @param string $s
2986          *
2987          * @return string
2988          */
2989         function normalize( $s ) {
2990                 global $wgAllUnicodeFixes;
2991                 $s = UtfNormal\Validator::cleanUp( $s );
2992                 if ( $wgAllUnicodeFixes ) {
2993                         $s = $this->transformUsingPairFile( 'normalize-ar.ser', $s );
2994                         $s = $this->transformUsingPairFile( 'normalize-ml.ser', $s );
2995                 }
2996
2997                 return $s;
2998         }
2999
3000         /**
3001          * Transform a string using serialized data stored in the given file (which
3002          * must be in the serialized subdirectory of $IP). The file contains pairs
3003          * mapping source characters to destination characters.
3004          *
3005          * The data is cached in process memory. This will go faster if you have the
3006          * FastStringSearch extension.
3007          *
3008          * @param string $file
3009          * @param string $string
3010          *
3011          * @throws MWException
3012          * @return string
3013          */
3014         function transformUsingPairFile( $file, $string ) {
3015                 if ( !isset( $this->transformData[$file] ) ) {
3016                         $data = wfGetPrecompiledData( $file );
3017                         if ( $data === false ) {
3018                                 throw new MWException( __METHOD__ . ": The transformation file $file is missing" );
3019                         }
3020                         $this->transformData[$file] = new ReplacementArray( $data );
3021                 }
3022                 return $this->transformData[$file]->replace( $string );
3023         }
3024
3025         /**
3026          * For right-to-left language support
3027          *
3028          * @return bool
3029          */
3030         function isRTL() {
3031                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3032         }
3033
3034         /**
3035          * Return the correct HTML 'dir' attribute value for this language.
3036          * @return string
3037          */
3038         function getDir() {
3039                 return $this->isRTL() ? 'rtl' : 'ltr';
3040         }
3041
3042         /**
3043          * Return 'left' or 'right' as appropriate alignment for line-start
3044          * for this language's text direction.
3045          *
3046          * Should be equivalent to CSS3 'start' text-align value....
3047          *
3048          * @return string
3049          */
3050         function alignStart() {
3051                 return $this->isRTL() ? 'right' : 'left';
3052         }
3053
3054         /**
3055          * Return 'right' or 'left' as appropriate alignment for line-end
3056          * for this language's text direction.
3057          *
3058          * Should be equivalent to CSS3 'end' text-align value....
3059          *
3060          * @return string
3061          */
3062         function alignEnd() {
3063                 return $this->isRTL() ? 'left' : 'right';
3064         }
3065
3066         /**
3067          * A hidden direction mark (LRM or RLM), depending on the language direction.
3068          * Unlike getDirMark(), this function returns the character as an HTML entity.
3069          * This function should be used when the output is guaranteed to be HTML,
3070          * because it makes the output HTML source code more readable. When
3071          * the output is plain text or can be escaped, getDirMark() should be used.
3072          *
3073          * @param bool $opposite Get the direction mark opposite to your language
3074          * @return string
3075          * @since 1.20
3076          */
3077         function getDirMarkEntity( $opposite = false ) {
3078                 if ( $opposite ) {
3079                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3080                 }
3081                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3082         }
3083
3084         /**
3085          * A hidden direction mark (LRM or RLM), depending on the language direction.
3086          * This function produces them as invisible Unicode characters and
3087          * the output may be hard to read and debug, so it should only be used
3088          * when the output is plain text or can be escaped. When the output is
3089          * HTML, use getDirMarkEntity() instead.
3090          *
3091          * @param bool $opposite Get the direction mark opposite to your language
3092          * @return string
3093          */
3094         function getDirMark( $opposite = false ) {
3095                 $lrm = "\xE2\x80\x8E"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3096                 $rlm = "\xE2\x80\x8F"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3097                 if ( $opposite ) {
3098                         return $this->isRTL() ? $lrm : $rlm;
3099                 }
3100                 return $this->isRTL() ? $rlm : $lrm;
3101         }
3102
3103         /**
3104          * @return array
3105          */
3106         function capitalizeAllNouns() {
3107                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3108         }
3109
3110         /**
3111          * An arrow, depending on the language direction.
3112          *
3113          * @param string $direction The direction of the arrow: forwards (default),
3114          *   backwards, left, right, up, down.
3115          * @return string
3116          */
3117         function getArrow( $direction = 'forwards' ) {
3118                 switch ( $direction ) {
3119                 case 'forwards':
3120                         return $this->isRTL() ? '←' : '→';
3121                 case 'backwards':
3122                         return $this->isRTL() ? '→' : '←';
3123                 case 'left':
3124                         return '←';
3125                 case 'right':
3126                         return '→';
3127                 case 'up':
3128                         return '↑';
3129                 case 'down':
3130                         return '↓';
3131                 }
3132         }
3133
3134         /**
3135          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3136          *
3137          * @return bool
3138          */
3139         function linkPrefixExtension() {
3140                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3141         }
3142
3143         /**
3144          * Get all magic words from cache.
3145          * @return array
3146          */
3147         function getMagicWords() {
3148                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3149         }
3150
3151         /**
3152          * Run the LanguageGetMagic hook once.
3153          */
3154         protected function doMagicHook() {
3155                 if ( $this->mMagicHookDone ) {
3156                         return;
3157                 }
3158                 $this->mMagicHookDone = true;
3159                 Hooks::run( 'LanguageGetMagic', [ &$this->mMagicExtensions, $this->getCode() ] );
3160         }
3161
3162         /**
3163          * Fill a MagicWord object with data from here
3164          *
3165          * @param MagicWord $mw
3166          */
3167         function getMagic( $mw ) {
3168                 // Saves a function call
3169                 if ( !$this->mMagicHookDone ) {
3170                         $this->doMagicHook();
3171                 }
3172
3173                 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
3174                         $rawEntry = $this->mMagicExtensions[$mw->mId];
3175                 } else {
3176                         $rawEntry = self::$dataCache->getSubitem(
3177                                 $this->mCode, 'magicWords', $mw->mId );
3178                 }
3179
3180                 if ( !is_array( $rawEntry ) ) {
3181                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3182                 } else {
3183                         $mw->mCaseSensitive = $rawEntry[0];
3184                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3185                 }
3186         }
3187
3188         /**
3189          * Add magic words to the extension array
3190          *
3191          * @param array $newWords
3192          */
3193         function addMagicWordsByLang( $newWords ) {
3194                 $fallbackChain = $this->getFallbackLanguages();
3195                 $fallbackChain = array_reverse( $fallbackChain );
3196                 foreach ( $fallbackChain as $code ) {
3197                         if ( isset( $newWords[$code] ) ) {
3198                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3199                         }
3200                 }
3201         }
3202
3203         /**
3204          * Get special page names, as an associative array
3205          *   canonical name => array of valid names, including aliases
3206          * @return array
3207          */
3208         function getSpecialPageAliases() {
3209                 // Cache aliases because it may be slow to load them
3210                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3211                         // Initialise array
3212                         $this->mExtendedSpecialPageAliases =
3213                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3214                         Hooks::run( 'LanguageGetSpecialPageAliases',
3215                                 [ &$this->mExtendedSpecialPageAliases, $this->getCode() ] );
3216                 }
3217
3218                 return $this->mExtendedSpecialPageAliases;
3219         }
3220
3221         /**
3222          * Italic is unsuitable for some languages
3223          *
3224          * @param string $text The text to be emphasized.
3225          * @return string
3226          */
3227         function emphasize( $text ) {
3228                 return "<em>$text</em>";
3229         }
3230
3231         /**
3232          * Normally we output all numbers in plain en_US style, that is
3233          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3234          * point twohundredthirtyfive. However this is not suitable for all
3235          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3236          * Icelandic just want to use commas instead of dots, and dots instead
3237          * of commas like "293.291,235".
3238          *
3239          * An example of this function being called:
3240          * <code>
3241          * wfMessage( 'message' )->numParams( $num )->text()
3242          * </code>
3243          *
3244          * See $separatorTransformTable on MessageIs.php for
3245          * the , => . and . => , implementation.
3246          *
3247          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3248          * @param int|float $number The string to be formatted, should be an integer
3249          *   or a floating point number.
3250          * @param bool $nocommafy Set to true for special numbers like dates
3251          * @return string
3252          */
3253         public function formatNum( $number, $nocommafy = false ) {
3254                 global $wgTranslateNumerals;
3255                 if ( !$nocommafy ) {
3256                         $number = $this->commafy( $number );
3257                         $s = $this->separatorTransformTable();
3258                         if ( $s ) {
3259                                 $number = strtr( $number, $s );
3260                         }
3261                 }
3262
3263                 if ( $wgTranslateNumerals ) {
3264                         $s = $this->digitTransformTable();
3265                         if ( $s ) {
3266                                 $number = strtr( $number, $s );
3267                         }
3268                 }
3269
3270                 return $number;
3271         }
3272
3273         /**
3274          * Front-end for non-commafied formatNum
3275          *
3276          * @param int|float $number The string to be formatted, should be an integer
3277          *        or a floating point number.
3278          * @since 1.21
3279          * @return string
3280          */
3281         public function formatNumNoSeparators( $number ) {
3282                 return $this->formatNum( $number, true );
3283         }
3284
3285         /**
3286          * @param string $number
3287          * @return string
3288          */
3289         public function parseFormattedNumber( $number ) {
3290                 $s = $this->digitTransformTable();
3291                 if ( $s ) {
3292                         // eliminate empty array values such as ''. (T66347)
3293                         $s = array_filter( $s );
3294                         $number = strtr( $number, array_flip( $s ) );
3295                 }
3296
3297                 $s = $this->separatorTransformTable();
3298                 if ( $s ) {
3299                         // eliminate empty array values such as ''. (T66347)
3300                         $s = array_filter( $s );
3301                         $number = strtr( $number, array_flip( $s ) );
3302                 }
3303
3304                 $number = strtr( $number, [ ',' => '' ] );
3305                 return $number;
3306         }
3307
3308         /**
3309          * Adds commas to a given number
3310          * @since 1.19
3311          * @param mixed $number
3312          * @return string
3313          */
3314         function commafy( $number ) {
3315                 $digitGroupingPattern = $this->digitGroupingPattern();
3316                 if ( $number === null ) {
3317                         return '';
3318                 }
3319
3320                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3321                         // default grouping is at thousands,  use the same for ###,###,### pattern too.
3322                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3323                 } else {
3324                         // Ref: http://cldr.unicode.org/translation/number-patterns
3325                         $sign = "";
3326                         if ( intval( $number ) < 0 ) {
3327                                 // For negative numbers apply the algorithm like positive number and add sign.
3328                                 $sign = "-";
3329                                 $number = substr( $number, 1 );
3330                         }
3331                         $integerPart = [];
3332                         $decimalPart = [];
3333                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3334                         preg_match( "/\d+/", $number, $integerPart );
3335                         preg_match( "/\.\d*/", $number, $decimalPart );
3336                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3337                         if ( $groupedNumber === $number ) {
3338                                 // the string does not have any number part. Eg: .12345
3339                                 return $sign . $groupedNumber;
3340                         }
3341                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3342                         while ( $start > 0 ) {
3343                                 $match = $matches[0][$numMatches - 1];
3344                                 $matchLen = strlen( $match );
3345                                 $start = $end - $matchLen;
3346                                 if ( $start < 0 ) {
3347                                         $start = 0;
3348                                 }
3349                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3350                                 $end = $start;
3351                                 if ( $numMatches > 1 ) {
3352                                         // use the last pattern for the rest of the number
3353                                         $numMatches--;
3354                                 }
3355                                 if ( $start > 0 ) {
3356                                         $groupedNumber = "," . $groupedNumber;
3357                                 }
3358                         }
3359                         return $sign . $groupedNumber;
3360                 }
3361         }
3362
3363         /**
3364          * @return string
3365          */
3366         function digitGroupingPattern() {
3367                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3368         }
3369
3370         /**
3371          * @return array
3372          */
3373         function digitTransformTable() {
3374                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3375         }
3376
3377         /**
3378          * @return array
3379          */
3380         function separatorTransformTable() {
3381                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3382         }
3383
3384         /**
3385          * Take a list of strings and build a locale-friendly comma-separated
3386          * list, using the local comma-separator message.
3387          * The last two strings are chained with an "and".
3388          * NOTE: This function will only work with standard numeric array keys (0, 1, 2…)
3389          *
3390          * @param string[] $l
3391          * @return string
3392          */
3393         function listToText( array $l ) {
3394                 $m = count( $l ) - 1;
3395                 if ( $m < 0 ) {
3396                         return '';
3397                 }
3398                 if ( $m > 0 ) {
3399                         $and = $this->msg( 'and' )->escaped();
3400                         $space = $this->msg( 'word-separator' )->escaped();
3401                         if ( $m > 1 ) {
3402                                 $comma = $this->msg( 'comma-separator' )->escaped();
3403                         }
3404                 }
3405                 $s = $l[$m];
3406                 for ( $i = $m - 1; $i >= 0; $i-- ) {
3407                         if ( $i == $m - 1 ) {
3408                                 $s = $l[$i] . $and . $space . $s;
3409                         } else {
3410                                 $s = $l[$i] . $comma . $s;
3411                         }
3412                 }
3413                 return $s;
3414         }
3415
3416         /**
3417          * Take a list of strings and build a locale-friendly comma-separated
3418          * list, using the local comma-separator message.
3419          * @param string[] $list Array of strings to put in a comma list
3420          * @return string
3421          */
3422         function commaList( array $list ) {
3423                 return implode(
3424                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3425                         $list
3426                 );
3427         }
3428
3429         /**
3430          * Take a list of strings and build a locale-friendly semicolon-separated
3431          * list, using the local semicolon-separator message.
3432          * @param string[] $list Array of strings to put in a semicolon list
3433          * @return string
3434          */
3435         function semicolonList( array $list ) {
3436                 return implode(
3437                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3438                         $list
3439                 );
3440         }
3441
3442         /**
3443          * Same as commaList, but separate it with the pipe instead.
3444          * @param string[] $list Array of strings to put in a pipe list
3445          * @return string
3446          */
3447         function pipeList( array $list ) {
3448                 return implode(
3449                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3450                         $list
3451                 );
3452         }
3453
3454         /**
3455          * Truncate a string to a specified length in bytes, appending an optional
3456          * string (e.g. for ellipses)
3457          *
3458          * The database offers limited byte lengths for some columns in the database;
3459          * multi-byte character sets mean we need to ensure that only whole characters
3460          * are included, otherwise broken characters can be passed to the user
3461          *
3462          * If $length is negative, the string will be truncated from the beginning
3463          *
3464          * @param string $string String to truncate
3465          * @param int $length Maximum length (including ellipses)
3466          * @param string $ellipsis String to append to the truncated text
3467          * @param bool $adjustLength Subtract length of ellipsis from $length.
3468          *      $adjustLength was introduced in 1.18, before that behaved as if false.
3469          * @return string
3470          */
3471         function truncate( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3472                 # Use the localized ellipsis character
3473                 if ( $ellipsis == '...' ) {
3474                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3475                 }
3476                 # Check if there is no need to truncate
3477                 if ( $length == 0 ) {
3478                         return $ellipsis; // convention
3479                 } elseif ( strlen( $string ) <= abs( $length ) ) {
3480                         return $string; // no need to truncate
3481                 }
3482                 $stringOriginal = $string;
3483                 # If ellipsis length is >= $length then we can't apply $adjustLength
3484                 if ( $adjustLength && strlen( $ellipsis ) >= abs( $length ) ) {
3485                         $string = $ellipsis; // this can be slightly unexpected
3486                 # Otherwise, truncate and add ellipsis...
3487                 } else {
3488                         $eLength = $adjustLength ? strlen( $ellipsis ) : 0;
3489                         if ( $length > 0 ) {
3490                                 $length -= $eLength;
3491                                 $string = substr( $string, 0, $length ); // xyz...
3492                                 $string = $this->removeBadCharLast( $string );
3493                                 $string = rtrim( $string );
3494                                 $string = $string . $ellipsis;
3495                         } else {
3496                                 $length += $eLength;
3497                                 $string = substr( $string, $length ); // ...xyz
3498                                 $string = $this->removeBadCharFirst( $string );
3499                                 $string = ltrim( $string );
3500                                 $string = $ellipsis . $string;
3501                         }
3502                 }
3503                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3504                 # This check is *not* redundant if $adjustLength, due to the single case where
3505                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3506                 if ( strlen( $string ) < strlen( $stringOriginal ) ) {
3507                         return $string;
3508                 } else {
3509                         return $stringOriginal;
3510                 }
3511         }
3512
3513         /**
3514          * Remove bytes that represent an incomplete Unicode character
3515          * at the end of string (e.g. bytes of the char are missing)
3516          *
3517          * @param string $string
3518          * @return string
3519          */
3520         protected function removeBadCharLast( $string ) {
3521                 if ( $string != '' ) {
3522                         $char = ord( $string[strlen( $string ) - 1] );
3523                         $m = [];
3524                         if ( $char >= 0xc0 ) {
3525                                 # We got the first byte only of a multibyte char; remove it.
3526                                 $string = substr( $string, 0, -1 );
3527                         } elseif ( $char >= 0x80 &&
3528                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3529                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3530                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3531                         ) {
3532                                 # We chopped in the middle of a character; remove it
3533                                 $string = $m[1];
3534                         }
3535                 }
3536                 return $string;
3537         }
3538
3539         /**
3540          * Remove bytes that represent an incomplete Unicode character
3541          * at the start of string (e.g. bytes of the char are missing)
3542          *
3543          * @param string $string
3544          * @return string
3545          */
3546         protected function removeBadCharFirst( $string ) {
3547                 if ( $string != '' ) {
3548                         $char = ord( $string[0] );
3549                         if ( $char >= 0x80 && $char < 0xc0 ) {
3550                                 # We chopped in the middle of a character; remove the whole thing
3551                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3552                         }
3553                 }
3554                 return $string;
3555         }
3556
3557         /**
3558          * Truncate a string of valid HTML to a specified length in bytes,
3559          * appending an optional string (e.g. for ellipses), and return valid HTML
3560          *
3561          * This is only intended for styled/linked text, such as HTML with
3562          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3563          * Also, this will not detect things like "display:none" CSS.
3564          *
3565          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3566          *
3567          * @param string $text HTML string to truncate
3568          * @param int $length (zero/positive) Maximum length (including ellipses)
3569          * @param string $ellipsis String to append to the truncated text
3570          * @return string
3571          */
3572         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3573                 # Use the localized ellipsis character
3574                 if ( $ellipsis == '...' ) {
3575                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3576                 }
3577                 # Check if there is clearly no need to truncate
3578                 if ( $length <= 0 ) {
3579                         return $ellipsis; // no text shown, nothing to format (convention)
3580                 } elseif ( strlen( $text ) <= $length ) {
3581                         return $text; // string short enough even *with* HTML (short-circuit)
3582                 }
3583
3584                 $dispLen = 0; // innerHTML legth so far
3585                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3586                 $tagType = 0; // 0-open, 1-close
3587                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3588                 $entityState = 0; // 0-not entity, 1-entity
3589                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3590                 $openTags = []; // open tag stack
3591                 $maybeState = null; // possible truncation state
3592
3593                 $textLen = strlen( $text );
3594                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3595                 for ( $pos = 0; true; ++$pos ) {
3596                         # Consider truncation once the display length has reached the maximim.
3597                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3598                         # Check that we're not in the middle of a bracket/entity...
3599                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3600                                 if ( !$testingEllipsis ) {
3601                                         $testingEllipsis = true;
3602                                         # Save where we are; we will truncate here unless there turn out to
3603                                         # be so few remaining characters that truncation is not necessary.
3604                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3605                                                 $maybeState = [ $ret, $openTags ]; // save state
3606                                         }
3607                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3608                                         # String in fact does need truncation, the truncation point was OK.
3609                                         list( $ret, $openTags ) = $maybeState; // reload state
3610                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3611                                         $ret .= $ellipsis; // add ellipsis
3612                                         break;
3613                                 }
3614                         }
3615                         if ( $pos >= $textLen ) {
3616                                 break; // extra iteration just for above checks
3617                         }
3618
3619                         # Read the next char...
3620                         $ch = $text[$pos];
3621                         $lastCh = $pos ? $text[$pos - 1] : '';
3622                         $ret .= $ch; // add to result string
3623                         if ( $ch == '<' ) {
3624                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3625                                 $entityState = 0; // for bad HTML
3626                                 $bracketState = 1; // tag started (checking for backslash)
3627                         } elseif ( $ch == '>' ) {
3628                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3629                                 $entityState = 0; // for bad HTML
3630                                 $bracketState = 0; // out of brackets
3631                         } elseif ( $bracketState == 1 ) {
3632                                 if ( $ch == '/' ) {
3633                                         $tagType = 1; // close tag (e.g. "</span>")
3634                                 } else {
3635                                         $tagType = 0; // open tag (e.g. "<span>")
3636                                         $tag .= $ch;
3637                                 }
3638                                 $bracketState = 2; // building tag name
3639                         } elseif ( $bracketState == 2 ) {
3640                                 if ( $ch != ' ' ) {
3641                                         $tag .= $ch;
3642                                 } else {
3643                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3644                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3645                                 }
3646                         } elseif ( $bracketState == 0 ) {
3647                                 if ( $entityState ) {
3648                                         if ( $ch == ';' ) {
3649                                                 $entityState = 0;
3650                                                 $dispLen++; // entity is one displayed char
3651                                         }
3652                                 } else {
3653                                         if ( $neLength == 0 && !$maybeState ) {
3654                                                 // Save state without $ch. We want to *hit* the first
3655                                                 // display char (to get tags) but not *use* it if truncating.
3656                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3657                                         }
3658                                         if ( $ch == '&' ) {
3659                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3660                                         } else {
3661                                                 $dispLen++; // this char is displayed
3662                                                 // Add the next $max display text chars after this in one swoop...
3663                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3664                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3665                                                 $dispLen += $skipped;
3666                                                 $pos += $skipped;
3667                                         }
3668                                 }
3669                         }
3670                 }
3671                 // Close the last tag if left unclosed by bad HTML
3672                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3673                 while ( count( $openTags ) > 0 ) {
3674                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3675                 }
3676                 return $ret;
3677         }
3678
3679         /**
3680          * truncateHtml() helper function
3681          * like strcspn() but adds the skipped chars to $ret
3682          *
3683          * @param string $ret
3684          * @param string $text
3685          * @param string $search
3686          * @param int $start
3687          * @param null|int $len
3688          * @return int
3689          */
3690         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3691                 if ( $len === null ) {
3692                         $len = -1; // -1 means "no limit" for strcspn
3693                 } elseif ( $len < 0 ) {
3694                         $len = 0; // sanity
3695                 }
3696                 $skipCount = 0;
3697                 if ( $start < strlen( $text ) ) {
3698                         $skipCount = strcspn( $text, $search, $start, $len );
3699                         $ret .= substr( $text, $start, $skipCount );
3700                 }
3701                 return $skipCount;
3702         }
3703
3704         /**
3705          * truncateHtml() helper function
3706          * (a) push or pop $tag from $openTags as needed
3707          * (b) clear $tag value
3708          * @param string &$tag Current HTML tag name we are looking at
3709          * @param int $tagType (0-open tag, 1-close tag)
3710          * @param string $lastCh Character before the '>' that ended this tag
3711          * @param array &$openTags Open tag stack (not accounting for $tag)
3712          */
3713         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3714                 $tag = ltrim( $tag );
3715                 if ( $tag != '' ) {
3716                         if ( $tagType == 0 && $lastCh != '/' ) {
3717                                 $openTags[] = $tag; // tag opened (didn't close itself)
3718                         } elseif ( $tagType == 1 ) {
3719                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3720                                         array_pop( $openTags ); // tag closed
3721                                 }
3722                         }
3723                         $tag = '';
3724                 }
3725         }
3726
3727         /**
3728          * Grammatical transformations, needed for inflected languages
3729          * Invoked by putting {{grammar:case|word}} in a message
3730          *
3731          * @param string $word
3732          * @param string $case
3733          * @return string
3734          */
3735         function convertGrammar( $word, $case ) {
3736                 global $wgGrammarForms;
3737                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3738                         return $wgGrammarForms[$this->getCode()][$case][$word];
3739                 }
3740
3741                 $grammarTransformations = $this->getGrammarTransformations();
3742
3743                 if ( isset( $grammarTransformations[$case] ) ) {
3744                         $forms = $grammarTransformations[$case];
3745
3746                         // Some names of grammar rules are aliases for other rules.
3747                         // In such cases the value is a string rather than object,
3748                         // so load the actual rules.
3749                         if ( is_string( $forms ) ) {
3750                                 $forms = $grammarTransformations[$forms];
3751                         }
3752
3753                         foreach ( array_values( $forms ) as $rule ) {
3754                                 $form = $rule[0];
3755
3756                                 if ( $form === '@metadata' ) {
3757                                         continue;
3758                                 }
3759
3760                                 $replacement = $rule[1];
3761
3762                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3763                                 $patternMatches = preg_match( $regex, $word );
3764
3765                                 if ( $patternMatches === false ) {
3766                                         wfLogWarning(
3767                                                 'An error occurred while processing grammar. ' .
3768                                                 "Word: '$word'. Regex: /$form/."
3769                                         );
3770                                 } elseif ( $patternMatches === 1 ) {
3771                                         $word = preg_replace( $regex, $replacement, $word );
3772
3773                                         break;
3774                                 }
3775                         }
3776                 }
3777
3778                 return $word;
3779         }
3780
3781         /**
3782          * Get the grammar forms for the content language
3783          * @return array Array of grammar forms
3784          * @since 1.20
3785          */
3786         function getGrammarForms() {
3787                 global $wgGrammarForms;
3788                 if ( isset( $wgGrammarForms[$this->getCode()] )
3789                         && is_array( $wgGrammarForms[$this->getCode()] )
3790                 ) {
3791                         return $wgGrammarForms[$this->getCode()];
3792                 }
3793
3794                 return [];
3795         }
3796
3797         /**
3798          * Get the grammar transformations data for the language.
3799          * Used like grammar forms, with {{GRAMMAR}} and cases,
3800          * but uses pairs of regexes and replacements instead of code.
3801          *
3802          * @return array[] Array of grammar transformations.
3803          * @throws MWException
3804          * @since 1.28
3805          */
3806         public function getGrammarTransformations() {
3807                 $languageCode = $this->getCode();
3808
3809                 if ( self::$grammarTransformations === null ) {
3810                         self::$grammarTransformations = new MapCacheLRU( 10 );
3811                 }
3812
3813                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3814                         return self::$grammarTransformations->get( $languageCode );
3815                 }
3816
3817                 $data = [];
3818
3819                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3820                 if ( is_readable( $grammarDataFile ) ) {
3821                         $data = FormatJson::decode(
3822                                 file_get_contents( $grammarDataFile ),
3823                                 true
3824                         );
3825
3826                         if ( $data === null ) {
3827                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3828                         }
3829
3830                         self::$grammarTransformations->set( $languageCode, $data );
3831                 }
3832
3833                 return $data;
3834         }
3835
3836         /**
3837          * Provides an alternative text depending on specified gender.
3838          * Usage {{gender:username|masculine|feminine|unknown}}.
3839          * username is optional, in which case the gender of current user is used,
3840          * but only in (some) interface messages; otherwise default gender is used.
3841          *
3842          * If no forms are given, an empty string is returned. If only one form is
3843          * given, it will be returned unconditionally. These details are implied by
3844          * the caller and cannot be overridden in subclasses.
3845          *
3846          * If three forms are given, the default is to use the third (unknown) form.
3847          * If fewer than three forms are given, the default is to use the first (masculine) form.
3848          * These details can be overridden in subclasses.
3849          *
3850          * @param string $gender
3851          * @param array $forms
3852          *
3853          * @return string
3854          */
3855         function gender( $gender, $forms ) {
3856                 if ( !count( $forms ) ) {
3857                         return '';
3858                 }
3859                 $forms = $this->preConvertPlural( $forms, 2 );
3860                 if ( $gender === 'male' ) {
3861                         return $forms[0];
3862                 }
3863                 if ( $gender === 'female' ) {
3864                         return $forms[1];
3865                 }
3866                 return isset( $forms[2] ) ? $forms[2] : $forms[0];
3867         }
3868
3869         /**
3870          * Plural form transformations, needed for some languages.
3871          * For example, there are 3 form of plural in Russian and Polish,
3872          * depending on "count mod 10". See [[w:Plural]]
3873          * For English it is pretty simple.
3874          *
3875          * Invoked by putting {{plural:count|wordform1|wordform2}}
3876          * or {{plural:count|wordform1|wordform2|wordform3}}
3877          *
3878          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3879          *
3880          * @param int $count Non-localized number
3881          * @param array $forms Different plural forms
3882          * @return string Correct form of plural for $count in this language
3883          */
3884         function convertPlural( $count, $forms ) {
3885                 // Handle explicit n=pluralform cases
3886                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3887                 if ( is_string( $forms ) ) {
3888                         return $forms;
3889                 }
3890                 if ( !count( $forms ) ) {
3891                         return '';
3892                 }
3893
3894                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3895                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3896                 return $forms[$pluralForm];
3897         }
3898
3899         /**
3900          * Handles explicit plural forms for Language::convertPlural()
3901          *
3902          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
3903          * If an explicitly defined plural form matches the $count, then
3904          * string value returned, otherwise array returned for further consideration
3905          * by CLDR rules or overridden convertPlural().
3906          *
3907          * @since 1.23
3908          *
3909          * @param int $count Non-localized number
3910          * @param array $forms Different plural forms
3911          *
3912          * @return array|string
3913          */
3914         protected function handleExplicitPluralForms( $count, array $forms ) {
3915                 foreach ( $forms as $index => $form ) {
3916                         if ( preg_match( '/\d+=/i', $form ) ) {
3917                                 $pos = strpos( $form, '=' );
3918                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
3919                                         return substr( $form, $pos + 1 );
3920                                 }
3921                                 unset( $forms[$index] );
3922                         }
3923                 }
3924                 return array_values( $forms );
3925         }
3926
3927         /**
3928          * Checks that convertPlural was given an array and pads it to requested
3929          * amount of forms by copying the last one.
3930          *
3931          * @param array $forms Array of forms given to convertPlural
3932          * @param int $count How many forms should there be at least
3933          * @return array Padded array of forms or an exception if not an array
3934          */
3935         protected function preConvertPlural( /* Array */ $forms, $count ) {
3936                 while ( count( $forms ) < $count ) {
3937                         $forms[] = $forms[count( $forms ) - 1];
3938                 }
3939                 return $forms;
3940         }
3941
3942         /**
3943          * Wraps argument with unicode control characters for directionality safety
3944          *
3945          * This solves the problem where directionality-neutral characters at the edge of
3946          * the argument string get interpreted with the wrong directionality from the
3947          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
3948          *
3949          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
3950          * directionality of the argument string, using the BIDI algorithm's own "First
3951          * strong directional codepoint" rule. Essentially, this works round the fact that
3952          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
3953          * direction inference). The latter is cleaner but still not widely supported.
3954          *
3955          * @param string $text Text to wrap
3956          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
3957          */
3958         public function embedBidi( $text = '' ) {
3959                 $dir = self::strongDirFromContent( $text );
3960                 if ( $dir === 'ltr' ) {
3961                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
3962                         return self::$lre . $text . self::$pdf;
3963                 }
3964                 if ( $dir === 'rtl' ) {
3965                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
3966                         return self::$rle . $text . self::$pdf;
3967                 }
3968                 // No strong directionality: do not wrap
3969                 return $text;
3970         }
3971
3972         /**
3973          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
3974          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
3975          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
3976          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
3977          * match up with it.
3978          *
3979          * @param string $str The validated block duration in English
3980          * @param User $user User object to use timezone from or null for $wgUser
3981          * @param int $now Current timestamp, for formatting relative block durations
3982          * @return string Somehow translated block duration
3983          * @see LanguageFi.php for example implementation
3984          */
3985         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
3986                 $duration = SpecialBlock::getSuggestedDurations( $this );
3987                 foreach ( $duration as $show => $value ) {
3988                         if ( strcmp( $str, $value ) == 0 ) {
3989                                 return htmlspecialchars( trim( $show ) );
3990                         }
3991                 }
3992
3993                 if ( wfIsInfinity( $str ) ) {
3994                         foreach ( $duration as $show => $value ) {
3995                                 if ( wfIsInfinity( $value ) ) {
3996                                         return htmlspecialchars( trim( $show ) );
3997                                 }
3998                         }
3999                 }
4000
4001                 // If all else fails, return a standard duration or timestamp description.
4002                 $time = strtotime( $str, $now );
4003                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4004                         return $str;
4005                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4006                         // The result differs based on current time, so the difference
4007                         // is a fixed duration length.
4008                         return $this->formatDuration( $time - $now );
4009                 } else { // It's an absolute timestamp.
4010                         if ( $time === 0 ) {
4011                                 // wfTimestamp() handles 0 as current time instead of epoch.
4012                                 $time = '19700101000000';
4013                         }
4014                         if ( $user ) {
4015                                 return $this->userTimeAndDate( $time, $user );
4016                         }
4017                         return $this->timeanddate( $time );
4018                 }
4019         }
4020
4021         /**
4022          * languages like Chinese need to be segmented in order for the diff
4023          * to be of any use
4024          *
4025          * @param string $text
4026          * @return string
4027          */
4028         public function segmentForDiff( $text ) {
4029                 return $text;
4030         }
4031
4032         /**
4033          * and unsegment to show the result
4034          *
4035          * @param string $text
4036          * @return string
4037          */
4038         public function unsegmentForDiff( $text ) {
4039                 return $text;
4040         }
4041
4042         /**
4043          * Return the LanguageConverter used in the Language
4044          *
4045          * @since 1.19
4046          * @return LanguageConverter
4047          */
4048         public function getConverter() {
4049                 return $this->mConverter;
4050         }
4051
4052         /**
4053          * convert text to all supported variants
4054          *
4055          * @param string $text
4056          * @return array
4057          */
4058         public function autoConvertToAllVariants( $text ) {
4059                 return $this->mConverter->autoConvertToAllVariants( $text );
4060         }
4061
4062         /**
4063          * convert text to different variants of a language.
4064          *
4065          * @param string $text
4066          * @return string
4067          */
4068         public function convert( $text ) {
4069                 return $this->mConverter->convert( $text );
4070         }
4071
4072         /**
4073          * Convert a Title object to a string in the preferred variant
4074          *
4075          * @param Title $title
4076          * @return string
4077          */
4078         public function convertTitle( $title ) {
4079                 return $this->mConverter->convertTitle( $title );
4080         }
4081
4082         /**
4083          * Convert a namespace index to a string in the preferred variant
4084          *
4085          * @param int $ns
4086          * @return string
4087          */
4088         public function convertNamespace( $ns ) {
4089                 return $this->mConverter->convertNamespace( $ns );
4090         }
4091
4092         /**
4093          * Check if this is a language with variants
4094          *
4095          * @return bool
4096          */
4097         public function hasVariants() {
4098                 return count( $this->getVariants() ) > 1;
4099         }
4100
4101         /**
4102          * Check if the language has the specific variant
4103          *
4104          * @since 1.19
4105          * @param string $variant
4106          * @return bool
4107          */
4108         public function hasVariant( $variant ) {
4109                 return (bool)$this->mConverter->validateVariant( $variant );
4110         }
4111
4112         /**
4113          * Perform output conversion on a string, and encode for safe HTML output.
4114          * @param string $text Text to be converted
4115          * @param bool $isTitle Whether this conversion is for the article title
4116          * @return string
4117          * @todo this should get integrated somewhere sane
4118          */
4119         public function convertHtml( $text, $isTitle = false ) {
4120                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
4121         }
4122
4123         /**
4124          * @param string $key
4125          * @return string
4126          */
4127         public function convertCategoryKey( $key ) {
4128                 return $this->mConverter->convertCategoryKey( $key );
4129         }
4130
4131         /**
4132          * Get the list of variants supported by this language
4133          * see sample implementation in LanguageZh.php
4134          *
4135          * @return string[] An array of language codes
4136          */
4137         public function getVariants() {
4138                 return $this->mConverter->getVariants();
4139         }
4140
4141         /**
4142          * @return string
4143          */
4144         public function getPreferredVariant() {
4145                 return $this->mConverter->getPreferredVariant();
4146         }
4147
4148         /**
4149          * @return string
4150          */
4151         public function getDefaultVariant() {
4152                 return $this->mConverter->getDefaultVariant();
4153         }
4154
4155         /**
4156          * @return string
4157          */
4158         public function getURLVariant() {
4159                 return $this->mConverter->getURLVariant();
4160         }
4161
4162         /**
4163          * If a language supports multiple variants, it is
4164          * possible that non-existing link in one variant
4165          * actually exists in another variant. this function
4166          * tries to find it. See e.g. LanguageZh.php
4167          * The input parameters may be modified upon return
4168          *
4169          * @param string &$link The name of the link
4170          * @param Title &$nt The title object of the link
4171          * @param bool $ignoreOtherCond To disable other conditions when
4172          *   we need to transclude a template or update a category's link
4173          */
4174         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4175                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4176         }
4177
4178         /**
4179          * returns language specific options used by User::getPageRenderHash()
4180          * for example, the preferred language variant
4181          *
4182          * @return string
4183          */
4184         function getExtraHashOptions() {
4185                 return $this->mConverter->getExtraHashOptions();
4186         }
4187
4188         /**
4189          * For languages that support multiple variants, the title of an
4190          * article may be displayed differently in different variants. this
4191          * function returns the apporiate title defined in the body of the article.
4192          *
4193          * @return string
4194          */
4195         public function getParsedTitle() {
4196                 return $this->mConverter->getParsedTitle();
4197         }
4198
4199         /**
4200          * Refresh the cache of conversion tables when
4201          * MediaWiki:Conversiontable* is updated.
4202          *
4203          * @param Title $title The Title of the page being updated
4204          */
4205         public function updateConversionTable( Title $title ) {
4206                 $this->mConverter->updateConversionTable( $title );
4207         }
4208
4209         /**
4210          * Prepare external link text for conversion. When the text is
4211          * a URL, it shouldn't be converted, and it'll be wrapped in
4212          * the "raw" tag (-{R| }-) to prevent conversion.
4213          *
4214          * This function is called "markNoConversion" for historical
4215          * reasons.
4216          *
4217          * @param string $text Text to be used for external link
4218          * @param bool $noParse Wrap it without confirming it's a real URL first
4219          * @return string The tagged text
4220          */
4221         public function markNoConversion( $text, $noParse = false ) {
4222                 // Excluding protocal-relative URLs may avoid many false positives.
4223                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
4224                         return $this->mConverter->markNoConversion( $text );
4225                 } else {
4226                         return $text;
4227                 }
4228         }
4229
4230         /**
4231          * A regular expression to match legal word-trailing characters
4232          * which should be merged onto a link of the form [[foo]]bar.
4233          *
4234          * @return string
4235          */
4236         public function linkTrail() {
4237                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4238         }
4239
4240         /**
4241          * A regular expression character set to match legal word-prefixing
4242          * characters which should be merged onto a link of the form foo[[bar]].
4243          *
4244          * @return string
4245          */
4246         public function linkPrefixCharset() {
4247                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4248         }
4249
4250         /**
4251          * Get the "parent" language which has a converter to convert a "compatible" language
4252          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4253          *
4254          * @return Language|null
4255          * @since 1.22
4256          */
4257         public function getParentLanguage() {
4258                 if ( $this->mParentLanguage !== false ) {
4259                         return $this->mParentLanguage;
4260                 }
4261
4262                 $code = explode( '-', $this->getCode() )[0];
4263                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4264                         $this->mParentLanguage = null;
4265                         return null;
4266                 }
4267                 $lang = self::factory( $code );
4268                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4269                         $this->mParentLanguage = null;
4270                         return null;
4271                 }
4272
4273                 $this->mParentLanguage = $lang;
4274                 return $lang;
4275         }
4276
4277         /**
4278          * Compare with an other language object
4279          *
4280          * @since 1.28
4281          * @param Language $lang
4282          * @return bool
4283          */
4284         public function equals( Language $lang ) {
4285                 return $lang->getCode() === $this->mCode;
4286         }
4287
4288         /**
4289          * Get the internal language code for this language object
4290          *
4291          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4292          * htmlspecialchars() or similar
4293          *
4294          * @return string
4295          */
4296         public function getCode() {
4297                 return $this->mCode;
4298         }
4299
4300         /**
4301          * Get the code in BCP 47 format which we can use
4302          * inside of html lang="" tags.
4303          *
4304          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4305          * htmlspecialchars() or similar.
4306          *
4307          * @since 1.19
4308          * @return string
4309          */
4310         public function getHtmlCode() {
4311                 if ( is_null( $this->mHtmlCode ) ) {
4312                         $this->mHtmlCode = wfBCP47( $this->getCode() );
4313                 }
4314                 return $this->mHtmlCode;
4315         }
4316
4317         /**
4318          * @param string $code
4319          */
4320         public function setCode( $code ) {
4321                 $this->mCode = $code;
4322                 // Ensure we don't leave incorrect cached data lying around
4323                 $this->mHtmlCode = null;
4324                 $this->mParentLanguage = false;
4325         }
4326
4327         /**
4328          * Get the language code from a file name. Inverse of getFileName()
4329          * @param string $filename $prefix . $languageCode . $suffix
4330          * @param string $prefix Prefix before the language code
4331          * @param string $suffix Suffix after the language code
4332          * @return string Language code, or false if $prefix or $suffix isn't found
4333          */
4334         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4335                 $m = null;
4336                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4337                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4338                 if ( !count( $m ) ) {
4339                         return false;
4340                 }
4341                 return str_replace( '_', '-', strtolower( $m[1] ) );
4342         }
4343
4344         /**
4345          * @param string $code
4346          * @param bool $fallback Whether we're going through language fallback chain
4347          * @return string Name of the language class
4348          */
4349         public static function classFromCode( $code, $fallback = true ) {
4350                 if ( $fallback && $code == 'en' ) {
4351                         return 'Language';
4352                 } else {
4353                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4354                 }
4355         }
4356
4357         /**
4358          * Get the name of a file for a certain language code
4359          * @param string $prefix Prepend this to the filename
4360          * @param string $code Language code
4361          * @param string $suffix Append this to the filename
4362          * @throws MWException
4363          * @return string $prefix . $mangledCode . $suffix
4364          */
4365         public static function getFileName( $prefix = 'Language', $code, $suffix = '.php' ) {
4366                 if ( !self::isValidBuiltInCode( $code ) ) {
4367                         throw new MWException( "Invalid language code \"$code\"" );
4368                 }
4369
4370                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4371         }
4372
4373         /**
4374          * @param string $code
4375          * @return string
4376          */
4377         public static function getMessagesFileName( $code ) {
4378                 global $IP;
4379                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4380                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4381                 return $file;
4382         }
4383
4384         /**
4385          * @param string $code
4386          * @return string
4387          * @throws MWException
4388          * @since 1.23
4389          */
4390         public static function getJsonMessagesFileName( $code ) {
4391                 global $IP;
4392
4393                 if ( !self::isValidBuiltInCode( $code ) ) {
4394                         throw new MWException( "Invalid language code \"$code\"" );
4395                 }
4396
4397                 return "$IP/languages/i18n/$code.json";
4398         }
4399
4400         /**
4401          * Get the first fallback for a given language.
4402          *
4403          * @param string $code
4404          *
4405          * @return bool|string
4406          */
4407         public static function getFallbackFor( $code ) {
4408                 $fallbacks = self::getFallbacksFor( $code );
4409                 if ( $fallbacks ) {
4410                         return $fallbacks[0];
4411                 }
4412                 return false;
4413         }
4414
4415         /**
4416          * Get the ordered list of fallback languages.
4417          *
4418          * @since 1.19
4419          * @param string $code Language code
4420          * @return array Non-empty array, ending in "en"
4421          */
4422         public static function getFallbacksFor( $code ) {
4423                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4424                         return [];
4425                 }
4426                 // For unknown languages, fallbackSequence returns an empty array,
4427                 // hardcode fallback to 'en' in that case.
4428                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4429         }
4430
4431         /**
4432          * Get the ordered list of fallback languages, ending with the fallback
4433          * language chain for the site language.
4434          *
4435          * @since 1.22
4436          * @param string $code Language code
4437          * @return array Array( fallbacks, site fallbacks )
4438          */
4439         public static function getFallbacksIncludingSiteLanguage( $code ) {
4440                 global $wgLanguageCode;
4441
4442                 // Usually, we will only store a tiny number of fallback chains, so we
4443                 // keep them in static memory.
4444                 $cacheKey = "{$code}-{$wgLanguageCode}";
4445
4446                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4447                         $fallbacks = self::getFallbacksFor( $code );
4448
4449                         // Append the site's fallback chain, including the site language itself
4450                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4451                         array_unshift( $siteFallbacks, $wgLanguageCode );
4452
4453                         // Eliminate any languages already included in the chain
4454                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4455
4456                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4457                 }
4458                 return self::$fallbackLanguageCache[$cacheKey];
4459         }
4460
4461         /**
4462          * Get all messages for a given language
4463          * WARNING: this may take a long time. If you just need all message *keys*
4464          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4465          *
4466          * @param string $code
4467          *
4468          * @return array
4469          */
4470         public static function getMessagesFor( $code ) {
4471                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4472         }
4473
4474         /**
4475          * Get a message for a given language
4476          *
4477          * @param string $key
4478          * @param string $code
4479          *
4480          * @return string
4481          */
4482         public static function getMessageFor( $key, $code ) {
4483                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4484         }
4485
4486         /**
4487          * Get all message keys for a given language. This is a faster alternative to
4488          * array_keys( Language::getMessagesFor( $code ) )
4489          *
4490          * @since 1.19
4491          * @param string $code Language code
4492          * @return array Array of message keys (strings)
4493          */
4494         public static function getMessageKeysFor( $code ) {
4495                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4496         }
4497
4498         /**
4499          * @param string $talk
4500          * @return mixed
4501          */
4502         function fixVariableInNamespace( $talk ) {
4503                 if ( strpos( $talk, '$1' ) === false ) {
4504                         return $talk;
4505                 }
4506
4507                 global $wgMetaNamespace;
4508                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4509
4510                 # Allow grammar transformations
4511                 # Allowing full message-style parsing would make simple requests
4512                 # such as action=raw much more expensive than they need to be.
4513                 # This will hopefully cover most cases.
4514                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4515                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4516                 return str_replace( ' ', '_', $talk );
4517         }
4518
4519         /**
4520          * @param string $m
4521          * @return string
4522          */
4523         function replaceGrammarInNamespace( $m ) {
4524                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4525         }
4526
4527         /**
4528          * Decode an expiry (block, protection, etc) which has come from the DB
4529          *
4530          * @param string $expiry Database expiry String
4531          * @param bool|int $format True to process using language functions, or TS_ constant
4532          *     to return the expiry in a given timestamp
4533          * @param string $infinity If $format is not true, use this string for infinite expiry
4534          * @return string
4535          * @since 1.18
4536          */
4537         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4538                 static $dbInfinity;
4539                 if ( $dbInfinity === null ) {
4540                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4541                 }
4542
4543                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4544                         return $format === true
4545                                 ? $this->getMessageFromDB( 'infiniteblock' )
4546                                 : $infinity;
4547                 } else {
4548                         return $format === true
4549                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4550                                 : wfTimestamp( $format, $expiry );
4551                 }
4552         }
4553
4554         /**
4555          * Formats a time given in seconds into a string representation of that time.
4556          *
4557          * @param int|float $seconds
4558          * @param array $format An optional argument that formats the returned string in different ways:
4559          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4560          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4561          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4562          *     and friends.
4563          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4564          *     or 'avoidminutes'.
4565          * @return string
4566          */
4567         function formatTimePeriod( $seconds, $format = [] ) {
4568                 if ( !is_array( $format ) ) {
4569                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4570                 }
4571                 if ( !isset( $format['avoid'] ) ) {
4572                         $format['avoid'] = false;
4573                 }
4574                 if ( !isset( $format['noabbrevs'] ) ) {
4575                         $format['noabbrevs'] = false;
4576                 }
4577                 $secondsMsg = wfMessage(
4578                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4579                 $minutesMsg = wfMessage(
4580                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4581                 $hoursMsg = wfMessage(
4582                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4583                 $daysMsg = wfMessage(
4584                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4585
4586                 if ( round( $seconds * 10 ) < 100 ) {
4587                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4588                         $s = $secondsMsg->params( $s )->text();
4589                 } elseif ( round( $seconds ) < 60 ) {
4590                         $s = $this->formatNum( round( $seconds ) );
4591                         $s = $secondsMsg->params( $s )->text();
4592                 } elseif ( round( $seconds ) < 3600 ) {
4593                         $minutes = floor( $seconds / 60 );
4594                         $secondsPart = round( fmod( $seconds, 60 ) );
4595                         if ( $secondsPart == 60 ) {
4596                                 $secondsPart = 0;
4597                                 $minutes++;
4598                         }
4599                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4600                         $s .= ' ';
4601                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4602                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4603                         $hours = floor( $seconds / 3600 );
4604                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4605                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4606                         if ( $secondsPart == 60 ) {
4607                                 $secondsPart = 0;
4608                                 $minutes++;
4609                         }
4610                         if ( $minutes == 60 ) {
4611                                 $minutes = 0;
4612                                 $hours++;
4613                         }
4614                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4615                         $s .= ' ';
4616                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4617                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4618                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4619                         }
4620                 } else {
4621                         $days = floor( $seconds / 86400 );
4622                         if ( $format['avoid'] === 'avoidminutes' ) {
4623                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4624                                 if ( $hours == 24 ) {
4625                                         $hours = 0;
4626                                         $days++;
4627                                 }
4628                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4629                                 $s .= ' ';
4630                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4631                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4632                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4633                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4634                                 if ( $minutes == 60 ) {
4635                                         $minutes = 0;
4636                                         $hours++;
4637                                 }
4638                                 if ( $hours == 24 ) {
4639                                         $hours = 0;
4640                                         $days++;
4641                                 }
4642                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4643                                 $s .= ' ';
4644                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4645                                 $s .= ' ';
4646                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4647                         } else {
4648                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4649                                 $s .= ' ';
4650                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4651                         }
4652                 }
4653                 return $s;
4654         }
4655
4656         /**
4657          * Format a bitrate for output, using an appropriate
4658          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4659          *   the magnitude in question.
4660          *
4661          * This use base 1000. For base 1024 use formatSize(), for another base
4662          * see formatComputingNumbers().
4663          *
4664          * @param int $bps
4665          * @return string
4666          */
4667         function formatBitrate( $bps ) {
4668                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4669         }
4670
4671         /**
4672          * @param int $size Size of the unit
4673          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4674          * @param string $messageKey Message key to be uesd
4675          * @return string
4676          */
4677         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4678                 if ( $size <= 0 ) {
4679                         return str_replace( '$1', $this->formatNum( $size ),
4680                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4681                         );
4682                 }
4683                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4684                 $index = 0;
4685
4686                 $maxIndex = count( $sizes ) - 1;
4687                 while ( $size >= $boundary && $index < $maxIndex ) {
4688                         $index++;
4689                         $size /= $boundary;
4690                 }
4691
4692                 // For small sizes no decimal places necessary
4693                 $round = 0;
4694                 if ( $index > 1 ) {
4695                         // For MB and bigger two decimal places are smarter
4696                         $round = 2;
4697                 }
4698                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4699
4700                 $size = round( $size, $round );
4701                 $text = $this->getMessageFromDB( $msg );
4702                 return str_replace( '$1', $this->formatNum( $size ), $text );
4703         }
4704
4705         /**
4706          * Format a size in bytes for output, using an appropriate
4707          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4708          *
4709          * This method use base 1024. For base 1000 use formatBitrate(), for
4710          * another base see formatComputingNumbers()
4711          *
4712          * @param int $size Size to format
4713          * @return string Plain text (not HTML)
4714          */
4715         function formatSize( $size ) {
4716                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4717         }
4718
4719         /**
4720          * Make a list item, used by various special pages
4721          *
4722          * @param string $page Page link
4723          * @param string $details HTML safe text between brackets
4724          * @param bool $oppositedm Add the direction mark opposite to your
4725          *   language, to display text properly
4726          * @return HTML escaped string
4727          */
4728         function specialList( $page, $details, $oppositedm = true ) {
4729                 if ( !$details ) {
4730                         return $page;
4731                 }
4732
4733                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4734                 return
4735                         $page .
4736                         $dirmark .
4737                         $this->msg( 'word-separator' )->escaped() .
4738                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4739         }
4740
4741         /**
4742          * Generate (prev x| next x) (20|50|100...) type links for paging
4743          *
4744          * @param Title $title Title object to link
4745          * @param int $offset
4746          * @param int $limit
4747          * @param array $query Optional URL query parameter string
4748          * @param bool $atend Optional param for specified if this is the last page
4749          * @return string
4750          */
4751         public function viewPrevNext( Title $title, $offset, $limit,
4752                 array $query = [], $atend = false
4753         ) {
4754                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4755
4756                 # Make 'previous' link
4757                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4758                 if ( $offset > 0 ) {
4759                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4760                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4761                 } else {
4762                         $plink = htmlspecialchars( $prev );
4763                 }
4764
4765                 # Make 'next' link
4766                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4767                 if ( $atend ) {
4768                         $nlink = htmlspecialchars( $next );
4769                 } else {
4770                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4771                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4772                 }
4773
4774                 # Make links to set number of items per page
4775                 $numLinks = [];
4776                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4777                         $numLinks[] = $this->numLink( $title, $offset, $num,
4778                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4779                 }
4780
4781                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4782                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4783         }
4784
4785         /**
4786          * Helper function for viewPrevNext() that generates links
4787          *
4788          * @param Title $title Title object to link
4789          * @param int $offset
4790          * @param int $limit
4791          * @param array $query Extra query parameters
4792          * @param string $link Text to use for the link; will be escaped
4793          * @param string $tooltipMsg Name of the message to use as tooltip
4794          * @param string $class Value of the "class" attribute of the link
4795          * @return string HTML fragment
4796          */
4797         private function numLink( Title $title, $offset, $limit, array $query, $link,
4798                 $tooltipMsg, $class
4799         ) {
4800                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4801                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4802                         ->numParams( $limit )->text();
4803
4804                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4805                         'title' => $tooltip, 'class' => $class ], $link );
4806         }
4807
4808         /**
4809          * Get the conversion rule title, if any.
4810          *
4811          * @return string
4812          */
4813         public function getConvRuleTitle() {
4814                 return $this->mConverter->getConvRuleTitle();
4815         }
4816
4817         /**
4818          * Get the compiled plural rules for the language
4819          * @since 1.20
4820          * @return array Associative array with plural form, and plural rule as key-value pairs
4821          */
4822         public function getCompiledPluralRules() {
4823                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4824                 $fallbacks = self::getFallbacksFor( $this->mCode );
4825                 if ( !$pluralRules ) {
4826                         foreach ( $fallbacks as $fallbackCode ) {
4827                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4828                                 if ( $pluralRules ) {
4829                                         break;
4830                                 }
4831                         }
4832                 }
4833                 return $pluralRules;
4834         }
4835
4836         /**
4837          * Get the plural rules for the language
4838          * @since 1.20
4839          * @return array Associative array with plural form number and plural rule as key-value pairs
4840          */
4841         public function getPluralRules() {
4842                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4843                 $fallbacks = self::getFallbacksFor( $this->mCode );
4844                 if ( !$pluralRules ) {
4845                         foreach ( $fallbacks as $fallbackCode ) {
4846                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4847                                 if ( $pluralRules ) {
4848                                         break;
4849                                 }
4850                         }
4851                 }
4852                 return $pluralRules;
4853         }
4854
4855         /**
4856          * Get the plural rule types for the language
4857          * @since 1.22
4858          * @return array Associative array with plural form number and plural rule type as key-value pairs
4859          */
4860         public function getPluralRuleTypes() {
4861                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4862                 $fallbacks = self::getFallbacksFor( $this->mCode );
4863                 if ( !$pluralRuleTypes ) {
4864                         foreach ( $fallbacks as $fallbackCode ) {
4865                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4866                                 if ( $pluralRuleTypes ) {
4867                                         break;
4868                                 }
4869                         }
4870                 }
4871                 return $pluralRuleTypes;
4872         }
4873
4874         /**
4875          * Find the index number of the plural rule appropriate for the given number
4876          * @param int $number
4877          * @return int The index number of the plural rule
4878          */
4879         public function getPluralRuleIndexNumber( $number ) {
4880                 $pluralRules = $this->getCompiledPluralRules();
4881                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4882                 return $form;
4883         }
4884
4885         /**
4886          * Find the plural rule type appropriate for the given number
4887          * For example, if the language is set to Arabic, getPluralType(5) should
4888          * return 'few'.
4889          * @since 1.22
4890          * @param int $number
4891          * @return string The name of the plural rule type, e.g. one, two, few, many
4892          */
4893         public function getPluralRuleType( $number ) {
4894                 $index = $this->getPluralRuleIndexNumber( $number );
4895                 $pluralRuleTypes = $this->getPluralRuleTypes();
4896                 if ( isset( $pluralRuleTypes[$index] ) ) {
4897                         return $pluralRuleTypes[$index];
4898                 } else {
4899                         return 'other';
4900                 }
4901         }
4902 }